diff --git a/analysis/current/analysis-report-hybrid.json b/analysis/current/analysis-report-hybrid.json index 8e247fe394d9df4eefa2c6509f663af3cf9c1964..cf5e948f363c7c813c78a701d757cf569d48e6f5 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": "20260421T060039Z", - "generated_at": "2026-04-21T06:06:35Z", + "snapshot_id": "20260421T120044Z", + "generated_at": "2026-04-21T12:17:08Z", "evidence_quality": "full", "llm_enrichment": true, "meta_bugs": [ diff --git a/analysis/current/analysis-report-hybrid.llm-reviews.json b/analysis/current/analysis-report-hybrid.llm-reviews.json index 77cb05758f7a406dbe0d9db7a56d9cbf7620ce3e..9f291c056a5e7cdb281d146b74e95c81d77bf7be 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": "20260421T060039Z", - "generated_at": "2026-04-21T06:06:35Z", + "snapshot_id": "20260421T120044Z", + "generated_at": "2026-04-21T12:17:08Z", "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": "31dd82c6fb7a9693f3ae7988bd88e9fed476e751189d484c1bbf29009a125ea4", + "prepared_review_unit_hash": "68c474652dbba90cfb989b3893065d0919301d3d16e3f077997780fd7862a7f5", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 2479, + "serialized_chars": 2478, "estimated_input_tokens": 620, "estimated_eval_tokens": 1496 }, @@ -39,7 +39,7 @@ "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 2479, + "serialized_chars": 2478, "estimated_input_tokens": 620, "estimated_eval_tokens": 1496 }, @@ -48,24 +48,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These do not look like duplicates: one is about Tapas tokenization differing from TensorFlow, one is about a bad login/auth error when downloading a private model, and one is a general failure to load bert-base-uncased. They touch different code paths and failure modes.", - "confidence": 0.96, - "canonical_issue_reason": "Issue #16618 is the broadest, most general model-loading report, so it is the best single issue to represent the cluster if one must be chosen.", + "summary": "The three issues describe different problems: Tapas tokenization mismatch, an auth/error-message problem for private models, and a generic failure to load bert-base-uncased. They do not look like duplicates or the same underlying bug.", + "confidence": 0.97, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue #16618 is the most general and user-facing of the three; the other two are narrower and clearly about different problems.", + "best_issue_reason": "No issue is a good canonical anchor for this cluster because the topics are unrelated; if one must be chosen, #16618 is the broadest user-facing failure, but it does not represent the others.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:13244", "right": "issue:16618", "accept": false, - "reason": "Tokenization mismatch for Tapas is a different problem from failing to load bert-base-uncased; they involve different subsystems and symptoms." + "reason": "Different symptoms and likely different code paths: Tapas tokenization parity vs a model-loading failure for bert-base-uncased." }, { "left": "issue:15980", "right": "issue:16618", "accept": false, - "reason": "Private-model download/auth error is a distinct login/authorization issue, not the same as a generic model-loading failure." + "reason": "One is about the error message/authorization path for private models, the other is a failure loading a public model; not the same bug." } ] }, @@ -154,8 +154,8 @@ "issue:44805|issue:45161", "issue:44805|issue:44918", "issue:44661|issue:44805", - "issue:30277|issue:43638", "issue:34634|issue:35545", + "issue:30277|issue:43638", "issue:43596|issue:43638", "issue:43975|issue:44457", "issue:43039|issue:45161", @@ -170,8 +170,8 @@ "issue:39290|issue:45081", "issue:43828|issue:45161", "issue:43425|issue:44292", - "issue:34928|issue:41720", "issue:44918|issue:45161", + "issue:34928|issue:41720", "issue:43541|issue:43716", "issue:43828|issue:44292", "issue:43716|issue:44292", @@ -188,8 +188,8 @@ "issue:41762|issue:43596", "issue:43901|issue:44512", "issue:39290|issue:41720", - "issue:43039|issue:44560", "issue:43866|issue:44863", + "issue:43039|issue:44560", "issue:43854|issue:43866", "issue:43425|issue:45070", "issue:39290|issue:43531", @@ -203,388 +203,352 @@ "issue:41093|issue:41762", "issue:43296|issue:43366" ], - "prepared_review_unit_hash": "538109df4c9a9e0dd7be883c0b25881fc90bb4a5eb48c1b663147c582af79d04", + "prepared_review_unit_hash": "15435980cf472e38716ef9ec097d81b497327fbc3b7a3170b1d298b85e91960a", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 63, - "serialized_chars": 46478, - "estimated_input_tokens": 11620, - "estimated_eval_tokens": 23496 + "serialized_chars": 46468, + "estimated_input_tokens": 11617, + "estimated_eval_tokens": 23490 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous: many issues share broad themes like model loading, DeepSpeed/ZeRO, or docs cleanup, but they point to different concrete failures. Only one soft pair looks like a true duplicate: the two ZeRO-3/Bert index-out-of-bounds reports.", - "confidence": 0.89, - "canonical_issue_reason": "issue:43596 is the cleanest representative of the ZeRO-3/Bert \"index 0 is out of bounds for dimension 0 with size 0\" bug; issue:43638 appears to be the same failure with a slightly different wording.", + "summary": "Mostly a heterogeneous cluster of unrelated model/runtime issues. The only clear duplicate families are the DeepSpeed ZeRO-3 index-out-of-bounds load failures and the near-identical mask-shape IndexError reports; the docs issues and the rest look like separate bugs.", + "confidence": 0.71, + "canonical_issue_reason": "Issue 43596 is the cleanest representative of the ZeRO-3 index-0-out-of-bounds load failure family: it describes the generic error pattern and matches the same root symptom seen in the close variants.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43596 is the strongest canonical issue for the only clear duplicate subcluster in this set; it is specific, reproducible, and has a near-identical twin in issue:43638.", + "best_issue_reason": "Issue 43596 is the strongest issue candidate because it most directly captures the shared ZeRO-3 load-time failure pattern and can subsume the closely related Bert/Gemma3 variants.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44322", "right": "issue:45464", "accept": false, - "reason": "Same Qwen3.5 area, but different failure modes: config attribute error vs streaming chat/completions failure." + "reason": "Different Qwen3.5 failures: one is a chat crash, the other is a streaming inference API failure." }, { "left": "issue:43716", "right": "issue:45237", "accept": false, - "reason": "Different models and different bugs; one is a dtype mismatch in Mistral-3 preprocessing, the other is an AMD GPU runtime issue for GPT-OSS." + "reason": "Both involve model/runtime problems, but one is a dtype mismatch and the other is an AMD GPU execution failure for GPT-OSS-20B; not the same bug." }, { "left": "issue:43828", "right": "issue:45237", "accept": false, - "reason": "Both are runtime failures, but the concrete code paths and models differ: autocast dtype mismatch vs AMD GPU incompatibility." + "reason": "Autocast dtype mismatch on Phi-tiny-MoE is unrelated to GPT-OSS-20B failing on AMD GPUs." }, { "left": "issue:45237", "right": "issue:45507", "accept": false, - "reason": "Different model families and different failure types; not the same underlying change or bug." + "reason": "Different models and failure modes: AMD GPU support issue vs invalid method call in GraniteMoEHybrid." }, { "left": "issue:44292", "right": "issue:45237", "accept": false, - "reason": "Different model-specific failures; NVFP4 runtime error is unrelated to AMD GPU support." + "reason": "Both are model runtime errors, but Qwen-3-8B-NVFP4 is a different code path from GPT-OSS-20B on AMD GPUs." }, { "left": "issue:30277", "right": "issue:35545", "accept": false, - "reason": "DeepSpeed collective mismatch and ONNX export are unrelated code paths." + "reason": "DeepSpeed collective mismatch for Jamba-v01 is unrelated to ModernBERT ONNX export failure." }, { "left": "issue:35545", "right": "issue:42915", "accept": false, - "reason": "Both mention model export/loading issues, but the concrete problems are unrelated." + "reason": "ModernBERT ONNX export and Qwen3Moe FineGrainedFP8 failures are different model/export paths." }, { "left": "issue:35545", "right": "issue:41720", "accept": false, - "reason": "ONNX export error versus auto device mapping cuda assert; different bugs." + "reason": "ONNX export error and Qwen3 auto device mapping cuda assert are distinct problems." }, { "left": "issue:44509", "right": "issue:44512", "accept": false, - "reason": "Both are v5 docs cleanups, but they target different removed features and would not merge into one fix." + "reason": "Both are stale v5 docs issues, but they concern different removed features (pipeline tasks vs run command)." }, { "left": "issue:35545", "right": "issue:36010", "accept": false, - "reason": "Model export error and GenerationMixin import failure are unrelated." + "reason": "ONNX export failure is unrelated to the GenerationMixin import error." }, { "left": "issue:43638", "right": "issue:44805", "accept": false, - "reason": "Both are shape/index errors, but the models and failure contexts differ." + "reason": "Both are IndexErrors, but one is DeepSpeed ZeRO-3/BERT load-time and the other is a mask-shape mismatch; different bugs." }, { "left": "issue:43596", "right": "issue:44805", "accept": false, - "reason": "Different symptom details and different model/setup; not the same concrete bug." + "reason": "Different failure signatures and code paths: ZeRO-3 index-out-of-bounds vs mask-shape mismatch." }, { "left": "issue:44805", "right": "issue:45161", "accept": false, - "reason": "Mask-shape IndexError is unrelated to GPT-OSS tensor-parallel behavior." + "reason": "Mask-shape IndexError and GPT-OSS MoE tensor-parallel issue are unrelated." }, { "left": "issue:44805", "right": "issue:44918", "accept": false, - "reason": "Different model families and failure points; not a duplicate." + "reason": "Mask/tensor shape mismatch is not the same as Qwen3.5 embedding unpacking under TRL SFT." }, { "left": "issue:44661", "right": "issue:44805", "accept": false, - "reason": "Tokenizer-mapping guardrail bug is unrelated to the mask/tensor shape error." + "reason": "Tokenizer mapping/add-new-model-like issue is unrelated to a mask-shape IndexError." }, { - "left": "issue:30277", - "right": "issue:43638", + "left": "issue:34634", + "right": "issue:35545", "accept": false, - "reason": "DeepSpeed collective mismatch and Bert zero-3 index error are different bugs." + "reason": "BarkProcessor voice_preset issue is unrelated to ModernBERT ONNX export." }, { - "left": "issue:34634", - "right": "issue:35545", + "left": "issue:30277", + "right": "issue:43638", "accept": false, - "reason": "Processor preset handling and ONNX export are unrelated." + "reason": "DeepSpeed collective mismatch for Jamba-v01 is unrelated to ZeRO-3 index-out-of-bounds in BertModel." }, { "left": "issue:43596", "right": "issue:43638", "accept": true, - "reason": "Same DeepSpeed ZeRO-3/Bert index-0-out-of-bounds failure; these look like the same underlying bug." + "reason": "Same DeepSpeed ZeRO-3 index-0-out-of-bounds load-time failure, just reported against different Bert model variants." }, { "left": "issue:43975", "right": "issue:44457", "accept": false, - "reason": "Different problems: detokenization regression versus LoRA merge/save mismatch." + "reason": "Detokenization bug and LoRA merge/save mismatch are different problems." }, { "left": "issue:43039", "right": "issue:45161", "accept": false, - "reason": "Liger Kernel cross-entropy dispatch and GPT-OSS tensor parallel failure are unrelated." + "reason": "Liger Kernel calling cross_entropy and GPT-OSS MoE tensor-parallel support are unrelated." }, { "left": "issue:44292", "right": "issue:44457", "accept": false, - "reason": "Different models and different failure modes." + "reason": "NVFP4 model runtime error and LoRA merge/save inconsistency are different issues." }, { "left": "issue:34928", "right": "issue:44805", "accept": false, - "reason": "Activation-checkpointing/FSDP tensor-size recomputation is unrelated to the mask shape IndexError." + "reason": "Activation-checkpointing/FSDP tensor-size mismatch is not the same as the mask-shape IndexError." }, { "left": "issue:24643", "right": "issue:30277", "accept": false, - "reason": "Both involve DeepSpeed, but the concrete failures differ completely." + "reason": "Training with DeepSpeed weight dimensionality error and Jamba collectives mismatch are different DeepSpeed failures." }, { "left": "issue:43975", "right": "issue:45237", "accept": false, - "reason": "Unrelated model/runtime issues." + "reason": "Detokenization bug and AMD GPU model loading failure are unrelated." }, { "left": "issue:43638", "right": "issue:45161", "accept": false, - "reason": "Bert ZeRO-3 index error and GPT-OSS TP failure are different bugs." + "reason": "ZeRO-3 BERT load failure and GPT-OSS MoE tensor-parallel issue are different bugs." }, { "left": "issue:34928", "right": "issue:36331", "accept": false, - "reason": "Training/checkpointing tensor mismatch and custom trainer signature mismatch are unrelated." + "reason": "FSDP activation-checkpointing tensor mismatch is unrelated to a CustomTrainer method signature error." }, { "left": "issue:43596", "right": "issue:45161", "accept": false, - "reason": "Different models and different failure classes." + "reason": "DeepSpeed ZeRO-3 load failure is unrelated to GPT-OSS MoE tensor-parallel behavior." }, { "left": "issue:43716", "right": "issue:45507", "accept": false, - "reason": "Image-preprocessor dtype mismatch and invalid method call in GraniteMoEHybrid are unrelated." + "reason": "Mistral dtype mismatch and GraniteMoEHybrid invalid method call do not share the same bug." }, { "left": "issue:39290", "right": "issue:45081", "accept": false, - "reason": "Gemma3 sliding_window_pattern missing and Mistral tokenizer regex patch crash are different model bugs." + "reason": "Gemma3/vLLM sliding-window attribute error is unrelated to the tokenizer regex patch crash." }, { "left": "issue:43828", "right": "issue:45161", "accept": false, - "reason": "Autocast dtype mismatch is unrelated to GPT-OSS MoE tensor parallel behavior." - }, - { - "left": "issue:43425", - "right": "issue:44292", - "accept": false, - "reason": "Torch version compatibility and Qwen NVFP4 runtime failure are unrelated." - }, - { - "left": "issue:34928", - "right": "issue:41720", - "accept": false, - "reason": "FSDP/activation-checkpointing recomputation issue is unrelated to Qwen auto device mapping." + "reason": "Autocast dtype mismatch in Phi-tiny-MoE is different from GPT-OSS MoE TP support." }, { "left": "issue:44918", "right": "issue:45161", "accept": false, - "reason": "TRL SFT embedding unpacking failure and GPT-OSS TP failure are different problems." - }, - { - "left": "issue:41762", - "right": "issue:43638", - "accept": false, - "reason": "Both are ZeRO-3 loading errors, but different models and not clearly the same bug." - }, - { - "left": "issue:43827", - "right": "issue:44512", - "accept": false, - "reason": "Both are docs issues, but they refer to different removed commands/features." - }, - { - "left": "issue:43716", - "right": "issue:44560", - "accept": false, - "reason": "Different models and different error types." - }, - { - "left": "issue:43716", - "right": "issue:44918", - "accept": false, - "reason": "Different model families and unrelated failure paths." + "reason": "Qwen3.5 embedding unpacking under TRL and GPT-OSS MoE tensor-parallel are different issues." }, { "left": "issue:43039", "right": "issue:43828", "accept": false, - "reason": "Different runtime problems; no shared concrete bug." + "reason": "Cross-entropy being called under Liger Kernel is unrelated to Phi-tiny-MoE autocast dtype mismatch." }, { "left": "issue:30277", "right": "issue:36331", "accept": false, - "reason": "DeepSpeed collective mismatch and trainer API signature mismatch are unrelated." + "reason": "Collective mismatch in DeepSpeed and CustomTrainer.compute_loss signature change are unrelated." }, { "left": "issue:43039", "right": "issue:44805", "accept": false, - "reason": "Liger Kernel cross-entropy call and mask-shape IndexError are unrelated." + "reason": "Liger Kernel cross-entropy call and mask-shape IndexError are different bugs." }, { "left": "issue:34928", "right": "issue:36010", "accept": false, - "reason": "Training shape-recompute bug and GenerationMixin import error are unrelated." + "reason": "FSDP/activation-checkpointing mismatch is unrelated to the GenerationMixin import error." }, { "left": "issue:41762", "right": "issue:43596", - "accept": false, - "reason": "Similar ZeRO-3 symptoms, but different models and insufficient evidence of the same underlying bug." + "accept": true, + "reason": "Same underlying DeepSpeed ZeRO-3 index-out-of-bounds load failure; only the affected model differs." }, { "left": "issue:43901", "right": "issue:44512", "accept": false, - "reason": "Different docs updates: text classification docs versus run command removal." + "reason": "TextClassificationPipeline docs mention and transformers run-command docs mention are separate stale-doc issues." }, { "left": "issue:39290", "right": "issue:41720", "accept": false, - "reason": "Gemma3/vLLM config attribute error and Qwen device-mapping cuda assert are different issues." + "reason": "Gemma3/vLLM sliding-window error is unrelated to Qwen3 auto device mapping cuda assert." }, { - "left": "issue:43039", - "right": "issue:44560", + "left": "issue:43866", + "right": "issue:44863", "accept": false, - "reason": "Liger Kernel cross-entropy behavior and Qwen3-vl video StopIteration are unrelated." + "reason": "Corrupted Ovis2 checkpoint and NemotronH load failure are different problems." }, { - "left": "issue:43866", - "right": "issue:44863", + "left": "issue:43039", + "right": "issue:44560", "accept": false, - "reason": "Corrupted checkpoint and NemotronH loading implementation bug are not the same failure." + "reason": "Liger Kernel cross-entropy issue and Qwen3-vl StopIteration video error are unrelated." }, { "left": "issue:43854", "right": "issue:43866", "accept": false, - "reason": "Unit-test model-loading failure and checkpoint corruption are different problems." + "reason": "GLM-4.7-Flash unit-test load failure and Ovis2 checkpoint corruption are different issues." }, { "left": "issue:43425", "right": "issue:45070", "accept": false, - "reason": "Torch version compatibility and pydantic PretrainedConfig field regression are unrelated." + "reason": "Torch compatibility and PretrainedConfig pydantic-field breakage are unrelated." }, { "left": "issue:39290", "right": "issue:43531", "accept": false, - "reason": "Both mention sliding_window-like config issues, but they affect different model families and code paths." + "reason": "Gemma3 sliding-window attribute error and Qwen3-MoE sliding_window issue are different model families and likely different fixes." }, { "left": "issue:41093", "right": "issue:43638", "accept": false, - "reason": "Mask/tensor shape mismatch and Bert ZeRO-3 index error are different failures." + "reason": "Although both are IndexErrors, the mask/tensor shape mismatch is a different failure mode from the ZeRO-3 BERT index-out-of-bounds bug." }, { "left": "issue:43404", "right": "issue:43866", "accept": false, - "reason": "Tied-weight bug in Mistral3 and corrupted Ovis2 checkpoint are unrelated." + "reason": "Mistral3 tied-weight bug and Ovis2 checkpoint corruption are not the same issue." }, { "left": "issue:45362", "right": "issue:45464", "accept": false, - "reason": "Both are Qwen3.5 chat-related, but one is a chat-template crash and the other is a streaming API failure." + "reason": "Qwen3.5 chat crash and Qwen3.5 streaming inference failure are separate symptoms and code paths." }, { "left": "issue:43792", "right": "issue:45070", "accept": false, - "reason": "Whisper runtime failure and pydantic model regression are unrelated." + "reason": "Whisper loading failure is unrelated to the pydantic PretrainedConfig regression." }, { "left": "issue:43541", "right": "issue:43828", "accept": false, - "reason": "Different MoE dtype/tracing failures with different models and code paths." + "reason": "Mixtral grouped_mm tracing error and Phi-tiny-MoE autocast dtype mismatch are distinct runtime issues." }, { "left": "issue:36010", "right": "issue:39290", "accept": false, - "reason": "GenerationMixin import failure and Gemma3 sliding_window config regression are unrelated." + "reason": "GenerationMixin import failure is unrelated to Gemma3/vLLM sliding-window attribute error." }, { "left": "issue:44841", "right": "issue:45084", "accept": false, - "reason": "Processor failure for Voxtral and template compilation error are unrelated." + "reason": "Voxtral processor failure and template-compilation TypeError are different bugs." }, { "left": "issue:41093", "right": "issue:41762", "accept": false, - "reason": "Both are IndexErrors, but the model/setup and actual failure context differ." + "reason": "The mask/tensor shape mismatch and the ZeRO-3 index-out-of-bounds load failure are different error classes." }, { "left": "issue:43296", "right": "issue:43366", "accept": false, - "reason": "PaddleOCR-VL vLLM loading failure and GGUF gpt-oss support request are different issues." + "reason": "PaddleOCR-VL load failure and GGUF gpt-oss support request are not the same issue." } ] }, "evaluator_result": { "accept": true, - "feedback": "Grounded and conservative overall. The duplicate call on issue:43596 / issue:43638 is well supported by the very similar ZeRO-3/Bert index-out-of-bounds reports, and the remaining soft-edge rejections stay cautious without overstating equivalence." + "feedback": "The cluster summary is grounded in the packet, and the soft-edge calls are conservative overall. The two accepted ZeRO-3/BERT-family pairs are well-supported by the titles, while the rest of the rejects stay appropriately cautious about model/failure-mode differences." }, "evaluator_used": true, "retried": false, @@ -655,18 +619,18 @@ "issue:45507" ], "soft_pairs": [ - "issue:43572|issue:45070", "issue:44387|issue:45005", + "issue:43572|issue:45070", "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:39290|issue:43742", "issue:43054|issue:43994", - "issue:43541|issue:44589", + "issue:39290|issue:43742", "issue:41093|issue:43596", + "issue:43541|issue:44589", "issue:42915|issue:43541", "issue:40990|issue:44960", "issue:39186|issue:44918", @@ -690,244 +654,244 @@ "issue:36331|issue:41762", "issue:44560|issue:45507" ], - "prepared_review_unit_hash": "48e9e32137015ae86a38cf6980c8aead85328ce7c25a78c9a89a6869484eb6c8", + "prepared_review_unit_hash": "84d7bb72d647c1bb8ca560704f137bc45ceb248477607a7fe242d2d2471003de", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 34, - "serialized_chars": 37563, - "estimated_input_tokens": 9391, - "estimated_eval_tokens": 19038 + "serialized_chars": 37560, + "estimated_input_tokens": 9390, + "estimated_eval_tokens": 19036 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly unrelated issue candidates from different regressions; only the non-persistent-buffer corruption reports are clear duplicates. I selected that regression as the canonical issue set.", - "confidence": 0.92, - "canonical_issue_reason": "Issue 43950 is the most explicit statement of the bug: `from_pretrained()` silently corrupts non-persistent buffers in transformers 5.x, which matches the closely related reports in 43644 and 44534.", + "summary": "The cluster is mostly heterogeneous, with one clear duplicate theme: the Transformers 5.x regression that corrupts non-persistent buffers during model loading. The rest of the issues cover unrelated model/config/runtime bugs.", + "confidence": 0.89, + "canonical_issue_reason": "Issue 43950 is the clearest and most specific report of the shared regression: `from_pretrained()` corrupting `register_buffer(persistent=False)` state in Transformers 5.x.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43950 is the best representative because it names the exact failing API and the corruption mechanism, making triage and fixing easier than the shorter duplicate titles.", + "best_issue_reason": "Issue 43950 best represents the duplicate pair because it names the concrete loading path and the exact non-persistent buffer corruption behavior.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43572", - "right": "issue:45070", + "left": "issue:44387", + "right": "issue:45005", "accept": false, - "reason": "Both involve v5 config-related breakage, but one is a missing `pad_token_idx` field and the other is a pydantic/PretrainedConfig field regression; not the same concrete bug." + "reason": "Both mention v5/quantization or tied weights, but they describe different failure modes and different code paths; not the same bug." }, { - "left": "issue:44387", - "right": "issue:45005", + "left": "issue:43572", + "right": "issue:45070", "accept": false, - "reason": "Different failures: int4 quantization memory growth vs tied-weights handling for translation models." + "reason": "One is a missing config field (`pad_token_idx`) and the other is a `PretrainedConfig`/pydantic field regression; too different to be a duplicate." }, { "left": "issue:43541", "right": "issue:45161", "accept": false, - "reason": "Different code paths and symptoms: torch dynamo/grouped_mm tracing failure vs tensor-parallel-only failure on GPT-OSS MoE." + "reason": "Grouped_mm torch-dynamo tracing errors and GPT-OSS tensor-parallel issues are distinct runtime problems." }, { "left": "issue:43828", "right": "issue:45464", "accept": false, - "reason": "Different model families and failure modes: autocast dtype mismatch vs streaming chat/completions API failure." + "reason": "Autocast dtype mismatch in Phi-tiny-MoE is unrelated to Qwen3.5 streaming inference failures." }, { "left": "issue:44291", "right": "issue:45070", "accept": false, - "reason": "`init_empty_weights` unexpected-argument bug is unrelated to the pydantic `PretrainedConfig` field issue." + "reason": "Unexpected `_is_hf_initialized` during `init_empty_weights` is a different initialization bug than the pydantic `PretrainedConfig` issue." }, { "left": "issue:43366", "right": "issue:45464", "accept": false, - "reason": "GGUF architecture support request is not the same as a Qwen3.5 streaming inference bug." + "reason": "GGUF architecture support for gpt-oss is unrelated to Qwen3.5 chat/completions streaming failures." }, { "left": "issue:43531", "right": "issue:45070", "accept": false, - "reason": "Qwen3-MoE sliding-window behavior is a model-specific runtime issue, not the config-field regression in 45070." + "reason": "A Qwen3-MoE sliding-window issue is not the same as a `PretrainedConfig`/pydantic field regression." }, { "left": "issue:44661", "right": "issue:45464", "accept": false, - "reason": "Tokenizer mapping/add-new-model-like failure is unrelated to the Qwen3.5 streaming inference error." - }, - { - "left": "issue:39290", - "right": "issue:43742", - "accept": false, - "reason": "Different models and failures: Gemma3/vLLM missing attribute vs MobileLLM key error." + "reason": "Tokenizer mapping failure when adding a new model is unrelated to Qwen3.5 streaming inference." }, { "left": "issue:43054", "right": "issue:43994", "accept": false, - "reason": "Both mention SigLIP2, but one is about worse text embeddings while the other is a broader nonsensical AutoModel/pipeline result; the concrete bug is not established as the same." + "reason": "Both involve SigLIP2, but one reports poor text embeddings while the other reports bad model/pipeline outputs; not clearly the same underlying bug." }, { - "left": "issue:43541", - "right": "issue:44589", + "left": "issue:39290", + "right": "issue:43742", "accept": false, - "reason": "Grouped_mm tracing failure and missing Float8 storage are distinct runtime errors." + "reason": "Gemma3 sliding-window/vLLM incompatibility and MobileLLM key loading errors are different model-specific problems." }, { "left": "issue:41093", "right": "issue:43596", "accept": false, - "reason": "Both are IndexErrors, but they happen in different loading paths and involve different shapes/components." + "reason": "These are both IndexErrors, but the mask-shape mismatch and zero3 empty-tensor initialization failure are different root causes." + }, + { + "left": "issue:43541", + "right": "issue:44589", + "accept": false, + "reason": "Torch-dynamo grouped_mm tracing on Mixtral is unrelated to a float8 storage lookup error." }, { "left": "issue:42915", "right": "issue:43541", "accept": false, - "reason": "FP8 config failure on Qwen3-MoE is unrelated to Mixtral dynamo tracing/grouped_mm." + "reason": "FineGrainedFP8 Qwen3-MoE failure and Mixtral grouped_mm tracing failure are different model/runtime issues." }, { "left": "issue:40990", "right": "issue:44960", "accept": false, - "reason": "No evidence they share a bug; perplexity regression on GPT-OSS and a generic GLM5 issue are different." + "reason": "High perplexity on GPT-OSS and a GLM5 issue are not the same bug." }, { "left": "issue:39186", "right": "issue:44918", "accept": false, - "reason": "FSDP `'weight' must be 2-D` is unrelated to TRL embedding unpacking." + "reason": "FSDP `'weight' must be 2-D` and TRL SFT input-unpacking failures are unrelated." }, { "left": "issue:43404", "right": "issue:45005", "accept": false, - "reason": "Both concern tied weights, but the affected models and failure modes differ; this is too broad to treat as the same concrete bug." + "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; not enough to treat as duplicates." }, { "left": "issue:43645", "right": "issue:45399", "accept": false, - "reason": "Custom model/Jupyter notebook initialization and flash-attn2 fallback logic are separate issues." + "reason": "Custom-model initialization in notebooks and flash-attn fallback gating are unrelated." }, { "left": "issue:44530", "right": "issue:45005", "accept": false, - "reason": "PagedAttentionCache linear_attention crash is unrelated to tied-weight regressions." + "reason": "PagedAttentionCache linear_attention crash is unrelated to tied-weights regressions." }, { "left": "issue:44863", "right": "issue:45005", "accept": false, - "reason": "NemotronH checkpoint loading failure is not the same as translation-model tied-weight issues." + "reason": "NemotronH checkpoint loading failure is unrelated to tied weights on translation models." }, { "left": "issue:43873", "right": "issue:45005", "accept": false, - "reason": "Quantization/offloading behavior and tied-weight handling are different problems." + "reason": "Quantization offloading behavior is a different subsystem from tied-weight handling." }, { "left": "issue:44360", "right": "issue:44512", "accept": false, - "reason": "A DSA indexer design discussion is unrelated to stale docs mentioning a removed command." + "reason": "A DSA indexer ReLU discussion and a docs typo are unrelated." }, { "left": "issue:42915", "right": "issue:45399", "accept": false, - "reason": "FP8 config support and flash-attn2 fallback checks are separate subsystems and failures." + "reason": "Qwen3-MoE FP8 failure and flash-attn fallback blocking are different problems." }, { "left": "issue:44918", "right": "issue:45507", "accept": false, - "reason": "TRL embedding unpacking and GraniteMoEHybrid invalid method calls do not look like the same bug." + "reason": "TRl SFT unpacking failure and GraniteMoEHybrid invalid-method calls are unrelated." }, { "left": "issue:43975", "right": "issue:44292", "accept": false, - "reason": "Incorrect detokenization and Qwen-3-8B-NVFP4 runtime error are different issues." + "reason": "Detokenization regression and Qwen-3-8B-NVFP4 runtime error are unrelated." }, { "left": "issue:44492", "right": "issue:44512", "accept": false, - "reason": "A typo in cache-strategy docs is unrelated to removed CLI-command documentation drift." + "reason": "A cache-strategy typo and removed CLI docs mention are unrelated documentation issues." }, { "left": "issue:43950", "right": "issue:44534", "accept": true, - "reason": "Both report the same non-persistent-buffer regression in transformers v5, with the latter phrased as the same junk-filled-buffer symptom." + "reason": "Both report the same Transformers v5 regression where non-persistent buffers are mishandled/corrupted during loading." }, { "left": "issue:43644", "right": "issue:43950", "accept": true, - "reason": "These describe the same v5 regression where non-persistent buffers are corrupted/fill with junk during `from_pretrained()`." + "reason": "Same bug: Transformers v5 corrupts non-persistent buffers; one report is the broader symptom, the other is the concrete `from_pretrained()` corruption path." }, { "left": "issue:44292", "right": "issue:45507", "accept": false, - "reason": "Qwen-3-8B-NVFP4 execution failure and GraniteMoEHybrid invalid-method calls are different model/runtime issues." + "reason": "Qwen-3-8B-NVFP4 runtime errors and GraniteMoEHybrid invalid-method calls are unrelated." }, { "left": "issue:43828", "right": "issue:44928", "accept": false, - "reason": "Autocast dtype mismatch and RLHF gradient explosion from 3D `position_ids` are different failure mechanisms." + "reason": "Different Qwen3.5-related failures: autocast dtype mismatch versus 3D position_ids / SDPA math fallback causing NaNs." }, { "left": "issue:39692", "right": "issue:43054", "accept": false, - "reason": "SigLIP2 docs-example errors and SigLIP2 text-embedding quality complaints are related by model family but not clearly the same concrete bug." + "reason": "SigLIP2 documentation example errors and SigLIP2 text-embedding quality regression are related only by model family, not the same concrete bug." }, { "left": "issue:43866", "right": "issue:44403", "accept": false, - "reason": "A corrupted checkpoint and an unrelated loading-noise issue are not the same underlying defect." + "reason": "Corrupted Ovis2 checkpoint and loading noise are not the same issue." }, { "left": "issue:41553", "right": "issue:43883", "accept": false, - "reason": "Bad AutoTokenizer error messaging and missing `all_tied_weights_keys` are unrelated." + "reason": "Bad AutoTokenizer error messaging for Voxtral is unrelated to missing `all_tied_weights_keys` on Molmo." }, { "left": "issue:36331", "right": "issue:41762", "accept": false, - "reason": "`compute_loss(num_items_in_batch)` compatibility and DeepSpeed ZeRO-3 loading `IndexError` are different bugs." + "reason": "A custom trainer signature mismatch and a DeepSpeed zero3 index error are unrelated." }, { "left": "issue:44560", "right": "issue:45507", "accept": false, - "reason": "Qwen3-vl video StopIteration and GraniteMoEHybrid invalid method calls are separate failures." + "reason": "Qwen3-vl embedding video StopIteration and GraniteMoEHybrid invalid-method calls are different runtime failures." } ] }, "evaluator_result": { "accept": true, - "feedback": "Grounded in the packet. The canonical issue choice is supported by the titles/body themes, and the only accepted soft-edge pairs are the clear non-persistent-buffer regression duplicates (43950/43644/44534). The remaining verdicts stay conservative and do not overclaim sameness." + "feedback": "The summary is grounded in the packet: the only clearly cohesive theme is the Transformers 5.x non-persistent-buffer regression, and the accepted duplicates (43644/43950/44534) are supported by the issue titles. The other soft-edge rejections stay conservative and do not overclaim duplication from the available evidence." }, "evaluator_used": true, "retried": false, @@ -999,8 +963,8 @@ ], "soft_pairs": [ "issue:43872|issue:44589", - "issue:41553|issue:43881", "issue:38175|issue:43054", + "issue:41553|issue:43881", "issue:42947|issue:44387", "issue:43054|issue:43493", "issue:43881|issue:43883", @@ -1008,11 +972,11 @@ "issue:42947|issue:43856", "issue:43975|issue:45520", "issue:43643|issue:45020", - "issue:44403|issue:44488", "issue:43782|issue:44960", + "issue:44403|issue:44488", "issue:42491|issue:44368", - "issue:41762|issue:42915", "issue:43975|issue:44451", + "issue:41762|issue:42915", "issue:45020|issue:45356", "issue:44661|issue:45020", "issue:43054|issue:43278", @@ -1020,8 +984,8 @@ "issue:44991|issue:45399", "issue:43827|issue:44509", "issue:43792|issue:44220", - "issue:45081|issue:45399", "issue:43404|issue:43950", + "issue:45081|issue:45399", "issue:43257|issue:43866", "issue:42915|issue:45020", "issue:43742|issue:45399", @@ -1031,252 +995,258 @@ "issue:43872|issue:44291", "issue:43854|issue:44863", "issue:43257|issue:45305", - "issue:42915|issue:43872", "issue:41553|issue:44488", + "issue:42915|issue:43872", "issue:36331|issue:41093" ], - "prepared_review_unit_hash": "cd54162ac02442011d7480b3d90ca64caf687ea69765ee8e33f62b638851dc20", + "prepared_review_unit_hash": "b32c67e71cd42b377a5acf5b391f40c620114610b9237e317ad6b682b8cbb1ff", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 36, - "serialized_chars": 38579, - "estimated_input_tokens": 9645, - "estimated_eval_tokens": 19546 + "serialized_chars": 38576, + "estimated_input_tokens": 9644, + "estimated_eval_tokens": 19544 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is mostly heterogeneous: many pairs share a subsystem or model family but not the same concrete bug. The clearest duplicate clusters are the stale v5 docs references, the `_is_hf_initialized` loading regression, and the SigLIP2 implementation discrepancy pair.", - "confidence": 0.74, - "canonical_issue_reason": "No single canonical issue spans the whole list; the items split into several unrelated bug reports. Among the true duplicate-like pairs, issue 43827 is the cleanest canonical anchor for the docs/pipeline-removal cluster.", + "summary": "Loose cluster of mostly model-loading, tokenizer, and training-regression issues across different models/subsystems. Issue 45020 is still the broadest issue anchor, but the soft pairs are mostly title-level similarities and should be rejected without stronger code-path evidence.", + "confidence": 0.66, + "canonical_issue_reason": "issue:45020 is the widest umbrella here: it describes a general recent-version regression breaking `remote_code` models, which best matches the cluster\u2019s repeated load/config failure pattern.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43827 is the strongest representative issue in this set because 44509 is essentially the same v5 docs problem with outdated pipeline/task references.", + "best_issue_reason": "issue:45020 is the most reusable anchor because it is broad, open, and centered on the shared regression theme rather than a single model-specific symptom.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43872", "right": "issue:44589", "accept": false, - "reason": "Both are loading-time type errors, but one is an `_is_hf_initialized`/bitsandbytes incompatibility and the other is a float8 storage lookup failure; different bugs." + "reason": "Different failure modes (`_is_hf_initialized` TypeError vs Float8 storage lookup error); title similarity is not enough to call them the same bug." }, { - "left": "issue:41553", - "right": "issue:43881", + "left": "issue:38175", + "right": "issue:43054", "accept": false, - "reason": "Different models and different failure modes; one is an AutoTokenizer error message issue, the other is a model loading failure." + "reason": "Both mention SigLIP/SigLIP2, but one is a zero-probability symptom and the other is degraded text embedding quality; no concrete shared code-path evidence." }, { - "left": "issue:38175", - "right": "issue:43054", + "left": "issue:41553", + "right": "issue:43881", "accept": false, - "reason": "Both mention SigLIP2, but one is zero probabilities and the other is lower embedding quality; not enough evidence they are the same concrete defect." + "reason": "Different model-loading problems with different symptoms and no indication of the same underlying defect." }, { "left": "issue:42947", "right": "issue:44387", "accept": false, - "reason": "Both are memory-related, but one is LoRA/gradient-checkpointing ineffectiveness and the other is int4 CUDA reserved-memory growth causing OOM." + "reason": "Both are training regressions, but one is LoRA checkpointing behavior and the other is int4/OOM memory usage; too different to merge." }, { "left": "issue:43054", "right": "issue:43493", - "accept": true, - "reason": "Both point to the same SigLIP2 HF-vs-JAX implementation discrepancy, with degraded text embeddings as the symptom." + "accept": false, + "reason": "Strong title similarity around SigLIP2, but the packet doesn\u2019t show enough to prove the same bug beyond broad topic overlap." }, { "left": "issue:43881", "right": "issue:43883", "accept": false, - "reason": "Different models and different issues: glm-4v loading vs a missing `all_tied_weights_keys` attribute in Molmo." + "reason": "Both are load failures, but they affect different models and different code paths." }, { "left": "issue:44403", "right": "issue:44451", "accept": false, - "reason": "One is about noisy loading warnings, the other is an actual inability to load a specific model." + "reason": "One is about noisy transformer loading output, the other is a specific model load failure; not the same issue." }, { "left": "issue:42947", "right": "issue:43856", "accept": false, - "reason": "Both concern memory, but one is checkpointing/LoRA behavior and the other is Qwen3 MoE training memory usage." + "reason": "Both involve Qwen3 training, but one is gradient-checkpointing/LoRA behavior and the other is memory efficiency in MoE training." }, { "left": "issue:43975", "right": "issue:45520", "accept": false, - "reason": "Unrelated problems: tokenizer detokenization regression vs Python 3.13 flash_attn import bookkeeping." + "reason": "Tokenizer detokenization regression and Python 3.13 import-utils KeyError are unrelated subsystems." }, { "left": "issue:43643", "right": "issue:45020", "accept": false, - "reason": "Both involve `remote_code`, but one is missing fields from `AutoConfig` and the other is broad breakage in recent versions." + "reason": "Related to remote-code loading, but one is missing fields in `AutoConfig` output and the other is a broader breakage report; not enough to merge." }, { - "left": "issue:44403", - "right": "issue:44488", + "left": "issue:43782", + "right": "issue:44960", "accept": false, - "reason": "Both are loading-related, but one is a warning/noise report and the other is a specific model load failure." + "reason": "Different models and different failures; no shared concrete bug is visible." }, { - "left": "issue:43782", - "right": "issue:44960", + "left": "issue:44403", + "right": "issue:44488", "accept": false, - "reason": "Different model families and errors; no clear shared code-path bug." + "reason": "Both are loading complaints, but they describe separate model-specific failures." }, { "left": "issue:42491", "right": "issue:44368", "accept": false, - "reason": "Both mention Qwen LoRA/tie-word-embeddings, but they describe different behaviors and code paths." + "reason": "A LoRA compatibility break and a tie-word-embeddings warning are not the same defect." }, { - "left": "issue:41762", - "right": "issue:42915", + "left": "issue:43975", + "right": "issue:44451", "accept": false, - "reason": "Different model-specific DeepSpeed failures; one is a Gemma3 ZeRO-3 load error, the other is Qwen3 MoE plus FP8 config." + "reason": "Tokenizer/detokenization regression versus inability to load a specific model; different bugs." }, { - "left": "issue:43975", - "right": "issue:44451", + "left": "issue:41762", + "right": "issue:42915", "accept": false, - "reason": "Different issues entirely: tokenizer detokenization regression vs model loading failure." + "reason": "Both touch distributed loading, but the model and failure modes differ too much." }, { "left": "issue:45020", "right": "issue:45356", "accept": false, - "reason": "Both involve recent-version regressions, but one is broad `remote_code` breakage and the other is a Kimi-K2.5 tokenizer codec/warning regression." + "reason": "Both are regressions in newer versions, but one is remote-code loading and the other is tokenizer/codec handling; different code paths." }, { "left": "issue:44661", "right": "issue:45020", "accept": false, - "reason": "`add-new-model-like`/tokenizer mapping failure is a different bug from the broader remote_code regressions." + "reason": "Different loading/registration failures; the shared theme is too broad to treat as one bug." }, { "left": "issue:43054", "right": "issue:43278", "accept": false, - "reason": "Both mention embeddings, but one is a SigLIP2 quality discrepancy and the other is a dtype change between train and eval." + "reason": "SigLIP2 embedding quality vs training/evaluate dtype mismatch are distinct symptoms." }, { "left": "issue:43824", "right": "issue:43872", "accept": false, - "reason": "Different load errors: import failure for Qwen2.5-VL vs bitsandbytes `_is_hf_initialized` incompatibility." + "reason": "Missing import and bitsandbytes TypeError are different errors with no shared underlying fix shown." }, { "left": "issue:44991", "right": "issue:45399", "accept": false, - "reason": "Unrelated: tokenizer loading regression vs flash-attn fallback gating logic." + "reason": "Tokenizer load failure and flash-attn fallback gating are unrelated." }, { "left": "issue:43827", "right": "issue:44509", - "accept": true, - "reason": "Same docs bug: stale v5 references to removed text-generation/summarization/translation pipeline tasks." + "accept": false, + "reason": "Both are docs-related and mention removed pipelines, but the packet only gives title-level overlap; that is not enough to confidently merge." }, { "left": "issue:43792", "right": "issue:44220", "accept": false, - "reason": "Likely related to Whisper audio preprocessing, but the reports are too different to confidently call the same concrete bug." + "reason": "The Whisper audio frontend association is plausible, but the visible evidence is still too thin to assert the same underlying bug." }, { - "left": "issue:45081", - "right": "issue:45399", + "left": "issue:43404", + "right": "issue:43950", "accept": false, - "reason": "Tokenizer regex crash and flash-attn fallback gating are separate issues." + "reason": "Tied lm_head weights and non-persistent buffer corruption are different model-state issues." }, { - "left": "issue:43404", - "right": "issue:43950", + "left": "issue:45081", + "right": "issue:45399", "accept": false, - "reason": "Different problems: Mistral3 weight tying vs silent corruption of non-persistent buffers." + "reason": "Tokenizer regex patch crash and flash-attn fallback checks are unrelated." }, { "left": "issue:43257", "right": "issue:43866", "accept": false, - "reason": "Different model and failure mode; Qwen3 MoE weight conversion vs a corrupted Ovis2 checkpoint." + "reason": "Different model families and different load failures; no shared code-path bug is visible." }, { - "left": "issue:43872", - "right": "issue:44291", - "accept": true, - "reason": "Same `_is_hf_initialized` loading regression; one report surfaces through bitsandbytes, the other through `init_empty_weights`." + "left": "issue:42915", + "right": "issue:45305", + "accept": false, + "reason": "Both involve DeepSpeed training behavior, but they describe different regressions and should not be merged." }, { - "left": "issue:43854", - "right": "issue:44863", + "left": "issue:43742", + "right": "issue:45399", "accept": false, - "reason": "Different models and different loading failures; not the same bug." + "reason": "A specific model-loading key error and a flash-attn fallback issue are not the same bug." }, { - "left": "issue:43257", - "right": "issue:45305", + "left": "issue:43854", + "right": "issue:44863", "accept": false, - "reason": "Both involve Qwen/DeepSpeed, but one is weight conversion and the other is gradient averaging with GAS/model_accepts_loss_kwargs." + "reason": "Different model-specific load failures; no evidence they come from one underlying defect." }, { "left": "issue:41553", - "right": "issue:44488", + "right": "issue:44991", "accept": false, - "reason": "Different models and different failure modes; one is a bad tokenizer error, the other a model load failure." + "reason": "One is a bad error message for tokenizer loading, the other is a tokenizer load regression; similar domain, different bug." }, { "left": "issue:43940", "right": "issue:45313", "accept": false, - "reason": "Both are DeepSpeed ZeRO-3 load failures, but they affect different Qwen model variants and do not read as the same concrete bug." + "reason": "Both are Qwen3/ZeRO-3-related, but the weight-loading failures are not shown to be the same code-path problem." }, { - "left": "issue:43742", - "right": "issue:45399", + "left": "issue:43872", + "right": "issue:44291", "accept": false, - "reason": "KeyError during model loading vs flash-attn fallback gating; unrelated." + "reason": "Both involve unexpected argument/init-empty-weights style loading errors, but the packet doesn\u2019t show enough to call them the same bug." }, { - "left": "issue:43742", - "right": "issue:45020", + "left": "issue:43854", + "right": "issue:45305", + "accept": false, + "reason": "Model load failure versus gradient averaging under DeepSpeed are unrelated." + }, + { + "left": "issue:43257", + "right": "issue:45305", "accept": false, - "reason": "Different bugs: a specific MobileLLM key error versus broad remote_code regressions." + "reason": "Different Qwen3 regressions affecting different parts of training/loading." }, { "left": "issue:41553", "right": "issue:44488", "accept": false, - "reason": "Different models and different loading paths; no shared concrete defect." + "reason": "Different model-loading issues with no clear shared defect." }, { "left": "issue:36331", "right": "issue:41093", "accept": false, - "reason": "Trainer loss-signature mismatch and tensor-mask shape mismatch are unrelated." + "reason": "Trainer API signature mismatch and tensor mask shape mismatch are completely different failures." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet, and the accepted soft-edge pairs are conservative enough to look mergeable into single bugfix PRs. The rejected pairs are appropriately cautious about shared subsystems without enough evidence of the same concrete defect." + "accept": false, + "feedback": "The summary is mostly grounded, but the soft-edge handling is too assertive for this packet. The analyst marked several pairs as clear duplicates/mergeable bugfixes even though the packet\u2019s soft-pair labels are negative and the visible evidence is only title-level similarity. In particular, the accepted pairs `43054\u219443493`, `43827\u219444509`, and `43792\u219444220` are stated too strongly as the same bug without enough support from the packet. The conservative move here would be to keep these as tentative or reject them. Because of those over-accepts, I can\u2019t accept the evaluation." }, "evaluator_used": true, - "retried": false, - "accepted_nontrivial_soft_edge": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -1345,16 +1315,16 @@ "soft_pairs": [ "issue:41093|issue:44805", "issue:44479|issue:44560", - "issue:43872|issue:45538", "issue:44530|issue:44863", + "issue:43872|issue:45538", "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:43366|issue:44589", "issue:43824|issue:43957", "issue:42915|issue:43957", "issue:42617|issue:43366", @@ -1363,8 +1333,8 @@ "issue:44560|issue:44805", "issue:36010|issue:42915", "issue:43638|issue:44661", - "issue:43596|issue:44661", "issue:44877|issue:45030", + "issue:43596|issue:44661", "issue:43644|issue:44534", "issue:43317|issue:44756", "issue:38175|issue:39692", @@ -1374,20 +1344,20 @@ "issue:44451|issue:44488", "issue:43708|issue:44075" ], - "prepared_review_unit_hash": "82a22d2b9d6bea5bebaa223468763440acac526f44b2909f668d5be4c321fc28", + "prepared_review_unit_hash": "b23da9691ecaec6332738cccfc81f99ccd6b6bbed5e41de592d3d6fbad79fae6", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36645, + "serialized_chars": 36647, "estimated_input_tokens": 9162, "estimated_eval_tokens": 18580 }, @@ -1396,201 +1366,201 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous, with several unrelated model-loading/regression reports. Only a few soft pairs look like true duplicates; the strongest ones are the Transformers 5 non-persistent-buffer regression and the custom model initialization regression.", - "confidence": 0.52, - "canonical_issue_reason": "issue:43644 is the clearest broad regression in the set and has an obvious duplicate in issue:44534; it best represents a concrete, actionable underlying bug.", + "summary": "The cluster is mostly well-triaged, but the ScandiBERT/sleng-bert pair is too weak to merge: both are BERT-style model loading failures, yet the packet does not show enough evidence that they share the same concrete regression. That pair should be rejected, while the rest of the conservative duplicate decisions remain reasonable.", + "confidence": 0.76, + "canonical_issue_reason": "issue:43644 is still the cleanest canonical issue in this packet: it is a specific Transformers v5 non-persistent-buffer regression and has a near-verbatim duplicate in issue:44534.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43644 is the best single representative of the duplicate-prone Transformers 5 regression subset, with a clear and general failure mode.", + "best_issue_reason": "issue:43644 is the strongest representative issue because it describes a concrete, regression-style bug with clear scope and a close duplicate that matches the same failure mode.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41093", "right": "issue:44805", - "accept": true, - "reason": "Same IndexError class and the same mask-vs-tensor shape mismatch pattern; these look like the same underlying bug." + "accept": false, + "reason": "Both are mask/tensor shape mismatch errors, but the packet does not show enough detail to prove they are the same underlying bug." }, { "left": "issue:44479", "right": "issue:44560", "accept": false, - "reason": "Both are Qwen video-related, but the affected models and failure modes differ enough that they do not clearly point to the same concrete bug." + "reason": "Both touch Qwen video-related paths, but the failures are different enough that they are not clearly the same bug." }, { - "left": "issue:43872", - "right": "issue:45538", + "left": "issue:44530", + "right": "issue:44863", "accept": false, - "reason": "Completely different subsystems: bitsandbytes quantization error vs CLIP tokenizer max length behavior." + "reason": "Different models and different failure modes: PagedAttentionCache crash versus NemotronH checkpoint loading." }, { - "left": "issue:44530", - "right": "issue:44863", + "left": "issue:43872", + "right": "issue:45538", "accept": false, - "reason": "Different models and different runtime failures; too broad to treat as the same bug." + "reason": "Unrelated issues: bitsandbytes Int8Params initialization versus CLIPTokenizer max length behavior." }, { "left": "issue:43366", "right": "issue:43828", "accept": false, - "reason": "GGUF/gpt-oss support is unrelated to the Phi-tiny-MoE autocast dtype mismatch." + "reason": "Different code paths and symptoms: GGUF/gpt-oss support versus dtype mismatch under autocast." }, { "left": "issue:24643", "right": "issue:30064", "accept": false, - "reason": "DeepSpeed training weight-shape error vs image processor void segmentation map handling are unrelated." + "reason": "DeepSpeed training weight-shape error is unrelated to image processor segmentation-map handling." }, { "left": "issue:43366", "right": "issue:43531", "accept": false, - "reason": "Both mention Qwen-related areas, but one is GGUF/gpt-oss support and the other is a sliding_window issue; not the same bug." + "reason": "No concrete shared bug: GGUF model support is unrelated to the sliding_window Qwen3-MoE issue." }, { "left": "issue:43550", "right": "issue:44910", "accept": false, - "reason": "Different model families and different code paths: torch.compile/SDPA on Bamba vs flash-attention position_ids on Qwen3.5." + "reason": "Different models and different failures: Bamba torch.compile/SDPA versus Qwen3.5 flash-attention memory access." }, { "left": "issue:43645", "right": "issue:43646", "accept": true, - "reason": "These describe the same Transformers 5 custom model initialization regression; the notebook mention in 43645 is just a narrower reproduction." + "reason": "These are the same Transformers 5.0.0 custom-model initialization regression, with one being a notebook-specific manifestation." }, { "left": "issue:43950", "right": "issue:45325", "accept": false, - "reason": "Both are regression-style loading issues, but one is non-persistent buffers in from_pretrained and the other is a Qwen2.5-VL rope/position_ids bug." + "reason": "Non-persistent buffer corruption and Qwen2.5-VL rope-index scaling are different bugs." }, { - "left": "issue:43366", - "right": "issue:44589", + "left": "issue:44560", + "right": "issue:44918", "accept": false, - "reason": "GGUF gpt-oss support is unrelated to the Float8 storage lookup failure." + "reason": "Both mention Qwen video or embedding behavior, but the concrete failures and affected paths differ." }, { - "left": "issue:44560", - "right": "issue:44918", + "left": "issue:43366", + "right": "issue:44589", "accept": false, - "reason": "Both involve Qwen3.5 video/input handling, but the observed failures and affected paths are distinct." + "reason": "No shared underlying bug; gpt-oss support is unrelated to the missing Float8 storage type." }, { "left": "issue:43824", "right": "issue:43957", "accept": false, - "reason": "Importing a missing class is not the same issue as meta-device loading regressions." + "reason": "Different import/load problems involving different models and mechanisms." }, { "left": "issue:42915", "right": "issue:43957", "accept": false, - "reason": "Qwen3Moe FP8 config failure and meta-device loading regressions are different code-path problems." + "reason": "Qwen3Moe FP8 config failure is unrelated to meta-device model loading regressions." }, { "left": "issue:42617", "right": "issue:43366", "accept": false, - "reason": "3d_parallel.py execution failure is unrelated to GGUF gpt-oss support." + "reason": "3d_parallel.py runtime failure does not match GGUF/gpt-oss support." }, { "left": "issue:38175", "right": "issue:43994", - "accept": false, - "reason": "Both are SigLIP2-related, but one is zero probabilities while the other is an AutoModel/pipeline loading issue; not clearly the same bug." + "accept": true, + "reason": "Same SigLIP2 bad-output symptom for the same model family; the later report is a near-duplicate manifestation." }, { "left": "issue:43646", "right": "issue:43950", "accept": false, - "reason": "Custom model initialization regression is different from non-persistent buffer corruption during from_pretrained." + "reason": "Custom model initialization regression is not the same as non-persistent buffer corruption." }, { "left": "issue:44560", "right": "issue:44805", "accept": false, - "reason": "Both surface as runtime errors, but one is a Qwen video StopIteration issue and the other is a generic mask/tensor shape mismatch." + "reason": "Different errors and domains: Qwen3-vl video StopIteration versus generic mask/tensor shape mismatch." }, { "left": "issue:36010", "right": "issue:42915", "accept": false, - "reason": "ImportError for GenerationMixin and Qwen3Moe FP8 failure are unrelated." + "reason": "GenerationMixin ImportError is unrelated to the Qwen3Moe FP8 loading issue." }, { "left": "issue:43638", "right": "issue:44661", "accept": false, - "reason": "Different failures: DeepSpeed zero3 index error vs add-new-model-like tokenizer mapping problem." + "reason": "Zero3/BertModel index error is unrelated to the TOKENIZER_MAPPING_NAMES limitation." }, { - "left": "issue:43596", - "right": "issue:44661", + "left": "issue:44877", + "right": "issue:45030", "accept": false, - "reason": "Both are initialization/config-related, but the concrete bugs and code paths differ." + "reason": "Both are config-related, but they are different model/config regressions." }, { - "left": "issue:44877", - "right": "issue:45030", + "left": "issue:43596", + "right": "issue:44661", "accept": false, - "reason": "Strict config loading for granite_speech and glm4v config validation are separate config-regression issues." + "reason": "Different failures: DeepSpeed ZeRO-3 empty tensor index error versus tokenizer-mapping limitation." }, { "left": "issue:43644", "right": "issue:44534", "accept": true, - "reason": "Near-identical reports of the Transformers v5 non-persistent-buffer corruption regression." + "reason": "Same Transformers v5 non-persistent-buffer regression; the titles are effectively describing the same bug." }, { "left": "issue:43317", "right": "issue:44756", "accept": false, - "reason": "device_map/offload loading failure and Strix Halo mmap OOM are different loading/performance issues." + "reason": "device_map/offload loading failure is unrelated to the Strix Halo mmap/OOM request." }, { "left": "issue:38175", "right": "issue:39692", "accept": false, - "reason": "Same model family, but one is wrong outputs and the other is documentation/example errors plus quantization failure." + "reason": "Related model family, but not the same concrete bug: bad probabilities versus docs example mismatch/quantization failure." }, { "left": "issue:45313", "right": "issue:45357", "accept": false, - "reason": "Both are Qwen3.5 regressions, but one is DeepSpeed ZeRO-3 weight loading and the other is save_pretrained key persistence." + "reason": "Different Qwen3.5 issues: ZeRO-3 weight loading versus incorrect visual encoder save_pretrained keys." }, { "left": "issue:42491", "right": "issue:43720", "accept": false, - "reason": "LoRA qwen3_moe compatibility on hf5.x and BitNet packed-weight loading are unrelated." + "reason": "LoRA/Qwen3_moe compatibility across versions is unrelated to BitNet packed-weight loading." }, { "left": "issue:44960", "right": "issue:44977", "accept": false, - "reason": "GLM5 placeholder issue and Qwen3.5 flash-attention generation failure are unrelated." + "reason": "GLM5 and Qwen3.5 flash-attention problems are unrelated." }, { "left": "issue:44451", "right": "issue:44488", "accept": false, - "reason": "Both are model-loading complaints, but the affected checkpoints and likely root causes are different; not enough evidence for a duplicate." + "reason": "The packet only shows two different BERT-style checkpoints that fail to load; that suggests a possible shared loader regression, but not enough evidence to confidently say it is the same bug." }, { "left": "issue:43708", "right": "issue:44075", "accept": false, - "reason": "Trainer checkpoint resume batch-size math and SGD optimizer args usage are different training bugs." + "reason": "Trainer max_steps resume bug is unrelated to SGD argument handling." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative about duplicate claims. The accepted soft pairs are well-supported by highly similar titles and failure modes, especially the 43644/44534 non-persistent-buffer regression. No pair is overstated, and the cluster-level characterization as mostly heterogeneous is reasonable." + "accept": false, + "feedback": "Most verdicts are conservative and grounded in the packet, but the accepted pair issue:44451 \u2194 issue:44488 is too strong from the available evidence. The titles only show two different BERT-style checkpoints that both fail to load; that suggests a possible shared loading regression, but not enough to confidently mark them as the same bug. Because the analyst overstates that link, the overall result should be rejected." }, "evaluator_used": true, - "retried": false, + "retried": true, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -1661,16 +1631,16 @@ "issue:43854|issue:43931", "issue:43856|issue:43873", "issue:42491|issue:45276", - "issue:45362|issue:45406", - "issue:41720|issue:42915", "issue:39692|issue:43994", "issue:43257|issue:43854", + "issue:41720|issue:42915", "issue:43856|issue:44387", + "issue:45362|issue:45406", "issue:44488|issue:44991", "issue:43644|issue:43645", - "issue:44661|issue:44683", "issue:42915|issue:44756", "issue:42491|issue:43931", + "issue:44661|issue:44683", "issue:42491|issue:43856", "issue:43404|issue:43454", "issue:45276|issue:45335", @@ -1689,8 +1659,8 @@ "issue:45276|issue:45292", "issue:44743|issue:44960", "issue:43381|issue:44928", - "issue:43957|issue:44756", "issue:44164|issue:44756", + "issue:43957|issue:44756", "issue:43646|issue:43708", "issue:45292|issue:45335", "issue:44479|issue:44877", @@ -1700,20 +1670,20 @@ "issue:43425|issue:43828", "issue:43828|issue:43957" ], - "prepared_review_unit_hash": "b911ae06f8b97c8d9e384f2c02aca318c3a9d47f25989add48262015a6eb3b9c", + "prepared_review_unit_hash": "d48da9f8fd85f8d80fc86562a838683c54e94a870feb1a7cc2553a304b912357", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 41, - "serialized_chars": 39810, + "serialized_chars": 39812, "estimated_input_tokens": 9953, "estimated_eval_tokens": 20162 }, @@ -1722,264 +1692,264 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most items are unrelated, but there is a small duplicate set around `resize_token_embeddings` not updating model output/decoder embeddings. The generic open issue is the best representative; no PRs are present.", - "confidence": 0.82, - "canonical_issue_reason": "`issue:45292` is the broadest and most general report of the shared bug: `resize_token_embeddings` not affecting output embeddings. It cleanly covers the model-specific reports in the accepted soft pairs and is still open.", + "summary": "Most pairs are false positives: they share broad themes like loading, quantization, or memory, but not the same concrete bug. The only convincing duplicate family is the `resize_token_embeddings` regression across Gemma/T5Gemma variants, where multiple reports point to the same missing propagation into nested embedding/output modules.", + "confidence": 0.86, + "canonical_issue_reason": "Issue 45292 is the broadest and most central report in the `resize_token_embeddings` regression family, and it best covers the model-specific follow-ups about nested output/decoder embeddings.", "canonical_pr_reason": null, - "best_issue_reason": "`issue:45292` is the best cluster representative because it states the underlying behavior generically, rather than for one specific model family, and matches the duplicate resize-token-embedding reports.", + "best_issue_reason": "Issue 45292 is the best representative issue because it states the generic symptom clearly and matches the closely related Gemma/T5Gemma reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43854", "right": "issue:43931", "accept": false, - "reason": "Both are model-loading failures, but they involve different models and different failure modes (`GLM-4.7-Flash` unit-test loading vs `Qwen3-VL-30B` weight-shape mismatch)." + "reason": "Different models and different failure modes: GLM-4.7 test loading vs Qwen3-VL weight-shape mismatch." }, { "left": "issue:43856", "right": "issue:43873", "accept": false, - "reason": "Both mention memory/quantization, but one is about Qwen3 MoE training memory usage and the other about offloading behavior with quantization; not the same concrete bug." + "reason": "Both mention Qwen3/MoE and memory, but one is training-memory inefficiency and the other is quantization/offloading behavior." }, { "left": "issue:42491", "right": "issue:45276", "accept": false, - "reason": "Different problems: a Qwen3 MoE LoRA compatibility break vs a `resize_token_embeddings` embedding-update bug." - }, - { - "left": "issue:45362", - "right": "issue:45406", - "accept": false, - "reason": "Different regressions in different entry points: `transformers chat` crash for Qwen3.5-35B vs `serve` crash because `Gemma4Processor` lacks `_tokenizer`." - }, - { - "left": "issue:41720", - "right": "issue:42915", - "accept": false, - "reason": "Both concern Qwen3/Qwen3-MoE, but one is a CUDA assert during auto device mapping and the other is a FineGrainedFP8Config failure; different code paths." + "reason": "Unrelated problems: LoRA checkpoint compatibility across branches vs Gemma4 token embedding resizing." }, { "left": "issue:39692", "right": "issue:43994", "accept": false, - "reason": "Both involve SigLIP2-style usage, but one is a documentation example with model/processor and quantization errors, while the other is nonsensical outputs in AutoModel/pipeline usage." + "reason": "Both involve SigLIP2, but one is a documentation/example error and the other is a runtime inference quality issue; not the same concrete bug." }, { "left": "issue:43257", "right": "issue:43854", "accept": false, - "reason": "Different loading failures: Qwen3 MoE weights not converted under accelerate+deepspeed vs a unit-test load failure for GLM-4.7-Flash." + "reason": "Qwen3 MoE loading/conversion bug vs GLM-4.7 test loading bug; different code paths." + }, + { + "left": "issue:41720", + "right": "issue:42915", + "accept": false, + "reason": "Both are Qwen3-related, but one is device-mapping CUDA assert and the other is FP8 config failure." }, { "left": "issue:43856", "right": "issue:44387", "accept": false, - "reason": "One is Qwen3 MoE training memory usage; the other is increased CUDA reserved memory under int4 quantization. Same theme, but not the same bug." + "reason": "Both are memory-related, but the causes differ: training memory usage vs int4 quantization OOM regression." + }, + { + "left": "issue:45362", + "right": "issue:45406", + "accept": false, + "reason": "Different models and subsystems: Qwen3.5 chat failure vs Gemma4 processor attribute error in serve." }, { "left": "issue:44488", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer/model loading issues, but they concern different models and different failure causes." + "reason": "Both are tokenizer/model loading issues, but they affect different models and there is no concrete shared code-path evidence." }, { "left": "issue:43644", "right": "issue:43645", "accept": false, - "reason": "One is about non-persistent buffers being junk-filled; the other is a notebook-specific custom-model initialization regression." - }, - { - "left": "issue:44661", - "right": "issue:44683", - "accept": false, - "reason": "Different features: `add-new-model-like` failing in tokenizer mapping vs compiled flex_attention failing on newer torch." + "reason": "Transformers 5 buffer initialization bug vs Jupyter custom-model initialization regression; not the same underlying problem." }, { "left": "issue:42915", "right": "issue:44756", "accept": false, - "reason": "Different bug classes: Qwen3 MoE FP8 loading failure vs disabling mmap on Strix Halo to avoid OOM." + "reason": "FP8 quantization failure vs mmap/OOM mitigation on Strix Halo; unrelated bugs." }, { "left": "issue:42491", "right": "issue:43931", "accept": false, - "reason": "Different models and different failure surfaces: LoRA incompatibility for Qwen3 MoE vs Qwen3-VL weight-shape mismatch." + "reason": "Different families: Qwen3 MoE LoRA compatibility vs Qwen3-VL checkpoint shape mismatch." + }, + { + "left": "issue:44661", + "right": "issue:44683", + "accept": false, + "reason": "Tokenizer-mapping/model-addition bug vs compiled flex_attention on torch>=2.9; unrelated." }, { "left": "issue:42491", "right": "issue:43856", "accept": false, - "reason": "Different Qwen3 MoE reports, but one is LoRA compatibility and the other is training memory usage; not the same concrete bug." + "reason": "Same broad Qwen3/MoE area, but one is checkpoint compatibility and the other is training memory usage." }, { "left": "issue:43404", "right": "issue:43454", "accept": false, - "reason": "Both mention lm_head tying in multimodal models, but they are separate model implementations (Mistral3 vs AyaVision) with different reports." + "reason": "Similar symptom class (untied lm_head), but different model implementations and no evidence of one shared code path." }, { "left": "issue:45276", "right": "issue:45335", "accept": true, - "reason": "Both report the same underlying bug: `resize_token_embeddings` does not update model-specific output/decoder embeddings. The model families differ, but the concrete failure is the same." + "reason": "Both report the same `resize_token_embeddings` regression failing to update nested embedding/output modules in model-specific wrappers; a single fix could plausibly cover both." }, { "left": "issue:36010", "right": "issue:39404", "accept": false, - "reason": "An import error for `GenerationMixin` is unrelated to Whisper pipeline `return_language` behavior." + "reason": "ImportError for GenerationMixin vs Whisper pipeline return_language regression; unrelated." }, { "left": "issue:44164", "right": "issue:44291", "accept": false, - "reason": "Both are loading/saving related, but `extra_state` handling and `init_empty_weights` argument errors are different code paths." + "reason": "Save/load extra_state handling vs init_empty_weights unexpected argument; different serialization/loading bugs." }, { "left": "issue:41720", "right": "issue:44155", "accept": false, - "reason": "CUDA assert on Qwen3 auto device mapping is unrelated to AudioFlamingo3 batched inference token/embedding leakage." + "reason": "CUDA assert during Qwen3 device mapping vs AudioFlamingo3 batched inference token leakage; unrelated." }, { "left": "issue:44402", "right": "issue:44488", "accept": false, - "reason": "Tokenizer vocab-size mismatch for one model is not the same as the broader model-loading failure in the other issue." + "reason": "Different tokenizer/model loading failures for different BERT models; not the same bug." }, { "left": "issue:43873", "right": "issue:44387", "accept": false, - "reason": "Both are memory-related, but offloading-with-quantization and int4 reserved-memory inflation are different defects." + "reason": "Quantization/offloading behavior vs int4 reserved-memory OOM; similar theme, different root cause." }, { "left": "issue:39692", "right": "issue:44960", "accept": false, - "reason": "SigLIP2 doc/example issues are unrelated to the generic GLM5 issue." + "reason": "SigLIP2 doc/example problem vs generic GLM5 issue; unrelated." }, { "left": "issue:44387", "right": "issue:44756", "accept": false, - "reason": "Int4 quantization OOM and disabling mmap on Strix Halo are different memory-management problems." + "reason": "Both can lead to OOM, but one is a Transformers 5 quantization regression and the other is a platform-specific mmap workaround." }, { "left": "issue:43425", "right": "issue:43957", "accept": false, - "reason": "Torch 2.10 incompatibility is unrelated to meta-device model-loading breakage." + "reason": "Torch version incompatibility vs meta-device loading regression; different problems." }, { "left": "issue:44292", "right": "issue:44912", "accept": false, - "reason": "Both are quantization/model-loading issues, but they affect different models and formats (NVFP4 vs MXFP4) with different symptoms." + "reason": "Different quantization formats and models: NVFP4 load error vs MXFP4 fallback to bf16." }, { "left": "issue:42947", "right": "issue:43381", "accept": false, - "reason": "Both mention gradient checkpointing, but one is ineffective under PEFT LoRA and the other says it cannot be used in eval mode." + "reason": "Gradient checkpointing complaints, but one is ineffective under PEFT LoRA and the other is eval-mode prohibition." }, { "left": "issue:43854", "right": "issue:45216", "accept": false, - "reason": "Different regressions: GLM-4.7-Flash loading in tests vs Qwen3.5 save_pretrained checkpoint correctness." + "reason": "GLM-4.7 test loading failure vs Qwen3.5 save_pretrained regression; unrelated." }, { "left": "issue:44291", "right": "issue:44756", "accept": false, - "reason": "`init_empty_weights` argument handling and Strix Halo mmap OOM are unrelated loading/memory issues." + "reason": "init_empty_weights/_is_hf_initialized TypeError vs mmap/OOM issue; unrelated." }, { "left": "issue:45276", - "right": "issue:45292", + "right": "issue:45335", "accept": true, - "reason": "Same bug family: `resize_token_embeddings` fails to propagate to output/decoder embeddings. `issue:45292` is the generic version of the same defect." + "reason": "Both are the same `resize_token_embeddings` family of bugs affecting nested embedding/output attributes across different model wrappers." }, { "left": "issue:44743", "right": "issue:44960", "accept": false, - "reason": "A recurrent-state/cache bug in modular_qwen3_5.py is unrelated to the generic GLM5 issue." + "reason": "Qwen3.5 recurrent-state cache bug vs GLM5 issue; unrelated." }, { "left": "issue:43381", "right": "issue:44928", "accept": false, - "reason": "Eval-mode gradient checkpointing and RLHF NaN/position_ids issues are different failure modes." + "reason": "Eval-mode gradient checkpointing rule vs Qwen3.5 RLHF gradient explosion from 3D position_ids; unrelated." }, { - "left": "issue:43957", + "left": "issue:44164", "right": "issue:44756", "accept": false, - "reason": "Meta-device loading breakage is unrelated to the mmap/OOM workaround issue." + "reason": "Save/from_pretrained extra_state handling vs mmap/OOM workaround; different code paths." }, { - "left": "issue:44164", + "left": "issue:43957", "right": "issue:44756", "accept": false, - "reason": "`extra_state` save/load handling is unrelated to Strix Halo mmap OOM." + "reason": "Meta-device model loading regression vs Strix Halo mmap/OOM issue; not the same bug." }, { "left": "issue:43646", "right": "issue:43708", "accept": false, - "reason": "Custom model initialization regressions and `resume_from_checkpoint` max_steps miscalculation are different trainer/model lifecycle bugs." + "reason": "Custom model init regression vs Trainer resume/max_steps calculation bug; unrelated." }, { "left": "issue:45292", "right": "issue:45335", "accept": true, - "reason": "Both describe the same `resize_token_embeddings` regression; `issue:45292` is the generic umbrella report and `issue:45335` is a model-specific instance." + "reason": "Same `resize_token_embeddings` regression: a generic output_embeddings report and a T5Gemma decoder.embed_tokens report of the same missing propagation behavior." }, { "left": "issue:44479", "right": "issue:44877", "accept": false, - "reason": "Different model families and different problems: video-input regression for Qwen models vs strict config blocking `granite_speech`." + "reason": "Video-input regression for several Qwen VL variants vs strict config blocking granite_speech loading; unrelated." }, { "left": "issue:44402", "right": "issue:44451", "accept": false, - "reason": "A tokenizer vocab-size mismatch is not the same as the broader inability to load `vesteinn/ScandiBERT`." + "reason": "Tokenizer vocab mismatch vs inability to load a different BERT model; not the same underlying issue." }, { "left": "issue:43645", "right": "issue:44164", "accept": false, - "reason": "Notebook/custom-model initialization problems are unrelated to `extra_state` save/load failures." + "reason": "Notebook custom-model initialization vs extra_state save/load handling; different bugs." }, { "left": "issue:43257", "right": "issue:44863", "accept": false, - "reason": "Qwen3 MoE weight conversion under accelerate+deepspeed is unrelated to NemotronH checkpoint loading." + "reason": "Qwen3 MoE weight conversion/loading bug vs NemotronH checkpoint loading bug; unrelated." }, { "left": "issue:43425", "right": "issue:43828", "accept": false, - "reason": "Torch version incompatibility is unrelated to the Phi-tiny-MoE autocast dtype mismatch." + "reason": "Torch version incompatibility vs autocast dtype mismatch on Phi-tiny-MoE; different root causes." }, { "left": "issue:43828", "right": "issue:43957", "accept": false, - "reason": "Autocast dtype mismatch and meta-device loading failures are different code paths and symptoms." + "reason": "Autocast dtype mismatch vs meta-device loading regression; unrelated." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet: the only accepted soft edges are the three `resize_token_embeddings` pairs, and `issue:45292` is indeed the broadest open representative of that cluster. The rejections are conservative and do not overclaim equivalence across clearly different bug classes. No PRs are present in the packet." + "feedback": "The analysis is conservative and grounded in the packet. The rejected pairs mostly rely on broad thematic similarity without clear shared bug evidence, and the accepted `resize_token_embeddings` pairs are plausibly the same fix family with consistent symptoms across Gemma/T5Gemma wrappers. The canonical issue choice is reasonable, and there is no obvious overstatement." }, "evaluator_used": true, "retried": false, @@ -2053,10 +2023,10 @@ "issue:44589|issue:44756", "issue:43906|issue:44530", "issue:43931|issue:44863", - "issue:43611|issue:43645", "issue:39404|issue:44912", - "issue:43425|issue:43716", + "issue:43611|issue:43645", "issue:44530|issue:45084", + "issue:43425|issue:43716", "issue:44265|issue:44898", "issue:43299|issue:43931", "issue:43901|issue:44509", @@ -2064,21 +2034,21 @@ "issue:43994|issue:44079", "issue:43856|issue:44756", "issue:44671|issue:44991", - "issue:40990|issue:42947", "issue:39404|issue:44671", - "issue:43596|issue:44683", + "issue:40990|issue:42947", "issue:44589|issue:45084", + "issue:43596|issue:44683", "issue:43425|issue:44918", "issue:36032|issue:43452", - "issue:43582|issue:43632", "issue:32090|issue:39404", + "issue:43582|issue:43632", "issue:44991|issue:45081", "issue:43957|issue:44164", - "issue:44164|issue:44479", "issue:40990|issue:43421", + "issue:44164|issue:44479", "issue:44488|issue:45081", - "issue:43425|issue:45161", "issue:43883|issue:44488", + "issue:43425|issue:45161", "issue:43611|issue:43644", "issue:44206|issue:44220", "issue:42947|issue:44756", @@ -2086,244 +2056,244 @@ "issue:34567|issue:40444", "issue:43716|issue:45084" ], - "prepared_review_unit_hash": "bbfd16582678dfdae27099a970c828b67b60ec8491bd8d6a4a68c5fd42e92d7d", + "prepared_review_unit_hash": "0c0437ee679256831b7fcec157acaca472f34f79d97a02532c4c818413a98a8e", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 35, - "serialized_chars": 38185, - "estimated_input_tokens": 9547, - "estimated_eval_tokens": 19350 + "serialized_chars": 38180, + "estimated_input_tokens": 9545, + "estimated_eval_tokens": 19346 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is mostly a collection of distinct regression reports across unrelated subsystems. I do not see any pair that is safe to merge as the same concrete bug; the closest audio-fbank pair is still too underspecified to treat as a duplicate with confidence.", - "confidence": 0.52, - "canonical_issue_reason": "Issue 43611 is the best representative anchor because it is a clear, concrete v5 model-loading regression and is broader than the very model-specific reports around it.", + "summary": "This cluster is thematically loose: many items are Transformers v5 / model-loading regressions, but the candidate pairs mostly point to different concrete bugs, model families, or subsystems, so they should not be merged.", + "confidence": 0.38, + "canonical_issue_reason": "issue:43957 is the broadest representative of the recurring v5 model-loading regression pattern in this set, so it works best as the cluster anchor.", "canonical_pr_reason": null, - "best_issue_reason": "43611 is the strongest standalone issue for triage: specific enough to be actionable, but broad enough to represent the larger v5-loading breakage theme in this set.", + "best_issue_reason": "issue:43957 is the most reusable issue to represent the cluster because it is version-scoped, loading-focused, and broad enough to cover the repeated regression theme.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44589", "right": "issue:44756", "accept": false, - "reason": "Different failures: float8 storage lookup vs mmap/OOM mitigation on Apple Silicon." + "reason": "Different failures: missing Float8 storage vs Apple Silicon mmap/OOM; no shared code-path." }, { "left": "issue:43906", "right": "issue:44530", "accept": false, - "reason": "Unrelated domains: isolated reproduction of another issue vs Qwen3.5 PagedAttentionCache crash." + "reason": "Isolated reproduction of one issue vs a Qwen3.5 PagedAttentionCache crash; unrelated symptoms and fixes." }, { "left": "issue:43931", "right": "issue:44863", "accept": false, - "reason": "Both are load failures, but for different model families and different symptoms (shape mismatch vs checkpoint loading support)." - }, - { - "left": "issue:43611", - "right": "issue:43645", - "accept": false, - "reason": "Both involve v5 model loading, but one is base_model_prefix handling and the other is custom-model initialization in notebooks; not the same bug." + "reason": "Both are model-loading problems, but for different model families and different failure modes (shape mismatch vs NemotronH checkpoint loading)." }, { "left": "issue:39404", "right": "issue:44912", "accept": false, - "reason": "Whisper pipeline return_language regression is unrelated to MXFP4 quantization fallback." + "reason": "Whisper pipeline return_language regression is unrelated to GPT-OSS MXFP4 quantization fallback." }, { - "left": "issue:43425", - "right": "issue:43716", + "left": "issue:43611", + "right": "issue:43645", "accept": false, - "reason": "Torch-version compatibility and image preprocessor dtype mismatch are unrelated issues." + "reason": "Both are v5 regressions, but one is base_model_prefix loading and the other is notebook custom-model initialization; not the same bug." }, { "left": "issue:44530", "right": "issue:45084", "accept": false, - "reason": "Different code paths: Qwen3.5 cache/group-type failure vs template compilation error." + "reason": "Qwen3.5 cache group-type crash vs template compilation error; different code paths." + }, + { + "left": "issue:43425", + "right": "issue:43716", + "accept": false, + "reason": "Torch 2.10 compatibility issue is not the same as a Mistral-3 image preprocessor dtype mismatch." }, { "left": "issue:44265", "right": "issue:44898", "accept": false, - "reason": "torch.export/torch_compilable_check failure is unrelated to Perceiver interpolation on non-default image sizes." + "reason": "torch.export / torch_compilable_check failure is unrelated to Perceiver non-default-resolution interpolation behavior." }, { "left": "issue:43299", "right": "issue:43931", "accept": false, - "reason": "Both are Qwen3/VL loading complaints, but one is MoE model loading and the other is checkpoint weight-shape mismatch; not the same concrete bug." + "reason": "Both concern Qwen3-VL loading, but one is MoE model loading in v5 dev0 and the other is a specific weight-shape mismatch; not clearly the same bug." }, { "left": "issue:43901", "right": "issue:44509", "accept": false, - "reason": "Different documentation regressions: return_all_scores vs removed pipeline task docs." + "reason": "Both are docs issues, but they concern different pipeline/task deprecations and different documentation fixes." }, { "left": "issue:44479", "right": "issue:44805", "accept": false, - "reason": "Video-input regression for Qwen models is unrelated to a masking shape mismatch error." + "reason": "Video-input regression for Qwen VL models is unrelated to a generic mask-shape IndexError." }, { "left": "issue:43994", "right": "issue:44079", "accept": false, - "reason": "One is a model/pipeline quality regression, the other is a ModelOutput key-assignment bug; too different." + "reason": "Siglip2 wrong outputs vs ModelOutput key assignment bug; different behavior and likely different fix." }, { "left": "issue:43856", "right": "issue:44756", "accept": false, - "reason": "Qwen3 MoE training memory usage and Strix Halo mmap/OOM are different problems." + "reason": "Qwen3 MoE training memory inefficiency is not the same as disabling mmap on Strix Halo to avoid OOM." }, { "left": "issue:44671", "right": "issue:44991", "accept": false, - "reason": "CamemBERT masked-LM prediction regression is unrelated to est-roberta tokenizer loading." - }, - { - "left": "issue:40990", - "right": "issue:42947", - "accept": false, - "reason": "High perplexity on a model is unrelated to LoRA gradient-checkpointing ineffectiveness." + "reason": "CamemBERT masked-LM regression and est-roberta tokenizer loading failure are separate model-specific issues." }, { "left": "issue:39404", "right": "issue:44671", "accept": false, - "reason": "Whisper return_language and CamemBERT masked-LM prediction issues do not share the same bug." + "reason": "Whisper pipeline behavior and CamemBERT masked-LM predictions are unrelated." }, { - "left": "issue:43596", - "right": "issue:44683", + "left": "issue:40990", + "right": "issue:42947", "accept": false, - "reason": "DeepSpeed ZeRO-3 init index error is unrelated to compiled flex_attention on newer torch." + "reason": "High perplexity on gpt-oss is unrelated to LoRA gradient checkpointing ineffectiveness." }, { "left": "issue:44589", "right": "issue:45084", "accept": false, - "reason": "Float8 storage lookup failure and template compilation error are unrelated." + "reason": "Float8 storage lookup failure and template-node compilation failure are different bugs." + }, + { + "left": "issue:43596", + "right": "issue:44683", + "accept": false, + "reason": "DeepSpeed ZeRO-3 init IndexError is unrelated to compiled flex_attention failing on newer torch." }, { "left": "issue:43425", "right": "issue:44918", "accept": false, - "reason": "Torch 2.10 compatibility and Qwen3.5 embedding unpacking with TRL are different regressions." + "reason": "Torch version incompatibility is not the same as Qwen3.5 embedding unpacking with TRL SFT trainer." }, { "left": "issue:36032", "right": "issue:43452", "accept": false, - "reason": "Tokenizer special-tokens API conflict is unrelated to gguf_file loading breakage." + "reason": "T5 tokenizer special-token conflict and gguf_file loading breakage are different loading paths." }, { - "left": "issue:43582", - "right": "issue:43632", + "left": "issue:32090", + "right": "issue:39404", "accept": false, - "reason": "Apple Silicon warmup TypeError and _is_hf_initialized flag breakage are different code paths." + "reason": "Trainer GPU broadcast NoneType error is unrelated to Whisper return_language pipeline regression." }, { - "left": "issue:32090", - "right": "issue:39404", + "left": "issue:43582", + "right": "issue:43632", "accept": false, - "reason": "Trainer GPU broadcast NoneType error is unrelated to Whisper pipeline return_language behavior." + "reason": "Apple Silicon warmup TypeError and _is_hf_initialized flag regression are separate issues." }, { "left": "issue:44991", "right": "issue:45081", "accept": false, - "reason": "Both affect tokenizer loading, but they hit different models and different failure modes." + "reason": "Both involve tokenizer loading, but they affect different models and different internal failure points." }, { "left": "issue:43957", "right": "issue:44164", "accept": false, - "reason": "Meta-device loading regression is not the same as extra_state save/from_pretrained handling." + "reason": "meta-device loading breakage and save/from_pretrained extra_state handling are distinct serialization/loading bugs." }, { - "left": "issue:44164", - "right": "issue:44479", + "left": "issue:40990", + "right": "issue:43421", "accept": false, - "reason": "Save/load extra_state handling is unrelated to video-input regression." + "reason": "Perplexity regression and runtime special-token post-processor update bug are unrelated." }, { - "left": "issue:40990", - "right": "issue:43421", + "left": "issue:44164", + "right": "issue:44479", "accept": false, - "reason": "Perplexity issue and runtime special-token post-processor update bug are unrelated." + "reason": "extra_state save/load handling is not the same as the video-input regression." }, { "left": "issue:44488", "right": "issue:45081", "accept": false, - "reason": "Different tokenizer-loading failures for different checkpoints." + "reason": "Different tokenizer-loading failures for different models; no evidence of a single underlying bug." }, { - "left": "issue:43425", - "right": "issue:45161", + "left": "issue:43883", + "right": "issue:44488", "accept": false, - "reason": "Torch compatibility does not match the GPT-OSS tensor-parallel loading problem." + "reason": "Molmo tied-weights attribute error and sleng-bert loading failure are unrelated." }, { - "left": "issue:43883", - "right": "issue:44488", + "left": "issue:43425", + "right": "issue:45161", "accept": false, - "reason": "Missing tied-weights metadata and sleng-bert tokenizer loading are unrelated." + "reason": "Torch 2.10 incompatibility is not the same as TP failing for GPT-OSS MoE." }, { "left": "issue:43611", "right": "issue:43644", "accept": false, - "reason": "Both are v5 loading regressions, but one is base_model_prefix and the other is non-persistent buffer initialization; not the same bug." + "reason": "Model-loading with base_model_prefix and non-persistent buffer initialization are different v5 regressions." }, { "left": "issue:44206", "right": "issue:44220", "accept": false, - "reason": "Both mention fbank/audio extraction, but the evidence is too thin to say they are the same concrete regression." + "reason": "Unsupported center arg in feature extractor and _torch_extract_fbank_features issues may be related, but the evidence is not strong enough to treat them as the same bug." }, { "left": "issue:42947", "right": "issue:44756", "accept": false, - "reason": "LoRA gradient checkpointing and Strix Halo mmap/OOM are unrelated." + "reason": "LoRA gradient-checkpointing behavior is unrelated to Strix Halo mmap/OOM." }, { "left": "issue:44488", "right": "issue:44987", "accept": false, - "reason": "Different model-loading failures with no shared concrete code-path bug." + "reason": "Two separate model-loading regressions for different repositories and failure modes." }, { "left": "issue:34567", "right": "issue:40444", "accept": false, - "reason": "TrainerState token counting and multi-image iterable dataset finetuning are unrelated." + "reason": "TrainerState token counter and Qwen2.5-VL iterable multi-image finetuning are unrelated." }, { "left": "issue:43716", "right": "issue:45084", "accept": false, - "reason": "Image-preprocessor dtype mismatch and non-template-node compilation error are unrelated." + "reason": "Mistral-3 image preprocessor dtype mismatch is unrelated to the template compilation error." } ] }, @@ -2349,6 +2319,7 @@ "nodes": [ "issue:36032", "issue:39404", + "issue:40990", "issue:41553", "issue:42915", "issue:43262", @@ -2372,7 +2343,6 @@ "issue:43825", "issue:43883", "issue:43994", - "issue:44062", "issue:44220", "issue:44265", "issue:44403", @@ -2384,6 +2354,7 @@ "issue:44610", "issue:44683", "issue:44756", + "issue:44811", "issue:44849", "issue:44987", "issue:44991", @@ -2400,18 +2371,18 @@ "issue:43883|issue:44991", "issue:43643|issue:44530", "issue:43452|issue:43482", - "issue:44451|issue:44987", - "issue:36032|issue:42915", "issue:43317|issue:43482", - "issue:43541|issue:45084", + "issue:36032|issue:42915", "issue:43262|issue:43377", + "issue:44451|issue:44987", + "issue:43541|issue:45084", "issue:39404|issue:43825", - "issue:42915|issue:44530", "issue:44849|issue:45325", + "issue:42915|issue:44530", "issue:43404|issue:44466", "issue:36032|issue:43645", - "issue:44987|issue:44991", "issue:43646|issue:44534", + "issue:44987|issue:44991", "issue:45084|issue:45507", "issue:43994|issue:44610", "issue:43377|issue:43819", @@ -2420,253 +2391,256 @@ "issue:45216|issue:45357", "issue:44403|issue:45081", "issue:44451|issue:45081", - "issue:43632|issue:44756", - "issue:44220|issue:44683", "issue:44265|issue:44610", + "issue:43632|issue:44756", "issue:41553|issue:43582", + "issue:44220|issue:44683", "issue:43404|issue:45127", "issue:42915|issue:45084", - "issue:44987|issue:45081", - "issue:43296|issue:43531", "issue:44479|issue:45325", - "issue:44062|issue:44987" + "issue:43296|issue:43531", + "issue:44987|issue:45081", + "issue:40990|issue:44811" ], - "prepared_review_unit_hash": "1a8adf42b51a7510d31f03d60bea4b54a5fada270f229535353ca0fe27fdcaa5", + "prepared_review_unit_hash": "0e2623970186c894501b5a18ec079a221a810a75226097db3ab8987ce911138f", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 34, - "serialized_chars": 37326, - "estimated_input_tokens": 9332, - "estimated_eval_tokens": 18920 + "serialized_chars": 37887, + "estimated_input_tokens": 9472, + "estimated_eval_tokens": 19200 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most pairs are clearly different bugs and should stay separate. The only strong duplicate-looking pair is the Qwen3.5 `save_pretrained` regression pair.", - "confidence": 0.76, - "canonical_issue_reason": "issue:43644 is the clearest, most concrete regression in the set and best serves as the representative issue for the v5 buffer/state corruption theme.", + "summary": "The cluster is loosely centered on Transformers v5/v5.0 regression reports, but the soft pairs mostly describe different models, APIs, or failure modes rather than the same underlying bug/change. No PRs are present.", + "confidence": 0.62, + "canonical_issue_reason": "Issue 43646 is the broadest, most central v5 regression report in the set: it targets custom model initialization, a root area that overlaps several other initialization-related complaints.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43644 is the most actionable standalone issue here: it states a specific, user-visible regression cleanly and is broad enough to represent the cluster better than the more model-specific follow-ups.", + "best_issue_reason": "Issue 43646 is the best representative issue for this cluster because it is general, version-specific, and closest to the shared v5 regression theme.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43644", "right": "issue:43646", "accept": false, - "reason": "Different failures: non-persistent buffer corruption vs custom model initialization breakage." + "reason": "Both are v5 initialization regressions, but 43644 is specifically about non-persistent buffers getting junk values, while 43646 is about custom model initialization broadly; not the same concrete bug." }, { "left": "issue:43883", "right": "issue:44991", "accept": false, - "reason": "Unrelated bugs: missing `all_tied_weights_keys` on one model vs tokenizer loading failure for a different model." + "reason": "Different failures in different models: Molmo missing an attribute vs tokenizer loading for EMBEDDIA/est-roberta." }, { "left": "issue:43643", "right": "issue:44530", "accept": false, - "reason": "Different code paths: `AutoConfig.from_pretrained` field loss vs PagedAttentionCache crashing on `linear_attention`." + "reason": "Missing fields from AutoConfig with trust_remote_code is unrelated to a PagedAttentionCache group-type crash in Qwen3.5 models." }, { "left": "issue:43452", "right": "issue:43482", "accept": false, - "reason": "Both involve loading, but one is a generic `gguf_file` API break while the other is a specific Qwen2.5-GGUF load failure; not clearly the same concrete bug." + "reason": "Both mention GGUF/loading, but one is about gguf_file breaking tokenizer/model loading APIs and the other is a Qwen2.5-GGUF v5 loading failure; too different to be the same bug." }, { - "left": "issue:44451", - "right": "issue:44987", + "left": "issue:43317", + "right": "issue:43482", "accept": false, - "reason": "Different models and likely different loading regressions; not the same underlying issue." + "reason": "device_map/offload dequantized-model loading is a different code path from GGUF loading." }, { "left": "issue:36032", "right": "issue:42915", "accept": false, - "reason": "Tokenizer `add_special_tokens` conflict is unrelated to Qwen3Moe FP8 config failure." + "reason": "T5 tokenizer special-tokens conflict and Qwen3MoE FP8 configuration failure are unrelated." }, { - "left": "issue:43317", - "right": "issue:43482", + "left": "issue:43262", + "right": "issue:43377", "accept": false, - "reason": "Dequantized offload/device-map failure is a different path from GGUF loading failure." + "reason": "Audio chat-template sampling-rate defaulting and MIMI batched-vs-single output mismatch are different audio bugs." }, { - "left": "issue:43541", - "right": "issue:45084", + "left": "issue:44451", + "right": "issue:44987", "accept": false, - "reason": "Both mention tracing/compilation, but they are different errors in different code paths and models." + "reason": "Both are loading regressions, but for different models and different failure surfaces; not one underlying issue." }, { - "left": "issue:43262", - "right": "issue:43377", + "left": "issue:43541", + "right": "issue:45084", "accept": false, - "reason": "Audio chat-template sampling-rate default bug is unrelated to MIMI batched-vs-single padding-mask behavior." + "reason": "Grouped_mm tracing failure in Mixtral is unrelated to template-node compilation errors." }, { "left": "issue:39404", "right": "issue:43825", "accept": false, - "reason": "One is a pipeline `return_language` regression; the other is an incorrect error message about translation support." + "reason": "Whisper pipeline return_language behavior and a pipeline error-message regression are not the same defect." }, { - "left": "issue:42915", - "right": "issue:44530", + "left": "issue:44849", + "right": "issue:45325", "accept": false, - "reason": "Different bugs: FineGrainedFP8Config failure vs PagedAttentionCache crash on `linear_attention`." + "reason": "Qwen3.5 hidden-states behavior and Qwen2.5-VL rope-index scaling are different model bugs." }, { - "left": "issue:44849", - "right": "issue:45325", + "left": "issue:42915", + "right": "issue:44530", "accept": false, - "reason": "Different model behaviors: `output_hidden_states=True` regression vs rope-index scaling for still images." + "reason": "FineGrainedFP8Config loading and PagedAttentionCache invalid group type are separate Qwen3-related issues." }, { "left": "issue:43404", "right": "issue:44466", "accept": false, - "reason": "Untied `lm_head` on one model is not the same as device-dependent serialization of tied weights." + "reason": "One is an untied lm_head in Mistral3ForConditionalGeneration; the other is serialization inconsistency for tied weights. Related area, but different bugs." }, { "left": "issue:36032", "right": "issue:43645", "accept": false, - "reason": "Tokenizer special-token conflict is unrelated to custom model initialization breaking in notebooks." + "reason": "Tokenizer-method conflict vs custom model initialization in Jupyter are not the same issue." }, { - "left": "issue:44987", - "right": "issue:44991", + "left": "issue:43646", + "right": "issue:44534", "accept": false, - "reason": "Different failures on different artifacts: model loading regression vs tokenizer loading regression." + "reason": "Both involve v5 initialization semantics, but one is custom model init and the other is non-persistent buffer contents; different concrete bugs." }, { - "left": "issue:43646", - "right": "issue:44534", + "left": "issue:44987", + "right": "issue:44991", "accept": false, - "reason": "Custom model initialization breakage is unrelated to non-persistent buffers being filled with junk." + "reason": "Both are v5 load failures, but they affect different repositories/models and different loading code paths." }, { "left": "issue:45084", "right": "issue:45507", "accept": false, - "reason": "Different errors: template compilation failure vs GraniteMoEHybrid calling an invalid method." + "reason": "Template compilation error and GraniteMoEHybrid invalid method call are unrelated." }, { "left": "issue:43994", "right": "issue:44610", "accept": false, - "reason": "Nonsensical model outputs are not the same as a processor/model input-size mismatch." + "reason": "Nonsensical outputs with SigLIP2 and a processor/model input-size mismatch in OmDet-Turbo are different problems." }, { "left": "issue:43377", "right": "issue:43819", "accept": false, - "reason": "MIMI padding-mask inconsistency is unrelated to DAC `from_latents` mismatch without STE." + "reason": "MIMI encoder padding-mask batching bug and DAC from_latents/STE mismatch are different codec issues." }, { "left": "issue:43645", "right": "issue:44534", "accept": false, - "reason": "Custom model initialization in notebooks is unrelated to non-persistent buffer junk serialization." + "reason": "Jupyter custom-model initialization and non-persistent buffer junk are separate v5 regressions." }, { "left": "issue:43322", "right": "issue:43540", "accept": false, - "reason": "Segmentation fault loading Llava Next is a different bug from Qwen3OmniMoe video-input validation." + "reason": "Llava Next segmentation fault and Qwen3OmniMoe video-input ValueError are different multimodal failures." }, { "left": "issue:45216", "right": "issue:45357", - "accept": true, - "reason": "Same underlying Qwen3.5 `save_pretrained` regression: both report incorrect saved checkpoints/visual encoder keys after v5 changes." + "accept": false, + "reason": "Both concern Qwen3.5 save_pretrained regressions, but one is incorrect checkpoint saving in general and the other is specifically wrong visual encoder keys." }, { "left": "issue:44403", "right": "issue:45081", "accept": false, - "reason": "Unnecessary load-time noise is not the same as a tokenizer crash in `_patch_mistral_regex`." + "reason": "General loading noise and a Mistral regex patch crash are unrelated." }, { "left": "issue:44451", "right": "issue:45081", "accept": false, - "reason": "Different model-loading failure vs Mistral tokenizer regex-patching crash." + "reason": "ScandiBERT loading failure and Mistral tokenizer regex patch failure are different tokenizer problems." }, { - "left": "issue:43632", - "right": "issue:44756", + "left": "issue:44265", + "right": "issue:44610", "accept": false, - "reason": "Broken `_is_hf_initialized` handling is unrelated to disabling mmap on Strix Halo to avoid OOM." + "reason": "torch.export failures with torch_compilable_check are unrelated to OmDet-Turbo processor resolution mismatch." }, { - "left": "issue:44220", - "right": "issue:44683", + "left": "issue:43632", + "right": "issue:44756", "accept": false, - "reason": "Feature extraction failure and compiled flex_attention failure are unrelated." + "reason": "_is_hf_initialized flag breakage and disabling mmap on Strix Halo are unrelated." }, { - "left": "issue:44265", - "right": "issue:44610", + "left": "issue:41553", + "right": "issue:43582", "accept": false, - "reason": "`torch.export`/`torch_compilable_check` failure is unrelated to OmDet-Turbo processor producing the wrong input size." + "reason": "Bad AutoTokenizer error message for Voxtral and AppleSilicon caching_allocator_warmup TypeError are unrelated." }, { - "left": "issue:41553", - "right": "issue:43582", + "left": "issue:44220", + "right": "issue:44683", "accept": false, - "reason": "Bad error messaging for AutoTokenizer loading Voxtral is unrelated to the Apple Silicon `caching_allocator_warmup` TypeError." + "reason": "_torch_extract_fbank_features and compiled flex_attention on torch>=2.9 are different low-level issues." }, { "left": "issue:43404", "right": "issue:45127", "accept": false, - "reason": "Untied `lm_head` on Mistral3 is not the same bug as LoRA merge collapse with extended vocabulary." + "reason": "lm_head tying in Mistral3ForConditionalGeneration and LoRA merge collapse with extended vocab are different weight-tying bugs." }, { - "left": "issue:44987", - "right": "issue:45081", + "left": "issue:42915", + "right": "issue:45084", + "accept": false, + "reason": "Qwen3Moe FP8 config failure and template compilation errors have no shared code-path." + }, + { + "left": "issue:44479", + "right": "issue:45325", "accept": false, - "reason": "Different models and different failures: loading `physical-intelligence/fast` vs Mistral tokenizer regex patch crash." + "reason": "Qwen3.5 video-input regression and Qwen2.5-VL rope-position scaling are different multimodal regressions." }, { "left": "issue:43296", "right": "issue:43531", "accept": false, - "reason": "PaddleOCR-VL load failure and Qwen3-MoE sliding-window issue are separate regressions." + "reason": "PaddleOCR-VL model loading in vLLM and a sliding_window issue in Qwen3-MoE are unrelated." }, { - "left": "issue:44479", - "right": "issue:45325", + "left": "issue:44987", + "right": "issue:45081", "accept": false, - "reason": "Video-input regression in several Qwen models is unrelated to still-image temporal position scaling in Qwen2.5-VL." + "reason": "Loading physical-intelligence/fast under v5 and Mistral tokenizer regex patching are different failures." }, { - "left": "issue:44062", - "right": "issue:44987", + "left": "issue:40990", + "right": "issue:44811", "accept": false, - "reason": "`AddedToken(..., special=...)` constructor conflict is unrelated to the `physical-intelligence/fast` loading regression." + "reason": "High perplexity on gpt-oss-20b and Whisper batch_decode skip_special_tokens behavior are unrelated." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet, and the soft-edge verdicts stay conservative overall. The only accepted duplicate-looking pair, issue:45216 vs issue:45357, is well supported by the titles. The other rejections mostly distinguish issues at the title/code-path level without overclaiming shared root cause." - }, - "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 }, @@ -2687,11 +2661,9 @@ "issue:35707", "issue:36032", "issue:40444", - "issue:40990", "issue:41720", "issue:42175", "issue:42915", - "issue:43296", "issue:43317", "issue:43377", "issue:43421", @@ -2708,17 +2680,20 @@ "issue:43792", "issue:43818", "issue:43856", + "issue:43881", + "issue:44062", "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:44987", "issue:44991", "issue:45020", "issue:45072", @@ -2733,296 +2708,293 @@ ], "soft_pairs": [ "issue:42175|issue:45542", - "issue:40990|issue:44811", - "issue:42915|issue:44291", "issue:43611|issue:44534", "issue:43317|issue:43856", + "issue:44062|issue:44987", + "issue:42915|issue:44291", "issue:36032|issue:43742", - "issue:43632|issue:44534", "issue:40444|issue:45325", - "issue:42915|issue:44589", + "issue:43632|issue:44534", "issue:43531|issue:44877", - "issue:43421|issue:44568", "issue:43818|issue:45357", + "issue:43421|issue:44568", + "issue:42915|issue:44589", "issue:36032|issue:41720", + "issue:44829|issue:45479", "issue:44683|issue:45020", "issue:43482|issue:45538", - "issue:44829|issue:45479", "issue:33666|issue:40444", "issue:43425|issue:44403", - "issue:44403|issue:44991", "issue:44079|issue:44534", - "issue:43792|issue:44589", + "issue:44403|issue:44991", "issue:35707|issue:40444", "issue:43377|issue:45491", + "issue:43792|issue:44589", "issue:43482|issue:43525", - "issue:44291|issue:44589", "issue:43749|issue:44079", + "issue:44291|issue:44589", "issue:43644|issue:44079", "issue:44079|issue:45072", - "issue:43632|issue:45072", "issue:44265|issue:45072", + "issue:43632|issue:45072", "issue:40444|issue:45381", "issue:45325|issue:45381", - "issue:43531|issue:44291", "issue:43611|issue:43646", + "issue:43531|issue:44291", "issue:41720|issue:45081", - "issue:42175|issue:43296" + "issue:43881|issue:44488" ], - "prepared_review_unit_hash": "5037f19fbae2871559be5d330eb59fef7f30d894589e07d9c000d5c797577624", + "prepared_review_unit_hash": "caabd7ea5fb120e6655eb355329b7fe4b3de044aea07858e7d0257a90969a653", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 36, - "serialized_chars": 37918, - "estimated_input_tokens": 9480, - "estimated_eval_tokens": 19216 + "serialized_chars": 38431, + "estimated_input_tokens": 9608, + "estimated_eval_tokens": 19472 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a mixed cluster of unrelated bugs, with one clear duplicate pair around missing TensorFlow backend support in torch-only installs. Most other pairs share broad subsystems or version contexts but not the same concrete defect.", - "confidence": 0.72, - "canonical_issue_reason": "issue:43644 is the clearest representative of the Transformers v5 non-persistent-buffer regression and is the same report as 44534, with earlier discussion and an inbound reference.", + "summary": "This cluster is mostly a set of unrelated Transformers issues/regressions across tokenizers, model loading, vision/video, device mapping, and training; none of the soft pairs look like true duplicates.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue stands out: the items span many different bug classes and model families, so grouping them as one underlying defect would be misleading.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43644 is the strongest anchor in this set: it is a concrete, well-scoped regression report and matches the duplicate report 44534 closely.", + "best_issue_reason": "If forced to pick an umbrella, issue 45020 is the broadest regression report, but it is still too generic and does not actually subsume the other distinct failures.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:42175", "right": "issue:45542", - "accept": true, - "reason": "Both point to the same packaging gap: the torch-only install path leaves the TensorFlow backend unavailable, producing backend/import errors." - }, - { - "left": "issue:40990", - "right": "issue:44811", - "accept": false, - "reason": "Completely different bugs: perplexity evaluation on GPT-OSS vs Whisper batch_decode skipping special tokens." - }, - { - "left": "issue:42915", - "right": "issue:44291", "accept": false, - "reason": "Both involve Transformers v5 loading paths, but one is a Qwen3MoE FP8 config failure and the other is an init_empty_weights/_is_hf_initialized TypeError; different root causes." + "reason": "Different backend/install problems: missing TensorFlow in a torch-only install vs tensorboard being present without TensorFlow." }, { "left": "issue:43611", "right": "issue:44534", "accept": false, - "reason": "Both are v5 regressions, but one is about base_model_prefix loading and the other about non-persistent buffers being filled with junk." + "reason": "Different breakages: base_model_prefix loading failure vs non-persistent buffer junk in v5." }, { "left": "issue:43317", "right": "issue:43856", "accept": false, - "reason": "Different symptoms and code paths: dequantized model offload/device_map loading vs Qwen3 MoE training memory usage." + "reason": "Different load/offload path vs MoE training memory-usage complaint; not the same bug." }, { - "left": "issue:36032", - "right": "issue:43742", + "left": "issue:44062", + "right": "issue:44987", "accept": false, - "reason": "Unrelated tokenizer load error vs a MobileLLM key error when loading a specific model." + "reason": "Tokenizer AddedToken conflict vs failure loading a specific model under v5; unrelated symptoms and code paths." }, { - "left": "issue:43632", - "right": "issue:44534", + "left": "issue:42915", + "right": "issue:44291", + "accept": false, + "reason": "Both mention v5-era type issues, but one is FineGrainedFP8Config and the other is init_empty_weights/_is_hf_initialized; distinct bugs." + }, + { + "left": "issue:36032", + "right": "issue:43742", "accept": false, - "reason": "Both are Transformers v5 breakages, but one concerns the _is_hf_initialized flag and the other non-persistent buffer initialization." + "reason": "T5 tokenizer method-name conflict vs key error loading MobileLLM; different tokenizer/model failures." }, { "left": "issue:40444", "right": "issue:45325", "accept": false, - "reason": "Both are Qwen2.5-VL bugs, but one is multi-image iterable dataset finetuning and the other is still-image temporal position_id scaling." + "reason": "Different Qwen2.5-VL bugs: multi-image IterableDataset failure vs still-image rope-index scaling." }, { - "left": "issue:42915", - "right": "issue:44589", + "left": "issue:43632", + "right": "issue:44534", "accept": false, - "reason": "Different failures: Qwen3MoE FP8 config handling vs missing Float8 storage object." + "reason": "_is_hf_initialized flag regression vs junk-filled non-persistent buffers; not the same underlying defect." }, { "left": "issue:43531", "right": "issue:44877", "accept": false, - "reason": "Different model/config issues: Qwen3-MoE sliding_window behavior vs granite_speech strict config loading." + "reason": "Sliding-window attention issue in Qwen3-MoE vs strict config preventing granite_speech loading." + }, + { + "left": "issue:43818", + "right": "issue:45357", + "accept": false, + "reason": "Video-LLaVA temporal-attention/weight-sharing bug vs Qwen3.5 save_pretrained key mismatch; different model subsystems." }, { "left": "issue:43421", "right": "issue:44568", "accept": false, - "reason": "Both involve special tokens, but one is runtime post-processor updates and the other is add_special_tokens not adding BOS/EOS for a specific tokenizer." + "reason": "Runtime post-processor update bug vs tokenizer add_special_tokens regression; different tokenizer behaviors." }, { - "left": "issue:43818", - "right": "issue:45357", + "left": "issue:42915", + "right": "issue:44589", "accept": false, - "reason": "Different multimodal model problems: Video-LLaVA weight/temporal-attention issue vs Qwen3.5 save_pretrained key regression." + "reason": "Different failure modes: FP8 config handling vs missing Float8 storage object." }, { "left": "issue:36032", "right": "issue:41720", "accept": false, - "reason": "Tokenizer construction error vs Qwen3 auto device mapping CUDA assert; unrelated bugs." + "reason": "Tokenizer class-name conflict vs Qwen3 auto device-map CUDA assert; unrelated." + }, + { + "left": "issue:44829", + "right": "issue:45479", + "accept": false, + "reason": "Different sequence-classification regressions and likely different root causes." }, { "left": "issue:44683", "right": "issue:45020", "accept": false, - "reason": "Different areas: compiled flex_attention on torch>=2.9 vs remote_code model loading regressions." + "reason": "Compiled flex_attention on torch>=2.9 is a specific runtime incompatibility, not the same as the broad remote_code loading regression." }, { "left": "issue:43482", "right": "issue:45538", "accept": false, - "reason": "Different tokenizer/loading issues: Qwen2.5-GGUF on transformers v5 vs CLIPTokenizer model_max_length behavior." - }, - { - "left": "issue:44829", - "right": "issue:45479", - "accept": false, - "reason": "Both are classification-training degeneracy reports, but one is tied to flash_attention_3 and the other to num_labels=1 loss semantics." + "reason": "Qwen2.5-GGUF loading failure vs CLIPTokenizer model_max_length behavior; no overlap in bug surface." }, { "left": "issue:33666", "right": "issue:40444", "accept": false, - "reason": "Both are Qwen2-VL/Qwen2.5-VL training issues, but they describe different failure modes and data setups." + "reason": "General multi-GPU training request vs a concrete multi-image IterableDataset bug; not duplicates." }, { "left": "issue:43425", "right": "issue:44403", "accept": false, - "reason": "Unrelated: Torch 2.10 compatibility vs noisy transformer loading messages." - }, - { - "left": "issue:44403", - "right": "issue:44991", - "accept": false, - "reason": "Different bugs: generic loading noise vs tokenizer loading failure for a specific model." + "reason": "Torch 2.10 compatibility report vs generic loading noise complaint; different issues." }, { "left": "issue:44079", "right": "issue:44534", "accept": false, - "reason": "ModelOutput key assignment bug vs non-persistent buffer junk initialization; not the same defect." + "reason": "ModelOutput key assignment bug vs non-persistent buffer initialization bug; distinct code paths." }, { - "left": "issue:43792", - "right": "issue:44589", + "left": "issue:44403", + "right": "issue:44991", "accept": false, - "reason": "Whisper runtime failure vs missing Float8 storage type; unrelated." + "reason": "Loading noise / UX issue vs tokenizer load failure for EMBEDDIA/est-roberta; unrelated." }, { "left": "issue:35707", "right": "issue:40444", "accept": false, - "reason": "Progressive generation with inputs_embeds/past_key_values is unrelated to Qwen2.5-VL finetuning with multiple images." + "reason": "Progressive generation with inputs_embeds/past_key_values is unrelated to Qwen2.5-VL multi-image dataset failures." }, { "left": "issue:43377", "right": "issue:45491", "accept": false, - "reason": "Both are batching/padding issues, but the affected models and mechanisms differ: MIMI padding-mask support vs Gemma3 sliding-window NaNs." + "reason": "MIMI encoder padding-mask bug vs Gemma3 sliding-window NaN embeddings; different models and mechanics." }, { - "left": "issue:43482", - "right": "issue:43525", + "left": "issue:43792", + "right": "issue:44589", "accept": false, - "reason": "Different loading regressions in unrelated models: Qwen2.5-GGUF vs Llama4Config pad_token_id." + "reason": "Whisper model can't run vs Float8 storage lookup failure; no shared underlying defect." }, { - "left": "issue:44291", - "right": "issue:44589", + "left": "issue:43482", + "right": "issue:43525", "accept": false, - "reason": "init_empty_weights/_is_hf_initialized TypeError is unrelated to the Float8 storage lookup failure." + "reason": "GGUF/v5 loading problem vs Llama4Config missing pad_token_id; unrelated model-loading issues." }, { "left": "issue:43749", "right": "issue:44079", "accept": false, - "reason": "FSDP CPU RAM efficient loading and ModelOutput key assignment are unrelated bugs." + "reason": "FSDP CPU RAM efficient loading regression vs ModelOutput key handling; different subsystems." + }, + { + "left": "issue:44291", + "right": "issue:44589", + "accept": false, + "reason": "Unexpected _is_hf_initialized argument vs missing Float8 storage object; different type errors." }, { "left": "issue:43644", "right": "issue:44079", "accept": false, - "reason": "Although both mention Transformers internals, one is about non-persistent buffers being corrupted and the other about ModelOutput key handling." + "reason": "Non-persistent buffer junk vs ModelOutput key assignment; unrelated regressions." }, { "left": "issue:44079", "right": "issue:45072", "accept": false, - "reason": "ModelOutput key assignment bug vs bf16 dtype mismatches in CI; different defects." + "reason": "ModelOutput key handling vs dtype mismatch in bfloat16 inference; different bugs." }, { - "left": "issue:43632", + "left": "issue:44265", "right": "issue:45072", "accept": false, - "reason": "_is_hf_initialized flag regression is unrelated to dtype mismatch failures in bf16 inference." + "reason": "torch.export.export failure with torch_compilable_check vs inference dtype mismatch; not the same change." }, { - "left": "issue:44265", + "left": "issue:43632", "right": "issue:45072", "accept": false, - "reason": "torch.export / torch_compilable_check failure is unrelated to SwitchTransformers/TimmWrapperModel dtype mismatches." + "reason": "_is_hf_initialized regression vs bfloat16 dtype mismatch in inference; unrelated." }, { "left": "issue:40444", "right": "issue:45381", "accept": false, - "reason": "Both are Qwen2.5-VL vision-input bugs, but one is multi-image finetuning and the other is video vision_position_ids." + "reason": "Multi-image IterableDataset failure vs Qwen2.5-VL video position-id regression; different modalities and symptoms." }, { "left": "issue:45325", "right": "issue:45381", "accept": false, - "reason": "Closely related subsystem, but one is still-image temporal position scaling and the other is video position_id computation; not the same concrete bug." + "reason": "Both are Qwen2.5-VL position-id bugs, but one is still-image temporal scaling and the other is video-input vision_position_ids; not clearly the same concrete defect." }, { - "left": "issue:43531", - "right": "issue:44291", + "left": "issue:43611", + "right": "issue:43646", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior is unrelated to init_empty_weights/_is_hf_initialized argument handling." + "reason": "base_model_prefix loading regression vs custom model initialization regression; different failures." }, { - "left": "issue:43611", - "right": "issue:43646", + "left": "issue:43531", + "right": "issue:44291", "accept": false, - "reason": "Both are Transformers v5 loading regressions, but they describe different failure points: base_model_prefix loading vs custom model initialization." + "reason": "Qwen3-MoE sliding_window issue vs init_empty_weights/_is_hf_initialized TypeError; unrelated." }, { "left": "issue:41720", "right": "issue:45081", "accept": false, - "reason": "Different components: Qwen3 auto device mapping CUDA assert vs Mistral tokenizer regex patch crash." + "reason": "Qwen3 auto device-mapping CUDA assert vs Mistral tokenizer regex patch crash; different code paths." }, { - "left": "issue:42175", - "right": "issue:43296", + "left": "issue:43881", + "right": "issue:44488", "accept": false, - "reason": "Both involve model/backend setup, but one is missing TensorFlow in a torch install and the other is PaddleOCR-VL load failure in vLLM." + "reason": "glm-4v-9b loading failure vs sleng-bert tokenizer loading issue; unrelated models and causes." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "Grounded overall and conservative. The canonical choice around issue:43644/44534 is well supported by the identical titles and the inbound reference on 43644. The only accepted soft edge, issue:42175 vs issue:45542, is a plausible same-failure packaging/backend-installation match. The remaining rejected pairs mostly distinguish different root causes despite shared subsystems or model families, which is appropriately cautious." - }, - "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 }, @@ -3044,6 +3016,7 @@ "issue:39692", "issue:40444", "issue:41720", + "issue:42175", "issue:42915", "issue:43262", "issue:43296", @@ -3075,7 +3048,6 @@ "issue:44291", "issue:44387", "issue:44466", - "issue:44488", "issue:44534", "issue:44610", "issue:44617", @@ -3089,16 +3061,16 @@ "issue:45357" ], "soft_pairs": [ - "issue:43881|issue:44488", "issue:43454|issue:44466", "issue:43994|issue:44898", + "issue:42175|issue:43296", "issue:44387|issue:44617", "issue:43828|issue:44683", "issue:42915|issue:43482", "issue:43526|issue:43883", "issue:43701|issue:45325", - "issue:43792|issue:43957", "issue:43262|issue:43746", + "issue:43792|issue:43957", "issue:43317|issue:43873", "issue:34634|issue:35707", "issue:43606|issue:44265", @@ -3106,239 +3078,260 @@ "issue:44186|issue:44265", "issue:44291|issue:44877", "issue:43606|issue:43819", - "issue:43632|issue:44164", "issue:35707|issue:41720", + "issue:43632|issue:44164", "issue:43377|issue:44898", - "issue:43454|issue:45072", "issue:44610|issue:45325", "issue:43454|issue:43550", + "issue:43454|issue:45072", "issue:43377|issue:44610", "issue:39692|issue:40444", "issue:43749|issue:44534", "issue:44466|issue:45357", "issue:43296|issue:44991", - "issue:43296|issue:44877", - "issue:43792|issue:44683", "issue:43454|issue:43720", + "issue:43296|issue:44877", "issue:41720|issue:43322", + "issue:43792|issue:44683", "issue:41720|issue:44977", "issue:43404|issue:45357" ], - "prepared_review_unit_hash": "7c276a7e3b7a4d8a8b48aef0c934eba5e2d18bb038ebba88d4d194b4ec238d0b", + "prepared_review_unit_hash": "5e19d8b6374bce1f12c581cc4881edf0d1d6e9e9dadbbdab196f1a46770c7a2b", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 34, - "serialized_chars": 37834, - "estimated_input_tokens": 9459, - "estimated_eval_tokens": 19174 + "serialized_chars": 37859, + "estimated_input_tokens": 9465, + "estimated_eval_tokens": 19186 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is a heterogeneous cluster of mostly unrelated issues with only loose similarity around Transformers v5 loading/serialization regressions. I do not see any soft pair that should be merged as a true duplicate; the closest subtheme is tied-weights / model-loading behavior, where issue 43454 is the best representative.", - "confidence": 0.43, - "canonical_issue_reason": "No exact duplicate cluster emerges; issue 43454 is the closest representative because it is the most central concrete bug in the tied-weights/model-loading subtheme.", + "summary": "This cluster is mostly a grab bag of model-loading, quantization, and serialization regressions, not one clean duplicate set. The only strong duplicate-looking pair is the tied lm_head / serialization bug in 43454 and 44466; the rest are related only at a broad subsystem level.", + "confidence": 0.74, + "canonical_issue_reason": "issue:44466 is the most general and technical representative of the clear duplicate pair: it describes the tied-weight serialization problem across devices in v5/main, which subsumes the more model-specific 43454 report.", "canonical_pr_reason": null, - "best_issue_reason": "43454 is the most actionable anchor issue among the set, but the cluster is too mixed to serve as a broad canonical for all items.", + "best_issue_reason": "issue:44466 best represents the actionable core bug in the cluster because it captures the broader tied-weight serialization defect rather than a single model-specific symptom.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43881", - "right": "issue:44488", - "accept": false, - "reason": "Both are load failures, but they involve different models and no shared concrete failure mode." - }, { "left": "issue:43454", "right": "issue:44466", - "accept": false, - "reason": "Both mention lm_head/tied weights, but one is a generation bug and the other is a device-dependent serialization issue." + "accept": true, + "reason": "Same underlying tied lm_head weight handling bug: one reports failed tying causing bad generation, the other reports inconsistent serialization of the same tied weights across devices." }, { "left": "issue:43994", "right": "issue:44898", "accept": false, - "reason": "Different model families and different bugs; they only share a broad vision-model failure theme." + "reason": "Both involve vision models, but they describe different failures on different models and code paths; too broad to treat as the same bug." + }, + { + "left": "issue:42175", + "right": "issue:43296", + "accept": false, + "reason": "Completely unrelated: backend packaging/extras vs an audio processor sampling-rate default bug." }, { "left": "issue:44387", "right": "issue:44617", "accept": false, - "reason": "Both mention OOM, but one is an int4 quantization memory regression and the other is a model-specific CUDA OOM." + "reason": "Both mention memory pressure, but one is a quantization reserved-memory regression and the other is a model-specific CUDA OOM symptom." }, { "left": "issue:43828", "right": "issue:44683", "accept": false, - "reason": "Autocast dtype mismatch and compiled flex_attention failure are distinct code-path problems." + "reason": "Different failures: autocast dtype mismatch versus torch 2.9 compiled flex_attention breakage." }, { "left": "issue:42915", "right": "issue:43482", "accept": false, - "reason": "Different models and different failure classes: FP8 config failure versus GGUF loading failure." + "reason": "Different model families and different problems: fine-grained FP8 config failure vs GGUF loading failure." }, { "left": "issue:43526", "right": "issue:43883", "accept": false, - "reason": "Label-reduction behavior and missing tied-weights metadata are unrelated bugs." + "reason": "Unrelated bugs: image label reduction logic versus a missing attribute during model loading." }, { "left": "issue:43701", "right": "issue:45325", "accept": false, - "reason": "Checkpoint key mismatch is not the same bug as RoPE position-id scaling." + "reason": "Different areas: checkpoint resume key mismatch versus Qwen2.5-VL rope index scaling." }, { - "left": "issue:43792", - "right": "issue:43957", + "left": "issue:43262", + "right": "issue:43746", "accept": false, - "reason": "These are separate loading/runtime failures with different underlying mechanisms." + "reason": "Both are audio-related, but one is a chat-template sampling-rate default and the other is PEFT/local checkpoint loading." }, { - "left": "issue:43262", - "right": "issue:43746", + "left": "issue:43792", + "right": "issue:43957", "accept": false, - "reason": "Audio chat-template sample-rate defaults and PEFT checkpoint loading are unrelated." + "reason": "Both are loading/runtime issues, but the concrete bugs and model paths differ." }, { "left": "issue:43317", "right": "issue:43873", "accept": false, - "reason": "Both involve offloading/quantization, but the specific failures and code paths differ." + "reason": "Both concern offload/quantization, but the titles point to different concrete failures and not one specific shared bug." }, { "left": "issue:34634", "right": "issue:35707", "accept": false, - "reason": "Voice preset handling and progressive generation with past key values are different issues." + "reason": "No clear overlap: BarkProcessor voice_preset versus generation with inputs_embeds/past_key_values." }, { "left": "issue:43606", "right": "issue:44265", "accept": false, - "reason": "CPU offload device mismatch and torch.export/torch_compilable_check are separate bugs." + "reason": "Different code paths: CPU offload device mismatch versus torch.export with torch_compilable_check." }, { "left": "issue:43881", "right": "issue:44991", "accept": false, - "reason": "One is a model load failure, the other a tokenizer-loading failure for a different model." + "reason": "Different failures: a model loading error versus tokenizer loading for a specific repo." }, { "left": "issue:44186", "right": "issue:44265", "accept": false, - "reason": "Tokenizer padding/truncation crashes and export-time compilers checks are unrelated." + "reason": "Tokenizer crash and export failure are unrelated despite both being regressions." }, { "left": "issue:44291", "right": "issue:44877", "accept": false, - "reason": "Init-empty-weights argument handling and strict config validation are different failure modes." + "reason": "Both are loading/config issues, but one is an init_empty_weights argument problem and the other is strict config validation." }, { "left": "issue:43606", "right": "issue:43819", "accept": false, - "reason": "Device mismatch on offload and DAC latent/forward mismatch are unrelated." - }, - { - "left": "issue:43632", - "right": "issue:44164", - "accept": false, - "reason": "Both are v5-related loader/save regressions, but one is `_is_hf_initialized` flag handling and the other is extra_state serialization." + "reason": "Different bugs: device mismatch in offload versus DAC latent conversion mismatch." }, { "left": "issue:35707", "right": "issue:41720", "accept": false, - "reason": "Progressive generation with embeds/past-key-values is unrelated to an auto-device-map CUDA assert." + "reason": "No shared underlying bug; generation API behavior versus device-mapping CUDA assertion." }, { - "left": "issue:43377", - "right": "issue:44898", + "left": "issue:43632", + "right": "issue:44164", "accept": false, - "reason": "Batched-vs-single encoder mismatch and image-resolution mismatch are different bugs." + "reason": "Both touch v5 loading/saving internals, but one is an _is_hf_initialized flag regression and the other is extra_state handling." }, { - "left": "issue:43454", - "right": "issue:45072", + "left": "issue:43377", + "right": "issue:44898", "accept": false, - "reason": "Tied-weights/generation failure and bfloat16 dtype mismatch are different problems." + "reason": "Different models and symptoms: batched-vs-single MIMI outputs versus Perceiver image-size handling." }, { "left": "issue:44610", "right": "issue:45325", "accept": false, - "reason": "Processor image-size mismatch and temporal position-id scaling are not the same bug." + "reason": "Unrelated visual processor size mismatch versus Qwen2.5-VL temporal position-id scaling." }, { "left": "issue:43454", "right": "issue:43550", "accept": false, - "reason": "Tied-weight generation issues are unrelated to the torch.compile SDPA failure." + "reason": "One is tied-weight generation/serialization, the other is a torch.compile SDPA issue." + }, + { + "left": "issue:43454", + "right": "issue:45072", + "accept": false, + "reason": "Different model code paths: tied lm_head weights versus dtype mismatch in SwitchTransformers/TimmWrapperModel." }, { "left": "issue:43377", "right": "issue:44610", "accept": false, - "reason": "MIMI batching/padding-mask behavior and OmDet-Turbo input-size mismatch are unrelated." + "reason": "Different multimodal preprocessing bugs with no concrete shared root cause." }, { "left": "issue:39692", "right": "issue:40444", "accept": false, - "reason": "A documentation example error does not match a finetuning failure on multi-image IterableDataset inputs." + "reason": "Distinct issues: documentation/example errors for SigLIP2 versus Qwen2.5-VL dataset handling." }, { "left": "issue:43749", "right": "issue:44534", "accept": false, - "reason": "FSDP CPU RAM efficient loading and non-persistent buffer initialization are different regressions." + "reason": "Both are v5-related loading regressions, but the specific failure modes are different." }, { "left": "issue:44466", "right": "issue:45357", "accept": false, - "reason": "Both are serialization issues, but one concerns lm_head.weight and the other visual encoder keys in a different model." + "reason": "Both are serialization bugs, but they involve different models and different serialized keys/paths." }, { "left": "issue:43296", "right": "issue:44991", "accept": false, - "reason": "PaddleOCR-VL loading in vLLM and tokenizer loading for EMBEDDIA/est-roberta are unrelated." + "reason": "Different failure types: vLLM model loading versus tokenizer loading regression." + }, + { + "left": "issue:43454", + "right": "issue:43720", + "accept": false, + "reason": "Different bugs: tied-weight handling versus packed-weight unpacking during accelerate loading." + }, + { + "left": "issue:43296", + "right": "issue:44877", + "accept": false, + "reason": "Different models and mechanisms: model loading failure versus strict config rejection." + }, + { + "left": "issue:41720", + "right": "issue:43322", + "accept": false, + "reason": "Both are model-loading crashes, but they are separate models and separate failure modes." }, { "left": "issue:43792", "right": "issue:44683", "accept": false, - "reason": "Whisper load/run failure and compiled flex_attention on torch 2.9 are distinct bugs." + "reason": "Completely unrelated: Whisper runtime failure versus compiled flex_attention breakage." }, { "left": "issue:43404", "right": "issue:45357", "accept": false, - "reason": "Mistral3 tied-weight serialization and Qwen3.5 visual-encoder-key serialization are different model-specific issues." + "reason": "Both involve weight/save behavior, but the model families and concrete bugs differ." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays mostly conservative: it treats the cluster as heterogeneous and only singles out 43454/44466 as the plausible duplicate-like pair. The other soft-edge rejections are appropriately cautious. If anything, the 43454/44466 linkage could be phrased a bit more tentatively (\"likely related\" rather than definitive same bug), but it is still supportable from the titles." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -3375,6 +3368,7 @@ "issue:43572", "issue:43577", "issue:43606", + "issue:43653", "issue:43708", "issue:43720", "issue:43746", @@ -3413,12 +3407,12 @@ "issue:42617|issue:42915", "issue:43550|issue:43720", "issue:44610|issue:44898", + "issue:44186|issue:44610", "issue:44877|issue:44991", "issue:44479|issue:44918", - "issue:44186|issue:44610", "issue:36331|issue:42915", - "issue:36331|issue:42617", "issue:44671|issue:44743", + "issue:36331|issue:42617", "issue:43316|issue:44913", "issue:43550|issue:43606", "issue:42915|issue:43317", @@ -3426,257 +3420,264 @@ "issue:44292|issue:44479", "issue:41720|issue:45161", "issue:43531|issue:43572", - "issue:43540|issue:43782", - "issue:44186|issue:44361", "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:45406|issue:45464", "issue:36010|issue:42617", "issue:36331|issue:41720", "issue:44514|issue:44545", "issue:33357|issue:35707", "issue:43606|issue:43720", - "issue:41720|issue:43746" + "issue:41720|issue:43746", + "issue:43550|issue:43653" ], - "prepared_review_unit_hash": "ad5846cd3f9798fe24a4744026f15f426dd6ea94f419aaf2d150b3298e770742", + "prepared_review_unit_hash": "2ba8ae2280dda0f28363e1e08bba48be5287b285e9c80a1927e06d57ccd7f806", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 34, - "serialized_chars": 37226, - "estimated_input_tokens": 9307, - "estimated_eval_tokens": 18870 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 35, + "serialized_chars": 38116, + "estimated_input_tokens": 9529, + "estimated_eval_tokens": 19314 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a set of unrelated issues; only one pair looks like a true duplicate. The two Qwen2_5_VLProcessor.apply_chat_template reports describe the same batched-input/padding=False crash.", - "confidence": 0.97, - "canonical_issue_reason": "issue:44545 is the best canonical issue: it states the bug clearly, matches the duplicate pair exactly, and has slightly more supporting references.", + "summary": "Mostly a heterogeneous set of unrelated closed issues. The only clear duplicate pair is the Qwen2_5_VLProcessor.apply_chat_template padding=False crash (44514/44545); everything else should stay separate.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 44545 is the cleanest canonical report for the only true duplicate pair: same function, same batched-input/padding=False crash, just a cleaner title than 44514.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44545 is the best overall issue representative for this cluster because it is the only clear duplicate target and has the cleanest title.", + "best_issue_reason": "Issue 44545 is the best representative issue here because it most clearly describes the duplicated apply_chat_template padding=False bug and is the cleaner canonical reference.