diff --git a/analysis/current/analysis-report-hybrid.json b/analysis/current/analysis-report-hybrid.json index 329346e184eeb01201797026225ac926b681ec66..413e13b0d61eecd8b47dacfcc442f43bfb879abf 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": "20260518T060213Z", - "generated_at": "2026-05-18T06:15:33Z", + "snapshot_id": "20260518T120024Z", + "generated_at": "2026-05-18T12:15:50Z", "evidence_quality": "full", "llm_enrichment": true, "meta_bugs": [ @@ -1346,278 +1346,6 @@ } ] }, - { - "cluster_id": "cluster-45081-4", - "summary": "Cluster of 2 issues and 2 PRs centered on issue #45081.", - "status": "open", - "confidence": 0.8, - "canonical_issue_number": 45081, - "canonical_pr_number": 45317, - "issue_numbers": [ - 45081, - 45356 - ], - "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-45987-4", - "summary": "Cluster of 1 issues and 3 PRs centered on issue #45987.", - "status": "open", - "confidence": 0.8, - "canonical_issue_number": 45987, - "canonical_pr_number": 46005, - "issue_numbers": [ - 45987 - ], - "pr_numbers": [ - 45997, - 46005, - 46010 - ], - "evidence_types": [ - "closing_reference", - "shared_issue_target", - "soft_similarity" - ], - "pr_comparisons": [ - { - "left_pr_number": 45997, - "right_pr_number": 46005, - "code_similarity": 0.597, - "size_similarity": 0.737, - "file_overlap": 1.0, - "area_overlap": 0.35, - "patch_similarity": 0.182, - "shared_filenames": [ - "src/transformers/cache_utils.py" - ], - "shared_file_areas": [ - { - "filename": "src/transformers/cache_utils.py", - "left_ranges": [ - [ - 290, - 302 - ] - ], - "right_ranges": [ - [ - 291, - 297 - ], - [ - 377, - 383 - ] - ] - } - ] - }, - { - "left_pr_number": 45997, - "right_pr_number": 46010, - "code_similarity": 0.306, - "size_similarity": 0.288, - "file_overlap": 0.333, - "area_overlap": 0.35, - "patch_similarity": 0.17, - "shared_filenames": [ - "src/transformers/cache_utils.py" - ], - "shared_file_areas": [ - { - "filename": "src/transformers/cache_utils.py", - "left_ranges": [ - [ - 290, - 302 - ] - ], - "right_ranges": [ - [ - 291, - 297 - ], - [ - 377, - 383 - ] - ] - } - ] - }, - { - "left_pr_number": 46005, - "right_pr_number": 46010, - "code_similarity": 0.516, - "size_similarity": 0.212, - "file_overlap": 0.333, - "area_overlap": 1.0, - "patch_similarity": 0.156, - "shared_filenames": [ - "src/transformers/cache_utils.py" - ], - "shared_file_areas": [ - { - "filename": "src/transformers/cache_utils.py", - "left_ranges": [ - [ - 291, - 297 - ], - [ - 377, - 383 - ] - ], - "right_ranges": [ - [ - 291, - 297 - ], - [ - 377, - 383 - ] - ] - } - ] - } - ] - }, - { - "cluster_id": "cluster-44001-4", - "summary": "Cluster of 4 related pull requests linked by soft_similarity.", - "status": "open", - "confidence": 0.5, - "canonical_issue_number": null, - "canonical_pr_number": 44002, - "issue_numbers": [], - "pr_numbers": [ - 44001, - 44002, - 44004, - 44029 - ], - "evidence_types": [ - "soft_similarity" - ], - "pr_comparisons": [ - { - "left_pr_number": 44001, - "right_pr_number": 44002, - "code_similarity": 0.175, - "size_similarity": 0.595, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.375, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 44001, - "right_pr_number": 44004, - "code_similarity": 0.087, - "size_similarity": 0.237, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.262, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 44001, - "right_pr_number": 44029, - "code_similarity": 0.083, - "size_similarity": 0.256, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.212, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 44002, - "right_pr_number": 44004, - "code_similarity": 0.125, - "size_similarity": 0.398, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.304, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 44002, - "right_pr_number": 44029, - "code_similarity": 0.135, - "size_similarity": 0.43, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.329, - "shared_filenames": [], - "shared_file_areas": [] - }, - { - "left_pr_number": 44004, - "right_pr_number": 44029, - "code_similarity": 0.253, - "size_similarity": 0.925, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.456, - "shared_filenames": [], - "shared_file_areas": [] - } - ] - }, { "cluster_id": "cluster-43240-3", "summary": "Cluster of 1 issues and 2 PRs centered on issue #43240.", @@ -1892,6 +1620,65 @@ } ] }, + { + "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-45561-3", "summary": "Cluster of 1 issues and 2 PRs centered on issue #45561.", @@ -2301,16 +2088,7 @@ ] } ], - "duplicate_issues": [ - { - "cluster_id": "cluster-45081-4", - "canonical_issue_number": 45081, - "duplicate_issue_numbers": [ - 45356 - ], - "reason": "Issues in cluster-45081-4 are treated as duplicates because they share closing_reference, shared_issue_target, soft_similarity evidence." - } - ], + "duplicate_issues": [], "duplicate_prs": [ { "cluster_id": "cluster-41115-3", @@ -2375,17 +2153,6 @@ "target_issue_number": 43979, "reason": "PRs in cluster-43979-11 are treated as duplicates because they converge on issue #43979 with closing_reference, shared_issue_target, soft_similarity evidence." }, - { - "cluster_id": "cluster-44001-4", - "canonical_pr_number": 44002, - "duplicate_pr_numbers": [ - 44001, - 44004, - 44029 - ], - "target_issue_number": null, - "reason": "PRs in cluster-44001-4 are treated as duplicates because they share soft_similarity evidence." - }, { "cluster_id": "cluster-44018-2", "canonical_pr_number": 44068, @@ -2396,13 +2163,13 @@ "reason": "PRs in cluster-44018-2 are treated as duplicates because they share soft_similarity evidence." }, { - "cluster_id": "cluster-45081-4", + "cluster_id": "cluster-45081-3", "canonical_pr_number": 45317, "duplicate_pr_numbers": [ 45086 ], "target_issue_number": 45081, - "reason": "PRs in cluster-45081-4 are treated as duplicates because they converge on issue #45081 with closing_reference, shared_issue_target, soft_similarity evidence." + "reason": "PRs in cluster-45081-3 are treated as duplicates because they converge on issue #45081 with closing_reference, shared_issue_target, soft_similarity evidence." }, { "cluster_id": "cluster-45561-3", @@ -2421,23 +2188,13 @@ ], "target_issue_number": 45647, "reason": "PRs in cluster-45647-3 are treated as duplicates because they converge on issue #45647 with closing_reference, shared_issue_target evidence." - }, - { - "cluster_id": "cluster-45987-4", - "canonical_pr_number": 46005, - "duplicate_pr_numbers": [ - 45997, - 46010 - ], - "target_issue_number": 45987, - "reason": "PRs in cluster-45987-4 are treated as duplicates because they converge on issue #45987 with closing_reference, shared_issue_target, soft_similarity evidence." } ], "best_issue": { - "cluster_id": "cluster-45987-4", - "issue_number": 45987, - "reason": "Issue #45987 is the strongest global issue candidate because it is open, belongs to a cluster with 4 artifacts, and carries 3 discussion comments plus 5 inbound references.", - "score": 151.0 + "cluster_id": "cluster-45647-3", + "issue_number": 45647, + "reason": "Issue #45647 is the strongest global issue candidate because it is open, belongs to a cluster with 3 artifacts, and carries 12 discussion comments plus 8 inbound references.", + "score": 162.0 }, "best_pr": { "cluster_id": "cluster-43979-11", diff --git a/analysis/current/analysis-report-hybrid.llm-reviews.json b/analysis/current/analysis-report-hybrid.llm-reviews.json index 03a923e674cf6dd780bd37bfd7a21a705142529c..135dcc9ce0e39263c393d8877c40bd5ce3f1c3d5 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": "20260518T060213Z", - "generated_at": "2026-05-18T06:15:33Z", + "snapshot_id": "20260518T120024Z", + "generated_at": "2026-05-18T12:15:50Z", "model": "gpt-5.4-mini?service_tier=flex", "reviews": [ { @@ -135,20 +135,20 @@ "issue:43854|issue:43866", "issue:41093|issue:43638" ], - "prepared_review_unit_hash": "c601e7f24a975b98b03931622c9e9d9a69ae663ad75e646abebf4445223f996a", + "prepared_review_unit_hash": "79ed1f2392eac2d8c62cced47b776ec6fac3df1e51400e86f8dff6456431c81f", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 66, - "serialized_chars": 47558, + "serialized_chars": 47560, "estimated_input_tokens": 11890, "estimated_eval_tokens": 24036 }, @@ -157,381 +157,321 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Conservative triage: the ZeRO-3 index-0/out-of-bounds cluster is solid (43596/43638/41762), and the v5 AutoTokenizer wrong-token-ID regression cluster is also solid (45812/45920). The Qwen3_5Moe rotary/offload pair is too under-evidenced from titles alone and should remain tentative/separate rather than merged as a duplicate.", - "confidence": 0.83, - "canonical_issue_reason": "Issue 43596 is the best anchor for the ZeRO-3 duplicate cluster: it names the shared index-0/out-of-bounds failure clearly and is broad enough to cover the Bert/Gemma3 load/init variants.", + "summary": "Mostly a loose topical cluster, not one real duplicate set. The only strong duplicate-looking groups are the Zero-3/BERT IndexError pair and the v5 AutoTokenizer token-ID regression pair; the rest are distinct model/back-end/doc issues.", + "confidence": 0.71, + "canonical_issue_reason": "issue:45920 is the broadest, clearest representative of the shared v5 AutoTokenizer wrong-token-ID regression, and it subsumes the narrower Granite-only report in issue:45812.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43596 is the clearest representative of the shared ZeRO-3 empty-tensor/index-0 OOB bug family.", + "best_issue_reason": "issue:45920 is the strongest standalone issue in this cluster: broad impact, clear regression framing, and it captures the common tokenizer code path better than the narrower instance.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44322", "right": "issue:45464", "accept": false, - "reason": "Missing config attribute vs streaming API failure; different bugs." + "reason": "Same model family, but different failures: missing config field vs streaming chat/completions crash." }, { "left": "issue:45237", "right": "issue:45685", "accept": false, - "reason": "Different backend/model problems, not the same code path." + "reason": "Different backends and symptoms: AMD GPU failure vs MPS histogram unsupported for Int." }, { "left": "issue:43716", "right": "issue:45237", "accept": false, - "reason": "Only broad model/runtime similarity; failure modes differ." + "reason": "Different model and error path: image preprocessor/model dtype mismatch vs AMD GPU runtime failure." }, { "left": "issue:43828", "right": "issue:45237", "accept": false, - "reason": "Both are runtime issues, but not the same underlying bug." + "reason": "Both are runtime errors, but the underlying bugs are unrelated." }, { "left": "issue:45237", "right": "issue:45507", "accept": false, - "reason": "Unrelated model/backend errors." + "reason": "Different model-specific issues; no shared concrete code path." }, { "left": "issue:44292", "right": "issue:45237", "accept": false, - "reason": "Different model and failure mode." + "reason": "Different model families and failure modes; not the same bug." }, { "left": "issue:30277", "right": "issue:35545", "accept": false, - "reason": "DeepSpeed collective mismatch vs ONNX export; unrelated." + "reason": "DeepSpeed collective mismatch is unrelated to ONNX export failure." }, { "left": "issue:35545", "right": "issue:42915", "accept": false, - "reason": "Different model/export issues." + "reason": "Both involve model support, but one is ONNX export and the other is FineGrainedFP8 loading." }, { "left": "issue:35545", "right": "issue:41720", "accept": false, - "reason": "ONNX export bug vs device-map CUDA assert; not the same issue." + "reason": "Different issues: export-to-ONNX vs device-map CUDA assert." }, { "left": "issue:45901", "right": "issue:46015", "accept": false, - "reason": "Table QA crash vs outdated docs; unrelated." + "reason": "Unrelated: table-question-answering crash vs outdated RAG examples." }, { "left": "issue:44509", "right": "issue:44512", "accept": false, - "reason": "Both docs regressions, but they cover different removed features." + "reason": "Both are stale docs, but they target different removed features and pages." }, { "left": "issue:35545", "right": "issue:36010", "accept": false, - "reason": "ONNX export error vs GenerationMixin import error." + "reason": "Different problems: export error vs GenerationMixin import breakage." }, { "left": "issue:44805", "right": "issue:45823", "accept": false, - "reason": "Different symptoms and code paths." + "reason": "Similar-looking runtime errors, but different models and failure modes." }, { "left": "issue:43638", "right": "issue:44805", "accept": false, - "reason": "Shape/index errors in different contexts; not clearly the same bug." + "reason": "Both are IndexError reports, but one is zero3/BERT init and the other is a mask/tensor shape mismatch." }, { "left": "issue:43596", "right": "issue:44805", "accept": false, - "reason": "ZeRO-3 empty-tensor/OOB vs mask-shape mismatch; different bugs." + "reason": "Different underlying bugs: zero3/BERT out-of-bounds vs mask/tensor shape mismatch." }, { "left": "issue:44805", "right": "issue:45161", "accept": false, - "reason": "Different runtime issues." + "reason": "Different model paths: masking/indexing error vs GPT-OSS tensor-parallel issue." }, { "left": "issue:44805", "right": "issue:44918", "accept": false, - "reason": "Embedding/trainer issue vs generic mask-shape mismatch." + "reason": "Different subsystems and errors; not the same bug." }, { "left": "issue:34634", "right": "issue:35545", "accept": false, - "reason": "Unrelated processor bug vs export bug." + "reason": "Processor voice_preset bug is unrelated to ModernBERT ONNX export." }, { "left": "issue:30277", "right": "issue:43638", "accept": false, - "reason": "Both involve DeepSpeed, but the failure modes differ." + "reason": "Both involve DeepSpeed, but the actual failures are different." }, { "left": "issue:43596", "right": "issue:43638", "accept": true, - "reason": "Same ZeRO-3 empty-index/index-0 out-of-bounds failure during load/init." + "reason": "Same Zero-3/BERT out-of-bounds IndexError on model loading/training; looks like one bug with the same code path." }, { "left": "issue:43039", "right": "issue:45161", "accept": false, - "reason": "Liger Kernel loss-path issue vs GPT-OSS TP-only failure." + "reason": "Liger Kernel cross-entropy routing is unrelated to GPT-OSS tensor-parallel failure." }, { "left": "issue:43975", "right": "issue:44457", "accept": false, - "reason": "Detokenization regression vs LoRA save/reload mismatch." + "reason": "Different issues: detokenization regression vs LoRA merge/save mismatch." }, { "left": "issue:44292", "right": "issue:44457", "accept": false, - "reason": "Different models and failure modes." + "reason": "Different models and symptoms; not a duplicate." }, { "left": "issue:34928", "right": "issue:44805", "accept": false, - "reason": "Activation-checkpointing/FSDP recompute mismatch is not the mask-shape error." + "reason": "Activation-checkpointing/FSDP tensor recomputation issue is unrelated to the mask-shape IndexError." }, { "left": "issue:24643", "right": "issue:30277", "accept": false, - "reason": "Different DeepSpeed errors and code paths." + "reason": "Both mention DeepSpeed, but the runtime errors come from different code paths." }, { "left": "issue:45859", "right": "issue:45902", "accept": false, - "reason": "Related Qwen3_5Moe rotary/offload theme, but the titles describe different failures and there is not enough evidence they are the same bug." + "reason": "Same class name, but different bugs: CPU offload incompatibility vs uninitialized-memory after materialization." }, { "left": "issue:43638", "right": "issue:45161", "accept": false, - "reason": "Different model families and error types." - }, - { - "left": "issue:34928", - "right": "issue:36331", - "accept": false, - "reason": "Different training API regressions." + "reason": "Different models and failure modes; not the same bug." }, { - "left": "issue:43596", - "right": "issue:45161", + "left": "issue:44509", + "right": "issue:44512", "accept": false, - "reason": "Unrelated DeepSpeed load bug vs GPT-OSS TP issue." + "reason": "Both are doc cleanup topics, but they concern different removed CLI/docs features." }, { - "left": "issue:43716", - "right": "issue:45507", + "left": "issue:44841", + "right": "issue:45823", "accept": false, - "reason": "Different model-specific implementation bugs." + "reason": "Processor failure and device-mapping mismatch are unrelated." }, { - "left": "issue:43828", + "left": "issue:43425", "right": "issue:44292", "accept": false, - "reason": "Different runtime/backend issues." - }, - { - "left": "issue:43039", - "right": "issue:44918", - "accept": false, - "reason": "Loss/kernel issue vs SFT embedding unpacking issue." - }, - { - "left": "issue:41762", - "right": "issue:43638", - "accept": true, - "reason": "Same ZeRO-3 empty-tensor/index-0 out-of-bounds failure during load/init." - }, - { - "left": "issue:43827", - "right": "issue:44512", - "accept": false, - "reason": "Both are docs removals, but they target different removed items." - }, - { - "left": "issue:43716", - "right": "issue:44560", - "accept": false, - "reason": "Different multimodal/model-processing failures." + "reason": "Torch version compatibility is unrelated to Qwen-3-8B-NVFP4 runtime failure." }, { "left": "issue:43828", "right": "issue:45685", "accept": false, - "reason": "Autocast dtype mismatch vs MPS kernel missing; unrelated." - }, - { - "left": "issue:39290", - "right": "issue:45081", - "accept": false, - "reason": "Different model/config issues." + "reason": "Autocast dtype mismatch vs MPS histogram unsupported are unrelated." }, { - "left": "issue:43425", - "right": "issue:44292", + "left": "issue:44560", + "right": "issue:45685", "accept": false, - "reason": "Torch version compatibility vs model runtime error." + "reason": "Video StopIteration bug is unrelated to the MPS Int histogram error." }, { - "left": "issue:45823", - "right": "issue:45874", + "left": "issue:43039", + "right": "issue:43828", "accept": false, - "reason": "Both Gemma4-related, but one is device mismatch and the other is reverse-embedding overhead." + "reason": "Different model/integration bugs; no shared concrete code path." }, { - "left": "issue:44918", - "right": "issue:45161", + "left": "issue:30277", + "right": "issue:36331", "accept": false, - "reason": "Different models and failure modes." + "reason": "DeepSpeed collective mismatch is unrelated to custom trainer signature breakage." }, { - "left": "issue:43716", - "right": "issue:45507", + "left": "issue:45237", + "right": "issue:46015", "accept": false, - "reason": "Different model-specific bugs." + "reason": "Model-on-AMD-GPU failure is unrelated to outdated RAG docs." }, { - "left": "issue:43828", - "right": "issue:45161", + "left": "issue:34928", + "right": "issue:36010", "accept": false, - "reason": "Different runtime/backend issues." + "reason": "Training/checkpointing issue is unrelated to GenerationMixin import error." }, { - "left": "issue:44805", - "right": "issue:46015", + "left": "issue:43039", + "right": "issue:44805", "accept": false, - "reason": "Mask-shape runtime error vs outdated docs; unrelated." + "reason": "Different subsystems and errors; not the same bug." }, { - "left": "issue:44560", - "right": "issue:45685", + "left": "issue:41762", + "right": "issue:43638", "accept": false, - "reason": "Different model/backend issues." + "reason": "Both are loading/runtime errors, but one is Qwen3 auto device mapping and the other is BERT zero3 init/training." }, { "left": "issue:43901", "right": "issue:44512", "accept": false, - "reason": "Different docs regressions." + "reason": "Both are docs issues, but they target different outdated API references." }, { "left": "issue:39290", "right": "issue:41720", "accept": false, - "reason": "Different model/backend problems." + "reason": "Different Qwen3/Gemma3-vs-Qwen3-vl style failures; not the same issue." }, { "left": "issue:45237", "right": "issue:45901", "accept": false, - "reason": "Different issues with no concrete overlap." + "reason": "Different models/features: AMD GPU support vs table-question-answering crash." }, { "left": "issue:44841", "right": "issue:45823", "accept": false, - "reason": "Processor failure vs device-map mismatch." + "reason": "Processor failure and device-map forward mismatch are unrelated." }, { "left": "issue:43039", "right": "issue:44560", "accept": false, - "reason": "Different model/runtime failures." + "reason": "Liger Kernel cross-entropy issue is unrelated to Qwen3-vl video StopIteration." }, { "left": "issue:43975", "right": "issue:46015", "accept": false, - "reason": "Model detokenization regression vs docs cleanup." + "reason": "Detokenization regression is unrelated to stale RAG docs." }, { "left": "issue:45812", "right": "issue:45920", "accept": true, - "reason": "Same v5 AutoTokenizer wrong-token-ID regression across model families." + "reason": "Same underlying v5 AutoTokenizer wrong-token-ID regression; 45920 is the broader umbrella and 45812 is a narrower instance." }, { "left": "issue:43866", "right": "issue:44863", "accept": false, - "reason": "Corrupted checkpoint vs checkpoint-loading implementation bug." + "reason": "Both are checkpoint/model-loading reports, but they concern different models and different failures." }, { "left": "issue:44863", "right": "issue:46015", "accept": false, - "reason": "Model loading bug vs outdated docs." + "reason": "NemotronH checkpoint loading bug is unrelated to docs cleanup." + }, + { + "left": "issue:44292", + "right": "issue:46015", + "accept": false, + "reason": "Qwen-3-8B-NVFP4 runtime failure is unrelated to outdated RAG docs." }, { "left": "issue:43854", "right": "issue:43866", "accept": false, - "reason": "Different model-loading problems." + "reason": "Unit-test model-loading issue and corrupted checkpoint report are different problems." }, { "left": "issue:41093", "right": "issue:43638", "accept": false, - "reason": "Mask index-shape mismatch vs ZeRO-3 empty-tensor/OOB error." - }, - { - "left": "issue:41762", - "right": "issue:43596", - "accept": true, - "reason": "Same ZeRO-3 index-0 out-of-bounds failure during model init/load." - }, - { - "left": "issue:43901", - "right": "issue:44512", - "accept": false, - "reason": "Different docs issues." - }, - { - "left": "issue:39290", - "right": "issue:41720", - "accept": false, - "reason": "Different model/backend problems." - }, - { - "left": "issue:45237", - "right": "issue:45901", - "accept": false, - "reason": "Unrelated runtime bugs." - }, - { - "left": "issue:44841", - "right": "issue:45823", - "accept": false, - "reason": "Different model-processing bug vs device-map mismatch." + "reason": "Similar-looking IndexErrors, but the mask/tensor mismatch and zero3/BERT out-of-bounds are different bugs." } ] }, "evaluator_result": { - "accept": false, - "feedback": "Most of the clustering is conservative and grounded, especially the ZeRO-3 index-0/OOB group and the AutoTokenizer regression group. However, one soft-edge accept is overstated: `issue:45859` vs `issue:45902` share a Qwen3_5Moe rotary/offload theme, but the packet only gives titles, and those titles describe different failure manifestations (`forward` CPU-offload incompatibility vs `inv_freq` uninitialized memory after materialization). That is not enough evidence to confidently mark them as the same duplicate bug. The summary should keep that relationship tentative or separate." + "accept": true, + "feedback": "Summary is grounded in the packet and stays conservative. The two accepted duplicate pairs are plausible from the titles/body framing, and the broader cluster characterization as mostly loose topical similarity is supported. Canonical issue choice for the v5 AutoTokenizer regression is reasonable. No mergeability concerns apply here since there is no PR pair." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -639,250 +579,274 @@ "issue:44530|issue:45005", "issue:42915|issue:45399" ], - "prepared_review_unit_hash": "faf1d7764b93b6fc090d8e16244a74ef61d539d80ca540d8ed5f22be98a90715", + "prepared_review_unit_hash": "b4d759ca2c2da0c69a39553ab66b552df56b66b52501b8432dadf66c223f2ece", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 39, - "serialized_chars": 39065, - "estimated_input_tokens": 9767, - "estimated_eval_tokens": 19790 + "serialized_chars": 39064, + "estimated_input_tokens": 9766, + "estimated_eval_tokens": 19788 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is mostly heterogeneous regression reports across unrelated models, configs, and runtime paths. Only the custom-module loading/cache-path pair looks like the same underlying bug; the rest should stay separate.", - "confidence": 0.58, - "canonical_issue_reason": "issue:45632 is the clearest root-cause report: it identifies the cache/path collision for local models sharing a leaf directory name, which also explains the wrong custom-module symptom in issue:45698.", + "summary": "This cluster is mostly a heterogeneous mix of unrelated Transformers regressions across different models, loaders, precision modes, and runtime backends. The only strong duplicate-like pair is the SigLIP2 output-quality reports.", + "confidence": 0.76, + "canonical_issue_reason": "Use issue 43994 as the canonical issue for the SigLIP2 regression: it states the broader user-visible failure more directly, while 43054 reads as a narrower comparative symptom report.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45632 is the best representative issue in this cluster because it states the underlying defect directly rather than just the downstream failure mode.", + "best_issue_reason": "If a single issue must represent the one clear duplicate subgroup here, 43994 is the clearest and most general report. There is no strong cluster-wide canonical because the rest of the items are not the same bug.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:39290", "right": "issue:43531", "accept": false, - "reason": "Both mention sliding_window, but they affect different model families and code paths (Gemma3/vLLM vs Qwen3-MoE), so they do not look like the same bug." + "reason": "Both mention sliding_window, but they involve different model families and different failure contexts; not the same concrete bug." }, { "left": "issue:43792", "right": "issue:45901", "accept": false, - "reason": "Different tasks, models, and failure modes; no evidence they share a concrete code-path bug." + "reason": "Whisper loading failure and table QA crash are unrelated code paths and symptoms." }, { "left": "issue:43425", "right": "issue:45070", "accept": false, - "reason": "Both are version/regression complaints, but one is torch compatibility and the other is a pydantic PretrainedConfig field break; too broad to merge." + "reason": "Different regressions: one is a torch version incompatibility, the other is a PretrainedConfig/pydantic breakage." }, { "left": "issue:43596", "right": "issue:45800", "accept": false, - "reason": "Different errors and contexts: DeepSpeed ZeRO-3/BertModel loading versus a torch/transformers version incompatibility report." + "reason": "Both are version/runtime errors, but the repros and causes differ substantially (Deepspeed/Bert vs torch-transformers incompatibility)." }, { "left": "issue:43975", "right": "issue:45901", "accept": false, - "reason": "Unrelated symptoms and subsystems; no shared bug is evident." - }, - { - "left": "issue:43792", - "right": "issue:45070", - "accept": false, - "reason": "Whisper runtime failure and PretrainedConfig/pydantic breakage are unrelated." + "reason": "Detokenization bug and table-question-answering crash are unrelated." }, { "left": "issue:45362", "right": "issue:45464", "accept": false, - "reason": "Both involve Qwen3.5, but one is a transformers chat crash and the other is a streaming chat/completions API failure; the concrete failing paths differ." + "reason": "Same broad model family, but one is a transformers chat crash and the other is a streaming chat/completions API failure; not enough evidence of one bug." }, { "left": "issue:44292", "right": "issue:45901", "accept": false, - "reason": "Different model/runtime issues with no shared underlying defect." + "reason": "Qwen-3 NVFP4 runtime failure is unrelated to the table QA crash." }, { "left": "issue:39290", "right": "issue:45823", "accept": false, - "reason": "Both are model-specific runtime regressions, but the models and failure modes are unrelated." + "reason": "Different models and different symptoms: sliding_window attribute error vs device mismatch in forward." }, { "left": "issue:43404", "right": "issue:43866", "accept": false, - "reason": "Mistral3 weight-tying bug versus an Ovis2 checkpoint corruption report; not the same issue." + "reason": "lm_head weight tying in Mistral3 and a corrupted Ovis2 checkpoint are different problems." }, { "left": "issue:45464", "right": "issue:45685", "accept": false, - "reason": "Chat API streaming failure and an MPS histogram/int issue are unrelated." + "reason": "Streaming inference API failure and MPS histogram error are unrelated." }, { "left": "issue:43541", "right": "issue:45685", "accept": false, - "reason": "Different backends and failures; one is torch dynamo/grouped_mm tracing on Mixtral, the other is an MPS integer histogram error." + "reason": "Different backends and ops; no shared underlying code-path bug is apparent." }, { "left": "issue:43541", "right": "issue:43828", "accept": false, - "reason": "Both mention dtype-like behavior in MoE models, but the concrete models, triggers, and failing operations are different." + "reason": "Both are precision/runtime issues, but they affect different models and failure mechanisms." }, { "left": "issue:44841", "right": "issue:45084", "accept": false, - "reason": "Processor failure for Voxtral is unrelated to a template-node compilation error." + "reason": "Processor loading failure and template compilation error are not the same issue." }, { "left": "issue:36010", "right": "issue:39290", "accept": false, - "reason": "ImportError for GenerationMixin and a sliding_window_pattern regression are separate issues." + "reason": "ImportError for GenerationMixin and a sliding_window config error are unrelated." }, { "left": "issue:41093", "right": "issue:41762", "accept": false, - "reason": "Both are index errors, but they occur in different loading scenarios and model families; the titles do not point to one bug." + "reason": "Both are index/shape errors during loading, but the repro paths and underlying tensors differ." }, { "left": "issue:45800", "right": "issue:45823", "accept": false, - "reason": "Torch/transformers version incompatibility and a Gemma4 device-map forward mismatch are distinct problems." + "reason": "Torch version incompatibility and a Gemma4 device mismatch are different bugs." }, { "left": "issue:43296", "right": "issue:43366", "accept": false, - "reason": "PaddleOCR-VL loading in vLLM and GGUF gpt-oss support are unrelated." + "reason": "PaddleOCR-VL loading in vLLM and GPT-OSS GGUF support are unrelated requests/failures." }, { "left": "issue:43572", "right": "issue:45070", "accept": false, - "reason": "Missing pad_token_idx in StableLmConfig is a different config regression from the pydantic PretrainedConfig field issue." + "reason": "Both are v5 regressions, but they hit different config fields and different failure modes." }, { "left": "issue:44387", "right": "issue:45005", "accept": false, - "reason": "Int4 quantization memory/OOM regression and tied-weights translation-model issues are different bugs." + "reason": "Quantization OOM and tied-weights behavior are unrelated." }, { "left": "issue:43541", "right": "issue:45161", "accept": false, - "reason": "Both touch MoE models, but one is a torch dynamo tracing failure and the other is a tensor-parallel support issue for GPT-OSS." + "reason": "Both mention MoE/parallelism, but the concrete failures differ too much to be duplicates." }, { "left": "issue:43828", "right": "issue:45464", "accept": false, - "reason": "Different models, APIs, and failure modes; no concrete shared code-path problem." + "reason": "Autocast dtype mismatch and chat/completions streaming failure are not the same bug." }, { "left": "issue:44291", "right": "issue:45070", "accept": false, - "reason": "init_empty_weights/_is_hf_initialized argument handling is unrelated to the PretrainedConfig pydantic field break." + "reason": "Unexpected _is_hf_initialized argument and a pydantic PretrainedConfig breakage are different regressions." }, { "left": "issue:43366", "right": "issue:45464", "accept": false, - "reason": "GGUF architecture support and Qwen3.5 streaming inference failures are unrelated." + "reason": "GPT-OSS GGUF architecture support and Qwen streaming inference are unrelated." }, { "left": "issue:45020", "right": "issue:45800", "accept": false, - "reason": "Remote_code breakage and torch version incompatibility are too broad and not the same defect." + "reason": "Generic remote_code breakage and torch-transformers incompatibility are different root causes." }, { "left": "issue:43531", "right": "issue:45070", "accept": false, - "reason": "A Qwen3-MoE sliding_window regression is unrelated to the PretrainedConfig/pydantic issue." + "reason": "Sliding_window handling and PretrainedConfig field regression do not look like the same bug." }, { "left": "issue:41093", "right": "issue:43596", "accept": false, - "reason": "Both are index errors during model setup, but the specific shapes, stacks, and affected models differ." + "reason": "Both are index errors with Deepspeed-related loading, but the affected models and failure shapes differ." }, { "left": "issue:43541", "right": "issue:44589", "accept": false, - "reason": "Grouped_mm tracing failure and Float8 storage lookup failure are different runtime bugs." + "reason": "Grouped_mm tracing failure and missing Float8 storage are different low-level errors." }, { "left": "issue:43054", "right": "issue:43994", + "accept": true, + "reason": "Both reports describe SigLIP2 producing bad/nonsensical text-embedding results across interfaces, which looks like the same underlying model bug." + }, + { + "left": "issue:43645", + "right": "issue:45399", "accept": false, - "reason": "Both concern SigLIP2 output quality, but one is about text embedding quality and the other about nonsensical pipeline results; not enough evidence of one concrete bug." + "reason": "Custom model initialization in notebooks and flash-attn fallback blocking are unrelated." }, { - "left": "issue:45542", - "right": "issue:45800", + "left": "issue:42915", + "right": "issue:45399", "accept": false, - "reason": "TensorBoard-without-TensorFlow backend detection and torch/transformers incompatibility are unrelated." + "reason": "FineGrainedFP8Config failure and flash-attn fallback logic are different issues." }, { - "left": "issue:45759", - "right": "issue:45902", + "left": "issue:45632", + "right": "issue:45698", "accept": false, - "reason": "One is about unwrapping text_config for composite multimodal configs; the other is an uninitialized-memory issue after to_empty(cuda). Different bugs." + "reason": "Both concern custom-module loading, but one is a leaf-directory cache collision and the other is wrong module loading after save_pretrained; not enough evidence they are the same bug." }, { "left": "issue:40990", "right": "issue:44960", "accept": false, - "reason": "Extremely high perplexity on gpt-oss and a GLM5 issue do not appear to share the same underlying defect." + "reason": "High perplexity on gpt-oss and GLM5 are unrelated model-specific regressions." }, { "left": "issue:39186", "right": "issue:44918", "accept": false, - "reason": "FSDP weight-shape failure and Qwen3.5 embedding unpacking under TRL SFT are different code paths." + "reason": "FSDP weight shape error and Qwen3.5 embedding unpacking are different failures." }, { "left": "issue:43716", "right": "issue:44918", "accept": false, - "reason": "Mistral3 image-preprocessor dtype mismatch and Qwen3.5 SFT embedding unpacking are unrelated." + "reason": "Gemma-3 image/text dtype mismatch and Qwen3.5 input embedding unpacking do not share a concrete code path." }, { "left": "issue:45632", "right": "issue:45698", - "accept": true, - "reason": "These look like the same custom-module loading bug: a cache/path collision causes from_pretrained to resolve the wrong custom module, and the second issue describes that symptom directly." + "accept": false, + "reason": "Both are about custom code loading, but the titles point to different mechanisms and different failure conditions." + }, + { + "left": "issue:43404", + "right": "issue:45005", + "accept": false, + "reason": "Tied lm_head in Mistral3 and translation-model tied-weights issues are not the same bug." + }, + { + "left": "issue:43645", + "right": "issue:45399", + "accept": false, + "reason": "Notebook custom-model initialization and flash-attn fallback behavior are unrelated." + }, + { + "left": "issue:44530", + "right": "issue:45005", + "accept": false, + "reason": "PagedAttentionCache linear_attention failure and tied-weights translation issues are unrelated." + }, + { + "left": "issue:42915", + "right": "issue:45399", + "accept": false, + "reason": "Qwen3Moe FP8 failure and flash-attn fallback gating are different issues." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. The only accepted soft pair is well supported by the titles (`trust_remote_code` cache/path collision vs. wrong custom module after `save_pretrained`), and the remaining pairs are reasonably rejected as distinct bugs or code paths. No obvious evidence overstatement." + "feedback": "The summary is grounded in the packet: the cluster is largely heterogeneous by model and failure mode, and the only accepted duplicate pair (43054/43994) is plausibly the same SigLIP2 regression based on the titles. The soft-edge rejections are generally conservative and well-justified." }, "evaluator_used": true, "retried": false, @@ -966,8 +930,8 @@ "issue:43828|issue:44928", "issue:36331|issue:41762", "issue:41628|issue:45800", - "issue:43866|issue:44403", "issue:43950|issue:45902", + "issue:43866|issue:44403", "issue:43872|issue:44589", "issue:43054|issue:45701", "issue:38175|issue:43054", @@ -987,14 +951,14 @@ "issue:43881|issue:43883", "issue:45685|issue:45823" ], - "prepared_review_unit_hash": "b1f01f172de00e66140a489508cf08431d1ddfdc209fa2f4519ced03c5355d76", + "prepared_review_unit_hash": "4168fccd8f77e3ede6f9154eb56cdbf01b1d47310a77866f4557ff249063b56e", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 48, @@ -1009,228 +973,216 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is mostly heterogeneous. The only clear duplicate cluster is the non-persistent-buffer regression around `from_pretrained()` in Transformers 5.x; a couple of SigLIP2 reports are related but not strong enough to merge confidently without more evidence.", + "summary": "This cluster is mostly heterogeneous. The only clear duplicate subgroup is the Transformers v5 non-persistent-buffer corruption regression; most other pairs are related by subsystem or model family only, not the same underlying bug.", "confidence": 0.86, - "canonical_issue_reason": "issue:43950 is the clearest, most specific report of the non-persistent-buffer regression and best represents the duplicate cluster with 43644/44534.", + "canonical_issue_reason": "Issue 43950 is the best canonical issue for the only true duplicate set here: it clearly describes the non-persistent-buffer corruption regression in v5, which 43644 and 44534 restate.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43950 is the strongest overall issue candidate: concrete symptom, clear regression framing, and directly captures the underlying bug class.", + "best_issue_reason": "43950 is the strongest standalone issue in the cluster because it is specific about the failure mode and regression, making it the best anchor for deduping the matching buffer-corruption reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43873", "right": "issue:45005", "accept": false, - "reason": "Different problems: quantization/offloading behavior vs tied-weights handling in v5." + "reason": "Different problems: quantization/offloading behavior vs tied-weights issues in translation models." }, { "left": "issue:44360", "right": "issue:44512", "accept": false, - "reason": "Unrelated: DSA indexer/ReLU discussion vs a docs wording issue." + "reason": "Unrelated: a DSA indexer bug versus a docs typo about a removed run command." }, { "left": "issue:44863", "right": "issue:45005", "accept": false, - "reason": "Different code paths: NemotronH checkpoint loading vs tied-weights regression." + "reason": "Different model-loading failure than the tied-weights v5 issue." }, { "left": "issue:44492", "right": "issue:44512", "accept": false, - "reason": "A typo in cache strategies is not the same as removed CLI command docs." + "reason": "Both are documentation-related, but they are distinct doc issues with different content." }, { "left": "issue:44918", "right": "issue:45507", "accept": false, - "reason": "Different models and failures: TRL SFT embedding unpacking vs GraniteMoE invalid method." + "reason": "Different code paths and failures: TRL embedding unpacking versus an invalid method call in GraniteMoEHybrid." }, { "left": "issue:43975", "right": "issue:44292", "accept": false, - "reason": "Distinct issues: detokenization bug vs Qwen-3 NVFP4 load/runtime error." + "reason": "Different runtime failures affecting different models; no clear shared root cause." }, { "left": "issue:43950", "right": "issue:44534", "accept": true, - "reason": "Same underlying regression: `from_pretrained()`/v5 fills non-persistent buffers with junk or corrupts them." + "reason": "Same underlying regression: v5/from_pretrained mishandles non-persistent buffers and leaves them as junk/corrupted." }, { "left": "issue:44292", "right": "issue:45507", "accept": false, - "reason": "Unrelated model-specific runtime errors." + "reason": "Distinct model-specific loading/runtime errors, not the same bug." }, { "left": "issue:39692", "right": "issue:43054", "accept": false, - "reason": "Both mention SigLIP2, but one is a broken docs example and the other is a model-quality complaint." + "reason": "Both mention SigLIP2, but one is a docs/example problem and the other is an embedding-quality complaint." }, { "left": "issue:43644", "right": "issue:43950", "accept": true, - "reason": "Same non-persistent-buffer regression in Transformers 5.x; wording differs but the bug is the same." + "reason": "Same bug: non-persistent buffers are being filled with junk/corrupted during v5 loading." }, { "left": "issue:43828", "right": "issue:44928", "accept": false, - "reason": "Different failures: autocast dtype mismatch vs 3D position_ids/flash-attn crash." + "reason": "Different issues: autocast dtype mismatch versus Qwen3.5 position-id/SDPA behavior." }, { "left": "issue:36331", "right": "issue:41762", "accept": false, - "reason": "Training API signature error is unrelated to Gemma3 ZeRO-3 loading failure." + "reason": "Completely different failures: trainer API mismatch versus ZeRO-3 loading index error." }, { "left": "issue:41628", "right": "issue:45800", "accept": false, - "reason": "Import/export issue vs torch/transformers version compatibility problem; not the same bug." + "reason": "Different bugs: missing import/export versus a torch/transformers compatibility problem." }, { - "left": "issue:43866", - "right": "issue:44403", + "left": "issue:43950", + "right": "issue:45902", "accept": false, - "reason": "Checkpoint corruption and load-time noise are distinct issues." + "reason": "Related buffer/memory symptoms, but not the same concrete bug or code path." }, { - "left": "issue:43950", - "right": "issue:45902", + "left": "issue:43866", + "right": "issue:44403", "accept": false, - "reason": "Related class of buffer/materialization bugs, but 45902 is specifically `inv_freq` after `to_empty`, not the same regression as generic non-persistent-buffer corruption." + "reason": "Checkpoint corruption and noisy loading warnings are not the same underlying issue." }, { "left": "issue:43872", "right": "issue:44589", "accept": false, - "reason": "Different incompatibilities: bitsandbytes constructor args vs missing float8 storage type." + "reason": "Different quantization/storage incompatibilities with different error surfaces." }, { "left": "issue:43054", "right": "issue:45701", "accept": false, - "reason": "SigLIP2 embedding quality vs tokenization changes; different subsystems and symptoms." + "reason": "SigLIP2 embedding quality versus a general tokenization-version change; no direct duplicate evidence." }, { "left": "issue:38175", "right": "issue:43054", "accept": false, - "reason": "Same model family, but zero probabilities and degraded text embeddings are not clearly the same underlying bug." + "reason": "Same model family, but the reported symptoms are too different to treat as one bug." }, { "left": "issue:41553", "right": "issue:43883", "accept": false, - "reason": "Tokenizer loading error message issue vs Molmo missing attribute during model loading." + "reason": "Unrelated loading/error-reporting issues for different models." }, { "left": "issue:44910", "right": "issue:45859", "accept": false, - "reason": "Qwen3.5 flash-attn crash due to packed-sequence interpretation is not the same as CPU-offload incompatibility in rotary embedding." + "reason": "Different Qwen3.5-related failures: flash-attn crash versus CPU-offload incompatibility." }, { "left": "issue:45663", "right": "issue:45902", "accept": false, - "reason": "Different model paths: Gemma4 shared_kv_state bug vs Qwen3_5Moe uninitialized `inv_freq` after materialization." + "reason": "Different Gemma4 vs Qwen3.5 code paths and failure modes." }, { "left": "issue:41553", "right": "issue:43881", "accept": false, - "reason": "Different models and failures: Voxtral tokenizer loading error vs glm-4v loading failure." + "reason": "Distinct model-loading problems with different causes." }, { "left": "issue:42947", "right": "issue:44387", "accept": false, - "reason": "Both are memory-related but one is gradient checkpointing/LoRA, the other is int4 reserved-memory/OOM regression." + "reason": "Both are memory-related, but one is LoRA checkpointing behavior and the other is int4 quantization OOM." }, { "left": "issue:43054", "right": "issue:43493", - "accept": true, - "reason": "Both are about the same SigLIP2 implementation discrepancy/regression; the JAX-vs-HF mismatch plausibly explains the degraded embeddings." + "accept": false, + "reason": "Closely related SigLIP2 complaints, but not clearly the same concrete bug." }, { "left": "issue:44403", "right": "issue:44451", "accept": false, - "reason": "Generic loading noise vs inability to load a specific checkpoint; not the same bug." + "reason": "Generic loading noise versus a specific model-loading failure; not the same issue." }, { "left": "issue:43643", "right": "issue:45020", "accept": false, - "reason": "Remote-code field loss in `AutoConfig` is different from a broader version-breakage report." + "reason": "Remote-code regressions are related, but one is missing fields in AutoConfig and the other is a broader compatibility complaint." }, { "left": "issue:43975", "right": "issue:45520", "accept": false, - "reason": "Detokenization bug and Python 3.13 flash-attn import issue are unrelated." + "reason": "Different failure classes: detokenization regression versus flash_attn import error." }, { "left": "issue:44403", "right": "issue:44488", "accept": false, - "reason": "Noise on load is not the same as a specific model failing to load." + "reason": "Both are model-loading problems, but they concern different models and likely different causes." }, { "left": "issue:42947", "right": "issue:43856", "accept": false, - "reason": "LoRA gradient checkpointing inefficiency is not the same as Qwen3 MoE memory use." + "reason": "Both mention memory, but one is gradient-checkpointing/LoRA and the other is MoE training efficiency." }, { "left": "issue:41762", "right": "issue:42915", "accept": false, - "reason": "Gemma3 ZeRO-3 loading failure and Qwen3Moe FP8 failure are separate model-specific issues." + "reason": "Different loading failures for different models/configurations." }, { "left": "issue:43782", "right": "issue:44960", "accept": false, - "reason": "Qwen3VL weight-only load error is unrelated to GLM5." + "reason": "Different models and unrelated errors." }, { "left": "issue:43881", "right": "issue:43883", "accept": false, - "reason": "Different failure surfaces: glm-4v loading vs missing `all_tied_weights_keys` on Molmo." + "reason": "Separate model-loading bugs with different error signatures." }, { "left": "issue:45685", "right": "issue:45823", "accept": false, - "reason": "MPS int histogram support issue is unrelated to Gemma4 device mismatch with `device_map=\"auto\"`." - }, - { - "left": "issue:45701", - "right": "issue:45823", - "accept": false, - "reason": "Tokenization differences and Gemma4 device placement bugs are unrelated." - }, - { - "left": "issue:44910", - "right": "issue:45859", - "accept": false, - "reason": "Flash-attn packed-sequence crash and CPU-offload rotary embedding incompatibility are different bugs." + "reason": "Different issues: MPS histogram support versus Gemma4 device placement mismatch." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary stays grounded in the packet and remains conservative overall. The non-persistent-buffer regression cluster is well supported, and the SigLIP2 linkage is stated cautiously enough. The accepted soft-edge pairs (43950/44534, 43644/43950, and 43054/43493) are plausible from the issue titles and do not overstate the evidence." + "feedback": "Grounded and conservative overall. The only clear duplicate subgroup is the Transformers v5 non-persistent-buffer corruption regression centered on issue 43950, and the other soft-edge calls are phrased cautiously and match the packet titles/body themes without overclaiming. " }, "evaluator_used": true, "retried": false, @@ -1335,241 +1287,235 @@ "issue:43366|issue:43828", "issue:45663|issue:45823" ], - "prepared_review_unit_hash": "68937ac43f461e540dc12503b3842e2a45ed8849d48d98fff2e79ffe7700f8c4", + "prepared_review_unit_hash": "ee69c055fddf0a4a383ec368cd53ce1436ce65428fcad8bb753988d6a409af46", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 33, - "serialized_chars": 38123, - "estimated_input_tokens": 9531, - "estimated_eval_tokens": 19318 + "serialized_chars": 38120, + "estimated_input_tokens": 9530, + "estimated_eval_tokens": 19316 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a loose similarity cluster of transformer loading/regression reports; only a few pairs look like true duplicates, with the strongest match in the docs cleanup subset.", - "confidence": 0.62, - "canonical_issue_reason": "Issue 45020 is the broadest, clearest umbrella for the cluster\u2019s dominant theme of recent model/tokenizer loading regressions, especially remote_code-related breakage.", + "summary": "The cluster is mostly a loose family of model-loading/regression reports, not one tight duplicate set. A few soft edges are solid duplicates, especially the docs pipeline-removal pair, the `_is_hf_initialized` / `init_empty_weights` compatibility pair, the Qwen VL video regression pair, and the mask-shape IndexError pair for 41093/44805. The 36331/41093 link should be rejected: those are different bugs.", + "confidence": 0.72, + "canonical_issue_reason": "issue:45020 is the broadest representative issue in the cluster: it captures the recent-transformers/model-loading regression theme that many of the narrower reports orbit, even though the cluster is not a single exact-duplicate family.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45020 is the most representative anchor for the cluster because it is broad, concrete, and plausibly subsumes several of the narrower loading-failure reports.", + "best_issue_reason": "issue:45020 is the best anchor because it is the most umbrella-like and representative of the cluster\u2019s shared regression pattern, rather than one narrow symptom report.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:42491", "right": "issue:44368", "accept": false, - "reason": "Related Qwen/LoRA wording, but one is a training incompatibility across branches and the other is a warning about `tie_word_embeddings`; not the same bug." + "reason": "Related Qwen3/LoRA themes, but one is a cross-version LoRA compatibility break and the other is a tie-word-embeddings warning; different failure modes." }, { "left": "issue:43824", "right": "issue:43872", "accept": false, - "reason": "Different failure modes: a Qwen2.5-VL import error vs a bitsandbytes constructor-kwarg incompatibility." + "reason": "Missing Qwen2_5_VL import vs bitsandbytes `Int8Params.__new__` incompatibility; unrelated code paths." }, { "left": "issue:45020", "right": "issue:45823", "accept": false, - "reason": "Broad remote_code loading regressions vs a Gemma4 device-map forward mismatch; not the same code-path." + "reason": "Broad remote_code/model-loading regression vs Gemma4 PLE device mismatch during forward; not the same bug." }, { "left": "issue:44991", "right": "issue:45399", "accept": false, - "reason": "Tokenizer loading failure for one model vs flash-attn fallback gating; different subsystems and fixes." + "reason": "Tokenizer load failure for one model vs flash-attention fallback gating; same broad area, different root cause." }, { "left": "issue:44661", "right": "issue:45020", "accept": false, - "reason": "Both involve model setup/loading, but one is a tokenizer-mapping edge case and the other is a broad remote_code regression." + "reason": "Tokenizer-mapping edge case in `add-new-model-like` vs broad remote_code regression; different defects." }, { "left": "issue:43828", "right": "issue:45823", "accept": false, - "reason": "Different runtime errors: autocast dtype mismatch vs device placement mismatch in Gemma4." + "reason": "Autocast dtype mismatch in Phi MoE vs Gemma4 PLE device placement mismatch; distinct bugs." }, { "left": "issue:45081", "right": "issue:45399", "accept": false, - "reason": "Tokenizer regex crash vs flash-attn fallback selection; unrelated concrete bugs." + "reason": "Tokenizer regex patch crash vs flash-attn fallback blocked by other checks; different subsystems." }, { "left": "issue:43827", "right": "issue:44509", "accept": true, - "reason": "Both describe the same v5 docs regression: deprecated pipeline task names remain referenced and should be removed/updated." + "reason": "Same docs regression after v5 pipeline removals; both report stale references to removed summarization/translation pipeline tasks." }, { "left": "issue:45020", "right": "issue:45356", "accept": false, - "reason": "Remote_code loading breakage is broader and unrelated to the Kimi-K2.5 tokenizer codec regression." + "reason": "General remote_code breakage vs Kimi-K2.5 tokenizer codec / regex-warning regression; different failure paths." }, { "left": "issue:43404", "right": "issue:43950", "accept": false, - "reason": "Weight tying in Mistral3 is a model-architecture bug, while non-persistent buffer corruption is a different `from_pretrained()` regression." + "reason": "Weight-tying bug in Mistral3 vs non-persistent buffer corruption on `from_pretrained()`; different loading mechanisms." }, { "left": "issue:43742", "right": "issue:45399", "accept": false, - "reason": "A MobileLLM load/key error is unrelated to the flash-attn fallback path." + "reason": "MobileLLM key error vs flash-attention fallback gating; both load-time issues, but not the same defect." }, { "left": "issue:42915", "right": "issue:45020", "accept": false, - "reason": "Qwen3 MoE + accelerate/deepspeed weight loading issue is much narrower than the general remote_code regression." + "reason": "Specific Qwen3 MoE FP8 config failure vs broad remote_code/model-loading regression; too different in scope and cause." }, { "left": "issue:43638", "right": "issue:45823", "accept": false, - "reason": "Different DeepSpeed and model-architecture failures; no shared underlying bug." + "reason": "ZeRO-3 index error on a Bert model vs Gemma4 PLE device mismatch; unrelated root causes." }, { "left": "issue:43257", "right": "issue:43866", "accept": false, - "reason": "Both are model-specific loading problems, but one is Qwen3 MoE weight conversion and the other is an Ovis2 checkpoint corruption report." + "reason": "Qwen3 MoE weight conversion under accelerate+deepspeed vs corrupted Ovis2 checkpoint; not the same bug." }, { "left": "issue:43792", "right": "issue:44220", - "accept": true, - "reason": "Both point to the same Whisper frontend failure in `_torch_extract_fbank_features`, with one report surfaced through `whisper-large-v2`." + "accept": false, + "reason": "Whisper runtime failure vs `_torch_extract_fbank_features()` issue; both audio-related but distinct." }, { "left": "issue:43872", "right": "issue:44291", "accept": true, - "reason": "Same `_is_hf_initialized` constructor mismatch during meta/init-empty loading; one is a bitsandbytes manifestation of the same kwarg regression." + "reason": "Same compatibility regression around `init_empty_weights` / `_is_hf_initialized` being passed into `Int8Params` and breaking model loading." }, { "left": "issue:42915", "right": "issue:43872", "accept": false, - "reason": "Both touch loading, but one is Qwen3 MoE weight conversion and the other is a constructor incompatibility involving bitsandbytes." + "reason": "Qwen3 MoE FP8 failure and bitsandbytes parameter-construction incompatibility are different loading/configuration bugs." }, { "left": "issue:43940", "right": "issue:45313", "accept": false, - "reason": "Both are DeepSpeed ZeRO-3 load failures, but the affected model paths and missing-weight symptoms differ." + "reason": "Both are DeepSpeed ZeRO-3 loading issues, but they describe different Qwen3-family failures and different missing-weight scenarios." }, { "left": "issue:41553", "right": "issue:44991", "accept": false, - "reason": "Tokenizer load issues, but one is a bad error message for Voxtral and the other is an actual tokenizer failure for EMBEDDIA/est-roberta." + "reason": "Bad tokenizer error for Voxtral vs tokenizer load failure for EMBEDDIA/est-roberta; same area, different bug." }, { "left": "issue:43854", "right": "issue:44863", "accept": false, - "reason": "A unit-test model-loading problem and a NemotronH checkpoint compatibility issue are not the same bug." + "reason": "GLM-4.7-Flash unit-test loading issue vs NemotronH checkpoint-loading failure; different model and code path." }, { "left": "issue:36331", "right": "issue:41093", "accept": false, - "reason": "Both are trainer/runtime exceptions, but they involve different call sites and different root causes." + "reason": "`compute_loss()` got unexpected `num_items_in_batch` TypeError vs mask-shape IndexError; the titles and symptoms do not support the same bug." }, { "left": "issue:43257", "right": "issue:45305", "accept": false, - "reason": "Qwen3 MoE weight conversion under accelerate/deepspeed is unrelated to gradient averaging under GAS with `model_accepts_loss_kwargs=True`." + "reason": "Missing weight conversion under DeepSpeed vs gradient averaging not happening with GAS and `model_accepts_loss_kwargs=True`; different behaviors." }, { "left": "issue:42947", "right": "issue:45923", "accept": false, - "reason": "Gradient checkpointing with PEFT LoRA is unrelated to a missing `supports_gradient_checkpointing` flag on a different trust-remote-code model." + "reason": "Both mention gradient checkpointing, but one is ineffective PEFT LoRA checkpointing and the other is a missing capability flag on a trust_remote_code model." }, { "left": "issue:44928", "right": "issue:45588", "accept": false, - "reason": "RLHF NaNs from position-id/SDPA behavior are unrelated to the `s_aux=None` flash-attention crash." + "reason": "RLHF gradient explosion from 3D position_ids / SDPA fallback vs flash_attention crash on `s_aux=None`; different bugs." }, { "left": "issue:43872", "right": "issue:45538", "accept": false, - "reason": "Constructor kwarg incompatibility vs CLIPTokenizer max-length default; different issues entirely." + "reason": "bitsandbytes parameter-construction regression vs CLIPTokenizer max-length sentinel; unrelated." }, { "left": "issue:41093", "right": "issue:44805", - "accept": false, - "reason": "Same broad error pattern, but the report details indicate different examples and likely different underlying causes." + "accept": true, + "reason": "Same mask-shape IndexError, just with slightly different tensor sizes and wording." }, { "left": "issue:41553", "right": "issue:44488", "accept": false, - "reason": "Both are tokenizer-loading complaints, but the affected models and reported failure mode are too different to treat as one bug." + "reason": "Both are loading-related, but one is a Voxtral tokenizer error and the other is a separate failure loading `cjvt/sleng-bert`." }, { "left": "issue:44479", "right": "issue:44560", - "accept": false, - "reason": "Both are video-input regressions, but one is a broader model-family input bug and the other is a specific Qwen3-VL embedding StopIteration case." + "accept": true, + "reason": "Both describe the same v5.3.0 video-input regression in the Qwen VL family; `44560` looks like a concrete manifestation of the broader bug." }, { "left": "issue:44530", "right": "issue:44863", "accept": false, - "reason": "PagedAttentionCache linear_attention crash and NemotronH checkpoint loading are unrelated." + "reason": "PagedAttentionCache `linear_attention` crash vs NemotronH checkpoint-loading failure; unrelated code paths." }, { "left": "issue:24643", "right": "issue:30064", "accept": false, - "reason": "DeepSpeed training weight-shape error vs segmentation-map processing failure; no shared underlying bug." + "reason": "DeepSpeed training weight-shape RuntimeError vs image-processor void-segmentation-map failure; no shared root cause." }, { "left": "issue:45325", "right": "issue:45902", "accept": false, - "reason": "Both concern Qwen-family internals, but one is a RoPE scaling issue and the other is uninitialized memory after meta-to-empty materialization." + "reason": "Qwen2.5-VL rope-index scaling bug vs uninitialized-memory after `to_empty(cuda)` materialization; different internals." }, { "left": "issue:43366", "right": "issue:43828", "accept": false, - "reason": "GGUF architecture support and autocast dtype mismatch are different problems." - }, - { - "left": "issue:45663", - "right": "issue:45823", - "accept": false, - "reason": "Both involve Gemma4, but one is an FSDP2 shared-state bug and the other is a device-map placement mismatch." + "reason": "GGUF gpt-oss support request vs Phi MoE autocast dtype mismatch; not the same bug." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The summary stays grounded in the packet and is appropriately cautious about the cluster being loose. The accepted soft-edge matches are the ones that are most directly supported by the titles, and the rejected pairs are framed conservatively without overclaiming duplicates." + "accept": false, + "feedback": "Reject: one accepted soft-edge verdict is not grounded in the packet. `issue:36331` is a `compute_loss()` unexpected `num_items_in_batch` TypeError, while `issue:41093` is a mask-shape IndexError; the titles do not support treating them as the same bug. Because of that, the summary also overstates the cluster\u2019s duplicate strength by including this pair among the strongest matches. The other acceptances (`43827`/`44509`, `43872`/`44291`, `44479`/`44560`, `41093`/`44805`) are more conservative and plausible." }, "evaluator_used": true, - "retried": false, + "retried": true, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -1668,208 +1614,208 @@ "issue:43708|issue:44075", "issue:43854|issue:43931" ], - "prepared_review_unit_hash": "891d64d76f162324b8ad66df1b8f80a67231a581142057fda2f26bf7b619f7a2", + "prepared_review_unit_hash": "98cae0867a07c7d3aa5eae99f04da9e45b7e5fa2653a3b1d5efedc3f3229c6c5", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 31, - "serialized_chars": 36268, - "estimated_input_tokens": 9067, - "estimated_eval_tokens": 18390 + "serialized_chars": 36262, + "estimated_input_tokens": 9066, + "estimated_eval_tokens": 18388 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is mostly heterogeneous and only a few soft pairs look like true duplicates. Clear duplicate groups are the transformers v5 non-persistent-buffer regression, the custom-model initialization regression, and the SigLIP2 output corruption reports. Most other pairs are merely same-subsystem or same-model-family false positives.", + "summary": "This set is mostly heterogeneous, but a few soft pairs look like true duplicates: the non-persistent buffer regression, the custom model initialization regression, and the SigLIP2 output regression. The rest are too model- or symptom-specific to merge safely.", "confidence": 0.83, - "canonical_issue_reason": "issue:43644 is the clearest, broadest regression report in the cluster and is duplicated by issue:44534; it names the concrete transformers v5 buffer-corruption bug directly.", + "canonical_issue_reason": "Issue 43644 is the clearest canonical issue among the near-duplicates: 44534 reports the same Transformers 5.x non-persistent-buffer corruption regression with essentially the same symptom.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43644 is the best representative issue because it captures the underlying regression at a general code-path level, making it the strongest anchor for deduping the related reports.", + "best_issue_reason": "Issue 43644 is the strongest issue to keep as the representative tracker because it is concrete, broadly applicable, and has an exact near-duplicate.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44155", "right": "issue:45584", "accept": false, - "reason": "Both are audio-generation bugs, but they affect different models and different failure modes; not the same underlying issue." + "reason": "Different models and failure modes: batched AudioFlamingo track leakage vs Whisper empty-transcription generation." }, { "left": "issue:43366", "right": "issue:43531", "accept": false, - "reason": "Both mention Qwen/GGUF-style model behavior, but the bug reports are unrelated and target different code paths." + "reason": "Both mention Qwen/GGUF-style model issues, but they describe different concrete bugs and code paths." }, { "left": "issue:43950", "right": "issue:45834", "accept": false, - "reason": "Non-persistent-buffer corruption and a Kosmos2.5 OCR index error are unrelated bugs." + "reason": "Non-persistent buffer corruption is unrelated to Kosmos2.5 OCR index errors." }, { "left": "issue:43550", "right": "issue:44910", "accept": false, - "reason": "Both involve attention kernels, but the model, trigger, and failure mode differ too much to be the same bug." + "reason": "Different regressions: Bamba torch.compile/SDPA vs Qwen3.5 flash-attention position_ids handling." }, { "left": "issue:43824", "right": "issue:43957", "accept": false, - "reason": "Importing a specific class and meta-device loading failures are distinct model-loading problems." + "reason": "Importing a missing class and meta-device loading failures are distinct problems." }, { "left": "issue:43645", "right": "issue:43646", "accept": true, - "reason": "These describe the same Transformers 5 custom-model initialization regression, with one being a notebook-specific manifestation." + "reason": "These read like the same custom-model initialization regression in Transformers 5.0.0, with the notebook mention just a narrower manifestation." }, { "left": "issue:42915", "right": "issue:43957", "accept": false, - "reason": "Qwen3MoE FP8 failure and meta-device loading issues are different defects." + "reason": "Qwen3Moe FP8 failure and generic meta-device loading issues are different bugs." }, { "left": "issue:43366", "right": "issue:44589", "accept": false, - "reason": "GGUF support for gpt-oss is unrelated to the Float8 storage error." + "reason": "GGUF support request and Float8 storage lookup failure are unrelated." }, { "left": "issue:44560", "right": "issue:44918", "accept": false, - "reason": "Both touch Qwen3.5, but one is a video StopIteration bug and the other is an SFT embedding-unpacking failure." + "reason": "Both involve Qwen3.x, but one is a video StopIteration error and the other is an input-embedding unpacking failure in TRL." }, { "left": "issue:42617", "right": "issue:43366", "accept": false, - "reason": "A 3d_parallel.py runtime issue is unrelated to GGUF model support." + "reason": "3d_parallel.py execution failure is not the same as GGUF architecture support." }, { "left": "issue:36010", "right": "issue:42915", "accept": false, - "reason": "GenerationMixin import failure and Qwen3MoE FP8 loading failure are separate issues." + "reason": "GenerationMixin import error and Qwen3Moe FP8 failure are unrelated." }, { "left": "issue:44661", "right": "issue:45823", "accept": false, - "reason": "Tokenizer-mapping validation and Gemma4 device-mismatch behavior are different bugs." + "reason": "Tokenizer mapping/model registration logic is unrelated to Gemma4 PLE device placement during forward." }, { "left": "issue:38175", "right": "issue:43994", "accept": true, - "reason": "Both report the same SigLIP2 model producing bad outputs; the later issue is the same underlying problem with a different description." + "reason": "Same underlying SigLIP2 regression: both report incorrect/degenerate outputs for google/siglip2-base-patch16-224 with AutoModel/pipeline usage." }, { "left": "issue:43632", - "right": "issue:44534", + "right": "issue:45834", "accept": false, - "reason": "Both are Transformers v5 regressions, but one is about _is_hf_initialized and the other about non-persistent buffers; distinct root causes." + "reason": "HF initialization flag regression and Kosmos2.5 OCR index errors are not the same bug." }, { - "left": "issue:44960", - "right": "issue:44977", + "left": "issue:44560", + "right": "issue:44805", "accept": false, - "reason": "Different model families and failure modes, despite both being generation issues." + "reason": "Different errors and model paths: video StopIteration vs mask/tensor shape mismatch." }, { - "left": "issue:42491", - "right": "issue:43720", + "left": "issue:43638", + "right": "issue:44661", "accept": false, - "reason": "Qwen3-MoE LoRA compatibility and BitNet packed-weight loading are unrelated." + "reason": "Zero3/Bert indexing failure is unrelated to tokenizer mapping registration constraints." }, { - "left": "issue:45313", - "right": "issue:45357", + "left": "issue:43596", + "right": "issue:44661", "accept": false, - "reason": "One is ZeRO-3 weight loading for language_model, the other is save_pretrained key serialization for visual encoder; not the same bug." + "reason": "A BertModel Zero3 init IndexError is not the same as the TOKENIZER_MAPPING_NAMES add-new-model-like failure." }, { - "left": "issue:41720", - "right": "issue:42915", + "left": "issue:44661", + "right": "issue:44805", "accept": false, - "reason": "Different Qwen3-related failures: device-mapping assert vs FP8 config failure." + "reason": "Tokenizer mapping validation and mask-shape IndexErrors are different code paths." }, { - "left": "issue:44451", - "right": "issue:44488", + "left": "issue:44877", + "right": "issue:45030", "accept": false, - "reason": "Both are model loading failures, but the reports are for different BERT models without evidence of a shared root cause." + "reason": "Both mention config validation, but they are distinct model/config failures with different triggers." }, { - "left": "issue:43644", - "right": "issue:44534", - "accept": true, - "reason": "These are the same Transformers v5 non-persistent-buffer corruption regression, phrased as identical symptoms." + "left": "issue:43317", + "right": "issue:44756", + "accept": false, + "reason": "GPU/CPU offload dequantization failure and Strix Halo mmap OOM are different loading problems." }, { - "left": "issue:44560", - "right": "issue:44805", + "left": "issue:38175", + "right": "issue:39692", "accept": false, - "reason": "A Qwen3.5 video-generation error and a mask-shape IndexError are not the same bug." + "reason": "Same model family, but one is a degenerate output bug and the other is a documentation example with model/processor mismatch and quantization failure." }, { - "left": "issue:43638", - "right": "issue:44661", - "accept": false, - "reason": "Zero3 training on non-pretrained BertModel and add-new-model-like tokenizer mapping failure are unrelated." + "left": "issue:43644", + "right": "issue:44534", + "accept": true, + "reason": "Near-exact duplicate: both report Transformers 5.x corrupting non-persistent buffers; 44534 is essentially a reworded repeat." }, { - "left": "issue:43596", - "right": "issue:44661", + "left": "issue:44960", + "right": "issue:44977", "accept": false, - "reason": "DeepSpeed ZeRO-3 BertModel index error is unrelated to tokenizer mapping validation." + "reason": "GLM5 issue and Qwen3.5 flash-attention generation failure are unrelated." }, { - "left": "issue:44661", - "right": "issue:44805", + "left": "issue:42491", + "right": "issue:43720", "accept": false, - "reason": "Tokenizer mapping validation and a tensor mask-shape IndexError are distinct defects." + "reason": "LoRA/Qwen3_moe hf4.x compatibility and BitNet packed-weight loading are separate regressions." }, { - "left": "issue:44877", - "right": "issue:45030", + "left": "issue:45313", + "right": "issue:45357", "accept": false, - "reason": "Strict config blocking granite_speech and glm4v config validation are separate config issues." + "reason": "ZeRO-3 language-model weight loading and save_pretrained visual-encoder key serialization are different problems." }, { - "left": "issue:43317", - "right": "issue:44756", + "left": "issue:41720", + "right": "issue:42915", "accept": false, - "reason": "Both concern loading/offload behavior, but one is dequantized-model offload and the other is mmap OOM on Strix Halo." + "reason": "Qwen3 auto device mapping cuda assert and Qwen3Moe FP8 failure are not the same bug." }, { - "left": "issue:38175", - "right": "issue:39692", + "left": "issue:44451", + "right": "issue:44488", "accept": false, - "reason": "Both are SigLIP2 reports, but one is inference output corruption and the other is documentation/example errors with processor/model mismatch and quantization." + "reason": "Both are BERT model-loading complaints, but the titles point to different models and no shared concrete failure is evident." }, { "left": "issue:43854", "right": "issue:43931", "accept": false, - "reason": "Both are model-loading errors, but for different models and different underlying mismatch conditions." + "reason": "GLM-4.7-Flash unit-test loading and Qwen3-VL-30B-A3B weight-shape mismatch are different model-specific issues." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet, and the soft-edge calls are mostly conservative. The duplicate anchors around 43644/44534 and 43645/43646 are well supported by the titles, and the SigLIP2 pair is plausible from the packet alone. If anything, the narrative around the SigLIP2 group should stay slightly softer (e.g. \u201clikely duplicate\u201d rather than fully \u201cclear\u201d), but this does not warrant rejection." + "feedback": "The judgments are mostly conservative and well grounded in the packet. The accepted soft pairs are plausible same-bug matches based on closely aligned titles and symptoms, especially 43644/44534, 43645/43646, and 38175/43994. The rejected pairs generally stay on the safe side by requiring more than superficial model-family overlap. The canonical issue choice and summary are supported by the evidence provided." }, "evaluator_used": true, "retried": false, @@ -1974,238 +1920,238 @@ "issue:45276|issue:45292", "issue:45684|issue:45698" ], - "prepared_review_unit_hash": "c04bdbe724f929fec665544ea8b20afd12f59dcf9973eb561a8152fb12bc7e12", + "prepared_review_unit_hash": "813abd841e93cbee47f5e7109ee5c44197c7a16f99b69cbfc67ba7fdb9494fc3", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 33, - "serialized_chars": 37770, - "estimated_input_tokens": 9443, - "estimated_eval_tokens": 19142 + "serialized_chars": 37775, + "estimated_input_tokens": 9444, + "estimated_eval_tokens": 19144 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly unrelated closed issues spanning tokenizer, quantization, model-loading, and memory regressions. The only clear duplicate set is the resize_token_embeddings regression across Gemma/T5Gemma variants, centered on issue 45276.", - "confidence": 0.92, - "canonical_issue_reason": "Issue 45276 is the best representative because it describes the broader resize_token_embeddings failure and names multiple affected outputs (embed_tokens_per_layer and output_embeddings), which also subsumes the closely related Gemma/T5Gemma reports.", + "summary": "The set is mostly a loose collection of unrelated Transformers issues. The only clear duplicate-like cluster is around `resize_token_embeddings` not propagating to nested/output embeddings across related model variants.", + "confidence": 0.69, + "canonical_issue_reason": "`issue:45292` is the most generic and reusable statement of the shared `resize_token_embeddings` regression, so it best serves as the representative issue for the small embedding-resize cluster.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45276 is the strongest global issue for this cluster because it most completely captures the shared root bug behind the resize_token_embeddings duplicates, while the rest of the items are largely unrelated.", + "best_issue_reason": "`issue:45292` is the broadest, least model-specific report among the candidates, and it cleanly subsumes the more specific Gemma4/T5Gemma variants.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45488", "right": "issue:45812", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different model families and different failure modes (pre-tokenizer override vs wrong token IDs)." + "reason": "Different tokenizer regressions in different model families; same broad area, not the same bug." }, { "left": "issue:39692", "right": "issue:43994", "accept": false, - "reason": "Both mention SigLIP2, but one is a doc example/quantization problem and the other is a bad inference result with AutoModel/pipeline; not the same bug." + "reason": "Both involve SigLIP2, but one is a docs/example problem and the other is a model loading/output issue." }, { "left": "issue:43257", "right": "issue:43854", "accept": false, - "reason": "Different models and different failure classes: Qwen3 MoE weight conversion vs GLM-4.7-Flash loading in tests." + "reason": "Different models and failure modes during loading; no concrete shared code-path is evident." }, { "left": "issue:44488", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer-loading complaints, but they concern different checkpoints and do not show the same concrete code-path failure." + "reason": "Both are tokenizer-loading complaints, but for different models and likely different root causes." }, { "left": "issue:42491", "right": "issue:43931", "accept": false, - "reason": "Different Qwen3-related problems: LoRA compatibility across versions vs a weight-shape mismatch when loading a VL model." + "reason": "Both mention Qwen3 loading issues, but one is LoRA/version compatibility and the other is a shape mismatch in a different model." }, { "left": "issue:43644", "right": "issue:43645", "accept": false, - "reason": "Both are Transformers 5 regressions, but one is about junk non-persistent buffers and the other about notebook-defined custom models." + "reason": "Distinct regressions: junk-filled buffers vs Jupyter custom-model initialization." }, { "left": "issue:43856", "right": "issue:43873", "accept": false, - "reason": "Both mention memory, but one is MoE training efficiency and the other is quantization/offloading behavior." + "reason": "Both are memory-related, but one is training memory efficiency and the other is offload behavior under quantization." }, { "left": "issue:45362", "right": "issue:45406", "accept": false, - "reason": "Different model/code paths: Qwen3.5 chat crashes vs Gemma4Processor missing _tokenizer in transformers serve." + "reason": "Different models and different crash sites; not the same bug." }, { "left": "issue:43856", "right": "issue:44387", "accept": false, - "reason": "Both are memory-related, but one is training inefficiency and the other is int4 CUDA reserved-memory/OOM regression." + "reason": "Both mention memory, but the reported conditions and failure modes differ materially." }, { "left": "issue:44910", "right": "issue:45588", "accept": false, - "reason": "Different flash-attention failures with different triggers and different offending inputs." + "reason": "Both touch flash attention, but the failures are unrelated: packed-sequence misread vs missing `s_aux` handling." }, { "left": "issue:42491", "right": "issue:45292", "accept": false, - "reason": "LoRA/Qwen3 MoE version incompatibility is unrelated to resize_token_embeddings behavior." + "reason": "A Qwen3 MoE LoRA compatibility issue is unrelated to embedding resizing behavior." }, { "left": "issue:45276", "right": "issue:45335", "accept": true, - "reason": "Same underlying bug: resize_token_embeddings does not propagate to tied/output embeddings across related model variants." + "reason": "Both report `resize_token_embeddings` failing to update nested model embeddings; these look like the same underlying regression across closely related model variants." }, { "left": "issue:44164", "right": "issue:44291", "accept": false, - "reason": "Different save/load problems: extra_state handling vs init_empty_weights/_is_hf_initialized TypeError." + "reason": "Separate serialization/loading bugs with different triggers and error types." }, { "left": "issue:36010", "right": "issue:39404", "accept": false, - "reason": "ImportError for GenerationMixin and Whisper return_language pipeline regression are unrelated." + "reason": "Completely different import vs Whisper pipeline behavior issues." }, { "left": "issue:43404", "right": "issue:43454", "accept": false, - "reason": "Both concern untied lm_head weights, but they target different models and do not look like the same exact bug." + "reason": "Both are untied `lm_head` reports, but they target different model families and there isn't enough evidence of one shared code-path bug." }, { "left": "issue:41720", "right": "issue:44155", "accept": false, - "reason": "Qwen3 auto device mapping CUDA assert and AudioFlamingo3 batched track leakage are unrelated." + "reason": "Different subsystems and symptoms: device-mapping crash vs batched inference token leak." }, { "left": "issue:39692", "right": "issue:44960", "accept": false, - "reason": "SigLIP2 doc-example errors are unrelated to the GLM5 issue." + "reason": "Unrelated model-specific issue vs generic GLM5 loading report." }, { "left": "issue:43425", "right": "issue:43957", "accept": false, - "reason": "Torch version incompatibility is unrelated to meta-device model loading regressions." + "reason": "Torch version compatibility and `meta`-device loading are different problems." }, { "left": "issue:44402", "right": "issue:44488", "accept": false, - "reason": "Different tokenizer/vocab issues affecting different models." + "reason": "Tokenizer vocab mismatch is not the same as a model loading failure." }, { "left": "issue:42947", "right": "issue:45704", "accept": false, - "reason": "Gradient checkpointing with PEFT LoRA is unrelated to the T5 apex RMSNorm memory leak." + "reason": "Gradient checkpointing/LoRA and an apex RMSNorm leak are unrelated." }, { "left": "issue:45663", "right": "issue:45759", "accept": false, - "reason": "Gemma4 FSDP2 shared_kv_states and Qwen composite config unwrapping are different failures." + "reason": "Different model architecture bugs: FSDP/shared KV state vs composite config unwrapping." }, { "left": "issue:43873", "right": "issue:44387", "accept": false, - "reason": "Quantization offloading behavior and CUDA reserved-memory/OOM are different bugs." + "reason": "Both concern memory, but one is offloading behavior and the other is int4 reserved-memory growth." }, { "left": "issue:43335", "right": "issue:45663", "accept": false, - "reason": "SwitchTransformers sparse-layer config bug is unrelated to Gemma4 FSDP2 attention state rebuilding." + "reason": "Sparse-layer construction bug is unrelated to Gemma4 attention state rebuilding." }, { "left": "issue:45663", "right": "issue:45859", "accept": false, - "reason": "Different Qwen/Gemma issues with different execution paths (FSDP2 attention vs CPU offload rotary embedding)." + "reason": "Different Qwen3/Gemma4 issues with different execution paths." }, { "left": "issue:44291", "right": "issue:44756", "accept": false, - "reason": "init_empty_weights argument handling is unrelated to disabling mmap on Strix Halo." + "reason": "`init_empty_weights` argument handling and Strix Halo mmap OOM are unrelated." }, { "left": "issue:44155", "right": "issue:45823", "accept": false, - "reason": "AudioFlamingo3 track leakage and Gemma4 PLE device mismatch are unrelated." + "reason": "Different model families and different failure modes." }, { "left": "issue:44292", "right": "issue:44912", "accept": false, - "reason": "Different quantized-model loading failures with different models and symptoms." + "reason": "Different quantized-model loading failures with no clear shared root cause." }, { "left": "issue:44387", "right": "issue:44756", "accept": false, - "reason": "Both mention OOM, but the causes are different and not mergeable into one fix." + "reason": "Both are OOM-related, but one is quantization memory growth and the other is platform-specific mmap behavior." }, { "left": "issue:43854", "right": "issue:45216", "accept": false, - "reason": "GLM-4.7 loading and Qwen3.5 save_pretrained regression are unrelated." + "reason": "Different regressions: GLM-4.7 unit-test loading vs Qwen3.5 checkpoint save correctness." }, { "left": "issue:44743", "right": "issue:44960", "accept": false, - "reason": "Qwen3 recurrent-state cache reset and GLM5 are unrelated issues." + "reason": "Recurrent-state cache behavior and a GLM5 loading issue are unrelated." }, { "left": "issue:43957", "right": "issue:44756", "accept": false, - "reason": "Meta-device model loading regressions and Strix Halo mmap OOM are different problems." + "reason": "`meta`-device model loading and Strix Halo OOM mitigation are different issues." }, { "left": "issue:45276", "right": "issue:45292", "accept": true, - "reason": "Same root regression: resize_token_embeddings fails to update output embeddings/tied weights, with one report adding a model-specific detail." + "reason": "Same `resize_token_embeddings` regression; one is a model-specific report and the other is the generic version of the same bug." }, { "left": "issue:45684", "right": "issue:45698", "accept": false, - "reason": "Both concern custom-model save/load, but one is about permissions propagation and the other about loading the wrong module." + "reason": "One is file permission propagation in `save_pretrained`, the other is wrong custom-module resolution on reload." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. It correctly isolates the resize_token_embeddings duplicates (45276/45292/45335) while treating the rest as mostly unrelated, and the soft-edge verdicts do not overclaim beyond the evidence in the issue titles/descriptions." + "feedback": "The summary is grounded in the packet and stays conservative. The representative issue choice for the resize_token_embeddings cluster is supported by the titles, and the accepted 45276/45335 pair looks like the same bugfix area and mergeable into one PR. The remaining soft-edge rejections are phrased cautiously and do not overstate shared root cause beyond what the packet shows." }, "evaluator_used": true, "retried": false, @@ -2280,8 +2226,8 @@ "issue:44164|issue:44756", "issue:42947|issue:43381", "issue:44479|issue:44877", - "issue:43646|issue:43708", "issue:45588|issue:45663", + "issue:43646|issue:43708", "issue:45292|issue:45335", "issue:43645|issue:44164", "issue:43257|issue:44863", @@ -2291,8 +2237,8 @@ "issue:44589|issue:44756", "issue:44683|issue:45823", "issue:39404|issue:45701", - "issue:43425|issue:45823", "issue:43906|issue:44530", + "issue:43425|issue:45823", "issue:43611|issue:43645", "issue:43425|issue:43716", "issue:44530|issue:45084", @@ -2305,243 +2251,243 @@ "issue:39404|issue:44671", "issue:43425|issue:44918", "issue:44589|issue:45084", - "issue:36032|issue:43452", "issue:44164|issue:45759", + "issue:36032|issue:43452", "issue:43596|issue:44683", "issue:44671|issue:44991" ], - "prepared_review_unit_hash": "22098e8bdff3ab091b11d80ad19f6878c719681cc84eb533e1172b389a1bb50f", + "prepared_review_unit_hash": "02858d2dff42c29ba3926b4ac434b4cefbf62e53887ff642c9a7e90832b2d150", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 33, - "serialized_chars": 37672, - "estimated_input_tokens": 9418, - "estimated_eval_tokens": 19092 + "serialized_chars": 37674, + "estimated_input_tokens": 9419, + "estimated_eval_tokens": 19094 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a loose Transformers v5 bug cluster with only two convincing duplicate pairs: resize_token_embeddings propagation and Qwen3-VL MoE load failures. The rest are same-subsystem or same-version reports but not the same concrete bug.", - "confidence": 0.77, - "canonical_issue_reason": "issue:45292 is the broader, cleaner statement of the resize_token_embeddings bug; issue:45335 is a model-specific manifestation of the same failure to update downstream embeddings.", + "summary": "This cluster is mostly a loose collection of unrelated Transformers regressions, largely around model loading and v5 compatibility. Only the token-embedding resize reports (45292/45335) look like the same underlying bug; the rest are better kept separate.", + "confidence": 0.88, + "canonical_issue_reason": "Issue 43611 is the best anchor: it is a broad, concrete Transformers 5.0 model-loading regression and is representative of the cluster\u2019s main theme without being overly model-specific.