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44423", "right": "issue:45406", "accept": false, - "reason": "Both are serve crashes, but they affect different processors/models and different AttributeErrors, so they are not the same bug." + "reason": "Both involve serving/multimodal failures, but the errors and affected components differ (`'str' object has no attribute 'to'` vs missing `_tokenizer`)." }, { "left": "issue:44186", "right": "issue:45072", "accept": false, - "reason": "Tokenizer NER/padding crash vs bfloat16 dtype mismatch in a different model path; unrelated underlying causes." + "reason": "Different bugs: tokenizer crashes on NER padding/truncation vs dtype mismatch in bfloat16 inference." }, { "left": "issue:34567", "right": "issue:40990", "accept": false, - "reason": "Trainer token-count tracking and model perplexity are unrelated problems." + "reason": "Unrelated: TrainerState token counting bug vs high perplexity on a specific model/eval setup." }, { "left": "issue:43526", "right": "issue:43881", "accept": false, - "reason": "BEiT label reduction bug and GLM-4V-9B load failure do not share the same code path." + "reason": "Different subsystems and failures: `reduce_labels` in BEiT image processing vs loading failure for glm-4v-9b." }, { "left": "issue:43873", "right": "issue:44756", "accept": false, - "reason": "Both mention memory/offload, but one is quantization offloading behavior and the other is mmap/OOM on Strix Halo; different issues." + "reason": "One is a generic quantization offload question, the other is a Strix Halo mmap/OOM workaround; not the same bug." }, { "left": "issue:43577", "right": "issue:43708", "accept": false, - "reason": "Model dtype loading bug vs Trainer checkpoint step calculation; no overlap in the concrete defect." + "reason": "Different problems: BLIP2 dtype propagation vs Trainer resume/max_steps calculation." }, { "left": "issue:42617", "right": "issue:42915", "accept": false, - "reason": "3d_parallel execution failure and Qwen3-MoE FineGrainedFP8Config failure are separate problems." + "reason": "Both are run/load failures, but `3d_parallel.py` and Qwen3Moe FineGrainedFP8Config point to different code paths." }, { "left": "issue:43550", "right": "issue:43720", "accept": false, - "reason": "torch.compile/SDPA failure in Bamba is unrelated to packed-weight unpacking during accelerate loading in BitNet." + "reason": "Distinct failures: torch.compile + SDPA on Bamba vs unpacking packed weights during accelerate loading for BitNet." }, { "left": "issue:44610", "right": "issue:44898", "accept": false, - "reason": "Different models and different image-size/resolution bugs; not the same underlying change." + "reason": "Both are vision-shape issues, but they affect different models and different mechanisms (processor output size vs interpolation path)." + }, + { + "left": "issue:44186", + "right": "issue:44610", + "accept": false, + "reason": "Tokenizer crash on NER inputs is unrelated to OmDet-Turbo processor/model image-size mismatch." }, { "left": "issue:44877", "right": "issue:44991", "accept": false, - "reason": "Strict config loading and tokenizer loading failures are different bugs in different subsystems." + "reason": "Different config/loading bugs: strict config rejects granite_speech, while tokenizer loading fails for EMBEDDIA/est-roberta." }, { "left": "issue:44479", "right": "issue:44918", "accept": false, - "reason": "Video-input regression for several Qwen VL models vs TRL SFT embedding unpacking; same family name, different failure mode." - }, - { - "left": "issue:44186", - "right": "issue:44610", - "accept": false, - "reason": "Tokenizer crash vs processor output-size mismatch; unrelated defects." + "reason": "Both mention Qwen and input handling, but one is a video-input regression and the other is a trainer embedding unpacking issue." }, { "left": "issue:36331", "right": "issue:42915", "accept": false, - "reason": "CustomTrainer argument mismatch and Qwen3-MoE FP8 loading failure are unrelated." + "reason": "Trainer `compute_loss` signature mismatch vs Qwen3Moe FineGrainedFP8Config loading failure." }, { - "left": "issue:36331", - "right": "issue:42617", + "left": "issue:44671", + "right": "issue:44743", "accept": false, - "reason": "Trainer API breakage is not the same as the 3d_parallel runtime failure." + "reason": "CamemBERT masked-LM prediction issue is unrelated to Qwen3.5 recurrent state reset in modular code." }, { - "left": "issue:44671", - "right": "issue:44743", + "left": "issue:36331", + "right": "issue:42617", "accept": false, - "reason": "CamemBERT masked-LM prediction issue and Qwen3.5 recurrent-state reset bug are unrelated." + "reason": "Different areas entirely: custom trainer API breakage vs `3d_parallel.py` execution failure." }, { "left": "issue:43316", "right": "issue:44913", "accept": false, - "reason": "Both are config-related, but they concern different model configs and different persistence behaviors." + "reason": "Both are config-related, but they concern different models and different API/serialization behavior." }, { "left": "issue:43550", "right": "issue:43606", "accept": false, - "reason": "Bamba compile/SDPA failure and Bark CPU-offload device mismatch are different concrete bugs." + "reason": "Both are runtime/device failures, but one is SDPA/compile and the other is CPU offload device mismatch." }, { "left": "issue:42915", "right": "issue:43317", "accept": false, - "reason": "Qwen3-MoE FP8 loading and dequantized model CPU/GPU offload loading are distinct issues." + "reason": "Different loading/offload issues: Qwen3Moe FineGrainedFP8Config vs dequantized model loading with `device_map=auto`." }, { "left": "issue:41720", "right": "issue:42617", "accept": false, - "reason": "Different models and different failures: auto device mapping assert vs 3d_parallel execution." + "reason": "Qwen3 auto-device-map CUDA assert is unrelated to the `3d_parallel.py` failure." }, { "left": "issue:44292", "right": "issue:44479", "accept": false, - "reason": "NVFP4 runtime failure and Qwen VL video regression are not the same bug." + "reason": "Different regressions: Qwen-3-8B-NVFP4 load/run failure vs video-input regression across Qwen VL variants." }, { "left": "issue:41720", "right": "issue:45161", "accept": false, - "reason": "Both involve parallelism, but the models and failure modes differ; not mergeable as one fix." + "reason": "Both are Qwen/GPT-OSS MoE scaling issues, but one is auto device mapping CUDA assert and the other is tensor-parallel only behavior." }, { "left": "issue:43531", "right": "issue:43572", "accept": false, - "reason": "Sliding-window behavior in Qwen3-MoE and missing pad_token_idx in StableLmConfig are unrelated." + "reason": "Different models and bugs: Qwen3-MoE sliding-window behavior vs missing `pad_token_idx` in StableLmConfig." }, { - "left": "issue:43540", - "right": "issue:43782", + "left": "issue:43388", + "right": "issue:45059", "accept": false, - "reason": "Qwen3OmniMoe video-processing ValueError and Qwen3VL weight_only load error are different code paths." + "reason": "Label gathering in metrics is unrelated to SAM3 text/bounding-box behavior." }, { "left": "issue:44186", "right": "issue:44361", "accept": false, - "reason": "Two tokenizer/task crashes, but on different tokenizers and different failure conditions." + "reason": "Both are tokenizer-related, but they affect different tokenizers and different failure modes." }, { - "left": "issue:43388", - "right": "issue:45059", + "left": "issue:43540", + "right": "issue:43782", "accept": false, - "reason": "Metric label gathering bug and SAM3 text/bbox behavior are not the same underlying defect." + "reason": "Different multimodal/Qwen failures: video input processing vs `weight_only=True` load error." }, { "left": "issue:41720", "right": "issue:43873", "accept": false, - "reason": "Auto device mapping CUDA assert and quantization offloading behavior are separate problems." + "reason": "Specific CUDA assert in Qwen3 auto device mapping is not the same as the broader quantization offload issue." }, { - "left": "issue:45406", - "right": "issue:45464", + "left": "issue:42915", + "right": "issue:43322", "accept": false, - "reason": "Different serve-time failures on different models/processors and different execution paths." + "reason": "Qwen3Moe FP8 config loading and Llava Next segmentation fault are different bugs." }, { - "left": "issue:42915", - "right": "issue:43322", + "left": "issue:45406", + "right": "issue:45464", "accept": false, - "reason": "Qwen3-MoE FP8 loading and Llava Next segfault are unrelated." + "reason": "Both touch serving, but one is a missing `_tokenizer` attribute and the other is a streaming chat/completions failure; different code paths." }, { "left": "issue:36010", "right": "issue:42617", "accept": false, - "reason": "GenerationMixin import error and 3d_parallel runtime failure do not match." + "reason": "Importing `GenerationMixin` and running `3d_parallel.py` are unrelated failures." }, { "left": "issue:36331", "right": "issue:41720", "accept": false, - "reason": "Trainer loss signature mismatch is unrelated to Qwen3 auto device-mapping failure." + "reason": "Trainer `compute_loss` kwarg breakage is unrelated to Qwen3 device-mapping CUDA asserts." }, { "left": "issue:44514", "right": "issue:44545", "accept": true, - "reason": "Same exact Qwen2_5_VLProcessor.apply_chat_template crash on batched input with padding=False; clearly the same bug." + "reason": "Same bug: `Qwen2_5_VLProcessor.apply_chat_template` crashes on batched input when `padding=False`, with matching symptom and scope." }, { "left": "issue:33357", "right": "issue:35707", "accept": false, - "reason": "MacOS CLIP bus error and progressive generation with inputs_embeds/past_key_values are unrelated." + "reason": "MacOS bus error loading a CLIP model 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 and BitNet packed-weight unpacking are different issues." + "reason": "Both involve offload/loading, but the concrete failures are different: device mismatch in CPU offload vs packed-weight unpacking during accelerate loading." }, { "left": "issue:41720", "right": "issue:43746", "accept": false, - "reason": "Qwen3 auto device mapping CUDA assert and GraniteSpeech PEFT checkpoint loading are unrelated." + "reason": "CUDA assert on Qwen3 auto device mapping is unrelated to GraniteSpeech PEFT local checkpoint loading." + }, + { + "left": "issue:43550", + "right": "issue:43653", + "accept": false, + "reason": "Different issues: Bamba torch.compile+SDPA failure vs BigBirdTokenizer mask token registration/decode bug." } ] }, "evaluator_result": { "accept": true, - "feedback": "The cluster summary is grounded in the packet and stays conservative. The only accepted duplicate pair, issue:44514 and issue:44545, is well supported by the matching Qwen2_5_VLProcessor.apply_chat_template batched-input/padding=False crash, and the canonical choice is reasonable given the slightly stronger references on issue:44545. The remaining soft-edge rejections are cautious and consistent with the titles." + "feedback": "Grounded and conservative overall. The only accepted pair, issue:44514 \u2194 issue:44545, is well supported by the packet as the same apply_chat_template batched-input/padding=False crash. The rest of the soft-edge rejections stay appropriately cautious and do not overclaim duplication." }, "evaluator_used": true, "retried": false, @@ -3707,7 +3708,7 @@ "issue:43381", "issue:43421", "issue:43531", - "issue:43550", + "issue:43540", "issue:43596", "issue:43606", "issue:43645", @@ -3718,6 +3719,7 @@ "issue:43873", "issue:43927", "issue:43994", + "issue:44008", "issue:44062", "issue:44077", "issue:44186", @@ -3747,238 +3749,266 @@ ], "soft_pairs": [ "issue:30064|issue:34634", - "issue:43550|issue:43653", "issue:44368|issue:45357", "issue:44492|issue:44509", - "issue:43606|issue:43653", "issue:44625|issue:45479", + "issue:43606|issue:43653", "issue:43421|issue:43927", "issue:44898|issue:45072", "issue:43782|issue:45381", - "issue:34634|issue:41720", "issue:34567|issue:35707", - "issue:44077|issue:44625", - "issue:42886|issue:44843", - "issue:44062|issue:44451", + "issue:34634|issue:41720", "issue:43381|issue:43844", - "issue:44451|issue:44991", + "issue:42886|issue:44843", + "issue:44077|issue:44625", "issue:43421|issue:43653", + "issue:44451|issue:44991", "issue:44918|issue:44936", - "issue:43531|issue:44589", - "issue:44987|issue:45042", + "issue:44062|issue:44451", "issue:43994|issue:44671", + "issue:44987|issue:45042", "issue:39692|issue:44610", + "issue:43531|issue:44589", "issue:43421|issue:44186", - "issue:43596|issue:44220", "issue:45081|issue:45356", - "issue:43531|issue:43645", "issue:45183|issue:45479", "issue:43746|issue:45381", - "issue:39692|issue:43873" + "issue:39692|issue:43873", + "issue:43531|issue:43645", + "issue:43596|issue:44220", + "issue:43746|issue:43873", + "issue:43540|issue:44008" ], - "prepared_review_unit_hash": "2c82192f7daa3034568ab3ca3efe3ba761146cf4373b2846c60473f206c91079", + "prepared_review_unit_hash": "fdc0f11bb5c20218728c0d6d5e0797b0a1db9e9018adcc461a35d61ace1dcbf2", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 29, - "serialized_chars": 36079, - "estimated_input_tokens": 9020, - "estimated_eval_tokens": 18296 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 37008, + "estimated_input_tokens": 9252, + "estimated_eval_tokens": 18760 }, "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/special-token subset. No pair looks like a true duplicate from titles alone, so I would not merge any of the soft edges.", - "confidence": 0.56, - "canonical_issue_reason": "issue:43421 is the most central and actionable issue in the tokenizer/special-token subset: it describes a concrete backend/post-processor update bug with a clear fix path.", + "summary": "This set is mostly unrelated issue reports; only the HF_HUB_OFFLINE tokenizer-loading reports (42886 and 44843) look like the same underlying offline-mode bug. The rest are broad-similarity false positives across different models, components, or failure modes.", + "confidence": 0.86, + "canonical_issue_reason": "issue:44843 is the most actionable canonical issue: it names the concrete root cause (`_patch_mistral_regex` calling `model_info()` unconditionally) and matches the offline-tokenizer failure seen in 42886.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43421 is the best representative issue overall because it is a specific, code-path-level bug rather than a model-specific symptom or documentation problem.", + "best_issue_reason": "issue:44843 is the strongest overall issue candidate because it is specific, reproducible, and root-cause oriented; the paired report 42886 is the broader symptom description.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:30064", "right": "issue:34634", "accept": false, - "reason": "Different components and failures: image processor segmentation maps vs Bark voice_preset handling." - }, - { - "left": "issue:43550", - "right": "issue:43653", - "accept": false, - "reason": "Unrelated bugs: torch.compile/SDPA model execution vs tokenizer special-token registration." + "reason": "Different processors and unrelated bugs: void segmentation maps vs Bark `voice_preset` handling." }, { "left": "issue:44368", "right": "issue:45357", "accept": false, - "reason": "Different problems: Qwen3.5 LoRA tie_word_embeddings warning vs incorrect visual encoder keys on save_pretrained." + "reason": "Both involve Qwen3.5, but one is a LoRA warning and the other is a save/load key regression; not the same bug." }, { "left": "issue:44492", "right": "issue:44509", "accept": false, - "reason": "Both are docs/wording-related, but they are not the same underlying bug or change." + "reason": "Both are docs-related, but one is a cache-strategy typo and the other is stale pipeline-task documentation." }, { - "left": "issue:43606", - "right": "issue:43653", + "left": "issue:44625", + "right": "issue:45479", "accept": false, - "reason": "Device-mismatch offload bug vs BigBirdTokenizer special-token decode bug; different code paths." + "reason": "Both mention `num_labels`, but one is Qwen3.5 config propagation and the other is a generic classification-loss edge case." }, { - "left": "issue:44625", - "right": "issue:45479", + "left": "issue:43606", + "right": "issue:43653", "accept": false, - "reason": "Both involve classification config, but one is propagation of num_labels and the other is a zero-loss problem; not the same issue." + "reason": "Unrelated subsystems: CPU offload/device mismatch vs BigBird tokenizer special-token decoding." }, { "left": "issue:43421", "right": "issue:43927", "accept": false, - "reason": "Both touch token/config state, but one is runtime post-processor updates and the other is save/load losing custom token IDs." + "reason": "Both touch tokenization/special-token behavior, but one is runtime post-processor refresh and the other is save/load ID loss." }, { "left": "issue:44898", "right": "issue:45072", "accept": false, - "reason": "Different failure classes: image resolution/interpolation vs bfloat16 dtype mismatch." + "reason": "Different models and failures: Perceiver resolution handling vs bf16 dtype mismatch." }, { "left": "issue:43782", "right": "issue:45381", "accept": false, - "reason": "Unrelated multimodal loading error vs video-position-id regression." + "reason": "Same broad Qwen VL family, but different concrete paths: `from_pretrained(weight_only=True)` vs video `vision_position_ids`." }, { - "left": "issue:34634", - "right": "issue:41720", + "left": "issue:34567", + "right": "issue:35707", "accept": false, - "reason": "Bark voice_preset bug is unrelated to Qwen3 auto device-mapping CUDA assert." + "reason": "Trainer token-count tracking is unrelated to progressive generation with `inputs_embeds`/`past_key_values`." }, { - "left": "issue:34567", - "right": "issue:35707", + "left": "issue:34634", + "right": "issue:41720", "accept": false, - "reason": "Trainer token-count state bug vs generation with inputs_embeds/past_key_values; different subsystems." + "reason": "Bark voice preset loading and Qwen3 auto device mapping are unrelated failures." }, { - "left": "issue:44077", - "right": "issue:44625", + "left": "issue:43381", + "right": "issue:43844", "accept": false, - "reason": "Patchtsmixer post_init policy issue is unrelated to Qwen3.5 num_labels propagation." + "reason": "Both mention gradients/checkpointing, but one is eval-mode checkpointing misuse and the other is ZeRO-3 gradient abnormality." }, { "left": "issue:42886", "right": "issue:44843", - "accept": false, - "reason": "Both involve tokenizer loading/offline behavior, but one is cache fallback and the other is an unconditional model_info call in a regex patch." + "accept": true, + "reason": "Both are offline tokenizer-loading failures; 44843 identifies the concrete root cause in `_patch_mistral_regex`/`model_info()`." }, { - "left": "issue:44062", - "right": "issue:44451", + "left": "issue:44077", + "right": "issue:44625", "accept": false, - "reason": "AddedToken keyword collision is a tokenizer-construction bug; ScandiBERT load failure is a different model-loading issue." + "reason": "`post_init` validation for patchtsmixer is unrelated to Qwen3.5 `num_labels` propagation." }, { - "left": "issue:43381", - "right": "issue:43844", + "left": "issue:43421", + "right": "issue:43653", "accept": false, - "reason": "Gradient checkpointing eval-mode limitation vs ZeRO-3 gradient growth; different training-path bugs." + "reason": "Special-token handling is too broad here; one is tokenizer backend post-processing, the other is a BigBird mask-token decode bug." }, { "left": "issue:44451", "right": "issue:44991", "accept": false, - "reason": "Both are load failures, but they concern different models and different root causes." + "reason": "Both are tokenizer/model loading failures, but the affected models and likely root causes differ." }, { - "left": "issue:43421", - "right": "issue:43653", + "left": "issue:44918", + "right": "issue:44936", "accept": false, - "reason": "Related tokenizer/special-token area, but not the same concrete bug: runtime post-processor refresh vs missing special-token registration." + "reason": "Different trainer bugs: SFT unpacking of embeddings vs evaluate-after-train failure." }, { - "left": "issue:44918", - "right": "issue:44936", + "left": "issue:44062", + "right": "issue:44451", "accept": false, - "reason": "Different trainer failures: SFT embedding unpacking vs evaluate-after-train behavior." + "reason": "AddedToken argument duplication is unrelated to ScandiBERT loading failure." }, { - "left": "issue:43531", - "right": "issue:44589", + "left": "issue:43994", + "right": "issue:44671", "accept": false, - "reason": "Sliding-window Qwen3-MoE bug is unrelated to Float8 storage object lookup." + "reason": "Both are incorrect model outputs, but for different models and different mechanisms." }, { "left": "issue:44987", "right": "issue:45042", "accept": false, - "reason": "Model-loading regression vs image processor torchvision dependency regression; different code paths." + "reason": "A model-loading regression is unrelated to PIL image processors wrongly requiring torchvision." }, { - "left": "issue:43994", - "right": "issue:44671", + "left": "issue:39692", + "right": "issue:44610", "accept": false, - "reason": "Both are incorrect model outputs, but they are different models and different symptoms." + "reason": "Both are processor/model mismatch style reports, but they concern different models and different failure modes." }, { - "left": "issue:39692", - "right": "issue:44610", + "left": "issue:43531", + "right": "issue:44589", "accept": false, - "reason": "Both mention processor/model mismatches, but one is a docs example error and the other is an OmDet-Turbo resolution mismatch." + "reason": "Qwen3-MoE sliding-window behavior is unrelated to the Float8 storage lookup error." }, { "left": "issue:43421", "right": "issue:44186", "accept": false, - "reason": "Tokenizer-related, but one is special-token post-processing and the other is NER/batched padding truncation crash." - }, - { - "left": "issue:43596", - "right": "issue:44220", - "accept": false, - "reason": "IndexError under ZeRO-3 init is unrelated to the fbank feature extraction issue." + "reason": "TokenizersBackend special-token refresh is unrelated to LayoutLMv2 NER/padding crash." }, { "left": "issue:45081", "right": "issue:45356", "accept": false, - "reason": "Both mention fix_mistral_regex, but one is an AttributeError crash and the other is a different tokenizer regression/warning; not clearly the same concrete bug." + "reason": "Both involve tokenizer regex patching, but one is a crash in `_patch_mistral_regex` and the other is a Kimi-K2.5 regression with codec/warning issues." }, { "left": "issue:45183", "right": "issue:45479", "accept": false, - "reason": "Validation error messaging bug vs zero-loss classification bug; different underlying causes." + "reason": "Validation-error messaging across models is not the same as the `num_labels=1` zero-loss classification bug." }, { "left": "issue:43746", "right": "issue:45381", "accept": false, - "reason": "PEFT local checkpoint loading bug is unrelated to Qwen2.5-VL video input position IDs." + "reason": "PEFT local checkpoint loading and Qwen2.5-VL video position IDs are unrelated." }, { "left": "issue:39692", "right": "issue:43873", "accept": false, - "reason": "SigLIP2 docs/model mismatch and quantization offloading problems are different issues." + "reason": "Documentation/example problems are unrelated to quantization/offloading behavior." + }, + { + "left": "issue:43531", + "right": "issue:43645", + "accept": false, + "reason": "Sliding-window Qwen3-MoE behavior is unrelated to Jupyter custom-model initialization." + }, + { + "left": "issue:43596", + "right": "issue:44220", + "accept": false, + "reason": "BertModel ZeRO-3 init crash and `_torch_extract_fbank_features()` are unrelated." + }, + { + "left": "issue:45081", + "right": "issue:45356", + "accept": false, + "reason": "Shared tokenizer-patch area is not enough; the reported failure modes do not match closely enough to merge." + }, + { + "left": "issue:45183", + "right": "issue:45479", + "accept": false, + "reason": "Different bugs in input validation vs classification loss behavior." + }, + { + "left": "issue:43746", + "right": "issue:43873", + "accept": false, + "reason": "PEFT checkpoint loading and quantization/offloading are separate concerns." + }, + { + "left": "issue:43540", + "right": "issue:44008", + "accept": false, + "reason": "Video processing ValueError in Qwen3OmniMoe is unrelated to Gemma 3n variable-name collision in forward()." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The analysis is grounded in the packet and stays conservative overall. The only clearly supported same-bug pair is 42886/44843, which both describe HF_HUB_OFFLINE tokenizer-loading failures and the latter gives a concrete root cause. The remaining soft-edge rejections are plausible from the titles and do not overclaim mergeability." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -4008,6 +4038,7 @@ "issue:43618", "issue:43643", "issue:43644", + "issue:43650", "issue:43653", "issue:43697", "issue:43716", @@ -4020,7 +4051,7 @@ "issue:43819", "issue:43825", "issue:43827", - "issue:43873", + "issue:43906", "issue:43927", "issue:43994", "issue:44008", @@ -4044,245 +4075,238 @@ "issue:45381" ], "soft_pairs": [ - "issue:43540|issue:44008", - "issue:43746|issue:43873", "issue:40444|issue:43746", "issue:43618|issue:43697", - "issue:44220|issue:44661", "issue:43716|issue:44008", "issue:43720|issue:43819", - "issue:44843|issue:45081", + "issue:44220|issue:44661", "issue:43653|issue:43720", "issue:43335|issue:43756", - "issue:44062|issue:44488", "issue:43421|issue:44811", "issue:44186|issue:44898", - "issue:30064|issue:36010", + "issue:44843|issue:45081", "issue:43550|issue:45072", + "issue:43550|issue:43927", + "issue:30064|issue:36010", "issue:43540|issue:45381", "issue:43577|issue:43818", - "issue:43550|issue:43927", + "issue:44062|issue:44488", "issue:43643|issue:44877", - "issue:43577|issue:43582", "issue:43531|issue:43643", + "issue:43577|issue:43582", "issue:43782|issue:44977", "issue:43644|issue:43749", - "issue:44448|issue:44671", "issue:43653|issue:43819", + "issue:44448|issue:44671", "issue:44561|issue:45020", "issue:43825|issue:43994", "issue:43550|issue:43819", "issue:44186|issue:44442", - "issue:43825|issue:43827" + "issue:43825|issue:43827", + "issue:43650|issue:43906" ], - "prepared_review_unit_hash": "38fd18ffe6a92ad3f54bcb445cdaddbf0df954c5e8a3836077129f979c5a7b47", + "prepared_review_unit_hash": "9ebbe33777049c14fd1fdc861c20bdc60e31fbe20cc51ec63b29caabbdf9f259", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 30, - "serialized_chars": 36314, - "estimated_input_tokens": 9079, - "estimated_eval_tokens": 18414 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36561, + "estimated_input_tokens": 9141, + "estimated_eval_tokens": 18538 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is very heterogeneous; most items are unrelated model-specific bugs. No soft pair looks like the same concrete underlying bug/change, so all soft edges should be rejected. The broadest and most representative issue is the open remote_code regression report.", - "confidence": 0.49, - "canonical_issue_reason": "issue:45020 is the broadest, most central report in the set: it describes recent-version breakage across models using `remote_code`, which matches the general loading/regression theme better than the narrower single-model reports.", + "summary": "This cluster is mostly a set of unrelated issues that only share broad Transformers topics (tokenizers, multimodal models, config/loading, v5 regressions). No soft pair looks like the same underlying bug or change, so all candidate links should be rejected.", + "confidence": 0.95, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:45020 is the best global representative because it is current, open, and describes an umbrella regression affecting multiple remote-code models rather than one narrow model-specific failure.", + "best_issue_reason": "No single issue is a good canonical representative because the cluster is heterogeneous; if forced, the most triage-worthy/general recent regression is issue:45020, but it is not a true cluster anchor.", "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 Gemma3n variable-name collision in forward()." - }, - { - "left": "issue:43746", - "right": "issue:43873", - "accept": false, - "reason": "Both touch loading/offloading, but one is PEFT local checkpoint loading and the other is quantization offloading behavior; not the same bug." - }, { "left": "issue:40444", "right": "issue:43746", "accept": false, - "reason": "IterableDataset multi-image finetuning failure is unrelated to GraniteSpeech PEFT checkpoint loading." + "reason": "Different problems: Qwen2.5-VL iterable multi-image finetuning failure vs GraniteSpeech PEFT/local checkpoint loading." }, { "left": "issue:43618", "right": "issue:43697", "accept": false, - "reason": "CLIP output attentions assignment and RTDetr output drift are different code paths and symptoms." - }, - { - "left": "issue:44220", - "right": "issue:44661", - "accept": false, - "reason": "Audio fbank extraction issue is unrelated to tokenizer mapping / add-new-model-like failure." + "reason": "CLIPOutput attentions regression and RTDetrV2 output mismatch are unrelated model-specific issues." }, { "left": "issue:43716", "right": "issue:44008", "accept": false, - "reason": "Different modality bugs: image-preprocessor dtype mismatch vs audio/video tensor name collision." + "reason": "Both are dtype-related, but one is Mistral image preprocessor mismatch and the other is a Gemma3n variable-name collision causing AttributeError." }, { "left": "issue:43720", "right": "issue:43819", "accept": false, - "reason": "BitNet packed-weight loading and DAC STE mismatch are unrelated implementation bugs." + "reason": "BitNet packed-weight loading bug and DAC forward/from_latents mismatch are different code paths and failures." }, { - "left": "issue:44843", - "right": "issue:45081", + "left": "issue:44220", + "right": "issue:44661", "accept": false, - "reason": "Both involve `_patch_mistral_regex`, but one is offline `model_info()` access and the other is a backend_tokenizer attribute crash; distinct failure modes." + "reason": "Audio fbank extraction bug vs tokenizer mapping/add-new-model-like failure; no shared underlying defect." }, { "left": "issue:43653", "right": "issue:43720", "accept": false, - "reason": "BigBirdTokenizer special-token decode bug is unrelated to BitNet accelerate loading." + "reason": "BigBirdTokenizer special-token registration bug is unrelated to BitNet packed-weight loading." }, { "left": "issue:43335", "right": "issue:43756", "accept": false, - "reason": "SwitchTransformers sparse-layer construction and Smollm3 RoPE-layer dropping are different model-specific configuration bugs." - }, - { - "left": "issue:44062", - "right": "issue:44488", - "accept": false, - "reason": "AddedToken constructor conflict is not the same as a model failing to load a specific checkpoint." + "reason": "SwitchTransformers sparse-layer config bug is not the same as Smollm3 RoPE-layer dropping." }, { "left": "issue:43421", "right": "issue:44811", "accept": false, - "reason": "Runtime post-processor update for special tokens is different from Whisper batch_decode skipping special tokens." + "reason": "Runtime post-processor update bug vs Whisper batch_decode skip_special_tokens behavior; different tokenizer subsystems." }, { "left": "issue:44186", "right": "issue:44898", "accept": false, - "reason": "LayoutLMv2 tokenizer padding/NER crash is unrelated to Perceiver non-default resolution inference." + "reason": "LayoutLMv2 tokenizer padding/NER crash is unrelated to Perceiver image-classification resolution handling." }, { - "left": "issue:30064", - "right": "issue:36010", + "left": "issue:44843", + "right": "issue:45081", "accept": false, - "reason": "Image processor segmentation-map handling is unrelated to GenerationMixin import failure." + "reason": "HF Hub offline/model_info call in tokenizer patch is distinct from a tokenizer.backend_tokenizer AttributeError." }, { "left": "issue:43550", "right": "issue:45072", "accept": false, - "reason": "Bamba torch.compile+SDPA failure and dtype mismatches in other models are different regressions." + "reason": "Both mention dtype/SDPA, but one is a torch.compile+SDPA model bug and the other is bfloat16 inference dtype mismatches across models." + }, + { + "left": "issue:43550", + "right": "issue:43927", + "accept": false, + "reason": "Bamba torch.compile/SDPA failure is unrelated to DiaConfig save/load losing custom token IDs." + }, + { + "left": "issue:30064", + "right": "issue:36010", + "accept": false, + "reason": "ImageProcessor void segmentation maps and GenerationMixin import error are unrelated." }, { "left": "issue:43540", "right": "issue:45381", "accept": false, - "reason": "Qwen3OmniMoe video-input failure is unrelated to Qwen2.5-VL video position id issues." + "reason": "Qwen3OmniMoe video-input ValueError and Qwen2.5-VL vision_position_ids issue are different multimodal bugs." }, { "left": "issue:43577", "right": "issue:43818", "accept": false, - "reason": "BLIP2 dtype persistence and Video-LLaVA temporal-attention/weight sharing are unrelated." + "reason": "BLIP2 dtype loading issue is unrelated to Video-LLaVA missing temporal attention / weight sharing." }, { - "left": "issue:43550", - "right": "issue:43927", + "left": "issue:44062", + "right": "issue:44488", "accept": false, - "reason": "torch.compile+SDPA failure is unrelated to DiaConfig save/load token-id loss." + "reason": "AddedToken special-argument TypeError and model loading failure for cjvt/sleng-bert are unrelated tokenizer/model issues." }, { "left": "issue:43643", "right": "issue:44877", "accept": false, - "reason": "Both involve config loading, but one is missing fields from `trust_remote_code` and the other is strict config blocking granite_speech; not the same defect." + "reason": "AutoConfig trust_remote_code missing fields is a different strict-config problem than granite_speech config loading." }, { - "left": "issue:43577", - "right": "issue:43582", + "left": "issue:43531", + "right": "issue:43643", "accept": false, - "reason": "BLIP2 dtype mismatch and Apple Silicon caching allocator TypeError are unrelated." + "reason": "Qwen3-MoE sliding_window behavior is unrelated to AutoConfig trust_remote_code field loss." }, { - "left": "issue:43531", - "right": "issue:43643", + "left": "issue:43577", + "right": "issue:43582", "accept": false, - "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to missing fields in AutoConfig with `trust_remote_code`." + "reason": "Both are dtype-related, but one is BLIP2 loading dtype and the other is an Apple Silicon caching_allocator_warmup TypeError." }, { "left": "issue:43782", "right": "issue:44977", "accept": false, - "reason": "Qwen3VL weight_only loading error is different from Qwen3.5 flash-attention generation behavior." + "reason": "Qwen3VL weight_only loading error is distinct from Qwen3.5 flash-attention generation failure." }, { "left": "issue:43644", "right": "issue:43749", "accept": false, - "reason": "Non-persistent buffer junk in v5 and FSDP CPU RAM efficient loading are different regressions." + "reason": "Non-persistent buffer junk in v5 is unrelated to FSDP_CPU_RAM_EFFICIENT_LOADING being broken." }, { - "left": "issue:44448", - "right": "issue:44671", + "left": "issue:43653", + "right": "issue:43819", "accept": false, - "reason": "Both are output regressions in v5, but they affect different models and code paths." + "reason": "BigBirdTokenizer mask token registration bug is unrelated to DAC latent/forward mismatch." }, { - "left": "issue:43653", - "right": "issue:43819", + "left": "issue:44448", + "right": "issue:44671", "accept": false, - "reason": "Tokenizer special-token registration bug is unrelated to DAC latent reconstruction mismatch." + "reason": "Pegasus v4/v5 output differences and CamemBERT masked LM prediction changes are separate model regressions." }, { "left": "issue:44561", "right": "issue:45020", "accept": false, - "reason": "44561 is one specific remote_code breakage cause, while 45020 is a broad umbrella report; not the same concrete bug." + "reason": "Both involve recent v5/remote_code breakage, but one is is_torch_fx_available removal and the other is broader remote_code failures." }, { "left": "issue:43825", "right": "issue:43994", "accept": false, - "reason": "Pipeline error-message wording is unrelated to SigLIP2 nonsensical outputs." + "reason": "pipeline() translation messaging bug and SigLIP2 nonsensical outputs are not the same defect." }, { "left": "issue:43550", "right": "issue:43819", "accept": false, - "reason": "SDPA/torch.compile failure and DAC from_latents mismatch are unrelated model implementations." + "reason": "Bamba torch.compile+SDPA and DAC forward/from_latents mismatch are unrelated model behaviors." }, { "left": "issue:44186", "right": "issue:44442", "accept": false, - "reason": "LayoutLMv2 tokenizer crashes and FastSpeech2Conformer tokenizer loading failure are different tokenizer bugs." + "reason": "LayoutLMv2 tokenizer crash and AutoTokenizer failing to load FastSpeech2ConformerTokenizer are unrelated tokenizer paths." }, { "left": "issue:43825", "right": "issue:43827", "accept": false, - "reason": "One is an incorrect pipeline error message; the other is stale docs referencing removed pipelines. Related area, but not the same bug." + "reason": "Both concern translation/pipeline v5 removals, but one is an error message mismatch and the other is stale docs; not the same bug/change." + }, + { + "left": "issue:43650", + "right": "issue:43906", + "accept": false, + "reason": "Placeholder issue \"ADD THE DATA\" is unrelated to the isolated reproduction of issue 38071." } ] }, @@ -4322,7 +4346,6 @@ "issue:43582", "issue:43606", "issue:43638", - "issue:43650", "issue:43653", "issue:43697", "issue:43720", @@ -4330,7 +4353,7 @@ "issue:43749", "issue:43782", "issue:43819", - "issue:43906", + "issue:43854", "issue:44060", "issue:44291", "issue:44368", @@ -4346,6 +4369,7 @@ "issue:44857", "issue:44912", "issue:44936", + "issue:44977", "issue:44991", "issue:45072", "issue:45081", @@ -4357,225 +4381,237 @@ ], "soft_pairs": [ "issue:44821|issue:45084", - "issue:43650|issue:43906", "issue:41720|issue:44991", - "issue:44805|issue:44936", - "issue:45081|issue:45406", "issue:43606|issue:45072", + "issue:44805|issue:44936", "issue:30064|issue:36331", "issue:42175|issue:45081", "issue:44610|issue:45072", "issue:43782|issue:45325", "issue:44485|issue:44509", + "issue:44060|issue:44368", "issue:34634|issue:36010", "issue:43720|issue:45072", - "issue:44060|issue:44368", "issue:43262|issue:43819", - "issue:43582|issue:44857", + "issue:45081|issue:45406", "issue:43749|issue:44291", + "issue:43582|issue:44857", "issue:43653|issue:44671", "issue:43526|issue:43697", "issue:43526|issue:44554", - "issue:45084|issue:45538", "issue:43653|issue:44912", - "issue:43454|issue:43577", + "issue:45084|issue:45538", "issue:44568|issue:45488", + "issue:43454|issue:43577", "issue:34634|issue:36331", - "issue:43452|issue:44991", - "issue:43638|issue:44936", + "issue:43746|issue:44811", "issue:43257|issue:43329", - "issue:43746|issue:44811" + "issue:43638|issue:44936", + "issue:43452|issue:44991", + "issue:43854|issue:44977" ], - "prepared_review_unit_hash": "2fa2ced6f8b03f58f5db9f2131400b92a7e881cdd9be57f2f98359e9e132ded2", + "prepared_review_unit_hash": "3881be57a9c574d953353a41981739d5f105a7073ce506d0305c8dc8a5ccb619", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36899, - "estimated_input_tokens": 9225, - "estimated_eval_tokens": 18706 + "serialized_chars": 36933, + "estimated_input_tokens": 9234, + "estimated_eval_tokens": 18724 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are a heterogeneous mix of unrelated bug reports; none of the soft-similarity pairs look like true duplicates or the same fixable code-path change. I would not merge any pair.", - "confidence": 0.97, - "canonical_issue_reason": "No single true canonical duplicate stands out because the cluster is heterogeneous. If one issue must be chosen as the representative anchor, issue #45072 is a concrete, well-scoped loading/inference bug report rather than a vague or documentation-only report.", + "summary": "This cluster is heterogeneous: the items span unrelated bugs in tokenizers, multimodal processors, loading paths, training code, and model-specific runtime failures. None of the soft-pair candidates look like the same underlying defect or change.", + "confidence": 0.95, + "canonical_issue_reason": "No single canonical issue fits this set; the reports are about different models and code paths rather than one shared bug.", "canonical_pr_reason": null, - "best_issue_reason": "Issue #45072 is the best representative issue in this set because it is a specific, reproducible runtime bug with a clear failure mode and actionable description; several other items are similarly about model loading/inference, but none are close enough to be duplicates.", + "best_issue_reason": "No issue is a clear cluster representative because there is no common underlying problem to normalize around.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44821", "right": "issue:45084", "accept": false, - "reason": "Different failures: loading AutoImageProcessor from URL versus a Jinja/template compilation TypeError. No shared code-path." - }, - { - "left": "issue:43650", - "right": "issue:43906", - "accept": false, - "reason": "Unrelated titles and problems; one is a filler issue, the other is a reproduced bug report." + "reason": "Different failures: URL image processor loading vs a template compilation TypeError." }, { "left": "issue:41720", "right": "issue:44991", "accept": false, - "reason": "Different models and subsystems: Qwen3 device-mapping CUDA assert versus tokenizer loading for EMBEDDIA/est-roberta." + "reason": "Unrelated model/runtime bug vs tokenizer-loading regression." }, { - "left": "issue:44805", - "right": "issue:44936", - "accept": false, - "reason": "Mask/index shape error in model forward path versus trainer evaluate failing after train; not the same bug." - }, - { - "left": "issue:45081", - "right": "issue:45406", + "left": "issue:43606", + "right": "issue:45072", "accept": false, - "reason": "Both are loading-time attribute errors, but they hit different objects and code paths (_patch_mistral_regex/tokenizer backend vs Gemma4Processor _tokenizer)." + "reason": "Both are inference mismatches, but they affect different subsystems and models; not the same bug." }, { - "left": "issue:43606", - "right": "issue:45072", + "left": "issue:44805", + "right": "issue:44936", "accept": false, - "reason": "Both involve inference/loading mismatches, but the concrete failures are different: CPU offload device mismatch versus bfloat16 dtype mismatches." + "reason": "Mask shape IndexError during model use vs trainer lifecycle failure after training." }, { "left": "issue:30064", "right": "issue:36331", "accept": false, - "reason": "Different components and symptoms: image processor segmentation maps versus custom trainer loss signature." + "reason": "Image segmentation-map processing bug is unrelated to CustomTrainer loss-signature incompatibility." }, { "left": "issue:42175", "right": "issue:45081", "accept": false, - "reason": "Package install/backend dependency issue versus tokenizer regex patch crash; unrelated." + "reason": "Packaging/backend selection issue vs Mistral tokenizer regex patch crash." }, { "left": "issue:44610", "right": "issue:45072", "accept": false, - "reason": "Processor output size mismatch versus dtype mismatch in inference; not the same underlying bug." + "reason": "Processor image-size mismatch vs dtype mismatch; different code paths." }, { "left": "issue:43782", "right": "issue:45325", "accept": false, - "reason": "Different Qwen VL bugs: weight_only load error versus RoPE position-id scaling regression." + "reason": "Qwen3VL load error is unrelated to Qwen2.5-VL rope-position scaling." }, { "left": "issue:44485", "right": "issue:44509", "accept": false, - "reason": "One is a model implementation discussion, the other is a docs cleanup issue; no duplicate relation." + "reason": "RoPE implementation discussion is unrelated to docs mentioning removed pipeline tasks." + }, + { + "left": "issue:44060", + "right": "issue:44368", + "accept": false, + "reason": "Both mention tied-weights warnings, but one is a concrete wrong tie mapping and the other is a fine-tuning warning behavior; not clearly the same defect." }, { "left": "issue:34634", "right": "issue:36010", "accept": false, - "reason": "Different processors and failures: Bark voice_preset versus GenerationMixin import error." + "reason": "Voice preset handling in BarkProcessor is unrelated to GenerationMixin import failure." }, { "left": "issue:43720", "right": "issue:45072", "accept": false, - "reason": "Accelerate loading of packed weights versus bfloat16 inference dtype mismatch; related theme but different concrete bugs." - }, - { - "left": "issue:44060", - "right": "issue:44368", - "accept": false, - "reason": "Both mention tied-weights warnings, but one reports an incorrect tie to a bias tensor while the other reports a warning emitted during LoRA fine-tuning." + "reason": "Packed-weight unpacking bug is unrelated to bfloat16 dtype mismatches." }, { "left": "issue:43262", "right": "issue:43819", "accept": false, - "reason": "Audio chat-template sampling-rate default bug versus DAC latent/forward mismatch; unrelated." + "reason": "Audio chat-template sampling-rate default bug is unrelated to DAC STE mismatch." }, { - "left": "issue:43582", - "right": "issue:44857", + "left": "issue:45081", + "right": "issue:45406", "accept": false, - "reason": "Apple Silicon TypeError in allocator warmup versus CUDA AMP crash in a loss function; different platforms and code paths." + "reason": "Mistral tokenizer backend access crash is unrelated to Gemma4Processor missing _tokenizer." }, { "left": "issue:43749", "right": "issue:44291", "accept": false, - "reason": "Both are loading-related, but FSDP CPU RAM efficient loading and init_empty_weights unexpected argument are distinct compatibility issues." + "reason": "Different loading failures: FSDP RAM-efficient loading vs init_empty_weights extra-argument handling." + }, + { + "left": "issue:43582", + "right": "issue:44857", + "accept": false, + "reason": "Apple Silicon warmup TypeError is unrelated to CUDA float16 AMP crash." }, { "left": "issue:43653", "right": "issue:44671", "accept": false, - "reason": "Tokenizer special-token registration bug versus masked LM prediction regression; not the same issue." + "reason": "BigBirdTokenizer special-token registration bug is not the same as CamemBERT MLM prediction regression." }, { "left": "issue:43526", "right": "issue:43697", "accept": false, - "reason": "Processor label reduction bug versus RTDetr output divergence in v5; different layers of the stack." + "reason": "BeitImageProcessorFast label reduction bug and RTDetrV2 output differences are distinct issues." }, { "left": "issue:43526", "right": "issue:44554", "accept": false, - "reason": "Vision processor label handling versus MPS attention correctness issue; unrelated." + "reason": "Image label reduction bug is unrelated to an MPS attention correctness issue." + }, + { + "left": "issue:43653", + "right": "issue:44912", + "accept": false, + "reason": "Tokenizer special-token decode bug is unrelated to MXFP4 quantization loading fallback." }, { "left": "issue:45084", "right": "issue:45538", "accept": false, - "reason": "Template-node compilation error versus CLIPTokenizer max-length constant bug; unrelated." + "reason": "Template-node compilation error is unrelated to CLIPTokenizer model_max_length behavior." }, { - "left": "issue:43653", - "right": "issue:44912", + "left": "issue:44568", + "right": "issue:45488", "accept": false, - "reason": "Tokenizer special token decoding issue versus MXFP4 quantization fallback; different subsystems." + "reason": "Different tokenizer regressions on different model families and code paths." }, { "left": "issue:43454", "right": "issue:43577", "accept": false, - "reason": "Weight tying bug in AyaVision versus dtype propagation bug in Blip2; different fixes." + "reason": "AyaVision weight-tying bug is unrelated to Blip2 dtype propagation." }, { - "left": "issue:44568", - "right": "issue:45488", + "left": "issue:34634", + "right": "issue:36331", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different tokenizer implementations and failure modes." + "reason": "Duplicate of unrelated pairing; no shared underlying bug." }, { - "left": "issue:34634", - "right": "issue:36331", + "left": "issue:43746", + "right": "issue:44811", "accept": false, - "reason": "BarkProcessor voice preset issue is unrelated to CustomTrainer.compute_loss signature incompatibility." + "reason": "Local checkpoint loading with PEFT adapters is unrelated to Whisper batch_decode skip_special_tokens handling." }, { - "left": "issue:43452", - "right": "issue:44991", + "left": "issue:43257", + "right": "issue:43329", "accept": false, - "reason": "gguf_file/from_pretrained handling versus a specific tokenizer loading regression; not the same code-path." + "reason": "Qwen3 MOE loading/weight conversion bug is unrelated to multimodal token counting." }, { "left": "issue:43638", "right": "issue:44936", "accept": false, - "reason": "DeepSpeed ZeRO-3 empty-index error versus trainer.evaluate failing after train; different behaviors and likely causes." + "reason": "Zero3 empty-index crash is not the same as evaluate-after-train behavior." + }, + { + "left": "issue:43452", + "right": "issue:44991", + "accept": false, + "reason": "gguf_file path handling and est-roberta tokenizer loading are different failures." + }, + { + "left": "issue:43854", + "right": "issue:44977", + "accept": false, + "reason": "GLM-4.7-Flash unit-test loading issue is unrelated to Qwen3.5 flash-attention generation." } ] }, @@ -4599,6 +4635,8 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-364", "nodes": [ + "issue:33357", + "issue:34567", "issue:36010", "issue:36331", "issue:38175", @@ -4625,7 +4663,7 @@ "issue:43749", "issue:43756", "issue:43819", - "issue:43854", + "issue:43906", "issue:43927", "issue:43931", "issue:44291", @@ -4638,7 +4676,6 @@ "issue:44811", "issue:44912", "issue:44938", - "issue:44977", "issue:44987", "issue:45042", "issue:45072", @@ -4648,259 +4685,269 @@ "issue:45538" ], "soft_pairs": [ - "issue:43854|issue:44977", "issue:38175|issue:44610", "issue:44589|issue:45042", - "issue:39692|issue:44912", "issue:43257|issue:43696", + "issue:39692|issue:44912", "issue:43819|issue:43927", "issue:43262|issue:44811", - "issue:36010|issue:43645", "issue:43720|issue:44610", "issue:43550|issue:44912", + "issue:36010|issue:43645", "issue:44568|issue:45356", "issue:43749|issue:44756", "issue:43329|issue:44479", - "issue:43531|issue:45042", "issue:43257|issue:43931", + "issue:43531|issue:45042", "issue:36010|issue:36331", "issue:43329|issue:43450", "issue:43257|issue:43756", "issue:43577|issue:45072", - "issue:43452|issue:43582", "issue:43697|issue:45357", - "issue:41720|issue:43540", + "issue:43452|issue:43582", "issue:42175|issue:44938", + "issue:41720|issue:43540", "issue:44479|issue:45381", - "issue:44987|issue:45538", "issue:43927|issue:44912", "issue:43540|issue:44560", - "issue:43317|issue:45538", - "issue:42915|issue:43742", "issue:43540|issue:44479", - "issue:44291|issue:45042" + "issue:33357|issue:34567", + "issue:44987|issue:45538", + "issue:42915|issue:43742", + "issue:43317|issue:45538", + "issue:44291|issue:45042", + "issue:43906|issue:45381" ], - "prepared_review_unit_hash": "39fe55d49ec512f122f63e69482b751bcdf4bda86d08e1c76a19ef63b3beae1f", + "prepared_review_unit_hash": "cc6f5f98bea4fdf34eafd804d61fe03d873f30c87031b4dbf93473a6d70784e1", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 31, - "serialized_chars": 36508, - "estimated_input_tokens": 9127, - "estimated_eval_tokens": 18510 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37400, + "estimated_input_tokens": 9350, + "estimated_eval_tokens": 18956 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is heterogeneous: the soft-similarity pairs mostly connect unrelated bugs across loading, quantization, video/multimodal processing, tokenizer behavior, and environment issues. I reject all soft edges and treat this as a broad mixed cluster rather than a duplicate set. If a single issue must anchor the cluster, 43749 is the strongest representative.", - "confidence": 0.79, - "canonical_issue_reason": "No true duplicate canonical spans this cluster; if one representative is needed, issue 43749 is the best anchor because it has the highest activity/inbound references and describes a broad loading/regression problem rather than a narrow model-specific symptom.", + "summary": "This cluster is mostly a set of unrelated issues; the only clear duplicate-style area is the Qwen video-input regression family. I\u2019d anchor the cluster on the broad v5.3.0 video regression report and only link the closest Qwen video-processing reports to it.", + "confidence": 0.7, + "canonical_issue_reason": "issue:44479 is the broadest and most central report in the only coherent subcluster here: a Transformers v5.3.0 video-input regression affecting multiple Qwen VL/MoE variants.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43749 is the best cluster representative: high discussion/inbound references and a broad infrastructure loading failure make it a stronger triage anchor than the many isolated model- or feature-specific bugs.", + "best_issue_reason": "issue:44479 best represents the shared underlying problem space because it describes a versioned regression with multiple affected model families and the same video-input code path.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43854", - "right": "issue:44977", - "accept": false, - "reason": "Different failures: GLM-4.7 load/test problem vs Qwen3.5 flash-attention generation regression." - }, { "left": "issue:38175", "right": "issue:44610", "accept": false, - "reason": "Unrelated symptoms: zero probabilities in SigLIP2 vs processor/model input-size mismatch in OmDet-Turbo." + "reason": "Different models and failures: SIGLIP2 zero probabilities vs OmDet-Turbo image-size mismatch." }, { "left": "issue:44589", "right": "issue:45042", "accept": false, - "reason": "Different layers: Float8 storage/type loading error vs PIL image processor incorrectly requiring torchvision." + "reason": "Unrelated backend issues: storage-object/float8 error vs unnecessary torchvision requirement in PIL image processors." }, { - "left": "issue:39692", - "right": "issue:44912", + "left": "issue:43257", + "right": "issue:43696", "accept": false, - "reason": "Docs/example mistakes for SigLIP2 vs MXFP4 quantization fallback for git-oss-20b." + "reason": "Different problems: Qwen3 MoE weight conversion under accelerate/deepspeed vs GPT-oss CUDA OOM." }, { - "left": "issue:43257", - "right": "issue:43696", + "left": "issue:39692", + "right": "issue:44912", "accept": false, - "reason": "Different problems: Qwen3 MoE loading with accelerate+deepspeed vs plain CUDA out-of-memory on GPT-oss-20b." + "reason": "Docs/example mistakes for SigLIP2 are not the same bug as MXFP4 loading falling back to bf16." }, { "left": "issue:43819", "right": "issue:43927", "accept": false, - "reason": "DAC forward/STE mismatch is unrelated to DiaConfig save/load token ID corruption." + "reason": "DAC latent conversion mismatch and DiaConfig save/load token-ID loss are separate model/config bugs." }, { "left": "issue:43262", "right": "issue:44811", "accept": false, - "reason": "Audio chat-template sampling-rate default bug is unrelated to Whisper batch_decode skip_special_tokens handling." - }, - { - "left": "issue:36010", - "right": "issue:43645", - "accept": false, - "reason": "Importing GenerationMixin and Jupyter custom-model initialization are separate breakages." + "reason": "Both are audio-related, but one is chat-template sampling-rate defaulting and the other is Whisper batch_decode token skipping." }, { "left": "issue:43720", "right": "issue:44610", "accept": false, - "reason": "Packed-weight unpacking during accelerate loading is unrelated to OmDet processor output-shape mismatch." + "reason": "Packed-weight unloading during accelerate loading is unrelated to the OmDet-Turbo processor size mismatch." }, { "left": "issue:43550", "right": "issue:44912", "accept": false, - "reason": "torch.compile/SDPA failure in Bamba is not the same as MXFP4 loading fallback in git-oss-20b." + "reason": "torch.compile/SDPA on Bamba is a different code path from MXFP4 quantization loading." + }, + { + "left": "issue:36010", + "right": "issue:43645", + "accept": false, + "reason": "Importing GenerationMixin from transformers.generation is not the same as the Jupyter/custom-model init regression." }, { "left": "issue:44568", "right": "issue:45356", "accept": false, - "reason": "Tokenizer add_special_tokens regression is distinct from Kimi-K2.5 codec/warning behavior." + "reason": "Both are tokenizer regressions, but they affect different models and different failure modes." }, { "left": "issue:43749", "right": "issue:44756", "accept": false, - "reason": "Both involve loading/performance, but one is FSDP CPU RAM efficient loading and the other is mmap on Strix Halo OOM; not the same bug." + "reason": "One is FSDP RAM-efficient loading, the other is Apple Silicon mmap/OOM behavior; related theme, not same bug." }, { "left": "issue:43329", "right": "issue:44479", - "accept": false, - "reason": "Undefined video-branch helper variables are a code bug; the other is a broader v5.3 video regression across Qwen models." + "accept": true, + "reason": "Both concern Qwen video-input handling in the same multimodal code path; one exposes the video-branch bug, the other reports the broader v5.3.0 regression." }, { - "left": "issue:43531", - "right": "issue:45042", + "left": "issue:43257", + "right": "issue:43931", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior has no overlap with PIL backend torchvision dependency regression." + "reason": "Different loading problems: MoE weight conversion under distributed loading vs Qwen3-VL shape mismatch." }, { - "left": "issue:43257", - "right": "issue:43931", + "left": "issue:43531", + "right": "issue:45042", "accept": false, - "reason": "MoE weights not converted under accelerate+deepspeed is different from Qwen3-VL-30B shape-mismatch loading." + "reason": "Qwen3-MoE sliding_window behavior is unrelated to PIL image processors requiring torchvision." }, { "left": "issue:36010", "right": "issue:36331", "accept": false, - "reason": "GenerationMixin import path breakage is unrelated to CustomTrainer.compute_loss signature mismatch." + "reason": "ImportError for GenerationMixin and compute_loss signature mismatch are separate API breakages." }, { "left": "issue:43329", "right": "issue:43450", "accept": false, - "reason": "Undefined video token-count helpers differ from batched video processor output-shape issues." + "reason": "Undefined video-token helper names and batched video processor output shape are different defects." }, { "left": "issue:43257", "right": "issue:43756", "accept": false, - "reason": "Qwen3 MoE loading/conversion failure is unrelated to Smollm3 RoPE-layer drop." + "reason": "Qwen3 MoE weight conversion and Smollm3 RoPE-layer count are unrelated model-specific issues." }, { "left": "issue:43577", "right": "issue:45072", "accept": false, - "reason": "Blip2 dtype staying float32 is a different dtype bug than SwitchTransformers/TimmWrapperModel bfloat16 mismatches." - }, - { - "left": "issue:43452", - "right": "issue:43582", - "accept": false, - "reason": "gguf_file tokenizer/model loading failure is unrelated to Apple Silicon caching_allocator_warmup TypeError." + "reason": "BLIP2 dtype propagation and SwitchTransformers/TimmWrapper bfloat16 mismatches are not the same bug." }, { "left": "issue:43697", "right": "issue:45357", "accept": false, - "reason": "RTDetrV2 output drift is not the same as Qwen3.5 save_pretrained visual-encoder key regression." + "reason": "RTDetrV2 output divergence and Qwen3.5 save_pretrained key regression are unrelated." }, { - "left": "issue:41720", - "right": "issue:43540", + "left": "issue:43452", + "right": "issue:43582", "accept": false, - "reason": "Qwen3 auto device-map CUDA assert on A800 and Qwen3OmniMoe video ValueError are different failures." + "reason": "gguf_file loader breakage and Apple Silicon caching_allocator_warmup TypeError are separate issues." }, { "left": "issue:42175", "right": "issue:44938", "accept": false, - "reason": "Missing TensorFlow in extras install and Python 3.14 import failure are unrelated environment/package issues." + "reason": "Missing TensorFlow in torch-only install and Python 3.14 load failure do not share the same root cause." }, { - "left": "issue:44479", - "right": "issue:45381", + "left": "issue:41720", + "right": "issue:43540", "accept": false, - "reason": "The former is a broad video-input regression across several Qwen models; the latter is a specific vision_position_ids bug in qwen2.5-vl." + "reason": "Qwen3 auto device-map cuda assert and Qwen3OmniMoe video ValueError are different failure paths." }, { - "left": "issue:44987", - "right": "issue:45538", - "accept": false, - "reason": "Loading physical-intelligence/fast on transformers>=5.1.0 and CLIPTokenizer model_max_length are unrelated." + "left": "issue:44479", + "right": "issue:45381", + "accept": true, + "reason": "Both are the same Qwen2.5-VL video-input regression class in v5.3.0, with 45381 specifically calling out wrong vision_position_ids." }, { "left": "issue:43927", "right": "issue:44912", "accept": false, - "reason": "DiaConfig token-ID persistence and MXFP4 quantization fallback affect different code paths and model families." + "reason": "DiaConfig token-ID persistence and MXFP4 loading fallback are unrelated." }, { "left": "issue:43540", "right": "issue:44560", "accept": false, - "reason": "Qwen3OmniMoe video processing ValueError is distinct from Qwen3-vl-embedding StopIteration." + "reason": "Different Qwen video bugs: one is a ValueError in Qwen3OmniMoe video processing, the other is a StopIteration regression in Qwen3-VL embedding." }, { - "left": "issue:43317", + "left": "issue:43540", + "right": "issue:44479", + "accept": false, + "reason": "Both involve Qwen video input, but the reported failures and model families are different enough to avoid merging." + }, + { + "left": "issue:33357", + "right": "issue:34567", + "accept": false, + "reason": "MacOS bus error with CLIP and TrainerState token counting are unrelated." + }, + { + "left": "issue:44987", "right": "issue:45538", "accept": false, - "reason": "dequantized model offload/device_map=auto failure is unrelated to CLIPTokenizer max-length handling." + "reason": "Loading physical-intelligence/fast and CLIPTokenizer max-length handling are separate problems." }, { "left": "issue:42915", "right": "issue:43742", "accept": false, - "reason": "Qwen3Moe FineGrainedFP8Config failure and MobileLLM key error are different model-loading bugs." + "reason": "Qwen3MoE FP8 config failure and MobileLLM key error do not share a concrete code path." }, { - "left": "issue:43540", - "right": "issue:44479", + "left": "issue:43317", + "right": "issue:45538", "accept": false, - "reason": "A specific Qwen3OmniMoe video ValueError is not the same as the broader v5.3 video regression across multiple Qwen models." + "reason": "Dequantized model loading with device_map=auto is unrelated to CLIPTokenizer max_length." }, { "left": "issue:44291", "right": "issue:45042", "accept": false, - "reason": "init_empty_weights unexpected argument error is unrelated to the torchvision requirement regression in image processors." + "reason": "init_empty_weights/_is_hf_initialized and PIL image processor torchvision dependency are unrelated." + }, + { + "left": "issue:43906", + "right": "issue:45381", + "accept": false, + "reason": "A reproduction of another issue is not the same as the Qwen2.5-VL video-position regression unless the underlying bug is shown to match, which isn\u2019t evident here." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: most pairs are indeed unrelated, and the two accepted Qwen video-input links are the only ones that plausibly share a root cause. The soft-edge rejections stay conservative and do not overstate equivalence." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -4919,7 +4966,6 @@ "nodes": [ "issue:28282", "issue:29127", - "issue:33357", "issue:34567", "issue:34689", "issue:36010", @@ -4935,12 +4981,12 @@ "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", @@ -4958,6 +5004,7 @@ "issue:45042", "issue:45081", "issue:45137", + "issue:45216", "issue:45290", "issue:45325", "issue:45375", @@ -4967,269 +5014,259 @@ "issue:45538" ], "soft_pairs": [ - "issue:33357|issue:34567", "issue:45137|issue:45405", - "issue:44062|issue:44991", - "issue:43906|issue:45381", "issue:42915|issue:45538", - "issue:42915|issue:44821", - "issue:42175|issue:43824", - "issue:42175|issue:44488", "issue:43526|issue:45479", + "issue:42175|issue:44488", + "issue:42175|issue:43824", + "issue:42915|issue:44821", + "issue:44062|issue:44991", "issue:38175|issue:44898", - "issue:43582|issue:43782", "issue:28282|issue:34689", + "issue:43582|issue:43782", "issue:43646|issue:43749", "issue:43493|issue:43994", - "issue:43742|issue:44291", - "issue:36010|issue:42175", "issue:43643|issue:45375", - "issue:28282|issue:29127", "issue:43756|issue:45325", + "issue:28282|issue:29127", + "issue:43742|issue:44291", + "issue:36010|issue:42175", "issue:44484|issue:44704", "issue:34567|issue:36331", - "issue:44062|issue:45081", - "issue:43452|issue:44488", "issue:44291|issue:44821", "issue:44792|issue:44964", + "issue:43452|issue:44488", + "issue:44062|issue:45081", "issue:43688|issue:43749", "issue:44514|issue:45290", - "issue:43452|issue:43577", - "issue:43742|issue:44589", "issue:43643|issue:44561", + "issue:43452|issue:43577", "issue:44514|issue:45381", + "issue:43742|issue:44589", + "issue:43697|issue:45216", "issue:45042|issue:45081" ], - "prepared_review_unit_hash": "beb48c37b6c02c005b90f9cdecc8251348a8ff7d32cb9792226d58c9399aa3b2", + "prepared_review_unit_hash": "5c3027d17cb4ca17045ff57f6e82d29720a89d5b6ebf9693b341f155b3813b39", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 37451, - "estimated_input_tokens": 9363, - "estimated_eval_tokens": 18982 + "soft_pair_count": 31, + "serialized_chars": 37197, + "estimated_input_tokens": 9300, + "estimated_eval_tokens": 18856 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most items are unrelated issue reports, mostly around distinct model-loading, tokenizer, or backend regressions. The only strong duplicate-looking pair is the Qwen2.5-VL apply_chat_template crash reports (44514/45290); the SigLIP2 pair is a plausible same-bug match but less certain. The rest look like false-positive similarity matches.", - "confidence": 0.66, - "canonical_issue_reason": "Issue 44514 is the better canonical representative for the apply_chat_template crash pair: both 44514 and 45290 describe crashes in the Qwen2.5-VL chat-template/tokenization path, and 44514 is the broader repro while 45290 is a more specific trigger.", + "summary": "This cluster is heterogeneous: most items are isolated Transformer bugs/regressions around different models, loaders, and training paths rather than duplicates of one underlying issue. I would not merge any of the soft pairs as duplicates.", + "confidence": 0.94, + "canonical_issue_reason": "If one issue must anchor the cluster, issue 43646 is the broadest and clearest regression report (`Transformers 5.0.0 breaks custom model initialization`). It is still only a weak canonical because the rest of the set spans unrelated model-loading, tokenizer, training, and processor bugs.", "canonical_pr_reason": null, - "best_issue_reason": "44514 is the clearest issue to keep as canonical because it directly names the shared apply_chat_template crash and has the broader repro description.", + "best_issue_reason": "Issue 43646 is the best representative issue because it is a general version-regression report with a concrete failure mode and broader applicability than the more model-specific reports. That said, it is not a strong duplicate hub for the whole set.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:33357", - "right": "issue:34567", - "accept": false, - "reason": "Different subsystems and failures: a CLIP Mac bus error vs TrainerState token-count tracking." - }, { "left": "issue:45137", "right": "issue:45405", "accept": false, - "reason": "DeepSpeed ZeRO3 deque errors and a PEFT version bump are unrelated." + "reason": "DeepSpeed ZeRO3 deque error and PEFT version pinning are unrelated problems." }, { - "left": "issue:44062", - "right": "issue:44991", - "accept": false, - "reason": "Both involve tokenizer/model loading, but the reported failures and likely causes differ." - }, - { - "left": "issue:43906", - "right": "issue:45381", + "left": "issue:42915", + "right": "issue:45538", "accept": false, - "reason": "Different Qwen2.5-VL regressions affecting different behaviors and code paths." + "reason": "FP8 config failure for Qwen3Moe is unrelated to CLIPTokenizer's extreme max_length value." }, { - "left": "issue:42915", - "right": "issue:45538", + "left": "issue:43526", + "right": "issue:45479", "accept": false, - "reason": "FP8 MoE loading and CLIP tokenizer max length are unrelated." + "reason": "BeitImageProcessorFast label reduction bug and sequence-classification zero-loss bug are different code paths." }, { - "left": "issue:42915", - "right": "issue:44821", + "left": "issue:42175", + "right": "issue:44488", "accept": false, - "reason": "Different model/load paths; no clear shared underlying bug." + "reason": "TensorFlow backend packaging issue is not the same as a model-loading failure for `cjvt/sleng-bert`." }, { "left": "issue:42175", "right": "issue:43824", "accept": false, - "reason": "Backend packaging vs missing model-class import are unrelated." + "reason": "Missing TensorFlow backend and missing Qwen2_5_VL export are unrelated import/setup issues." }, { - "left": "issue:42175", - "right": "issue:44488", + "left": "issue:42915", + "right": "issue:44821", "accept": false, - "reason": "Dependency packaging and a specific model load failure are not the same bug." + "reason": "FP8 config handling and AutoImageProcessor URL loading are different bugs." }, { - "left": "issue:43526", - "right": "issue:45479", + "left": "issue:44062", + "right": "issue:44991", "accept": false, - "reason": "Beit label reduction and sequence-classification zero loss are different problems." + "reason": "AddedToken keyword collision and tokenizer loading failure for est-roberta are separate tokenizer issues." }, { "left": "issue:38175", "right": "issue:44898", "accept": false, - "reason": "Same broad vision area, but different models and failure modes." + "reason": "SigLIP2 zero probabilities and Perceiver non-default resolution failure are different model-specific inference bugs." }, { - "left": "issue:43582", - "right": "issue:43782", + "left": "issue:28282", + "right": "issue:34689", "accept": false, - "reason": "Apple Silicon allocator TypeError and Qwen3VL weight_only loading error are unrelated." + "reason": "PyTorch missing import error and Llama 3.2 Vision loading regression are unrelated." }, { - "left": "issue:28282", - "right": "issue:34689", + "left": "issue:43582", + "right": "issue:43782", "accept": false, - "reason": "Missing PyTorch ImportError and Llama 3.2 Vision load breakage are different issues." + "reason": "AppleSilicon warmup TypeError and Qwen3VL weight_only loading error do not share the same failure mechanism." }, { "left": "issue:43646", "right": "issue:43749", "accept": false, - "reason": "Custom model initialization breakage and FSDP CPU RAM-efficient loading are separate changes." + "reason": "Custom model initialization regression and FSDP CPU RAM efficient loading breakage are different issues." }, { "left": "issue:43493", "right": "issue:43994", - "accept": true, - "reason": "Both report SigLIP2 correctness regressions on the same checkpoint, so they plausibly describe the same underlying model bug." - }, - { - "left": "issue:43742", - "right": "issue:44291", "accept": false, - "reason": "A MobileLLM key error and an init_empty_weights TypeError are not the same failure." + "reason": "SigLIP2 implementation discrepancy and bad AutoModel/pipeline outputs are both about the model, but not the same concrete bug." }, { - "left": "issue:36010", - "right": "issue:42175", + "left": "issue:43643", + "right": "issue:45375", "accept": false, - "reason": "GenerationMixin import failure and TensorFlow backend packaging are unrelated." + "reason": "Both concern missing config fields, but one is trust_remote_code object hydration and the other is a specific Qwen config field dropped by strict parsing." }, { - "left": "issue:43643", - "right": "issue:45375", + "left": "issue:43756", + "right": "issue:45325", "accept": false, - "reason": "Both mention missing config fields, but the model/config contexts differ too much to treat as one bug." + "reason": "Smollm3 RoPE layer count mismatch and Qwen2.5-VL temporal position scaling are different architecture/position encoding problems." }, { "left": "issue:28282", "right": "issue:29127", "accept": false, - "reason": "Missing PyTorch dependency error vs LayoutLMv3 message clarity are unrelated." + "reason": "PyTorch import error and LayoutLMv3 box-content error are unrelated error-message/validation issues." }, { - "left": "issue:43756", - "right": "issue:45325", + "left": "issue:43742", + "right": "issue:44291", "accept": false, - "reason": "RoPE layer dropping and Qwen2.5-VL temporal position scaling are distinct model-specific bugs." + "reason": "MobileLLM loading key error and init_empty_weights unexpected `_is_hf_initialized` argument are different loading regressions." + }, + { + "left": "issue:36010", + "right": "issue:42175", + "accept": false, + "reason": "GenerationMixin import failure and TensorFlow backend packaging issue are separate import/export problems." }, { "left": "issue:44484", "right": "issue:44704", "accept": false, - "reason": "save_pretrained shard sizing and AutoProcessor kwargs propagation are different code paths." + "reason": "max_shard_size default question and AutoProcessor kwargs propagation bug are unrelated save/load behavior issues." }, { "left": "issue:34567", "right": "issue:36331", "accept": false, - "reason": "TrainerState token counting and compute_loss signature compatibility are separate Trainer issues." + "reason": "TrainerState token counting and CustomTrainer.compute_loss signature mismatch are different training-trainer bugs." }, { - "left": "issue:44062", - "right": "issue:45081", + "left": "issue:44291", + "right": "issue:44821", "accept": false, - "reason": "AddedToken special-kw handling and Mistral regex patching are unrelated tokenizer bugs." + "reason": "Unexpected init_empty_weights argument and AutoImageProcessor URL loading do not describe the same bug." }, { - "left": "issue:43452", - "right": "issue:44488", + "left": "issue:44792", + "right": "issue:44964", "accept": false, - "reason": "gguf_file loading breakage and a specific sleng-bert load failure are not clearly the same bug." + "reason": "Janus test failure and Phi-4 multimodal loading failure are separate model-specific regressions." }, { - "left": "issue:44291", - "right": "issue:44821", + "left": "issue:43452", + "right": "issue:44488", "accept": false, - "reason": "init_empty_weights TypeError and loading AutoImageProcessor from URL are unrelated." + "reason": "gguf_file support breakage and inability to load `cjvt/sleng-bert` are different load-path failures." }, { - "left": "issue:44792", - "right": "issue:44964", + "left": "issue:44062", + "right": "issue:45081", "accept": false, - "reason": "Janus test failure and Phi-4 multimodal model load failure are different multimodal issues." + "reason": "Tokenizer AddedToken duplication and Mistral regex patch backend-tokenizer crash are distinct tokenizer bugs." }, { "left": "issue:43688", "right": "issue:43749", "accept": false, - "reason": "Auxiliary-loss normalization and FSDP loading are unrelated." + "reason": "Auxiliary loss normalization and FSDP CPU RAM efficient loading are unrelated training-time bugs." }, { "left": "issue:44514", "right": "issue:45290", - "accept": true, - "reason": "Both are apply_chat_template/tokenize crashes in the Qwen2.5-VL path; the triggers differ, but they look close enough to be the same underlying bug." + "accept": false, + "reason": "Batched Qwen2_5_VL chat template crash and assistant tool-call crash are both chat-template issues but not the same concrete edge case." + }, + { + "left": "issue:43643", + "right": "issue:44561", + "accept": false, + "reason": "Missing config fields under trust_remote_code and v5 removal of `is_torch_fx_available` are separate upgrade breakages." }, { "left": "issue:43452", "right": "issue:43577", "accept": false, - "reason": "gguf_file loading and BLIP-2 dtype handling are different bugs." + "reason": "gguf loading breakage and BLIP2 dtype not propagating are different model-loading defects." }, { - "left": "issue:43742", - "right": "issue:44589", + "left": "issue:44514", + "right": "issue:45381", "accept": false, - "reason": "Model-load key errors and missing Float8 storage errors are unrelated." + "reason": "Both mention Qwen2.5-VL chat/video behavior, but one is chat-template batching and the other is incorrect vision position IDs." }, - { - "left": "issue:43643", - "right": "issue:44561", + { + "left": "issue:43742", + "right": "issue:44589", "accept": false, - "reason": "Missing config fields and removal of is_torch_fx_available are different breakages." + "reason": "MobileLLM key error and missing Float8 storage are unrelated load/runtime compatibility problems." }, { - "left": "issue:44514", - "right": "issue:45381", + "left": "issue:43697", + "right": "issue:45216", "accept": false, - "reason": "An apply_chat_template crash and a position-id scaling bug are unrelated." + "reason": "RTDetrV2 output drift and Qwen3.5 checkpoint save corruption are different regression classes." }, { "left": "issue:45042", "right": "issue:45081", "accept": false, - "reason": "Image-processor torchvision dependency regression and Mistral regex patching are unrelated." + "reason": "PIL backend requiring torchvision and Mistral regex backend-tokenizer crash are not the same issue." } ] }, - "evaluator_result": { - "accept": false, - "feedback": "The writeup overstates the evidence in the packet. In particular, issue 44514 vs 45290 are not shown as exact duplicates by the titles alone: one is about batched input with `padding=False`, the other about assistant tool-call messages with no content. The canonical-issue reason also claims they match \"almost verbatim\" and specifies a concrete shared failure mode that is not grounded in the packet text. The SigLIP2 pair may be plausible, but the overall summary and the accepted soft-edge verdicts should be phrased more conservatively." - }, - "evaluator_used": true, - "retried": true, - "accepted_nontrivial_soft_edge": true, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -5260,6 +5297,7 @@ "issue:41762", "issue:43122", "issue:43295", + "issue:43329", "issue:43452", "issue:43525", "issue:43531", @@ -5269,7 +5307,6 @@ "issue:43688", "issue:43697", "issue:43720", - "issue:43742", "issue:43819", "issue:43906", "issue:43927", @@ -5291,264 +5328,257 @@ "issue:44977", "issue:44991", "issue:45005", - "issue:45216", + "issue:45325", "issue:45356", "issue:45440" ], "soft_pairs": [ - "issue:43697|issue:45216", "issue:43531|issue:44821", "issue:44625|issue:44877", - "issue:43452|issue:44843", "issue:43819|issue:44811", "issue:43720|issue:44898", "issue:43688|issue:45440", "issue:43653|issue:44568", - "issue:41762|issue:44936", "issue:44464|issue:44743", + "issue:43452|issue:44843", "issue:44466|issue:45005", + "issue:41762|issue:44936", "issue:43295|issue:45356", "issue:24643|issue:29127", - "issue:41720|issue:43906", "issue:40444|issue:44811", + "issue:41720|issue:43906", "issue:43525|issue:43643", "issue:44361|issue:44442", "issue:43577|issue:43697", "issue:38175|issue:43295", "issue:43653|issue:44361", "issue:44843|issue:44991", - "issue:44479|issue:44936", "issue:28282|issue:32090", "issue:28282|issue:33357", "issue:28282|issue:35141", + "issue:44479|issue:44936", "issue:43122|issue:43906", "issue:44977|issue:45356", - "issue:30064|issue:33666", "issue:44464|issue:44977", - "issue:43531|issue:43742", + "issue:30064|issue:33666", "issue:43653|issue:43927", - "issue:43577|issue:44554" + "issue:43577|issue:44554", + "issue:43329|issue:45325" ], - "prepared_review_unit_hash": "94ed7c71b5d7e2361920365a8b2a93d9a93b88c255acd0f0d5b410e3b6333623", + "prepared_review_unit_hash": "3a9d58759f0771d81ff339fd4bdde70b0f7b1987bccab8a44715aed6ffb45fef", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 37547, - "estimated_input_tokens": 9387, - "estimated_eval_tokens": 19030 + "soft_pair_count": 31, + "serialized_chars": 37383, + "estimated_input_tokens": 9346, + "estimated_eval_tokens": 18948 }, "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-style themes (tokenizers, configs, model loading, regressions), but they describe different models, code paths, and failure modes. I would not merge any of the proposed pairs as duplicates.", - "confidence": 0.95, - "canonical_issue_reason": "No single issue cleanly represents the cluster; the items span unrelated bugs across tokenization, model loading, training, vision, and generation.", + "summary": "This cluster is a false-positive mix of unrelated issues across tokenizers, configs, multimodal processors, training, and device/runtime regressions. The soft pairs only share broad subsystem words, not the same concrete bug/change, so all soft edges should be rejected.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue fits this cluster; the items are heterogeneous and do not share one underlying defect.", "canonical_pr_reason": null, - "best_issue_reason": "No issue is a strong global duplicate target because the cluster lacks a single underlying bug.", + "best_issue_reason": "No strong best issue representative exists because the cluster is not a true duplicate set.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43697", - "right": "issue:45216", - "accept": false, - "reason": "Both are regressions, but one is RTDetr output mismatch and the other is Qwen3.5 save_pretrained corruption; different code paths and fixes." - }, { "left": "issue:43531", "right": "issue:44821", "accept": false, - "reason": "Qwen3-MoE sliding window recurrence bug vs AutoImageProcessor URL loading; unrelated subsystems." + "reason": "Qwen3-MoE sliding-window behavior and AutoImageProcessor URL loading are unrelated bugs." }, { "left": "issue:44625", "right": "issue:44877", "accept": false, - "reason": "Different config propagation issues for Qwen3.5 versus strict Granite speech config loading." - }, - { - "left": "issue:43452", - "right": "issue:44843", - "accept": false, - "reason": "GGUF tokenizer/model loading breakage is not the same as Mistral regex/offline model_info() behavior." + "reason": "num_labels propagation in Qwen3.5 is separate from granite_speech strict-config loading." }, { "left": "issue:43819", "right": "issue:44811", "accept": false, - "reason": "DAC latent reconstruction mismatch is unrelated to Whisper batch_decode skip_special_tokens handling." + "reason": "DAC latent decoding and Whisper batch_decode special-token handling are different code paths." }, { "left": "issue:43720", "right": "issue:44898", "accept": false, - "reason": "BitNet packed-weight loading and Perceiver interpolation are distinct model-specific bugs." + "reason": "BitNet packed-weight loading and Perceiver interpolation failure are distinct model bugs." }, { "left": "issue:43688", "right": "issue:45440", "accept": false, - "reason": "Auxiliary loss normalization in MoE models is unrelated to DeepSeekV3 native-vs-remote divergence." + "reason": "Auxiliary-loss normalization and DeepseekV3MoE divergence are not the same defect." }, { "left": "issue:43653", "right": "issue:44568", "accept": false, - "reason": "BigBirdTokenizer special-token registration is unrelated to mdeberta-v3 BOS/EOS add_special_tokens behavior." + "reason": "Both involve tokens, but BigBird mask-token registration and mdeberta BOS/EOS insertion are different issues." }, { - "left": "issue:41762", - "right": "issue:44936", + "left": "issue:44464", + "right": "issue:44743", "accept": false, - "reason": "Gemma3 ZeRO-3 load failure and trainer.evaluate() after train() are different failure modes." + "reason": "Compiled-forward chunked generation and Qwen3_5 recurrent-state reset are separate generation bugs." }, { - "left": "issue:44464", - "right": "issue:44743", + "left": "issue:43452", + "right": "issue:44843", "accept": false, - "reason": "Chunked generation with compiled forward is not the same as Qwen3.5 recurrent-state reset in modular_qwen3_5.py." + "reason": "gguf_file loading breakage and offline model_info calls in mistral regex patch are different tokenizer-loading failures." }, { "left": "issue:44466", "right": "issue:45005", "accept": false, - "reason": "lm_head serialization/tied-weights issue and translation-model tied-weights issues are related only at a broad level, not the same bug." + "reason": "Device-dependent lm_head serialization and translation-model tied-weight problems are not clearly the same bug." + }, + { + "left": "issue:41762", + "right": "issue:44936", + "accept": false, + "reason": "ZeRO-3 model loading IndexError and trainer.evaluate after train are unrelated trainer/runtime issues." }, { "left": "issue:43295", "right": "issue:45356", "accept": false, - "reason": "Custom model processor/tokenizer regression differs from Kimi-K2.5 tokenizer codec/fix_mistral_regex regression." + "reason": "Processor.tokenizer regression and Kimi-K2.5 codec/tokenizer regression are separate changes." }, { "left": "issue:24643", "right": "issue:29127", "accept": false, - "reason": "DeepSpeed training weight-dimension error is not the same as LayoutLMv3 box validation messaging." + "reason": "DeepSpeed weight shape RuntimeError and LayoutLMV3 box error messaging are unrelated." }, { - "left": "issue:41720", - "right": "issue:43906", + "left": "issue:40444", + "right": "issue:44811", "accept": false, - "reason": "Qwen3 auto device mapping cuda assert and an isolated reproduction of another issue are not the same concrete bug." + "reason": "IterableDataset multi-image finetuning failure is unrelated to Whisper batch_decode behavior." }, { - "left": "issue:40444", - "right": "issue:44811", + "left": "issue:41720", + "right": "issue:43906", "accept": false, - "reason": "IterableDataset multi-image finetuning failure is unrelated to Whisper batch_decode special-token decoding." + "reason": "Qwen3 auto device-mapping cudaErrorAssert and an isolated reproduction reference are not the same concrete bug." }, { "left": "issue:43525", "right": "issue:43643", "accept": false, - "reason": "Missing pad_token_id on Llama4Config is distinct from trust_remote_code returning incomplete AutoConfig fields." + "reason": "Missing pad_token_id in Llama4Config and missing fields from trust_remote_code config loading are different config issues." }, { "left": "issue:44361", "right": "issue:44442", "accept": false, - "reason": "MLukeTokenizer task AttributeError and FastSpeech2ConformerTokenizer loading failure are different tokenizer implementations." + "reason": "MLukeTokenizer task AttributeError and FastSpeech2ConformerTokenizer loading failure are unrelated tokenizer bugs." }, { "left": "issue:43577", "right": "issue:43697", "accept": false, - "reason": "BLIP2 dtype preservation and RTDetr output differences are unrelated model behavior bugs." + "reason": "Blip2 dtype casting and RTDetrV2 output drift in v5 are different model regressions." }, { "left": "issue:38175", "right": "issue:43295", "accept": false, - "reason": "SigLIP2 zero probabilities and processor/tokenizer regression involve different models and symptoms." + "reason": "Siglip2 zero probabilities and processor.tokenizer regression are different problem classes." }, { "left": "issue:43653", - "right": "issue:44361", + "right": "issue:43927", "accept": false, - "reason": "BigBirdTokenizer special-token registration and MLukeTokenizer AttributeError are separate tokenizer bugs." + "reason": "BigBird special-token registration and DiaConfig custom token IDs are separate token-management bugs." }, { "left": "issue:44843", "right": "issue:44991", "accept": false, - "reason": "Offline model_info() behavior in _patch_mistral_regex is not the same as tokenizer loading failure for est-roberta." - }, - { - "left": "issue:44479", - "right": "issue:44936", - "accept": false, - "reason": "Qwen video-input regression and trainer lifecycle failure are unrelated." + "reason": "Offline model_info misuse and EMBEDDIA/est-roberta tokenizer loading failure are not the same underlying issue." }, { "left": "issue:28282", "right": "issue:32090", "accept": false, - "reason": "Generic PyTorch import failure and Trainer _gpu_broadcast_one NoneType error are clearly different." + "reason": "Missing PyTorch import and Trainer NoneType broadcast error are unrelated." }, { "left": "issue:28282", "right": "issue:33357", "accept": false, - "reason": "Missing PyTorch import and MacOS bus error with community CLIP are unrelated." + "reason": "PyTorch not found and MacOS bus error in CLIP are unrelated runtime failures." }, { "left": "issue:28282", "right": "issue:35141", "accept": false, - "reason": "PyTorch missing import versus embedding reinitialization after resizing are different layers of failure." + "reason": "PyTorch import error and tied embedding reinitialization are different bugs." + }, + { + "left": "issue:44479", + "right": "issue:44936", + "accept": false, + "reason": "Qwen video-input regression and trainer.evaluate after train are unrelated." }, { "left": "issue:43122", "right": "issue:43906", "accept": false, - "reason": "Tokenizer version drift and an isolated reproduction of another issue are not the same bug." + "reason": "Tokenizer output changes across versions and an isolated reproduction link do not establish the same bug." }, { "left": "issue:44977", "right": "issue:45356", "accept": false, - "reason": "Qwen3.5 flash-attention generation regression and Kimi-K2.5 tokenizer regression involve different model families and paths." - }, - { - "left": "issue:30064", - "right": "issue:33666", - "accept": false, - "reason": "Void segmentation map processing and Qwen2-VL multi-GPU training are unrelated vision/training issues." + "reason": "Qwen3.5 flash-attention generation issue and Kimi-K2.5 tokenizer regression are different failures." }, { "left": "issue:44464", "right": "issue:44977", "accept": false, - "reason": "Compiled-forward chunked generation inconsistency is not the same as Qwen3.5 flash-attention generation failure." + "reason": "Chunked-generation compiled-forward inconsistency and flash-attention generation failure are separate runtime issues." }, { - "left": "issue:43531", - "right": "issue:43742", + "left": "issue:30064", + "right": "issue:33666", "accept": false, - "reason": "Qwen3-MoE sliding-window state handling is unrelated to MobileLLM-125M key error on load." + "reason": "Void segmentation map processing and Qwen2-VL multi-GPU training are unrelated." }, { "left": "issue:43653", "right": "issue:43927", "accept": false, - "reason": "BigBirdTokenizer special-token registration and DiaConfig custom token IDs save/load corruption are different tokenizer/config issues." + "reason": "This pair is still not the same bug: special-token registration vs config token ID persistence." }, { "left": "issue:43577", "right": "issue:44554", "accept": false, - "reason": "BLIP2 dtype loading bug and MPS attention correctness issue are unrelated." + "reason": "BLIP2 dtype loading and MPS attention correctness on mismatched head dims are unrelated." + }, + { + "left": "issue:43329", + "right": "issue:45325", + "accept": false, + "reason": "Undefined video-branch helper variables and still-image rope-index scaling are different multimodal bugs." } ] }, @@ -5589,6 +5619,7 @@ "issue:43493", "issue:43504", "issue:43525", + "issue:43531", "issue:43540", "issue:43550", "issue:43572", @@ -5617,18 +5648,16 @@ "issue:45084", "issue:45200", "issue:45229", - "issue:45325", "issue:45357", "issue:45405" ], "soft_pairs": [ - "issue:43329|issue:45325", + "issue:43531|issue:43742", "issue:44479|issue:44625", - "issue:43012|issue:43577", - "issue:41720|issue:44464", "issue:43550|issue:44361", + "issue:41720|issue:44464", "issue:43504|issue:43720", - "issue:43452|issue:44451", + "issue:43012|issue:43577", "issue:43504|issue:43550", "issue:32090|issue:33666", "issue:32090|issue:34567", @@ -5636,245 +5665,246 @@ "issue:43504|issue:44610", "issue:43278|issue:43688", "issue:44849|issue:45357", + "issue:43452|issue:44451", "issue:45200|issue:45229", "issue:43334|issue:43572", "issue:43756|issue:44913", - "issue:43540|issue:43716", "issue:43493|issue:43825", - "issue:43742|issue:44821", + "issue:43540|issue:43716", "issue:33666|issue:36010", - "issue:43525|issue:43572", - "issue:43720|issue:43927", + "issue:43742|issue:44821", "issue:43606|issue:43927", + "issue:43720|issue:43927", + "issue:43525|issue:43572", "issue:43688|issue:43873", "issue:43278|issue:43844", - "issue:43540|issue:44977", - "issue:43720|issue:44361", "issue:43257|issue:43299", - "issue:44964|issue:45405", + "issue:43720|issue:44361", + "issue:43540|issue:44977", "issue:39692|issue:43493", + "issue:44964|issue:45405", "issue:43329|issue:43606" ], - "prepared_review_unit_hash": "ef75676425bce8efaec28f2219f33b26c1e20460a0e842dbe7cbb5e499dddd27", + "prepared_review_unit_hash": "571b94dd7510861df4962f309e4d828f0876b3ea44551413acc84c11a8e90bd4", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 37319, - "estimated_input_tokens": 9330, - "estimated_eval_tokens": 18916 + "serialized_chars": 37256, + "estimated_input_tokens": 9314, + "estimated_eval_tokens": 18884 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose semantic grouping of unrelated Transformers regressions (config-loading, multimodal/video bugs, dtype/offload issues, and compile/runtime errors). None of the soft pairs look like exact duplicate reports; at best they are related by broad theme, not the same concrete bug.", - "confidence": 0.41, - "canonical_issue_reason": "No single exact duplicate stands out, but issue 43334 is the best anchor because it cleanly represents the recurring v5 config-loading regression where a model-specific config is missing a pad token field and loading fails with an AttributeError.", + "summary": "This cluster is not a duplicate set; it\u2019s a broad mix of unrelated Transformers bugs across loading, dtype, multimodal, training, and config serialization paths. All soft pairs should be rejected.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue stands out because the items span distinct models and code paths rather than one underlying bug.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43334 is the most representative issue for the dominant configuration-regression theme in this set: it is specific, reproducible, and clearly tied to a loader breakage in dev versions.", + "best_issue_reason": "No strong global representative issue: the apparent similarities are mostly broad subsystem overlap, not the same defect.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43329", - "right": "issue:45325", + "left": "issue:43531", + "right": "issue:43742", "accept": false, - "reason": "Both are Qwen VL/video-related, but they describe different bugs: undefined video-token helpers in multimodal token counting vs wrong still-image temporal position scaling." + "reason": "Both are load/config failures, but they affect different models and symptoms; no shared code-path bug is evident." }, { "left": "issue:44479", "right": "issue:44625", "accept": false, - "reason": "One is a video-input regression across several Qwen VL models; the other is a Qwen3.5 config propagation bug for num_labels. Different failure modes and fixes." + "reason": "Video-input regression vs num_labels propagation bug; unrelated behavior and fixes." }, { - "left": "issue:43012", - "right": "issue:43577", + "left": "issue:43550", + "right": "issue:44361", "accept": false, - "reason": "Both mention dtype/precision, but one is a PyTorch warning during compile and the other is model/qformer staying float32 after load. Not the same code-path problem." + "reason": "torch.compile/SDPA failure is unrelated to a tokenizer AttributeError on tasks." }, { "left": "issue:41720", "right": "issue:44464", "accept": false, - "reason": "Different problems: CUDA assert with auto device mapping for Qwen3 vs inconsistent outputs from chunked generation with compiled forward." - }, - { - "left": "issue:43550", - "right": "issue:44361", - "accept": false, - "reason": "Unrelated subsystems: SDPA/torch.compile failure in Bamba vs tokenizer AttributeError on tasks in MLukeTokenizer." + "reason": "CUDA assert during auto device mapping is a different bug from inconsistent chunked generation under compiled forward." }, { "left": "issue:43504", "right": "issue:43720", "accept": false, - "reason": "Both are loading-related, but one is a legacy field in Beit semantic segmentation config and the other is packed-weight unpacking for BitNet during accelerate loading." + "reason": "Legacy-field loading issue in Beit is unrelated to BitNet packed-weight unpacking during accelerate loading." }, { - "left": "issue:43452", - "right": "issue:44451", + "left": "issue:43012", + "right": "issue:43577", "accept": false, - "reason": "Both concern model loading, but the concrete bugs differ: gguf_file handling in tokenizer/model loaders vs failure to load a specific pretrained model due to changed behavior." + "reason": "Both mention dtype, but one is a compile-time warning and the other is a load-time dtype propagation bug for a different model." }, { "left": "issue:43504", "right": "issue:43550", "accept": false, - "reason": "Different models and different failures: legacy config field loading vs compile/SDPA runtime failure." + "reason": "Different models and failure modes; no same underlying defect." }, { "left": "issue:32090", "right": "issue:33666", "accept": false, - "reason": "Trainer broadcast TypeError and Qwen2-VL multi-GPU training are too broad and do not describe the same underlying issue." + "reason": "Both touch multi-GPU/training, but one is a Trainer broadcast TypeError and the other is a Qwen2-VL training topic." }, { "left": "issue:32090", "right": "issue:34567", "accept": false, - "reason": "Trainer broadcast NoneType error and TrainerState token counter not updating are unrelated Trainer bugs." + "reason": "Trainer broadcast NoneType error and token-count tracking bug are separate Trainer issues." }, { "left": "issue:43262", "right": "issue:45084", "accept": false, - "reason": "Audio chat-template sampling-rate defaulting and template-node compilation failure are different components and symptoms." + "reason": "Audio chat-template sampling-rate default is unrelated to compile-time non-template-node errors." }, { "left": "issue:43504", "right": "issue:44610", "accept": false, - "reason": "One is a legacy preset-loading issue in BEiT; the other is an OmDet-Turbo processor/model input-size mismatch." + "reason": "Pretrained loading with a legacy field is unrelated to processor/model resolution-size mismatch." }, { "left": "issue:43278", "right": "issue:43688", "accept": false, - "reason": "Both involve training/inference behavior, but one is dtype drift between training and evaluate while the other is auxiliary loss normalization in MoE models." + "reason": "Dtype drift in eval is unrelated to auxiliary-loss normalization." }, { "left": "issue:44849", "right": "issue:45357", "accept": false, - "reason": "Qwen3.5 hidden-states bug and incorrect visual encoder keys on save_pretrained are distinct regressions." + "reason": "Hidden-states output bug and incorrect save_pretrained visual keys are different Qwen3.5 issues." + }, + { + "left": "issue:43452", + "right": "issue:44451", + "accept": false, + "reason": "Both are model-loading failures, but one is gguf_file-specific and the other is a different model regression." }, { "left": "issue:45200", "right": "issue:45229", "accept": false, - "reason": "Different Gemma 4 issues: text-only mm_token_type_ids defaults vs multi-GPU inference OOM." + "reason": "mm_token_type_ids defaulting and multi-GPU CUDA OOM are unrelated Gemma 4 problems." }, { "left": "issue:43334", "right": "issue:43572", "accept": false, - "reason": "Both are missing-pad-token-field config regressions, but they affect different config classes and attribute names; the evidence is too broad to treat them as the same bug." + "reason": "Missing pad_token_id vs missing pad_token_idx are similar in wording, but they are different config fields on different models." }, { "left": "issue:43756", "right": "issue:44913", "accept": false, - "reason": "Smollm3 RoPE-layer behavior and GPTNeoX rotary_pct reload behavior are separate architecture/config issues." - }, - { - "left": "issue:43540", - "right": "issue:43716", - "accept": false, - "reason": "Different multimodal bugs: Qwen3OmniMoe video processing ValueError vs Mistral-3 image preprocessor/model dtype mismatch." + "reason": "RoPE-layer dropping and rotary_pct reload regression both involve configuration, but they are not the same bug." }, { "left": "issue:43493", "right": "issue:43825", "accept": false, - "reason": "SigLIP2 HF-vs-JAX discrepancy and pipeline translation-task error message are unrelated." + "reason": "SigLIP2 implementation discrepancy is unrelated to a pipeline error message about supported translation tasks." }, { - "left": "issue:43742", - "right": "issue:44821", + "left": "issue:43540", + "right": "issue:43716", "accept": false, - "reason": "MobileLLM key error on load and AutoImageProcessor-from-URL failure are different loader paths with different root causes." + "reason": "Qwen3OmniMoe video-processing failure and Mistral-3 dtype mismatch are separate multimodal issues." }, { "left": "issue:33666", "right": "issue:36010", "accept": false, - "reason": "Qwen2-VL multi-GPU training and GenerationMixin import failure are not the same underlying bug." + "reason": "Qwen2-VL multi-GPU training and GenerationMixin import failure have no shared code-path." }, { - "left": "issue:43525", - "right": "issue:43572", + "left": "issue:43742", + "right": "issue:44821", "accept": false, - "reason": "Both are config attribute regressions after v5, but they involve different models and fields; not enough evidence of one shared fix." + "reason": "MobileLLM loading key error and AutoImageProcessor URL loading are unrelated." }, { - "left": "issue:43720", + "left": "issue:43606", "right": "issue:43927", "accept": false, - "reason": "Packed-weight unpacking during accelerate loading and DiaConfig custom token IDs lost on save/load are different bugs." + "reason": "CPU offload device mismatch is unrelated to DiaConfig save/load token-ID loss." }, { - "left": "issue:43606", + "left": "issue:43720", "right": "issue:43927", "accept": false, - "reason": "CPU offload device mismatch and DiaConfig save/load token ID loss are unrelated." + "reason": "Packed-weight unpacking during accelerate loading is unrelated to DiaConfig custom-token-ID persistence." + }, + { + "left": "issue:43525", + "right": "issue:43572", + "accept": false, + "reason": "Both are missing-token-id config regressions, but they affect different fields and models; not clearly one defect." }, { "left": "issue:43688", "right": "issue:43873", "accept": false, - "reason": "Aux-loss normalization in MoE models and quantization/offloading behavior are different change areas." + "reason": "Aux-loss normalization and quantization/offload behavior are unrelated." }, { "left": "issue:43278", "right": "issue:43844", "accept": false, - "reason": "Embedding dtype drift in evaluate and abnormal gradient growth under DeepSpeed ZeRO-3 are different training issues." + "reason": "Eval dtype drift is unrelated to ZeRO-3 gradient inflation." }, { - "left": "issue:43540", - "right": "issue:44977", + "left": "issue:43257", + "right": "issue:43299", "accept": false, - "reason": "Both concern Qwen3.5-family behavior, but one is video-input processing and the other is flash-attention generation correctness." + "reason": "Both involve Qwen3 MoE loading, but one is accelerate/deepspeed weight conversion and the other is a transformers-dev regression on Qwen3VL MoE; too different to merge." }, { "left": "issue:43720", "right": "issue:44361", "accept": false, - "reason": "BitNet packed-weight loading and MLukeTokenizer task AttributeError are unrelated." + "reason": "BitNet loading bug and MLukeTokenizer task AttributeError are unrelated." }, { - "left": "issue:43257", - "right": "issue:43299", + "left": "issue:43540", + "right": "issue:44977", "accept": false, - "reason": "Both involve Qwen3/MoE loading, but one is accelerate+deepspeed weight conversion and the other is a dev-version loading regression in Qwen3VL MoE. Too different to merge." + "reason": "Video-input processing failure and flash-attention generation bug are different Qwen3-family issues." }, { - "left": "issue:44964", - "right": "issue:45405", + "left": "issue:39692", + "right": "issue:43493", "accept": false, - "reason": "Model loading failure and a version-pin release issue are not the same bug or change." + "reason": "Both are SigLIP2-related, but one is a docs/example mismatch and the other is an implementation discrepancy vs JAX." }, { - "left": "issue:39692", - "right": "issue:43493", + "left": "issue:44964", + "right": "issue:45405", "accept": false, - "reason": "SigLIP2 doc example errors and HF-vs-JAX implementation discrepancy are different issues; one is documentation/example correctness, the other is model parity." + "reason": "Phi-4 multimodal loading failure is unrelated to the PEFT minimum-version packaging issue." }, { "left": "issue:43329", "right": "issue:43606", "accept": false, - "reason": "Video token counting bug and Bark CPU offload device mismatch are unrelated." + "reason": "Undefined video-token helper/merge-size bug is unrelated to Bark CPU-offload device mismatch." } ] }, @@ -5947,48 +5977,48 @@ "issue:45520" ], "soft_pairs": [ - "issue:44488|issue:44843", - "issue:44589|issue:45520", "issue:33666|issue:36331", "issue:43493|issue:43844", + "issue:44589|issue:45520", + "issue:44488|issue:44843", "issue:43756|issue:44060", "issue:43646|issue:44466", - "issue:44423|issue:44734", "issue:43688|issue:43746", "issue:44521|issue:45290", - "issue:43824|issue:44589", + "issue:44423|issue:44734", "issue:33357|issue:33666", + "issue:43824|issue:44589", "issue:39692|issue:43295", "issue:43994|issue:44493", "issue:43504|issue:43606", "issue:43504|issue:43653", + "issue:43927|issue:44361", "issue:43575|issue:45405", "issue:43572|issue:44291", - "issue:43927|issue:44361", "issue:43278|issue:43701", "issue:44792|issue:45405", "issue:43257|issue:45405", "issue:43606|issue:44112", "issue:15354|issue:38175", - "issue:45084|issue:45520", "issue:44554|issue:45446", - "issue:43540|issue:45200", - "issue:43582|issue:44977" + "issue:43582|issue:44977", + "issue:45084|issue:45520", + "issue:43540|issue:45200" ], - "prepared_review_unit_hash": "60577a898ffde87ef031e4d79b015b747d8b15dfea12936bb0d1afc557ba0e81", + "prepared_review_unit_hash": "ee560bf93c09400dc64cea186b4e338c70173e41e2d42536e016b6913be9ddae", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 27, - "serialized_chars": 35425, + "serialized_chars": 35428, "estimated_input_tokens": 8857, "estimated_eval_tokens": 17970 }, @@ -5997,174 +6027,174 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous. A few pairs share a subsystem or API surface, but none look like the same underlying bug/change, so I reject all soft edges. No PRs are present.", - "confidence": 0.89, - "canonical_issue_reason": "Issue 44521 is the closest representative of the only coherent subtopic in the set (multimodal/chat-template behavior) and is an active user-facing regression, but it is not a duplicate of the other items.", + "summary": "The cluster is largely heterogeneous: most items are unrelated model-specific bugs, regressions, or docs/CI issues. None of the soft pairs look like true duplicates, though a few are superficially similar by subsystem.", + "confidence": 0.93, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue 44521 is the best umbrella issue because it is concrete, current, and in a broadly used path; the rest of the set is too mixed to form a better canonical duplicate target.", + "best_issue_reason": "Issue 43646 is the broadest regression-style report in the set, but even it does not subsume the other problems; there is no strong single canonical issue for this cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44488", - "right": "issue:44843", + "left": "issue:33666", + "right": "issue:36331", "accept": false, - "reason": "Both involve loading/tokenizer-related behavior, but one is a model-loading failure and the other is an offline-mode tokenizer patch bug; different failure modes and code paths." + "reason": "Different bugs: multi-GPU training behavior vs a Trainer API signature mismatch." }, { - "left": "issue:44589", - "right": "issue:45520", + "left": "issue:43493", + "right": "issue:43844", "accept": false, - "reason": "These are unrelated import/storage/type errors in different areas; no shared bug or change." + "reason": "Unrelated failures: SigLIP2 output discrepancy vs ZeRO-3 gradient growth during training." }, { - "left": "issue:33666", - "right": "issue:36331", + "left": "issue:44589", + "right": "issue:45520", "accept": false, - "reason": "Multi-GPU training for Qwen2-VL and a CustomTrainer signature regression are unrelated." + "reason": "Different code paths: missing Float8 storage handling vs flash_attn import lookup on Python 3.13." }, { - "left": "issue:43493", - "right": "issue:43844", + "left": "issue:44488", + "right": "issue:44843", "accept": false, - "reason": "SigLIP2 implementation discrepancy and ZeRO-3 gradient growth are different correctness issues." + "reason": "Both involve model loading/tokenizer flow, but one is a specific model load failure and the other is offline hub access in a tokenizer patch." }, { "left": "issue:43756", "right": "issue:44060", "accept": false, - "reason": "Both mention model internals, but one is RoPE layer dropping and the other is a tied-weights warning involving Qwen3-Next; not the same bug." + "reason": "Different model-architecture issues: RoPE layer dropping vs a tied-weights warning bug." }, { "left": "issue:43646", "right": "issue:44466", "accept": false, - "reason": "Both touch serialization/tied weights, but one is custom model initialization breakage and the other is device-dependent lm_head serialization; different concrete problems." - }, - { - "left": "issue:44423", - "right": "issue:44734", - "accept": false, - "reason": "Both are server crashes, but one is a multimodal string/.to error and the other is KV-cache tensor indexing; different code paths." + "reason": "Both are v5-related, but one is custom model initialization and the other is device-dependent serialization of lm_head.weight." }, { "left": "issue:43688", "right": "issue:43746", "accept": false, - "reason": "Auxiliary-loss normalization and PEFT checkpoint loading are unrelated." + "reason": "Completely different failures: auxiliary-loss normalization vs PEFT adapter checkpoint loading." }, { "left": "issue:44521", "right": "issue:45290", "accept": false, - "reason": "Both use apply_chat_template, but one is wrong assistant mask output for multimodal inputs and the other is a tool-call/no-content crash; distinct bugs." + "reason": "Same API surface, but different bugs: zero assistant masks in multimodal inputs vs tool-call messages with no content crashing." }, { - "left": "issue:43824", - "right": "issue:44589", + "left": "issue:44423", + "right": "issue:44734", "accept": false, - "reason": "An import error for Qwen2.5-VL and a missing Float8 storage type are unrelated." + "reason": "Both are serving crashes, but one is multimodal input handling and the other is KV-cache continuation indexing." }, { "left": "issue:33357", "right": "issue:33666", "accept": false, - "reason": "MacOS bus error on a CLIP model and Qwen2-VL multi-GPU training are different issues." + "reason": "Different problem domains: MacOS bus error in CLIP loading vs multi-GPU training." + }, + { + "left": "issue:43824", + "right": "issue:44589", + "accept": false, + "reason": "Import error for a missing model class is unrelated to the Float8 storage-object lookup failure." }, { "left": "issue:39692", "right": "issue:43295", "accept": false, - "reason": "A SigLIP2 docs example bug and a regression in processor/tokenizer access are not the same underlying problem." + "reason": "Docs/example failures and processor/tokenizer mismatch are not the same as the v4.57.5 processor.tokenizer regression." }, { "left": "issue:43994", "right": "issue:44493", "accept": false, - "reason": "Nonsensical SigLIP2 outputs and unexpected position-id keys point to different model/config issues." + "reason": "Different symptoms and likely causes: SigLIP2 nonsensical outputs vs a position-id key warning affecting many models." }, { "left": "issue:43504", "right": "issue:43606", "accept": false, - "reason": "Legacy-field load failure for Beit and a CPU offload device mismatch in bark-small are unrelated." + "reason": "One is a legacy-field loading bug in BEiT, the other is a CPU offload device-mismatch in Bark." }, { "left": "issue:43504", "right": "issue:43653", "accept": false, - "reason": "Legacy field loading for Beit and BigBirdTokenizer special-token registration are different bugs." + "reason": "Legacy-field loading for BEiT is unrelated to BigBirdTokenizer special-token registration." + }, + { + "left": "issue:43927", + "right": "issue:44361", + "accept": false, + "reason": "Different tokenizer/config issues: saved token IDs being lost vs MLukeTokenizer AttributeError on tasks." }, { "left": "issue:43575", "right": "issue:45405", "accept": false, - "reason": "Tensor-parallel OOM for a model load and a PEFT version bump are not the same issue." + "reason": "OOM when loading a large model with tensor parallelism is unrelated to a PEFT version pin issue." }, { "left": "issue:43572", "right": "issue:44291", "accept": false, - "reason": "Missing pad_token_idx in a config and an init_empty_weights/_is_hf_initialized argument mismatch are different compatibility problems." - }, - { - "left": "issue:43927", - "right": "issue:44361", - "accept": false, - "reason": "Config token-ID persistence and MLukeTokenizer AttributeError are unrelated." + "reason": "Missing config field after update vs an init_empty_weights signature change are different regressions." }, { "left": "issue:43278", "right": "issue:43701", "accept": false, - "reason": "Embedding dtype drift between train/eval and resume_from_checkpoint key mismatch are different regressions." + "reason": "Different workflows: dtype change during evaluate vs resume_from_checkpoint key mismatch." }, { "left": "issue:44792", "right": "issue:45405", "accept": false, - "reason": "A janus image-generation test failure and a PEFT version pin issue are unrelated." + "reason": "A janus image-generation test failure is unrelated to the PEFT minimum-version issue." }, { "left": "issue:43257", "right": "issue:45405", "accept": false, - "reason": "Qwen3 MOE weight conversion under accelerate/deepspeed and a dependency-version bump are not the same bug." + "reason": "Qwen3 MoE weight conversion with accelerate/deepspeed is unrelated to the PEFT version bump." }, { "left": "issue:43606", "right": "issue:44112", "accept": false, - "reason": "Both are CI/device-mismatch themed, but one is real CPU offload mismatch and the other is a stale GraniteSpeech test; not duplicates." + "reason": "A real runtime device-mismatch bug is not the same as a stale GraniteSpeech CI test." }, { "left": "issue:15354", "right": "issue:38175", "accept": false, - "reason": "JIT scripting/export failure and zero probabilities on SigLIP2 are unrelated." + "reason": "TorchScript GeneratorExp export failure and SigLIP2 zero-probabilities are unrelated bugs." }, { - "left": "issue:45084", - "right": "issue:45520", + "left": "issue:44554", + "right": "issue:45446", "accept": false, - "reason": "Template-node compilation error and flash_attn import failure on Python 3.13 are different issues." + "reason": "MPS attention correctness and a PyTorch version check/import guard are different issues." }, { - "left": "issue:44554", - "right": "issue:45446", + "left": "issue:43582", + "right": "issue:44977", "accept": false, - "reason": "MPS attention correctness with mismatched head dims and a PyTorch version check bug for AuxRequest import are unrelated." + "reason": "Apple Silicon allocator TypeError is unrelated to Qwen3.5 flash-attention generation behavior." }, { - "left": "issue:43540", - "right": "issue:45200", + "left": "issue:45084", + "right": "issue:45520", "accept": false, - "reason": "Qwen3OmniMoe video processing error and Gemma 4 mm_token_type_ids defaulting are different multimodal bugs." + "reason": "Compilation of non-template nodes and flash_attn import lookup are unrelated failures." }, { - "left": "issue:43582", - "right": "issue:44977", + "left": "issue:43540", + "right": "issue:45200", "accept": false, - "reason": "AppleSilicon caching_allocator_warmup TypeError and Qwen3.5 flash-attention generation failure are not the same underlying defect." + "reason": "Video-input processing in Qwen3OmniMoe is unrelated to mm_token_type_ids defaults for text-only fine-tuning." } ] }, @@ -6238,15 +6268,15 @@ "issue:45412" ], "soft_pairs": [ - "issue:45072|issue:45198", "issue:43232|issue:44849", + "issue:45072|issue:45198", "issue:43844|issue:44964", "issue:39401|issue:44568", "issue:44315|issue:44849", "issue:37428|issue:43957", "issue:44554|issue:45356", - "issue:43334|issue:43525", "issue:43295|issue:44610", + "issue:43334|issue:43525", "issue:43493|issue:43701", "issue:45200|issue:45325", "issue:44898|issue:45412", @@ -6254,14 +6284,14 @@ "issue:44466|issue:45127", "issue:40990|issue:44625", "issue:43388|issue:43526", - "issue:44514|issue:44521", "issue:43844|issue:44936", - "issue:37428|issue:43824", + "issue:44514|issue:44521", "issue:43575|issue:44464", - "issue:43572|issue:44987", + "issue:37428|issue:43824", "issue:43295|issue:43450", - "issue:43572|issue:44589", + "issue:43572|issue:44987", "issue:44442|issue:45198", + "issue:43572|issue:44589", "issue:43824|issue:43975", "issue:43295|issue:43873", "issue:38175|issue:45412", @@ -6269,214 +6299,214 @@ "issue:43295|issue:43994", "issue:44360|issue:44561" ], - "prepared_review_unit_hash": "9ccc3ce2f4f2b3714fe60080e04f08dc563b68bb5843857dfb2ddc3070e85c87", + "prepared_review_unit_hash": "788281c3c88c96a744b42a5e889fd43e1014bc81949fad8c8f70e5c5aed94b04", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36888, - "estimated_input_tokens": 9222, - "estimated_eval_tokens": 18700 + "serialized_chars": 36892, + "estimated_input_tokens": 9223, + "estimated_eval_tokens": 18702 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is very heterogeneous: it mixes unrelated tokenizer, multimodal, loading, training, and memory issues. None of the soft pairs looks like the same underlying bug or change, so I rejected all of them.", - "confidence": 0.96, - "canonical_issue_reason": "issue:44521 is the best cluster representative: it is active, well-discussed, and describes a concrete regression in a core multimodal API with clear repro details.", - "canonical_pr_reason": null, - "best_issue_reason": "issue:44521 has the strongest combination of impact, discussion, and reproducibility among this set, even though it does not duplicate the other issues.", - "best_pr_reason": null, + "summary": "This cluster is a heterogeneous mix of unrelated issue reports across tokenizers, processors, loading, training, and dtype/memory behavior. None of the soft pairs looks like the same underlying bug/change, so there is no clear canonical item.", + "confidence": 0.91, + "canonical_issue_reason": "No single issue clearly subsumes the others; the items span distinct components and failure modes.", + "canonical_pr_reason": "No PRs are present in this cluster.", + "best_issue_reason": "No issue is a strong global representative because the cluster is not centered on one bug class or code path.", + "best_pr_reason": "No PRs are present in this cluster.", "soft_edge_verdicts": [ { - "left": "issue:45072", - "right": "issue:45198", + "left": "issue:43232", + "right": "issue:44849", "accept": false, - "reason": "Different failures: dtype mismatch in bfloat16 inference vs save_pretrained/tokenization problems in Wav2Vec2." + "reason": "Different problems: generation kwarg sync after sync_gpus vs Qwen3.5 hidden-states behavior." }, { - "left": "issue:43232", - "right": "issue:44849", + "left": "issue:45072", + "right": "issue:45198", "accept": false, - "reason": "Different code paths: generation kwargs sync issue vs Qwen3.5 hidden-states output bug." + "reason": "Different model paths and symptoms: bfloat16 dtype mismatches vs Wav2Vec2 save/tokenization failure." }, { "left": "issue:43844", "right": "issue:44964", "accept": false, - "reason": "Unrelated problems: ZeRO-3 gradient growth vs model-loading failure for Phi-4 multimodal." + "reason": "ZeRO-3 gradient growth and a Phi-4 loading failure are unrelated bugs." }, { "left": "issue:39401", "right": "issue:44568", "accept": false, - "reason": "Both are tokenizer bugs, but they affect different models and behaviors: offset mapping vs missing BOS/EOS on add_special_tokens." + "reason": "Different tokenizer bugs: wrong offset mapping vs add_special_tokens not inserting BOS/EOS." }, { "left": "issue:44315", "right": "issue:44849", "accept": false, - "reason": "Liger kernel not applied during model_init is unrelated to Qwen3.5 output_hidden_states behavior." + "reason": "Liger Kernel application in Trainer setup is unrelated to a Qwen3.5 hidden-states bug." }, { "left": "issue:37428", "right": "issue:43957", "accept": false, - "reason": "ImportError for a flash-attention helper is not the same as torch.device('meta') model-loading breakage." + "reason": "Importing a flash-attention helper and meta-device model loading are separate failure modes." }, { "left": "issue:44554", "right": "issue:45356", "accept": false, - "reason": "MPS attention correctness issue is unrelated to Kimi-K2.5 tokenizer codec/regression handling." + "reason": "MPS attention correctness and Kimi-K2.5 tokenizer codec handling are not the same bug." }, { - "left": "issue:43334", - "right": "issue:43525", + "left": "issue:43295", + "right": "issue:44610", "accept": false, - "reason": "Both mention missing pad_token_id, but they are different model families and distinct config objects, not the same bug." + "reason": "Processor/tokenizer API regression vs incorrect image size emitted by a video/vision processor." }, { - "left": "issue:43295", - "right": "issue:44610", + "left": "issue:43334", + "right": "issue:43525", "accept": false, - "reason": "Processor.tokenizer/images-to-tokenizer regression is unrelated to OmDet-Turbo image-size mismatch." + "reason": "Both mention missing pad_token_id, but for different models/config paths and likely separate fixes." }, { "left": "issue:43493", "right": "issue:43701", "accept": false, - "reason": "SigLIP2 implementation discrepancy is unrelated to resume_from_checkpoint key mismatch." + "reason": "SigLIP2 implementation discrepancy and resume_from_checkpoint key mismatch are unrelated." }, { "left": "issue:45200", "right": "issue:45325", "accept": false, - "reason": "Gemma 4 token-type defaults and Qwen2.5-VL rope-index scaling are different multimodal config bugs." + "reason": "Gemma 4 mm_token_type_ids defaulting and Qwen2.5-VL temporal position scaling are different issues." }, { "left": "issue:44898", "right": "issue:45412", "accept": false, - "reason": "Perceiver interpolation failure is unrelated to RT-DETR memory not being released." + "reason": "Perceiver resize failure and RT-DETR memory release are unrelated." }, { "left": "issue:43873", "right": "issue:44464", "accept": false, - "reason": "Quantization/offloading behavior is a different problem from chunked generation with compiled forward." + "reason": "Quantization offloading behavior and compiled-forward chunked generation are different code paths." }, { "left": "issue:44466", "right": "issue:45127", "accept": false, - "reason": "lm_head serialization inconsistency is unrelated to LoRA merge collapse with extended vocabulary." + "reason": "Serialization of lm_head.weight is not the same as LoRA merge collapse with extended vocab." }, { "left": "issue:40990", "right": "issue:44625", "accept": false, - "reason": "High perplexity on gpt-oss-20b is unrelated to Qwen3.5 num_labels propagation." + "reason": "High perplexity on gpt-oss-20b and Qwen3.5 num_labels propagation are unrelated." }, { "left": "issue:43388", "right": "issue:43526", "accept": false, - "reason": "Both involve labels, but one drops tuple labels in gather_for_metrics while the other reduces BEiT labels incorrectly." - }, - { - "left": "issue:44514", - "right": "issue:44521", - "accept": false, - "reason": "Same API area, but different symptoms: batched padding=False crash vs all-zero assistant masks for multimodal inputs." + "reason": "Trainer metric-gathering tuple handling differs from BeitImageProcessorFast reduce_labels behavior." }, { "left": "issue:43844", "right": "issue:44936", "accept": false, - "reason": "ZeRO-3 gradient inflation and trainer.evaluate-after-train failure are unrelated training issues." + "reason": "ZeRO-3 gradient instability is unrelated to trainer.evaluate() failing after train()." }, { - "left": "issue:37428", - "right": "issue:43824", + "left": "issue:44514", + "right": "issue:44521", "accept": false, - "reason": "Different import/loading failures: flash-attention helper ImportError vs missing Qwen2_5_VLForConditionalGeneration export." + "reason": "Both touch Qwen2.5-VL chat templating, but one is a batched padding=False crash and the other all-zero assistant masks; not clearly one fix." }, { "left": "issue:43575", "right": "issue:44464", "accept": false, - "reason": "OOM on loading Qwen2-57B with tp is not the same as chunked generation inconsistency with compiled forward." + "reason": "TP load OOM for Qwen2-57B is unrelated to compiled-forward chunked generation inconsistencies." }, { - "left": "issue:43572", - "right": "issue:44987", + "left": "issue:37428", + "right": "issue:43824", "accept": false, - "reason": "StableLmConfig pad_token_idx regression is unrelated to loading physical-intelligence/fast on transformers>=5.1.0." + "reason": "Different import errors for different symbols and models; no shared underlying change." }, { "left": "issue:43295", "right": "issue:43450", "accept": false, - "reason": "Custom processor/tokenizer regression is not the same as batched video processor shape handling." + "reason": "Processor/tokenizer regression and batched video-shape handling are distinct issues." }, { "left": "issue:43572", - "right": "issue:44589", + "right": "issue:44987", "accept": false, - "reason": "Missing pad_token_idx in StableLmConfig is unrelated to Float8 storage lookup failure." + "reason": "StableLmConfig pad_token_idx regression and a generic model-loading failure are not the same bug." }, { "left": "issue:44442", "right": "issue:45198", "accept": false, - "reason": "AutoTokenizer loading failure for FastSpeech2ConformerTokenizer is unrelated to Wav2Vec2 save/tokenization failure." + "reason": "FastSpeech2ConformerTokenizer loading and Wav2Vec2 save/tokenization failure are unrelated." }, { - "left": "issue:43824", - "right": "issue:43975", + "left": "issue:38175", + "right": "issue:45412", "accept": false, - "reason": "Qwen2_5_VL import failure is unrelated to deepseek-coder detokenization regression." + "reason": "SigLIP2 zero-probability outputs and RT-DETR memory cleanup are unrelated." }, { - "left": "issue:43295", - "right": "issue:43873", + "left": "issue:43278", + "right": "issue:43856", "accept": false, - "reason": "Processor/tokenizer regression and quantization offloading bug are different subsystems and different failures." + "reason": "Embedding dtype drift between train/eval is not the same as Qwen3 MoE memory inefficiency." }, { - "left": "issue:38175", - "right": "issue:45412", + "left": "issue:43572", + "right": "issue:44589", "accept": false, - "reason": "SigLIP2 zero-probabilities bug is unrelated to RT-DETR memory release behavior." + "reason": "Missing pad_token_idx in StableLmConfig and Float8 storage lookup failure are different regressions." }, { - "left": "issue:43278", - "right": "issue:43856", + "left": "issue:43824", + "right": "issue:43975", "accept": false, - "reason": "Embedding dtype drift in eval is not the same as Qwen3 MoE training memory inefficiency." + "reason": "Importing Qwen2.5-VL classes and DeepSeek-Coder detokenization are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:43873", + "accept": false, + "reason": "Processor/tokenizer regression and quantization offloading behavior are different subsystems." }, { "left": "issue:43295", "right": "issue:43994", "accept": false, - "reason": "Custom processor/tokenizer regression is unrelated to SigLIP2 nonsensical outputs." + "reason": "Processor/tokenizer regression and SigLIP2 nonsensical outputs are unrelated." }, { "left": "issue:44360", "right": "issue:44561", "accept": false, - "reason": "DSA indexer/ReLU discussion is unrelated to removal of is_torch_fx_available breaking trust_remote_code models." + "reason": "The DSA indexer ReLU discussion and removal of is_torch_fx_available breaking trust_remote_code models are unrelated." } ] }, @@ -6550,13 +6580,13 @@ "issue:45478" ], "soft_pairs": [ - "issue:43827|issue:43901", "issue:44849|issue:45003", + "issue:43827|issue:43901", "issue:38175|issue:43493", - "issue:42915|issue:43824", "issue:43653|issue:44112", - "issue:44568|issue:44991", + "issue:42915|issue:43824", "issue:43450|issue:43493", + "issue:44568|issue:44991", "issue:45127|issue:45245", "issue:45290|issue:45381", "issue:43819|issue:45083", @@ -6569,263 +6599,256 @@ "issue:44448|issue:44743", "issue:43606|issue:45083", "issue:43278|issue:43825", - "issue:37428|issue:43425", "issue:29942|issue:32090", "issue:45310|issue:45478", - "issue:43819|issue:44493", "issue:43611|issue:43931", - "issue:39692|issue:43526", + "issue:37428|issue:43425", "issue:43335|issue:43550", - "issue:37428|issue:44351", - "issue:43450|issue:43526", + "issue:43819|issue:44493", + "issue:39692|issue:43526", "issue:44855|issue:45083", + "issue:43450|issue:43526", + "issue:37428|issue:44351", "issue:45200|issue:45381", "issue:43404|issue:45200", "issue:39692|issue:44493", - "issue:43335|issue:43653", - "issue:44448|issue:44779" + "issue:43335|issue:43653" ], - "prepared_review_unit_hash": "97499c222d4baee33989bbca5f48ebb47a1e9a6d038cbaf1b4b0e10d6e95b63d", + "prepared_review_unit_hash": "3a02c0a826efc7240621ce8e20845cec08d60545a5cc7061a1e5c5413abe69c3", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 34, - "serialized_chars": 37953, - "estimated_input_tokens": 9489, - "estimated_eval_tokens": 19234 + "soft_pair_count": 33, + "serialized_chars": 37640, + "estimated_input_tokens": 9410, + "estimated_eval_tokens": 19076 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly heterogeneous issue set. The only soft-edge duplicate that is well supported from the packet is the Qwen3.5 MoE from_pretrained pair (issue:45310 / issue:45478); the other candidate pairs look like broader thematic similarities rather than the same underlying bug.", + "summary": "This set is mostly heterogeneous and should not be collapsed as one cluster. The only clear duplicate is the Qwen3.5 MoE from_pretrained regression pair (issue:45310 and issue:45478); the rest span unrelated bugs, docs issues, and model-specific failures.", "confidence": 0.94, - "canonical_issue_reason": "issue:45310 is the best canonical issue because it is the earlier report of the same Qwen3.5 MoE from_pretrained failure and has the same concrete title/scope as issue:45478.", + "canonical_issue_reason": "issue:45478 is the best canonical candidate for the only real duplicate pair: it reports the same Qwen3.5 MoE from_pretrained failure as issue:45310, with matching title/symptom and the later report likely being the duplicate.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45310 is the best representative issue for this duplicate pair: it states the concrete failure directly, predates issue:45478, and the packet supports treating 45478 as the later near-identical report.", + "best_issue_reason": "issue:45478 is the strongest issue representative in this set because it matches the concrete failure mode most directly; however, the broader collection is not a single deduplicable bug cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43827", - "right": "issue:43901", - "accept": false, - "reason": "Both are v5/docs-related, but they describe different concrete problems: leftover pipeline references versus return_all_scores docs/behavior." - }, { "left": "issue:44849", "right": "issue:45003", "accept": false, - "reason": "Different models and different code paths; one is a Qwen3.5 hidden-states bug, the other is unsafe sys.modules access." + "reason": "Different problems: Qwen3.5 hidden-state output bug vs unsafe sys.modules access in modeling_utils." }, { - "left": "issue:38175", - "right": "issue:43493", + "left": "issue:43827", + "right": "issue:43901", "accept": false, - "reason": "Both mention SigLIP2, but the reported failures are different: zero probabilities versus HF/JAX implementation discrepancy." + "reason": "Both are docs-related, but they concern different pipeline/deprecation topics and different sections." }, { - "left": "issue:42915", - "right": "issue:43824", + "left": "issue:38175", + "right": "issue:43493", "accept": false, - "reason": "Different model families and failure modes; FP8 config failure is unrelated to the Qwen2.5-VL import error." + "reason": "Both mention SigLIP2, but one is unexpected zero probabilities and the other is a broader HF vs JAX implementation discrepancy." }, { "left": "issue:43653", "right": "issue:44112", "accept": false, - "reason": "BigBirdTokenizer special-token registration and GraniteSpeech CI device override are unrelated bugs." + "reason": "Unrelated CI failures: BigBirdTokenizer special-token decode issue vs stale device override test in GraniteSpeech." }, { - "left": "issue:44568", - "right": "issue:44991", + "left": "issue:42915", + "right": "issue:43824", "accept": false, - "reason": "These are separate tokenizer regressions affecting different models and different behaviors." + "reason": "Different failures: FineGrainedFP8Config for Qwen3Moe vs missing Qwen2_5_VL import." }, { "left": "issue:43450", "right": "issue:43493", "accept": false, - "reason": "A batched video-shape bug is not the same as the SigLIP2 implementation mismatch." + "reason": "Different subsystems: batched video processor shape bug vs SigLIP2 implementation mismatch." + }, + { + "left": "issue:44568", + "right": "issue:44991", + "accept": false, + "reason": "Both tokenizer-related, but one is missing BOS/EOS addition and the other is tokenizer loading for a specific repo." }, { "left": "issue:45127", "right": "issue:45245", "accept": false, - "reason": "LoRA merge/vocab collapse and the category-count runtime error are different defects." + "reason": "Unrelated issues: LoRA merge collapse with extended vocabulary vs category cardinality runtime error." }, { "left": "issue:45290", "right": "issue:45381", "accept": false, - "reason": "Different multimodal bugs: chat-template tool-call handling versus Qwen2.5-VL video position ids." + "reason": "Different chat-template/video bugs: tool-call assistant content crash vs qwen2.5-vl video vision_position_ids mismatch." }, { "left": "issue:43819", "right": "issue:45083", "accept": false, - "reason": "DAC latent/STE mismatch and qwen3_omni_moe feature-length behavior do not describe the same code-path problem." + "reason": "Different DAC/qwen3_omni_moe bugs with no shared concrete code path." }, { "left": "issue:43450", "right": "issue:43575", "accept": false, - "reason": "Video processor batch-shape handling and TP OOM on model load are unrelated." + "reason": "Unrelated: video processor batch shape bug vs tensor-parallel OOM when loading Qwen2-57B-A14B-Instruct." }, { "left": "issue:44462", "right": "issue:44561", "accept": false, - "reason": "Tokenizer.json loading and removal of is_torch_fx_available affect different subsystems." + "reason": "Different regressions: tokenizer.json ignored vs removal of is_torch_fx_available breaking trust_remote_code models." }, { "left": "issue:44561", "right": "issue:44779", "accept": false, - "reason": "Both are v5 regressions, but one is trust_remote_code compatibility and the other is Deepseek tokenizer correctness." + "reason": "Both tokenizer/model compatibility issues in v5, but they affect different components and failure modes." }, { "left": "issue:44610", "right": "issue:44857", "accept": false, - "reason": "OmDet-Turbo processor sizing and LwDetrImageLoss AMP crash are different problems." + "reason": "Different model/loss paths: OmDet-Turbo processor input sizing vs LwDetrImageLoss AMP crash." }, { "left": "issue:43819", "right": "issue:44112", "accept": false, - "reason": "Audio/DAC latent handling is unrelated to GraniteSpeech CI device override behavior." + "reason": "No common bug: DAC latent mismatch vs GraniteSpeech CI device override failure." }, { "left": "issue:43262", "right": "issue:43450", "accept": false, - "reason": "Different processor bugs: audio sampling-rate defaulting versus batched video output shape." + "reason": "Audio chat-template sampling-rate default issue vs video processor batch-shape bug." }, { "left": "issue:44448", "right": "issue:44743", "accept": false, - "reason": "Pegasus v4/v5 output difference and qwen3_5 recurrent-state reset are unrelated." + "reason": "Unrelated: Pegasus output change across versions vs qwen3_5 modular recurrent-state reset bug." }, { "left": "issue:43606", "right": "issue:45083", "accept": false, - "reason": "CPU offload device mismatch is not the same as qwen3_omni_moe feature-length behavior." + "reason": "Device-mismatch CI bug from CPU offload vs qwen3_omni_moe helper output-length behavior." }, { "left": "issue:43278", "right": "issue:43825", "accept": false, - "reason": "Embedding dtype regression and pipeline error-message behavior are different issues." - }, - { - "left": "issue:37428", - "right": "issue:43425", - "accept": false, - "reason": "Two version-related errors, but with different missing symbols / incompatibility surfaces." + "reason": "Different failures: embedding dtype drift in eval vs pipeline() error message/deprecation behavior." }, { "left": "issue:29942", "right": "issue:32090", "accept": false, - "reason": "Flash Attention 2 test failures and a Trainer _gpu_broadcast_one NoneType error are unrelated." + "reason": "Both are old training/attention issues, but the symptoms and code paths differ completely." }, { "left": "issue:45310", "right": "issue:45478", "accept": true, - "reason": "Same title and same concrete Qwen3.5 MoE from_pretrained failure; 45478 reads as a later duplicate report of 45310." - }, - { - "left": "issue:43819", - "right": "issue:44493", - "accept": false, - "reason": "DAC latent/STE mismatch and position-id warnings across many models are different bugs." + "reason": "Same concrete bug: both report the Qwen3.5 MoE from_pretrained error on transformers>=5.4.0, with matching symptom and scope." }, { "left": "issue:43611", "right": "issue:43931", "accept": false, - "reason": "base_model_prefix loading breakage and Qwen3-VL weight-shape mismatch are distinct model-loading problems." + "reason": "Different loading regressions: base_model_prefix breakage vs weight-shape mismatch for Qwen3-VL-30B-A3B-Instruct." }, { - "left": "issue:39692", - "right": "issue:43526", + "left": "issue:37428", + "right": "issue:43425", "accept": false, - "reason": "SigLIP2 documentation/model-processor issues and BEiT fast label reduction are unrelated." + "reason": "Both are import/version compatibility issues, but they concern different missing symbols and different release breakages." }, { "left": "issue:43335", "right": "issue:43550", "accept": false, - "reason": "SwitchTransformers sparse-layer config and Bamba SDPA torch.compile failure are different defects." + "reason": "Different model/config bug classes: sparse layer construction vs torch.compile SDPA failure." }, { - "left": "issue:37428", - "right": "issue:44351", + "left": "issue:43819", + "right": "issue:44493", "accept": false, - "reason": "Both are import errors, but for different missing symbols and likely different regressions." + "reason": "Unrelated issues: DAC latent/forward mismatch vs unexpected position-id key warnings across models." }, { - "left": "issue:43450", + "left": "issue:39692", "right": "issue:43526", "accept": false, - "reason": "Video batch-shape handling and BEiT label reduction are unrelated." + "reason": "Different processor bugs: SigLIP2 docs/model mismatch vs BeitImageProcessorFast label reduction returning only one label." }, { "left": "issue:44855", "right": "issue:45083", "accept": false, - "reason": "Python 3.13 torch.jit parsing and qwen3_omni_moe helper behavior are unrelated." + "reason": "Python 3.13 import parsing problem in DebertaV2Model vs qwen3_omni_moe feature-extraction helper behavior." + }, + { + "left": "issue:43450", + "right": "issue:43526", + "accept": false, + "reason": "Different processor issues with no shared underlying code path." + }, + { + "left": "issue:37428", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but for different missing names and different modules." }, { "left": "issue:45200", "right": "issue:45381", "accept": false, - "reason": "Gemma 4 mm_token_type_ids defaulting and Qwen2.5-VL video position ids are different multimodal bugs." + "reason": "Different Gemma 4 text-finetuning token-type handling vs qwen2.5-vl video position-id bug." }, { "left": "issue:43404", "right": "issue:45200", "accept": false, - "reason": "Mistral3 weight tying and Gemma 4 text-only fine-tuning inputs are unrelated." + "reason": "Different model classes and failures: tied lm_head in Mistral3 vs Gemma 4 mm_token_type_ids defaults." }, { "left": "issue:39692", "right": "issue:44493", "accept": false, - "reason": "SigLIP2 example errors and widespread position-id warnings are not the same underlying issue." + "reason": "Distinct issues: SigLIP2 example/model mismatch vs generic position-id warnings." }, { "left": "issue:43335", "right": "issue:43653", "accept": false, - "reason": "SwitchTransformers sparse-layer creation and BigBirdTokenizer special-token registration are unrelated." - }, - { - "left": "issue:44448", - "right": "issue:44779", - "accept": false, - "reason": "Pegasus generation differences and Deepseek tokenizer regressions are distinct." + "reason": "Different bugs: sparse layer creation in SwitchTransformers vs BigBirdTokenizer special-token decode issue." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The main duplicate call is plausible, but the summary is not fully grounded: issue:45310 is not open in the packet (it is closed), so the canonical-issue rationale overstates the evidence. Please revise any status claims and keep the duplicate verdict limited to what the packet supports." + "accept": true, + "feedback": "Grounded and conservative. The only accepted duplicate pair is well-supported by identical Qwen3.5 MoE from_pretrained symptoms in issue:45310 and issue:45478, and the other soft-edge rejections stay on the safe side by distinguishing different model/component failure modes." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -6843,7 +6866,6 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-364", "nodes": [ - "issue:25251", "issue:29127", "issue:29942", "issue:32090", @@ -6878,6 +6900,7 @@ "issue:44493", "issue:44554", "issue:44568", + "issue:44779", "issue:44857", "issue:44977", "issue:44987", @@ -6895,278 +6918,278 @@ "soft_pairs": [ "issue:44977|issue:45229", "issue:29127|issue:32090", - "issue:44292|issue:45520", + "issue:44448|issue:44779", "issue:38175|issue:44493", + "issue:44292|issue:45520", "issue:43335|issue:43504", "issue:43335|issue:45083", "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:44448|issue:44977", "issue:43723|issue:44554", "issue:43335|issue:44112", "issue:34689|issue:35141", - "issue:43606|issue:44857", "issue:44857|issue:45072", + "issue:43606|issue:44857", "issue:45372|issue:45406", - "issue:43873|issue:44493", "issue:43066|issue:44462", + "issue:43873|issue:44493", "issue:43873|issue:45229", "issue:43278|issue:43329", "issue:43701|issue:45083", "issue:29942|issue:35141", "issue:43550|issue:45083", "issue:41720|issue:45229", - "issue:37428|issue:44987", "issue:32090|issue:33357", "issue:43976|issue:45440", + "issue:37428|issue:44987", "issue:44451|issue:45520", "issue:45372|issue:45459", "issue:44448|issue:44466", - "issue:43065|issue:45412", "issue:43065|issue:43994", - "issue:25251|issue:32090" + "issue:43065|issue:45412" ], - "prepared_review_unit_hash": "c729571a62cc493c123410759a22d1ffce878ede92ddd5ec3b4a66b72674cd2f", + "prepared_review_unit_hash": "ae007732c4b90060721950d68784da7ee2e627f9dde439f8f486bae921c83b5a", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 35, - "serialized_chars": 38319, - "estimated_input_tokens": 9580, - "estimated_eval_tokens": 19416 + "serialized_chars": 38314, + "estimated_input_tokens": 9579, + "estimated_eval_tokens": 19414 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "All suggested soft edges look like false positives: they share broad language (model loading, attention, quantization, CI, etc.) but describe different failures, code paths, or model families. No duplicate cluster is strong enough to merge.", - "confidence": 0.93, - "canonical_issue_reason": "No single canonical issue emerges; the candidate pairs are mostly unrelated bugs with only superficial subsystem overlap.", + "summary": "The soft pairs are mostly broad semantic neighbors, not true duplicates: they cover unrelated model families, tokenizer regressions, device/offload bugs, and CI/test failures. I found no pair that clearly shares the same underlying bug or change.", + "confidence": 0.84, + "canonical_issue_reason": "No strong duplicate anchor exists because the cluster is heterogeneous; if a representative must be chosen, issue 45229 is the most detailed and active, but it is not a good dedupe hub for the rest.", "canonical_pr_reason": null, - "best_issue_reason": "No issue is a safe global representative because the set is heterogeneous across distinct models, loaders, and runtime errors.", + "best_issue_reason": "Issue 45229 is the best representative only in the sense that it is active and well-described; it is not semantically close enough to the others to serve as a true canonical duplicate.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44977", "right": "issue:45229", "accept": false, - "reason": "Both are model/inference problems, but one is flash-attention generation corruption and the other is multi-GPU CUDA OOM; different failure modes and fixes." + "reason": "Different models and failure modes: Qwen3.5 flash-attention generation vs Gemma4 multi-GPU OOM." }, { "left": "issue:29127", "right": "issue:32090", "accept": false, - "reason": "One is a LayoutLMv3 validation/error-message issue, the other is a Trainer broadcast TypeError; unrelated code paths." + "reason": "Different code paths: LayoutLMv3 box validation messaging vs Trainer GPU broadcast NoneType error." }, { - "left": "issue:44292", - "right": "issue:45520", + "left": "issue:44448", + "right": "issue:44779", "accept": false, - "reason": "Qwen-3 NVFP4 runtime failure and a Python 3.13 flash_attn import KeyError are distinct issues with different causes." + "reason": "Both are v5 behavior regressions, but they affect different tokenization/generation flows and different models." }, { "left": "issue:38175", "right": "issue:44493", "accept": false, - "reason": "SigLIP2 zero probabilities and unexpected position-id keys are separate model-behavior bugs with no clear shared root cause." + "reason": "Different symptoms: zero probabilities in SigLIP2 vs unexpected position-id key serialization/deserialization issue." + }, + { + "left": "issue:44292", + "right": "issue:45520", + "accept": false, + "reason": "Different root causes: NVFP4 model runtime error vs Python 3.13 flash_attn import-key error." }, { "left": "issue:43335", "right": "issue:43504", "accept": false, - "reason": "A SwitchTransformers config edge case and a BEiT pretrained-load legacy-field issue are unrelated." + "reason": "Different subsystems and bugs: SwitchTransformers sparse-layer config vs BEiT pretrained-load legacy field." }, { "left": "issue:43335", "right": "issue:45083", "accept": false, - "reason": "Sparse-layer construction in SwitchTransformers and a qwen3_omni_moe feature-length helper bug are different code paths." + "reason": "Different models and logic: SwitchTransformers config bug vs qwen3_omni_moe feature-length helper bug." }, { "left": "issue:43550", "right": "issue:44112", "accept": false, - "reason": "torch.compile/SDPA failure in Bamba is unrelated to a GraniteSpeech CI test stale-device override." + "reason": "Unrelated: Bamba SDPA/torch.compile failure vs GraniteSpeech stale CI test." }, { "left": "issue:43335", "right": "issue:45072", "accept": false, - "reason": "SwitchTransformers sparse-layer config and dtype mismatch CI failures in other models are not the same bug." + "reason": "Different issue types: SwitchTransformers layer construction vs dtype mismatch in inference wrappers." }, { "left": "issue:44568", "right": "issue:45491", "accept": false, - "reason": "Tokenizer special-token behavior and Gemma3 NaN embeddings under sliding-window attention are unrelated." + "reason": "Different regressions: tokenizer special-token handling vs Gemma3 NaN embeddings from sliding-window attention." }, { "left": "issue:43720", "right": "issue:44112", "accept": false, - "reason": "Packed-weight loading failure in BitNet does not match a GraniteSpeech CI device-override test issue." + "reason": "Different code paths: packed-weight unpacking during accelerate loading vs GraniteSpeech CI device-override test." }, { - "left": "issue:44448", - "right": "issue:44977", + "left": "issue:40444", + "right": "issue:43526", "accept": false, - "reason": "Pegasus v4/v5 output drift is a model regression, while the other issue is Qwen3.5 flash-attention generation failure." + "reason": "Different processors and failures: Qwen2.5-VL iterable dataset multi-image finetuning vs BEiT Fast label reduction bug." }, { - "left": "issue:40444", - "right": "issue:43526", + "left": "issue:44448", + "right": "issue:44977", "accept": false, - "reason": "Qwen2.5-VL multi-image iterable dataset finetuning and BEiT reduce_labels returning one label are different preprocessing bugs." + "reason": "Different models and symptoms: Pegasus v4/v5 output difference vs Qwen3.5 flash-attention generation failure." }, { "left": "issue:43723", "right": "issue:44554", "accept": false, - "reason": "AutoTokenizer loading in v5 and MPS attention correctness with mismatched head dims are unrelated." + "reason": "Tokenizer loading regression vs MPS attention correctness issue; no shared bug path." }, { "left": "issue:43335", "right": "issue:44112", "accept": false, - "reason": "SwitchTransformers sparse-layer config and GraniteSpeech CI test failure are unrelated." + "reason": "Different subsystems: SwitchTransformers config construction vs GraniteSpeech CI test flake." }, { "left": "issue:34689", "right": "issue:35141", "accept": false, - "reason": "Llama 3.2 vision model loading breakage and embedding reinitialization after resizing are distinct issues." + "reason": "Both involve loading/init behavior, but one is Llama 3.2 Vision model loading and the other is embedding reinitialization after resize." }, { - "left": "issue:43606", - "right": "issue:44857", + "left": "issue:44857", + "right": "issue:45072", "accept": false, - "reason": "CPU offload device mismatch and float16 AMP CUDA crash are different runtime/device problems." + "reason": "Different failures: CUDA AMP crash in LwDetrImageLoss vs bfloat16 dtype mismatch in model wrappers." }, { - "left": "issue:44857", - "right": "issue:45072", + "left": "issue:43606", + "right": "issue:44857", "accept": false, - "reason": "LwDetrImageLoss AMP crash and SwitchTransformers/TimmWrapperModel bfloat16 dtype mismatch are separate failures." + "reason": "Different runtime problems: CPU offload device mismatch vs float16 AMP CUDA crash." }, { "left": "issue:45372", "right": "issue:45406", "accept": false, - "reason": "Gemma4 processor loading blocked by a mistral_common import error is not the same as Gemma4Processor missing _tokenizer in serve." + "reason": "Both touch Gemma4, but one is a dependency import error and the other is a missing processor attribute at serve time." }, { - "left": "issue:43873", - "right": "issue:44493", + "left": "issue:43066", + "right": "issue:44462", "accept": false, - "reason": "Quantization offloading behavior and unexpected position-id keys are not the same underlying bug." + "reason": "Tokenizer-related, but one is decoder-type selection in v5 and the other is AutoTokenizer ignoring repository tokenizer.json." }, { - "left": "issue:43066", - "right": "issue:44462", + "left": "issue:43873", + "right": "issue:44493", "accept": false, - "reason": "Wrong tokenizer decoder type in v5 and AutoTokenizer ignoring tokenizer.json are related to tokenizers broadly but describe different failures." + "reason": "Unrelated: quantization/offload behavior vs unexpected position-id key warnings." }, { "left": "issue:43873", "right": "issue:45229", "accept": false, - "reason": "Quantization offloading and Gemma4 multi-GPU OOM are both memory-related at a high level, but not the same concrete bug." + "reason": "Different failure modes: offloading with quantization vs Gemma4 multi-GPU OOM." }, { "left": "issue:43278", "right": "issue:43329", "accept": false, - "reason": "Embedding dtype drift in evaluate and a multimodal token-count helper using undefined video variables are unrelated." + "reason": "Different model internals: embedding dtype drift vs undefined video-token helper variables." }, { "left": "issue:43701", "right": "issue:45083", "accept": false, - "reason": "resume_from_checkpoint key mismatch and qwen3_omni_moe feature-length behavior are different parts of the codebase." + "reason": "Resume checkpoint key mismatch is unrelated to qwen3_omni_moe token-length helper behavior." }, { "left": "issue:29942", "right": "issue:35141", "accept": false, - "reason": "Flash Attention 2 test failures and embedding reinitialization after resize are unrelated." + "reason": "Flash Attention 2 test failures are not the same bug as embedding reinitialization after resize." }, { "left": "issue:43550", "right": "issue:45083", "accept": false, - "reason": "Bamba SDPA compile failure and qwen3_omni_moe feature-length helper behavior are distinct bugs." + "reason": "Different code paths and symptoms: Bamba compile/SDPA vs qwen3_omni_moe helper bug." }, { "left": "issue:41720", "right": "issue:45229", "accept": false, - "reason": "Qwen3 auto device mapping cudaErrorAssert and Gemma4 multi-GPU OOM are different runtime failures." - }, - { - "left": "issue:37428", - "right": "issue:44987", - "accept": false, - "reason": "Flash attention utility import breakage and loading physical-intelligence/fast are not the same issue." + "reason": "Different runtime failures: Qwen3 auto-device cudaErrorAssert vs Gemma4 multi-GPU OOM." }, { "left": "issue:32090", "right": "issue:33357", "accept": false, - "reason": "Trainer _gpu_broadcast_one NoneType error and MacOS bus error with community CLIP are unrelated." + "reason": "Different environments and bugs: Trainer broadcast NoneType vs MacOS bus error on community CLIP." }, { "left": "issue:43976", "right": "issue:45440", "accept": false, - "reason": "Python version compatibility issue and DeepseekV3MoE divergence from remote implementation are different classes of bugs." + "reason": "Python version compatibility issue is unrelated to DeepSeekV3MoE implementation divergence." + }, + { + "left": "issue:37428", + "right": "issue:44987", + "accept": false, + "reason": "Import error for flash-attention utility is not the same as loading a specific model repo on transformers>=5.1.0." }, { "left": "issue:44451", "right": "issue:45520", "accept": false, - "reason": "Cannot load a specific tokenizer/model and flash_attn KeyError on Python 3.13 are unrelated." + "reason": "Different cause: model loading failure for ScandiBERT vs flash_attn import lookup error on Python 3.13." }, { "left": "issue:45372", "right": "issue:45459", "accept": false, - "reason": "Gemma4 processor import problem and tokenizer error masking when protobuf is absent are different error paths." + "reason": "Gemma4 processor import failure and tokenizer error-masking via protobuf handling are different bugs." }, { "left": "issue:44448", "right": "issue:44466", "accept": false, - "reason": "Pegasus output regression and lm_head.weight serialization inconsistency are distinct v4/v5 issues." - }, - { - "left": "issue:43065", - "right": "issue:45412", - "accept": false, - "reason": "Dummy Conv2d in Sam3PixelDecoder and RT-DETR memory not being released are unrelated model implementation issues." + "reason": "Both concern v5 behavior changes, but one is generation output drift and the other is lm_head.weight serialization inconsistency." }, { "left": "issue:43065", "right": "issue:43994", "accept": false, - "reason": "Sam3PixelDecoder dummy Conv2d and SigLIP2 nonsensical outputs are not the same bug." + "reason": "Different model issues: dummy Conv2d in Sam3PixelDecoder vs nonsensical SigLIP2 outputs." }, { - "left": "issue:25251", - "right": "issue:32090", + "left": "issue:43065", + "right": "issue:45412", "accept": false, - "reason": "Pipeline top_k nesting change and Trainer GPU broadcast TypeError are unrelated." + "reason": "Different symptoms and code paths: dummy module in Sam3PixelDecoder vs RT-DETR memory not being released." } ] }, @@ -7190,6 +7213,7 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-364", "nodes": [ + "issue:25251", "issue:29127", "issue:29942", "issue:30064", @@ -7197,7 +7221,6 @@ "issue:33290", "issue:33357", "issue:34567", - "issue:35141", "issue:36010", "issue:39692", "issue:42915", @@ -7240,15 +7263,16 @@ "issue:45491" ], "soft_pairs": [ + "issue:25251|issue:32090", "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:44871|issue:44977", - "issue:43994|issue:45412", + "issue:44361|issue:45198", "issue:43421|issue:45491", + "issue:43994|issue:45412", + "issue:44871|issue:44977", "issue:42915|issue:42994", "issue:44206|issue:44479", "issue:43756|issue:45440", @@ -7263,229 +7287,228 @@ "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:43334|issue:43531", "issue:43262|issue:44206", "issue:44977|issue:45245", "issue:29942|issue:36010", - "issue:29127|issue:33357", - "issue:29127|issue:35141" + "issue:29127|issue:33357" ], - "prepared_review_unit_hash": "2bfec5282f4862bf48952da2cb856cb58505e8bf595199b75bb75aab0a579968", + "prepared_review_unit_hash": "6451945188bb0565265422141523e9ddb811e28861da1368e54c722df9856782", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 37065, - "estimated_input_tokens": 9267, - "estimated_eval_tokens": 18790 + "serialized_chars": 37021, + "estimated_input_tokens": 9256, + "estimated_eval_tokens": 18768 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous: the soft-similarity pairs share broad themes (tokenizers, Qwen, vision, trainer, quantization) but not the same concrete code-path bug. I would not merge any of the proposed soft pairs as duplicates.", - "confidence": 0.91, - "canonical_issue_reason": "issue:45198 is the most self-contained, concrete bug report in the set (clear failure mode plus save_pretrained/tokenization reproduction), so it is the best single representative if one must be chosen, though it does not subsume the rest.", + "summary": "This cluster is highly heterogeneous: it mixes unrelated bugs across tokenizers, processors, training, quantization, model loading, and inference. None of the soft pairs look like the same underlying issue or change, so there is no clean canonical duplicate target.", + "confidence": 0.97, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:45198 is a strong issue to anchor triage because it describes a specific, reproducible runtime failure with a narrow surface area and clear user impact.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:25251", + "right": "issue:32090", + "accept": false, + "reason": "Different failure modes: pipeline output shape change vs Trainer broadcast TypeError." + }, { "left": "issue:44186", "right": "issue:45198", "accept": false, - "reason": "Both are tokenizer failures, but they hit different models and code paths: LayoutLMv2 NER/padding vs Wav2Vec2 save_pretrained/tokenization." + "reason": "Both are tokenizer-related, but they affect different models and different bugs (LayoutLMv2 NER/padding crash vs Wav2Vec2 save/tokenization failure)." }, { "left": "issue:43278", "right": "issue:43381", "accept": false, - "reason": "One is a dtype change between train/eval, the other is a gradient-checkpointing restriction in eval mode; related context, not the same bug." + "reason": "One is a dtype mismatch between train/eval, the other is checkpointing disallowed in eval mode; not the same code path." }, { "left": "issue:43653", "right": "issue:44743", "accept": false, - "reason": "BigBird tokenizer special-token registration and Qwen3-MoE recurrent-state resetting are unrelated subsystems and failure modes." + "reason": "Tokenizer special-token registration vs Qwen3.5 recurrent-state reset in cached generation are unrelated." }, { "left": "issue:44265", "right": "issue:45198", "accept": false, - "reason": "torch.export with torch_compilable_check is a model export problem; 45198 is a tokenizer/save_pretrained issue." + "reason": "torch.export failure from torch_compilable_check is unrelated to Wav2Vec2 save_pretrained/tokenization errors." }, { - "left": "issue:44361", - "right": "issue:45198", + "left": "issue:43257", + "right": "issue:43278", "accept": false, - "reason": "Different tokenizer/model families and different failures: MLuke task AttributeError vs Wav2Vec2 save/tokenization failure." + "reason": "Qwen3 MoE weight conversion/loading with accelerate+deepspeed is unrelated to BF16-to-FP32 evaluation dtype drift." }, { - "left": "issue:43257", - "right": "issue:43278", + "left": "issue:44361", + "right": "issue:45198", "accept": false, - "reason": "Both involve Qwen/precision-adjacent complaints, but one is MoE weight conversion under accelerate+deepspeed and the other is eval-time dtype drift." + "reason": "Different tokenizer/model failures with no shared concrete bug." }, { - "left": "issue:44871", - "right": "issue:44977", + "left": "issue:43421", + "right": "issue:45491", "accept": false, - "reason": "Gemma EOS config mismatch and Qwen3.5 flash-attention generation issues are separate model-specific bugs." + "reason": "Runtime post-processor updates for special tokens are unrelated to Gemma3 NaN embeddings in mixed-length batching." }, { "left": "issue:43994", "right": "issue:45412", "accept": false, - "reason": "SigLIP2 inference nonsense and RT-DETR memory not released are unrelated symptoms and code paths." + "reason": "SigLIP2 incorrect outputs and RT-DETR memory release are distinct issues in different subsystems." }, { - "left": "issue:43421", - "right": "issue:45491", + "left": "issue:44871", + "right": "issue:44977", "accept": false, - "reason": "Runtime special-token post-processor updates and Gemma3 NaN embeddings from sliding-window attention are different tokenizer/model bugs." + "reason": "Gemma-3 eos_token_id inconsistency is not the same as Qwen3.5 flash-attention generation failure." }, { "left": "issue:42915", "right": "issue:42994", "accept": false, - "reason": "Both mention quantization, but one is Qwen3Moe with FineGrainedFP8Config and the other is quantized model saving; not the same concrete defect." + "reason": "Both mention quantization, but one is FP8 config failure and the other is quantized model saving; different problems." }, { "left": "issue:44206", "right": "issue:44479", "accept": false, - "reason": "Feature extractor unsupported-arg crash and Qwen VL video regression are distinct input/architecture problems." + "reason": "Audio feature extractor center-arg crash is unrelated to Qwen VL video-input regression." }, { "left": "issue:43756", "right": "issue:45440", "accept": false, - "reason": "Both are model implementation divergences, but for different models and different missing behaviors; not mergeable as one fix." + "reason": "Smollm3 RoPE layer mismatch and DeepseekV3MoE divergence are different model-specific implementation bugs." }, { "left": "issue:43504", "right": "issue:44857", "accept": false, - "reason": "Beit pretrained-load legacy field handling and LwDetr AMP/CUDA loss crash are unrelated." + "reason": "Beit pretrained-load legacy field issue is unrelated to LwDetrImageLoss AMP/CUDA crash." }, { "left": "issue:44610", "right": "issue:45412", "accept": false, - "reason": "Processor image-size mismatch and RT-DETR memory retention are different bugs with no shared code-path." + "reason": "Processor image-size mismatch and RT-DETR memory leak are unrelated." }, { "left": "issue:29942", "right": "issue:30064", "accept": false, - "reason": "Flash Attention 2 test failures and void segmentation-map processing are separate CI/runtime issues." + "reason": "Failing Flash Attention 2 tests and void segmentation map processing are separate regressions." }, { "left": "issue:33357", "right": "issue:39692", "accept": false, - "reason": "Both involve vision-language models, but one is a MacOS bus error on CLIP and the other is a SigLIP2 doc example with model/processor and quantization errors." + "reason": "MacOS CLIP bus error and SigLIP2 docs example errors are different model and reproduction contexts." }, { "left": "issue:45103", "right": "issue:45372", "accept": false, - "reason": "auto_docstring annotation handling and Gemma 4 processor loading/import errors are unrelated packaging/runtime problems." + "reason": "auto_docstring annotation handling and ReasoningEffort import failure are unrelated." }, { "left": "issue:32090", "right": "issue:33290", "accept": false, - "reason": "Trainer broadcast NoneType and deepspeed+adafactor OOM are different training failures." + "reason": "Trainer NoneType broadcast error and deepspeed Adafactor OOM are different training failures." }, { "left": "issue:29127", "right": "issue:34567", "accept": false, - "reason": "LayoutLMv3 error messaging and TrainerState token-count tracking are unrelated issues." + "reason": "LayoutLMv3 error-message clarity and TrainerState token counting bug do not share a concrete code path." }, { "left": "issue:45072", "right": "issue:45229", "accept": false, - "reason": "dtype mismatch in bfloat16 inference and multi-GPU CUDA OOM are not the same bug." + "reason": "bf16 dtype mismatch in inference and Gemma4 multi-GPU OOM are unrelated." }, { "left": "issue:43824", "right": "issue:44279", "accept": false, - "reason": "Missing Qwen2.5-VL import and a generic transformers dependency issue are too broad to treat as the same failure." + "reason": "A specific import error for Qwen2.5-VL is not the same as a general dependency issue." }, { "left": "issue:43381", "right": "issue:43701", "accept": false, - "reason": "Gradient checkpointing in eval mode and resume-from-checkpoint key mismatch do not share the same root cause." + "reason": "Eval-mode checkpointing restriction and resume_from_checkpoint key mismatch are different checkpoint-related bugs." }, { "left": "issue:43653", "right": "issue:44442", "accept": false, - "reason": "BigBird tokenizer special-token registration and AutoTokenizer failing to load FastSpeech2ConformerTokenizer are different tokenizer-loading bugs." - }, - { - "left": "issue:43334", - "right": "issue:43531", - "accept": false, - "reason": "Qwen3-VL pad_token_id loading failure and Qwen3-MoE sliding-window behavior are separate model/config issues." + "reason": "BigBird special-token decode issue and AutoTokenizer loading FastSpeech2Conformer are unrelated tokenizer bugs." }, { "left": "issue:43526", "right": "issue:44186", "accept": false, - "reason": "BeitImageProcessorFast label reduction and LayoutLMv2 tokenizer crashes are unrelated preprocessing bugs." + "reason": "BeitImageProcessorFast label reduction and LayoutLMv2Tokenizer NER/padding crash affect different preprocessing paths." }, { "left": "issue:44464", "right": "issue:45491", "accept": false, - "reason": "Compiled-forward chunked generation inconsistency and Gemma3 NaN embeddings from mixed-length batching are different generation paths." + "reason": "Compiled-forward chunked generation inconsistency is unrelated to Gemma3 NaN embeddings." + }, + { + "left": "issue:43334", + "right": "issue:43531", + "accept": false, + "reason": "Qwen3-VL pad_token_id import failure and Qwen3-MoE sliding_window behavior are separate model issues." }, { "left": "issue:43262", "right": "issue:44206", "accept": false, - "reason": "Audio chat-template sampling-rate defaulting and unsupported center-arg crash are unrelated processor bugs." + "reason": "Audio chat-template sample-rate default and LasrFeatureExtractor center-arg crash are different bugs." }, { "left": "issue:44977", "right": "issue:45245", "accept": false, - "reason": "Qwen3.5 flash-attention generation and category-count overflow are entirely different failures." + "reason": "Qwen3.5 flash-attention generation failure is unrelated to the categories-cardinality runtime error." }, { "left": "issue:29942", "right": "issue:36010", "accept": false, - "reason": "Flash Attention test failures and GenerationMixin import failure are separate issues." + "reason": "Flash Attention 2 test failures and GenerationMixin import breakage are unrelated." }, { "left": "issue:29127", "right": "issue:33357", "accept": false, - "reason": "LayoutLMv3 error messaging and CLIP MacOS bus error are unrelated." - }, - { - "left": "issue:29127", - "right": "issue:35141", - "accept": false, - "reason": "LayoutLMv3 box-validation messaging and embedding reinitialization after resize are different code paths and symptoms." + "reason": "LayoutLMv3 err_handle messaging and CLIP MacOS bus error are unrelated issues." } ] }, @@ -7509,6 +7532,7 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-364", "nodes": [ + "issue:29127", "issue:29942", "issue:33290", "issue:33666", @@ -7516,7 +7540,6 @@ "issue:34689", "issue:35141", "issue:39692", - "issue:41628", "issue:42915", "issue:43232", "issue:43316", @@ -7530,6 +7553,7 @@ "issue:43646", "issue:43673", "issue:43701", + "issue:43720", "issue:43723", "issue:43761", "issue:43827", @@ -7553,17 +7577,17 @@ "issue:45072", "issue:45216", "issue:45290", - "issue:45372", "issue:45440", "issue:45479", "issue:45491" ], "soft_pairs": [ + "issue:29127|issue:35141", "issue:43761|issue:44079", - "issue:44442|issue:45072", "issue:42915|issue:43931", - "issue:43526|issue:44857", + "issue:44442|issue:45072", "issue:43232|issue:43388", + "issue:43526|issue:44857", "issue:43526|issue:44610", "issue:44479|issue:45290", "issue:44743|issue:44811", @@ -7573,231 +7597,237 @@ "issue:43931|issue:45216", "issue:43646|issue:44077", "issue:43761|issue:44188", - "issue:43316|issue:43525", - "issue:43575|issue:44387", - "issue:43994|issue:44206", "issue:44964|issue:45440", + "issue:43994|issue:44206", + "issue:43575|issue:44387", + "issue:43316|issue:43525", "issue:29942|issue:33666", "issue:39692|issue:45491", "issue:44442|issue:44898", - "issue:33290|issue:34567", "issue:43673|issue:44743", + "issue:33290|issue:34567", "issue:34689|issue:43611", - "issue:43381|issue:43873", "issue:43493|issue:43827", - "issue:44077|issue:44479", + "issue:43381|issue:43873", "issue:33290|issue:35141", + "issue:44077|issue:44479", "issue:33290|issue:34689", - "issue:41628|issue:45372" + "issue:43720|issue:44442" ], - "prepared_review_unit_hash": "5ea9d4784489ce40c2dbdf05b0015ee7a815f5979c09f5a438c9eaf9d68a8ff0", + "prepared_review_unit_hash": "6d16e79fa3443f3c5b234b9c1cc168d09fcab5ff17c18feed051694db1558af7", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 36717, - "estimated_input_tokens": 9180, - "estimated_eval_tokens": 18616 + "soft_pair_count": 31, + "serialized_chars": 37033, + "estimated_input_tokens": 9259, + "estimated_eval_tokens": 18774 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: the soft pairs mostly connect unrelated issues that share only broad keywords (loading, OOM, config, cache, etc.) rather than the same bug. I would not merge any of the candidate pairs.", - "confidence": 0.98, - "canonical_issue_reason": null, + "summary": "No clear duplicate cluster here: the items span unrelated issues across model loading, training, generation, quantization, docs, and processor behavior. All soft pairs should be rejected.", + "confidence": 0.94, + "canonical_issue_reason": "No single canonical issue is supported by the titles/descriptions; the set is heterogeneous and does not center on one underlying bug.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "issue:43873 is the strongest representative only by activity/ref count and broad impact, but it is not semantically a true canonical for this cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43761", - "right": "issue:44079", + "left": "issue:29127", + "right": "issue:35141", "accept": false, - "reason": "Different bugs: CLIPVisionModel hidden_states regression vs ModelOutput key assignment behavior." + "reason": "Different problems: LayoutLMv3 box-info error handling vs token embedding reinit after resize." }, { - "left": "issue:44442", - "right": "issue:45072", + "left": "issue:43761", + "right": "issue:44079", "accept": false, - "reason": "Tokenizer loading failure and bfloat16 dtype mismatch are unrelated problems." + "reason": "Different code paths: CLIPVisionModel hidden_states regression vs ModelOutput key assignment bug." }, { "left": "issue:42915", "right": "issue:43931", "accept": false, - "reason": "Different models and failures: Qwen3Moe FP8 config vs Qwen3-VL weight shape mismatch." + "reason": "Both are loading-related, but one is FineGrainedFP8 with Qwen3Moe and the other is a shape-mismatch on Qwen3-VL; not the same bug." }, { - "left": "issue:43526", - "right": "issue:44857", + "left": "issue:44442", + "right": "issue:45072", "accept": false, - "reason": "BeitImageProcessorFast label reduction bug is unrelated to LwDetrImageLoss AMP crash." + "reason": "Tokenizer loading failure vs bfloat16 dtype mismatch in CI; unrelated failures." }, { "left": "issue:43232", "right": "issue:43388", "accept": false, - "reason": "Generation kwargs handling and metric label gathering are distinct code paths." + "reason": "Generation kwargs update after sync_gpus is unrelated to gather_for_metrics label truncation." + }, + { + "left": "issue:43526", + "right": "issue:44857", + "accept": false, + "reason": "BEiT image processor label reduction bug vs LwDetrImageLoss AMP crash; different components and symptoms." }, { "left": "issue:43526", "right": "issue:44610", "accept": false, - "reason": "Different processor bugs: label reduction vs image size mismatch." + "reason": "ImageProcessor label handling vs OmDet-Turbo processor size mismatch; not the same issue." }, { "left": "issue:44479", "right": "issue:45290", "accept": false, - "reason": "Video input regression and chat template tool-call crash are separate issues." + "reason": "Video input regression for Qwen VL models vs chat template crash on tool-call assistant messages." }, { "left": "issue:44743", "right": "issue:44811", "accept": false, - "reason": "Qwen recurrent-state reset and Whisper batch_decode token skipping are unrelated." + "reason": "Qwen3.5 recurrent state reset bug vs Whisper batch_decode skip_special_tokens handling." }, { "left": "issue:43723", "right": "issue:45479", "accept": false, - "reason": "Tokenizer loading error and sequence-classification zero-loss bug do not match." + "reason": "Tokenizer loading issue vs sequence-classification zero-loss bug; unrelated." }, { "left": "issue:45003", "right": "issue:45440", "accept": false, - "reason": "sys.modules access bug and DeepseekV3MoE divergence are unrelated." + "reason": "Unsafe sys.modules access in modeling_utils vs DeepseekV3MoE implementation divergence; different layers and failures." }, { "left": "issue:43575", "right": "issue:43701", "accept": false, - "reason": "Tensor-parallel OOM and resume_from_checkpoint key mismatch are different failures." + "reason": "Tensor-parallel OOM when loading Qwen2-57B vs resume_from_checkpoint key mismatch; not the same underlying bug." }, { "left": "issue:43931", "right": "issue:45216", "accept": false, - "reason": "Both involve model loading/saving, but the concrete bugs differ: shape mismatch vs bad checkpoint save." + "reason": "Model load shape mismatch vs save_pretrained checkpoint regression; different directions of failure." }, { "left": "issue:43646", "right": "issue:44077", "accept": false, - "reason": "Custom model initialization regression is not the same as patchtsmixer post_init handling." + "reason": "Custom model initialization breakage vs patchtsmixer post_init API constraint; too different." }, { "left": "issue:43761", "right": "issue:44188", "accept": false, - "reason": "Hidden_states regression and attention-kernel divergence under compile are unrelated." + "reason": "CLIPVisionModel hidden_states regression vs attention-kernel divergence under torch.compile; unrelated." }, { - "left": "issue:43316", - "right": "issue:43525", + "left": "issue:44964", + "right": "issue:45440", "accept": false, - "reason": "Different config APIs for different models; not the same missing-attribute bug." + "reason": "Phi-4 multimodal loading error vs DeepSeekV3MoE implementation mismatch; different models and problems." }, { - "left": "issue:43575", - "right": "issue:44387", + "left": "issue:43994", + "right": "issue:44206", "accept": false, - "reason": "Both mention OOM, but one is tensor-parallel load OOM and the other is int4 reserved-memory growth." + "reason": "SigLIP2 nonsensical outputs vs LasrFeatureExtractor unsupported center arg; different regressions." }, { - "left": "issue:43994", - "right": "issue:44206", + "left": "issue:43575", + "right": "issue:44387", "accept": false, - "reason": "SigLIP2 bad outputs and LasrFeatureExtractor center-arg crash are separate issues." + "reason": "Qwen2 TP loading OOM vs increased reserved memory under int4 quantization; both memory-related but not the same concrete bug." }, { - "left": "issue:44964", - "right": "issue:45440", + "left": "issue:43316", + "right": "issue:43525", "accept": false, - "reason": "Phi-4 multimodal loading and DeepseekV3MoE divergence are unrelated." + "reason": "Gemma3TextConfig API discrepancy vs Llama4Config missing pad_token_id; distinct API surface issues." }, { "left": "issue:29942", "right": "issue:33666", "accept": false, - "reason": "Flash Attention 2 test failures and Qwen2-VL multi-GPU training are not the same bug." + "reason": "Flash Attention 2 test failures vs Qwen2-VL multi-GPU training; separate concerns." }, { "left": "issue:39692", "right": "issue:45491", "accept": false, - "reason": "Documentation/example errors are unrelated to Gemma3 NaN embeddings." + "reason": "SigLIP2 docs example errors vs Gemma3 NaN embeddings on mixed-length batches; docs vs runtime model bug." }, { "left": "issue:44442", "right": "issue:44898", "accept": false, - "reason": "Tokenizer load failure and Perceiver non-default-resolution failure are different problems." - }, - { - "left": "issue:33290", - "right": "issue:34567", - "accept": false, - "reason": "Adafactor/DeepSpeed OOM is unrelated to TrainerState token counter not updating." + "reason": "FastSpeech2ConformerTokenizer loading failure vs Perceiver non-default resolution failure; unrelated model families." }, { "left": "issue:43673", "right": "issue:44743", "accept": false, - "reason": "Both mention cache, but they concern different models and different cache/state logic." + "reason": "GenerationMixin cache missing in chunked_prefill vs recurrent state reset in modular_qwen3_5.py; different generation bugs." }, { - "left": "issue:34689", - "right": "issue:43611", + "left": "issue:33290", + "right": "issue:34567", "accept": false, - "reason": "Two separate model-loading regressions with different causes and versions." + "reason": "AdaFactor/deepspeed OOM vs TrainerState token counter not updating; not the same issue." }, { - "left": "issue:43381", - "right": "issue:43873", + "left": "issue:34689", + "right": "issue:43611", "accept": false, - "reason": "Gradient checkpointing eval-mode restriction is unrelated to quantization offloading." + "reason": "Llama 3.2 Vision loading breakage vs base_model_prefix loading breakage in v5.0.0; both loading-related but different failure mechanisms." }, { "left": "issue:43493", "right": "issue:43827", "accept": false, - "reason": "SigLIP2 JAX parity and docs still using pipeline() are unrelated." + "reason": "SigLIP2 HF/JAX divergence vs docs still referencing removed pipeline(); different categories." }, { - "left": "issue:44077", - "right": "issue:44479", + "left": "issue:43381", + "right": "issue:43873", "accept": false, - "reason": "patchtsmixer post_init handling and Qwen video-input regression do not share a code path." + "reason": "Gradient checkpointing eval-mode restriction vs quantization/offloading behavior; unrelated." }, { "left": "issue:33290", "right": "issue:35141", "accept": false, - "reason": "OOM in DeepSpeed AdaFactor and embedding reinit after resize are unrelated." + "reason": "AdaFactor/deepspeed OOM vs embedding reinitialization after resize; different bugs." + }, + { + "left": "issue:44077", + "right": "issue:44479", + "accept": false, + "reason": "patchtsmixer post_init policy vs Qwen VL video-input regression; unrelated." }, { "left": "issue:33290", "right": "issue:34689", "accept": false, - "reason": "Different failures: optimizer OOM vs model-loading breakage." + "reason": "AdaFactor OOM vs Llama 3.2 Vision loading regression; different symptoms and code paths." }, { - "left": "issue:41628", - "right": "issue:45372", + "left": "issue:43720", + "right": "issue:44442", "accept": false, - "reason": "Both are import errors, but for different symbols, modules, and processor-loading flows." + "reason": "Accelerate loading/unpacking packed weights in BitNet vs tokenizer loading failure; unrelated." } ] }, @@ -7830,6 +7860,7 @@ "issue:36010", "issue:36331", "issue:39692", + "issue:41628", "issue:41950", "issue:42175", "issue:43065", @@ -7844,7 +7875,6 @@ "issue:43643", "issue:43688", "issue:43716", - "issue:43720", "issue:43749", "issue:43761", "issue:43825", @@ -7855,7 +7885,6 @@ "issue:44162", "issue:44190", "issue:44206", - "issue:44442", "issue:44479", "issue:44496", "issue:44514", @@ -7868,17 +7897,18 @@ "issue:44913", "issue:45127", "issue:45137", - "issue:45200" + "issue:45200", + "issue:45372" ], "soft_pairs": [ - "issue:43720|issue:44442", + "issue:41628|issue:45372", "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:33290|issue:36331", "issue:44855|issue:45137", "issue:34689|issue:36010", "issue:43316|issue:44877", @@ -7889,192 +7919,192 @@ "issue:43404|issue:43441", "issue:16998|issue:22355", "issue:25251|issue:28282", - "issue:43316|issue:43716", "issue:43065|issue:44617", + "issue:43316|issue:43716", "issue:41950|issue:43475", "issue:43295|issue:43825", "issue:45127|issue:45200", "issue:43761|issue:44479", "issue:44077|issue:44534" ], - "prepared_review_unit_hash": "0f3b53f93390986d3682abf2eaa5de210c407358b4438bcb9354c12892273305", + "prepared_review_unit_hash": "d6ac68ca46ff1185ac2d81b4753d875ff5fc85eb03244cdcc061d98a84563e6d", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 25, - "serialized_chars": 35154, - "estimated_input_tokens": 8789, - "estimated_eval_tokens": 17834 + "serialized_chars": 35181, + "estimated_input_tokens": 8796, + "estimated_eval_tokens": 17848 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a heterogeneous grab-bag of unrelated Transformers issues; none of the soft pairs look like the same underlying bug/change closely enough to merge. The most representative issue is the high-activity core loading regression around FSDP CPU RAM efficient loading.", - "confidence": 0.34, - "canonical_issue_reason": "issue:43749 is the strongest anchor: it has the highest discussion/inbound activity and describes a core loading-path regression with broad impact and a clear bug report.", + "summary": "These soft pairs are mostly near-miss topic similarities, but they do not look like the same underlying bug or change. The cluster is heterogeneous across unrelated import, config, pipeline, model-loading, and training issues.", + "confidence": 0.97, + "canonical_issue_reason": "No single issue clearly subsumes the others; the items describe distinct failures across different models, APIs, and subsystems.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43749 is the best overall representative because it is a substantive, widely impactful runtime bug rather than a niche model-specific, docs, or API-edge case report.", + "best_issue_reason": "If forced to pick a representative, issue 34689 is a clear regression report with a concrete model/version and reproducible loading failure, but it still does not match the rest of the set.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43720", - "right": "issue:44442", + "left": "issue:41628", + "right": "issue:45372", "accept": false, - "reason": "Different subsystems and failures: BitNet packed-weight loading vs AutoTokenizer loading a specific tokenizer class." + "reason": "Both are import/loading failures, but one is about AutoImageProcessor exports and the other about mistral_common ReasoningEffort missing during Gemma 4 processor loading." }, { "left": "issue:43749", "right": "issue:43931", "accept": false, - "reason": "Both involve loading, but one is FSDP CPU-efficient loading and the other is a Qwen3-VL shape mismatch; not the same bug." + "reason": "Both concern loading, but one is broken FSDP CPU RAM efficient loading while the other is a Qwen3-VL shape-mismatch model load error." }, { "left": "issue:43643", "right": "issue:44913", "accept": false, - "reason": "Unrelated config issues: trust_remote_code missing fields vs GPTNeoX rotary_pct persistence on reload." + "reason": "Both involve config behavior on reload, but they are different bugs in different configs: trust_remote_code missing fields vs GPTNeoX rotary_pct reverting." }, { "left": "issue:43873", "right": "issue:44871", "accept": false, - "reason": "Quantization/offloading behavior is unrelated to Gemma-3 eos_token_id config inconsistency." + "reason": "Quantization/offloading behavior and Gemma eos_token_id inconsistency are unrelated bugs." }, { "left": "issue:39692", "right": "issue:44206", "accept": false, - "reason": "Docs example errors for SigLIP2 are not the same as LasrFeatureExtractor passing an unsupported argument." + "reason": "One is a SigLIP2 documentation example with multiple errors; the other is a feature extractor API regression passing an unsupported center arg." }, { "left": "issue:44190", "right": "issue:44496", "accept": false, - "reason": "Local dataset loading in a training script is unrelated to an unrecognized model/config loading failure." + "reason": "A local dataset loading failure in a training script is not the same as an unrecognized model/config.json model_type error." }, { - "left": "issue:33290", - "right": "issue:36331", + "left": "issue:43381", + "right": "issue:43856", "accept": false, - "reason": "OOM during Adafactor+DeepSpeed training is a different problem from a CustomTrainer signature mismatch." + "reason": "Gradient checkpointing in eval mode and Qwen3 MoE memory inefficiency are different training-time issues." }, { - "left": "issue:43381", - "right": "issue:43856", + "left": "issue:33290", + "right": "issue:36331", "accept": false, - "reason": "Gradient checkpointing in eval mode and Qwen3 MoE memory usage are different code paths and symptoms." + "reason": "OOM with Adafactor in DeepSpeed is unrelated to a CustomTrainer compute_loss signature mismatch." }, { "left": "issue:44855", "right": "issue:45137", "accept": false, - "reason": "Python 3.13 DeBERTa import parsing bug is unrelated to DeepSpeed ZeRO3 deque underflow." + "reason": "Python 3.13 Deberta import parsing error and DeepSpeed ZeRO3 deque underflow are different failure modes." }, { "left": "issue:34689", "right": "issue:36010", "accept": false, - "reason": "Llama 3.2 model loading regression is not the same as failing to import GenerationMixin from a moved module." + "reason": "A Llama 3.2 Vision loading regression is not the same as missing GenerationMixin import from transformers.generation." }, { "left": "issue:43316", "right": "issue:44877", "accept": false, - "reason": "Gemma3TextConfig API inconsistency is different from strict config rejection for granite_speech." + "reason": "Gemma3TextConfig API mismatch and strict config blocking granite_speech loading are different configuration issues." }, { "left": "issue:43540", "right": "issue:44514", "accept": false, - "reason": "Qwen3OmniMoe video-processing ValueError is unrelated to Qwen2_5_VL batched chat-template padding behavior." + "reason": "Qwen3OmniMoe video processing ValueError and batched Qwen2_5_VL apply_chat_template crash are distinct processor bugs." }, { "left": "issue:44162", "right": "issue:44792", "accept": false, - "reason": "ESM2 model breakage is not the same as a Janus test failure around image generation." + "reason": "ESM2 breaking scientific workflows and a janus test failure are unrelated." }, { "left": "issue:43388", "right": "issue:43688", "accept": false, - "reason": "Metric gathering drops tuple labels; auxiliary-loss normalization in OLMoE/GPT Oss is a different training-loss issue." + "reason": "A metrics-gathering last-batch label truncation bug is not the same as auxiliary loss normalization in OLMoE/GPT Oss." }, { "left": "issue:42175", "right": "issue:43316", "accept": false, - "reason": "Missing TensorFlow from a torch extra install has nothing to do with Gemma3TextConfig API discrepancies." + "reason": "Missing tensorflow in a torch extra install and a Gemma3TextConfig API discrepancy are unrelated." }, { "left": "issue:43404", "right": "issue:43441", "accept": false, - "reason": "These are different model bugs: lm_head tying in Mistral3 versus FlashAttention failure in Ministral-3." + "reason": "Weight tying in Mistral3ForConditionalGeneration and FlashAttention failure in Ministral-3 are different concrete code-path problems." }, { "left": "issue:16998", "right": "issue:22355", "accept": false, - "reason": "A model_max_length question for DeBERTa-V3 is unrelated to a missing transformers.onnx module import." + "reason": "Model_max_length questions for DeBERTa-V3 and a transformers.onnx import error have no shared underlying bug." }, { "left": "issue:25251", "right": "issue:28282", "accept": false, - "reason": "Pipeline top_k output-shape behavior is unrelated to an AutoModel import error requiring PyTorch." + "reason": "Pipeline top_k output nesting and missing PyTorch for AutoModel are unrelated." }, { - "left": "issue:43316", - "right": "issue:43716", + "left": "issue:43065", + "right": "issue:44617", "accept": false, - "reason": "Config API discrepancy and Mistral-3 image-preprocessor dtype mismatch are distinct issues." + "reason": "A dummy Conv2d in Sam3PixelDecoder and Sam3Video CUDA OOM are not the same bug." }, { - "left": "issue:43065", - "right": "issue:44617", + "left": "issue:43316", + "right": "issue:43716", "accept": false, - "reason": "A dummy Conv2d in Sam3PixelDecoder is not the same as CUDA OOM in Sam3Video." + "reason": "Gemma3TextConfig API differences and Mistral-3 image preprocessor dtype mismatch are different issues." }, { "left": "issue:41950", "right": "issue:43475", "accept": false, - "reason": "Video-classification pipeline processor lookup and missing fpn_position_embeddings are different Sam3/video bugs." + "reason": "Video-classification selecting image processors and Sam3VisionEncoderOutput missing fpn_position_embeddings are unrelated." }, { "left": "issue:43295", "right": "issue:43825", "accept": false, - "reason": "Custom processor.tokenizer regression is unrelated to an incorrect translation-task error message." + "reason": "A processor.tokenizer regression is not the same as an incorrect v5 translation error message." }, { "left": "issue:45127", "right": "issue:45200", "accept": false, - "reason": "LoRA merge collapse with extended vocab is a different model-training issue from Gemma 4 mm_token_type_ids defaults." + "reason": "LoRA merge collapse with tied embeddings and Gemma 4 mm_token_type_ids defaults are different Gemma-related problems." }, { "left": "issue:43761", "right": "issue:44479", "accept": false, - "reason": "CLIPVisionModel hidden_states regression is unrelated to Qwen video-input regressions." + "reason": "CLIPVisionModel hidden_states regression and Qwen video input regression are distinct model-specific bugs." }, { "left": "issue:44077", "right": "issue:44534", "accept": false, - "reason": "patchtsmixer post_init validation and non-persistent buffer initialization are different v5 behavior changes." + "reason": "patchtsmixer post_init policy and non-persistent buffer junk in v5 are unrelated framework issues." } ] }, @@ -8149,9 +8179,9 @@ ], "soft_pairs": [ "issue:43550|issue:43701", + "issue:43606|issue:43701", "issue:44734|issue:45406", "issue:43295|issue:44871", - "issue:43606|issue:43701", "issue:43452|issue:44496", "issue:43646|issue:44913", "issue:43906|issue:45290", @@ -8161,14 +8191,14 @@ "issue:43931|issue:44060", "issue:44521|issue:45381", "issue:43575|issue:43873", - "issue:44479|issue:44734", "issue:44792|issue:44977", - "issue:32090|issue:35141", + "issue:44479|issue:44734", "issue:44464|issue:44871", + "issue:32090|issue:35141", "issue:43525|issue:43531", "issue:44933|issue:45362", - "issue:43065|issue:44610", "issue:43827|issue:43994", + "issue:43065|issue:44610", "issue:44534|issue:44913", "issue:41720|issue:45245", "issue:44938|issue:45542", @@ -8176,196 +8206,196 @@ "issue:44190|issue:44964", "issue:43388|issue:43582" ], - "prepared_review_unit_hash": "7e46caf48aa19e73a253f83321bbcfb32838b198fd003a6b30db31a00cd1de64", + "prepared_review_unit_hash": "c10b4b5da3a80ffc019077a8e6188a71d96fe80ce40151d38f8c55d0ee3923e4", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 27, - "serialized_chars": 35604, - "estimated_input_tokens": 8901, - "estimated_eval_tokens": 18058 + "serialized_chars": 35606, + "estimated_input_tokens": 8902, + "estimated_eval_tokens": 18060 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is heterogeneous: the issues span unrelated models, regressions, and subsystems. None of the suggested soft pairs look like the same underlying bug or change, so all should be rejected.", - "confidence": 0.96, - "canonical_issue_reason": "No clear canonical issue: the items are distinct bug reports across different code paths, models, and failure modes, with no strong duplicate center.", + "summary": "This cluster is mostly a false-positive mix of unrelated issues spanning model loading, multimodal/chat-template regressions, config serialization, distributed training, and docs/environment failures. I do not see a single underlying bug or change tying these together, so all soft edges are rejected.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue fits this cluster: the items are heterogeneous and do not share a common root cause.", "canonical_pr_reason": null, - "best_issue_reason": "No single issue stands out as a global representative because the set is not a cohesive duplicate cluster.", + "best_issue_reason": "If forced to pick a representative issue, issue:44521 is the strongest candidate because it has the most cross-references and describes a broadly impactful multimodal chat-template bug, but it still does not represent the whole set.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43550", "right": "issue:43701", "accept": false, - "reason": "Different failures: torch.compile+SDPA crash vs resume_from_checkpoint key mismatch." + "reason": "Different failures: torch.compile SDPA on Bamba vs checkpoint key mismatch during resume." + }, + { + "left": "issue:43606", + "right": "issue:43701", + "accept": false, + "reason": "CPU offload device mismatch is unrelated to resume_from_checkpoint key handling." }, { "left": "issue:44734", "right": "issue:45406", "accept": false, - "reason": "Different areas: KV-cache continuation indexing vs Gemma4Processor missing _tokenizer." + "reason": "Both mention serve, but one is KV-cache indexing and the other is a missing processor attribute." }, { "left": "issue:43295", "right": "issue:44871", "accept": false, - "reason": "Both are regression/config issues, but one is processor/tokenizer API breakage and the other is eos_token_id mismatch." - }, - { - "left": "issue:43606", - "right": "issue:43701", - "accept": false, - "reason": "CPU offload device mismatch is unrelated to checkpoint resume key mismatch." + "reason": "Different regressions: processor/tokenizer API break vs eos_token_id config inconsistency." }, { "left": "issue:43452", "right": "issue:44496", "accept": false, - "reason": "gguf_file loading paths and unrecognized-model config errors are different model-loading problems." + "reason": "Both are loading problems, but one is gguf tokenizer/model loading and the other is unrecognized model config." }, { "left": "issue:43646", "right": "issue:44913", "accept": false, - "reason": "Custom initialization breakage is not the same as GPTNeoX rotary_pct not persisting on reload." + "reason": "Custom model init breakage and GPTNeoX config reload behavior are separate code paths." }, { "left": "issue:43906", "right": "issue:45290", "accept": false, - "reason": "An isolated reproduction of another issue is not the same as apply_chat_template crashing on tool-call messages." + "reason": "Isolated repro of an older issue is not the same bug as chat template crashes with tool calls." }, { "left": "issue:43761", "right": "issue:44242", "accept": false, - "reason": "CLIPVision hidden_states regression is unrelated to load balancing loss handling in MoE routing." + "reason": "CLIPVision hidden_states regression and missing load-balancing loss are unrelated model behaviors." }, { "left": "issue:43575", "right": "issue:43742", "accept": false, - "reason": "OOM on tensor-parallel load and key error on loading MobileLLM are different load-time problems." + "reason": "TP OOM on a large Qwen model is not the same as a MobileLLM key error." }, { "left": "issue:42617", "right": "issue:43575", "accept": false, - "reason": "3d_parallel.py failure is unrelated to Qwen2-57B tensor-parallel OOM." + "reason": "Both are distributed/model-parallel related, but one is a 3d_parallel.py runtime issue and the other is TP memory OOM." }, { "left": "issue:43931", "right": "issue:44060", "accept": false, - "reason": "Weight-shape mismatch loading Qwen3-VL differs from tied-weights warning on Qwen3-Next." + "reason": "Weight-shape mismatch in Qwen3-VL is unrelated to a tied-weights warning in Qwen3-Next." }, { "left": "issue:44521", "right": "issue:45381", "accept": false, - "reason": "All-zero assistant masks in chat templating is distinct from Qwen2.5-VL video vision_position_ids regression." + "reason": "Both are multimodal/Qwen-VL related, but assistant_mask generation and vision_position_ids are different bugs." }, { "left": "issue:43575", "right": "issue:43873", "accept": false, - "reason": "Tensor-parallel OOM and quantization offload behavior are not the same bug." - }, - { - "left": "issue:44479", - "right": "issue:44734", - "accept": false, - "reason": "Video-input regression for Qwen VL models is unrelated to serve KV-cache continuation indexing." + "reason": "TP loading OOM and offloading with quantization are different memory-management problems." }, { "left": "issue:44792", "right": "issue:44977", "accept": false, - "reason": "Failed janus image-generation test is not the same as Qwen3.5 flash-attention generation bug." + "reason": "Failed janus image-generation test and Qwen3.5 flash-attention generation issue are unrelated." }, { - "left": "issue:32090", - "right": "issue:35141", + "left": "issue:44479", + "right": "issue:44734", "accept": false, - "reason": "Trainer _gpu_broadcast_one NoneType error and embedding reinit on post_init are different bugs." + "reason": "Video-input regression and serve KV-cache continuation crash affect different paths and symptoms." }, { "left": "issue:44464", "right": "issue:44871", "accept": false, - "reason": "Chunked generation with compiled forward is unrelated to eos_token_id configuration inconsistency." + "reason": "Compiled-forward chunked generation inconsistency is unrelated to eos_token_id configuration drift." + }, + { + "left": "issue:32090", + "right": "issue:35141", + "accept": false, + "reason": "Trainer GPU broadcast TypeError and embedding reinitialization in post_init are different bugs." }, { "left": "issue:43525", "right": "issue:43531", "accept": false, - "reason": "Missing pad_token_id in Llama4Config and sliding_window issues in Qwen3-MoE are different model config bugs." + "reason": "Missing pad_token_id on Llama4Config and Qwen3-MoE sliding_window behavior are separate config/model issues." }, { "left": "issue:44933", "right": "issue:45362", "accept": false, - "reason": "Missing import from image_utils is unrelated to Qwen3.5 chat crashes." + "reason": "A missing import from image_utils is not the same as a Qwen3.5 chat-generation crash." }, { - "left": "issue:43065", - "right": "issue:44610", + "left": "issue:43827", + "right": "issue:43994", "accept": false, - "reason": "Dummy Conv2d in Sam3PixelDecoder and OmDet-Turbo input-size mismatch are different model-specific issues." + "reason": "Docs still referencing pipeline() is a documentation issue, not the SigLIP2 inference bug." }, { - "left": "issue:43827", - "right": "issue:43994", + "left": "issue:43065", + "right": "issue:44610", "accept": false, - "reason": "Docs still referencing pipeline() is unrelated to SigLIP2 nonsensical outputs." + "reason": "Sam3PixelDecoder dummy Conv2d and OmDet-Turbo input-size mismatch are unrelated vision-model issues." }, { "left": "issue:44534", "right": "issue:44913", "accept": false, - "reason": "Non-persistent buffer junk in v5 is not the same as rotary_pct resetting on reload." + "reason": "Non-persistent buffer corruption in v5 is not the same as GPTNeoX rotary_pct reload loss." }, { "left": "issue:41720", "right": "issue:45245", "accept": false, - "reason": "Qwen3 auto device mapping cuda assert and category-count limit error are unrelated runtime failures." + "reason": "Qwen3 auto-device-map cuda assert and category-cardinality RuntimeError are unrelated." }, { "left": "issue:44938", "right": "issue:45542", "accept": false, - "reason": "Python 3.14 load failure is not the same as missing TensorFlow backend when only tensorboard is installed." + "reason": "Python 3.14 import/load failure and missing TensorFlow backend due to tensorboard-only install are different environment issues." }, { "left": "issue:43650", "right": "issue:44220", "accept": false, - "reason": "ADD THE DATA is not a meaningful duplicate of _torch_extract_fbank_features()." + "reason": "'ADD THE DATA' is unrelated to _torch_extract_fbank_features." }, { "left": "issue:44190", "right": "issue:44964", "accept": false, - "reason": "Local dataset loading in an example script is unrelated to loading Phi-4-multimodal-instruct." + "reason": "Local dataset loading with a no-trainer script is unrelated to loading Phi-4 multimodal weights." }, { "left": "issue:43388", "right": "issue:43582", "accept": false, - "reason": "gather_for_metrics tuple-label truncation is unrelated to Apple Silicon warmup TypeError." + "reason": "Metric gathering label truncation and Apple Silicon caching allocator TypeError do not share a code path." } ] }, @@ -8411,6 +8441,7 @@ "issue:43976", "issue:44016", "issue:44077", + "issue:44079", "issue:44112", "issue:44186", "issue:44190", @@ -8426,7 +8457,6 @@ "issue:44792", "issue:44855", "issue:44871", - "issue:44877", "issue:44913", "issue:45003", "issue:45127", @@ -8443,9 +8473,9 @@ "issue:45127|issue:45357", "issue:44355|issue:44855", "issue:43644|issue:44913", - "issue:44077|issue:44877", - "issue:43540|issue:44734", + "issue:44077|issue:44079", "issue:43976|issue:44855", + "issue:43540|issue:44734", "issue:44496|issue:45372", "issue:43295|issue:45276", "issue:43550|issue:44190", @@ -8455,14 +8485,14 @@ "issue:41950|issue:44016", "issue:39692|issue:44190", "issue:43644|issue:44871", - "issue:36246|issue:41628", "issue:43479|issue:45375", "issue:44186|issue:44734", - "issue:36246|issue:38617", + "issue:36246|issue:41628", "issue:38175|issue:43065", "issue:44246|issue:45003", - "issue:44016|issue:44220", "issue:44466|issue:44913", + "issue:44016|issue:44220", + "issue:36246|issue:38617", "issue:44683|issue:45446", "issue:42175|issue:44279", "issue:44112|issue:44792", @@ -8470,221 +8500,224 @@ "issue:43295|issue:45335", "issue:43643|issue:44496" ], - "prepared_review_unit_hash": "6612d019d4a883a74825af21ff3eadfe6a609fa6f817e46079492f5915863ae8", + "prepared_review_unit_hash": "c5ffc118e4f869b7684b05ce36363e9621882d615f2435809e88dd1beeac42d5", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36997, - "estimated_input_tokens": 9250, - "estimated_eval_tokens": 18756 + "serialized_chars": 37010, + "estimated_input_tokens": 9253, + "estimated_eval_tokens": 18762 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: most pairs share only broad subsystem similarity (imports, serialization, multimodal, config handling) but not the same concrete bug. I would not merge any of the soft edges, and there is no clear canonical issue/PR for the whole set.", - "confidence": 0.96, - "canonical_issue_reason": null, + "summary": "Most pairs are unrelated one-off issues. Only 44683\u219445446 looks like a likely true duplicate; the Qwen3.5 save_pretrained pair is only a tentative related cluster and not strong enough to accept as a duplicate.", + "confidence": 0.82, + "canonical_issue_reason": "44683 is the best canonical anchor because it describes the broader flex_attention failure on torch>=2.9, while 45446 reads like a narrower version-check/root-cause report for the same code path.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "44683 is the strongest representative issue overall because it cleanly captures the only accepted duplicate cluster and states the user-visible regression at the right level of generality.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43761", "right": "issue:44743", "accept": false, - "reason": "Both are regression-style model bugs, but one is CLIPVision hidden states and the other is Qwen3.5 recurrent state reset; different code paths." + "reason": "Different models and failure modes: CLIPVision hidden states vs Qwen recurrent-state reset." }, { "left": "issue:45127", "right": "issue:45357", "accept": false, - "reason": "Both involve tied/serialized weights, but they affect different models and different save/merge paths; not the same bug." + "reason": "Related Qwen3.5 save_pretrained area, but the concrete bugs differ: LoRA/tied-weights collapse vs incorrect visual encoder keys." }, { "left": "issue:44355", "right": "issue:44855", "accept": false, - "reason": "One is compiled Python file errors, the other is a Python 3.13 torch.jit indentation parsing issue; unrelated failures." + "reason": "Both are import/runtime errors, but they come from unrelated causes and code paths." }, { "left": "issue:43644", "right": "issue:44913", "accept": false, - "reason": "Both concern state persistence on reload, but one is non-persistent buffers and the other is a specific GPTNeoX config field; too different to merge." + "reason": "Both involve persistence/reload, but one is junk non-persistent buffers and the other is a GPTNeoX config field reset." }, { "left": "issue:44077", - "right": "issue:44877", + "right": "issue:44079", "accept": false, - "reason": "Both mention config handling, but one is a patchtsmixer post_init policy issue and the other is granite_speech loading under strict config; different bugs." + "reason": "Unrelated issues: patchtsmixer post_init handling vs ModelOutput key assignment." }, { - "left": "issue:43540", - "right": "issue:44734", + "left": "issue:43976", + "right": "issue:44855", "accept": false, - "reason": "Qwen3OmniMoe video-input validation and transformers serve KV-cache indexing are unrelated code paths." + "reason": "Both mention Python version, but the underlying bugs are unrelated." }, { - "left": "issue:43976", - "right": "issue:44855", + "left": "issue:43540", + "right": "issue:44734", "accept": false, - "reason": "One is a Python/version compatibility report, the other is a JIT parsing/import syntax problem; not the same underlying issue." + "reason": "Different subsystems: Qwen3OmniMoe video input processing vs serving KV-cache continuation indexing." }, { "left": "issue:44496", "right": "issue:45372", "accept": false, - "reason": "Both are model-loading failures, but one is missing model_type and the other is a processor import dependency break; different root causes." + "reason": "Both are model-loading/import problems, but the missing symbols and root causes differ." }, { "left": "issue:43295", "right": "issue:45276", "accept": false, - "reason": "Both mention embedding/token resizing, but one is a processor/tokenizer regression and the other is Gemma4 embedding propagation; not the same concrete change." + "reason": "Both are regressions touching model APIs, but processor/tokenizer behavior and embedding resize behavior are different bugs." }, { "left": "issue:43550", "right": "issue:44190", "accept": false, - "reason": "Torch.compile/SDPA model failure and local dataset loading in an image-classification script are unrelated." + "reason": "Completely different bugs: torch.compile SDPA on Bamba vs dataset loading in an example script." }, { "left": "issue:43335", "right": "issue:44855", "accept": false, - "reason": "SwitchTransformers sparse-layer config bug is unrelated to the Python 3.13 torch.jit parsing issue." + "reason": "Different areas entirely: SwitchTransformers config bug vs DebertaV2 import parsing error." }, { "left": "issue:43404", "right": "issue:45216", "accept": false, - "reason": "Both involve model weight handling, but one is Mistral3 lm_head tying and the other is Qwen3.5 save_pretrained checkpoint contents; not mergeable as one fix." + "reason": "Both mention tied weights/checkpointing, but they affect different models and code paths." }, { "left": "issue:43116", "right": "issue:44190", "accept": false, - "reason": "A multi-label classification example bug and a local dataset loading issue in a different script are unrelated example problems." + "reason": "Different example scripts and different failures; not the same underlying issue." }, { "left": "issue:41950", "right": "issue:44016", "accept": false, - "reason": "Video-classification pipeline processor lookup and a notebook syntax error are completely different classes of bug." + "reason": "Unrelated: video-classification pipeline processor lookup vs notebook syntax error." }, { "left": "issue:39692", "right": "issue:44190", "accept": false, - "reason": "SigLIP2 docs/example issues and image-classification script dataset loading are unrelated example failures." + "reason": "Different example/documentation issues and different failure types." }, { "left": "issue:43644", "right": "issue:44871", "accept": false, - "reason": "One is a global buffer serialization regression; the other is a Gemma3 eos_token_id config mismatch. Same broad area, different bugs." - }, - { - "left": "issue:36246", - "right": "issue:41628", - "accept": false, - "reason": "Both are import errors, but one is a missing class in a model submodule and the other is a missing top-level AutoImageProcessor export; different symbols and causes." + "reason": "Serialization/reload bug vs Gemma eos_token_id mismatch; not the same defect." }, { "left": "issue:43479", "right": "issue:45375", "accept": false, - "reason": "Both are config-field handling issues, but one is default vision/audio config initialization and the other is a missing deepstack_visual_indexes field in Qwen3_5MoeVisionConfig." + "reason": "Both are multimodal config issues, but the affected models and actual config problems differ." }, { "left": "issue:44186", "right": "issue:44734", "accept": false, - "reason": "Tokenizer padding/crash in LayoutLMv2 and KV-cache continuation in transformers serve are unrelated." + "reason": "Tokenizer padding/truncation crash is unrelated to serving KV-cache continuation indexing." }, { "left": "issue:36246", - "right": "issue:38617", + "right": "issue:41628", "accept": false, - "reason": "Both are import errors, but they affect different modules and different missing names; no shared underlying fix is evident." + "reason": "Both are ImportError reports, but for different missing names in different modules." }, { "left": "issue:38175", "right": "issue:43065", "accept": false, - "reason": "Unexpected zero probabilities in SigLIP2 and a dummy Conv2d in Sam3PixelDecoder are unrelated model-specific bugs." + "reason": "Different model-specific functional bugs with no shared code path." }, { "left": "issue:44246", "right": "issue:45003", "accept": false, - "reason": "Import-time slowness and unsafe sys.modules access are related only loosely; not the same concrete bug or fix." + "reason": "Import performance issue vs unsafe sys.modules access; at most indirectly related." + }, + { + "left": "issue:44466", + "right": "issue:44913", + "accept": false, + "reason": "Both concern config/state persistence, but one is lm_head serialization and the other is a GPTNeoX config reload default." }, { "left": "issue:44016", "right": "issue:44220", "accept": false, - "reason": "Notebook syntax error and _torch_extract_fbank_features() failure are unrelated." + "reason": "Notebook syntax error and audio feature extraction issue are unrelated." }, { - "left": "issue:44466", - "right": "issue:44913", + "left": "issue:36246", + "right": "issue:38617", "accept": false, - "reason": "Both concern serialization/reload behavior, but one is tied-weight serialization and the other is a GPTNeoX config default reset; different fixes." + "reason": "Both are ImportError reports, but they involve different missing names in different modules." }, { "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 different compatibility bugs." + "accept": true, + "reason": "Same flex_attention code path and same torch-version-guard problem; 45446 appears to isolate the root cause behind the broader torch>=2.9 failure." }, { "left": "issue:42175", "right": "issue:44279", "accept": false, - "reason": "TensorFlow missing from a torch extra install and a generic dependency issue are not the same dependency bug." + "reason": "Dependency/back-end installation complaints, but not the same specific bug." }, { "left": "issue:44112", "right": "issue:44792", "accept": false, - "reason": "CI stale device override failure in GraniteSpeech and janus image-generation test failure are different test issues." + "reason": "Both are failing tests, but they cover unrelated models and failures." }, { "left": "issue:43854", "right": "issue:44792", "accept": false, - "reason": "Model loading failure for GLM-4.7-Flash and a janus generate-images test failure are unrelated." + "reason": "Different model test failures with no shared concrete bug." }, { "left": "issue:43295", "right": "issue:45335", "accept": false, - "reason": "Both are resize_token_embeddings regressions, but they hit different model architectures and different internal embeddings." + "reason": "Both are regressions around embeddings/tokenizers, but the affected behaviors and models are different." }, { "left": "issue:43643", "right": "issue:44496", "accept": false, - "reason": "A trust_remote_code config-field loss and an unrecognized-model error are separate loading issues with different root causes." + "reason": "Different config-loading failures: missing fields with trust_remote_code vs unrecognized model_type." } ] }, - "evaluator_result": null, - "evaluator_used": false, - "retried": false, - "accepted_nontrivial_soft_edge": false, + "evaluator_result": { + "accept": false, + "feedback": "The write-up is a bit too strong for the packet. It claims \u2018only two soft pairs look like true duplicates,\u2019 but the provided soft-edge verdicts accept only one pair (44683\u219445446). The Qwen3.5 save_pretrained relationship (45216\u219445357) may be plausible, but it is not an accepted soft pair in the packet and should be phrased more conservatively. I\u2019d accept a summary that says most pairs are unrelated and one soft pair is a likely duplicate, with Qwen3.5 noted as a tentative related cluster rather than a settled duplicate." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -8701,9 +8734,9 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-364", "nodes": [ + "issue:38175", "issue:41950", "issue:42907", - "issue:43295", "issue:43299", "issue:43441", "issue:43479", @@ -8731,8 +8764,8 @@ "issue:44464", "issue:44496", "issue:44521", + "issue:44617", "issue:44683", - "issue:44779", "issue:44871", "issue:44877", "issue:44933", @@ -8768,249 +8801,242 @@ "issue:44075|issue:44368", "issue:44521|issue:45325", "issue:43854|issue:44964", - "issue:44683|issue:45542", "issue:44242|issue:45200", "issue:44464|issue:45245", + "issue:44683|issue:45542", "issue:43976|issue:45405", - "issue:44351|issue:45278", "issue:41950|issue:43650", "issue:43479|issue:43525", - "issue:41950|issue:45020", + "issue:44351|issue:45278", "issue:44496|issue:44877", + "issue:41950|issue:45020", "issue:44462|issue:45081", - "issue:44297|issue:44462", "issue:43550|issue:44964", "issue:44279|issue:45405", + "issue:44297|issue:44462", "issue:43299|issue:43854", - "issue:43295|issue:44779", - "issue:44462|issue:44779" + "issue:38175|issue:44617" ], - "prepared_review_unit_hash": "5212e2e145c5ba8d397fad818a7bc701447cdd0a5d1bc0842a59f4bedc4bdd31", + "prepared_review_unit_hash": "abc27508d52b6e26e45ccc19fecd6a59b4b689542798fe06e2ce5157472d6568", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 33, - "serialized_chars": 37579, - "estimated_input_tokens": 9395, - "estimated_eval_tokens": 19046 + "soft_pair_count": 32, + "serialized_chars": 37204, + "estimated_input_tokens": 9301, + "estimated_eval_tokens": 18858 }, "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, mostly version-specific regressions and model-loading bugs. None of the soft pairs look like true duplicates or mergeable PR-equivalents, so I would reject all soft edges.", - "confidence": 0.2, - "canonical_issue_reason": "issue:43824 is the strongest representative issue here: it is concrete, well-scoped, has the most discussion/inbound references, and clearly describes a reproducible import failure with a specific model class.", + "summary": "This cluster is a heterogeneous grab-bag of unrelated Transformers issues across many models/features; the soft edges are mostly superficial text similarity and should not be merged as duplicates.", + "confidence": 0.97, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:43824 is the best single issue to keep as the representative bug report because it is precise, high-signal, and appears to have broader community impact than the other items.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43299", "right": "issue:45216", "accept": false, - "reason": "Different regressions on different Qwen model families: one is loading Qwen3VL MoE models, the other is save_pretrained corruption for Qwen3.5." + "reason": "Different bugs: Qwen3VL MoE loading regression vs Qwen3.5 save_pretrained checkpoint serialization." }, { "left": "issue:42907", "right": "issue:43441", "accept": false, - "reason": "Both mention Ministral, but one is about saving dequantized models while the other is a FlashAttention loading/runtime failure; not the same bug." + "reason": "Ministral/Devstral save failure is unrelated to FlashAttention runtime failure in Ministral-3." }, { "left": "issue:44016", "right": "issue:44062", "accept": false, - "reason": "Notebook syntax error and AddedToken keyword collision are unrelated failures in different code paths." + "reason": "Notebook syntax error and tokenizers AddedToken keyword conflict are unrelated." }, { "left": "issue:44112", "right": "issue:44964", "accept": false, - "reason": "A stale CI test in GraniteSpeech and a Phi-4 multimodal loading failure are distinct issues." + "reason": "CI stale override test failure is unrelated to Phi-4-multimodal loading." }, { "left": "issue:44871", "right": "issue:45245", "accept": false, - "reason": "Gemma eos_token_id mismatch and a categories-limit runtime error are unrelated." + "reason": "Gemma eos_token_id mismatch is unrelated to category-count runtime error." }, { "left": "issue:43673", "right": "issue:44336", "accept": false, - "reason": "Chunked_prefill cache handling and ANSI formatting in loading_report are different subsystems and symptoms." + "reason": "Chunked_prefill cache missing is unrelated to ANSI codes in loading_report." }, { "left": "issue:43976", "right": "issue:44987", "accept": false, - "reason": "Python version compatibility and loading physical-intelligence/fast are not the same underlying bug." + "reason": "Python version compatibility issue is unrelated to loading failure for physical-intelligence/fast." }, { "left": "issue:43526", "right": "issue:43901", "accept": false, - "reason": "BeitImageProcessorFast label reduction is unrelated to a docs mismatch about return_all_scores." + "reason": "BeitImageProcessorFast reduce_labels bug is unrelated to TextClassificationPipeline docs wording." }, { "left": "issue:44496", "right": "issue:45042", "accept": false, - "reason": "One is an unrecognized model/config issue; the other is an image processor torchvision dependency regression." + "reason": "Olmo model config loading failure is unrelated to PIL backend torchvision dependency regression." }, { "left": "issue:43824", "right": "issue:44933", "accept": false, - "reason": "Different import failures: missing Qwen2_5_VL export versus a nonexistent image_utils import." + "reason": "Qwen2.5-VL import error and missing image_utils import are different symbols and code paths." }, { "left": "issue:43976", "right": "issue:44279", "accept": false, - "reason": "A Python support issue and a generic dependency issue are too broad and distinct to be duplicates." + "reason": "Python version support problem is unrelated to a generic dependency issue." }, { "left": "issue:43299", "right": "issue:45357", "accept": false, - "reason": "Qwen3VL loading and Qwen3.5 visual-encoder save_pretrained corruption are separate regressions." + "reason": "Qwen3VL MoE loading bug is unrelated to Qwen3.5 visual-encoder save_pretrained regression." }, { "left": "issue:43577", "right": "issue:45216", "accept": false, - "reason": "Blip2 dtype propagation and Qwen3.5 checkpoint saving are unrelated." + "reason": "Blip2 dtype propagation bug is unrelated to Qwen3.5 save_pretrained serialization." }, { "left": "issue:43901", "right": "issue:44496", "accept": false, - "reason": "Documentation drift for return_all_scores is not the same as a model auto-mapping/load failure." + "reason": "Pipeline docs mismatch is unrelated to model config loading failure." }, { "left": "issue:44075", "right": "issue:44368", "accept": false, - "reason": "SGD optimizer args being ignored and a tie_word_embeddings warning are unrelated." + "reason": "SGD optimizer args ignored and tie_word_embeddings warning are unrelated." }, { "left": "issue:44521", "right": "issue:45325", "accept": false, - "reason": "Assistant mask generation for multimodal chat templates is unrelated to Qwen2.5-VL rope index scaling." + "reason": "assistant_masks being zero for multimodal chat templates is unrelated to Qwen2.5-VL rope index scaling." }, { "left": "issue:43854", "right": "issue:44964", "accept": false, - "reason": "Both are model-loading issues, but they affect different models and failure points; no evidence they share the same code path." - }, - { - "left": "issue:44683", - "right": "issue:45542", - "accept": false, - "reason": "Compiled flex_attention on torch>=2.9 and a missing TensorFlow backend import are unrelated." + "reason": "GLM-4.7-Flash unit test loading issue and Phi-4-multimodal loading issue are different model-specific failures." }, { "left": "issue:44242", "right": "issue:45200", "accept": false, - "reason": "Load-balancing loss omission and Gemma 4 mm_token_type_ids defaulting are different training-time issues." + "reason": "Missing load balancing loss with output_router_logits=False is unrelated to Gemma 4 mm_token_type_ids defaults." }, { "left": "issue:44464", "right": "issue:45245", "accept": false, - "reason": "Chunked generation inconsistency and a categories cardinality limit error are unrelated." + "reason": "Compiled-forward chunked generation inconsistency is unrelated to category-count limit error." }, { - "left": "issue:43976", - "right": "issue:45405", + "left": "issue:44683", + "right": "issue:45542", "accept": false, - "reason": "Python compatibility and an unreleased PEFT version bump are different dependency problems." + "reason": "Compiled flex_attention on torch>=2.9 is unrelated to TensorFlow/tensorboard backend installation." }, { - "left": "issue:44351", - "right": "issue:45278", + "left": "issue:43976", + "right": "issue:45405", "accept": false, - "reason": "A specific missing HybridCache export may be one symptom of broader import breakage, but the evidence is too weak to call it the same bug as the generic multi-import failure." + "reason": "Python version support issue is unrelated to PEFT minimum version bump." }, { "left": "issue:41950", "right": "issue:43650", "accept": false, - "reason": "Video-classification processor lookup and a vague 'ADD THE DATA' issue are not related." + "reason": "Video-classification processor lookup bug is unrelated to the unrelated 'ADD THE DATA' issue." }, { "left": "issue:43479", "right": "issue:43525", "accept": false, - "reason": "Phi4MultimodalConfig default initialization and a Llama4Config pad_token_id AttributeError are separate model config bugs." + "reason": "Phi4Multimodal default-config initialization and Llama4Config missing pad_token_id are different config bugs." }, { - "left": "issue:41950", - "right": "issue:45020", + "left": "issue:44351", + "right": "issue:45278", "accept": false, - "reason": "A specific video-classification pipeline bug and a broad remote_code regression are not the same concrete issue." + "reason": "HybridCache import error and broad import errors after version upgrade are not the same concrete bug." }, { "left": "issue:44496", "right": "issue:44877", "accept": false, - "reason": "Unrecognized model loading and strict granite_speech config validation are different failures." + "reason": "Unrecognized model loading and strict granite_speech config loading are different config/registration problems." }, { - "left": "issue:44462", - "right": "issue:45081", + "left": "issue:41950", + "right": "issue:45020", "accept": false, - "reason": "Tokenizer JSON loading and a Mistral regex patch crash are unrelated tokenizer issues." + "reason": "Image processor lookup in video pipeline is unrelated to broad remote_code breakage." }, { - "left": "issue:44297", - "right": "issue:44462", + "left": "issue:44462", + "right": "issue:45081", "accept": false, - "reason": "Tokenizer class metadata mismatch on save and AutoTokenizer ignoring tokenizer.json are different load/save bugs." + "reason": "AutoTokenizer ignoring tokenizer.json is unrelated to Mistral regex patch crashing on backend_tokenizer access." }, { "left": "issue:43550", "right": "issue:44964", "accept": false, - "reason": "Bamba torch.compile/SDPA failure and Phi-4 multimodal loading are unrelated." + "reason": "Bamba torch.compile SDPA failure is unrelated to Phi-4-multimodal loading." }, { "left": "issue:44279", "right": "issue:45405", "accept": false, - "reason": "A generic transformers dependency complaint and a specific unreleased PEFT minimum version bump are not duplicates." + "reason": "Generic dependency issue is unrelated to the PEFT version pin problem." + }, + { + "left": "issue:44297", + "right": "issue:44462", + "accept": false, + "reason": "tokenizer_class mismatch on save_pretrained is unrelated to AutoTokenizer ignoring tokenizer.json on load." }, { "left": "issue:43299", "right": "issue:43854", "accept": false, - "reason": "Different model families and different failure modes; no shared code path is evident." + "reason": "Qwen3VL MoE loading and Qwen2.5-VL import/export symbol issues are different model families and failure modes." }, { - "left": "issue:43295", - "right": "issue:44779", - "accept": false, - "reason": "Both are tokenizer regressions, but they affect different models and report different incorrect behaviors." - }, - { - "left": "issue:44462", - "right": "issue:44779", + "left": "issue:38175", + "right": "issue:44617", "accept": false, - "reason": "Deepseek tokenizer correctness regression is not the same as AutoTokenizer ignoring repository tokenizer.json." + "reason": "Siglip2 zero probabilities bug is unrelated to Sam3Video CUDA OOM." } ] }, @@ -9034,7 +9060,6 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-364", "nodes": [ - "issue:38175", "issue:39692", "issue:42175", "issue:42994", @@ -9047,14 +9072,12 @@ "issue:43441", "issue:43450", "issue:43475", - "issue:43525", "issue:43650", "issue:43673", "issue:43723", "issue:43761", "issue:43784", "issue:43824", - "issue:43881", "issue:43901", "issue:43976", "issue:43994", @@ -9066,8 +9089,10 @@ "issue:44279", "issue:44291", "issue:44297", + "issue:44462", "issue:44466", "issue:44521", + "issue:44589", "issue:44617", "issue:44623", "issue:44671", @@ -9079,21 +9104,23 @@ "issue:45250", "issue:45276", "issue:45290", + "issue:45362", "issue:45397", "issue:45542" ], "soft_pairs": [ - "issue:38175|issue:44617", + "issue:43295|issue:44779", "issue:39692|issue:43901", + "issue:44462|issue:44779", "issue:43650|issue:45290", "issue:44016|issue:44291", "issue:44466|issue:45276", "issue:43441|issue:44683", "issue:43450|issue:44279", - "issue:42994|issue:43317", - "issue:43065|issue:44898", - "issue:43723|issue:44162", "issue:43262|issue:44521", + "issue:43723|issue:44162", + "issue:43065|issue:44898", + "issue:42994|issue:43317", "issue:44261|issue:44263", "issue:44623|issue:45397", "issue:43122|issue:43650", @@ -9101,213 +9128,219 @@ "issue:43784|issue:45542", "issue:44671|issue:44779", "issue:43673|issue:43761", - "issue:44987|issue:45542", "issue:43901|issue:43994", + "issue:44987|issue:45542", "issue:44297|issue:45397", "issue:42175|issue:43976", "issue:43475|issue:44016", "issue:43824|issue:43976", "issue:43901|issue:44246", "issue:43352|issue:45250", - "issue:43525|issue:43881", - "issue:43295|issue:44617" + "issue:43295|issue:44617", + "issue:44589|issue:45362" ], - "prepared_review_unit_hash": "34618173c6223d8165cb271fc63780ab31b2be3e6568b867fcdde7c6efadd74d", + "prepared_review_unit_hash": "6bd5636fa9ff10d1ecb98bf8476075b965196a37a027b0d40fe5ab31fadfb3c2", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 28, - "serialized_chars": 35249, - "estimated_input_tokens": 8813, - "estimated_eval_tokens": 17882 + "soft_pair_count": 29, + "serialized_chars": 35550, + "estimated_input_tokens": 8888, + "estimated_eval_tokens": 18032 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "All soft pairs are rejected: they share broad keywords or subsystems, but each pair points to a different concrete bug, regression, or docs issue. The most representative issue in this set is the SigLIP2 output-quality report, which is the clearest self-contained user-facing bug.", - "confidence": 0.86, - "canonical_issue_reason": "issue:43994 is the clearest, most self-contained bug report: it names the model, the wrong behavior, and the affected entry points (AutoModel and pipeline).", + "summary": "The candidate pairs are mostly loose thematic matches (tokenizers, chat templates, quantization, FlashAttention, processors) but they describe different bugs, different code paths, or different model families. I would not merge any of them as duplicates.", + "confidence": 0.93, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:43994 is the strongest standalone issue in this set because it gives a concrete end-user symptom and a broad repro surface, making it the best representative issue.", + "best_issue_reason": "issue:43994 is the strongest standalone issue here: it is concrete, reproducible, and describes a clear user-facing regression in SigLIP2 inference with AutoModel/pipeline.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:38175", - "right": "issue:44617", + "left": "issue:43295", + "right": "issue:44779", "accept": false, - "reason": "SigLIP2 zero-probability outputs vs SAM3 video CUDA OOM are unrelated failures." + "reason": "Both are v5 tokenizer/processor regressions, but one is about processor.tokenizer/image handling and the other is Deepseek tokenizer output; different failures and code paths." }, { "left": "issue:39692", "right": "issue:43901", "accept": false, - "reason": "Both mention docs/examples, but one is a SigLIP2 example bug and the other is a pipeline-docs behavior mismatch." + "reason": "Both mention docs, but one is a SigLIP2 example with model/processor mismatch and quantization failure, while the other is a pipeline docs mismatch about return_all_scores." + }, + { + "left": "issue:44462", + "right": "issue:44779", + "accept": false, + "reason": "AutoTokenizer ignoring tokenizer.json and Deepseek tokenizer producing wrong results are separate tokenizer bugs, not the same underlying defect." }, { "left": "issue:43650", "right": "issue:45290", "accept": false, - "reason": "Generic placeholder issue vs apply_chat_template crash on tool-call assistant messages; no shared underlying bug." + "reason": "Completely unrelated: one is an empty/titled issue, the other is a specific apply_chat_template crash with tool calls." }, { "left": "issue:44016", "right": "issue:44291", "accept": false, - "reason": "Notebook syntax error is unrelated to the init_empty_weights TypeError regression." + "reason": "Notebook syntax error vs init_empty_weights/_is_hf_initialized loading TypeError; different surfaces and root causes." }, { "left": "issue:44466", "right": "issue:45276", "accept": false, - "reason": "Both touch weights/serialization-adjacent behavior, but one is inconsistent lm_head serialization and the other is resize_token_embeddings not propagating." + "reason": "Serialization of lm_head.weight depending on device is unrelated to gemma4 resize_token_embeddings not updating embeddings." }, { "left": "issue:43441", "right": "issue:44683", "accept": false, - "reason": "Both involve Flash Attention, but one is model support gating and the other is a compiled flex_attention failure on newer torch." + "reason": "Both involve attention backends, but one is a model not supporting FlashAttention 2.0 and the other is compiled flex_attention failing on torch>=2.9; not the same bug." }, { "left": "issue:43450", "right": "issue:44279", "accept": false, - "reason": "Video processor batching shape bug vs dependency issue; different layers and symptoms." + "reason": "Video processor batching shape bug vs a generic dependency issue; no shared code-path or concrete duplicate evidence." }, { - "left": "issue:42994", - "right": "issue:43317", + "left": "issue:43262", + "right": "issue:44521", "accept": false, - "reason": "Quantized model saving failure and dequantized offload loading failure are different code paths." + "reason": "Audio chat-template sampling-rate default issue and multimodal assistant_mask issue are distinct chat-template bugs." }, { - "left": "issue:43065", - "right": "issue:44898", + "left": "issue:43723", + "right": "issue:44162", "accept": false, - "reason": "Dummy Conv2d in Sam3PixelDecoder is unrelated to Perceiver interpolation behavior." + "reason": "AutoTokenizer.from_pretrained loading problem and ESM2 broken model behavior are unrelated." }, { - "left": "issue:43723", - "right": "issue:44162", + "left": "issue:43065", + "right": "issue:44898", "accept": false, - "reason": "Tokenizer loading regression vs ESM2 model breakage; not the same underlying issue." + "reason": "SAM3 pixel decoder dummy conv and Perceiver non-default resolution failure are different model-specific issues." }, { - "left": "issue:43262", - "right": "issue:44521", + "left": "issue:42994", + "right": "issue:43317", "accept": false, - "reason": "Audio chat-template sampling-rate default and multimodal assistant_mask all-zero bug are different apply_chat_template problems." + "reason": "Saving a quantized model and loading a dequantized model with device_map=auto are different quantization/offload paths." }, { "left": "issue:44261", "right": "issue:44263", "accept": false, - "reason": "MLA layernorm precision issue and GlmMoeDsaIndexer torch.split return-value issue are unrelated." + "reason": "MLA rms_norm_eps precision issue and GlmMoeDsaIndexer torch.split return values are unrelated." }, { "left": "issue:44623", "right": "issue:45397", "accept": false, - "reason": "processor.save_pretrained missing files is distinct from gemma-4 zero3 from_pretrained failure." + "reason": "processor.save_pretrained missing files and gemma-4 zero3 from_pretrained are different save/load bugs." }, { "left": "issue:43122", "right": "issue:43650", "accept": false, - "reason": "Tokenizer behavior regression is unrelated to the placeholder 'ADD THE DATA' issue." + "reason": "Tokenizer regression vs an unrelated empty issue title; no duplicate relationship." }, { "left": "issue:44623", "right": "issue:44749", "accept": false, - "reason": "Save-pretrained file omission vs Chinese-reported performance slowdown after upgrade are different bugs." + "reason": "Saving processor files and slow filtering after upgrading to 5.3.0 are unrelated." }, { "left": "issue:43784", "right": "issue:45542", "accept": false, - "reason": "Missing nn import in sentence-transformers is unrelated to TensorBoard/TensorFlow backend installation behavior." + "reason": "Import-time NameError in sentence-transformers integration vs missing TensorFlow backend when only tensorboard is installed; different dependency failures." }, { "left": "issue:44671", "right": "issue:44779", "accept": false, - "reason": "CamemBERT masked-LM prediction regression and Deepseek tokenizer regression are different model/tokenizer bugs." + "reason": "CamemBERT MLM prediction regression and Deepseek tokenizer regression are model-specific and distinct." }, { "left": "issue:43673", "right": "issue:43761", "accept": false, - "reason": "Generation cache missing in chunked_prefill vs CLIP hidden_states output regression are separate code paths." + "reason": "Generation cache missing during chunked_prefill and CLIPVisionModel hidden_states=None are different regressions." }, { - "left": "issue:44987", - "right": "issue:45542", + "left": "issue:43901", + "right": "issue:43994", "accept": false, - "reason": "Model loading failure on transformers>=5.1.0 is unrelated to missing TensorFlow backend after pip install '.[torch]'." + "reason": "Docs note about return_all_scores does not match a runtime SigLIP2 inference correctness bug." }, { - "left": "issue:43901", - "right": "issue:43994", + "left": "issue:44987", + "right": "issue:45542", "accept": false, - "reason": "Docs mention of return_all_scores is not the same as SigLIP2 producing nonsensical outputs." + "reason": "Model loading failure for physical-intelligence/fast and TensorFlow backend installation issue are unrelated." }, { "left": "issue:44297", "right": "issue:45397", "accept": false, - "reason": "Tokenizer save metadata mismatch and gemma-4 zero3 loading failure do not describe the same bug." + "reason": "Tokenizer save_pretrained metadata mismatch and gemma-4 zero3 from_pretrained are different problems." }, { "left": "issue:42175", "right": "issue:43976", "accept": false, - "reason": "Missing TensorFlow in the torch extra and Python-version compatibility failure are different dependency issues." + "reason": "Missing TensorFlow in the torch extra and a Python version compatibility regression are unrelated." }, { "left": "issue:43475", "right": "issue:44016", "accept": false, - "reason": "SAM3 video encoder attribute error is unrelated to a notebook syntax error." + "reason": "SAM3 video encoder attribute error and notebook syntax error are unrelated." }, { "left": "issue:43824", "right": "issue:43976", "accept": false, - "reason": "Missing Qwen2_5_VL import and Python version support regression are unrelated." + "reason": "Missing Qwen2_5_VL import and Python 3.9/3.10 support regression are unrelated." }, { "left": "issue:43901", "right": "issue:44246", "accept": false, - "reason": "Pipeline docs mismatch vs import being slow sometimes are not the same problem." + "reason": "Docs mismatch about return_all_scores and slow import behavior are different issues." }, { "left": "issue:43352", "right": "issue:45250", "accept": false, - "reason": "Specific model lacking Flash Attention 2.0 support is not the same as the generic Flash Attention 2.0 issue/request." + "reason": "Generic Flash Attention 2.0 discussion is too broad to treat as the same bug as NemotronHForCausalLM explicitly not supporting it." }, { - "left": "issue:43525", - "right": "issue:43881", + "left": "issue:43295", + "right": "issue:44617", "accept": false, - "reason": "Llama4Config missing pad_token_id and glm-4v-9b loading failed are different attribute/loading failures." + "reason": "Custom model processor/tokenizer regression and SAM3Video CUDA OOM are unrelated." }, { - "left": "issue:43295", - "right": "issue:44617", + "left": "issue:44589", + "right": "issue:45362", "accept": false, - "reason": "Processor.tokenizer regression with image passing is unrelated to SAM3 video CUDA OOM." + "reason": "Float8 storage lookup TypeError and Qwen3.5 chat crash are different failure modes." } ] }, @@ -9340,11 +9373,13 @@ "issue:43299", "issue:43441", "issue:43450", + "issue:43525", "issue:43526", "issue:43531", "issue:43540", "issue:43596", "issue:43673", + "issue:43881", "issue:43976", "issue:43994", "issue:44162", @@ -9360,7 +9395,6 @@ "issue:44485", "issue:44496", "issue:44561", - "issue:44589", "issue:44617", "issue:44661", "issue:44749", @@ -9368,7 +9402,6 @@ "issue:44857", "issue:44871", "issue:44877", - "issue:44987", "issue:45003", "issue:45200", "issue:45276", @@ -9377,235 +9410,227 @@ "issue:45362", "issue:45375", "issue:45397", - "issue:45459", "issue:45542" ], "soft_pairs": [ - "issue:44589|issue:45362", + "issue:43525|issue:43881", "issue:43450|issue:44617", "issue:44188|issue:44336", "issue:41628|issue:43976", - "issue:41950|issue:44220", "issue:44295|issue:45397", "issue:42491|issue:43299", - "issue:44246|issue:44496", + "issue:41950|issue:44220", "issue:44336|issue:44373", - "issue:44661|issue:45542", + "issue:44246|issue:44496", "issue:44462|issue:45356", + "issue:44661|issue:45542", "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:36246|issue:44351", "issue:44336|issue:44485", "issue:44162|issue:45200", "issue:44242|issue:44857", "issue:44246|issue:44871", - "issue:44295|issue:44749", "issue:44779|issue:45292", + "issue:44295|issue:44749", "issue:44877|issue:45375", - "issue:36246|issue:43596", "issue:43441|issue:43540", - "issue:44987|issue:45459", - "issue:43066|issue:44779" + "issue:43066|issue:44779", + "issue:36246|issue:43596" ], - "prepared_review_unit_hash": "c2ba531fa00f668233b267b3f6d2f1ffb8696230ac230bd6865849558a52c2a4", + "prepared_review_unit_hash": "aa62a45472f92db1c768edda18d567610ec0babc84427da082a8c948cae68f24", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 28, - "serialized_chars": 35863, - "estimated_input_tokens": 8966, - "estimated_eval_tokens": 18188 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 34940, + "estimated_input_tokens": 8735, + "estimated_eval_tokens": 17726 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is highly heterogeneous: most pairs are only broad Transformers v5 regressions or entirely unrelated model/pipeline bugs, so no soft edge looks like the same concrete issue. If forced to pick a representative, issue 41628 is the broadest user-facing import-regression report.", - "confidence": 0.27, - "canonical_issue_reason": "Issue 41628 is the cleanest, most generic top-level API breakage report in the set (a direct import failure from `transformers`), so it is the least-wrong representative of the cluster\u2019s loose v5 import/regression theme.", + "summary": "This cluster is heterogeneous: most items are distinct Transformers v5 regressions, import errors, or model-specific loading bugs rather than duplicates. No soft pair looks like the same concrete underlying bug/change with enough confidence to merge.", + "confidence": 0.24, + "canonical_issue_reason": "No clear duplicate emerges from the set. If one issue must represent the cluster, issue 44561 is the best umbrella because it captures a broad v5 API-breakage/regression theme affecting downstream models.", "canonical_pr_reason": null, - "best_issue_reason": "41628 is the most broadly applicable and user-visible failure mode among these issues, making it the best canonical issue candidate despite the overall heterogeneity.", + "best_issue_reason": "Issue 44561 is the most representative general regression report in this set; it is broad, version-specific, and plausibly connected to multiple downstream breakages. Still, it should not be treated as a duplicate of the others.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44589", - "right": "issue:45362", + "left": "issue:43525", + "right": "issue:43881", "accept": false, - "reason": "Different failures: a storage/type import/runtime error versus a Qwen chat crash. Same ecosystem, not the same bug." + "reason": "Both are loading failures, but they affect different models and describe different failures: missing Llama4Config.pad_token_id vs a generic glm-4v-9b loading error." }, { "left": "issue:43450", "right": "issue:44617", "accept": false, - "reason": "One is a batched video processor shape bug; the other is CUDA OOM in Sam3Video. Not the same underlying issue." + "reason": "Different problem classes: batched video processor shape bug versus Sam3Video CUDA OOM." }, { "left": "issue:44188", "right": "issue:44336", "accept": false, - "reason": "Diverging attention kernels under `torch.compile` is unrelated to ANSI code output from loading_report." + "reason": "Unrelated subsystems: torch.compile attention kernel divergence versus ANSI codes in loading_report." }, { "left": "issue:41628", "right": "issue:43976", "accept": false, - "reason": "ImportError for `AutoImageProcessor` and a Python version compatibility report are different problem classes." - }, - { - "left": "issue:41950", - "right": "issue:44220", - "accept": false, - "reason": "Video pipeline choosing image processors and `_torch_extract_fbank_features()` failures are unrelated code paths." + "reason": "Both mention import/runtime compatibility, but one is a missing AutoImageProcessor import and the other is a Python version compatibility complaint; not the same bug." }, { "left": "issue:44295", "right": "issue:45397", "accept": false, - "reason": "Buffer access on `position_ids` and Zero3 `from_pretrained` breakage are different bugs." + "reason": "Different code paths and symptoms: position_ids buffer access error versus gemma-4 zero3 from_pretrained loading." }, { "left": "issue:42491", "right": "issue:43299", "accept": false, - "reason": "Both involve Qwen3 MoE, but one is HF4.x LoRA compatibility and the other is v5 loading breakage for Qwen3VL MoE; not mergeable as one bug." + "reason": "Both involve Qwen3 MoE family regressions, but one is LoRA compatibility across branches and the other is Qwen3VL model loading; too different to merge." }, { - "left": "issue:44246", - "right": "issue:44496", + "left": "issue:41950", + "right": "issue:44220", "accept": false, - "reason": "Slow import and unrecognized model/config loading are not the same underlying failure." + "reason": "Video-classification pipeline selecting image processors is unrelated to _torch_extract_fbank_features()." }, { "left": "issue:44336", "right": "issue:44373", "accept": false, - "reason": "Terminal ANSI escape output is unrelated to a wrong docstring for `position_ids`." + "reason": "Formatting in loading_report and a wrong docstring for position_ids are unrelated." }, { - "left": "issue:44661", - "right": "issue:45542", + "left": "issue:44246", + "right": "issue:44496", "accept": false, - "reason": "Tokenizer mapping-name handling and missing TensorFlow backend installation are unrelated." + "reason": "Import slowdown and unrecognized model config are different failures with no shared concrete code path." }, { "left": "issue:44462", "right": "issue:45356", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different models and different failure modes; too broad to treat as one bug." + "reason": "Both are tokenizer regressions, but one ignores tokenizer.json and the other has codec/warning behavior in Kimi-K2.5; not the same bug." + }, + { + "left": "issue:44661", + "right": "issue:45542", + "accept": false, + "reason": "Tokenizer mapping registration failure is unrelated to the tensorboard-without-TF backend error." }, { "left": "issue:43526", "right": "issue:45003", "accept": false, - "reason": "BeitImageProcessorFast label reduction and unsafe `sys.modules` access are unrelated." + "reason": "BeitImageProcessorFast reduce_labels bug and sys.modules access in modeling_utils are unrelated." }, { "left": "issue:43994", "right": "issue:45276", "accept": false, - "reason": "Nonsensical model outputs and resize-token-embedding propagation are different issues." + "reason": "Different model-specific regressions: SigLIP2 output quality versus gemma4 resize_token_embeddings behavior." }, { "left": "issue:43531", "right": "issue:45362", "accept": false, - "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to a Qwen3.5-35B chat crash." + "reason": "Both are Qwen-related, but sliding_window handling and chat crashes are distinct concrete failures." }, { "left": "issue:36246", "right": "issue:42757", "accept": false, - "reason": "Both are import errors, but for different missing symbols and different root causes." + "reason": "Different missing imports from different packages; same broad symptom, not the same underlying issue." }, { - "left": "issue:36246", - "right": "issue:44351", + "left": "issue:43673", + "right": "issue:44561", "accept": false, - "reason": "Missing `Qwen2_5_VLImageProcessor` and missing `HybridCache` are separate export regressions." + "reason": "Both are v5 regressions, but one is cache handling during chunked_prefill and the other is a removed helper breaking trust_remote_code." }, { - "left": "issue:43673", - "right": "issue:44561", + "left": "issue:36246", + "right": "issue:44351", "accept": false, - "reason": "Chunked-prefill cache behavior and removal of `is_torch_fx_available` break different code paths." + "reason": "Both are import errors, but they concern different symbols and likely different breakages." }, { "left": "issue:44336", "right": "issue:44485", "accept": false, - "reason": "ANSI output noise and GLM-5 RoPE implementation are unrelated." + "reason": "ANSI output behavior and GLM-5 RoPE implementation are unrelated." }, { "left": "issue:44162", "right": "issue:45200", "accept": false, - "reason": "ESM2 breakage and Gemma 4 `mm_token_type_ids` defaults are different model-specific bugs." + "reason": "ESM2 breakage and Gemma 4 mm_token_type_ids defaulting are separate model-specific issues." }, { "left": "issue:44242", "right": "issue:44857", "accept": false, - "reason": "Missing load-balancing loss and float16 AMP crash are unrelated training/runtime failures." + "reason": "MoE router loss behavior and LwDetrImageLoss AMP crash are different code paths." }, { "left": "issue:44246", "right": "issue:44871", "accept": false, - "reason": "Import slowness and Gemma eos-token-id inconsistency are not the same bug." - }, - { - "left": "issue:44295", - "right": "issue:44749", - "accept": false, - "reason": "`position_ids` buffer access and data-filtering slowdown after upgrade are unrelated." + "reason": "Import performance regression is unrelated to Gemma-3 eos_token_id inconsistency." }, { "left": "issue:44779", "right": "issue:45292", "accept": false, - "reason": "Tokenizer regression and `resize_token_embeddings` not updating output embeddings are different code paths." + "reason": "Tokenizer incorrect results and resize_token_embeddings not updating output embeddings are different bugs." }, { - "left": "issue:44877", - "right": "issue:45375", + "left": "issue:44295", + "right": "issue:44749", "accept": false, - "reason": "Strict config rejection for granite_speech and a silently dropped Qwen config field are similar only at a high level, not the same concrete defect." + "reason": "Position_ids buffer access error and Chinese-language performance regression are unrelated." }, { - "left": "issue:36246", - "right": "issue:43596", + "left": "issue:44877", + "right": "issue:45375", "accept": false, - "reason": "An import error and a DeepSpeed ZeRO3/BertModel index error are unrelated." + "reason": "Both involve config strictness, but they affect different model configs and fields; not the same underlying defect." }, { "left": "issue:43441", "right": "issue:43540", "accept": false, - "reason": "Ministral flash-attention failure and Qwen3OmniMoe video-input ValueError are different bugs." + "reason": "FlashAttention failure in Ministral-3 and Qwen3OmniMoe video ValueError are unrelated." }, { - "left": "issue:44987", - "right": "issue:45459", + "left": "issue:43066", + "right": "issue:44779", "accept": false, - "reason": "Loading a specific model and masking tokenizer decode errors when protobuf is absent are unrelated." + "reason": "Both are tokenizer regressions in v5, but they are different model families and different failure modes." }, { - "left": "issue:43066", - "right": "issue:44779", + "left": "issue:36246", + "right": "issue:43596", "accept": false, - "reason": "Both are tokenizer regressions in v5, but they concern different models and different decoder/codec issues, so they should not be merged." + "reason": "Missing Qwen2_5_VLImageProcessor import and DeepSpeed zero3/BertModel IndexError are unrelated." } ] }, @@ -9629,6 +9654,8 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-364", "nodes": [ + "issue:28282", + "issue:30990", "issue:41628", "issue:41950", "issue:42673", @@ -9656,16 +9683,15 @@ "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:44987", "issue:44995", "issue:45003", "issue:45137", @@ -9678,10 +9704,11 @@ "issue:45459" ], "soft_pairs": [ + "issue:44987|issue:45459", "issue:43901|issue:45003", - "issue:41628|issue:41950", "issue:42886|issue:45229", "issue:44995|issue:45200", + "issue:41628|issue:41950", "issue:43257|issue:45137", "issue:43065|issue:43408", "issue:43232|issue:44336", @@ -9700,215 +9727,209 @@ "issue:43232|issue:43673", "issue:43931|issue:44368", "issue:44297|issue:44623", - "issue:43646|issue:44162", "issue:43673|issue:45278", + "issue:43646|issue:44162", "issue:43381|issue:45292", "issue:43381|issue:45335", "issue:42673|issue:43704", - "issue:44361|issue:44749" + "issue:28282|issue:30990" ], - "prepared_review_unit_hash": "ae9788d5886e100cfb50f20a3865eadf72b0c49c1e740f1075b739b94d0932d9", + "prepared_review_unit_hash": "6d27097b48b3df46830f9265dca57dc522e091369ddd903124e72e6910c4197b", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 28, - "serialized_chars": 35530, - "estimated_input_tokens": 8883, - "estimated_eval_tokens": 18022 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36342, + "estimated_input_tokens": 9086, + "estimated_eval_tokens": 18428 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose topical grouping, not a real duplicate set. The only clear duplicate pair is the Qwen3ForCausalLM VRAM leak report; the rest differ by model, API surface, or failure mode.", - "confidence": 0.89, - "canonical_issue_reason": "issue:42673 is the original, more detailed report of the Qwen3ForCausalLM VRAM leak; issue:43704 repeats the same title/symptoms and reads as the duplicate.", + "summary": "This cluster is mostly a set of unrelated issue reports with a few loose topical overlaps. The only clear duplicate is the Qwen3ForCausalLM VRAM leak report, which appears in two identical-titled issues.", + "confidence": 0.96, + "canonical_issue_reason": "issue:42673 is the older, fuller report and the exact duplicate anchor for issue:43704.", "canonical_pr_reason": null, - "best_issue_reason": "issue:42673 is the clearest canonical issue in the cluster because it is the true duplicate anchor and has the fuller report.", + "best_issue_reason": "issue:42673 is the strongest representative issue here: it is specific, well-scoped, and has an exact duplicate counterpart.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43901", - "right": "issue:45003", + "left": "issue:44987", + "right": "issue:45459", "accept": false, - "reason": "Docs mismatch for text classification vs unsafe sys.modules access in modeling_utils; unrelated bugs." + "reason": "Different failures: model loading/version compatibility vs tokenizer protobuf error masking." }, { - "left": "issue:41628", - "right": "issue:41950", + "left": "issue:43901", + "right": "issue:45003", "accept": false, - "reason": "Both mention image processors, but one is an import error and the other is a video pipeline lookup bug; different failure modes." + "reason": "Docs mismatch for a pipeline option is unrelated to unsafe sys.modules access." }, { "left": "issue:42886", "right": "issue:45229", "accept": false, - "reason": "Offline cache loading for tokenizers is unrelated to Gemma4 multi-GPU CUDA OOM." + "reason": "Offline tokenizer cache loading and Gemma4 multi-GPU OOM are unrelated." }, { "left": "issue:44995", "right": "issue:45200", "accept": false, - "reason": "GlmMoeDsa stale indexer cache on second forward is not the same as Gemma4 mm_token_type_ids defaults." + "reason": "A forward-pass crash in GlmMoeDsa is not the same as Gemma4 token-id defaults." + }, + { + "left": "issue:41628", + "right": "issue:41950", + "accept": false, + "reason": "Both mention image processors, but one is an import error and the other is a pipeline selection bug." }, { "left": "issue:43257", "right": "issue:45137", "accept": false, - "reason": "Both involve DeepSpeed, but one is MOE weight conversion and the other is a ZeRO3 deque error; different code paths." + "reason": "Different DeepSpeed problems: weight conversion under accelerate+deepspeed vs a ZeRO3 deque error." }, { "left": "issue:43065", "right": "issue:43408", "accept": false, - "reason": "SAM3 dummy Conv2d vs sam3_video/sam3_tracker config warning are distinct model issues." + "reason": "SAM3 config/type warning and a dummy Conv2d implementation are separate bugs." }, { "left": "issue:43232", "right": "issue:44336", "accept": false, - "reason": "Generation kwargs after sync_gpus is unrelated to ANSI output in loading_report." + "reason": "Generation kwargs handling after sync_gpus is unrelated to ANSI code output in loading_report." }, { "left": "issue:44496", "right": "issue:45003", "accept": false, - "reason": "Unrecognized model/config loading error is unrelated to sys.modules access in modeling_utils." + "reason": "Model config recognition failure is unrelated to unsafe sys.modules access." }, { "left": "issue:43262", "right": "issue:44246", "accept": false, - "reason": "Audio processor sampling-rate default is unrelated to slow import timing." + "reason": "Audio chat-template sampling-rate defaulting is unrelated to import slowdowns." }, { "left": "issue:43295", "right": "issue:45292", "accept": false, - "reason": "Custom processor.tokenizer regression is not the same bug as resize_token_embeddings not updating output embeddings." + "reason": "Processor.tokenizer regression and resize_token_embeddings behavior affect different code paths." }, { "left": "issue:44060", "right": "issue:44077", "accept": false, - "reason": "Tied-weights warning bug and patchtsmixer post_init deprecation are unrelated." + "reason": "Tied-weights warning in Qwen3-Next is not the same as patchtsmixer post_init policy." }, { "left": "issue:43475", "right": "issue:43540", "accept": false, - "reason": "SAM 3 Video missing attribute vs Qwen3OmniMoe video ValueError are different model failures." + "reason": "SAM3 vision encoder output attribute loss and Qwen3OmniMoe video processing are different bugs." }, { "left": "issue:44466", "right": "issue:45335", "accept": false, - "reason": "Serialization inconsistency of tied weights is different from resize_token_embeddings not affecting decoder.embed_tokens." + "reason": "Both involve weight handling, but they are different symptoms on different model families." }, { "left": "issue:43653", "right": "issue:44162", "accept": false, - "reason": "BigBirdTokenizer special-token registration bug is unrelated to ESM2 breakage." + "reason": "BigBirdTokenizer special-token registration and ESM2 breakage are unrelated." }, { "left": "issue:44368", "right": "issue:44977", "accept": false, - "reason": "Tie_word_embeddings warning during LoRA fine-tuning is not the same as flash-attention generation failure." + "reason": "A tied-embeddings warning is not the same as a flash-attention generation failure." }, { "left": "issue:44933", "right": "issue:45459", "accept": false, - "reason": "Missing import from image_utils and protobuf decode-error masking are different import/exception issues." + "reason": "Missing image_utils import and protobuf decode-error handling are unrelated tokenizer/import issues." }, { "left": "issue:43441", "right": "issue:43606", "accept": false, - "reason": "FlashAttention failure in Ministral-3 is unrelated to Bark-small CPU offload device mismatch." + "reason": "FlashAttention failure in Ministral-3 is unrelated to a CPU offload device mismatch in bark-small." }, { "left": "issue:43673", "right": "issue:44734", "accept": false, - "reason": "Missing GenerationMixin cache in chunked_prefill and server KV-cache tensor indexing are both cache-related but not the same concrete bug." + "reason": "Both involve generation/cache behavior, but they hit different paths and failure modes." }, { "left": "issue:44869", "right": "issue:45356", "accept": false, - "reason": "Whisper trailing replacement-character crash is unrelated to Kimi-K2.5 codec handling." + "reason": "Whisper timestamp decoding crash and Kimi tokenizer codec regression are not the same bug." }, { "left": "issue:43232", "right": "issue:43673", "accept": false, - "reason": "Generation kwargs sync_gpus issue and missing cache in chunked_prefill are separate generation bugs." + "reason": "Generation kwargs syncing and chunked-prefill cache missing are distinct generation issues." }, { "left": "issue:43931", "right": "issue:44368", "accept": false, - "reason": "Qwen3-VL weight-shape mismatch is unrelated to tie_word_embeddings warnings." + "reason": "Weight-shape mismatch on model load is unrelated to tie_word_embeddings warnings." }, { "left": "issue:44297", "right": "issue:44623", "accept": false, - "reason": "Tokenizer save_pretrained class mismatch and processor save_pretrained missing files are different serialization defects." + "reason": "Tokenizer save_pretrained metadata mismatch and processor save_pretrained missing files are different serialization bugs." }, { "left": "issue:43646", "right": "issue:44162", "accept": false, - "reason": "Generic custom model initialization breakage does not match the ESM2-specific failure." - }, - { - "left": "issue:43673", - "right": "issue:45278", - "accept": false, - "reason": "Generation cache bug is not the same as broad import errors after version upgrade." + "reason": "Generic v5 custom model initialization breakage is not the same as the ESM2-specific failure." }, { "left": "issue:43381", "right": "issue:45292", "accept": false, - "reason": "Gradient checkpointing in eval mode is unrelated to resize_token_embeddings behavior." - }, - { - "left": "issue:43381", - "right": "issue:45335", - "accept": false, - "reason": "Gradient checkpointing in eval mode is unrelated to t5gemma embedding resizing." + "reason": "Gradient checkpointing in eval mode and resize_token_embeddings behavior do not share a concrete bug." }, { "left": "issue:42673", "right": "issue:43704", "accept": true, - "reason": "Same Qwen3ForCausalLM VRAM leak, same title, same symptom, and the later issue is a duplicate report." + "reason": "Exact same title and same VRAM leak description; clear duplicate pair." }, { - "left": "issue:44361", - "right": "issue:44749", + "left": "issue:28282", + "right": "issue:30990", "accept": false, - "reason": "MLukeTokenizer AttributeError and Chinese import/filtering slowdown are unrelated." + "reason": "PyTorch import absence and Sentence Transformers loading hangs are unrelated." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is well grounded in the packet. The canonical duplicate choice for issue:42673 vs issue:43704 is supported by identical titles and the fuller earlier report, and the soft-edge verdicts stay conservative overall. I don\u2019t see any overstated evidence that would require rejection." + "feedback": "The summary is grounded in the packet: the only clearly duplicated pair is issue:42673 / issue:43704, which match on title and are consistent with the older/fuller canonical choice. The remaining soft-edge verdicts are conservative false calls and the explanations stay within what the packet supports." }, "evaluator_used": true, "retried": false, @@ -9929,7 +9950,6 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-364", "nodes": [ - "issue:28282", "issue:30990", "issue:35141", "issue:36010", @@ -9957,6 +9977,7 @@ "issue:44222", "issue:44242", "issue:44351", + "issue:44361", "issue:44568", "issue:44623", "issue:44661", @@ -9979,273 +10000,273 @@ "issue:45478" ], "soft_pairs": [ - "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:42994|issue:43482", "issue:43475|issue:45381", + "issue:42994|issue:43482", "issue:43475|issue:43643", "issue:43701|issue:44749", - "issue:45081|issue:45103", "issue:42371|issue:45092", + "issue:45081|issue:45103", "issue:45375|issue:45478", "issue:41628|issue:44933", "issue:45310|issue:45375", - "issue:42994|issue:43525", - "issue:43475|issue:44661", "issue:43873|issue:44945", "issue:42222|issue:45200", + "issue:43475|issue:44661", + "issue:42994|issue:43525", "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:41628|issue:44351", "issue:42994|issue:44164", "issue:43232|issue:44792", "issue:44038|issue:45310", "issue:30990|issue:35141", - "issue:43479|issue:44079", "issue:43377|issue:44242", - "issue:36010|issue:41628", + "issue:43479|issue:44079", "issue:43673|issue:44188", - "issue:42898|issue:44568", - "issue:43479|issue:44869" + "issue:36010|issue:41628", + "issue:43479|issue:44869", + "issue:42898|issue:44568" ], - "prepared_review_unit_hash": "2e1580dc417ca862b9b6712231ac31fa5693f1a23d2fe69909c16c95c79e8d50", + "prepared_review_unit_hash": "7062516464194a64f2c5045aee5daadba05a2c3200154981c54d730907c23753", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 34, - "serialized_chars": 37538, - "estimated_input_tokens": 9385, - "estimated_eval_tokens": 19026 + "serialized_chars": 37493, + "estimated_input_tokens": 9374, + "estimated_eval_tokens": 19004 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is not a duplicate set: it mixes many unrelated import, config, generation, tokenizer, quantization, and model-specific regressions. No soft pair is close enough to merge as the same underlying bug/change.", - "confidence": 0.96, - "canonical_issue_reason": "No single issue is a safe canonical duplicate because the items span distinct failure modes and different code paths.", + "summary": "This cluster is mostly heterogeneous: unrelated model-loading, tokenizer, generation, quantization, and multimodal bugs are mixed together. There is one clear near-duplicate theme around Qwen3.5 MoE from_pretrained failures, but the provided soft pairs do not include the strongest duplicate edge.", + "confidence": 0.68, + "canonical_issue_reason": "Best canonical issue is the most concrete and representative bug report in the only clear duplicate theme: the Qwen3.5 MoE from_pretrained failure.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45310 is the most narrowly scoped, concrete loader regression in the set, so it is the best standalone representative, but it does not subsume the others.", + "best_issue_reason": "Strongest overall issue to keep as the cluster representative because it is specific, reproducible, and closest to the only apparent duplicate family in this set.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:28282", - "right": "issue:30990", + "left": "issue:44361", + "right": "issue:44749", "accept": false, - "reason": "PyTorch-missing import error vs Sentence Transformers hanging during load; different symptoms and code paths." + "reason": "Different failures: MLukeTokenizer task AttributeError vs a Chinese-language performance slowdown report; no shared code-path bug." }, { "left": "issue:44077", "right": "issue:45003", "accept": false, - "reason": "Optional post_init policy change vs unsafe sys.modules access; unrelated bugs." + "reason": "One is a patchtsmixer post_init policy issue, the other is a sys.modules access bug; unrelated symptoms and fixes." }, { "left": "issue:44749", "right": "issue:45310", "accept": false, - "reason": "Data filtering slowdown after upgrade vs Qwen3.5 from_pretrained failure; not the same issue." + "reason": "Performance regression in data filtering vs Qwen3.5 MoE from_pretrained error; not the same bug." }, { "left": "issue:45230", "right": "issue:45397", "accept": false, - "reason": "Generic bug report vs Gemma-4 zero3 from_pretrained failure; no shared concrete defect." + "reason": "Generic bug report vs Gemma-4 zero3 from_pretrained failure; insufficiently related." }, { "left": "issue:44038", "right": "issue:44623", "accept": false, - "reason": "Qwen3-VL-Moe loading bug vs processor.save_pretrained missing files; different subsystems." + "reason": "Qwen3-VL-Moe loading bug vs processor.save_pretrained missing files; different code paths." }, { - "left": "issue:42994", - "right": "issue:43482", + "left": "issue:43475", + "right": "issue:45381", "accept": false, - "reason": "Quantized model saving failure vs Qwen2.5-GGUF loading failure; unrelated." + "reason": "Both are video/VL issues, but for different models and different attributes/position-id logic." }, { - "left": "issue:43475", - "right": "issue:45381", + "left": "issue:42994", + "right": "issue:43482", "accept": false, - "reason": "SAM 3 video attribute error vs qwen2.5-vl video position-id bug; different models and paths." + "reason": "Quantized saving failure vs GGUF loading failure; opposite directions and different model formats." }, { "left": "issue:43475", "right": "issue:43643", "accept": false, - "reason": "SAM 3 VisionEncoder output bug vs AutoConfig trust_remote_code field loss; unrelated." + "reason": "SAM 3 Video attribute error vs AutoConfig missing fields with trust_remote_code; unrelated root causes." }, { "left": "issue:43701", "right": "issue:44749", "accept": false, - "reason": "resume_from_checkpoint key mismatch vs performance regression in data filtering; not the same bug." + "reason": "Checkpoint key mismatch vs data filtering slowdown; no overlap in bug mechanism." }, { - "left": "issue:45081", - "right": "issue:45103", + "left": "issue:42371", + "right": "issue:45092", "accept": false, - "reason": "Mistral tokenizer patch crash vs auto_docstring annotation crash; unrelated code paths." + "reason": "TF32 settings warning vs InternVL2 meta-init incompatibility; different subsystems." }, { - "left": "issue:42371", - "right": "issue:45092", + "left": "issue:45081", + "right": "issue:45103", "accept": false, - "reason": "TF32 settings warning vs remote-code meta-init incompatibility; different problems." + "reason": "Tokenizer regex patch crash vs auto_docstring annotations crash; different failures and files." }, { "left": "issue:45375", "right": "issue:45478", "accept": false, - "reason": "Missing vision-config field in strict handling vs Qwen3.5 Moe from_pretrained error; related family but not the same defect." + "reason": "Missing vision-config field vs generic Qwen3.5 MoE from_pretrained error; related model family but not the same concrete bug." }, { "left": "issue:41628", "right": "issue:44933", "accept": false, - "reason": "AutoImageProcessor export error vs missing image_utils import; similar theme, different symbols and fixes." + "reason": "Both are import problems, but they concern different symbols and likely different export mistakes." }, { "left": "issue:45310", "right": "issue:45375", "accept": false, - "reason": "Both concern Qwen3.5-family models, but one is a from_pretrained loader failure and the other is a missing config field; not mergeable as one bug." + "reason": "Qwen3.5 MoE from_pretrained error and a missing vision-config field are related only by model family, not the same bug." }, { - "left": "issue:42994", - "right": "issue:43525", + "left": "issue:43873", + "right": "issue:44945", "accept": false, - "reason": "Quantized save failure vs Llama4Config missing pad_token_id; unrelated." + "reason": "Quantization offloading issue vs pipeline parallelism output correctness; different runtime paths." }, { - "left": "issue:43475", - "right": "issue:44661", + "left": "issue:42222", + "right": "issue:45200", "accept": false, - "reason": "SAM 3 attribute error vs add-new-model-like/tokenizer mapping failure; no shared code path." + "reason": "Vitpose model breakage vs Gemma 4 token-type-id behavior; unrelated models and features." }, { - "left": "issue:43873", - "right": "issue:44945", + "left": "issue:43475", + "right": "issue:44661", "accept": false, - "reason": "Quantization offloading issue vs pipeline-parallel output mismatch; different runtime behaviors." + "reason": "SAM 3 Video attribute error vs tokenizer mapping registration failure; unrelated." }, { - "left": "issue:42222", - "right": "issue:45200", + "left": "issue:42994", + "right": "issue:43525", "accept": false, - "reason": "vitpose model breakage vs Gemma 4 token-type defaulting; unrelated model families." + "reason": "Quantized save failure vs Llama4Config missing pad_token_id; different objects and symptoms." }, { "left": "issue:42831", "right": "issue:44222", "accept": false, - "reason": "FP8 accuracy regression vs FP8 save_pretrained moe; different failure modes." + "reason": "FP8 accuracy issue vs FP8 save_pretrained/moe issue; both mention FP8 but not the same concrete bug." }, { "left": "issue:45362", "right": "issue:45375", "accept": false, - "reason": "Qwen3.5-35B chat crash vs missing deepstack_visual_indexes field; not the same bug." + "reason": "Qwen3.5 chat crash vs missing deepstack_visual_indexes field; different failure modes." }, { "left": "issue:45310", "right": "issue:45397", "accept": false, - "reason": "Qwen3.5 Moe from_pretrained error vs Gemma-4 zero3 from_pretrained issue; unrelated models." + "reason": "Both mention from_pretrained, but for different models and different incompatibilities." }, { "left": "issue:43232", "right": "issue:44188", "accept": false, - "reason": "Generation kwargs update bug vs torch.compile attention-kernel divergence; different code paths." + "reason": "Generation kwargs update bug vs attention kernel divergence under torch.compile; unrelated." }, { - "left": "issue:41628", - "right": "issue:44351", + "left": "issue:44222", + "right": "issue:45310", "accept": false, - "reason": "AutoImageProcessor import error vs HybridCache import error; both imports, but not the same missing symbol or fix." + "reason": "FP8 save_pretrained moe issue vs Qwen3.5 MoE from_pretrained error; same broad family, different concrete path." }, { - "left": "issue:44222", - "right": "issue:45310", + "left": "issue:41628", + "right": "issue:44351", "accept": false, - "reason": "FP8 save_pretrained moe vs Qwen3.5 Moe from_pretrained error; unrelated." + "reason": "Both are missing-import reports, but for different classes and different API surfaces." }, { "left": "issue:42994", "right": "issue:44164", "accept": false, - "reason": "Quantized saving failure vs extra_state save/from_pretrained handling; different persistence bugs." + "reason": "Quantized model saving failure vs extra_state serialization/deserialization failure; not the same bug." }, { "left": "issue:43232", "right": "issue:44792", "accept": false, - "reason": "Generation sync_gpus kwargs bug vs janus test failure; no shared underlying defect." + "reason": "Generation kwargs handling vs a failed janus image-generation test; no shared underlying issue." }, { "left": "issue:44038", "right": "issue:45310", "accept": false, - "reason": "Qwen3-VL-Moe bug vs Qwen3.5 Moe from_pretrained error; same broad family, but not the same concrete failure." + "reason": "Both are Qwen MoE-related, but one is a Qwen3-VL-Moe bug and the other is a Qwen3.5 MoE from_pretrained failure; too different to merge." }, { "left": "issue:30990", "right": "issue:35141", "accept": false, - "reason": "Loading stall vs embedding resizing/post_init reinitialization; unrelated." - }, - { - "left": "issue:43479", - "right": "issue:44079", - "accept": false, - "reason": "Phi4Multimodal config defaults vs ModelOutput key assignment; different object semantics." + "reason": "Sentence Transformers loading hang vs embedding resize/post_init reinitialization; unrelated." }, { "left": "issue:43377", "right": "issue:44242", "accept": false, - "reason": "MIMI batched-vs-single padding-mask bug vs MoE load-balancing loss omission; unrelated." + "reason": "MIMI padding-mask mismatch vs MoE load-balancing loss omission; different model components and symptoms." }, { - "left": "issue:36010", - "right": "issue:41628", + "left": "issue:43479", + "right": "issue:44079", "accept": false, - "reason": "GenerationMixin import missing vs AutoImageProcessor import missing; both are import regressions, but they concern different exports and fixes." + "reason": "Config initialization with None vs ModelOutput key assignment bug; different object layers." }, { "left": "issue:43673", "right": "issue:44188", "accept": false, - "reason": "GenerationMixin cache missing in chunked prefill vs torch.compile branching divergence; not the same code-path problem." + "reason": "Missing GenerationMixin cache in v5 chunked prefill vs attention-kernel divergence under compile; unrelated." }, { - "left": "issue:42898", - "right": "issue:44568", + "left": "issue:36010", + "right": "issue:41628", "accept": false, - "reason": "v5 clean_up_tokenization_spaces behavior change vs missing BOS/EOS when add_special_tokens=True; different tokenizer regressions." + "reason": "Both are import errors, but they involve different removed/missing symbols and different fixes." }, { "left": "issue:43479", "right": "issue:44869", "accept": false, - "reason": "Config initialization bug vs Whisper word timestamp decode crash; unrelated." + "reason": "Multimodal config defaulting bug vs Whisper timestamp decode crash; unrelated." + }, + { + "left": "issue:42898", + "right": "issue:44568", + "accept": false, + "reason": "Both are tokenizer behavior changes in v5, but they are different regressions affecting different APIs and models." } ] }, @@ -10329,263 +10350,269 @@ "issue:44060|issue:45125", "issue:44734|issue:45103", "issue:42898|issue:45356", - "issue:43723|issue:44568", "issue:45137|issue:45276", + "issue:43723|issue:44568", "issue:42898|issue:44779", "issue:43874|issue:45092", "issue:44077|issue:44242", "issue:42898|issue:44448", "issue:41628|issue:45459", "issue:44222|issue:44297", - "issue:43479|issue:43644", "issue:42898|issue:43295", + "issue:43479|issue:43644", "issue:44442|issue:45459", "issue:44222|issue:44623", "issue:42831|issue:44623", "issue:44206|issue:45103", "issue:22355|issue:28282", "issue:44297|issue:45310", - "issue:41628|issue:42757", "issue:44483|issue:45406", + "issue:41628|issue:42757", "issue:44222|issue:44484", "issue:44263|issue:44360", "issue:43723|issue:44977", "issue:44263|issue:45092", - "issue:44514|issue:45103", - "issue:44295|issue:44623" + "issue:44295|issue:44623", + "issue:44514|issue:45103" ], - "prepared_review_unit_hash": "44cde44f61b66aba35f1964e9338ffb340a6e9acfd71528fb171ec472ba92471", + "prepared_review_unit_hash": "69ad125856676450611c28f1a50bf050cd8281176920b04d5710d56f19de4c2b", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 34, - "serialized_chars": 37829, - "estimated_input_tokens": 9458, - "estimated_eval_tokens": 19172 + "serialized_chars": 37842, + "estimated_input_tokens": 9461, + "estimated_eval_tokens": 19178 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose semantic neighborhood of unrelated Transformers issues, not true duplicates. The shared wording (tokenizer, save_pretrained, import, v5 regression, multimodal) is superficial and usually points to different code paths or different models/components.", - "confidence": 0.96, - "canonical_issue_reason": "issue:42898 is the broadest, most general regression in the set and has the clearest cross-cutting impact; it is also more central than the highly model-specific or packaging-only reports.", + "summary": "This cluster is largely heterogeneous: mostly unrelated Transformers import, tokenizer, processor, and model-regression issues. None of the soft pairs look like the same concrete bug/change, so I would not merge any of them as duplicates.", + "confidence": 0.94, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:42898 is the best representative issue for the cluster because it describes a concrete Transformers v5 behavior regression with wider applicability than the other narrowly scoped bugs.", + "best_issue_reason": "issue:42898 is the strongest representative of the cluster\u2019s recurring theme of Transformers v5 tokenizer behavior regressions, but it is still not a true duplicate anchor for the whole set.