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45292 is the best representative issue because it states the underlying API bug in a general way and cleanly subsumes the t5gemma-specific report.", + "best_issue_reason": "Issue 43611 is the strongest standalone issue to keep: clear symptom, broad impact, and a central regression category that matches the surrounding cluster better than the narrow model-specific reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43381", "right": "issue:44928", "accept": false, - "reason": "Different failures: eval-mode gradient checkpointing vs RLHF NaNs from Qwen3.5 SDPA/BF16 behavior." + "reason": "Both mention gradient/checkpointing, but the failures are different: eval-mode checkpointing vs RLHF NaN/SDPA fallback on Qwen3.5." }, { "left": "issue:44164", "right": "issue:44756", "accept": false, - "reason": "One is save/from_pretrained extra_state handling; the other is mmap/OOM on Strix Halo." + "reason": "Unrelated bugs: serialization of extra_state vs disabling mmap on Strix Halo to avoid OOM." }, { "left": "issue:42947", "right": "issue:43381", "accept": false, - "reason": "Both mention gradient checkpointing, but one is PEFT LoRA ineffectiveness and the other is eval-mode incompatibility." + "reason": "Both involve gradient checkpointing, but one is about LoRA effectiveness and the other about eval-mode incompatibility." }, { "left": "issue:44479", "right": "issue:44877", "accept": false, - "reason": "Video-input regression for Qwen VL models is unrelated to granite_speech strict-config loading." + "reason": "Different subsystems and failures: Qwen VL video-input regression vs strict config blocking granite_speech loading." }, { - "left": "issue:43646", - "right": "issue:43708", + "left": "issue:45588", + "right": "issue:45663", "accept": false, - "reason": "Custom model init in notebooks is unrelated to Trainer max_steps resumption math." + "reason": "Different code paths: flash_attention sink-less crash vs Gemma4 FSDP2 shared_kv_states rebuild bug." }, { - "left": "issue:45588", - "right": "issue:45663", + "left": "issue:43646", + "right": "issue:43708", "accept": false, - "reason": "Different code paths: flash_attention sink handling vs Gemma-4 FSDP2 shared_kv_states rebuilding." + "reason": "Custom model init in notebooks is unrelated to resume_from_checkpoint max_steps calculation." }, { "left": "issue:45292", "right": "issue:45335", "accept": true, - "reason": "Same underlying bug: resize_token_embeddings does not propagate to tied/output embeddings, with t5gemma as a specific case." + "reason": "Same underlying resize_token_embeddings bug; one report names output_embeddings and the other decoder.embed_tokens, both saying resizing does not propagate." }, { "left": "issue:43645", "right": "issue:44164", "accept": false, - "reason": "Notebook-defined custom model initialization is not the same as extra_state save/load failure." + "reason": "Notebook custom-model initialization and extra_state serialization are different issues." }, { "left": "issue:43257", "right": "issue:44863", "accept": false, - "reason": "Both are checkpoint-loading complaints, but they affect different model families and mechanisms." + "reason": "Both are model-loading problems, but they affect different architectures and failure modes; not the same bug." }, { "left": "issue:44402", "right": "issue:44451", "accept": false, - "reason": "Different tokenizer/model loading problems; shared symptom of load failure is too broad." + "reason": "Tokenizer vocab-size mismatch is not enough to equate it with a separate model-loading failure for ScandiBERT." }, { "left": "issue:43828", "right": "issue:43957", "accept": false, - "reason": "Autocast dtype mismatch is distinct from meta-device loading breakage." + "reason": "Different failures: autocast dtype mismatch vs meta-device loading breakage." }, { "left": "issue:43425", "right": "issue:43828", "accept": false, - "reason": "Torch version incompatibility is not the same bug as Phi MoE autocast dtype mismatch." + "reason": "Torch version incompatibility and a model-specific autocast dtype mismatch are not the same issue." }, { "left": "issue:44589", "right": "issue:44756", "accept": false, - "reason": "Float8 storage lookup error is unrelated to mmap-related OOM on Strix Halo." + "reason": "Float8 storage lookup failure is unrelated to the Strix Halo mmap OOM report." }, { "left": "issue:44683", "right": "issue:45823", "accept": false, - "reason": "Compiled flex_attention on torch>=2.9 is unrelated to Gemma4 device-map forward mismatch." + "reason": "Compiled flex_attention on torch>=2.9 and Gemma4 device-mapping mismatch are separate bugs." }, { "left": "issue:39404", "right": "issue:45701", "accept": false, - "reason": "Whisper pipeline return_language regression is unrelated to tokenization-version changes." + "reason": "Whisper pipeline return_language regression is unrelated to a generic tokenization-change report." }, { - "left": "issue:43425", - "right": "issue:45823", + "left": "issue:43906", + "right": "issue:44530", "accept": false, - "reason": "Torch 2.10 incompatibility and Gemma4 device mismatch are different issues." + "reason": "The isolated reproduction issue does not describe the same failure as the Qwen3.5 PagedAttentionCache crash." }, { - "left": "issue:43906", - "right": "issue:44530", + "left": "issue:43425", + "right": "issue:45823", "accept": false, - "reason": "Isolated reproduction of an earlier issue is not the same as the Qwen3.5 PagedAttentionCache linear_attention crash." + "reason": "Torch 2.10 incompatibility and Gemma4 PLE device mismatch are different problems." }, { "left": "issue:43611", "right": "issue:43645", "accept": false, - "reason": "Both involve v5 custom/loading breakage, but base_model_prefix loading and notebook-defined model init are different concrete failures." + "reason": "Base_model_prefix loading regression is not the same as custom model initialization in notebooks." }, { "left": "issue:43425", "right": "issue:43716", "accept": false, - "reason": "Torch compatibility issue is unrelated to Mistral-3 dtype mismatch between image preprocessor and model." + "reason": "A Torch version incompatibility is unrelated to a Mistral-3 image preprocessor/model dtype mismatch." }, { "left": "issue:44530", "right": "issue:45084", "accept": false, - "reason": "PagedAttentionCache linear_attention crash is not the same as the template-node compilation TypeError." + "reason": "Linear-attention group crash and non-template-node compilation error are different code paths." }, { "left": "issue:39404", "right": "issue:44912", "accept": false, - "reason": "Whisper pipeline language return regression is unrelated to MXFP4 quantization fallback." + "reason": "Whisper pipeline language return and MXFP4 quantization fallback are unrelated." }, { "left": "issue:43901", "right": "issue:44509", "accept": false, - "reason": "Both are docs-related, but they refer to different pipeline/task deprecations." + "reason": "Both are docs-related, but they cover different removed/changed pipeline docs and should not be merged as duplicates." }, { "left": "issue:44265", "right": "issue:44898", "accept": false, - "reason": "torch.export with torch_compilable_check is a different failure from Perceiver interpolation resizing." + "reason": "torch.export/torch_compilable_check failure and Perceiver non-default-resolution failure are separate issues." }, { "left": "issue:44479", "right": "issue:44805", "accept": false, - "reason": "Qwen VL video-input regression and generic mask-shape IndexError are not the same bug." + "reason": "Qwen video-input regression and a mask-shape IndexError are not the same bug." }, { "left": "issue:43994", "right": "issue:44079", "accept": false, - "reason": "Incorrect model/pipeline outputs are unrelated to ModelOutput key assignment when a key was previously None." + "reason": "Siglip2 incorrect outputs and ModelOutput key assignment are different behaviors." }, { "left": "issue:43299", "right": "issue:43931", - "accept": true, - "reason": "Same Qwen3-VL loading problem family; 43931 looks like a specific weight-shape manifestation of the broader MoE load failure." + "accept": false, + "reason": "Both involve Qwen3-VL loading, but one is a version-specific MoE loading regression and the other is a concrete weight-shape mismatch; too different to merge." }, { "left": "issue:39404", "right": "issue:44671", "accept": false, - "reason": "Whisper pipeline regression is unrelated to CamemBERT masked-LM prediction changes." + "reason": "Whisper pipeline behavior and CamemBERT masked-LM predictions are unrelated." }, { "left": "issue:43425", "right": "issue:44918", "accept": false, - "reason": "Torch version incompatibility is not the same issue as Qwen3.5 input-embedding unpacking in TRL SFT." + "reason": "Torch incompatibility is unrelated to Qwen3.5 embedding unpacking in TRL SFT." }, { "left": "issue:44589", "right": "issue:45084", "accept": false, - "reason": "Float8 storage lookup and non-template-node compilation failures are unrelated." + "reason": "Float8 storage lookup and non-template-node compilation are unrelated." }, { - "left": "issue:36032", - "right": "issue:43452", + "left": "issue:44164", + "right": "issue:45759", "accept": false, - "reason": "T5Tokenizer special-token method conflict is unrelated to gguf_file loading breakage in AutoTokenizer/AutoModel." + "reason": "extra_state serialization failure is unrelated to unwrapping text_config in composite Qwen multimodal configs." }, { - "left": "issue:44164", - "right": "issue:45759", + "left": "issue:36032", + "right": "issue:43452", "accept": false, - "reason": "extra_state save/load handling is unrelated to unwrapping text_config for composite Qwen configs." + "reason": "Tokenizer special-token method conflict and gguf_file loading breakage are different bugs." }, { "left": "issue:43596", "right": "issue:44683", "accept": false, - "reason": "Deepspeed ZeRO3 BertModel initialization failure is unrelated to compiled flex_attention on newer torch." + "reason": "Zero3 init/BertModel index error is unrelated to compiled flex_attention failing on torch>=2.9." }, { "left": "issue:44671", "right": "issue:44991", "accept": false, - "reason": "CamemBERT masked-LM prediction regression is not the same as tokenizer loading failure for EMBEDDIA/est-roberta." + "reason": "CamemBERT prediction regression and tokenizer loading failure for est-roberta are not the same issue." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary stays within the evidence in the packet: it identifies the two strongest duplicate pairs and treats the rest as related-but-distinct reports. The soft-edge rejections are conservative, and the accepted pairs are reasonably supported by the issue titles." + "feedback": "The summary is well grounded in the packet and stays conservative: it treats most pairs as separate regressions, and the one accepted pair (45292/45335) is a plausible same-bug resize_token_embeddings issue. The chosen anchor issue and cluster characterization are subjective but supported by the issue titles." }, "evaluator_used": true, "retried": false, @@ -2630,8 +2576,8 @@ "issue:43883|issue:44488", "issue:44811|issue:45584", "issue:44464|issue:45987", - "issue:43644|issue:43646", "issue:44451|issue:44987", + "issue:43644|issue:43646", "issue:43452|issue:43482", "issue:42915|issue:44530", "issue:43541|issue:45084", @@ -2639,217 +2585,217 @@ "issue:44849|issue:45325", "issue:36032|issue:43645", "issue:44987|issue:44991", - "issue:43089|issue:45987", - "issue:43404|issue:44466" + "issue:43404|issue:44466", + "issue:43089|issue:45987" ], - "prepared_review_unit_hash": "0e84dd37745fb3a5504dc462073c89ee66adce156c4111ba4959201b9a32e24c", + "prepared_review_unit_hash": "369bd14dd10ded0fc164eb90b068d049ea45d385c908e597db72bdaa60ac13f0", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 30, - "serialized_chars": 36131, - "estimated_input_tokens": 9033, - "estimated_eval_tokens": 18322 + "serialized_chars": 36126, + "estimated_input_tokens": 9032, + "estimated_eval_tokens": 18320 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These issues are only superficially similar; several are Transformers v5 regressions, but they describe different code paths and failure modes. No soft pair looks like the same underlying bug/change, so none should be merged as duplicates.", - "confidence": 0.82, - "canonical_issue_reason": "Issue 43611 is the closest broad representative: it captures a Transformers v5 model-loading regression, which is the main recurring theme among the more similar items. Still, the overall set is too heterogeneous for a true canonical duplicate.", + "summary": "The candidate set is highly heterogeneous; most pairs are only loosely similar by subsystem or regression theme, not the same underlying bug/change. No PRs are present. I reject all soft-edge pairs.", + "confidence": 0.9, + "canonical_issue_reason": "No single canonical issue: the items span unrelated bugs and regressions, with only broad thematic overlap in a few loading/tokenizer/model cases.", "canonical_pr_reason": null, - "best_issue_reason": "43611 is the strongest single anchor if one must pick one issue: it is broad, actionable, and closer to the repeated v5 loading breakages than the more model-specific reports. It is not a perfect cluster canonical, just the least bad representative.", + "best_issue_reason": "No issue is a strong global canonical for this cluster; if forced, issue:43452 is the broadest loading-regression report, but it still does not represent the whole set.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:32090", "right": "issue:39404", "accept": false, - "reason": "Different subsystems and failures: Trainer GPU broadcast NoneType vs Whisper pipeline return_language regression." + "reason": "Different components and failures: Trainer GPU broadcast TypeError vs Whisper pipeline return_language regression." }, { "left": "issue:43856", "right": "issue:44756", "accept": false, - "reason": "Both are memory-related, but one is Qwen3 MoE training inefficiency and the other is Strix Halo mmap OOM; not the same bug." + "reason": "Both mention memory/OOM, but one is Qwen3 MoE training memory use and the other is mmap on Strix Halo; not the same bug." }, { "left": "issue:44991", "right": "issue:45081", "accept": false, - "reason": "Both involve loading/tokenizers, but one is a specific tokenizer loading regression and the other is a Mistral regex patch crash." + "reason": "Both are tokenizer-loading regressions, but for different models and different code paths (EST-RoBERTa vs Mistral regex patching)." }, { "left": "issue:44737", "right": "issue:45859", "accept": false, - "reason": "XLNet CPU device placement bug vs Qwen3_5Moe CPU offload incompatibility; unrelated code paths." + "reason": "Both involve device placement, but XLNet positional encoding CPU allocation and Qwen3.5 MoE CPU-offload incompatibility are distinct bugs." }, { "left": "issue:44164", "right": "issue:44479", "accept": false, - "reason": "save/from_pretrained extra_state handling and a video input regression are distinct problems." + "reason": "Save/from_pretrained extra_state handling and a Qwen VL video-input regression are unrelated." }, { "left": "issue:43582", "right": "issue:43632", "accept": false, - "reason": "Apple Silicon warmup TypeError and _is_hf_initialized flag breakage are separate regressions." + "reason": "Apple Silicon allocator warmup TypeError is unrelated to the _is_hf_initialized v5 regression." }, { "left": "issue:44488", "right": "issue:45081", "accept": false, - "reason": "Different model-loading failures: one is a specific BERT load issue, the other is a tokenizer regex patch crash." + "reason": "Different failures: loading cjvt/sleng-bert vs a Mistral tokenizer regex crash." }, { "left": "issue:44661", "right": "issue:44683", "accept": false, - "reason": "Tokenizer mapping extension failure vs compiled flex_attention failure on torch>=2.9; unrelated." + "reason": "Tokenizer mapping/model registration bug vs compiled flex_attention on torch>=2.9; unrelated." }, { "left": "issue:40990", "right": "issue:43421", "accept": false, - "reason": "Perplexity anomaly on gpt-oss-20b is not the same as runtime special-token post-processor updates." + "reason": "Perplexity issue on gpt-oss-20b is unrelated to runtime special-token post-processor updates." }, { "left": "issue:44488", "right": "issue:44987", "accept": false, - "reason": "Both are load failures, but they affect different models and appear to stem from different breakages." + "reason": "Both are model-loading failures, but for different repos and different reported breakages." }, { "left": "issue:43611", "right": "issue:43644", "accept": false, - "reason": "Both are v5 regressions, but one is model-loading/base_model_prefix and the other is non-persistent buffer serialization; different fixes." + "reason": "Both are v5 regressions, but base_model_prefix loading and non-persistent buffer junk are separate issues." }, { "left": "issue:44206", "right": "issue:44220", "accept": false, - "reason": "Different audio-processing bugs: unsupported center arg vs _torch_extract_fbank_features behavior." + "reason": "Both are audio-feature related, but the unsupported center arg and _torch_extract_fbank_features crash are not clearly the same bug." }, { "left": "issue:43716", "right": "issue:45084", "accept": false, - "reason": "Mistral image-preprocessor dtype mismatch is unrelated to the template compilation TypeError." + "reason": "Mistral-3 dtype mismatch in image preprocessing is unrelated to a template-node compilation TypeError." }, { "left": "issue:43643", "right": "issue:44530", "accept": false, - "reason": "AutoConfig trust_remote_code missing fields and PagedAttentionCache linear_attention crash are separate issues." + "reason": "AutoConfig trust_remote_code missing fields is unrelated to PagedAttentionCache linear_attention crash." }, { "left": "issue:36032", "right": "issue:42915", "accept": false, - "reason": "Tokenizer special-token conflict and Qwen3Moe FineGrainedFP8Config failure are unrelated." + "reason": "Tokenizer add_special_tokens conflict and Qwen3 MoE FP8 config failure are unrelated." }, { "left": "issue:34567", "right": "issue:40444", "accept": false, - "reason": "TrainerState token counting and iterable dataset multi-image fine-tuning are different bugs." + "reason": "TrainerState token counting bug is unrelated to iterable multi-image Qwen2.5-VL finetuning failure." }, { "left": "issue:43883", "right": "issue:44488", "accept": false, - "reason": "Missing all_tied_weights_keys on Molmo is not the same as a model loading failure for a specific checkpoint." + "reason": "Molmo tied-weights attribute error is unrelated to loading cjvt/sleng-bert." }, { "left": "issue:44811", "right": "issue:45584", "accept": false, - "reason": "Whisper batch_decode skip_special_tokens handling and empty transcription after align_special_tokens are related to Whisper, but not the same failure." + "reason": "Both are Whisper-related, but batch_decode skip_special_tokens handling and empty-transcription generation after align_special_tokens are different code paths." }, { "left": "issue:44464", "right": "issue:45987", "accept": false, - "reason": "Chunked generation inconsistency with compiled forward is unrelated to StaticCache.get_seq_length returning the wrong type." + "reason": "Chunked generation with compiled forward is not the same as StaticCache.get_seq_length returning the wrong type." }, { - "left": "issue:43644", - "right": "issue:43646", + "left": "issue:44451", + "right": "issue:44987", "accept": false, - "reason": "Non-persistent buffer junk and custom model initialization regressions share a release but not a concrete code path." + "reason": "Separate model-loading failures for different models; no shared underlying bug is evident." }, { - "left": "issue:44451", - "right": "issue:44987", + "left": "issue:43644", + "right": "issue:43646", "accept": false, - "reason": "Different model-loading reports with different failure signatures and targets." + "reason": "Non-persistent buffer initialization junk vs custom model initialization breakage are distinct regressions." }, { "left": "issue:43452", "right": "issue:43482", "accept": false, - "reason": "GGUF-related loading failures, but one is the gguf_file API path and the other is a Qwen2.5-GGUF v5 regression." + "reason": "Both involve GGUF loading, but one is a broad gguf_file/from_pretrained regression and the other is a specific model failure; not enough evidence they are the same bug." }, { "left": "issue:42915", "right": "issue:44530", "accept": false, - "reason": "Qwen3 MoE FineGrainedFP8Config and PagedAttentionCache invalid group type are distinct architectural failures." + "reason": "Qwen3 MoE FP8 config failure is unrelated to PagedAttentionCache linear_attention crash." }, { "left": "issue:43541", "right": "issue:45084", "accept": false, - "reason": "Torch dynamo tracing failure in Mixtral grouped_mm and template-node compilation are unrelated." + "reason": "Torch dynamo grouped_mm tracing failure is unrelated to template-node compilation errors." }, { "left": "issue:43317", "right": "issue:43482", "accept": false, - "reason": "gpu+cpu offload dequantized-model loading and Qwen2.5-GGUF loading are different loading paths." + "reason": "Device-map dequantized-model offload failure is unrelated to Qwen2.5-GGUF loading." }, { "left": "issue:44849", "right": "issue:45325", "accept": false, - "reason": "Qwen3.5 hidden-state output bug and Qwen2.5-VL rope-index scaling bug are separate model-specific regressions." + "reason": "Qwen3.5 output_hidden_states bug and Qwen2.5-VL rope-index scaling bug are different issues." }, { "left": "issue:36032", "right": "issue:43645", "accept": false, - "reason": "Tokenizer method-name conflict and Jupyter custom-model initialization failure are not the same bug." + "reason": "Tokenizer special-token conflict is unrelated to Jupyter/custom-model initialization breakage." }, { "left": "issue:44987", "right": "issue:44991", "accept": false, - "reason": "Both are model/tokenizer load failures, but they target different checkpoints and likely different regression points." + "reason": "Both are loading regressions, but they affect different artifacts and likely different failure causes." }, { - "left": "issue:43089", - "right": "issue:45987", + "left": "issue:43404", + "right": "issue:44466", "accept": false, - "reason": "Generation performance overhead vs StaticCache sequence-length contract bug; different problems." + "reason": "lm_head tying in Mistral3 and inconsistent lm_head serialization by device are related in theme but not the same concrete bug." }, { - "left": "issue:43404", - "right": "issue:44466", + "left": "issue:43089", + "right": "issue:45987", "accept": false, - "reason": "Untied lm_head weight in Mistral3 and inconsistent serialization by device are related to weights, but not the same concrete issue." + "reason": "Generation GPU-sync overhead is unrelated to StaticCache.get_seq_length returning the wrong type." } ] }, @@ -2956,236 +2902,233 @@ "issue:43632|issue:44534", "issue:42175|issue:45542" ], - "prepared_review_unit_hash": "69d09afdca9fd2fdaaf5045be1e69cce67a45c5aee6f3a960342455fc05bbbbd", + "prepared_review_unit_hash": "8670bd5d7d4997cd76018ed0431ecf55d8eb95506d557c048c5ef63ca36bae9e", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 37272, - "estimated_input_tokens": 9318, - "estimated_eval_tokens": 18892 + "serialized_chars": 37278, + "estimated_input_tokens": 9320, + "estimated_eval_tokens": 18896 }, "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 bug reports; the only strong duplicate candidate is the pair about Qwen3.5 `save_pretrained` writing incorrect checkpoint/visual-encoder keys.", - "confidence": 0.68, - "canonical_issue_reason": "Issue 43611 is the best umbrella representative of the shared v5 model-loading regression theme: it is broad, concrete, and closer to the center of the cluster than the more model-specific reports.", + "summary": "This cluster is mostly heterogeneous, with one loose subtheme around Transformers v5 loading/initialization regressions, but the soft-edge candidates are not true duplicates and span different code paths, models, or modalities.", + "confidence": 0.74, + "canonical_issue_reason": "Issue 43632 is the broadest and most central Transformers v5 regression in this set, touching model initialization/loading behavior that overlaps conceptually with several other issues here.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43611 is the strongest overall representative because it captures the widest and most central failure mode in the set: Transformers v5 breaking model loading for affected classes.", + "best_issue_reason": "43632 is the best representative of the cluster\u2019s main theme: a core Transformers v5 initialization regression with broad downstream impact, unlike the more model-specific or modality-specific reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43262", "right": "issue:43377", "accept": false, - "reason": "Different bug classes: audio chat-template sampling-rate defaulting vs MIMI batching/padding-mask mismatch." + "reason": "Audio processor chat-template sampling-rate default vs MIMI encoder batch-padding-mask mismatch; different components and bugs." }, { "left": "issue:45084", "right": "issue:45507", "accept": false, - "reason": "Unrelated failures: template compilation error vs GraniteMoEHybrid calling an invalid method." + "reason": "Template compilation error vs GraniteMoEHybrid calling an invalid method; unrelated failure modes." }, { "left": "issue:43883", "right": "issue:44991", "accept": false, - "reason": "Both involve loading, but one is a missing tied-weights attribute and the other is tokenizer loading; not the same concrete bug." + "reason": "Missing `all_tied_weights_keys` on MolmoForCausalLM vs tokenizer loading failure for EMBEDDIA/est-roberta; different subsystems." }, { "left": "issue:43994", "right": "issue:44610", "accept": false, - "reason": "Both are vision model input/processor issues, but the models and failure mechanisms differ." + "reason": "SigLIP2 bad outputs via AutoModel/pipeline vs OmDet-Turbo processor/model input-size mismatch; not the same bug." }, { "left": "issue:43646", "right": "issue:44534", "accept": false, - "reason": "Custom model initialization in notebooks is unrelated to non-persistent buffer corruption." + "reason": "Custom model initialization breakage vs non-persistent buffers being filled with junk; both v5 regressions, but different code paths." }, { "left": "issue:43322", "right": "issue:43540", "accept": false, - "reason": "Different multimodal models and different failure paths: segfault on Llava Next vs video-processing ValueError on Qwen3OmniMoe." + "reason": "Llava Next segfault vs Qwen3OmniMoe video ValueError; both multimodal, but distinct crashes and models." }, { "left": "issue:43645", "right": "issue:44534", "accept": false, - "reason": "Notebook custom-model init regression is not the same as buffer initialization junk." + "reason": "Jupyter/custom model init regression vs buffer initialization regression; not the same underlying issue." }, { "left": "issue:44403", "right": "issue:45081", "accept": false, - "reason": "Generic loading noise issue vs a specific Mistral regex patch crash." + "reason": "Generic loading noise vs Mistral regex patch crash on tokenizer backend access; different behaviors and fixes." }, { "left": "issue:43377", "right": "issue:43819", "accept": false, - "reason": "One is padding-mask batching in MIMI; the other is DAC `from_latents` vs forward mismatch." + "reason": "MIMI encoder padding-mask discrepancy vs DAC STE mismatch in `from_latents`; different model internals and symptoms." }, { "left": "issue:44265", "right": "issue:44610", "accept": false, - "reason": "Torch export compiler checks and image-size processor mismatch are unrelated." + "reason": "`torch.export.export` failure with `torch_compilable_check` vs processor output-size mismatch; unrelated export vs preprocessing bugs." }, { "left": "issue:42915", "right": "issue:45084", "accept": false, - "reason": "FP8 config failure is unrelated to template-node compilation." + "reason": "Qwen3Moe FP8 config failure vs non-template compilation error; different model/config and failure mechanism." }, { "left": "issue:45216", "right": "issue:45357", - "accept": true, - "reason": "Both are the same Qwen3.5 `save_pretrained` serialization regression, with the second narrowing it to incorrect visual-encoder keys." + "accept": false, + "reason": "Both are Qwen3.5 save_pretrained regressions, but one concerns checkpoint correctness and the other visual encoder key naming; related area but not the same concrete bug." }, { "left": "issue:44451", "right": "issue:45081", "accept": false, - "reason": "Model load failure for a specific checkpoint is not the same as a Mistral tokenizer regex crash." + "reason": "Model load failure for vesteinn/ScandiBERT vs tokenizer backend attribute crash in Mistral regex patch; different root causes." }, { "left": "issue:44987", "right": "issue:45081", "accept": false, - "reason": "Tokenizer/model loading failure for one repo vs a specific Mistral regex patch bug." + "reason": "Model loading failure for physical-intelligence/fast vs Mistral tokenizer regex patch crash; different models and code paths." }, { "left": "issue:44220", "right": "issue:44683", "accept": false, - "reason": "Feature extraction and flex-attention compilation are different subsystems and different bugs." + "reason": "`_torch_extract_fbank_features()` issue vs compiled flex_attention failure on newer torch; unrelated audio feature extraction vs attention compilation." }, { "left": "issue:42915", "right": "issue:44291", "accept": false, - "reason": "FP8 config handling and `init_empty_weights` argument incompatibility are different regressions." + "reason": "Qwen3Moe FP8 config issue vs `init_empty_weights`/_is_hf_initialized argument handling; distinct initialization/configuration problems." }, { "left": "issue:43404", "right": "issue:45127", "accept": false, - "reason": "Both mention tied weights, but one is a missing `lm_head` tie and the other is LoRA merge collapse after vocab extension." + "reason": "Mistral3 untied lm_head weights vs LoRA merge collapse with extended vocab; both involve weights, but not the same bug." }, { "left": "issue:43632", "right": "issue:44756", "accept": false, - "reason": "HF initialization flag handling is unrelated to Strix Halo mmap/OOM behavior." + "reason": "`_is_hf_initialized` regression vs mmap/OOM workaround on Strix Halo; unrelated loading/runtime concerns." }, { "left": "issue:44062", - "right": "issue:44991", + "right": "issue:44987", "accept": false, - "reason": "Tokenizer `AddedToken` keyword collision is not clearly the same concrete bug as the EMBIDDIA tokenizer load failure." + "reason": "Duplicate `special` keyword in AddedToken construction vs tokenizer/model loading failure for a specific checkpoint; different tokenizer-layer bugs." }, { "left": "issue:44479", "right": "issue:45325", "accept": false, - "reason": "Related Qwen-VL family, but one is a broad video-input regression and the other is still-image temporal position-id scaling." + "reason": "Video input regression across Qwen variants vs Qwen2.5-VL rope index scaling bug; both visual models, but different preprocessing/math errors." }, { "left": "issue:36032", "right": "issue:43742", "accept": false, - "reason": "Different load-time errors for different models: T5 tokenizer conflict vs MobileLLM key error." + "reason": "T5 tokenizer special-token method conflict vs MobileLLM key error on load; unrelated tokenizer vs model-loading issues." }, { "left": "issue:42915", "right": "issue:44589", "accept": false, - "reason": "FP8 config failure and missing Float8 storage type are distinct low-level issues." + "reason": "Qwen3Moe FP8 config failure vs missing Float8 storage type on load; both FP8-adjacent, but different error surfaces and fixes." }, { "left": "issue:40444", "right": "issue:45325", "accept": false, - "reason": "IterableDataset multi-image training failure is unrelated to Qwen2.5-VL RoPE temporal scaling." + "reason": "IterableDataset multi-image finetuning failure vs Qwen2.5-VL rope index scaling bug; different multimodal training vs inference issues." }, { "left": "issue:43296", "right": "issue:43531", "accept": false, - "reason": "PaddleOCR-VL load failure and Qwen3-MoE sliding-window bug are different models and code paths." + "reason": "PaddleOCR-VL load failure in vLLM vs Qwen3-MoE sliding_window behavior; different models and code paths." }, { "left": "issue:43611", "right": "issue:44534", "accept": false, - "reason": "Both are v5 regressions, but one is base-model-prefix loading and the other is non-persistent buffer initialization." + "reason": "`base_model_prefix` loading regression vs non-persistent buffer junk values; both v5-related, but not the same bug." }, { "left": "issue:36032", "right": "issue:41720", "accept": false, - "reason": "Tokenizer conflict vs CUDA assert from auto device mapping are unrelated." + "reason": "T5 tokenizer load error vs Qwen3 auto device mapping cuda assert; different model families and failure layers." }, { "left": "issue:43531", "right": "issue:44877", "accept": false, - "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to granite_speech config strictness." + "reason": "Qwen3-MoE sliding_window issue vs strict config blocking granite_speech load; unrelated model behavior vs config validation." }, { "left": "issue:43425", "right": "issue:44403", "accept": false, - "reason": "Torch version incompatibility is not the same as noisy transformer loading." + "reason": "Torch 2.10 incompatibility vs generic transformer-loading noise; one is dependency compatibility, the other a runtime loading symptom." }, { "left": "issue:44079", "right": "issue:45663", "accept": false, - "reason": "`ModelOutput` key assignment and Gemma4 FSDP2 shared-kv KeyError are distinct bugs." + "reason": "ModelOutput key assignment bug vs Gemma4 FSDP2 shared_kv_states KeyError; different abstractions and failure sites." }, { "left": "issue:40990", "right": "issue:44811", "accept": false, - "reason": "Perplexity anomaly on a language model is unrelated to Whisper batch_decode special-token handling." + "reason": "High perplexity on GPT-OSS/WikiText-2 vs Whisper batch_decode skip_special_tokens regression; unrelated evaluation vs decoding behavior." }, { "left": "issue:43632", "right": "issue:44534", "accept": false, - "reason": "HF-initialization flag regression is distinct from junk-filled non-persistent buffers." + "reason": "_is_hf_initialized flag regression vs non-persistent buffer initialization bug; both Transformers v5 issues, but distinct code paths and symptoms." }, { "left": "issue:42175", "right": "issue:45542", "accept": false, - "reason": "Both touch TensorFlow/backend packaging, but one is missing TF in the torch extra and the other is an undefined backend error when only tensorboard is present." + "reason": "Packaging/install backend inclusion issue vs tensorboard-only install causing undefined tf backend; different dependency resolution problems." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative overall. The chosen representative issue (43611) matches the broad v5 model-loading regression theme, and the only accepted soft pair (45216/45357) is plausibly the same Qwen3.5 save_pretrained bug family. The remaining soft-edge rejections are cautious rather than overstated." - }, - "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 }, @@ -3280,20 +3223,20 @@ "issue:43611|issue:43646", "issue:42175|issue:43296" ], - "prepared_review_unit_hash": "b3da27a5267265b1921e18bb0276d9a811c9ce8083a188f9c98eb7a4b503c7cb", + "prepared_review_unit_hash": "a20bb808612ae457568959432161281640b615db802f178ae684aaafb3d0eb56", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 35637, + "serialized_chars": 35640, "estimated_input_tokens": 8910, "estimated_eval_tokens": 18076 }, @@ -3302,180 +3245,198 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose collection of unrelated or only broadly similar Transformer regressions; there are no convincing duplicate pairs to merge. The best representative issue is the core v5 initialization regression, but it still does not unify the rest into a single duplicate thread.", - "confidence": 0.79, - "canonical_issue_reason": "issue:43632 is the most central umbrella for the set because it describes a core Transformers v5 initialization regression that overlaps several other loading/init failures more than any other candidate.", + "summary": "This cluster is not a duplicate set; it\u2019s a loose collection of mostly unrelated transformer regression reports, many centered on Transformers v5 compatibility but covering different code paths and model families.", + "confidence": 0.83, + "canonical_issue_reason": "issue:43646 is the best representative anchor: it is a broad, concrete Transformers v5 compatibility regression about custom model initialization, which matches the dominant theme of the cluster better than the narrower tokenizer/model-specific reports.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43632 is the strongest cluster anchor: broad, high-impact, and closest to the recurring v5 model-loading/initialization breakages seen elsewhere in the set.", + "best_issue_reason": "issue:43646 is the strongest standalone issue to keep: it\u2019s specific enough to be actionable yet broad enough to represent the main v5 regression pattern seen here.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41553", "right": "issue:43582", "accept": false, - "reason": "Voxtral tokenizer error and Apple Silicon allocator TypeError are unrelated subsystems and failure modes." + "reason": "Different subsystems and failures: tokenizer error messaging vs Apple Silicon allocator warmup TypeError." }, { "left": "issue:43421", "right": "issue:44568", "accept": false, - "reason": "Both concern token special handling, but one is runtime post-processor sync while the other is add_special_tokens behavior for a specific tokenizer." + "reason": "Both touch tokenizers, but one is runtime special-token post-processing and the other is a specific BOS/EOS regression for one tokenizer; not the same bug." }, { "left": "issue:43792", "right": "issue:44589", "accept": false, - "reason": "Whisper model runtime failure and Float8 storage lookup error are unrelated." + "reason": "Completely different failure modes: Whisper model loading/runtime vs Float8 storage lookup." }, { "left": "issue:43818", "right": "issue:45357", "accept": false, - "reason": "Video-LLaVA video_tower architecture bug and Qwen3.5 save_pretrained key regression are different code paths and symptoms." + "reason": "Video-LLaVA weight/temporal-attention issue is unrelated to Qwen3.5 save_pretrained key names." }, { "left": "issue:44683", "right": "issue:45020", "accept": false, - "reason": "Compiled flex_attention on torch>=2.9 is unrelated to remote_code model-loading regressions." + "reason": "One is torch >=2.9 flex_attention compilation, the other is remote_code model loading in recent Transformers." }, { "left": "issue:33666", "right": "issue:40444", "accept": false, - "reason": "Both involve Qwen VL training, but multi-GPU training and iterable-dataset multi-image failure are different bugs." + "reason": "Both are Qwen VL training issues, but one is multi-GPU training and the other is IterableDataset multi-image prompt failures; not the same bug." }, { "left": "issue:44403", "right": "issue:44991", "accept": false, - "reason": "Generic loading noise is not the same as tokenizer loading failure for a specific model." + "reason": "Generic loading noise vs tokenizer loading failure for a specific model; different symptoms and code paths." }, { "left": "issue:44462", "right": "issue:45812", "accept": false, - "reason": "Both are AutoTokenizer regressions, but ignoring tokenizer.json and producing wrong Granite token IDs are distinct issues." + "reason": "Tokenizer.json discovery bug and Granite token-id regression are both tokenizer-related but not the same concrete failure." }, { "left": "issue:43317", "right": "issue:43856", "accept": false, - "reason": "GPU/CPU offload loading failure and Qwen3 MoE training memory inefficiency are unrelated." + "reason": "device_map/offload loading failure is unrelated to Qwen3 MoE memory efficiency during training." }, { "left": "issue:43482", "right": "issue:45538", "accept": false, - "reason": "GGUF loading failure and CLIPTokenizer max_length behavior are unrelated tokenizer/model problems." + "reason": "GGUF loading under Transformers v5 and CLIPTokenizer max-length behavior are unrelated." }, { "left": "issue:44829", "right": "issue:45479", "accept": false, - "reason": "Flash-attention training degeneration and num_labels=1 zero-loss classification are different root causes." + "reason": "Flash-attention-3 training degeneration is not the same as zero-loss classification with num_labels=1." }, { "left": "issue:43381", "right": "issue:45923", "accept": false, - "reason": "Eval-mode gradient checkpointing restriction and missing supports_gradient_checkpointing flag are not the same bug." + "reason": "Eval-mode gradient checkpointing misuse vs missing supports_gradient_checkpointing flag on a trust_remote_code model." }, { "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 failure." + "reason": "Progressive generation with inputs_embeds/past_key_values is unrelated to VL finetuning with multiple images per prompt." }, { "left": "issue:44079", "right": "issue:44534", "accept": false, - "reason": "ModelOutput key assignment bug and non-persistent buffer junk regression are separate internal issues." + "reason": "ModelOutput key assignment bug and non-persistent buffer junk are different internals with different effects." }, { "left": "issue:44291", "right": "issue:44589", "accept": false, - "reason": "init_empty_weights/_is_hf_initialized argument error is unrelated to Float8 storage resolution." + "reason": "Empty-weights initialization arg mismatch is unrelated to Float8 storage resolution." }, { "left": "issue:43482", "right": "issue:43525", "accept": false, - "reason": "Both are loading regressions in v5, but GGUF loading and missing pad_token_id on Llama4Config are different problems." + "reason": "GGUF loading failure and missing pad_token_id on Llama4Config are different model/config regressions." }, { "left": "issue:43377", "right": "issue:45491", "accept": false, - "reason": "MIMI padding-mask discrepancy and Gemma3 sliding-window NaNs are different batching/masking bugs." + "reason": "MIMI encoder padding-mask inconsistency is unrelated to Gemma3 sliding-window NaN embeddings." }, { "left": "issue:43749", "right": "issue:44079", "accept": false, - "reason": "FSDP CPU RAM efficient loading and ModelOutput key assignment are unrelated." + "reason": "FSDP CPU RAM efficient loading is unrelated to ModelOutput key bookkeeping." }, { "left": "issue:40444", "right": "issue:45381", "accept": false, - "reason": "Qwen2.5-VL multi-image training failure and video vision_position_ids regression are different multimodal paths." + "reason": "Multi-image prompt finetuning failure is not the same as Qwen2.5-VL video position-id regression." }, { "left": "issue:41720", "right": "issue:45081", "accept": false, - "reason": "Qwen3 device-map cuda assert and Mistral tokenizer regex patch crash are unrelated." + "reason": "A800 cudaErrorAssert with auto device mapping is unrelated to Mistral tokenizer regex patch crashes." }, { "left": "issue:43644", "right": "issue:44079", "accept": false, - "reason": "Both are v5 regressions, but non-persistent buffer corruption and ModelOutput key assignment are distinct behaviors." + "reason": "Non-persistent buffer corruption and ModelOutput key handling are separate regressions." }, { "left": "issue:44265", "right": "issue:45072", "accept": false, - "reason": "torch.export/check failure and bfloat16 dtype mismatch are unrelated." + "reason": "torch.export/torch_compilable_check failure is unrelated to dtype mismatches in inference." }, { "left": "issue:43531", "right": "issue:44291", "accept": false, - "reason": "Sliding-window behavior in Qwen3-MoE and init_empty_weights argument handling are unrelated." + "reason": "Qwen3-MoE sliding-window behavior does not match the init_empty_weights/_is_hf_initialized issue." }, { "left": "issue:45325", "right": "issue:45381", "accept": false, - "reason": "Both involve Qwen2.5-VL position IDs, but one is still-image temporal scaling and the other is video input positional handling; not the same concrete bug." + "reason": "Both are Qwen2.5-VL positional-id bugs, but one is still-image temporal scaling and the other is video vision_position_ids; too different to merge." }, { "left": "issue:43632", "right": "issue:45072", "accept": false, - "reason": "_is_hf_initialized regression and bfloat16 dtype mismatch are unrelated." + "reason": "_is_hf_initialized breakage is a model initialization regression, not a dtype mismatch issue." }, { "left": "issue:43994", "right": "issue:44898", "accept": false, - "reason": "Different vision models and different failure modes; only a broad image-resolution theme overlaps." + "reason": "SigLIP2 AutoModel/pipeline nonsense results and Perceiver non-default resolution failure are unrelated vision-model bugs." }, { "left": "issue:43611", "right": "issue:43646", "accept": false, - "reason": "Both are Transformers v5 model-init regressions, but base_model_prefix handling and custom model initialization are distinct enough that they should not be merged." + "reason": "Both are Transformers v5 compatibility regressions, but one is base_model_prefix loading and the other is custom model initialization; different concrete bugs." }, { "left": "issue:42175", "right": "issue:43296", "accept": false, - "reason": "Package backend selection and vLLM model-loading failure are unrelated." + "reason": "Pip extras backend packaging is unrelated to loading PaddleOCR-VL in vLLM." + }, + { + "left": "issue:44534", + "right": "issue:43644", + "accept": false, + "reason": "Both mention v5-era regressions, but one concerns buffer contents and the other concerns non-persistent buffer junk specifically; not the same artifact." + }, + { + "left": "issue:44568", + "right": "issue:43421", + "accept": false, + "reason": "A specific add_special_tokens regression on one tokenizer is not the same as runtime post-processor update behavior." + }, + { + "left": "issue:45020", + "right": "issue:43632", + "accept": false, + "reason": "Remote_code model breakage and _is_hf_initialized handling are different compatibility regressions." } ] }, @@ -3538,6 +3499,7 @@ "issue:44898", "issue:45072", "issue:45325", + "issue:45357", "issue:45584", "issue:45701", "issue:45759", @@ -3565,8 +3527,8 @@ "issue:34634|issue:35707", "issue:43873|issue:45701", "issue:44186|issue:44265", - "issue:35707|issue:41720", "issue:43606|issue:43819", + "issue:35707|issue:41720", "issue:41720|issue:45823", "issue:43454|issue:43550", "issue:43377|issue:44610", @@ -3576,210 +3538,217 @@ "issue:43257|issue:45923", "issue:43454|issue:45072", "issue:43572|issue:45759", - "issue:40990|issue:42947" + "issue:40990|issue:42947", + "issue:44466|issue:45357" ], - "prepared_review_unit_hash": "28eba6e828c993326e10d7f15b8a60962727597b1e0b83599814cbdf6e4ed10a", + "prepared_review_unit_hash": "0effb5cb84a1b0f44a53ebd87a83ecd5bad739755840caa06027a72e3b2f7766", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 29, - "serialized_chars": 35900, - "estimated_input_tokens": 8975, - "estimated_eval_tokens": 18206 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36787, + "estimated_input_tokens": 9197, + "estimated_eval_tokens": 18650 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster looks noisy and mostly consists of superficially similar but distinct issues across tokenization, offloading, multimodal loading, and model-specific regressions. I would use issue 43873 as the best umbrella representative because it is open, broad, and has the strongest offloading/quantization framing, but I do not see any PRs in the cluster.", - "confidence": 0.36, - "canonical_issue_reason": "issue:43873 is the broadest open report with active discussion and inbound references, and its offloading/quantization framing makes it the most plausible umbrella issue in this noisy cluster.", + "summary": "This set is a grab-bag of mostly unrelated bug reports; the soft pairs share broad themes like loading, offloading, tokenization, or multimodal regressions, but not the same concrete root cause. No soft pair looks safe to merge as a duplicate or same-change match.", + "confidence": 0.89, + "canonical_issue_reason": "If a single issue must represent the set, issue 45920 is the broadest systemic regression: it affects multiple model families, has clear impact, and is the most generally reusable reference point among these otherwise unrelated reports.