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44964", "right": "issue:45137", "accept": false, - "reason": "Different problems: Phi-4 multimodal loading vs a DeepSpeed ZeRO3 deque/indexing failure." + "reason": "Different failures: Phi-4 multimodal model loading vs DeepSpeed ZeRO3 deque underflow." }, { "left": "issue:43976", "right": "issue:44933", "accept": false, - "reason": "Python version/backend support issue vs an import name missing from image_utils; unrelated." + "reason": "Python-version incompatibility is unrelated to the missing image_utils import." }, { "left": "issue:42175", "right": "issue:44933", "accept": false, - "reason": "Packaging/backend dependency omission vs missing image_utils import; not the same bug." + "reason": "Backend packaging for torch/tensorflow is unrelated to an import from image_utils." }, { "left": "issue:43784", "right": "issue:43824", "accept": false, - "reason": "Sentence-transformers import NameError vs missing Qwen2.5-VL class export; different failure modes." + "reason": "Sentence-transformers import NameError and missing Qwen2.5-VL export are different code paths." }, { "left": "issue:43278", "right": "issue:45137", "accept": false, - "reason": "Training/eval dtype mismatch vs DeepSpeed ZeRO3 empty deque error; unrelated code paths." + "reason": "BF16->FP32 eval dtype drift is unrelated to ZeRO3 deque errors." }, { "left": "issue:44623", "right": "issue:45310", "accept": false, - "reason": "Processor save_pretrained file emission bug vs Qwen3.5 MoE from_pretrained regression." + "reason": "Processor save_pretrained missing files is not the same as Qwen3.5-MoE from_pretrained failure." }, { "left": "issue:44188", "right": "issue:44373", "accept": false, - "reason": "Attention kernel divergence under torch.compile vs a docstring issue; not a duplicate." + "reason": "torch.compile branching bug and a docstring issue are unrelated." }, { "left": "issue:44060", "right": "issue:45125", "accept": false, - "reason": "Incorrect tied-weights warning vs missing _tp_plan for tensor parallelism; separate model bugs." + "reason": "Wrong tied-weights warning is not the same as a missing _tp_plan attribute." }, { "left": "issue:44734", "right": "issue:45103", "accept": false, - "reason": "Serving KV-cache indexing crash vs auto_docstring crash from future annotations; different subsystems." + "reason": "KV-cache indexing crash and auto_docstring AttributeError are unrelated." }, { "left": "issue:42898", "right": "issue:45356", "accept": false, - "reason": "Tokenizer cleanup behavior change vs Kimi codec handling / warning regression; not the same underlying bug." + "reason": "Both are tokenizer regressions, but they are different bugs with different symptoms and fixes." }, { - "left": "issue:43723", - "right": "issue:44568", + "left": "issue:45137", + "right": "issue:45276", "accept": false, - "reason": "AutoTokenizer loading failure vs BOS/EOS not being added; both tokenizer-related but distinct." + "reason": "ZeRO3 deque underflow is unrelated to gemma4 resize_token_embeddings propagation." }, { - "left": "issue:45137", - "right": "issue:45276", + "left": "issue:43723", + "right": "issue:44568", "accept": false, - "reason": "ZeRO3 deque/index failure vs resize_token_embeddings not updating Gemma4 embeddings; unrelated." + "reason": "Tokenizer loading failure and BOS/EOS add_special_tokens behavior are distinct issues." }, { "left": "issue:42898", "right": "issue:44779", "accept": false, - "reason": "clean_up_tokenization_spaces regression vs Deepseek tokenizer output regression; different tokenizer bugs." + "reason": "clean_up_tokenization_spaces behavior change is not the same as Deepseek tokenizer output corruption." }, { "left": "issue:43874", "right": "issue:45092", "accept": false, - "reason": "Missing image-patch method vs remote-code/meta-init incompatibility; different multimodal issues." + "reason": "Missing image patch count method is unrelated to remote-code checkpoint meta-init compatibility." }, { "left": "issue:44077", "right": "issue:44242", "accept": false, - "reason": "patchtsmixer post_init API restriction vs MoE load-balancing loss omission; unrelated." + "reason": "Optional post_init policy and missing MoE load-balancing loss are different implementation bugs." }, { "left": "issue:42898", "right": "issue:44448", "accept": false, - "reason": "Tokenization cleanup behavior change vs Pegasus output mismatch; not the same defect." + "reason": "Token cleanup regression and Pegasus output drift are not the same concrete bug." }, { "left": "issue:41628", "right": "issue:45459", "accept": false, - "reason": "Missing AutoImageProcessor import vs protobuf-related tokenizer error masking; different import paths." + "reason": "Missing AutoImageProcessor import and protobuf error masking are unrelated." }, { "left": "issue:44222", "right": "issue:44297", "accept": false, - "reason": "FP8 save_pretrained/MoE issue vs tokenizer_class metadata mismatch; unrelated." + "reason": "FP8 save_pretrained/moe issues do not match tokenizer_class metadata mismatch on save_pretrained." }, { - "left": "issue:43479", - "right": "issue:43644", + "left": "issue:42898", + "right": "issue:43295", "accept": false, - "reason": "Phi4MultimodalConfig default initialization bug vs non-persistent buffer junk in v5; different layers of the stack." + "reason": "Tokenizer cleanup change is unrelated to processor.tokenizer regression in v4.57.5." }, { - "left": "issue:42898", - "right": "issue:43295", + "left": "issue:43479", + "right": "issue:43644", "accept": false, - "reason": "Tokenization cleanup regression vs processor.tokenizer / image passing regression; separate processor changes." + "reason": "Default multimodal config initialization and junk non-persistent buffers are different bugs." }, { "left": "issue:44442", "right": "issue:45459", "accept": false, - "reason": "FastSpeech2ConformerTokenizer loading failure vs hidden tokenizer errors when protobuf is absent; not a duplicate." + "reason": "FastSpeech2ConformerTokenizer loading and protobuf error handling are different failure modes." }, { "left": "issue:44222", "right": "issue:44623", "accept": false, - "reason": "FP8 save_pretrained/MoE bug vs processor.save_pretrained missing files; different save paths and objects." + "reason": "MoE FP8 save_pretrained and processor.save_pretrained missing files are not the same bug." }, { "left": "issue:42831", "right": "issue:44623", "accept": false, - "reason": "FineGrainedFP8 accuracy problem vs processor.save_pretrained file omission; unrelated." + "reason": "FP8 accuracy regression is unrelated to processor.save_pretrained file omission." }, { "left": "issue:44206", "right": "issue:45103", "accept": false, - "reason": "Unsupported center argument in feature extractor vs auto_docstring annotations crash; different failures." + "reason": "FeatureExtractor center-arg crash and auto_docstring annotation handling are unrelated." }, { "left": "issue:22355", "right": "issue:28282", "accept": false, - "reason": "Missing transformers.onnx module import vs AutoModel requiring PyTorch; not the same import error." + "reason": "onnx module import failure is unrelated to missing PyTorch backend detection." }, { "left": "issue:44297", "right": "issue:45310", "accept": false, - "reason": "Tokenizer config metadata mismatch vs Qwen3.5 MoE from_pretrained error; separate issues." + "reason": "Tokenizer config mismatch on save and Qwen3.5-MoE from_pretrained failure are different issues." }, { - "left": "issue:41628", - "right": "issue:42757", + "left": "issue:44483", + "right": "issue:45406", "accept": false, - "reason": "AutoImageProcessor import failure vs is_offline_mode import failure from huggingface_hub; unrelated imports." + "reason": "Chat/completions request rejection and Gemma4Processor _tokenizer crash are different server/processor bugs." }, { - "left": "issue:44483", - "right": "issue:45406", + "left": "issue:41628", + "right": "issue:42757", "accept": false, - "reason": "chat/completions request handling vs Gemma4Processor missing _tokenizer; different serve failures." + "reason": "Transformer export import error and huggingface_hub offline_mode import error are unrelated." }, { "left": "issue:44222", "right": "issue:44484", "accept": false, - "reason": "FP8 save_pretrained/MoE regression vs max_shard_size default question; not a duplicate." + "reason": "FP8 save_pretrained bug and a max_shard_size question are not the same change." }, { "left": "issue:44263", "right": "issue:44360", "accept": false, - "reason": "torch.split return values issue vs DSA indexer lacking ReLU; same area, but different concrete bug." + "reason": "These are both DSA-indexer-related, but one is about torch.split outputs and the other about a missing ReLU; not the same bug." }, { "left": "issue:44295", - "right": "issue:45459", + "right": "issue:45310", "accept": false, - "reason": "position_ids buffer read error vs protobuf import masking tokenizer errors; unrelated." + "reason": "position_ids buffer access error is unrelated to Qwen3.5-MoE from_pretrained failure." }, { - "left": "issue:44263", - "right": "issue:45092", + "left": "issue:43723", + "right": "issue:44977", "accept": false, - "reason": "GlmMoeDsaIndexer split behavior vs InternVL2 meta-initialization incompatibility; different components." + "reason": "Tokenizer loading issue is unrelated to flash-attention generation regressions." }, { - "left": "issue:44514", - "right": "issue:45103", + "left": "issue:44263", + "right": "issue:45092", "accept": false, - "reason": "Batched apply_chat_template padding bug vs auto_docstring future-annotations crash; unrelated." + "reason": "GlmMoeDsaIndexer split handling and InternVL2 meta-init incompatibility are unrelated." }, { "left": "issue:44295", "right": "issue:44623", "accept": false, - "reason": "buffer registration/read error vs missing processor save files; different bugs." + "reason": "position_ids buffer error is unrelated to processor.save_pretrained missing files." + }, + { + "left": "issue:44514", + "right": "issue:45103", + "accept": false, + "reason": "Qwen2_5_VLProcessor batched apply_chat_template crash and auto_docstring AttributeError are unrelated." } ] }, @@ -10659,11 +10686,11 @@ ], "soft_pairs": [ "issue:43479|issue:44913", - "issue:43723|issue:44871", "issue:44373|issue:44485", + "issue:43723|issue:44871", "issue:44261|issue:45092", - "issue:43723|issue:44991", "issue:30990|issue:38175", + "issue:43723|issue:44991", "issue:43723|issue:44442", "issue:45092|issue:45357", "issue:42886|issue:42994", @@ -10672,231 +10699,225 @@ "issue:42831|issue:44038", "issue:43576|issue:43784", "issue:43122|issue:44521", - "issue:42757|issue:45278", "issue:44263|issue:44373", + "issue:42757|issue:45278", "issue:44117|issue:44704", "issue:42886|issue:43066", "issue:43122|issue:44514", "issue:30990|issue:33357", "issue:44704|issue:45290", "issue:43232|issue:45245", + "issue:43232|issue:44743", "issue:42371|issue:44263", "issue:42907|issue:44945", - "issue:43232|issue:44743", "issue:44295|issue:44297", "issue:43976|issue:44945", "issue:39401|issue:44488", "issue:42907|issue:43937", "issue:43408|issue:43475" ], - "prepared_review_unit_hash": "09652ddf11c18f0946e8e7e087db6ed28985595bd08192c20c6f5ab5045f217e", + "prepared_review_unit_hash": "3a9cc870cdf2e8bd41d9efcc058a302167f345f6a8ddfd1b97df4264991d11b5", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 30, - "serialized_chars": 35658, - "estimated_input_tokens": 8915, - "estimated_eval_tokens": 18086 + "serialized_chars": 35656, + "estimated_input_tokens": 8914, + "estimated_eval_tokens": 18084 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose theme of Transformers v5 regressions, but the paired items point to different concrete bugs, models, or API surfaces. I would not merge any of the soft-edge pairs as duplicates.", - "confidence": 0.93, - "canonical_issue_reason": "Issue 43723 is the best representative issue for the cluster because it captures the broad tokenizer-loading regression theme seen in several nearby reports, even though the overall cluster is not a true duplicate set.", + "summary": "These items are mostly unrelated single-issue reports across tokenizer loading, model config regressions, quantization/saving, import errors, and a few model-specific bugs. No soft pair is a clear same-bug duplicate; most share only a broad subsystem or symptom family.", + "confidence": 0.82, + "canonical_issue_reason": "issue:43723 is the best general anchor: it is a concrete, version-scoped tokenizer-loading regression with a broad symptom that matches the cluster\u2019s repeated loading/import complaints more than the other, narrower reports.", "canonical_pr_reason": null, - "best_issue_reason": "43723 is the most generally representative issue title in this set for v5 loading regressions, especially around tokenizer / from_pretrained behavior.", + "best_issue_reason": "issue:43723 is the strongest standalone issue to represent this set because it is specific, reproducible, and broadly relevant to the tokenizer-loading regressions seen here.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43479", "right": "issue:44913", "accept": false, - "reason": "Different bugs: default sub-config initialization vs rotary_pct not persisting on reload." + "reason": "Different bugs: one is wrong default vision/audio config initialization in Phi4MultimodalConfig, the other is GPTNeoX rotary_pct not persisting on reload." }, { - "left": "issue:43723", - "right": "issue:44871", + "left": "issue:44373", + "right": "issue:44485", "accept": false, - "reason": "Both involve config/loading behavior, but one is a tokenizer-loading regression and the other is an eos_token_id inconsistency in Gemma-3." + "reason": "Docstring typo versus RoPE implementation behavior; no shared concrete failure." }, { - "left": "issue:44373", - "right": "issue:44485", + "left": "issue:43723", + "right": "issue:44871", "accept": false, - "reason": "Docstring fix for position_ids is unrelated to GLM-5 RoPE implementation behavior." + "reason": "Both concern tokenizer/model config behavior, but one is a tokenizer loading failure and the other is an eos_token_id mismatch." }, { "left": "issue:44261", "right": "issue:45092", "accept": false, - "reason": "Different model families and code paths: MLA layernorm epsilon precision issue vs InternVL2 meta-init compatibility." + "reason": "Different code paths and symptoms: MLA rms_norm_eps precision issue versus remote-code/meta-init incompatibility." }, { - "left": "issue:43723", - "right": "issue:44991", + "left": "issue:30990", + "right": "issue:38175", "accept": false, - "reason": "Both are tokenizer load failures, but the reports describe different models and likely different root causes." + "reason": "Both are model-related user-facing failures, but one is a loading hang and the other is unexpected zero probabilities on a specific model." }, { - "left": "issue:30990", - "right": "issue:38175", + "left": "issue:43723", + "right": "issue:44991", "accept": false, - "reason": "Sentence Transformers loading hang and SIGLIP2 zero probabilities are unrelated." + "reason": "Both are tokenizer-loading regressions, but they appear model-specific and not the same concrete bug." }, { "left": "issue:43723", "right": "issue:44442", "accept": false, - "reason": "AutoTokenizer failing for a specific tokenizer is not the same as the broader tokenizer-loading regression in 43723." + "reason": "AutoTokenizer loading regression versus a specific FastSpeech2ConformerTokenizer load failure; too little evidence they are the same defect." }, { "left": "issue:45092", "right": "issue:45357", "accept": false, - "reason": "Both are Qwen-related, but one is meta initialization incompatibility and the other is save_pretrained writing wrong visual encoder keys." + "reason": "Both involve model compatibility/regression, but one is meta initialization for old InternVL2 checkpoints and the other is incorrect save_pretrained key handling in Qwen3.5." }, { "left": "issue:42886", "right": "issue:42994", "accept": false, - "reason": "Offline cache tokenizer loading and quantized model saving are different operations and failure modes." + "reason": "Offline tokenizer cache loading failure is not the same as quantized model saving failure." }, { "left": "issue:42831", "right": "issue:45310", "accept": false, - "reason": "FineGrainedFP8 accuracy regression is not the same bug as Qwen3.5 from_pretrained loading failure." + "reason": "FineGrainedFP8 accuracy degradation and Qwen3.5 from_pretrained errors are different problems." }, { "left": "issue:41628", "right": "issue:43824", "accept": false, - "reason": "Different missing imports for different symbols; same broad API-surface area, but not the same underlying bug." + "reason": "Both are import errors, but they concern different missing symbols and different model/component paths." }, { "left": "issue:42831", "right": "issue:44038", "accept": false, - "reason": "These are different Qwen/transformers issues with different symptoms and likely different fixes." + "reason": "Shared Qwen context is too broad; one is FP8 accuracy and the other is a model-specific from_pretrained bug." }, { "left": "issue:43576", "right": "issue:43784", "accept": false, - "reason": "transformers env command and sentence-transformers import NameError are unrelated." + "reason": "transformers env command breakage and sentence-transformers import NameError are distinct failures." }, { "left": "issue:43122", "right": "issue:44521", "accept": false, - "reason": "Tokenization differences across versions and all-zero assistant masks are distinct behaviors in different code paths." + "reason": "Different tokenizer/template issues: changed tokenization behavior versus all-zero assistant masks for multimodal inputs." }, { - "left": "issue:42757", - "right": "issue:45278", + "left": "issue:44263", + "right": "issue:44373", "accept": false, - "reason": "A missing huggingface_hub symbol and a broad set of import errors after upgrade are not the same concrete issue." + "reason": "Tensor split/indexing bug versus a docstring problem; not the same issue." }, { - "left": "issue:44263", - "right": "issue:44373", + "left": "issue:42757", + "right": "issue:45278", "accept": false, - "reason": "torch.split return handling in GlmMoeDsaIndexer is unrelated to a wrong docstring for position_ids." + "reason": "Specific missing import from huggingface_hub versus a broad set of import errors after upgrade." }, { "left": "issue:44117", "right": "issue:44704", "accept": false, - "reason": "TOKENIZER_MAPPING_NAMES returning None and AutoProcessor kwargs forwarding are different loading bugs." + "reason": "Tokenizer mapping returning None and AutoProcessor kwargs passthrough are different code paths." }, { "left": "issue:42886", "right": "issue:43066", "accept": false, - "reason": "Offline tokenizer cache loading is unrelated to wrong tokenizer decoder type in v5." + "reason": "Offline cache loading failure and wrong tokenizer decoder type are unrelated tokenizer issues." }, { "left": "issue:43122", "right": "issue:44514", "accept": false, - "reason": "General tokenization changes are not the same as batched chat template crash with padding=False." - }, - { - "left": "issue:30990", - "right": "issue:33357", - "accept": false, - "reason": "Sentence Transformers loading hang and MacOS bus error on CLIP model are unrelated failures." + "reason": "Tokenization differences across versions versus apply_chat_template crashing on batched padded=False input." }, { "left": "issue:44704", "right": "issue:45290", "accept": false, - "reason": "AutoProcessor kwargs forwarding and chat_template crash with tool-call messages affect different code paths." + "reason": "AutoProcessor cached_file kwargs bug and assistant/tool-call chat-template crash are different concrete failures." }, { "left": "issue:43232", "right": "issue:45245", "accept": false, - "reason": "Generation kwargs update logic and category-cardinality runtime error are unrelated." + "reason": "Generation kwargs/sync_gpus bug versus a categories-limit runtime error." + }, + { + "left": "issue:43232", + "right": "issue:44743", + "accept": false, + "reason": "Both are generation/model-runtime bugs, but they affect different code paths and symptoms." }, { "left": "issue:42371", "right": "issue:44263", "accept": false, - "reason": "TF32 settings guidance and GlmMoeDsaIndexer split behavior are unrelated." + "reason": "TF32 configuration warning and torch.split return-value issue are unrelated." }, { "left": "issue:42907", "right": "issue:44945", "accept": false, - "reason": "Saving dequantized models and incorrect output under pipeline parallelism are different bugs." - }, - { - "left": "issue:43232", - "right": "issue:44743", - "accept": false, - "reason": "Both touch generation/caching, but one is a sync_gpus kwargs issue and the other is recurrent state reset in modular_qwen3_5.py." + "reason": "Saving dequantized models and incorrect output under pipeline parallelism are not the same bug." }, { "left": "issue:44295", "right": "issue:44297", "accept": false, - "reason": "Reading position_ids after buffer registration and tokenizer_class mismatch on save_pretrained are unrelated." + "reason": "Reading a registered position_ids buffer and tokenizer_class mismatch in tokenizer_config.json are separate issues." }, { "left": "issue:43976", "right": "issue:44945", "accept": false, - "reason": "Python version compatibility and pipeline-parallel output correctness are not the same issue." + "reason": "Python version compatibility and pipeline-parallel output corruption are unrelated." }, { "left": "issue:39401", "right": "issue:44488", "accept": false, - "reason": "Offset_mapping bug in Qwen3 tokenizer and failure to load a specific model checkpoint are not the same underlying problem." + "reason": "Tokenizer offset_mapping bug and failure to load a specific model are different problems." }, { "left": "issue:42907", "right": "issue:43937", "accept": false, - "reason": "Quantized model saving and invalid GenerationConfig for GLM-5 are unrelated." + "reason": "Model saving failure and invalid GenerationConfig are unrelated." }, { "left": "issue:43408", "right": "issue:43475", "accept": false, - "reason": "Model-type warning on load and missing fpn_position_embeddings are different SAM 3 issues." + "reason": "Model-type warning during instantiation and missing fpn_position_embeddings attribute are different SAM3 video bugs." } ] }, @@ -10972,243 +10993,225 @@ "soft_pairs": [ "issue:44625|issue:44829", "issue:29127|issue:30990", - "issue:41628|issue:45278", "issue:44360|issue:44485", + "issue:41628|issue:45278", "issue:43116|issue:45200", "issue:39401|issue:44521", "issue:44038|issue:44297", - "issue:43122|issue:44779", "issue:44261|issue:44373", + "issue:43122|issue:44779", "issue:42371|issue:42831", "issue:44038|issue:44295", + "issue:36296|issue:43116", "issue:42907|issue:43452", "issue:43992|issue:44704", - "issue:36296|issue:43116", "issue:44038|issue:44222", - "issue:43122|issue:44977", "issue:43278|issue:44945", + "issue:43122|issue:44977", "issue:45216|issue:45478", "issue:43824|issue:44351", "issue:42907|issue:44222", "issue:42757|issue:44351", - "issue:42907|issue:44912", "issue:42222|issue:43441", - "issue:43784|issue:45446", + "issue:42907|issue:44912", "issue:42673|issue:45072", + "issue:43784|issue:45446", "issue:43064|issue:43673", + "issue:22355|issue:29127", "issue:43550|issue:43784", "issue:38617|issue:45278", - "issue:22355|issue:29127", "issue:43937|issue:44485" ], - "prepared_review_unit_hash": "02fef8d29295dea7d9c150fa3e082e69e8aaee1c808fe7307d971b1d6117ea8d", + "prepared_review_unit_hash": "5a78981a98a71c85d5e70a9665cf595b20e3d0275e045b5b4fc7f7814d5d4199", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36239, - "estimated_input_tokens": 9060, - "estimated_eval_tokens": 18376 + "serialized_chars": 36231, + "estimated_input_tokens": 9058, + "estimated_eval_tokens": 18372 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "No true duplicate pairs stand out here; this is a noisy mix of unrelated Transformers bugs. I selected the broadest umbrella issue as the best representative of the cluster.", - "confidence": 0.62, - "canonical_issue_reason": "Issue 45278 is the broadest umbrella in the set, covering multiple import errors after upgrading to 5.5.0, so it is the closest representative issue for this noisy cluster.", + "summary": "This set is very heterogeneous: most items are unrelated model-specific bugs, import regressions, or API/documentation issues. None of the soft pairs look like true duplicates; they mostly share only a broad theme (imports, tokenizers, quantization, or training) rather than the same concrete failure mode.", + "confidence": 0.89, + "canonical_issue_reason": "issue:45278 is the broadest regression report in the set and the most plausible umbrella representative, but it still does not actually unify the other issues.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45278 is the most general and cluster-representative report; it can subsume several of the import-error style items better than the more specific bugs.", + "best_issue_reason": "issue:45278 is the strongest cluster representative because it describes a wide import-regression symptom after upgrade, which is more general than the other narrowly scoped reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44625", "right": "issue:44829", "accept": false, - "reason": "Different model/task bugs: num_labels propagation vs flash-attention training collapse." + "reason": "Different bugs: Qwen3.5 config propagation vs flash-attention-3 training degeneration." }, { "left": "issue:29127", "right": "issue:30990", "accept": false, - "reason": "Different subsystems and symptoms: layoutlmv3 error messaging vs sentence-transformers loading hang." + "reason": "Unrelated: layoutlmv3 error-message clarity vs Sentence Transformers loading hang." }, { - "left": "issue:41628", - "right": "issue:45278", + "left": "issue:44360", + "right": "issue:44485", "accept": false, - "reason": "Both are import-related, but one is a specific missing symbol and the other is a broad multi-import regression; not the same concrete bug." + "reason": "Different code paths: DSA indexer/ReLU concern vs GLM-5 RoPE implementation." }, { - "left": "issue:44360", - "right": "issue:44485", + "left": "issue:41628", + "right": "issue:45278", "accept": false, - "reason": "Unrelated topics: DSA indexer/ReLU vs GLM-5 RoPE implementation." + "reason": "One is a specific missing import, the other is a broad upgrade-time import regression; not the same underlying failure." }, { "left": "issue:43116", "right": "issue:45200", "accept": false, - "reason": "Different functionality: example-script multi-label empty results vs Gemma 4 multimodal token-id defaults." + "reason": "Different symptoms and subsystems: example-script multi-label output vs Gemma 4 multimodal token IDs." }, { "left": "issue:39401", "right": "issue:44521", "accept": false, - "reason": "Different tokenizer bugs: offset_mapping for Qwen3 vs multimodal assistant_masks in apply_chat_template." + "reason": "Tokenizer offset mapping bug vs chat-template assistant mask bug." }, { "left": "issue:44038", "right": "issue:44297", "accept": false, - "reason": "One is a Qwen3-VL-Moe model issue; the other is tokenizer save metadata mismatch. Different bugs." + "reason": "Qwen3-VL-Moe runtime/import behavior vs tokenizer save metadata mismatch." }, { - "left": "issue:43122", - "right": "issue:44779", + "left": "issue:44261", + "right": "issue:44373", "accept": false, - "reason": "Both mention tokenization regressions, but they affect different models and likely different code paths." + "reason": "MLA RMS-norm precision/config bug vs a docstring issue for position_ids." }, { - "left": "issue:44261", - "right": "issue:44373", + "left": "issue:43122", + "right": "issue:44779", "accept": false, - "reason": "Precision/config bug vs a docstring issue; not the same underlying change." + "reason": "Both are tokenizer-related, but one is a general version regression and the other is DeepSeek-specific incorrect tokenization." }, { "left": "issue:42371", "right": "issue:42831", "accept": false, - "reason": "TF32 API guidance and FineGrainedFP8 accuracy are unrelated." + "reason": "TF32 API/deprecation issue is unrelated to FineGrainedFP8 accuracy." }, { "left": "issue:44038", "right": "issue:44295", "accept": false, - "reason": "Different failures: Qwen3-VL-Moe compatibility vs position_ids buffer access." + "reason": "Qwen3-VL-Moe bug is not the same as reading position_ids after buffer registration." + }, + { + "left": "issue:36296", + "right": "issue:43116", + "accept": false, + "reason": "Tensor-parallel training bug and classification-example output bug are unrelated." }, { "left": "issue:42907", "right": "issue:43452", "accept": false, - "reason": "Both involve save/load workflows, but one is dequantized MoE saving and the other is gguf_file loading with Auto* APIs." + "reason": "Saving dequantized models is a different path from gguf_file loading failures." }, { "left": "issue:43992", "right": "issue:44704", "accept": false, - "reason": "Different call sites and bugs: missing embed_tokens.weight in UMT5 vs kwargs not forwarded in AutoProcessor." - }, - { - "left": "issue:36296", - "right": "issue:43116", - "accept": false, - "reason": "Tensor-parallel training bug is unrelated to the multi-label classification example issue." + "reason": "Missing embed_tokens.weight on load is unrelated to AutoProcessor kwargs forwarding." }, { "left": "issue:44038", "right": "issue:44222", "accept": false, - "reason": "Both touch MoE/FP8 areas, but one is a general Qwen3-VL-Moe issue and the other is an FP8 save_pretrained bug; not clearly the same bug." + "reason": "Both mention MoE, but one is a Qwen3-VL-Moe failure and the other is FP8 save_pretrained behavior." }, { - "left": "issue:43122", - "right": "issue:44977", + "left": "issue:43278", + "right": "issue:44945", "accept": false, - "reason": "Both are tokenizer regressions in v5, but they concern different models and do not look like one concrete fix." + "reason": "Embedding dtype drift in eval is not the same as pipeline-parallel incorrect output." }, { - "left": "issue:43278", - "right": "issue:44945", + "left": "issue:43122", + "right": "issue:44977", "accept": false, - "reason": "Embedding dtype mismatch during evaluate is unrelated to incorrect outputs under pipeline parallelism." + "reason": "Tokenizer regression and flash-attention generation failure are different bugs." }, { "left": "issue:45216", "right": "issue:45478", "accept": false, - "reason": "Same model family, but one is a save_pretrained regression and the other is a from_pretrained load error; different code paths." + "reason": "One concerns save_pretrained checkpoint correctness; the other is a from_pretrained import error." }, { "left": "issue:43824", "right": "issue:44351", "accept": false, - "reason": "Two separate import errors for different symbols; broad API breakage, but not the same underlying bug." + "reason": "Both are import failures, but they are different missing symbols from different modules." }, { "left": "issue:42907", "right": "issue:44222", "accept": false, - "reason": "Both are save_pretrained-related MoE/quantization reports, but the model and failure modes differ too much to treat as duplicates." + "reason": "Different model/formats and different save paths: dequantized Ministral/Devstral vs FP8 MoE save_pretrained." }, { "left": "issue:42757", "right": "issue:44351", "accept": false, - "reason": "Different missing imports from different packages/symbols." - }, - { - "left": "issue:42907", - "right": "issue:44912", - "accept": false, - "reason": "Save/dequantize issues versus MXFP4 load fallback are different problems." + "reason": "External huggingface_hub import failure is not the same bug as missing HybridCache in transformers." }, { "left": "issue:42222", "right": "issue:43441", "accept": false, - "reason": "Vitpose model breakage is unrelated to Ministral-3 FlashAttention failures." - }, - { - "left": "issue:43784", - "right": "issue:45446", - "accept": false, - "reason": "Importing sentence-transformers NameError is unrelated to the flex_attention version-check bug." + "reason": "VitPose model breakage and Ministral flash-attention failure are unrelated." }, { - "left": "issue:42673", - "right": "issue:45072", + "left": "issue:42907", + "right": "issue:44912", "accept": false, - "reason": "VRAM leak in Qwen3ForCausalLM threads is unrelated to dtype mismatches in inference." + "reason": "Saving dequantized models and MXFP4 quantization fallback are different problems." }, { - "left": "issue:43064", - "right": "issue:43673", + "left": "issue:22355", + "right": "issue:29127", "accept": false, - "reason": "Different training/runtime areas: FSDP2 optimizer state corruption vs generation cache missing in chunked prefill." + "reason": "ONNX module import error is unrelated to layoutlmv3 box-content error messaging." }, { "left": "issue:43550", "right": "issue:43784", "accept": false, - "reason": "torch.compile/SDPA model failure is unrelated to the sentence-transformers import NameError." + "reason": "torch.compile/SDPA training bug vs sentence-transformers import error; different code paths." }, { "left": "issue:38617", "right": "issue:45278", "accept": false, - "reason": "Both are import errors, but one is a specific missing internal symbol and the other is a broad upgrade regression; not a duplicate pair." - }, - { - "left": "issue:22355", - "right": "issue:29127", - "accept": false, - "reason": "No overlap in bug area: ONNX import failure vs LayoutLMv3 error-message clarity." + "reason": "Specific missing layer_type_validation import is not the same as a broad multi-import regression after upgrade." }, { "left": "issue:43937", "right": "issue:44485", "accept": false, - "reason": "GLM-5 generation config validation and GLM-5 RoPE implementation are related only by model name, not the same bug." + "reason": "GLM-5 generation-config validation and GLM-5 RoPE implementation are different defects." } ] }, @@ -11258,11 +11261,11 @@ "issue:44038", "issue:44075", "issue:44117", - "issue:44220", "issue:44222", "issue:44261", "issue:44291", "issue:44410", + "issue:44483", "issue:44484", "issue:44496", "issue:44779", @@ -11277,6 +11280,7 @@ "issue:45310", "issue:45405", "issue:45446", + "issue:45464", "issue:45478", "issue:45542" ], @@ -11290,8 +11294,8 @@ "issue:43976|issue:45446", "issue:39401|issue:44779", "issue:43931|issue:44945", - "issue:33453|issue:39401", "issue:45083|issue:45446", + "issue:33453|issue:39401", "issue:43937|issue:44977", "issue:43827|issue:45446", "issue:42673|issue:43906", @@ -11308,218 +11312,218 @@ "issue:43408|issue:43742", "issue:42673|issue:43577", "issue:42222|issue:45542", - "issue:38617|issue:44220", "issue:42371|issue:44261", - "issue:43937|issue:44496" + "issue:43937|issue:44496", + "issue:44483|issue:45464" ], - "prepared_review_unit_hash": "c4d9d1f92af415baa20bd048b2ecb62fce42c9b599670e7a70da04b83a226089", + "prepared_review_unit_hash": "d0d6e5a494ce664ac3171acc31de4cd3b32270434ecf20afeb6242803f004693", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36030, - "estimated_input_tokens": 9008, - "estimated_eval_tokens": 18272 + "serialized_chars": 36622, + "estimated_input_tokens": 9156, + "estimated_eval_tokens": 18568 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a false-positive grouping: the items share broad themes like tokenizers, Qwen, loading, or v5 regressions, but the concrete bugs and code paths are different. No soft edge looks safe to merge, and there are no PRs in the set.", + "summary": "This cluster is very heterogeneous: mostly unrelated Transformers bug reports spanning tokenizers, model loading, processors, API behavior, and version regressions. A few pairs are thematically close, but none look like the same concrete issue/change with enough confidence to merge as duplicates.", "confidence": 0.83, - "canonical_issue_reason": "No single canonical issue emerges; the set is a loose mix of unrelated bugs, so any canonical choice would be arbitrary.", + "canonical_issue_reason": "No single issue safely represents the whole cluster; the items are too diverse. If forced to pick one representative regression report, issue 45216 is the clearest and most concrete.", "canonical_pr_reason": null, - "best_issue_reason": "If forced, issue:45216 is the closest thing to a representative item because it is a concrete version-bound regression report, but it does not unify the cluster.", + "best_issue_reason": "Issue 45216 is the strongest representative candidate because it is a specific, version-bounded regression report with a clear model family and behavior.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43116", "right": "issue:43493", "accept": false, - "reason": "Different bugs and models: example script multi-label output vs SigLIP2 implementation discrepancy." + "reason": "Different bugs: one is multi-label classification example output, the other is a SigLIP2 implementation discrepancy." }, { "left": "issue:42371", "right": "issue:43874", "accept": false, - "reason": "Unrelated topics: TF32 settings vs a missing GLM46V image-processor method." + "reason": "Unrelated subsystems: TF32 API settings versus a missing image-processor method in GLM46V." }, { "left": "issue:44075", "right": "issue:44936", "accept": false, - "reason": "Different failure modes: optimizer args ignored vs trainer.evaluate() failing after train()." + "reason": "Different code paths: optimizer argument handling versus trainer evaluate/train lifecycle." }, { "left": "issue:44222", "right": "issue:45478", "accept": false, - "reason": "Both mention MoE, but one is FP8 save_pretrained and the other is a from_pretrained loading error; not the same concrete bug." + "reason": "Same broad Qwen3.5/MoE area, but one is save_pretrained FP8 behavior and the other is from_pretrained loading failure; not the same concrete bug." }, { "left": "issue:43116", "right": "issue:43825", "accept": false, - "reason": "Example-script classification bug vs pipeline deprecation/error-message issue; unrelated paths." + "reason": "One is an example script classification bug; the other is a pipeline() v5 error-message issue." }, { "left": "issue:44038", "right": "issue:45478", "accept": false, - "reason": "Same family name only; the reported symptoms and likely code paths are different." + "reason": "Both mention Qwen3.5/MoE, but they describe different failures and likely different code paths." }, { "left": "issue:43976", "right": "issue:45446", "accept": false, - "reason": "Python version compatibility issue vs a PyTorch version check bug; unrelated." + "reason": "Different compatibility problems: Python version support versus a PyTorch version check bug." }, { "left": "issue:39401", "right": "issue:44779", "accept": false, - "reason": "Both tokenizer issues, but different models and distinct symptoms (offset mapping vs incorrect results)." + "reason": "Both are tokenizer regressions, but on different models and symptoms (offset mapping vs incorrect generation output)." }, { "left": "issue:43931", "right": "issue:44945", "accept": false, - "reason": "Model weight-shape mismatch vs pipeline parallelism producing wrong outputs; different problems." + "reason": "Model weight-shape loading mismatch is unrelated to incorrect outputs under pipeline parallelism." }, { - "left": "issue:33453", - "right": "issue:39401", + "left": "issue:45083", + "right": "issue:45446", "accept": false, - "reason": "Generic tokenizer loading regression vs Qwen3 offset_mapping bug; not the same underlying defect." + "reason": "Different areas: qwen3_omni_moe feature extraction helper versus flex_attention import/version checking." }, { - "left": "issue:45083", - "right": "issue:45446", + "left": "issue:33453", + "right": "issue:39401", "accept": false, - "reason": "Feature-extraction helper behavior vs a flex_attention import/version check; unrelated." + "reason": "Both involve tokenizers, but one is loading regression and the other is offset_mapping correctness on Qwen3." }, { "left": "issue:43937", "right": "issue:44977", "accept": false, - "reason": "GenerationConfig validation error vs flash-attention generation malfunction; different code paths." + "reason": "GenerationConfig validation error is not the same as flash-attention generation corruption." }, { "left": "issue:43827", "right": "issue:45446", "accept": false, - "reason": "Docs still mentioning pipeline() vs a runtime import/version check bug; unrelated." + "reason": "Documentation references to removed pipeline() API are unrelated to flex_attention import gating." }, { "left": "issue:42673", "right": "issue:43906", "accept": false, - "reason": "VRAM leak in multithreaded inference vs an isolated reproduction of a different issue; no evidence they are the same bug." + "reason": "No clear shared bug; one is a VRAM leak in dataloader threads, the other is a reproduction of a different issue." }, { "left": "issue:44291", "right": "issue:45071", "accept": false, - "reason": "init_empty_weights/_is_hf_initialized TypeError vs PretrainedConfig type-checking breakage; different failures." + "reason": "Both are loading/type-related regressions, but they concern different APIs and error causes." }, { "left": "issue:43450", "right": "issue:43986", "accept": false, - "reason": "Batched video processor shape bug vs crash from missing torchvision during AutoProcessor loading; unrelated." + "reason": "A batched video-shape bug is not the same as a missing- torchvision crash on AutoProcessor load." }, { "left": "issue:45405", "right": "issue:45446", "accept": false, - "reason": "Released-version bump concern vs PyTorch import/version check bug; unrelated." + "reason": "Version pinning for PEFT is unrelated to PyTorch version checks in flex_attention." }, { "left": "issue:44829", "right": "issue:45216", "accept": false, - "reason": "Training degeneracy under flash_attention_3 vs a save_pretrained checkpoint regression; different behaviors." + "reason": "Different failure modes: degenerate training with flash_attention_3 versus incorrect checkpoint saving." }, { "left": "issue:44484", "right": "issue:45310", "accept": false, - "reason": "save_pretrained shard-size question vs Qwen3.5 MoE from_pretrained error; unrelated." + "reason": "Shard-size question is unrelated to Qwen3.5 MoE from_pretrained errors." }, { "left": "issue:39401", "right": "issue:43421", "accept": false, - "reason": "Qwen3 offset mapping bug vs runtime post-processor update behavior; related area only, not same bug." + "reason": "Tokenizer offset_mapping regression is not the same as runtime special-token post-processor updates." }, { "left": "issue:38617", "right": "issue:41628", "accept": false, - "reason": "Different import errors with different missing symbols and likely different fixes." + "reason": "Both are import errors, but for different symbols and likely different export paths." }, { "left": "issue:44117", "right": "issue:45081", "accept": false, - "reason": "Tokenizer registry None handling vs Mistral regex patch crash; different root causes." + "reason": "Tokenizer mapping None-handling is a different bug from the Mistral regex patch crash." }, { "left": "issue:42673", "right": "issue:43582", "accept": false, - "reason": "VRAM leak in model use vs Apple Silicon TypeError in warmup helper; unrelated." + "reason": "VRAM leak in dataloader threads and Apple Silicon TypeError are unrelated." }, { "left": "issue:44410", "right": "issue:45310", "accept": false, - "reason": "Missing projections in qwen3next vs Qwen3.5 MoE loading error; not the same defect." + "reason": "Missing projections in qwen3next is a model-architecture issue, not the same as Qwen3.5 MoE loading failure." }, { "left": "issue:43408", "right": "issue:43742", "accept": false, - "reason": "SAM3 tracker/video model warning vs MobileLLM loading key error; unrelated." + "reason": "SAM3 tracker warning and MobileLLM key error are distinct loading/config issues." }, { "left": "issue:42673", "right": "issue:43577", "accept": false, - "reason": "VRAM leak vs incorrect dtype propagation in Blip2 loading; different bugs." + "reason": "Different bugs: VRAM leak versus dtype not propagating on BLIP2 load." }, { "left": "issue:42222", "right": "issue:45542", "accept": false, - "reason": "VitPose model breakage vs undefined tf backend error; no connection." - }, - { - "left": "issue:38617", - "right": "issue:44220", - "accept": false, - "reason": "ImportError for a config symbol vs an audio feature-extraction helper issue; unrelated." + "reason": "VitPose model breakage and TensorFlow/tensorboard backend detection are unrelated." }, { "left": "issue:42371", "right": "issue:44261", "accept": false, - "reason": "TF32 behavior settings vs missing rms_norm_eps in MLA layernorm; different subsystems." + "reason": "TF32 API control and MLA rms_norm_eps precision mismatch do not share the same bug." }, { "left": "issue:43937", "right": "issue:44496", "accept": false, - "reason": "Invalid generation config vs unrecognized model/config loading failure; not the same bug." + "reason": "Invalid GenerationConfig and unrecognized-model config.json errors are different loading/validation failures." + }, + { + "left": "issue:44483", + "right": "issue:45464", + "accept": false, + "reason": "Both involve chat/completions, but one is a v5.3 request acceptance issue and the other is a streaming failure on a specific model." } ] }, @@ -11559,6 +11563,7 @@ "issue:43441", "issue:43452", "issue:43531", + "issue:43582", "issue:43673", "issue:43874", "issue:43931", @@ -11566,12 +11571,12 @@ "issue:43986", "issue:44038", "issue:44117", + "issue:44220", "issue:44261", "issue:44263", "issue:44315", "issue:44368", "issue:44410", - "issue:44483", "issue:44484", "issue:44485", "issue:44493", @@ -11588,14 +11593,13 @@ "issue:45372", "issue:45375", "issue:45459", - "issue:45464", "issue:45478", "issue:45542" ], "soft_pairs": [ - "issue:44483|issue:45464", - "issue:42175|issue:42222", + "issue:38617|issue:44220", "issue:42617|issue:45341", + "issue:42175|issue:42222", "issue:44933|issue:45278", "issue:44484|issue:45478", "issue:43066|issue:43122", @@ -11611,262 +11615,276 @@ "issue:41628|issue:45071", "issue:45071|issue:45542", "issue:45003|issue:45341", - "issue:43097|issue:43452", + "issue:43931|issue:44410", "issue:43986|issue:45341", + "issue:43097|issue:43452", "issue:44117|issue:44987", - "issue:43931|issue:44410", "issue:43335|issue:45071", "issue:43408|issue:44368", "issue:43931|issue:44038", "issue:43874|issue:44263", - "issue:44829|issue:45245", "issue:43986|issue:45372", + "issue:44829|issue:45245", "issue:43976|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:44704|issue:45103", + "issue:43986|issue:44117", + "issue:43582|issue:45341" ], - "prepared_review_unit_hash": "7407466d1c0568032398e29de58392d79961549310deaeff42c48d9fb4813de7", + "prepared_review_unit_hash": "543b655ccc2d7b275cd0dc3057279674cd43acf8e3ffab84a4f05cd1492557cc", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 35, - "serialized_chars": 38050, - "estimated_input_tokens": 9513, - "estimated_eval_tokens": 19282 + "soft_pair_count": 37, + "serialized_chars": 38627, + "estimated_input_tokens": 9657, + "estimated_eval_tokens": 19570 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly a set of similarly worded but distinct regressions/import failures. I would not merge any of the soft-edge pairs; the best triage anchor is the broad import-regression report, not a true duplicate set.", + "summary": "This cluster is mostly a loose mix of Transformers regression reports: import errors, tokenizer/config changes, and several model-specific bugs. The soft links look similarity-based rather than true duplicates, so I would not merge any of the candidate pairs.", "confidence": 0.89, - "canonical_issue_reason": null, + "canonical_issue_reason": "issue:45278 is the broadest umbrella for the cluster\u2019s recurring theme of import regressions after upgrading Transformers; it best represents the shared high-level failure mode even though the cluster itself is heterogeneous.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the broadest, most umbrella-style report of upgrade-related import failures, so it is the best triage anchor, but it does not duplicate the other issues.", + "best_issue_reason": "issue:45278 is the most general and cluster-representative issue, especially for the import/error-after-upgrade subset; it is the best single issue to anchor this group.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44483", - "right": "issue:45464", + "left": "issue:38617", + "right": "issue:44220", "accept": false, - "reason": "Both mention chat/completions failures, but one is a generic v5.3 API break and the other is Qwen3.5-0.8B streaming behavior; not the same bug." + "reason": "Different failures: a missing configuration utility import vs an audio feature-extraction issue." }, { - "left": "issue:42175", - "right": "issue:42222", + "left": "issue:42617", + "right": "issue:45341", "accept": false, - "reason": "TensorFlow backend packaging vs broken VitPose model files; unrelated problems." + "reason": "Unrelated: one is a runtime script failure, the other is a testing utility bug." }, { - "left": "issue:42617", - "right": "issue:45341", + "left": "issue:42175", + "right": "issue:42222", "accept": false, - "reason": "A script execution failure and a testing utility bug are unrelated." + "reason": "Different subsystems and symptoms: backend install packaging vs broken ViTPose model files." }, { "left": "issue:44933", "right": "issue:45278", "accept": false, - "reason": "A specific missing import from image_utils is not the same as a broad post-upgrade import-error report." + "reason": "Both mention imports, but one is a specific missing image_utils export and the other is a broad upgrade regression; not the same concrete bug." }, { "left": "issue:44484", "right": "issue:45478", "accept": false, - "reason": "Save/sharding behavior and Qwen3.5-Moe loading errors are different code paths." + "reason": "Unrelated: shard-size behavior in save_pretrained vs a Qwen3.5 MoE from_pretrained error." }, { "left": "issue:43066", "right": "issue:43122", "accept": false, - "reason": "Both are tokenizer regressions in v5, but one is decoder-type metadata and the other is changed tokenization output; too different to merge." + "reason": "Both concern tokenizer behavior in v5, but one is decoder type mismatch and the other is different tokenization output; too broad to be the same bug." }, { "left": "issue:43673", "right": "issue:45341", "accept": false, - "reason": "Generation cache behavior and a testing utility bug are unrelated." + "reason": "Different code paths: generation cache handling vs a testing utility bug." }, { "left": "issue:43097", "right": "issue:44493", "accept": false, - "reason": "Removed config field vs unexpected position-id keys are different symptoms and code paths." + "reason": "Different issues: a removed config field vs unexpected position-id keys during model loading." }, { "left": "issue:42222", "right": "issue:45459", "accept": false, - "reason": "VitPose model breakage and protobuf-related tokenizer error masking are unrelated." + "reason": "One is a broken model family, the other is tokenizer error handling with protobuf; not the same underlying problem." }, { "left": "issue:44410", "right": "issue:45478", "accept": false, - "reason": "Different Qwen model families and different failure modes." + "reason": "Both are Qwen-family model bugs, but one is missing projections and the other is a from_pretrained failure; not one fixable change." }, { "left": "issue:43874", "right": "issue:44261", "accept": false, - "reason": "Image-token counting bug in GLM46V vs RMS-epsilon precision issue in MLA are distinct." + "reason": "Different models and failures: missing image patch helper vs missing rms_norm_eps in MLA config." }, { "left": "issue:43441", "right": "issue:44829", "accept": false, - "reason": "Both involve flash attention, but one is a model-loading failure and the other is degenerate training; not the same concrete bug." + "reason": "Both involve attention implementations, but one is model loading with FlashAttention and the other is training degeneration under flash_attention_3." }, { "left": "issue:43986", "right": "issue:44484", "accept": false, - "reason": "AutoProcessor/torchvision crash is unrelated to max_shard_size behavior." + "reason": "Unrelated: AutoProcessor crash without torchvision vs max shard size default behavior." }, { "left": "issue:43012", "right": "issue:43408", "accept": false, - "reason": "A numerical warning during bfloat16 compilation is unrelated to a model-type mismatch warning." + "reason": "Both are warnings, but they concern entirely different models and causes." }, { "left": "issue:44263", "right": "issue:44485", "accept": false, - "reason": "Torch.split indexing bug and RoPE implementation concern are separate GLM issues." + "reason": "GLM-related but different concrete bugs: torch.split indexing vs RoPE implementation." }, { "left": "issue:41628", "right": "issue:45071", "accept": false, - "reason": "Missing AutoImageProcessor import and PretrainedConfig type-checking regression are unrelated." + "reason": "Different regressions: missing AutoImageProcessor export vs PretrainedConfig type-checking breakage." }, { "left": "issue:45071", "right": "issue:45542", "accept": false, - "reason": "Type-checking regression and missing TensorFlow backend wiring are unrelated." + "reason": "Different failure modes: type checking regression vs backend selection error when TensorFlow is absent." }, { "left": "issue:45003", "right": "issue:45341", "accept": false, - "reason": "Unsafe sys.modules access and a testing_utils bug do not describe the same fault." + "reason": "No shared code path: sys.modules access bug vs testing utility issue." }, { - "left": "issue:43097", - "right": "issue:43452", + "left": "issue:43931", + "right": "issue:44410", "accept": false, - "reason": "Removed tie_embeddings behavior and gguf_file loading failures are different issues." + "reason": "Both are Qwen model load issues, but one is a weight-shape mismatch and the other is missing projections; not the same bug." }, { "left": "issue:43986", "right": "issue:45341", "accept": false, - "reason": "Video-model processor crash and testing_utils bug are unrelated." + "reason": "Unrelated processor-loading crash and testing utility bug." }, { - "left": "issue:44117", - "right": "issue:44987", + "left": "issue:43097", + "right": "issue:43452", "accept": false, - "reason": "Tokenizer mapping returning None may contribute to loading failures, but the model-specific load error is not clearly the same bug." + "reason": "Different topics: removed tie_embeddings config vs gguf_file tokenizer/model loading failure." }, { - "left": "issue:43931", - "right": "issue:44410", + "left": "issue:44117", + "right": "issue:44987", "accept": false, - "reason": "Weight-shape mismatch for Qwen3-VL vs missing projections in Qwen3Next are different model/code-path problems." + "reason": "Both involve tokenizer/model loading, but the concrete failures differ and there is no clear same code-path bug." }, { "left": "issue:43335", "right": "issue:45071", "accept": false, - "reason": "SwitchTransformers sparse-layer construction and PretrainedConfig type checking are unrelated." + "reason": "Different regressions: SwitchTransformers config construction vs config type-checking." }, { "left": "issue:43408", "right": "issue:44368", "accept": false, - "reason": "Model-type warning and tie_word_embeddings warning are different warnings with different causes." + "reason": "Both are warnings, but they concern different models and unrelated config messages." }, { "left": "issue:43931", "right": "issue:44038", "accept": false, - "reason": "Both are Qwen3-VL related, but one is a shape mismatch and the other is a transformers-version/MoE compatibility issue." + "reason": "Qwen3-VL load mismatch vs Qwen3-VL-Moe compatibility issue; similar family, not same defect." }, { "left": "issue:43874", "right": "issue:44263", "accept": false, - "reason": "Different GLM bugs: missing image-patch helper vs torch.split return handling." + "reason": "Different GLM bugs affecting different components; not mergeable into one fix." }, { - "left": "issue:44829", - "right": "issue:45245", + "left": "issue:43986", + "right": "issue:45372", "accept": false, - "reason": "FlashAttention training degradation and category-cardinality limit are unrelated." + "reason": "Different import/loading failures: missing torchvision vs missing ReasoningEffort import." }, { - "left": "issue:43986", - "right": "issue:45372", + "left": "issue:44829", + "right": "issue:45245", "accept": false, - "reason": "Processor loading without torchvision and mistral_common import failure are unrelated dependency issues." + "reason": "FlashAttention training degradation is unrelated to categorical-feature cardinality limits." }, { "left": "issue:43976", "right": "issue:45071", "accept": false, - "reason": "Python-version incompatibility and PretrainedConfig type checking are different regressions." + "reason": "Python/version support issue vs a type-checking regression; not the same underlying bug." }, { "left": "issue:43531", "right": "issue:45341", "accept": false, - "reason": "Qwen3-MoE sliding-window behavior and a testing utility bug are unrelated." + "reason": "Different subsystems: MoE sliding_window behavior vs testing utilities." }, { "left": "issue:43064", "right": "issue:44315", "accept": false, - "reason": "Distributed optimizer-state corruption and Liger Kernel application timing are different training issues." + "reason": "Training-state corruption under FSDP/PEFT is unrelated to Liger Kernel application with model_init." }, { "left": "issue:38617", "right": "issue:42757", "accept": false, - "reason": "Both are import errors, but they are missing different symbols from different packages." + "reason": "Both are ImportError reports, but for different missing symbols in different packages." }, { "left": "issue:44261", "right": "issue:44485", "accept": false, - "reason": "RMS-epsilon precision bug and GLM-5 RoPE implementation concern are separate issues." + "reason": "Both mention GLM/MLA-related settings, but one is config eps precision and the other is RoPE implementation." }, { "left": "issue:44410", "right": "issue:45375", "accept": false, - "reason": "Missing projections in Qwen3Next and a missing vision-config field are unrelated." + "reason": "Same broad Qwen ecosystem, but one is missing layer projections and the other is a config field dropped by strict parsing." }, { "left": "issue:44704", "right": "issue:45103", "accept": false, - "reason": "Passing kwargs to cached_file and auto_docstring annotation handling are different helper bugs." + "reason": "Different code paths: cached_file kwargs propagation vs docstring processing with future annotations." + }, + { + "left": "issue:43986", + "right": "issue:44117", + "accept": false, + "reason": "Processor loading without torchvision is unrelated to tokenizer mapping returning None." + }, + { + "left": "issue:43582", + "right": "issue:45341", + "accept": false, + "reason": "Apple Silicon allocator TypeError is unrelated to the testing utility bug." } ] }, @@ -11934,33 +11952,30 @@ "issue:44655", "issue:44683", "issue:45290", - "issue:45341", "issue:45381", "issue:45479", "issue:45538" ], "soft_pairs": [ "issue:44062|issue:44117", - "issue:43582|issue:45341", - "issue:38617|issue:44351", - "issue:43986|issue:44117", "issue:44038|issue:44410", + "issue:38617|issue:44351", "issue:43502|issue:43992", "issue:43097|issue:43873", "issue:43992|issue:44484", - "issue:43986|issue:45290", "issue:43937|issue:45538", + "issue:43824|issue:43986", "issue:43066|issue:43906", - "issue:38617|issue:43097", "issue:43825|issue:44655", "issue:43352|issue:44380", + "issue:38617|issue:43097", "issue:43097|issue:43901", "issue:43012|issue:43550", "issue:43012|issue:44368", "issue:42886|issue:43381", "issue:43208|issue:43550", - "issue:42907|issue:44655", "issue:44380|issue:45479", + "issue:42907|issue:44655", "issue:43525|issue:44117", "issue:43352|issue:43976", "issue:43064|issue:44380", @@ -11974,140 +11989,128 @@ "issue:42757|issue:43824", "issue:44038|issue:44230", "issue:44483|issue:45381", - "issue:38617|issue:43502", "issue:43299|issue:43867", - "issue:43901|issue:44655" + "issue:43901|issue:44655", + "issue:38617|issue:43502" ], - "prepared_review_unit_hash": "f7a729ad872cee7bf1c3403eab50404b1bf78cd5b8254cc2770efcabd9faa567", + "prepared_review_unit_hash": "297c0a5e8300c1f479a15760830bc74bff526109d0f9c9a2c51bb9b7d9d55b5d", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 37, - "serialized_chars": 38528, - "estimated_input_tokens": 9632, - "estimated_eval_tokens": 19520 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 35, + "serialized_chars": 37404, + "estimated_input_tokens": 9351, + "estimated_eval_tokens": 18958 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set does not form a single duplicate cluster; the items span unrelated import errors, model-specific regressions, tokenizer/pipeline issues, and training/runtime bugs. The soft-similarity links are mostly superficial topic overlaps, not the same underlying defect.", - "confidence": 0.97, - "canonical_issue_reason": "No clear canonical issue: the issues are heterogeneous and do not share one concrete bug or fix path.", + "summary": "Mostly heterogeneous issue set with many v5 regressions and model-specific breakages; almost all soft pairs are non-duplicates. One CLIP output regression pair looks like the same underlying bug, but there are no pull requests in this cluster.", + "confidence": 0.67, + "canonical_issue_reason": "issue:43097 is the broadest, most central v5-breaking-change report in the set (API removal with downstream impact) and is a reasonable anchor for the otherwise heterogeneous cluster.", "canonical_pr_reason": null, - "best_issue_reason": "No single issue is a good global representative because the cluster is not cohesive enough to anchor on one bug.", + "best_issue_reason": "issue:43097 is the strongest representative issue: concrete regression, wider blast radius, and more likely to subsume related downstream breakages than the more model-specific reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44062", "right": "issue:44117", "accept": false, - "reason": "Different failures: AddedToken keyword collision vs tokenizer mapping returning None during from_pretrained." + "reason": "Both are tokenizer-related, but one is an AddedToken TypeError and the other is a mapping-name/import assumption bug; different failure modes and fixes." }, { - "left": "issue:43582", - "right": "issue:45341", + "left": "issue:44038", + "right": "issue:44410", "accept": false, - "reason": "Unrelated: Apple Silicon caching allocator TypeError vs a testing_utils bug." + "reason": "Same broad Qwen family, but one is a model-loading regression and the other is missing layer projections; not the same bug." }, { "left": "issue:38617", "right": "issue:44351", "accept": false, - "reason": "Both are import errors, but for different missing symbols and different call paths." - }, - { - "left": "issue:43986", - "right": "issue:44117", - "accept": false, - "reason": "Different areas: AutoProcessor/video-model crash without torchvision vs tokenizer mapping assumptions." - }, - { - "left": "issue:44038", - "right": "issue:44410", - "accept": false, - "reason": "Both involve Qwen-family models, but they describe different model loading/projection issues." + "reason": "Both are import errors, but for unrelated symbols from different modules; no evidence of one shared underlying breakage." }, { "left": "issue:43502", "right": "issue:43992", "accept": false, - "reason": "Different symptoms: unintended API calls with local_files_only vs missing embed_tokens.weight in UMT5Encoder loading." + "reason": "One is offline/network behavior, the other is missing weights during model load; these are distinct issues." }, { "left": "issue:43097", "right": "issue:43873", "accept": false, - "reason": "Tie-embeddings API removal and quantized offloading behavior are unrelated." + "reason": "API removal versus quantization/offloading behavior; unrelated code paths and fixes." }, { "left": "issue:43992", "right": "issue:44484", "accept": false, - "reason": "Model weight loading bug vs shard-size question for save_pretrained; no shared code-path defect." - }, - { - "left": "issue:43986", - "right": "issue:45290", - "accept": false, - "reason": "Both touch chat/processor flows, but one is missing torchvision and the other is a tool-call templating crash." + "reason": "Missing checkpoint weights and shard-size default behavior are different save/load concerns." }, { "left": "issue:43937", "right": "issue:45538", "accept": false, - "reason": "GenerationConfig validation for GLM-5 is unrelated to CLIPTokenizer max length." + "reason": "GenerationConfig validation for GLM-5 is unrelated to CLIP tokenizer max-length handling." }, { - "left": "issue:43066", - "right": "issue:43906", + "left": "issue:43824", + "right": "issue:43986", "accept": false, - "reason": "Wrong tokenizer decoder type is not the same as an isolated reproduction of a different issue." + "reason": "Both mention model loading, but one is a missing top-level class export and the other is an AutoProcessor/torchvision crash." }, { - "left": "issue:38617", - "right": "issue:43097", + "left": "issue:43066", + "right": "issue:43906", "accept": false, - "reason": "Different deprecations/import surface: missing configuration_utils symbol vs removed tie_embeddings_and_encoder_decoder." + "reason": "The titles suggest tokenizer behavior, but there is not enough evidence they are the same concrete bug." }, { "left": "issue:43825", "right": "issue:44655", "accept": false, - "reason": "Pipeline docs/message issue vs inability to save Pipeline objects; different behavior and fix." + "reason": "Error messaging in pipeline versus pipeline serialization are separate concerns." }, { "left": "issue:43352", "right": "issue:44380", "accept": false, - "reason": "FlashAttention support error for a Nemotron model is unrelated to GPT2 attention scaling under SDPA/FlashAttention." + "reason": "Different attention-backend problems on different models; not mergeable as one fix." + }, + { + "left": "issue:38617", + "right": "issue:43097", + "accept": false, + "reason": "Importing a removed utility and removing a config helper are different API breakages." }, { "left": "issue:43097", "right": "issue:43901", "accept": false, - "reason": "Removed config API vs stale TextClassificationPipeline docs; not the same bug." + "reason": "A v5 API removal versus a docs/behavior mismatch in text classification pipeline." }, { "left": "issue:43012", "right": "issue:43550", "accept": false, - "reason": "Different warnings/errors: bf16 compile warning vs Bamba torch.compile SDPA failure." + "reason": "A precision warning during compile and a compile failure under SDPA are distinct symptoms and likely distinct fixes." }, { "left": "issue:43012", "right": "issue:44368", "accept": false, - "reason": "Both mention warnings, but one is a PyTorch precision warning and the other is a tie_word_embeddings config warning." + "reason": "Float32 warning on bf16 compile is unrelated to a tie_word_embeddings warning." }, { "left": "issue:42886", @@ -12119,122 +12122,125 @@ "left": "issue:43208", "right": "issue:43550", "accept": false, - "reason": "xLSTM training bugs and Bamba SDPA compile failure affect different models and code paths." + "reason": "xLSTM training bugs and Bamba torch.compile+SDPA failure are different model-specific issues." }, { - "left": "issue:42907", - "right": "issue:44655", + "left": "issue:44380", + "right": "issue:45479", "accept": false, - "reason": "Saving dequantized models and saving Pipeline objects are different persistence problems." + "reason": "Attention-scaling backend behavior is unrelated to sequence-classification loss degeneracy." }, { - "left": "issue:44380", - "right": "issue:45479", + "left": "issue:42907", + "right": "issue:44655", "accept": false, - "reason": "Attention scaling backend bug is unrelated to sequence-classification zero-loss behavior." + "reason": "Saving dequantized model weights is not the same as saving Pipeline objects." }, { "left": "issue:43525", "right": "issue:44117", "accept": false, - "reason": "Missing Llama4Config attribute and tokenizer mapping None are different configuration/loading defects." + "reason": "Missing config attribute for Llama4 and tokenizer mapping returning None are unrelated loader bugs." }, { "left": "issue:43352", "right": "issue:43976", "accept": false, - "reason": "FlashAttention support for one model is unrelated to Python version compatibility." + "reason": "Flash Attention support for a model and Python-version compatibility are different problems." }, { "left": "issue:43064", "right": "issue:44380", "accept": false, - "reason": "Distributed optimizer-state corruption and GPT2 attention scaling are unrelated." + "reason": "Distributed optimizer-state corruption is unrelated to attention-scaling behavior." }, { "left": "issue:43208", "right": "issue:43526", "accept": false, - "reason": "xLSTM training bugs and BeitImageProcessorFast label reduction are unrelated." + "reason": "Training-model-size bugs in xLSTM versus a BEiT image-processor label bug are not the same issue." }, { "left": "issue:44483", "right": "issue:44514", "accept": false, - "reason": "OpenAI-style chat/completions endpoint rejection is not the same as apply_chat_template crashing on batched tool-call input." + "reason": "OpenAI-compatible chat API rejection and Qwen2_5_VL chat-template batching are separate request-handling bugs." }, { "left": "issue:43618", "right": "issue:44380", "accept": false, - "reason": "CLIP attentions regression and GPT2 attention scaling are different model-specific issues." + "reason": "CLIP output-field regression and GPT2 attention scaling under SDPA are unrelated." }, { "left": "issue:43618", "right": "issue:43761", - "accept": false, - "reason": "Both are CLIP regressions, but one drops attentions and the other drops hidden_states; not enough evidence of one shared bug." + "accept": true, + "reason": "Both describe the same CLIP v5 regression family: expected vision-model outputs are not populated correctly (attentions/hidden_states)." }, { "left": "issue:36683", "right": "issue:43582", "accept": false, - "reason": "Gemma3Config missing vocab_size and Apple Silicon allocator TypeError are unrelated." + "reason": "Gemma3 config missing vocab_size and Apple Silicon allocator TypeError are unrelated." }, { "left": "issue:43012", "right": "issue:44683", "accept": false, - "reason": "bf16 warning vs compiled flex_attention failure are separate backend/precision issues." + "reason": "A compile warning about bf16 precision is not the same as a flex_attention compile failure on newer torch." }, { "left": "issue:44483", "right": "issue:45290", "accept": false, - "reason": "Different request paths: /v1/chat/completions API handling vs tokenizer/template crash on tool-call messages." + "reason": "Chat-completions API acceptance and tool-call chat-template crashes are different layers of functionality." }, { "left": "issue:42757", "right": "issue:43824", "accept": false, - "reason": "Import errors for different missing names from different modules and packages." + "reason": "Different missing imports from different packages/symbols; no shared root cause shown." }, { "left": "issue:44038", "right": "issue:44230", "accept": false, - "reason": "Qwen3-VL-Moe loading issue and fp8 support request are not the same defect." + "reason": "One is a loading bug for Qwen3-VL-Moe, the other is an fp8 inference support request; not duplicates." }, { "left": "issue:44483", "right": "issue:45381", "accept": false, - "reason": "Chat/completions request handling is unrelated to Qwen2.5-VL video vision_position_ids." - }, - { - "left": "issue:38617", - "right": "issue:43502", - "accept": false, - "reason": "Missing configuration_utils symbol and unwanted API calls despite local_files_only concern different mechanisms." + "reason": "API request rejection and Qwen2.5-VL video position-id issues are different bugs." }, { "left": "issue:43299", "right": "issue:43867", "accept": false, - "reason": "Qwen3VL-Moe loading regression and sorted-state_dict load error are different model-loading problems." + "reason": "Qwen3-VL-Moe loading regression and sorted-state_dict load error are different failure modes." }, { "left": "issue:43901", "right": "issue:44655", "accept": false, - "reason": "Docs mismatch for return_all_scores is unrelated to Pipeline save_pretrained behavior." + "reason": "TextClassificationPipeline docs mismatch and pipeline save_pretrained failure are unrelated." + }, + { + "left": "issue:38617", + "right": "issue:43502", + "accept": false, + "reason": "Removed import from configuration_utils versus local_files_only network requests are different breakages." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays appropriately conservative overall. The lone accepted soft pair (issue:43618 / issue:43761) is plausibly the same CLIP output-regression family, while the rest of the rejects are cautious and well-justified. No PR-pair concerns since this cluster contains no PRs." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -12275,7 +12281,6 @@ "issue:43976", "issue:43992", "issue:44077", - "issue:44112", "issue:44188", "issue:44220", "issue:44242", @@ -12290,6 +12295,7 @@ "issue:44857", "issue:44938", "issue:44964", + "issue:44977", "issue:45071", "issue:45216", "issue:45278", @@ -12306,16 +12312,16 @@ "issue:36683|issue:43475", "issue:43867|issue:45216", "issue:44220|issue:44655", - "issue:43066|issue:45381", "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:36010|issue:42757", "issue:43618|issue:45216", "issue:36683|issue:44242", "issue:43992|issue:45310", @@ -12337,258 +12343,258 @@ "issue:43618|issue:44857", "issue:42890|issue:44964", "issue:28282|issue:33453", - "issue:42890|issue:44112" + "issue:43618|issue:44977" ], - "prepared_review_unit_hash": "aa39ae4710c11f5f1aafc12c005651cef821deefac86a2c357340618d99c93ca", + "prepared_review_unit_hash": "7366a3976dabe88e16375bf56c3932e21e5389af29e61045facfed56fbbb95da", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 37, - "serialized_chars": 38190, - "estimated_input_tokens": 9548, - "estimated_eval_tokens": 19352 + "serialized_chars": 38181, + "estimated_input_tokens": 9546, + "estimated_eval_tokens": 19348 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These are mostly unrelated bug reports across model loading, saving, attention, config, and CI. The soft pairs share surface keywords but not the same concrete defect, so none should be merged as duplicates. No PRs were provided.", - "confidence": 0.79, - "canonical_issue_reason": "No single true duplicate target stands out; if one issue must serve as an umbrella, issue:45278 is the broadest open report about widespread import regressions after an upgrade.", + "summary": "These are mostly distinct issue reports across different models and subsystems. The soft-similarity pairs are generally topic-alikes rather than true duplicates; I would reject them all. The clearest representative issue in the set is the Qwen3.5 MoE from_pretrained regression (45310), which also appears to have a later near-identical follow-up report.", + "confidence": 0.86, + "canonical_issue_reason": "issue:45310 is the clearest self-contained regression report in the set and appears to represent the Qwen3.5 MoE from_pretrained failure family better than the later follow-up report.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the best global issue candidate only as a broad tracker for import failures, but it is not a real canonical duplicate for this set.", + "best_issue_reason": "issue:45310 is the best representative issue because it has a concrete, reproducible model-loading failure and is the most likely anchor for the related Qwen3.5 MoE reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43756", "right": "issue:44485", "accept": false, - "reason": "Different bugs: Smollm3 RoPE-layer mismatch vs a save_pretrained shard-size question." + "reason": "Both mention RoPE, but one is a model-specific layer-count bug and the other is a general GLM-5 RoPE implementation discussion; not the same defect." }, { "left": "issue:42994", "right": "issue:44655", "accept": false, - "reason": "Both involve saving, but one is quantized model saving and the other is Pipeline.save_pretrained; different code paths." + "reason": "Both are save-related, but one is quantized-model saving failure and the other is Pipeline.save_pretrained() support; different code paths and failure modes." }, { "left": "issue:36683", "right": "issue:43475", "accept": false, - "reason": "Missing config attribute in Gemma3 vs missing output field in SAM 3 Video; different models and failures." + "reason": "Both are missing-attribute errors, but on different model classes and different missing fields; not a duplicate bug." }, { "left": "issue:43867", "right": "issue:45216", "accept": false, - "reason": "Load-time state_dict ordering error vs a save_pretrained regression; opposite operations and different root causes." + "reason": "Both concern loading/saving correctness, but one is a state_dict ordering load error and the other is a Qwen3.5 save_pretrained regression; different underlying problems." }, { "left": "issue:44220", "right": "issue:44655", "accept": false, - "reason": "Feature extraction bug vs pipeline saving bug; no shared concrete defect." + "reason": "Audio feature extraction helper bug vs Pipeline serialization bug; unrelated functionality." }, { - "left": "issue:43066", - "right": "issue:45381", + "left": "issue:43352", + "right": "issue:43937", "accept": false, - "reason": "Tokenizer decoder type regression vs Qwen2.5-VL video position-id issue; unrelated." + "reason": "FlashAttention support error for a model vs invalid GenerationConfig for GLM-5; different failures." }, { - "left": "issue:43352", - "right": "issue:43937", + "left": "issue:43066", + "right": "issue:45381", "accept": false, - "reason": "FlashAttention support check in Nemotron vs invalid GenerationConfig in GLM-5; different validation failures." + "reason": "Tokenizer decoder type mismatch in v5 vs qwen2.5-vl video position-id bug; different components and symptoms." }, { "left": "issue:44077", "right": "issue:44380", "accept": false, - "reason": "patchtsmixer post_init policy vs GPT2 attention scaling under SDPA/FlashAttention; different subsystems and bugs." + "reason": "One is a model-class post_init policy issue, the other is GPT2 attention scaling being ignored in SDPA/FlashAttention; not the same bug." }, { "left": "issue:36683", "right": "issue:44077", "accept": false, - "reason": "Different model/config issues: Gemma3 missing vocab_size vs patchtsmixer post_init handling." + "reason": "Gemma3Config missing vocab_size and patchtsmixer post_init policy are unrelated model/config issues." }, { "left": "issue:44263", "right": "issue:45440", "accept": false, - "reason": "GlmMoeDsaIndexer torch.split issue vs DeepseekV3MoE implementation divergence; not the same defect." + "reason": "torch.split return values in an indexer versus DeepseekV3 MoE implementation divergence; different subsystems." }, { "left": "issue:43064", "right": "issue:44492", "accept": false, - "reason": "FSDP2/PEFT optimizer-state bug vs a typo in cache strategies; no overlap in underlying change." + "reason": "FSDP/PEFT optimizer-state corruption versus a cache-strategy typo; clearly unrelated." }, { "left": "issue:43976", "right": "issue:44938", "accept": false, - "reason": "Both are version-compatibility reports, but one is Python 3.9+ and the other Python 3.14; different compatibility breaks." + "reason": "Python version compatibility in 5.1.0 versus Python 3.14 load failure; related only at a very broad packaging level, not the same bug." }, { "left": "issue:36683", "right": "issue:43525", "accept": false, - "reason": "Both mention missing config fields, but Gemma3 vocab_size and Llama4 pad_token_id are different models and different attributes." + "reason": "Different missing-attribute regressions on different configs (vocab_size vs pad_token_id); separate issues." }, { - "left": "issue:36010", - "right": "issue:42757", + "left": "issue:44188", + "right": "issue:44380", "accept": false, - "reason": "Both are import errors, but for different symbols from different modules; not the same underlying regression." + "reason": "torch.compile branching divergence in attention kernels is not the same as GPT2 scaling being ignored in SDPA/FlashAttention." }, { - "left": "issue:44188", - "right": "issue:44380", + "left": "issue:36010", + "right": "issue:42757", "accept": false, - "reason": "Attention-related, but one is torch.compile branching divergence and the other is GPT2 scaling ignored under backend dispatch." + "reason": "Different import errors from different dependencies: GenerationMixin import vs huggingface_hub offline helper import." }, { "left": "issue:43618", "right": "issue:45216", "accept": false, - "reason": "CLIPOutput attentions missing and Qwen3.5 save_pretrained regression are unrelated bugs." + "reason": "CLIPOutput attentions assignment regression versus Qwen3.5 save_pretrained regression; unrelated model areas." }, { "left": "issue:36683", "right": "issue:44242", "accept": false, - "reason": "Gemma3 config attribute error vs load-balancing loss omission; different features and code paths." + "reason": "Gemma3Config attribute bug versus MoE load-balancing loss omission; different model families and mechanisms." }, { "left": "issue:43992", "right": "issue:45310", "accept": false, - "reason": "UMT5Encoder missing embed_tokens weight vs Qwen3.5 MoE from_pretrained error; different loading failures." + "reason": "UMT5Encoder missing embed_tokens.weight and Qwen3.5 MoE from_pretrained error are different loading bugs." }, { "left": "issue:43010", "right": "issue:45278", "accept": false, - "reason": "Cache update no_grad decoration request vs broad import errors after upgrade; unrelated." + "reason": "Cache update no_grad decoration proposal versus a broad import-error report after upgrade; not the same change." }, { "left": "issue:36683", "right": "issue:43352", "accept": false, - "reason": "Different model-specific issues: Gemma3 config field missing vs Nemotron FlashAttention support error." + "reason": "Missing config attribute in Gemma3 versus FlashAttention support limitation for NemotronH; different problems." }, { "left": "issue:43352", "right": "issue:43931", "accept": false, - "reason": "Unsupported FlashAttention vs weight-shape mismatch during load; distinct bugs." + "reason": "FlashAttention support error for one model versus weight-shape mismatch when loading Qwen3-VL; not duplicates." }, { "left": "issue:41669", "right": "issue:44246", "accept": false, - "reason": "One is about removing import * for startup cost, the other about intermittent import slowness; related topic but not the same defect." + "reason": "Import-star performance cleanup versus import latency sometimes being slow; the latter is a symptom report, not the same code change." }, { "left": "issue:43867", "right": "issue:45375", "accept": false, - "reason": "Sorted state_dict loading error vs missing deepstack_visual_indexes being dropped by strict config handling; different root causes." + "reason": "Both are configuration/loading issues, but one is state_dict ordering and the other is a missing field silently dropped by strict config handling." }, { "left": "issue:36683", "right": "issue:43064", "accept": false, - "reason": "Gemma3 config attribute issue vs FSDP2 optimizer-state corruption on nonzero ranks; unrelated." + "reason": "Unrelated: config missing vocab_size versus distributed training optimizer-state corruption." }, { "left": "issue:43992", "right": "issue:45478", "accept": false, - "reason": "Both are Qwen3.5 loading problems, but they concern different model variants and different failure modes." + "reason": "UMT5Encoder missing weight and Qwen3.5 MoE from_pretrained error are separate model-loading failures." }, { "left": "issue:44938", "right": "issue:45071", "accept": false, - "reason": "Python 3.14 import failure vs PretrainedConfig type-checking regression; not the same bug." + "reason": "Python 3.14 load failure versus type-checking breakage in v5.4.0; different regression classes." }, { "left": "issue:43867", "right": "issue:45357", "accept": false, - "reason": "One is load failure from sorted state_dict, the other is save_pretrained writing wrong visual encoder keys." + "reason": "State_dict ordering load bug versus incorrect visual encoder keys in save_pretrained; related only at a high level." }, { "left": "issue:43010", "right": "issue:43976", "accept": false, - "reason": "Cache no_grad API change vs Python version incompatibility; unrelated." + "reason": "Cache no_grad decoration change is unrelated to Python-version compatibility." }, { "left": "issue:44938", "right": "issue:45381", "accept": false, - "reason": "Python 3.14 load failure and Qwen2.5-VL vision_position_ids bug are separate regressions." + "reason": "Python 3.14 import/load failure versus qwen2.5-vl video position-id bug; different failure modes." }, { "left": "issue:44360", "right": "issue:45440", "accept": false, - "reason": "DSA indexer missing ReLU vs DeepseekV3MoE divergence from remote implementation; different functionality." + "reason": "DSA indexer missing ReLU versus DeepseekV3 MoE implementation divergence; unrelated code paths." }, { "left": "issue:43618", "right": "issue:44077", "accept": false, - "reason": "CLIPOutput attentions regression vs patchtsmixer post_init restriction; unrelated." + "reason": "CLIP attentions regression versus patchtsmixer post_init policy; not the same bug." }, { "left": "issue:36296", "right": "issue:43575", "accept": false, - "reason": "Both mention tensor parallelism, but one is a training correctness bug and the other is expected OOM for a very large model." + "reason": "Tensor-parallel training bug versus OOM when loading a specific Qwen2-57B model with tp; different symptoms and likely different root causes." }, { "left": "issue:43867", "right": "issue:44484", "accept": false, - "reason": "Load error when state_dict is sorted vs a question about max_shard_size default; no shared defect." + "reason": "State_dict load ordering bug versus a question about max_shard_size default; not duplicates." }, { "left": "issue:43618", "right": "issue:44857", "accept": false, - "reason": "Missing attentions in CLIPOutput vs float16 AMP crash in LwDetrImageLoss; different areas entirely." + "reason": "CLIPOutput attentions assignment issue versus LwDetrImageLoss AMP/CUDA crash; different areas." }, { "left": "issue:42890", "right": "issue:44964", "accept": false, - "reason": "Stale device override CI test vs Phi-4 multimodal load failure; unrelated." + "reason": "Test flakiness from missing set_seed versus model load failure for Phi-4 multimodal; unrelated." }, { "left": "issue:28282", "right": "issue:33453", "accept": false, - "reason": "PyTorch-missing ImportError for AutoModel vs tokenizer-loading regression; not the same issue." + "reason": "ImportError for missing PyTorch dependency versus tokenizer loading regression; only a superficial import-related similarity." }, { - "left": "issue:42890", - "right": "issue:44112", + "left": "issue:43618", + "right": "issue:44977", "accept": false, - "reason": "Both mention stale device override tests, but the actual failing models and test cases are different." + "reason": "CLIP output attentions regression versus Qwen3.5 flash-attention generation bug; different models and code paths." } ] }, @@ -12630,7 +12636,6 @@ "issue:43299", "issue:43502", "issue:43576", - "issue:43618", "issue:43673", "issue:43723", "issue:43824", @@ -12638,6 +12643,7 @@ "issue:43874", "issue:44062", "issue:44075", + "issue:44112", "issue:44222", "issue:44230", "issue:44315", @@ -12654,27 +12660,27 @@ "issue:44749", "issue:44871", "issue:44961", - "issue:44977", "issue:45071", "issue:45125", "issue:45341", "issue:45375", + "issue:45468", "issue:45478" ], "soft_pairs": [ - "issue:43618|issue:44977", - "issue:43502|issue:44062", + "issue:42890|issue:44112", "issue:43010|issue:44075", "issue:42890|issue:45341", "issue:43723|issue:44556", - "issue:44222|issue:45125", + "issue:43502|issue:44062", "issue:44387|issue:44556", + "issue:44222|issue:45125", "issue:43010|issue:43065", "issue:43010|issue:43232", - "issue:42548|issue:44351", "issue:45125|issue:45478", - "issue:44355|issue:44749", "issue:45125|issue:45375", + "issue:42548|issue:44351", + "issue:44355|issue:44749", "issue:30990|issue:31795", "issue:43502|issue:44485", "issue:44075|issue:44871", @@ -12697,246 +12703,273 @@ "issue:43576|issue:45071", "issue:43673|issue:44492", "issue:38617|issue:43576", - "issue:33357|issue:33453" + "issue:33357|issue:33453", + "issue:44961|issue:45468", + "issue:31795|issue:38617", + "issue:43576|issue:44351" ], - "prepared_review_unit_hash": "102bd00af91a52fed393764fcd2a8f11a9e55fd5e987b99f836cd37f3d1f048b", + "prepared_review_unit_hash": "5f49b31a26a7e2462be7fe077a1ce91815482ceb94dc648b1a1880d9cd3b4729", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 36, - "serialized_chars": 37891, - "estimated_input_tokens": 9473, - "estimated_eval_tokens": 19202 + "soft_pair_count": 39, + "serialized_chars": 38798, + "estimated_input_tokens": 9700, + "estimated_eval_tokens": 19656 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly a grab-bag of Transformers v5 regressions and unrelated model/test/docs issues. The only clear common thread is broad loading/import breakage, but none of the soft pairs look like true duplicates.", - "confidence": 0.69, - "canonical_issue_reason": "44556 is the broadest representative of the shared theme: checkpoint reload breaks after upgrading from v4.57 to v5.x. It best serves as an umbrella issue for the loading-regression subset.", + "summary": "These are mostly unrelated issue reports; I did not find any soft pair that clearly describes the same underlying bug/change. The strongest single issue is the concrete Qwen3.5 MoE v5.4.0 loading regression, but there is no PR in the set.", + "confidence": 0.81, + "canonical_issue_reason": "issue:45478 is the most concrete and actionable report: it names the affected version, model family, and failure mode (`from_pretrained` error), making it a good anchor issue.", "canonical_pr_reason": null, - "best_issue_reason": "44556 is the most globally representative issue in this set because it captures the main v5 migration/loading regression pattern without being overly model-specific.", + "best_issue_reason": "issue:45478 has a specific regression description and a narrow code path, so it is the best standalone issue candidate here.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43618", - "right": "issue:44977", - "accept": false, - "reason": "Different failures: missing CLIPOutput attentions vs Qwen3.5 generation with flash-attention." - }, - { - "left": "issue:43502", - "right": "issue:44062", + "left": "issue:42890", + "right": "issue:44112", "accept": false, - "reason": "Unrelated bugs: offline-mode network requests vs AddedToken constructor argument collision." + "reason": "Both are test failures, but they concern different models and different failure modes; not the same bug." }, { "left": "issue:43010", "right": "issue:44075", "accept": false, - "reason": "Different code paths: cache update no_grad behavior vs SGD optimizer args not being used." + "reason": "Cache update decoration and SGD argument handling are unrelated code paths." }, { "left": "issue:42890", "right": "issue:45341", "accept": false, - "reason": "Both are test-related, but one is flaky seeds in SamHQ integration tests and the other is a bug in testing_utils.py." + "reason": "One is a specific missing set_seed in a model integration test; the other is a generic testing_utils bug with no clear shared defect." }, { "left": "issue:43723", "right": "issue:44556", "accept": false, - "reason": "Both are loading regressions, but one is tokenizer loading and the other is checkpoint reload after v5 upgrade; not the same concrete bug." + "reason": "Both involve loading regressions in v5, but one is tokenizer loading and the other is checkpoint reload compatibility; too broad and not the same concrete bug." }, { - "left": "issue:44222", - "right": "issue:45125", + "left": "issue:43502", + "right": "issue:44062", "accept": false, - "reason": "Different MoE/FP8 issues: save_pretrained behavior vs missing tensor-parallel plan." + "reason": "Local-files-only network requests and AddedToken keyword duplication are unrelated." }, { "left": "issue:44387", "right": "issue:44556", "accept": false, - "reason": "Int4 quantization OOM is a memory/quantization issue, not a checkpoint reload regression." + "reason": "Quantization OOM and checkpoint reload failure are different problems." + }, + { + "left": "issue:44222", + "right": "issue:45125", + "accept": false, + "reason": "Both mention Qwen3.5 MoE/FP8-related work, but one is save_pretrained FP8 and the other is tensor-parallel metadata; not the same issue." }, { "left": "issue:43010", "right": "issue:43065", "accept": false, - "reason": "Cache update semantics and a dummy Conv2d in Sam3PixelDecoder are unrelated." + "reason": "Cache update no_grad and Sam3PixelDecoder dummy Conv2d are unrelated." }, { "left": "issue:43010", "right": "issue:43232", "accept": false, - "reason": "Both touch generation utilities, but the specific bugs are different: no_grad on cache update vs sync_gpus kwargs handling." + "reason": "Cache update no_grad and generation kwargs sync behavior are different code paths." }, { - "left": "issue:42548", - "right": "issue:44351", + "left": "issue:45125", + "right": "issue:45478", "accept": false, - "reason": "Both are import errors, but for different symbols and modules." + "reason": "Both concern Qwen3.5 MoE loading, but one is missing tensor-parallel plan and the other is a broader from_pretrained regression; not clearly the same concrete bug." }, { "left": "issue:45125", - "right": "issue:45478", + "right": "issue:45375", "accept": false, - "reason": "Same model family, but one is missing _tp_plan and the other is a broader from_pretrained failure; not the same bug." + "reason": "Tensor parallel metadata and a missing vision config field are unrelated." }, { - "left": "issue:44355", - "right": "issue:44749", + "left": "issue:42548", + "right": "issue:44351", "accept": false, - "reason": "Compiled-file errors and a Chinese-described performance slowdown are unrelated." + "reason": "Both are import errors, but for different symbols and likely different export breaks." }, { - "left": "issue:45125", - "right": "issue:45375", + "left": "issue:44355", + "right": "issue:44749", "accept": false, - "reason": "Both are Qwen3.5 MoE-related, but one is tensor-parallel metadata and the other is a missing config field dropped by strict validation." + "reason": "Running compiled Python files and slower filtering after upgrade are unrelated." }, { "left": "issue:30990", "right": "issue:31795", "accept": false, - "reason": "Sentence Transformers loading stall and a forward-docstring issue are unrelated." + "reason": "Sentence Transformers loading hangs versus forward doc confusion are unrelated." }, { "left": "issue:43502", "right": "issue:44485", "accept": false, - "reason": "Offline network requests and GLM-5 RoPE implementation are different subsystems and symptoms." + "reason": "Offline/local-file request behavior and GLM-5 RoPE implementation are unrelated." }, { "left": "issue:44075", "right": "issue:44871", "accept": false, - "reason": "Optimizer argument handling and Gemma eos_token_id config mismatch are unrelated." + "reason": "SGD optimizer arguments and eos_token_id config mismatch are different bugs." }, { "left": "issue:43010", "right": "issue:44368", "accept": false, - "reason": "Cache update decoration and tie_word_embeddings warning are different issues." + "reason": "Cache update no_grad and a tie_word_embeddings warning are unrelated." }, { "left": "issue:44315", "right": "issue:44360", "accept": false, - "reason": "Liger Kernel application in model_init and DSA indexer ReLU are unrelated." + "reason": "Liger kernel application and DSA indexer activation absence are unrelated." }, { "left": "issue:42890", "right": "issue:44373", "accept": false, - "reason": "Test seed flakiness and a wrong docstring are not the same defect." + "reason": "A flaky integration test and a docstring typo are not the same issue." }, { "left": "issue:38175", "right": "issue:43502", "accept": false, - "reason": "Unexpected zero probabilities in SigLIP2 is unrelated to local_files_only network access." + "reason": "Zero probabilities in a model and unwanted API requests are unrelated." }, { "left": "issue:43299", "right": "issue:44556", "accept": false, - "reason": "Both are load regressions, but for different model families and code paths." + "reason": "Both are version-related loading regressions, but they affect different models and failure modes." }, { "left": "issue:43502", "right": "issue:44961", "accept": false, - "reason": "Local-files-only network leakage and the unrelated 'racoon' issue do not match." + "reason": "Local-files-only networking and a one-word title are unrelated." }, { "left": "issue:44355", "right": "issue:44655", "accept": false, - "reason": "Compiled Python-file errors and pipeline save_pretrained support are unrelated." + "reason": "Compiled Python file errors and pipeline save_pretrained failures are different code paths." }, { "left": "issue:44222", "right": "issue:44230", "accept": false, - "reason": "Both mention FP8/MoE, but one is save_pretrained and the other is inference support; different bugs." + "reason": "Both are FP8/Qwen3.5 MoE themed, but one is save_pretrained and the other is inference support; not the same bug." }, { "left": "issue:44355", "right": "issue:45071", "accept": false, - "reason": "Compiled-file runtime errors and PretrainedConfig type-checking are unrelated." + "reason": "Compiled file execution errors and PretrainedConfig typing issues are unrelated." }, { "left": "issue:42831", "right": "issue:43874", "accept": false, - "reason": "FineGrainedFP8 accuracy drift and a missing image-processor method are different defects." + "reason": "FineGrainedFP8 accuracy and GLM46V image processor token counting are unrelated." }, { "left": "issue:42890", "right": "issue:43232", "accept": false, - "reason": "Flaky tests due to seeding and generation kwargs after sync_gpus are unrelated." + "reason": "A failing seed-dependent test and generation cache kwargs handling are unrelated." }, { "left": "issue:33453", "right": "issue:34689", "accept": false, - "reason": "Both are model-loading regressions, but for different models and different failure modes." + "reason": "Tokenizer loading regression and Llama 3.2 Vision model loading regression are different problems." }, { "left": "issue:42757", "right": "issue:43576", "accept": false, - "reason": "ImportError for is_offline_mode and a broken transformers env command are unrelated." + "reason": "Missing `is_offline_mode` import and broken `transformers env` CLI 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 issue." + "reason": "Tokenizer loading regression and tensor parallel training bug are unrelated." }, { "left": "issue:43827", "right": "issue:44961", "accept": false, - "reason": "Docs still referencing pipeline() and 'racoon' are unrelated." + "reason": "Docs still referencing pipeline() and a one-word issue title are not duplicates." }, { "left": "issue:43576", "right": "issue:43824", "accept": false, - "reason": "Broken env command and missing Qwen2_5_VLForConditionalGeneration import are different failures." + "reason": "Broken env command and missing Qwen2_5_VL export are both v5 regressions but not the same concrete change." + }, + { + "left": "issue:43576", + "right": "issue:45071", + "accept": false, + "reason": "CLI breakage and PretrainedConfig type checking are unrelated." }, { "left": "issue:43673", "right": "issue:44492", "accept": false, - "reason": "Generation cache missing in chunked_prefill and a cache-strategy typo are not the same bug." + "reason": "Missing cache in generation and a typo in cache strategies are unrelated." }, { "left": "issue:38617", "right": "issue:43576", "accept": false, - "reason": "Importing layer_type_validation and the env command failure are unrelated." + "reason": "Missing import in configuration_utils and broken env command are different failures." }, { "left": "issue:33357", "right": "issue:33453", "accept": false, - "reason": "Different regressions: MacOS bus error on CLIP vs tokenizer loading regression." + "reason": "MacOS bus error in CLIP and tokenizer loading regression are unrelated." + }, + { + "left": "issue:44961", + "right": "issue:45468", + "accept": false, + "reason": "No substantive signal in the one-word issue; not the same as Gemma-4 audio positional encoding." + }, + { + "left": "issue:31795", + "right": "issue:38617", + "accept": false, + "reason": "Doc confusion about forward args and an import error are unrelated." + }, + { + "left": "issue:43576", + "right": "issue:44351", + "accept": false, + "reason": "Broken env command and missing HybridCache import are different v5 export/CLI issues." } ] }, @@ -12963,7 +12996,6 @@ "issue:31795", "issue:33453", "issue:36296", - "issue:38617", "issue:41628", "issue:41669", "issue:42371", @@ -12975,6 +13007,8 @@ "issue:43526", "issue:43576", "issue:43673", + "issue:43723", + "issue:43827", "issue:43901", "issue:43976", "issue:44038", @@ -13005,14 +13039,10 @@ "issue:45125", "issue:45230", "issue:45341", - "issue:45468", "issue:45478", "issue:45542" ], "soft_pairs": [ - "issue:44961|issue:45468", - "issue:31795|issue:38617", - "issue:43576|issue:44351", "issue:44908|issue:44961", "issue:41628|issue:42548", "issue:36296|issue:43901", @@ -13047,246 +13077,267 @@ "issue:44315|issue:44829", "issue:44393|issue:45478", "issue:44077|issue:44861", - "issue:31795|issue:33453" + "issue:43827|issue:44908", + "issue:31795|issue:33453", + "issue:41669|issue:44908", + "issue:43723|issue:44393" ], - "prepared_review_unit_hash": "ae419bdddcfe4001e061f812d8f51e72be5ffe7a7fd5aa86252487dc5e24613b", + "prepared_review_unit_hash": "584dfc293d0aaafafb6643e650f8a72c036679f28ad65d9fee03c8918ff4a21d", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 38, - "serialized_chars": 38384, - "estimated_input_tokens": 9596, - "estimated_eval_tokens": 19448 + "serialized_chars": 38388, + "estimated_input_tokens": 9597, + "estimated_eval_tokens": 19450 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous: it mixes import/export regressions, lazy-loading/performance problems, model-specific bugs, docs issues, and unrelated training/runtime errors. None of the soft pairs look like the same concrete bug or change.", - "confidence": 0.91, - "canonical_issue_reason": "No single issue is representative; the items do not form one duplicate set and span unrelated subsystems and failure modes.", + "summary": "This cluster is mostly heterogeneous: a few issues are in the import/lazy-loading area, but most pairs are unrelated regressions, model bugs, docs problems, or version-compatibility reports. No PRs are present.", + "confidence": 0.71, + "canonical_issue_reason": "issue:41669 is the most concrete and representative item in the cluster, with a clear root cause around import-time slowdown and module loading behavior; it best matches the small import/lazy-loading theme that appears across a few nearby issues.", "canonical_pr_reason": null, - "best_issue_reason": "No issue is a good global canonical for this cluster because the themes are too diverse and the pairwise similarities are mostly superficial.", + "best_issue_reason": "Best triage anchor because it is specific, actionable, and broad enough to represent the limited shared import-related surface in this set.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44961", - "right": "issue:45468", - "accept": false, - "reason": "Unrelated titles: a nonsensical issue title vs a Gemma-4 audio positional-encoding bug." - }, - { - "left": "issue:31795", - "right": "issue:38617", - "accept": false, - "reason": "Documentation confusion about forward args vs an import error for a missing symbol." - }, - { - "left": "issue:43576", - "right": "issue:44351", - "accept": false, - "reason": "Both involve v5 imports, but one is a broken env command and the other is a missing cache class export." - }, { "left": "issue:44908", "right": "issue:44961", "accept": false, - "reason": "Completely unrelated: scheduler kwargs vs an unrelated placeholder issue title." + "reason": "Different domains: scheduler kwargs bug vs. an unrelated placeholder issue." }, { "left": "issue:41628", "right": "issue:42548", "accept": false, - "reason": "Both are import errors, but for different top-level symbols and likely different missing exports." + "reason": "Both are import errors, but they concern different missing symbols and no concrete shared code-path is shown." }, { "left": "issue:36296", "right": "issue:43901", "accept": false, - "reason": "Tensor-parallel training bug vs pipeline docs mismatch; different subsystem and failure mode." + "reason": "Tensor-parallel training bug vs. pipeline docs mismatch; unrelated." }, { "left": "issue:44485", "right": "issue:44961", "accept": false, - "reason": "RoPE implementation discussion is unrelated to the placeholder issue." + "reason": "RoPE implementation discussion vs. unrelated placeholder issue." }, { "left": "issue:36296", "right": "issue:43526", "accept": false, - "reason": "Tensor parallelism bug vs BeitImageProcessorFast label-reduction bug." + "reason": "Tensor parallel training bug and a BeitImageProcessorFast label bug are unrelated." }, { "left": "issue:36296", "right": "issue:43012", "accept": false, - "reason": "Training parallelism bug vs bfloat16 compilation warning; no shared code-path." + "reason": "Different subsystems and symptoms: tensor-parallel training vs. bfloat16 compile warning." }, { "left": "issue:44273", "right": "issue:45003", "accept": false, - "reason": "Both mention loading/import behavior, but one is a vague lazy-loading failure and the other is a specific sys.modules access bug; not clearly the same defect." + "reason": "Both touch loading behavior, but one is a general lazy-loading report and the other a specific sys.modules access issue; not enough to treat as the same bug." }, { "left": "issue:41628", "right": "issue:44355", "accept": false, - "reason": "Importing a symbol from transformers vs errors in compiled Python files; unrelated." + "reason": "Top-level import failure vs. compiled-file runtime error; no shared bug evidence." }, { "left": "issue:44261", "right": "issue:44360", "accept": false, - "reason": "MLA RMS norm epsilon precision issue vs a DSA indexer ReLU omission." + "reason": "MLA precision/config issue vs. DSA indexer/ReLU discussion; unrelated." }, { "left": "issue:44273", "right": "issue:45542", "accept": false, - "reason": "Lazy-loading problem vs undefined TF backend when only tensorboard is installed." + "reason": "Lazy loading issue vs. TensorFlow backend detection issue; different code paths." }, { "left": "issue:44623", "right": "issue:45230", "accept": false, - "reason": "Processor save_pretrained missing files vs a generic bug report with no concrete overlap." + "reason": "Processor save_pretrained missing files vs. a generic bug report; no match." }, { "left": "issue:43576", "right": "issue:44556", "accept": false, - "reason": "v5 env-command breakage vs checkpoint reload incompatibility; similar release era, but different concrete failures." + "reason": "transformers env command breakage vs. checkpoint reload incompatibility; different failures." }, { "left": "issue:41669", "right": "issue:43673", "accept": false, - "reason": "Import-time slowdown from model imports vs missing GenerationMixin cache in chunked prefill." + "reason": "Import-slowdown / star-import issue vs. GenerationMixin cache missing during chunked prefill; not the same bug." }, { "left": "issue:44393", "right": "issue:45071", "accept": false, - "reason": "Qwen3-VL bbox output issue vs PretrainedConfig type-checking regression; different code paths." + "reason": "Qwen3-VL bounding-box output issue vs. PretrainedConfig type-checking regression; unrelated." }, { "left": "issue:44038", "right": "issue:45125", "accept": false, - "reason": "Both mention Qwen3-Moe, but one is a general v5/Qwen3-VL-Moe bug and the other is a missing tensor-parallel plan for Qwen3_5Moe." + "reason": "Both mention Qwen models, but one is a general bug report and the other is a tensor-parallel metadata omission; not the same issue." }, { "left": "issue:44556", "right": "issue:45092", "accept": false, - "reason": "Both are checkpoint/loading compatibility issues, but one is generic version upgrade reload failure and the other is remote-code/meta-init incompatibility; too broad to call duplicates." + "reason": "Upgrade/reload incompatibility vs. meta-init remote-code incompatibility; related timeframe, different concrete problems." }, { "left": "issue:42371", "right": "issue:44492", "accept": false, - "reason": "TF32 API guidance vs a typo in cache strategies." + "reason": "TF32 settings warning vs. a cache-strategy typo; unrelated." }, { "left": "issue:42890", "right": "issue:44908", "accept": false, - "reason": "Test flakiness from missing set_seed vs scheduler kwargs handling." + "reason": "Test seeding flakiness vs. scheduler kwargs bug; unrelated." }, { "left": "issue:44704", "right": "issue:44908", "accept": false, - "reason": "AutoProcessor cached_file kwargs bug vs scheduler kwargs bug; different APIs and behavior." + "reason": "AutoProcessor kwargs forwarding vs. scheduler kwarg handling; different code paths." }, { "left": "issue:44246", "right": "issue:44273", "accept": false, - "reason": "Import performance complaint vs lazy-loading failure; related theme, not the same bug." + "reason": "Import slowness and lazy loading are adjacent, but these are still distinct reported failures without concrete duplicate evidence." }, { "left": "issue:43408", "right": "issue:44315", "accept": false, - "reason": "Model-type warning in SAM3 vs Liger Kernel not applied with model_init." + "reason": "sam3_tracker type warning vs. Liger Kernel model_init behavior; unrelated." }, { "left": "issue:44297", "right": "issue:45230", "accept": false, - "reason": "Tokenizer save_pretrained mismatch vs generic bug report." + "reason": "Tokenizer_class metadata mismatch vs. generic bug report; no duplicate signal." + }, + { + "left": "issue:43576", + "right": "issue:44908", + "accept": false, + "reason": "Broken env command vs. inverse_sqrt scheduler kwargs bug; unrelated." }, { "left": "issue:43352", "right": "issue:44393", "accept": false, - "reason": "Flash Attention 2 unsupported in Nemotron vs Qwen3-VL bbox hallucination/error." + "reason": "Flash Attention support error vs. Qwen3-VL bounding-box hallucination; unrelated." + }, + { + "left": "issue:44908", + "right": "issue:45071", + "accept": false, + "reason": "Scheduler bug vs. config type-checking regression; unrelated." }, { "left": "issue:43976", "right": "issue:44393", "accept": false, - "reason": "Python version compatibility issue vs a vision-language output bug." + "reason": "Python-version compatibility vs. vision-language output issue; unrelated." }, { "left": "issue:44295", "right": "issue:45230", "accept": false, - "reason": "position_ids buffer access error vs generic bug report." + "reason": "position_ids buffer access bug vs. generic bug report; no match." + }, + { + "left": "issue:41669", + "right": "issue:44351", + "accept": false, + "reason": "Import * slowdown and missing HybridCache import are both import-related, but they are different missing-export problems with no proof of a single root cause." + }, + { + "left": "issue:41669", + "right": "issue:45341", + "accept": false, + "reason": "Import slowdown issue vs. a testing_utils bug report; unrelated." + }, + { + "left": "issue:44492", + "right": "issue:44945", + "accept": false, + "reason": "Typo in cache strategies vs. incorrect pipeline-parallel LLM output; unrelated." + }, + { + "left": "issue:41669", + "right": "issue:44492", + "accept": false, + "reason": "Import-time performance regression vs. a typo; not the same underlying bug." }, { "left": "issue:44315", "right": "issue:44829", "accept": false, - "reason": "Liger Kernel application problem vs degenerate training with flash_attention_3; distinct training issues." + "reason": "Liger Kernel model_init behavior vs. flash_attention_3 training degeneration; different failures." }, { - "left": "issue:44351", - "right": "issue:44861", + "left": "issue:44393", + "right": "issue:45478", "accept": false, - "reason": "Missing HybridCache export vs tied-weights key AttributeError." + "reason": "Different Qwen3-VL/Qwen3.5-Moe bugs with different symptoms and code paths." }, { - "left": "issue:44360", - "right": "issue:45003", + "left": "issue:44077", + "right": "issue:44861", "accept": false, - "reason": "DSA indexer implementation issue vs sys.modules access in modeling_utils." + "reason": "Optional post_init handling vs. tied-weights key crash; unrelated." }, { - "left": "issue:44945", - "right": "issue:44961", + "left": "issue:43827", + "right": "issue:44908", "accept": false, - "reason": "Pipeline parallelism output corruption vs placeholder issue title." + "reason": "Docs still reference pipeline() vs. scheduler kwargs bug; unrelated." }, { - "left": "issue:45003", - "right": "issue:45071", + "left": "issue:31795", + "right": "issue:33453", "accept": false, - "reason": "Unsafe sys.modules access vs PretrainedConfig typing regression." + "reason": "Documentation ambiguity vs. tokenizer-loading regression; unrelated." }, { "left": "issue:41669", - "right": "issue:44492", + "right": "issue:44908", "accept": false, - "reason": "Import-star performance regression vs a typo in cache strategy docs/code." + "reason": "Import-time slowdown vs. scheduler kwargs bug; no shared code-path." }, { - "left": "issue:31795", - "right": "issue:33453", + "left": "issue:43723", + "right": "issue:44393", "accept": false, - "reason": "Docs confusion in model.forward vs tokenizer loading regression; no shared bug." + "reason": "Tokenizer loading in v5 vs. Qwen3-VL bounding-box output issue; unrelated." } ] }, @@ -13332,15 +13383,12 @@ "issue:43606", "issue:43618", "issue:43704", - "issue:43723", - "issue:43827", "issue:44230", "issue:44261", "issue:44263", "issue:44273", "issue:44315", "issue:44360", - "issue:44393", "issue:44485", "issue:44492", "issue:44556", @@ -13357,9 +13405,6 @@ "issue:45468" ], "soft_pairs": [ - "issue:43827|issue:44908", - "issue:41669|issue:44908", - "issue:43723|issue:44393", "issue:44655|issue:44861", "issue:43576|issue:44273", "issue:44360|issue:45468", @@ -13399,268 +13444,262 @@ "issue:31515|issue:36683", "issue:43298|issue:44230" ], - "prepared_review_unit_hash": "91b8a4b00e517b7f2d11a8fcd7dd8383ff091f008d7a742f64cbed4a58c7a950", + "prepared_review_unit_hash": "6b21982e9ae1be24051b1e48ff3456fe30fbad620907e97fe52d8baf20c3806c", "component_budget": { "node_count": 364, "item_count": 364, "soft_pair_count": 1360, - "serialized_chars": 701640, - "estimated_input_tokens": 175410, - "estimated_eval_tokens": 351076 + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 }, "budget": { - "node_count": 45, - "item_count": 45, - "soft_pair_count": 41, - "serialized_chars": 37507, - "estimated_input_tokens": 9377, - "estimated_eval_tokens": 19010 + "node_count": 42, + "item_count": 42, + "soft_pair_count": 38, + "serialized_chars": 34909, + "estimated_input_tokens": 8728, + "estimated_eval_tokens": 17712 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a noisy mix of Transformers v5 loading/tokenizer/config regressions and several unrelated model-specific bugs. There is no strong exact-duplicate pair; the best anchor is a broad checkpoint reload compatibility issue, but confidence is low-moderate.", - "confidence": 0.43, - "canonical_issue_reason": "Issue 44556 is the broadest, clearest representative of the main cluster theme: checkpoint reload incompatibility after upgrading to v5. It is more general than the model-specific loading reports and closer to a canonical tracker issue.", + "summary": "This cluster is highly heterogeneous: mostly unrelated issue reports around loading, tokenizer/config regressions, model-specific bugs, and docs/test noise. No soft-edge pair looks like the same concrete underlying bug or change, and there are no PRs to consolidate.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 44556 is the broadest loading/v5 compatibility report and is the closest thing to an umbrella issue here; however, the set is not actually a true duplicate cluster.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 44556 is the best single issue to anchor this set because it captures the overarching v4-to-v5 loading regression theme that recurs across several nearby reports, even though the cluster itself is heterogeneous.", + "best_issue_reason": "44556 best represents the recurring checkpoint/model-loading regression theme among otherwise unrelated items.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43827", - "right": "issue:44908", - "accept": false, - "reason": "Different subsystems: pipeline-removal docs issue vs scheduler keyword handling." - }, - { - "left": "issue:41669", - "right": "issue:44908", - "accept": false, - "reason": "Import-star performance/cleanup issue is unrelated to inverse_sqrt scheduler kwargs." - }, - { - "left": "issue:43723", - "right": "issue:44393", - "accept": false, - "reason": "Tokenizer loading regression and Qwen3-VL bbox output bug are different code paths." - }, { "left": "issue:44655", "right": "issue:44861", "accept": false, - "reason": "Saving Pipeline objects and tied-weight-key crash are separate save-path bugs." + "reason": "Both mention save/pretrained-adjacent code, but one is Pipeline serialization and the other is tied-weight key handling; different bugs." }, { "left": "issue:43576", "right": "issue:44273", "accept": false, - "reason": "Broken env command and lazy loading failure are different features and failure modes." + "reason": "Different areas: CLI/env command breakage vs lazy loading behavior; no shared code-path issue." }, { "left": "issue:44360", "right": "issue:45468", "accept": false, - "reason": "Unrelated model internals: DSA indexer ReLU vs Gemma audio positional encoding." + "reason": "Unrelated model internals: DSA indexer/ReLU vs Gemma-4 audio positional encoding." }, { "left": "issue:45230", "right": "issue:45310", "accept": false, - "reason": "Generic bug report vs concrete Qwen3.5 MoE from_pretrained error; not the same bug." + "reason": "Generic bug report vs a specific Qwen3.5 MoE from_pretrained regression; not the same issue." }, { "left": "issue:44273", "right": "issue:44908", "accept": false, - "reason": "Lazy loading problem is unrelated to lr_scheduler_kwargs handling." + "reason": "Lazy loading and inverse_sqrt scheduler kwargs are unrelated subsystems." }, { "left": "issue:41669", "right": "issue:43704", "accept": false, - "reason": "Model import cleanup and VRAM leak in dataloader threads are unrelated." + "reason": "Import-star cleanup/perf issue vs VRAM leak in dataloader threads; different change and failure mode." }, { "left": "issue:43519", "right": "issue:44485", "accept": false, - "reason": "Timestamp calculation in Qwen3VL and GLM-5 RoPE implementation are different model-specific bugs." + "reason": "Qwen3VL timestamp calculation vs GLM-5 RoPE implementation; unrelated model logic." }, { "left": "issue:43519", "right": "issue:44908", "accept": false, - "reason": "Timestamp calculation and scheduler kwargs are unrelated." + "reason": "Processor timestamp math and scheduler kwargs are unrelated bugs." }, { "left": "issue:31795", "right": "issue:34689", "accept": false, - "reason": "Doc confusion in forward args does not match a model-loading regression." + "reason": "Documentation confusion for forward args vs a concrete model-loading regression." }, { "left": "issue:31795", "right": "issue:37428", "accept": false, - "reason": "Documentation issue and flash-attention import error are unrelated." + "reason": "Docs issue vs ImportError for flash attention utility; not the same bug." }, { "left": "issue:44315", "right": "issue:45092", "accept": false, - "reason": "Both involve model creation/loading, but Liger kernel application and remote-code meta-init incompatibility are distinct code paths." + "reason": "Liger kernel application in model_init is not the same as remote-code checkpoint/meta-init incompatibility." }, { "left": "issue:43704", "right": "issue:44485", "accept": false, - "reason": "VRAM leak in multi-threaded inference is unrelated to RoPE implementation." + "reason": "VRAM leak in multi-threaded inference vs RoPE implementation bug." }, { "left": "issue:44485", "right": "issue:45468", "accept": false, - "reason": "Both mention positional-style internals, but they are different models and different bugs." + "reason": "Different model families and different math/encoding paths." }, { "left": "issue:42371", "right": "issue:43704", "accept": false, - "reason": "TF32 API guidance and VRAM leak are unrelated." + "reason": "TF32 API guidance issue vs memory leak bug; no overlap in underlying change." }, { "left": "issue:43502", "right": "issue:43519", "accept": false, - "reason": "Local-files-only network leakage and timestamp math are unrelated." + "reason": "local_files_only network leakage vs timestamp calculation in a processor." }, { "left": "issue:43519", "right": "issue:44623", "accept": false, - "reason": "Timestamp calculation bug and processor.save_pretrained missing files are different features." + "reason": "Processor timestamp math vs processor.save_pretrained missing files; different code paths." }, { "left": "issue:43519", "right": "issue:43704", "accept": false, - "reason": "Different model bugs: timestamp math vs VRAM leak." + "reason": "Processor math bug vs VRAM leak; unrelated." }, { "left": "issue:44263", "right": "issue:45468", "accept": false, - "reason": "torch.split return handling and Gemma audio encoding are unrelated." + "reason": "torch.split return values in one indexer vs Gemma-4 audio positional encoding." }, { "left": "issue:44261", "right": "issue:45468", "accept": false, - "reason": "Missing rms_norm_eps in MLA and Gemma audio positional encoding are different implementation bugs." + "reason": "Missing rms_norm_eps in MLA layernorm vs Gemma-4 audio positional encoding; different models." }, { "left": "issue:43606", "right": "issue:45468", "accept": false, - "reason": "CPU offload device mismatch is unrelated to Gemma audio positional encoding." + "reason": "CPU offload device mismatch in bark-small vs Gemma-4 audio positional encoding." }, { "left": "issue:42890", "right": "issue:43704", "accept": false, - "reason": "Flaky integration test seeding and VRAM leak are unrelated." + "reason": "Flaky integration test seeding issue vs a runtime VRAM leak." }, { "left": "issue:42913", "right": "issue:43066", "accept": false, - "reason": "Both are tokenizer-related v5 reports, but they describe different symptoms and likely different fixes." + "reason": "Both are tokenizer v5-related, but one is a broad behavior change and the other is a specific decoder-type bug; not clearly the same underlying defect." }, { "left": "issue:44908", "right": "issue:44998", "accept": false, - "reason": "Scheduler kwargs bug and unrelated placeholder issue." + "reason": "Scheduler kwargs bug vs an unrelated placeholder issue." }, { "left": "issue:44861", "right": "issue:45290", "accept": false, - "reason": "Tied-weight-key crash and chat template tool-call crash are distinct code paths." + "reason": "Tied-weight key crash and chat template tool-call crash are separate failures." }, { "left": "issue:44998", "right": "issue:45341", "accept": false, - "reason": "Placeholder unemployment issue is unrelated to testing_utils." + "reason": "Placeholder issue vs a testing_utils bug; unrelated." }, { "left": "issue:44492", "right": "issue:44998", "accept": false, - "reason": "Cache-strategy typo and unrelated placeholder issue." + "reason": "Typo in cache strategies vs unrelated placeholder issue." }, { "left": "issue:43618", "right": "issue:44998", "accept": false, - "reason": "CLIPOutput attentions regression is unrelated to the placeholder issue." + "reason": "CLIPOutput attention assignment regression vs unrelated placeholder issue." }, { "left": "issue:44556", "right": "issue:44998", "accept": false, - "reason": "Checkpoint reload regression and unrelated placeholder issue." + "reason": "Checkpoint reload regression vs unrelated placeholder issue." }, { - "left": "issue:44230", - "right": "issue:44623", + "left": "issue:43525", + "right": "issue:44861", + "accept": false, + "reason": "Llama4Config pad_token_id missing vs tied-weight key handling; different config fields and code paths." + }, + { + "left": "issue:43572", + "right": "issue:44861", "accept": false, - "reason": "FP8 inference support request and processor.save_pretrained missing files are unrelated." + "reason": "StableLmConfig pad_token_idx regression vs tied-weight key AttributeError; unrelated." }, { "left": "issue:31515", "right": "issue:38617", "accept": false, - "reason": "Checkpoint loading slowness and missing layer_type_validation import are different failures." + "reason": "Checkpoint loading slow vs ImportError for layer_type_validation; different failures." }, { "left": "issue:44230", - "right": "issue:45310", + "right": "issue:44623", "accept": false, - "reason": "FP8 support request is unrelated to Qwen3.5 MoE from_pretrained error." + "reason": "FP8 inference support request vs missing files on processor.save_pretrained; not the same bug." }, { "left": "issue:31515", "right": "issue:36296", "accept": false, - "reason": "Slow from_pretrained loading and tensor-parallel training bug are unrelated." + "reason": "from_pretrained slowdown vs tensor-parallel training bug." }, { "left": "issue:31515", "right": "issue:31795", "accept": false, - "reason": "Checkpoint load performance and forward-argument docs confusion are unrelated." + "reason": "Loading performance issue vs forward doc confusion." + }, + { + "left": "issue:44230", + "right": "issue:45310", + "accept": false, + "reason": "Both mention Qwen3.5 MoE, but one is fp8 inference support and the other is a from_pretrained regression; different changes." }, { "left": "issue:31515", "right": "issue:33453", "accept": false, - "reason": "Slow loading and tokenizer-loading regression are different symptoms, not the same bug." + "reason": "Both are loading-related, but one is checkpoint load slowness and the other is tokenizer loading regression." }, { "left": "issue:31515", "right": "issue:36683", "accept": false, - "reason": "Slow checkpoint loading and Gemma3Config missing vocab_size are unrelated." + "reason": "Checkpoint loading slowdown vs Gemma3Config missing vocab_size; unrelated." }, { "left": "issue:43298", "right": "issue:44230", "accept": false, - "reason": "Audio-model batching support and FP8 inference support are separate requests." + "reason": "Continuous batching for audio models vs fp8 support for Qwen models; different feature gaps." } ] }, @@ -13690,12 +13729,12 @@ "soft_pairs": [ "issue:26350|issue:33260" ], - "prepared_review_unit_hash": "614238314873a4bb7aa5a575268b45ebd5ff33c459a8a435a18facfb0a911fd9", + "prepared_review_unit_hash": "b8668984fa84439432f32317baa962fab8616fa98bd9c2bcecce906078d2ca4b", "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 }, @@ -13703,7 +13742,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1589, + "serialized_chars": 1590, "estimated_input_tokens": 398, "estimated_eval_tokens": 1052 }, @@ -13712,18 +13751,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are both community-contribution issues, but they target different features (Flash Attention 2 vs GGUF) and different code paths, so they should not be deduplicated.", - "confidence": 0.97, + "summary": "These are unrelated feature-request issues: one asks for Flash Attention 2 support on more architectures, the other asks for GGUF support on more architectures. They target different subsystems and different deliverables, so they should not be clustered as duplicates.", + "confidence": 0.98, "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "No single issue is a canonical representative here; the two issues are unrelated feature requests despite similar wording.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:26350", "right": "issue:33260", "accept": false, - "reason": "Different requested features with different underlying implementations; the shared phrasing about supporting more architectures is too generic to imply the same bug/change." + "reason": "Different features and code paths: Flash Attention 2 support vs GGUF support. They are both broad architecture-support requests but not the same underlying bug or change." } ] }, @@ -13760,12 +13799,12 @@ "issue:43834|issue:44365", "issue:44365|issue:44370" ], - "prepared_review_unit_hash": "adc76524be98365e6a2c4a601a5dc3947fc7c5ea63d96672bffec0a5377b18cf", + "prepared_review_unit_hash": "0aa0d148f039a45db1e8a688654b0d5beaa690844ef0fe5ec6683fb5b835b75a", "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 }, @@ -13773,7 +13812,7 @@ "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4455, + "serialized_chars": 4454, "estimated_input_tokens": 1114, "estimated_eval_tokens": 2484 }, @@ -13782,58 +13821,55 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The cluster is dominated by near-duplicate i18n translation-request issues. The three placeholder-titled issues look effectively identical; the Turkish-specific issue is related thematically but not the same concrete request.", - "confidence": 0.91, - "canonical_issue_reason": "Issue #43834 is the earliest of the near-identical placeholder issues, with the same title/body pattern as the others, so it is the cleanest representative of the duplicate set.", + "summary": "These look like the same issue template family, but only #27088 is a concrete Turkish translation request; the other three are generic placeholder variants and don\u2019t clearly represent the same actionable issue.", + "confidence": 0.84, + "canonical_issue_reason": "#27088 is the only concrete, language-specific issue title ([i18n-TR] Turkish). The others are generic placeholders, so this is the best representative issue.", "canonical_pr_reason": null, - "best_issue_reason": "#43834 is the best global issue candidate because it most clearly represents the repeated template issue and is earlier than the other near-identical copies.", + "best_issue_reason": "#27088 is the most actionable and specific issue in the cluster; it is clearly about Turkish docs translation rather than a templated placeholder.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:27088", "right": "issue:44370", "accept": false, - "reason": "Turkish-specific translation request vs a generic placeholder translation issue; same theme, but not the same concrete request." + "reason": "Same broad translation-template wording, but #44370 is generic and not clearly the same concrete issue as the Turkish-specific request." }, { "left": "issue:27088", "right": "issue:44365", "accept": false, - "reason": "Different concrete target language/request: Turkish versus a generic placeholder template, so this is not a true duplicate." + "reason": "High textual similarity, but the second item is a placeholder template rather than a specific Turkish docs translation issue." }, { "left": "issue:27088", "right": "issue:43834", "accept": false, - "reason": "The Turkish issue is more specific than the generic placeholder issue; they match on broad topic only, not the same underlying request." + "reason": "The titles are template-related, but #43834 is generic and does not clearly duplicate the concrete Turkish translation request." }, { "left": "issue:43834", "right": "issue:44370", - "accept": true, - "reason": "These two look like the same template issue: same placeholder title and near-identical body, with no evidence of a meaningful difference." + "accept": false, + "reason": "Both are generic template placeholders; similarity alone is not enough to confirm the same underlying issue." }, { "left": "issue:43834", "right": "issue:44365", - "accept": true, - "reason": "Near-identical placeholder translation issues with no visible substantive distinction." + "accept": false, + "reason": "Generic placeholder titles with no concrete language or unique issue details; not safe to merge as duplicates." }, { "left": "issue:44365", "right": "issue:44370", - "accept": true, - "reason": "Same placeholder title/body pattern and extremely high similarity indicate the same underlying issue." + "accept": false, + "reason": "Both look like the same template class, but there is no evidence they are the same specific issue." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet and the soft-edge verdicts are conservative. The three placeholder-titled issues are supported as near-identical by very high similarity scores and matching titles, and selecting the earliest as canonical is reasonable. The Turkish-specific issue is treated as related but not identical, which is appropriately cautious given the visible title difference and lack of direct evidence tying it to the same concrete request." - }, - "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 }, @@ -13891,12 +13927,12 @@ "issue:43837|issue:45027", "issue:43837|issue:45177" ], - "prepared_review_unit_hash": "d17088d4f578a396881eee9dc4a2740d788586f17c38678bff8400e5529ffa78", + "prepared_review_unit_hash": "9f53b89a9d1a3218b91f99047b8d0383600d32aeea36ef74aab699563d2f60a6", "component_budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 25, - "serialized_chars": 16009, + "serialized_chars": 16011, "estimated_input_tokens": 4003, "estimated_eval_tokens": 8262 }, @@ -13904,7 +13940,7 @@ "node_count": 13, "item_count": 13, "soft_pair_count": 25, - "serialized_chars": 16009, + "serialized_chars": 16011, "estimated_input_tokens": 4003, "estimated_eval_tokens": 8262 }, @@ -13913,162 +13949,162 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are not duplicates; they are separate model-request issues spanning different model families and even different modalities (text, audio, OCR, vision).", - "confidence": 0.98, - "canonical_issue_reason": "No single issue cleanly represents the whole cluster as a duplicate set. If one must be chosen as the broadest umbrella, issue 42738 is the most generic, but it still only covers a subset of the requests.", + "summary": "This cluster is not a duplicate set: every item is a separate model-request issue for different models, tasks, or modalities (BERT variants, TTS/ASR, audio, OCR, vision, sparse transformer). No PRs are present.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue fits the cluster because the requests are for different underlying models and capabilities, not the same bug or change.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 42738 is the broadest and least specific title ('BERT-like models with RoPE'), so it is the closest to an umbrella representative, though the cluster is too heterogeneous for true deduplication.", + "best_issue_reason": "If forced to pick one representative, issue 37015 is the most generic and active 'Add X' request, but it still only covers NeoBERT and does not generalize to the others.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:36817", "right": "issue:41211", "accept": false, - "reason": "Different model requests: EuroBert vs DEIMv2; same broad area, but not the same underlying change." + "reason": "Different model requests: EuroBERT vs DEIMv2; same broad category, not the same change." }, { "left": "issue:36817", "right": "issue:43671", "accept": false, - "reason": "EuroBert config support vs Qwen3-TTS support are unrelated models and modalities." + "reason": "EuroBERT config support vs Qwen3-TTS support are unrelated model additions." }, { "left": "issue:37015", "right": "issue:43837", "accept": false, - "reason": "NeoBERT and Qwen3-ASR are different model additions with different code paths." + "reason": "NeoBERT and Qwen3-ASR are different models and different code paths." }, { "left": "issue:37015", "right": "issue:44121", "accept": false, - "reason": "NeoBERT vs OpenAI weight-sparse transformer are distinct architectures and requests." + "reason": "NeoBERT vs a weight-sparse transformer request; not the same underlying feature." }, { "left": "issue:37015", "right": "issue:43909", "accept": false, - "reason": "NeoBERT and LFM2.5 Audio 1.5B are not the same model or feature." + "reason": "NeoBERT and LFM2.5 Audio are distinct model families/modalities." }, { "left": "issue:37015", "right": "issue:42503", "accept": false, - "reason": "NeoBERT and ModernVBERT are separate model requests; related theme, not duplicate." + "reason": "NeoBERT vs ModernVBERT are separate model requests, not duplicates." }, { "left": "issue:37015", "right": "issue:45027", "accept": false, - "reason": "NeoBERT vs Voxtral TTS support are different models and different functionality." + "reason": "NeoBERT and Voxtral TTS support target different models and integration work." }, { "left": "issue:41211", "right": "issue:43671", "accept": false, - "reason": "DEIMv2 and Qwen3-TTS are unrelated model additions." + "reason": "DEIMv2 and Qwen3-TTS are unrelated additions." }, { "left": "issue:42503", "right": "issue:43837", "accept": false, - "reason": "ModernVBERT models and Qwen3-ASR support are distinct requests, not one concrete fix." + "reason": "ModernVBERT models vs Qwen3-ASR; different model families and tasks." }, { "left": "issue:42503", "right": "issue:43671", "accept": false, - "reason": "ModernVBERT vs Qwen3-TTS spans different model families and modality." + "reason": "ModernVBERT and Qwen3-TTS are not the same model/change." }, { "left": "issue:42503", "right": "issue:42738", "accept": false, - "reason": "Both mention BERT-like ideas, but ModernVBERT models and generic RoPE BERT-like support are not the same request." + "reason": "ModernVBERT support and BERT-like RoPE models are different requests." }, { "left": "issue:42503", "right": "issue:45027", "accept": false, - "reason": "ModernVBERT models and Voxtral TTS support are unrelated." + "reason": "ModernVBERT models vs Voxtral-4B-TTS are unrelated." }, { "left": "issue:42738", "right": "issue:45027", "accept": false, - "reason": "BERT-like RoPE models vs Voxtral TTS support are different classes of model changes." + "reason": "A generic BERT-like RoPE request is not the same as Voxtral TTS support." }, { "left": "issue:43671", "right": "issue:43837", "accept": false, - "reason": "Both are Qwen3 proposals, but TTS and ASR are different modalities and code paths." + "reason": "Qwen3-TTS and Qwen3-ASR share a family name but request different capabilities and code paths." }, { "left": "issue:43671", "right": "issue:45027", "accept": false, - "reason": "Qwen3-TTS and Voxtral TTS are different model integrations, not one duplicate request." + "reason": "Qwen3-TTS support and Voxtral TTS support are different model integrations." }, { "left": "issue:43671", "right": "issue:45177", "accept": false, - "reason": "Audio TTS support and DeepSeek-OCR2 are unrelated model requests." + "reason": "Qwen3-TTS and DeepSeek-OCR2 are unrelated modalities and changes." }, { "left": "issue:43671", "right": "issue:44121", "accept": false, - "reason": "Qwen3-TTS and weight-sparse transformer support are different changes." + "reason": "Qwen3-TTS support does not overlap with a sparse-transformer architecture request." }, { "left": "issue:43671", "right": "issue:43909", "accept": false, - "reason": "Qwen3-TTS and LFM2.5 Audio are different audio models, not the same bug/change." + "reason": "Qwen3-TTS and LFM2.5 Audio are separate audio-model requests." }, { "left": "issue:43837", "right": "issue:43909", "accept": false, - "reason": "Qwen3-ASR and LFM2.5 Audio are different speech-model requests." + "reason": "Qwen3-ASR support and LFM2.5 Audio are distinct model additions." }, { "left": "issue:43837", "right": "issue:45322", "accept": false, - "reason": "Qwen3-ASR support and EUPE encoder support are unrelated model additions." + "reason": "Qwen3-ASR and EUPE are different tasks and model families." }, { "left": "issue:43837", "right": "issue:44121", "accept": false, - "reason": "ASR support and weight-sparse transformer support are not the same code-path problem." + "reason": "Qwen3-ASR support and weight-sparse transformer support are not the same change." }, { "left": "issue:43909", "right": "issue:44205", "accept": false, - "reason": "LFM2.5 Audio and SAM3-LiteText are different model families and modalities." + "reason": "LFM2.5 Audio 1.5B and SAM3-LiteText are unrelated model requests." }, { "left": "issue:44121", "right": "issue:45322", "accept": false, - "reason": "Weight-sparse transformer support and EUPE encoder support are separate architecture requests." + "reason": "OpenAI weight-sparse transformer vs EUPE are different architectures and requests." }, { "left": "issue:43837", "right": "issue:45027", "accept": false, - "reason": "Qwen3-ASR and Voxtral-4B-TTS-2603 are different speech-model integrations." + "reason": "Qwen3-ASR and Voxtral-4B-TTS are different models and integration paths." }, { "left": "issue:43837", "right": "issue:45177", "accept": false, - "reason": "ASR support and DeepSeek-OCR2 are unrelated model requests." + "reason": "Qwen3-ASR and DeepSeek-OCR2 are unrelated tasks." } ] }, @@ -14109,12 +14145,12 @@ "issue:43668|issue:44031", "issue:43668|issue:43813" ], - "prepared_review_unit_hash": "d01e8031a2f96939f16383ae9fec1210926b19760a394a383cfb608fb6d94ffe", + "prepared_review_unit_hash": "166e51d89266207bc6ed9a94480f8d5239393cd6a85cb673806ac2bddf3dc779", "component_budget": { "node_count": 6, "item_count": 6, "soft_pair_count": 8, - "serialized_chars": 6291, + "serialized_chars": 6292, "estimated_input_tokens": 1573, "estimated_eval_tokens": 3402 }, @@ -14122,7 +14158,7 @@ "node_count": 6, "item_count": 6, "soft_pair_count": 8, - "serialized_chars": 6291, + "serialized_chars": 6292, "estimated_input_tokens": 1573, "estimated_eval_tokens": 3402 }, @@ -14131,24 +14167,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These issues are only superficially related (mostly around configs, norms, losses, or tokenizer warnings), but they describe different bugs/questions in different code paths. I would not merge any of the soft-edge pairs as duplicates.", - "confidence": 0.93, - "canonical_issue_reason": null, + "summary": "These issues are only superficially similar by theme (norms, kwargs, typos, warnings) but do not describe the same underlying bug or change. I would not deduplicate any of the soft pairs.", + "confidence": 0.89, + "canonical_issue_reason": "If one representative is needed, issue:43240 is the most concrete and broadly actionable runtime bug report; however, the set does not contain a true duplicate cluster.", "canonical_pr_reason": null, - "best_issue_reason": "None of these is a clear canonical duplicate target; if one must be chosen as the most concrete user-facing bug report, issue 44031 stands out because it describes a specific regression affecting all tokenizers after a version change.", + "best_issue_reason": "issue:43240 is the strongest global issue candidate because it reports an actual behavior bug with broader applicability than the more narrow typo/type-hint/model-design reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:37934", "right": "issue:43668", "accept": false, - "reason": "One asks whether Llama4TextL2Norm should actually be RMSNorm; the other reports an incorrect type hint for ModernBERTConfig.norm_eps. Different models and different defect types." + "reason": "Both mention normalization, but one questions model design (RMS norm vs Llama4TextL2Norm) while the other is a config type-hint typo; not the same bug." }, { "left": "issue:42936", "right": "issue:43240", "accept": false, - "reason": "Mask2Former ignore_value not being used is a model-specific loss-parameter bug; the other is a broader kwargs-forwarding issue in loss calculation. Related area, but not the same bug." + "reason": "Both involve loss-related arguments, but 42936 is a specific Mask2Former ignore_value bug and 43240 is a general kwargs passthrough issue; close, but not clearly the same underlying defect." }, { "left": "issue:42936", @@ -14160,31 +14196,31 @@ "left": "issue:42936", "right": "issue:43813", "accept": false, - "reason": "Mask2Former ignore_value behavior and a typo in a quantization attribute name are completely different issues." + "reason": "Loss ignore_value handling and a quantization-operation typo are unrelated." }, { "left": "issue:42936", "right": "issue:43668", "accept": false, - "reason": "Different model/components and different defect classes: unused ignore_value vs an incorrect type hint." + "reason": "Model loss behavior vs config type-hint correctness; no shared bug." }, { "left": "issue:43240", "right": "issue:43668", "accept": false, - "reason": "A generic kwargs-loss-forwarding bug is not the same as a ModernBERTConfig type annotation error." + "reason": "General loss kwargs forwarding and a ModernBERTConfig type-hint typo are different issues." }, { "left": "issue:43668", "right": "issue:44031", "accept": false, - "reason": "ModernBERTConfig norm_eps typing and tokenizer regex warnings affect different subsystems with different root causes." + "reason": "Config type-hint bug and tokenizer warning regression are unrelated." }, { "left": "issue:43668", "right": "issue:43813", "accept": false, - "reason": "An incorrect type hint and a typo in a quantization operation name are unrelated." + "reason": "A type-hint error and a naming typo in quantization code are different problems." } ] }, @@ -14214,12 +14250,12 @@ "soft_pairs": [ "issue:41084|issue:43037" ], - "prepared_review_unit_hash": "6b41eb3638e1b4e28bb5458b24ab4af0d1c99e63e0b298d439579ad33d321e1b", + "prepared_review_unit_hash": "b8b1e5db0ffc5dba1f4dd070dd307c1ec1609f27bee2acef0859f5f09c91834d", "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 }, @@ -14227,7 +14263,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1479, + "serialized_chars": 1480, "estimated_input_tokens": 370, "estimated_eval_tokens": 996 }, @@ -14236,18 +14272,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues are not duplicates: one is a generic request about block decoding, while the other is a model-specific request for DeepSeek v3.2 support.", + "summary": "The two issues appear unrelated: one is about Set Block Decoding, while the other requests DeepSeek v3.2 support. No shared targets, filenames, or concrete overlap suggests duplication.", "confidence": 0.96, - "canonical_issue_reason": null, + "canonical_issue_reason": "Neither issue is a good canonical choice for the other because they describe different feature areas and different underlying changes.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "Neither issue is suitable as a global canonical issue; the topics do not align closely enough to represent the same bug or request.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41084", "right": "issue:43037", "accept": false, - "reason": "Different underlying topics: decoding behavior vs. support for a specific model/version. No indication they refer to the same bug or change." + "reason": "These are different requests with no concrete overlap: one concerns block decoding behavior, the other requests model support for DeepSeek v3.2." } ] }, @@ -14277,7 +14313,7 @@ "soft_pairs": [ "issue:43598|issue:43600" ], - "prepared_review_unit_hash": "dc9e5582b81ded1c85adc936663f831ea11c5a88f8e424f5f149599cf7b596e7", + "prepared_review_unit_hash": "696262ffa69273f8c062a80b6a7230eb522a782c32281f346fa0061246b45d08", "component_budget": { "node_count": 2, "item_count": 2, @@ -14299,8 +14335,8 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues are superficially related through Trainer/custom trainer behavior, but they describe different underlying problems: one is about `remove_unused_column` customizability, the other about FLOPS estimation for custom trainers. They should not be treated as duplicates.", - "confidence": 0.92, + "summary": "These two issues are thematically adjacent but not duplicates: one is about making `Trainer.remove_unused_column` more customizable, while the other is about FLOPS estimation handling custom trainers.", + "confidence": 0.95, "canonical_issue_reason": null, "canonical_pr_reason": null, "best_issue_reason": null, @@ -14310,7 +14346,7 @@ "left": "issue:43598", "right": "issue:43600", "accept": false, - "reason": "Different bug reports with different goals: one asks to revisit `remove_unused_column` handling, the other asks to account for custom trainers in FLOPS estimation. Shared Trainer context is too broad to merge them as the same issue." + "reason": "Different underlying problems and code paths: custom column dropping in training data handling vs FLOPS counting logic for custom trainers. Shared mention of custom trainers is too broad to treat as the same issue." } ] }, @@ -14340,12 +14376,12 @@ "soft_pairs": [ "issue:43684|issue:44405" ], - "prepared_review_unit_hash": "6ab6852a2579e0cdbdbd428acca630119553107b42cbaf7989979c93280a98fc", + "prepared_review_unit_hash": "12827ebdeff7fb3990b5867466140042aadb1132c77a175886c44487e3a6c3a3", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1611, + "serialized_chars": 1612, "estimated_input_tokens": 403, "estimated_eval_tokens": 1062 }, @@ -14353,7 +14389,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1611, + "serialized_chars": 1612, "estimated_input_tokens": 403, "estimated_eval_tokens": 1062 }, @@ -14362,18 +14398,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related only by the Qwen naming family, but they request different capabilities: one adds Qwen3-Omni registration for AutoModel/AutoModelForConditionalGeneration, the other asks for AutoModelForSequenceClassification support for Qwen3.5. They do not look like the same underlying change.", - "confidence": 0.97, - "canonical_issue_reason": "No canonical issue: the issues target different model variants and different auto classes, so they are separate feature requests.", + "summary": "These are related model-registration requests for Qwen variants, but they target different AutoModel classes and different concrete capabilities, so they should not be treated as duplicates.", + "confidence": 0.88, + "canonical_issue_reason": "Issue 43684 is the broader request: add Qwen3-Omni registration for AutoModel and AutoModelForConditionalGeneration, which is more likely to subsume a library registration gap.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43684 is the broader model-registration request and is the closest representative of the cluster, but it is not a duplicate of 44405.", + "best_issue_reason": "Issue 43684 is the best representative because it is the more general registration task and aligns with the broader API surface; 44405 is a narrower sequence-classification-specific support request for a different config.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43684", "right": "issue:44405", "accept": false, - "reason": "Different Qwen variants and different target heads (conditional generation vs sequence classification); not the same underlying bug/change." + "reason": "Both are about Qwen model support, but one is for Qwen3-Omni registration in AutoModel/AutoModelForConditionalGeneration and the other is for AutoModelForSequenceClassification support in Qwen3.5. They affect different code paths and are not the same underlying change." } ] }, @@ -14410,77 +14446,80 @@ "issue:43845|issue:43859", "issue:43846|issue:43859" ], - "prepared_review_unit_hash": "98aee86260290997919fd829372c8126987b9bca726e983070a6802d80bf190e", + "prepared_review_unit_hash": "b3d98e982a19f3c41e59fdb39650bdc8ea6c396a125ee9085dbbbc22f7185b22", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4265, - "estimated_input_tokens": 1067, - "estimated_eval_tokens": 2390 + "serialized_chars": 4263, + "estimated_input_tokens": 1066, + "estimated_eval_tokens": 2388 }, "budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4265, - "estimated_input_tokens": 1067, - "estimated_eval_tokens": 2390 + "serialized_chars": 4263, + "estimated_input_tokens": 1066, + "estimated_eval_tokens": 2388 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Three issues share the same very vague title (\u201chuggingface \u201d), but there isn\u2019t enough concrete content to confirm they are the same underlying bug. The lone outlier (\u201cFett\u201d) looks unrelated. No PRs are present.", - "confidence": 0.78, - "canonical_issue_reason": "issue:43845 is the best canonical issue by default: it is one of the duplicate-looking \u201chuggingface \u201d reports, is the earliest among them, and has a fuller body than 43859.", + "summary": "Three issues share the exact same generic title \"huggingface \" and are very likely duplicates; the separate \"Fett\" issue looks unrelated. I\u2019d keep the earliest huggingface report as the canonical issue.", + "confidence": 0.8, + "canonical_issue_reason": "Issue 43845 is the earliest of the three highly similar \"huggingface \" reports, so it is the best canonical representative for that duplicate set.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43845 is the best cluster representative because it is the earliest of the near-identical huggingface-titled issues and appears at least as informative as the others.", + "best_issue_reason": "Issue 43845 is the best overall issue candidate because it is the earliest and most likely to capture the shared huggingface report cluster; 43835 has a different title and appears unrelated.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43835", "right": "issue:43845", "accept": false, - "reason": "Different titles (\u201cFett\u201d vs \u201chuggingface \u201d) and no shared targets/files or other concrete evidence of the same bug." + "reason": "Different titles and no concrete shared details beyond weak similarity; likely unrelated." }, { "left": "issue:43835", "right": "issue:43846", "accept": false, - "reason": "Different titles and no substantive body/detail evidence tying them to the same underlying issue." + "reason": "Different titles and no evidence they describe the same underlying problem." }, { "left": "issue:43835", "right": "issue:43859", "accept": false, - "reason": "Different titles and no concrete overlap beyond generic soft similarity." + "reason": "Different titles and no concrete overlap indicating the same issue." }, { "left": "issue:43845", "right": "issue:43846", - "accept": false, - "reason": "Titles match, but the available metadata is too thin to prove they describe the same specific bug rather than two separate vague reports." + "accept": true, + "reason": "Same exact title and very close timestamps/body size; these look like duplicate reports of the same huggingface issue." }, { "left": "issue:43845", "right": "issue:43859", - "accept": false, - "reason": "Same broad title, but no explicit issue target, filename, or detailed description to confirm a shared underlying problem." + "accept": true, + "reason": "Same exact title and strong similarity; likely duplicate reports of the same underlying issue." }, { "left": "issue:43846", "right": "issue:43859", - "accept": false, - "reason": "Same broad title, but the cluster evidence is insufficient to safely merge them as duplicates of one concrete issue." + "accept": true, + "reason": "Same exact title and strong similarity; likely duplicate reports of the same underlying issue." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: the three issues with identical \"huggingface \" titles do form a plausible duplicate cluster, and 43835 has only weak similarity plus a different title. The canonical choice of the earliest huggingface report is conservative, and the soft-edge accepts are supported by the repeated title and high soft-similarity scores without overstating certainty." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -14514,88 +14553,88 @@ "issue:44572|issue:44937", "issue:44573|issue:44937" ], - "prepared_review_unit_hash": "6d883e2697d2ef0dd0c36fd597ac8d3c143ba2fb4c2368c8bbb87e57341f76c3", + "prepared_review_unit_hash": "3e81cee2e8719bb9a008dcd53f3041fc16b22bbc28d14f5670a6703f90c71ca8", "component_budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 9, - "serialized_chars": 5794, - "estimated_input_tokens": 1449, - "estimated_eval_tokens": 3154 + "serialized_chars": 5788, + "estimated_input_tokens": 1447, + "estimated_eval_tokens": 3150 }, "budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 9, - "serialized_chars": 5794, - "estimated_input_tokens": 1449, - "estimated_eval_tokens": 3154 + "serialized_chars": 5788, + "estimated_input_tokens": 1447, + "estimated_eval_tokens": 3150 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is mostly low-information/spam issues with no clear shared bug or feature request; they do not look like duplicates of one another.", - "confidence": 0.97, - "canonical_issue_reason": "Issue #44371 is a reasonable anchor only because it is one of the earliest low-information spam tickets in the set; there is no substantive duplicate signal tying the issues together.", + "summary": "These are not duplicates of one substantive issue; the cluster looks like mostly spam/placeholder issues with one vague documentation suggestion, so there is no clear canonical bug report here.", + "confidence": 0.98, + "canonical_issue_reason": "No issue has enough concrete problem description to serve as a true canonical report.", "canonical_pr_reason": null, - "best_issue_reason": "No issue here is a strong canonical duplicate target; #44371 is the least-bad representative by chronology, but the cluster does not contain a meaningful underlying issue.", + "best_issue_reason": "Issue 44418 is the least spam-like and the only one with a somewhat descriptive title, but it still lacks actionable detail.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44371", "right": "issue:44572", "accept": false, - "reason": "Both are spam-like placeholders, but there is no evidence they describe the same underlying bug or change." + "reason": "Both look like spam placeholders and provide no evidence of the same underlying problem." }, { "left": "issue:44371", "right": "issue:44573", "accept": false, - "reason": "Same as above: identical low-quality titles are not enough to treat them as duplicates of a concrete issue." + "reason": "Both look like spam placeholders and provide no evidence of the same underlying problem." }, { "left": "issue:44371", "right": "issue:44418", "accept": false, - "reason": "One is spam-like and the other is a generic documentation suggestion; they are not the same issue." + "reason": "One is spam-like and the other is a generic documentation suggestion; not the same concrete issue." }, { "left": "issue:44371", "right": "issue:44937", "accept": false, - "reason": "The titles are unrelated; one is spam and the other is a strange named-item suggestion, not a duplicate bug report." + "reason": "Spam placeholder versus unrelated title about Google LLC; no shared bug/change." }, { "left": "issue:44418", "right": "issue:44572", "accept": false, - "reason": "A documentation enhancement request and a spam placeholder do not indicate the same underlying problem." + "reason": "Generic documentation suggestion versus spam placeholder; not a duplicate underlying report." }, { "left": "issue:44418", "right": "issue:44573", "accept": false, - "reason": "Generic documentation request versus spam placeholder; no shared concrete issue." + "reason": "Generic documentation suggestion versus spam placeholder; not a duplicate underlying report." }, { "left": "issue:44572", "right": "issue:44573", "accept": false, - "reason": "Both are spam-like placeholders, but there is no concrete shared bug/change to merge them on." + "reason": "Both are spam-like placeholders, but there is no concrete issue content tying them to the same underlying bug." }, { "left": "issue:44572", "right": "issue:44937", "accept": false, - "reason": "Unrelated titles and no shared issue target or context." + "reason": "Spam placeholder versus unrelated Google LLC title; no evidence they describe the same issue." }, { "left": "issue:44573", "right": "issue:44937", "accept": false, - "reason": "Unrelated titles and no shared issue target or context." + "reason": "Spam placeholder versus unrelated Google LLC title; no evidence they describe the same issue." } ] }, @@ -14625,46 +14664,46 @@ "soft_pairs": [ "issue:45145|issue:45151" ], - "prepared_review_unit_hash": "a39b788154ed82dba4a68484e50f7f6fb4df5a6478ec99a494a29eb832b866a7", + "prepared_review_unit_hash": "01374fd53b5e155cf20472def2ded764b9ffb42d9e1c1f6a91f5e742d8da3abe", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1645, - "estimated_input_tokens": 412, - "estimated_eval_tokens": 1080 + "serialized_chars": 1644, + "estimated_input_tokens": 411, + "estimated_eval_tokens": 1078 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1645, - "estimated_input_tokens": 412, - "estimated_eval_tokens": 1080 + "serialized_chars": 1644, + "estimated_input_tokens": 411, + "estimated_eval_tokens": 1078 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two closed issues are highly similar and likely refer to the same proposed energy-optimization idea, but the evidence is limited to title/body similarity with no explicit linkage, so the relationship should be treated as unresolved rather than a firm duplicate.", - "confidence": 0.87, - "canonical_issue_reason": "#45145 is the better canonical issue because it has the same title and slightly stronger linkage signal (one inbound reference), making it the more established thread.", + "summary": "The issues are highly similar and share the exact title, but the available evidence is still soft rather than definitive. I would keep 45145 as the canonical issue if these are later confirmed duplicates, while treating the pair as likely duplicates rather than a certain duplicate match.", + "confidence": 0.83, + "canonical_issue_reason": "Issue 45145 is earlier and already has an inbound reference, so it is the best canonical representative among these highly similar reports.", "canonical_pr_reason": null, - "best_issue_reason": "#45145 is the best global issue representative for the same reason: it appears to be the more established of two near-identical issue threads.", + "best_issue_reason": "45145 is the most stable thread to keep because it predates 45151 and has incoming linkage, making it the better anchor for consolidation.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45145", "right": "issue:45151", "accept": false, - "reason": "The issues have identical titles and similar bodies, but the packet only provides soft_similarity evidence, no explicit cross-links, and deterministic_accept is false. That supports a likely duplicate, not a confident duplicate classification, so this edge should remain unresolved or be treated as low-confidence duplicate-likely." + "reason": "The titles are identical and the similarity signal is strong, but the packet does not provide enough body-level evidence to confidently say they are the same underlying report; this is better treated as likely duplicates / highly similar issues rather than a confirmed duplicate." } ] }, "evaluator_result": { "accept": false, - "feedback": "The canonical issue choice is reasonably grounded, but the soft-edge verdict is too strong for the evidence in the packet. The pair only has soft_similarity support, no explicit linkage, and deterministic_accept is false, so calling them a \"clear duplicate\" / \"almost certainly duplicates\" overstates confidence. Please downgrade to a more conservative duplicate-likelihood assessment or leave the edge unresolved." + "feedback": "The canonical choice is grounded, but the duplicate verdict is a bit too strong for the evidence available. The packet only shows identical titles plus a soft-similarity signal (with deterministic_accept=false), not enough to call the bodies 'identical' or the reports unquestionably the same. A more conservative phrasing would be that they are likely duplicates / highly similar reports, with 45145 as the better canonical issue because it is earlier and has an inbound reference." }, "evaluator_used": true, "retried": true, @@ -14778,172 +14817,178 @@ "pull_request:43823|pull_request:45489", "pull_request:43251|pull_request:43254" ], - "prepared_review_unit_hash": "483dbb8c46708baf8cb704092db786ffbce9f3fca891548ec90f9a7ef6dc3f5f", + "prepared_review_unit_hash": "2e9cd6aee97978de2cd5db127181f4b6a14de0146f227e8140a12191c26d06da", "component_budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 23, - "serialized_chars": 37543, - "estimated_input_tokens": 9386, - "estimated_eval_tokens": 19028 + "serialized_chars": 37419, + "estimated_input_tokens": 9355, + "estimated_eval_tokens": 18966 }, "budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 23, - "serialized_chars": 37543, - "estimated_input_tokens": 9386, - "estimated_eval_tokens": 19028 + "serialized_chars": 37419, + "estimated_input_tokens": 9355, + "estimated_eval_tokens": 18966 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The cluster is mostly a mix of unrelated PRs spanning model additions, docs, tests, and refactors. The only clear duplicate pair is the two `fixed_cross_entropy` PRs targeting the same issue and code path; the rest should stay separate.", - "confidence": 0.97, + "summary": "This is not a single duplicate cluster: most PRs are unrelated model additions, refactors, or test/docs changes. The only clear duplicate pair is the two fixed_cross_entropy kwargs fixes (43251 and 43254).", + "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43251 is the best canonical choice for the `fixed_cross_entropy` fix: it explicitly targets issue 43240, updates the same loss helper as 43254, and has the more complete/intentional implementation and discussion. The other items are unrelated to each other.", + "canonical_pr_reason": "PR 43251 is the best canonical representative of the only true duplicate pair because it explicitly targets issue 43240 and has the more complete patch for forwarding cross-entropy kwargs.", "best_issue_reason": null, - "best_pr_reason": "PR 43251 is the strongest overall PR in the set because it is a focused, well-scoped fix with an explicit issue link and clear behavior change. Most other PRs are unrelated feature or refactor work, or test/doc-only changes.", + "best_pr_reason": "43251 is the best overall PR to anchor this cluster only because it is the clearest, issue-linked implementation among the near-duplicates; the rest are not duplicates of it.", "soft_edge_verdicts": [ { "left": "pull_request:39895", "right": "pull_request:43424", "accept": false, - "reason": "One adds VideoPrism support; the other adds an Executorch dynamic-shape test. Different models and different code paths." + "reason": "VideoPrism model addition vs ExecuTorch dynamic-shape export test; different code paths and different goals." }, { "left": "pull_request:42668", "right": "pull_request:43823", "accept": false, - "reason": "Processor robustness changes for existing models versus adding a new MobileLLM model. Not the same bug or change." + "reason": "Processor-pretrained robustness fixes vs adding MobileLLM model support; unrelated changes." }, { "left": "pull_request:43251", "right": "pull_request:43823", "accept": false, - "reason": "Loss helper kwargs fix versus MobileLLM model creation. Unrelated." + "reason": "cross_entropy kwargs fix vs MobileLLM model addition; no shared underlying bug/change." }, { "left": "pull_request:43251", "right": "pull_request:44827", "accept": false, - "reason": "`fixed_cross_entropy` kwargs support is unrelated to Mistral4 test/model refactoring." + "reason": "loss utility kwargs fix vs Mistral4 test cleanup; not the same issue." }, { "left": "pull_request:43424", "right": "pull_request:44827", "accept": false, - "reason": "Executorch dynamic-shape export tests and Mistral4 fixes are different features in different subsystems." + "reason": "ExecuTorch dynamic-shape export test vs Mistral4 test fix; both are tests but for different behaviors." }, { "left": "pull_request:43424", "right": "pull_request:45435", "accept": false, - "reason": "Dynamic-shape export testing is unrelated to Whisper tokenizer bounds checking." + "reason": "ExecuTorch exportability test vs Whisper tokenizer special-token indexing bug; unrelated." }, { "left": "pull_request:43424", "right": "pull_request:43823", "accept": false, - "reason": "Different model/export areas with no shared underlying bug." + "reason": "ExecuTorch export test vs MobileLLM addition; different features." }, { "left": "pull_request:43823", "right": "pull_request:45454", "accept": false, - "reason": "MobileLLM model addition versus Gemma4 training behavior fix; no shared code-path problem." + "reason": "MobileLLM model addition vs Gemma4 training-mask change; not the same change." }, { "left": "pull_request:43823", "right": "pull_request:45435", "accept": false, - "reason": "MobileLLM model creation and Whisper tokenizer indexing are unrelated." + "reason": "MobileLLM model addition vs Whisper tokenizer bugfix; unrelated." }, { "left": "pull_request:43823", "right": "pull_request:43995", "accept": false, - "reason": "New model addition versus Falcon output-interface refactor; different changes." + "reason": "MobileLLM addition vs Falcon output-collection refactor; different models and code paths." }, { "left": "pull_request:43995", "right": "pull_request:45435", "accept": false, - "reason": "Falcon output-capture refactor is not the same issue as Whisper tokenization bounds handling." + "reason": "Falcon refactor vs Whisper tokenization bugfix; no overlap beyond being library code." }, { "left": "pull_request:43995", "right": "pull_request:45213", "accept": false, - "reason": "Falcon refactor and model-creation skill content are unrelated." + "reason": "Falcon model refactor vs repository skill/docs PR; clearly unrelated." }, { "left": "pull_request:43995", "right": "pull_request:44827", "accept": false, - "reason": "Falcon output-interface refactor and Mistral4 test fixes are different code paths." + "reason": "Falcon output interface refactor vs Mistral4 tests; different models and problems." }, { "left": "pull_request:44827", "right": "pull_request:45489", "accept": false, - "reason": "Mistral4 fixes and Gemma3n/Gemma4 cache-sharing alignment are different model families and bugs." + "reason": "Mistral4 test fixes vs Gemma3n/Gemma4 cache-sharing alignment; distinct underlying changes." }, { "left": "pull_request:44827", "right": "pull_request:45170", "accept": false, - "reason": "Mistral4 test fixes do not match the CLIP-like `layrnorm` rename/correction PR." + "reason": "Mistral4 test fixes vs layrnorm/layernorm rename fix; unrelated." }, { "left": "pull_request:43823", "right": "pull_request:44827", "accept": false, - "reason": "MobileLLM model addition and Mistral4 fixes are unrelated." + "reason": "MobileLLM model addition vs Mistral4 test fixes; different scope and code paths." }, { "left": "pull_request:45213", "right": "pull_request:45435", "accept": false, - "reason": "A new-model skill PR and a Whisper tokenizer bugfix are not the same underlying change." + "reason": "model-creation skill/docs PR vs Whisper tokenizer bugfix; not comparable as the same change." }, { "left": "pull_request:44827", "right": "pull_request:45213", "accept": false, - "reason": "Mistral4 bugfixes and a model-creation skill are unrelated." + "reason": "Mistral4 tests vs model-creation skill; unrelated." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45213", + "accept": false, + "reason": "MobileLLM model addition vs skill file for creating models; not the same artifact or fix." }, { "left": "pull_request:44827", "right": "pull_request:45435", "accept": false, - "reason": "Mistral4 test fixes and Whisper tokenizer indexing fixes are unrelated." + "reason": "Mistral4 tests vs Whisper tokenizer fix; different models and behavior." }, { "left": "pull_request:45454", "right": "pull_request:45489", "accept": false, - "reason": "Both touch Gemma4-related files, but one is about training text-only samples and the other is about cache-sharing alignment; they fix different concrete problems." + "reason": "Both touch Gemma4-adjacent files, but one is text-only training mask behavior and the other is cache-sharing alignment; they are not the same concrete fix." }, { "left": "pull_request:43823", "right": "pull_request:45489", "accept": false, - "reason": "MobileLLM addition and Gemma3n/Gemma4 cache-sharing alignment are unrelated." + "reason": "MobileLLM addition vs Gemma3n/Gemma4 cache-sharing refactor; unrelated." }, { "left": "pull_request:43251", "right": "pull_request:43254", "accept": true, - "reason": "Both PRs fix the same `fixed_cross_entropy` helper to support additional kwargs/weight/label_smoothing for issue 43240, and they modify the same function on the same code path." + "reason": "Both PRs fix the same fixed_cross_entropy/nn.functional.cross_entropy kwargs handling in the same file, with the same issue target 43240; they are duplicate implementations of the same bugfix." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. The only clearly mergeable duplicate pair is 43251/43254, which both target the same `fixed_cross_entropy` helper and the same issue. The other pair rejections are consistent with the listed titles/files and do not overclaim shared bugs or mergeability." + "feedback": "The summary is grounded in the packet: most PRs are clearly unrelated model additions, refactors, or test/docs changes, and the only defensible duplicate pair is 43251/43254, which both modify src/transformers/loss/loss_utils.py to address the shared 43240 cross-entropy kwargs issue. The soft-edge verdicts are conservative and do not overstate mergeability." }, "evaluator_used": true, "retried": false, @@ -14970,12 +15015,12 @@ "soft_pairs": [ "pull_request:40546|pull_request:42781" ], - "prepared_review_unit_hash": "7d5f44780d748776aece7817aedfcd212a3ef371ed53752c634e79f5fc8bc80b", + "prepared_review_unit_hash": "1a56cb953af464b15869e54803db9a9d3a87bad3bcb459220f3113802dc75940", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 7225, + "serialized_chars": 7226, "estimated_input_tokens": 1807, "estimated_eval_tokens": 3870 }, @@ -14983,7 +15028,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 7225, + "serialized_chars": 7226, "estimated_input_tokens": 1807, "estimated_eval_tokens": 3870 }, @@ -14992,18 +15037,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs are about VibeVoice, but they are not the same change: one is the base model integration and the other is a later realtime extension with additional tokenizer/generation work. They overlap heavily in shared model scaffolding, but the concrete goals differ enough that they should not be merged as duplicates.", + "summary": "Both PRs are about VibeVoice, but they target different scopes: #40546 is the base VibeVoice integration, while #42781 is a later realtime-oriented extension with additional tokenizer/feature-extraction work. They are related, not duplicates.", "confidence": 0.84, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #42781 is the broader, newer VibeVoice-related PR and appears to extend the base implementation with realtime support and the acoustic tokenizer, so it is the better representative of the cluster.", + "canonical_pr_reason": "PR #40546 is the original, non-draft VibeVoice implementation with substantial discussion and core integration work; it is the better canonical representative of the base project change.", "best_issue_reason": null, - "best_pr_reason": "PR #42781 is the most complete and current VibeVoice PR in the set, covering the broader feature set beyond the initial implementation.", + "best_pr_reason": "PR #40546 is the most suitable representative because it covers the foundational VibeVoice addition and is more mature than the draft realtime follow-up.", "soft_edge_verdicts": [ { "left": "pull_request:40546", "right": "pull_request:42781", "accept": false, - "reason": "40546 implements core VibeVoice support, while 42781 adds VibeVoice Realtime plus acoustic tokenizer and generation changes. They share the same model family and some files, but they are different concrete changes and not safely mergeable as one PR." + "reason": "Shared VibeVoice files indicate the same model family, but the diffs suggest different concrete changes: #40546 adds the initial VibeVoice model, while #42781 adds realtime support plus acoustic tokenizer/feature-extraction changes. These could not safely be merged as one duplicate PR." } ] }, @@ -15033,12 +15078,12 @@ "soft_pairs": [ "pull_request:40755|pull_request:41224" ], - "prepared_review_unit_hash": "ac8c1b7a26caaa5d1d6cb181c5fd7df7885f8086155ce24b695016620f15f8d3", + "prepared_review_unit_hash": "97cc9e31c8d6088c37a1e0371488fd633cc9fdeed958fdbdbabb9ab0e6604277", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4675, + "serialized_chars": 4674, "estimated_input_tokens": 1169, "estimated_eval_tokens": 2594 }, @@ -15046,7 +15091,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4675, + "serialized_chars": 4674, "estimated_input_tokens": 1169, "estimated_eval_tokens": 2594 }, @@ -15055,18 +15100,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two pull requests are unrelated: one adds TimesFM covariate-based forecasting, while the other adds DINOv3 image-classification support. They do not appear to fix the same bug or implement the same change.", + "summary": "These are unrelated pull requests in different model families: TimesFM covariate forecasting vs. DINOv3 image classification support. They should not be treated as duplicates.", "confidence": 0.98, - "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: the PRs target different model families, different code paths, and different features.", + "canonical_issue_reason": "No canonical issue: there are no issues in the cluster, and the PRs do not describe the same bug or change.", + "canonical_pr_reason": "No canonical PR for duplication purposes: the PRs target different code paths, different models, and different feature additions.", "best_issue_reason": null, - "best_pr_reason": "PR 41224 is the better standalone representative because it is a narrower, self-contained feature addition with a single clear model capability.", + "best_pr_reason": "If a representative PR is needed, 41224 is the more self-contained feature addition, but it is not a duplicate of 40755.", "soft_edge_verdicts": [ { "left": "pull_request:40755", "right": "pull_request:41224", "accept": false, - "reason": "Different models and different functionality: TimesFM covariate forecasting versus DINOv3 image classification. They are not the same underlying change and should not be merged as duplicates." + "reason": "Different features in different subsystems: TimesFM forecasting with covariates vs. DINOv3 image classification support. They do not fix the same underlying problem and could not plausibly be merged into one PR." } ] }, @@ -15162,12 +15207,12 @@ "soft_pairs": [ "pull_request:41797|pull_request:45186" ], - "prepared_review_unit_hash": "a03af63a220b332b5d92c9a22f4904217612bf2f7168cc89b8c71fd7603b6c74", + "prepared_review_unit_hash": "852241ff3ed83c7a470cea6912b56d1287e74c29cfafabbfe03422cfd3f1d5ca", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6722, + "serialized_chars": 6723, "estimated_input_tokens": 1681, "estimated_eval_tokens": 3618 }, @@ -15175,7 +15220,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6722, + "serialized_chars": 6723, "estimated_input_tokens": 1681, "estimated_eval_tokens": 3618 }, @@ -15184,18 +15229,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two separate model-addition PRs (DeepSeek OCR vs Isaac) that share only generic registration/docs scaffolding, not the same underlying change.", - "confidence": 0.98, + "summary": "These are two separate model-integration PRs (DeepSeek OCR vs Isaac) that only overlap on shared transformer scaffold files, not the underlying model or code path. They should not be clustered as duplicates.", + "confidence": 0.99, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: the diffs implement different models with different configs, processors, tests, and docs, so they are not duplicates.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "Neither PR is a global duplicate representative; each is a distinct model implementation and should be triaged independently.", + "best_pr_reason": "Neither PR is a duplicate of the other; if one must be chosen as a representative, 45186 is a more complete standalone model addition, but it does not subsume 41797.", "soft_edge_verdicts": [ { "left": "pull_request:41797", "right": "pull_request:45186", "accept": false, - "reason": "Different model families and code paths (DeepSeek OCR vs Isaac); overlap is limited to shared Transformers registration/doc plumbing, which is not enough to treat them as the same change." + "reason": "Different models and implementations; shared auto/registry/docs files are generic integration boilerplate, not evidence of the same fix or change." } ] }, @@ -15225,12 +15270,12 @@ "soft_pairs": [ "pull_request:42166|pull_request:45426" ], - "prepared_review_unit_hash": "4e67818c4df0b90c59cfcccedacdc9b6215e7976cc4b08f75f1d6af750cad2d5", + "prepared_review_unit_hash": "b91cdb431e7a1d26466351e225592ee2bbf3fecd30c682a15dfa7f0cb18a9649", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5408, + "serialized_chars": 5407, "estimated_input_tokens": 1352, "estimated_eval_tokens": 2960 }, @@ -15238,7 +15283,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5408, + "serialized_chars": 5407, "estimated_input_tokens": 1352, "estimated_eval_tokens": 2960 }, @@ -15247,18 +15292,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both items are pull requests adding different model integrations: InternVLFlash vs AXK1. They share only common registration/boilerplate files, not the same underlying change.", - "confidence": 0.98, + "summary": "The two pull requests are unrelated model-addition changes: one introduces InternVLFlash, the other introduces AXK1. They overlap only in shared auto-registration boilerplate, not in the underlying code change or bug fix.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: these are unrelated feature additions for different models, so they should not be clustered as duplicates.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "Neither PR is a good global representative of the other; if forced to pick a cluster anchor, PR #42166 has more review activity, but it is not a duplicate of #45426.", + "best_pr_reason": "PR 42166 is the more complete exemplar of a model-integration change because it includes docs and tests, but it is not a duplicate of PR 45426; the two PRs target different models and code paths.", "soft_edge_verdicts": [ { "left": "pull_request:42166", "right": "pull_request:45426", "accept": false, - "reason": "Reject: they add different models and different codepaths (InternVLFlash vs AXK1). Shared auto-registration files are generic boilerplate, not evidence of the same change." + "reason": "Different model integrations (InternVLFlash vs AXK1); shared registry files are incidental boilerplate, not the same change." } ] }, @@ -15295,7 +15340,7 @@ "pull_request:43291|pull_request:44815", "pull_request:43665|pull_request:44815" ], - "prepared_review_unit_hash": "128e96617927b4e100cff7e108ba48b6b7e4ecd9f70126e4ab65c49174dce506", + "prepared_review_unit_hash": "4a7cff3be0bd0d748c76a78b472c445ad062fbf6c5b621b2977013f18b41e2e0", "component_budget": { "node_count": 4, "item_count": 4, @@ -15317,48 +15362,48 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The four PRs are not duplicates of one another. Two are Whisper-related, but they fix different areas (ASR pipeline language return vs tokenizer/test behavior). The CLIP/ViT test PR and the FP8 dequant PR are unrelated to the others.", - "confidence": 0.95, + "summary": "The cluster is heterogeneous: one Whisper ASR pipeline fix, one Whisper tokenizer/test refactor, one obviously unrelated CLIP/ViT test-only change, and one FP8 dequant/model-loading PR. None are duplicate candidates.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 42227 is the cleanest representative: it has an explicit issue target, a focused code-path fix in the Whisper ASR pipeline, and matching test coverage.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR 42227 is the strongest standalone PR in the set because it is a concrete bug fix with a clear user-facing behavior change and direct issue linkage.", + "best_pr_reason": "44815 is the strongest standalone PR: it contains a coherent functional FP8 dequant/model-loading change, while the others are either narrow Whisper fixes or clearly unrelated test edits.", "soft_edge_verdicts": [ { "left": "pull_request:42227", "right": "pull_request:43291", "accept": false, - "reason": "Both are Whisper-related, but they fix different problems in different code paths: ASR pipeline return-language plumbing vs tokenizer decode/test behavior." + "reason": "Both are Whisper-related, but they fix different code paths: ASR pipeline return_language forwarding vs tokenizer decode/test behavior. Not the same underlying bug or mergeable into one PR." }, { "left": "pull_request:42227", "right": "pull_request:43665", "accept": false, - "reason": "Unrelated areas: Whisper ASR pipeline vs CLIP/ViT test-file changes. No shared underlying bug or mergeable combined fix." + "reason": "Unrelated changes: a real Whisper pipeline fix versus dummy assertions in CLIP/ViT tests. Not duplicate or same change." }, { "left": "pull_request:42227", "right": "pull_request:44815", "accept": false, - "reason": "Different subsystems and bugs: Whisper return-language handling vs FP8 dequant/loading logic. Not the same change." + "reason": "Different subsystems and bugs: Whisper ASR pipeline vs FP8 dequant/model-loading logic. No shared concrete code-path problem." }, { "left": "pull_request:43291", "right": "pull_request:43665", "accept": false, - "reason": "Whisper tokenizer/test work does not match the CLIP/ViT test edits; these are separate and not mergeable as one fix." + "reason": "Whisper tokenizer/test refactor is unrelated to CLIP/ViT test edits. Not the same underlying issue." }, { "left": "pull_request:43291", "right": "pull_request:44815", "accept": false, - "reason": "No common concrete bug or code path: Whisper decode behavior vs FP8 quantization/loading. Unrelated PRs." + "reason": "Different areas and fixes: Whisper tokenizer decode behavior vs FP8 dequant/model-loading. No basis to merge as duplicates." }, { "left": "pull_request:43665", "right": "pull_request:44815", "accept": false, - "reason": "CLIP/ViT test changes are unrelated to the FP8 dequant fix; they do not address the same issue." + "reason": "Completely unrelated: placeholder test failures in CLIP/ViT versus a functional FP8 dequant change." } ] }, @@ -15388,12 +15433,12 @@ "soft_pairs": [ "pull_request:42975|pull_request:42976" ], - "prepared_review_unit_hash": "458ead408e6ff51ed837d143544655806020a3b67970cdb8757b0856add250ec", + "prepared_review_unit_hash": "8b7a368b2e951532511d9429a03917d7d69f3dc0427fc7ed59c5f7c517c3f282", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5935, + "serialized_chars": 5934, "estimated_input_tokens": 1484, "estimated_eval_tokens": 3224 }, @@ -15401,7 +15446,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5935, + "serialized_chars": 5934, "estimated_input_tokens": 1484, "estimated_eval_tokens": 3224 }, @@ -15410,18 +15455,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two maintenance PRs for GitHub Actions upgrades, but they target different action families and change different workflows. They overlap in broad CI infrastructure, not in the same concrete bug or mergeable change set.", - "confidence": 0.89, + "summary": "These are related workflow-maintenance PRs, but they are not the same change: one targets Node 24 compatibility for GitHub Actions, the other is a broader \u201clatest versions\u201d bump, mainly for docker/build-push-action and a few workflow actions.", + "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 42975 is the better representative because it has the clearer, more specific goal (Node 24 compatibility) and covers a broader set of workflow action-version bumps.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR 42975 is the most suitable canonical PR for the cluster: its title and diff make the maintenance intent explicit, while the other PR is a separate dependency bump focused on docker/build-push-action and related workflow updates.", + "best_pr_reason": "PR 42975 is the better representative of the cluster because it has a specific compatibility goal and a more coherent scope; PR 42976 is a broader dependency refresh with different intent.", "soft_edge_verdicts": [ { "left": "pull_request:42975", "right": "pull_request:42976", "accept": false, - "reason": "Do not merge as duplicates: 42975 updates checkout/cache/upload-artifact versions for Node 24 compatibility, while 42976 bumps docker/build-push-action and other workflow actions to latest versions. They are related maintenance changes but not the same concrete change or code-path fix." + "reason": "Do not merge: they overlap in workflow files, but they are not fixing the same concrete problem. 42975 is about Node 24 compatibility for GitHub Actions, while 42976 is a general upgrade of action versions, especially docker/build-push-action. Shared subsystem alone is insufficient for duplicate treatment." } ] }, @@ -15472,118 +15517,118 @@ "pull_request:45283|pull_request:45532", "pull_request:45490|pull_request:45532" ], - "prepared_review_unit_hash": "3495eeacad70709a7e9b98b84524250dae72090dbbf87f752acf60563196a188", + "prepared_review_unit_hash": "2736eda2d1bb1be84d547aa98f8bebfcaef6da8ef7697bdf1003a12d3d14f5d8", "component_budget": { "node_count": 10, "item_count": 10, "soft_pair_count": 14, - "serialized_chars": 32085, - "estimated_input_tokens": 8022, - "estimated_eval_tokens": 16300 + "serialized_chars": 32133, + "estimated_input_tokens": 8034, + "estimated_eval_tokens": 16324 }, "budget": { "node_count": 10, "item_count": 10, "soft_pair_count": 14, - "serialized_chars": 32085, - "estimated_input_tokens": 8022, - "estimated_eval_tokens": 16300 + "serialized_chars": 32133, + "estimated_input_tokens": 8034, + "estimated_eval_tokens": 16324 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is mostly a set of unrelated model-addition and infrastructure PRs that share boilerplate docs/auto-registry edits. I found no true duplicate pairs worth merging.", + "summary": "This cluster is a mix of unrelated model-addition PRs and one distributed-loading infrastructure PR. The soft overlaps are mostly boilerplate in docs/auto registration files, not the same underlying bug or feature. No duplicate pair should be merged.", "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR fits the cluster: the substantive changes span different models and unrelated infra work, and the overlaps are mostly shared documentation/registry boilerplate.", + "canonical_pr_reason": "No single canonical PR: these PRs implement different models or features (ViT NEPA, Molmo, Molmo2, Qwen3-ASR, Lfm2Audio, PP loading, PrismML GGUF, Qwen3.5 GGUF, CTSM, SLANet) and are not duplicates.", "best_issue_reason": null, - "best_pr_reason": "45532 is the closest thing to a representative only because it has the strongest soft link and active discussion, but it still does not duplicate the rest of the cluster.", + "best_pr_reason": "PR 45532 is the strongest representative of the cluster\u2019s model-support pattern, with a substantial model implementation and the clearest end-to-end addition, but it still targets SLANet specifically and is not interchangeable with the others.", "soft_edge_verdicts": [ { "left": "pull_request:42978", "right": "pull_request:43448", "accept": false, - "reason": "Different model additions: ViT NEPA vs Molmo. Shared docs/auto-registry files are boilerplate, not the same change." + "reason": "Different model ports: ViT NEPA vs Molmo. Shared docs/auto-registration files are generic scaffolding, not the same change." }, { "left": "pull_request:42978", "right": "pull_request:43451", "accept": false, - "reason": "Different model additions: ViT NEPA vs Molmo2. No shared concrete code-path or bug fix." + "reason": "Different model ports: ViT NEPA vs Molmo2. Similar repository boilerplate, but separate architectures and codepaths." }, { "left": "pull_request:42978", "right": "pull_request:45490", "accept": false, - "reason": "Different model additions: ViT NEPA vs CTSM. Overlap is limited to generic docs/auto-registration edits." + "reason": "Different model ports: ViT NEPA vs CTSM. The overlap is only in shared registration/docs files." }, { "left": "pull_request:43448", "right": "pull_request:43451", "accept": false, - "reason": "Molmo and Molmo2 are distinct model implementations with different files and behavior; not the same underlying change." + "reason": "Molmo and Molmo2 are related by name but are different model generations with different files and support code." }, { "left": "pull_request:43448", "right": "pull_request:45490", "accept": false, - "reason": "Molmo vs CTSM are unrelated model additions; shared boilerplate files do not indicate duplication." + "reason": "Unrelated model additions: Molmo vs CTSM. Shared auto/docs touches are boilerplate, not a duplicate change." }, { "left": "pull_request:43448", "right": "pull_request:45532", "accept": false, - "reason": "Molmo vs SLANet are different model support PRs; they only share registry/docs plumbing." + "reason": "Different model additions: Molmo vs SLANet. The common files are registry/docs plumbing only." }, { "left": "pull_request:43451", "right": "pull_request:45490", "accept": false, - "reason": "Molmo2 and CTSM are separate model additions with different implementations and targets." + "reason": "Different model additions: Molmo2 vs CTSM. No shared bug or concrete codepath." }, { "left": "pull_request:43451", "right": "pull_request:45532", "accept": false, - "reason": "Molmo2 vs SLANet are unrelated model support changes despite some shared auto-mapping boilerplate." + "reason": "Different model additions: Molmo2 vs SLANet. Similar scaffolding does not make them duplicates." }, { "left": "pull_request:43838", "right": "pull_request:45283", "accept": false, - "reason": "Qwen3-ASR model support and Qwen3.5 GGUF loading support are different changes: one adds a model, the other adds loader mapping/quantization support." + "reason": "Qwen3-ASR model support and Qwen3.5 GGUF loading are different features affecting different codepaths." }, { "left": "pull_request:43973", "right": "pull_request:45283", "accept": false, - "reason": "Lfm2 Audio model support vs Qwen3.5 GGUF loading support are different concrete code paths and features." + "reason": "Lfm2Audio model support is unrelated to Qwen3.5 GGUF loader mapping; same broad audio/modeling area only." }, { "left": "pull_request:44601", "right": "pull_request:45532", "accept": false, - "reason": "Pipeline-parallel loading infrastructure and SLANet model support are unrelated; shared files are incidental registry/util updates." + "reason": "Pipeline-parallel loading infrastructure is unrelated to SLANet model support." }, { "left": "pull_request:45157", "right": "pull_request:45532", "accept": false, - "reason": "PrismML GGUF dequantization support vs SLANet model support are different feature areas and code paths." + "reason": "PrismML GGUF dequantization support and SLANet model support fix different problems and codepaths." }, { "left": "pull_request:45283", "right": "pull_request:45532", "accept": false, - "reason": "Qwen3.5 GGUF loading support and SLANet model support do not fix the same underlying problem." + "reason": "Qwen3.5 GGUF loading support and SLANet model support are different changes with no shared underlying bug." }, { "left": "pull_request:45490", "right": "pull_request:45532", "accept": false, - "reason": "CTSM and SLANet are distinct model-addition PRs; the shared files are generic docs/auto-registration boilerplate, not a duplicate fix." + "reason": "CTSM and SLANet are different model-support PRs; they only overlap on generic docs/model registration files, not the same feature." } ] }, @@ -15616,12 +15661,12 @@ "pull_request:43488|pull_request:45476", "pull_request:43532|pull_request:45476" ], - "prepared_review_unit_hash": "42f5c34ba74a7caf1287e059191967d4327d2221323a541a38d492f4270b4d21", + "prepared_review_unit_hash": "c2221f8667941beb44318f487c8d9c9a5e52b7ddf6578898f44736b99d90b0ba", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 7468, + "serialized_chars": 7467, "estimated_input_tokens": 1867, "estimated_eval_tokens": 3990 }, @@ -15629,7 +15674,7 @@ "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 7468, + "serialized_chars": 7467, "estimated_input_tokens": 1867, "estimated_eval_tokens": 3990 }, @@ -15638,30 +15683,30 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are three unrelated throwaway PRs that all share the same \"[don't merge]\"/test-noise pattern, but they change different code paths: packaging/version checks, CI workflow plumbing, and a PR-CI caller plus tokenizer/test junk. The shared vit test file looks incidental rather than a common underlying fix.", - "confidence": 0.89, + "summary": "These three PRs are unrelated synthetic \"don't merge\" changes with only incidental overlap in shared test filenames. They do not fix the same bug or implement the same change, so none should be clustered as duplicates.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43532 is the most coherent representative of the cluster: it has a single CI/workflow-oriented theme and more substantial, connected edits than the others.", + "canonical_pr_reason": "No canonical PR: each PR edits a different area and change intent (packaging/versioning, workflow/test plumbing, and CI workflow/tokenizer/test noise).", "best_issue_reason": null, - "best_pr_reason": "PR 43532 is the best representative because its changes are most internally consistent around CI workflow behavior, whereas the other two are narrower or more obviously synthetic checks.", + "best_pr_reason": "PR 43532 is the least representative as a cluster anchor because it has the broadest workflow/test plumbing changes, but it still is not a true duplicate of the others.", "soft_edge_verdicts": [ { "left": "pull_request:43488", "right": "pull_request:43532", "accept": false, - "reason": "Different underlying changes: 43488 is about setup/dependency/version formatting, while 43532 rewires self-comment CI and related utilities. Shared vit test edits look incidental." + "reason": "Different underlying changes: setup/version/dependency edits vs CI/workflow and test plumbing; shared test filename is incidental." }, { "left": "pull_request:43488", "right": "pull_request:45476", "accept": false, - "reason": "43488 targets packaging/version/dependency metadata; 45476 adds a PR-CI workflow caller and unrelated tokenizer/test edits. Not the same concrete bug or change." + "reason": "Different code paths and intent: packaging/version tweaks vs workflow + tokenizer/test noise; not the same change." }, { "left": "pull_request:43532", "right": "pull_request:45476", "accept": false, - "reason": "Both touch CI-adjacent files, but they address different mechanisms: self-comment CI versus PR CI caller setup. The overlap in tests is not enough to treat them as one change." + "reason": "Both touch test/workflow files, but the concrete edits are unrelated and would not plausibly merge into one PR." } ] }, @@ -15691,12 +15736,12 @@ "soft_pairs": [ "pull_request:43612|pull_request:43613" ], - "prepared_review_unit_hash": "06899421bbc6a9f495591f0a33989aac0c0d17bd8e1ce2002360b382f478d260", + "prepared_review_unit_hash": "e40ccb5c930499f9742b47b2bb59374536733f1625c54c514fb69895af195380", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6083, + "serialized_chars": 6081, "estimated_input_tokens": 1521, "estimated_eval_tokens": 3298 }, @@ -15704,7 +15749,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6083, + "serialized_chars": 6081, "estimated_input_tokens": 1521, "estimated_eval_tokens": 3298 }, @@ -15713,18 +15758,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both pull requests add new, similarly named segmentation pipelines and share a doc/registry scaffolding pattern, but they implement different features and code paths: one is for Promptable Concept Segmentation, the other for Promptable Visual Segmentation. They are related, not duplicates.", - "confidence": 0.95, + "summary": "These are two separate open PRs adding different pipeline features: promptable concept segmentation vs promptable visual segmentation. They share scaffolding and docs, but they target different tasks/models and should not be deduplicated.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "If a representative PR must be chosen, #43613 is the broader one: it adds the visual segmentation pipeline plus more model/test integrations and touches a wider set of supporting files.", + "canonical_pr_reason": "No single canonical PR: the PRs implement different segmentation pipelines with different model integrations and documentation.", "best_issue_reason": null, - "best_pr_reason": "#43613 is the better cluster representative because it is broader and more complete, but it should not be treated as a duplicate of #43612 since it targets a different pipeline and model path.", + "best_pr_reason": "If a representative is needed, 43613 is slightly broader, but it is still a distinct feature and not a substitute for 43612.", "soft_edge_verdicts": [ { "left": "pull_request:43612", "right": "pull_request:43613", "accept": false, - "reason": "Different underlying feature sets and code paths: concept segmentation vs visual segmentation. Shared docs/registry files and similar pipeline scaffolding are not enough to merge these as duplicates." + "reason": "Different underlying changes: one adds Promptable Concept Segmentation, the other Promptable Visual Segmentation. Shared pipeline/docs scaffolding is not enough to treat them as the same fix/change." } ] }, @@ -15754,7 +15799,7 @@ "soft_pairs": [ "pull_request:43651|pull_request:43663" ], - "prepared_review_unit_hash": "003548d77de3a1260dd8af1fdb16080c8ae8bd408f6713e4a0f1695bc16bf598", + "prepared_review_unit_hash": "386457bdd478d8aa4918a472de2787ebe80e477f50e4859d2377474e37fe1298", "component_budget": { "node_count": 2, "item_count": 2, @@ -15776,18 +15821,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two separate Trainer extensions in the same file: one adds a hook for gradient-accumulation loss scaling, the other adds a hook for signature-column filtering. They touch different code paths and are not duplicates.", - "confidence": 0.97, + "summary": "Both PRs touch `src/transformers/trainer.py`, but they address different extension points: one changes gradient-accumulation loss scaling, the other extracts signature-column filtering for custom trainers. They are not the same underlying change.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #43663 is the cleaner, self-contained API extension and is the better representative of this cluster if one must be chosen.", + "canonical_pr_reason": "PR #43651 is the better canonical representative because it makes a concrete behavior change in `training_step` and introduces `_loss_is_scaled_for_ga` to control loss scaling semantics.", "best_issue_reason": null, - "best_pr_reason": "PR #43663 is the most suitable standalone change: it refactors column-signature handling into an overrideable method without altering training behavior.", + "best_pr_reason": "PR #43651 is the strongest standalone PR in the cluster: it fixes a specific training-loss scaling path with a clear user-facing effect and is more directly substantive than the helper-method extraction in #43663.", "soft_edge_verdicts": [ { "left": "pull_request:43651", "right": "pull_request:43663", "accept": false, - "reason": "Both modify trainer.py, but they address unrelated concerns: gradient-accumulation loss scaling vs. dataloader column filtering. They do not fix the same concrete bug or change the same behavior." + "reason": "Different code paths and goals: #43651 changes gradient accumulation loss normalization, while #43663 refactors signature-column inspection for custom trainers. Shared file/subsystem is not enough to treat them as duplicates." } ] }, @@ -15819,52 +15864,52 @@ "pull_request:43656|pull_request:43842", "pull_request:43836|pull_request:43842" ], - "prepared_review_unit_hash": "079e389db8d8793a091a9ed93cc581121bb98faf128b46629530ee8e281698b9", + "prepared_review_unit_hash": "fbd1bcbf912d76cd72e921d798e552e89d2ff4c1191e259ed849a87626a3f297", "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 point at the same serve.py TypeAdapter runtime-annotation crash, with 43842 being the most direct and narrowly scoped fix.", - "confidence": 0.73, + "summary": "All three PRs target the same import-time annotation problem in `src/transformers/cli/serve.py`, with overlapping fixes around `TypeAdapter`/future annotations. PR #43842 is the cleanest, most focused representative of that bug fix.", + "confidence": 0.71, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43842 most directly fixes the reported NameError by keeping TypeAdapter out of runtime evaluation when pydantic is absent, with a minimal targeted diff and matching issue linkage.", + "canonical_pr_reason": "#43842 most directly addresses the concrete `TypeAdapter` NameError in the CLI serve path with the smallest, clearest change set.", "best_issue_reason": null, - "best_pr_reason": "PR 43842 is the best standalone fix: it addresses the concrete crash path in serve.py with the least extra churn and the clearest bug-specific intent.", + "best_pr_reason": "#43842 is the best cluster representative because it fixes the specific runtime failure directly and is narrower than the broader annotation cleanup in the other PRs.", "soft_edge_verdicts": [ { "left": "pull_request:43656", "right": "pull_request:43842", "accept": true, - "reason": "Both appear to fix the same serve.py TypeAdapter NameError in the CLI, using alternative annotation/runtime-import handling for the same crash path." + "reason": "Both are fixing the same serve CLI annotation/`TypeAdapter` NameError path in the same file, and their changes are plausibly mergeable into one PR." }, { "left": "pull_request:43836", "right": "pull_request:43842", "accept": true, - "reason": "Same file, same TypeAdapter NameError, and both change the import/annotation strategy to avoid evaluating TypeAdapter at runtime when pydantic is absent." + "reason": "Both target the same `TypeAdapter` annotation failure in `serve.py`; the differences look like implementation details around the same bug fix rather than separate changes." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet: all three PRs target src/transformers/cli/serve.py and the TypeAdapter NameError/runtime-annotation issue, with 43842 clearly the narrowest fix. The soft-edge verdicts are conservative enough, and the pairs look mergeable as alternative fixes for the same bug." + "feedback": "The cluster summary is grounded in the packet: all three PRs touch `src/transformers/cli/serve.py` and the diffs are consistent with the same import-time annotation/`TypeAdapter` NameError class of bug. The chosen representative, #43842, is also the smallest and most focused fix. The soft-edge accepts are reasonable and the paired changes look mergeable as overlapping implementations of the same bugfix." }, "evaluator_used": true, "retried": false, @@ -15894,7 +15939,7 @@ "pull_request:43757|pull_request:45500", "pull_request:43757|pull_request:45506" ], - "prepared_review_unit_hash": "89e25286137d69b5eece3fecc1d687959f9116531fdbbe39cef479e20daed9bf", + "prepared_review_unit_hash": "b5b9cb84d0a4ccee3281d75ffa388baeac95980e02907c19f7d51ac394ea6f55", "component_budget": { "node_count": 3, "item_count": 3, @@ -15916,30 +15961,30 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "All three PRs are about GPT-OSS GGUF loading, but only 45506 looks like the complete end-state implementation. 45500 is a draft follow-up and 43757 is a narrower fallback-only stopgap.", - "confidence": 0.95, + "summary": "Two PRs are near-exact duplicates for full GPT-OSS GGUF support; the earlier PR is only a partial fallback and should not be merged with the fuller implementation. The latest non-draft PR is the best canonical PR.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "45506 is the strongest canonical PR: it adds the GGUF metadata mapping, loader support, and tests, is not draft, and has review activity. 45500 is draft and 43757 only provides a fallback.", + "canonical_pr_reason": "PR #45506 is the most complete and current version: it adds the GGUF architecture mapping, loader handling, and tests, and it is explicitly positioned as the latest superseding work.", "best_issue_reason": null, - "best_pr_reason": "45506 best represents the full fix for GPT-OSS GGUF support; it is the most complete and mergeable implementation among the three.", + "best_pr_reason": "PR #45506 is the strongest representative of the cluster because it contains the full end-to-end GGUF support changes, while #45500 is a draft variant and #43757 is only a fallback workaround.", "soft_edge_verdicts": [ { "left": "pull_request:45500", "right": "pull_request:45506", "accept": true, - "reason": "Same underlying GPT-OSS GGUF loading work, same tracking issue, and nearly identical scope; 45506 is the fuller version of 45500." + "reason": "Same underlying GPT-OSS GGUF loading change: same title, same issue target, same core files, and the later PR appears to be the finalized version of the earlier draft." }, { "left": "pull_request:43757", "right": "pull_request:45500", "accept": false, - "reason": "Related but not the same concrete change: 43757 only adds a fallback to avoid hard failure, while 45500 implements full GGUF support." + "reason": "43757 is only a narrow fallback-to-gpt-neox workaround, while 45500 is a broader full GGUF-loading implementation; they overlap on one file and issue, but not the same concrete fix." }, { "left": "pull_request:43757", "right": "pull_request:45506", "accept": false, - "reason": "43757 is a narrow compatibility fallback, whereas 45506 is a broader full-support PR; they are not the same fix and would not plausibly merge as one PR." + "reason": "43757 is a partial stopgap and 45506 is a full support PR; they address the same broad topic but not the same mergeable change set or exact code-path fix." } ] }, @@ -16037,6 +16082,7 @@ "pull_request:44017", "pull_request:44018", "pull_request:44019", + "pull_request:44024", "pull_request:44044", "pull_request:44056", "pull_request:44059", @@ -16066,8 +16112,8 @@ "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:44018|pull_request:44068", "pull_request:44066|pull_request:44068", "pull_request:44018|pull_request:44066", "pull_request:44019|pull_request:44722", @@ -16077,12 +16123,12 @@ "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:44116", "pull_request:44018|pull_request:44086", "pull_request:44018|pull_request:44085", - "pull_request:44068|pull_request:44086", "pull_request:44071|pull_request:44085", + "pull_request:44068|pull_request:44086", "pull_request:44059|pull_request:44086", "pull_request:44085|pull_request:44086", "pull_request:44059|pull_request:44722", @@ -16110,7 +16156,6 @@ "pull_request:43997|pull_request:44018", "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", @@ -16129,354 +16174,415 @@ "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:44044|pull_request:44154", + "pull_request:44024|pull_request:44076" ], - "prepared_review_unit_hash": "8e4bbb82040c6f2fadbbd9d411e63ae54af34b8876ae39d8893f031653b99f67", + "prepared_review_unit_hash": "4adf64f07e74dccaa041a5263441a9d2095b26d5eac2780a425656d7a78eedac", "component_budget": { "node_count": 28, "item_count": 28, - "soft_pair_count": 78, - "serialized_chars": 84898, - "estimated_input_tokens": 21225, - "estimated_eval_tokens": 42706 + "soft_pair_count": 77, + "serialized_chars": 84500, + "estimated_input_tokens": 21125, + "estimated_eval_tokens": 42506 }, "budget": { - "node_count": 27, - "item_count": 27, + "node_count": 28, + "item_count": 28, "soft_pair_count": 72, - "serialized_chars": 75093, - "estimated_input_tokens": 18774, - "estimated_eval_tokens": 37804 + "serialized_chars": 76934, + "estimated_input_tokens": 19234, + "estimated_eval_tokens": 38724 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a set of model-specific refactors to standardized output tracing, not one single duplicate change. Only the GPTJ/CodeGen variants and the GPT-Neo variants look like true overlaps; the rest share the umbrella issue but not the same code path.", - "confidence": 0.95, + "summary": "This cluster is mostly a broad family of model-by-model output-tracing refactors, but only two duplicate groups look real: the GPT-Neo pair and the GPT-J/CodeGen pair. Most other edges share only the umbrella issue #43979 or the same refactor theme, which is too broad to merge as duplicates.", + "confidence": 0.82, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44722 is the best canonical PR for the GPTJ/CodeGen subcluster: it is the latest and most complete instance of that concrete output-tracing refactor.", + "canonical_pr_reason": "PR 44722 is the best representative of the duplicate subset: it is the latest and broadest GPT-J/CodeGen standardized-output-tracing refactor, and it appears to subsume the earlier GPT-J-only iterations.", "best_issue_reason": null, - "best_pr_reason": "44722 is the strongest representative of the shared decorator-based output-tracing migration, with the broadest concrete GPTJ/CodeGen implementation among the overlapping PRs.", + "best_pr_reason": "PR 44722 is the strongest global representative because it most clearly captures the repeated GPT-J output-tracing change, with the cleanest and broadest scope among the duplicate candidates.", "soft_edge_verdicts": [ { "left": "pull_request:43996", "right": "pull_request:44085", "accept": false, - "reason": "Different model families and code paths; only the umbrella tracking issue matches." + "reason": "Different models/files; same umbrella output-tracing theme is too broad to call these duplicates." }, { "left": "pull_request:43996", "right": "pull_request:44044", "accept": false, - "reason": "Different model families and code paths; same issue target is too broad to merge." + "reason": "Different model code paths; both are standardized-output-tracing refactors, but not the same change." }, { "left": "pull_request:44066", "right": "pull_request:44085", "accept": true, - "reason": "Both are GPTJ output-tracing refactors on the same code path and could plausibly be one PR." + "reason": "Same GPT-J output-tracing refactor in the same file; these look like overlapping iterations of the same change." }, { "left": "pull_request:44007", "right": "pull_request:44072", "accept": false, - "reason": "ResNet/RT-DETR ResNet versus EfficientNet; same theme, but not the same concrete change." + "reason": "ResNet and EfficientNet are separate implementations; same refactor pattern is not enough for duplicate status." }, { "left": "pull_request:44072", "right": "pull_request:44722", "accept": false, - "reason": "EfficientNet refactor versus GPTJ/CodeGen refactor; unrelated code paths." + "reason": "Different model families and code paths; only the high-level output-tracing theme overlaps." }, { "left": "pull_request:44066", "right": "pull_request:44072", "accept": false, - "reason": "GPTJ/CodeGen versus EfficientNet; shared output-tracing theme only." + "reason": "Different models/files; same standardized output-tracing umbrella is too broad." }, { "left": "pull_request:44013", "right": "pull_request:44044", "accept": false, - "reason": "MobileNetV2 versus DeBERTa; separate model-specific migrations." + "reason": "MobileNetV2 and DeBERTaV2 are unrelated code paths; only the refactor style matches." }, { "left": "pull_request:44066", "right": "pull_request:44086", "accept": false, - "reason": "GPTJ/CodeGen versus MGP-STR; not the same underlying bug or change." - }, - { - "left": "pull_request:44018", - "right": "pull_request:44068", - "accept": true, - "reason": "Both are GPT-Neo output-tracing refactors for the same model and code path." + "reason": "GPT-J/CodeGen and MGP-STR are different model changes, not the same underlying fix." }, { "left": "pull_request:44066", "right": "pull_request:44071", "accept": false, - "reason": "GPTJ/CodeGen versus MPT; only the shared decorator pattern matches." + "reason": "Different model families and implementations; same output-tracing modernization is not a duplicate signal." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44068", + "accept": true, + "reason": "Same GPT-Neo file and same output-tracing refactor; these are likely duplicate iterations of the same change." }, { "left": "pull_request:44066", "right": "pull_request:44068", "accept": false, - "reason": "GPTJ versus GPT-Neo; similar infrastructure work but different model code paths." + "reason": "GPT-J/CodeGen vs GPT-Neo are different code paths, so this is only a thematic match." }, { "left": "pull_request:44018", "right": "pull_request:44066", "accept": false, - "reason": "GPT-Neo versus GPTJ/CodeGen; not the same refactor." + "reason": "Different model families; shared issue target alone does not make them duplicates." }, { "left": "pull_request:44019", "right": "pull_request:44722", "accept": false, - "reason": "ResNet versus GPTJ/CodeGen; same umbrella issue, different implementations." + "reason": "ResNet and GPT-J/CodeGen are unrelated model changes." }, { "left": "pull_request:44086", "right": "pull_request:44722", "accept": false, - "reason": "MGP-STR versus GPTJ/CodeGen; no shared concrete code path." + "reason": "Different models and concrete code paths; same tracking issue is not enough." }, { "left": "pull_request:44019", "right": "pull_request:44085", "accept": false, - "reason": "ResNet versus GPTJ; distinct model changes." + "reason": "ResNet and GPT-J are separate refactors, not the same underlying bug or change." }, { "left": "pull_request:44018", "right": "pull_request:44019", "accept": false, - "reason": "GPT-Neo versus ResNet; related pattern, but not a duplicate." + "reason": "GPT-Neo and ResNet are different models; no shared concrete code path." }, { "left": "pull_request:44019", "right": "pull_request:44071", "accept": false, - "reason": "ResNet versus MPT; different model-specific output handling." + "reason": "ResNet and MPT are unrelated implementations, despite the same general refactor theme." }, { - "left": "pull_request:44068", - "right": "pull_request:44116", + "left": "pull_request:44018", + "right": "pull_request:44086", "accept": false, - "reason": "GPT-Neo versus Flaubert/XLM; not the same underlying fix." + "reason": "GPT-Neo and MGP-STR are different model code paths." }, { "left": "pull_request:44068", "right": "pull_request:44722", "accept": false, - "reason": "GPT-Neo versus GPTJ/CodeGen; same refactor style, different code paths." + "reason": "GPT-Neo and GPT-J/CodeGen are different models; only the refactor pattern overlaps." }, { - "left": "pull_request:44018", - "right": "pull_request:44116", + "left": "pull_request:44071", + "right": "pull_request:44722", "accept": false, - "reason": "GPT-Neo versus Flaubert/XLM; separate model migrations." + "reason": "MPT and GPT-J/CodeGen are separate implementations; not the same concrete fix." }, { - "left": "pull_request:44071", - "right": "pull_request:44722", + "left": "pull_request:44018", + "right": "pull_request:44116", "accept": false, - "reason": "MPT versus GPTJ/CodeGen; only the output-tracing framework is shared." + "reason": "GPT-Neo and Flaubert are different model code paths." }, { "left": "pull_request:44018", "right": "pull_request:44086", "accept": false, - "reason": "GPT-Neo versus MGP-STR; not the same concrete change." + "reason": "Different models/files; same standardized-output-tracing theme is too broad." }, { "left": "pull_request:44018", "right": "pull_request:44085", "accept": false, - "reason": "GPT-Neo versus GPTJ; similar theme, different implementation." + "reason": "GPT-Neo and GPT-J are distinct refactors, not duplicates." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44085", + "accept": false, + "reason": "MPT and GPT-J are unrelated concrete changes." }, { "left": "pull_request:44068", "right": "pull_request:44086", "accept": false, - "reason": "GPT-Neo versus MGP-STR; unrelated model code paths." + "reason": "GPT-Neo and MGP-STR are different code paths." }, { "left": "pull_request:44059", "right": "pull_request:44086", "accept": false, - "reason": "GPT2 versus MGP-STR; only the umbrella output-tracing work overlaps." + "reason": "GPT2 and MGP-STR are separate model refactors." }, { "left": "pull_request:44085", "right": "pull_request:44086", "accept": false, - "reason": "GPTJ versus MGP-STR; not the same bug or refactor." + "reason": "GPT-J and MGP-STR are unrelated changes." }, { "left": "pull_request:44059", "right": "pull_request:44722", "accept": false, - "reason": "GPT2 versus GPTJ/CodeGen; distinct model-specific refactors." + "reason": "GPT2 and GPT-J/CodeGen are different model implementations." }, { "left": "pull_request:44098", "right": "pull_request:44101", "accept": false, - "reason": "ViLT versus Flaubert/XLM; different model families." + "reason": "ViLT and Flaubert/XLM are different model code paths." }, { - "left": "pull_request:44073", - "right": "pull_request:44101", + "left": "pull_request:44018", + "right": "pull_request:44071", "accept": false, - "reason": "VisualBert versus Flaubert/XLM; not the same code path." + "reason": "GPT-Neo and MPT are separate refactors with no shared concrete bug." }, { "left": "pull_request:44017", "right": "pull_request:44722", "accept": false, - "reason": "Segformer versus GPTJ/CodeGen; separate migrations." + "reason": "SegFormer and GPT-J/CodeGen are unrelated model changes." }, { "left": "pull_request:44017", "right": "pull_request:44161", "accept": false, - "reason": "Segformer versus LongT5; same pattern, different models." + "reason": "SegFormer and LongT5 are different implementations; only the umbrella refactor theme matches." }, { "left": "pull_request:44017", "right": "pull_request:44073", "accept": false, - "reason": "Segformer versus VisualBert; unrelated implementations." + "reason": "SegFormer and VisualBert are distinct code paths." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-Neo and MPT are different model changes." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44086", + "accept": false, + "reason": "MPT and MGP-STR are unrelated model implementations." }, { "left": "pull_request:43997", "right": "pull_request:44073", "accept": false, - "reason": "RegNet versus VisualBert; different model-specific code." + "reason": "RegNet and VisualBert are separate code paths." }, { - "left": "pull_request:44010", - "right": "pull_request:44071", + "left": "pull_request:44056", + "right": "pull_request:44101", "accept": false, - "reason": "SqueezeBert versus MPT; not a duplicate change." + "reason": "MPNet and Flaubert/XLM are different model refactors." }, { - "left": "pull_request:44010", - "right": "pull_request:44073", + "left": "pull_request:44073", + "right": "pull_request:44161", "accept": false, - "reason": "SqueezeBert versus VisualBert; separate output-tracing refactors." + "reason": "VisualBert and LongT5 are unrelated model implementations." }, { - "left": "pull_request:44010", + "left": "pull_request:44101", + "right": "pull_request:44161", + "accept": false, + "reason": "Flaubert/XLM and LongT5 are separate model code paths." + }, + { + "left": "pull_request:44017", "right": "pull_request:44018", "accept": false, - "reason": "SqueezeBert versus GPT-Neo; same theme only." + "reason": "SegFormer and GPT-Neo are different models; same refactor style is not enough." }, { "left": "pull_request:43997", "right": "pull_request:44056", "accept": false, - "reason": "RegNet versus MPNet; different models and code paths." + "reason": "RegNet and MPNet are separate implementations." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44076", + "accept": false, + "reason": "MPT and ImageGPT are different model code paths." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44101", + "accept": false, + "reason": "TextNet and Flaubert/XLM are unrelated changes." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44161", + "accept": false, + "reason": "MPNet and LongT5 are different model implementations." }, { "left": "pull_request:43997", "right": "pull_request:44018", "accept": false, - "reason": "RegNet versus GPT-Neo; not the same underlying fix." + "reason": "RegNet and GPT-Neo are unrelated model refactors." }, { "left": "pull_request:43997", "right": "pull_request:44161", "accept": false, - "reason": "RegNet versus LongT5; separate refactors." + "reason": "RegNet and LongT5 are different code paths." }, { "left": "pull_request:44074", - "right": "pull_request:44101", + "right": "pull_request:44161", "accept": false, - "reason": "TextNet versus Flaubert/XLM; different model implementations." + "reason": "TextNet and LongT5 are separate model changes." }, { - "left": "pull_request:44056", - "right": "pull_request:44161", + "left": "pull_request:43997", + "right": "pull_request:44071", "accept": false, - "reason": "MPNet versus LongT5; same umbrella issue, different code paths." + "reason": "RegNet and MPT are different model implementations." }, { - "left": "pull_request:44074", + "left": "pull_request:44010", + "right": "pull_request:44071", + "accept": false, + "reason": "SqueezeBert and MPT are unrelated concrete refactors." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44073", + "accept": false, + "reason": "SqueezeBert and VisualBert are different model code paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44129", + "accept": false, + "reason": "GPT-Neo and SpeechT5 are separate model changes." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44129", + "accept": false, + "reason": "MPT and SpeechT5 are different concrete refactors." + }, + { + "left": "pull_request:43996", "right": "pull_request:44161", "accept": false, - "reason": "TextNet versus LongT5; not the same concrete change." + "reason": "CVT/FNet and LongT5 are unrelated model implementations." }, { "left": "pull_request:43996", "right": "pull_request:44017", "accept": false, - "reason": "FNet/CVT versus Segformer; unrelated model-specific work." + "reason": "CVT/FNet and SegFormer are different model code paths." }, { "left": "pull_request:44010", "right": "pull_request:44044", "accept": false, - "reason": "SqueezeBert versus DeBERTa; distinct changes." + "reason": "SqueezeBert and DeBERTaV2 are different models; only the refactor style overlaps." }, { "left": "pull_request:44044", "right": "pull_request:44074", "accept": false, - "reason": "DeBERTa versus TextNet; same pattern, different code paths." + "reason": "DeBERTaV2 and TextNet are unrelated model changes." }, { "left": "pull_request:44066", "right": "pull_request:44722", "accept": true, - "reason": "Both cover the same GPTJ/CodeGen decorator migration and could be merged into one PR." + "reason": "Same GPT-J/CodeGen output-tracing refactor; 44722 looks like the later, broader version of the same change." }, { "left": "pull_request:44085", "right": "pull_request:44722", "accept": true, - "reason": "Both are GPTJ output-tracing refactors with overlapping concrete changes." + "reason": "Both target the same GPT-J output-tracing code path; these look like duplicate iterations." }, { "left": "pull_request:44085", "right": "pull_request:44129", "accept": false, - "reason": "GPTJ versus SpeechT5; same umbrella issue, not the same bug." - }, - { - "left": "pull_request:43996", - "right": "pull_request:44007", - "accept": false, - "reason": "FNet/CVT versus ResNet/RT-DETR ResNet; unrelated model refactors." + "reason": "GPT-J and SpeechT5 are different models and concrete code paths." }, { - "left": "pull_request:43996", + "left": "pull_request:44044", "right": "pull_request:44129", "accept": false, - "reason": "FNet/CVT versus SpeechT5; different code paths." + "reason": "DeBERTaV2 and SpeechT5 are unrelated refactors." }, { "left": "pull_request:44044", - "right": "pull_request:44129", + "right": "pull_request:44154", "accept": false, - "reason": "DeBERTa versus SpeechT5; separate model-specific changes." + "reason": "DeBERTaV2 and ViTS are different model implementations." }, { - "left": "pull_request:44044", - "right": "pull_request:44154", + "left": "pull_request:44024", + "right": "pull_request:44076", "accept": false, - "reason": "DeBERTa versus ViTS; not the same underlying refactor." + "reason": "FocalNet and ImageGPT are separate model code paths." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative about overlap versus shared umbrella issue. The accepted soft-edge pairs are all within the same model-family refactor and look plausibly mergeable into a single PR for the same bugfix, while the rejected pairs are appropriately cautious about only sharing the broader standardized output-tracing theme." + "feedback": "The summary is conservative and stays close to the packet: it identifies the broad output-tracing refactor theme and only elevates the clearest same-code-path pairs as duplicates. The accepted soft-edge verdicts are also restrained and mergeable for the same bugfix. I don\u2019t see an overstatement of evidence." }, "evaluator_used": true, "retried": false, @@ -16499,85 +16605,76 @@ "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": "a39be23eb4eeaf89398285982d8c83e5ee8f2f9c5677e4d1af2a76e2e9da9b0b", + "prepared_review_unit_hash": "ed6ec968cf7aaba71a9aad14a8aa970e7f83343c5ebf46a3bac4eb4edd254fae", "component_budget": { "node_count": 28, "item_count": 28, - "soft_pair_count": 78, - "serialized_chars": 84898, - "estimated_input_tokens": 21225, - "estimated_eval_tokens": 42706 + "soft_pair_count": 77, + "serialized_chars": 84500, + "estimated_input_tokens": 21125, + "estimated_eval_tokens": 42506 }, "budget": { - "node_count": 8, - "item_count": 8, - "soft_pair_count": 6, - "serialized_chars": 17126, - "estimated_input_tokens": 4282, - "estimated_eval_tokens": 8820 + "node_count": 6, + "item_count": 6, + "soft_pair_count": 5, + "serialized_chars": 13066, + "estimated_input_tokens": 3267, + "estimated_eval_tokens": 6790 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These PRs are all model-specific refactors to standardized output tracing/capture_outputs, but they target different architectures and code paths, so they are not duplicates of one another.", - "confidence": 0.91, + "summary": "These are model-specific refactors to standardized output tracing/capture infrastructure across different architectures, not duplicates of the same bug or change. They share a theme, but each PR touches a distinct code path and model implementation.", + "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44129 is the strongest representative of the cluster: it has an explicit issue link, a clear output-tracing refactor, and enough surrounding changes to show the intended migration pattern.", + "canonical_pr_reason": "No single canonical PR cleanly represents the whole set because each PR is a separate model migration. If one must be chosen as the most representative, 44073 is the clearest self-contained example of the standardized output-tracing refactor.", "best_issue_reason": null, - "best_pr_reason": "PR 44129 is the best single PR to anchor the cluster because it is the most explicitly scoped and documented, while the others are similar migrations in unrelated models.", + "best_pr_reason": "44073 is the most representative PR for this cluster: it explicitly introduces capture_outputs and _can_record_outputs for one model, making it a clean exemplar of the broader tracing migration.", "soft_edge_verdicts": [ - { - "left": "pull_request:44024", - "right": "pull_request:44076", - "accept": false, - "reason": "Same broad refactor theme, but different models (FocalNet vs ImageGPT) and different code paths; not the same underlying change." - }, { "left": "pull_request:44073", "right": "pull_request:44074", "accept": false, - "reason": "Both migrate output tracing, but one is VisualBert and the other TextNet; they are separate model implementations and cannot be merged as one concrete fix." + "reason": "Both are output-tracing refactors, but they apply to different models (VisualBert vs TextNet) with different forward paths and tests; not the same concrete change." }, { "left": "pull_request:44056", "right": "pull_request:44073", "accept": false, - "reason": "MPNet and VisualBert are unrelated model code paths; the similarity is only the shared standardized-output refactor pattern." + "reason": "Same broad migration pattern, but MPNet and VisualBert have separate model-specific tracing hooks and code paths; they are not mergeable as one PR." }, { "left": "pull_request:44056", "right": "pull_request:44074", "accept": false, - "reason": "MPNet and TextNet touch different architectures and forward paths, so this is not the same underlying bug or change." + "reason": "These are distinct refactors for different model implementations, not the same underlying bug or fix." }, { "left": "pull_request:43997", "right": "pull_request:44010", "accept": false, - "reason": "RegNet and SqueezeBert are unrelated implementations; these are separate output-tracing migrations, not one shared fix." + "reason": "RegNet and SqueezeBert each migrate output collection in their own model code; the overlap is thematic only." }, { "left": "pull_request:44010", "right": "pull_request:44129", "accept": false, - "reason": "SqueezeBert and SpeechT5 both adjust output capture, but they affect different model internals and cannot plausibly be merged into one PR as the same concrete code-path fix." + "reason": "They target different models and different tracing logic; sharing a standardized output-capture pattern is not enough to treat them as duplicates." } ] }, @@ -16656,268 +16753,268 @@ "pull_request:44028|pull_request:44030", "pull_request:44029|pull_request:44030" ], - "prepared_review_unit_hash": "1ed95264fa2191610891d8373501d4dee45e5e4cbed692316150e57dc1dc82b1", + "prepared_review_unit_hash": "fc8e6f5dfcdb6c901a766e39719bbc319fbeb9a828e0bb9ef5f75829a237a9ee", "component_budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 39, - "serialized_chars": 39774, - "estimated_input_tokens": 9944, - "estimated_eval_tokens": 20144 + "serialized_chars": 39784, + "estimated_input_tokens": 9946, + "estimated_eval_tokens": 20148 }, "budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 39, - "serialized_chars": 39774, - "estimated_input_tokens": 9944, - "estimated_eval_tokens": 20144 + "serialized_chars": 39784, + "estimated_input_tokens": 9946, + "estimated_eval_tokens": 20148 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are all similar output-tracing refactors, but they target different model implementations and code paths. They are related by theme, not by the same underlying bug/change, so the soft edges should be rejected. No issue is present; one PR can serve as the cluster representative.", - "confidence": 0.84, + "summary": "These PRs all use similar 'refactor output tracing' wording, but they are model-specific changes across unrelated code paths (backbones, encoder-decoder wrappers, depth models, Mamba/RWKV, etc.). I would not dedupe them on title similarity alone.", + "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44003 is the strongest representative because it is a broader output-tracing refactor that touches two related model files and introduces the new capturing plumbing, making it more central than the single-model variants.", + "canonical_pr_reason": "No canonical PR: the set spans many independent model implementations rather than one duplicated fix.", "best_issue_reason": null, - "best_pr_reason": "PR 44003 best represents the cluster\u2019s common pattern while still being a concrete, nontrivial code change across multiple files.", + "best_pr_reason": "No single best PR for deduplication; the PRs are separate, self-contained refactors in different models.", "soft_edge_verdicts": [ { "left": "pull_request:43998", "right": "pull_request:44002", "accept": false, - "reason": "Different model families and files (timm_backbone vs upernet); same refactor theme, not the same code-path bug." + "reason": "Different model code paths; same tracing refactor theme is too broad to be a duplicate." }, { "left": "pull_request:43998", "right": "pull_request:44000", "accept": false, - "reason": "Timm backbone and vision-text dual encoder are unrelated implementations; this is only a shared output-tracing pattern." + "reason": "Distinct architectures and files; not the same underlying bug or change." }, { "left": "pull_request:43998", "right": "pull_request:44001", "accept": false, - "reason": "Different model code paths (timm_backbone vs univnet); no evidence of one shared defect." + "reason": "Only shares the generic output-tracing pattern, not a concrete shared fix." }, { "left": "pull_request:43998", "right": "pull_request:44003", "accept": false, - "reason": "These touch different model stacks (timm_backbone vs mamba/falcon_mamba) and are separate refactors." + "reason": "Unrelated model family and implementation; not mergeable as one PR." }, { "left": "pull_request:43998", "right": "pull_request:44029", "accept": false, - "reason": "Different model families and mechanics; same refactor motif, not mergeable as one fix." + "reason": "Different code paths in different models; similarity is just thematic." }, { "left": "pull_request:43999", "right": "pull_request:44001", "accept": false, - "reason": "MobileNetV1 and UnivNet are unrelated model paths; the overlap is only in output-tracing cleanup." + "reason": "MobileNet vs UnivNet are unrelated models; no shared concrete defect." }, { "left": "pull_request:43999", "right": "pull_request:44002", "accept": false, - "reason": "MobileNetV1 vs UperNet are separate implementations; no shared underlying bug is shown." + "reason": "These touch different model implementations, not one duplicate fix." }, { "left": "pull_request:43999", "right": "pull_request:44025", "accept": false, - "reason": "Different model families and files; both are refactors, but not the same concrete change." + "reason": "Different architectures and files; only the refactor style matches." }, { "left": "pull_request:43999", "right": "pull_request:44026", "accept": false, - "reason": "MobileNetV1 and VisionEncoderDecoder are unrelated code paths despite similar signature cleanup." + "reason": "Separate model-specific refactors; not the same underlying change." }, { "left": "pull_request:43999", "right": "pull_request:44027", "accept": false, - "reason": "MobileNetV1 and SpeechEncoderDecoder are separate model stacks; same pattern, different fixes." + "reason": "Only broad subsystem overlap; the code paths are different." }, { "left": "pull_request:44000", "right": "pull_request:44001", "accept": false, - "reason": "Vision-Text Dual Encoder and UnivNet are different model paths; no shared bug or mergeable PR pair." + "reason": "Vision-text dual encoder and UnivNet are unrelated; not a duplicate." }, { "left": "pull_request:44000", "right": "pull_request:44002", "accept": false, - "reason": "Different model families (vision-text dual encoder vs UperNet); only the refactor style matches." + "reason": "Same high-level theme, but different model code and behavior." }, { "left": "pull_request:44000", "right": "pull_request:44003", "accept": false, - "reason": "Vision-text dual encoder and mamba/falcon_mamba are unrelated concrete changes." + "reason": "Mamba/FalconMamba refactor is a different concrete change from vision-text dual encoder." }, { "left": "pull_request:44000", "right": "pull_request:44025", "accept": false, - "reason": "Different targets and implementations; both are output-tracing refactors, but not the same fix." + "reason": "Different models and files; too generic to treat as the same fix." }, { "left": "pull_request:44001", "right": "pull_request:44002", "accept": false, - "reason": "UnivNet and UperNet are distinct model paths; no evidence they address one underlying bug." + "reason": "UnivNet and UperNet are unrelated code paths; not a duplicate change." }, { "left": "pull_request:44001", "right": "pull_request:44003", "accept": false, - "reason": "UnivNet vs mamba/falcon_mamba are separate changes with only superficial similarity." + "reason": "Different architectures and refactor targets; not the same bug or change." }, { "left": "pull_request:44001", "right": "pull_request:44025", "accept": false, - "reason": "Different model stacks and changed files; not the same concrete issue." + "reason": "No shared concrete code-path problem, only similar output-tracing wording." }, { "left": "pull_request:44002", "right": "pull_request:44029", "accept": false, - "reason": "UperNet and RWKV are unrelated model implementations; shared theme is too broad." + "reason": "UperNet and RWKV are unrelated model implementations." }, { "left": "pull_request:44002", "right": "pull_request:44003", "accept": false, - "reason": "Different model code paths and refactor scopes; not plausibly one merged PR." + "reason": "Different model families and concrete edits; not dedupable." }, { "left": "pull_request:44003", "right": "pull_request:44004", "accept": false, - "reason": "Mamba/falcon_mamba output tracing and CodeGen output tracing are separate model-specific changes." + "reason": "Mamba and CodeGen changes affect different subsystems and code paths." }, { "left": "pull_request:43998", "right": "pull_request:44004", "accept": false, - "reason": "Timm backbone and CodeGen are unrelated subsystems; the similarity is only generic refactoring." + "reason": "Timm backbone and CodeGen are unrelated; same refactor language is insufficient." }, { "left": "pull_request:44001", "right": "pull_request:44004", "accept": false, - "reason": "UnivNet and CodeGen fix different model code paths, so they should not be merged as duplicates." + "reason": "UnivNet vs CodeGen are different models with different outputs and call paths." }, { "left": "pull_request:44002", "right": "pull_request:44004", "accept": false, - "reason": "UperNet and CodeGen are distinct implementations with no shared concrete defect." + "reason": "UperNet and CodeGen do not share a concrete underlying bug." }, { "left": "pull_request:44000", "right": "pull_request:44004", "accept": false, - "reason": "Vision-text dual encoder vs CodeGen are different model paths; same broad theme is insufficient." + "reason": "Vision-text dual encoder and CodeGen are not the same change." }, { "left": "pull_request:44025", "right": "pull_request:44026", "accept": false, - "reason": "Depth Anything and VisionEncoderDecoder are unrelated model families; not the same bug or change." + "reason": "Depth Anything and Vision-Encoder-Decoder are unrelated; same helper usage is not enough." }, { "left": "pull_request:44025", "right": "pull_request:44027", "accept": false, - "reason": "Depth Anything and SpeechEncoderDecoder are different code paths; only the output-tracing refactor matches." + "reason": "Different model families and implementation details; not a duplicate." }, { "left": "pull_request:44025", "right": "pull_request:44028", "accept": false, - "reason": "Depth Anything and SuperPoint are separate models; no shared underlying issue is evident." + "reason": "Depth model refactor vs SuperPoint refactor are separate changes." }, { "left": "pull_request:44025", "right": "pull_request:44029", "accept": false, - "reason": "Depth Anything and RWKV are unrelated implementations, so this is not one duplicate change." + "reason": "Depth Anything and RWKV touch entirely different code paths." }, { "left": "pull_request:44025", "right": "pull_request:44030", "accept": false, - "reason": "Depth Anything and DPR touch different model stacks and should remain separate PRs." + "reason": "Different model types and files; no shared concrete fix." }, { "left": "pull_request:44026", "right": "pull_request:44029", "accept": false, - "reason": "VisionEncoderDecoder and RWKV are different code paths; common output-tracing refactor is too broad." + "reason": "Vision-Encoder-Decoder and RWKV are unrelated implementations." }, { "left": "pull_request:44026", "right": "pull_request:44027", "accept": false, - "reason": "VisionEncoderDecoder and SpeechEncoderDecoder are separate encoder-decoder variants with different implementations." + "reason": "Encoder-decoder wrappers are different model classes here, not one duplicate PR." }, { "left": "pull_request:44026", "right": "pull_request:44028", "accept": false, - "reason": "Different model families (vision encoder-decoder vs SuperPoint), so no same-bug basis." + "reason": "Only generic tracing refactor similarity; not the same code path." }, { "left": "pull_request:44026", "right": "pull_request:44030", "accept": false, - "reason": "VisionEncoderDecoder and DPR are unrelated model code paths; not mergeable as one fix." + "reason": "Different models and behavior; not plausibly one merged fix." }, { "left": "pull_request:44027", "right": "pull_request:44028", "accept": false, - "reason": "SpeechEncoderDecoder and SuperPoint are unrelated model implementations." + "reason": "Speech-encoder-decoder and SuperPoint are unrelated changes." }, { "left": "pull_request:44027", "right": "pull_request:44029", "accept": false, - "reason": "SpeechEncoderDecoder and RWKV are different concrete code paths; only the refactor style overlaps." + "reason": "Different model families and concrete edits." }, { "left": "pull_request:44027", "right": "pull_request:44030", "accept": false, - "reason": "SpeechEncoderDecoder and DPR are unrelated targets, so this is not a duplicate change." + "reason": "No shared underlying bug; just similar refactor phrasing." }, { "left": "pull_request:44028", "right": "pull_request:44029", "accept": false, - "reason": "SuperPoint and RWKV address different model logic; same theme does not justify merging." + "reason": "SuperPoint and RWKV are unrelated model implementations." }, { "left": "pull_request:44028", "right": "pull_request:44030", "accept": false, - "reason": "SuperPoint and DPR are distinct model code paths with no shared bug." + "reason": "Different model code paths; not the same change." }, { "left": "pull_request:44029", "right": "pull_request:44030", "accept": false, - "reason": "RWKV and DPR are unrelated implementations; these are separate model-specific refactors." + "reason": "RWKV and DPR are separate models with separate fixes." } ] }, @@ -16950,7 +17047,7 @@ "pull_request:44662|pull_request:45075", "pull_request:45075|pull_request:45401" ], - "prepared_review_unit_hash": "89e02490afbfc5586ddd71e25599998ba6dac42e01bab8a9ce87a708572a10ae", + "prepared_review_unit_hash": "9c56da9ee5142f48da9654b381d7986dd73a126d5d2adf78ad9f1da5644564cc", "component_budget": { "node_count": 3, "item_count": 3, @@ -16972,30 +17069,30 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are three separate model-addition pull requests for different architectures (PenguinVL, DeepSeek-OCR-2, Voxtral TTS). They share only boilerplate registry/docs updates, not the same underlying bug or code change, so none should be clustered as duplicates.", - "confidence": 0.97, + "summary": "These are three separate model-addition PRs (PenguinVL, DeepSeek-OCR-2, Voxtral TTS) with only boilerplate overlap in auto-registration/docs. They do not appear to fix the same underlying change, so none should be clustered as duplicates.", + "confidence": 0.99, "canonical_issue_reason": null, "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": null, + "best_pr_reason": "No PR is a true duplicate of the others; if one must be kept as a representative, 45075 is the most self-contained with an explicit implementation scope and broad test coverage, but it is not equivalent to the others.", "soft_edge_verdicts": [ { "left": "pull_request:44662", "right": "pull_request:45401", "accept": false, - "reason": "Different model implementations and code paths: PenguinVL vision-language model vs Voxtral TTS speech model. Shared auto-registration/doc scaffolding is generic boilerplate, not the same change." + "reason": "Different model integrations for different products; shared files are just common registration/docs boilerplate, not the same code-path change." }, { "left": "pull_request:44662", "right": "pull_request:45075", "accept": false, - "reason": "Different targets and functionality: PenguinVL vs DeepSeek-OCR-2. The overlap is limited to standard Transformers integration files, which is insufficient to treat them as one PR." + "reason": "PenguinVL and DeepSeek-OCR-2 are unrelated model implementations with only generic transformer scaffolding in common." }, { "left": "pull_request:45075", "right": "pull_request:45401", "accept": false, - "reason": "These add unrelated models for OCR vs text-to-speech. They touch similar framework plumbing but do not fix or implement the same concrete feature or bug." + "reason": "DeepSeek-OCR-2 vs Voxtral TTS are distinct model additions touching different model logic and tests, so they are not the same underlying PR." } ] }, @@ -17025,12 +17122,12 @@ "soft_pairs": [ "pull_request:44979|pull_request:45363" ], - "prepared_review_unit_hash": "7302f7c67423962ccec9fef113101cdd411e5b09ed6f79626d8ff25f77b1b004", + "prepared_review_unit_hash": "d85351de58336264d2bf247c1b3e5f731d8fbb64090fccbfef64b4ac9a70f9ab", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4407, + "serialized_chars": 4408, "estimated_input_tokens": 1102, "estimated_eval_tokens": 2460 }, @@ -17038,7 +17135,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4407, + "serialized_chars": 4408, "estimated_input_tokens": 1102, "estimated_eval_tokens": 2460 }, @@ -17047,18 +17144,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are related feature work around fusion, but they target different layers and code paths: one introduces a generic Module Fusion API, the other wires KernelConfig-based n-to-1 kernel fusion into model loading and adds an example.", - "confidence": 0.87, + "summary": "These PRs are related in the broad area of module/kernel fusion, but they are not the same change: #44979 adds a new Module Fusion API scaffold, while #45363 implements n-to-1 kernel fusion through KernelConfig and model loading integrations. The overlap is thematic rather than a shared concrete code-path fix.", + "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR: 44979 is the foundational module-fusion API, while 45363 is a separate integration of KernelConfig-driven n-to-1 fusion built on top of that idea.", + "canonical_pr_reason": "If a single PR must represent the cluster, #45363 is the stronger canonical choice because it is the more complete, end-to-end user-facing implementation; #44979 is an enabling API foundation rather than the same change.", "best_issue_reason": null, - "best_pr_reason": "44979 is the broader foundational abstraction and the cleaner cluster representative; 45363 is a more concrete integration/example change rather than the core API surface.", + "best_pr_reason": "#45363 is the best representative PR in this cluster because it shows a concrete application of kernel fusion in the loading/integration path, whereas #44979 is a separate API feature scaffold with different scope.", "soft_edge_verdicts": [ { "left": "pull_request:44979", "right": "pull_request:45363", "accept": false, - "reason": "Related but not duplicates: 44979 adds a new generic Module Fusion API (`src/transformers/module_fusion.py` and tests), while 45363 implements a specific `KernelConfig`-based n-to-1 fusion path across integration/model-loading files. They address different abstractions and code paths." + "reason": "Do not merge: #44979 introduces a generic Module Fusion API, while #45363 wires KernelConfig-based n-to-1 kernel fusion into model loading. They are related infrastructure, but not the same underlying bug or change and would not plausibly be merged into one PR." } ] }, @@ -17088,7 +17185,7 @@ "soft_pairs": [ "pull_request:45086|pull_request:45317" ], - "prepared_review_unit_hash": "14470d6c689f3a4471d4ec01fd794d0a50221ba7eb0186a5054faee4d0eaf933", + "prepared_review_unit_hash": "cd787b1e8640f6eb1143dbadd6efc5aa58d4681e13ff077eee8b7caeeb5a2e33", "component_budget": { "node_count": 2, "item_count": 2, @@ -17110,18 +17207,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both pull requests target the same Mistral regex AttributeError in `_patch_mistral_regex` and make the same code fix in `tokenization_utils_tokenizers.py`; the later PR also adds a test, so it is the stronger canonical PR candidate.", - "confidence": 0.97, + "summary": "Both PRs patch the same Mistral regex pre-tokenizer AttributeError in the same file and target the same issue; PR 45317 looks like the fuller duplicate with added tests.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45317 is the more complete version of the same fix: it applies the same code change as 45086 and also adds test coverage, making it the better canonical PR.", + "canonical_pr_reason": "45317 is the more complete fix for the same AttributeError path, with the same code change plus tests, so it is the better canonical PR among the duplicates.", "best_issue_reason": null, - "best_pr_reason": "PR 45317 is the best representative because it fixes the same concrete bug as 45086 but includes additional validation via tests and has more discussion activity.", + "best_pr_reason": "45317 is the best representative because it directly fixes the bug and includes test coverage, whereas 45086 is the narrower earlier variant.", "soft_edge_verdicts": [ { "left": "pull_request:45086", "right": "pull_request:45317", "accept": true, - "reason": "Accept: both PRs address the same AttributeError in `_patch_mistral_regex` with the same underlying code-path fix in the same file, and 45317 is a superset with tests." + "reason": "Same underlying bug and same code-path fix in _patch_mistral_regex; 45317 is essentially a superset of 45086 with tests, so these are duplicate PRs." } ] }, @@ -17151,40 +17248,40 @@ "soft_pairs": [ "pull_request:45332|pull_request:45333" ], - "prepared_review_unit_hash": "81cd4f0525f766dac1bba01b28a0b2a2db1e228b0e84f248d286c1357244d986", + "prepared_review_unit_hash": "abc0f79201563ba925e1b2e47774ba47d45ec0e3899a871d1e742a7fbceab1b0", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5398, - "estimated_input_tokens": 1350, - "estimated_eval_tokens": 2956 + "serialized_chars": 5396, + "estimated_input_tokens": 1349, + "estimated_eval_tokens": 2954 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5398, - "estimated_input_tokens": 1350, - "estimated_eval_tokens": 2956 + "serialized_chars": 5396, + "estimated_input_tokens": 1349, + "estimated_eval_tokens": 2954 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open PRs add adjacent heterogeneous support: #45333 introduces per-layer config infrastructure, while #45332 adds model/cache handling on top of it. They are related, but not duplicates.", - "confidence": 0.87, + "summary": "These are related but distinct PRs for the same heterogeneous-modeling effort: one adds per-layer config plumbing, the other adds runtime/model integration on top of it. They share infrastructure files, but they do not look like the same concrete change.", + "confidence": 0.89, "canonical_issue_reason": null, - "canonical_pr_reason": "Choose #45333 as the canonical PR because it introduces the foundational per-layer heterogeneous config API and helper utilities that the other PR builds on.", + "canonical_pr_reason": "PR 45332 is the broader end-to-end feature PR and appears to build on the per-layer config plumbing, so it is the better representative of the cluster.", "best_issue_reason": null, - "best_pr_reason": "Choose #45332 as the best representative PR because it covers the broader end-to-end heterogeneous model and cache integration, making it the more complete user-facing change.", + "best_pr_reason": "PR 45332 is the strongest standalone representative because it covers the actual model/runtime support in addition to the heterogeneous config groundwork.", "soft_edge_verdicts": [ { "left": "pull_request:45332", "right": "pull_request:45333", "accept": false, - "reason": "Reject: these are related prerequisite/follow-up PRs, not the same underlying change. #45333 adds config infrastructure; #45332 adds modeling/cache support and broader integration work." + "reason": "Reject: 45333 is focused on per-layer configuration support, while 45332 is model/runtime integration (cache, masking, GPT-OSS/modeling changes). They share a feature area and helper files, but they are complementary stackable PRs rather than one duplicate concrete change." } ] }, @@ -17214,7 +17311,7 @@ "soft_pairs": [ "pull_request:45415|pull_request:45425" ], - "prepared_review_unit_hash": "10750f0dc6fa0cb26f7c1ccc490a0f34e97e22ebf74b2d58a691713e74fa429c", + "prepared_review_unit_hash": "af87c1bb4ddeb249b217b263f0353f6e560ed79df4c36a100637ce42eee9b30d", "component_budget": { "node_count": 2, "item_count": 2, @@ -17236,18 +17333,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open PRs in the same typing-cleanup area, but they are not the same underlying change. PR 45415 is a broad sweep adding type checking across many transformers modules and relaxing type-ignore warnings, while PR 45425 is a narrower typing helper update around `modeling_utils`/PEFT-related protocols and aliases.", - "confidence": 0.86, + "summary": "These are related only at a high level (typing cleanup), but they are not the same concrete change. PR 45415 is a broad sweep adding type checking across many transformer modules, while PR 45425 adds specific typing aliases/protocols for modeling_utils/peft and a check-types update. I would not cluster them as duplicates.", + "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45415 is the best representative of the cluster because it is the broader, more central typing/type-checking sweep affecting many files and build config.", + "canonical_pr_reason": "PR 45425 is the more focused, concrete typing change: it introduces reusable protocol/alias definitions and targets a small set of files, making it a better representative than the broad cross-module typing sweep in PR 45415.", "best_issue_reason": null, - "best_pr_reason": "PR 45415 is the strongest global representative: it covers the larger cross-cutting type-checking initiative, whereas PR 45425 is a narrower follow-up/helper change.", + "best_pr_reason": "PR 45425 is the best representative because it addresses a specific typing gap with clear scope, whereas PR 45415 is a wide-ranging refactor-like typing pass that is less suitable as a canonical match for this pair.", "soft_edge_verdicts": [ { "left": "pull_request:45415", "right": "pull_request:45425", "accept": false, - "reason": "Both are typing-related, but they address different scopes: 45415 is a broad repository-wide type-checking pass, while 45425 adds specific typing aliases/protocols for `modeling_utils`/PEFT. They are not the same concrete fix and would not naturally merge into one PR." + "reason": "Reject: both are typing-related, but they do not appear to fix the same underlying bug/change. 45415 is a broad type-checking sweep across many files; 45425 adds targeted typing protocols/aliases for modeling_utils and peft. Shared files alone are not enough to merge them as duplicates." } ] }, @@ -17277,7 +17374,7 @@ "soft_pairs": [ "pull_request:45470|pull_request:45487" ], - "prepared_review_unit_hash": "1661d1c946f4c4e43067e21000404becf0f43d59ec1853e950583fc551f1d05b", + "prepared_review_unit_hash": "bfa28de9a896c2584f7fbf945093b2da81f07d67c3e2fe69d5e69f6457123447", "component_budget": { "node_count": 2, "item_count": 2, @@ -17299,18 +17396,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two pull requests are unrelated: one only skips flash-attn composite-model tests for Sam3LiteText, while the other fixes a real model-parallel/device-placement bug across several text models. They should not be clustered as duplicates.", - "confidence": 0.96, + "summary": "These pull requests are not duplicates: one only skips flash-attn composite-model tests for Sam3LiteText, while the other fixes a real model-parallel/device bug in AltCLIP/ChineseCLIP and related text models.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45487 is the strongest canonical PR because it contains the substantive code fix (device placement for buffered token type ids and related model-parallel adjustments) across multiple affected models, unlike PR 45470 which only adds test skips.", + "canonical_pr_reason": "PR #45487 is the stronger representative change because it fixes a concrete runtime bug in model-parallel code paths, whereas PR #45470 is only a test skip for unsupported flash-attn coverage.", "best_issue_reason": null, - "best_pr_reason": "PR 45487 is the best representative artifact for this cluster since it addresses an actual implementation bug with concrete code changes and broader impact; PR 45470 is just a test skip and not a comparable fix.", + "best_pr_reason": "PR #45487 is the best fit for the cluster as a substantive code fix with clear user-facing impact; PR #45470 is a narrow test-only workaround and not the same underlying change.", "soft_edge_verdicts": [ { "left": "pull_request:45470", "right": "pull_request:45487", "accept": false, - "reason": "Different problems and change types: 45470 skips flash-attn tests for Sam3LiteText due to unsupported flash attention, while 45487 fixes a model-parallel/device mismatch in AltCLIP/ChineseCLIP and related models. They are not the same underlying bug or mergeable change." + "reason": "Different problems and code paths: 45470 skips flash-attn tests for Sam3LiteText due to unsupported flash attention, while 45487 fixes token_type_ids device handling and parallelism issues in AltCLIP/ChineseCLIP-related models." } ] }, diff --git a/analysis/current/manifest.json b/analysis/current/manifest.json index 7dec0a15a6d72a70c6b4458797f6fe7a3f193048..b6bc939124b282877cca7a84c16a6b0433e238d2 100644 --- a/analysis/current/manifest.json +++ b/analysis/current/manifest.json @@ -1,8 +1,8 @@ { - "analysis_id": "hybrid-model-20260421t060039z", + "analysis_id": "hybrid-model-20260421t120044z", "archived_artifacts": { - "hybrid": "snapshots/20260421T060039Z/analysis-runs/hybrid-model-20260421t060039z/analysis-report-hybrid.json", - "hybrid_reviews": "snapshots/20260421T060039Z/analysis-runs/hybrid-model-20260421t060039z/analysis-report-hybrid.llm-reviews.json" + "hybrid": "snapshots/20260421T120044Z/analysis-runs/hybrid-model-20260421t120044z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260421T120044Z/analysis-runs/hybrid-model-20260421t120044z/analysis-report-hybrid.llm-reviews.json" }, "artifacts": { "hybrid": "analysis/current/analysis-report-hybrid.json", @@ -10,9 +10,9 @@ }, "channel": "canonical", "model": null, - "published_at": "2026-04-21T06:06:36Z", + "published_at": "2026-04-21T12:17:10Z", "repo": "huggingface/transformers", "schema_version": 1, - "snapshot_id": "20260421T060039Z", + "snapshot_id": "20260421T120044Z", "variant": "hybrid" } diff --git a/snapshots/20260421T120044Z/analysis-runs/hybrid-model-20260421t120044z/analysis-report-hybrid.json b/snapshots/20260421T120044Z/analysis-runs/hybrid-model-20260421t120044z/analysis-report-hybrid.json new file mode 100644 index 0000000000000000000000000000000000000000..cf5e948f363c7c813c78a701d757cf569d48e6f5 --- /dev/null +++ b/snapshots/20260421T120044Z/analysis-runs/hybrid-model-20260421t120044z/analysis-report-hybrid.json @@ -0,0 +1,2104 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260421T120044Z", + "generated_at": "2026-04-21T12:17:08Z", + "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": 0.116, + "size_similarity": 0.17, + "file_overlap": 0.25, + "area_overlap": 0.005, + "patch_similarity": 0.034, + "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": 0.886, + "size_similarity": 0.877, + "file_overlap": 0.75, + "area_overlap": 0.996, + "patch_similarity": 0.912, + "shared_filenames": [ + "src/transformers/integrations/ggml.py", + "src/transformers/modeling_gguf_pytorch_utils.py", + "tests/quantization/ggml/test_ggml.py" + ], + "shared_file_areas": [ + { + "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-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." + } + ], + "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/20260421T120044Z/analysis-runs/hybrid-model-20260421t120044z/analysis-report-hybrid.llm-reviews.json b/snapshots/20260421T120044Z/analysis-runs/hybrid-model-20260421t120044z/analysis-report-hybrid.llm-reviews.json new file mode 100644 index 0000000000000000000000000000000000000000..9f291c056a5e7cdb281d146b74e95c81d77bf7be --- /dev/null +++ b/snapshots/20260421T120044Z/analysis-runs/hybrid-model-20260421t120044z/analysis-report-hybrid.llm-reviews.json @@ -0,0 +1,17422 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260421T120044Z", + "generated_at": "2026-04-21T12:17:08Z", + "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": "68c474652dbba90cfb989b3893065d0919301d3d16e3f077997780fd7862a7f5", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 2478, + "estimated_input_tokens": 620, + "estimated_eval_tokens": 1496 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 2478, + "estimated_input_tokens": 620, + "estimated_eval_tokens": 1496 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The three issues describe different problems: Tapas tokenization mismatch, an auth/error-message problem for private models, and a generic failure to load bert-base-uncased. They do not look like duplicates or the same underlying bug.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "No issue is a good canonical anchor for this cluster because the topics are unrelated; if one must be chosen, #16618 is the broadest user-facing failure, but it does not represent the others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:13244", + "right": "issue:16618", + "accept": false, + "reason": "Different symptoms and likely different code paths: Tapas tokenization parity vs a model-loading failure for bert-base-uncased." + }, + { + "left": "issue:15980", + "right": "issue:16618", + "accept": false, + "reason": "One is about the error message/authorization path for private models, the other is a failure loading a public model; not the same bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 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-15354-364", + "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:35545|issue:36010", + "issue:43638|issue:44805", + "issue:43596|issue:44805", + "issue:44805|issue:45161", + "issue:44805|issue:44918", + "issue:44661|issue:44805", + "issue:34634|issue:35545", + "issue:30277|issue:43638", + "issue:43596|issue:43638", + "issue:43975|issue:44457", + "issue:43039|issue:45161", + "issue:44292|issue:44457", + "issue:34928|issue:44805", + "issue:24643|issue:30277", + "issue: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:44918|issue:45161", + "issue:34928|issue:41720", + "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:43866|issue:44863", + "issue:43039|issue:44560", + "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": "15435980cf472e38716ef9ec097d81b497327fbc3b7a3170b1d298b85e91960a", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 63, + "serialized_chars": 46468, + "estimated_input_tokens": 11617, + "estimated_eval_tokens": 23490 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly a heterogeneous cluster of unrelated model/runtime issues. The only clear duplicate families are the DeepSpeed ZeRO-3 index-out-of-bounds load failures and the near-identical mask-shape IndexError reports; the docs issues and the rest look like separate bugs.", + "confidence": 0.71, + "canonical_issue_reason": "Issue 43596 is the cleanest representative of the ZeRO-3 index-0-out-of-bounds load failure family: it describes the generic error pattern and matches the same root symptom seen in the close variants.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43596 is the strongest issue candidate because it most directly captures the shared ZeRO-3 load-time failure pattern and can subsume the closely related Bert/Gemma3 variants.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44322", + "right": "issue:45464", + "accept": false, + "reason": "Different Qwen3.5 failures: one is a chat crash, the other is a streaming inference API failure." + }, + { + "left": "issue:43716", + "right": "issue:45237", + "accept": false, + "reason": "Both involve model/runtime problems, but one is a dtype mismatch and the other is an AMD GPU execution failure for GPT-OSS-20B; not the same bug." + }, + { + "left": "issue:43828", + "right": "issue:45237", + "accept": false, + "reason": "Autocast dtype mismatch on Phi-tiny-MoE is unrelated to GPT-OSS-20B failing on AMD GPUs." + }, + { + "left": "issue:45237", + "right": "issue:45507", + "accept": false, + "reason": "Different models and failure modes: AMD GPU support issue vs invalid method call in GraniteMoEHybrid." + }, + { + "left": "issue:44292", + "right": "issue:45237", + "accept": false, + "reason": "Both are model runtime errors, but Qwen-3-8B-NVFP4 is a different code path from GPT-OSS-20B on AMD GPUs." + }, + { + "left": "issue:30277", + "right": "issue:35545", + "accept": false, + "reason": "DeepSpeed collective mismatch for Jamba-v01 is unrelated to ModernBERT ONNX export failure." + }, + { + "left": "issue:35545", + "right": "issue:42915", + "accept": false, + "reason": "ModernBERT ONNX export and Qwen3Moe FineGrainedFP8 failures are different model/export paths." + }, + { + "left": "issue:35545", + "right": "issue:41720", + "accept": false, + "reason": "ONNX export error and Qwen3 auto device mapping cuda assert are distinct problems." + }, + { + "left": "issue:44509", + "right": "issue:44512", + "accept": false, + "reason": "Both are stale v5 docs issues, but they concern different removed features (pipeline tasks vs run command)." + }, + { + "left": "issue:35545", + "right": "issue:36010", + "accept": false, + "reason": "ONNX export failure is unrelated to the GenerationMixin import error." + }, + { + "left": "issue:43638", + "right": "issue:44805", + "accept": false, + "reason": "Both are IndexErrors, but one is DeepSpeed ZeRO-3/BERT load-time and the other is a mask-shape mismatch; different bugs." + }, + { + "left": "issue:43596", + "right": "issue:44805", + "accept": false, + "reason": "Different failure signatures and code paths: ZeRO-3 index-out-of-bounds vs mask-shape mismatch." + }, + { + "left": "issue:44805", + "right": "issue:45161", + "accept": false, + "reason": "Mask-shape IndexError and GPT-OSS MoE tensor-parallel issue are unrelated." + }, + { + "left": "issue:44805", + "right": "issue:44918", + "accept": false, + "reason": "Mask/tensor shape mismatch is not the same as Qwen3.5 embedding unpacking under TRL SFT." + }, + { + "left": "issue:44661", + "right": "issue:44805", + "accept": false, + "reason": "Tokenizer mapping/add-new-model-like issue is unrelated to a mask-shape IndexError." + }, + { + "left": "issue:34634", + "right": "issue:35545", + "accept": false, + "reason": "BarkProcessor voice_preset issue is unrelated to ModernBERT ONNX export." + }, + { + "left": "issue:30277", + "right": "issue:43638", + "accept": false, + "reason": "DeepSpeed collective mismatch for Jamba-v01 is unrelated to ZeRO-3 index-out-of-bounds in BertModel." + }, + { + "left": "issue:43596", + "right": "issue:43638", + "accept": true, + "reason": "Same DeepSpeed ZeRO-3 index-0-out-of-bounds load-time failure, just reported against different Bert model variants." + }, + { + "left": "issue:43975", + "right": "issue:44457", + "accept": false, + "reason": "Detokenization bug and LoRA merge/save mismatch are different problems." + }, + { + "left": "issue:43039", + "right": "issue:45161", + "accept": false, + "reason": "Liger Kernel calling cross_entropy and GPT-OSS MoE tensor-parallel support are unrelated." + }, + { + "left": "issue:44292", + "right": "issue:44457", + "accept": false, + "reason": "NVFP4 model runtime error and LoRA merge/save inconsistency are different issues." + }, + { + "left": "issue:34928", + "right": "issue:44805", + "accept": false, + "reason": "Activation-checkpointing/FSDP tensor-size mismatch is not the same as the mask-shape IndexError." + }, + { + "left": "issue:24643", + "right": "issue:30277", + "accept": false, + "reason": "Training with DeepSpeed weight dimensionality error and Jamba collectives mismatch are different DeepSpeed failures." + }, + { + "left": "issue:43975", + "right": "issue:45237", + "accept": false, + "reason": "Detokenization bug and AMD GPU model loading failure are unrelated." + }, + { + "left": "issue:43638", + "right": "issue:45161", + "accept": false, + "reason": "ZeRO-3 BERT load failure and GPT-OSS MoE tensor-parallel issue are different bugs." + }, + { + "left": "issue:34928", + "right": "issue:36331", + "accept": false, + "reason": "FSDP activation-checkpointing tensor mismatch is unrelated to a CustomTrainer method signature error." + }, + { + "left": "issue:43596", + "right": "issue:45161", + "accept": false, + "reason": "DeepSpeed ZeRO-3 load failure is unrelated to GPT-OSS MoE tensor-parallel behavior." + }, + { + "left": "issue:43716", + "right": "issue:45507", + "accept": false, + "reason": "Mistral dtype mismatch and GraniteMoEHybrid invalid method call do not share the same bug." + }, + { + "left": "issue:39290", + "right": "issue:45081", + "accept": false, + "reason": "Gemma3/vLLM sliding-window attribute error is unrelated to the tokenizer regex patch crash." + }, + { + "left": "issue:43828", + "right": "issue:45161", + "accept": false, + "reason": "Autocast dtype mismatch in Phi-tiny-MoE is different from GPT-OSS MoE TP support." + }, + { + "left": "issue:44918", + "right": "issue:45161", + "accept": false, + "reason": "Qwen3.5 embedding unpacking under TRL and GPT-OSS MoE tensor-parallel are different issues." + }, + { + "left": "issue:43039", + "right": "issue:43828", + "accept": false, + "reason": "Cross-entropy being called under Liger Kernel is unrelated to Phi-tiny-MoE autocast dtype mismatch." + }, + { + "left": "issue:30277", + "right": "issue:36331", + "accept": false, + "reason": "Collective mismatch in DeepSpeed and CustomTrainer.compute_loss signature change are unrelated." + }, + { + "left": "issue:43039", + "right": "issue:44805", + "accept": false, + "reason": "Liger Kernel cross-entropy call and mask-shape IndexError are different bugs." + }, + { + "left": "issue:34928", + "right": "issue:36010", + "accept": false, + "reason": "FSDP/activation-checkpointing mismatch is unrelated to the GenerationMixin import error." + }, + { + "left": "issue:41762", + "right": "issue:43596", + "accept": true, + "reason": "Same underlying DeepSpeed ZeRO-3 index-out-of-bounds load failure; only the affected model differs." + }, + { + "left": "issue:43901", + "right": "issue:44512", + "accept": false, + "reason": "TextClassificationPipeline docs mention and transformers run-command docs mention are separate stale-doc issues." + }, + { + "left": "issue:39290", + "right": "issue:41720", + "accept": false, + "reason": "Gemma3/vLLM sliding-window error is unrelated to Qwen3 auto device mapping cuda assert." + }, + { + "left": "issue:43866", + "right": "issue:44863", + "accept": false, + "reason": "Corrupted Ovis2 checkpoint and NemotronH load failure are different problems." + }, + { + "left": "issue:43039", + "right": "issue:44560", + "accept": false, + "reason": "Liger Kernel cross-entropy issue and Qwen3-vl StopIteration video error are unrelated." + }, + { + "left": "issue:43854", + "right": "issue:43866", + "accept": false, + "reason": "GLM-4.7-Flash unit-test load failure and Ovis2 checkpoint corruption are different issues." + }, + { + "left": "issue:43425", + "right": "issue:45070", + "accept": false, + "reason": "Torch compatibility and PretrainedConfig pydantic-field breakage are unrelated." + }, + { + "left": "issue:39290", + "right": "issue:43531", + "accept": false, + "reason": "Gemma3 sliding-window attribute error and Qwen3-MoE sliding_window issue are different model families and likely different fixes." + }, + { + "left": "issue:41093", + "right": "issue:43638", + "accept": false, + "reason": "Although both are IndexErrors, the mask/tensor shape mismatch is a different failure mode from the ZeRO-3 BERT index-out-of-bounds bug." + }, + { + "left": "issue:43404", + "right": "issue:43866", + "accept": false, + "reason": "Mistral3 tied-weight bug and Ovis2 checkpoint corruption are not the same issue." + }, + { + "left": "issue:45362", + "right": "issue:45464", + "accept": false, + "reason": "Qwen3.5 chat crash and Qwen3.5 streaming inference failure are separate symptoms and code paths." + }, + { + "left": "issue:43792", + "right": "issue:45070", + "accept": false, + "reason": "Whisper loading failure is unrelated to the pydantic PretrainedConfig regression." + }, + { + "left": "issue:43541", + "right": "issue:43828", + "accept": false, + "reason": "Mixtral grouped_mm tracing error and Phi-tiny-MoE autocast dtype mismatch are distinct runtime issues." + }, + { + "left": "issue:36010", + "right": "issue:39290", + "accept": false, + "reason": "GenerationMixin import failure is unrelated to Gemma3/vLLM sliding-window attribute error." + }, + { + "left": "issue:44841", + "right": "issue:45084", + "accept": false, + "reason": "Voxtral processor failure and template-compilation TypeError are different bugs." + }, + { + "left": "issue:41093", + "right": "issue:41762", + "accept": false, + "reason": "The mask/tensor shape mismatch and the ZeRO-3 index-out-of-bounds load failure are different error classes." + }, + { + "left": "issue:43296", + "right": "issue:43366", + "accept": false, + "reason": "PaddleOCR-VL load failure and GGUF gpt-oss support request are not the same issue." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The cluster summary is grounded in the packet, and the soft-edge calls are conservative overall. The two accepted ZeRO-3/BERT-family pairs are well-supported by the titles, while the rest of the rejects stay appropriately cautious about model/failure-mode differences." + }, + "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-15354-364", + "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:43644", + "issue:43645", + "issue:43742", + "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:44387|issue:45005", + "issue:43572|issue:45070", + "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:43054|issue:43994", + "issue:39290|issue:43742", + "issue:41093|issue:43596", + "issue:43541|issue:44589", + "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:44360|issue:44512", + "issue:42915|issue:45399", + "issue:44918|issue:45507", + "issue:43975|issue:44292", + "issue:44492|issue:44512", + "issue:43950|issue:44534", + "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:36331|issue:41762", + "issue:44560|issue:45507" + ], + "prepared_review_unit_hash": "84d7bb72d647c1bb8ca560704f137bc45ceb248477607a7fe242d2d2471003de", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37560, + "estimated_input_tokens": 9390, + "estimated_eval_tokens": 19036 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly heterogeneous, with one clear duplicate theme: the Transformers 5.x regression that corrupts non-persistent buffers during model loading. The rest of the issues cover unrelated model/config/runtime bugs.", + "confidence": 0.89, + "canonical_issue_reason": "Issue 43950 is the clearest and most specific report of the shared regression: `from_pretrained()` corrupting `register_buffer(persistent=False)` state in Transformers 5.x.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43950 best represents the duplicate pair because it names the concrete loading path and the exact non-persistent buffer corruption behavior.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44387", + "right": "issue:45005", + "accept": false, + "reason": "Both mention v5/quantization or tied weights, but they describe different failure modes and different code paths; not the same bug." + }, + { + "left": "issue:43572", + "right": "issue:45070", + "accept": false, + "reason": "One is a missing config field (`pad_token_idx`) and the other is a `PretrainedConfig`/pydantic field regression; too different to be a duplicate." + }, + { + "left": "issue:43541", + "right": "issue:45161", + "accept": false, + "reason": "Grouped_mm torch-dynamo tracing errors and GPT-OSS tensor-parallel issues are distinct runtime problems." + }, + { + "left": "issue:43828", + "right": "issue:45464", + "accept": false, + "reason": "Autocast dtype mismatch in Phi-tiny-MoE is unrelated to Qwen3.5 streaming inference failures." + }, + { + "left": "issue:44291", + "right": "issue:45070", + "accept": false, + "reason": "Unexpected `_is_hf_initialized` during `init_empty_weights` is a different initialization bug than the pydantic `PretrainedConfig` issue." + }, + { + "left": "issue:43366", + "right": "issue:45464", + "accept": false, + "reason": "GGUF architecture support for gpt-oss is unrelated to Qwen3.5 chat/completions streaming failures." + }, + { + "left": "issue:43531", + "right": "issue:45070", + "accept": false, + "reason": "A Qwen3-MoE sliding-window issue is not the same as a `PretrainedConfig`/pydantic field regression." + }, + { + "left": "issue:44661", + "right": "issue:45464", + "accept": false, + "reason": "Tokenizer mapping failure when adding a new model is unrelated to Qwen3.5 streaming inference." + }, + { + "left": "issue:43054", + "right": "issue:43994", + "accept": false, + "reason": "Both involve SigLIP2, but one reports poor text embeddings while the other reports bad model/pipeline outputs; not clearly the same underlying bug." + }, + { + "left": "issue:39290", + "right": "issue:43742", + "accept": false, + "reason": "Gemma3 sliding-window/vLLM incompatibility and MobileLLM key loading errors are different model-specific problems." + }, + { + "left": "issue:41093", + "right": "issue:43596", + "accept": false, + "reason": "These are both IndexErrors, but the mask-shape mismatch and zero3 empty-tensor initialization failure are different root causes." + }, + { + "left": "issue:43541", + "right": "issue:44589", + "accept": false, + "reason": "Torch-dynamo grouped_mm tracing on Mixtral is unrelated to a float8 storage lookup error." + }, + { + "left": "issue:42915", + "right": "issue:43541", + "accept": false, + "reason": "FineGrainedFP8 Qwen3-MoE failure and Mixtral grouped_mm tracing failure are different model/runtime issues." + }, + { + "left": "issue:40990", + "right": "issue:44960", + "accept": false, + "reason": "High perplexity on GPT-OSS and a GLM5 issue are not the same bug." + }, + { + "left": "issue:39186", + "right": "issue:44918", + "accept": false, + "reason": "FSDP `'weight' must be 2-D` and TRL SFT input-unpacking failures are unrelated." + }, + { + "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; not enough to treat as duplicates." + }, + { + "left": "issue:43645", + "right": "issue:45399", + "accept": false, + "reason": "Custom-model initialization in notebooks and flash-attn fallback gating are unrelated." + }, + { + "left": "issue:44530", + "right": "issue:45005", + "accept": false, + "reason": "PagedAttentionCache linear_attention crash is unrelated to tied-weights regressions." + }, + { + "left": "issue:44863", + "right": "issue:45005", + "accept": false, + "reason": "NemotronH checkpoint loading failure is unrelated to tied weights on translation models." + }, + { + "left": "issue:43873", + "right": "issue:45005", + "accept": false, + "reason": "Quantization offloading behavior is a different subsystem from tied-weight handling." + }, + { + "left": "issue:44360", + "right": "issue:44512", + "accept": false, + "reason": "A DSA indexer ReLU discussion and a docs typo are unrelated." + }, + { + "left": "issue:42915", + "right": "issue:45399", + "accept": false, + "reason": "Qwen3-MoE FP8 failure and flash-attn fallback blocking are different problems." + }, + { + "left": "issue:44918", + "right": "issue:45507", + "accept": false, + "reason": "TRl SFT unpacking failure and GraniteMoEHybrid invalid-method calls are unrelated." + }, + { + "left": "issue:43975", + "right": "issue:44292", + "accept": false, + "reason": "Detokenization regression and Qwen-3-8B-NVFP4 runtime error are unrelated." + }, + { + "left": "issue:44492", + "right": "issue:44512", + "accept": false, + "reason": "A cache-strategy typo and removed CLI docs mention are unrelated documentation issues." + }, + { + "left": "issue:43950", + "right": "issue:44534", + "accept": true, + "reason": "Both report the same Transformers v5 regression where non-persistent buffers are mishandled/corrupted during loading." + }, + { + "left": "issue:43644", + "right": "issue:43950", + "accept": true, + "reason": "Same bug: Transformers v5 corrupts non-persistent buffers; one report is the broader symptom, the other is the concrete `from_pretrained()` corruption path." + }, + { + "left": "issue:44292", + "right": "issue:45507", + "accept": false, + "reason": "Qwen-3-8B-NVFP4 runtime errors and GraniteMoEHybrid invalid-method calls are unrelated." + }, + { + "left": "issue:43828", + "right": "issue:44928", + "accept": false, + "reason": "Different Qwen3.5-related failures: autocast dtype mismatch versus 3D position_ids / SDPA math fallback causing NaNs." + }, + { + "left": "issue:39692", + "right": "issue:43054", + "accept": false, + "reason": "SigLIP2 documentation example errors and SigLIP2 text-embedding quality regression are related only by model family, not the same concrete bug." + }, + { + "left": "issue:43866", + "right": "issue:44403", + "accept": false, + "reason": "Corrupted Ovis2 checkpoint and loading noise are not the same issue." + }, + { + "left": "issue:41553", + "right": "issue:43883", + "accept": false, + "reason": "Bad AutoTokenizer error messaging for Voxtral is unrelated to missing `all_tied_weights_keys` on Molmo." + }, + { + "left": "issue:36331", + "right": "issue:41762", + "accept": false, + "reason": "A custom trainer signature mismatch and a DeepSpeed zero3 index error are unrelated." + }, + { + "left": "issue:44560", + "right": "issue:45507", + "accept": false, + "reason": "Qwen3-vl embedding video StopIteration and GraniteMoEHybrid invalid-method calls are different runtime failures." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: the only clearly cohesive theme is the Transformers 5.x non-persistent-buffer regression, and the accepted duplicates (43644/43950/44534) are supported by the issue titles. The other soft-edge rejections stay conservative and do not overclaim duplication 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": 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-15354-364", + "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:38175|issue:43054", + "issue:41553|issue:43881", + "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:43782|issue:44960", + "issue:44403|issue:44488", + "issue:42491|issue:44368", + "issue:43975|issue:44451", + "issue:41762|issue:42915", + "issue:45020|issue:45356", + "issue:44661|issue:45020", + "issue:43054|issue:43278", + "issue:43824|issue:43872", + "issue:44991|issue:45399", + "issue:43827|issue:44509", + "issue:43792|issue:44220", + "issue:43404|issue:43950", + "issue:45081|issue:45399", + "issue:43257|issue:43866", + "issue:42915|issue:45020", + "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:43257|issue:45305", + "issue:41553|issue:44488", + "issue:42915|issue:43872", + "issue:36331|issue:41093" + ], + "prepared_review_unit_hash": "b32c67e71cd42b377a5acf5b391f40c620114610b9237e317ad6b682b8cbb1ff", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 36, + "serialized_chars": 38576, + "estimated_input_tokens": 9644, + "estimated_eval_tokens": 19544 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Loose cluster of mostly model-loading, tokenizer, and training-regression issues across different models/subsystems. Issue 45020 is still the broadest issue anchor, but the soft pairs are mostly title-level similarities and should be rejected without stronger code-path evidence.", + "confidence": 0.66, + "canonical_issue_reason": "issue:45020 is the widest umbrella here: it describes a general recent-version regression breaking `remote_code` models, which best matches the cluster\u2019s repeated load/config failure pattern.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45020 is the most reusable anchor because it is broad, open, and centered on the shared regression theme rather than a single model-specific symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43872", + "right": "issue:44589", + "accept": false, + "reason": "Different failure modes (`_is_hf_initialized` TypeError vs Float8 storage lookup error); title similarity is not enough to call them the same bug." + }, + { + "left": "issue:38175", + "right": "issue:43054", + "accept": false, + "reason": "Both mention SigLIP/SigLIP2, but one is a zero-probability symptom and the other is degraded text embedding quality; no concrete shared code-path evidence." + }, + { + "left": "issue:41553", + "right": "issue:43881", + "accept": false, + "reason": "Different model-loading problems with different symptoms and no indication of the same underlying defect." + }, + { + "left": "issue:42947", + "right": "issue:44387", + "accept": false, + "reason": "Both are training regressions, but one is LoRA checkpointing behavior and the other is int4/OOM memory usage; too different to merge." + }, + { + "left": "issue:43054", + "right": "issue:43493", + "accept": false, + "reason": "Strong title similarity around SigLIP2, but the packet doesn\u2019t show enough to prove the same bug beyond broad topic overlap." + }, + { + "left": "issue:43881", + "right": "issue:43883", + "accept": false, + "reason": "Both are load failures, but they affect different models and different code paths." + }, + { + "left": "issue:44403", + "right": "issue:44451", + "accept": false, + "reason": "One is about noisy transformer loading output, the other is a specific model load failure; not the same issue." + }, + { + "left": "issue:42947", + "right": "issue:43856", + "accept": false, + "reason": "Both involve Qwen3 training, but one is gradient-checkpointing/LoRA behavior and the other is memory efficiency in MoE training." + }, + { + "left": "issue:43975", + "right": "issue:45520", + "accept": false, + "reason": "Tokenizer detokenization regression and Python 3.13 import-utils KeyError are unrelated subsystems." + }, + { + "left": "issue:43643", + "right": "issue:45020", + "accept": false, + "reason": "Related to remote-code loading, but one is missing fields in `AutoConfig` output and the other is a broader breakage report; not enough to merge." + }, + { + "left": "issue:43782", + "right": "issue:44960", + "accept": false, + "reason": "Different models and different failures; no shared concrete bug is visible." + }, + { + "left": "issue:44403", + "right": "issue:44488", + "accept": false, + "reason": "Both are loading complaints, but they describe separate model-specific failures." + }, + { + "left": "issue:42491", + "right": "issue:44368", + "accept": false, + "reason": "A LoRA compatibility break and a tie-word-embeddings warning are not the same defect." + }, + { + "left": "issue:43975", + "right": "issue:44451", + "accept": false, + "reason": "Tokenizer/detokenization regression versus inability to load a specific model; different bugs." + }, + { + "left": "issue:41762", + "right": "issue:42915", + "accept": false, + "reason": "Both touch distributed loading, but the model and failure modes differ too much." + }, + { + "left": "issue:45020", + "right": "issue:45356", + "accept": false, + "reason": "Both are regressions in newer versions, but one is remote-code loading and the other is tokenizer/codec handling; different code paths." + }, + { + "left": "issue:44661", + "right": "issue:45020", + "accept": false, + "reason": "Different loading/registration failures; the shared theme is too broad to treat as one bug." + }, + { + "left": "issue:43054", + "right": "issue:43278", + "accept": false, + "reason": "SigLIP2 embedding quality vs training/evaluate dtype mismatch are distinct symptoms." + }, + { + "left": "issue:43824", + "right": "issue:43872", + "accept": false, + "reason": "Missing import and bitsandbytes TypeError are different errors with no shared underlying fix shown." + }, + { + "left": "issue:44991", + "right": "issue:45399", + "accept": false, + "reason": "Tokenizer load failure and flash-attn fallback gating are unrelated." + }, + { + "left": "issue:43827", + "right": "issue:44509", + "accept": false, + "reason": "Both are docs-related and mention removed pipelines, but the packet only gives title-level overlap; that is not enough to confidently merge." + }, + { + "left": "issue:43792", + "right": "issue:44220", + "accept": false, + "reason": "The Whisper audio frontend association is plausible, but the visible evidence is still too thin to assert the same underlying bug." + }, + { + "left": "issue:43404", + "right": "issue:43950", + "accept": false, + "reason": "Tied lm_head weights and non-persistent buffer corruption are different model-state issues." + }, + { + "left": "issue:45081", + "right": "issue:45399", + "accept": false, + "reason": "Tokenizer regex patch crash and flash-attn fallback checks are unrelated." + }, + { + "left": "issue:43257", + "right": "issue:43866", + "accept": false, + "reason": "Different model families and different load failures; no shared code-path bug is visible." + }, + { + "left": "issue:42915", + "right": "issue:45305", + "accept": false, + "reason": "Both involve DeepSpeed training behavior, but they describe different regressions and should not be merged." + }, + { + "left": "issue:43742", + "right": "issue:45399", + "accept": false, + "reason": "A specific model-loading key error and a flash-attn fallback issue are not the same bug." + }, + { + "left": "issue:43854", + "right": "issue:44863", + "accept": false, + "reason": "Different model-specific load failures; no evidence they come from one underlying defect." + }, + { + "left": "issue:41553", + "right": "issue:44991", + "accept": false, + "reason": "One is a bad error message for tokenizer loading, the other is a tokenizer load regression; similar domain, different bug." + }, + { + "left": "issue:43940", + "right": "issue:45313", + "accept": false, + "reason": "Both are Qwen3/ZeRO-3-related, but the weight-loading failures are not shown to be the same code-path problem." + }, + { + "left": "issue:43872", + "right": "issue:44291", + "accept": false, + "reason": "Both involve unexpected argument/init-empty-weights style loading errors, but the packet doesn\u2019t show enough to call them the same bug." + }, + { + "left": "issue:43854", + "right": "issue:45305", + "accept": false, + "reason": "Model load failure versus gradient averaging under DeepSpeed are unrelated." + }, + { + "left": "issue:43257", + "right": "issue:45305", + "accept": false, + "reason": "Different Qwen3 regressions affecting different parts of training/loading." + }, + { + "left": "issue:41553", + "right": "issue:44488", + "accept": false, + "reason": "Different model-loading issues with no clear shared defect." + }, + { + "left": "issue:36331", + "right": "issue:41093", + "accept": false, + "reason": "Trainer API signature mismatch and tensor mask shape mismatch are completely different failures." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The summary is mostly grounded, but the soft-edge handling is too assertive for this packet. The analyst marked several pairs as clear duplicates/mergeable bugfixes even though the packet\u2019s soft-pair labels are negative and the visible evidence is only title-level similarity. In particular, the accepted pairs `43054\u219443493`, `43827\u219444509`, and `43792\u219444220` are stated too strongly as the same bug without enough support from the packet. The conservative move here would be to keep these as tentative or reject them. Because of those over-accepts, I can\u2019t accept the evaluation." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, + "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-15354-364", + "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:45313", + "issue:45325", + "issue:45357", + "issue:45538" + ], + "soft_pairs": [ + "issue:41093|issue:44805", + "issue:44479|issue:44560", + "issue:44530|issue:44863", + "issue:43872|issue:45538", + "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:44560|issue:44918", + "issue:43366|issue:44589", + "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:44877|issue:45030", + "issue:43596|issue:44661", + "issue:43644|issue:44534", + "issue:43317|issue:44756", + "issue:38175|issue:39692", + "issue:45313|issue:45357", + "issue:42491|issue:43720", + "issue:44960|issue:44977", + "issue:44451|issue:44488", + "issue:43708|issue:44075" + ], + "prepared_review_unit_hash": "b23da9691ecaec6332738cccfc81f99ccd6b6bbed5e41de592d3d6fbad79fae6", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36647, + "estimated_input_tokens": 9162, + "estimated_eval_tokens": 18580 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly well-triaged, but the ScandiBERT/sleng-bert pair is too weak to merge: both are BERT-style model loading failures, yet the packet does not show enough evidence that they share the same concrete regression. That pair should be rejected, while the rest of the conservative duplicate decisions remain reasonable.", + "confidence": 0.76, + "canonical_issue_reason": "issue:43644 is still the cleanest canonical issue in this packet: it is a specific Transformers v5 non-persistent-buffer regression and has a near-verbatim duplicate in issue:44534.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43644 is the strongest representative issue because it describes a concrete, regression-style bug with clear scope and a close duplicate that matches the same failure mode.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41093", + "right": "issue:44805", + "accept": false, + "reason": "Both are mask/tensor shape mismatch errors, but the packet does not show enough detail to prove they are the same underlying bug." + }, + { + "left": "issue:44479", + "right": "issue:44560", + "accept": false, + "reason": "Both touch Qwen video-related paths, but the failures are different enough that they are not clearly the same bug." + }, + { + "left": "issue:44530", + "right": "issue:44863", + "accept": false, + "reason": "Different models and different failure modes: PagedAttentionCache crash versus NemotronH checkpoint loading." + }, + { + "left": "issue:43872", + "right": "issue:45538", + "accept": false, + "reason": "Unrelated issues: bitsandbytes Int8Params initialization versus CLIPTokenizer max length behavior." + }, + { + "left": "issue:43366", + "right": "issue:43828", + "accept": false, + "reason": "Different code paths and symptoms: GGUF/gpt-oss support versus dtype mismatch under autocast." + }, + { + "left": "issue:24643", + "right": "issue:30064", + "accept": false, + "reason": "DeepSpeed training weight-shape error is unrelated to image processor segmentation-map handling." + }, + { + "left": "issue:43366", + "right": "issue:43531", + "accept": false, + "reason": "No concrete shared bug: GGUF model support is unrelated to the sliding_window Qwen3-MoE issue." + }, + { + "left": "issue:43550", + "right": "issue:44910", + "accept": false, + "reason": "Different models and different failures: Bamba torch.compile/SDPA versus Qwen3.5 flash-attention memory access." + }, + { + "left": "issue:43645", + "right": "issue:43646", + "accept": true, + "reason": "These are the same Transformers 5.0.0 custom-model initialization regression, with one being a notebook-specific manifestation." + }, + { + "left": "issue:43950", + "right": "issue:45325", + "accept": false, + "reason": "Non-persistent buffer corruption and Qwen2.5-VL rope-index scaling are different bugs." + }, + { + "left": "issue:44560", + "right": "issue:44918", + "accept": false, + "reason": "Both mention Qwen video or embedding behavior, but the concrete failures and affected paths differ." + }, + { + "left": "issue:43366", + "right": "issue:44589", + "accept": false, + "reason": "No shared underlying bug; gpt-oss support is unrelated to the missing Float8 storage type." + }, + { + "left": "issue:43824", + "right": "issue:43957", + "accept": false, + "reason": "Different import/load problems involving different models and mechanisms." + }, + { + "left": "issue:42915", + "right": "issue:43957", + "accept": false, + "reason": "Qwen3Moe FP8 config failure is unrelated to meta-device model loading regressions." + }, + { + "left": "issue:42617", + "right": "issue:43366", + "accept": false, + "reason": "3d_parallel.py runtime failure does not match GGUF/gpt-oss support." + }, + { + "left": "issue:38175", + "right": "issue:43994", + "accept": true, + "reason": "Same SigLIP2 bad-output symptom for the same model family; the later report is a near-duplicate manifestation." + }, + { + "left": "issue:43646", + "right": "issue:43950", + "accept": false, + "reason": "Custom model initialization regression is not the same as non-persistent buffer corruption." + }, + { + "left": "issue:44560", + "right": "issue:44805", + "accept": false, + "reason": "Different errors and domains: Qwen3-vl video StopIteration versus generic mask/tensor shape mismatch." + }, + { + "left": "issue:36010", + "right": "issue:42915", + "accept": false, + "reason": "GenerationMixin ImportError is unrelated to the Qwen3Moe FP8 loading issue." + }, + { + "left": "issue:43638", + "right": "issue:44661", + "accept": false, + "reason": "Zero3/BertModel index error is unrelated to the TOKENIZER_MAPPING_NAMES limitation." + }, + { + "left": "issue:44877", + "right": "issue:45030", + "accept": false, + "reason": "Both are config-related, but they are different model/config regressions." + }, + { + "left": "issue:43596", + "right": "issue:44661", + "accept": false, + "reason": "Different failures: DeepSpeed ZeRO-3 empty tensor index error versus tokenizer-mapping limitation." + }, + { + "left": "issue:43644", + "right": "issue:44534", + "accept": true, + "reason": "Same Transformers v5 non-persistent-buffer regression; the titles are effectively describing the same bug." + }, + { + "left": "issue:43317", + "right": "issue:44756", + "accept": false, + "reason": "device_map/offload loading failure is unrelated to the Strix Halo mmap/OOM request." + }, + { + "left": "issue:38175", + "right": "issue:39692", + "accept": false, + "reason": "Related model family, but not the same concrete bug: bad probabilities versus docs example mismatch/quantization failure." + }, + { + "left": "issue:45313", + "right": "issue:45357", + "accept": false, + "reason": "Different Qwen3.5 issues: ZeRO-3 weight loading versus incorrect visual encoder save_pretrained keys." + }, + { + "left": "issue:42491", + "right": "issue:43720", + "accept": false, + "reason": "LoRA/Qwen3_moe compatibility across versions is unrelated to BitNet packed-weight loading." + }, + { + "left": "issue:44960", + "right": "issue:44977", + "accept": false, + "reason": "GLM5 and Qwen3.5 flash-attention problems are unrelated." + }, + { + "left": "issue:44451", + "right": "issue:44488", + "accept": false, + "reason": "The packet only shows two different BERT-style checkpoints that fail to load; that suggests a possible shared loader regression, but not enough evidence to confidently say it is the same bug." + }, + { + "left": "issue:43708", + "right": "issue:44075", + "accept": false, + "reason": "Trainer max_steps resume bug is unrelated to SGD argument handling." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "Most verdicts are conservative and grounded in the packet, but the accepted pair issue:44451 \u2194 issue:44488 is too strong from the available evidence. The titles only show two different BERT-style checkpoints that both fail to load; that suggests a possible shared loading regression, but not enough to confidently mark them as the same bug. Because the analyst overstates that link, the overall result should be rejected." + }, + "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": 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-15354-364", + "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:42491|issue:45276", + "issue:39692|issue:43994", + "issue:43257|issue:43854", + "issue:41720|issue:42915", + "issue:43856|issue:44387", + "issue:45362|issue:45406", + "issue:44488|issue:44991", + "issue:43644|issue:43645", + "issue:42915|issue:44756", + "issue:42491|issue:43931", + "issue:44661|issue:44683", + "issue:42491|issue:43856", + "issue:43404|issue:43454", + "issue:45276|issue:45335", + "issue:36010|issue:39404", + "issue:44164|issue:44291", + "issue:41720|issue:44155", + "issue:44402|issue:44488", + "issue:43873|issue:44387", + "issue:39692|issue:44960", + "issue:44387|issue:44756", + "issue:43425|issue:43957", + "issue:44292|issue:44912", + "issue:42947|issue:43381", + "issue:43854|issue:45216", + "issue:44291|issue:44756", + "issue:45276|issue:45292", + "issue:44743|issue:44960", + "issue:43381|issue:44928", + "issue:44164|issue:44756", + "issue:43957|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": "d48da9f8fd85f8d80fc86562a838683c54e94a870feb1a7cc2553a304b912357", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 41, + "serialized_chars": 39812, + "estimated_input_tokens": 9953, + "estimated_eval_tokens": 20162 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most pairs are false positives: they share broad themes like loading, quantization, or memory, but not the same concrete bug. The only convincing duplicate family is the `resize_token_embeddings` regression across Gemma/T5Gemma variants, where multiple reports point to the same missing propagation into nested embedding/output modules.", + "confidence": 0.86, + "canonical_issue_reason": "Issue 45292 is the broadest and most central report in the `resize_token_embeddings` regression family, and it best covers the model-specific follow-ups about nested output/decoder embeddings.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45292 is the best representative issue because it states the generic symptom clearly and matches the closely related Gemma/T5Gemma reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43854", + "right": "issue:43931", + "accept": false, + "reason": "Different models and different failure modes: GLM-4.7 test loading vs Qwen3-VL weight-shape mismatch." + }, + { + "left": "issue:43856", + "right": "issue:43873", + "accept": false, + "reason": "Both mention Qwen3/MoE and memory, but one is training-memory inefficiency and the other is quantization/offloading behavior." + }, + { + "left": "issue:42491", + "right": "issue:45276", + "accept": false, + "reason": "Unrelated problems: LoRA checkpoint compatibility across branches vs Gemma4 token embedding resizing." + }, + { + "left": "issue:39692", + "right": "issue:43994", + "accept": false, + "reason": "Both involve SigLIP2, but one is a documentation/example error and the other is a runtime inference quality issue; not the same concrete bug." + }, + { + "left": "issue:43257", + "right": "issue:43854", + "accept": false, + "reason": "Qwen3 MoE loading/conversion bug vs GLM-4.7 test loading bug; different code paths." + }, + { + "left": "issue:41720", + "right": "issue:42915", + "accept": false, + "reason": "Both are Qwen3-related, but one is device-mapping CUDA assert and the other is FP8 config failure." + }, + { + "left": "issue:43856", + "right": "issue:44387", + "accept": false, + "reason": "Both are memory-related, but the causes differ: training memory usage vs int4 quantization OOM regression." + }, + { + "left": "issue:45362", + "right": "issue:45406", + "accept": false, + "reason": "Different models and subsystems: Qwen3.5 chat failure vs Gemma4 processor attribute error in serve." + }, + { + "left": "issue:44488", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer/model loading issues, but they affect different models and there is no concrete shared code-path evidence." + }, + { + "left": "issue:43644", + "right": "issue:43645", + "accept": false, + "reason": "Transformers 5 buffer initialization bug vs Jupyter custom-model initialization regression; not the same underlying problem." + }, + { + "left": "issue:42915", + "right": "issue:44756", + "accept": false, + "reason": "FP8 quantization failure vs mmap/OOM mitigation on Strix Halo; unrelated bugs." + }, + { + "left": "issue:42491", + "right": "issue:43931", + "accept": false, + "reason": "Different families: Qwen3 MoE LoRA compatibility vs Qwen3-VL checkpoint shape mismatch." + }, + { + "left": "issue:44661", + "right": "issue:44683", + "accept": false, + "reason": "Tokenizer-mapping/model-addition bug vs compiled flex_attention on torch>=2.9; unrelated." + }, + { + "left": "issue:42491", + "right": "issue:43856", + "accept": false, + "reason": "Same broad Qwen3/MoE area, but one is checkpoint compatibility and the other is training memory usage." + }, + { + "left": "issue:43404", + "right": "issue:43454", + "accept": false, + "reason": "Similar symptom class (untied lm_head), but different model implementations and no evidence of one shared code path." + }, + { + "left": "issue:45276", + "right": "issue:45335", + "accept": true, + "reason": "Both report the same `resize_token_embeddings` regression failing to update nested embedding/output modules in model-specific wrappers; a single fix could plausibly cover both." + }, + { + "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/load extra_state handling vs init_empty_weights unexpected argument; different serialization/loading bugs." + }, + { + "left": "issue:41720", + "right": "issue:44155", + "accept": false, + "reason": "CUDA assert during Qwen3 device mapping vs AudioFlamingo3 batched inference token leakage; unrelated." + }, + { + "left": "issue:44402", + "right": "issue:44488", + "accept": false, + "reason": "Different tokenizer/model loading failures for different BERT models; not the same bug." + }, + { + "left": "issue:43873", + "right": "issue:44387", + "accept": false, + "reason": "Quantization/offloading behavior vs int4 reserved-memory OOM; similar theme, different root cause." + }, + { + "left": "issue:39692", + "right": "issue:44960", + "accept": false, + "reason": "SigLIP2 doc/example problem vs generic GLM5 issue; unrelated." + }, + { + "left": "issue:44387", + "right": "issue:44756", + "accept": false, + "reason": "Both can lead to OOM, but one is a Transformers 5 quantization regression and the other is a platform-specific mmap workaround." + }, + { + "left": "issue:43425", + "right": "issue:43957", + "accept": false, + "reason": "Torch version incompatibility vs meta-device loading regression; different problems." + }, + { + "left": "issue:44292", + "right": "issue:44912", + "accept": false, + "reason": "Different quantization formats and models: NVFP4 load error vs MXFP4 fallback to bf16." + }, + { + "left": "issue:42947", + "right": "issue:43381", + "accept": false, + "reason": "Gradient checkpointing complaints, but one is ineffective under PEFT LoRA and the other is eval-mode prohibition." + }, + { + "left": "issue:43854", + "right": "issue:45216", + "accept": false, + "reason": "GLM-4.7 test loading failure vs Qwen3.5 save_pretrained regression; unrelated." + }, + { + "left": "issue:44291", + "right": "issue:44756", + "accept": false, + "reason": "init_empty_weights/_is_hf_initialized TypeError vs mmap/OOM issue; unrelated." + }, + { + "left": "issue:45276", + "right": "issue:45335", + "accept": true, + "reason": "Both are the same `resize_token_embeddings` family of bugs affecting nested embedding/output attributes across different model wrappers." + }, + { + "left": "issue:44743", + "right": "issue:44960", + "accept": false, + "reason": "Qwen3.5 recurrent-state cache bug vs GLM5 issue; unrelated." + }, + { + "left": "issue:43381", + "right": "issue:44928", + "accept": false, + "reason": "Eval-mode gradient checkpointing rule vs Qwen3.5 RLHF gradient explosion from 3D position_ids; unrelated." + }, + { + "left": "issue:44164", + "right": "issue:44756", + "accept": false, + "reason": "Save/from_pretrained extra_state handling vs mmap/OOM workaround; different code paths." + }, + { + "left": "issue:43957", + "right": "issue:44756", + "accept": false, + "reason": "Meta-device model loading regression vs Strix Halo mmap/OOM issue; not the same bug." + }, + { + "left": "issue:43646", + "right": "issue:43708", + "accept": false, + "reason": "Custom model init regression vs Trainer resume/max_steps calculation bug; unrelated." + }, + { + "left": "issue:45292", + "right": "issue:45335", + "accept": true, + "reason": "Same `resize_token_embeddings` regression: a generic output_embeddings report and a T5Gemma decoder.embed_tokens report of the same missing propagation behavior." + }, + { + "left": "issue:44479", + "right": "issue:44877", + "accept": false, + "reason": "Video-input regression for several Qwen VL variants vs strict config blocking granite_speech loading; unrelated." + }, + { + "left": "issue:44402", + "right": "issue:44451", + "accept": false, + "reason": "Tokenizer vocab mismatch vs inability to load a different BERT model; not the same underlying issue." + }, + { + "left": "issue:43645", + "right": "issue:44164", + "accept": false, + "reason": "Notebook custom-model initialization vs extra_state save/load handling; different bugs." + }, + { + "left": "issue:43257", + "right": "issue:44863", + "accept": false, + "reason": "Qwen3 MoE weight conversion/loading bug vs NemotronH checkpoint loading bug; unrelated." + }, + { + "left": "issue:43425", + "right": "issue:43828", + "accept": false, + "reason": "Torch version incompatibility vs autocast dtype mismatch on Phi-tiny-MoE; different root causes." + }, + { + "left": "issue:43828", + "right": "issue:43957", + "accept": false, + "reason": "Autocast dtype mismatch vs meta-device loading regression; unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The analysis is conservative and grounded in the packet. The rejected pairs mostly rely on broad thematic similarity without clear shared bug evidence, and the accepted `resize_token_embeddings` pairs are plausibly the same fix family with consistent symptoms across Gemma/T5Gemma wrappers. The canonical issue choice is reasonable, and there is no obvious 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": 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-15354-364", + "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:44530|issue:45084", + "issue:43425|issue:43716", + "issue:44265|issue:44898", + "issue:43299|issue:43931", + "issue:43901|issue:44509", + "issue:44479|issue:44805", + "issue:43994|issue:44079", + "issue:43856|issue:44756", + "issue:44671|issue:44991", + "issue:39404|issue:44671", + "issue:40990|issue:42947", + "issue:44589|issue:45084", + "issue:43596|issue:44683", + "issue:43425|issue:44918", + "issue:36032|issue:43452", + "issue:32090|issue:39404", + "issue:43582|issue:43632", + "issue:44991|issue:45081", + "issue:43957|issue:44164", + "issue:40990|issue:43421", + "issue:44164|issue:44479", + "issue:44488|issue:45081", + "issue:43883|issue:44488", + "issue:43425|issue:45161", + "issue:43611|issue:43644", + "issue:44206|issue:44220", + "issue:42947|issue:44756", + "issue:44488|issue:44987", + "issue:34567|issue:40444", + "issue:43716|issue:45084" + ], + "prepared_review_unit_hash": "0c0437ee679256831b7fcec157acaca472f34f79d97a02532c4c818413a98a8e", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 35, + "serialized_chars": 38180, + "estimated_input_tokens": 9545, + "estimated_eval_tokens": 19346 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is thematically loose: many items are Transformers v5 / model-loading regressions, but the candidate pairs mostly point to different concrete bugs, model families, or subsystems, so they should not be merged.", + "confidence": 0.38, + "canonical_issue_reason": "issue:43957 is the broadest representative of the recurring v5 model-loading regression pattern in this set, so it works best as the cluster anchor.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43957 is the most reusable issue to represent the cluster because it is version-scoped, loading-focused, and broad enough to cover the repeated regression theme.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44589", + "right": "issue:44756", + "accept": false, + "reason": "Different failures: missing Float8 storage vs Apple Silicon mmap/OOM; no shared code-path." + }, + { + "left": "issue:43906", + "right": "issue:44530", + "accept": false, + "reason": "Isolated reproduction of one issue vs a Qwen3.5 PagedAttentionCache crash; unrelated symptoms and fixes." + }, + { + "left": "issue:43931", + "right": "issue:44863", + "accept": false, + "reason": "Both are model-loading problems, but for different model families and different failure modes (shape mismatch vs NemotronH checkpoint loading)." + }, + { + "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 notebook custom-model initialization; not the same bug." + }, + { + "left": "issue:44530", + "right": "issue:45084", + "accept": false, + "reason": "Qwen3.5 cache group-type crash vs template compilation error; different code paths." + }, + { + "left": "issue:43425", + "right": "issue:43716", + "accept": false, + "reason": "Torch 2.10 compatibility issue is not the same as a Mistral-3 image preprocessor dtype mismatch." + }, + { + "left": "issue:44265", + "right": "issue:44898", + "accept": false, + "reason": "torch.export / torch_compilable_check failure is unrelated to Perceiver non-default-resolution interpolation behavior." + }, + { + "left": "issue:43299", + "right": "issue:43931", + "accept": false, + "reason": "Both concern Qwen3-VL loading, but one is MoE model loading in v5 dev0 and the other is a specific weight-shape mismatch; not clearly the same bug." + }, + { + "left": "issue:43901", + "right": "issue:44509", + "accept": false, + "reason": "Both are docs issues, but they concern different pipeline/task deprecations and different documentation fixes." + }, + { + "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:43994", + "right": "issue:44079", + "accept": false, + "reason": "Siglip2 wrong outputs vs ModelOutput key assignment bug; different behavior and likely different fix." + }, + { + "left": "issue:43856", + "right": "issue:44756", + "accept": false, + "reason": "Qwen3 MoE training memory inefficiency is not the same as disabling mmap on Strix Halo to avoid OOM." + }, + { + "left": "issue:44671", + "right": "issue:44991", + "accept": false, + "reason": "CamemBERT masked-LM regression and est-roberta tokenizer loading failure are separate model-specific issues." + }, + { + "left": "issue:39404", + "right": "issue:44671", + "accept": false, + "reason": "Whisper pipeline behavior and CamemBERT masked-LM predictions are unrelated." + }, + { + "left": "issue:40990", + "right": "issue:42947", + "accept": false, + "reason": "High perplexity on gpt-oss is unrelated to LoRA gradient checkpointing ineffectiveness." + }, + { + "left": "issue:44589", + "right": "issue:45084", + "accept": false, + "reason": "Float8 storage lookup failure and template-node compilation failure are different bugs." + }, + { + "left": "issue:43596", + "right": "issue:44683", + "accept": false, + "reason": "DeepSpeed ZeRO-3 init IndexError is unrelated to compiled flex_attention failing on newer torch." + }, + { + "left": "issue:43425", + "right": "issue:44918", + "accept": false, + "reason": "Torch version incompatibility is not the same as Qwen3.5 embedding unpacking with TRL SFT trainer." + }, + { + "left": "issue:36032", + "right": "issue:43452", + "accept": false, + "reason": "T5 tokenizer special-token conflict and gguf_file loading breakage are different loading paths." + }, + { + "left": "issue:32090", + "right": "issue:39404", + "accept": false, + "reason": "Trainer GPU broadcast NoneType error is unrelated to Whisper return_language pipeline regression." + }, + { + "left": "issue:43582", + "right": "issue:43632", + "accept": false, + "reason": "Apple Silicon warmup TypeError and _is_hf_initialized flag regression are separate issues." + }, + { + "left": "issue:44991", + "right": "issue:45081", + "accept": false, + "reason": "Both involve tokenizer loading, but they affect different models and different internal failure points." + }, + { + "left": "issue:43957", + "right": "issue:44164", + "accept": false, + "reason": "meta-device loading breakage and save/from_pretrained extra_state handling are distinct serialization/loading bugs." + }, + { + "left": "issue:40990", + "right": "issue:43421", + "accept": false, + "reason": "Perplexity regression and runtime special-token post-processor update bug are unrelated." + }, + { + "left": "issue:44164", + "right": "issue:44479", + "accept": false, + "reason": "extra_state save/load handling is not the same as the video-input regression." + }, + { + "left": "issue:44488", + "right": "issue:45081", + "accept": false, + "reason": "Different tokenizer-loading failures for different models; no evidence of a single underlying bug." + }, + { + "left": "issue:43883", + "right": "issue:44488", + "accept": false, + "reason": "Molmo tied-weights attribute error and sleng-bert loading failure are unrelated." + }, + { + "left": "issue:43425", + "right": "issue:45161", + "accept": false, + "reason": "Torch 2.10 incompatibility is not the same as TP failing for GPT-OSS MoE." + }, + { + "left": "issue:43611", + "right": "issue:43644", + "accept": false, + "reason": "Model-loading with base_model_prefix and non-persistent buffer initialization are different v5 regressions." + }, + { + "left": "issue:44206", + "right": "issue:44220", + "accept": false, + "reason": "Unsupported center arg in feature extractor and _torch_extract_fbank_features issues may be related, but the evidence is not strong enough to treat them as the same bug." + }, + { + "left": "issue:42947", + "right": "issue:44756", + "accept": false, + "reason": "LoRA gradient-checkpointing behavior is unrelated to Strix Halo mmap/OOM." + }, + { + "left": "issue:44488", + "right": "issue:44987", + "accept": false, + "reason": "Two separate model-loading regressions for different repositories and failure modes." + }, + { + "left": "issue:34567", + "right": "issue:40444", + "accept": false, + "reason": "TrainerState token counter and Qwen2.5-VL iterable multi-image finetuning are unrelated." + }, + { + "left": "issue:43716", + "right": "issue:45084", + "accept": false, + "reason": "Mistral-3 image preprocessor dtype mismatch is unrelated to the template compilation error." + } + ] + }, + "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-15354-364", + "nodes": [ + "issue:36032", + "issue:39404", + "issue:40990", + "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:43883", + "issue:43994", + "issue:44220", + "issue:44265", + "issue:44403", + "issue:44451", + "issue:44466", + "issue:44479", + "issue:44530", + "issue:44534", + "issue:44610", + "issue:44683", + "issue:44756", + "issue:44811", + "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:43883|issue:44991", + "issue:43643|issue:44530", + "issue:43452|issue:43482", + "issue:43317|issue:43482", + "issue:36032|issue:42915", + "issue:43262|issue:43377", + "issue:44451|issue:44987", + "issue:43541|issue:45084", + "issue:39404|issue:43825", + "issue:44849|issue:45325", + "issue:42915|issue:44530", + "issue:43404|issue:44466", + "issue:36032|issue:43645", + "issue:43646|issue:44534", + "issue:44987|issue:44991", + "issue:45084|issue:45507", + "issue:43994|issue:44610", + "issue:43377|issue:43819", + "issue:43645|issue:44534", + "issue:43322|issue:43540", + "issue:45216|issue:45357", + "issue:44403|issue:45081", + "issue:44451|issue:45081", + "issue:44265|issue:44610", + "issue:43632|issue:44756", + "issue:41553|issue:43582", + "issue:44220|issue:44683", + "issue:43404|issue:45127", + "issue:42915|issue:45084", + "issue:44479|issue:45325", + "issue:43296|issue:43531", + "issue:44987|issue:45081", + "issue:40990|issue:44811" + ], + "prepared_review_unit_hash": "0e2623970186c894501b5a18ec079a221a810a75226097db3ab8987ce911138f", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37887, + "estimated_input_tokens": 9472, + "estimated_eval_tokens": 19200 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is loosely centered on Transformers v5/v5.0 regression reports, but the soft pairs mostly describe different models, APIs, or failure modes rather than the same underlying bug/change. No PRs are present.", + "confidence": 0.62, + "canonical_issue_reason": "Issue 43646 is the broadest, most central v5 regression report in the set: it targets custom model initialization, a root area that overlaps several other initialization-related complaints.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43646 is the best representative issue for this cluster because it is general, version-specific, and closest to the shared v5 regression theme.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43644", + "right": "issue:43646", + "accept": false, + "reason": "Both are v5 initialization regressions, but 43644 is specifically about non-persistent buffers getting junk values, while 43646 is about custom model initialization broadly; not the same concrete bug." + }, + { + "left": "issue:43883", + "right": "issue:44991", + "accept": false, + "reason": "Different failures in different models: Molmo missing an attribute vs tokenizer loading for EMBEDDIA/est-roberta." + }, + { + "left": "issue:43643", + "right": "issue:44530", + "accept": false, + "reason": "Missing fields from AutoConfig with trust_remote_code is unrelated to a PagedAttentionCache group-type crash in Qwen3.5 models." + }, + { + "left": "issue:43452", + "right": "issue:43482", + "accept": false, + "reason": "Both mention GGUF/loading, but one is about gguf_file breaking tokenizer/model loading APIs and the other is a Qwen2.5-GGUF v5 loading failure; too different to be the same bug." + }, + { + "left": "issue:43317", + "right": "issue:43482", + "accept": false, + "reason": "device_map/offload dequantized-model loading is a different code path from GGUF loading." + }, + { + "left": "issue:36032", + "right": "issue:42915", + "accept": false, + "reason": "T5 tokenizer special-tokens conflict and Qwen3MoE FP8 configuration failure are unrelated." + }, + { + "left": "issue:43262", + "right": "issue:43377", + "accept": false, + "reason": "Audio chat-template sampling-rate defaulting and MIMI batched-vs-single output mismatch are different audio bugs." + }, + { + "left": "issue:44451", + "right": "issue:44987", + "accept": false, + "reason": "Both are loading regressions, but for different models and different failure surfaces; not one underlying issue." + }, + { + "left": "issue:43541", + "right": "issue:45084", + "accept": false, + "reason": "Grouped_mm tracing failure in Mixtral is unrelated to template-node compilation errors." + }, + { + "left": "issue:39404", + "right": "issue:43825", + "accept": false, + "reason": "Whisper pipeline return_language behavior and a pipeline error-message regression are not the same defect." + }, + { + "left": "issue:44849", + "right": "issue:45325", + "accept": false, + "reason": "Qwen3.5 hidden-states behavior and Qwen2.5-VL rope-index scaling are different model bugs." + }, + { + "left": "issue:42915", + "right": "issue:44530", + "accept": false, + "reason": "FineGrainedFP8Config loading and PagedAttentionCache invalid group type are separate Qwen3-related issues." + }, + { + "left": "issue:43404", + "right": "issue:44466", + "accept": false, + "reason": "One is an untied lm_head in Mistral3ForConditionalGeneration; the other is serialization inconsistency for tied weights. Related area, but different bugs." + }, + { + "left": "issue:36032", + "right": "issue:43645", + "accept": false, + "reason": "Tokenizer-method conflict vs custom model initialization in Jupyter are not the same issue." + }, + { + "left": "issue:43646", + "right": "issue:44534", + "accept": false, + "reason": "Both involve v5 initialization semantics, but one is custom model init and the other is non-persistent buffer contents; different concrete bugs." + }, + { + "left": "issue:44987", + "right": "issue:44991", + "accept": false, + "reason": "Both are v5 load failures, but they affect different repositories/models and different loading code paths." + }, + { + "left": "issue:45084", + "right": "issue:45507", + "accept": false, + "reason": "Template compilation error and GraniteMoEHybrid invalid method call are unrelated." + }, + { + "left": "issue:43994", + "right": "issue:44610", + "accept": false, + "reason": "Nonsensical outputs with SigLIP2 and a processor/model input-size mismatch in OmDet-Turbo are different problems." + }, + { + "left": "issue:43377", + "right": "issue:43819", + "accept": false, + "reason": "MIMI encoder padding-mask batching bug and DAC from_latents/STE mismatch are different codec issues." + }, + { + "left": "issue:43645", + "right": "issue:44534", + "accept": false, + "reason": "Jupyter custom-model initialization and non-persistent buffer junk are separate v5 regressions." + }, + { + "left": "issue:43322", + "right": "issue:43540", + "accept": false, + "reason": "Llava Next segmentation fault and Qwen3OmniMoe video-input ValueError are different multimodal failures." + }, + { + "left": "issue:45216", + "right": "issue:45357", + "accept": false, + "reason": "Both concern Qwen3.5 save_pretrained regressions, but one is incorrect checkpoint saving in general and the other is specifically wrong visual encoder keys." + }, + { + "left": "issue:44403", + "right": "issue:45081", + "accept": false, + "reason": "General loading noise and a Mistral regex patch crash are unrelated." + }, + { + "left": "issue:44451", + "right": "issue:45081", + "accept": false, + "reason": "ScandiBERT loading failure and Mistral tokenizer regex patch failure are different tokenizer problems." + }, + { + "left": "issue:44265", + "right": "issue:44610", + "accept": false, + "reason": "torch.export failures with torch_compilable_check are unrelated to OmDet-Turbo processor resolution mismatch." + }, + { + "left": "issue:43632", + "right": "issue:44756", + "accept": false, + "reason": "_is_hf_initialized flag breakage and disabling mmap on Strix Halo are unrelated." + }, + { + "left": "issue:41553", + "right": "issue:43582", + "accept": false, + "reason": "Bad AutoTokenizer error message for Voxtral and AppleSilicon caching_allocator_warmup TypeError are unrelated." + }, + { + "left": "issue:44220", + "right": "issue:44683", + "accept": false, + "reason": "_torch_extract_fbank_features and compiled flex_attention on torch>=2.9 are different low-level issues." + }, + { + "left": "issue:43404", + "right": "issue:45127", + "accept": false, + "reason": "lm_head tying in Mistral3ForConditionalGeneration and LoRA merge collapse with extended vocab are different weight-tying bugs." + }, + { + "left": "issue:42915", + "right": "issue:45084", + "accept": false, + "reason": "Qwen3Moe FP8 config failure and template compilation errors have no shared code-path." + }, + { + "left": "issue:44479", + "right": "issue:45325", + "accept": false, + "reason": "Qwen3.5 video-input regression and Qwen2.5-VL rope-position scaling are different multimodal regressions." + }, + { + "left": "issue:43296", + "right": "issue:43531", + "accept": false, + "reason": "PaddleOCR-VL model loading in vLLM and a sliding_window issue in Qwen3-MoE are unrelated." + }, + { + "left": "issue:44987", + "right": "issue:45081", + "accept": false, + "reason": "Loading physical-intelligence/fast under v5 and Mistral tokenizer regex patching are different failures." + }, + { + "left": "issue:40990", + "right": "issue:44811", + "accept": false, + "reason": "High perplexity on gpt-oss-20b and Whisper batch_decode skip_special_tokens behavior are unrelated." + } + ] + }, + "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": 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-15354-364", + "nodes": [ + "issue:33666", + "issue:35707", + "issue:36032", + "issue:40444", + "issue:41720", + "issue:42175", + "issue:42915", + "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:44062", + "issue:44079", + "issue:44265", + "issue:44291", + "issue:44403", + "issue:44488", + "issue:44534", + "issue:44568", + "issue:44589", + "issue:44683", + "issue:44829", + "issue:44877", + "issue:44987", + "issue:44991", + "issue:45020", + "issue:45072", + "issue:45081", + "issue:45325", + "issue:45357", + "issue:45381", + "issue:45479", + "issue:45491", + "issue:45538", + "issue:45542" + ], + "soft_pairs": [ + "issue:42175|issue:45542", + "issue:43611|issue:44534", + "issue:43317|issue:43856", + "issue:44062|issue:44987", + "issue:42915|issue:44291", + "issue:36032|issue:43742", + "issue:40444|issue:45325", + "issue:43632|issue:44534", + "issue:43531|issue:44877", + "issue:43818|issue:45357", + "issue:43421|issue:44568", + "issue:42915|issue:44589", + "issue:36032|issue:41720", + "issue:44829|issue:45479", + "issue:44683|issue:45020", + "issue:43482|issue:45538", + "issue:33666|issue:40444", + "issue:43425|issue:44403", + "issue:44079|issue:44534", + "issue:44403|issue:44991", + "issue:35707|issue:40444", + "issue:43377|issue:45491", + "issue:43792|issue:44589", + "issue:43482|issue:43525", + "issue:43749|issue:44079", + "issue:44291|issue:44589", + "issue:43644|issue:44079", + "issue:44079|issue:45072", + "issue:44265|issue:45072", + "issue:43632|issue:45072", + "issue:40444|issue:45381", + "issue:45325|issue:45381", + "issue:43611|issue:43646", + "issue:43531|issue:44291", + "issue:41720|issue:45081", + "issue:43881|issue:44488" + ], + "prepared_review_unit_hash": "caabd7ea5fb120e6655eb355329b7fe4b3de044aea07858e7d0257a90969a653", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 36, + "serialized_chars": 38431, + "estimated_input_tokens": 9608, + "estimated_eval_tokens": 19472 + }, + "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/regressions across tokenizers, model loading, vision/video, device mapping, and training; none of the soft pairs look like true duplicates.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue stands out: the items span many different bug classes and model families, so grouping them as one underlying defect would be misleading.", + "canonical_pr_reason": null, + "best_issue_reason": "If forced to pick an umbrella, issue 45020 is the broadest regression report, but it is still too generic and does not actually subsume the other distinct failures.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42175", + "right": "issue:45542", + "accept": false, + "reason": "Different backend/install problems: missing TensorFlow in a torch-only install vs tensorboard being present without TensorFlow." + }, + { + "left": "issue:43611", + "right": "issue:44534", + "accept": false, + "reason": "Different breakages: base_model_prefix loading failure vs non-persistent buffer junk in v5." + }, + { + "left": "issue:43317", + "right": "issue:43856", + "accept": false, + "reason": "Different load/offload path vs MoE training memory-usage complaint; not the same bug." + }, + { + "left": "issue:44062", + "right": "issue:44987", + "accept": false, + "reason": "Tokenizer AddedToken conflict vs failure loading a specific model under v5; unrelated symptoms and code paths." + }, + { + "left": "issue:42915", + "right": "issue:44291", + "accept": false, + "reason": "Both mention v5-era type issues, but one is FineGrainedFP8Config and the other is init_empty_weights/_is_hf_initialized; distinct bugs." + }, + { + "left": "issue:36032", + "right": "issue:43742", + "accept": false, + "reason": "T5 tokenizer method-name conflict vs key error loading MobileLLM; different tokenizer/model failures." + }, + { + "left": "issue:40444", + "right": "issue:45325", + "accept": false, + "reason": "Different Qwen2.5-VL bugs: multi-image IterableDataset failure vs still-image rope-index scaling." + }, + { + "left": "issue:43632", + "right": "issue:44534", + "accept": false, + "reason": "_is_hf_initialized flag regression vs junk-filled non-persistent buffers; not the same underlying defect." + }, + { + "left": "issue:43531", + "right": "issue:44877", + "accept": false, + "reason": "Sliding-window attention issue in Qwen3-MoE vs strict config preventing granite_speech loading." + }, + { + "left": "issue:43818", + "right": "issue:45357", + "accept": false, + "reason": "Video-LLaVA temporal-attention/weight-sharing bug vs Qwen3.5 save_pretrained key mismatch; different model subsystems." + }, + { + "left": "issue:43421", + "right": "issue:44568", + "accept": false, + "reason": "Runtime post-processor update bug vs tokenizer add_special_tokens regression; different tokenizer behaviors." + }, + { + "left": "issue:42915", + "right": "issue:44589", + "accept": false, + "reason": "Different failure modes: FP8 config handling vs missing Float8 storage object." + }, + { + "left": "issue:36032", + "right": "issue:41720", + "accept": false, + "reason": "Tokenizer class-name conflict vs Qwen3 auto device-map CUDA assert; unrelated." + }, + { + "left": "issue:44829", + "right": "issue:45479", + "accept": false, + "reason": "Different sequence-classification regressions and likely different root causes." + }, + { + "left": "issue:44683", + "right": "issue:45020", + "accept": false, + "reason": "Compiled flex_attention on torch>=2.9 is a specific runtime incompatibility, not the same as the broad remote_code loading regression." + }, + { + "left": "issue:43482", + "right": "issue:45538", + "accept": false, + "reason": "Qwen2.5-GGUF loading failure vs CLIPTokenizer model_max_length behavior; no overlap in bug surface." + }, + { + "left": "issue:33666", + "right": "issue:40444", + "accept": false, + "reason": "General multi-GPU training request vs a concrete multi-image IterableDataset bug; not duplicates." + }, + { + "left": "issue:43425", + "right": "issue:44403", + "accept": false, + "reason": "Torch 2.10 compatibility report vs generic loading noise complaint; different issues." + }, + { + "left": "issue:44079", + "right": "issue:44534", + "accept": false, + "reason": "ModelOutput key assignment bug vs non-persistent buffer initialization bug; distinct code paths." + }, + { + "left": "issue:44403", + "right": "issue:44991", + "accept": false, + "reason": "Loading noise / UX issue vs tokenizer load failure for EMBEDDIA/est-roberta; unrelated." + }, + { + "left": "issue:35707", + "right": "issue:40444", + "accept": false, + "reason": "Progressive generation with inputs_embeds/past_key_values is unrelated to Qwen2.5-VL multi-image dataset failures." + }, + { + "left": "issue:43377", + "right": "issue:45491", + "accept": false, + "reason": "MIMI encoder padding-mask bug vs Gemma3 sliding-window NaN embeddings; different models and mechanics." + }, + { + "left": "issue:43792", + "right": "issue:44589", + "accept": false, + "reason": "Whisper model can't run vs Float8 storage lookup failure; no shared underlying defect." + }, + { + "left": "issue:43482", + "right": "issue:43525", + "accept": false, + "reason": "GGUF/v5 loading problem vs Llama4Config missing pad_token_id; unrelated model-loading issues." + }, + { + "left": "issue:43749", + "right": "issue:44079", + "accept": false, + "reason": "FSDP CPU RAM efficient loading regression vs ModelOutput key handling; different subsystems." + }, + { + "left": "issue:44291", + "right": "issue:44589", + "accept": false, + "reason": "Unexpected _is_hf_initialized argument vs missing Float8 storage object; different type errors." + }, + { + "left": "issue:43644", + "right": "issue:44079", + "accept": false, + "reason": "Non-persistent buffer junk vs ModelOutput key assignment; unrelated regressions." + }, + { + "left": "issue:44079", + "right": "issue:45072", + "accept": false, + "reason": "ModelOutput key handling vs dtype mismatch in bfloat16 inference; different bugs." + }, + { + "left": "issue:44265", + "right": "issue:45072", + "accept": false, + "reason": "torch.export.export failure with torch_compilable_check vs inference dtype mismatch; not the same change." + }, + { + "left": "issue:43632", + "right": "issue:45072", + "accept": false, + "reason": "_is_hf_initialized regression vs bfloat16 dtype mismatch in inference; unrelated." + }, + { + "left": "issue:40444", + "right": "issue:45381", + "accept": false, + "reason": "Multi-image IterableDataset failure vs Qwen2.5-VL video position-id regression; different modalities and symptoms." + }, + { + "left": "issue:45325", + "right": "issue:45381", + "accept": false, + "reason": "Both are Qwen2.5-VL position-id bugs, but one is still-image temporal scaling and the other is video-input vision_position_ids; not clearly the same concrete defect." + }, + { + "left": "issue:43611", + "right": "issue:43646", + "accept": false, + "reason": "base_model_prefix loading regression vs custom model initialization regression; different failures." + }, + { + "left": "issue:43531", + "right": "issue:44291", + "accept": false, + "reason": "Qwen3-MoE sliding_window issue vs init_empty_weights/_is_hf_initialized TypeError; unrelated." + }, + { + "left": "issue:41720", + "right": "issue:45081", + "accept": false, + "reason": "Qwen3 auto device-mapping CUDA assert vs Mistral tokenizer regex patch crash; different code paths." + }, + { + "left": "issue:43881", + "right": "issue:44488", + "accept": false, + "reason": "glm-4v-9b loading failure vs sleng-bert tokenizer loading issue; unrelated models and causes." + } + ] + }, + "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": 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-15354-364", + "nodes": [ + "issue:34634", + "issue:35707", + "issue:39692", + "issue:40444", + "issue:41720", + "issue:42175", + "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:43454|issue:44466", + "issue:43994|issue:44898", + "issue:42175|issue:43296", + "issue:44387|issue:44617", + "issue:43828|issue:44683", + "issue:42915|issue:43482", + "issue:43526|issue:43883", + "issue:43701|issue:45325", + "issue:43262|issue:43746", + "issue:43792|issue:43957", + "issue:43317|issue:43873", + "issue:34634|issue:35707", + "issue:43606|issue:44265", + "issue:43881|issue:44991", + "issue:44186|issue:44265", + "issue:44291|issue:44877", + "issue:43606|issue:43819", + "issue:35707|issue:41720", + "issue:43632|issue:44164", + "issue:43377|issue:44898", + "issue:44610|issue:45325", + "issue:43454|issue:43550", + "issue:43454|issue:45072", + "issue:43377|issue:44610", + "issue:39692|issue:40444", + "issue:43749|issue:44534", + "issue:44466|issue:45357", + "issue:43296|issue:44991", + "issue:43454|issue:43720", + "issue:43296|issue:44877", + "issue:41720|issue:43322", + "issue:43792|issue:44683", + "issue:41720|issue:44977", + "issue:43404|issue:45357" + ], + "prepared_review_unit_hash": "5e19d8b6374bce1f12c581cc4881edf0d1d6e9e9dadbbdab196f1a46770c7a2b", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37859, + "estimated_input_tokens": 9465, + "estimated_eval_tokens": 19186 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab bag of model-loading, quantization, and serialization regressions, not one clean duplicate set. The only strong duplicate-looking pair is the tied lm_head / serialization bug in 43454 and 44466; the rest are related only at a broad subsystem level.", + "confidence": 0.74, + "canonical_issue_reason": "issue:44466 is the most general and technical representative of the clear duplicate pair: it describes the tied-weight serialization problem across devices in v5/main, which subsumes the more model-specific 43454 report.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44466 best represents the actionable core bug in the cluster because it captures the broader tied-weight serialization defect rather than a single model-specific symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43454", + "right": "issue:44466", + "accept": true, + "reason": "Same underlying tied lm_head weight handling bug: one reports failed tying causing bad generation, the other reports inconsistent serialization of the same tied weights across devices." + }, + { + "left": "issue:43994", + "right": "issue:44898", + "accept": false, + "reason": "Both involve vision models, but they describe different failures on different models and code paths; too broad to treat as the same bug." + }, + { + "left": "issue:42175", + "right": "issue:43296", + "accept": false, + "reason": "Completely unrelated: backend packaging/extras vs an audio processor sampling-rate default bug." + }, + { + "left": "issue:44387", + "right": "issue:44617", + "accept": false, + "reason": "Both mention memory pressure, but one is a quantization reserved-memory regression and the other is a model-specific CUDA OOM symptom." + }, + { + "left": "issue:43828", + "right": "issue:44683", + "accept": false, + "reason": "Different failures: autocast dtype mismatch versus torch 2.9 compiled flex_attention breakage." + }, + { + "left": "issue:42915", + "right": "issue:43482", + "accept": false, + "reason": "Different model families and different problems: fine-grained FP8 config failure vs GGUF loading failure." + }, + { + "left": "issue:43526", + "right": "issue:43883", + "accept": false, + "reason": "Unrelated bugs: image label reduction logic versus a missing attribute during model loading." + }, + { + "left": "issue:43701", + "right": "issue:45325", + "accept": false, + "reason": "Different areas: checkpoint resume key mismatch versus Qwen2.5-VL rope index scaling." + }, + { + "left": "issue:43262", + "right": "issue:43746", + "accept": false, + "reason": "Both are audio-related, but one is a chat-template sampling-rate default and the other is PEFT/local checkpoint loading." + }, + { + "left": "issue:43792", + "right": "issue:43957", + "accept": false, + "reason": "Both are loading/runtime issues, but the concrete bugs and model paths differ." + }, + { + "left": "issue:43317", + "right": "issue:43873", + "accept": false, + "reason": "Both concern offload/quantization, but the titles point to different concrete failures and not one specific shared bug." + }, + { + "left": "issue:34634", + "right": "issue:35707", + "accept": false, + "reason": "No clear overlap: BarkProcessor voice_preset versus generation with inputs_embeds/past_key_values." + }, + { + "left": "issue:43606", + "right": "issue:44265", + "accept": false, + "reason": "Different code paths: CPU offload device mismatch versus torch.export with torch_compilable_check." + }, + { + "left": "issue:43881", + "right": "issue:44991", + "accept": false, + "reason": "Different failures: a model loading error versus tokenizer loading for a specific repo." + }, + { + "left": "issue:44186", + "right": "issue:44265", + "accept": false, + "reason": "Tokenizer crash and export failure are unrelated despite both being regressions." + }, + { + "left": "issue:44291", + "right": "issue:44877", + "accept": false, + "reason": "Both are loading/config issues, but one is an init_empty_weights argument problem and the other is strict config validation." + }, + { + "left": "issue:43606", + "right": "issue:43819", + "accept": false, + "reason": "Different bugs: device mismatch in offload versus DAC latent conversion mismatch." + }, + { + "left": "issue:35707", + "right": "issue:41720", + "accept": false, + "reason": "No shared underlying bug; generation API behavior versus device-mapping CUDA assertion." + }, + { + "left": "issue:43632", + "right": "issue:44164", + "accept": false, + "reason": "Both touch v5 loading/saving internals, but one is an _is_hf_initialized flag regression and the other is extra_state handling." + }, + { + "left": "issue:43377", + "right": "issue:44898", + "accept": false, + "reason": "Different models and symptoms: batched-vs-single MIMI outputs versus Perceiver image-size handling." + }, + { + "left": "issue:44610", + "right": "issue:45325", + "accept": false, + "reason": "Unrelated visual processor size mismatch versus Qwen2.5-VL temporal position-id scaling." + }, + { + "left": "issue:43454", + "right": "issue:43550", + "accept": false, + "reason": "One is tied-weight generation/serialization, the other is a torch.compile SDPA issue." + }, + { + "left": "issue:43454", + "right": "issue:45072", + "accept": false, + "reason": "Different model code paths: tied lm_head weights versus dtype mismatch in SwitchTransformers/TimmWrapperModel." + }, + { + "left": "issue:43377", + "right": "issue:44610", + "accept": false, + "reason": "Different multimodal preprocessing bugs with no concrete shared root cause." + }, + { + "left": "issue:39692", + "right": "issue:40444", + "accept": false, + "reason": "Distinct issues: documentation/example errors for SigLIP2 versus Qwen2.5-VL dataset handling." + }, + { + "left": "issue:43749", + "right": "issue:44534", + "accept": false, + "reason": "Both are v5-related loading regressions, but the specific failure modes are different." + }, + { + "left": "issue:44466", + "right": "issue:45357", + "accept": false, + "reason": "Both are serialization bugs, but they involve different models and different serialized keys/paths." + }, + { + "left": "issue:43296", + "right": "issue:44991", + "accept": false, + "reason": "Different failure types: vLLM model loading versus tokenizer loading regression." + }, + { + "left": "issue:43454", + "right": "issue:43720", + "accept": false, + "reason": "Different bugs: tied-weight handling versus packed-weight unpacking during accelerate loading." + }, + { + "left": "issue:43296", + "right": "issue:44877", + "accept": false, + "reason": "Different models and mechanisms: model loading failure versus strict config rejection." + }, + { + "left": "issue:41720", + "right": "issue:43322", + "accept": false, + "reason": "Both are model-loading crashes, but they are separate models and separate failure modes." + }, + { + "left": "issue:43792", + "right": "issue:44683", + "accept": false, + "reason": "Completely unrelated: Whisper runtime failure versus compiled flex_attention breakage." + }, + { + "left": "issue:43404", + "right": "issue:45357", + "accept": false, + "reason": "Both involve weight/save behavior, but the model families and concrete bugs differ." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays mostly conservative: it treats the cluster as heterogeneous and only singles out 43454/44466 as the plausible duplicate-like pair. The other soft-edge rejections are appropriately cautious. If anything, the 43454/44466 linkage could be phrased a bit more tentatively (\"likely related\" rather than definitive same bug), but it is still supportable from the titles." + }, + "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": 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-15354-364", + "nodes": [ + "issue:33357", + "issue:34567", + "issue:35707", + "issue:36010", + "issue:36331", + "issue:40990", + "issue:41720", + "issue:42617", + "issue:42915", + "issue:43316", + "issue:43317", + "issue:43322", + "issue:43388", + "issue:43526", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43572", + "issue:43577", + "issue:43606", + "issue:43653", + "issue:43708", + "issue:43720", + "issue:43746", + "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:43526|issue:43881", + "issue:43873|issue:44756", + "issue:43577|issue:43708", + "issue:42617|issue:42915", + "issue:43550|issue:43720", + "issue:44610|issue:44898", + "issue:44186|issue:44610", + "issue:44877|issue:44991", + "issue:44479|issue:44918", + "issue:36331|issue:42915", + "issue:44671|issue:44743", + "issue:36331|issue:42617", + "issue:43316|issue:44913", + "issue:43550|issue:43606", + "issue:42915|issue:43317", + "issue:41720|issue:42617", + "issue:44292|issue:44479", + "issue:41720|issue:45161", + "issue:43531|issue:43572", + "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:36010|issue:42617", + "issue:36331|issue:41720", + "issue:44514|issue:44545", + "issue:33357|issue:35707", + "issue:43606|issue:43720", + "issue:41720|issue:43746", + "issue:43550|issue:43653" + ], + "prepared_review_unit_hash": "2ba8ae2280dda0f28363e1e08bba48be5287b285e9c80a1927e06d57ccd7f806", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 35, + "serialized_chars": 38116, + "estimated_input_tokens": 9529, + "estimated_eval_tokens": 19314 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly a heterogeneous set of unrelated closed issues. The only clear duplicate pair is the Qwen2_5_VLProcessor.apply_chat_template padding=False crash (44514/44545); everything else should stay separate.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 44545 is the cleanest canonical report for the only true duplicate pair: same function, same batched-input/padding=False crash, just a cleaner title than 44514.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44545 is the best representative issue here because it most clearly describes the duplicated apply_chat_template padding=False bug and is the cleaner canonical reference.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44423", + "right": "issue:45406", + "accept": false, + "reason": "Both involve serving/multimodal failures, but the errors and affected components differ (`'str' object has no attribute 'to'` vs missing `_tokenizer`)." + }, + { + "left": "issue:44186", + "right": "issue:45072", + "accept": false, + "reason": "Different bugs: tokenizer crashes on NER padding/truncation vs dtype mismatch in bfloat16 inference." + }, + { + "left": "issue:34567", + "right": "issue:40990", + "accept": false, + "reason": "Unrelated: TrainerState token counting bug vs high perplexity on a specific model/eval setup." + }, + { + "left": "issue:43526", + "right": "issue:43881", + "accept": false, + "reason": "Different subsystems and failures: `reduce_labels` in BEiT image processing vs loading failure for glm-4v-9b." + }, + { + "left": "issue:43873", + "right": "issue:44756", + "accept": false, + "reason": "One is a generic quantization offload question, the other is a Strix Halo mmap/OOM workaround; not the same bug." + }, + { + "left": "issue:43577", + "right": "issue:43708", + "accept": false, + "reason": "Different problems: BLIP2 dtype propagation vs Trainer resume/max_steps calculation." + }, + { + "left": "issue:42617", + "right": "issue:42915", + "accept": false, + "reason": "Both are run/load failures, but `3d_parallel.py` and Qwen3Moe FineGrainedFP8Config point to different code paths." + }, + { + "left": "issue:43550", + "right": "issue:43720", + "accept": false, + "reason": "Distinct failures: torch.compile + SDPA on Bamba vs unpacking packed weights during accelerate loading for BitNet." + }, + { + "left": "issue:44610", + "right": "issue:44898", + "accept": false, + "reason": "Both are vision-shape issues, but they affect different models and different mechanisms (processor output size vs interpolation path)." + }, + { + "left": "issue:44186", + "right": "issue:44610", + "accept": false, + "reason": "Tokenizer crash on NER inputs is unrelated to OmDet-Turbo processor/model image-size mismatch." + }, + { + "left": "issue:44877", + "right": "issue:44991", + "accept": false, + "reason": "Different config/loading bugs: strict config rejects granite_speech, while tokenizer loading fails for EMBEDDIA/est-roberta." + }, + { + "left": "issue:44479", + "right": "issue:44918", + "accept": false, + "reason": "Both mention Qwen and input handling, but one is a video-input regression and the other is a trainer embedding unpacking issue." + }, + { + "left": "issue:36331", + "right": "issue:42915", + "accept": false, + "reason": "Trainer `compute_loss` signature mismatch vs Qwen3Moe FineGrainedFP8Config loading failure." + }, + { + "left": "issue:44671", + "right": "issue:44743", + "accept": false, + "reason": "CamemBERT masked-LM prediction issue is unrelated to Qwen3.5 recurrent state reset in modular code." + }, + { + "left": "issue:36331", + "right": "issue:42617", + "accept": false, + "reason": "Different areas entirely: custom trainer API breakage vs `3d_parallel.py` execution failure." + }, + { + "left": "issue:43316", + "right": "issue:44913", + "accept": false, + "reason": "Both are config-related, but they concern different models and different API/serialization behavior." + }, + { + "left": "issue:43550", + "right": "issue:43606", + "accept": false, + "reason": "Both are runtime/device failures, but one is SDPA/compile and the other is CPU offload device mismatch." + }, + { + "left": "issue:42915", + "right": "issue:43317", + "accept": false, + "reason": "Different loading/offload issues: Qwen3Moe FineGrainedFP8Config vs dequantized model loading with `device_map=auto`." + }, + { + "left": "issue:41720", + "right": "issue:42617", + "accept": false, + "reason": "Qwen3 auto-device-map CUDA assert is unrelated to the `3d_parallel.py` failure." + }, + { + "left": "issue:44292", + "right": "issue:44479", + "accept": false, + "reason": "Different regressions: Qwen-3-8B-NVFP4 load/run failure vs video-input regression across Qwen VL variants." + }, + { + "left": "issue:41720", + "right": "issue:45161", + "accept": false, + "reason": "Both are Qwen/GPT-OSS MoE scaling issues, but one is auto device mapping CUDA assert and the other is tensor-parallel only behavior." + }, + { + "left": "issue:43531", + "right": "issue:43572", + "accept": false, + "reason": "Different models and bugs: Qwen3-MoE sliding-window behavior vs missing `pad_token_idx` in StableLmConfig." + }, + { + "left": "issue:43388", + "right": "issue:45059", + "accept": false, + "reason": "Label gathering in metrics is unrelated to SAM3 text/bounding-box behavior." + }, + { + "left": "issue:44186", + "right": "issue:44361", + "accept": false, + "reason": "Both are tokenizer-related, but they affect different tokenizers and different failure modes." + }, + { + "left": "issue:43540", + "right": "issue:43782", + "accept": false, + "reason": "Different multimodal/Qwen failures: video input processing vs `weight_only=True` load error." + }, + { + "left": "issue:41720", + "right": "issue:43873", + "accept": false, + "reason": "Specific CUDA assert in Qwen3 auto device mapping is not the same as the broader quantization offload issue." + }, + { + "left": "issue:42915", + "right": "issue:43322", + "accept": false, + "reason": "Qwen3Moe FP8 config loading and Llava Next segmentation fault are different bugs." + }, + { + "left": "issue:45406", + "right": "issue:45464", + "accept": false, + "reason": "Both touch serving, but one is a missing `_tokenizer` attribute and the other is a streaming chat/completions failure; different code paths." + }, + { + "left": "issue:36010", + "right": "issue:42617", + "accept": false, + "reason": "Importing `GenerationMixin` and running `3d_parallel.py` are unrelated failures." + }, + { + "left": "issue:36331", + "right": "issue:41720", + "accept": false, + "reason": "Trainer `compute_loss` kwarg breakage is unrelated to Qwen3 device-mapping CUDA asserts." + }, + { + "left": "issue:44514", + "right": "issue:44545", + "accept": true, + "reason": "Same bug: `Qwen2_5_VLProcessor.apply_chat_template` crashes on batched input when `padding=False`, with matching symptom and scope." + }, + { + "left": "issue:33357", + "right": "issue:35707", + "accept": false, + "reason": "MacOS bus error loading a CLIP model is unrelated to progressive generation with `inputs_embeds` and `past_key_values`." + }, + { + "left": "issue:43606", + "right": "issue:43720", + "accept": false, + "reason": "Both involve offload/loading, but the concrete failures are different: device mismatch in CPU offload vs packed-weight unpacking during accelerate loading." + }, + { + "left": "issue:41720", + "right": "issue:43746", + "accept": false, + "reason": "CUDA assert on Qwen3 auto device mapping is unrelated to GraniteSpeech PEFT local checkpoint loading." + }, + { + "left": "issue:43550", + "right": "issue:43653", + "accept": false, + "reason": "Different issues: Bamba torch.compile+SDPA failure vs BigBirdTokenizer mask token registration/decode bug." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded and conservative overall. The only accepted pair, issue:44514 \u2194 issue:44545, is well supported by the packet as the same apply_chat_template batched-input/padding=False crash. The rest of the soft-edge rejections stay appropriately cautious and do not overclaim duplication." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 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-15354-364", + "nodes": [ + "issue:30064", + "issue:34567", + "issue:34634", + "issue:35707", + "issue:39692", + "issue:41720", + "issue:42886", + "issue:43381", + "issue:43421", + "issue:43531", + "issue:43540", + "issue:43596", + "issue:43606", + "issue:43645", + "issue:43653", + "issue:43746", + "issue:43782", + "issue:43844", + "issue:43873", + "issue:43927", + "issue:43994", + "issue:44008", + "issue:44062", + "issue:44077", + "issue:44186", + "issue:44220", + "issue:44368", + "issue:44451", + "issue:44492", + "issue:44509", + "issue:44589", + "issue:44610", + "issue:44625", + "issue:44671", + "issue:44843", + "issue:44898", + "issue:44918", + "issue:44936", + "issue:44987", + "issue:44991", + "issue:45042", + "issue:45072", + "issue:45081", + "issue:45183", + "issue:45356", + "issue:45357", + "issue:45381", + "issue:45479" + ], + "soft_pairs": [ + "issue:30064|issue:34634", + "issue:44368|issue:45357", + "issue:44492|issue:44509", + "issue:44625|issue:45479", + "issue:43606|issue:43653", + "issue:43421|issue:43927", + "issue:44898|issue:45072", + "issue:43782|issue:45381", + "issue:34567|issue:35707", + "issue:34634|issue:41720", + "issue:43381|issue:43844", + "issue:42886|issue:44843", + "issue:44077|issue:44625", + "issue:43421|issue:43653", + "issue:44451|issue:44991", + "issue:44918|issue:44936", + "issue:44062|issue:44451", + "issue:43994|issue:44671", + "issue:44987|issue:45042", + "issue:39692|issue:44610", + "issue:43531|issue:44589", + "issue:43421|issue:44186", + "issue:45081|issue:45356", + "issue:45183|issue:45479", + "issue:43746|issue:45381", + "issue:39692|issue:43873", + "issue:43531|issue:43645", + "issue:43596|issue:44220", + "issue:43746|issue:43873", + "issue:43540|issue:44008" + ], + "prepared_review_unit_hash": "fdc0f11bb5c20218728c0d6d5e0797b0a1db9e9018adcc461a35d61ace1dcbf2", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 37008, + "estimated_input_tokens": 9252, + "estimated_eval_tokens": 18760 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly unrelated issue reports; only the HF_HUB_OFFLINE tokenizer-loading reports (42886 and 44843) look like the same underlying offline-mode bug. The rest are broad-similarity false positives across different models, components, or failure modes.", + "confidence": 0.86, + "canonical_issue_reason": "issue:44843 is the most actionable canonical issue: it names the concrete root cause (`_patch_mistral_regex` calling `model_info()` unconditionally) and matches the offline-tokenizer failure seen in 42886.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44843 is the strongest overall issue candidate because it is specific, reproducible, and root-cause oriented; the paired report 42886 is the broader symptom description.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:30064", + "right": "issue:34634", + "accept": false, + "reason": "Different processors and unrelated bugs: void segmentation maps vs Bark `voice_preset` handling." + }, + { + "left": "issue:44368", + "right": "issue:45357", + "accept": false, + "reason": "Both involve Qwen3.5, but one is a LoRA warning and the other is a save/load key regression; not the same bug." + }, + { + "left": "issue:44492", + "right": "issue:44509", + "accept": false, + "reason": "Both are docs-related, but one is a cache-strategy typo and the other is stale pipeline-task documentation." + }, + { + "left": "issue:44625", + "right": "issue:45479", + "accept": false, + "reason": "Both mention `num_labels`, but one is Qwen3.5 config propagation and the other is a generic classification-loss edge case." + }, + { + "left": "issue:43606", + "right": "issue:43653", + "accept": false, + "reason": "Unrelated subsystems: CPU offload/device mismatch vs BigBird tokenizer special-token decoding." + }, + { + "left": "issue:43421", + "right": "issue:43927", + "accept": false, + "reason": "Both touch tokenization/special-token behavior, but one is runtime post-processor refresh and the other is save/load ID loss." + }, + { + "left": "issue:44898", + "right": "issue:45072", + "accept": false, + "reason": "Different models and failures: Perceiver resolution handling vs bf16 dtype mismatch." + }, + { + "left": "issue:43782", + "right": "issue:45381", + "accept": false, + "reason": "Same broad Qwen VL family, but different concrete paths: `from_pretrained(weight_only=True)` vs video `vision_position_ids`." + }, + { + "left": "issue:34567", + "right": "issue:35707", + "accept": false, + "reason": "Trainer token-count tracking is unrelated to progressive generation with `inputs_embeds`/`past_key_values`." + }, + { + "left": "issue:34634", + "right": "issue:41720", + "accept": false, + "reason": "Bark voice preset loading and Qwen3 auto device mapping are unrelated failures." + }, + { + "left": "issue:43381", + "right": "issue:43844", + "accept": false, + "reason": "Both mention gradients/checkpointing, but one is eval-mode checkpointing misuse and the other is ZeRO-3 gradient abnormality." + }, + { + "left": "issue:42886", + "right": "issue:44843", + "accept": true, + "reason": "Both are offline tokenizer-loading failures; 44843 identifies the concrete root cause in `_patch_mistral_regex`/`model_info()`." + }, + { + "left": "issue:44077", + "right": "issue:44625", + "accept": false, + "reason": "`post_init` validation for patchtsmixer is unrelated to Qwen3.5 `num_labels` propagation." + }, + { + "left": "issue:43421", + "right": "issue:43653", + "accept": false, + "reason": "Special-token handling is too broad here; one is tokenizer backend post-processing, the other is a BigBird mask-token decode bug." + }, + { + "left": "issue:44451", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer/model loading failures, but the affected models and likely root causes differ." + }, + { + "left": "issue:44918", + "right": "issue:44936", + "accept": false, + "reason": "Different trainer bugs: SFT unpacking of embeddings vs evaluate-after-train failure." + }, + { + "left": "issue:44062", + "right": "issue:44451", + "accept": false, + "reason": "AddedToken argument duplication is unrelated to ScandiBERT loading failure." + }, + { + "left": "issue:43994", + "right": "issue:44671", + "accept": false, + "reason": "Both are incorrect model outputs, but for different models and different mechanisms." + }, + { + "left": "issue:44987", + "right": "issue:45042", + "accept": false, + "reason": "A model-loading regression is unrelated to PIL image processors wrongly requiring torchvision." + }, + { + "left": "issue:39692", + "right": "issue:44610", + "accept": false, + "reason": "Both are processor/model mismatch style reports, but they concern different models and different failure modes." + }, + { + "left": "issue:43531", + "right": "issue:44589", + "accept": false, + "reason": "Qwen3-MoE sliding-window behavior is unrelated to the Float8 storage lookup error." + }, + { + "left": "issue:43421", + "right": "issue:44186", + "accept": false, + "reason": "TokenizersBackend special-token refresh is unrelated to LayoutLMv2 NER/padding crash." + }, + { + "left": "issue:45081", + "right": "issue:45356", + "accept": false, + "reason": "Both involve tokenizer regex patching, but one is a crash in `_patch_mistral_regex` and the other is a Kimi-K2.5 regression with codec/warning issues." + }, + { + "left": "issue:45183", + "right": "issue:45479", + "accept": false, + "reason": "Validation-error messaging across models is not the same as the `num_labels=1` zero-loss classification bug." + }, + { + "left": "issue:43746", + "right": "issue:45381", + "accept": false, + "reason": "PEFT local checkpoint loading and Qwen2.5-VL video position IDs are unrelated." + }, + { + "left": "issue:39692", + "right": "issue:43873", + "accept": false, + "reason": "Documentation/example problems are unrelated to quantization/offloading behavior." + }, + { + "left": "issue:43531", + "right": "issue:43645", + "accept": false, + "reason": "Sliding-window Qwen3-MoE behavior is unrelated to Jupyter custom-model initialization." + }, + { + "left": "issue:43596", + "right": "issue:44220", + "accept": false, + "reason": "BertModel ZeRO-3 init crash and `_torch_extract_fbank_features()` are unrelated." + }, + { + "left": "issue:45081", + "right": "issue:45356", + "accept": false, + "reason": "Shared tokenizer-patch area is not enough; the reported failure modes do not match closely enough to merge." + }, + { + "left": "issue:45183", + "right": "issue:45479", + "accept": false, + "reason": "Different bugs in input validation vs classification loss behavior." + }, + { + "left": "issue:43746", + "right": "issue:43873", + "accept": false, + "reason": "PEFT checkpoint loading and quantization/offloading are separate concerns." + }, + { + "left": "issue:43540", + "right": "issue:44008", + "accept": false, + "reason": "Video processing ValueError in Qwen3OmniMoe is unrelated to Gemma 3n variable-name collision in forward()." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The analysis is grounded in the packet and stays conservative overall. The only clearly supported same-bug pair is 42886/44843, which both describe HF_HUB_OFFLINE tokenizer-loading failures and the latter gives a concrete root cause. The remaining soft-edge rejections are plausible from the titles and do not overclaim mergeability." + }, + "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": 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-15354-364", + "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:43650", + "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:43906", + "issue:43927", + "issue:43994", + "issue:44008", + "issue:44062", + "issue:44186", + "issue:44220", + "issue:44442", + "issue:44448", + "issue:44488", + "issue:44561", + "issue:44661", + "issue:44671", + "issue:44811", + "issue:44843", + "issue:44877", + "issue:44898", + "issue:44977", + "issue:45020", + "issue:45072", + "issue:45081", + "issue:45381" + ], + "soft_pairs": [ + "issue:40444|issue:43746", + "issue:43618|issue:43697", + "issue:43716|issue:44008", + "issue:43720|issue:43819", + "issue:44220|issue:44661", + "issue:43653|issue:43720", + "issue:43335|issue:43756", + "issue:43421|issue:44811", + "issue:44186|issue:44898", + "issue:44843|issue:45081", + "issue:43550|issue:45072", + "issue:43550|issue:43927", + "issue:30064|issue:36010", + "issue:43540|issue:45381", + "issue:43577|issue:43818", + "issue:44062|issue:44488", + "issue:43643|issue:44877", + "issue:43531|issue:43643", + "issue:43577|issue:43582", + "issue:43782|issue:44977", + "issue:43644|issue:43749", + "issue:43653|issue:43819", + "issue:44448|issue:44671", + "issue:44561|issue:45020", + "issue:43825|issue:43994", + "issue:43550|issue:43819", + "issue:44186|issue:44442", + "issue:43825|issue:43827", + "issue:43650|issue:43906" + ], + "prepared_review_unit_hash": "9ebbe33777049c14fd1fdc861c20bdc60e31fbe20cc51ec63b29caabbdf9f259", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36561, + "estimated_input_tokens": 9141, + "estimated_eval_tokens": 18538 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a set of unrelated issues that only share broad Transformers topics (tokenizers, multimodal models, config/loading, v5 regressions). No soft pair looks like the same underlying bug or change, so all candidate links should be rejected.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "No single issue is a good canonical representative because the cluster is heterogeneous; if forced, the most triage-worthy/general recent regression is issue:45020, but it is not a true cluster anchor.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:40444", + "right": "issue:43746", + "accept": false, + "reason": "Different problems: Qwen2.5-VL iterable multi-image finetuning failure vs GraniteSpeech PEFT/local checkpoint loading." + }, + { + "left": "issue:43618", + "right": "issue:43697", + "accept": false, + "reason": "CLIPOutput attentions regression and RTDetrV2 output mismatch are unrelated model-specific issues." + }, + { + "left": "issue:43716", + "right": "issue:44008", + "accept": false, + "reason": "Both are dtype-related, but one is Mistral image preprocessor mismatch and the other is a Gemma3n variable-name collision causing AttributeError." + }, + { + "left": "issue:43720", + "right": "issue:43819", + "accept": false, + "reason": "BitNet packed-weight loading bug and DAC forward/from_latents mismatch are different code paths and failures." + }, + { + "left": "issue:44220", + "right": "issue:44661", + "accept": false, + "reason": "Audio fbank extraction bug vs tokenizer mapping/add-new-model-like failure; no shared underlying defect." + }, + { + "left": "issue:43653", + "right": "issue:43720", + "accept": false, + "reason": "BigBirdTokenizer special-token registration bug is unrelated to BitNet packed-weight loading." + }, + { + "left": "issue:43335", + "right": "issue:43756", + "accept": false, + "reason": "SwitchTransformers sparse-layer config bug is not the same as Smollm3 RoPE-layer dropping." + }, + { + "left": "issue:43421", + "right": "issue:44811", + "accept": false, + "reason": "Runtime post-processor update bug vs Whisper batch_decode skip_special_tokens behavior; different tokenizer subsystems." + }, + { + "left": "issue:44186", + "right": "issue:44898", + "accept": false, + "reason": "LayoutLMv2 tokenizer padding/NER crash is unrelated to Perceiver image-classification resolution handling." + }, + { + "left": "issue:44843", + "right": "issue:45081", + "accept": false, + "reason": "HF Hub offline/model_info call in tokenizer patch is distinct from a tokenizer.backend_tokenizer AttributeError." + }, + { + "left": "issue:43550", + "right": "issue:45072", + "accept": false, + "reason": "Both mention dtype/SDPA, but one is a torch.compile+SDPA model bug and the other is bfloat16 inference dtype mismatches across models." + }, + { + "left": "issue:43550", + "right": "issue:43927", + "accept": false, + "reason": "Bamba torch.compile/SDPA failure is unrelated to DiaConfig save/load losing custom token IDs." + }, + { + "left": "issue:30064", + "right": "issue:36010", + "accept": false, + "reason": "ImageProcessor void segmentation maps and GenerationMixin import error are unrelated." + }, + { + "left": "issue:43540", + "right": "issue:45381", + "accept": false, + "reason": "Qwen3OmniMoe video-input ValueError and Qwen2.5-VL vision_position_ids issue are different multimodal bugs." + }, + { + "left": "issue:43577", + "right": "issue:43818", + "accept": false, + "reason": "BLIP2 dtype loading issue is unrelated to Video-LLaVA missing temporal attention / weight sharing." + }, + { + "left": "issue:44062", + "right": "issue:44488", + "accept": false, + "reason": "AddedToken special-argument TypeError and model loading failure for cjvt/sleng-bert are unrelated tokenizer/model issues." + }, + { + "left": "issue:43643", + "right": "issue:44877", + "accept": false, + "reason": "AutoConfig trust_remote_code missing fields is a different strict-config problem than granite_speech config loading." + }, + { + "left": "issue:43531", + "right": "issue:43643", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior is unrelated to AutoConfig trust_remote_code field loss." + }, + { + "left": "issue:43577", + "right": "issue:43582", + "accept": false, + "reason": "Both are dtype-related, but one is BLIP2 loading dtype and the other is an Apple Silicon caching_allocator_warmup TypeError." + }, + { + "left": "issue:43782", + "right": "issue:44977", + "accept": false, + "reason": "Qwen3VL weight_only loading error is distinct from Qwen3.5 flash-attention generation failure." + }, + { + "left": "issue:43644", + "right": "issue:43749", + "accept": false, + "reason": "Non-persistent buffer junk in v5 is unrelated to FSDP_CPU_RAM_EFFICIENT_LOADING being broken." + }, + { + "left": "issue:43653", + "right": "issue:43819", + "accept": false, + "reason": "BigBirdTokenizer mask token registration bug is unrelated to DAC latent/forward mismatch." + }, + { + "left": "issue:44448", + "right": "issue:44671", + "accept": false, + "reason": "Pegasus v4/v5 output differences and CamemBERT masked LM prediction changes are separate model regressions." + }, + { + "left": "issue:44561", + "right": "issue:45020", + "accept": false, + "reason": "Both involve recent v5/remote_code breakage, but one is is_torch_fx_available removal and the other is broader remote_code failures." + }, + { + "left": "issue:43825", + "right": "issue:43994", + "accept": false, + "reason": "pipeline() translation messaging bug and SigLIP2 nonsensical outputs are not the same defect." + }, + { + "left": "issue:43550", + "right": "issue:43819", + "accept": false, + "reason": "Bamba torch.compile+SDPA and DAC forward/from_latents mismatch are unrelated model behaviors." + }, + { + "left": "issue:44186", + "right": "issue:44442", + "accept": false, + "reason": "LayoutLMv2 tokenizer crash and AutoTokenizer failing to load FastSpeech2ConformerTokenizer are unrelated tokenizer paths." + }, + { + "left": "issue:43825", + "right": "issue:43827", + "accept": false, + "reason": "Both concern translation/pipeline v5 removals, but one is an error message mismatch and the other is stale docs; not the same bug/change." + }, + { + "left": "issue:43650", + "right": "issue:43906", + "accept": false, + "reason": "Placeholder issue \"ADD THE DATA\" is unrelated to the isolated reproduction of issue 38071." + } + ] + }, + "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-15354-364", + "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:43653", + "issue:43697", + "issue:43720", + "issue:43746", + "issue:43749", + "issue:43782", + "issue:43819", + "issue:43854", + "issue:44060", + "issue:44291", + "issue:44368", + "issue:44485", + "issue:44509", + "issue:44554", + "issue:44568", + "issue:44610", + "issue:44671", + "issue:44805", + "issue:44811", + "issue:44821", + "issue:44857", + "issue:44912", + "issue:44936", + "issue:44977", + "issue:44991", + "issue:45072", + "issue:45081", + "issue:45084", + "issue:45325", + "issue:45406", + "issue:45488", + "issue:45538" + ], + "soft_pairs": [ + "issue:44821|issue:45084", + "issue:41720|issue:44991", + "issue:43606|issue:45072", + "issue:44805|issue:44936", + "issue:30064|issue:36331", + "issue:42175|issue:45081", + "issue:44610|issue:45072", + "issue:43782|issue:45325", + "issue:44485|issue:44509", + "issue:44060|issue:44368", + "issue:34634|issue:36010", + "issue:43720|issue:45072", + "issue:43262|issue:43819", + "issue:45081|issue:45406", + "issue:43749|issue:44291", + "issue:43582|issue:44857", + "issue:43653|issue:44671", + "issue:43526|issue:43697", + "issue:43526|issue:44554", + "issue:43653|issue:44912", + "issue:45084|issue:45538", + "issue:44568|issue:45488", + "issue:43454|issue:43577", + "issue:34634|issue:36331", + "issue:43746|issue:44811", + "issue:43257|issue:43329", + "issue:43638|issue:44936", + "issue:43452|issue:44991", + "issue:43854|issue:44977" + ], + "prepared_review_unit_hash": "3881be57a9c574d953353a41981739d5f105a7073ce506d0305c8dc8a5ccb619", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36933, + "estimated_input_tokens": 9234, + "estimated_eval_tokens": 18724 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: the items span unrelated bugs in tokenizers, multimodal processors, loading paths, training code, and model-specific runtime failures. None of the soft-pair candidates look like the same underlying defect or change.", + "confidence": 0.95, + "canonical_issue_reason": "No single canonical issue fits this set; the reports are about different models and code paths rather than one shared bug.", + "canonical_pr_reason": null, + "best_issue_reason": "No issue is a clear cluster representative because there is no common underlying problem to normalize around.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44821", + "right": "issue:45084", + "accept": false, + "reason": "Different failures: URL image processor loading vs a template compilation TypeError." + }, + { + "left": "issue:41720", + "right": "issue:44991", + "accept": false, + "reason": "Unrelated model/runtime bug vs tokenizer-loading regression." + }, + { + "left": "issue:43606", + "right": "issue:45072", + "accept": false, + "reason": "Both are inference mismatches, but they affect different subsystems and models; not the same bug." + }, + { + "left": "issue:44805", + "right": "issue:44936", + "accept": false, + "reason": "Mask shape IndexError during model use vs trainer lifecycle failure after training." + }, + { + "left": "issue:30064", + "right": "issue:36331", + "accept": false, + "reason": "Image segmentation-map processing bug is unrelated to CustomTrainer loss-signature incompatibility." + }, + { + "left": "issue:42175", + "right": "issue:45081", + "accept": false, + "reason": "Packaging/backend selection issue vs Mistral tokenizer regex patch crash." + }, + { + "left": "issue:44610", + "right": "issue:45072", + "accept": false, + "reason": "Processor image-size mismatch vs dtype mismatch; different code paths." + }, + { + "left": "issue:43782", + "right": "issue:45325", + "accept": false, + "reason": "Qwen3VL load error is unrelated to Qwen2.5-VL rope-position scaling." + }, + { + "left": "issue:44485", + "right": "issue:44509", + "accept": false, + "reason": "RoPE implementation discussion is unrelated to docs mentioning removed pipeline tasks." + }, + { + "left": "issue:44060", + "right": "issue:44368", + "accept": false, + "reason": "Both mention tied-weights warnings, but one is a concrete wrong tie mapping and the other is a fine-tuning warning behavior; not clearly the same defect." + }, + { + "left": "issue:34634", + "right": "issue:36010", + "accept": false, + "reason": "Voice preset handling in BarkProcessor is unrelated to GenerationMixin import failure." + }, + { + "left": "issue:43720", + "right": "issue:45072", + "accept": false, + "reason": "Packed-weight unpacking bug is unrelated to bfloat16 dtype mismatches." + }, + { + "left": "issue:43262", + "right": "issue:43819", + "accept": false, + "reason": "Audio chat-template sampling-rate default bug is unrelated to DAC STE mismatch." + }, + { + "left": "issue:45081", + "right": "issue:45406", + "accept": false, + "reason": "Mistral tokenizer backend access crash is unrelated to Gemma4Processor missing _tokenizer." + }, + { + "left": "issue:43749", + "right": "issue:44291", + "accept": false, + "reason": "Different loading failures: FSDP RAM-efficient loading vs init_empty_weights extra-argument handling." + }, + { + "left": "issue:43582", + "right": "issue:44857", + "accept": false, + "reason": "Apple Silicon warmup TypeError is unrelated to CUDA float16 AMP crash." + }, + { + "left": "issue:43653", + "right": "issue:44671", + "accept": false, + "reason": "BigBirdTokenizer special-token registration bug is not the same as CamemBERT MLM prediction regression." + }, + { + "left": "issue:43526", + "right": "issue:43697", + "accept": false, + "reason": "BeitImageProcessorFast label reduction bug and RTDetrV2 output differences are distinct issues." + }, + { + "left": "issue:43526", + "right": "issue:44554", + "accept": false, + "reason": "Image label reduction bug is unrelated to an MPS attention correctness issue." + }, + { + "left": "issue:43653", + "right": "issue:44912", + "accept": false, + "reason": "Tokenizer special-token decode bug is unrelated to MXFP4 quantization loading fallback." + }, + { + "left": "issue:45084", + "right": "issue:45538", + "accept": false, + "reason": "Template-node compilation error is unrelated to CLIPTokenizer model_max_length behavior." + }, + { + "left": "issue:44568", + "right": "issue:45488", + "accept": false, + "reason": "Different tokenizer regressions on different model families and code paths." + }, + { + "left": "issue:43454", + "right": "issue:43577", + "accept": false, + "reason": "AyaVision weight-tying bug is unrelated to Blip2 dtype propagation." + }, + { + "left": "issue:34634", + "right": "issue:36331", + "accept": false, + "reason": "Duplicate of unrelated pairing; no shared underlying bug." + }, + { + "left": "issue:43746", + "right": "issue:44811", + "accept": false, + "reason": "Local checkpoint loading with PEFT adapters is unrelated to Whisper batch_decode skip_special_tokens handling." + }, + { + "left": "issue:43257", + "right": "issue:43329", + "accept": false, + "reason": "Qwen3 MOE loading/weight conversion bug is unrelated to multimodal token counting." + }, + { + "left": "issue:43638", + "right": "issue:44936", + "accept": false, + "reason": "Zero3 empty-index crash is not the same as evaluate-after-train behavior." + }, + { + "left": "issue:43452", + "right": "issue:44991", + "accept": false, + "reason": "gguf_file path handling and est-roberta tokenizer loading are different failures." + }, + { + "left": "issue:43854", + "right": "issue:44977", + "accept": false, + "reason": "GLM-4.7-Flash unit-test loading issue is unrelated to Qwen3.5 flash-attention generation." + } + ] + }, + "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-15354-364", + "nodes": [ + "issue:33357", + "issue:34567", + "issue:36010", + "issue:36331", + "issue:38175", + "issue:39692", + "issue:41720", + "issue:42175", + "issue:42915", + "issue:43257", + "issue:43262", + "issue:43317", + "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:43906", + "issue:43927", + "issue:43931", + "issue:44291", + "issue:44479", + "issue:44560", + "issue:44568", + "issue:44589", + "issue:44610", + "issue:44756", + "issue:44811", + "issue:44912", + "issue:44938", + "issue:44987", + "issue:45042", + "issue:45072", + "issue:45356", + "issue:45357", + "issue:45381", + "issue:45538" + ], + "soft_pairs": [ + "issue:38175|issue:44610", + "issue:44589|issue:45042", + "issue:43257|issue:43696", + "issue:39692|issue:44912", + "issue:43819|issue:43927", + "issue:43262|issue:44811", + "issue:43720|issue:44610", + "issue:43550|issue:44912", + "issue:36010|issue:43645", + "issue:44568|issue:45356", + "issue:43749|issue:44756", + "issue:43329|issue:44479", + "issue:43257|issue:43931", + "issue:43531|issue:45042", + "issue:36010|issue:36331", + "issue:43329|issue:43450", + "issue:43257|issue:43756", + "issue:43577|issue:45072", + "issue:43697|issue:45357", + "issue:43452|issue:43582", + "issue:42175|issue:44938", + "issue:41720|issue:43540", + "issue:44479|issue:45381", + "issue:43927|issue:44912", + "issue:43540|issue:44560", + "issue:43540|issue:44479", + "issue:33357|issue:34567", + "issue:44987|issue:45538", + "issue:42915|issue:43742", + "issue:43317|issue:45538", + "issue:44291|issue:45042", + "issue:43906|issue:45381" + ], + "prepared_review_unit_hash": "cc6f5f98bea4fdf34eafd804d61fe03d873f30c87031b4dbf93473a6d70784e1", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37400, + "estimated_input_tokens": 9350, + "estimated_eval_tokens": 18956 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a set of unrelated issues; the only clear duplicate-style area is the Qwen video-input regression family. I\u2019d anchor the cluster on the broad v5.3.0 video regression report and only link the closest Qwen video-processing reports to it.", + "confidence": 0.7, + "canonical_issue_reason": "issue:44479 is the broadest and most central report in the only coherent subcluster here: a Transformers v5.3.0 video-input regression affecting multiple Qwen VL/MoE variants.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44479 best represents the shared underlying problem space because it describes a versioned regression with multiple affected model families and the same video-input code path.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:38175", + "right": "issue:44610", + "accept": false, + "reason": "Different models and failures: SIGLIP2 zero probabilities vs OmDet-Turbo image-size mismatch." + }, + { + "left": "issue:44589", + "right": "issue:45042", + "accept": false, + "reason": "Unrelated backend issues: storage-object/float8 error vs unnecessary torchvision requirement in PIL image processors." + }, + { + "left": "issue:43257", + "right": "issue:43696", + "accept": false, + "reason": "Different problems: Qwen3 MoE weight conversion under accelerate/deepspeed vs GPT-oss CUDA OOM." + }, + { + "left": "issue:39692", + "right": "issue:44912", + "accept": false, + "reason": "Docs/example mistakes for SigLIP2 are not the same bug as MXFP4 loading falling back to bf16." + }, + { + "left": "issue:43819", + "right": "issue:43927", + "accept": false, + "reason": "DAC latent conversion mismatch and DiaConfig save/load token-ID loss are separate model/config bugs." + }, + { + "left": "issue:43262", + "right": "issue:44811", + "accept": false, + "reason": "Both are audio-related, but one is chat-template sampling-rate defaulting and the other is Whisper batch_decode token skipping." + }, + { + "left": "issue:43720", + "right": "issue:44610", + "accept": false, + "reason": "Packed-weight unloading during accelerate loading is unrelated to the OmDet-Turbo processor size mismatch." + }, + { + "left": "issue:43550", + "right": "issue:44912", + "accept": false, + "reason": "torch.compile/SDPA on Bamba is a different code path from MXFP4 quantization loading." + }, + { + "left": "issue:36010", + "right": "issue:43645", + "accept": false, + "reason": "Importing GenerationMixin from transformers.generation is not the same as the Jupyter/custom-model init regression." + }, + { + "left": "issue:44568", + "right": "issue:45356", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different models and different failure modes." + }, + { + "left": "issue:43749", + "right": "issue:44756", + "accept": false, + "reason": "One is FSDP RAM-efficient loading, the other is Apple Silicon mmap/OOM behavior; related theme, not same bug." + }, + { + "left": "issue:43329", + "right": "issue:44479", + "accept": true, + "reason": "Both concern Qwen video-input handling in the same multimodal code path; one exposes the video-branch bug, the other reports the broader v5.3.0 regression." + }, + { + "left": "issue:43257", + "right": "issue:43931", + "accept": false, + "reason": "Different loading problems: MoE weight conversion under distributed loading vs Qwen3-VL shape mismatch." + }, + { + "left": "issue:43531", + "right": "issue:45042", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior is unrelated to PIL image processors requiring torchvision." + }, + { + "left": "issue:36010", + "right": "issue:36331", + "accept": false, + "reason": "ImportError for GenerationMixin and compute_loss signature mismatch are separate API breakages." + }, + { + "left": "issue:43329", + "right": "issue:43450", + "accept": false, + "reason": "Undefined video-token helper names and batched video processor output shape are different defects." + }, + { + "left": "issue:43257", + "right": "issue:43756", + "accept": false, + "reason": "Qwen3 MoE weight conversion and Smollm3 RoPE-layer count are unrelated model-specific issues." + }, + { + "left": "issue:43577", + "right": "issue:45072", + "accept": false, + "reason": "BLIP2 dtype propagation and SwitchTransformers/TimmWrapper bfloat16 mismatches are not the same bug." + }, + { + "left": "issue:43697", + "right": "issue:45357", + "accept": false, + "reason": "RTDetrV2 output divergence and Qwen3.5 save_pretrained key regression are unrelated." + }, + { + "left": "issue:43452", + "right": "issue:43582", + "accept": false, + "reason": "gguf_file loader breakage and Apple Silicon caching_allocator_warmup TypeError are separate issues." + }, + { + "left": "issue:42175", + "right": "issue:44938", + "accept": false, + "reason": "Missing TensorFlow in torch-only install and Python 3.14 load failure do not share the same root cause." + }, + { + "left": "issue:41720", + "right": "issue:43540", + "accept": false, + "reason": "Qwen3 auto device-map cuda assert and Qwen3OmniMoe video ValueError are different failure paths." + }, + { + "left": "issue:44479", + "right": "issue:45381", + "accept": true, + "reason": "Both are the same Qwen2.5-VL video-input regression class in v5.3.0, with 45381 specifically calling out wrong vision_position_ids." + }, + { + "left": "issue:43927", + "right": "issue:44912", + "accept": false, + "reason": "DiaConfig token-ID persistence and MXFP4 loading fallback are unrelated." + }, + { + "left": "issue:43540", + "right": "issue:44560", + "accept": false, + "reason": "Different Qwen video bugs: one is a ValueError in Qwen3OmniMoe video processing, the other is a StopIteration regression in Qwen3-VL embedding." + }, + { + "left": "issue:43540", + "right": "issue:44479", + "accept": false, + "reason": "Both involve Qwen video input, but the reported failures and model families are different enough to avoid merging." + }, + { + "left": "issue:33357", + "right": "issue:34567", + "accept": false, + "reason": "MacOS bus error with CLIP and TrainerState token counting are unrelated." + }, + { + "left": "issue:44987", + "right": "issue:45538", + "accept": false, + "reason": "Loading physical-intelligence/fast and CLIPTokenizer max-length handling are separate problems." + }, + { + "left": "issue:42915", + "right": "issue:43742", + "accept": false, + "reason": "Qwen3MoE FP8 config failure and MobileLLM key error do not share a concrete code path." + }, + { + "left": "issue:43317", + "right": "issue:45538", + "accept": false, + "reason": "Dequantized model loading with device_map=auto is unrelated to CLIPTokenizer max_length." + }, + { + "left": "issue:44291", + "right": "issue:45042", + "accept": false, + "reason": "init_empty_weights/_is_hf_initialized and PIL image processor torchvision dependency are unrelated." + }, + { + "left": "issue:43906", + "right": "issue:45381", + "accept": false, + "reason": "A reproduction of another issue is not the same as the Qwen2.5-VL video-position regression unless the underlying bug is shown to match, which isn\u2019t evident here." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: most pairs are indeed unrelated, and the two accepted Qwen video-input links are the only ones that plausibly share a root cause. The soft-edge rejections stay conservative and do not overstate equivalence." + }, + "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-15354-364", + "nodes": [ + "issue:28282", + "issue:29127", + "issue:34567", + "issue:34689", + "issue:36010", + "issue:36331", + "issue:38175", + "issue:42175", + "issue:42915", + "issue:43452", + "issue:43493", + "issue:43526", + "issue:43577", + "issue:43582", + "issue:43643", + "issue:43646", + "issue:43688", + "issue:43697", + "issue:43742", + "issue:43749", + "issue:43756", + "issue:43782", + "issue:43824", + "issue:43994", + "issue:44062", + "issue:44291", + "issue:44484", + "issue:44488", + "issue:44514", + "issue:44561", + "issue:44589", + "issue:44704", + "issue:44792", + "issue:44821", + "issue:44898", + "issue:44964", + "issue:44991", + "issue:45042", + "issue:45081", + "issue:45137", + "issue:45216", + "issue:45290", + "issue:45325", + "issue:45375", + "issue:45381", + "issue:45405", + "issue:45479", + "issue:45538" + ], + "soft_pairs": [ + "issue:45137|issue:45405", + "issue:42915|issue:45538", + "issue:43526|issue:45479", + "issue:42175|issue:44488", + "issue:42175|issue:43824", + "issue:42915|issue:44821", + "issue:44062|issue:44991", + "issue:38175|issue:44898", + "issue:28282|issue:34689", + "issue:43582|issue:43782", + "issue:43646|issue:43749", + "issue:43493|issue:43994", + "issue:43643|issue:45375", + "issue:43756|issue:45325", + "issue:28282|issue:29127", + "issue:43742|issue:44291", + "issue:36010|issue:42175", + "issue:44484|issue:44704", + "issue:34567|issue:36331", + "issue:44291|issue:44821", + "issue:44792|issue:44964", + "issue:43452|issue:44488", + "issue:44062|issue:45081", + "issue:43688|issue:43749", + "issue:44514|issue:45290", + "issue:43643|issue:44561", + "issue:43452|issue:43577", + "issue:44514|issue:45381", + "issue:43742|issue:44589", + "issue:43697|issue:45216", + "issue:45042|issue:45081" + ], + "prepared_review_unit_hash": "5c3027d17cb4ca17045ff57f6e82d29720a89d5b6ebf9693b341f155b3813b39", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37197, + "estimated_input_tokens": 9300, + "estimated_eval_tokens": 18856 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: most items are isolated Transformer bugs/regressions around different models, loaders, and training paths rather than duplicates of one underlying issue. I would not merge any of the soft pairs as duplicates.", + "confidence": 0.94, + "canonical_issue_reason": "If one issue must anchor the cluster, issue 43646 is the broadest and clearest regression report (`Transformers 5.0.0 breaks custom model initialization`). It is still only a weak canonical because the rest of the set spans unrelated model-loading, tokenizer, training, and processor bugs.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43646 is the best representative issue because it is a general version-regression report with a concrete failure mode and broader applicability than the more model-specific reports. That said, it is not a strong duplicate hub for the whole set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45137", + "right": "issue:45405", + "accept": false, + "reason": "DeepSpeed ZeRO3 deque error and PEFT version pinning are unrelated problems." + }, + { + "left": "issue:42915", + "right": "issue:45538", + "accept": false, + "reason": "FP8 config failure for Qwen3Moe is unrelated to CLIPTokenizer's extreme max_length value." + }, + { + "left": "issue:43526", + "right": "issue:45479", + "accept": false, + "reason": "BeitImageProcessorFast label reduction bug and sequence-classification zero-loss bug are different code paths." + }, + { + "left": "issue:42175", + "right": "issue:44488", + "accept": false, + "reason": "TensorFlow backend packaging issue is not the same as a model-loading failure for `cjvt/sleng-bert`." + }, + { + "left": "issue:42175", + "right": "issue:43824", + "accept": false, + "reason": "Missing TensorFlow backend and missing Qwen2_5_VL export are unrelated import/setup issues." + }, + { + "left": "issue:42915", + "right": "issue:44821", + "accept": false, + "reason": "FP8 config handling and AutoImageProcessor URL loading are different bugs." + }, + { + "left": "issue:44062", + "right": "issue:44991", + "accept": false, + "reason": "AddedToken keyword collision and tokenizer loading failure for est-roberta are separate tokenizer issues." + }, + { + "left": "issue:38175", + "right": "issue:44898", + "accept": false, + "reason": "SigLIP2 zero probabilities and Perceiver non-default resolution failure are different model-specific inference bugs." + }, + { + "left": "issue:28282", + "right": "issue:34689", + "accept": false, + "reason": "PyTorch missing import error and Llama 3.2 Vision loading regression are unrelated." + }, + { + "left": "issue:43582", + "right": "issue:43782", + "accept": false, + "reason": "AppleSilicon warmup TypeError and Qwen3VL weight_only loading error do not share the same failure mechanism." + }, + { + "left": "issue:43646", + "right": "issue:43749", + "accept": false, + "reason": "Custom model initialization regression and FSDP CPU RAM efficient loading breakage are different issues." + }, + { + "left": "issue:43493", + "right": "issue:43994", + "accept": false, + "reason": "SigLIP2 implementation discrepancy and bad AutoModel/pipeline outputs are both about the model, but not the same concrete bug." + }, + { + "left": "issue:43643", + "right": "issue:45375", + "accept": false, + "reason": "Both concern missing config fields, but one is trust_remote_code object hydration and the other is a specific Qwen config field dropped by strict parsing." + }, + { + "left": "issue:43756", + "right": "issue:45325", + "accept": false, + "reason": "Smollm3 RoPE layer count mismatch and Qwen2.5-VL temporal position scaling are different architecture/position encoding problems." + }, + { + "left": "issue:28282", + "right": "issue:29127", + "accept": false, + "reason": "PyTorch import error and LayoutLMv3 box-content error are unrelated error-message/validation issues." + }, + { + "left": "issue:43742", + "right": "issue:44291", + "accept": false, + "reason": "MobileLLM loading key error and init_empty_weights unexpected `_is_hf_initialized` argument are different loading regressions." + }, + { + "left": "issue:36010", + "right": "issue:42175", + "accept": false, + "reason": "GenerationMixin import failure and TensorFlow backend packaging issue are separate import/export problems." + }, + { + "left": "issue:44484", + "right": "issue:44704", + "accept": false, + "reason": "max_shard_size default question and AutoProcessor kwargs propagation bug are unrelated save/load behavior issues." + }, + { + "left": "issue:34567", + "right": "issue:36331", + "accept": false, + "reason": "TrainerState token counting and CustomTrainer.compute_loss signature mismatch are different training-trainer bugs." + }, + { + "left": "issue:44291", + "right": "issue:44821", + "accept": false, + "reason": "Unexpected init_empty_weights argument and AutoImageProcessor URL loading do not describe the same bug." + }, + { + "left": "issue:44792", + "right": "issue:44964", + "accept": false, + "reason": "Janus test failure and Phi-4 multimodal loading failure are separate model-specific regressions." + }, + { + "left": "issue:43452", + "right": "issue:44488", + "accept": false, + "reason": "gguf_file support breakage and inability to load `cjvt/sleng-bert` are different load-path failures." + }, + { + "left": "issue:44062", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer AddedToken duplication and Mistral regex patch backend-tokenizer crash are distinct tokenizer bugs." + }, + { + "left": "issue:43688", + "right": "issue:43749", + "accept": false, + "reason": "Auxiliary loss normalization and FSDP CPU RAM efficient loading are unrelated training-time bugs." + }, + { + "left": "issue:44514", + "right": "issue:45290", + "accept": false, + "reason": "Batched Qwen2_5_VL chat template crash and assistant tool-call crash are both chat-template issues but not the same concrete edge case." + }, + { + "left": "issue:43643", + "right": "issue:44561", + "accept": false, + "reason": "Missing config fields under trust_remote_code and v5 removal of `is_torch_fx_available` are separate upgrade breakages." + }, + { + "left": "issue:43452", + "right": "issue:43577", + "accept": false, + "reason": "gguf loading breakage and BLIP2 dtype not propagating are different model-loading defects." + }, + { + "left": "issue:44514", + "right": "issue:45381", + "accept": false, + "reason": "Both mention Qwen2.5-VL chat/video behavior, but one is chat-template batching and the other is incorrect vision position IDs." + }, + { + "left": "issue:43742", + "right": "issue:44589", + "accept": false, + "reason": "MobileLLM key error and missing Float8 storage are unrelated load/runtime compatibility problems." + }, + { + "left": "issue:43697", + "right": "issue:45216", + "accept": false, + "reason": "RTDetrV2 output drift and Qwen3.5 checkpoint save corruption are different regression classes." + }, + { + "left": "issue:45042", + "right": "issue:45081", + "accept": false, + "reason": "PIL backend requiring torchvision and Mistral regex backend-tokenizer crash are not 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": 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-15354-364", + "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:43452", + "issue:43525", + "issue:43531", + "issue:43577", + "issue:43643", + "issue:43653", + "issue:43688", + "issue:43697", + "issue:43720", + "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:44821", + "issue:44843", + "issue:44877", + "issue:44898", + "issue:44936", + "issue:44977", + "issue:44991", + "issue:45005", + "issue:45325", + "issue:45356", + "issue:45440" + ], + "soft_pairs": [ + "issue:43531|issue:44821", + "issue:44625|issue:44877", + "issue:43819|issue:44811", + "issue:43720|issue:44898", + "issue:43688|issue:45440", + "issue:43653|issue:44568", + "issue:44464|issue:44743", + "issue:43452|issue:44843", + "issue:44466|issue:45005", + "issue:41762|issue:44936", + "issue:43295|issue:45356", + "issue:24643|issue:29127", + "issue:40444|issue:44811", + "issue:41720|issue:43906", + "issue:43525|issue:43643", + "issue:44361|issue:44442", + "issue:43577|issue:43697", + "issue:38175|issue:43295", + "issue:43653|issue:44361", + "issue:44843|issue:44991", + "issue:28282|issue:32090", + "issue:28282|issue:33357", + "issue:28282|issue:35141", + "issue:44479|issue:44936", + "issue:43122|issue:43906", + "issue:44977|issue:45356", + "issue:44464|issue:44977", + "issue:30064|issue:33666", + "issue:43653|issue:43927", + "issue:43577|issue:44554", + "issue:43329|issue:45325" + ], + "prepared_review_unit_hash": "3a9d58759f0771d81ff339fd4bdde70b0f7b1987bccab8a44715aed6ffb45fef", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37383, + "estimated_input_tokens": 9346, + "estimated_eval_tokens": 18948 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a false-positive mix of unrelated issues across tokenizers, configs, multimodal processors, training, and device/runtime regressions. The soft pairs only share broad subsystem words, not the same concrete bug/change, so all soft edges should be rejected.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue fits this cluster; the items are heterogeneous and do not share one underlying defect.", + "canonical_pr_reason": null, + "best_issue_reason": "No strong best issue representative exists because the cluster is not a true duplicate set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43531", + "right": "issue:44821", + "accept": false, + "reason": "Qwen3-MoE sliding-window behavior and AutoImageProcessor URL loading are unrelated bugs." + }, + { + "left": "issue:44625", + "right": "issue:44877", + "accept": false, + "reason": "num_labels propagation in Qwen3.5 is separate from granite_speech strict-config loading." + }, + { + "left": "issue:43819", + "right": "issue:44811", + "accept": false, + "reason": "DAC latent decoding and Whisper batch_decode special-token handling are different code paths." + }, + { + "left": "issue:43720", + "right": "issue:44898", + "accept": false, + "reason": "BitNet packed-weight loading and Perceiver interpolation failure are distinct model bugs." + }, + { + "left": "issue:43688", + "right": "issue:45440", + "accept": false, + "reason": "Auxiliary-loss normalization and DeepseekV3MoE divergence are not the same defect." + }, + { + "left": "issue:43653", + "right": "issue:44568", + "accept": false, + "reason": "Both involve tokens, but BigBird mask-token registration and mdeberta BOS/EOS insertion are different issues." + }, + { + "left": "issue:44464", + "right": "issue:44743", + "accept": false, + "reason": "Compiled-forward chunked generation and Qwen3_5 recurrent-state reset are separate generation bugs." + }, + { + "left": "issue:43452", + "right": "issue:44843", + "accept": false, + "reason": "gguf_file loading breakage and offline model_info calls in mistral regex patch are different tokenizer-loading failures." + }, + { + "left": "issue:44466", + "right": "issue:45005", + "accept": false, + "reason": "Device-dependent lm_head serialization and translation-model tied-weight problems are not clearly the same bug." + }, + { + "left": "issue:41762", + "right": "issue:44936", + "accept": false, + "reason": "ZeRO-3 model loading IndexError and trainer.evaluate after train are unrelated trainer/runtime issues." + }, + { + "left": "issue:43295", + "right": "issue:45356", + "accept": false, + "reason": "Processor.tokenizer regression and Kimi-K2.5 codec/tokenizer regression are separate changes." + }, + { + "left": "issue:24643", + "right": "issue:29127", + "accept": false, + "reason": "DeepSpeed weight shape RuntimeError and LayoutLMV3 box error messaging are unrelated." + }, + { + "left": "issue:40444", + "right": "issue:44811", + "accept": false, + "reason": "IterableDataset multi-image finetuning failure is unrelated to Whisper batch_decode behavior." + }, + { + "left": "issue:41720", + "right": "issue:43906", + "accept": false, + "reason": "Qwen3 auto device-mapping cudaErrorAssert and an isolated reproduction reference are not the same concrete bug." + }, + { + "left": "issue:43525", + "right": "issue:43643", + "accept": false, + "reason": "Missing pad_token_id in Llama4Config and missing fields from trust_remote_code config loading are different config issues." + }, + { + "left": "issue:44361", + "right": "issue:44442", + "accept": false, + "reason": "MLukeTokenizer task AttributeError and FastSpeech2ConformerTokenizer loading failure are unrelated tokenizer bugs." + }, + { + "left": "issue:43577", + "right": "issue:43697", + "accept": false, + "reason": "Blip2 dtype casting and RTDetrV2 output drift in v5 are different model regressions." + }, + { + "left": "issue:38175", + "right": "issue:43295", + "accept": false, + "reason": "Siglip2 zero probabilities and processor.tokenizer regression are different problem classes." + }, + { + "left": "issue:43653", + "right": "issue:43927", + "accept": false, + "reason": "BigBird special-token registration and DiaConfig custom token IDs are separate token-management bugs." + }, + { + "left": "issue:44843", + "right": "issue:44991", + "accept": false, + "reason": "Offline model_info misuse and EMBEDDIA/est-roberta tokenizer loading failure are not the same underlying issue." + }, + { + "left": "issue:28282", + "right": "issue:32090", + "accept": false, + "reason": "Missing PyTorch import and Trainer NoneType broadcast error are unrelated." + }, + { + "left": "issue:28282", + "right": "issue:33357", + "accept": false, + "reason": "PyTorch not found and MacOS bus error in CLIP are unrelated runtime failures." + }, + { + "left": "issue:28282", + "right": "issue:35141", + "accept": false, + "reason": "PyTorch import error and tied embedding reinitialization are different bugs." + }, + { + "left": "issue:44479", + "right": "issue:44936", + "accept": false, + "reason": "Qwen video-input regression and trainer.evaluate after train are unrelated." + }, + { + "left": "issue:43122", + "right": "issue:43906", + "accept": false, + "reason": "Tokenizer output changes across versions and an isolated reproduction link do not establish the same bug." + }, + { + "left": "issue:44977", + "right": "issue:45356", + "accept": false, + "reason": "Qwen3.5 flash-attention generation issue and Kimi-K2.5 tokenizer regression are different failures." + }, + { + "left": "issue:44464", + "right": "issue:44977", + "accept": false, + "reason": "Chunked-generation compiled-forward inconsistency and flash-attention generation failure are separate runtime issues." + }, + { + "left": "issue:30064", + "right": "issue:33666", + "accept": false, + "reason": "Void segmentation map processing and Qwen2-VL multi-GPU training are unrelated." + }, + { + "left": "issue:43653", + "right": "issue:43927", + "accept": false, + "reason": "This pair is still not the same bug: special-token registration vs config token ID persistence." + }, + { + "left": "issue:43577", + "right": "issue:44554", + "accept": false, + "reason": "BLIP2 dtype loading and MPS attention correctness on mismatched head dims are unrelated." + }, + { + "left": "issue:43329", + "right": "issue:45325", + "accept": false, + "reason": "Undefined video-branch helper variables and still-image rope-index scaling are different multimodal 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": 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-15354-364", + "nodes": [ + "issue:32090", + "issue:33666", + "issue:34567", + "issue:36010", + "issue:39692", + "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:43531", + "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:44610", + "issue:44625", + "issue:44821", + "issue:44849", + "issue:44913", + "issue:44964", + "issue:44977", + "issue:45084", + "issue:45200", + "issue:45229", + "issue:45357", + "issue:45405" + ], + "soft_pairs": [ + "issue:43531|issue:43742", + "issue:44479|issue:44625", + "issue:43550|issue:44361", + "issue:41720|issue:44464", + "issue:43504|issue:43720", + "issue:43012|issue:43577", + "issue:43504|issue:43550", + "issue:32090|issue:33666", + "issue:32090|issue:34567", + "issue:43262|issue:45084", + "issue:43504|issue:44610", + "issue:43278|issue:43688", + "issue:44849|issue:45357", + "issue:43452|issue:44451", + "issue:45200|issue:45229", + "issue:43334|issue:43572", + "issue:43756|issue:44913", + "issue:43493|issue:43825", + "issue:43540|issue:43716", + "issue:33666|issue:36010", + "issue:43742|issue:44821", + "issue:43606|issue:43927", + "issue:43720|issue:43927", + "issue:43525|issue:43572", + "issue:43688|issue:43873", + "issue:43278|issue:43844", + "issue:43257|issue:43299", + "issue:43720|issue:44361", + "issue:43540|issue:44977", + "issue:39692|issue:43493", + "issue:44964|issue:45405", + "issue:43329|issue:43606" + ], + "prepared_review_unit_hash": "571b94dd7510861df4962f309e4d828f0876b3ea44551413acc84c11a8e90bd4", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37256, + "estimated_input_tokens": 9314, + "estimated_eval_tokens": 18884 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is not a duplicate set; it\u2019s a broad mix of unrelated Transformers bugs across loading, dtype, multimodal, training, and config serialization paths. All soft pairs should be rejected.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue stands out because the items span distinct models and code paths rather than one underlying bug.", + "canonical_pr_reason": null, + "best_issue_reason": "No strong global representative issue: the apparent similarities are mostly broad subsystem overlap, not the same defect.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43531", + "right": "issue:43742", + "accept": false, + "reason": "Both are load/config failures, but they affect different models and symptoms; no shared code-path bug is evident." + }, + { + "left": "issue:44479", + "right": "issue:44625", + "accept": false, + "reason": "Video-input regression vs num_labels propagation bug; unrelated behavior and fixes." + }, + { + "left": "issue:43550", + "right": "issue:44361", + "accept": false, + "reason": "torch.compile/SDPA failure is unrelated to a tokenizer AttributeError on tasks." + }, + { + "left": "issue:41720", + "right": "issue:44464", + "accept": false, + "reason": "CUDA assert during auto device mapping is a different bug from inconsistent chunked generation under compiled forward." + }, + { + "left": "issue:43504", + "right": "issue:43720", + "accept": false, + "reason": "Legacy-field loading issue in Beit is unrelated to BitNet packed-weight unpacking during accelerate loading." + }, + { + "left": "issue:43012", + "right": "issue:43577", + "accept": false, + "reason": "Both mention dtype, but one is a compile-time warning and the other is a load-time dtype propagation bug for a different model." + }, + { + "left": "issue:43504", + "right": "issue:43550", + "accept": false, + "reason": "Different models and failure modes; no same underlying defect." + }, + { + "left": "issue:32090", + "right": "issue:33666", + "accept": false, + "reason": "Both touch multi-GPU/training, but one is a Trainer broadcast TypeError and the other is a Qwen2-VL training topic." + }, + { + "left": "issue:32090", + "right": "issue:34567", + "accept": false, + "reason": "Trainer broadcast NoneType error and token-count tracking bug are separate Trainer issues." + }, + { + "left": "issue:43262", + "right": "issue:45084", + "accept": false, + "reason": "Audio chat-template sampling-rate default is unrelated to compile-time non-template-node errors." + }, + { + "left": "issue:43504", + "right": "issue:44610", + "accept": false, + "reason": "Pretrained loading with a legacy field is unrelated to processor/model resolution-size mismatch." + }, + { + "left": "issue:43278", + "right": "issue:43688", + "accept": false, + "reason": "Dtype drift in eval is unrelated to auxiliary-loss normalization." + }, + { + "left": "issue:44849", + "right": "issue:45357", + "accept": false, + "reason": "Hidden-states output bug and incorrect save_pretrained visual keys are different Qwen3.5 issues." + }, + { + "left": "issue:43452", + "right": "issue:44451", + "accept": false, + "reason": "Both are model-loading failures, but one is gguf_file-specific and the other is a different model regression." + }, + { + "left": "issue:45200", + "right": "issue:45229", + "accept": false, + "reason": "mm_token_type_ids defaulting and multi-GPU CUDA OOM are unrelated Gemma 4 problems." + }, + { + "left": "issue:43334", + "right": "issue:43572", + "accept": false, + "reason": "Missing pad_token_id vs missing pad_token_idx are similar in wording, but they are different config fields on different models." + }, + { + "left": "issue:43756", + "right": "issue:44913", + "accept": false, + "reason": "RoPE-layer dropping and rotary_pct reload regression both involve configuration, but they are not the same bug." + }, + { + "left": "issue:43493", + "right": "issue:43825", + "accept": false, + "reason": "SigLIP2 implementation discrepancy is unrelated to a pipeline error message about supported translation tasks." + }, + { + "left": "issue:43540", + "right": "issue:43716", + "accept": false, + "reason": "Qwen3OmniMoe video-processing failure and Mistral-3 dtype mismatch are separate multimodal issues." + }, + { + "left": "issue:33666", + "right": "issue:36010", + "accept": false, + "reason": "Qwen2-VL multi-GPU training and GenerationMixin import failure have no shared code-path." + }, + { + "left": "issue:43742", + "right": "issue:44821", + "accept": false, + "reason": "MobileLLM loading key error and AutoImageProcessor URL loading are unrelated." + }, + { + "left": "issue:43606", + "right": "issue:43927", + "accept": false, + "reason": "CPU offload device mismatch is unrelated to DiaConfig save/load token-ID loss." + }, + { + "left": "issue:43720", + "right": "issue:43927", + "accept": false, + "reason": "Packed-weight unpacking during accelerate loading is unrelated to DiaConfig custom-token-ID persistence." + }, + { + "left": "issue:43525", + "right": "issue:43572", + "accept": false, + "reason": "Both are missing-token-id config regressions, but they affect different fields and models; not clearly one defect." + }, + { + "left": "issue:43688", + "right": "issue:43873", + "accept": false, + "reason": "Aux-loss normalization and quantization/offload behavior are unrelated." + }, + { + "left": "issue:43278", + "right": "issue:43844", + "accept": false, + "reason": "Eval dtype drift is unrelated to ZeRO-3 gradient inflation." + }, + { + "left": "issue:43257", + "right": "issue:43299", + "accept": false, + "reason": "Both involve Qwen3 MoE loading, but one is accelerate/deepspeed weight conversion and the other is a transformers-dev regression on Qwen3VL MoE; too different to merge." + }, + { + "left": "issue:43720", + "right": "issue:44361", + "accept": false, + "reason": "BitNet loading bug and MLukeTokenizer task AttributeError are unrelated." + }, + { + "left": "issue:43540", + "right": "issue:44977", + "accept": false, + "reason": "Video-input processing failure and flash-attention generation bug are different Qwen3-family issues." + }, + { + "left": "issue:39692", + "right": "issue:43493", + "accept": false, + "reason": "Both are SigLIP2-related, but one is a docs/example mismatch and the other is an implementation discrepancy vs JAX." + }, + { + "left": "issue:44964", + "right": "issue:45405", + "accept": false, + "reason": "Phi-4 multimodal loading failure is unrelated to the PEFT minimum-version packaging issue." + }, + { + "left": "issue:43329", + "right": "issue:43606", + "accept": false, + "reason": "Undefined video-token helper/merge-size bug is unrelated to Bark CPU-offload device mismatch." + } + ] + }, + "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-15354-364", + "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:44488", + "issue:44493", + "issue:44521", + "issue:44554", + "issue:44589", + "issue:44734", + "issue:44792", + "issue:44843", + "issue:44977", + "issue:45084", + "issue:45200", + "issue:45290", + "issue:45405", + "issue:45446", + "issue:45520" + ], + "soft_pairs": [ + "issue:33666|issue:36331", + "issue:43493|issue:43844", + "issue:44589|issue:45520", + "issue:44488|issue:44843", + "issue:43756|issue:44060", + "issue:43646|issue:44466", + "issue:43688|issue:43746", + "issue:44521|issue:45290", + "issue:44423|issue:44734", + "issue:33357|issue:33666", + "issue:43824|issue:44589", + "issue:39692|issue:43295", + "issue:43994|issue:44493", + "issue:43504|issue:43606", + "issue:43504|issue:43653", + "issue:43927|issue:44361", + "issue:43575|issue:45405", + "issue:43572|issue:44291", + "issue:43278|issue:43701", + "issue:44792|issue:45405", + "issue:43257|issue:45405", + "issue:43606|issue:44112", + "issue:15354|issue:38175", + "issue:44554|issue:45446", + "issue:43582|issue:44977", + "issue:45084|issue:45520", + "issue:43540|issue:45200" + ], + "prepared_review_unit_hash": "ee560bf93c09400dc64cea186b4e338c70173e41e2d42536e016b6913be9ddae", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35428, + "estimated_input_tokens": 8857, + "estimated_eval_tokens": 17970 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is largely heterogeneous: most items are unrelated model-specific bugs, regressions, or docs/CI issues. None of the soft pairs look like true duplicates, though a few are superficially similar by subsystem.