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43873 is the best representative because it is open, broadly scoped, and thematically central to one of the few recurring patterns in the cluster (offloading/quantization).", + "best_issue_reason": "Issue 45920 is the strongest overall issue candidate because it describes a wide tokenizer regression across many models, is open, and has the best chance of being a useful umbrella reference compared with the narrower, model-specific reports here.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43454", "right": "issue:44466", "accept": false, - "reason": "Both mention lm_head/tied weights, but one is a model-specific failure to tie weights and the other is a device-dependent serialization inconsistency; related area, not the same bug." + "reason": "Both mention lm_head/tied weights, but one is a generation bug in AyaVision while the other is a serialization inconsistency tied to device state; different concrete failure modes." }, { "left": "issue:45325", "right": "issue:45834", "accept": false, - "reason": "These are different multimodal bugs: Qwen2.5-VL rope/index scaling vs Kosmos2.5 OCR long-input indexing. No concrete shared code-path is evident." + "reason": "Both are vision/multimodal indexing problems, but they affect different models and different code paths; no evidence they stem from the same bug." }, { "left": "issue:43792", "right": "issue:43957", "accept": false, - "reason": "Whisper runtime failure and meta-device loading regressions are different symptoms and likely different code paths." + "reason": "Whisper runtime failure and meta-device loading breakage are different model-loading problems with different triggers and symptoms." }, { "left": "issue:42915", "right": "issue:43482", "accept": false, - "reason": "Qwen3MoE FP8 config failure and Qwen2.5 GGUF loading failure are unrelated model-loading problems." + "reason": "FineGrainedFP8Config failure and GGUF loading failure are unrelated quantization/loading issues." }, { "left": "issue:43296", "right": "issue:45800", "accept": false, - "reason": "PaddleOCR-VL/vLLM loading failure and a torch-transformers version incompatibility are too broad and not the same concrete bug." + "reason": "These are different compatibility problems: one is a vLLM/PaddleOCR-VL load failure, the other is a torch/transformers version incompatibility." }, { "left": "issue:44387", "right": "issue:44617", "accept": false, - "reason": "Reserved-memory OOM under int4 quantization and Sam3Video OOM are both memory issues, but the titles point to different failure modes and models." + "reason": "Both mention OOM/memory, but one is a general int4 quantization reserved-memory regression and the other is a specific model running out of memory." }, { "left": "issue:43701", "right": "issue:45325", "accept": false, - "reason": "Checkpoint resume key mismatch is unrelated to the Qwen2.5-VL rope-position bug." + "reason": "Checkpoint resume key mismatch is unrelated to the Qwen2.5-VL RoPE position-id bug." }, { "left": "issue:44479", "right": "issue:45823", "accept": false, - "reason": "Video input regression in Qwen VL models and Gemma4 PLE device mismatch are different model families and different failure mechanisms." + "reason": "Video-input regression for Qwen VL models and Gemma4 PLE device mismatch are different forward-pass bugs in different models." }, { "left": "issue:43606", "right": "issue:44265", "accept": false, - "reason": "Bark CPU offload device mismatch and torch.export failure on torch_compilable_check are unrelated code paths." + "reason": "CPU offload device mismatch for Bark is unrelated to torch.export failing on torch_compilable_check." }, { "left": "issue:43262", "right": "issue:43746", "accept": false, - "reason": "Audio chat-template sampling-rate defaulting and GraniteSpeech PEFT checkpoint loading are separate audio/model-loading problems." + "reason": "Audio chat-template sampling-rate defaulting and GraniteSpeech PEFT checkpoint loading are separate audio/model-loading issues." }, { "left": "issue:43881", "right": "issue:44488", "accept": false, - "reason": "glm-4v-9b load failure and sleng-bert load failure are different models with no obvious shared root cause." + "reason": "Different models, different load failures, and no shared concrete failure mechanism." }, { "left": "issue:44291", "right": "issue:44877", "accept": false, - "reason": "init_empty_weights/_is_hf_initialized and strict config loading are both v5-loading regressions, but the concrete failures differ." + "reason": "Unexpected _is_hf_initialized argument with init_empty_weights is a constructor/load API issue, while strict config rejection is a config-schema problem." }, { "left": "issue:40990", "right": "issue:45584", "accept": false, - "reason": "High perplexity on gpt-oss and Whisper empty-transcription generation are unrelated generation bugs." + "reason": "Perplexity on GPT-OSS and Whisper empty-transcription generation failure are unrelated inference bugs." }, { "left": "issue:44568", "right": "issue:45920", "accept": false, - "reason": "Both concern tokenization in v5, but one is a specific add_special_tokens/BOS-EOS regression and the other is broader wrong-token-ID behavior across multiple models." + "reason": "One is a missing BOS/EOS addition for a specific tokenizer; the other is a broader wrong-token-IDs regression across multiple tokenizers." }, { "left": "issue:34634", "right": "issue:35707", "accept": false, - "reason": "Bark voice_preset and progressive generation with inputs_embeds/past_key_values are unrelated." + "reason": "Bark voice preset handling and progressive generation with inputs_embeds/past_key_values are unrelated features." }, { "left": "issue:43873", "right": "issue:45701", "accept": false, - "reason": "Offloading/quantization issues and tokenizer-version changes are different subsystems and different bug classes." + "reason": "Offloading/quantization behavior and version-dependent tokenization are different subsystems and failure modes." }, { "left": "issue:44186", "right": "issue:44265", "accept": false, - "reason": "LayoutLMv2 tokenizer padding/NER crashes and torch.export torch_compilable_check failures do not point to the same underlying defect." + "reason": "Tokenizer crashes on NER/batched padding is unrelated to export failures involving torch_compilable_check." }, { - "left": "issue:35707", - "right": "issue:41720", + "left": "issue:43606", + "right": "issue:43819", "accept": false, - "reason": "These are both generation/runtime issues, but one is about inputs_embeds/past_key_values and the other is a Qwen3 device-map CUDA assert." + "reason": "CPU offload device mismatch for Bark and DAC latent/forward mismatch are different models and different bugs." }, { - "left": "issue:43606", - "right": "issue:43819", + "left": "issue:35707", + "right": "issue:41720", "accept": false, - "reason": "Bark CPU offload mismatch and DAC.from_latents vs forward missing-STE mismatch are distinct audio-model bugs." + "reason": "Inputs_embeds generation and Qwen3 auto device-map CUDA asserts are not the same code-path problem." }, { "left": "issue:41720", "right": "issue:45823", "accept": false, - "reason": "Qwen3 auto device mapping CUDA assert and Gemma4 PLE forward device mismatch are different models and different device-placement failures." + "reason": "Both involve device placement, but one is a CUDA assert with Qwen3 and the other is a Gemma4 PLE device mismatch; not the same bug." }, { "left": "issue:43454", "right": "issue:43550", "accept": false, - "reason": "Tied-weight lm_head bug in AyaVision and Bamba torch.compile SDPA failure are unrelated." + "reason": "lm_head weight tying/serialization and torch.compile+SDPA failure are unrelated issues." }, { "left": "issue:43377", "right": "issue:44610", "accept": false, - "reason": "MIMI batch-vs-single padding-mask mismatch and OmDet-Turbo image-size mismatch are separate processor/model contract bugs." + "reason": "Missing padding-mask support in MIMI encoder is unrelated to OmDet-Turbo producing the wrong input resolution." }, { "left": "issue:39692", "right": "issue:40444", "accept": false, - "reason": "SigLIP2 doc example errors and Qwen2.5-VL iterable-dataset multi-image finetuning failure are different multimodal issues." + "reason": "Doc-example model/processor mismatch and Qwen2.5-VL finetuning with multiple images are different problems." }, { "left": "issue:43317", "right": "issue:43873", "accept": false, - "reason": "Both involve offloading/quantization, but one is a device_map loading failure for a dequantized model and the other is a broader offloading regression; too little evidence they are the same bug." + "reason": "Both mention offload/quantization, but one is a dequantized model load failure on gpu+cpu offload and the other is a broader offloading regression; insufficient evidence of the same root cause." }, { "left": "issue:43377", "right": "issue:44898", "accept": false, - "reason": "Missing padding-mask support in MIMI encoder and Perceiver non-default-resolution failure are unrelated input-handling bugs." + "reason": "MIMI padding-mask batching bug and Perceiver non-default resolution failure are different model-specific shape/vision issues." }, { "left": "issue:43257", "right": "issue:45923", "accept": false, - "reason": "Qwen3 MoE weights not converted with accelerate+deepspeed and missing supports_gradient_checkpointing on a trust_remote_code model are unrelated." + "reason": "Qwen3 MOE weight conversion under accelerate+deepspeed and missing supports_gradient_checkpointing on a trust_remote_code model are unrelated load/config issues." }, { "left": "issue:43454", "right": "issue:45072", "accept": false, - "reason": "lm_head tying/serialization and bfloat16 dtype mismatches in SwitchTransformers/TimmWrapperModel are different issues." + "reason": "AyaVision tied-weight generation issue and dtype mismatches in SwitchTransformers/TimmWrapperModel are different bugs." }, { "left": "issue:43572", "right": "issue:45759", "accept": false, - "reason": "StableLmConfig pad_token_idx regression and AutoModelForCausalLM.from_config unwrapping text_config are not the same bug." + "reason": "Missing pad_token_idx in StableLmConfig and failure to unwrap text_config for composite Qwen multimodal configs are different configuration issues." }, { "left": "issue:40990", "right": "issue:42947", "accept": false, - "reason": "High perplexity on a language model and ineffective gradient checkpointing with PEFT LoRA are unrelated." + "reason": "High perplexity on a model evaluation task and gradient checkpointing ineffectiveness with LoRA are unrelated." + }, + { + "left": "issue:44466", + "right": "issue:45357", + "accept": false, + "reason": "Both concern serialization, but one is lm_head weight inconsistency by device and the other is incorrect visual encoder keys for Qwen3.5; not the same fix." } ] }, @@ -3830,10 +3799,11 @@ "issue:44186", "issue:44387", "issue:44423", - "issue:44466", "issue:44479", "issue:44534", "issue:44610", + "issue:44671", + "issue:44743", "issue:44756", "issue:44877", "issue:44898", @@ -3853,18 +3823,17 @@ ], "soft_pairs": [ "issue:43526|issue:43883", - "issue:44466|issue:45357", "issue:43454|issue:43720", "issue:43749|issue:44534", "issue:41720|issue:43322", "issue:43881|issue:44991", - "issue:43296|issue:44991", + "issue:43296|issue:44877", "issue:41720|issue:44977", "issue:44008|issue:45823", "issue:42617|issue:42915", "issue:40990|issue:45701", - "issue:44186|issue:45072", "issue:43404|issue:45357", + "issue:44186|issue:45072", "issue:45632|issue:45684", "issue:43577|issue:43708", "issue:44387|issue:45704", @@ -3872,8 +3841,8 @@ "issue:43873|issue:44756", "issue:36331|issue:42617", "issue:44877|issue:44991", - "issue:36331|issue:42915", "issue:43550|issue:43720", + "issue:36331|issue:42915", "issue:43450|issue:45701", "issue:44610|issue:44898", "issue:44186|issue:44610", @@ -3884,234 +3853,235 @@ "issue:42915|issue:43317", "issue:43295|issue:45701", "issue:41720|issue:45161", - "issue:43526|issue:43881" + "issue:43526|issue:43881", + "issue:44671|issue:44743" ], - "prepared_review_unit_hash": "3df6d4c613af69fd9f80f2135ccb5056e588253381b80cc78fdc19684b7f4f53", + "prepared_review_unit_hash": "dc525243b0ba7a56c2544a691fc2f3adb2c68022740f65cd72cae25995e216a5", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 33, - "serialized_chars": 37092, - "estimated_input_tokens": 9273, - "estimated_eval_tokens": 18802 + "serialized_chars": 37608, + "estimated_input_tokens": 9402, + "estimated_eval_tokens": 19060 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is very noisy: most items are distinct model-, processor-, quantization-, or trainer-specific bugs, not duplicates of one underlying defect. No soft-edge pair looks safe to merge into a single canonical artifact.", - "confidence": 0.88, - "canonical_issue_reason": "Issue 45701 is the broadest active report in the set and the best umbrella anchor for the cluster's recurring version/regression theme, with the most general user-facing symptom among the candidates.", + "summary": "This is a heterogeneous cluster of Transformers bug reports, mostly around model loading, offloading, quantization, multimodal processors, and a few trainer/config regressions. None of the soft pairs look like the same underlying bug or change, so I would not merge any of them as duplicates.", + "confidence": 0.89, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "45701 is the strongest standalone issue to keep as the canonical reference because it is open, broadly phrased, and most likely to attract related regression reports.", + "best_issue_reason": "issue:43873 is the broadest and most central report in the cluster: it is open, has inbound references, and covers the shared offloading/quantization failure mode that several other issues only loosely resemble.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43526", "right": "issue:43883", "accept": false, - "reason": "Different subsystems and symptoms: image label reduction bug vs missing tied-weights metadata on a causal LM." - }, - { - "left": "issue:44466", - "right": "issue:45357", - "accept": false, - "reason": "Both involve save_pretrained, but one is a device-dependent lm_head serialization issue and the other is incorrect visual encoder keys in Qwen3.5." + "reason": "Different bugs: one is a BeitImageProcessorFast label-reduction bug, the other is a Molmo tied-weights attribute error." }, { "left": "issue:43454", "right": "issue:43720", "accept": false, - "reason": "Different failure modes: tied lm_head weights on AyaVision vs packed-weight unpacking during accelerate loading for BitNet." + "reason": "Different model paths and failures: AyaVision lm_head tying vs BitNet packed-weight unpacking during accelerate loading." }, { "left": "issue:43749", "right": "issue:44534", "accept": false, - "reason": "FSDP CPU-efficient loading and non-persistent buffer initialization are separate bugs with different code paths." + "reason": "Both mention loading/runtime stability, but the concrete problems differ: FSDP CPU RAM-efficient loading vs non-persistent buffers being filled with junk." }, { "left": "issue:41720", "right": "issue:43322", "accept": false, - "reason": "Both are load-time failures, but they affect different models and different errors (cuda assert vs segfault)." + "reason": "Unrelated symptoms and models: Qwen3 auto device mapping CUDA assert vs Llava Next segmentation fault." }, { "left": "issue:43881", "right": "issue:44991", "accept": false, - "reason": "Unrelated model loading vs tokenizer loading failures; no shared concrete code path." + "reason": "Different scope: a specific glm-4v-9b load failure vs a tokenizer-loading regression for EMBEDDIA/est-roberta." }, { "left": "issue:43296", - "right": "issue:44991", + "right": "issue:44877", "accept": false, - "reason": "PaddleOCR-VL loading in vLLM and est-roberta tokenizer loading are unrelated issues." + "reason": "Different root causes: PaddleOCR-VL load failure in vLLM vs strict config blocking granite_speech config loading." }, { "left": "issue:41720", "right": "issue:44977", "accept": false, - "reason": "Same model family, but one is auto device-mapping load failure and the other is flash-attention generation regression." + "reason": "Both involve Qwen3.x, but one is an auto device mapping CUDA assert and the other is a flash-attention generation bug." }, { "left": "issue:44008", "right": "issue:45823", "accept": false, - "reason": "Different Gemma models and different bugs: variable-name collision in forward vs device mismatch in PLE." + "reason": "Both are Gemma-related multimodal issues, but one is a variable-name collision in Gemma3nModel.forward and the other is a PLE device mismatch in Gemma4." }, { "left": "issue:42617", "right": "issue:42915", "accept": false, - "reason": "3d_parallel.py runtime failure and Qwen3Moe FP8 loading failure are unrelated." + "reason": "Different failure modes and models: 3d_parallel.py execution problem vs Qwen3Moe failing with FineGrainedFP8Config." }, { "left": "issue:40990", "right": "issue:45701", "accept": false, - "reason": "High perplexity on GPT-OSS and tokenization changes are not the same bug without evidence tying the symptom to tokenizer regression." + "reason": "The titles are too general to match: high perplexity on GPT-OSS/WikiText-2 vs a broad tokenization-version complaint." }, { - "left": "issue:44186", - "right": "issue:45072", + "left": "issue:43404", + "right": "issue:45357", "accept": false, - "reason": "Tokenizer crash on NER/batched padding is unrelated to dtype mismatch in bfloat16 inference." + "reason": "Both concern multimodal weight handling, but one is untied lm_head weights in Mistral3ForConditionalGeneration and the other is incorrect visual encoder keys on save_pretrained for Qwen3.5." }, { - "left": "issue:43404", - "right": "issue:45357", + "left": "issue:44186", + "right": "issue:45072", "accept": false, - "reason": "Both concern model serialization/tied weights, but they describe different models and different broken artifacts." + "reason": "Different classes of bugs: tokenizer crashes on NER/batched padding vs dtype mismatches in SwitchTransformers and TimmWrapperModel under bf16 inference." }, { "left": "issue:45632", "right": "issue:45684", "accept": false, - "reason": "Cache-path collision for trust_remote_code and permission propagation in save_pretrained are separate filesystem concerns." + "reason": "One is a cache-path collision for trust_remote_code; the other is save_pretrained copying read-only permissions." }, { "left": "issue:43577", "right": "issue:43708", "accept": false, - "reason": "BLIP2 dtype propagation and Trainer checkpoint step calculation are entirely different bugs." + "reason": "Different problems: Blip2 dtype not propagating to submodules vs resume_from_checkpoint miscomputing max_steps after batch-size changes." }, { "left": "issue:44387", "right": "issue:45704", "accept": false, - "reason": "Quantization memory growth and an apex RMSNorm memory leak are different performance issues with different causes." + "reason": "Both mention memory, but one is int4 quantization OOM in Transformers 5.x and the other is a T5 RMSNorm memory leak from apex." }, { "left": "issue:44423", "right": "issue:45406", "accept": false, - "reason": "Both affect serve/multimodal flows, but they fail on different missing attributes and different processor code paths." + "reason": "Both are serve-related multimodal crashes, but the concrete errors differ: AttributeError on str.to vs Gemma4Processor missing _tokenizer." }, { "left": "issue:43873", "right": "issue:44756", "accept": false, - "reason": "Quantization offload behavior and mmap OOM on Strix Halo are not the same underlying defect." + "reason": "Related offloading/memory theme, but not the same bug: general quantization offloading behavior vs a Strix Halo mmap OOM workaround." }, { "left": "issue:36331", "right": "issue:42617", "accept": false, - "reason": "Trainer compute_loss signature break and 3d_parallel.py execution failure are unrelated." + "reason": "Completely different areas: Trainer.compute_loss signature regression vs a 3d_parallel.py runtime failure." }, { "left": "issue:44877", "right": "issue:44991", "accept": false, - "reason": "Strict config loading for granite_speech and tokenizer loading for est-roberta are distinct problems." + "reason": "Different kinds of loading regressions: strict config rejection vs tokenizer loading failure for a specific model." }, { - "left": "issue:36331", - "right": "issue:42915", + "left": "issue:43550", + "right": "issue:43720", "accept": false, - "reason": "CustomTrainer kwargs regression and Qwen3Moe FP8 failure are unrelated trainer/model bugs." + "reason": "Different concrete failures: Bamba torch.compile with SDPA vs BitNet packed-weight unpacking during accelerate loading." }, { - "left": "issue:43550", - "right": "issue:43720", + "left": "issue:36331", + "right": "issue:42915", "accept": false, - "reason": "torch.compile SDPA failure in Bamba and packed-weight loading failure in BitNet are different code paths." + "reason": "Unrelated trainer API breakage vs Qwen3Moe FP8 loading failure." }, { "left": "issue:43450", "right": "issue:45701", "accept": false, - "reason": "Batched video shape bug is unrelated to tokenization/version behavior." + "reason": "Video processor batched-shape bug does not match a broad tokenization version-change report." }, { "left": "issue:44610", "right": "issue:44898", "accept": false, - "reason": "Both are image-size related, but they affect different processors/models and different normalization/resize logic." + "reason": "Both are image-size/preprocessing issues, but the models and root causes differ: OmDet-Turbo processor output size mismatch vs Perceiver interpolation failure." }, { "left": "issue:44186", "right": "issue:44610", "accept": false, - "reason": "NER tokenizer crashes and OmDet-Turbo image size mismatch are unrelated." + "reason": "Different domains: LayoutLMv2 tokenizer crash on NER/batched padding vs OmDet-Turbo processor producing the wrong image resolution." }, { "left": "issue:44479", "right": "issue:44918", "accept": false, - "reason": "Video-input regression in Qwen models and SFT embedding unpacking are different issues." + "reason": "Both involve Qwen3.5-family issues, but one is a video-input regression and the other is unpacking Qwen3.5 input embeddings with trl SFT trainer." }, { "left": "issue:41720", "right": "issue:42617", "accept": false, - "reason": "Different models and unrelated failures: CUDA assert during auto device mapping vs running 3d_parallel.py." + "reason": "Different models and failure surfaces: Qwen3 device-mapping CUDA assert vs inability to run 3d_parallel.py." }, { "left": "issue:43550", "right": "issue:43606", "accept": false, - "reason": "Different failures: compile/SDPA regression vs CPU-offload device mismatch." + "reason": "Both are device/compile-related, but one is Bamba SDPA under torch.compile and the other is bark-small CPU offload device mismatch." }, { "left": "issue:43316", "right": "issue:44913", "accept": false, - "reason": "Config API discrepancy in Gemma3TextConfig and GPTNeoX rotary_pct reload bug are unrelated." + "reason": "Different API issues: Gemma3TextConfig discrepancy vs GPTNeoXConfig rotary_pct not persisting on reload." }, { "left": "issue:42915", "right": "issue:43317", "accept": false, - "reason": "Qwen3Moe FP8 config failure and dequantized model offload failure are different loading paths." + "reason": "Different loading failures: Qwen3Moe FineGrainedFP8Config vs dequantized model failing with device_map=auto on GPU+CPU offload." }, { "left": "issue:43295", "right": "issue:45701", "accept": false, - "reason": "Processor.tokenizer regression in multimodal custom code and generic tokenization-change report are not the same concrete bug." + "reason": "A regression in processor.tokenizer/image handling is unrelated to a generic tokenization-version complaint." }, { "left": "issue:41720", "right": "issue:45161", "accept": false, - "reason": "Auto device mapping CUDA assert and TP-only GPT-OSS MoE issue are separate distributed/model-loading problems." + "reason": "Both touch Qwen/GPT-OSS-style MoE topics, but one is a CUDA assert on auto device mapping and the other is a TP-only problem with GPT-OSS MoE." }, { "left": "issue:43526", "right": "issue:43881", "accept": false, - "reason": "ImageProcessorFast label reduction and glm-4v-9b load failure are unrelated." + "reason": "Unrelated bugs: BeitImageProcessorFast label reduction vs glm-4v-9b model loading failure." + }, + { + "left": "issue:44671", + "right": "issue:44743", + "accept": false, + "reason": "Both are Qwen-related language-model bugs, but one is CamemBERT masked-LM predictions and the other is recurrent-state reset behavior in modular_qwen3_5.py." } ] }, @@ -4163,14 +4133,14 @@ "issue:44186", "issue:44361", "issue:44368", + "issue:44451", "issue:44492", "issue:44509", "issue:44514", "issue:44545", "issue:44625", - "issue:44671", - "issue:44743", "issue:44898", + "issue:44991", "issue:45059", "issue:45072", "issue:45081", @@ -4186,7 +4156,6 @@ ], "soft_pairs": [ "issue:45081|issue:45812", - "issue:44671|issue:44743", "issue:43531|issue:43572", "issue:43540|issue:43782", "issue:36010|issue:42617", @@ -4197,8 +4166,8 @@ "issue:45406|issue:45464", "issue:33357|issue:35707", "issue:43421|issue:45701", - "issue:41720|issue:43746", "issue:43606|issue:43720", + "issue:41720|issue:43746", "issue:30064|issue:34634", "issue:43550|issue:43653", "issue:44492|issue:44509", @@ -4209,223 +4178,224 @@ "issue:43531|issue:45759", "issue:34634|issue:41720", "issue:34567|issue:35707", - "issue:44625|issue:45479", "issue:43421|issue:43927", + "issue:44625|issue:45479", "issue:44368|issue:45357", + "issue:44898|issue:45072", "issue:43782|issue:45381", - "issue:44898|issue:45072" + "issue:44451|issue:44991" ], - "prepared_review_unit_hash": "457a3b71e8f538d55a8247f428160c98327e2f98311d318ad541669fd00ffef7", + "prepared_review_unit_hash": "f06099e90e732cb1aedd024ea9552bf398c0abb5bd60ea234476addc5ae4c003", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36688, - "estimated_input_tokens": 9172, - "estimated_eval_tokens": 18600 + "serialized_chars": 36658, + "estimated_input_tokens": 9165, + "estimated_eval_tokens": 18586 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly unrelated issue set; only one pair is a clear duplicate. The rest share broad subsystem/theme similarity but describe different failure modes or code paths.", + "summary": "Most pairs are unrelated. The only likely duplicate in this packet is issue:44514 / issue:44545: they have identical titles and very similar timing, so 44514 is the best canonical issue, but the packet only supports a conservative title-level duplicate call.", "confidence": 0.92, - "canonical_issue_reason": "Issue 45812 is the strongest representative overall: it is open, has multiple inbound references, and describes a broad, high-impact tokenizer regression affecting all Granite models.", + "canonical_issue_reason": "issue:44514 is the earlier of the two near-identical reports, and both items share the same crash title, making it the best canonical issue candidate on the evidence shown.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45812 is the best global issue candidate because it appears central, user-impacting, and widely referenced; it is more representative than the narrow duplicate pair.", + "best_issue_reason": "issue:44514 is the strongest issue candidate because it is the earlier report in the likely duplicate pair and the titles match exactly; the packet does not provide stronger body-level proof.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45081", "right": "issue:45812", "accept": false, - "reason": "Both involve tokenization-related regressions, but one is a Mistral regex patch crash and the other is Granite token ID corruption; different models and code paths." - }, - { - "left": "issue:44671", - "right": "issue:44743", - "accept": false, - "reason": "CamemBERT masked-LM prediction errors and Qwen3.5 recurrent-state reset are unrelated bugs in different model families." + "reason": "Different tokenizer bugs in different models: Mistral regex patch crash vs Granite token ID regression." }, { "left": "issue:43531", "right": "issue:43572", "accept": false, - "reason": "Qwen3-MoE sliding window behavior and StableLmConfig pad_token_idx regression are separate config/model issues." + "reason": "Unrelated config issues: Qwen3-MoE sliding_window behavior vs StableLmConfig missing pad_token_idx." }, { "left": "issue:43540", "right": "issue:43782", "accept": false, - "reason": "Video-input processing failure in Qwen3OmniMoe is not the same as weight_only loading failure in Qwen3VL." + "reason": "Different Qwen3 multimodal failures with different symptoms and code paths." }, { "left": "issue:36010", "right": "issue:42617", "accept": false, - "reason": "GenerationMixin import error and failure to run a 3D parallel script are unrelated." + "reason": "ImportError for GenerationMixin is unrelated to a 3d_parallel.py runtime failure." }, { "left": "issue:43421", "right": "issue:45812", "accept": false, - "reason": "Runtime special-token post-processing update and Granite token ID regression are different tokenizer lifecycle bugs." + "reason": "Runtime tokenizer post-processor updates are not the same bug as Granite token ID regression." }, { "left": "issue:44186", "right": "issue:44361", "accept": false, - "reason": "LayoutLMv2 NER/padding crash and MLukeTokenizer AttributeError are distinct tokenizer issues." + "reason": "Different tokenizer crashes affecting different models and tasks." }, { "left": "issue:42915", "right": "issue:43322", "accept": false, - "reason": "Qwen3Moe FP8 failure and Llava Next segfault have different triggers and model paths." + "reason": "Qwen3Moe FP8 config failure and Llava Next segmentation fault are distinct issues." }, { "left": "issue:43388", "right": "issue:45059", "accept": false, - "reason": "Tuple-label metric gathering bug and SAM3 text/box weirdness are unrelated training/evaluation problems." + "reason": "Different bugs: metric gathering label truncation vs SAM3 text/bounding-box behavior." }, { "left": "issue:45406", "right": "issue:45464", "accept": false, - "reason": "Gemma4Processor missing _tokenizer in serve is a server-side processor bug, not the Qwen3.5 streaming API failure." + "reason": "Gemma4Processor missing _tokenizer in serve is unrelated to Qwen3.5 streaming inference failure." }, { "left": "issue:33357", "right": "issue:35707", "accept": false, - "reason": "MacOS CLIP bus error and progressive generation with inputs_embeds/past_key_values are different runtime failures." + "reason": "MacOS bus error with a CLIP model is unrelated to progressive generation with inputs_embeds and past_key_values." }, { "left": "issue:43421", "right": "issue:45701", "accept": false, - "reason": "Both mention tokenization, but one is about special-token post-processing and the other is a broader version-dependent tokenization change." + "reason": "Broad tokenization-version behavior is not the same as a specific post-processor update bug." }, { - "left": "issue:41720", - "right": "issue:43746", + "left": "issue:43606", + "right": "issue:43720", "accept": false, - "reason": "Qwen3 auto device mapping CUDA assert and GraniteSpeech local checkpoint loading with PEFT adapters are unrelated." + "reason": "CPU offload device mismatch and BitNet packed-weight unpacking are different failure mechanisms." }, { - "left": "issue:43606", - "right": "issue:43720", + "left": "issue:41720", + "right": "issue:43746", "accept": false, - "reason": "CPU offload device mismatch in Bark-small and BitNet packed-weight unpacking during accelerate loading are different loading bugs." + "reason": "Qwen3 auto-device-map cuda assert is unrelated to GraniteSpeech PEFT checkpoint loading." }, { "left": "issue:30064", "right": "issue:34634", "accept": false, - "reason": "Image Processor void segmentation maps and BarkProcessor voice_preset are unrelated processor issues." + "reason": "Void segmentation map processing and Bark voice_preset handling are unrelated." }, { "left": "issue:43550", "right": "issue:43653", "accept": false, - "reason": "torch.compile SDPA failure in Bamba-9B-v2 and BigBirdTokenizer mask token registration are different subsystems." + "reason": "torch.compile SDPA failure is unrelated to BigBirdTokenizer special-token registration." }, { "left": "issue:44492", "right": "issue:44509", "accept": false, - "reason": "A typo in cache strategies and outdated docs mentioning removed pipeline tasks are not the same underlying bug." + "reason": "A cache-strategy typo is not the same as stale docs for removed pipeline tasks." }, { "left": "issue:43873", "right": "issue:45704", "accept": false, - "reason": "Quantization/offloading behavior and T5 apex.FusedRMSNorm memory leak are separate issues." + "reason": "Quantization/offloading behavior and an apex RMSNorm memory leak are unrelated." }, { "left": "issue:44514", "right": "issue:44545", "accept": true, - "reason": "Same title, same processor method, same batched padding=False crash; this is a clear duplicate pair." + "reason": "The titles are identical and the reports are adjacent, so they very likely describe the same crash; the packet only supports a conservative duplicate call." }, { "left": "issue:43606", "right": "issue:43653", "accept": false, - "reason": "CPU offload device mismatch in Bark-small and BigBirdTokenizer special-token registration are unrelated." + "reason": "Different subsystems and failure modes: CPU offload device mismatch vs tokenizer special-token registration." }, { "left": "issue:42490", "right": "issue:43329", "accept": false, - "reason": "Trailing-slash local model import warning and undefined video-branch variables in multimodal token counting are different bugs." + "reason": "Trailing-slash import warning is unrelated to an undefined helper in multimodal token counting." }, { "left": "issue:43531", "right": "issue:45759", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and AutoModelForCausalLM config unwrapping for Qwen 3.5/3.6 multimodal configs are distinct." + "reason": "Sliding-window behavior in Qwen3-MoE is not the same as from_config failing to unwrap text_config." }, { "left": "issue:34634", "right": "issue:41720", "accept": false, - "reason": "BarkProcessor voice_preset and Qwen3 device-mapping CUDA assert are unrelated." + "reason": "BarkProcessor voice_preset and Qwen3 auto device mapping errors are unrelated." }, { "left": "issue:34567", "right": "issue:35707", "accept": false, - "reason": "TrainerState token counting and progressive generation cache handling are different areas." + "reason": "TrainerState token counter issues are unrelated to generation with inputs_embeds and past_key_values." }, { - "left": "issue:44625", - "right": "issue:45479", + "left": "issue:43421", + "right": "issue:43927", "accept": false, - "reason": "Qwen3.5 num_labels propagation and zero-loss classification with num_labels=1 are not the same defect." + "reason": "Tokenizer backend post-processor updates and DiaConfig save/load token IDs are different bugs." }, { - "left": "issue:43421", - "right": "issue:43927", + "left": "issue:44625", + "right": "issue:45479", "accept": false, - "reason": "Special-token post-processor update and DiaConfig custom token ID loss on save/load are related to token IDs but not the same bug." + "reason": "Config propagation of num_labels is not the same as degenerate zero loss with num_labels=1." }, { "left": "issue:44368", "right": "issue:45357", "accept": false, - "reason": "Tie-word-embeddings warning in Qwen3.5 fine-tuning and incorrect visual encoder keys on save_pretrained are different config persistence issues." + "reason": "A tie_word_embeddings warning is unrelated to incorrect visual encoder keys during save_pretrained." + }, + { + "left": "issue:44898", + "right": "issue:45072", + "accept": false, + "reason": "Perceiver interpolation at non-default resolution and bfloat16 dtype mismatches are distinct failures." }, { "left": "issue:43782", "right": "issue:45381", "accept": false, - "reason": "Qwen3VL weight_only loading error and qwen2.5-vl video vision_position_ids mismatch are different multimodal bugs." + "reason": "Qwen3VL weight_only loading error and qwen2.5-vl video vision_position_ids bug are different code paths." }, { - "left": "issue:44898", - "right": "issue:45072", + "left": "issue:44451", + "right": "issue:44991", "accept": false, - "reason": "Perceiver interpolation failure and dtype mismatches in SwitchTransformers/TimmWrapperModel are separate inference issues." + "reason": "Different tokenizer-loading regressions affecting different models." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The assessment is grounded in the packet and stays conservative overall. The canonical issue choice is well supported by the title and inbound references for issue 45812, and the soft-edge verdicts mostly avoid overclaiming. The only accepted pair, 44514/44545, is clearly duplicate from the matching title/function/crash description." + "accept": false, + "feedback": "The duplicate-pair conclusion for issue:44514 / issue:44545 is plausible from the titles, but the writeup overstates what is supported by the packet. Phrases like \"precise repro,\" \"narrow code path,\" and \"exact duplicate\" imply body-level evidence that is not shown here. Keep the conclusion more conservative: the titles are nearly identical and the pair looks very likely to be the same bug, but the packet alone does not prove the full repro details." }, "evaluator_used": true, - "retried": false, + "retried": true, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -4480,7 +4450,6 @@ "issue:44918", "issue:44936", "issue:44987", - "issue:44991", "issue:45042", "issue:45081", "issue:45356", @@ -4493,7 +4462,6 @@ "issue:45923" ], "soft_pairs": [ - "issue:44451|issue:44991", "issue:44077|issue:44625", "issue:44987|issue:45042", "issue:45759|issue:45859", @@ -4523,239 +4491,230 @@ "issue:43540|issue:44008", "issue:44220|issue:44661", "issue:30064|issue:36010", - "issue:43618|issue:43697", - "issue:43653|issue:43720" + "issue:43653|issue:43720", + "issue:43618|issue:43697" ], - "prepared_review_unit_hash": "9d2c0141deffb3d9755e9ea96936de89a73b183d3155d4b47fe5a55a45331852", + "prepared_review_unit_hash": "a40f43c9a5ba1917d9ca968e9bb5a156e672bfe1cbef09e806c3fa04ddfb9df9", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 37508, - "estimated_input_tokens": 9377, - "estimated_eval_tokens": 19010 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36639, + "estimated_input_tokens": 9160, + "estimated_eval_tokens": 18576 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous: nearly all pairs are different model/config/tokenizer bugs with only superficial similarity. The only plausible duplicate subgroup is the Mistral/Kimi tokenizer regex patch area (_patch_mistral_regex / fix_mistral_regex).", - "confidence": 0.86, - "canonical_issue_reason": "No single issue cleanly represents the whole cluster; if a representative is needed, issue 45081 is the best anchor because it is the most concrete manifestation of the shared tokenizer-regex failure mode with 45356.", + "summary": "This cluster is a heterogeneous mix of unrelated Transformer bugs and regressions; the soft pairs mostly share only broad subsystem or keyword similarity, not the same concrete failing code path. I did not find any pair that should be merged as duplicates.", + "confidence": 0.92, + "canonical_issue_reason": "issue:44843 is the clearest, most actionable bug report here: it names a specific failing path (_patch_mistral_regex), an explicit offline-mode regression, and a reproducible symptom.