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43646 is the broadest regression-style report in the set, but even it does not subsume the other problems; there is no strong single canonical issue for this cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:33666", + "right": "issue:36331", + "accept": false, + "reason": "Different bugs: multi-GPU training behavior vs a Trainer API signature mismatch." + }, + { + "left": "issue:43493", + "right": "issue:43844", + "accept": false, + "reason": "Unrelated failures: SigLIP2 output discrepancy vs ZeRO-3 gradient growth during training." + }, + { + "left": "issue:44589", + "right": "issue:45520", + "accept": false, + "reason": "Different code paths: missing Float8 storage handling vs flash_attn import lookup on Python 3.13." + }, + { + "left": "issue:44488", + "right": "issue:44843", + "accept": false, + "reason": "Both involve model loading/tokenizer flow, but one is a specific model load failure and the other is offline hub access in a tokenizer patch." + }, + { + "left": "issue:43756", + "right": "issue:44060", + "accept": false, + "reason": "Different model-architecture issues: RoPE layer dropping vs a tied-weights warning bug." + }, + { + "left": "issue:43646", + "right": "issue:44466", + "accept": false, + "reason": "Both are v5-related, but one is custom model initialization and the other is device-dependent serialization of lm_head.weight." + }, + { + "left": "issue:43688", + "right": "issue:43746", + "accept": false, + "reason": "Completely different failures: auxiliary-loss normalization vs PEFT adapter checkpoint loading." + }, + { + "left": "issue:44521", + "right": "issue:45290", + "accept": false, + "reason": "Same API surface, but different bugs: zero assistant masks in multimodal inputs vs tool-call messages with no content crashing." + }, + { + "left": "issue:44423", + "right": "issue:44734", + "accept": false, + "reason": "Both are serving crashes, but one is multimodal input handling and the other is KV-cache continuation indexing." + }, + { + "left": "issue:33357", + "right": "issue:33666", + "accept": false, + "reason": "Different problem domains: MacOS bus error in CLIP loading vs multi-GPU training." + }, + { + "left": "issue:43824", + "right": "issue:44589", + "accept": false, + "reason": "Import error for a missing model class is unrelated to the Float8 storage-object lookup failure." + }, + { + "left": "issue:39692", + "right": "issue:43295", + "accept": false, + "reason": "Docs/example failures and processor/tokenizer mismatch are not the same as the v4.57.5 processor.tokenizer regression." + }, + { + "left": "issue:43994", + "right": "issue:44493", + "accept": false, + "reason": "Different symptoms and likely causes: SigLIP2 nonsensical outputs vs a position-id key warning affecting many models." + }, + { + "left": "issue:43504", + "right": "issue:43606", + "accept": false, + "reason": "One is a legacy-field loading bug in BEiT, the other is a CPU offload device-mismatch in Bark." + }, + { + "left": "issue:43504", + "right": "issue:43653", + "accept": false, + "reason": "Legacy-field loading for BEiT is unrelated to BigBirdTokenizer special-token registration." + }, + { + "left": "issue:43927", + "right": "issue:44361", + "accept": false, + "reason": "Different tokenizer/config issues: saved token IDs being lost vs MLukeTokenizer AttributeError on tasks." + }, + { + "left": "issue:43575", + "right": "issue:45405", + "accept": false, + "reason": "OOM when loading a large model with tensor parallelism is unrelated to a PEFT version pin issue." + }, + { + "left": "issue:43572", + "right": "issue:44291", + "accept": false, + "reason": "Missing config field after update vs an init_empty_weights signature change are different regressions." + }, + { + "left": "issue:43278", + "right": "issue:43701", + "accept": false, + "reason": "Different workflows: dtype change during evaluate vs resume_from_checkpoint key mismatch." + }, + { + "left": "issue:44792", + "right": "issue:45405", + "accept": false, + "reason": "A janus image-generation test failure is unrelated to the PEFT minimum-version issue." + }, + { + "left": "issue:43257", + "right": "issue:45405", + "accept": false, + "reason": "Qwen3 MoE weight conversion with accelerate/deepspeed is unrelated to the PEFT version bump." + }, + { + "left": "issue:43606", + "right": "issue:44112", + "accept": false, + "reason": "A real runtime device-mismatch bug is not the same as a stale GraniteSpeech CI test." + }, + { + "left": "issue:15354", + "right": "issue:38175", + "accept": false, + "reason": "TorchScript GeneratorExp export failure and SigLIP2 zero-probabilities are unrelated bugs." + }, + { + "left": "issue:44554", + "right": "issue:45446", + "accept": false, + "reason": "MPS attention correctness and a PyTorch version check/import guard are different issues." + }, + { + "left": "issue:43582", + "right": "issue:44977", + "accept": false, + "reason": "Apple Silicon allocator TypeError is unrelated to Qwen3.5 flash-attention generation behavior." + }, + { + "left": "issue:45084", + "right": "issue:45520", + "accept": false, + "reason": "Compilation of non-template nodes and flash_attn import lookup are unrelated failures." + }, + { + "left": "issue:43540", + "right": "issue:45200", + "accept": false, + "reason": "Video-input processing in Qwen3OmniMoe is unrelated to mm_token_type_ids defaults for text-only fine-tuning." + } + ] + }, + "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-15354-364", + "nodes": [ + "issue:37428", + "issue:38175", + "issue:39401", + "issue:40990", + "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:45072", + "issue:45127", + "issue:45198", + "issue:45200", + "issue:45325", + "issue:45356", + "issue:45412" + ], + "soft_pairs": [ + "issue:43232|issue:44849", + "issue:45072|issue:45198", + "issue:43844|issue:44964", + "issue:39401|issue:44568", + "issue:44315|issue:44849", + "issue:37428|issue:43957", + "issue:44554|issue:45356", + "issue:43295|issue:44610", + "issue:43334|issue:43525", + "issue:43493|issue:43701", + "issue:45200|issue:45325", + "issue:44898|issue:45412", + "issue:43873|issue:44464", + "issue:44466|issue:45127", + "issue:40990|issue:44625", + "issue:43388|issue:43526", + "issue:43844|issue:44936", + "issue:44514|issue:44521", + "issue:43575|issue:44464", + "issue:37428|issue:43824", + "issue:43295|issue:43450", + "issue:43572|issue:44987", + "issue:44442|issue:45198", + "issue:43572|issue:44589", + "issue:43824|issue:43975", + "issue:43295|issue:43873", + "issue:38175|issue:45412", + "issue:43278|issue:43856", + "issue:43295|issue:43994", + "issue:44360|issue:44561" + ], + "prepared_review_unit_hash": "788281c3c88c96a744b42a5e889fd43e1014bc81949fad8c8f70e5c5aed94b04", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36892, + "estimated_input_tokens": 9223, + "estimated_eval_tokens": 18702 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a heterogeneous mix of unrelated issue reports across tokenizers, processors, loading, training, and dtype/memory behavior. None of the soft pairs looks like the same underlying bug/change, so there is no clear canonical item.", + "confidence": 0.91, + "canonical_issue_reason": "No single issue clearly subsumes the others; the items span distinct components and failure modes.", + "canonical_pr_reason": "No PRs are present in this cluster.", + "best_issue_reason": "No issue is a strong global representative because the cluster is not centered on one bug class or code path.", + "best_pr_reason": "No PRs are present in this cluster.", + "soft_edge_verdicts": [ + { + "left": "issue:43232", + "right": "issue:44849", + "accept": false, + "reason": "Different problems: generation kwarg sync after sync_gpus vs Qwen3.5 hidden-states behavior." + }, + { + "left": "issue:45072", + "right": "issue:45198", + "accept": false, + "reason": "Different model paths and symptoms: bfloat16 dtype mismatches vs Wav2Vec2 save/tokenization failure." + }, + { + "left": "issue:43844", + "right": "issue:44964", + "accept": false, + "reason": "ZeRO-3 gradient growth and a Phi-4 loading failure are unrelated bugs." + }, + { + "left": "issue:39401", + "right": "issue:44568", + "accept": false, + "reason": "Different tokenizer bugs: wrong offset mapping vs add_special_tokens not inserting BOS/EOS." + }, + { + "left": "issue:44315", + "right": "issue:44849", + "accept": false, + "reason": "Liger Kernel application in Trainer setup is unrelated to a Qwen3.5 hidden-states bug." + }, + { + "left": "issue:37428", + "right": "issue:43957", + "accept": false, + "reason": "Importing a flash-attention helper and meta-device model loading are separate failure modes." + }, + { + "left": "issue:44554", + "right": "issue:45356", + "accept": false, + "reason": "MPS attention correctness and Kimi-K2.5 tokenizer codec handling are not the same bug." + }, + { + "left": "issue:43295", + "right": "issue:44610", + "accept": false, + "reason": "Processor/tokenizer API regression vs incorrect image size emitted by a video/vision processor." + }, + { + "left": "issue:43334", + "right": "issue:43525", + "accept": false, + "reason": "Both mention missing pad_token_id, but for different models/config paths and likely separate fixes." + }, + { + "left": "issue:43493", + "right": "issue:43701", + "accept": false, + "reason": "SigLIP2 implementation discrepancy and resume_from_checkpoint key mismatch are unrelated." + }, + { + "left": "issue:45200", + "right": "issue:45325", + "accept": false, + "reason": "Gemma 4 mm_token_type_ids defaulting and Qwen2.5-VL temporal position scaling are different issues." + }, + { + "left": "issue:44898", + "right": "issue:45412", + "accept": false, + "reason": "Perceiver resize failure and RT-DETR memory release are unrelated." + }, + { + "left": "issue:43873", + "right": "issue:44464", + "accept": false, + "reason": "Quantization offloading behavior and compiled-forward chunked generation are different code paths." + }, + { + "left": "issue:44466", + "right": "issue:45127", + "accept": false, + "reason": "Serialization of lm_head.weight is not the same as LoRA merge collapse with extended vocab." + }, + { + "left": "issue:40990", + "right": "issue:44625", + "accept": false, + "reason": "High perplexity on gpt-oss-20b and Qwen3.5 num_labels propagation are unrelated." + }, + { + "left": "issue:43388", + "right": "issue:43526", + "accept": false, + "reason": "Trainer metric-gathering tuple handling differs from BeitImageProcessorFast reduce_labels behavior." + }, + { + "left": "issue:43844", + "right": "issue:44936", + "accept": false, + "reason": "ZeRO-3 gradient instability is unrelated to trainer.evaluate() failing after train()." + }, + { + "left": "issue:44514", + "right": "issue:44521", + "accept": false, + "reason": "Both touch Qwen2.5-VL chat templating, but one is a batched padding=False crash and the other all-zero assistant masks; not clearly one fix." + }, + { + "left": "issue:43575", + "right": "issue:44464", + "accept": false, + "reason": "TP load OOM for Qwen2-57B is unrelated to compiled-forward chunked generation inconsistencies." + }, + { + "left": "issue:37428", + "right": "issue:43824", + "accept": false, + "reason": "Different import errors for different symbols and models; no shared underlying change." + }, + { + "left": "issue:43295", + "right": "issue:43450", + "accept": false, + "reason": "Processor/tokenizer regression and batched video-shape handling are distinct issues." + }, + { + "left": "issue:43572", + "right": "issue:44987", + "accept": false, + "reason": "StableLmConfig pad_token_idx regression and a generic model-loading failure are not the same bug." + }, + { + "left": "issue:44442", + "right": "issue:45198", + "accept": false, + "reason": "FastSpeech2ConformerTokenizer loading and Wav2Vec2 save/tokenization failure are unrelated." + }, + { + "left": "issue:38175", + "right": "issue:45412", + "accept": false, + "reason": "SigLIP2 zero-probability outputs and RT-DETR memory cleanup are unrelated." + }, + { + "left": "issue:43278", + "right": "issue:43856", + "accept": false, + "reason": "Embedding dtype drift between train/eval is not the same as Qwen3 MoE memory inefficiency." + }, + { + "left": "issue:43572", + "right": "issue:44589", + "accept": false, + "reason": "Missing pad_token_idx in StableLmConfig and Float8 storage lookup failure are different regressions." + }, + { + "left": "issue:43824", + "right": "issue:43975", + "accept": false, + "reason": "Importing Qwen2.5-VL classes and DeepSeek-Coder detokenization are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:43873", + "accept": false, + "reason": "Processor/tokenizer regression and quantization offloading behavior are different subsystems." + }, + { + "left": "issue:43295", + "right": "issue:43994", + "accept": false, + "reason": "Processor/tokenizer regression and SigLIP2 nonsensical outputs are unrelated." + }, + { + "left": "issue:44360", + "right": "issue:44561", + "accept": false, + "reason": "The DSA indexer ReLU discussion and removal of is_torch_fx_available breaking trust_remote_code models 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": 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-15354-364", + "nodes": [ + "issue:29942", + "issue:32090", + "issue:37428", + "issue:38175", + "issue:39692", + "issue:42915", + "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:43824", + "issue:43825", + "issue:43827", + "issue:43901", + "issue:43931", + "issue:44112", + "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" + ], + "soft_pairs": [ + "issue:44849|issue:45003", + "issue:43827|issue:43901", + "issue:38175|issue:43493", + "issue:43653|issue:44112", + "issue:42915|issue:43824", + "issue:43450|issue:43493", + "issue:44568|issue:44991", + "issue:45127|issue:45245", + "issue:45290|issue:45381", + "issue:43819|issue:45083", + "issue:43450|issue:43575", + "issue:44462|issue:44561", + "issue:44561|issue:44779", + "issue:44610|issue:44857", + "issue:43819|issue:44112", + "issue:43262|issue:43450", + "issue:44448|issue:44743", + "issue:43606|issue:45083", + "issue:43278|issue:43825", + "issue:29942|issue:32090", + "issue:45310|issue:45478", + "issue:43611|issue:43931", + "issue:37428|issue:43425", + "issue:43335|issue:43550", + "issue:43819|issue:44493", + "issue:39692|issue:43526", + "issue:44855|issue:45083", + "issue:43450|issue:43526", + "issue:37428|issue:44351", + "issue:45200|issue:45381", + "issue:43404|issue:45200", + "issue:39692|issue:44493", + "issue:43335|issue:43653" + ], + "prepared_review_unit_hash": "3a02c0a826efc7240621ce8e20845cec08d60545a5cc7061a1e5c5413abe69c3", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37640, + "estimated_input_tokens": 9410, + "estimated_eval_tokens": 19076 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly heterogeneous and should not be collapsed as one cluster. The only clear duplicate is the Qwen3.5 MoE from_pretrained regression pair (issue:45310 and issue:45478); the rest span unrelated bugs, docs issues, and model-specific failures.", + "confidence": 0.94, + "canonical_issue_reason": "issue:45478 is the best canonical candidate for the only real duplicate pair: it reports the same Qwen3.5 MoE from_pretrained failure as issue:45310, with matching title/symptom and the later report likely being the duplicate.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45478 is the strongest issue representative in this set because it matches the concrete failure mode most directly; however, the broader collection is not a single deduplicable bug cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44849", + "right": "issue:45003", + "accept": false, + "reason": "Different problems: Qwen3.5 hidden-state output bug vs unsafe sys.modules access in modeling_utils." + }, + { + "left": "issue:43827", + "right": "issue:43901", + "accept": false, + "reason": "Both are docs-related, but they concern different pipeline/deprecation topics and different sections." + }, + { + "left": "issue:38175", + "right": "issue:43493", + "accept": false, + "reason": "Both mention SigLIP2, but one is unexpected zero probabilities and the other is a broader HF vs JAX implementation discrepancy." + }, + { + "left": "issue:43653", + "right": "issue:44112", + "accept": false, + "reason": "Unrelated CI failures: BigBirdTokenizer special-token decode issue vs stale device override test in GraniteSpeech." + }, + { + "left": "issue:42915", + "right": "issue:43824", + "accept": false, + "reason": "Different failures: FineGrainedFP8Config for Qwen3Moe vs missing Qwen2_5_VL import." + }, + { + "left": "issue:43450", + "right": "issue:43493", + "accept": false, + "reason": "Different subsystems: batched video processor shape bug vs SigLIP2 implementation mismatch." + }, + { + "left": "issue:44568", + "right": "issue:44991", + "accept": false, + "reason": "Both tokenizer-related, but one is missing BOS/EOS addition and the other is tokenizer loading for a specific repo." + }, + { + "left": "issue:45127", + "right": "issue:45245", + "accept": false, + "reason": "Unrelated issues: LoRA merge collapse with extended vocabulary vs category cardinality runtime error." + }, + { + "left": "issue:45290", + "right": "issue:45381", + "accept": false, + "reason": "Different chat-template/video bugs: tool-call assistant content crash vs qwen2.5-vl video vision_position_ids mismatch." + }, + { + "left": "issue:43819", + "right": "issue:45083", + "accept": false, + "reason": "Different DAC/qwen3_omni_moe bugs with no shared concrete code path." + }, + { + "left": "issue:43450", + "right": "issue:43575", + "accept": false, + "reason": "Unrelated: video processor batch shape bug vs tensor-parallel OOM when loading Qwen2-57B-A14B-Instruct." + }, + { + "left": "issue:44462", + "right": "issue:44561", + "accept": false, + "reason": "Different regressions: tokenizer.json ignored vs removal of is_torch_fx_available breaking trust_remote_code models." + }, + { + "left": "issue:44561", + "right": "issue:44779", + "accept": false, + "reason": "Both tokenizer/model compatibility issues in v5, but they affect different components and failure modes." + }, + { + "left": "issue:44610", + "right": "issue:44857", + "accept": false, + "reason": "Different model/loss paths: OmDet-Turbo processor input sizing vs LwDetrImageLoss AMP crash." + }, + { + "left": "issue:43819", + "right": "issue:44112", + "accept": false, + "reason": "No common bug: DAC latent mismatch vs GraniteSpeech CI device override failure." + }, + { + "left": "issue:43262", + "right": "issue:43450", + "accept": false, + "reason": "Audio chat-template sampling-rate default issue vs video processor batch-shape bug." + }, + { + "left": "issue:44448", + "right": "issue:44743", + "accept": false, + "reason": "Unrelated: Pegasus output change across versions vs qwen3_5 modular recurrent-state reset bug." + }, + { + "left": "issue:43606", + "right": "issue:45083", + "accept": false, + "reason": "Device-mismatch CI bug from CPU offload vs qwen3_omni_moe helper output-length behavior." + }, + { + "left": "issue:43278", + "right": "issue:43825", + "accept": false, + "reason": "Different failures: embedding dtype drift in eval vs pipeline() error message/deprecation behavior." + }, + { + "left": "issue:29942", + "right": "issue:32090", + "accept": false, + "reason": "Both are old training/attention issues, but the symptoms and code paths differ completely." + }, + { + "left": "issue:45310", + "right": "issue:45478", + "accept": true, + "reason": "Same concrete bug: both report the Qwen3.5 MoE from_pretrained error on transformers>=5.4.0, with matching symptom and scope." + }, + { + "left": "issue:43611", + "right": "issue:43931", + "accept": false, + "reason": "Different loading regressions: base_model_prefix breakage vs weight-shape mismatch for Qwen3-VL-30B-A3B-Instruct." + }, + { + "left": "issue:37428", + "right": "issue:43425", + "accept": false, + "reason": "Both are import/version compatibility issues, but they concern different missing symbols and different release breakages." + }, + { + "left": "issue:43335", + "right": "issue:43550", + "accept": false, + "reason": "Different model/config bug classes: sparse layer construction vs torch.compile SDPA failure." + }, + { + "left": "issue:43819", + "right": "issue:44493", + "accept": false, + "reason": "Unrelated issues: DAC latent/forward mismatch vs unexpected position-id key warnings across models." + }, + { + "left": "issue:39692", + "right": "issue:43526", + "accept": false, + "reason": "Different processor bugs: SigLIP2 docs/model mismatch vs BeitImageProcessorFast label reduction returning only one label." + }, + { + "left": "issue:44855", + "right": "issue:45083", + "accept": false, + "reason": "Python 3.13 import parsing problem in DebertaV2Model vs qwen3_omni_moe feature-extraction helper behavior." + }, + { + "left": "issue:43450", + "right": "issue:43526", + "accept": false, + "reason": "Different processor issues with no shared underlying code path." + }, + { + "left": "issue:37428", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but for different missing names and different modules." + }, + { + "left": "issue:45200", + "right": "issue:45381", + "accept": false, + "reason": "Different Gemma 4 text-finetuning token-type handling vs qwen2.5-vl video position-id bug." + }, + { + "left": "issue:43404", + "right": "issue:45200", + "accept": false, + "reason": "Different model classes and failures: tied lm_head in Mistral3 vs Gemma 4 mm_token_type_ids defaults." + }, + { + "left": "issue:39692", + "right": "issue:44493", + "accept": false, + "reason": "Distinct issues: SigLIP2 example/model mismatch vs generic position-id warnings." + }, + { + "left": "issue:43335", + "right": "issue:43653", + "accept": false, + "reason": "Different bugs: sparse layer creation in SwitchTransformers vs BigBirdTokenizer special-token decode issue." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded and conservative. The only accepted duplicate pair is well-supported by identical Qwen3.5 MoE from_pretrained symptoms in issue:45310 and issue:45478, and the other soft-edge rejections stay on the safe side by distinguishing different model/component failure modes." + }, + "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-15354-364", + "nodes": [ + "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:43701", + "issue:43720", + "issue:43723", + "issue:43873", + "issue:43976", + "issue:43994", + "issue:44112", + "issue:44292", + "issue:44448", + "issue:44451", + "issue:44462", + "issue:44466", + "issue:44493", + "issue:44554", + "issue:44568", + "issue:44779", + "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:44977|issue:45229", + "issue:29127|issue:32090", + "issue:44448|issue:44779", + "issue:38175|issue:44493", + "issue:44292|issue:45520", + "issue:43335|issue:43504", + "issue:43335|issue:45083", + "issue:43550|issue:44112", + "issue:43335|issue:45072", + "issue:44568|issue:45491", + "issue:43720|issue:44112", + "issue:40444|issue:43526", + "issue:44448|issue:44977", + "issue:43723|issue:44554", + "issue:43335|issue:44112", + "issue:34689|issue:35141", + "issue:44857|issue:45072", + "issue:43606|issue:44857", + "issue:45372|issue:45406", + "issue:43066|issue:44462", + "issue:43873|issue:44493", + "issue:43873|issue:45229", + "issue:43278|issue:43329", + "issue:43701|issue:45083", + "issue:29942|issue:35141", + "issue:43550|issue:45083", + "issue:41720|issue:45229", + "issue:32090|issue:33357", + "issue:43976|issue:45440", + "issue:37428|issue:44987", + "issue:44451|issue:45520", + "issue:45372|issue:45459", + "issue:44448|issue:44466", + "issue:43065|issue:43994", + "issue:43065|issue:45412" + ], + "prepared_review_unit_hash": "ae007732c4b90060721950d68784da7ee2e627f9dde439f8f486bae921c83b5a", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 35, + "serialized_chars": 38314, + "estimated_input_tokens": 9579, + "estimated_eval_tokens": 19414 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The soft pairs are mostly broad semantic neighbors, not true duplicates: they cover unrelated model families, tokenizer regressions, device/offload bugs, and CI/test failures. I found no pair that clearly shares the same underlying bug or change.", + "confidence": 0.84, + "canonical_issue_reason": "No strong duplicate anchor exists because the cluster is heterogeneous; if a representative must be chosen, issue 45229 is the most detailed and active, but it is not a good dedupe hub for the rest.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45229 is the best representative only in the sense that it is active and well-described; it is not semantically close enough to the others to serve as a true canonical duplicate.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44977", + "right": "issue:45229", + "accept": false, + "reason": "Different models and failure modes: Qwen3.5 flash-attention generation vs Gemma4 multi-GPU OOM." + }, + { + "left": "issue:29127", + "right": "issue:32090", + "accept": false, + "reason": "Different code paths: LayoutLMv3 box validation messaging vs Trainer GPU broadcast NoneType error." + }, + { + "left": "issue:44448", + "right": "issue:44779", + "accept": false, + "reason": "Both are v5 behavior regressions, but they affect different tokenization/generation flows and different models." + }, + { + "left": "issue:38175", + "right": "issue:44493", + "accept": false, + "reason": "Different symptoms: zero probabilities in SigLIP2 vs unexpected position-id key serialization/deserialization issue." + }, + { + "left": "issue:44292", + "right": "issue:45520", + "accept": false, + "reason": "Different root causes: NVFP4 model runtime error vs Python 3.13 flash_attn import-key error." + }, + { + "left": "issue:43335", + "right": "issue:43504", + "accept": false, + "reason": "Different subsystems and bugs: SwitchTransformers sparse-layer config vs BEiT pretrained-load legacy field." + }, + { + "left": "issue:43335", + "right": "issue:45083", + "accept": false, + "reason": "Different models and logic: SwitchTransformers config bug vs qwen3_omni_moe feature-length helper bug." + }, + { + "left": "issue:43550", + "right": "issue:44112", + "accept": false, + "reason": "Unrelated: Bamba SDPA/torch.compile failure vs GraniteSpeech stale CI test." + }, + { + "left": "issue:43335", + "right": "issue:45072", + "accept": false, + "reason": "Different issue types: SwitchTransformers layer construction vs dtype mismatch in inference wrappers." + }, + { + "left": "issue:44568", + "right": "issue:45491", + "accept": false, + "reason": "Different regressions: tokenizer special-token handling vs Gemma3 NaN embeddings from sliding-window attention." + }, + { + "left": "issue:43720", + "right": "issue:44112", + "accept": false, + "reason": "Different code paths: packed-weight unpacking during accelerate loading vs GraniteSpeech CI device-override test." + }, + { + "left": "issue:40444", + "right": "issue:43526", + "accept": false, + "reason": "Different processors and failures: Qwen2.5-VL iterable dataset multi-image finetuning vs BEiT Fast label reduction bug." + }, + { + "left": "issue:44448", + "right": "issue:44977", + "accept": false, + "reason": "Different models and symptoms: Pegasus v4/v5 output difference vs Qwen3.5 flash-attention generation failure." + }, + { + "left": "issue:43723", + "right": "issue:44554", + "accept": false, + "reason": "Tokenizer loading regression vs MPS attention correctness issue; no shared bug path." + }, + { + "left": "issue:43335", + "right": "issue:44112", + "accept": false, + "reason": "Different subsystems: SwitchTransformers config construction vs GraniteSpeech CI test flake." + }, + { + "left": "issue:34689", + "right": "issue:35141", + "accept": false, + "reason": "Both involve loading/init behavior, but one is Llama 3.2 Vision model loading and the other is embedding reinitialization after resize." + }, + { + "left": "issue:44857", + "right": "issue:45072", + "accept": false, + "reason": "Different failures: CUDA AMP crash in LwDetrImageLoss vs bfloat16 dtype mismatch in model wrappers." + }, + { + "left": "issue:43606", + "right": "issue:44857", + "accept": false, + "reason": "Different runtime problems: CPU offload device mismatch vs float16 AMP CUDA crash." + }, + { + "left": "issue:45372", + "right": "issue:45406", + "accept": false, + "reason": "Both touch Gemma4, but one is a dependency import error and the other is a missing processor attribute at serve time." + }, + { + "left": "issue:43066", + "right": "issue:44462", + "accept": false, + "reason": "Tokenizer-related, but one is decoder-type selection in v5 and the other is AutoTokenizer ignoring repository tokenizer.json." + }, + { + "left": "issue:43873", + "right": "issue:44493", + "accept": false, + "reason": "Unrelated: quantization/offload behavior vs unexpected position-id key warnings." + }, + { + "left": "issue:43873", + "right": "issue:45229", + "accept": false, + "reason": "Different failure modes: offloading with quantization vs Gemma4 multi-GPU OOM." + }, + { + "left": "issue:43278", + "right": "issue:43329", + "accept": false, + "reason": "Different model internals: embedding dtype drift vs undefined video-token helper variables." + }, + { + "left": "issue:43701", + "right": "issue:45083", + "accept": false, + "reason": "Resume checkpoint key mismatch is unrelated to qwen3_omni_moe token-length helper behavior." + }, + { + "left": "issue:29942", + "right": "issue:35141", + "accept": false, + "reason": "Flash Attention 2 test failures are not the same bug as embedding reinitialization after resize." + }, + { + "left": "issue:43550", + "right": "issue:45083", + "accept": false, + "reason": "Different code paths and symptoms: Bamba compile/SDPA vs qwen3_omni_moe helper bug." + }, + { + "left": "issue:41720", + "right": "issue:45229", + "accept": false, + "reason": "Different runtime failures: Qwen3 auto-device cudaErrorAssert vs Gemma4 multi-GPU OOM." + }, + { + "left": "issue:32090", + "right": "issue:33357", + "accept": false, + "reason": "Different environments and bugs: Trainer broadcast NoneType vs MacOS bus error on community CLIP." + }, + { + "left": "issue:43976", + "right": "issue:45440", + "accept": false, + "reason": "Python version compatibility issue is unrelated to DeepSeekV3MoE implementation divergence." + }, + { + "left": "issue:37428", + "right": "issue:44987", + "accept": false, + "reason": "Import error for flash-attention utility is not the same as loading a specific model repo on transformers>=5.1.0." + }, + { + "left": "issue:44451", + "right": "issue:45520", + "accept": false, + "reason": "Different cause: model loading failure for ScandiBERT vs flash_attn import lookup error on Python 3.13." + }, + { + "left": "issue:45372", + "right": "issue:45459", + "accept": false, + "reason": "Gemma4 processor import failure and tokenizer error-masking via protobuf handling are different bugs." + }, + { + "left": "issue:44448", + "right": "issue:44466", + "accept": false, + "reason": "Both concern v5 behavior changes, but one is generation output drift and the other is lm_head.weight serialization inconsistency." + }, + { + "left": "issue:43065", + "right": "issue:43994", + "accept": false, + "reason": "Different model issues: dummy Conv2d in Sam3PixelDecoder vs nonsensical SigLIP2 outputs." + }, + { + "left": "issue:43065", + "right": "issue:45412", + "accept": false, + "reason": "Different symptoms and code paths: dummy module in Sam3PixelDecoder vs RT-DETR memory not being released." + } + ] + }, + "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-15354-364", + "nodes": [ + "issue:25251", + "issue:29127", + "issue:29942", + "issue:30064", + "issue:32090", + "issue:33290", + "issue:33357", + "issue:34567", + "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:25251|issue:32090", + "issue:44186|issue:45198", + "issue:43278|issue:43381", + "issue:43653|issue:44743", + "issue:44265|issue:45198", + "issue:43257|issue:43278", + "issue:44361|issue:45198", + "issue:43421|issue:45491", + "issue:43994|issue:45412", + "issue:44871|issue:44977", + "issue:42915|issue:42994", + "issue:44206|issue:44479", + "issue:43756|issue:45440", + "issue:43504|issue:44857", + "issue:44610|issue:45412", + "issue:29942|issue:30064", + "issue:33357|issue:39692", + "issue:45103|issue:45372", + "issue:32090|issue:33290", + "issue:29127|issue:34567", + "issue:45072|issue:45229", + "issue:43824|issue:44279", + "issue:43381|issue:43701", + "issue:43653|issue:44442", + "issue:43526|issue:44186", + "issue:44464|issue:45491", + "issue:43334|issue:43531", + "issue:43262|issue:44206", + "issue:44977|issue:45245", + "issue:29942|issue:36010", + "issue:29127|issue:33357" + ], + "prepared_review_unit_hash": "6451945188bb0565265422141523e9ddb811e28861da1368e54c722df9856782", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37021, + "estimated_input_tokens": 9256, + "estimated_eval_tokens": 18768 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous: it mixes unrelated bugs across tokenizers, processors, training, quantization, model loading, and inference. None of the soft pairs look like the same underlying issue or change, so there is no clean canonical duplicate target.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:25251", + "right": "issue:32090", + "accept": false, + "reason": "Different failure modes: pipeline output shape change vs Trainer broadcast TypeError." + }, + { + "left": "issue:44186", + "right": "issue:45198", + "accept": false, + "reason": "Both are tokenizer-related, but they affect different models and different bugs (LayoutLMv2 NER/padding crash vs Wav2Vec2 save/tokenization failure)." + }, + { + "left": "issue:43278", + "right": "issue:43381", + "accept": false, + "reason": "One is a dtype mismatch between train/eval, the other is checkpointing disallowed in eval mode; not the same code path." + }, + { + "left": "issue:43653", + "right": "issue:44743", + "accept": false, + "reason": "Tokenizer special-token registration vs Qwen3.5 recurrent-state reset in cached generation are unrelated." + }, + { + "left": "issue:44265", + "right": "issue:45198", + "accept": false, + "reason": "torch.export failure from torch_compilable_check is unrelated to Wav2Vec2 save_pretrained/tokenization errors." + }, + { + "left": "issue:43257", + "right": "issue:43278", + "accept": false, + "reason": "Qwen3 MoE weight conversion/loading with accelerate+deepspeed is unrelated to BF16-to-FP32 evaluation dtype drift." + }, + { + "left": "issue:44361", + "right": "issue:45198", + "accept": false, + "reason": "Different tokenizer/model failures with no shared concrete bug." + }, + { + "left": "issue:43421", + "right": "issue:45491", + "accept": false, + "reason": "Runtime post-processor updates for special tokens are unrelated to Gemma3 NaN embeddings in mixed-length batching." + }, + { + "left": "issue:43994", + "right": "issue:45412", + "accept": false, + "reason": "SigLIP2 incorrect outputs and RT-DETR memory release are distinct issues in different subsystems." + }, + { + "left": "issue:44871", + "right": "issue:44977", + "accept": false, + "reason": "Gemma-3 eos_token_id inconsistency is not the same as Qwen3.5 flash-attention generation failure." + }, + { + "left": "issue:42915", + "right": "issue:42994", + "accept": false, + "reason": "Both mention quantization, but one is FP8 config failure and the other is quantized model saving; different problems." + }, + { + "left": "issue:44206", + "right": "issue:44479", + "accept": false, + "reason": "Audio feature extractor center-arg crash is unrelated to Qwen VL video-input regression." + }, + { + "left": "issue:43756", + "right": "issue:45440", + "accept": false, + "reason": "Smollm3 RoPE layer mismatch and DeepseekV3MoE divergence are different model-specific implementation bugs." + }, + { + "left": "issue:43504", + "right": "issue:44857", + "accept": false, + "reason": "Beit pretrained-load legacy field issue is unrelated to LwDetrImageLoss AMP/CUDA crash." + }, + { + "left": "issue:44610", + "right": "issue:45412", + "accept": false, + "reason": "Processor image-size mismatch and RT-DETR memory leak are unrelated." + }, + { + "left": "issue:29942", + "right": "issue:30064", + "accept": false, + "reason": "Failing Flash Attention 2 tests and void segmentation map processing are separate regressions." + }, + { + "left": "issue:33357", + "right": "issue:39692", + "accept": false, + "reason": "MacOS CLIP bus error and SigLIP2 docs example errors are different model and reproduction contexts." + }, + { + "left": "issue:45103", + "right": "issue:45372", + "accept": false, + "reason": "auto_docstring annotation handling and ReasoningEffort import failure are unrelated." + }, + { + "left": "issue:32090", + "right": "issue:33290", + "accept": false, + "reason": "Trainer NoneType broadcast error and deepspeed Adafactor OOM are different training failures." + }, + { + "left": "issue:29127", + "right": "issue:34567", + "accept": false, + "reason": "LayoutLMv3 error-message clarity and TrainerState token counting bug do not share a concrete code path." + }, + { + "left": "issue:45072", + "right": "issue:45229", + "accept": false, + "reason": "bf16 dtype mismatch in inference and Gemma4 multi-GPU OOM are unrelated." + }, + { + "left": "issue:43824", + "right": "issue:44279", + "accept": false, + "reason": "A specific import error for Qwen2.5-VL is not the same as a general dependency issue." + }, + { + "left": "issue:43381", + "right": "issue:43701", + "accept": false, + "reason": "Eval-mode checkpointing restriction and resume_from_checkpoint key mismatch are different checkpoint-related bugs." + }, + { + "left": "issue:43653", + "right": "issue:44442", + "accept": false, + "reason": "BigBird special-token decode issue and AutoTokenizer loading FastSpeech2Conformer are unrelated tokenizer bugs." + }, + { + "left": "issue:43526", + "right": "issue:44186", + "accept": false, + "reason": "BeitImageProcessorFast label reduction and LayoutLMv2Tokenizer NER/padding crash affect different preprocessing paths." + }, + { + "left": "issue:44464", + "right": "issue:45491", + "accept": false, + "reason": "Compiled-forward chunked generation inconsistency is unrelated to Gemma3 NaN embeddings." + }, + { + "left": "issue:43334", + "right": "issue:43531", + "accept": false, + "reason": "Qwen3-VL pad_token_id import failure and Qwen3-MoE sliding_window behavior are separate model issues." + }, + { + "left": "issue:43262", + "right": "issue:44206", + "accept": false, + "reason": "Audio chat-template sample-rate default and LasrFeatureExtractor center-arg crash are different bugs." + }, + { + "left": "issue:44977", + "right": "issue:45245", + "accept": false, + "reason": "Qwen3.5 flash-attention generation failure is unrelated to the categories-cardinality runtime error." + }, + { + "left": "issue:29942", + "right": "issue:36010", + "accept": false, + "reason": "Flash Attention 2 test failures and GenerationMixin import breakage are unrelated." + }, + { + "left": "issue:29127", + "right": "issue:33357", + "accept": false, + "reason": "LayoutLMv3 err_handle messaging and CLIP MacOS bus error are unrelated issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 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-15354-364", + "nodes": [ + "issue:29127", + "issue:29942", + "issue:33290", + "issue:33666", + "issue:34567", + "issue:34689", + "issue:35141", + "issue:39692", + "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:43720", + "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:45440", + "issue:45479", + "issue:45491" + ], + "soft_pairs": [ + "issue:29127|issue:35141", + "issue:43761|issue:44079", + "issue:42915|issue:43931", + "issue:44442|issue:45072", + "issue:43232|issue:43388", + "issue:43526|issue:44857", + "issue:43526|issue:44610", + "issue:44479|issue:45290", + "issue:44743|issue:44811", + "issue:43723|issue:45479", + "issue:45003|issue:45440", + "issue:43575|issue:43701", + "issue:43931|issue:45216", + "issue:43646|issue:44077", + "issue:43761|issue:44188", + "issue:44964|issue:45440", + "issue:43994|issue:44206", + "issue:43575|issue:44387", + "issue:43316|issue:43525", + "issue:29942|issue:33666", + "issue:39692|issue:45491", + "issue:44442|issue:44898", + "issue:43673|issue:44743", + "issue:33290|issue:34567", + "issue:34689|issue:43611", + "issue:43493|issue:43827", + "issue:43381|issue:43873", + "issue:33290|issue:35141", + "issue:44077|issue:44479", + "issue:33290|issue:34689", + "issue:43720|issue:44442" + ], + "prepared_review_unit_hash": "6d16e79fa3443f3c5b234b9c1cc168d09fcab5ff17c18feed051694db1558af7", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37033, + "estimated_input_tokens": 9259, + "estimated_eval_tokens": 18774 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "No clear duplicate cluster here: the items span unrelated issues across model loading, training, generation, quantization, docs, and processor behavior. All soft pairs should be rejected.", + "confidence": 0.94, + "canonical_issue_reason": "No single canonical issue is supported by the titles/descriptions; the set is heterogeneous and does not center on one underlying bug.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43873 is the strongest representative only by activity/ref count and broad impact, but it is not semantically a true canonical for this cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:29127", + "right": "issue:35141", + "accept": false, + "reason": "Different problems: LayoutLMv3 box-info error handling vs token embedding reinit after resize." + }, + { + "left": "issue:43761", + "right": "issue:44079", + "accept": false, + "reason": "Different code paths: CLIPVisionModel hidden_states regression vs ModelOutput key assignment bug." + }, + { + "left": "issue:42915", + "right": "issue:43931", + "accept": false, + "reason": "Both are loading-related, but one is FineGrainedFP8 with Qwen3Moe and the other is a shape-mismatch on Qwen3-VL; not the same bug." + }, + { + "left": "issue:44442", + "right": "issue:45072", + "accept": false, + "reason": "Tokenizer loading failure vs bfloat16 dtype mismatch in CI; unrelated failures." + }, + { + "left": "issue:43232", + "right": "issue:43388", + "accept": false, + "reason": "Generation kwargs update after sync_gpus is unrelated to gather_for_metrics label truncation." + }, + { + "left": "issue:43526", + "right": "issue:44857", + "accept": false, + "reason": "BEiT image processor label reduction bug vs LwDetrImageLoss AMP crash; different components and symptoms." + }, + { + "left": "issue:43526", + "right": "issue:44610", + "accept": false, + "reason": "ImageProcessor label handling vs OmDet-Turbo processor size mismatch; not the same issue." + }, + { + "left": "issue:44479", + "right": "issue:45290", + "accept": false, + "reason": "Video input regression for Qwen VL models vs chat template crash on tool-call assistant messages." + }, + { + "left": "issue:44743", + "right": "issue:44811", + "accept": false, + "reason": "Qwen3.5 recurrent state reset bug vs Whisper batch_decode skip_special_tokens handling." + }, + { + "left": "issue:43723", + "right": "issue:45479", + "accept": false, + "reason": "Tokenizer loading issue vs sequence-classification zero-loss bug; unrelated." + }, + { + "left": "issue:45003", + "right": "issue:45440", + "accept": false, + "reason": "Unsafe sys.modules access in modeling_utils vs DeepseekV3MoE implementation divergence; different layers and failures." + }, + { + "left": "issue:43575", + "right": "issue:43701", + "accept": false, + "reason": "Tensor-parallel OOM when loading Qwen2-57B vs resume_from_checkpoint key mismatch; not the same underlying bug." + }, + { + "left": "issue:43931", + "right": "issue:45216", + "accept": false, + "reason": "Model load shape mismatch vs save_pretrained checkpoint regression; different directions of failure." + }, + { + "left": "issue:43646", + "right": "issue:44077", + "accept": false, + "reason": "Custom model initialization breakage vs patchtsmixer post_init API constraint; too different." + }, + { + "left": "issue:43761", + "right": "issue:44188", + "accept": false, + "reason": "CLIPVisionModel hidden_states regression vs attention-kernel divergence under torch.compile; unrelated." + }, + { + "left": "issue:44964", + "right": "issue:45440", + "accept": false, + "reason": "Phi-4 multimodal loading error vs DeepSeekV3MoE implementation mismatch; different models and problems." + }, + { + "left": "issue:43994", + "right": "issue:44206", + "accept": false, + "reason": "SigLIP2 nonsensical outputs vs LasrFeatureExtractor unsupported center arg; different regressions." + }, + { + "left": "issue:43575", + "right": "issue:44387", + "accept": false, + "reason": "Qwen2 TP loading OOM vs increased reserved memory under int4 quantization; both memory-related but not the same concrete bug." + }, + { + "left": "issue:43316", + "right": "issue:43525", + "accept": false, + "reason": "Gemma3TextConfig API discrepancy vs Llama4Config missing pad_token_id; distinct API surface issues." + }, + { + "left": "issue:29942", + "right": "issue:33666", + "accept": false, + "reason": "Flash Attention 2 test failures vs Qwen2-VL multi-GPU training; separate concerns." + }, + { + "left": "issue:39692", + "right": "issue:45491", + "accept": false, + "reason": "SigLIP2 docs example errors vs Gemma3 NaN embeddings on mixed-length batches; docs vs runtime model bug." + }, + { + "left": "issue:44442", + "right": "issue:44898", + "accept": false, + "reason": "FastSpeech2ConformerTokenizer loading failure vs Perceiver non-default resolution failure; unrelated model families." + }, + { + "left": "issue:43673", + "right": "issue:44743", + "accept": false, + "reason": "GenerationMixin cache missing in chunked_prefill vs recurrent state reset in modular_qwen3_5.py; different generation bugs." + }, + { + "left": "issue:33290", + "right": "issue:34567", + "accept": false, + "reason": "AdaFactor/deepspeed OOM vs TrainerState token counter not updating; not the same issue." + }, + { + "left": "issue:34689", + "right": "issue:43611", + "accept": false, + "reason": "Llama 3.2 Vision loading breakage vs base_model_prefix loading breakage in v5.0.0; both loading-related but different failure mechanisms." + }, + { + "left": "issue:43493", + "right": "issue:43827", + "accept": false, + "reason": "SigLIP2 HF/JAX divergence vs docs still referencing removed pipeline(); different categories." + }, + { + "left": "issue:43381", + "right": "issue:43873", + "accept": false, + "reason": "Gradient checkpointing eval-mode restriction vs quantization/offloading behavior; unrelated." + }, + { + "left": "issue:33290", + "right": "issue:35141", + "accept": false, + "reason": "AdaFactor/deepspeed OOM vs embedding reinitialization after resize; different bugs." + }, + { + "left": "issue:44077", + "right": "issue:44479", + "accept": false, + "reason": "patchtsmixer post_init policy vs Qwen VL video-input regression; unrelated." + }, + { + "left": "issue:33290", + "right": "issue:34689", + "accept": false, + "reason": "AdaFactor OOM vs Llama 3.2 Vision loading regression; different symptoms and code paths." + }, + { + "left": "issue:43720", + "right": "issue:44442", + "accept": false, + "reason": "Accelerate loading/unpacking packed weights in BitNet vs tokenizer loading failure; unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 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-15354-364", + "nodes": [ + "issue:16998", + "issue:22355", + "issue:25251", + "issue:28282", + "issue:33290", + "issue:34689", + "issue:36010", + "issue:36331", + "issue:39692", + "issue:41628", + "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:43749", + "issue:43761", + "issue:43825", + "issue:43856", + "issue:43873", + "issue:43931", + "issue:44077", + "issue:44162", + "issue:44190", + "issue:44206", + "issue:44479", + "issue:44496", + "issue:44514", + "issue:44534", + "issue:44617", + "issue:44792", + "issue:44855", + "issue:44871", + "issue:44877", + "issue:44913", + "issue:45127", + "issue:45137", + "issue:45200", + "issue:45372" + ], + "soft_pairs": [ + "issue:41628|issue:45372", + "issue:43749|issue:43931", + "issue:43643|issue:44913", + "issue:43873|issue:44871", + "issue:39692|issue:44206", + "issue:44190|issue:44496", + "issue:43381|issue:43856", + "issue:33290|issue:36331", + "issue:44855|issue:45137", + "issue:34689|issue:36010", + "issue:43316|issue:44877", + "issue:43540|issue:44514", + "issue:44162|issue:44792", + "issue:43388|issue:43688", + "issue:42175|issue:43316", + "issue:43404|issue:43441", + "issue:16998|issue:22355", + "issue:25251|issue:28282", + "issue:43065|issue:44617", + "issue:43316|issue:43716", + "issue:41950|issue:43475", + "issue:43295|issue:43825", + "issue:45127|issue:45200", + "issue:43761|issue:44479", + "issue:44077|issue:44534" + ], + "prepared_review_unit_hash": "d6ac68ca46ff1185ac2d81b4753d875ff5fc85eb03244cdcc061d98a84563e6d", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 25, + "serialized_chars": 35181, + "estimated_input_tokens": 8796, + "estimated_eval_tokens": 17848 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These soft pairs are mostly near-miss topic similarities, but they do not look like the same underlying bug or change. The cluster is heterogeneous across unrelated import, config, pipeline, model-loading, and training issues.", + "confidence": 0.97, + "canonical_issue_reason": "No single issue clearly subsumes the others; the items describe distinct failures across different models, APIs, and subsystems.", + "canonical_pr_reason": null, + "best_issue_reason": "If forced to pick a representative, issue 34689 is a clear regression report with a concrete model/version and reproducible loading failure, but it still does not match the rest of the set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41628", + "right": "issue:45372", + "accept": false, + "reason": "Both are import/loading failures, but one is about AutoImageProcessor exports and the other about mistral_common ReasoningEffort missing during Gemma 4 processor loading." + }, + { + "left": "issue:43749", + "right": "issue:43931", + "accept": false, + "reason": "Both concern loading, but one is broken FSDP CPU RAM efficient loading while the other is a Qwen3-VL shape-mismatch model load error." + }, + { + "left": "issue:43643", + "right": "issue:44913", + "accept": false, + "reason": "Both involve config behavior on reload, but they are different bugs in different configs: trust_remote_code missing fields vs GPTNeoX rotary_pct reverting." + }, + { + "left": "issue:43873", + "right": "issue:44871", + "accept": false, + "reason": "Quantization/offloading behavior and Gemma eos_token_id inconsistency are unrelated bugs." + }, + { + "left": "issue:39692", + "right": "issue:44206", + "accept": false, + "reason": "One is a SigLIP2 documentation example with multiple errors; the other is a feature extractor API regression passing an unsupported center arg." + }, + { + "left": "issue:44190", + "right": "issue:44496", + "accept": false, + "reason": "A local dataset loading failure in a training script is not the same as an unrecognized model/config.json model_type error." + }, + { + "left": "issue:43381", + "right": "issue:43856", + "accept": false, + "reason": "Gradient checkpointing in eval mode and Qwen3 MoE memory inefficiency are different training-time issues." + }, + { + "left": "issue:33290", + "right": "issue:36331", + "accept": false, + "reason": "OOM with Adafactor in DeepSpeed is unrelated to a CustomTrainer compute_loss signature mismatch." + }, + { + "left": "issue:44855", + "right": "issue:45137", + "accept": false, + "reason": "Python 3.13 Deberta import parsing error and DeepSpeed ZeRO3 deque underflow are different failure modes." + }, + { + "left": "issue:34689", + "right": "issue:36010", + "accept": false, + "reason": "A Llama 3.2 Vision loading regression is not the same as missing GenerationMixin import from transformers.generation." + }, + { + "left": "issue:43316", + "right": "issue:44877", + "accept": false, + "reason": "Gemma3TextConfig API mismatch and strict config blocking granite_speech loading are different configuration issues." + }, + { + "left": "issue:43540", + "right": "issue:44514", + "accept": false, + "reason": "Qwen3OmniMoe video processing ValueError and batched Qwen2_5_VL apply_chat_template crash are distinct processor bugs." + }, + { + "left": "issue:44162", + "right": "issue:44792", + "accept": false, + "reason": "ESM2 breaking scientific workflows and a janus test failure are unrelated." + }, + { + "left": "issue:43388", + "right": "issue:43688", + "accept": false, + "reason": "A metrics-gathering last-batch label truncation bug is not the same as auxiliary loss normalization in OLMoE/GPT Oss." + }, + { + "left": "issue:42175", + "right": "issue:43316", + "accept": false, + "reason": "Missing tensorflow in a torch extra install and a Gemma3TextConfig API discrepancy are unrelated." + }, + { + "left": "issue:43404", + "right": "issue:43441", + "accept": false, + "reason": "Weight tying in Mistral3ForConditionalGeneration and FlashAttention failure in Ministral-3 are different concrete code-path problems." + }, + { + "left": "issue:16998", + "right": "issue:22355", + "accept": false, + "reason": "Model_max_length questions for DeBERTa-V3 and a transformers.onnx import error have no shared underlying bug." + }, + { + "left": "issue:25251", + "right": "issue:28282", + "accept": false, + "reason": "Pipeline top_k output nesting and missing PyTorch for AutoModel are unrelated." + }, + { + "left": "issue:43065", + "right": "issue:44617", + "accept": false, + "reason": "A dummy Conv2d in Sam3PixelDecoder and Sam3Video CUDA OOM are not the same bug." + }, + { + "left": "issue:43316", + "right": "issue:43716", + "accept": false, + "reason": "Gemma3TextConfig API differences and Mistral-3 image preprocessor dtype mismatch are different issues." + }, + { + "left": "issue:41950", + "right": "issue:43475", + "accept": false, + "reason": "Video-classification selecting image processors and Sam3VisionEncoderOutput missing fpn_position_embeddings are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:43825", + "accept": false, + "reason": "A processor.tokenizer regression is not the same as an incorrect v5 translation error message." + }, + { + "left": "issue:45127", + "right": "issue:45200", + "accept": false, + "reason": "LoRA merge collapse with tied embeddings and Gemma 4 mm_token_type_ids defaults are different Gemma-related problems." + }, + { + "left": "issue:43761", + "right": "issue:44479", + "accept": false, + "reason": "CLIPVisionModel hidden_states regression and Qwen video input regression are distinct model-specific bugs." + }, + { + "left": "issue:44077", + "right": "issue:44534", + "accept": false, + "reason": "patchtsmixer post_init policy and non-persistent buffer junk in v5 are unrelated framework issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 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-15354-364", + "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:43582", + "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: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:44938", + "issue:44964", + "issue:44977", + "issue:45245", + "issue:45290", + "issue:45362", + "issue:45381", + "issue:45406", + "issue:45542" + ], + "soft_pairs": [ + "issue:43550|issue:43701", + "issue:43606|issue:43701", + "issue:44734|issue:45406", + "issue:43295|issue:44871", + "issue:43452|issue:44496", + "issue:43646|issue:44913", + "issue:43906|issue:45290", + "issue:43761|issue:44242", + "issue:43575|issue:43742", + "issue:42617|issue:43575", + "issue:43931|issue:44060", + "issue:44521|issue:45381", + "issue:43575|issue:43873", + "issue:44792|issue:44977", + "issue:44479|issue:44734", + "issue:44464|issue:44871", + "issue:32090|issue:35141", + "issue:43525|issue:43531", + "issue:44933|issue:45362", + "issue:43827|issue:43994", + "issue:43065|issue:44610", + "issue:44534|issue:44913", + "issue:41720|issue:45245", + "issue:44938|issue:45542", + "issue:43650|issue:44220", + "issue:44190|issue:44964", + "issue:43388|issue:43582" + ], + "prepared_review_unit_hash": "c10b4b5da3a80ffc019077a8e6188a71d96fe80ce40151d38f8c55d0ee3923e4", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 35606, + "estimated_input_tokens": 8902, + "estimated_eval_tokens": 18060 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a false-positive mix of unrelated issues spanning model loading, multimodal/chat-template regressions, config serialization, distributed training, and docs/environment failures. I do not see a single underlying bug or change tying these together, so all soft edges are rejected.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue fits this cluster: the items are heterogeneous and do not share a common root cause.", + "canonical_pr_reason": null, + "best_issue_reason": "If forced to pick a representative issue, issue:44521 is the strongest candidate because it has the most cross-references and describes a broadly impactful multimodal chat-template bug, but it still does not represent the whole set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43550", + "right": "issue:43701", + "accept": false, + "reason": "Different failures: torch.compile SDPA on Bamba vs checkpoint key mismatch during resume." + }, + { + "left": "issue:43606", + "right": "issue:43701", + "accept": false, + "reason": "CPU offload device mismatch is unrelated to resume_from_checkpoint key handling." + }, + { + "left": "issue:44734", + "right": "issue:45406", + "accept": false, + "reason": "Both mention serve, but one is KV-cache indexing and the other is a missing processor attribute." + }, + { + "left": "issue:43295", + "right": "issue:44871", + "accept": false, + "reason": "Different regressions: processor/tokenizer API break vs eos_token_id config inconsistency." + }, + { + "left": "issue:43452", + "right": "issue:44496", + "accept": false, + "reason": "Both are loading problems, but one is gguf tokenizer/model loading and the other is unrecognized model config." + }, + { + "left": "issue:43646", + "right": "issue:44913", + "accept": false, + "reason": "Custom model init breakage and GPTNeoX config reload behavior are separate code paths." + }, + { + "left": "issue:43906", + "right": "issue:45290", + "accept": false, + "reason": "Isolated repro of an older issue is not the same bug as chat template crashes with tool calls." + }, + { + "left": "issue:43761", + "right": "issue:44242", + "accept": false, + "reason": "CLIPVision hidden_states regression and missing load-balancing loss are unrelated model behaviors." + }, + { + "left": "issue:43575", + "right": "issue:43742", + "accept": false, + "reason": "TP OOM on a large Qwen model is not the same as a MobileLLM key error." + }, + { + "left": "issue:42617", + "right": "issue:43575", + "accept": false, + "reason": "Both are distributed/model-parallel related, but one is a 3d_parallel.py runtime issue and the other is TP memory OOM." + }, + { + "left": "issue:43931", + "right": "issue:44060", + "accept": false, + "reason": "Weight-shape mismatch in Qwen3-VL is unrelated to a tied-weights warning in Qwen3-Next." + }, + { + "left": "issue:44521", + "right": "issue:45381", + "accept": false, + "reason": "Both are multimodal/Qwen-VL related, but assistant_mask generation and vision_position_ids are different bugs." + }, + { + "left": "issue:43575", + "right": "issue:43873", + "accept": false, + "reason": "TP loading OOM and offloading with quantization are different memory-management problems." + }, + { + "left": "issue:44792", + "right": "issue:44977", + "accept": false, + "reason": "Failed janus image-generation test and Qwen3.5 flash-attention generation issue are unrelated." + }, + { + "left": "issue:44479", + "right": "issue:44734", + "accept": false, + "reason": "Video-input regression and serve KV-cache continuation crash affect different paths and symptoms." + }, + { + "left": "issue:44464", + "right": "issue:44871", + "accept": false, + "reason": "Compiled-forward chunked generation inconsistency is unrelated to eos_token_id configuration drift." + }, + { + "left": "issue:32090", + "right": "issue:35141", + "accept": false, + "reason": "Trainer GPU broadcast TypeError and embedding reinitialization in post_init are different bugs." + }, + { + "left": "issue:43525", + "right": "issue:43531", + "accept": false, + "reason": "Missing pad_token_id on Llama4Config and Qwen3-MoE sliding_window behavior are separate config/model issues." + }, + { + "left": "issue:44933", + "right": "issue:45362", + "accept": false, + "reason": "A missing import from image_utils is not the same as a Qwen3.5 chat-generation crash." + }, + { + "left": "issue:43827", + "right": "issue:43994", + "accept": false, + "reason": "Docs still referencing pipeline() is a documentation issue, not the SigLIP2 inference bug." + }, + { + "left": "issue:43065", + "right": "issue:44610", + "accept": false, + "reason": "Sam3PixelDecoder dummy Conv2d and OmDet-Turbo input-size mismatch are unrelated vision-model issues." + }, + { + "left": "issue:44534", + "right": "issue:44913", + "accept": false, + "reason": "Non-persistent buffer corruption in v5 is not the same as GPTNeoX rotary_pct reload loss." + }, + { + "left": "issue:41720", + "right": "issue:45245", + "accept": false, + "reason": "Qwen3 auto-device-map cuda assert and category-cardinality RuntimeError are unrelated." + }, + { + "left": "issue:44938", + "right": "issue:45542", + "accept": false, + "reason": "Python 3.14 import/load failure and missing TensorFlow backend due to tensorboard-only install are different environment issues." + }, + { + "left": "issue:43650", + "right": "issue:44220", + "accept": false, + "reason": "'ADD THE DATA' is unrelated to _torch_extract_fbank_features." + }, + { + "left": "issue:44190", + "right": "issue:44964", + "accept": false, + "reason": "Local dataset loading with a no-trainer script is unrelated to loading Phi-4 multimodal weights." + }, + { + "left": "issue:43388", + "right": "issue:43582", + "accept": false, + "reason": "Metric gathering label truncation and Apple Silicon caching allocator TypeError do not share a code path." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 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-15354-364", + "nodes": [ + "issue:36246", + "issue:38175", + "issue:38617", + "issue:39692", + "issue:41628", + "issue:41950", + "issue:42175", + "issue:43065", + "issue:43116", + "issue:43295", + "issue:43335", + "issue:43404", + "issue:43479", + "issue:43540", + "issue:43550", + "issue:43643", + "issue:43644", + "issue: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:45127", + "issue:45216", + "issue:45276", + "issue:45335", + "issue:45357", + "issue:45372", + "issue:45375", + "issue:45446" + ], + "soft_pairs": [ + "issue:43761|issue:44743", + "issue:45127|issue:45357", + "issue:44355|issue:44855", + "issue:43644|issue:44913", + "issue:44077|issue:44079", + "issue:43976|issue:44855", + "issue:43540|issue:44734", + "issue:44496|issue:45372", + "issue:43295|issue:45276", + "issue:43550|issue:44190", + "issue:43335|issue:44855", + "issue:43404|issue:45216", + "issue:43116|issue:44190", + "issue:41950|issue:44016", + "issue:39692|issue:44190", + "issue:43644|issue:44871", + "issue:43479|issue:45375", + "issue:44186|issue:44734", + "issue:36246|issue:41628", + "issue:38175|issue:43065", + "issue:44246|issue:45003", + "issue:44466|issue:44913", + "issue:44016|issue:44220", + "issue:36246|issue:38617", + "issue:44683|issue:45446", + "issue:42175|issue:44279", + "issue:44112|issue:44792", + "issue:43854|issue:44792", + "issue:43295|issue:45335", + "issue:43643|issue:44496" + ], + "prepared_review_unit_hash": "c5ffc118e4f869b7684b05ce36363e9621882d615f2435809e88dd1beeac42d5", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 37010, + "estimated_input_tokens": 9253, + "estimated_eval_tokens": 18762 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most pairs are unrelated one-off issues. Only 44683\u219445446 looks like a likely true duplicate; the Qwen3.5 save_pretrained pair is only a tentative related cluster and not strong enough to accept as a duplicate.", + "confidence": 0.82, + "canonical_issue_reason": "44683 is the best canonical anchor because it describes the broader flex_attention failure on torch>=2.9, while 45446 reads like a narrower version-check/root-cause report for the same code path.", + "canonical_pr_reason": null, + "best_issue_reason": "44683 is the strongest representative issue overall because it cleanly captures the only accepted duplicate cluster and states the user-visible regression at the right level of generality.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43761", + "right": "issue:44743", + "accept": false, + "reason": "Different models and failure modes: CLIPVision hidden states vs Qwen recurrent-state reset." + }, + { + "left": "issue:45127", + "right": "issue:45357", + "accept": false, + "reason": "Related Qwen3.5 save_pretrained area, but the concrete bugs differ: LoRA/tied-weights collapse vs incorrect visual encoder keys." + }, + { + "left": "issue:44355", + "right": "issue:44855", + "accept": false, + "reason": "Both are import/runtime errors, but they come from unrelated causes and code paths." + }, + { + "left": "issue:43644", + "right": "issue:44913", + "accept": false, + "reason": "Both involve persistence/reload, but one is junk non-persistent buffers and the other is a GPTNeoX config field reset." + }, + { + "left": "issue:44077", + "right": "issue:44079", + "accept": false, + "reason": "Unrelated issues: patchtsmixer post_init handling vs ModelOutput key assignment." + }, + { + "left": "issue:43976", + "right": "issue:44855", + "accept": false, + "reason": "Both mention Python version, but the underlying bugs are unrelated." + }, + { + "left": "issue:43540", + "right": "issue:44734", + "accept": false, + "reason": "Different subsystems: Qwen3OmniMoe video input processing vs serving KV-cache continuation indexing." + }, + { + "left": "issue:44496", + "right": "issue:45372", + "accept": false, + "reason": "Both are model-loading/import problems, but the missing symbols and root causes differ." + }, + { + "left": "issue:43295", + "right": "issue:45276", + "accept": false, + "reason": "Both are regressions touching model APIs, but processor/tokenizer behavior and embedding resize behavior are different bugs." + }, + { + "left": "issue:43550", + "right": "issue:44190", + "accept": false, + "reason": "Completely different bugs: torch.compile SDPA on Bamba vs dataset loading in an example script." + }, + { + "left": "issue:43335", + "right": "issue:44855", + "accept": false, + "reason": "Different areas entirely: SwitchTransformers config bug vs DebertaV2 import parsing error." + }, + { + "left": "issue:43404", + "right": "issue:45216", + "accept": false, + "reason": "Both mention tied weights/checkpointing, but they affect different models and code paths." + }, + { + "left": "issue:43116", + "right": "issue:44190", + "accept": false, + "reason": "Different example scripts and different failures; not the same underlying issue." + }, + { + "left": "issue:41950", + "right": "issue:44016", + "accept": false, + "reason": "Unrelated: video-classification pipeline processor lookup vs notebook syntax error." + }, + { + "left": "issue:39692", + "right": "issue:44190", + "accept": false, + "reason": "Different example/documentation issues and different failure types." + }, + { + "left": "issue:43644", + "right": "issue:44871", + "accept": false, + "reason": "Serialization/reload bug vs Gemma eos_token_id mismatch; not the same defect." + }, + { + "left": "issue:43479", + "right": "issue:45375", + "accept": false, + "reason": "Both are multimodal config issues, but the affected models and actual config problems differ." + }, + { + "left": "issue:44186", + "right": "issue:44734", + "accept": false, + "reason": "Tokenizer padding/truncation crash is unrelated to serving KV-cache continuation indexing." + }, + { + "left": "issue:36246", + "right": "issue:41628", + "accept": false, + "reason": "Both are ImportError reports, but for different missing names in different modules." + }, + { + "left": "issue:38175", + "right": "issue:43065", + "accept": false, + "reason": "Different model-specific functional bugs with no shared code path." + }, + { + "left": "issue:44246", + "right": "issue:45003", + "accept": false, + "reason": "Import performance issue vs unsafe sys.modules access; at most indirectly related." + }, + { + "left": "issue:44466", + "right": "issue:44913", + "accept": false, + "reason": "Both concern config/state persistence, but one is lm_head serialization and the other is a GPTNeoX config reload default." + }, + { + "left": "issue:44016", + "right": "issue:44220", + "accept": false, + "reason": "Notebook syntax error and audio feature extraction issue are unrelated." + }, + { + "left": "issue:36246", + "right": "issue:38617", + "accept": false, + "reason": "Both are ImportError reports, but they involve different missing names in different modules." + }, + { + "left": "issue:44683", + "right": "issue:45446", + "accept": true, + "reason": "Same flex_attention code path and same torch-version-guard problem; 45446 appears to isolate the root cause behind the broader torch>=2.9 failure." + }, + { + "left": "issue:42175", + "right": "issue:44279", + "accept": false, + "reason": "Dependency/back-end installation complaints, but not the same specific bug." + }, + { + "left": "issue:44112", + "right": "issue:44792", + "accept": false, + "reason": "Both are failing tests, but they cover unrelated models and failures." + }, + { + "left": "issue:43854", + "right": "issue:44792", + "accept": false, + "reason": "Different model test failures with no shared concrete bug." + }, + { + "left": "issue:43295", + "right": "issue:45335", + "accept": false, + "reason": "Both are regressions around embeddings/tokenizers, but the affected behaviors and models are different." + }, + { + "left": "issue:43643", + "right": "issue:44496", + "accept": false, + "reason": "Different config-loading failures: missing fields with trust_remote_code vs unrecognized model_type." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The write-up is a bit too strong for the packet. It claims \u2018only two soft pairs look like true duplicates,\u2019 but the provided soft-edge verdicts accept only one pair (44683\u219445446). The Qwen3.5 save_pretrained relationship (45216\u219445357) may be plausible, but it is not an accepted soft pair in the packet and should be phrased more conservatively. I\u2019d accept a summary that says most pairs are unrelated and one soft pair is a likely duplicate, with Qwen3.5 noted as a tentative related cluster rather than a settled duplicate." + }, + "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": 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-15354-364", + "nodes": [ + "issue:38175", + "issue:41950", + "issue:42907", + "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:44683", + "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", + "issue:45542" + ], + "soft_pairs": [ + "issue:43299|issue:45216", + "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:43526|issue:43901", + "issue:44496|issue:45042", + "issue:43824|issue:44933", + "issue:43976|issue:44279", + "issue:43299|issue:45357", + "issue:43577|issue:45216", + "issue:43901|issue:44496", + "issue:44075|issue:44368", + "issue:44521|issue:45325", + "issue:43854|issue:44964", + "issue:44242|issue:45200", + "issue:44464|issue:45245", + "issue:44683|issue:45542", + "issue:43976|issue:45405", + "issue:41950|issue:43650", + "issue:43479|issue:43525", + "issue:44351|issue:45278", + "issue:44496|issue:44877", + "issue:41950|issue:45020", + "issue:44462|issue:45081", + "issue:43550|issue:44964", + "issue:44279|issue:45405", + "issue:44297|issue:44462", + "issue:43299|issue:43854", + "issue:38175|issue:44617" + ], + "prepared_review_unit_hash": "abc27508d52b6e26e45ccc19fecd6a59b4b689542798fe06e2ce5157472d6568", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37204, + "estimated_input_tokens": 9301, + "estimated_eval_tokens": 18858 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a heterogeneous grab-bag of unrelated Transformers issues across many models/features; the soft edges are mostly superficial text similarity and should not be merged as duplicates.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43299", + "right": "issue:45216", + "accept": false, + "reason": "Different bugs: Qwen3VL MoE loading regression vs Qwen3.5 save_pretrained checkpoint serialization." + }, + { + "left": "issue:42907", + "right": "issue:43441", + "accept": false, + "reason": "Ministral/Devstral save failure is unrelated to FlashAttention runtime failure in Ministral-3." + }, + { + "left": "issue:44016", + "right": "issue:44062", + "accept": false, + "reason": "Notebook syntax error and tokenizers AddedToken keyword conflict are unrelated." + }, + { + "left": "issue:44112", + "right": "issue:44964", + "accept": false, + "reason": "CI stale override test failure is unrelated to Phi-4-multimodal loading." + }, + { + "left": "issue:44871", + "right": "issue:45245", + "accept": false, + "reason": "Gemma eos_token_id mismatch is unrelated to category-count runtime error." + }, + { + "left": "issue:43673", + "right": "issue:44336", + "accept": false, + "reason": "Chunked_prefill cache missing is unrelated to ANSI codes in loading_report." + }, + { + "left": "issue:43976", + "right": "issue:44987", + "accept": false, + "reason": "Python version compatibility issue is unrelated to loading failure for physical-intelligence/fast." + }, + { + "left": "issue:43526", + "right": "issue:43901", + "accept": false, + "reason": "BeitImageProcessorFast reduce_labels bug is unrelated to TextClassificationPipeline docs wording." + }, + { + "left": "issue:44496", + "right": "issue:45042", + "accept": false, + "reason": "Olmo model config loading failure is unrelated to PIL backend torchvision dependency regression." + }, + { + "left": "issue:43824", + "right": "issue:44933", + "accept": false, + "reason": "Qwen2.5-VL import error and missing image_utils import are different symbols and code paths." + }, + { + "left": "issue:43976", + "right": "issue:44279", + "accept": false, + "reason": "Python version support problem is unrelated to a generic dependency issue." + }, + { + "left": "issue:43299", + "right": "issue:45357", + "accept": false, + "reason": "Qwen3VL MoE loading bug is unrelated to Qwen3.5 visual-encoder save_pretrained regression." + }, + { + "left": "issue:43577", + "right": "issue:45216", + "accept": false, + "reason": "Blip2 dtype propagation bug is unrelated to Qwen3.5 save_pretrained serialization." + }, + { + "left": "issue:43901", + "right": "issue:44496", + "accept": false, + "reason": "Pipeline docs mismatch is unrelated to model config loading failure." + }, + { + "left": "issue:44075", + "right": "issue:44368", + "accept": false, + "reason": "SGD optimizer args ignored and tie_word_embeddings warning are unrelated." + }, + { + "left": "issue:44521", + "right": "issue:45325", + "accept": false, + "reason": "assistant_masks being zero for multimodal chat templates is unrelated to Qwen2.5-VL rope index scaling." + }, + { + "left": "issue:43854", + "right": "issue:44964", + "accept": false, + "reason": "GLM-4.7-Flash unit test loading issue and Phi-4-multimodal loading issue are different model-specific failures." + }, + { + "left": "issue:44242", + "right": "issue:45200", + "accept": false, + "reason": "Missing load balancing loss with output_router_logits=False is unrelated to Gemma 4 mm_token_type_ids defaults." + }, + { + "left": "issue:44464", + "right": "issue:45245", + "accept": false, + "reason": "Compiled-forward chunked generation inconsistency is unrelated to category-count limit error." + }, + { + "left": "issue:44683", + "right": "issue:45542", + "accept": false, + "reason": "Compiled flex_attention on torch>=2.9 is unrelated to TensorFlow/tensorboard backend installation." + }, + { + "left": "issue:43976", + "right": "issue:45405", + "accept": false, + "reason": "Python version support issue is unrelated to PEFT minimum version bump." + }, + { + "left": "issue:41950", + "right": "issue:43650", + "accept": false, + "reason": "Video-classification processor lookup bug is unrelated to the unrelated 'ADD THE DATA' issue." + }, + { + "left": "issue:43479", + "right": "issue:43525", + "accept": false, + "reason": "Phi4Multimodal default-config initialization and Llama4Config missing pad_token_id are different config bugs." + }, + { + "left": "issue:44351", + "right": "issue:45278", + "accept": false, + "reason": "HybridCache import error and broad import errors after version upgrade are not the same concrete bug." + }, + { + "left": "issue:44496", + "right": "issue:44877", + "accept": false, + "reason": "Unrecognized model loading and strict granite_speech config loading are different config/registration problems." + }, + { + "left": "issue:41950", + "right": "issue:45020", + "accept": false, + "reason": "Image processor lookup in video pipeline is unrelated to broad remote_code breakage." + }, + { + "left": "issue:44462", + "right": "issue:45081", + "accept": false, + "reason": "AutoTokenizer ignoring tokenizer.json is unrelated to Mistral regex patch crashing on backend_tokenizer access." + }, + { + "left": "issue:43550", + "right": "issue:44964", + "accept": false, + "reason": "Bamba torch.compile SDPA failure is unrelated to Phi-4-multimodal loading." + }, + { + "left": "issue:44279", + "right": "issue:45405", + "accept": false, + "reason": "Generic dependency issue is unrelated to the PEFT version pin problem." + }, + { + "left": "issue:44297", + "right": "issue:44462", + "accept": false, + "reason": "tokenizer_class mismatch on save_pretrained is unrelated to AutoTokenizer ignoring tokenizer.json on load." + }, + { + "left": "issue:43299", + "right": "issue:43854", + "accept": false, + "reason": "Qwen3VL MoE loading and Qwen2.5-VL import/export symbol issues are different model families and failure modes." + }, + { + "left": "issue:38175", + "right": "issue:44617", + "accept": false, + "reason": "Siglip2 zero probabilities bug is unrelated to Sam3Video CUDA OOM." + } + ] + }, + "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-15354-364", + "nodes": [ + "issue:39692", + "issue:42175", + "issue:42994", + "issue:43065", + "issue:43122", + "issue:43262", + "issue:43295", + "issue:43317", + "issue:43352", + "issue:43441", + "issue:43450", + "issue:43475", + "issue:43650", + "issue:43673", + "issue:43723", + "issue:43761", + "issue:43784", + "issue:43824", + "issue:43901", + "issue:43976", + "issue:43994", + "issue:44016", + "issue:44162", + "issue:44246", + "issue:44261", + "issue:44263", + "issue:44279", + "issue:44291", + "issue:44297", + "issue:44462", + "issue:44466", + "issue:44521", + "issue:44589", + "issue:44617", + "issue:44623", + "issue:44671", + "issue:44683", + "issue:44749", + "issue:44779", + "issue:44898", + "issue:44987", + "issue:45250", + "issue:45276", + "issue:45290", + "issue:45362", + "issue:45397", + "issue:45542" + ], + "soft_pairs": [ + "issue:43295|issue:44779", + "issue:39692|issue:43901", + "issue:44462|issue:44779", + "issue:43650|issue:45290", + "issue:44016|issue:44291", + "issue:44466|issue:45276", + "issue:43441|issue:44683", + "issue:43450|issue:44279", + "issue:43262|issue:44521", + "issue:43723|issue:44162", + "issue:43065|issue:44898", + "issue:42994|issue:43317", + "issue:44261|issue:44263", + "issue:44623|issue:45397", + "issue:43122|issue:43650", + "issue:44623|issue:44749", + "issue:43784|issue:45542", + "issue:44671|issue:44779", + "issue:43673|issue:43761", + "issue:43901|issue:43994", + "issue:44987|issue:45542", + "issue:44297|issue:45397", + "issue:42175|issue:43976", + "issue:43475|issue:44016", + "issue:43824|issue:43976", + "issue:43901|issue:44246", + "issue:43352|issue:45250", + "issue:43295|issue:44617", + "issue:44589|issue:45362" + ], + "prepared_review_unit_hash": "6bd5636fa9ff10d1ecb98bf8476075b965196a37a027b0d40fe5ab31fadfb3c2", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35550, + "estimated_input_tokens": 8888, + "estimated_eval_tokens": 18032 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The candidate pairs are mostly loose thematic matches (tokenizers, chat templates, quantization, FlashAttention, processors) but they describe different bugs, different code paths, or different model families. I would not merge any of them as duplicates.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "issue:43994 is the strongest standalone issue here: it is concrete, reproducible, and describes a clear user-facing regression in SigLIP2 inference with AutoModel/pipeline.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43295", + "right": "issue:44779", + "accept": false, + "reason": "Both are v5 tokenizer/processor regressions, but one is about processor.tokenizer/image handling and the other is Deepseek tokenizer output; different failures and code paths." + }, + { + "left": "issue:39692", + "right": "issue:43901", + "accept": false, + "reason": "Both mention docs, but one is a SigLIP2 example with model/processor mismatch and quantization failure, while the other is a pipeline docs mismatch about return_all_scores." + }, + { + "left": "issue:44462", + "right": "issue:44779", + "accept": false, + "reason": "AutoTokenizer ignoring tokenizer.json and Deepseek tokenizer producing wrong results are separate tokenizer bugs, not the same underlying defect." + }, + { + "left": "issue:43650", + "right": "issue:45290", + "accept": false, + "reason": "Completely unrelated: one is an empty/titled issue, the other is a specific apply_chat_template crash with tool calls." + }, + { + "left": "issue:44016", + "right": "issue:44291", + "accept": false, + "reason": "Notebook syntax error vs init_empty_weights/_is_hf_initialized loading TypeError; different surfaces and root causes." + }, + { + "left": "issue:44466", + "right": "issue:45276", + "accept": false, + "reason": "Serialization of lm_head.weight depending on device is unrelated to gemma4 resize_token_embeddings not updating embeddings." + }, + { + "left": "issue:43441", + "right": "issue:44683", + "accept": false, + "reason": "Both involve attention backends, but one is a model not supporting FlashAttention 2.0 and the other is compiled flex_attention failing on torch>=2.9; not the same bug." + }, + { + "left": "issue:43450", + "right": "issue:44279", + "accept": false, + "reason": "Video processor batching shape bug vs a generic dependency issue; no shared code-path or concrete duplicate evidence." + }, + { + "left": "issue:43262", + "right": "issue:44521", + "accept": false, + "reason": "Audio chat-template sampling-rate default issue and multimodal assistant_mask issue are distinct chat-template bugs." + }, + { + "left": "issue:43723", + "right": "issue:44162", + "accept": false, + "reason": "AutoTokenizer.from_pretrained loading problem and ESM2 broken model behavior are unrelated." + }, + { + "left": "issue:43065", + "right": "issue:44898", + "accept": false, + "reason": "SAM3 pixel decoder dummy conv and Perceiver non-default resolution failure are different model-specific issues." + }, + { + "left": "issue:42994", + "right": "issue:43317", + "accept": false, + "reason": "Saving a quantized model and loading a dequantized model with device_map=auto are different quantization/offload paths." + }, + { + "left": "issue:44261", + "right": "issue:44263", + "accept": false, + "reason": "MLA rms_norm_eps precision issue and GlmMoeDsaIndexer torch.split return values are unrelated." + }, + { + "left": "issue:44623", + "right": "issue:45397", + "accept": false, + "reason": "processor.save_pretrained missing files and gemma-4 zero3 from_pretrained are different save/load bugs." + }, + { + "left": "issue:43122", + "right": "issue:43650", + "accept": false, + "reason": "Tokenizer regression vs an unrelated empty issue title; no duplicate relationship." + }, + { + "left": "issue:44623", + "right": "issue:44749", + "accept": false, + "reason": "Saving processor files and slow filtering after upgrading to 5.3.0 are unrelated." + }, + { + "left": "issue:43784", + "right": "issue:45542", + "accept": false, + "reason": "Import-time NameError in sentence-transformers integration vs missing TensorFlow backend when only tensorboard is installed; different dependency failures." + }, + { + "left": "issue:44671", + "right": "issue:44779", + "accept": false, + "reason": "CamemBERT MLM prediction regression and Deepseek tokenizer regression are model-specific and distinct." + }, + { + "left": "issue:43673", + "right": "issue:43761", + "accept": false, + "reason": "Generation cache missing during chunked_prefill and CLIPVisionModel hidden_states=None are different regressions." + }, + { + "left": "issue:43901", + "right": "issue:43994", + "accept": false, + "reason": "Docs note about return_all_scores does not match a runtime SigLIP2 inference correctness bug." + }, + { + "left": "issue:44987", + "right": "issue:45542", + "accept": false, + "reason": "Model loading failure for physical-intelligence/fast and TensorFlow backend installation issue are unrelated." + }, + { + "left": "issue:44297", + "right": "issue:45397", + "accept": false, + "reason": "Tokenizer save_pretrained metadata mismatch and gemma-4 zero3 from_pretrained are different problems." + }, + { + "left": "issue:42175", + "right": "issue:43976", + "accept": false, + "reason": "Missing TensorFlow in the torch extra and a Python version compatibility regression are unrelated." + }, + { + "left": "issue:43475", + "right": "issue:44016", + "accept": false, + "reason": "SAM3 video encoder attribute error and notebook syntax error are unrelated." + }, + { + "left": "issue:43824", + "right": "issue:43976", + "accept": false, + "reason": "Missing Qwen2_5_VL import and Python 3.9/3.10 support regression are unrelated." + }, + { + "left": "issue:43901", + "right": "issue:44246", + "accept": false, + "reason": "Docs mismatch about return_all_scores and slow import behavior are different issues." + }, + { + "left": "issue:43352", + "right": "issue:45250", + "accept": false, + "reason": "Generic Flash Attention 2.0 discussion is too broad to treat as the same bug as NemotronHForCausalLM explicitly not supporting it." + }, + { + "left": "issue:43295", + "right": "issue:44617", + "accept": false, + "reason": "Custom model processor/tokenizer regression and SAM3Video CUDA OOM are unrelated." + }, + { + "left": "issue:44589", + "right": "issue:45362", + "accept": false, + "reason": "Float8 storage lookup TypeError and Qwen3.5 chat crash are different failure modes." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 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-15354-364", + "nodes": [ + "issue:36246", + "issue:41628", + "issue:41950", + "issue:42491", + "issue:42757", + "issue:43066", + "issue:43299", + "issue:43441", + "issue:43450", + "issue:43525", + "issue:43526", + "issue:43531", + "issue:43540", + "issue:43596", + "issue:43673", + "issue:43881", + "issue:43976", + "issue:43994", + "issue:44162", + "issue:44188", + "issue:44220", + "issue:44242", + "issue:44246", + "issue:44295", + "issue:44336", + "issue:44351", + "issue:44373", + "issue:44462", + "issue:44485", + "issue:44496", + "issue:44561", + "issue:44617", + "issue:44661", + "issue:44749", + "issue:44779", + "issue:44857", + "issue:44871", + "issue:44877", + "issue:45003", + "issue:45200", + "issue:45276", + "issue:45292", + "issue:45356", + "issue:45362", + "issue:45375", + "issue:45397", + "issue:45542" + ], + "soft_pairs": [ + "issue:43525|issue:43881", + "issue:43450|issue:44617", + "issue:44188|issue:44336", + "issue:41628|issue:43976", + "issue:44295|issue:45397", + "issue:42491|issue:43299", + "issue:41950|issue:44220", + "issue:44336|issue:44373", + "issue:44246|issue:44496", + "issue:44462|issue:45356", + "issue:44661|issue:45542", + "issue:43526|issue:45003", + "issue:43994|issue:45276", + "issue:43531|issue:45362", + "issue:36246|issue:42757", + "issue:43673|issue:44561", + "issue:36246|issue:44351", + "issue:44336|issue:44485", + "issue:44162|issue:45200", + "issue:44242|issue:44857", + "issue:44246|issue:44871", + "issue:44779|issue:45292", + "issue:44295|issue:44749", + "issue:44877|issue:45375", + "issue:43441|issue:43540", + "issue:43066|issue:44779", + "issue:36246|issue:43596" + ], + "prepared_review_unit_hash": "aa62a45472f92db1c768edda18d567610ec0babc84427da082a8c948cae68f24", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 34940, + "estimated_input_tokens": 8735, + "estimated_eval_tokens": 17726 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: most items are distinct Transformers v5 regressions, import errors, or model-specific loading bugs rather than duplicates. No soft pair looks like the same concrete underlying bug/change with enough confidence to merge.", + "confidence": 0.24, + "canonical_issue_reason": "No clear duplicate emerges from the set. If one issue must represent the cluster, issue 44561 is the best umbrella because it captures a broad v5 API-breakage/regression theme affecting downstream models.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44561 is the most representative general regression report in this set; it is broad, version-specific, and plausibly connected to multiple downstream breakages. Still, it should not be treated as a duplicate of the others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43525", + "right": "issue:43881", + "accept": false, + "reason": "Both are loading failures, but they affect different models and describe different failures: missing Llama4Config.pad_token_id vs a generic glm-4v-9b loading error." + }, + { + "left": "issue:43450", + "right": "issue:44617", + "accept": false, + "reason": "Different problem classes: batched video processor shape bug versus Sam3Video CUDA OOM." + }, + { + "left": "issue:44188", + "right": "issue:44336", + "accept": false, + "reason": "Unrelated subsystems: torch.compile attention kernel divergence versus ANSI codes in loading_report." + }, + { + "left": "issue:41628", + "right": "issue:43976", + "accept": false, + "reason": "Both mention import/runtime compatibility, but one is a missing AutoImageProcessor import and the other is a Python version compatibility complaint; not the same bug." + }, + { + "left": "issue:44295", + "right": "issue:45397", + "accept": false, + "reason": "Different code paths and symptoms: position_ids buffer access error versus gemma-4 zero3 from_pretrained loading." + }, + { + "left": "issue:42491", + "right": "issue:43299", + "accept": false, + "reason": "Both involve Qwen3 MoE family regressions, but one is LoRA compatibility across branches and the other is Qwen3VL model loading; too different to merge." + }, + { + "left": "issue:41950", + "right": "issue:44220", + "accept": false, + "reason": "Video-classification pipeline selecting image processors is unrelated to _torch_extract_fbank_features()." + }, + { + "left": "issue:44336", + "right": "issue:44373", + "accept": false, + "reason": "Formatting in loading_report and a wrong docstring for position_ids are unrelated." + }, + { + "left": "issue:44246", + "right": "issue:44496", + "accept": false, + "reason": "Import slowdown and unrecognized model config are different failures with no shared concrete code path." + }, + { + "left": "issue:44462", + "right": "issue:45356", + "accept": false, + "reason": "Both are tokenizer regressions, but one ignores tokenizer.json and the other has codec/warning behavior in Kimi-K2.5; not the same bug." + }, + { + "left": "issue:44661", + "right": "issue:45542", + "accept": false, + "reason": "Tokenizer mapping registration failure is unrelated to the tensorboard-without-TF backend error." + }, + { + "left": "issue:43526", + "right": "issue:45003", + "accept": false, + "reason": "BeitImageProcessorFast reduce_labels bug and sys.modules access in modeling_utils are unrelated." + }, + { + "left": "issue:43994", + "right": "issue:45276", + "accept": false, + "reason": "Different model-specific regressions: SigLIP2 output quality versus gemma4 resize_token_embeddings behavior." + }, + { + "left": "issue:43531", + "right": "issue:45362", + "accept": false, + "reason": "Both are Qwen-related, but sliding_window handling and chat crashes are distinct concrete failures." + }, + { + "left": "issue:36246", + "right": "issue:42757", + "accept": false, + "reason": "Different missing imports from different packages; same broad symptom, not the same underlying issue." + }, + { + "left": "issue:43673", + "right": "issue:44561", + "accept": false, + "reason": "Both are v5 regressions, but one is cache handling during chunked_prefill and the other is a removed helper breaking trust_remote_code." + }, + { + "left": "issue:36246", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they concern different symbols and likely different breakages." + }, + { + "left": "issue:44336", + "right": "issue:44485", + "accept": false, + "reason": "ANSI output behavior and GLM-5 RoPE implementation are unrelated." + }, + { + "left": "issue:44162", + "right": "issue:45200", + "accept": false, + "reason": "ESM2 breakage and Gemma 4 mm_token_type_ids defaulting are separate model-specific issues." + }, + { + "left": "issue:44242", + "right": "issue:44857", + "accept": false, + "reason": "MoE router loss behavior and LwDetrImageLoss AMP crash are different code paths." + }, + { + "left": "issue:44246", + "right": "issue:44871", + "accept": false, + "reason": "Import performance regression is unrelated to Gemma-3 eos_token_id inconsistency." + }, + { + "left": "issue:44779", + "right": "issue:45292", + "accept": false, + "reason": "Tokenizer incorrect results and resize_token_embeddings not updating output embeddings are different bugs." + }, + { + "left": "issue:44295", + "right": "issue:44749", + "accept": false, + "reason": "Position_ids buffer access error and Chinese-language performance regression are unrelated." + }, + { + "left": "issue:44877", + "right": "issue:45375", + "accept": false, + "reason": "Both involve config strictness, but they affect different model configs and fields; not the same underlying defect." + }, + { + "left": "issue:43441", + "right": "issue:43540", + "accept": false, + "reason": "FlashAttention failure in Ministral-3 and Qwen3OmniMoe video ValueError are unrelated." + }, + { + "left": "issue:43066", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions in v5, but they are different model families and different failure modes." + }, + { + "left": "issue:36246", + "right": "issue:43596", + "accept": false, + "reason": "Missing Qwen2_5_VLImageProcessor import and DeepSpeed zero3/BertModel IndexError are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 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-15354-364", + "nodes": [ + "issue:28282", + "issue:30990", + "issue:41628", + "issue:41950", + "issue:42673", + "issue:42886", + "issue:43065", + "issue:43232", + "issue:43257", + "issue:43262", + "issue:43295", + "issue:43381", + "issue:43408", + "issue:43441", + "issue:43475", + "issue:43540", + "issue:43606", + "issue:43646", + "issue:43653", + "issue:43673", + "issue:43704", + "issue:43901", + "issue:43931", + "issue:44060", + "issue:44077", + "issue:44162", + "issue:44246", + "issue:44297", + "issue:44336", + "issue:44368", + "issue:44466", + "issue:44496", + "issue:44623", + "issue:44734", + "issue:44869", + "issue:44933", + "issue:44977", + "issue:44987", + "issue:44995", + "issue:45003", + "issue:45137", + "issue:45200", + "issue:45229", + "issue:45278", + "issue:45292", + "issue:45335", + "issue:45356", + "issue:45459" + ], + "soft_pairs": [ + "issue:44987|issue:45459", + "issue:43901|issue:45003", + "issue:42886|issue:45229", + "issue:44995|issue:45200", + "issue:41628|issue:41950", + "issue:43257|issue:45137", + "issue:43065|issue:43408", + "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:44933|issue:45459", + "issue:43441|issue:43606", + "issue:43673|issue:44734", + "issue:44869|issue:45356", + "issue:43232|issue:43673", + "issue:43931|issue:44368", + "issue:44297|issue:44623", + "issue:43673|issue:45278", + "issue:43646|issue:44162", + "issue:43381|issue:45292", + "issue:43381|issue:45335", + "issue:42673|issue:43704", + "issue:28282|issue:30990" + ], + "prepared_review_unit_hash": "6d27097b48b3df46830f9265dca57dc522e091369ddd903124e72e6910c4197b", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36342, + "estimated_input_tokens": 9086, + "estimated_eval_tokens": 18428 + }, + "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 with a few loose topical overlaps. The only clear duplicate is the Qwen3ForCausalLM VRAM leak report, which appears in two identical-titled issues.", + "confidence": 0.96, + "canonical_issue_reason": "issue:42673 is the older, fuller report and the exact duplicate anchor for issue:43704.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:42673 is the strongest representative issue here: it is specific, well-scoped, and has an exact duplicate counterpart.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44987", + "right": "issue:45459", + "accept": false, + "reason": "Different failures: model loading/version compatibility vs tokenizer protobuf error masking." + }, + { + "left": "issue:43901", + "right": "issue:45003", + "accept": false, + "reason": "Docs mismatch for a pipeline option is unrelated to unsafe sys.modules access." + }, + { + "left": "issue:42886", + "right": "issue:45229", + "accept": false, + "reason": "Offline tokenizer cache loading and Gemma4 multi-GPU OOM are unrelated." + }, + { + "left": "issue:44995", + "right": "issue:45200", + "accept": false, + "reason": "A forward-pass crash in GlmMoeDsa is not the same as Gemma4 token-id defaults." + }, + { + "left": "issue:41628", + "right": "issue:41950", + "accept": false, + "reason": "Both mention image processors, but one is an import error and the other is a pipeline selection bug." + }, + { + "left": "issue:43257", + "right": "issue:45137", + "accept": false, + "reason": "Different DeepSpeed problems: weight conversion under accelerate+deepspeed vs a ZeRO3 deque error." + }, + { + "left": "issue:43065", + "right": "issue:43408", + "accept": false, + "reason": "SAM3 config/type warning and a dummy Conv2d implementation are separate bugs." + }, + { + "left": "issue:43232", + "right": "issue:44336", + "accept": false, + "reason": "Generation kwargs handling after sync_gpus is unrelated to ANSI code output in loading_report." + }, + { + "left": "issue:44496", + "right": "issue:45003", + "accept": false, + "reason": "Model config recognition failure is unrelated to unsafe sys.modules access." + }, + { + "left": "issue:43262", + "right": "issue:44246", + "accept": false, + "reason": "Audio chat-template sampling-rate defaulting is unrelated to import slowdowns." + }, + { + "left": "issue:43295", + "right": "issue:45292", + "accept": false, + "reason": "Processor.tokenizer regression and resize_token_embeddings behavior affect different code paths." + }, + { + "left": "issue:44060", + "right": "issue:44077", + "accept": false, + "reason": "Tied-weights warning in Qwen3-Next is not the same as patchtsmixer post_init policy." + }, + { + "left": "issue:43475", + "right": "issue:43540", + "accept": false, + "reason": "SAM3 vision encoder output attribute loss and Qwen3OmniMoe video processing are different bugs." + }, + { + "left": "issue:44466", + "right": "issue:45335", + "accept": false, + "reason": "Both involve weight handling, but they are different symptoms on different model families." + }, + { + "left": "issue:43653", + "right": "issue:44162", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and ESM2 breakage are unrelated." + }, + { + "left": "issue:44368", + "right": "issue:44977", + "accept": false, + "reason": "A tied-embeddings warning is not the same as a flash-attention generation failure." + }, + { + "left": "issue:44933", + "right": "issue:45459", + "accept": false, + "reason": "Missing image_utils import and protobuf decode-error handling are unrelated tokenizer/import issues." + }, + { + "left": "issue:43441", + "right": "issue:43606", + "accept": false, + "reason": "FlashAttention failure in Ministral-3 is unrelated to a CPU offload device mismatch in bark-small." + }, + { + "left": "issue:43673", + "right": "issue:44734", + "accept": false, + "reason": "Both involve generation/cache behavior, but they hit different paths and failure modes." + }, + { + "left": "issue:44869", + "right": "issue:45356", + "accept": false, + "reason": "Whisper timestamp decoding crash and Kimi tokenizer codec regression are not the same bug." + }, + { + "left": "issue:43232", + "right": "issue:43673", + "accept": false, + "reason": "Generation kwargs syncing and chunked-prefill cache missing are distinct generation issues." + }, + { + "left": "issue:43931", + "right": "issue:44368", + "accept": false, + "reason": "Weight-shape mismatch on model load is unrelated to tie_word_embeddings warnings." + }, + { + "left": "issue:44297", + "right": "issue:44623", + "accept": false, + "reason": "Tokenizer save_pretrained metadata mismatch and processor save_pretrained missing files are different serialization bugs." + }, + { + "left": "issue:43646", + "right": "issue:44162", + "accept": false, + "reason": "Generic v5 custom model initialization breakage is not the same as the ESM2-specific failure." + }, + { + "left": "issue:43381", + "right": "issue:45292", + "accept": false, + "reason": "Gradient checkpointing in eval mode and resize_token_embeddings behavior do not share a concrete bug." + }, + { + "left": "issue:42673", + "right": "issue:43704", + "accept": true, + "reason": "Exact same title and same VRAM leak description; clear duplicate pair." + }, + { + "left": "issue:28282", + "right": "issue:30990", + "accept": false, + "reason": "PyTorch import absence and Sentence Transformers loading hangs are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: the only clearly duplicated pair is issue:42673 / issue:43704, which match on title and are consistent with the older/fuller canonical choice. The remaining soft-edge verdicts are conservative false calls and the explanations stay within what the packet supports." + }, + "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-15354-364", + "nodes": [ + "issue:30990", + "issue:35141", + "issue:36010", + "issue:41628", + "issue:42222", + "issue:42371", + "issue:42831", + "issue:42898", + "issue:42994", + "issue:43232", + "issue:43377", + "issue:43475", + "issue:43479", + "issue:43482", + "issue:43525", + "issue:43643", + "issue:43673", + "issue:43701", + "issue:43873", + "issue:44038", + "issue:44077", + "issue:44079", + "issue:44164", + "issue:44188", + "issue:44222", + "issue:44242", + "issue:44351", + "issue:44361", + "issue:44568", + "issue:44623", + "issue:44661", + "issue:44749", + "issue:44792", + "issue:44869", + "issue:44933", + "issue:44945", + "issue:45003", + "issue:45081", + "issue:45092", + "issue:45103", + "issue:45200", + "issue:45230", + "issue:45310", + "issue:45362", + "issue:45375", + "issue:45381", + "issue:45397", + "issue:45478" + ], + "soft_pairs": [ + "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:42994|issue:43482", + "issue:43475|issue:43643", + "issue:43701|issue:44749", + "issue:42371|issue:45092", + "issue:45081|issue:45103", + "issue:45375|issue:45478", + "issue:41628|issue:44933", + "issue:45310|issue:45375", + "issue:43873|issue:44945", + "issue:42222|issue:45200", + "issue:43475|issue:44661", + "issue:42994|issue:43525", + "issue:42831|issue:44222", + "issue:45362|issue:45375", + "issue:45310|issue:45397", + "issue:43232|issue:44188", + "issue:44222|issue:45310", + "issue:41628|issue:44351", + "issue:42994|issue:44164", + "issue:43232|issue:44792", + "issue:44038|issue:45310", + "issue:30990|issue:35141", + "issue:43377|issue:44242", + "issue:43479|issue:44079", + "issue:43673|issue:44188", + "issue:36010|issue:41628", + "issue:43479|issue:44869", + "issue:42898|issue:44568" + ], + "prepared_review_unit_hash": "7062516464194a64f2c5045aee5daadba05a2c3200154981c54d730907c23753", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37493, + "estimated_input_tokens": 9374, + "estimated_eval_tokens": 19004 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: unrelated model-loading, tokenizer, generation, quantization, and multimodal bugs are mixed together. There is one clear near-duplicate theme around Qwen3.5 MoE from_pretrained failures, but the provided soft pairs do not include the strongest duplicate edge.", + "confidence": 0.68, + "canonical_issue_reason": "Best canonical issue is the most concrete and representative bug report in the only clear duplicate theme: the Qwen3.5 MoE from_pretrained failure.", + "canonical_pr_reason": null, + "best_issue_reason": "Strongest overall issue to keep as the cluster representative because it is specific, reproducible, and closest to the only apparent duplicate family in this set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44361", + "right": "issue:44749", + "accept": false, + "reason": "Different failures: MLukeTokenizer task AttributeError vs a Chinese-language performance slowdown report; no shared code-path bug." + }, + { + "left": "issue:44077", + "right": "issue:45003", + "accept": false, + "reason": "One is a patchtsmixer post_init policy issue, the other is a sys.modules access bug; unrelated symptoms and fixes." + }, + { + "left": "issue:44749", + "right": "issue:45310", + "accept": false, + "reason": "Performance regression in data filtering vs Qwen3.5 MoE from_pretrained error; not the same bug." + }, + { + "left": "issue:45230", + "right": "issue:45397", + "accept": false, + "reason": "Generic bug report vs Gemma-4 zero3 from_pretrained failure; insufficiently related." + }, + { + "left": "issue:44038", + "right": "issue:44623", + "accept": false, + "reason": "Qwen3-VL-Moe loading bug vs processor.save_pretrained missing files; different code paths." + }, + { + "left": "issue:43475", + "right": "issue:45381", + "accept": false, + "reason": "Both are video/VL issues, but for different models and different attributes/position-id logic." + }, + { + "left": "issue:42994", + "right": "issue:43482", + "accept": false, + "reason": "Quantized saving failure vs GGUF loading failure; opposite directions and different model formats." + }, + { + "left": "issue:43475", + "right": "issue:43643", + "accept": false, + "reason": "SAM 3 Video attribute error vs AutoConfig missing fields with trust_remote_code; unrelated root causes." + }, + { + "left": "issue:43701", + "right": "issue:44749", + "accept": false, + "reason": "Checkpoint key mismatch vs data filtering slowdown; no overlap in bug mechanism." + }, + { + "left": "issue:42371", + "right": "issue:45092", + "accept": false, + "reason": "TF32 settings warning vs InternVL2 meta-init incompatibility; different subsystems." + }, + { + "left": "issue:45081", + "right": "issue:45103", + "accept": false, + "reason": "Tokenizer regex patch crash vs auto_docstring annotations crash; different failures and files." + }, + { + "left": "issue:45375", + "right": "issue:45478", + "accept": false, + "reason": "Missing vision-config field vs generic Qwen3.5 MoE from_pretrained error; related model family but not the same concrete bug." + }, + { + "left": "issue:41628", + "right": "issue:44933", + "accept": false, + "reason": "Both are import problems, but they concern different symbols and likely different export mistakes." + }, + { + "left": "issue:45310", + "right": "issue:45375", + "accept": false, + "reason": "Qwen3.5 MoE from_pretrained error and a missing vision-config field are related only by model family, not the same bug." + }, + { + "left": "issue:43873", + "right": "issue:44945", + "accept": false, + "reason": "Quantization offloading issue vs pipeline parallelism output correctness; different runtime paths." + }, + { + "left": "issue:42222", + "right": "issue:45200", + "accept": false, + "reason": "Vitpose model breakage vs Gemma 4 token-type-id behavior; unrelated models and features." + }, + { + "left": "issue:43475", + "right": "issue:44661", + "accept": false, + "reason": "SAM 3 Video attribute error vs tokenizer mapping registration failure; unrelated." + }, + { + "left": "issue:42994", + "right": "issue:43525", + "accept": false, + "reason": "Quantized save failure vs Llama4Config missing pad_token_id; different objects and symptoms." + }, + { + "left": "issue:42831", + "right": "issue:44222", + "accept": false, + "reason": "FP8 accuracy issue vs FP8 save_pretrained/moe issue; both mention FP8 but not the same concrete bug." + }, + { + "left": "issue:45362", + "right": "issue:45375", + "accept": false, + "reason": "Qwen3.5 chat crash vs missing deepstack_visual_indexes field; different failure modes." + }, + { + "left": "issue:45310", + "right": "issue:45397", + "accept": false, + "reason": "Both mention from_pretrained, but for different models and different incompatibilities." + }, + { + "left": "issue:43232", + "right": "issue:44188", + "accept": false, + "reason": "Generation kwargs update bug vs attention kernel divergence under torch.compile; unrelated." + }, + { + "left": "issue:44222", + "right": "issue:45310", + "accept": false, + "reason": "FP8 save_pretrained moe issue vs Qwen3.5 MoE from_pretrained error; same broad family, different concrete path." + }, + { + "left": "issue:41628", + "right": "issue:44351", + "accept": false, + "reason": "Both are missing-import reports, but for different classes and different API surfaces." + }, + { + "left": "issue:42994", + "right": "issue:44164", + "accept": false, + "reason": "Quantized model saving failure vs extra_state serialization/deserialization failure; not the same bug." + }, + { + "left": "issue:43232", + "right": "issue:44792", + "accept": false, + "reason": "Generation kwargs handling vs a failed janus image-generation test; no shared underlying issue." + }, + { + "left": "issue:44038", + "right": "issue:45310", + "accept": false, + "reason": "Both are Qwen MoE-related, but one is a Qwen3-VL-Moe bug and the other is a Qwen3.5 MoE from_pretrained failure; too different to merge." + }, + { + "left": "issue:30990", + "right": "issue:35141", + "accept": false, + "reason": "Sentence Transformers loading hang vs embedding resize/post_init reinitialization; unrelated." + }, + { + "left": "issue:43377", + "right": "issue:44242", + "accept": false, + "reason": "MIMI padding-mask mismatch vs MoE load-balancing loss omission; different model components and symptoms." + }, + { + "left": "issue:43479", + "right": "issue:44079", + "accept": false, + "reason": "Config initialization with None vs ModelOutput key assignment bug; different object layers." + }, + { + "left": "issue:43673", + "right": "issue:44188", + "accept": false, + "reason": "Missing GenerationMixin cache in v5 chunked prefill vs attention-kernel divergence under compile; unrelated." + }, + { + "left": "issue:36010", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but they involve different removed/missing symbols and different fixes." + }, + { + "left": "issue:43479", + "right": "issue:44869", + "accept": false, + "reason": "Multimodal config defaulting bug vs Whisper timestamp decode crash; unrelated." + }, + { + "left": "issue:42898", + "right": "issue:44568", + "accept": false, + "reason": "Both are tokenizer behavior changes in v5, but they are different regressions affecting different APIs and models." + } + ] + }, + "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-15354-364", + "nodes": [ + "issue:22355", + "issue:28282", + "issue:41628", + "issue:42175", + "issue:42757", + "issue:42831", + "issue:42898", + "issue:43278", + "issue:43295", + "issue:43479", + "issue:43644", + "issue:43723", + "issue:43784", + "issue:43824", + "issue:43874", + "issue:43976", + "issue:44060", + "issue:44077", + "issue:44188", + "issue:44206", + "issue:44222", + "issue:44242", + "issue:44263", + "issue:44295", + "issue:44297", + "issue:44360", + "issue:44373", + "issue:44442", + "issue:44448", + "issue:44483", + "issue:44484", + "issue:44514", + "issue:44568", + "issue:44623", + "issue:44734", + "issue:44779", + "issue:44933", + "issue:44964", + "issue:44977", + "issue:45092", + "issue:45103", + "issue:45125", + "issue:45137", + "issue:45276", + "issue:45310", + "issue:45356", + "issue:45406", + "issue:45459" + ], + "soft_pairs": [ + "issue:44964|issue:45137", + "issue:43976|issue:44933", + "issue:42175|issue:44933", + "issue:43784|issue:43824", + "issue:43278|issue:45137", + "issue:44623|issue:45310", + "issue:44188|issue:44373", + "issue:44060|issue:45125", + "issue:44734|issue:45103", + "issue:42898|issue:45356", + "issue:45137|issue:45276", + "issue:43723|issue:44568", + "issue:42898|issue:44779", + "issue:43874|issue:45092", + "issue:44077|issue:44242", + "issue:42898|issue:44448", + "issue:41628|issue:45459", + "issue:44222|issue:44297", + "issue:42898|issue:43295", + "issue:43479|issue:43644", + "issue:44442|issue:45459", + "issue:44222|issue:44623", + "issue:42831|issue:44623", + "issue:44206|issue:45103", + "issue:22355|issue:28282", + "issue:44297|issue:45310", + "issue:44483|issue:45406", + "issue:41628|issue:42757", + "issue:44222|issue:44484", + "issue:44263|issue:44360", + "issue:43723|issue:44977", + "issue:44263|issue:45092", + "issue:44295|issue:44623", + "issue:44514|issue:45103" + ], + "prepared_review_unit_hash": "69ad125856676450611c28f1a50bf050cd8281176920b04d5710d56f19de4c2b", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37842, + "estimated_input_tokens": 9461, + "estimated_eval_tokens": 19178 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is largely heterogeneous: mostly unrelated Transformers import, tokenizer, processor, and model-regression issues. None of the soft pairs look like the same concrete bug/change, so I would not merge any of them as duplicates.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "issue:42898 is the strongest representative of the cluster\u2019s recurring theme of Transformers v5 tokenizer behavior regressions, but it is still not a true duplicate anchor for the whole set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44964", + "right": "issue:45137", + "accept": false, + "reason": "Different failures: Phi-4 multimodal model loading vs DeepSpeed ZeRO3 deque underflow." + }, + { + "left": "issue:43976", + "right": "issue:44933", + "accept": false, + "reason": "Python-version incompatibility is unrelated to the missing image_utils import." + }, + { + "left": "issue:42175", + "right": "issue:44933", + "accept": false, + "reason": "Backend packaging for torch/tensorflow is unrelated to an import from image_utils." + }, + { + "left": "issue:43784", + "right": "issue:43824", + "accept": false, + "reason": "Sentence-transformers import NameError and missing Qwen2.5-VL export are different code paths." + }, + { + "left": "issue:43278", + "right": "issue:45137", + "accept": false, + "reason": "BF16->FP32 eval dtype drift is unrelated to ZeRO3 deque errors." + }, + { + "left": "issue:44623", + "right": "issue:45310", + "accept": false, + "reason": "Processor save_pretrained missing files is not the same as Qwen3.5-MoE from_pretrained failure." + }, + { + "left": "issue:44188", + "right": "issue:44373", + "accept": false, + "reason": "torch.compile branching bug and a docstring issue are unrelated." + }, + { + "left": "issue:44060", + "right": "issue:45125", + "accept": false, + "reason": "Wrong tied-weights warning is not the same as a missing _tp_plan attribute." + }, + { + "left": "issue:44734", + "right": "issue:45103", + "accept": false, + "reason": "KV-cache indexing crash and auto_docstring AttributeError are unrelated." + }, + { + "left": "issue:42898", + "right": "issue:45356", + "accept": false, + "reason": "Both are tokenizer regressions, but they are different bugs with different symptoms and fixes." + }, + { + "left": "issue:45137", + "right": "issue:45276", + "accept": false, + "reason": "ZeRO3 deque underflow is unrelated to gemma4 resize_token_embeddings propagation." + }, + { + "left": "issue:43723", + "right": "issue:44568", + "accept": false, + "reason": "Tokenizer loading failure and BOS/EOS add_special_tokens behavior are distinct issues." + }, + { + "left": "issue:42898", + "right": "issue:44779", + "accept": false, + "reason": "clean_up_tokenization_spaces behavior change is not the same as Deepseek tokenizer output corruption." + }, + { + "left": "issue:43874", + "right": "issue:45092", + "accept": false, + "reason": "Missing image patch count method is unrelated to remote-code checkpoint meta-init compatibility." + }, + { + "left": "issue:44077", + "right": "issue:44242", + "accept": false, + "reason": "Optional post_init policy and missing MoE load-balancing loss are different implementation bugs." + }, + { + "left": "issue:42898", + "right": "issue:44448", + "accept": false, + "reason": "Token cleanup regression and Pegasus output drift are not the same concrete bug." + }, + { + "left": "issue:41628", + "right": "issue:45459", + "accept": false, + "reason": "Missing AutoImageProcessor import and protobuf error masking are unrelated." + }, + { + "left": "issue:44222", + "right": "issue:44297", + "accept": false, + "reason": "FP8 save_pretrained/moe issues do not match tokenizer_class metadata mismatch on save_pretrained." + }, + { + "left": "issue:42898", + "right": "issue:43295", + "accept": false, + "reason": "Tokenizer cleanup change is unrelated to processor.tokenizer regression in v4.57.5." + }, + { + "left": "issue:43479", + "right": "issue:43644", + "accept": false, + "reason": "Default multimodal config initialization and junk non-persistent buffers are different bugs." + }, + { + "left": "issue:44442", + "right": "issue:45459", + "accept": false, + "reason": "FastSpeech2ConformerTokenizer loading and protobuf error handling are different failure modes." + }, + { + "left": "issue:44222", + "right": "issue:44623", + "accept": false, + "reason": "MoE FP8 save_pretrained and processor.save_pretrained missing files are not the same bug." + }, + { + "left": "issue:42831", + "right": "issue:44623", + "accept": false, + "reason": "FP8 accuracy regression is unrelated to processor.save_pretrained file omission." + }, + { + "left": "issue:44206", + "right": "issue:45103", + "accept": false, + "reason": "FeatureExtractor center-arg crash and auto_docstring annotation handling are unrelated." + }, + { + "left": "issue:22355", + "right": "issue:28282", + "accept": false, + "reason": "onnx module import failure is unrelated to missing PyTorch backend detection." + }, + { + "left": "issue:44297", + "right": "issue:45310", + "accept": false, + "reason": "Tokenizer config mismatch on save and Qwen3.5-MoE from_pretrained failure are different issues." + }, + { + "left": "issue:44483", + "right": "issue:45406", + "accept": false, + "reason": "Chat/completions request rejection and Gemma4Processor _tokenizer crash are different server/processor bugs." + }, + { + "left": "issue:41628", + "right": "issue:42757", + "accept": false, + "reason": "Transformer export import error and huggingface_hub offline_mode import error are unrelated." + }, + { + "left": "issue:44222", + "right": "issue:44484", + "accept": false, + "reason": "FP8 save_pretrained bug and a max_shard_size question are not the same change." + }, + { + "left": "issue:44263", + "right": "issue:44360", + "accept": false, + "reason": "These are both DSA-indexer-related, but one is about torch.split outputs and the other about a missing ReLU; not the same bug." + }, + { + "left": "issue:44295", + "right": "issue:45310", + "accept": false, + "reason": "position_ids buffer access error is unrelated to Qwen3.5-MoE from_pretrained failure." + }, + { + "left": "issue:43723", + "right": "issue:44977", + "accept": false, + "reason": "Tokenizer loading issue is unrelated to flash-attention generation regressions." + }, + { + "left": "issue:44263", + "right": "issue:45092", + "accept": false, + "reason": "GlmMoeDsaIndexer split handling and InternVL2 meta-init incompatibility are unrelated." + }, + { + "left": "issue:44295", + "right": "issue:44623", + "accept": false, + "reason": "position_ids buffer error is unrelated to processor.save_pretrained missing files." + }, + { + "left": "issue:44514", + "right": "issue:45103", + "accept": false, + "reason": "Qwen2_5_VLProcessor batched apply_chat_template crash and auto_docstring AttributeError 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": 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-15354-364", + "nodes": [ + "issue:30990", + "issue:33357", + "issue:38175", + "issue:39401", + "issue:41628", + "issue:42371", + "issue:42757", + "issue:42831", + "issue:42886", + "issue:42907", + "issue:42994", + "issue:43066", + "issue:43122", + "issue:43232", + "issue:43408", + "issue:43475", + "issue:43479", + "issue:43576", + "issue:43723", + "issue:43784", + "issue:43824", + "issue:43937", + "issue:43976", + "issue:44038", + "issue:44117", + "issue:44261", + "issue:44263", + "issue:44295", + "issue:44297", + "issue:44373", + "issue:44442", + "issue:44485", + "issue:44488", + "issue:44514", + "issue:44521", + "issue:44704", + "issue:44743", + "issue:44871", + "issue:44913", + "issue:44945", + "issue:44991", + "issue:45092", + "issue:45245", + "issue:45278", + "issue:45290", + "issue:45310", + "issue:45357" + ], + "soft_pairs": [ + "issue:43479|issue:44913", + "issue:44373|issue:44485", + "issue:43723|issue:44871", + "issue:44261|issue:45092", + "issue:30990|issue:38175", + "issue:43723|issue:44991", + "issue:43723|issue:44442", + "issue:45092|issue:45357", + "issue:42886|issue:42994", + "issue:42831|issue:45310", + "issue:41628|issue:43824", + "issue:42831|issue:44038", + "issue:43576|issue:43784", + "issue:43122|issue:44521", + "issue:44263|issue:44373", + "issue:42757|issue:45278", + "issue:44117|issue:44704", + "issue:42886|issue:43066", + "issue:43122|issue:44514", + "issue:30990|issue:33357", + "issue:44704|issue:45290", + "issue:43232|issue:45245", + "issue:43232|issue:44743", + "issue:42371|issue:44263", + "issue:42907|issue:44945", + "issue:44295|issue:44297", + "issue:43976|issue:44945", + "issue:39401|issue:44488", + "issue:42907|issue:43937", + "issue:43408|issue:43475" + ], + "prepared_review_unit_hash": "3a9cc870cdf2e8bd41d9efcc058a302167f345f6a8ddfd1b97df4264991d11b5", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 35656, + "estimated_input_tokens": 8914, + "estimated_eval_tokens": 18084 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These items are mostly unrelated single-issue reports across tokenizer loading, model config regressions, quantization/saving, import errors, and a few model-specific bugs. No soft pair is a clear same-bug duplicate; most share only a broad subsystem or symptom family.", + "confidence": 0.82, + "canonical_issue_reason": "issue:43723 is the best general anchor: it is a concrete, version-scoped tokenizer-loading regression with a broad symptom that matches the cluster\u2019s repeated loading/import complaints more than the other, narrower reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43723 is the strongest standalone issue to represent this set because it is specific, reproducible, and broadly relevant to the tokenizer-loading regressions seen here.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43479", + "right": "issue:44913", + "accept": false, + "reason": "Different bugs: one is wrong default vision/audio config initialization in Phi4MultimodalConfig, the other is GPTNeoX rotary_pct not persisting on reload." + }, + { + "left": "issue:44373", + "right": "issue:44485", + "accept": false, + "reason": "Docstring typo versus RoPE implementation behavior; no shared concrete failure." + }, + { + "left": "issue:43723", + "right": "issue:44871", + "accept": false, + "reason": "Both concern tokenizer/model config behavior, but one is a tokenizer loading failure and the other is an eos_token_id mismatch." + }, + { + "left": "issue:44261", + "right": "issue:45092", + "accept": false, + "reason": "Different code paths and symptoms: MLA rms_norm_eps precision issue versus remote-code/meta-init incompatibility." + }, + { + "left": "issue:30990", + "right": "issue:38175", + "accept": false, + "reason": "Both are model-related user-facing failures, but one is a loading hang and the other is unexpected zero probabilities on a specific model." + }, + { + "left": "issue:43723", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer-loading regressions, but they appear model-specific and not the same concrete bug." + }, + { + "left": "issue:43723", + "right": "issue:44442", + "accept": false, + "reason": "AutoTokenizer loading regression versus a specific FastSpeech2ConformerTokenizer load failure; too little evidence they are the same defect." + }, + { + "left": "issue:45092", + "right": "issue:45357", + "accept": false, + "reason": "Both involve model compatibility/regression, but one is meta initialization for old InternVL2 checkpoints and the other is incorrect save_pretrained key handling in Qwen3.5." + }, + { + "left": "issue:42886", + "right": "issue:42994", + "accept": false, + "reason": "Offline tokenizer cache loading failure is not the same as quantized model saving failure." + }, + { + "left": "issue:42831", + "right": "issue:45310", + "accept": false, + "reason": "FineGrainedFP8 accuracy degradation and Qwen3.5 from_pretrained errors are different problems." + }, + { + "left": "issue:41628", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but they concern different missing symbols and different model/component paths." + }, + { + "left": "issue:42831", + "right": "issue:44038", + "accept": false, + "reason": "Shared Qwen context is too broad; one is FP8 accuracy and the other is a model-specific from_pretrained bug." + }, + { + "left": "issue:43576", + "right": "issue:43784", + "accept": false, + "reason": "transformers env command breakage and sentence-transformers import NameError are distinct failures." + }, + { + "left": "issue:43122", + "right": "issue:44521", + "accept": false, + "reason": "Different tokenizer/template issues: changed tokenization behavior versus all-zero assistant masks for multimodal inputs." + }, + { + "left": "issue:44263", + "right": "issue:44373", + "accept": false, + "reason": "Tensor split/indexing bug versus a docstring problem; not the same issue." + }, + { + "left": "issue:42757", + "right": "issue:45278", + "accept": false, + "reason": "Specific missing import from huggingface_hub versus a broad set of import errors after upgrade." + }, + { + "left": "issue:44117", + "right": "issue:44704", + "accept": false, + "reason": "Tokenizer mapping returning None and AutoProcessor kwargs passthrough are different code paths." + }, + { + "left": "issue:42886", + "right": "issue:43066", + "accept": false, + "reason": "Offline cache loading failure and wrong tokenizer decoder type are unrelated tokenizer issues." + }, + { + "left": "issue:43122", + "right": "issue:44514", + "accept": false, + "reason": "Tokenization differences across versions versus apply_chat_template crashing on batched padded=False input." + }, + { + "left": "issue:44704", + "right": "issue:45290", + "accept": false, + "reason": "AutoProcessor cached_file kwargs bug and assistant/tool-call chat-template crash are different concrete failures." + }, + { + "left": "issue:43232", + "right": "issue:45245", + "accept": false, + "reason": "Generation kwargs/sync_gpus bug versus a categories-limit runtime error." + }, + { + "left": "issue:43232", + "right": "issue:44743", + "accept": false, + "reason": "Both are generation/model-runtime bugs, but they affect different code paths and symptoms." + }, + { + "left": "issue:42371", + "right": "issue:44263", + "accept": false, + "reason": "TF32 configuration warning and torch.split return-value issue are unrelated." + }, + { + "left": "issue:42907", + "right": "issue:44945", + "accept": false, + "reason": "Saving dequantized models and incorrect output under pipeline parallelism are not the same bug." + }, + { + "left": "issue:44295", + "right": "issue:44297", + "accept": false, + "reason": "Reading a registered position_ids buffer and tokenizer_class mismatch in tokenizer_config.json are separate issues." + }, + { + "left": "issue:43976", + "right": "issue:44945", + "accept": false, + "reason": "Python version compatibility and pipeline-parallel output corruption are unrelated." + }, + { + "left": "issue:39401", + "right": "issue:44488", + "accept": false, + "reason": "Tokenizer offset_mapping bug and failure to load a specific model are different problems." + }, + { + "left": "issue:42907", + "right": "issue:43937", + "accept": false, + "reason": "Model saving failure and invalid GenerationConfig are unrelated." + }, + { + "left": "issue:43408", + "right": "issue:43475", + "accept": false, + "reason": "Model-type warning during instantiation and missing fpn_position_embeddings attribute are different SAM3 video 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-15354-364", + "nodes": [ + "issue:22355", + "issue:29127", + "issue:30990", + "issue:36296", + "issue:38617", + "issue:39401", + "issue:41628", + "issue:42222", + "issue:42371", + "issue:42673", + "issue:42757", + "issue:42831", + "issue:42907", + "issue:43064", + "issue:43116", + "issue:43122", + "issue:43278", + "issue:43441", + "issue:43452", + "issue:43550", + "issue:43673", + "issue:43784", + "issue:43824", + "issue:43937", + "issue:43992", + "issue:44038", + "issue:44222", + "issue:44261", + "issue:44295", + "issue:44297", + "issue:44351", + "issue:44360", + "issue:44373", + "issue:44485", + "issue:44521", + "issue:44625", + "issue:44704", + "issue:44779", + "issue:44829", + "issue:44912", + "issue:44945", + "issue:44977", + "issue:45072", + "issue:45200", + "issue:45216", + "issue:45278", + "issue:45446", + "issue:45478" + ], + "soft_pairs": [ + "issue:44625|issue:44829", + "issue:29127|issue:30990", + "issue:44360|issue:44485", + "issue:41628|issue:45278", + "issue:43116|issue:45200", + "issue:39401|issue:44521", + "issue:44038|issue:44297", + "issue:44261|issue:44373", + "issue:43122|issue:44779", + "issue:42371|issue:42831", + "issue:44038|issue:44295", + "issue:36296|issue:43116", + "issue:42907|issue:43452", + "issue:43992|issue:44704", + "issue:44038|issue:44222", + "issue:43278|issue:44945", + "issue:43122|issue:44977", + "issue:45216|issue:45478", + "issue:43824|issue:44351", + "issue:42907|issue:44222", + "issue:42757|issue:44351", + "issue:42222|issue:43441", + "issue:42907|issue:44912", + "issue:42673|issue:45072", + "issue:43784|issue:45446", + "issue:43064|issue:43673", + "issue:22355|issue:29127", + "issue:43550|issue:43784", + "issue:38617|issue:45278", + "issue:43937|issue:44485" + ], + "prepared_review_unit_hash": "5a78981a98a71c85d5e70a9665cf595b20e3d0275e045b5b4fc7f7814d5d4199", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36231, + "estimated_input_tokens": 9058, + "estimated_eval_tokens": 18372 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is very heterogeneous: most items are unrelated model-specific bugs, import regressions, or API/documentation issues. None of the soft pairs look like true duplicates; they mostly share only a broad theme (imports, tokenizers, quantization, or training) rather than the same concrete failure mode.", + "confidence": 0.89, + "canonical_issue_reason": "issue:45278 is the broadest regression report in the set and the most plausible umbrella representative, but it still does not actually unify the other issues.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 is the strongest cluster representative because it describes a wide import-regression symptom after upgrade, which is more general than the other narrowly scoped reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44625", + "right": "issue:44829", + "accept": false, + "reason": "Different bugs: Qwen3.5 config propagation vs flash-attention-3 training degeneration." + }, + { + "left": "issue:29127", + "right": "issue:30990", + "accept": false, + "reason": "Unrelated: layoutlmv3 error-message clarity vs Sentence Transformers loading hang." + }, + { + "left": "issue:44360", + "right": "issue:44485", + "accept": false, + "reason": "Different code paths: DSA indexer/ReLU concern vs GLM-5 RoPE implementation." + }, + { + "left": "issue:41628", + "right": "issue:45278", + "accept": false, + "reason": "One is a specific missing import, the other is a broad upgrade-time import regression; not the same underlying failure." + }, + { + "left": "issue:43116", + "right": "issue:45200", + "accept": false, + "reason": "Different symptoms and subsystems: example-script multi-label output vs Gemma 4 multimodal token IDs." + }, + { + "left": "issue:39401", + "right": "issue:44521", + "accept": false, + "reason": "Tokenizer offset mapping bug vs chat-template assistant mask bug." + }, + { + "left": "issue:44038", + "right": "issue:44297", + "accept": false, + "reason": "Qwen3-VL-Moe runtime/import behavior vs tokenizer save metadata mismatch." + }, + { + "left": "issue:44261", + "right": "issue:44373", + "accept": false, + "reason": "MLA RMS-norm precision/config bug vs a docstring issue for position_ids." + }, + { + "left": "issue:43122", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer-related, but one is a general version regression and the other is DeepSeek-specific incorrect tokenization." + }, + { + "left": "issue:42371", + "right": "issue:42831", + "accept": false, + "reason": "TF32 API/deprecation issue is unrelated to FineGrainedFP8 accuracy." + }, + { + "left": "issue:44038", + "right": "issue:44295", + "accept": false, + "reason": "Qwen3-VL-Moe bug is not the same as reading position_ids after buffer registration." + }, + { + "left": "issue:36296", + "right": "issue:43116", + "accept": false, + "reason": "Tensor-parallel training bug and classification-example output bug are unrelated." + }, + { + "left": "issue:42907", + "right": "issue:43452", + "accept": false, + "reason": "Saving dequantized models is a different path from gguf_file loading failures." + }, + { + "left": "issue:43992", + "right": "issue:44704", + "accept": false, + "reason": "Missing embed_tokens.weight on load is unrelated to AutoProcessor kwargs forwarding." + }, + { + "left": "issue:44038", + "right": "issue:44222", + "accept": false, + "reason": "Both mention MoE, but one is a Qwen3-VL-Moe failure and the other is FP8 save_pretrained behavior." + }, + { + "left": "issue:43278", + "right": "issue:44945", + "accept": false, + "reason": "Embedding dtype drift in eval is not the same as pipeline-parallel incorrect output." + }, + { + "left": "issue:43122", + "right": "issue:44977", + "accept": false, + "reason": "Tokenizer regression and flash-attention generation failure are different bugs." + }, + { + "left": "issue:45216", + "right": "issue:45478", + "accept": false, + "reason": "One concerns save_pretrained checkpoint correctness; the other is a from_pretrained import error." + }, + { + "left": "issue:43824", + "right": "issue:44351", + "accept": false, + "reason": "Both are import failures, but they are different missing symbols from different modules." + }, + { + "left": "issue:42907", + "right": "issue:44222", + "accept": false, + "reason": "Different model/formats and different save paths: dequantized Ministral/Devstral vs FP8 MoE save_pretrained." + }, + { + "left": "issue:42757", + "right": "issue:44351", + "accept": false, + "reason": "External huggingface_hub import failure is not the same bug as missing HybridCache in transformers." + }, + { + "left": "issue:42222", + "right": "issue:43441", + "accept": false, + "reason": "VitPose model breakage and Ministral flash-attention failure are unrelated." + }, + { + "left": "issue:42907", + "right": "issue:44912", + "accept": false, + "reason": "Saving dequantized models and MXFP4 quantization fallback are different problems." + }, + { + "left": "issue:22355", + "right": "issue:29127", + "accept": false, + "reason": "ONNX module import error is unrelated to layoutlmv3 box-content error messaging." + }, + { + "left": "issue:43550", + "right": "issue:43784", + "accept": false, + "reason": "torch.compile/SDPA training bug vs sentence-transformers import error; different code paths." + }, + { + "left": "issue:38617", + "right": "issue:45278", + "accept": false, + "reason": "Specific missing layer_type_validation import is not the same as a broad multi-import regression after upgrade." + }, + { + "left": "issue:43937", + "right": "issue:44485", + "accept": false, + "reason": "GLM-5 generation-config validation and GLM-5 RoPE implementation are different defects." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 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-15354-364", + "nodes": [ + "issue:33453", + "issue:38617", + "issue:39401", + "issue:41628", + "issue:42222", + "issue:42371", + "issue:42673", + "issue:43116", + "issue:43408", + "issue:43421", + "issue:43450", + "issue:43493", + "issue:43577", + "issue:43582", + "issue:43742", + "issue:43825", + "issue:43827", + "issue:43874", + "issue:43906", + "issue:43931", + "issue:43937", + "issue:43976", + "issue:43986", + "issue:44038", + "issue:44075", + "issue:44117", + "issue:44222", + "issue:44261", + "issue:44291", + "issue:44410", + "issue:44483", + "issue:44484", + "issue:44496", + "issue:44779", + "issue:44829", + "issue:44936", + "issue:44945", + "issue:44977", + "issue:45071", + "issue:45081", + "issue:45083", + "issue:45216", + "issue:45310", + "issue:45405", + "issue:45446", + "issue:45464", + "issue:45478", + "issue:45542" + ], + "soft_pairs": [ + "issue:43116|issue:43493", + "issue:42371|issue:43874", + "issue:44075|issue:44936", + "issue:44222|issue:45478", + "issue:43116|issue:43825", + "issue:44038|issue:45478", + "issue:43976|issue:45446", + "issue:39401|issue:44779", + "issue:43931|issue:44945", + "issue:45083|issue:45446", + "issue:33453|issue:39401", + "issue:43937|issue:44977", + "issue:43827|issue:45446", + "issue:42673|issue:43906", + "issue:44291|issue:45071", + "issue:43450|issue:43986", + "issue:45405|issue:45446", + "issue:44829|issue:45216", + "issue:44484|issue:45310", + "issue:39401|issue:43421", + "issue:38617|issue:41628", + "issue:44117|issue:45081", + "issue:42673|issue:43582", + "issue:44410|issue:45310", + "issue:43408|issue:43742", + "issue:42673|issue:43577", + "issue:42222|issue:45542", + "issue:42371|issue:44261", + "issue:43937|issue:44496", + "issue:44483|issue:45464" + ], + "prepared_review_unit_hash": "d0d6e5a494ce664ac3171acc31de4cd3b32270434ecf20afeb6242803f004693", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36622, + "estimated_input_tokens": 9156, + "estimated_eval_tokens": 18568 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is very heterogeneous: mostly unrelated Transformers bug reports spanning tokenizers, model loading, processors, API behavior, and version regressions. A few pairs are thematically close, but none look like the same concrete issue/change with enough confidence to merge as duplicates.", + "confidence": 0.83, + "canonical_issue_reason": "No single issue safely represents the whole cluster; the items are too diverse. If forced to pick one representative regression report, issue 45216 is the clearest and most concrete.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45216 is the strongest representative candidate because it is a specific, version-bounded regression report with a clear model family and behavior.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43116", + "right": "issue:43493", + "accept": false, + "reason": "Different bugs: one is multi-label classification example output, the other is a SigLIP2 implementation discrepancy." + }, + { + "left": "issue:42371", + "right": "issue:43874", + "accept": false, + "reason": "Unrelated subsystems: TF32 API settings versus a missing image-processor method in GLM46V." + }, + { + "left": "issue:44075", + "right": "issue:44936", + "accept": false, + "reason": "Different code paths: optimizer argument handling versus trainer evaluate/train lifecycle." + }, + { + "left": "issue:44222", + "right": "issue:45478", + "accept": false, + "reason": "Same broad Qwen3.5/MoE area, but one is save_pretrained FP8 behavior and the other is from_pretrained loading failure; not the same concrete bug." + }, + { + "left": "issue:43116", + "right": "issue:43825", + "accept": false, + "reason": "One is an example script classification bug; the other is a pipeline() v5 error-message issue." + }, + { + "left": "issue:44038", + "right": "issue:45478", + "accept": false, + "reason": "Both mention Qwen3.5/MoE, but they describe different failures and likely different code paths." + }, + { + "left": "issue:43976", + "right": "issue:45446", + "accept": false, + "reason": "Different compatibility problems: Python version support versus a PyTorch version check bug." + }, + { + "left": "issue:39401", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions, but on different models and symptoms (offset mapping vs incorrect generation output)." + }, + { + "left": "issue:43931", + "right": "issue:44945", + "accept": false, + "reason": "Model weight-shape loading mismatch is unrelated to incorrect outputs under pipeline parallelism." + }, + { + "left": "issue:45083", + "right": "issue:45446", + "accept": false, + "reason": "Different areas: qwen3_omni_moe feature extraction helper versus flex_attention import/version checking." + }, + { + "left": "issue:33453", + "right": "issue:39401", + "accept": false, + "reason": "Both involve tokenizers, but one is loading regression and the other is offset_mapping correctness on Qwen3." + }, + { + "left": "issue:43937", + "right": "issue:44977", + "accept": false, + "reason": "GenerationConfig validation error is not the same as flash-attention generation corruption." + }, + { + "left": "issue:43827", + "right": "issue:45446", + "accept": false, + "reason": "Documentation references to removed pipeline() API are unrelated to flex_attention import gating." + }, + { + "left": "issue:42673", + "right": "issue:43906", + "accept": false, + "reason": "No clear shared bug; one is a VRAM leak in dataloader threads, the other is a reproduction of a different issue." + }, + { + "left": "issue:44291", + "right": "issue:45071", + "accept": false, + "reason": "Both are loading/type-related regressions, but they concern different APIs and error causes." + }, + { + "left": "issue:43450", + "right": "issue:43986", + "accept": false, + "reason": "A batched video-shape bug is not the same as a missing- torchvision crash on AutoProcessor load." + }, + { + "left": "issue:45405", + "right": "issue:45446", + "accept": false, + "reason": "Version pinning for PEFT is unrelated to PyTorch version checks in flex_attention." + }, + { + "left": "issue:44829", + "right": "issue:45216", + "accept": false, + "reason": "Different failure modes: degenerate training with flash_attention_3 versus incorrect checkpoint saving." + }, + { + "left": "issue:44484", + "right": "issue:45310", + "accept": false, + "reason": "Shard-size question is unrelated to Qwen3.5 MoE from_pretrained errors." + }, + { + "left": "issue:39401", + "right": "issue:43421", + "accept": false, + "reason": "Tokenizer offset_mapping regression is not the same as runtime special-token post-processor updates." + }, + { + "left": "issue:38617", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but for different symbols and likely different export paths." + }, + { + "left": "issue:44117", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer mapping None-handling is a different bug from the Mistral regex patch crash." + }, + { + "left": "issue:42673", + "right": "issue:43582", + "accept": false, + "reason": "VRAM leak in dataloader threads and Apple Silicon TypeError are unrelated." + }, + { + "left": "issue:44410", + "right": "issue:45310", + "accept": false, + "reason": "Missing projections in qwen3next is a model-architecture issue, not the same as Qwen3.5 MoE loading failure." + }, + { + "left": "issue:43408", + "right": "issue:43742", + "accept": false, + "reason": "SAM3 tracker warning and MobileLLM key error are distinct loading/config issues." + }, + { + "left": "issue:42673", + "right": "issue:43577", + "accept": false, + "reason": "Different bugs: VRAM leak versus dtype not propagating on BLIP2 load." + }, + { + "left": "issue:42222", + "right": "issue:45542", + "accept": false, + "reason": "VitPose model breakage and TensorFlow/tensorboard backend detection are unrelated." + }, + { + "left": "issue:42371", + "right": "issue:44261", + "accept": false, + "reason": "TF32 API control and MLA rms_norm_eps precision mismatch do not share the same bug." + }, + { + "left": "issue:43937", + "right": "issue:44496", + "accept": false, + "reason": "Invalid GenerationConfig and unrecognized-model config.json errors are different loading/validation failures." + }, + { + "left": "issue:44483", + "right": "issue:45464", + "accept": false, + "reason": "Both involve chat/completions, but one is a v5.3 request acceptance issue and the other is a streaming failure on a specific model." + } + ] + }, + "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": 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-15354-364", + "nodes": [ + "issue:38617", + "issue:41628", + "issue:42175", + "issue:42222", + "issue:42617", + "issue:42757", + "issue:43012", + "issue:43064", + "issue:43066", + "issue:43097", + "issue:43122", + "issue:43335", + "issue:43408", + "issue:43441", + "issue:43452", + "issue:43531", + "issue:43582", + "issue:43673", + "issue:43874", + "issue:43931", + "issue:43976", + "issue:43986", + "issue:44038", + "issue:44117", + "issue:44220", + "issue:44261", + "issue:44263", + "issue:44315", + "issue:44368", + "issue:44410", + "issue:44484", + "issue:44485", + "issue:44493", + "issue:44704", + "issue:44829", + "issue:44933", + "issue:44987", + "issue:45003", + "issue:45071", + "issue:45103", + "issue:45245", + "issue:45278", + "issue:45341", + "issue:45372", + "issue:45375", + "issue:45459", + "issue:45478", + "issue:45542" + ], + "soft_pairs": [ + "issue:38617|issue:44220", + "issue:42617|issue:45341", + "issue:42175|issue:42222", + "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:44410|issue:45478", + "issue:43874|issue:44261", + "issue:43441|issue:44829", + "issue:43986|issue:44484", + "issue:43012|issue:43408", + "issue:44263|issue:44485", + "issue:41628|issue:45071", + "issue:45071|issue:45542", + "issue:45003|issue:45341", + "issue:43931|issue:44410", + "issue:43986|issue:45341", + "issue:43097|issue:43452", + "issue:44117|issue:44987", + "issue:43335|issue:45071", + "issue:43408|issue:44368", + "issue:43931|issue:44038", + "issue:43874|issue:44263", + "issue:43986|issue:45372", + "issue:44829|issue:45245", + "issue:43976|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:43986|issue:44117", + "issue:43582|issue:45341" + ], + "prepared_review_unit_hash": "543b655ccc2d7b275cd0dc3057279674cd43acf8e3ffab84a4f05cd1492557cc", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 37, + "serialized_chars": 38627, + "estimated_input_tokens": 9657, + "estimated_eval_tokens": 19570 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a loose mix of Transformers regression reports: import errors, tokenizer/config changes, and several model-specific bugs. The soft links look similarity-based rather than true duplicates, so I would not merge any of the candidate pairs.", + "confidence": 0.89, + "canonical_issue_reason": "issue:45278 is the broadest umbrella for the cluster\u2019s recurring theme of import regressions after upgrading Transformers; it best represents the shared high-level failure mode even though the cluster itself is heterogeneous.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 is the most general and cluster-representative issue, especially for the import/error-after-upgrade subset; it is the best single issue to anchor this group.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:38617", + "right": "issue:44220", + "accept": false, + "reason": "Different failures: a missing configuration utility import vs an audio feature-extraction issue." + }, + { + "left": "issue:42617", + "right": "issue:45341", + "accept": false, + "reason": "Unrelated: one is a runtime script failure, the other is a testing utility bug." + }, + { + "left": "issue:42175", + "right": "issue:42222", + "accept": false, + "reason": "Different subsystems and symptoms: backend install packaging vs broken ViTPose model files." + }, + { + "left": "issue:44933", + "right": "issue:45278", + "accept": false, + "reason": "Both mention imports, but one is a specific missing image_utils export and the other is a broad upgrade regression; not the same concrete bug." + }, + { + "left": "issue:44484", + "right": "issue:45478", + "accept": false, + "reason": "Unrelated: shard-size behavior in save_pretrained vs a Qwen3.5 MoE from_pretrained error." + }, + { + "left": "issue:43066", + "right": "issue:43122", + "accept": false, + "reason": "Both concern tokenizer behavior in v5, but one is decoder type mismatch and the other is different tokenization output; too broad to be the same bug." + }, + { + "left": "issue:43673", + "right": "issue:45341", + "accept": false, + "reason": "Different code paths: generation cache handling vs a testing utility bug." + }, + { + "left": "issue:43097", + "right": "issue:44493", + "accept": false, + "reason": "Different issues: a removed config field vs unexpected position-id keys during model loading." + }, + { + "left": "issue:42222", + "right": "issue:45459", + "accept": false, + "reason": "One is a broken model family, the other is tokenizer error handling with protobuf; not the same underlying problem." + }, + { + "left": "issue:44410", + "right": "issue:45478", + "accept": false, + "reason": "Both are Qwen-family model bugs, but one is missing projections and the other is a from_pretrained failure; not one fixable change." + }, + { + "left": "issue:43874", + "right": "issue:44261", + "accept": false, + "reason": "Different models and failures: missing image patch helper vs missing rms_norm_eps in MLA config." + }, + { + "left": "issue:43441", + "right": "issue:44829", + "accept": false, + "reason": "Both involve attention implementations, but one is model loading with FlashAttention and the other is training degeneration under flash_attention_3." + }, + { + "left": "issue:43986", + "right": "issue:44484", + "accept": false, + "reason": "Unrelated: AutoProcessor crash without torchvision vs max shard size default behavior." + }, + { + "left": "issue:43012", + "right": "issue:43408", + "accept": false, + "reason": "Both are warnings, but they concern entirely different models and causes." + }, + { + "left": "issue:44263", + "right": "issue:44485", + "accept": false, + "reason": "GLM-related but different concrete bugs: torch.split indexing vs RoPE implementation." + }, + { + "left": "issue:41628", + "right": "issue:45071", + "accept": false, + "reason": "Different regressions: missing AutoImageProcessor export vs PretrainedConfig type-checking breakage." + }, + { + "left": "issue:45071", + "right": "issue:45542", + "accept": false, + "reason": "Different failure modes: type checking regression vs backend selection error when TensorFlow is absent." + }, + { + "left": "issue:45003", + "right": "issue:45341", + "accept": false, + "reason": "No shared code path: sys.modules access bug vs testing utility issue." + }, + { + "left": "issue:43931", + "right": "issue:44410", + "accept": false, + "reason": "Both are Qwen model load issues, but one is a weight-shape mismatch and the other is missing projections; not the same bug." + }, + { + "left": "issue:43986", + "right": "issue:45341", + "accept": false, + "reason": "Unrelated processor-loading crash and testing utility bug." + }, + { + "left": "issue:43097", + "right": "issue:43452", + "accept": false, + "reason": "Different topics: removed tie_embeddings config vs gguf_file tokenizer/model loading failure." + }, + { + "left": "issue:44117", + "right": "issue:44987", + "accept": false, + "reason": "Both involve tokenizer/model loading, but the concrete failures differ and there is no clear same code-path bug." + }, + { + "left": "issue:43335", + "right": "issue:45071", + "accept": false, + "reason": "Different regressions: SwitchTransformers config construction vs config type-checking." + }, + { + "left": "issue:43408", + "right": "issue:44368", + "accept": false, + "reason": "Both are warnings, but they concern different models and unrelated config messages." + }, + { + "left": "issue:43931", + "right": "issue:44038", + "accept": false, + "reason": "Qwen3-VL load mismatch vs Qwen3-VL-Moe compatibility issue; similar family, not same defect." + }, + { + "left": "issue:43874", + "right": "issue:44263", + "accept": false, + "reason": "Different GLM bugs affecting different components; not mergeable into one fix." + }, + { + "left": "issue:43986", + "right": "issue:45372", + "accept": false, + "reason": "Different import/loading failures: missing torchvision vs missing ReasoningEffort import." + }, + { + "left": "issue:44829", + "right": "issue:45245", + "accept": false, + "reason": "FlashAttention training degradation is unrelated to categorical-feature cardinality limits." + }, + { + "left": "issue:43976", + "right": "issue:45071", + "accept": false, + "reason": "Python/version support issue vs a type-checking regression; not the same underlying bug." + }, + { + "left": "issue:43531", + "right": "issue:45341", + "accept": false, + "reason": "Different subsystems: MoE sliding_window behavior vs testing utilities." + }, + { + "left": "issue:43064", + "right": "issue:44315", + "accept": false, + "reason": "Training-state corruption under FSDP/PEFT is unrelated to Liger Kernel application with model_init." + }, + { + "left": "issue:38617", + "right": "issue:42757", + "accept": false, + "reason": "Both are ImportError reports, but for different missing symbols in different packages." + }, + { + "left": "issue:44261", + "right": "issue:44485", + "accept": false, + "reason": "Both mention GLM/MLA-related settings, but one is config eps precision and the other is RoPE implementation." + }, + { + "left": "issue:44410", + "right": "issue:45375", + "accept": false, + "reason": "Same broad Qwen ecosystem, but one is missing layer projections and the other is a config field dropped by strict parsing." + }, + { + "left": "issue:44704", + "right": "issue:45103", + "accept": false, + "reason": "Different code paths: cached_file kwargs propagation vs docstring processing with future annotations." + }, + { + "left": "issue:43986", + "right": "issue:44117", + "accept": false, + "reason": "Processor loading without torchvision is unrelated to tokenizer mapping returning None." + }, + { + "left": "issue:43582", + "right": "issue:45341", + "accept": false, + "reason": "Apple Silicon allocator TypeError is unrelated to the testing utility 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": 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-15354-364", + "nodes": [ + "issue:36683", + "issue:38617", + "issue:42757", + "issue:42886", + "issue:42907", + "issue:43012", + "issue:43064", + "issue:43066", + "issue:43097", + "issue:43208", + "issue:43299", + "issue:43352", + "issue:43381", + "issue:43502", + "issue:43525", + "issue:43526", + "issue:43550", + "issue:43582", + "issue:43618", + "issue:43761", + "issue:43824", + "issue:43825", + "issue:43867", + "issue:43873", + "issue:43901", + "issue:43906", + "issue:43937", + "issue:43976", + "issue:43986", + "issue:43992", + "issue:44038", + "issue:44062", + "issue:44117", + "issue:44230", + "issue:44351", + "issue:44368", + "issue:44380", + "issue:44410", + "issue:44483", + "issue:44484", + "issue:44514", + "issue:44655", + "issue:44683", + "issue:45290", + "issue:45381", + "issue:45479", + "issue:45538" + ], + "soft_pairs": [ + "issue:44062|issue:44117", + "issue:44038|issue:44410", + "issue:38617|issue:44351", + "issue:43502|issue:43992", + "issue:43097|issue:43873", + "issue:43992|issue:44484", + "issue:43937|issue:45538", + "issue:43824|issue:43986", + "issue:43066|issue:43906", + "issue:43825|issue:44655", + "issue:43352|issue:44380", + "issue:38617|issue:43097", + "issue:43097|issue:43901", + "issue:43012|issue:43550", + "issue:43012|issue:44368", + "issue:42886|issue:43381", + "issue:43208|issue:43550", + "issue:44380|issue:45479", + "issue:42907|issue:44655", + "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:43618|issue:43761", + "issue:36683|issue:43582", + "issue:43012|issue:44683", + "issue:44483|issue:45290", + "issue:42757|issue:43824", + "issue:44038|issue:44230", + "issue:44483|issue:45381", + "issue:43299|issue:43867", + "issue:43901|issue:44655", + "issue:38617|issue:43502" + ], + "prepared_review_unit_hash": "297c0a5e8300c1f479a15760830bc74bff526109d0f9c9a2c51bb9b7d9d55b5d", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 35, + "serialized_chars": 37404, + "estimated_input_tokens": 9351, + "estimated_eval_tokens": 18958 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly heterogeneous issue set with many v5 regressions and model-specific breakages; almost all soft pairs are non-duplicates. One CLIP output regression pair looks like the same underlying bug, but there are no pull requests in this cluster.", + "confidence": 0.67, + "canonical_issue_reason": "issue:43097 is the broadest, most central v5-breaking-change report in the set (API removal with downstream impact) and is a reasonable anchor for the otherwise heterogeneous cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43097 is the strongest representative issue: concrete regression, wider blast radius, and more likely to subsume related downstream breakages than the more model-specific reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44062", + "right": "issue:44117", + "accept": false, + "reason": "Both are tokenizer-related, but one is an AddedToken TypeError and the other is a mapping-name/import assumption bug; different failure modes and fixes." + }, + { + "left": "issue:44038", + "right": "issue:44410", + "accept": false, + "reason": "Same broad Qwen family, but one is a model-loading regression and the other is missing layer projections; not the same bug." + }, + { + "left": "issue:38617", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but for unrelated symbols from different modules; no evidence of one shared underlying breakage." + }, + { + "left": "issue:43502", + "right": "issue:43992", + "accept": false, + "reason": "One is offline/network behavior, the other is missing weights during model load; these are distinct issues." + }, + { + "left": "issue:43097", + "right": "issue:43873", + "accept": false, + "reason": "API removal versus quantization/offloading behavior; unrelated code paths and fixes." + }, + { + "left": "issue:43992", + "right": "issue:44484", + "accept": false, + "reason": "Missing checkpoint weights and shard-size default behavior are different save/load concerns." + }, + { + "left": "issue:43937", + "right": "issue:45538", + "accept": false, + "reason": "GenerationConfig validation for GLM-5 is unrelated to CLIP tokenizer max-length handling." + }, + { + "left": "issue:43824", + "right": "issue:43986", + "accept": false, + "reason": "Both mention model loading, but one is a missing top-level class export and the other is an AutoProcessor/torchvision crash." + }, + { + "left": "issue:43066", + "right": "issue:43906", + "accept": false, + "reason": "The titles suggest tokenizer behavior, but there is not enough evidence they are the same concrete bug." + }, + { + "left": "issue:43825", + "right": "issue:44655", + "accept": false, + "reason": "Error messaging in pipeline versus pipeline serialization are separate concerns." + }, + { + "left": "issue:43352", + "right": "issue:44380", + "accept": false, + "reason": "Different attention-backend problems on different models; not mergeable as one fix." + }, + { + "left": "issue:38617", + "right": "issue:43097", + "accept": false, + "reason": "Importing a removed utility and removing a config helper are different API breakages." + }, + { + "left": "issue:43097", + "right": "issue:43901", + "accept": false, + "reason": "A v5 API removal versus a docs/behavior mismatch in text classification pipeline." + }, + { + "left": "issue:43012", + "right": "issue:43550", + "accept": false, + "reason": "A precision warning during compile and a compile failure under SDPA are distinct symptoms and likely distinct fixes." + }, + { + "left": "issue:43012", + "right": "issue:44368", + "accept": false, + "reason": "Float32 warning on bf16 compile is unrelated to a tie_word_embeddings warning." + }, + { + "left": "issue:42886", + "right": "issue:43381", + "accept": false, + "reason": "Offline cache loading and gradient checkpointing in eval mode are unrelated." + }, + { + "left": "issue:43208", + "right": "issue:43550", + "accept": false, + "reason": "xLSTM training bugs and Bamba torch.compile+SDPA failure are different model-specific issues." + }, + { + "left": "issue:44380", + "right": "issue:45479", + "accept": false, + "reason": "Attention-scaling backend behavior is unrelated to sequence-classification loss degeneracy." + }, + { + "left": "issue:42907", + "right": "issue:44655", + "accept": false, + "reason": "Saving dequantized model weights is not the same as saving Pipeline objects." + }, + { + "left": "issue:43525", + "right": "issue:44117", + "accept": false, + "reason": "Missing config attribute for Llama4 and tokenizer mapping returning None are unrelated loader bugs." + }, + { + "left": "issue:43352", + "right": "issue:43976", + "accept": false, + "reason": "Flash Attention support for a model and Python-version compatibility are different problems." + }, + { + "left": "issue:43064", + "right": "issue:44380", + "accept": false, + "reason": "Distributed optimizer-state corruption is unrelated to attention-scaling behavior." + }, + { + "left": "issue:43208", + "right": "issue:43526", + "accept": false, + "reason": "Training-model-size bugs in xLSTM versus a BEiT image-processor label bug are not the same issue." + }, + { + "left": "issue:44483", + "right": "issue:44514", + "accept": false, + "reason": "OpenAI-compatible chat API rejection and Qwen2_5_VL chat-template batching are separate request-handling bugs." + }, + { + "left": "issue:43618", + "right": "issue:44380", + "accept": false, + "reason": "CLIP output-field regression and GPT2 attention scaling under SDPA are unrelated." + }, + { + "left": "issue:43618", + "right": "issue:43761", + "accept": true, + "reason": "Both describe the same CLIP v5 regression family: expected vision-model outputs are not populated correctly (attentions/hidden_states)." + }, + { + "left": "issue:36683", + "right": "issue:43582", + "accept": false, + "reason": "Gemma3 config missing vocab_size and Apple Silicon allocator TypeError are unrelated." + }, + { + "left": "issue:43012", + "right": "issue:44683", + "accept": false, + "reason": "A compile warning about bf16 precision is not the same as a flex_attention compile failure on newer torch." + }, + { + "left": "issue:44483", + "right": "issue:45290", + "accept": false, + "reason": "Chat-completions API acceptance and tool-call chat-template crashes are different layers of functionality." + }, + { + "left": "issue:42757", + "right": "issue:43824", + "accept": false, + "reason": "Different missing imports from different packages/symbols; no shared root cause shown." + }, + { + "left": "issue:44038", + "right": "issue:44230", + "accept": false, + "reason": "One is a loading bug for Qwen3-VL-Moe, the other is an fp8 inference support request; not duplicates." + }, + { + "left": "issue:44483", + "right": "issue:45381", + "accept": false, + "reason": "API request rejection and Qwen2.5-VL video position-id issues are different bugs." + }, + { + "left": "issue:43299", + "right": "issue:43867", + "accept": false, + "reason": "Qwen3-VL-Moe loading regression and sorted-state_dict load error are different failure modes." + }, + { + "left": "issue:43901", + "right": "issue:44655", + "accept": false, + "reason": "TextClassificationPipeline docs mismatch and pipeline save_pretrained failure are unrelated." + }, + { + "left": "issue:38617", + "right": "issue:43502", + "accept": false, + "reason": "Removed import from configuration_utils versus local_files_only network requests are different breakages." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays appropriately conservative overall. The lone accepted soft pair (issue:43618 / issue:43761) is plausibly the same CLIP output-regression family, while the rest of the rejects are cautious and well-justified. No PR-pair concerns since this cluster contains no PRs." + }, + "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": 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-15354-364", + "nodes": [ + "issue:28282", + "issue:33453", + "issue:36010", + "issue:36296", + "issue:36683", + "issue:41669", + "issue:42757", + "issue:42890", + "issue:42994", + "issue:43010", + "issue:43064", + "issue:43066", + "issue:43352", + "issue:43475", + "issue:43525", + "issue:43575", + "issue:43618", + "issue:43756", + "issue:43867", + "issue:43931", + "issue:43937", + "issue:43976", + "issue:43992", + "issue:44077", + "issue:44188", + "issue:44220", + "issue:44242", + "issue:44246", + "issue:44263", + "issue:44360", + "issue:44380", + "issue:44484", + "issue:44485", + "issue:44492", + "issue:44655", + "issue:44857", + "issue:44938", + "issue:44964", + "issue:44977", + "issue:45071", + "issue:45216", + "issue:45278", + "issue:45310", + "issue:45357", + "issue:45375", + "issue:45381", + "issue:45440", + "issue:45478" + ], + "soft_pairs": [ + "issue:43756|issue:44485", + "issue:42994|issue:44655", + "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:44188|issue:44380", + "issue:36010|issue:42757", + "issue:43618|issue:45216", + "issue:36683|issue:44242", + "issue:43992|issue:45310", + "issue:43010|issue:45278", + "issue:36683|issue:43352", + "issue:43352|issue:43931", + "issue:41669|issue:44246", + "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:36296|issue:43575", + "issue:43867|issue:44484", + "issue:43618|issue:44857", + "issue:42890|issue:44964", + "issue:28282|issue:33453", + "issue:43618|issue:44977" + ], + "prepared_review_unit_hash": "7366a3976dabe88e16375bf56c3932e21e5389af29e61045facfed56fbbb95da", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 37, + "serialized_chars": 38181, + "estimated_input_tokens": 9546, + "estimated_eval_tokens": 19348 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These are mostly distinct issue reports across different models and subsystems. The soft-similarity pairs are generally topic-alikes rather than true duplicates; I would reject them all. The clearest representative issue in the set is the Qwen3.5 MoE from_pretrained regression (45310), which also appears to have a later near-identical follow-up report.", + "confidence": 0.86, + "canonical_issue_reason": "issue:45310 is the clearest self-contained regression report in the set and appears to represent the Qwen3.5 MoE from_pretrained failure family better than the later follow-up report.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45310 is the best representative issue because it has a concrete, reproducible model-loading failure and is the most likely anchor for the related Qwen3.5 MoE reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43756", + "right": "issue:44485", + "accept": false, + "reason": "Both mention RoPE, but one is a model-specific layer-count bug and the other is a general GLM-5 RoPE implementation discussion; not the same defect." + }, + { + "left": "issue:42994", + "right": "issue:44655", + "accept": false, + "reason": "Both are save-related, but one is quantized-model saving failure and the other is Pipeline.save_pretrained() support; different code paths and failure modes." + }, + { + "left": "issue:36683", + "right": "issue:43475", + "accept": false, + "reason": "Both are missing-attribute errors, but on different model classes and different missing fields; not a duplicate bug." + }, + { + "left": "issue:43867", + "right": "issue:45216", + "accept": false, + "reason": "Both concern loading/saving correctness, but one is a state_dict ordering load error and the other is a Qwen3.5 save_pretrained regression; different underlying problems." + }, + { + "left": "issue:44220", + "right": "issue:44655", + "accept": false, + "reason": "Audio feature extraction helper bug vs Pipeline serialization bug; unrelated functionality." + }, + { + "left": "issue:43352", + "right": "issue:43937", + "accept": false, + "reason": "FlashAttention support error for a model vs invalid GenerationConfig for GLM-5; different failures." + }, + { + "left": "issue:43066", + "right": "issue:45381", + "accept": false, + "reason": "Tokenizer decoder type mismatch in v5 vs qwen2.5-vl video position-id bug; different components and symptoms." + }, + { + "left": "issue:44077", + "right": "issue:44380", + "accept": false, + "reason": "One is a model-class post_init policy issue, the other is GPT2 attention scaling being ignored in SDPA/FlashAttention; not the same bug." + }, + { + "left": "issue:36683", + "right": "issue:44077", + "accept": false, + "reason": "Gemma3Config missing vocab_size and patchtsmixer post_init policy are unrelated model/config issues." + }, + { + "left": "issue:44263", + "right": "issue:45440", + "accept": false, + "reason": "torch.split return values in an indexer versus DeepseekV3 MoE implementation divergence; different subsystems." + }, + { + "left": "issue:43064", + "right": "issue:44492", + "accept": false, + "reason": "FSDP/PEFT optimizer-state corruption versus a cache-strategy typo; clearly unrelated." + }, + { + "left": "issue:43976", + "right": "issue:44938", + "accept": false, + "reason": "Python version compatibility in 5.1.0 versus Python 3.14 load failure; related only at a very broad packaging level, not the same bug." + }, + { + "left": "issue:36683", + "right": "issue:43525", + "accept": false, + "reason": "Different missing-attribute regressions on different configs (vocab_size vs pad_token_id); separate issues." + }, + { + "left": "issue:44188", + "right": "issue:44380", + "accept": false, + "reason": "torch.compile branching divergence in attention kernels is not the same as GPT2 scaling being ignored in SDPA/FlashAttention." + }, + { + "left": "issue:36010", + "right": "issue:42757", + "accept": false, + "reason": "Different import errors from different dependencies: GenerationMixin import vs huggingface_hub offline helper import." + }, + { + "left": "issue:43618", + "right": "issue:45216", + "accept": false, + "reason": "CLIPOutput attentions assignment regression versus Qwen3.5 save_pretrained regression; unrelated model areas." + }, + { + "left": "issue:36683", + "right": "issue:44242", + "accept": false, + "reason": "Gemma3Config attribute bug versus MoE load-balancing loss omission; different model families and mechanisms." + }, + { + "left": "issue:43992", + "right": "issue:45310", + "accept": false, + "reason": "UMT5Encoder missing embed_tokens.weight and Qwen3.5 MoE from_pretrained error are different loading bugs." + }, + { + "left": "issue:43010", + "right": "issue:45278", + "accept": false, + "reason": "Cache update no_grad decoration proposal versus a broad import-error report after upgrade; not the same change." + }, + { + "left": "issue:36683", + "right": "issue:43352", + "accept": false, + "reason": "Missing config attribute in Gemma3 versus FlashAttention support limitation for NemotronH; different problems." + }, + { + "left": "issue:43352", + "right": "issue:43931", + "accept": false, + "reason": "FlashAttention support error for one model versus weight-shape mismatch when loading Qwen3-VL; not duplicates." + }, + { + "left": "issue:41669", + "right": "issue:44246", + "accept": false, + "reason": "Import-star performance cleanup versus import latency sometimes being slow; the latter is a symptom report, not the same code change." + }, + { + "left": "issue:43867", + "right": "issue:45375", + "accept": false, + "reason": "Both are configuration/loading issues, but one is state_dict ordering and the other is a missing field silently dropped by strict config handling." + }, + { + "left": "issue:36683", + "right": "issue:43064", + "accept": false, + "reason": "Unrelated: config missing vocab_size versus distributed training optimizer-state corruption." + }, + { + "left": "issue:43992", + "right": "issue:45478", + "accept": false, + "reason": "UMT5Encoder missing weight and Qwen3.5 MoE from_pretrained error are separate model-loading failures." + }, + { + "left": "issue:44938", + "right": "issue:45071", + "accept": false, + "reason": "Python 3.14 load failure versus type-checking breakage in v5.4.0; different regression classes." + }, + { + "left": "issue:43867", + "right": "issue:45357", + "accept": false, + "reason": "State_dict ordering load bug versus incorrect visual encoder keys in save_pretrained; related only at a high level." + }, + { + "left": "issue:43010", + "right": "issue:43976", + "accept": false, + "reason": "Cache no_grad decoration change is unrelated to Python-version compatibility." + }, + { + "left": "issue:44938", + "right": "issue:45381", + "accept": false, + "reason": "Python 3.14 import/load failure versus qwen2.5-vl video position-id bug; different failure modes." + }, + { + "left": "issue:44360", + "right": "issue:45440", + "accept": false, + "reason": "DSA indexer missing ReLU versus DeepseekV3 MoE implementation divergence; unrelated code paths." + }, + { + "left": "issue:43618", + "right": "issue:44077", + "accept": false, + "reason": "CLIP attentions regression versus patchtsmixer post_init policy; not the same bug." + }, + { + "left": "issue:36296", + "right": "issue:43575", + "accept": false, + "reason": "Tensor-parallel training bug versus OOM when loading a specific Qwen2-57B model with tp; different symptoms and likely different root causes." + }, + { + "left": "issue:43867", + "right": "issue:44484", + "accept": false, + "reason": "State_dict load ordering bug versus a question about max_shard_size default; not duplicates." + }, + { + "left": "issue:43618", + "right": "issue:44857", + "accept": false, + "reason": "CLIPOutput attentions assignment issue versus LwDetrImageLoss AMP/CUDA crash; different areas." + }, + { + "left": "issue:42890", + "right": "issue:44964", + "accept": false, + "reason": "Test flakiness from missing set_seed versus model load failure for Phi-4 multimodal; unrelated." + }, + { + "left": "issue:28282", + "right": "issue:33453", + "accept": false, + "reason": "ImportError for missing PyTorch dependency versus tokenizer loading regression; only a superficial import-related similarity." + }, + { + "left": "issue:43618", + "right": "issue:44977", + "accept": false, + "reason": "CLIP output attentions regression versus Qwen3.5 flash-attention generation bug; different models and code paths." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 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-15354-364", + "nodes": [ + "issue:30990", + "issue:31795", + "issue:33357", + "issue:33453", + "issue:34689", + "issue:36296", + "issue:38175", + "issue:38617", + "issue:42548", + "issue:42757", + "issue:42831", + "issue:42890", + "issue:43010", + "issue:43065", + "issue:43232", + "issue:43299", + "issue:43502", + "issue:43576", + "issue:43673", + "issue:43723", + "issue:43824", + "issue:43827", + "issue:43874", + "issue:44062", + "issue:44075", + "issue:44112", + "issue:44222", + "issue:44230", + "issue:44315", + "issue:44351", + "issue:44355", + "issue:44360", + "issue:44368", + "issue:44373", + "issue:44387", + "issue:44485", + "issue:44492", + "issue:44556", + "issue:44655", + "issue:44749", + "issue:44871", + "issue:44961", + "issue:45071", + "issue:45125", + "issue:45341", + "issue:45375", + "issue:45468", + "issue:45478" + ], + "soft_pairs": [ + "issue:42890|issue:44112", + "issue:43010|issue:44075", + "issue:42890|issue:45341", + "issue:43723|issue:44556", + "issue:43502|issue:44062", + "issue:44387|issue:44556", + "issue:44222|issue:45125", + "issue:43010|issue:43065", + "issue:43010|issue:43232", + "issue:45125|issue:45478", + "issue:45125|issue:45375", + "issue:42548|issue:44351", + "issue:44355|issue:44749", + "issue:30990|issue:31795", + "issue:43502|issue:44485", + "issue:44075|issue:44871", + "issue:43010|issue:44368", + "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:42831|issue:43874", + "issue:42890|issue:43232", + "issue:33453|issue:34689", + "issue:42757|issue:43576", + "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:43576|issue:44351" + ], + "prepared_review_unit_hash": "5f49b31a26a7e2462be7fe077a1ce91815482ceb94dc648b1a1880d9cd3b4729", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 39, + "serialized_chars": 38798, + "estimated_input_tokens": 9700, + "estimated_eval_tokens": 19656 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These are mostly unrelated issue reports; I did not find any soft pair that clearly describes the same underlying bug/change. The strongest single issue is the concrete Qwen3.5 MoE v5.4.0 loading regression, but there is no PR in the set.", + "confidence": 0.81, + "canonical_issue_reason": "issue:45478 is the most concrete and actionable report: it names the affected version, model family, and failure mode (`from_pretrained` error), making it a good anchor issue.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45478 has a specific regression description and a narrow code path, so it is the best standalone issue candidate here.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42890", + "right": "issue:44112", + "accept": false, + "reason": "Both are test failures, but they concern different models and different failure modes; not the same bug." + }, + { + "left": "issue:43010", + "right": "issue:44075", + "accept": false, + "reason": "Cache update decoration and SGD argument handling are unrelated code paths." + }, + { + "left": "issue:42890", + "right": "issue:45341", + "accept": false, + "reason": "One is a specific missing set_seed in a model integration test; the other is a generic testing_utils bug with no clear shared defect." + }, + { + "left": "issue:43723", + "right": "issue:44556", + "accept": false, + "reason": "Both involve loading regressions in v5, but one is tokenizer loading and the other is checkpoint reload compatibility; too broad and not the same concrete bug." + }, + { + "left": "issue:43502", + "right": "issue:44062", + "accept": false, + "reason": "Local-files-only network requests and AddedToken keyword duplication are unrelated." + }, + { + "left": "issue:44387", + "right": "issue:44556", + "accept": false, + "reason": "Quantization OOM and checkpoint reload failure are different problems." + }, + { + "left": "issue:44222", + "right": "issue:45125", + "accept": false, + "reason": "Both mention Qwen3.5 MoE/FP8-related work, but one is save_pretrained FP8 and the other is tensor-parallel metadata; not the same issue." + }, + { + "left": "issue:43010", + "right": "issue:43065", + "accept": false, + "reason": "Cache update no_grad and Sam3PixelDecoder dummy Conv2d are unrelated." + }, + { + "left": "issue:43010", + "right": "issue:43232", + "accept": false, + "reason": "Cache update no_grad and generation kwargs sync behavior are different code paths." + }, + { + "left": "issue:45125", + "right": "issue:45478", + "accept": false, + "reason": "Both concern Qwen3.5 MoE loading, but one is missing tensor-parallel plan and the other is a broader from_pretrained regression; not clearly the same concrete bug." + }, + { + "left": "issue:45125", + "right": "issue:45375", + "accept": false, + "reason": "Tensor parallel metadata and a missing vision config field are unrelated." + }, + { + "left": "issue:42548", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but for different symbols and likely different export breaks." + }, + { + "left": "issue:44355", + "right": "issue:44749", + "accept": false, + "reason": "Running compiled Python files and slower filtering after upgrade are unrelated." + }, + { + "left": "issue:30990", + "right": "issue:31795", + "accept": false, + "reason": "Sentence Transformers loading hangs versus forward doc confusion are unrelated." + }, + { + "left": "issue:43502", + "right": "issue:44485", + "accept": false, + "reason": "Offline/local-file request behavior and GLM-5 RoPE implementation are unrelated." + }, + { + "left": "issue:44075", + "right": "issue:44871", + "accept": false, + "reason": "SGD optimizer arguments and eos_token_id config mismatch are different bugs." + }, + { + "left": "issue:43010", + "right": "issue:44368", + "accept": false, + "reason": "Cache update no_grad and a tie_word_embeddings warning are unrelated." + }, + { + "left": "issue:44315", + "right": "issue:44360", + "accept": false, + "reason": "Liger kernel application and DSA indexer activation absence are unrelated." + }, + { + "left": "issue:42890", + "right": "issue:44373", + "accept": false, + "reason": "A flaky integration test and a docstring typo are not the same issue." + }, + { + "left": "issue:38175", + "right": "issue:43502", + "accept": false, + "reason": "Zero probabilities in a model and unwanted API requests are unrelated." + }, + { + "left": "issue:43299", + "right": "issue:44556", + "accept": false, + "reason": "Both are version-related loading regressions, but they affect different models and failure modes." + }, + { + "left": "issue:43502", + "right": "issue:44961", + "accept": false, + "reason": "Local-files-only networking and a one-word title are unrelated." + }, + { + "left": "issue:44355", + "right": "issue:44655", + "accept": false, + "reason": "Compiled Python file errors and pipeline save_pretrained failures are different code paths." + }, + { + "left": "issue:44222", + "right": "issue:44230", + "accept": false, + "reason": "Both are FP8/Qwen3.5 MoE themed, but one is save_pretrained and the other is inference support; not the same bug." + }, + { + "left": "issue:44355", + "right": "issue:45071", + "accept": false, + "reason": "Compiled file execution errors and PretrainedConfig typing issues are unrelated." + }, + { + "left": "issue:42831", + "right": "issue:43874", + "accept": false, + "reason": "FineGrainedFP8 accuracy and GLM46V image processor token counting are unrelated." + }, + { + "left": "issue:42890", + "right": "issue:43232", + "accept": false, + "reason": "A failing seed-dependent test and generation cache kwargs handling are unrelated." + }, + { + "left": "issue:33453", + "right": "issue:34689", + "accept": false, + "reason": "Tokenizer loading regression and Llama 3.2 Vision model loading regression are different problems." + }, + { + "left": "issue:42757", + "right": "issue:43576", + "accept": false, + "reason": "Missing `is_offline_mode` import and broken `transformers env` CLI are not the same bug." + }, + { + "left": "issue:33453", + "right": "issue:36296", + "accept": false, + "reason": "Tokenizer loading regression and tensor parallel training bug are unrelated." + }, + { + "left": "issue:43827", + "right": "issue:44961", + "accept": false, + "reason": "Docs still referencing pipeline() and a one-word issue title are not duplicates." + }, + { + "left": "issue:43576", + "right": "issue:43824", + "accept": false, + "reason": "Broken env command and missing Qwen2_5_VL export are both v5 regressions but not the same concrete change." + }, + { + "left": "issue:43576", + "right": "issue:45071", + "accept": false, + "reason": "CLI breakage and PretrainedConfig type checking are unrelated." + }, + { + "left": "issue:43673", + "right": "issue:44492", + "accept": false, + "reason": "Missing cache in generation and a typo in cache strategies are unrelated." + }, + { + "left": "issue:38617", + "right": "issue:43576", + "accept": false, + "reason": "Missing import in configuration_utils and broken env command are different failures." + }, + { + "left": "issue:33357", + "right": "issue:33453", + "accept": false, + "reason": "MacOS bus error in CLIP and tokenizer loading regression are unrelated." + }, + { + "left": "issue:44961", + "right": "issue:45468", + "accept": false, + "reason": "No substantive signal in the one-word issue; not the same as Gemma-4 audio positional encoding." + }, + { + "left": "issue:31795", + "right": "issue:38617", + "accept": false, + "reason": "Doc confusion about forward args and an import error are unrelated." + }, + { + "left": "issue:43576", + "right": "issue:44351", + "accept": false, + "reason": "Broken env command and missing HybridCache import are different v5 export/CLI issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 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-15354-364", + "nodes": [ + "issue:31795", + "issue:33453", + "issue:36296", + "issue:41628", + "issue:41669", + "issue:42371", + "issue:42548", + "issue:42890", + "issue:43012", + "issue:43352", + "issue:43408", + "issue:43526", + "issue:43576", + "issue:43673", + "issue:43723", + "issue:43827", + "issue:43901", + "issue:43976", + "issue:44038", + "issue:44077", + "issue:44246", + "issue:44261", + "issue:44273", + "issue:44295", + "issue:44297", + "issue:44315", + "issue:44351", + "issue:44355", + "issue:44360", + "issue:44393", + "issue:44485", + "issue:44492", + "issue:44556", + "issue:44623", + "issue:44704", + "issue:44829", + "issue:44861", + "issue:44908", + "issue:44945", + "issue:44961", + "issue:45003", + "issue:45071", + "issue:45092", + "issue:45125", + "issue:45230", + "issue:45341", + "issue:45478", + "issue:45542" + ], + "soft_pairs": [ + "issue:44908|issue:44961", + "issue:41628|issue:42548", + "issue:36296|issue:43901", + "issue:44485|issue:44961", + "issue:36296|issue:43526", + "issue:36296|issue:43012", + "issue:44273|issue:45003", + "issue:41628|issue:44355", + "issue:44261|issue:44360", + "issue:44273|issue:45542", + "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:43827|issue:44908", + "issue:31795|issue:33453", + "issue:41669|issue:44908", + "issue:43723|issue:44393" + ], + "prepared_review_unit_hash": "584dfc293d0aaafafb6643e650f8a72c036679f28ad65d9fee03c8918ff4a21d", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 38, + "serialized_chars": 38388, + "estimated_input_tokens": 9597, + "estimated_eval_tokens": 19450 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: a few issues are in the import/lazy-loading area, but most pairs are unrelated regressions, model bugs, docs problems, or version-compatibility reports. No PRs are present.", + "confidence": 0.71, + "canonical_issue_reason": "issue:41669 is the most concrete and representative item in the cluster, with a clear root cause around import-time slowdown and module loading behavior; it best matches the small import/lazy-loading theme that appears across a few nearby issues.", + "canonical_pr_reason": null, + "best_issue_reason": "Best triage anchor because it is specific, actionable, and broad enough to represent the limited shared import-related surface in this set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44908", + "right": "issue:44961", + "accept": false, + "reason": "Different domains: scheduler kwargs bug vs. an unrelated placeholder issue." + }, + { + "left": "issue:41628", + "right": "issue:42548", + "accept": false, + "reason": "Both are import errors, but they concern different missing symbols and no concrete shared code-path is shown." + }, + { + "left": "issue:36296", + "right": "issue:43901", + "accept": false, + "reason": "Tensor-parallel training bug vs. pipeline docs mismatch; unrelated." + }, + { + "left": "issue:44485", + "right": "issue:44961", + "accept": false, + "reason": "RoPE implementation discussion vs. unrelated placeholder issue." + }, + { + "left": "issue:36296", + "right": "issue:43526", + "accept": false, + "reason": "Tensor parallel training bug and a BeitImageProcessorFast label bug are unrelated." + }, + { + "left": "issue:36296", + "right": "issue:43012", + "accept": false, + "reason": "Different subsystems and symptoms: tensor-parallel training vs. bfloat16 compile warning." + }, + { + "left": "issue:44273", + "right": "issue:45003", + "accept": false, + "reason": "Both touch loading behavior, but one is a general lazy-loading report and the other a specific sys.modules access issue; not enough to treat as the same bug." + }, + { + "left": "issue:41628", + "right": "issue:44355", + "accept": false, + "reason": "Top-level import failure vs. compiled-file runtime error; no shared bug evidence." + }, + { + "left": "issue:44261", + "right": "issue:44360", + "accept": false, + "reason": "MLA precision/config issue vs. DSA indexer/ReLU discussion; unrelated." + }, + { + "left": "issue:44273", + "right": "issue:45542", + "accept": false, + "reason": "Lazy loading issue vs. TensorFlow backend detection issue; different code paths." + }, + { + "left": "issue:44623", + "right": "issue:45230", + "accept": false, + "reason": "Processor save_pretrained missing files vs. a generic bug report; no match." + }, + { + "left": "issue:43576", + "right": "issue:44556", + "accept": false, + "reason": "transformers env command breakage vs. checkpoint reload incompatibility; different failures." + }, + { + "left": "issue:41669", + "right": "issue:43673", + "accept": false, + "reason": "Import-slowdown / star-import issue vs. GenerationMixin cache missing during chunked prefill; not the same bug." + }, + { + "left": "issue:44393", + "right": "issue:45071", + "accept": false, + "reason": "Qwen3-VL bounding-box output issue vs. PretrainedConfig type-checking regression; unrelated." + }, + { + "left": "issue:44038", + "right": "issue:45125", + "accept": false, + "reason": "Both mention Qwen models, but one is a general bug report and the other is a tensor-parallel metadata omission; not the same issue." + }, + { + "left": "issue:44556", + "right": "issue:45092", + "accept": false, + "reason": "Upgrade/reload incompatibility vs. meta-init remote-code incompatibility; related timeframe, different concrete problems." + }, + { + "left": "issue:42371", + "right": "issue:44492", + "accept": false, + "reason": "TF32 settings warning vs. a cache-strategy typo; unrelated." + }, + { + "left": "issue:42890", + "right": "issue:44908", + "accept": false, + "reason": "Test seeding flakiness vs. scheduler kwargs bug; unrelated." + }, + { + "left": "issue:44704", + "right": "issue:44908", + "accept": false, + "reason": "AutoProcessor kwargs forwarding vs. scheduler kwarg handling; different code paths." + }, + { + "left": "issue:44246", + "right": "issue:44273", + "accept": false, + "reason": "Import slowness and lazy loading are adjacent, but these are still distinct reported failures without concrete duplicate evidence." + }, + { + "left": "issue:43408", + "right": "issue:44315", + "accept": false, + "reason": "sam3_tracker type warning vs. Liger Kernel model_init behavior; unrelated." + }, + { + "left": "issue:44297", + "right": "issue:45230", + "accept": false, + "reason": "Tokenizer_class metadata mismatch vs. generic bug report; no duplicate signal." + }, + { + "left": "issue:43576", + "right": "issue:44908", + "accept": false, + "reason": "Broken env command vs. inverse_sqrt scheduler kwargs bug; unrelated." + }, + { + "left": "issue:43352", + "right": "issue:44393", + "accept": false, + "reason": "Flash Attention support error vs. Qwen3-VL bounding-box hallucination; unrelated." + }, + { + "left": "issue:44908", + "right": "issue:45071", + "accept": false, + "reason": "Scheduler bug vs. config type-checking regression; unrelated." + }, + { + "left": "issue:43976", + "right": "issue:44393", + "accept": false, + "reason": "Python-version compatibility vs. vision-language output issue; unrelated." + }, + { + "left": "issue:44295", + "right": "issue:45230", + "accept": false, + "reason": "position_ids buffer access bug vs. generic bug report; no match." + }, + { + "left": "issue:41669", + "right": "issue:44351", + "accept": false, + "reason": "Import * slowdown and missing HybridCache import are both import-related, but they are different missing-export problems with no proof of a single root cause." + }, + { + "left": "issue:41669", + "right": "issue:45341", + "accept": false, + "reason": "Import slowdown issue vs. a testing_utils bug report; unrelated." + }, + { + "left": "issue:44492", + "right": "issue:44945", + "accept": false, + "reason": "Typo in cache strategies vs. incorrect pipeline-parallel LLM output; unrelated." + }, + { + "left": "issue:41669", + "right": "issue:44492", + "accept": false, + "reason": "Import-time performance regression vs. a typo; not the same underlying bug." + }, + { + "left": "issue:44315", + "right": "issue:44829", + "accept": false, + "reason": "Liger Kernel model_init behavior vs. flash_attention_3 training degeneration; different failures." + }, + { + "left": "issue:44393", + "right": "issue:45478", + "accept": false, + "reason": "Different Qwen3-VL/Qwen3.5-Moe bugs with different symptoms and code paths." + }, + { + "left": "issue:44077", + "right": "issue:44861", + "accept": false, + "reason": "Optional post_init handling vs. tied-weights key crash; unrelated." + }, + { + "left": "issue:43827", + "right": "issue:44908", + "accept": false, + "reason": "Docs still reference pipeline() vs. scheduler kwargs bug; unrelated." + }, + { + "left": "issue:31795", + "right": "issue:33453", + "accept": false, + "reason": "Documentation ambiguity vs. tokenizer-loading regression; unrelated." + }, + { + "left": "issue:41669", + "right": "issue:44908", + "accept": false, + "reason": "Import-time slowdown vs. scheduler kwargs bug; no shared code-path." + }, + { + "left": "issue:43723", + "right": "issue:44393", + "accept": false, + "reason": "Tokenizer loading in v5 vs. Qwen3-VL bounding-box output issue; 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": 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-15354-364", + "nodes": [ + "issue:31515", + "issue:31795", + "issue:33453", + "issue:34689", + "issue:36296", + "issue:36683", + "issue:37428", + "issue:38617", + "issue:41669", + "issue:42371", + "issue:42890", + "issue:42913", + "issue:43066", + "issue:43298", + "issue:43502", + "issue:43519", + "issue:43525", + "issue:43572", + "issue:43576", + "issue:43606", + "issue:43618", + "issue:43704", + "issue:44230", + "issue:44261", + "issue:44263", + "issue:44273", + "issue:44315", + "issue:44360", + "issue:44485", + "issue:44492", + "issue:44556", + "issue:44623", + "issue:44655", + "issue:44861", + "issue:44908", + "issue:44998", + "issue:45092", + "issue:45230", + "issue:45290", + "issue:45310", + "issue:45341", + "issue:45468" + ], + "soft_pairs": [ + "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: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", + "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": "6b21982e9ae1be24051b1e48ff3456fe30fbad620907e97fe52d8baf20c3806c", + "component_budget": { + "node_count": 364, + "item_count": 364, + "soft_pair_count": 1360, + "serialized_chars": 701642, + "estimated_input_tokens": 175411, + "estimated_eval_tokens": 351078 + }, + "budget": { + "node_count": 42, + "item_count": 42, + "soft_pair_count": 38, + "serialized_chars": 34909, + "estimated_input_tokens": 8728, + "estimated_eval_tokens": 17712 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous: mostly unrelated issue reports around loading, tokenizer/config regressions, model-specific bugs, and docs/test noise. No soft-edge pair looks like the same concrete underlying bug or change, and there are no PRs to consolidate.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 44556 is the broadest loading/v5 compatibility report and is the closest thing to an umbrella issue here; however, the set is not actually a true duplicate cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "44556 best represents the recurring checkpoint/model-loading regression theme among otherwise unrelated items.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44655", + "right": "issue:44861", + "accept": false, + "reason": "Both mention save/pretrained-adjacent code, but one is Pipeline serialization and the other is tied-weight key handling; different bugs." + }, + { + "left": "issue:43576", + "right": "issue:44273", + "accept": false, + "reason": "Different areas: CLI/env command breakage vs lazy loading behavior; no shared code-path issue." + }, + { + "left": "issue:44360", + "right": "issue:45468", + "accept": false, + "reason": "Unrelated model internals: DSA indexer/ReLU vs Gemma-4 audio positional encoding." + }, + { + "left": "issue:45230", + "right": "issue:45310", + "accept": false, + "reason": "Generic bug report vs a specific Qwen3.5 MoE from_pretrained regression; not the same issue." + }, + { + "left": "issue:44273", + "right": "issue:44908", + "accept": false, + "reason": "Lazy loading and inverse_sqrt scheduler kwargs are unrelated subsystems." + }, + { + "left": "issue:41669", + "right": "issue:43704", + "accept": false, + "reason": "Import-star cleanup/perf issue vs VRAM leak in dataloader threads; different change and failure mode." + }, + { + "left": "issue:43519", + "right": "issue:44485", + "accept": false, + "reason": "Qwen3VL timestamp calculation vs GLM-5 RoPE implementation; unrelated model logic." + }, + { + "left": "issue:43519", + "right": "issue:44908", + "accept": false, + "reason": "Processor timestamp math and scheduler kwargs are unrelated bugs." + }, + { + "left": "issue:31795", + "right": "issue:34689", + "accept": false, + "reason": "Documentation confusion for forward args vs a concrete model-loading regression." + }, + { + "left": "issue:31795", + "right": "issue:37428", + "accept": false, + "reason": "Docs issue vs ImportError for flash attention utility; not the same bug." + }, + { + "left": "issue:44315", + "right": "issue:45092", + "accept": false, + "reason": "Liger kernel application in model_init is not the same as remote-code checkpoint/meta-init incompatibility." + }, + { + "left": "issue:43704", + "right": "issue:44485", + "accept": false, + "reason": "VRAM leak in multi-threaded inference vs RoPE implementation bug." + }, + { + "left": "issue:44485", + "right": "issue:45468", + "accept": false, + "reason": "Different model families and different math/encoding paths." + }, + { + "left": "issue:42371", + "right": "issue:43704", + "accept": false, + "reason": "TF32 API guidance issue vs memory leak bug; no overlap in underlying change." + }, + { + "left": "issue:43502", + "right": "issue:43519", + "accept": false, + "reason": "local_files_only network leakage vs timestamp calculation in a processor." + }, + { + "left": "issue:43519", + "right": "issue:44623", + "accept": false, + "reason": "Processor timestamp math vs processor.save_pretrained missing files; different code paths." + }, + { + "left": "issue:43519", + "right": "issue:43704", + "accept": false, + "reason": "Processor math bug vs VRAM leak; unrelated." + }, + { + "left": "issue:44263", + "right": "issue:45468", + "accept": false, + "reason": "torch.split return values in one indexer vs Gemma-4 audio positional encoding." + }, + { + "left": "issue:44261", + "right": "issue:45468", + "accept": false, + "reason": "Missing rms_norm_eps in MLA layernorm vs Gemma-4 audio positional encoding; different models." + }, + { + "left": "issue:43606", + "right": "issue:45468", + "accept": false, + "reason": "CPU offload device mismatch in bark-small vs Gemma-4 audio positional encoding." + }, + { + "left": "issue:42890", + "right": "issue:43704", + "accept": false, + "reason": "Flaky integration test seeding issue vs a runtime VRAM leak." + }, + { + "left": "issue:42913", + "right": "issue:43066", + "accept": false, + "reason": "Both are tokenizer v5-related, but one is a broad behavior change and the other is a specific decoder-type bug; not clearly the same underlying defect." + }, + { + "left": "issue:44908", + "right": "issue:44998", + "accept": false, + "reason": "Scheduler kwargs bug vs an unrelated placeholder issue." + }, + { + "left": "issue:44861", + "right": "issue:45290", + "accept": false, + "reason": "Tied-weight key crash and chat template tool-call crash are separate failures." + }, + { + "left": "issue:44998", + "right": "issue:45341", + "accept": false, + "reason": "Placeholder issue vs a testing_utils bug; unrelated." + }, + { + "left": "issue:44492", + "right": "issue:44998", + "accept": false, + "reason": "Typo in cache strategies vs unrelated placeholder issue." + }, + { + "left": "issue:43618", + "right": "issue:44998", + "accept": false, + "reason": "CLIPOutput attention assignment regression vs unrelated placeholder issue." + }, + { + "left": "issue:44556", + "right": "issue:44998", + "accept": false, + "reason": "Checkpoint reload regression vs unrelated placeholder issue." + }, + { + "left": "issue:43525", + "right": "issue:44861", + "accept": false, + "reason": "Llama4Config pad_token_id missing vs tied-weight key handling; different config fields and code paths." + }, + { + "left": "issue:43572", + "right": "issue:44861", + "accept": false, + "reason": "StableLmConfig pad_token_idx regression vs tied-weight key AttributeError; unrelated." + }, + { + "left": "issue:31515", + "right": "issue:38617", + "accept": false, + "reason": "Checkpoint loading slow vs ImportError for layer_type_validation; different failures." + }, + { + "left": "issue:44230", + "right": "issue:44623", + "accept": false, + "reason": "FP8 inference support request vs missing files on processor.save_pretrained; not the same bug." + }, + { + "left": "issue:31515", + "right": "issue:36296", + "accept": false, + "reason": "from_pretrained slowdown vs tensor-parallel training bug." + }, + { + "left": "issue:31515", + "right": "issue:31795", + "accept": false, + "reason": "Loading performance issue vs forward doc confusion." + }, + { + "left": "issue:44230", + "right": "issue:45310", + "accept": false, + "reason": "Both mention Qwen3.5 MoE, but one is fp8 inference support and the other is a from_pretrained regression; different changes." + }, + { + "left": "issue:31515", + "right": "issue:33453", + "accept": false, + "reason": "Both are loading-related, but one is checkpoint load slowness and the other is tokenizer loading regression." + }, + { + "left": "issue:31515", + "right": "issue:36683", + "accept": false, + "reason": "Checkpoint loading slowdown vs Gemma3Config missing vocab_size; unrelated." + }, + { + "left": "issue:43298", + "right": "issue:44230", + "accept": false, + "reason": "Continuous batching for audio models vs fp8 support for Qwen models; different feature gaps." + } + ] + }, + "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": "b8668984fa84439432f32317baa962fab8616fa98bd9c2bcecce906078d2ca4b", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1590, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1590, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are unrelated feature-request issues: one asks for Flash Attention 2 support on more architectures, the other asks for GGUF support on more architectures. They target different subsystems and different deliverables, so they should not be clustered as duplicates.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:26350", + "right": "issue:33260", + "accept": false, + "reason": "Different features and code paths: Flash Attention 2 support vs GGUF support. They are both broad architecture-support requests but not the same underlying bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 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": "0aa0d148f039a45db1e8a688654b0d5beaa690844ef0fe5ec6683fb5b835b75a", + "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 look like the same issue template family, but only #27088 is a concrete Turkish translation request; the other three are generic placeholder variants and don\u2019t clearly represent the same actionable issue.", + "confidence": 0.84, + "canonical_issue_reason": "#27088 is the only concrete, language-specific issue title ([i18n-TR] Turkish). The others are generic placeholders, so this is the best representative issue.", + "canonical_pr_reason": null, + "best_issue_reason": "#27088 is the most actionable and specific issue in the cluster; it is clearly about Turkish docs translation rather than a templated placeholder.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:27088", + "right": "issue:44370", + "accept": false, + "reason": "Same broad translation-template wording, but #44370 is generic and not clearly the same concrete issue as the Turkish-specific request." + }, + { + "left": "issue:27088", + "right": "issue:44365", + "accept": false, + "reason": "High textual similarity, but the second item is a placeholder template rather than a specific Turkish docs translation issue." + }, + { + "left": "issue:27088", + "right": "issue:43834", + "accept": false, + "reason": "The titles are template-related, but #43834 is generic and does not clearly duplicate the concrete Turkish translation request." + }, + { + "left": "issue:43834", + "right": "issue:44370", + "accept": false, + "reason": "Both are generic template placeholders; similarity alone is not enough to confirm the same underlying issue." + }, + { + "left": "issue:43834", + "right": "issue:44365", + "accept": false, + "reason": "Generic placeholder titles with no concrete language or unique issue details; not safe to merge as duplicates." + }, + { + "left": "issue:44365", + "right": "issue:44370", + "accept": false, + "reason": "Both look like the same template class, but there is no evidence they are the same specific issue." + } + ] + }, + "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": "9f53b89a9d1a3218b91f99047b8d0383600d32aeea36ef74aab699563d2f60a6", + "component_budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 25, + "serialized_chars": 16011, + "estimated_input_tokens": 4003, + "estimated_eval_tokens": 8262 + }, + "budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 25, + "serialized_chars": 16011, + "estimated_input_tokens": 4003, + "estimated_eval_tokens": 8262 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is not a duplicate set: every item is a separate model-request issue for different models, tasks, or modalities (BERT variants, TTS/ASR, audio, OCR, vision, sparse transformer). No PRs are present.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue fits the cluster because the requests are for different underlying models and capabilities, not the same bug or change.", + "canonical_pr_reason": null, + "best_issue_reason": "If forced to pick one representative, issue 37015 is the most generic and active 'Add X' request, but it still only covers NeoBERT and does not generalize to the others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36817", + "right": "issue:41211", + "accept": false, + "reason": "Different model requests: EuroBERT vs DEIMv2; same broad category, not the same change." + }, + { + "left": "issue:36817", + "right": "issue:43671", + "accept": false, + "reason": "EuroBERT config support vs Qwen3-TTS support are unrelated model additions." + }, + { + "left": "issue:37015", + "right": "issue:43837", + "accept": false, + "reason": "NeoBERT and Qwen3-ASR are different models and different code paths." + }, + { + "left": "issue:37015", + "right": "issue:44121", + "accept": false, + "reason": "NeoBERT vs a weight-sparse transformer request; not the same underlying feature." + }, + { + "left": "issue:37015", + "right": "issue:43909", + "accept": false, + "reason": "NeoBERT and LFM2.5 Audio are distinct model families/modalities." + }, + { + "left": "issue:37015", + "right": "issue:42503", + "accept": false, + "reason": "NeoBERT vs ModernVBERT are separate model requests, not duplicates." + }, + { + "left": "issue:37015", + "right": "issue:45027", + "accept": false, + "reason": "NeoBERT and Voxtral TTS support target different models and integration work." + }, + { + "left": "issue:41211", + "right": "issue:43671", + "accept": false, + "reason": "DEIMv2 and Qwen3-TTS are unrelated additions." + }, + { + "left": "issue:42503", + "right": "issue:43837", + "accept": false, + "reason": "ModernVBERT models vs Qwen3-ASR; different model families and tasks." + }, + { + "left": "issue:42503", + "right": "issue:43671", + "accept": false, + "reason": "ModernVBERT and Qwen3-TTS are not the same model/change." + }, + { + "left": "issue:42503", + "right": "issue:42738", + "accept": false, + "reason": "ModernVBERT support and BERT-like RoPE models are different requests." + }, + { + "left": "issue:42503", + "right": "issue:45027", + "accept": false, + "reason": "ModernVBERT models vs Voxtral-4B-TTS are unrelated." + }, + { + "left": "issue:42738", + "right": "issue:45027", + "accept": false, + "reason": "A generic BERT-like RoPE request is not the same as Voxtral TTS support." + }, + { + "left": "issue:43671", + "right": "issue:43837", + "accept": false, + "reason": "Qwen3-TTS and Qwen3-ASR share a family name but request different capabilities and code paths." + }, + { + "left": "issue:43671", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-TTS support and Voxtral TTS support are different model integrations." + }, + { + "left": "issue:43671", + "right": "issue:45177", + "accept": false, + "reason": "Qwen3-TTS and DeepSeek-OCR2 are unrelated modalities and changes." + }, + { + "left": "issue:43671", + "right": "issue:44121", + "accept": false, + "reason": "Qwen3-TTS support does not overlap with a sparse-transformer architecture request." + }, + { + "left": "issue:43671", + "right": "issue:43909", + "accept": false, + "reason": "Qwen3-TTS and LFM2.5 Audio are separate audio-model requests." + }, + { + "left": "issue:43837", + "right": "issue:43909", + "accept": false, + "reason": "Qwen3-ASR support and LFM2.5 Audio are distinct model additions." + }, + { + "left": "issue:43837", + "right": "issue:45322", + "accept": false, + "reason": "Qwen3-ASR and EUPE are different tasks and model families." + }, + { + "left": "issue:43837", + "right": "issue:44121", + "accept": false, + "reason": "Qwen3-ASR support and weight-sparse transformer support are not the same change." + }, + { + "left": "issue:43909", + "right": "issue:44205", + "accept": false, + "reason": "LFM2.5 Audio 1.5B and SAM3-LiteText are unrelated model requests." + }, + { + "left": "issue:44121", + "right": "issue:45322", + "accept": false, + "reason": "OpenAI weight-sparse transformer vs EUPE are different architectures and requests." + }, + { + "left": "issue:43837", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-ASR and Voxtral-4B-TTS are different models and integration paths." + }, + { + "left": "issue:43837", + "right": "issue:45177", + "accept": false, + "reason": "Qwen3-ASR and DeepSeek-OCR2 are unrelated tasks." + } + ] + }, + "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": "166e51d89266207bc6ed9a94480f8d5239393cd6a85cb673806ac2bddf3dc779", + "component_budget": { + "node_count": 6, + "item_count": 6, + "soft_pair_count": 8, + "serialized_chars": 6292, + "estimated_input_tokens": 1573, + "estimated_eval_tokens": 3402 + }, + "budget": { + "node_count": 6, + "item_count": 6, + "soft_pair_count": 8, + "serialized_chars": 6292, + "estimated_input_tokens": 1573, + "estimated_eval_tokens": 3402 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These issues are only superficially similar by theme (norms, kwargs, typos, warnings) but do not describe the same underlying bug or change. I would not deduplicate any of the soft pairs.", + "confidence": 0.89, + "canonical_issue_reason": "If one representative is needed, issue:43240 is the most concrete and broadly actionable runtime bug report; however, the set does not contain a true duplicate cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43240 is the strongest global issue candidate because it reports an actual behavior bug with broader applicability than the more narrow typo/type-hint/model-design reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:37934", + "right": "issue:43668", + "accept": false, + "reason": "Both mention normalization, but one questions model design (RMS norm vs Llama4TextL2Norm) while the other is a config type-hint typo; not the same bug." + }, + { + "left": "issue:42936", + "right": "issue:43240", + "accept": false, + "reason": "Both involve loss-related arguments, but 42936 is a specific Mask2Former ignore_value bug and 43240 is a general kwargs passthrough issue; close, but not clearly the same underlying defect." + }, + { + "left": "issue:42936", + "right": "issue:44031", + "accept": false, + "reason": "Mask2Former loss handling and tokenizer regex warnings are unrelated code paths and symptoms." + }, + { + "left": "issue:42936", + "right": "issue:43813", + "accept": false, + "reason": "Loss ignore_value handling and a quantization-operation typo are unrelated." + }, + { + "left": "issue:42936", + "right": "issue:43668", + "accept": false, + "reason": "Model loss behavior vs config type-hint correctness; no shared bug." + }, + { + "left": "issue:43240", + "right": "issue:43668", + "accept": false, + "reason": "General loss kwargs forwarding and a ModernBERTConfig type-hint typo are different issues." + }, + { + "left": "issue:43668", + "right": "issue:44031", + "accept": false, + "reason": "Config type-hint bug and tokenizer warning regression are unrelated." + }, + { + "left": "issue:43668", + "right": "issue:43813", + "accept": false, + "reason": "A type-hint error and a naming typo in quantization code 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": 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": "b8b1e5db0ffc5dba1f4dd070dd307c1ec1609f27bee2acef0859f5f09c91834d", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1480, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1480, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues appear unrelated: one is about Set Block Decoding, while the other requests DeepSeek v3.2 support. No shared targets, filenames, or concrete overlap suggests duplication.", + "confidence": 0.96, + "canonical_issue_reason": "Neither issue is a good canonical choice for the other because they describe different feature areas and different underlying changes.", + "canonical_pr_reason": null, + "best_issue_reason": "Neither issue is suitable as a global canonical issue; the topics do not align closely enough to represent the same bug or request.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41084", + "right": "issue:43037", + "accept": false, + "reason": "These are different requests with no concrete overlap: one concerns block decoding behavior, the other requests model support for DeepSeek v3.2." + } + ] + }, + "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": "696262ffa69273f8c062a80b6a7230eb522a782c32281f346fa0061246b45d08", + "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 two issues are thematically adjacent but not duplicates: one is about making `Trainer.remove_unused_column` more customizable, while the other is about FLOPS estimation handling custom trainers.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43598", + "right": "issue:43600", + "accept": false, + "reason": "Different underlying problems and code paths: custom column dropping in training data handling vs FLOPS counting logic for custom trainers. Shared mention of custom trainers 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": "12827ebdeff7fb3990b5867466140042aadb1132c77a175886c44487e3a6c3a3", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1612, + "estimated_input_tokens": 403, + "estimated_eval_tokens": 1062 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1612, + "estimated_input_tokens": 403, + "estimated_eval_tokens": 1062 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related model-registration requests for Qwen variants, but they target different AutoModel classes and different concrete capabilities, so they should not be treated as duplicates.", + "confidence": 0.88, + "canonical_issue_reason": "Issue 43684 is the broader request: add Qwen3-Omni registration for AutoModel and AutoModelForConditionalGeneration, which is more likely to subsume a library registration gap.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43684 is the best representative because it is the more general registration task and aligns with the broader API surface; 44405 is a narrower sequence-classification-specific support request for a different config.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43684", + "right": "issue:44405", + "accept": false, + "reason": "Both are about Qwen model support, but one is for Qwen3-Omni registration in AutoModel/AutoModelForConditionalGeneration and the other is for AutoModelForSequenceClassification support in Qwen3.5. They affect different code paths and 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": "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": "b3d98e982a19f3c41e59fdb39650bdc8ea6c396a125ee9085dbbbc22f7185b22", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4263, + "estimated_input_tokens": 1066, + "estimated_eval_tokens": 2388 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4263, + "estimated_input_tokens": 1066, + "estimated_eval_tokens": 2388 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Three issues share the exact same generic title \"huggingface \" and are very likely duplicates; the separate \"Fett\" issue looks unrelated. I\u2019d keep the earliest huggingface report as the canonical issue.", + "confidence": 0.8, + "canonical_issue_reason": "Issue 43845 is the earliest of the three highly similar \"huggingface \" reports, so it is the best canonical representative for that duplicate set.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43845 is the best overall issue candidate because it is the earliest and most likely to capture the shared huggingface report cluster; 43835 has a different title and appears unrelated.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43835", + "right": "issue:43845", + "accept": false, + "reason": "Different titles and no concrete shared details beyond weak similarity; likely unrelated." + }, + { + "left": "issue:43835", + "right": "issue:43846", + "accept": false, + "reason": "Different titles and no evidence they describe the same underlying problem." + }, + { + "left": "issue:43835", + "right": "issue:43859", + "accept": false, + "reason": "Different titles and no concrete overlap indicating the same issue." + }, + { + "left": "issue:43845", + "right": "issue:43846", + "accept": true, + "reason": "Same exact title and very close timestamps/body size; these look like duplicate reports of the same huggingface issue." + }, + { + "left": "issue:43845", + "right": "issue:43859", + "accept": true, + "reason": "Same exact title and strong similarity; likely duplicate reports of the same underlying issue." + }, + { + "left": "issue:43846", + "right": "issue:43859", + "accept": true, + "reason": "Same exact title and strong similarity; likely duplicate reports of the same underlying issue." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: the three issues with identical \"huggingface \" titles do form a plausible duplicate cluster, and 43835 has only weak similarity plus a different title. The canonical choice of the earliest huggingface report is conservative, and the soft-edge accepts are supported by the repeated title and high soft-similarity scores without overstating certainty." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "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": "3e81cee2e8719bb9a008dcd53f3041fc16b22bbc28d14f5670a6703f90c71ca8", + "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": "These are not duplicates of one substantive issue; the cluster looks like mostly spam/placeholder issues with one vague documentation suggestion, so there is no clear canonical bug report here.", + "confidence": 0.98, + "canonical_issue_reason": "No issue has enough concrete problem description to serve as a true canonical report.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44418 is the least spam-like and the only one with a somewhat descriptive title, but it still lacks actionable detail.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44371", + "right": "issue:44572", + "accept": false, + "reason": "Both look like spam placeholders and provide no evidence of the same underlying problem." + }, + { + "left": "issue:44371", + "right": "issue:44573", + "accept": false, + "reason": "Both look like spam placeholders and provide no evidence of the same underlying problem." + }, + { + "left": "issue:44371", + "right": "issue:44418", + "accept": false, + "reason": "One is spam-like and the other is a generic documentation suggestion; not the same concrete issue." + }, + { + "left": "issue:44371", + "right": "issue:44937", + "accept": false, + "reason": "Spam placeholder versus unrelated title about Google LLC; no shared bug/change." + }, + { + "left": "issue:44418", + "right": "issue:44572", + "accept": false, + "reason": "Generic documentation suggestion versus spam placeholder; not a duplicate underlying report." + }, + { + "left": "issue:44418", + "right": "issue:44573", + "accept": false, + "reason": "Generic documentation suggestion versus spam placeholder; not a duplicate underlying report." + }, + { + "left": "issue:44572", + "right": "issue:44573", + "accept": false, + "reason": "Both are spam-like placeholders, but there is no concrete issue content tying them to the same underlying bug." + }, + { + "left": "issue:44572", + "right": "issue:44937", + "accept": false, + "reason": "Spam placeholder versus unrelated Google LLC title; no evidence they describe the same issue." + }, + { + "left": "issue:44573", + "right": "issue:44937", + "accept": false, + "reason": "Spam placeholder versus unrelated Google LLC title; no evidence they describe the same issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 12, + "component_count": 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": "01374fd53b5e155cf20472def2ded764b9ffb42d9e1c1f6a91f5e742d8da3abe", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1644, + "estimated_input_tokens": 411, + "estimated_eval_tokens": 1078 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1644, + "estimated_input_tokens": 411, + "estimated_eval_tokens": 1078 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The issues are highly similar and share the exact title, but the available evidence is still soft rather than definitive. I would keep 45145 as the canonical issue if these are later confirmed duplicates, while treating the pair as likely duplicates rather than a certain duplicate match.", + "confidence": 0.83, + "canonical_issue_reason": "Issue 45145 is earlier and already has an inbound reference, so it is the best canonical representative among these highly similar reports.", + "canonical_pr_reason": null, + "best_issue_reason": "45145 is the most stable thread to keep because it predates 45151 and has incoming linkage, making it the better anchor for consolidation.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45145", + "right": "issue:45151", + "accept": false, + "reason": "The titles are identical and the similarity signal is strong, but the packet does not provide enough body-level evidence to confidently say they are the same underlying report; this is better treated as likely duplicates / highly similar issues rather than a confirmed duplicate." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The canonical choice is grounded, but the duplicate verdict is a bit too strong for the evidence available. The packet only shows identical titles plus a soft-similarity signal (with deterministic_accept=false), not enough to call the bodies 'identical' or the reports unquestionably the same. A more conservative phrasing would be that they are likely duplicates / highly similar reports, with 45145 as the better canonical issue because it is earlier and has an inbound reference." + }, + "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": 23, + "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-39895-13", + "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: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: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:45454", + "pull_request:43823|pull_request:45435", + "pull_request:43823|pull_request:43995", + "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:45213|pull_request:45435", + "pull_request:44827|pull_request:45213", + "pull_request:43823|pull_request:45213", + "pull_request:44827|pull_request:45435", + "pull_request:45454|pull_request:45489", + "pull_request:43823|pull_request:45489", + "pull_request:43251|pull_request:43254" + ], + "prepared_review_unit_hash": "2e9cd6aee97978de2cd5db127181f4b6a14de0146f227e8140a12191c26d06da", + "component_budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 23, + "serialized_chars": 37419, + "estimated_input_tokens": 9355, + "estimated_eval_tokens": 18966 + }, + "budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 23, + "serialized_chars": 37419, + "estimated_input_tokens": 9355, + "estimated_eval_tokens": 18966 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This is not a single duplicate cluster: most PRs are unrelated model additions, refactors, or test/docs changes. The only clear duplicate pair is the two fixed_cross_entropy kwargs fixes (43251 and 43254).", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43251 is the best canonical representative of the only true duplicate pair because it explicitly targets issue 43240 and has the more complete patch for forwarding cross-entropy kwargs.", + "best_issue_reason": null, + "best_pr_reason": "43251 is the best overall PR to anchor this cluster only because it is the clearest, issue-linked implementation among the near-duplicates; the rest are not duplicates of it.", + "soft_edge_verdicts": [ + { + "left": "pull_request:39895", + "right": "pull_request:43424", + "accept": false, + "reason": "VideoPrism model addition vs ExecuTorch dynamic-shape export test; different code paths and different goals." + }, + { + "left": "pull_request:42668", + "right": "pull_request:43823", + "accept": false, + "reason": "Processor-pretrained robustness fixes vs adding MobileLLM model support; unrelated changes." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43823", + "accept": false, + "reason": "cross_entropy kwargs fix vs MobileLLM model addition; no shared underlying bug/change." + }, + { + "left": "pull_request:43251", + "right": "pull_request:44827", + "accept": false, + "reason": "loss utility kwargs fix vs Mistral4 test cleanup; not the same issue." + }, + { + "left": "pull_request:43424", + "right": "pull_request:44827", + "accept": false, + "reason": "ExecuTorch dynamic-shape export test vs Mistral4 test fix; both are tests but for different behaviors." + }, + { + "left": "pull_request:43424", + "right": "pull_request:45435", + "accept": false, + "reason": "ExecuTorch exportability test vs Whisper tokenizer special-token indexing bug; unrelated." + }, + { + "left": "pull_request:43424", + "right": "pull_request:43823", + "accept": false, + "reason": "ExecuTorch export test vs MobileLLM addition; different features." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45454", + "accept": false, + "reason": "MobileLLM model addition vs Gemma4 training-mask change; not the same change." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45435", + "accept": false, + "reason": "MobileLLM model addition vs Whisper tokenizer bugfix; unrelated." + }, + { + "left": "pull_request:43823", + "right": "pull_request:43995", + "accept": false, + "reason": "MobileLLM addition vs Falcon output-collection refactor; different models and code paths." + }, + { + "left": "pull_request:43995", + "right": "pull_request:45435", + "accept": false, + "reason": "Falcon refactor vs Whisper tokenization bugfix; no overlap beyond being library code." + }, + { + "left": "pull_request:43995", + "right": "pull_request:45213", + "accept": false, + "reason": "Falcon model refactor vs repository skill/docs PR; clearly unrelated." + }, + { + "left": "pull_request:43995", + "right": "pull_request:44827", + "accept": false, + "reason": "Falcon output interface refactor vs Mistral4 tests; different models and problems." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45489", + "accept": false, + "reason": "Mistral4 test fixes vs Gemma3n/Gemma4 cache-sharing alignment; distinct underlying changes." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45170", + "accept": false, + "reason": "Mistral4 test fixes vs layrnorm/layernorm rename fix; unrelated." + }, + { + "left": "pull_request:43823", + "right": "pull_request:44827", + "accept": false, + "reason": "MobileLLM model addition vs Mistral4 test fixes; different scope and code paths." + }, + { + "left": "pull_request:45213", + "right": "pull_request:45435", + "accept": false, + "reason": "model-creation skill/docs PR vs Whisper tokenizer bugfix; not comparable as the same change." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45213", + "accept": false, + "reason": "Mistral4 tests vs model-creation skill; unrelated." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45213", + "accept": false, + "reason": "MobileLLM model addition vs skill file for creating models; not the same artifact or fix." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45435", + "accept": false, + "reason": "Mistral4 tests vs Whisper tokenizer fix; different models and behavior." + }, + { + "left": "pull_request:45454", + "right": "pull_request:45489", + "accept": false, + "reason": "Both touch Gemma4-adjacent files, but one is text-only training mask behavior and the other is cache-sharing alignment; they are not the same concrete fix." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45489", + "accept": false, + "reason": "MobileLLM addition vs Gemma3n/Gemma4 cache-sharing refactor; unrelated." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43254", + "accept": true, + "reason": "Both PRs fix the same fixed_cross_entropy/nn.functional.cross_entropy kwargs handling in the same file, with the same issue target 43240; they are duplicate implementations of the same bugfix." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: most PRs are clearly unrelated model additions, refactors, or test/docs changes, and the only defensible duplicate pair is 43251/43254, which both modify src/transformers/loss/loss_utils.py to address the shared 43240 cross-entropy kwargs issue. The soft-edge verdicts are conservative and do not overstate mergeability." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 2, + "component_count": 23, + "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": "1a56cb953af464b15869e54803db9a9d3a87bad3bcb459220f3113802dc75940", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7226, + "estimated_input_tokens": 1807, + "estimated_eval_tokens": 3870 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7226, + "estimated_input_tokens": 1807, + "estimated_eval_tokens": 3870 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs are about VibeVoice, but they target different scopes: #40546 is the base VibeVoice integration, while #42781 is a later realtime-oriented extension with additional tokenizer/feature-extraction work. They are related, not duplicates.", + "confidence": 0.84, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #40546 is the original, non-draft VibeVoice implementation with substantial discussion and core integration work; it is the better canonical representative of the base project change.", + "best_issue_reason": null, + "best_pr_reason": "PR #40546 is the most suitable representative because it covers the foundational VibeVoice addition and is more mature than the draft realtime follow-up.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40546", + "right": "pull_request:42781", + "accept": false, + "reason": "Shared VibeVoice files indicate the same model family, but the diffs suggest different concrete changes: #40546 adds the initial VibeVoice model, while #42781 adds realtime support plus acoustic tokenizer/feature-extraction changes. These could not safely be merged as one duplicate PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 3, + "component_count": 23, + "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": "97cc9e31c8d6088c37a1e0371488fd633cc9fdeed958fdbdbabb9ab0e6604277", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4674, + "estimated_input_tokens": 1169, + "estimated_eval_tokens": 2594 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4674, + "estimated_input_tokens": 1169, + "estimated_eval_tokens": 2594 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are unrelated pull requests in different model families: TimesFM covariate forecasting vs. DINOv3 image classification support. They should not be treated as duplicates.", + "confidence": 0.98, + "canonical_issue_reason": "No canonical issue: there are no issues in the cluster, and the PRs do not describe the same bug or change.", + "canonical_pr_reason": "No canonical PR for duplication purposes: the PRs target different code paths, different models, and different feature additions.", + "best_issue_reason": null, + "best_pr_reason": "If a representative PR is needed, 41224 is the more self-contained feature addition, but it is not a duplicate of 40755.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40755", + "right": "pull_request:41224", + "accept": false, + "reason": "Different features in different subsystems: TimesFM forecasting with covariates vs. DINOv3 image classification support. They do not fix the same underlying problem and could not plausibly be merged into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 4, + "component_count": 23, + "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": "88e60f926a84fb13a881f1700f827def51f6a2034775286ab5e8f054384e362c", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5929, + "estimated_input_tokens": 1483, + "estimated_eval_tokens": 3222 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5929, + "estimated_input_tokens": 1483, + "estimated_eval_tokens": 3222 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both pull requests are about adding DEIMv2 support to Transformers. PR #44339 is the more complete, later implementation and appears to supersede the smaller initial addition in #41356.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #44339 is the stronger canonical PR because it is the later, more complete DEIMv2 integration: it adds the model, loss, auto-mapping, conversion script, and related docs/tests, while #41356 is an earlier partial model/image-processor addition.", + "best_issue_reason": null, + "best_pr_reason": "PR #44339 is the best representative of the cluster because it most fully implements the DEIMv2 feature and includes the broader set of integration changes expected for a merged model addition.", + "soft_edge_verdicts": [ + { + "left": "pull_request:41356", + "right": "pull_request:44339", + "accept": true, + "reason": "Same underlying feature addition for DEIMv2 support, with substantial overlap in model files, docs, and tests; #44339 looks like the expanded successor to the earlier partial PR." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: both PRs target the same DEIMv2 issue, share several core files, and #44339 clearly adds broader integration pieces beyond #41356. The soft-edge verdict is conservative because it says the later PR 'looks like' an expanded successor rather than asserting a definitive relationship." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 5, + "component_count": 23, + "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": "852241ff3ed83c7a470cea6912b56d1287e74c29cfafabbfe03422cfd3f1d5ca", + "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-integration PRs (DeepSeek OCR vs Isaac) that only overlap on shared transformer scaffold files, not the underlying model or code path. They should not be clustered as duplicates.", + "confidence": 0.99, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "Neither PR is a duplicate of the other; if one must be chosen as a representative, 45186 is a more complete standalone model addition, but it does not subsume 41797.", + "soft_edge_verdicts": [ + { + "left": "pull_request:41797", + "right": "pull_request:45186", + "accept": false, + "reason": "Different models and implementations; shared auto/registry/docs files are generic integration boilerplate, not evidence of the same fix 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": 6, + "component_count": 23, + "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": "b91cdb431e7a1d26466351e225592ee2bbf3fecd30c682a15dfa7f0cb18a9649", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5407, + "estimated_input_tokens": 1352, + "estimated_eval_tokens": 2960 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5407, + "estimated_input_tokens": 1352, + "estimated_eval_tokens": 2960 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two pull requests are unrelated model-addition changes: one introduces InternVLFlash, the other introduces AXK1. They overlap only in shared auto-registration boilerplate, not in the underlying code change or bug fix.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "PR 42166 is the more complete exemplar of a model-integration change because it includes docs and tests, but it is not a duplicate of PR 45426; the two PRs target different models and code paths.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42166", + "right": "pull_request:45426", + "accept": false, + "reason": "Different model integrations (InternVLFlash vs AXK1); shared registry files are incidental boilerplate, 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": 7, + "component_count": 23, + "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": "4a7cff3be0bd0d748c76a78b472c445ad062fbf6c5b621b2977013f18b41e2e0", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 10170, + "estimated_input_tokens": 2543, + "estimated_eval_tokens": 5342 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 10170, + "estimated_input_tokens": 2543, + "estimated_eval_tokens": 5342 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The cluster is heterogeneous: one Whisper ASR pipeline fix, one Whisper tokenizer/test refactor, one obviously unrelated CLIP/ViT test-only change, and one FP8 dequant/model-loading PR. None are duplicate candidates.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "44815 is the strongest standalone PR: it contains a coherent functional FP8 dequant/model-loading change, while the others are either narrow Whisper fixes or clearly unrelated test edits.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42227", + "right": "pull_request:43291", + "accept": false, + "reason": "Both are Whisper-related, but they fix different code paths: ASR pipeline return_language forwarding vs tokenizer decode/test behavior. Not the same underlying bug or mergeable into one PR." + }, + { + "left": "pull_request:42227", + "right": "pull_request:43665", + "accept": false, + "reason": "Unrelated changes: a real Whisper pipeline fix versus dummy assertions in CLIP/ViT tests. Not duplicate or same change." + }, + { + "left": "pull_request:42227", + "right": "pull_request:44815", + "accept": false, + "reason": "Different subsystems and bugs: Whisper ASR pipeline vs FP8 dequant/model-loading logic. No shared concrete code-path problem." + }, + { + "left": "pull_request:43291", + "right": "pull_request:43665", + "accept": false, + "reason": "Whisper tokenizer/test refactor is unrelated to CLIP/ViT test edits. Not the same underlying issue." + }, + { + "left": "pull_request:43291", + "right": "pull_request:44815", + "accept": false, + "reason": "Different areas and fixes: Whisper tokenizer decode behavior vs FP8 dequant/model-loading. No basis to merge as duplicates." + }, + { + "left": "pull_request:43665", + "right": "pull_request:44815", + "accept": false, + "reason": "Completely unrelated: placeholder test failures in CLIP/ViT versus a functional FP8 dequant change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 8, + "component_count": 23, + "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": "8b7a368b2e951532511d9429a03917d7d69f3dc0427fc7ed59c5f7c517c3f282", + "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": "These are related workflow-maintenance PRs, but they are not the same change: one targets Node 24 compatibility for GitHub Actions, the other is a broader \u201clatest versions\u201d bump, mainly for docker/build-push-action and a few workflow actions.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "PR 42975 is the better representative of the cluster because it has a specific compatibility goal and a more coherent scope; PR 42976 is a broader dependency refresh with different intent.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42975", + "right": "pull_request:42976", + "accept": false, + "reason": "Do not merge: they overlap in workflow files, but they are not fixing the same concrete problem. 42975 is about Node 24 compatibility for GitHub Actions, while 42976 is a general upgrade of action versions, especially docker/build-push-action. Shared subsystem alone is insufficient for duplicate treatment." + } + ] + }, + "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": 23, + "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-42978-10", + "nodes": [ + "pull_request:42978", + "pull_request:43448", + "pull_request:43451", + "pull_request:43838", + "pull_request:43973", + "pull_request:44601", + "pull_request:45157", + "pull_request:45283", + "pull_request:45490", + "pull_request:45532" + ], + "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:43448|pull_request:45532", + "pull_request:43451|pull_request:45490", + "pull_request:43451|pull_request:45532", + "pull_request:43838|pull_request:45283", + "pull_request:43973|pull_request:45283", + "pull_request:44601|pull_request:45532", + "pull_request:45157|pull_request:45532", + "pull_request:45283|pull_request:45532", + "pull_request:45490|pull_request:45532" + ], + "prepared_review_unit_hash": "2736eda2d1bb1be84d547aa98f8bebfcaef6da8ef7697bdf1003a12d3d14f5d8", + "component_budget": { + "node_count": 10, + "item_count": 10, + "soft_pair_count": 14, + "serialized_chars": 32133, + "estimated_input_tokens": 8034, + "estimated_eval_tokens": 16324 + }, + "budget": { + "node_count": 10, + "item_count": 10, + "soft_pair_count": 14, + "serialized_chars": 32133, + "estimated_input_tokens": 8034, + "estimated_eval_tokens": 16324 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is a mix of unrelated model-addition PRs and one distributed-loading infrastructure PR. The soft overlaps are mostly boilerplate in docs/auto registration files, not the same underlying bug or feature. No duplicate pair should be merged.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR: these PRs implement different models or features (ViT NEPA, Molmo, Molmo2, Qwen3-ASR, Lfm2Audio, PP loading, PrismML GGUF, Qwen3.5 GGUF, CTSM, SLANet) and are not duplicates.", + "best_issue_reason": null, + "best_pr_reason": "PR 45532 is the strongest representative of the cluster\u2019s model-support pattern, with a substantial model implementation and the clearest end-to-end addition, but it still targets SLANet specifically and is not interchangeable with the others.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42978", + "right": "pull_request:43448", + "accept": false, + "reason": "Different model ports: ViT NEPA vs Molmo. Shared docs/auto-registration files are generic scaffolding, not the same change." + }, + { + "left": "pull_request:42978", + "right": "pull_request:43451", + "accept": false, + "reason": "Different model ports: ViT NEPA vs Molmo2. Similar repository boilerplate, but separate architectures and codepaths." + }, + { + "left": "pull_request:42978", + "right": "pull_request:45490", + "accept": false, + "reason": "Different model ports: ViT NEPA vs CTSM. The overlap is only in shared registration/docs files." + }, + { + "left": "pull_request:43448", + "right": "pull_request:43451", + "accept": false, + "reason": "Molmo and Molmo2 are related by name but are different model generations with different files and support code." + }, + { + "left": "pull_request:43448", + "right": "pull_request:45490", + "accept": false, + "reason": "Unrelated model additions: Molmo vs CTSM. Shared auto/docs touches are boilerplate, not a duplicate change." + }, + { + "left": "pull_request:43448", + "right": "pull_request:45532", + "accept": false, + "reason": "Different model additions: Molmo vs SLANet. The common files are registry/docs plumbing only." + }, + { + "left": "pull_request:43451", + "right": "pull_request:45490", + "accept": false, + "reason": "Different model additions: Molmo2 vs CTSM. No shared bug or concrete codepath." + }, + { + "left": "pull_request:43451", + "right": "pull_request:45532", + "accept": false, + "reason": "Different model additions: Molmo2 vs SLANet. Similar scaffolding does not make them duplicates." + }, + { + "left": "pull_request:43838", + "right": "pull_request:45283", + "accept": false, + "reason": "Qwen3-ASR model support and Qwen3.5 GGUF loading are different features affecting different codepaths." + }, + { + "left": "pull_request:43973", + "right": "pull_request:45283", + "accept": false, + "reason": "Lfm2Audio model support is unrelated to Qwen3.5 GGUF loader mapping; same broad audio/modeling area only." + }, + { + "left": "pull_request:44601", + "right": "pull_request:45532", + "accept": false, + "reason": "Pipeline-parallel loading infrastructure is unrelated to SLANet model support." + }, + { + "left": "pull_request:45157", + "right": "pull_request:45532", + "accept": false, + "reason": "PrismML GGUF dequantization support and SLANet model support fix different problems and codepaths." + }, + { + "left": "pull_request:45283", + "right": "pull_request:45532", + "accept": false, + "reason": "Qwen3.5 GGUF loading support and SLANet model support are different changes with no shared underlying bug." + }, + { + "left": "pull_request:45490", + "right": "pull_request:45532", + "accept": false, + "reason": "CTSM and SLANet are different model-support PRs; they only overlap on generic docs/model registration files, not the same feature." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 10, + "component_count": 23, + "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": "c2221f8667941beb44318f487c8d9c9a5e52b7ddf6578898f44736b99d90b0ba", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 7467, + "estimated_input_tokens": 1867, + "estimated_eval_tokens": 3990 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 7467, + "estimated_input_tokens": 1867, + "estimated_eval_tokens": 3990 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These three PRs are unrelated synthetic \"don't merge\" changes with only incidental overlap in shared test filenames. They do not fix the same bug or implement the same change, so none should be clustered as duplicates.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: each PR edits a different area and change intent (packaging/versioning, workflow/test plumbing, and CI workflow/tokenizer/test noise).", + "best_issue_reason": null, + "best_pr_reason": "PR 43532 is the least representative as a cluster anchor because it has the broadest workflow/test plumbing changes, but it still is not a true duplicate of the others.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43488", + "right": "pull_request:43532", + "accept": false, + "reason": "Different underlying changes: setup/version/dependency edits vs CI/workflow and test plumbing; shared test filename is incidental." + }, + { + "left": "pull_request:43488", + "right": "pull_request:45476", + "accept": false, + "reason": "Different code paths and intent: packaging/version tweaks vs workflow + tokenizer/test noise; not the same change." + }, + { + "left": "pull_request:43532", + "right": "pull_request:45476", + "accept": false, + "reason": "Both touch test/workflow files, but the concrete edits are unrelated 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": 11, + "component_count": 23, + "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": "e40ccb5c930499f9742b47b2bb59374536733f1625c54c514fb69895af195380", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6081, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6081, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two separate open PRs adding different pipeline features: promptable concept segmentation vs promptable visual segmentation. They share scaffolding and docs, but they target different tasks/models and should not be deduplicated.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR: the PRs implement different segmentation pipelines with different model integrations and documentation.", + "best_issue_reason": null, + "best_pr_reason": "If a representative is needed, 43613 is slightly broader, but it is still a distinct feature and not a substitute for 43612.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43612", + "right": "pull_request:43613", + "accept": false, + "reason": "Different underlying changes: one adds Promptable Concept Segmentation, the other Promptable Visual Segmentation. Shared pipeline/docs scaffolding is not enough to treat them as the same fix/change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 12, + "component_count": 23, + "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": "386457bdd478d8aa4918a472de2787ebe80e477f50e4859d2377474e37fe1298", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4292, + "estimated_input_tokens": 1073, + "estimated_eval_tokens": 2402 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4292, + "estimated_input_tokens": 1073, + "estimated_eval_tokens": 2402 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs touch `src/transformers/trainer.py`, but they address different extension points: one changes gradient-accumulation loss scaling, the other extracts signature-column filtering for custom trainers. They are not the same underlying change.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43651 is the better canonical representative because it makes a concrete behavior change in `training_step` and introduces `_loss_is_scaled_for_ga` to control loss scaling semantics.", + "best_issue_reason": null, + "best_pr_reason": "PR #43651 is the strongest standalone PR in the cluster: it fixes a specific training-loss scaling path with a clear user-facing effect and is more directly substantive than the helper-method extraction in #43663.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43651", + "right": "pull_request:43663", + "accept": false, + "reason": "Different code paths and goals: #43651 changes gradient accumulation loss normalization, while #43663 refactors signature-column inspection for custom trainers. Shared file/subsystem is not enough to treat them as duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 13, + "component_count": 23, + "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": "fbd1bcbf912d76cd72e921d798e552e89d2ff4c1191e259ed849a87626a3f297", + "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 PRs target the same import-time annotation problem in `src/transformers/cli/serve.py`, with overlapping fixes around `TypeAdapter`/future annotations. PR #43842 is the cleanest, most focused representative of that bug fix.", + "confidence": 0.71, + "canonical_issue_reason": null, + "canonical_pr_reason": "#43842 most directly addresses the concrete `TypeAdapter` NameError in the CLI serve path with the smallest, clearest change set.", + "best_issue_reason": null, + "best_pr_reason": "#43842 is the best cluster representative because it fixes the specific runtime failure directly and is narrower than the broader annotation cleanup in the other PRs.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43656", + "right": "pull_request:43842", + "accept": true, + "reason": "Both are fixing the same serve CLI annotation/`TypeAdapter` NameError path in the same file, and their changes are plausibly mergeable into one PR." + }, + { + "left": "pull_request:43836", + "right": "pull_request:43842", + "accept": true, + "reason": "Both target the same `TypeAdapter` annotation failure in `serve.py`; the differences look like implementation details around the same bug fix rather than separate changes." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The cluster summary is grounded in the packet: all three PRs touch `src/transformers/cli/serve.py` and the diffs are consistent with the same import-time annotation/`TypeAdapter` NameError class of bug. The chosen representative, #43842, is also the smallest and most focused fix. The soft-edge accepts are reasonable and the paired changes look mergeable as overlapping implementations of the same bugfix." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 14, + "component_count": 23, + "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": "b5b9cb84d0a4ccee3281d75ffa388baeac95980e02907c19f7d51ac394ea6f55", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 7756, + "estimated_input_tokens": 1939, + "estimated_eval_tokens": 4134 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 7756, + "estimated_input_tokens": 1939, + "estimated_eval_tokens": 4134 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two PRs are near-exact duplicates for full GPT-OSS GGUF support; the earlier PR is only a partial fallback and should not be merged with the fuller implementation. The latest non-draft PR is the best canonical PR.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45506 is the most complete and current version: it adds the GGUF architecture mapping, loader handling, and tests, and it is explicitly positioned as the latest superseding work.", + "best_issue_reason": null, + "best_pr_reason": "PR #45506 is the strongest representative of the cluster because it contains the full end-to-end GGUF support changes, while #45500 is a draft variant and #43757 is only a fallback workaround.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45500", + "right": "pull_request:45506", + "accept": true, + "reason": "Same underlying GPT-OSS GGUF loading change: same title, same issue target, same core files, and the later PR appears to be the finalized version of the earlier draft." + }, + { + "left": "pull_request:43757", + "right": "pull_request:45500", + "accept": false, + "reason": "43757 is only a narrow fallback-to-gpt-neox workaround, while 45500 is a broader full GGUF-loading implementation; they overlap on one file and issue, but not the same concrete fix." + }, + { + "left": "pull_request:43757", + "right": "pull_request:45506", + "accept": false, + "reason": "43757 is a partial stopgap and 45506 is a full support PR; they address the same broad topic but not the same mergeable change set or exact code-path fix." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 15, + "component_count": 23, + "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": 23, + "review_unit_index": 1, + "review_unit_count": 2, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43996-28", + "nodes": [ + "pull_request:43996", + "pull_request:43997", + "pull_request:44007", + "pull_request:44010", + "pull_request:44013", + "pull_request:44017", + "pull_request:44018", + "pull_request:44019", + "pull_request:44024", + "pull_request:44044", + "pull_request:44056", + "pull_request:44059", + "pull_request:44066", + "pull_request:44068", + "pull_request:44071", + "pull_request:44072", + "pull_request:44073", + "pull_request:44074", + "pull_request:44076", + "pull_request:44085", + "pull_request:44086", + "pull_request:44098", + "pull_request:44101", + "pull_request:44116", + "pull_request:44129", + "pull_request:44154", + "pull_request:44161", + "pull_request:44722" + ], + "soft_pairs": [ + "pull_request:43996|pull_request:44085", + "pull_request:43996|pull_request:44044", + "pull_request:44066|pull_request:44085", + "pull_request:44007|pull_request:44072", + "pull_request:44072|pull_request:44722", + "pull_request:44066|pull_request:44072", + "pull_request:44013|pull_request:44044", + "pull_request:44066|pull_request:44086", + "pull_request:44066|pull_request:44071", + "pull_request:44018|pull_request:44068", + "pull_request:44066|pull_request:44068", + "pull_request:44018|pull_request:44066", + "pull_request:44019|pull_request:44722", + "pull_request:44086|pull_request:44722", + "pull_request:44019|pull_request:44085", + "pull_request:44018|pull_request:44019", + "pull_request:44019|pull_request:44071", + "pull_request:44068|pull_request:44116", + "pull_request:44068|pull_request:44722", + "pull_request:44071|pull_request:44722", + "pull_request:44018|pull_request:44116", + "pull_request:44018|pull_request:44086", + "pull_request:44018|pull_request:44085", + "pull_request:44071|pull_request:44085", + "pull_request:44068|pull_request:44086", + "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:43997|pull_request:44073", + "pull_request:44017|pull_request:44071", + "pull_request:44059|pull_request:44071", + "pull_request:44073|pull_request:44101", + "pull_request:44059|pull_request:44068", + "pull_request:44018|pull_request:44059", + "pull_request:44056|pull_request:44101", + "pull_request:44073|pull_request:44161", + "pull_request:44101|pull_request:44161", + "pull_request: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:43997|pull_request:44161", + "pull_request:44074|pull_request:44161", + "pull_request:43997|pull_request:44071", + "pull_request:44010|pull_request:44071", + "pull_request:44010|pull_request:44073", + "pull_request:44018|pull_request:44129", + "pull_request:43997|pull_request:44074", + "pull_request:43997|pull_request:44129", + "pull_request:44010|pull_request:44018", + "pull_request:44071|pull_request:44129", + "pull_request:43996|pull_request:44161", + "pull_request:43996|pull_request:44017", + "pull_request:44010|pull_request:44044", + "pull_request:44044|pull_request:44074", + "pull_request:44066|pull_request:44722", + "pull_request:44085|pull_request:44722", + "pull_request:44085|pull_request:44129", + "pull_request:43996|pull_request:44007", + "pull_request:43996|pull_request:44129", + "pull_request:44044|pull_request:44129", + "pull_request:44044|pull_request:44154", + "pull_request:44024|pull_request:44076" + ], + "prepared_review_unit_hash": "4adf64f07e74dccaa041a5263441a9d2095b26d5eac2780a425656d7a78eedac", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 77, + "serialized_chars": 84500, + "estimated_input_tokens": 21125, + "estimated_eval_tokens": 42506 + }, + "budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 72, + "serialized_chars": 76934, + "estimated_input_tokens": 19234, + "estimated_eval_tokens": 38724 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a broad family of model-by-model output-tracing refactors, but only two duplicate groups look real: the GPT-Neo pair and the GPT-J/CodeGen pair. Most other edges share only the umbrella issue #43979 or the same refactor theme, which is too broad to merge as duplicates.", + "confidence": 0.82, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44722 is the best representative of the duplicate subset: it is the latest and broadest GPT-J/CodeGen standardized-output-tracing refactor, and it appears to subsume the earlier GPT-J-only iterations.", + "best_issue_reason": null, + "best_pr_reason": "PR 44722 is the strongest global representative because it most clearly captures the repeated GPT-J output-tracing change, with the cleanest and broadest scope among the duplicate candidates.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43996", + "right": "pull_request:44085", + "accept": false, + "reason": "Different models/files; same umbrella output-tracing theme is too broad to call these duplicates." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44044", + "accept": false, + "reason": "Different model code paths; both are standardized-output-tracing refactors, but not the same change." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44085", + "accept": true, + "reason": "Same GPT-J output-tracing refactor in the same file; these look like overlapping iterations of the same change." + }, + { + "left": "pull_request:44007", + "right": "pull_request:44072", + "accept": false, + "reason": "ResNet and EfficientNet are separate implementations; same refactor pattern is not enough for duplicate status." + }, + { + "left": "pull_request:44072", + "right": "pull_request:44722", + "accept": false, + "reason": "Different model families and code paths; only the high-level output-tracing theme overlaps." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44072", + "accept": false, + "reason": "Different models/files; same standardized output-tracing umbrella is too broad." + }, + { + "left": "pull_request:44013", + "right": "pull_request:44044", + "accept": false, + "reason": "MobileNetV2 and DeBERTaV2 are unrelated code paths; only the refactor style matches." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J/CodeGen and MGP-STR are different model changes, not the same underlying fix." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44071", + "accept": false, + "reason": "Different model families and implementations; same output-tracing modernization is not a duplicate signal." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44068", + "accept": true, + "reason": "Same GPT-Neo file and same output-tracing refactor; these are likely duplicate iterations of the same change." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44068", + "accept": false, + "reason": "GPT-J/CodeGen vs GPT-Neo are different code paths, so this is only a thematic match." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44066", + "accept": false, + "reason": "Different model families; shared issue target alone does not make them duplicates." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44722", + "accept": false, + "reason": "ResNet and GPT-J/CodeGen are unrelated model changes." + }, + { + "left": "pull_request:44086", + "right": "pull_request:44722", + "accept": false, + "reason": "Different models and concrete code paths; same tracking issue is not enough." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44085", + "accept": false, + "reason": "ResNet and GPT-J are separate refactors, not the same underlying bug or change." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44019", + "accept": false, + "reason": "GPT-Neo and ResNet are different models; no shared concrete code path." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44071", + "accept": false, + "reason": "ResNet and MPT are unrelated implementations, despite the same general refactor theme." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo and MGP-STR are different model code paths." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-Neo and GPT-J/CodeGen are different models; only the refactor pattern overlaps." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44722", + "accept": false, + "reason": "MPT and GPT-J/CodeGen are separate implementations; not the same concrete fix." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo and Flaubert are different model code paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44086", + "accept": false, + "reason": "Different models/files; same standardized-output-tracing theme is too broad." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44085", + "accept": false, + "reason": "GPT-Neo and GPT-J are distinct refactors, not duplicates." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44085", + "accept": false, + "reason": "MPT and GPT-J are unrelated concrete changes." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo and MGP-STR are different code paths." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT2 and MGP-STR are separate model refactors." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J and MGP-STR are unrelated changes." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT2 and GPT-J/CodeGen are different model implementations." + }, + { + "left": "pull_request:44098", + "right": "pull_request:44101", + "accept": false, + "reason": "ViLT and Flaubert/XLM are different model code paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-Neo and MPT are separate refactors with no shared concrete bug." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44722", + "accept": false, + "reason": "SegFormer and GPT-J/CodeGen are unrelated model changes." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44161", + "accept": false, + "reason": "SegFormer and LongT5 are different implementations; only the umbrella refactor theme matches." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44073", + "accept": false, + "reason": "SegFormer and VisualBert are distinct code paths." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-Neo and MPT are different model changes." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44086", + "accept": false, + "reason": "MPT and MGP-STR are unrelated model implementations." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44073", + "accept": false, + "reason": "RegNet and VisualBert are separate code paths." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44101", + "accept": false, + "reason": "MPNet and Flaubert/XLM are different model refactors." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44161", + "accept": false, + "reason": "VisualBert and LongT5 are unrelated model implementations." + }, + { + "left": "pull_request:44101", + "right": "pull_request:44161", + "accept": false, + "reason": "Flaubert/XLM and LongT5 are separate model code paths." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44018", + "accept": false, + "reason": "SegFormer and GPT-Neo are different models; same refactor style is not enough." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44056", + "accept": false, + "reason": "RegNet and MPNet are separate implementations." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44076", + "accept": false, + "reason": "MPT and ImageGPT are different model code paths." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44101", + "accept": false, + "reason": "TextNet and Flaubert/XLM are unrelated changes." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44161", + "accept": false, + "reason": "MPNet and LongT5 are different model implementations." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44018", + "accept": false, + "reason": "RegNet and GPT-Neo are unrelated model refactors." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44161", + "accept": false, + "reason": "RegNet and LongT5 are different code paths." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44161", + "accept": false, + "reason": "TextNet and LongT5 are separate model changes." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44071", + "accept": false, + "reason": "RegNet and MPT are different model implementations." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44071", + "accept": false, + "reason": "SqueezeBert and MPT are unrelated concrete refactors." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44073", + "accept": false, + "reason": "SqueezeBert and VisualBert are different model code paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44129", + "accept": false, + "reason": "GPT-Neo and SpeechT5 are separate model changes." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44129", + "accept": false, + "reason": "MPT and SpeechT5 are different concrete refactors." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44161", + "accept": false, + "reason": "CVT/FNet and LongT5 are unrelated model implementations." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44017", + "accept": false, + "reason": "CVT/FNet and SegFormer are different model code paths." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44044", + "accept": false, + "reason": "SqueezeBert and DeBERTaV2 are different models; only the refactor style overlaps." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44074", + "accept": false, + "reason": "DeBERTaV2 and TextNet are unrelated model changes." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44722", + "accept": true, + "reason": "Same GPT-J/CodeGen output-tracing refactor; 44722 looks like the later, broader version of the same change." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44722", + "accept": true, + "reason": "Both target the same GPT-J output-tracing code path; these look like duplicate iterations." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44129", + "accept": false, + "reason": "GPT-J and SpeechT5 are different models and concrete code paths." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44129", + "accept": false, + "reason": "DeBERTaV2 and SpeechT5 are unrelated refactors." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44154", + "accept": false, + "reason": "DeBERTaV2 and ViTS are different model implementations." + }, + { + "left": "pull_request:44024", + "right": "pull_request:44076", + "accept": false, + "reason": "FocalNet and ImageGPT are separate model code paths." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is conservative and stays close to the packet: it identifies the broad output-tracing refactor theme and only elevates the clearest same-code-path pairs as duplicates. The accepted soft-edge verdicts are also restrained and mergeable for the same bugfix. I don\u2019t see an overstatement of evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 16, + "component_count": 23, + "review_unit_index": 2, + "review_unit_count": 2, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43996-28", + "nodes": [ + "pull_request:43997", + "pull_request:44010", + "pull_request:44056", + "pull_request:44073", + "pull_request:44074", + "pull_request:44129" + ], + "soft_pairs": [ + "pull_request:44073|pull_request:44074", + "pull_request:44056|pull_request:44073", + "pull_request:44056|pull_request:44074", + "pull_request:43997|pull_request:44010", + "pull_request:44010|pull_request:44129" + ], + "prepared_review_unit_hash": "ed6ec968cf7aaba71a9aad14a8aa970e7f83343c5ebf46a3bac4eb4edd254fae", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 77, + "serialized_chars": 84500, + "estimated_input_tokens": 21125, + "estimated_eval_tokens": 42506 + }, + "budget": { + "node_count": 6, + "item_count": 6, + "soft_pair_count": 5, + "serialized_chars": 13066, + "estimated_input_tokens": 3267, + "estimated_eval_tokens": 6790 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These are model-specific refactors to standardized output tracing/capture infrastructure across different architectures, not duplicates of the same bug or change. They share a theme, but each PR touches a distinct code path and model implementation.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR cleanly represents the whole set because each PR is a separate model migration. If one must be chosen as the most representative, 44073 is the clearest self-contained example of the standardized output-tracing refactor.", + "best_issue_reason": null, + "best_pr_reason": "44073 is the most representative PR for this cluster: it explicitly introduces capture_outputs and _can_record_outputs for one model, making it a clean exemplar of the broader tracing migration.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44073", + "right": "pull_request:44074", + "accept": false, + "reason": "Both are output-tracing refactors, but they apply to different models (VisualBert vs TextNet) with different forward paths and tests; not the same concrete change." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44073", + "accept": false, + "reason": "Same broad migration pattern, but MPNet and VisualBert have separate model-specific tracing hooks and code paths; they are not mergeable as one PR." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44074", + "accept": false, + "reason": "These are distinct refactors for different model implementations, not the same underlying bug or fix." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44010", + "accept": false, + "reason": "RegNet and SqueezeBert each migrate output collection in their own model code; the overlap is thematic only." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44129", + "accept": false, + "reason": "They target different models and different tracing logic; sharing a standardized output-capture pattern is not enough to treat them as duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 17, + "component_count": 23, + "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": "fc8e6f5dfcdb6c901a766e39719bbc319fbeb9a828e0bb9ef5f75829a237a9ee", + "component_budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 39, + "serialized_chars": 39784, + "estimated_input_tokens": 9946, + "estimated_eval_tokens": 20148 + }, + "budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 39, + "serialized_chars": 39784, + "estimated_input_tokens": 9946, + "estimated_eval_tokens": 20148 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs all use similar 'refactor output tracing' wording, but they are model-specific changes across unrelated code paths (backbones, encoder-decoder wrappers, depth models, Mamba/RWKV, etc.). I would not dedupe them on title similarity alone.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: the set spans many independent model implementations rather than one duplicated fix.", + "best_issue_reason": null, + "best_pr_reason": "No single best PR for deduplication; the PRs are separate, self-contained refactors in different models.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43998", + "right": "pull_request:44002", + "accept": false, + "reason": "Different model code paths; same tracing refactor theme is too broad to be a duplicate." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44000", + "accept": false, + "reason": "Distinct architectures and files; not the same underlying bug or change." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44001", + "accept": false, + "reason": "Only shares the generic output-tracing pattern, not a concrete shared fix." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44003", + "accept": false, + "reason": "Unrelated model family and implementation; not mergeable as one PR." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44029", + "accept": false, + "reason": "Different code paths in different models; similarity is just thematic." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44001", + "accept": false, + "reason": "MobileNet vs UnivNet are unrelated models; no shared concrete defect." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44002", + "accept": false, + "reason": "These touch different model implementations, not one duplicate fix." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44025", + "accept": false, + "reason": "Different architectures and files; only the refactor style matches." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44026", + "accept": false, + "reason": "Separate model-specific refactors; not the same underlying change." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44027", + "accept": false, + "reason": "Only broad subsystem overlap; the code paths are different." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44001", + "accept": false, + "reason": "Vision-text dual encoder and UnivNet are unrelated; not a duplicate." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44002", + "accept": false, + "reason": "Same high-level theme, but different model code and behavior." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44003", + "accept": false, + "reason": "Mamba/FalconMamba refactor is a different concrete change from vision-text dual encoder." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44025", + "accept": false, + "reason": "Different models and files; too generic to treat as the same fix." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44002", + "accept": false, + "reason": "UnivNet and UperNet are unrelated code paths; not a duplicate change." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44003", + "accept": false, + "reason": "Different architectures and refactor targets; not the same bug or change." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44025", + "accept": false, + "reason": "No shared concrete code-path problem, only similar output-tracing wording." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44029", + "accept": false, + "reason": "UperNet and RWKV are unrelated model implementations." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44003", + "accept": false, + "reason": "Different model families and concrete edits; not dedupable." + }, + { + "left": "pull_request:44003", + "right": "pull_request:44004", + "accept": false, + "reason": "Mamba and CodeGen changes affect different subsystems and code paths." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44004", + "accept": false, + "reason": "Timm backbone and CodeGen are unrelated; same refactor language is insufficient." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44004", + "accept": false, + "reason": "UnivNet vs CodeGen are different models with different outputs and call paths." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44004", + "accept": false, + "reason": "UperNet and CodeGen do not share a concrete underlying bug." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44004", + "accept": false, + "reason": "Vision-text dual encoder and CodeGen are not the same change." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44026", + "accept": false, + "reason": "Depth Anything and Vision-Encoder-Decoder are unrelated; same helper usage is not enough." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44027", + "accept": false, + "reason": "Different model families and implementation details; not a duplicate." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44028", + "accept": false, + "reason": "Depth model refactor vs SuperPoint refactor are separate changes." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44029", + "accept": false, + "reason": "Depth Anything and RWKV touch entirely different code paths." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44030", + "accept": false, + "reason": "Different model types and files; no shared concrete fix." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44029", + "accept": false, + "reason": "Vision-Encoder-Decoder and RWKV are unrelated implementations." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44027", + "accept": false, + "reason": "Encoder-decoder wrappers are different model classes here, not one duplicate PR." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44028", + "accept": false, + "reason": "Only generic tracing refactor similarity; not the same code path." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44030", + "accept": false, + "reason": "Different models and behavior; not plausibly one merged fix." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44028", + "accept": false, + "reason": "Speech-encoder-decoder and SuperPoint are unrelated changes." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44029", + "accept": false, + "reason": "Different model families and concrete edits." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44030", + "accept": false, + "reason": "No shared underlying bug; just similar refactor phrasing." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44029", + "accept": false, + "reason": "SuperPoint and RWKV are unrelated model implementations." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44030", + "accept": false, + "reason": "Different model code paths; not the same change." + }, + { + "left": "pull_request:44029", + "right": "pull_request:44030", + "accept": false, + "reason": "RWKV and DPR are separate models with separate fixes." + } + ] + }, + "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": 23, + "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": "9c56da9ee5142f48da9654b381d7986dd73a126d5d2adf78ad9f1da5644564cc", + "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-addition PRs (PenguinVL, DeepSeek-OCR-2, Voxtral TTS) with only boilerplate overlap in auto-registration/docs. They do not appear to fix the same underlying change, so none should be clustered as duplicates.", + "confidence": 0.99, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "No PR is a true duplicate of the others; if one must be kept as a representative, 45075 is the most self-contained with an explicit implementation scope and broad test coverage, but it is not equivalent to the others.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44662", + "right": "pull_request:45401", + "accept": false, + "reason": "Different model integrations for different products; shared files are just common registration/docs boilerplate, not the same code-path change." + }, + { + "left": "pull_request:44662", + "right": "pull_request:45075", + "accept": false, + "reason": "PenguinVL and DeepSeek-OCR-2 are unrelated model implementations with only generic transformer scaffolding in common." + }, + { + "left": "pull_request:45075", + "right": "pull_request:45401", + "accept": false, + "reason": "DeepSeek-OCR-2 vs Voxtral TTS are distinct model additions touching different model logic and tests, so they are not the same underlying PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 19, + "component_count": 23, + "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": "d85351de58336264d2bf247c1b3e5f731d8fbb64090fccbfef64b4ac9a70f9ab", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4408, + "estimated_input_tokens": 1102, + "estimated_eval_tokens": 2460 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4408, + "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 in the broad area of module/kernel fusion, but they are not the same change: #44979 adds a new Module Fusion API scaffold, while #45363 implements n-to-1 kernel fusion through KernelConfig and model loading integrations. The overlap is thematic rather than a shared concrete code-path fix.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "If a single PR must represent the cluster, #45363 is the stronger canonical choice because it is the more complete, end-to-end user-facing implementation; #44979 is an enabling API foundation rather than the same change.", + "best_issue_reason": null, + "best_pr_reason": "#45363 is the best representative PR in this cluster because it shows a concrete application of kernel fusion in the loading/integration path, whereas #44979 is a separate API feature scaffold with different scope.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44979", + "right": "pull_request:45363", + "accept": false, + "reason": "Do not merge: #44979 introduces a generic Module Fusion API, while #45363 wires KernelConfig-based n-to-1 kernel fusion into model loading. They are related infrastructure, but not the same underlying bug or change and would not plausibly be merged into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 20, + "component_count": 23, + "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": "cd787b1e8640f6eb1143dbadd6efc5aa58d4681e13ff077eee8b7caeeb5a2e33", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4414, + "estimated_input_tokens": 1104, + "estimated_eval_tokens": 2464 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4414, + "estimated_input_tokens": 1104, + "estimated_eval_tokens": 2464 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs patch the same Mistral regex pre-tokenizer AttributeError in the same file and target the same issue; PR 45317 looks like the fuller duplicate with added tests.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "45317 is the more complete fix for the same AttributeError path, with the same code change plus tests, so it is the better canonical PR among the duplicates.", + "best_issue_reason": null, + "best_pr_reason": "45317 is the best representative because it directly fixes the bug and includes test coverage, whereas 45086 is the narrower earlier variant.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45086", + "right": "pull_request:45317", + "accept": true, + "reason": "Same underlying bug and same code-path fix in _patch_mistral_regex; 45317 is essentially a superset of 45086 with tests, so these are duplicate PRs." + } + ] + }, + "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": 23, + "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": "abc0f79201563ba925e1b2e47774ba47d45ec0e3899a871d1e742a7fbceab1b0", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5396, + "estimated_input_tokens": 1349, + "estimated_eval_tokens": 2954 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5396, + "estimated_input_tokens": 1349, + "estimated_eval_tokens": 2954 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related but distinct PRs for the same heterogeneous-modeling effort: one adds per-layer config plumbing, the other adds runtime/model integration on top of it. They share infrastructure files, but they do not look like the same concrete change.", + "confidence": 0.89, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45332 is the broader end-to-end feature PR and appears to build on the per-layer config plumbing, so it is the better representative of the cluster.", + "best_issue_reason": null, + "best_pr_reason": "PR 45332 is the strongest standalone representative because it covers the actual model/runtime support in addition to the heterogeneous config groundwork.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45332", + "right": "pull_request:45333", + "accept": false, + "reason": "Reject: 45333 is focused on per-layer configuration support, while 45332 is model/runtime integration (cache, masking, GPT-OSS/modeling changes). They share a feature area and helper files, but they are complementary stackable PRs rather than one duplicate 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": 22, + "component_count": 23, + "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": "af87c1bb4ddeb249b217b263f0353f6e560ed79df4c36a100637ce42eee9b30d", + "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 are related only at a high level (typing cleanup), but they are not the same concrete change. PR 45415 is a broad sweep adding type checking across many transformer modules, while PR 45425 adds specific typing aliases/protocols for modeling_utils/peft and a check-types update. I would not cluster them as duplicates.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45425 is the more focused, concrete typing change: it introduces reusable protocol/alias definitions and targets a small set of files, making it a better representative than the broad cross-module typing sweep in PR 45415.", + "best_issue_reason": null, + "best_pr_reason": "PR 45425 is the best representative because it addresses a specific typing gap with clear scope, whereas PR 45415 is a wide-ranging refactor-like typing pass that is less suitable as a canonical match for this pair.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45415", + "right": "pull_request:45425", + "accept": false, + "reason": "Reject: both are typing-related, but they do not appear to fix the same underlying bug/change. 45415 is a broad type-checking sweep across many files; 45425 adds targeted typing protocols/aliases for modeling_utils and peft. Shared files alone are not enough to merge them 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": 23, + "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": "bfa28de9a896c2584f7fbf945093b2da81f07d67c3e2fe69d5e69f6457123447", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5068, + "estimated_input_tokens": 1267, + "estimated_eval_tokens": 2790 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5068, + "estimated_input_tokens": 1267, + "estimated_eval_tokens": 2790 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These pull requests are not duplicates: one only skips flash-attn composite-model tests for Sam3LiteText, while the other fixes a real model-parallel/device bug in AltCLIP/ChineseCLIP and related text models.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45487 is the stronger representative change because it fixes a concrete runtime bug in model-parallel code paths, whereas PR #45470 is only a test skip for unsupported flash-attn coverage.", + "best_issue_reason": null, + "best_pr_reason": "PR #45487 is the best fit for the cluster as a substantive code fix with clear user-facing impact; PR #45470 is a narrow test-only workaround and not the same underlying change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45470", + "right": "pull_request:45487", + "accept": false, + "reason": "Different problems and code paths: 45470 skips flash-attn tests for Sam3LiteText due to unsupported flash attention, while 45487 fixes token_type_ids device handling and parallelism issues in AltCLIP/ChineseCLIP-related models." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + } + ] +} diff --git a/snapshots/20260421T120044Z/analysis-runs/hybrid-model-20260421t120044z/manifest.json b/snapshots/20260421T120044Z/analysis-runs/hybrid-model-20260421t120044z/manifest.json new file mode 100644 index 0000000000000000000000000000000000000000..2761465247f8614cc87f1fe5504ecfe87deb6077 --- /dev/null +++ b/snapshots/20260421T120044Z/analysis-runs/hybrid-model-20260421t120044z/manifest.json @@ -0,0 +1,14 @@ +{ + "analysis_id": "hybrid-model-20260421t120044z", + "artifacts": { + "hybrid": "snapshots/20260421T120044Z/analysis-runs/hybrid-model-20260421t120044z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260421T120044Z/analysis-runs/hybrid-model-20260421t120044z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "model": null, + "published_at": "2026-04-21T12:17:10Z", + "repo": "huggingface/transformers", + "schema_version": 1, + "snapshot_id": "20260421T120044Z", + "variant": "hybrid" +} diff --git a/snapshots/20260421T120044Z/manifest.json b/snapshots/20260421T120044Z/manifest.json index 4e79402522b76612b1819d54e9e46d7d2badade9..cc9d34117b617ae011573455546bd00912ac755a 100644 --- a/snapshots/20260421T120044Z/manifest.json +++ b/snapshots/20260421T120044Z/manifest.json @@ -32,6 +32,24 @@ "timeline_events": 722 }, "extracted_at": "2026-04-21T12:00:44Z", + "published_analysis": { + "canonical_analysis_id": "hybrid-model-20260421t120044z", + "runs": { + "hybrid-model-20260421t120044z": { + "analysis_id": "hybrid-model-20260421t120044z", + "artifacts": { + "hybrid": "snapshots/20260421T120044Z/analysis-runs/hybrid-model-20260421t120044z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260421T120044Z/analysis-runs/hybrid-model-20260421t120044z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "manifest_path": "snapshots/20260421T120044Z/analysis-runs/hybrid-model-20260421t120044z/manifest.json", + "model": null, + "published_at": "2026-04-21T12:17:10Z", + "variant": "hybrid" + } + }, + "schema_version": 1 + }, "repo": "huggingface/transformers", "snapshot_id": "20260421T120044Z", "watermark": {