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45081 is the best issue candidate because it describes a specific, reproducible tokenizer-loading crash in the shared fix_mistral_regex path, which is the only clearly duplicated theme here.", + "best_issue_reason": "issue:44843 is the strongest standalone issue in the set because it is narrowly scoped, reproducible, and points to a concrete code path with a likely fix surface.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44451", - "right": "issue:44991", - "accept": false, - "reason": "Both are tokenizer/model loading failures, but for different models and likely different root causes; not the same bug." - }, { "left": "issue:44077", "right": "issue:44625", "accept": false, - "reason": "Unrelated model-specific config bugs: PatchTST post-init vs Qwen3.5 num_labels propagation." + "reason": "Different model families and failures: PatchTSMixer post-init handling vs Qwen3.5 config propagation." }, { "left": "issue:44987", "right": "issue:45042", "accept": false, - "reason": "Different problems: loading a specific model vs PIL image processors incorrectly depending on torchvision." + "reason": "One is a model-loading regression; the other is an image-processor torchvision dependency issue." }, { "left": "issue:45759", "right": "issue:45859", "accept": false, - "reason": "Both involve Qwen 3.5, but one is config unwrapping and the other is CPU-offload behavior in rotary embedding; distinct code paths." + "reason": "Both involve Qwen 3.x, but one is config unwrapping for from_config and the other is CPU-offload incompatibility in rotary embedding." }, { "left": "issue:43531", "right": "issue:44589", "accept": false, - "reason": "Qwen3-MoE sliding window vs Float8 storage loading; unrelated failures." + "reason": "Unrelated symptoms: sliding-window behavior in Qwen3-MoE vs a Float8 storage loading error." }, { "left": "issue:39692", "right": "issue:44610", "accept": false, - "reason": "Doc/example mismatch in SigLIP2 vs OmDet-Turbo processor/model size mismatch; not the same issue." + "reason": "Both are multimodal/image-adjacent, but one is SigLIP2 docs/example errors and the other is an OmDet-Turbo processor/model size mismatch." }, { "left": "issue:44062", "right": "issue:44451", "accept": false, - "reason": "Tokenizer AddedToken argument handling vs loading a specific tokenizer/model; too different to be duplicates." + "reason": "Tokenizer construction error vs a specific model-load failure for ScandiBERT; no shared failing path is evident." }, { "left": "issue:45704", "right": "issue:45923", "accept": false, - "reason": "Apex RMSNorm memory leak in T5 is unrelated to a missing supports_gradient_checkpointing flag." + "reason": "T5 RMSNorm/apex leak is unrelated to a trust_remote_code gradient-checkpointing flag omission." }, { "left": "issue:43421", "right": "issue:43653", "accept": false, - "reason": "Both touch tokenizer special tokens, but one is runtime post-processor updating and the other is BigBird mask-token registration/decode behavior." + "reason": "Both touch tokenizer special tokens, but one is a backend post-processor update bug and the other is a BigBirdTokenizer registration/decode issue." }, { "left": "issue:45588", "right": "issue:45759", "accept": false, - "reason": "Flash-attention crash on sink-less models vs Qwen composite config unwrapping; different bugs." + "reason": "Flash-attention s_aux handling is unrelated to Qwen config unwrapping." }, { "left": "issue:45663", "right": "issue:45923", "accept": false, - "reason": "Gemma-4 FSDP2 shared_kv_states bug is unrelated to the trust_remote_code gradient-checkpointing flag issue." + "reason": "Gemma-4 FSDP2 shared_kv_states bug is a different training-path failure than the Nemotron gradient-checkpointing flag issue." }, { "left": "issue:44918", "right": "issue:44936", "accept": false, - "reason": "Both are trainer-related, but one is input-embedding unpacking and the other is evaluate-after-train state handling; not the same defect." + "reason": "Both involve trainer workflows, but one is SFT embedding unpacking and the other is evaluate-after-train state handling." }, { "left": "issue:42886", "right": "issue:44843", "accept": false, - "reason": "Both concern offline tokenizer loading, but 44843 is specifically the mistral regex patch calling model_info(); 42886 is broader cache-loading failure." + "reason": "Both mention offline/tokenizer loading, but one is cache loading in offline mode and the other is an unconditional model_info call in _patch_mistral_regex." }, { "left": "issue:43381", "right": "issue:43844", "accept": false, - "reason": "Eval-mode gradient checkpointing limitation vs abnormal gradient growth under ZeRO-3; unrelated training issues." + "reason": "Gradient checkpointing eval-mode restriction is unrelated to DeepSpeed ZeRO-3 gradient explosion." }, { "left": "issue:43421", "right": "issue:44186", "accept": false, - "reason": "General TokenizersBackend post-processor/runtime special-token update bug vs LayoutLMv2Tokenizer crash on NER padding/truncation." + "reason": "Shared tokenizer theme only; the concrete bugs are different (runtime special-token updates vs LayoutLMv2 NER/padding crashes)." }, { "left": "issue:43994", "right": "issue:44671", "accept": false, - "reason": "Different models and symptoms: SigLIP2 nonsensical outputs vs CamemBERT masked-LM regression." + "reason": "Different models and outcomes: SigLIP2 nonsensical outputs vs CamemBERT masked-LM prediction regression." }, { "left": "issue:45704", "right": "issue:45859", "accept": false, - "reason": "RMSNorm memory leak in T5 vs Qwen3.5 CPU-offload rotary embedding issue; not duplicates." + "reason": "RMSNorm memory leak in T5 is unrelated to Qwen rotary embedding CPU-offload behavior." }, { "left": "issue:43525", "right": "issue:43531", "accept": false, - "reason": "Llama4 pad_token_id attribute error vs Qwen3-MoE sliding_window issue; separate config/model bugs." + "reason": "Both are Qwen-related, but one is a missing pad_token_id attribute and the other is a sliding_window issue." }, { "left": "issue:45081", "right": "issue:45356", - "accept": true, - "reason": "Both point to the same tokenizer regex patch path (_patch_mistral_regex / fix_mistral_regex) and appear to be variants of the same regression in tokenizer loading/codec handling." + "accept": false, + "reason": "Both involve fix_mistral_regex, but one is an AttributeError crash while the other is a codec-handling regression with a warning issue." }, { "left": "issue:41720", "right": "issue:45704", "accept": false, - "reason": "Qwen3 auto device-mapping CUDA assert is unrelated to the T5 RMSNorm memory leak." + "reason": "Qwen3 auto device mapping CUDA assert is unrelated to T5/apex RMSNorm memory leak." }, { "left": "issue:43746", "right": "issue:43873", "accept": false, - "reason": "PEFT/local checkpoint loading for GraniteSpeech vs quantization offloading behavior; different bugs." + "reason": "PEFT local-checkpoint loading failure is not the same bug as quantization offloading behavior." }, { "left": "issue:39692", "right": "issue:43873", "accept": false, - "reason": "SigLIP2 docs/model mismatch vs quantization offloading; no shared root cause." + "reason": "SigLIP2 doc/example errors and quantization offloading problems are different issue classes." }, { "left": "issue:40444", "right": "issue:43746", "accept": false, - "reason": "IterableDataset multi-image finetuning failure vs GraniteSpeech PEFT local checkpoint loading; unrelated." + "reason": "Multi-image IterableDataset finetuning failure is unrelated to PEFT-adapter local checkpoint loading." }, { "left": "issue:45356", "right": "issue:45907", "accept": false, - "reason": "Tokenizer regex regression vs list_repo_files revision forwarding in CI; different layers and symptoms." + "reason": "Kimi tokenizer codec/fix_mistral_regex regression is unrelated to list_repo_files revision forwarding." }, { "left": "issue:43596", "right": "issue:44220", "accept": false, - "reason": "DeepSpeed ZeRO-3/BertModel index error vs audio fbank feature extraction issue; unrelated." + "reason": "DeepSpeed zero3/BertModel indexing error is unrelated to the audio fbank feature extraction issue." }, { "left": "issue:42915", "right": "issue:43746", "accept": false, - "reason": "Qwen3Moe FineGrainedFP8Config failure vs GraniteSpeech PEFT checkpoint loading; different bugs." + "reason": "Qwen3Moe FP8 config failure and GraniteSpeech PEFT checkpoint loading are different model-path bugs." }, { "left": "issue:43720", "right": "issue:43819", "accept": false, - "reason": "BitNet packed-weight loading issue vs DAC.from_latents/STE mismatch; distinct code paths." + "reason": "Packed-weight loading in BitNet is unrelated to DAC.from_latents / missing STE mismatch." }, { "left": "issue:43540", "right": "issue:44008", "accept": false, - "reason": "Video-input processing failure in Qwen3OmniMoe vs Gemma 3n audio mask name collision; unrelated." + "reason": "Video input processing error in Qwen3OmniMoe is unrelated to Gemma 3n audio_mel_mask variable collision." }, { "left": "issue:44220", "right": "issue:44661", "accept": false, - "reason": "Audio feature extraction vs add-new-model-like tokenizer mapping bug; no common underlying defect." + "reason": "Audio feature extraction bug and TOKENIZER_MAPPING_NAMES add-new-model-like failure are unrelated." }, { "left": "issue:30064", "right": "issue:36010", "accept": false, - "reason": "Void segmentation map processing and GenerationMixin import error are entirely unrelated." + "reason": "Image segmentation map processing failure and a GenerationMixin import error do not share the same code path." }, { - "left": "issue:43618", - "right": "issue:43697", + "left": "issue:43653", + "right": "issue:43720", "accept": false, - "reason": "CLIPOutput attentions assignment vs RTDetrV2 output differences in v5; different model regressions." + "reason": "BigBirdTokenizer special-token registration and BitNet packed-weight loading are unrelated issues." }, { - "left": "issue:43653", - "right": "issue:43720", + "left": "issue:43618", + "right": "issue:43697", "accept": false, - "reason": "BigBirdTokenizer special-token decode bug vs BitNet packed-weight loading; unrelated." + "reason": "CLIPOutput attentions assignment regression is unrelated to RTDetrV2 output differences in v5." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is well grounded in the packet and stays conservative. The only accepted soft-edge pair, issue:45081 with issue:45356, is plausibly the same Mistral/Kimi tokenizer-regex regression. The other pairwise rejections are appropriately cautious and do not overstate duplication from surface similarity alone." - }, - "evaluator_used": true, + "evaluator_result": null, + "evaluator_used": false, "retried": false, - "accepted_nontrivial_soft_edge": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -4833,8 +4792,8 @@ "issue:44186|issue:44898", "issue:44843|issue:45081", "issue:43653|issue:43819", - "issue:43421|issue:44811", "issue:43643|issue:44877", + "issue:43421|issue:44811", "issue:43577|issue:43818", "issue:43540|issue:45381", "issue:43782|issue:44977", @@ -4842,196 +4801,196 @@ "issue:44448|issue:44671", "issue:45584|issue:45834", "issue:39692|issue:45834", - "issue:44561|issue:45020", "issue:43825|issue:43994", + "issue:44561|issue:45020", "issue:30064|issue:36331", "issue:41720|issue:44991", "issue:44821|issue:45084" ], - "prepared_review_unit_hash": "c1f5c1e9a08c7c0a741fccae8e5f394592acf581dbd95bd959feb1316c69b006", + "prepared_review_unit_hash": "ff8cb69162c4eec612d7a438b59867e19c8dfc4e7bfdc671cdc1ebfeeba13f1c", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 26, - "serialized_chars": 35296, - "estimated_input_tokens": 8824, - "estimated_eval_tokens": 17904 + "serialized_chars": 35290, + "estimated_input_tokens": 8823, + "estimated_eval_tokens": 17902 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a loose semantic grouping of mostly unrelated issue reports, with no clear duplicate pair. The best representative issue is the broad remote-code regression report, but the soft edges are not strong enough to merge any pair.", - "confidence": 0.19, - "canonical_issue_reason": "Issue 45020 is the broadest umbrella-style report in the set; it summarizes a general regression affecting remote_code models, whereas the other issues are narrower, model-specific bugs.", + "summary": "These issues are mostly heterogeneous model- or subsystem-specific regressions, not true duplicates. I would reject every soft pair; the closest umbrella issue in the set is the broad remote_code regression report, but even that is only a loose representative, not a duplicate of the others.", + "confidence": 0.42, + "canonical_issue_reason": "issue:45020 is the broadest umbrella in the set: it covers recent-version regressions affecting remote_code models, which is the closest common thread among a few tokenizer/config-loading reports.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45020 is the most suitable cluster representative because it is the most general and user-facing regression report. It is not a duplicate of the others, just the best umbrella issue if one must be chosen.", + "best_issue_reason": "issue:45020 is the best representative issue because it is the most general and likely to gather related remote_code breakages, even though the rest of the cluster is too diverse to be a single duplicate set.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45183", "right": "issue:45479", "accept": false, - "reason": "Different bugs: input/embedding validation vs. sequence-classification loss logic. Same broad area, but not the same concrete failure." + "reason": "Different bugs: input validation error messaging vs a sequence-classification zero-loss regression." }, { "left": "issue:43295", "right": "issue:45834", "accept": false, - "reason": "Both involve multimodal processors, but one is a processor.tokenizer regression and the other is an OCR long-input index error. Not the same issue." + "reason": "Processor/tokenizer regression is unrelated to Kosmos2.5 long-OCR index errors." }, { "left": "issue:43335", "right": "issue:43756", "accept": false, - "reason": "Unrelated model internals: sparse-layer construction vs. RoPE layer count mismatch. No shared code-path bug." + "reason": "SwitchTransformers sparse-layer construction and Smollm3 RoPE-layer dropping are unrelated model-specific issues." }, { "left": "issue:43716", "right": "issue:44008", "accept": false, - "reason": "Different subsystems and failures: image preprocessor dtype mismatch vs. Gemma3n variable-name collision in forward()." + "reason": "Mistral-3 dtype mismatch and Gemma3n variable-name collision are different failure modes in different code paths." }, { "left": "issue:43550", "right": "issue:45072", "accept": false, - "reason": "Both mention model execution, but one is torch.compile/SDPA on Bamba and the other is bf16 dtype mismatch in SwitchTransformers/TimmWrapperModel. Not mergeable." + "reason": "Both are model/runtime regressions, but they affect different models and different mechanisms (SDPA compile vs dtype mismatch)." }, { "left": "issue:43550", "right": "issue:43819", "accept": false, - "reason": "Distinct problems: compile+SDPA failure vs. DAC from_latents/forward mismatch with missing STE." + "reason": "Bamba torch.compile/SDPA failure is unrelated to DAC.from_latents missing STE parity." }, { "left": "issue:43950", "right": "issue:45987", "accept": false, - "reason": "One is a from_pretrained buffer-loading regression; the other is a StaticCache return-type contract bug. Different code paths." + "reason": "from_pretrained buffer corruption and StaticCache.get_seq_length returning a tensor are separate API regressions." }, { "left": "issue:44062", "right": "issue:44488", "accept": false, - "reason": "Tokenizer construction TypeError vs. a specific model failing to load. The titles are tokenizer-related but not the same underlying defect." + "reason": "AddedToken special-argument duplication is unrelated to loading cjvt/sleng-bert." }, { "left": "issue:43322", "right": "issue:45685", "accept": false, - "reason": "Segfault on Llava Next loading and an MPS histogram limitation in MoE are unrelated failures." + "reason": "Llava Next loading segfault and MPS histogram-on-int error are unrelated." }, { "left": "issue:44186", "right": "issue:44898", "accept": false, - "reason": "Tokenizer crash on NER/padding versus image-classification resolution handling. Different components and symptoms." + "reason": "LayoutLMv2 tokenizer padding/truncation crashes and Perceiver image-size failures are different model-specific bugs." }, { "left": "issue:44843", "right": "issue:45081", "accept": false, - "reason": "Both touch _patch_mistral_regex, but one is an offline network call regression and the other is an AttributeError on backend_tokenizer. Related area, not the same bug." + "reason": "Both touch _patch_mistral_regex, but one is an offline hub call bug and the other is a backend_tokenizer attribute crash; not the same underlying defect." }, { "left": "issue:43653", "right": "issue:43819", "accept": false, - "reason": "BigBird special-token registration issue is unrelated to DAC latent reconstruction mismatch." + "reason": "BigBirdTokenizer special-token registration and DAC STE mismatch are unrelated." }, { - "left": "issue:43421", - "right": "issue:44811", + "left": "issue:43643", + "right": "issue:44877", "accept": false, - "reason": "Post-processor refresh on special-token edits vs. Whisper batch_decode skip_special_tokens handling. Similar tokenization surface, different bug." + "reason": "Missing fields from AutoConfig with trust_remote_code and granite_speech strict-config loading are not the same concrete bug." }, { - "left": "issue:43643", - "right": "issue:44877", + "left": "issue:43421", + "right": "issue:44811", "accept": false, - "reason": "Both concern config loading, but one is missing fields with trust_remote_code and the other is strict config rejection for granite_speech. Not the same defect." + "reason": "Runtime post-processor updates for special tokens and Whisper batch_decode skip_special_tokens behavior are distinct tokenizer issues." }, { "left": "issue:43577", "right": "issue:43818", "accept": false, - "reason": "Blip2 dtype propagation bug and Video-LLaVA temporal-attention/weight-sharing issue are unrelated." + "reason": "Blip2 dtype preservation and Video-LLaVA temporal-attention/shared-weight problems are unrelated." }, { "left": "issue:43540", "right": "issue:45381", "accept": false, - "reason": "Both involve Qwen video inputs, but the concrete failures differ: ValueError in Qwen3OmniMoe vs. wrong vision_position_ids in qwen2.5-vl." + "reason": "Both mention video inputs, but Qwen3OmniMoe ValueError and qwen2.5-vl vision_position_ids are different bugs." }, { "left": "issue:43782", "right": "issue:44977", "accept": false, - "reason": "Qwen3VL weight_only loading error and Qwen3.5 flash-attention generation issue are separate code paths." + "reason": "Qwen3VL weight_only loading error and Qwen3.5 flash-attention generation issue are separate failures." }, { "left": "issue:43531", "right": "issue:43643", "accept": false, - "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to trust_remote_code config field loss." + "reason": "Sliding-window handling in Qwen3-MoE is unrelated to AutoConfig trust_remote_code field loss." }, { "left": "issue:44448", "right": "issue:44671", "accept": false, - "reason": "Both are output regressions, but on different models and likely different causes. Not enough to treat as the same bug." + "reason": "Version-specific output changes for different models do not indicate the same underlying bug." }, { "left": "issue:45584", "right": "issue:45834", "accept": false, - "reason": "Whisper empty-transcription generation failure and Kosmos2.5 OCR index error are different model-specific issues." + "reason": "Whisper empty-transcription generation and Kosmos2.5 OCR index errors are unrelated." }, { "left": "issue:39692", "right": "issue:45834", "accept": false, - "reason": "Documentation/example problems and quantization mismatch are unrelated to long OCR input indexing." + "reason": "Documentation/example errors for SigLIP2 do not match Kosmos2.5 long-input indexing failures." }, { - "left": "issue:44561", - "right": "issue:45020", + "left": "issue:43825", + "right": "issue:43994", "accept": false, - "reason": "Both mention remote_code, but 45020 is a broad umbrella regression while 44561 is a specific is_torch_fx_available removal breakage. Not the same concrete bug." + "reason": "Pipeline error-message wording and SigLIP2 nonsensical outputs are unrelated." }, { - "left": "issue:43825", - "right": "issue:43994", + "left": "issue:44561", + "right": "issue:45020", "accept": false, - "reason": "Pipeline translation-task messaging issue vs. SigLIP2 nonsensical outputs. Different failure modes and code paths." + "reason": "44561 is a specific trust_remote_code regression caused by a removed helper; 45020 is a broader umbrella report, but not the same concrete bug." }, { "left": "issue:30064", "right": "issue:36331", "accept": false, - "reason": "Image segmentation-map processing and CustomTrainer loss-signature mismatch are unrelated." + "reason": "ImageProcessor void-segmentation-map failure and CustomTrainer compute_loss signature breakage are unrelated." }, { "left": "issue:41720", "right": "issue:44991", "accept": false, - "reason": "A800 cudaErrorAssert with auto device mapping and tokenizer loading for EMBEDDIA/est-roberta are unrelated." + "reason": "Qwen3 auto device mapping cuda assert and tokenizer loading for EMBEDDIA/est-roberta are unrelated." }, { "left": "issue:44821", "right": "issue:45084", "accept": false, - "reason": "Loading AutoImageProcessor from URL and template-node compilation errors do not share an underlying bug." + "reason": "AutoImageProcessor URL loading and non-template-node compilation errors are different issues." } ] }, @@ -5108,8 +5067,8 @@ "issue:44186|issue:44442", "issue:44485|issue:44509", "issue:44479|issue:45834", - "issue:43577|issue:43582", "issue:43606|issue:45072", + "issue:43577|issue:43582", "issue:43825|issue:43827", "issue:44805|issue:44936", "issue:34634|issue:36010", @@ -5119,8 +5078,8 @@ "issue:43720|issue:45072", "issue:42175|issue:45081", "issue:43749|issue:44291", - "issue:45081|issue:45406", "issue:43650|issue:43906", + "issue:45081|issue:45406", "issue:43582|issue:44857", "issue:44060|issue:44368", "issue:43262|issue:43819", @@ -5128,200 +5087,200 @@ "issue:43526|issue:44554", "issue:34634|issue:36331", "issue:45459|issue:45803", - "issue:43742|issue:45020", "issue:43329|issue:45600", + "issue:43742|issue:45020", "issue:43653|issue:44912" ], - "prepared_review_unit_hash": "117e82d9fc6cec8357a07038c268bd98b9d340de03e930d785bb444139021311", + "prepared_review_unit_hash": "0f311b30d34efcc0c6b393942136dda291b94bca6c6023720d4170d57cf6dae1", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 27, - "serialized_chars": 35610, - "estimated_input_tokens": 8903, - "estimated_eval_tokens": 18062 + "serialized_chars": 35616, + "estimated_input_tokens": 8904, + "estimated_eval_tokens": 18064 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These soft pairs are mostly false positives: they share broad Transformers themes (v5 regressions, loading, tokenizers, multimodal models) but describe different concrete bugs, code paths, or model families. No PRs are present. The most representative issue is the broad remote-code regression report, but the cluster is too heterogeneous to treat as one duplicate set.", - "confidence": 0.72, - "canonical_issue_reason": "issue:45020 is the broadest issue in the set and best captures the recurring 'recent versions break loading/runtime behavior' theme, even though the overall cluster is not a true duplicate group.", + "summary": "This set is mostly a grab bag of unrelated Transformers issues. No soft edge looks like a true duplicate with the same concrete bug path; the closest pair is the v5 pipeline-removal messaging/docs pair, but I\u2019d still keep it split.", + "confidence": 0.84, + "canonical_issue_reason": "Issue 43825 is the most self-contained, actionable bug report in the set: a concrete runtime regression in the error message around removed pipeline tasks.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45020 is the best umbrella representative because it spans a wide class of version-induced breakages rather than a single model-specific symptom.", + "best_issue_reason": "43825 is the best representative issue because it is specific and reproducible; most other items are broader environment/model regressions or unrelated subsystems.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43644", "right": "issue:43749", "accept": false, - "reason": "Both are loading/runtime regressions, but one is junk non-persistent buffers and the other is FSDP CPU RAM efficient loading; different failure modes and code paths." + "reason": "Both concern loading/runtime behavior, but they are different failures: junk non-persistent buffers in v5 vs broken FSDP CPU RAM efficient loading." }, { "left": "issue:44186", "right": "issue:44442", "accept": false, - "reason": "Different tokenizer bugs on different models: LayoutLMv2 NER/padding crash vs FastSpeech2Conformer AutoTokenizer load failure." + "reason": "Different tokenizer paths and failures: LayoutLMv2 NER/padding crash vs AutoTokenizer loading FastSpeech2ConformerTokenizer." }, { "left": "issue:44485", "right": "issue:44509", "accept": false, - "reason": "One is a GLM-5 RoPE implementation discussion, the other is docs still mentioning removed pipeline tasks; same area only at a very high level." + "reason": "One is a GLM-5 RoPE implementation discussion, the other is docs still mentioning removed pipeline tasks; not the same bug/change." }, { "left": "issue:44479", "right": "issue:45834", "accept": false, - "reason": "Both concern multimodal inputs, but the concrete issues differ: Qwen video regression vs Kosmos2.5 OCR long-input index error." + "reason": "Different models and symptoms: video-input regression for Qwen VL models vs Kosmos2.5 OCR index error." }, { - "left": "issue:43577", - "right": "issue:43582", + "left": "issue:43606", + "right": "issue:45072", "accept": false, - "reason": "Blip2 dtype propagation and Apple Silicon caching allocator TypeError are unrelated bugs." + "reason": "Both are inference/load-time issues, but device mismatch on Bark CPU offload is unrelated to dtype mismatch in SwitchTransformers/TimmWrapperModel." }, { - "left": "issue:43606", - "right": "issue:45072", + "left": "issue:43577", + "right": "issue:43582", "accept": false, - "reason": "CPU offload device mismatch for Bark-small is unrelated to bfloat16 dtype mismatches in SwitchTransformers/TimmWrapperModel." + "reason": "Blip2 dtype propagation and Apple Silicon warmup TypeError are separate bugs in different code paths." }, { "left": "issue:43825", "right": "issue:43827", "accept": false, - "reason": "One is an error-message wording bug, the other is stale docs referencing removed pipelines; related topic but not the same defect." + "reason": "They are related by the v5 pipeline removals, but one is an error-message bug and the other is a docs cleanup; not the same concrete artifact." }, { "left": "issue:44805", "right": "issue:44936", "accept": false, - "reason": "Both are trainer/runtime failures, but they affect different workflows and symptoms; no evidence of a shared root cause." + "reason": "Different failures: mask/index shape mismatch vs trainer.evaluate() breaking after training." }, { "left": "issue:34634", "right": "issue:36010", "accept": false, - "reason": "BarkProcessor voice_preset failure and GenerationMixin import error are separate issues in different subsystems." + "reason": "BarkProcessor voice_preset failure is unrelated to the GenerationMixin import error." }, { "left": "issue:44671", "right": "issue:45701", "accept": false, - "reason": "CamemBERT masked-LM regressions and tokenization changes across versions are not the same underlying bug." + "reason": "CamemBERT masked-LM regression and a general tokenization-change report are not the same underlying bug." }, { "left": "issue:44610", "right": "issue:45072", "accept": false, - "reason": "OmDet-Turbo processor size mismatch and dtype mismatch bugs are unrelated." + "reason": "OmDet-Turbo processor image-size mismatch is unrelated to SwitchTransformers/TimmWrapperModel dtype mismatches." }, { "left": "issue:43782", "right": "issue:45325", "accept": false, - "reason": "Both are Qwen VL issues, but one is a from_pretrained weight_only load error and the other is a rope-index scaling bug." + "reason": "Qwen3VL weight_only loading error and Qwen2.5-VL rope indexing bug affect different code paths." }, { "left": "issue:43720", "right": "issue:45072", "accept": false, - "reason": "BitNet packed-weight loading failure is a different code path from generic inference dtype mismatches." + "reason": "Packed-weight unpacking during accelerate loading is a different issue than bfloat16 dtype mismatches." }, { "left": "issue:42175", "right": "issue:45081", "accept": false, - "reason": "Backend installation packaging and Mistral tokenizer regex patching are unrelated." + "reason": "Backend-install dependency problems are unrelated to a tokenizer regex patch crash." }, { "left": "issue:43749", "right": "issue:44291", "accept": false, - "reason": "Both are model-loading regressions, but FSDP RAM-efficient loading and init_empty_weights argument handling are distinct problems." + "reason": "Both involve model loading, but FSDP CPU RAM efficient loading and init_empty_weights/_is_hf_initialized are distinct failures." }, { - "left": "issue:45081", - "right": "issue:45406", + "left": "issue:43650", + "right": "issue:43906", "accept": false, - "reason": "Mistral tokenizer backend-tokenizer access and Gemma4Processor missing _tokenizer are separate processor/tokenizer failures." + "reason": "The titles do not describe the same bug; one is a vague request, the other is a reproduction of another issue." }, { - "left": "issue:43650", - "right": "issue:43906", + "left": "issue:45081", + "right": "issue:45406", "accept": false, - "reason": "One is a nonsensical/abnormal issue title, the other is a reproduction for a specific upstream bug; no duplicate relation." + "reason": "Tokenizer regex patch crash and Gemma4Processor missing _tokenizer in serve are separate processor/tokenizer bugs." }, { "left": "issue:43582", "right": "issue:44857", "accept": false, - "reason": "Apple Silicon allocator TypeError and CUDA float16 AMP crash are different platform-specific issues." + "reason": "Apple Silicon caching allocator warmup TypeError is unrelated to LwDetrImageLoss float16 AMP crash on CUDA." }, { "left": "issue:44060", "right": "issue:44368", "accept": false, - "reason": "Both involve tied-weights warnings, but one reports an incorrect tying bug and the other a user warning about tie_word_embeddings configuration." + "reason": "Both mention tied-weights warnings in Qwen models, but they describe different warning surfaces and likely different fixes; not enough to treat as the same bug." }, { "left": "issue:43262", "right": "issue:43819", "accept": false, - "reason": "Audio sampling-rate defaulting and DAC from_latents/forward mismatch are unrelated audio bugs." + "reason": "Audio chat-template sampling-rate defaulting and DAC.from_latents missing STE are unrelated audio-processing issues." }, { "left": "issue:43526", "right": "issue:43697", "accept": false, - "reason": "BeitImageProcessorFast label reduction and RTDetrV2 output drift are different vision-processing bugs." + "reason": "BeitImageProcessorFast label reduction and RTDetrV2 output differences are different model/processor bugs." }, { "left": "issue:43526", "right": "issue:44554", "accept": false, - "reason": "Image label reduction and MPS attention correctness are unrelated." + "reason": "No shared failure mode: label reduction bug vs MPS attention correctness issue." }, { "left": "issue:34634", "right": "issue:36331", "accept": false, - "reason": "BarkProcessor voice_preset and CustomTrainer compute_loss signature mismatch do not share a code path." + "reason": "BarkProcessor voice_preset and CustomTrainer.compute_loss signature mismatch are unrelated." }, { "left": "issue:45459", "right": "issue:45803", "accept": false, - "reason": "Both involve exception handling, but they are in different functions and solve different error-hiding problems." + "reason": "Both involve exception handling, but they are different code paths and different bugs; not the same underlying issue." }, { - "left": "issue:43742", - "right": "issue:45020", + "left": "issue:43329", + "right": "issue:45600", "accept": false, - "reason": "A specific MobileLLM key error is not the same as broad remote_code breakage." + "reason": "A multimodal token-count video-branch bug is unrelated to stale auto-mapping references for removed configs." }, { - "left": "issue:43329", - "right": "issue:45600", + "left": "issue:43742", + "right": "issue:45020", "accept": false, - "reason": "Video-token counting bug in multimodal processors and stale auto_mappings config references are unrelated." + "reason": "KeyError loading MobileLLM-125M and generic remote_code breakage are too broad to be the same concrete issue." }, { "left": "issue:43653", "right": "issue:44912", "accept": false, - "reason": "BigBirdTokenizer special-token registration and MXFP4 quantization loading are separate CI/runtime regressions." + "reason": "BigBirdTokenizer special-token registration and MXFP4 quantization fallback are unrelated bugs." } ] }, @@ -5402,8 +5361,8 @@ "issue:44589|issue:45042", "issue:43854|issue:44977", "issue:43819|issue:43927", - "issue:36010|issue:43645", "issue:39692|issue:44912", + "issue:36010|issue:43645", "issue:43257|issue:43329", "issue:43720|issue:44610", "issue:43550|issue:44912", @@ -5420,221 +5379,218 @@ "issue:43643|issue:45759", "issue:43329|issue:43450", "issue:43257|issue:43931", - "issue:44568|issue:45488", "issue:43606|issue:45859", + "issue:44568|issue:45488", "issue:44849|issue:45701" ], - "prepared_review_unit_hash": "56ab62806ed466e72188a1b627140b671d19eb7ab183687abd632854301d171f", + "prepared_review_unit_hash": "5a80a8dab22d785c987dba14b20ac4ddef28b362d3ffb306dbb4e10b8e91f194", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 29, - "serialized_chars": 36010, - "estimated_input_tokens": 9003, - "estimated_eval_tokens": 18262 + "serialized_chars": 36008, + "estimated_input_tokens": 9002, + "estimated_eval_tokens": 18260 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is still mostly noisy and spans several unrelated bug classes. The safest anchor remains the broad tokenizer/version-regression issue, while the video pair should not be merged as the same bug without stronger evidence.", - "confidence": 0.33, - "canonical_issue_reason": "issue:45701 is the broadest representative of the tokenizer/version-regression portion of the cluster, which is the most coherent subgroup here.", + "summary": "This is a loose lexical-similarity cluster, not a duplicate set: the issues span training state bugs, import/load failures, tokenizer regressions, multimodal processor problems, quantization/offload issues, and model-specific generation errors.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue stands out; the cluster is too heterogeneous and the title similarities are superficial rather than describing one underlying bug.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45701 is the best global issue anchor because it is the most general report and can loosely cover several tokenizer-regression variants, even though the cluster also contains unrelated loading, multimodal, and runtime bugs.", + "best_issue_reason": "If forced to pick a representative, issue:45701 is the broadest/open-ended title, but it is still too generic and not a reliable duplicate anchor for the rest of the cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45084", "right": "issue:45538", "accept": false, - "reason": "Different failure modes: compiler/template-node error vs. CLIPTokenizer max-length behavior." + "reason": "Different problems: one is a compile-time template-node error, the other is a tokenizer max-length constant issue." }, { "left": "issue:43638", "right": "issue:44936", "accept": false, - "reason": "A deepspeed zero3 Bert training crash is not the same bug as evaluate-after-train failure." + "reason": "Both mention training/evaluation failures, but one is a DeepSpeed ZeRO-3 indexing crash and the other is a post-train evaluate() failure; different code paths." }, { "left": "issue:43452", "right": "issue:44991", "accept": false, - "reason": "gguf_file loading and a tokenizer regression for a specific model are different code paths." + "reason": "Both are model-loading/tokenizer-related, but one concerns gguf_file handling in model/tokenizer loaders and the other is a tokenizer regression for a specific checkpoint." }, { "left": "issue:43746", "right": "issue:44811", "accept": false, - "reason": "PEFT/local-checkpoint loading for GraniteSpeech is unrelated to Whisper batch_decode special-token handling." + "reason": "One is about PEFT adapter local checkpoint loading; the other is Whisper processor batch_decode skipping special tokens." }, { "left": "issue:38175", "right": "issue:44610", "accept": false, - "reason": "Different models and distinct problems: zero probabilities vs. image-size mismatch." + "reason": "Different modalities and symptoms: unexpected zero probabilities in a vision model versus an image processor size mismatch." }, { "left": "issue:44589", "right": "issue:45042", "accept": false, - "reason": "Float8 storage lookup failure is unrelated to the PIL backend processor torchvision dependency issue." + "reason": "These are unrelated loading/backend issues: a missing Float8 storage type versus PIL image processors unnecessarily requiring torchvision." }, { "left": "issue:43854", "right": "issue:44977", "accept": false, - "reason": "Both are model-specific inference issues, but the titles point to separate regressions with different models and symptoms." + "reason": "Different bugs: a unit-test model load problem for GLM-4.7-Flash versus Qwen3.5 flash-attention generation behavior." }, { "left": "issue:43819", "right": "issue:43927", "accept": false, - "reason": "DAC latent reconstruction mismatch and DiaConfig token-ID loss are unrelated bugs." + "reason": "Both involve model math during generation, but DAC from_latents/STE mismatch is unrelated to DiaConfig losing custom token IDs on save/load." }, { - "left": "issue:36010", - "right": "issue:43645", + "left": "issue:39692", + "right": "issue:44912", "accept": false, - "reason": "Importing GenerationMixin and custom-model initialization in Jupyter are separate API breakages." + "reason": "Both mention model-specific regressions, but one is a SigLIP2 docs example error and the other is MXFP4 quantization falling back to bf16." }, { - "left": "issue:39692", - "right": "issue:44912", + "left": "issue:36010", + "right": "issue:43645", "accept": false, - "reason": "SigLIP2 docs errors and gpt-oss MXFP4 fallback are unrelated model-specific issues." + "reason": "Both are import/configuration related, but one is a GenerationMixin import error and the other is a Jupyter custom-model initialization regression." }, { "left": "issue:43257", "right": "issue:43329", "accept": false, - "reason": "Qwen3 MoE loading/conversion and multimodal token counting are different problems." + "reason": "Different Qwen/MoE-related failures: weight conversion with accelerate+deepspeed versus undefined video token counting helpers." }, { "left": "issue:43720", "right": "issue:44610", "accept": false, - "reason": "Packed-weight handling during accelerate loading is unrelated to the OmDet-Turbo size mismatch." + "reason": "BitNet packed-weight unpacking during accelerate loading is a different failure mode than an OmDet processor/model input-size mismatch." }, { "left": "issue:43550", "right": "issue:44912", "accept": false, - "reason": "torch.compile/SDPA failure in Bamba and MXFP4 quantization fallback are different bugs." + "reason": "One is a torch.compile + SDPA failure in Bamba, the other is a quantization-format regression for gpt-oss." }, { "left": "issue:43454", "right": "issue:43577", "accept": false, - "reason": "Weight tying failure and dtype propagation are distinct multimodal-loading issues." + "reason": "AyaVision lm_head tying bug is unrelated to BLIP2 dtype propagation during loading." }, { "left": "issue:43257", "right": "issue:43696", "accept": false, - "reason": "Qwen3 MoE loading/conversion and GPT-oss OOM are not the same underlying issue." + "reason": "Different issues: Qwen3 MOE weight conversion/loading versus GPU OOM while loading gpt-oss-20b." }, { "left": "issue:43262", "right": "issue:44811", "accept": false, - "reason": "Audio chat-template sample-rate defaulting and Whisper special-token handling are separate processor behaviors." + "reason": "Audio processor sample-rate defaulting is unrelated to Whisper batch_decode skip_special_tokens handling." }, { "left": "issue:43329", "right": "issue:44479", "accept": false, - "reason": "They are related to video processing, but 43329 is a specific undefined-name bug in the video token-count branch while 44479 is a broader video-input regression across multiple Qwen VL variants; the linkage is not strong enough to treat as the same mergeable fix." + "reason": "The first is a missing video token-counting implementation detail; the second is a v5.3.0 video input regression across several Qwen variants." }, { "left": "issue:42490", "right": "issue:43257", "accept": false, - "reason": "Trailing-slash local-model loading warnings and Qwen3 MoE conversion are unrelated." + "reason": "Import warning for local paths with trailing slashes is unrelated to Qwen3 MOE loading/conversion under accelerate+deepspeed." }, { "left": "issue:43531", "right": "issue:45042", "accept": false, - "reason": "Qwen3-MoE sliding-window behavior and PIL backend torchvision dependency are unrelated." + "reason": "Sliding_window behavior in Qwen3-MoE is unrelated to a PIL backend dependency regression." }, { "left": "issue:36010", "right": "issue:36331", "accept": false, - "reason": "Importing GenerationMixin and compute_loss keyword mismatch are separate API regressions." + "reason": "An import error for GenerationMixin is not the same underlying bug as CustomTrainer.compute_loss receiving an unexpected num_items_in_batch kwarg." }, { "left": "issue:43749", "right": "issue:44756", "accept": false, - "reason": "FSDP CPU RAM efficient loading and disabling mmap on Strix Halo are both memory-related but not the same concrete bug." + "reason": "Both concern loading efficiency, but FSDP CPU RAM efficient loading and Strix Halo mmap OOM are distinct mechanisms." }, { "left": "issue:44568", "right": "issue:45356", "accept": false, - "reason": "Tokenizer special-token insertion and Kimi-K2.5 codec/warning regression are different tokenizer problems." + "reason": "Tokenizer special-token handling for mdeberta-v3-base differs from Kimi-K2.5 codec handling and warning behavior." }, { "left": "issue:34567", "right": "issue:40990", "accept": false, - "reason": "Trainer token-count tracking and GPT-oss perplexity are unrelated training/metric issues." + "reason": "TrainerState token accounting is unrelated to GPT-oss-20b high perplexity on WikiText-2." }, { "left": "issue:43643", "right": "issue:45759", "accept": false, - "reason": "trust_remote_code config field loss and unwrapping text_config for composite multimodal configs are different config-loading bugs." + "reason": "Missing fields from trust_remote_code AutoConfig and unwrapping text_config for composite Qwen multimodal configs are different config-loading bugs." }, { "left": "issue:43329", "right": "issue:43450", "accept": false, - "reason": "Both mention video processors, but one is an undefined-name token-count bug and the other is a batched-shape issue." + "reason": "Undefined multimodal token-counting helpers are a different bug from batched video processor output shape." }, { "left": "issue:43257", "right": "issue:43931", "accept": false, - "reason": "Accelerate/deepspeed conversion failure and Qwen3-VL weight-shape mismatch are different loading errors." + "reason": "Qwen3 MOE loading with accelerate+deepspeed is not the same as Qwen3-VL-30B-A3B weight-shape mismatch on load." }, { - "left": "issue:44568", - "right": "issue:45488", + "left": "issue:43606", + "right": "issue:45859", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different tokenizers and different mechanisms." + "reason": "CPU-offload device mismatch in bark-small is unrelated to Qwen3_5MoeTextRotaryEmbedding incompatibility with CPU offload." }, { - "left": "issue:43606", - "right": "issue:45859", + "left": "issue:44568", + "right": "issue:45488", "accept": false, - "reason": "CPU offload device mismatch in Bark and Qwen3_5MoE rotary embedding incompatibility are unrelated runtime bugs." + "reason": "Tokenizer BOS/EOS insertion regression is different from LlamaTokenizer overriding tokenizer.json pre-tokenizer settings." }, { "left": "issue:44849", "right": "issue:45701", "accept": false, - "reason": "Qwen3.5 hidden-state output bug and the broad tokenization-version regression are different failure modes." + "reason": "Qwen3.5 hidden-state bug is a specific model output issue, while 45701 is a broad tokenization-version complaint." } ] }, - "evaluator_result": { - "accept": false, - "feedback": "The overall clustering is plausible, but the one accepted soft pair is too strongly asserted from the packet alone. `issue:43329` is about undefined names in `_get_num_multimodal_tokens`' video branch, while `issue:44479` is a broader video-input regression across several Qwen VL variants; the titles support related video-processing failures, but not clearly the same underlying bug or a single mergeable fix. Because the positive linkage is overstated, I can\u2019t accept the summary as fully conservative." - }, - "evaluator_used": true, - "retried": true, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null @@ -5706,8 +5662,8 @@ "issue:42915|issue:43742", "issue:44291|issue:45042", "issue:33357|issue:34567", - "issue:43452|issue:43582", "issue:44479|issue:45381", + "issue:43452|issue:43582", "issue:44987|issue:45538", "issue:41720|issue:43540", "issue:42915|issue:44821", @@ -5725,9 +5681,9 @@ "issue:38175|issue:44898", "issue:43404|issue:45987", "issue:45137|issue:45405", + "issue:43906|issue:45381", "issue:28282|issue:34689", "issue:42175|issue:44488", - "issue:43906|issue:45381", "issue:42915|issue:43749", "issue:43742|issue:44291", "issue:44737|issue:45588", @@ -5735,14 +5691,14 @@ "issue:28282|issue:29127", "issue:43526|issue:45479" ], - "prepared_review_unit_hash": "6199b3d04228fe5d9f95c7ed9e7be38dcb5f030d9bf5f36c32544d506c4a451c", + "prepared_review_unit_hash": "280637db8ddf86aeff5d973ac98243de00a77d703a9fb85131b73c1a4ee33932", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 48, @@ -5757,204 +5713,204 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is largely heterogeneous; the soft pairs mostly share broad vocabulary but do not look like the same underlying bug/change. I\u2019d use the broad Qwen 5.3.0 video-input regression issue as the best representative issue, but confidence is only moderate because the cluster has several unrelated subthemes.", - "confidence": 0.73, - "canonical_issue_reason": "Issue 44479 is the broadest and most explicit concrete regression report among the few Qwen/video-input candidates, so it is the best anchor if one issue must represent the cluster.", + "summary": "This set is heterogeneous: it spans unrelated Transformers bugs across model loading, tokenizers, video inputs, dtype/device-map issues, training state, and backend/install problems. None of the soft pairs look like the same concrete bug or a mergeable duplicate PR case.", + "confidence": 0.98, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue 44479 has the clearest versioned regression framing and spans multiple affected Qwen VL variants, making it the strongest representative issue in this set.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43697", "right": "issue:45357", "accept": false, - "reason": "Different models and failure modes: RTDetrV2 output drift vs Qwen3.5 visual-encoder save_pretrained regression." + "reason": "Different bugs and code paths: RTDetrV2 output drift vs Qwen3.5 save_pretrained key serialization." }, { "left": "issue:42915", "right": "issue:43742", "accept": false, - "reason": "Both are model-loading related, but they describe different models and distinct root causes." + "reason": "Different model-loading failures: FineGrainedFP8Config handling vs MobileLLM-125M key error." }, { "left": "issue:44291", "right": "issue:45042", "accept": false, - "reason": "init_empty_weights/_is_hf_initialized is unrelated to PIL backend torchvision requirements." + "reason": "Unrelated issues: init_empty_weights argument handling vs PIL backend requiring torchvision." }, { "left": "issue:33357", "right": "issue:34567", "accept": false, - "reason": "MacOS bus error in CLIP is unrelated to TrainerState token-count tracking." + "reason": "Different subsystems and symptoms: MacOS bus error on CLIP vs TrainerState counter not updating." }, { - "left": "issue:43452", - "right": "issue:43582", + "left": "issue:44479", + "right": "issue:45381", "accept": false, - "reason": "Tokenizer/gguf loading and AppleSilicon allocator warmup are unrelated bugs." + "reason": "Both involve video input, but one is a broad multi-model regression and the other is a specific qwen2.5-vl vision_position_ids bug." }, { - "left": "issue:44479", - "right": "issue:45381", + "left": "issue:43452", + "right": "issue:43582", "accept": false, - "reason": "Both concern Qwen video input, but one is a broad multi-model regression and the other is a specific vision_position_ids issue; not enough evidence they are the same bug." + "reason": "Completely different failures: gguf_file tokenizer/model loading vs AppleSilicon warmup TypeError." }, { "left": "issue:44987", "right": "issue:45538", "accept": false, - "reason": "Loading physical-intelligence/fast and CLIPTokenizer max_length are unrelated issues." + "reason": "Different causes and targets: one model load regression vs CLIPTokenizer model_max_length behavior." }, { "left": "issue:41720", "right": "issue:43540", "accept": false, - "reason": "cudaErrorAssert with auto device mapping is unrelated to Qwen3OmniMoe video-input ValueError." + "reason": "Different Qwen3 problems: auto device-map cuda assert vs video-input ValueError." }, { "left": "issue:42915", "right": "issue:44821", "accept": false, - "reason": "Qwen3Moe FP8 loading and AutoImageProcessor URL loading are different code paths and failures." + "reason": "Different failure modes: FP8 config issue vs loading AutoImageProcessor from URL." }, { "left": "issue:44987", "right": "issue:45907", "accept": false, - "reason": "Model loading failure and list_repo_files revision forwarding are unrelated." + "reason": "Different issues: model loading regression vs list_repo_files not forwarding revision." }, { "left": "issue:43257", "right": "issue:43756", "accept": false, - "reason": "Qwen3 MoE conversion under accelerate/deepspeed is unrelated to Smollm3 RoPE-layer dropping." + "reason": "Different model families and bugs: Qwen3 MOE weight conversion vs Smollm3 RoPE-layer handling." }, { "left": "issue:42915", "right": "issue:45538", "accept": false, - "reason": "Qwen3Moe FP8 loading and CLIPTokenizer model_max_length are different bugs." + "reason": "Different topics: quantization/config loading vs tokenizer max_length default." }, { "left": "issue:43927", "right": "issue:44912", "accept": false, - "reason": "DiaConfig token-ID persistence and MXFP4 quantization fallback are unrelated." + "reason": "Different layers of the stack: DiaConfig save/load token IDs vs MXFP4 quantization fallback." }, { "left": "issue:43540", "right": "issue:44479", "accept": false, - "reason": "Both involve Qwen video inputs, but the titles point to different models and different symptoms; too little evidence for one duplicate." + "reason": "Related domain, but not the same bug: Qwen3OmniMoe video ValueError vs broader v5.3.0 video regression across several Qwen variants." }, { "left": "issue:42175", "right": "issue:43824", "accept": false, - "reason": "TensorFlow backend packaging and missing Qwen2.5-VL import are unrelated." + "reason": "Different problems: pip backend packaging vs missing Qwen2.5-VL import symbol." }, { "left": "issue:43540", "right": "issue:44560", "accept": false, - "reason": "Both are Qwen video-input bugs, but the failures differ and there is no clear sign of a single shared defect." + "reason": "Different video-input bugs in different Qwen models and code paths." }, { "left": "issue:43317", "right": "issue:45538", "accept": false, - "reason": "device_map/offload loading failure is unrelated to CLIPTokenizer max_length." + "reason": "Device-map/offload loading bug vs tokenizer max_length default issue." }, { "left": "issue:43577", "right": "issue:45072", "accept": false, - "reason": "Blip2 dtype propagation and SwitchTransformers/TimmWrapperModel bf16 mismatch are separate model-specific dtype issues." + "reason": "Both mention dtype, but they affect different architectures and likely different fixes." }, { "left": "issue:44062", "right": "issue:44991", "accept": false, - "reason": "AddedToken special-arg duplication and tokenizer loading for EMBEDDIA/est-roberta are not clearly the same bug." + "reason": "Tokenizer-related, but one is an AddedToken constructor error and the other is a specific model tokenizer load failure." }, { "left": "issue:42175", "right": "issue:44938", "accept": false, - "reason": "TensorFlow backend packaging and Python 3.14 import failure are unrelated." + "reason": "Unrelated: TensorFlow backend packaging vs Python 3.14 import/load failure." }, { "left": "issue:38175", "right": "issue:44898", "accept": false, - "reason": "Zero probabilities in SigLIP2 and Perceiver non-default-resolution failure are different image-model issues." + "reason": "Different vision-model bugs: zero probabilities in SigLIP2 vs Perceiver non-default resolution failure." }, { "left": "issue:43404", "right": "issue:45987", "accept": false, - "reason": "Mistral3 lm_head tying and StaticCache seq-length return type are unrelated code paths." + "reason": "Different functionality: Mistral3 weight tying bug vs StaticCache seq length return type bug." }, { "left": "issue:45137", "right": "issue:45405", "accept": false, - "reason": "DeepSpeed ZeRO3 deque underflow and unreleased MIN_PEFT_VERSION bump are unrelated." + "reason": "Different concerns entirely: DeepSpeed ZeRO3 deque error vs PEFT version bump/release mismatch." + }, + { + "left": "issue:43906", + "right": "issue:45381", + "accept": false, + "reason": "No concrete overlap: isolated reproduction of another issue vs qwen2.5-vl video position-id regression." }, { "left": "issue:28282", "right": "issue:34689", "accept": false, - "reason": "Missing PyTorch ImportError and Llama 3.2 Vision loading regression are different problems." + "reason": "Both are loading errors, but one is missing PyTorch dependency and the other is a model-loading regression for Llama 3.2 Vision." }, { "left": "issue:42175", "right": "issue:44488", "accept": false, - "reason": "TensorFlow backend packaging and model-loading failure for cjvt/sleng-bert are unrelated." - }, - { - "left": "issue:43906", - "right": "issue:45381", - "accept": false, - "reason": "A reproduction of another issue and a Qwen2.5-VL video-position regression are not the same concrete bug." + "reason": "Different issues: packaging/backends vs a specific model load failure for cjvt/sleng-bert." }, { "left": "issue:42915", "right": "issue:43749", "accept": false, - "reason": "Qwen3Moe FP8 loading and FSDP CPU RAM efficient loading are different loading paths with different failures." + "reason": "Different problem areas: FineGrainedFP8Config vs FSDP CPU RAM efficient loading." }, { "left": "issue:43742", "right": "issue:44291", "accept": false, - "reason": "MobileLLM-125M key error and init_empty_weights _is_hf_initialized TypeError are unrelated." + "reason": "Different load-time exceptions affecting different APIs and model families." }, { "left": "issue:44737", "right": "issue:45588", "accept": false, - "reason": "XLNet CPU-device inefficiency and flash_attention sink-less AttributeError are unrelated." + "reason": "Different bugs: XLNet CPU tensor allocation vs flash_attention s_aux=None crash." }, { "left": "issue:43540", "right": "issue:45823", "accept": false, - "reason": "Qwen3OmniMoe video-input ValueError and Gemma4 PLE device mismatch are different bugs." + "reason": "Different model families and symptoms: Qwen3OmniMoe video parsing vs Gemma4 PLE device mismatch." }, { "left": "issue:28282", "right": "issue:29127", "accept": false, - "reason": "PyTorch-missing ImportError and LayoutLMv3 box-information error are unrelated." + "reason": "Different issues: missing PyTorch dependency vs LayoutLMv3 error-message clarity for insufficient boxes." }, { "left": "issue:43526", "right": "issue:45479", "accept": false, - "reason": "BeitImageProcessorFast label reduction and sequence-classification zero-loss degeneration are unrelated." + "reason": "Different bug classes: BeitImageProcessorFast label reduction vs sequence-classification zero-loss behavior." } ] }, @@ -6039,8 +5995,8 @@ "issue:43452|issue:44488", "issue:45042|issue:45081", "issue:43742|issue:44589", - "issue:43531|issue:44821", "issue:45405|issue:45600", + "issue:43531|issue:44821", "issue:43697|issue:45216", "issue:44514|issue:45381", "issue:43688|issue:43749", @@ -6054,199 +6010,199 @@ "issue:43452|issue:43577", "issue:44466|issue:45005" ], - "prepared_review_unit_hash": "ad58a6ee9988efa233db3550a8e6f1e02f66c23fa7ed9967b5725b70f7279db0", + "prepared_review_unit_hash": "5ceb14cfb80aecb32f852d2f57ee5911b2f56cc7fe17526b025730de1e55cd79", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 26, - "serialized_chars": 35283, - "estimated_input_tokens": 8821, - "estimated_eval_tokens": 17898 + "serialized_chars": 35288, + "estimated_input_tokens": 8822, + "estimated_eval_tokens": 17900 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Conservative cluster overall: most soft edges are false positives, and the `apply_chat_template` pair should not be treated as an exact duplicate because the triggers differ. Real duplicate-like matches remain limited to the clearly same-model/same-regression pairs, while the Qwen/VL chat-template issues are related but distinct crash modes.", - "confidence": 0.84, - "canonical_issue_reason": "Issue 45290 is the cleaner anchor for the `apply_chat_template` crash family because it states a concrete failure mode, but it is not an exact duplicate of 44514.", + "summary": "Mostly heterogeneous issue set with one clear duplicate mini-cluster around tied-weights serialization in v5. I found one strong duplicate pair; the other soft matches are too broad or model-specific to merge confidently.", + "confidence": 0.41, + "canonical_issue_reason": "Issue 45005 is the best umbrella for the tied-weights regression: it states the broader v5 problem, while 44466 looks like a specific device-dependent manifestation of the same save/serialization bug.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45290 is the best issue anchor for the cluster\u2019s main `apply_chat_template` branch, but only as a related report; 44514 should remain separate because its batched `padding=False` trigger is different.", + "best_issue_reason": "45005 is the most representative issue to anchor triage because it captures the broader underlying regression rather than one symptom.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43582", "right": "issue:43782", "accept": false, - "reason": "Different failure modes and code paths: Apple Silicon allocator TypeError vs Qwen3VL weight_only loading error." + "reason": "Different subsystems and failures: Apple Silicon allocator warmup TypeError vs Qwen3VL weight_only loading error." }, { "left": "issue:43493", "right": "issue:43994", - "accept": true, - "reason": "Both reports point to the same SigLIP2 correctness regression for the same model family." + "accept": false, + "reason": "Both involve SigLIP2, but one is a HF-vs-JAX discrepancy and the other is a bad AutoModel/pipeline result; not clearly the same concrete bug." }, { "left": "issue:43756", "right": "issue:45325", "accept": false, - "reason": "Related vision/position-id themes, but different models and distinct RoPE/position encoding bugs." + "reason": "Different models and different RoPE/position-id bugs; no shared code-path evidence." }, { "left": "issue:34567", "right": "issue:36331", "accept": false, - "reason": "TrainerState token counting and CustomTrainer compute_loss signature incompatibility are unrelated issues." + "reason": "Trainer-state token accounting bug vs custom compute_loss signature mismatch; related area, not the same defect." }, { "left": "issue:44062", "right": "issue:45081", "accept": false, - "reason": "Both are tokenizer-related, but the concrete bugs differ: AddedToken kwarg duplication vs mistral regex backend_tokenizer crash." + "reason": "Different tokenizer failures: duplicate AddedToken kwargs vs mistral regex patch crashing on backend_tokenizer access." }, { "left": "issue:44291", "right": "issue:44821", "accept": false, - "reason": "Different loading paths and symptoms: init_empty_weights/_is_hf_initialized vs AutoImageProcessor URL loading." + "reason": "init_empty_weights/_is_hf_initialized loading bug is unrelated to AutoImageProcessor URL loading." }, { "left": "issue:45357", "right": "issue:45920", "accept": false, - "reason": "save_pretrained visual-encoder key regression and tokenizer ID regression are different code-path bugs." + "reason": "Different regressions: Qwen3.5 save_pretrained visual-key serialization vs generic AutoTokenizer token-id errors." }, { "left": "issue:43643", "right": "issue:45375", "accept": false, - "reason": "Both concern config fields, but one is a generic trust_remote_code AutoConfig omission and the other is a specific strict-field drop." + "reason": "Both are config-related, but one is remote-code config field loss and the other is a specific strict-field drop in Qwen3_5MoeVisionConfig." }, { "left": "issue:44792", "right": "issue:44964", "accept": false, - "reason": "Different multimodal model failures with no shared concrete bug." + "reason": "Both are multimodal/model-load failures, but they affect different models and likely different code paths." }, { "left": "issue:43452", "right": "issue:44488", "accept": false, - "reason": "Different loading regressions: gguf_file/autoload breakage vs a model-specific load failure." + "reason": "Generic gguf_file/from_pretrained breakage vs one model failing to load; not enough evidence they are the same bug." }, { "left": "issue:45042", "right": "issue:45081", "accept": false, - "reason": "Unrelated bugs: image processor torchvision dependency vs tokenizer regex crash." + "reason": "PIL/torchvision dependency regression is unrelated to Mistral regex patching." }, { "left": "issue:43742", "right": "issue:44589", "accept": false, - "reason": "Different model-loading errors with no shared concrete code path." + "reason": "KeyError on MobileLLM loading and missing Float8 storage are different loader/runtime issues." }, { - "left": "issue:43531", - "right": "issue:44821", + "left": "issue:45405", + "right": "issue:45600", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior is unrelated to AutoImageProcessor URL loading." + "reason": "PEFT version pin mismatch vs stale auto_mappings reference; both CI/package issues, but not the same change." }, { - "left": "issue:45405", - "right": "issue:45600", + "left": "issue:43531", + "right": "issue:44821", "accept": false, - "reason": "PEFT version bump and removed auto_mappings references are separate maintenance issues." + "reason": "Qwen3-MoE sliding_window behavior and AutoImageProcessor URL loading are unrelated." }, { "left": "issue:43697", "right": "issue:45216", "accept": false, - "reason": "RTDetrV2 output regression and Qwen3.5 save_pretrained regression affect different code paths." + "reason": "RTDetrV2 output regression and Qwen3.5 checkpoint save regression affect different paths and symptoms." }, { "left": "issue:44514", "right": "issue:45381", "accept": false, - "reason": "Both are Qwen/VL-related, but one is a chat-template/tool-call crash and the other is a vision_position_ids bug." + "reason": "Both are Qwen2.5-VL, but one is chat-template batching and the other is video position-id calculation; not the same bug." }, { "left": "issue:43688", "right": "issue:43749", "accept": false, - "reason": "Auxiliary-loss normalization and FSDP CPU RAM efficient loading are unrelated bugs." + "reason": "Aux-loss normalization bug vs FSDP CPU RAM-efficient loading breakage are unrelated." }, { "left": "issue:43720", "right": "issue:44898", "accept": false, - "reason": "BitNet packed-weight loading and Perceiver interpolation are different problems." + "reason": "BitNet packed-weight loading bug and Perceiver image-classification failure are different model-specific problems." }, { "left": "issue:44484", "right": "issue:44704", "accept": false, - "reason": "A save_pretrained sharding-size question is unrelated to AutoProcessor cached_file kwargs handling." + "reason": "save_pretrained shard-size behavior vs AutoProcessor cached_file kwargs forwarding are unrelated." }, { "left": "issue:24643", "right": "issue:29127", "accept": false, - "reason": "DeepSpeed weight-shape training error and LayoutLMv3 error-message clarity are different issues." + "reason": "DeepSpeed training weight-dimension runtime error vs LayoutLMv3 box validation message clarity are different issues." }, { "left": "issue:41762", "right": "issue:44936", "accept": false, - "reason": "Gemma3 ZeRO-3 loading failure and trainer.evaluate after train are unrelated." + "reason": "Gemma3 ZeRO-3 load crash and trainer evaluate-after-train failure do not look like the same underlying bug." }, { "left": "issue:44625", "right": "issue:44877", "accept": false, - "reason": "Different config bugs: num_labels propagation vs strict config loading for granite_speech." + "reason": "Qwen3.5 num_labels propagation and granite_speech strict-config loading are both config bugs, but not the same concrete issue." }, { "left": "issue:44514", "right": "issue:45290", "accept": false, - "reason": "Related `apply_chat_template` crashes, but the triggers differ (`padding=False` batched input vs assistant tool-call messages with no content), so there is no exact-duplicate evidence." + "reason": "Both touch apply_chat_template, but one is batched padding handling and the other is tool-call assistant message handling." }, { "left": "issue:43452", "right": "issue:44843", "accept": false, - "reason": "gguf_file loading breakage and HF_HUB_OFFLINE mistral-regex behavior are different regressions." + "reason": "gguf_file loading breakage is unrelated to the mistral-regex offline model_info call bug." }, { "left": "issue:43452", "right": "issue:43577", "accept": false, - "reason": "Model loading via gguf_file is unrelated to Blip2 dtype propagation." + "reason": "GGUF/from_pretrained loading issue and BLIP2 dtype propagation issue are different code paths." }, { "left": "issue:44466", "right": "issue:45005", "accept": true, - "reason": "Both describe the same tied-weights serialization regression in v5, centered on lm_head/translation-model weight handling." + "reason": "Both describe the same tied-weights serialization regression in v5/save_pretrained; 45005 is the broader umbrella and 44466 is a concrete manifestation." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The cluster is mostly conservative, but one soft-edge verdict overstates the evidence: treating issue:44514 and issue:45290 as an exact duplicate is not grounded in the packet. Their titles describe related `apply_chat_template` failures, but with different triggers (`padding=False` batched input vs assistant messages with tool calls and no content), and the packet provides no direct evidence they are the same bug. Because the canonical anchor and summary lean on that stronger claim, I can\u2019t accept this review." + "accept": true, + "feedback": "The summary is well grounded in the packet and stays conservative overall. The only accepted pair, 44466/45005, is plausibly the same tied-weights save/serialization regression in v5, with 45005 as the broader umbrella. The other soft matches are mostly model- or code-path-specific and were reasonably rejected." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -6318,9 +6274,9 @@ "issue:43819|issue:44811", "issue:44936|issue:45823", "issue:38175|issue:43295", - "issue:30064|issue:33666", "issue:44361|issue:44442", "issue:43653|issue:44361", + "issue:30064|issue:33666", "issue:44464|issue:44743", "issue:41720|issue:43906", "issue:43653|issue:44568", @@ -6337,250 +6293,250 @@ "issue:43688|issue:45701", "issue:43653|issue:43927", "issue:43531|issue:43742", - "issue:44464|issue:44977", "issue:43550|issue:43927", + "issue:44464|issue:44977", "issue:44479|issue:44625", "issue:43550|issue:44361", + "issue:43122|issue:43906", "issue:43577|issue:44554", "issue:43577|issue:45588", - "issue:43122|issue:43906", - "issue:32090|issue:34567", - "issue:41720|issue:44464", "issue:43504|issue:43550", - "issue:43504|issue:43720" + "issue:43504|issue:43720", + "issue:32090|issue:34567", + "issue:41720|issue:44464" ], - "prepared_review_unit_hash": "3b6677e02867faea7ea7c0a4134e811d72ae4de2f036f52eb713957a9c6564ab", + "prepared_review_unit_hash": "c8b48e762319516cc90deb7c21b36f281b94ad69389672bd6ca69e153f5a773b", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 34, - "serialized_chars": 37929, - "estimated_input_tokens": 9483, - "estimated_eval_tokens": 19222 + "serialized_chars": 37924, + "estimated_input_tokens": 9481, + "estimated_eval_tokens": 19218 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is broad and heterogeneous, with many unrelated regression reports. None of the soft-edge pairs look like the same underlying bug/change, though several share surface-level themes like tokenization or model loading. The broadest representative issue is the tokenizer-regression umbrella report.", - "confidence": 0.62, - "canonical_issue_reason": "issue:45701 is the broadest, most umbrella-like report in the set and has inbound references, so it is the best anchor for this loosely related cluster.", + "summary": "This cluster is mostly a noisy mix of unrelated bugs. The strongest recurring theme is tokenizer/version-regression reports, but the soft pairs still describe different concrete failures, so none should be merged.", + "confidence": 0.9, + "canonical_issue_reason": "issue:45701 is the best umbrella representative: it explicitly frames cross-version tokenization changes and has inbound references, which fits the main recurring theme better than the more specific symptom reports.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45701 is the most representative issue overall because it is broad, version-related, and closest to the recurring tokenizer-regression theme seen in the cluster.", + "best_issue_reason": "issue:45701 is the most suitable single issue to anchor this cluster because it is the broadest tokenizer-regression report and most likely to subsume the nearby version-dependent tokenization complaints.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43295", "right": "issue:45356", "accept": false, - "reason": "Both are tokenizer regressions, but one is about custom model code/processor.tokenizer handling and the other is a Kimi tokenizer codec/regression issue; different bugs." + "reason": "Both are tokenizer regressions, but one is about processor.tokenizer/image handling and the other is a Kimi-K2.5 codec/fix_mistral_regex regression; different failure modes." }, { "left": "issue:43819", "right": "issue:44811", "accept": false, - "reason": "DAC.from_latents vs Whisper batch_decode special-token handling are unrelated code paths and symptoms." + "reason": "Different code paths: DAC.from_latents vs Whisper batch_decode skip_special_tokens." }, { "left": "issue:44936", "right": "issue:45823", "accept": false, - "reason": "Trainer lifecycle failure after train() is unrelated to Gemma4 PLE device mismatch during forward." + "reason": "Trainer evaluate-after-train behavior is unrelated to Gemma4 PLE device_map forward mismatch." }, { "left": "issue:38175", "right": "issue:43295", "accept": false, - "reason": "Siglip2 zero probabilities and processor.tokenizer/image-passing regression are different model behaviors and failure modes." - }, - { - "left": "issue:30064", - "right": "issue:33666", - "accept": false, - "reason": "Image processor segmentation-map parsing and Qwen2-VL multi-GPU training are unrelated issues." + "reason": "siglip2 zero probabilities is a model/output issue, not the tokenizer regression described by 43295." }, { "left": "issue:44361", "right": "issue:44442", "accept": false, - "reason": "MLukeTokenizer task AttributeError and AutoTokenizer failing to load FastSpeech2ConformerTokenizer are different tokenizer failures." + "reason": "Both involve tokenizers, but they are different tokenizer classes and distinct loading failures." }, { "left": "issue:43653", "right": "issue:44361", "accept": false, - "reason": "BigBirdTokenizer special-token registration and MLukeTokenizer task AttributeError are unrelated bugs." + "reason": "BigBird special-token registration/empty decode is unrelated to MLukeTokenizer task AttributeError." + }, + { + "left": "issue:30064", + "right": "issue:33666", + "accept": false, + "reason": "Image processor void segmentation maps and Qwen2-VL multi-GPU training are unrelated subsystems." }, { "left": "issue:44464", "right": "issue:44743", "accept": false, - "reason": "Chunked generation inconsistency with compiled forward and Qwen3.5 recurrent-state reset are different generation-path problems." + "reason": "Both concern generation/state handling, but one is compiled-forward inconsistency and the other is recurrent-state reset with cache; not the same bug." }, { "left": "issue:41720", "right": "issue:43906", "accept": false, - "reason": "Qwen3 auto device-map cudaErrorAssert and an isolated reproduction of issue 38071 do not match the same concrete failure." + "reason": "Qwen3 auto-device-map cuda assert is not the same as the isolated reproduction of a different upstream issue." }, { "left": "issue:43653", "right": "issue:44568", "accept": false, - "reason": "Both involve special tokens, but the models and defects differ: BigBird mask token registration vs mDeBERTa BOS/EOS insertion." + "reason": "Both are tokenizer special-token issues, but on different tokenizers with different symptoms and code paths." }, { "left": "issue:28282", "right": "issue:32090", "accept": false, - "reason": "PyTorch-missing ImportError and Trainer _gpu_broadcast_one NoneType TypeError are unrelated." + "reason": "PyTorch missing ImportError is unrelated to Trainer _gpu_broadcast_one rejecting NoneType." }, { "left": "issue:28282", "right": "issue:33357", "accept": false, - "reason": "PyTorch-missing ImportError and Mac bus error on CLIP are different environments and root causes." + "reason": "Missing PyTorch import is unrelated to the MacOS bus error on a community CLIP model." }, { "left": "issue:43688", "right": "issue:45704", "accept": false, - "reason": "Auxiliary-loss normalization in MoE models and apex.FusedRMSNorm memory leak in T5 are unrelated." + "reason": "Aux-loss normalization in MoE models is unrelated to the T5 apex.FusedRMSNorm memory leak." }, { "left": "issue:43688", "right": "issue:45440", "accept": false, - "reason": "Both touch MoE models, but auxiliary-loss normalization and native DeepseekV3MoE divergence are different bugs." + "reason": "Both touch MoE models, but one is auxiliary-loss normalization and the other is parity divergence from a remote implementation." }, { "left": "issue:45491", "right": "issue:46002", "accept": false, - "reason": "Gemma3 mixed-length sliding-window NaNs and classification-head NaNs after checkout/init are different contexts and mechanisms." + "reason": "Both mention NaNs, but the triggers are different: sliding-window all-padding embeddings vs checkout/init classification heads." }, { "left": "issue:43577", "right": "issue:43697", "accept": false, - "reason": "BLIP2 dtype-casting issue and RTDetrV2 output-difference regression are unrelated." + "reason": "BLIP2 dtype preservation and RTDetrV2 output mismatch are separate model-specific bugs." }, { "left": "issue:28282", "right": "issue:35141", "accept": false, - "reason": "Missing PyTorch ImportError is unrelated to output-embedding reinitialization after resizing token embeddings." + "reason": "Missing PyTorch is unrelated to output embeddings being reinitialized during resize." }, { "left": "issue:43643", "right": "issue:44561", "accept": false, - "reason": "Missing fields from AutoConfig with trust_remote_code and removal of is_torch_fx_available are separate v5 breakages." + "reason": "AutoConfig missing fields under trust_remote_code and removal of is_torch_fx_available are different regressions." }, { "left": "issue:44843", "right": "issue:44991", "accept": false, - "reason": "HF_HUB_OFFLINE breakage in _patch_mistral_regex and tokenizer loading failure for est-roberta are different tokenizer-path bugs." + "reason": "Offline model_info() call in tokenizer patching is a different bug from failing to load EMBEDDIA/est-roberta's tokenizer." }, { "left": "issue:43646", "right": "issue:43749", "accept": false, - "reason": "Custom model initialization breakage in v5 and FSDP_CPU_RAM_EFFICIENT_LOADING failure are not the same issue." + "reason": "Custom model initialization breakage is unrelated to FSDP CPU RAM efficient loading." }, { "left": "issue:43688", "right": "issue:45701", "accept": false, - "reason": "MoE auxiliary-loss normalization is not the same as the broad tokenization-change regression." + "reason": "MoE auxiliary-loss normalization is unrelated to the tokenizer version-change reports." }, { "left": "issue:43653", "right": "issue:43927", "accept": false, - "reason": "BigBird special-token registration and DiaConfig losing custom token IDs are different serialization/tokenizer bugs." + "reason": "BigBird mask-token special-token registration and DiaConfig custom token IDs loss are different token-management bugs." }, { "left": "issue:43531", "right": "issue:43742", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and MobileLLM-125M load key error are unrelated." + "reason": "Qwen3-MoE sliding_window behavior is unrelated to the MobileLLM-125M key error on load." }, { - "left": "issue:44464", - "right": "issue:44977", + "left": "issue:43550", + "right": "issue:43927", "accept": false, - "reason": "Compiled-forward chunked generation inconsistency and Qwen3.5 flash-attention generation failure are different inference bugs." + "reason": "torch.compile SDPA failure is unrelated to config token IDs being lost after save/load." }, { - "left": "issue:43550", - "right": "issue:43927", + "left": "issue:44464", + "right": "issue:44977", "accept": false, - "reason": "Bamba-9B-v2 torch.compile SDPA failure and DiaConfig token-ID loss are unrelated." + "reason": "Compiled-forward chunked generation inconsistency and flash-attention generation failure are different concrete code paths." }, { "left": "issue:44479", "right": "issue:44625", "accept": false, - "reason": "Video-input regression for Qwen VL models and Qwen3.5 num_labels propagation are different defects." + "reason": "Video-input regression across Qwen VL variants is unrelated to num_labels propagation in Qwen3.5." }, { "left": "issue:43550", "right": "issue:44361", "accept": false, - "reason": "torch.compile/SDPA failure in Bamba is unrelated to MLukeTokenizer AttributeError." + "reason": "Bamba torch.compile/SDPA failure is unrelated to MLukeTokenizer AttributeError." + }, + { + "left": "issue:43122", + "right": "issue:43906", + "accept": false, + "reason": "Different tokenization across versions and an isolated reproduction of another issue are not the same bug." }, { "left": "issue:43577", "right": "issue:44554", "accept": false, - "reason": "BLIP2 dtype persistence and MPS attention correctness issue are unrelated." + "reason": "BLIP2 dtype loading and MPS attention correctness are unrelated." }, { "left": "issue:43577", "right": "issue:45588", "accept": false, - "reason": "BLIP2 dtype persistence and flash_attention.py s_aux=None crash are different code paths." + "reason": "dtype preservation on load is unrelated to flash_attention.py crashing on s_aux=None." }, { - "left": "issue:43122", - "right": "issue:43906", + "left": "issue:43504", + "right": "issue:43550", + "accept": false, + "reason": "Beit legacy-field loading failure is unrelated to Bamba torch.compile SDPA." + }, + { + "left": "issue:43504", + "right": "issue:43720", "accept": false, - "reason": "Tokenizer-output changes across versions and an isolated reproduction of issue 38071 do not demonstrate the same bug." + "reason": "Legacy preset loading for Beit is unrelated to BitNet packed-weight unpacking during accelerate loading." }, { "left": "issue:32090", "right": "issue:34567", "accept": false, - "reason": "Trainer broadcast TypeError and num_input_tokens_seen not updating are different Trainer-state issues." + "reason": "Trainer _gpu_broadcast_one NoneType and TrainerState num_input_tokens_seen not updating are separate trainer bugs." }, { "left": "issue:41720", "right": "issue:44464", "accept": false, - "reason": "Qwen3 auto device-mapping cuda assert and compiled-forward chunked generation inconsistency are unrelated." - }, - { - "left": "issue:43504", - "right": "issue:43550", - "accept": false, - "reason": "Beit legacy-field load failure and Bamba torch.compile SDPA failure are different model-loading/runtime bugs." - }, - { - "left": "issue:43504", - "right": "issue:43720", - "accept": false, - "reason": "Beit legacy-field loading issue and BitNet packed-weight unpacking failure are unrelated." + "reason": "Qwen3 auto-device-map cuda assert is unrelated to chunked generation with compiled forward." } ] }, @@ -6658,11 +6614,11 @@ "issue:32090|issue:33666", "issue:43452|issue:44451", "issue:43742|issue:44821", - "issue:43278|issue:43688", "issue:33666|issue:36010", - "issue:43262|issue:43295", + "issue:43278|issue:43688", "issue:43606|issue:43927", "issue:43493|issue:43825", + "issue:43262|issue:43295", "issue:43720|issue:43927", "issue:44849|issue:45357", "issue:43756|issue:44913", @@ -6674,218 +6630,218 @@ "issue:33666|issue:36331", "issue:39692|issue:43493", "issue:43540|issue:44977", + "issue:43493|issue:43844", "issue:44964|issue:45405", "issue:45803|issue:45987", - "issue:43493|issue:43844", "issue:43278|issue:43844", - "issue:43257|issue:43299", "issue:43756|issue:45083", - "issue:43334|issue:43572", + "issue:43257|issue:43299", "issue:45200|issue:45229", + "issue:43334|issue:43572", "issue:44488|issue:44843" ], - "prepared_review_unit_hash": "90d95324be902be1df5af617c0098a76343e846991e8eae14a6b86fff0de85c3", + "prepared_review_unit_hash": "c02a4a28804214324e434c2b3f5e8f667b1fc435a933db691dcd71e481ac3687", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36496, - "estimated_input_tokens": 9124, - "estimated_eval_tokens": 18504 + "serialized_chars": 36497, + "estimated_input_tokens": 9125, + "estimated_eval_tokens": 18506 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous transformer bugs/regressions across unrelated models and code paths. None of the suggested soft pairs look like true same-bug duplicates, so I would not merge any of them. The strongest single issue here is the Granite tokenizer regression, but there is no safe canonical duplicate anchor for the whole cluster.", - "confidence": 0.93, - "canonical_issue_reason": "No safe canonical issue: the cluster spans unrelated failures, and the soft pairs do not indicate a single underlying bug.", + "summary": "The cluster is mostly a title-similarity bucket, not a duplicate set; the soft pairs are related only at a very broad subsystem level and should be rejected. The strongest standalone issue is the Granite tokenizer regression, but there is no single canonical duplicate across the whole cluster.", + "confidence": 0.87, + "canonical_issue_reason": "issue:45812 is the most central, concrete regression in the set: it has broad impact, clear tokenizer misbehavior, and the strongest discussion/reference signal.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45812 is the strongest representative issue because it is an open, high-signal tokenizer regression with multiple inbound references; however, it still does not unify the cluster.", + "best_issue_reason": "issue:45812 is the best representative issue to anchor triage because it is a clearly described, high-impact, user-facing regression with multiple inbound references.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43504", "right": "issue:44610", "accept": false, - "reason": "Both are model-loading/processor failures, but they concern different models and different bugs: legacy field loading vs image size mismatch." + "reason": "Different models and failures: legacy-field load error vs processor/model input-size mismatch." }, { "left": "issue:32090", "right": "issue:33666", "accept": false, - "reason": "Trainer NoneType broadcast error and Qwen2-VL multi-GPU training are unrelated failures." + "reason": "Trainer GPU broadcast TypeError is unrelated to Qwen2-VL multi-GPU training behavior." }, { "left": "issue:43452", "right": "issue:44451", "accept": false, - "reason": "One is a gguf_file loading regression; the other is a specific model-loading failure. Same broad area, not the same bug." + "reason": "Both are loading failures, but one is gguf handling and the other is a specific model load regression." }, { "left": "issue:43742", "right": "issue:44821", "accept": false, - "reason": "KeyError loading MobileLLM and AutoImageProcessor URL loading are different code paths and different failure modes." - }, - { - "left": "issue:43278", - "right": "issue:43688", - "accept": false, - "reason": "Embedding dtype drift in eval and auxiliary loss normalization are unrelated training issues." + "reason": "KeyError on MobileLLM loading is not the same as AutoImageProcessor URL loading." }, { "left": "issue:33666", "right": "issue:36010", "accept": false, - "reason": "Multi-GPU training and GenerationMixin import failure are completely different problems." + "reason": "Multi-GPU training for Qwen2-VL is unrelated to a GenerationMixin import regression." }, { - "left": "issue:43262", - "right": "issue:43295", + "left": "issue:43278", + "right": "issue:43688", "accept": false, - "reason": "Both involve processors, but one is an audio sampling-rate default and the other is a regression in processor.tokenizer/image handling." + "reason": "Embedding dtype drift during eval is unrelated to auxiliary-loss normalization in MoE models." }, { "left": "issue:43606", "right": "issue:43927", "accept": false, - "reason": "CPU offload device mismatch for bark-small and DiaConfig token ID loss are unrelated bugs." + "reason": "CPU offload device mismatch and DiaConfig token-id persistence are different code paths and symptoms." }, { "left": "issue:43493", "right": "issue:43825", "accept": false, - "reason": "SigLIP2 implementation discrepancy and pipeline translation error messaging are different issues." + "reason": "SigLIP2 implementation mismatch is unrelated to a pipeline error-message regression." + }, + { + "left": "issue:43262", + "right": "issue:43295", + "accept": false, + "reason": "Audio chat-template sample-rate defaulting is unrelated to processor.tokenizer/image-passing regression." }, { "left": "issue:43720", "right": "issue:43927", "accept": false, - "reason": "Packed-weight unpacking during accelerate loading and DiaConfig token ID persistence are different model-loading bugs." + "reason": "BitNet packed-weight unpacking and DiaConfig token IDs are separate loading bugs." }, { "left": "issue:44849", "right": "issue:45357", "accept": false, - "reason": "Qwen3.5 hidden-state behavior and visual encoder key saving are separate regressions." + "reason": "Qwen3.5 hidden-state behavior and visual-encoder save_pretrained key corruption are different regressions." }, { "left": "issue:43756", "right": "issue:44913", "accept": false, - "reason": "Smollm3 RoPE-layer loss and GPTNeoX rotary_pct reload are both config-related but not the same defect." + "reason": "Smollm3 RoPE-layer omission is not the same as GPTNeoX rotary_pct reload loss." }, { "left": "issue:44625", "right": "issue:46002", "accept": false, - "reason": "num_labels propagation and NaNs in classification heads are different symptoms with different likely causes." + "reason": "num_labels propagation and NaNs in classification heads are distinct model init/config issues." }, { "left": "issue:44568", "right": "issue:45812", "accept": false, - "reason": "Both are tokenizer regressions, but one is BOS/EOS handling for mdeberta and the other is Granite token IDs; not the same bug." + "reason": "mDeBERTa add_special_tokens tokenization bug is unrelated to Granite token-ID corruption." }, { "left": "issue:43525", "right": "issue:43572", "accept": false, - "reason": "Missing pad_token_id and missing pad_token_idx are similar in theme, but they affect different configs and code paths." + "reason": "Both involve missing pad-token fields, but they affect different configs and failure modes." }, { "left": "issue:43720", "right": "issue:44361", "accept": false, - "reason": "BitNet packed-weight loading and MLukeTokenizer AttributeError are unrelated." + "reason": "BitNet accelerate loading and MLukeTokenizer AttributeError are not the same bug." }, { "left": "issue:28218", "right": "issue:33357", "accept": false, - "reason": "Tokenizer spacing after added token and a MacOS CLIP bus error are clearly unrelated." + "reason": "Tokenizer extra-space bug is unrelated to a MacOS bus error on a CLIP model." }, { "left": "issue:33666", "right": "issue:36331", "accept": false, - "reason": "Multi-GPU training and unexpected num_items_in_batch in compute_loss are different trainer issues." + "reason": "Qwen2-VL multi-GPU training is unrelated to CustomTrainer.compute_loss signature breakage." }, { "left": "issue:39692", "right": "issue:43493", "accept": false, - "reason": "Both mention SigLIP2, but one is a doc example problem and the other is an implementation discrepancy; not the same fix." + "reason": "SigLIP2 doc example errors are documentation/runtime issues, not the same as core implementation discrepancy." }, { "left": "issue:43540", "right": "issue:44977", "accept": false, - "reason": "Qwen3OmniMoe video-input processing and Qwen3.5 flash-attention generation are unrelated." + "reason": "Qwen3OmniMoe video processing error and Qwen3.5 flash-attention generation bug are different problems." + }, + { + "left": "issue:43493", + "right": "issue:43844", + "accept": false, + "reason": "SigLIP2 implementation mismatch is unrelated to DeepSpeed ZeRO-3 gradient growth." }, { "left": "issue:44964", "right": "issue:45405", "accept": false, - "reason": "Model loading failure and MIN_PEFT_VERSION packaging issue are unrelated." + "reason": "Phi-4 multimodal load failure is unrelated to a dependency version bump issue." }, { "left": "issue:45803", "right": "issue:45987", "accept": false, - "reason": "Bare except in FuyuBatchFeature and StaticCache return-type mismatch are unrelated bugs." + "reason": "Bare except in Fuyu tensor conversion and StaticCache seq-length return type are unrelated." }, { - "left": "issue:43493", + "left": "issue:43278", "right": "issue:43844", "accept": false, - "reason": "SigLIP2 discrepancy and ZeRO-3 gradient growth are unrelated training/model issues." + "reason": "Embedding dtype drift is unrelated to ZeRO-3 gradient instability." }, { - "left": "issue:43278", - "right": "issue:43844", + "left": "issue:43756", + "right": "issue:45083", "accept": false, - "reason": "Embedding dtype regression and ZeRO-3 gradient abnormality do not share the same underlying cause." + "reason": "Smollm3 RoPE-layer dropping and Qwen3 Omni helper length behavior are separate implementation bugs." }, { "left": "issue:43257", "right": "issue:43299", "accept": false, - "reason": "Both are Qwen3-family loading regressions, but the affected models and failure modes differ enough that they should not be merged." + "reason": "Qwen3 MOE accelerate/deepspeed conversion and Qwen3 VL loading regression are different model families and failures." }, { - "left": "issue:43756", - "right": "issue:45083", + "left": "issue:45200", + "right": "issue:45229", "accept": false, - "reason": "Smollm3 RoPE-layer issue and qwen3_omni_moe helper-length behavior are unrelated." + "reason": "Missing mm_token_type_ids default and Gemma 4 multi-GPU OOM are not the same bug." }, { "left": "issue:43334", "right": "issue:43572", "accept": false, - "reason": "Missing pad_token_id in Qwen3-VL and missing pad_token_idx in StableLmConfig are different model-specific regressions." - }, - { - "left": "issue:45200", - "right": "issue:45229", - "accept": false, - "reason": "Zero-default token type IDs and CUDA OOM in multi-GPU inference are unrelated." + "reason": "Missing pad_token_id in Qwen3-VL config and missing pad_token_idx in StableLmConfig are similar in shape but different defects." }, { "left": "issue:44488", "right": "issue:44843", "accept": false, - "reason": "Model load failure for sleng-bert and unconditional model_info() patching for offline mode are different tokenizer-loading issues." + "reason": "Two separate loading issues: model-specific tokenizer load failure vs offline-mode hub lookup bug." } ] }, @@ -6967,18 +6923,18 @@ "issue:43504|issue:43606", "issue:33357|issue:33666", "issue:43572|issue:44291", - "issue:44589|issue:45520", "issue:43927|issue:44361", + "issue:44589|issue:45520", "issue:43504|issue:43653", "issue:44464|issue:45584", - "issue:44423|issue:44734", "issue:43606|issue:44112", + "issue:44423|issue:44734", "issue:43575|issue:45405", "issue:43994|issue:44493", "issue:43278|issue:43701", "issue:44792|issue:45405", - "issue:29942|issue:30333", "issue:44521|issue:45290", + "issue:29942|issue:30333", "issue:42490|issue:43696", "issue:43844|issue:44964", "issue:28218|issue:34567", @@ -6986,190 +6942,190 @@ "issue:43582|issue:44977", "issue:43540|issue:45200" ], - "prepared_review_unit_hash": "a5517783574e88c1f190601e157fa2e84270973dc2c679fee8df8220e8a1acf8", + "prepared_review_unit_hash": "bc374d2c648e817aa69515b62092c75ae34463a00206c2c4570e3e0f06ec70dc", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 26, - "serialized_chars": 35511, - "estimated_input_tokens": 8878, - "estimated_eval_tokens": 18012 + "serialized_chars": 35513, + "estimated_input_tokens": 8879, + "estimated_eval_tokens": 18014 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a heterogeneous set of unrelated issue reports; none of the soft candidate pairs look like the same underlying bug or change, so all proposed duplicate edges should be rejected.", - "confidence": 0.98, - "canonical_issue_reason": "No canonical issue is appropriate here: the items cover many different models, APIs, and failure modes, so there is no single representative duplicate target.", + "summary": "This cluster is a grab bag of unrelated issue reports across tokenizers, multimodal loading, training/runtime errors, and CI; none of the soft pairs look like true duplicates.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue: the items describe different bugs in different code paths and subsystems, so there is no clear duplicate anchor for the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "No best issue for deduplication: the cluster is not centered on one concrete problem, and the candidates do not form a true duplicate set.", + "best_issue_reason": "No strong best issue representative; none of the issues subsume the others or appear to be the same underlying defect.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43334", "right": "issue:45759", "accept": false, - "reason": "Both involve Qwen multimodal loading/configuration, but one is a missing pad_token_id error and the other is AutoModelForCausalLM config unwrapping; different concrete failures." + "reason": "Both involve Qwen multimodal configs, but one is a load-time pad_token_id AttributeError and the other is AutoModelForCausalLM.from_config not unwrapping text_config; different failure modes." }, { "left": "issue:39692", "right": "issue:43295", "accept": false, - "reason": "One is a SigLIP2 documentation example issue, the other is a regression in processor/tokenizer image handling; unrelated bugs." + "reason": "Both are multimodal regression reports, but one is a SigLIP2 documentation example problem and the other is processor.tokenizer/image-passing custom model breakage." }, { "left": "issue:43646", "right": "issue:44466", "accept": false, - "reason": "Custom model initialization breakage and lm_head serialization inconsistency are different code paths and symptoms." + "reason": "Distinct custom-model issues: init_empty_weights initialization failure versus inconsistent lm_head.weight serialization tied to device placement." }, { "left": "issue:45229", "right": "issue:45823", "accept": false, - "reason": "Both are Gemma4-related, but one is CUDA OOM in multi-GPU inference and the other is a device mismatch during forward; not the same bug." + "reason": "Both concern Gemma4 and multi-device behavior, but one is CUDA OOM during inference and the other is a PLE device mismatch during forward." }, { "left": "issue:43824", "right": "issue:44589", "accept": false, - "reason": "ImportError for Qwen2.5-VL and a Float8 storage TypeError are unrelated failures." + "reason": "Different import/runtime problems: Qwen2_5_VL import failure versus missing Float8 storage type." }, { "left": "issue:43504", "right": "issue:43606", "accept": false, - "reason": "BEiT pretrained loading with a legacy field and bark-small CPU offload device mismatch are unrelated." + "reason": "Beit legacy-field loading bug and suno/bark-small CPU-offload device mismatch are unrelated code paths." }, { "left": "issue:33357", "right": "issue:33666", "accept": false, - "reason": "MacOS bus error on CLIP and Qwen2-VL multi-GPU training are different models, environments, and failure modes." + "reason": "Different models and failures: MacOS bus error for community CLIP versus Qwen2-VL multi-GPU training." }, { "left": "issue:43572", "right": "issue:44291", "accept": false, - "reason": "StableLmConfig pad_token_idx regression and init_empty_weights unexpected argument are distinct initialization issues." + "reason": "Both mention transformers 5/init changes, but one is a missing pad_token_idx config field and the other is an unexpected _is_hf_initialized argument TypeError." }, { - "left": "issue:44589", - "right": "issue:45520", + "left": "issue:43927", + "right": "issue:44361", "accept": false, - "reason": "Float8 storage lookup failure and Python 3.13 flash_attn KeyError are unrelated import/runtime problems." + "reason": "Tokenizer/config token-ID persistence bug versus MLukeTokenizer task AttributeError; not the same defect." }, { - "left": "issue:43927", - "right": "issue:44361", + "left": "issue:44589", + "right": "issue:45520", "accept": false, - "reason": "DiaConfig token ID persistence and MLukeTokenizer task AttributeError are different tokenizer/config bugs." + "reason": "Both are import-related, but one is a storage object error and the other is a flash_attn KeyError on Python 3.13." }, { "left": "issue:43504", "right": "issue:43653", "accept": false, - "reason": "BEiT legacy-field loading and BigBirdTokenizer special-token registration are unrelated." + "reason": "Different tokenizer/model-loading bugs: Beit legacy field preset loading versus BigBirdTokenizer special-token registration/empty decode." }, { "left": "issue:44464", "right": "issue:45584", "accept": false, - "reason": "Compiled-forward chunked generation inconsistency and whisper empty-transcription failure after align_special_tokens are different paths." + "reason": "Chunked generation with compiled forward is unrelated to empty transcription after align_special_tokens." }, { - "left": "issue:44423", - "right": "issue:44734", + "left": "issue:43606", + "right": "issue:44112", "accept": false, - "reason": "Both are serve-related crashes, but one is multimodal input handling and the other is KV-cache continuation indexing; not the same issue." + "reason": "One is a real CPU-offload device mismatch for bark-small; the other is a stale GraniteSpeech CI test issue." }, { - "left": "issue:43606", - "right": "issue:44112", + "left": "issue:44423", + "right": "issue:44734", "accept": false, - "reason": "A real CPU-offload device mismatch and a stale GraniteSpeech test failure are not the same bug." + "reason": "Both affect serving, but one crashes on multimodal continuous batching and the other on KV-cache continuation indexing in /v1/responses." }, { "left": "issue:43575", "right": "issue:45405", "accept": false, - "reason": "Qwen2-57B tensor-parallel OOM and a PEFT version bump problem are unrelated." + "reason": "Qwen2 model OOM with tensor parallelism is unrelated to a PEFT version pin bump." }, { "left": "issue:43994", "right": "issue:44493", "accept": false, - "reason": "SigLIP2 nonsensical outputs and unexpected position-id keys are different model/config issues." + "reason": "SigLIP2 nonsensical outputs and unexpected position-id keys are different regression symptoms with different causes." }, { "left": "issue:43278", "right": "issue:43701", "accept": false, - "reason": "Embedding dtype drift between train/eval and resume_from_checkpoint key mismatch are unrelated." + "reason": "BF16-to-FP32 embedding dtype drift in evaluate is unrelated to resume_from_checkpoint key mismatch." }, { "left": "issue:44792", "right": "issue:45405", "accept": false, - "reason": "A failing Janus image-generation test and an unreleased PEFT version bump are unrelated." + "reason": "Failed janus generation test and PEFT minimum-version bump do not describe the same bug." }, { - "left": "issue:29942", - "right": "issue:30333", + "left": "issue:44521", + "right": "issue:45290", "accept": false, - "reason": "Flash Attention 2 test failures and MLflow reporting on failed jobs are unrelated." + "reason": "Both involve chat templates, but all-zero assistant masks for multimodal inputs is a different bug from crashing on assistant tool-call messages with no content." }, { - "left": "issue:44521", - "right": "issue:45290", + "left": "issue:29942", + "right": "issue:30333", "accept": false, - "reason": "Both concern apply_chat_template, but one is zeroed assistant masks for multimodal inputs and the other is a crash on tool-call assistant messages; distinct bugs." + "reason": "Flash Attention 2 test failures and MLFlow reporting after training failure are unrelated." }, { "left": "issue:42490", "right": "issue:43696", "accept": false, - "reason": "Local-path trailing-slash import warning and GPT-oss CUDA OOM are unrelated." + "reason": "Trailing-slash local-model import warning is unrelated to GPT-oss-20b CUDA OOM." }, { "left": "issue:43844", "right": "issue:44964", "accept": false, - "reason": "ZeRO-3 gradient growth and Phi-4 multimodal load failure are different issues." + "reason": "ZeRO-3 gradient anomaly and Phi-4 multimodal loading failure are different issues." }, { "left": "issue:28218", "right": "issue:34567", "accept": false, - "reason": "Tokenizer spacing after added token and TrainerState num_input_tokens_seen not updating are unrelated." + "reason": "Tokenizer extra-space bug and TrainerState token-count tracking bug are unrelated." }, { "left": "issue:44554", "right": "issue:45446", "accept": false, - "reason": "An MPS attention correctness issue and a PyTorch version check bug in flex_attention are different code paths." + "reason": "MPS attention correctness issue and a PyTorch version check bug in flex_attention are different code paths." }, { "left": "issue:43582", "right": "issue:44977", "accept": false, - "reason": "AppleSilicon caching_allocator_warmup TypeError and Qwen3.5 flash-attention generation problems are unrelated." + "reason": "Apple Silicon caching_allocator_warmup TypeError and Qwen3.5 flash-attention generation problems are unrelated." }, { "left": "issue:43540", "right": "issue:45200", "accept": false, - "reason": "Qwen3OmniMoe video processing ValueError and Gemma4 mm_token_type_ids defaulting are different multimodal bugs." + "reason": "Qwen3OmniMoe video-processing ValueError and Gemma4 mm_token_type_ids defaulting are separate multimodal bugs." } ] }, @@ -7274,220 +7230,220 @@ "issue:43278|issue:43856", "issue:42490|issue:43493" ], - "prepared_review_unit_hash": "f9b574db28ade7620fb6718d3367457d5a8871d82f296b3feb628ecc662e46ba", + "prepared_review_unit_hash": "9c78cb0a48db868ea81eaa4ceaf810b1e3378992e3a013b81a07670bde75e844", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 31, - "serialized_chars": 36667, - "estimated_input_tokens": 9167, - "estimated_eval_tokens": 18590 + "serialized_chars": 36676, + "estimated_input_tokens": 9169, + "estimated_eval_tokens": 18594 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a similarity bucket of unrelated bug reports, not a true duplicate set. A few issues share surface-level themes (tokenizers, multimodal processors, config fields, model loading), but none look like the same concrete bug or change.", - "confidence": 0.86, - "canonical_issue_reason": "No single canonical issue cleanly represents the whole cluster because the reports are heterogeneous; if forced to pick a local hub, issue 43295 is the closest anchor among the processor-regression subset.", + "summary": "These items are a heterogeneous mix of unrelated Transformers bugs; none of the soft pairs look like true duplicates, though a few share broad subsystem or model-family overlap.", + "confidence": 0.75, + "canonical_issue_reason": "I\u2019d use issue:37428 as the canonical issue because it is a clear, broadly relevant core-runtime regression and the most representative of the mixed cluster.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43295 is the best representative only in the narrow sense that it has the strongest local connectivity to other processor-related reports, but it still does not unify the cluster into one duplicate bug.", + "best_issue_reason": "issue:37428 is the best representative issue here: concrete, actionable, and library-level, whereas most others are narrower model-specific reports or orthogonal bugs.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:15354", "right": "issue:38175", "accept": false, - "reason": "Different failures in different areas: TorchScript export of a ViT model vs SigLIP2 output-probability behavior." + "reason": "Different failures: torch.jit.script generator-expression support vs SigLIP2 zero-probability output." }, { "left": "issue:45084", "right": "issue:45520", "accept": false, - "reason": "Both mention import-related problems, but one is a template-node compilation error and the other is a Python 3.13 flash_attn import warning/key lookup issue." + "reason": "Unrelated code paths: template-node compilation error vs flash_attn import KeyError on Python 3.13." }, { "left": "issue:45072", "right": "issue:45198", "accept": false, - "reason": "Unrelated model paths and symptoms: dtype mismatch during bfloat16 inference vs save_pretrained/tokenization failure in Wav2Vec2." + "reason": "Different components and symptoms: bfloat16 dtype mismatch vs Wav2Vec2 save/tokenization failure." }, { "left": "issue:45200", "right": "issue:45823", "accept": false, - "reason": "Both involve Gemma4, but one is missing/default mm_token_type_ids and the other is a device-map forward mismatch." + "reason": "Both Gemma4, but one is missing mm_token_type_ids defaulting and the other is a device_map forward mismatch." }, { "left": "issue:37428", "right": "issue:43957", "accept": false, - "reason": "Different concrete bugs: missing import symbol in flash attention utils vs model loading with torch.device(\"meta\")." + "reason": "Both are loading/import regressions, but one is a missing flash attention helper import and the other is meta-device model loading." }, { "left": "issue:43232", "right": "issue:44849", "accept": false, - "reason": "Generation-kwargs update after sync_gpus is unrelated to the Qwen3.5 hidden-states output bug." + "reason": "Generation kwarg update bug vs Qwen3.5 hidden-state output bug; not the same code path." }, { "left": "issue:44315", "right": "issue:44849", "accept": false, - "reason": "Liger Kernel not applied in model_init is a training/integration issue, not the same as Qwen3.5 output_hidden_states behavior." + "reason": "Liger Kernel application in model_init is unrelated to the Qwen3.5 output_hidden_states regression." }, { "left": "issue:43493", "right": "issue:43701", "accept": false, - "reason": "SigLIP2 implementation parity is unrelated to resume_from_checkpoint key mismatch." + "reason": "SigLIP2 implementation discrepancy and resume_from_checkpoint key mismatch are distinct problems." }, { "left": "issue:43295", "right": "issue:44610", "accept": false, - "reason": "Both are processor/model integration regressions, but one is a processor-tokenizer API break and the other is an image-size mismatch for OmDet-Turbo." + "reason": "Processor API regression vs OmDet image-size mismatch; different failures in different processors/models." }, { "left": "issue:44554", "right": "issue:45356", "accept": false, - "reason": "MPS attention correctness is unrelated to the Kimi-K2.5 tokenizer codec/regression issue." + "reason": "MPS attention correctness issue and Kimi tokenizer regression are unrelated." }, { "left": "issue:44898", "right": "issue:45412", "accept": false, - "reason": "Perceiver interpolation failure and RT-DETR memory release are different bugs with different code paths." + "reason": "Perceiver resize/interpolation bug vs RT-DETR memory leak; no shared underlying defect." }, { "left": "issue:44466", "right": "issue:45127", "accept": false, - "reason": "Both touch tied weights, but one is device-dependent lm_head serialization and the other is LoRA merge collapse with extended vocabulary." + "reason": "Both touch tied-weights behavior, but one is serialization inconsistency and the other is LoRA merge collapse after vocab extension." }, { "left": "issue:43572", "right": "issue:44987", "accept": false, - "reason": "Missing pad_token_idx in StableLmConfig is not the same as loading the physical-intelligence/fast model." + "reason": "Missing pad_token_idx in one config is not the same as loading failure for physical-intelligence/fast on 5.1.0." }, { "left": "issue:37428", "right": "issue:43824", "accept": false, - "reason": "Flash-attention import symbol removal is unrelated to importing Qwen2_5_VLForConditionalGeneration." + "reason": "Both are import errors, but they involve different missing symbols/classes and different modules." }, { "left": "issue:45200", "right": "issue:45325", "accept": false, - "reason": "Gemma4 mm_token_type_ids defaults and Qwen2.5-VL rope-index scaling are different multimodal bugs." + "reason": "Both are Gemma/Qwen2.5-VL multimodal issues? No\u2014these are different models and different bugs." }, { "left": "issue:43572", - "right": "issue:43824", + "right": "issue:43957", "accept": false, - "reason": "StableLmConfig pad_token_idx regression is unrelated to the Qwen2_5_VL import error." + "reason": "StableLm config field regression and meta-device loading breakage are separate issues." }, { "left": "issue:43295", "right": "issue:43450", "accept": false, - "reason": "Custom processor/tokenizer regression and batched video processor shape bug are different failures." + "reason": "Processor.tokenizer regression is unrelated to batched video processor output shape." }, { "left": "issue:43844", "right": "issue:44936", "accept": false, - "reason": "ZeRO-3 gradient explosion and trainer.evaluate-after-train are distinct training issues." + "reason": "DeepSpeed/ZeRO gradient growth and trainer evaluate-after-train failure are different training issues." }, { "left": "issue:43295", "right": "issue:43873", "accept": false, - "reason": "Processor/tokenizer regression is unrelated to offloading/quantization behavior." + "reason": "Custom processor API regression and quantization/offloading behavior are not the same bug." }, { "left": "issue:43334", "right": "issue:43525", "accept": false, - "reason": "Both mention missing pad_token_id, but they affect different model configs and are not the same bug." + "reason": "Missing pad_token_id on different configs/models does not indicate the same underlying defect." }, { "left": "issue:44792", "right": "issue:45563", "accept": false, - "reason": "A janus test failure and a stale warning for num_return_sequences are unrelated." + "reason": "A failing Janus test and a stale generation warning are not the same change or bug." }, { "left": "issue:43824", "right": "issue:43975", "accept": false, - "reason": "ImportError for a model class is unrelated to the DeepSeek coder detokenization regression." + "reason": "Importing a missing class and incorrect detokenization are unrelated." }, { "left": "issue:43388", "right": "issue:43526", "accept": false, - "reason": "These are different label-handling bugs in different utilities (gather_for_metrics vs BeitImageProcessorFast)." + "reason": "Both involve labels, but one is metric-gathering on tuple labels and the other is image-processor label reduction." }, { "left": "issue:42915", "right": "issue:43824", "accept": false, - "reason": "Qwen3Moe FP8 config failure and Qwen2_5_VL import error are different model-loading problems." + "reason": "FineGrainedFP8Config failure and Qwen2.5-VL import error are entirely different problems." }, { "left": "issue:38175", "right": "issue:45412", "accept": false, - "reason": "SigLIP2 zero probabilities and RT-DETR memory leak are unrelated." + "reason": "SigLIP2 zero probabilities and RT-DETR memory not being released are unrelated." }, { "left": "issue:38175", "right": "issue:43493", "accept": false, - "reason": "Same model family, but one reports zero probabilities and the other an implementation-vs-JAX discrepancy; not the same concrete bug." + "reason": "Same model family, but one reports zero probabilities and the other a JAX/HF implementation discrepancy; not enough to merge." }, { "left": "issue:44514", "right": "issue:44521", "accept": false, - "reason": "Both are Qwen2_5_VLProcessor chat-template issues, but one is a batched padding crash and the other is all-zero assistant masks; different failure modes." + "reason": "Both concern apply_chat_template, but one is a batching crash and the other is all-zero assistant masks; different failure modes." }, { "left": "issue:44442", "right": "issue:45198", "accept": false, - "reason": "FastSpeech2ConformerTokenizer load failure and Wav2Vec2 save/tokenization failure are different tokenizer bugs." + "reason": "FastSpeech2ConformerTokenizer loading failure is unrelated to Wav2Vec2 save_pretrained/tokenization failure." }, { "left": "issue:28218", "right": "issue:35141", "accept": false, - "reason": "Tokenizer spacing after an added token is unrelated to embedding reinitialization after resizing." + "reason": "Tokenizer spacing after added tokens is unrelated to lm_head reinitialization after resizing embeddings." }, { "left": "issue:43278", "right": "issue:43856", "accept": false, - "reason": "Embedding dtype drift in evaluate is unrelated to Qwen3 MoE memory inefficiency." + "reason": "Embedding dtype drift in eval and Qwen3 MoE training memory inefficiency are different bugs." }, { "left": "issue:42490", "right": "issue:43493", "accept": false, - "reason": "Local-path import warning and SigLIP2 implementation discrepancy are unrelated issues." + "reason": "Local-path import warning and SigLIP2 JAX discrepancy are unrelated." } ] }, @@ -7571,9 +7527,9 @@ "issue:43819|issue:44112", "issue:44610|issue:44857", "issue:37428|issue:43425", + "issue:45290|issue:45381", "issue:43335|issue:43550", "issue:43606|issue:45083", - "issue:45290|issue:45381", "issue:44448|issue:44743", "issue:44857|issue:45593", "issue:45200|issue:45588", @@ -7582,265 +7538,265 @@ "issue:37428|issue:44351", "issue:43611|issue:43931", "issue:44206|issue:45823", - "issue:44360|issue:44561", "issue:43335|issue:43653", + "issue:44360|issue:44561", "issue:39692|issue:43526", "issue:43819|issue:44493", "issue:43550|issue:44112", "issue:43335|issue:43504", "issue:44855|issue:45083", "issue:45310|issue:45478", - "issue:29127|issue:32090", "issue:43335|issue:45083", - "issue:39692|issue:44493", + "issue:29127|issue:32090", "issue:45588|issue:45859", + "issue:39692|issue:44493", "issue:44448|issue:44779", "issue:38175|issue:44493" ], - "prepared_review_unit_hash": "6fa737c7b0bf021caf5d4e422f2691e3c91e92c942d618be152a2dc7ddd65d9c", + "prepared_review_unit_hash": "fc7b3bd01d4dee7b7bf81726766eb50391f31ab17ace2788f06ce5ed616ebde0", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 35, - "serialized_chars": 38289, - "estimated_input_tokens": 9573, - "estimated_eval_tokens": 19402 + "serialized_chars": 38281, + "estimated_input_tokens": 9571, + "estimated_eval_tokens": 19398 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Only one soft-edge pair looks like a true duplicate: the two Qwen3.5 MoE from_pretrained regression reports. The rest are related only at a broad subsystem/theme level or have clearly different failure modes.", + "summary": "This cluster is mostly heterogeneous; nearly all soft pairs are adjacent-but-distinct issues across different models, processors, or regressions. The only clear duplicate is the Qwen3.5-Moe `from_pretrained` error pair (45310/45478).", "confidence": 0.95, - "canonical_issue_reason": "issue:45310 is the better canonical issue: it is the earlier, more specific report of the same transformers>=5.4.0 / Qwen3.5 MoE from_pretrained failure, while issue:45478 reads like a later duplicate.", + "canonical_issue_reason": "issue:45310 is the clearest canonical issue because it is the earlier, more complete report of the same Qwen3.5-Moe `from_pretrained` failure also described in issue:45478.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45310 is the strongest issue candidate overall because it cleanly identifies the concrete regression and is the only clear duplicate anchor in this set.", + "best_issue_reason": "issue:45310 is the best representative of the cluster: it is the only near-certain duplicate target and has the most direct, actionable bug description.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43653", "right": "issue:44112", "accept": false, - "reason": "Different bugs: tokenizer special-token registration vs a GraniteSpeech CI stale device override test." + "reason": "Both are CI-related, but one is a tokenizer special-token decode bug and the other is a GraniteSpeech device-override test failure." }, { "left": "issue:43450", "right": "issue:43575", "accept": false, - "reason": "Different failures: batched video processor shape vs tensor-parallel OOM on Qwen2-57B." + "reason": "Video processor batch-shape bug vs Qwen2-57B tensor-parallel OOM; different subsystems and failure modes." }, { "left": "issue:43827", "right": "issue:43901", "accept": false, - "reason": "Both are docs-related, but they cover different pipelines and different behavior changes." + "reason": "One is about v5 pipeline removal in docs for summarization/translation, the other is TextClassificationPipeline docs mentioning `return_all_scores`." }, { "left": "issue:44206", "right": "issue:45584", "accept": false, - "reason": "Unrelated audio issues: unsupported feature-extractor arg vs Whisper empty-transcription generation after token alignment." + "reason": "Unsupported `center` arg in LasrFeatureExtractor is unrelated to Whisper empty-transcription failure after `align_special_tokens`." }, { "left": "issue:45127", "right": "issue:45245", "accept": false, - "reason": "Different code paths entirely: LoRA merge with tied embeddings vs a category-count limit error." + "reason": "LoRA merge / tied embeddings collapse is unrelated to the category-count limit error." }, { "left": "issue:43262", "right": "issue:43450", "accept": false, - "reason": "Different modalities and symptoms: audio chat-template sampling rate vs video batch shape." + "reason": "Audio chat-template sampling-rate default bug vs video processor batched-shape bug." }, { "left": "issue:43819", "right": "issue:45083", "accept": false, - "reason": "Different model areas: DAC latent reconstruction mismatch vs qwen3_omni_moe feature-length helper behavior." + "reason": "DAC latent/forward mismatch is a different code path from the qwen3_omni_moe feature-length helper behavior." }, { "left": "issue:43819", "right": "issue:44112", "accept": false, - "reason": "No concrete overlap beyond both being bugs; the model/audio paths are unrelated." + "reason": "Different models and failures: DAC missing-STE mismatch vs GraniteSpeech CI device override." }, { "left": "issue:44610", "right": "issue:44857", "accept": false, - "reason": "One is an image processor size mismatch, the other is an AMP/CUDA crash in a loss function." + "reason": "OmDet-Turbo input-size mismatch is unrelated to the LwDetrImageLoss float16/CUDA crash." }, { "left": "issue:37428", "right": "issue:43425", "accept": false, - "reason": "An import error for one helper is not the same as broad Torch 2.10 incompatibility." + "reason": "ImportError for `_flash_supports_window_size` is not the same bug as Torch 2.10 incompatibility." + }, + { + "left": "issue:45290", + "right": "issue:45381", + "accept": false, + "reason": "Chat-template tool-call crash and Qwen2.5-VL video `vision_position_ids` bug affect different paths." }, { "left": "issue:43335", "right": "issue:43550", "accept": false, - "reason": "Different subsystems: SwitchTransformers config bug vs Bamba torch.compile/SDPA failure." + "reason": "SwitchTransformers sparse-layer construction bug vs Bamba torch.compile/SDPA failure." }, { "left": "issue:43606", "right": "issue:45083", "accept": false, - "reason": "CPU offload device mismatch is unrelated to qwen3_omni_moe feature-length behavior." - }, - { - "left": "issue:45290", - "right": "issue:45381", - "accept": false, - "reason": "Different problems in different models: tool-call template crash vs qwen2.5-vl vision_position_ids issue." + "reason": "CPU offload device mismatch in `suno/bark-small` is unrelated to qwen3_omni_moe feature-length helper behavior." }, { "left": "issue:44448", "right": "issue:44743", "accept": false, - "reason": "Both are v4/v5 regression reports, but the affected model behavior is different." + "reason": "Pegasus v4/v5 output regression is unrelated to Qwen modular recurrent-state reset under cache." }, { "left": "issue:44857", "right": "issue:45593", "accept": false, - "reason": "Detection-loss AMP crash is not the same as missing auxiliary losses when denoising is off." + "reason": "AMP/CUDA crash in LwDetrImageLoss is not the same as D-FINE skipping auxiliary losses when denoising is off." }, { "left": "issue:45200", "right": "issue:45588", "accept": false, - "reason": "Gemma 4 token-type-id handling and flash_attention sink-less auxiliary-state handling are unrelated." + "reason": "Gemma 4 `mm_token_type_ids` defaulting issue is distinct from the flash-attention `s_aux=None` crash." }, { "left": "issue:43278", "right": "issue:43825", "accept": false, - "reason": "A dtype regression in training/eval is unrelated to a pipeline error-message bug." + "reason": "Training/evaluate embedding dtype drift is unrelated to the pipeline() v5 error-message regression." }, { "left": "issue:29942", "right": "issue:32090", "accept": false, - "reason": "Flash Attention 2 test failures and a Trainer broadcast NoneType error are different issues." + "reason": "Flash Attention 2 test failures and `_gpu_broadcast_one` TypeError are separate failures in different areas." }, { "left": "issue:37428", "right": "issue:44351", "accept": false, - "reason": "Both are import errors, but they concern different symbols and likely different breakages." + "reason": "Both are import errors, but they miss different symbols (`_flash_supports_window_size` vs `HybridCache`) and imply different fixes." }, { "left": "issue:43611", "right": "issue:43931", "accept": false, - "reason": "Generic model-loading breakage versus a specific Qwen3-VL weight-shape mismatch are not the same bug." + "reason": "`base_model_prefix` loading breakage and Qwen3-VL weight-shape mismatch are different model-loading bugs." }, { "left": "issue:44206", "right": "issue:45823", "accept": false, - "reason": "Both mention device-related problems, but they affect different models and different execution paths." + "reason": "LasrFeatureExtractor `center`-arg regression is unrelated to Gemma4 PLE device-mismatch during forward." }, { - "left": "issue:44360", - "right": "issue:44561", + "left": "issue:43335", + "right": "issue:43653", "accept": false, - "reason": "DSA indexer discussion is unrelated to trust_remote_code breakage from removing is_torch_fx_available." + "reason": "Sparse-layer config bug and tokenizer special-token registration bug do not share the same underlying code path." }, { - "left": "issue:43335", - "right": "issue:43653", + "left": "issue:44360", + "right": "issue:44561", "accept": false, - "reason": "Different models and failures: SwitchTransformers sparse-layer creation vs BigBirdTokenizer decode/special-token bug." + "reason": "DSA indexer missing ReLU is unrelated to removal of `is_torch_fx_available` breaking `trust_remote_code` models." }, { "left": "issue:39692", "right": "issue:43526", "accept": false, - "reason": "SigLIP2 docs/model mismatch is unrelated to BeitImageProcessorFast reduce_labels behavior." + "reason": "SigLIP2 docs/model mismatch and BEiT fast processor `reduce_labels` bug are different issues." }, { "left": "issue:43819", "right": "issue:44493", "accept": false, - "reason": "DAC latent mismatch is unrelated to position-id/key warnings across many models." + "reason": "DAC forward mismatch is not the same as the widespread unexpected position-id key issue." }, { "left": "issue:43550", "right": "issue:44112", "accept": false, - "reason": "torch.compile/SDPA failure is unrelated to the GraniteSpeech CI device-override test." + "reason": "Bamba compile/SDPA failure is unrelated to the stale GraniteSpeech device-override test." }, { "left": "issue:43335", "right": "issue:43504", "accept": false, - "reason": "A config bug in SwitchTransformers is not the same as Beit semantic-segmentation legacy-field loading failure." + "reason": "SwitchTransformers sparse-layer creation bug and BEiT semantic-segmentation legacy-field load failure are unrelated." }, { "left": "issue:44855", "right": "issue:45083", "accept": false, - "reason": "Python 3.13 indentation/parsing failure in DebertaV2 is unrelated to qwen3_omni_moe feature-length logic." + "reason": "Python 3.13 `IndentationError` in DebertaV2Model is unrelated to qwen3_omni_moe feature-length helper behavior." }, { "left": "issue:45310", "right": "issue:45478", "accept": true, - "reason": "Same underlying bug: both report the Qwen3.5 MoE from_pretrained failure on transformers>=5.4.0." - }, - { - "left": "issue:29127", - "right": "issue:32090", - "accept": false, - "reason": "LayoutLMv3 box-content error handling and Trainer GPU broadcast NoneType are unrelated." + "reason": "Same Qwen3.5-Moe `from_pretrained` failure; titles and symptoms match closely enough to be the same bug." }, { "left": "issue:43335", "right": "issue:45083", "accept": false, - "reason": "SwitchTransformers sparse-layer config bug does not match qwen3_omni_moe feature-length behavior." + "reason": "Sparse-layer configuration and qwen3_omni_moe feature-length helper are unrelated." }, { - "left": "issue:39692", - "right": "issue:44493", + "left": "issue:29127", + "right": "issue:32090", "accept": false, - "reason": "SigLIP2 documentation/model mismatch is unrelated to the generic position-id warning issue." + "reason": "LayoutLMv3 error-message clarity issue is unrelated to the Trainer `_gpu_broadcast_one` TypeError." }, { "left": "issue:45588", "right": "issue:45859", "accept": false, - "reason": "Flash-attention auxiliary-state handling and Qwen3_5Moe rotary embedding CPU-offload compatibility are different code paths." + "reason": "Flash-attention `s_aux=None` crash and Qwen3_5Moe CPU-offload incompatibility are different bugs." + }, + { + "left": "issue:39692", + "right": "issue:44493", + "accept": false, + "reason": "SigLIP2 docs/model mismatch is unrelated to the unexpected position-id key issue affecting many models." }, { "left": "issue:44448", "right": "issue:44779", "accept": false, - "reason": "Pegasus output divergence and DeepSeek tokenizer regression are separate model-specific issues." + "reason": "Pegasus output differences and Deepseek tokenizer regressions are distinct issues." }, { "left": "issue:38175", "right": "issue:44493", "accept": false, - "reason": "Unexpected zero probabilities in SigLIP2 are unrelated to the position-id/key warning issue." + "reason": "SigLIP2 zero-probabilities bug is not the same as the generic unexpected position-id key regression." } ] }, "evaluator_result": { "accept": true, - "feedback": "The assessment is grounded in the packet: 45310 and 45478 have the same title and clearly read as the same Qwen3.5 MoE from_pretrained regression, with 45310 earlier, so the canonical choice is reasonable. The other soft-edge verdicts are conservative and mostly rely on distinct model/subsystem and failure-mode differences visible in the titles. No overstatement stands out." + "feedback": "The cluster summary is well grounded in the packet: the only clearly duplicated pair is issue:45310/issue:45478, which have identical titles and closely matching Qwen3.5-Moe from_pretrained reports. The rest of the soft-pair rejections are conservative and consistent with the issue titles/subsystems; no pair is overstated as a duplicate." }, "evaluator_used": true, "retried": false, @@ -7912,13 +7868,13 @@ ], "soft_pairs": [ "issue:43335|issue:45072", - "issue:44292|issue:45520", - "issue:43335|issue:45593", "issue:43720|issue:44112", - "issue:42490|issue:43854", + "issue:43335|issue:45593", + "issue:44292|issue:45520", "issue:44977|issue:45229", - "issue:44561|issue:44779", + "issue:42490|issue:43854", "issue:44462|issue:44561", + "issue:44561|issue:44779", "issue:40444|issue:43526", "issue:34689|issue:35141", "issue:43606|issue:44857", @@ -7931,219 +7887,225 @@ "issue:44568|issue:45491", "issue:43550|issue:45083", "issue:45198|issue:45803", + "issue:29942|issue:35141", "issue:37428|issue:44987", "issue:43873|issue:44493", - "issue:29942|issue:35141", "issue:32090|issue:33357", "issue:45372|issue:45588", "issue:43873|issue:45229", "issue:45459|issue:45712", - "issue:44977|issue:45563", "issue:43412|issue:45588", + "issue:44977|issue:45563", "issue:43278|issue:43329" ], - "prepared_review_unit_hash": "181b5188160d4b143559e20803e5d3ac10384d7533bc4158a5c15da2029d2c1c", + "prepared_review_unit_hash": "194161df7b29dacd7a9bcd748d24aabe08b1c993c426f57912a933f7763c536a", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36911, - "estimated_input_tokens": 9228, - "estimated_eval_tokens": 18712 + "serialized_chars": 36905, + "estimated_input_tokens": 9227, + "estimated_eval_tokens": 18710 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous rather than duplicate-heavy: the soft pairs span unrelated tokenizer, loading, attention, training, quantization, CI, and model-specific regressions. I reject all soft-edge merges.", - "confidence": 0.98, - "canonical_issue_reason": "No single issue is a safe canonical here; the items describe unrelated bugs across different code paths and model families.", + "summary": "This cluster is highly heterogeneous: most items are unrelated tokenizer, model-loading, CI, attention, and serialization bugs. I found no soft pair that clearly describes the same underlying issue tightly enough to merge.", + "confidence": 0.91, + "canonical_issue_reason": "No single canonical issue stands out because the cluster is noisy and spans many unrelated bug classes; the items do not form one cohesive duplicate set.", "canonical_pr_reason": null, - "best_issue_reason": "No global best issue is suitable, because even the broadest-looking reports do not cover the rest of the cluster.", + "best_issue_reason": "issue:43873 is the strongest standalone issue by activity/inbound references and has a clear, reproducible bug report, but it is still not a duplicate hub for the rest of the cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43335", "right": "issue:45072", "accept": false, - "reason": "SwitchTransformers sparse-layer config bug vs Bamba torch.compile/SDPA inference failure; different model and failure mode." + "reason": "Different bugs: SwitchTransformers sparse-layer config vs dtype mismatch in a different model path." }, { - "left": "issue:44292", - "right": "issue:45520", + "left": "issue:43720", + "right": "issue:44112", "accept": false, - "reason": "Qwen-3 NVFP4 runtime error vs a Python 3.13 flash_attn import-utils KeyError; unrelated causes." + "reason": "Different failures: packed-weight loading vs a stale CI device-override test." }, { "left": "issue:43335", "right": "issue:45593", "accept": false, - "reason": "Sparse-layer construction bug vs D-FINE auxiliary-loss behavior; no shared code path." - }, - { - "left": "issue:43720", - "right": "issue:44112", - "accept": false, - "reason": "BitNet packed-weight loading bug vs GraniteSpeech CI device-override test flake; not the same issue." + "reason": "Unrelated model bugs: SwitchTransformers config creation vs D-FINE auxiliary-loss handling." }, { - "left": "issue:42490", - "right": "issue:43854", + "left": "issue:44292", + "right": "issue:45520", "accept": false, - "reason": "Trailing-slash import warning vs GLM-4.7 test loading failure; different loader problems." + "reason": "Both touch import/runtime environment issues, but one is NVFP4 model execution and the other is flash_attn import lookup on Python 3.13." }, { "left": "issue:44977", "right": "issue:45229", "accept": false, - "reason": "Qwen3.5 flash-attention generation bug vs Gemma4 multi-GPU OOM; different failure class." + "reason": "Different concrete problems: flash-attention generation issue vs multi-GPU CUDA OOM." }, { - "left": "issue:44561", - "right": "issue:44779", + "left": "issue:42490", + "right": "issue:43854", "accept": false, - "reason": "Removed helper breaking trust_remote_code vs Deepseek tokenizer regression; different components and symptoms." + "reason": "Local-path warning vs GLM-4.7-Flash model-loading test failure are not the same bug." }, { "left": "issue:44462", "right": "issue:44561", "accept": false, - "reason": "Tokenizer.json precedence bug vs v5 helper removal breakage; not the same underlying defect." + "reason": "Tokenizer repository loading vs v5 removal of is_torch_fx_available affect different code paths." + }, + { + "left": "issue:44561", + "right": "issue:44779", + "accept": false, + "reason": "Different regressions: trust_remote_code import breakage vs DeepSeek tokenizer output corruption." }, { "left": "issue:40444", "right": "issue:43526", "accept": false, - "reason": "Qwen2.5-VL iterable-dataset multi-image failure vs BEiT fast processor label reduction bug; unrelated." + "reason": "Both involve vision preprocessing, but one is iterable multi-image finetuning and the other is BEiT fast label reduction." }, { "left": "issue:34689", "right": "issue:35141", "accept": false, - "reason": "Llama 3.2 vision model loading breakage vs embedding reinitialization after resize; different code paths." + "reason": "Model loading regression vs embedding reinitialization after resize are separate issues." }, { "left": "issue:43606", "right": "issue:44857", "accept": false, - "reason": "CPU offload device mismatch vs float16 AMP crash in LwDetrImageLoss; unrelated bugs." + "reason": "Device-mismatch offload bug and AMP/CUDA loss crash are different failure modes." }, { "left": "issue:44448", "right": "issue:44977", "accept": false, - "reason": "Pegasus v4/v5 output drift vs Qwen3.5 flash-attention generation failure; not mergeable as one fix." + "reason": "Pegasus v4/v5 output drift is unrelated to the Qwen flash-attention generation bug." }, { "left": "issue:44857", "right": "issue:45072", "accept": false, - "reason": "LwDetrImageLoss AMP crash vs SwitchTransformers dtype mismatch; different subsystem and symptom." + "reason": "Different runtime paths: image loss AMP crash vs bfloat16 inference dtype mismatch." }, { "left": "issue:43723", "right": "issue:44554", "accept": false, - "reason": "Tokenizer loading regression in v5 vs MPS attention correctness issue; no shared bug." + "reason": "Tokenizer loading regression is unrelated to the MPS attention correctness issue." }, { "left": "issue:28218", "right": "issue:32090", "accept": false, - "reason": "Tokenizer spacing after added token vs Trainer _gpu_broadcast_one NoneType error; unrelated." + "reason": "Tokenizer spacing bug vs Trainer broadcast TypeError are unrelated." }, { "left": "issue:43701", "right": "issue:45083", "accept": false, - "reason": "resume_from_checkpoint key mismatch vs qwen3_omni_moe length helper bug; different areas." + "reason": "Checkpoint key mismatch is not the same as the qwen3_omni_moe feature-length helper bug." }, { "left": "issue:43927", "right": "issue:45698", "accept": false, - "reason": "Both involve save/load, but one loses custom token IDs and the other loads the wrong custom module; not the same defect." + "reason": "Both involve save/load, but custom token IDs being lost is different from loading the wrong custom module." }, { "left": "issue:44568", "right": "issue:45491", "accept": false, - "reason": "Special-token insertion regression vs Gemma3 NaN embeddings with sliding-window attention; unrelated failures." + "reason": "Special-token addition regression and Gemma3 NaN embeddings have different causes and code paths." }, { "left": "issue:43550", "right": "issue:45083", "accept": false, - "reason": "torch.compile SDPA failure vs feature-extraction length helper bug; different code paths." + "reason": "torch.compile/SDPA failure is unrelated to the feature-length helper behavior bug." }, { "left": "issue:45198", "right": "issue:45803", "accept": false, - "reason": "Wav2Vec2 save/tokenization failure vs bare except swallowing errors in FuyuBatchFeature; not duplicates." + "reason": "Wav2Vec2 save/tokenization failure vs bare-except swallowing errors in FuyuBatchFeature are not the same issue." + }, + { + "left": "issue:29942", + "right": "issue:35141", + "accept": false, + "reason": "Flash Attention 2 test failures are unrelated to embedding reinitialization after resize." }, { "left": "issue:37428", "right": "issue:44987", "accept": false, - "reason": "Missing flash_attention_utils symbol vs model loading failure for a specific repo; too different to merge." + "reason": "Importing a removed flash-attn helper is not the same as a model loading failure for physical-intelligence/fast." }, { "left": "issue:43873", "right": "issue:44493", "accept": false, - "reason": "Quantization offloading issue vs unexpected position-id keys; unrelated." - }, - { - "left": "issue:29942", - "right": "issue:35141", - "accept": false, - "reason": "Flash Attention 2 test failures vs output embedding reinit on resize; different bugs." + "reason": "Quantization/offloading behavior is unrelated to unexpected position-id keys." }, { "left": "issue:32090", "right": "issue:33357", "accept": false, - "reason": "Trainer broadcast NoneType error vs MacOS bus error on CLIP loading; unrelated." + "reason": "Trainer NoneType broadcast error and MacOS bus error with CLIP are different bugs." }, { "left": "issue:45372", "right": "issue:45588", "accept": false, - "reason": "Gemma processor import breakage vs flash_attention.py crash on s_aux=None; different components." + "reason": "Gemma processor import breakage is unrelated to flash_attention.py crashing on s_aux=None." }, { - "left": "issue:45459", - "right": "issue:45712", + "left": "issue:43873", + "right": "issue:45229", "accept": false, - "reason": "Tokenizer error masking when protobuf is absent vs leftover dummy classes; not the same underlying issue." + "reason": "Quantization/offloading issue is not the same as Gemma4 multi-GPU OOM." }, { - "left": "issue:44977", - "right": "issue:45563", + "left": "issue:45459", + "right": "issue:45712", "accept": false, - "reason": "Flash-attention generation bug vs stale paged-generate warning; one is behavioral, the other is warning cleanup." + "reason": "Tokenizer error masking from a bare except is unrelated to leftover dummy class cleanup." }, { "left": "issue:43412", "right": "issue:45588", "accept": false, - "reason": "Executorch export / erfinv support problem vs flash_attention.py AttributeError; unrelated." + "reason": "Executorch export missing training guard/erfinv support is unrelated to flash_attention.py s_aux=None handling." + }, + { + "left": "issue:44977", + "right": "issue:45563", + "accept": false, + "reason": "Flash-attention generation correctness and a stale paged-generate warning are not the same bug." }, { "left": "issue:43278", "right": "issue:43329", "accept": false, - "reason": "Eval/training dtype drift vs undefined video-branch variables in multimodal token counting; different bugs." + "reason": "Training/evaluate dtype drift is unrelated to the undefined video-branch variables in _get_num_multimodal_tokens." } ] }, @@ -8240,208 +8202,208 @@ "issue:44361|issue:45198", "issue:44871|issue:44977", "issue:44451|issue:45520", - "issue:42915|issue:42994", "issue:44265|issue:45198", + "issue:42915|issue:42994", "issue:43421|issue:45491", "issue:43756|issue:45440", "issue:29942|issue:30064" ], - "prepared_review_unit_hash": "a5cdefd389b4d392eb7bb73de36fe3739d6699ea42d849a43fec5a5ab4a14539", + "prepared_review_unit_hash": "6f233c0a52c90b5595e33a8b04601fd6ae19d2f578eb1cefa5bf0ddc7fee0bf1", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 36046, - "estimated_input_tokens": 9012, - "estimated_eval_tokens": 18280 + "serialized_chars": 36052, + "estimated_input_tokens": 9013, + "estimated_eval_tokens": 18282 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "No soft pair looks like a true duplicate or same fix; most are only loosely related by subsystem (tokenization, Gemma, CUDA, etc.) but differ in concrete failure mode and code path. No PRs are present in the cluster, so triage is issue-only.", - "confidence": 0.91, - "canonical_issue_reason": "issue:28218 is the clearest, high-signal user bug report in the set: a specific tokenizer regression with a reproducible output change and active discussion.", + "summary": "The cluster is highly heterogeneous; the soft pairs mostly share only broad subsystem overlap, not the same underlying bug or change. I would not merge any of the candidate pairs as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue stands out: the items span unrelated failures across tokenization, generation, quantization, vision, and model-loading paths.", "canonical_pr_reason": null, - "best_issue_reason": "issue:28218 is the best representative issue because it is specific, still open, and describes a concrete behavioral regression rather than a vague environment or model-specific failure.", + "best_issue_reason": "No strong global best issue for duplication canonicalization because there is no common root cause tying the set together.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41720", "right": "issue:45229", "accept": false, - "reason": "Both are CUDA-related, but one is a Qwen3 device-mapping assert and the other is a Gemma4 multi-GPU OOM; different models and failure modes." + "reason": "Different models and failure modes: device-mapping CUDA assert vs multi-GPU OOM." }, { "left": "issue:10000", "right": "issue:15354", "accept": false, - "reason": "Completely unrelated model/runtime failures: DistilBERT init issue vs a torch.jit scripting/export GeneratorExp limitation." + "reason": "Unrelated model/export problems: DistilBERT issue vs torch.jit GeneratorExp scripting limitation." }, { "left": "issue:43976", "right": "issue:45440", "accept": false, - "reason": "One is a Python-version compatibility issue, the other is a model implementation divergence for DeepSeekV3MoE; no shared underlying bug." + "reason": "Different bug classes: Python-version compatibility vs DeepSeek implementation divergence." }, { "left": "issue:42490", "right": "issue:43575", "accept": false, - "reason": "Local-path import warning and Qwen2 tensor-parallel OOM are unrelated." + "reason": "Local-path import warning and tensor-parallel OOM are unrelated." }, { "left": "issue:45372", "right": "issue:45406", "accept": false, - "reason": "Both mention Gemma4 processor loading, but one is a missing ReasoningEffort import and the other is a missing _tokenizer attribute in serve; different root causes." + "reason": "Both involve Gemma 4 loading, but one is a missing dependency symbol and the other is a missing processor attribute; not the same bug." }, { "left": "issue:28218", "right": "issue:29127", "accept": false, - "reason": "Tokenizer spacing regression vs a LayoutLMv3 error-message clarity issue; not the same bug." + "reason": "Tokenizer spacing behavior vs LayoutLMv3 error-message clarity are different issues." }, { "left": "issue:43065", "right": "issue:43994", "accept": false, - "reason": "Sam3PixelDecoder dummy Conv2d vs SigLIP2 bad outputs are separate model issues." + "reason": "Different subsystems and symptoms: dummy conv layer in Sam3PixelDecoder vs bad SigLIP2 outputs." }, { "left": "issue:45372", "right": "issue:45459", "accept": false, - "reason": "Gemma4 processor import failure and tokenizer protobuf error masking are different tokenizer/loading paths." + "reason": "One is an import dependency failure; the other is tokenizer error handling around protobuf." }, { "left": "issue:44857", "right": "issue:45588", "accept": false, - "reason": "Different components and bugs: LwDetrImageLoss AMP/CUDA crash vs flash_attention.py AttributeError on s_aux=None." + "reason": "Different code paths: LW-DETR AMP crash vs flash-attention integration crash on sink-less models." }, { "left": "issue:43065", "right": "issue:45412", "accept": false, - "reason": "Model architecture placeholder layer vs RT-DETR memory release/leak issue; unrelated." + "reason": "Dummy layer definition and RT-DETR memory release are unrelated." }, { "left": "issue:44448", "right": "issue:44466", "accept": false, - "reason": "Both concern output/serialization behavior, but one is generation output drift and the other is lm_head weight serialization depending on device." + "reason": "Both are v5 regressions, but one is generation output mismatch and the other is serialization behavior; not the same defect." }, { "left": "issue:45245", "right": "issue:45563", "accept": false, - "reason": "Category-count runtime limit and stale generate warning are unrelated." + "reason": "Category-limit runtime error vs stale warning in paged generate are unrelated." }, { "left": "issue:45588", "right": "issue:45823", "accept": false, - "reason": "Flash-attention sinkless-model crash vs Gemma4 PLE device mismatch; different code paths." + "reason": "Different Gemma4 issues: flash-attention s_aux handling vs device mismatch in forward." }, { "left": "issue:43653", "right": "issue:44743", "accept": false, - "reason": "BigBirdTokenizer special-token registration bug vs Qwen3.5 recurrent-state reset bug; not the same issue." + "reason": "Tokenizer special-token registration vs Qwen3.5 recurrent-state cache handling are unrelated." }, { "left": "issue:43278", "right": "issue:43381", "accept": false, - "reason": "Embedding dtype drift and gradient checkpointing in eval mode are distinct training/eval concerns." + "reason": "Eval-mode dtype change and checkpointing-in-eval are separate problems." }, { "left": "issue:25251", "right": "issue:32090", "accept": false, - "reason": "Pipeline top_k output-shape change vs Trainer broadcast NoneType error are unrelated." + "reason": "Pipeline output nesting and Trainer broadcast NoneType error are different failures." }, { "left": "issue:44186", "right": "issue:45198", "accept": false, - "reason": "Different tokenizer/model families and failures: LayoutLMv2 NER/padding crash vs Wav2Vec2 save/tokenization failure." + "reason": "Different tokenizers/models and different failure surfaces: NER padding/truncation crash vs save_pretrained/tokenization failure." }, { "left": "issue:44206", "right": "issue:44479", "accept": false, - "reason": "Feature-extractor center-arg regression and Qwen video-input regression are separate regressions." + "reason": "Different components and regressions: feature extractor center arg vs video input regression in Qwen VL models." }, { "left": "issue:43994", "right": "issue:45412", "accept": false, - "reason": "SigLIP2 incorrect outputs and RT-DETR memory retention are unrelated." + "reason": "SigLIP2 inference corruption and RT-DETR memory leak are unrelated." }, { "left": "issue:43066", "right": "issue:44462", "accept": false, - "reason": "Wrong decoder type in v5 vs AutoTokenizer ignoring tokenizer.json are different tokenizer-loading bugs." + "reason": "Wrong decoder type and ignoring tokenizer.json are both tokenizer-related but not the same underlying bug." }, { "left": "issue:44361", "right": "issue:45198", "accept": false, - "reason": "MLukeTokenizer AttributeError and Wav2Vec2 save/tokenization fail do not share the same failure path." + "reason": "Different tokenizer/model paths and different errors; no shared root cause." }, { "left": "issue:44871", "right": "issue:44977", "accept": false, - "reason": "Gemma3 eos_token_id config mismatch vs Qwen3.5 flash-attention generation failure are unrelated." + "reason": "EOS-token config inconsistency and flash-attention generation regression are unrelated." }, { "left": "issue:44451", "right": "issue:45520", "accept": false, - "reason": "ScandiBERT loading failure and Python 3.13 flash_attn import KeyError are separate compatibility/import problems." + "reason": "Model loading failure and Python 3.13 flash_attn import KeyError are different problems." }, { - "left": "issue:42915", - "right": "issue:42994", + "left": "issue:44265", + "right": "issue:45198", "accept": false, - "reason": "Both involve quantization, but one is a FineGrainedFP8Config load failure and the other is quantized model saving failure." + "reason": "torch.export check failure is unrelated to Wav2Vec2 save/tokenization failure." }, { - "left": "issue:44265", - "right": "issue:45198", + "left": "issue:42915", + "right": "issue:42994", "accept": false, - "reason": "torch.export.export incompatibility with torch_compilable_check is unrelated to Wav2Vec2 tokenization/save_pretrained failure." + "reason": "Both involve quantization, but one is FP8 config failure and the other is quantized-model saving." }, { "left": "issue:43421", "right": "issue:45491", "accept": false, - "reason": "Tokenizer post-processor update bug and Gemma3 NaN embeddings with mixed-length batching are different problems." + "reason": "Runtime post-processor update and Gemma3 NaN embeddings are unrelated." }, { "left": "issue:43756", "right": "issue:45440", "accept": false, - "reason": "RoPE-layer count mismatch in Smollm3 and DeepSeekV3MoE implementation divergence are not the same change." + "reason": "Different model families and defects: RoPE-layer dropping vs DeepSeekV3MoE divergence." }, { "left": "issue:29942", "right": "issue:30064", "accept": false, - "reason": "Flash Attention 2 test failures and void segmentation-map processing are unrelated." + "reason": "FlashAttention 2 test failures and void segmentation-map processing are different issues." } ] }, @@ -8490,7 +8452,6 @@ "issue:43824", "issue:43873", "issue:43931", - "issue:44062", "issue:44186", "issue:44206", "issue:44279", @@ -8498,6 +8459,8 @@ "issue:44464", "issue:44610", "issue:44625", + "issue:44743", + "issue:44811", "issue:44857", "issue:44977", "issue:45072", @@ -8510,8 +8473,7 @@ "issue:45584", "issue:45593", "issue:45834", - "issue:45859", - "issue:45907" + "issue:45859" ], "soft_pairs": [ "issue:33357|issue:39692", @@ -8520,12 +8482,12 @@ "issue:42490|issue:43873", "issue:32090|issue:33290", "issue:44610|issue:45412", - "issue:43824|issue:44279", "issue:45440|issue:45593", + "issue:43824|issue:44279", "issue:43653|issue:44442", "issue:43381|issue:43701", - "issue:43526|issue:45834", "issue:43761|issue:45859", + "issue:43526|issue:45834", "issue:45103|issue:45372", "issue:43526|issue:44186", "issue:29942|issue:36010", @@ -8533,212 +8495,212 @@ "issue:44625|issue:45584", "issue:29127|issue:35141", "issue:43334|issue:43531", - "issue:44977|issue:45245", "issue:44464|issue:45491", - "issue:42915|issue:43931", + "issue:44977|issue:45245", "issue:44442|issue:45072", + "issue:42915|issue:43931", "issue:43262|issue:44206", "issue:43526|issue:44857", "issue:42490|issue:43381", "issue:43575|issue:43701", - "issue:44062|issue:45907" + "issue:44743|issue:44811" ], - "prepared_review_unit_hash": "fbfe63df3c14708dac9246e9521ae2bc91ce91563b8d30666dc81f171037c1ed", + "prepared_review_unit_hash": "e40b762a77f3344e947354ce1e0b3fa67456a0570c3557061e8a8a2dedd556fe", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 35468, - "estimated_input_tokens": 8867, - "estimated_eval_tokens": 17990 + "serialized_chars": 35482, + "estimated_input_tokens": 8871, + "estimated_eval_tokens": 17998 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set spans many unrelated bugs, regressions, and loader/model-specific issues across different components. None of the soft-edge pairs look like the same underlying problem, so all candidate duplicates are rejected.", - "confidence": 0.94, - "canonical_issue_reason": null, + "summary": "The cluster is heterogeneous: the soft pairs share only broad themes (Qwen, processors, loading, training) but not the same concrete bug or change. I would not merge any of these pairs as duplicates.", + "confidence": 0.98, + "canonical_issue_reason": "No canonical issue: the items are mostly unrelated bugs across different models, processors, and trainer/runtime paths, so there is no clear duplicate hub.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43824 is the strongest standalone report here: a concrete import regression with clear impact and multiple inbound references.", + "best_issue_reason": "No single issue is a good global representative; issue 43824 is well-scoped and concrete, but the cluster is too heterogeneous for it to serve as a canonical anchor.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:33357", "right": "issue:39692", "accept": false, - "reason": "Different problems: a MacOS bus error in CLIP loading vs documentation example mistakes for SigLIP2." + "reason": "Both involve multimodal/model-loading problems, but one is a MacOS bus error for a community CLIP model and the other is a documentation example/model-processor mismatch with quantization failure; not the same bug." }, { "left": "issue:43257", "right": "issue:43278", "accept": false, - "reason": "Unrelated bugs: Qwen3 MoE weight conversion under accelerate+deepspeed vs embedding dtype drift in eval." + "reason": "Different failure modes: MOE weights not converted under accelerate+deepspeed versus embedding dtype drifting from BF16 to FP32 during evaluation." }, { "left": "issue:29127", "right": "issue:34567", "accept": false, - "reason": "Different areas: LayoutLMv3 error messaging vs TrainerState token-count tracking." + "reason": "LayoutLMv3 box-info error messaging and TrainerState token counting are unrelated code paths." }, { "left": "issue:42490", "right": "issue:43873", "accept": false, - "reason": "A local-path import warning is not the same as quantization/offloading behavior." + "reason": "Trailing-slash import warning for local models is unrelated to quantization/offloading behavior." }, { "left": "issue:32090", "right": "issue:33290", "accept": false, - "reason": "Trainer NoneType broadcast failure and deepspeed Adafactor OOM are distinct distributed-training issues." + "reason": "Trainer NoneType broadcast failure and AdaFactor/deepspeed OOM are distinct runtime issues." }, { "left": "issue:44610", "right": "issue:45412", "accept": false, - "reason": "Processor/model input-size mismatch is unrelated to RT-DETR memory being retained after deletion." + "reason": "Processor/model input-size mismatch for OmDet-Turbo is unrelated to RT-DETR memory release on deletion." }, { - "left": "issue:43824", - "right": "issue:44279", + "left": "issue:45440", + "right": "issue:45593", "accept": false, - "reason": "Specific missing export/import for Qwen2_5_VL is not the same as a generic dependency issue." + "reason": "DeepseekV3MoE implementation divergence and D-FINE auxiliary-loss behavior are different model bugs." }, { - "left": "issue:45440", - "right": "issue:45593", + "left": "issue:43824", + "right": "issue:44279", "accept": false, - "reason": "DeepSeekV3 implementation divergence and D-FINE auxiliary-loss behavior are different model bugs." + "reason": "Specific missing Qwen2_5_VLForConditionalGeneration import is not the same as a generic dependency issue." }, { "left": "issue:43653", "right": "issue:44442", "accept": false, - "reason": "Both involve tokenizers, but different tokenizers and failure modes; not the same bug." + "reason": "BigBirdTokenizer special-token registration and FastSpeech2ConformerTokenizer load failure are different tokenizer bugs." }, { "left": "issue:43381", "right": "issue:43701", "accept": false, - "reason": "Gradient checkpointing in eval mode and checkpoint key mismatch are separate training-flow issues." + "reason": "Gradient checkpointing in eval mode and resume_from_checkpoint key mismatch are unrelated trainer issues." }, { - "left": "issue:43526", - "right": "issue:45834", + "left": "issue:43761", + "right": "issue:45859", "accept": false, - "reason": "BEiT label reduction bug vs Kosmos2.5 long-OCR index error are unrelated." + "reason": "CLIPVisionModel hidden_states regression and Qwen3_5MoeTextRotaryEmbedding CPU offload incompatibility do not share the same code path." }, { - "left": "issue:43761", - "right": "issue:45859", + "left": "issue:43526", + "right": "issue:45834", "accept": false, - "reason": "CLIPVision hidden_states regression and Qwen3_5Moe CPU-offload incompatibility affect different code paths." + "reason": "BeitImageProcessorFast label reduction bug and Kosmos2.5 OCR long-input index error are unrelated." }, { "left": "issue:45103", "right": "issue:45372", "accept": false, - "reason": "auto_docstring annotation parsing crash and Gemma 4 processor import failure are unrelated." + "reason": "Auto-docstring annotation crash and Gemma 4 processor import failure are unrelated." }, { "left": "issue:43526", "right": "issue:44186", "accept": false, - "reason": "Both are tokenizer/processor reports, but they concern different models and different failure modes." + "reason": "Both are processor/tokenization bugs, but one is BeitImageProcessorFast label reduction and the other is LayoutLMv2Tokenizer NER/padding crashes; different implementations and symptoms." }, { "left": "issue:29942", "right": "issue:36010", "accept": false, - "reason": "Flash Attention 2 test failures are not the same as a GenerationMixin import error." + "reason": "Flash Attention 2 test failures and GenerationMixin import errors are unrelated." }, { "left": "issue:29127", "right": "issue:33357", "accept": false, - "reason": "LayoutLMv3 error clarity issue is unrelated to the CLIP MacOS bus error." + "reason": "LayoutLMv3 error-message clarity and CLIP MacOS bus error are unrelated." }, { "left": "issue:44625", "right": "issue:45584", "accept": false, - "reason": "Qwen3.5 num_labels propagation and Whisper empty-transcription generation are different regressions." + "reason": "Qwen3.5 num_labels propagation and Whisper empty-transcription generation failure are different bugs despite both touching processing/output behavior." }, { "left": "issue:29127", "right": "issue:35141", "accept": false, - "reason": "Error-message clarity for LayoutLMv3 is unrelated to embedding reinitialization after resize." + "reason": "A diagnostic message issue in LayoutLMv3 is unrelated to embedding reinitialization after resize in post_init." }, { "left": "issue:43334", "right": "issue:43531", "accept": false, - "reason": "Both are Qwen3-family issues, but one is a missing pad_token_id and the other is a sliding_window bug." - }, - { - "left": "issue:44977", - "right": "issue:45245", - "accept": false, - "reason": "Flash-attention generation regression and category-cardinality RuntimeError are unrelated." + "reason": "Missing pad_token_id in Qwen3-VL text config and sliding_window behavior in Qwen3-MoE are different configuration bugs." }, { "left": "issue:44464", "right": "issue:45491", "accept": false, - "reason": "Compiled-forward chunked generation inconsistency and Gemma3 NaN embeddings are different generation-path bugs." + "reason": "Compiled-forward chunked generation inconsistency and Gemma3 NaN embeddings in mixed-length batches are distinct numerical/attention issues." }, { - "left": "issue:42915", - "right": "issue:43931", + "left": "issue:44977", + "right": "issue:45245", "accept": false, - "reason": "FineGrainedFP8Config failure and Qwen3-VL weight-shape mismatch are distinct model-loading issues." + "reason": "Qwen3.5 flash-attention generation problems and the 2^24 category limit runtime error are unrelated." }, { "left": "issue:44442", "right": "issue:45072", "accept": false, - "reason": "Tokenizer loading failure and bfloat16 dtype mismatch in inference are unrelated." + "reason": "Tokenizer loading failure and bfloat16 dtype mismatch in inference are different problems." + }, + { + "left": "issue:42915", + "right": "issue:43931", + "accept": false, + "reason": "Both are Qwen loading issues, but one is a FineGrainedFP8Config failure and the other is a weight-shape mismatch for Qwen3-VL; not the same concrete bug." }, { "left": "issue:43262", "right": "issue:44206", "accept": false, - "reason": "Audio chat-template sampling-rate default and LasrFeatureExtractor center-arg crash are different processor bugs." + "reason": "Audio chat-template sample-rate defaulting and LasrFeatureExtractor center-arg crash are separate audio-processing bugs." }, { "left": "issue:43526", "right": "issue:44857", "accept": false, - "reason": "BEiT label reduction and LwDetr float16 AMP crash are unrelated vision issues." + "reason": "BeitImageProcessorFast reduce_labels and LwDetrImageLoss AMP/CUDA crash do not overlap." }, { "left": "issue:42490", "right": "issue:43381", "accept": false, - "reason": "Import warning on local paths is unrelated to gradient-checkpointing restrictions in eval mode." + "reason": "Import warning for local paths and gradient checkpointing eval-mode restrictions are unrelated." }, { "left": "issue:43575", "right": "issue:43701", "accept": false, - "reason": "Tensor-parallel OOM for Qwen2-57B and resume_from_checkpoint key mismatch are different problems." + "reason": "Tensor-parallel OOM for a Qwen2-57B model and checkpoint key mismatch are unrelated." }, { - "left": "issue:44062", - "right": "issue:45907", + "left": "issue:44743", + "right": "issue:44811", "accept": false, - "reason": "AddedToken keyword duplication and list_repo_files revision forwarding are unrelated API bugs." + "reason": "Qwen3_5 recurrent-state reset logic and Whisper batch_decode skip_special_tokens handling are different model-specific bugs." } ] }, @@ -8767,7 +8729,6 @@ "issue:34689", "issue:35141", "issue:36010", - "issue:36331", "issue:41628", "issue:42490", "issue:42907", @@ -8789,6 +8750,7 @@ "issue:43827", "issue:43873", "issue:43931", + "issue:44062", "issue:44077", "issue:44079", "issue:44188", @@ -8796,7 +8758,6 @@ "issue:44442", "issue:44479", "issue:44743", - "issue:44811", "issue:44871", "issue:44898", "issue:44913", @@ -8809,10 +8770,11 @@ "issue:45588", "issue:45698", "issue:45759", - "issue:45803" + "issue:45803", + "issue:45907" ], "soft_pairs": [ - "issue:44743|issue:44811", + "issue:44062|issue:45907", "issue:43232|issue:43388", "issue:43575|issue:44387", "issue:43646|issue:44077", @@ -8821,13 +8783,13 @@ "issue:45003|issue:45440", "issue:44442|issue:45803", "issue:43723|issue:45479", - "issue:43761|issue:44188", "issue:44479|issue:45290", + "issue:43761|issue:44188", "issue:43316|issue:43525", "issue:44442|issue:44898", "issue:33290|issue:34567", - "issue:42907|issue:45698", "issue:42490|issue:44743", + "issue:42907|issue:45698", "issue:34689|issue:43611", "issue:33290|issue:35141", "issue:33290|issue:34689", @@ -8841,85 +8803,84 @@ "issue:43643|issue:44913", "issue:43749|issue:43931", "issue:34689|issue:36010", - "issue:43873|issue:44871", - "issue:33290|issue:36331" + "issue:43873|issue:44871" ], - "prepared_review_unit_hash": "e058929b7a4e0b4ab837a2b89f5d493ed652a24d2c8f6835744ac9daf73f2ff9", + "prepared_review_unit_hash": "bb46484e2bdeaf964227c26dfcef48cda6a3910fcd4345ed298691ab852d27d6", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 31, - "serialized_chars": 37235, - "estimated_input_tokens": 9309, - "estimated_eval_tokens": 18874 + "soft_pair_count": 30, + "serialized_chars": 36916, + "estimated_input_tokens": 9229, + "estimated_eval_tokens": 18714 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is highly heterogeneous: most items are distinct framework regressions or model-specific bugs, not duplicates. I found no soft pair that looks safe to merge, and there are no PRs in the set.", - "confidence": 0.18, - "canonical_issue_reason": "Issue 43646 is the broadest framework-level regression in the set, but the cluster is too mixed to treat it as a true umbrella duplicate.", + "summary": "This cluster is heterogeneous: the items cover unrelated regressions, API breaks, loading/saving bugs, and docs issues. No soft pair looks like the same underlying bug/change, so all candidate merges are rejected.", + "confidence": 0.87, + "canonical_issue_reason": "No true duplicate cluster is evident; if a representative issue is needed, issue 43873 is the strongest signal item because it is open, broadly user-facing, and has the most inbound references/discussion.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43646 is the closest thing to a general representative because it concerns core model initialization behavior rather than a single model or peripheral API surface.", + "best_issue_reason": "Issue 43873 is the best standalone representative: active, high-impact, and clearly a real regression with multiple references.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44743", - "right": "issue:44811", + "left": "issue:44062", + "right": "issue:45907", "accept": false, - "reason": "Different code paths: Qwen3.5 recurrent-state caching vs Whisper processor decoding options." + "reason": "Different bugs: tokenizer AddedToken argument collision vs a repo file listing API regression." }, { "left": "issue:43232", "right": "issue:43388", "accept": false, - "reason": "One is generation bookkeeping after sync_gpus; the other is metric gathering dropping tuple labels." + "reason": "Different code paths: generation kwarg sync issue vs metric gathering dropping tuple labels." }, { "left": "issue:43575", "right": "issue:44387", "accept": false, - "reason": "Both mention OOM, but the causes differ: tensor-parallel loading vs extra CUDA reserved memory under int4 quantization." + "reason": "Both mention OOM, but one is tensor-parallel model loading and the other is int4 quantization reserved-memory growth." }, { "left": "issue:43646", "right": "issue:44077", "accept": false, - "reason": "Custom model initialization regression vs a patchtsmixer post_init API constraint; not the same bug." + "reason": "Related to initialization, but one is a broad v5 custom model init break and the other is a specific `patchtsmixer` post_init policy issue." }, { "left": "issue:43761", "right": "issue:44079", "accept": false, - "reason": "CLIP hidden_states output bug and ModelOutput key assignment bug are unrelated internals." + "reason": "One is CLIPVision hidden_states handling; the other is `ModelOutput` key assignment semantics." }, { "left": "issue:43441", "right": "issue:45588", "accept": false, - "reason": "Both touch flash attention, but one is a Ministral FlashAttention failure and the other is an s_aux=None crash in a different integration path." + "reason": "Both touch flash attention, but they fail in different models and different branches of the integration code." }, { "left": "issue:45003", "right": "issue:45440", "accept": false, - "reason": "Unsafe sys.modules access and DeepseekV3MoE divergence are unrelated issues." + "reason": "Unrelated: unsafe `sys.modules` access vs a DeepSeekV3 implementation divergence." }, { "left": "issue:44442", "right": "issue:45803", "accept": false, - "reason": "Tokenizer loading failure vs a broad exception swallowing errors in Fuyu batch feature conversion." + "reason": "Tokenizer loading failure is unrelated to a bare `except` in Fuyu tensor conversion." }, { "left": "issue:43723", @@ -8928,136 +8889,130 @@ "reason": "Tokenizer loading in v5 is unrelated to a sequence-classification zero-loss bug." }, { - "left": "issue:43761", - "right": "issue:44188", + "left": "issue:44479", + "right": "issue:45290", "accept": false, - "reason": "CLIPVision hidden_states regression and attention-kernel divergence under torch.compile are different problems." + "reason": "Different regressions: multimodal video input handling vs chat template tool-call tokenization." }, { - "left": "issue:44479", - "right": "issue:45290", + "left": "issue:43761", + "right": "issue:44188", "accept": false, - "reason": "Video-input regression for Qwen VL models vs chat template crashes on tool-call messages." + "reason": "CLIPVision hidden_states bug is unrelated to attention-kernel divergence under `torch.compile`." }, { "left": "issue:43316", "right": "issue:43525", "accept": false, - "reason": "Different config mismatches: Gemma3TextConfig API discrepancy vs Llama4 missing pad_token_id." + "reason": "Both are config/API discrepancies, but they concern different models and different missing fields." }, { "left": "issue:44442", "right": "issue:44898", "accept": false, - "reason": "AutoTokenizer loading failure and Perceiver positional interpolation failure do not share a bug path." + "reason": "Tokenizer auto-loading failure is unrelated to Perceiver image-resolution interpolation failure." }, { "left": "issue:33290", "right": "issue:34567", "accept": false, - "reason": "OOM in deepspeed Adafactor training vs TrainerState token-count not updating." + "reason": "OOM in DeepSpeed Adafactor is unrelated to `num_input_tokens_seen` not updating." }, { - "left": "issue:42907", - "right": "issue:45698", + "left": "issue:42490", + "right": "issue:44743", "accept": false, - "reason": "Save/dequantization behavior and wrong custom-module reload after save_pretrained are related only broadly, not the same defect." + "reason": "Local-path import warning is unrelated to recurrent state reset in Qwen3.5 cache usage." }, { - "left": "issue:42490", - "right": "issue:44743", + "left": "issue:42907", + "right": "issue:45698", "accept": false, - "reason": "Trailing-slash local-model import warning is unrelated to Qwen recurrent-state reset behavior." + "reason": "Saving dequantized models is different from loading the wrong custom module after save/load." }, { "left": "issue:34689", "right": "issue:43611", "accept": false, - "reason": "Llama 3.2 vision loading breakage and base_model_prefix handling are distinct loading regressions." + "reason": "Both are model-loading regressions, but they are version- and cause-specific and not the same bug." }, { "left": "issue:33290", "right": "issue:35141", "accept": false, - "reason": "Training OOM with Adafactor is unrelated to output embedding reinitialization in post_init." + "reason": "OOM from optimizer usage is unrelated to embedding reinitialization during `post_init`." }, { "left": "issue:33290", "right": "issue:34689", "accept": false, - "reason": "OOM in optimizer use vs model-loading breakage for Llama 3.2 Vision are different bugs." + "reason": "Different failure modes: training OOM vs model-loading breakage for a specific Llama vision model." }, { "left": "issue:43381", "right": "issue:43873", "accept": false, - "reason": "Gradient checkpointing eval-mode restriction and quantization offloading failure are separate runtime behaviors." + "reason": "Gradient checkpointing eval-mode behavior is unrelated to quantization offloading." }, { "left": "issue:43931", "right": "issue:45216", "accept": false, - "reason": "Weight-shape mismatch on load and incorrect checkpoint save_pretrained behavior are different failure modes." + "reason": "Weight-shape mismatch on load is different from a bad checkpoint produced by `save_pretrained`." }, { "left": "issue:43493", "right": "issue:43827", "accept": false, - "reason": "SigLIP2 implementation discrepancy and stale docs referencing pipeline() are not duplicates." + "reason": "Model implementation discrepancy is unrelated to docs still referencing removed pipeline APIs." }, { "left": "issue:44077", "right": "issue:44479", "accept": false, - "reason": "patchtsmixer post_init policy and Qwen video regression affect different models and code paths." + "reason": "A `post_init` policy issue is unrelated to the video-input regression in Qwen VL models." }, { "left": "issue:43720", "right": "issue:44442", "accept": false, - "reason": "BitNet packed-weight loading and AutoTokenizer loading are unrelated subsystems." + "reason": "BitNet accelerate-loading unpacking bug is unrelated to FastSpeech2ConformerTokenizer auto-loading." }, { "left": "issue:43316", "right": "issue:45759", "accept": false, - "reason": "Both involve config APIs, but Gemma3TextConfig discrepancy and Qwen multimodal text_config unwrapping are different concrete bugs." + "reason": "Gemma3TextConfig API mismatch is unrelated to `from_config` not unwrapping `text_config` for Qwen multimodal configs." }, { "left": "issue:41628", "right": "issue:45372", "accept": false, - "reason": "AutoImageProcessor import error and mistral_common ReasoningEffort import error are separate missing-symbol issues." + "reason": "Both are ImportErrors, but they involve different imports, packages, and loading paths." }, { "left": "issue:43643", "right": "issue:44913", "accept": false, - "reason": "trust_remote_code missing fields and GPTNeoX rotary_pct reset are different config serialization problems." + "reason": "Missing fields with `trust_remote_code` is distinct from `rotary_pct` not persisting on reload." }, { "left": "issue:43749", "right": "issue:43931", "accept": false, - "reason": "FSDP CPU RAM efficient loading and Qwen3-VL shape mismatch are not the same defect." + "reason": "FSDP CPU RAM efficient loading failure is unrelated to a Qwen3-VL shape mismatch on model load." }, { "left": "issue:34689", "right": "issue:36010", "accept": false, - "reason": "Model-loading regression for Llama 3.2 vision vs GenerationMixin import failure are unrelated." + "reason": "Different import/load regressions: a specific Llama vision loading break vs missing `GenerationMixin` import." }, { "left": "issue:43873", "right": "issue:44871", "accept": false, - "reason": "Quantization/offloading memory growth and Gemma-3 eos_token_id inconsistency do not share a code path." - }, - { - "left": "issue:33290", - "right": "issue:36331", - "accept": false, - "reason": "Adafactor OOM in deepspeed and CustomTrainer.compute_loss signature mismatch are unrelated training bugs." + "reason": "Quantization/offloading behavior is unrelated to an EOS token config inconsistency in Gemma-3." } ] }, @@ -9085,6 +9040,8 @@ "issue:22355", "issue:25251", "issue:28282", + "issue:33290", + "issue:36331", "issue:41950", "issue:42175", "issue:42617", @@ -9116,7 +9073,6 @@ "issue:44514", "issue:44534", "issue:44617", - "issue:44734", "issue:44743", "issue:44792", "issue:44855", @@ -9126,17 +9082,17 @@ "issue:45127", "issue:45137", "issue:45200", - "issue:45406", "issue:45584", "issue:45698" ], "soft_pairs": [ + "issue:33290|issue:36331", "issue:44190|issue:44496", "issue:44855|issue:45137", "issue:43540|issue:44514", "issue:43316|issue:44877", - "issue:42175|issue:43316", "issue:43065|issue:44617", + "issue:42175|issue:43316", "issue:25251|issue:28282", "issue:43295|issue:43825", "issue:43404|issue:43441", @@ -9156,128 +9112,132 @@ "issue:43575|issue:43742", "issue:43452|issue:44496", "issue:43646|issue:44913", - "issue:43504|issue:45698", - "issue:44734|issue:45406", - "issue:44617|issue:45584" + "issue:43504|issue:45698" ], - "prepared_review_unit_hash": "1fdb33ab47704883292917ebc42b40d45668f4b03157b55424017d9ca4c5c38d", + "prepared_review_unit_hash": "fc2ffce13f45d2c39d624aa0b57ecbff5bc9b677a417635fa5daaec6ea81a59f", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 28, - "serialized_chars": 36029, - "estimated_input_tokens": 9008, - "estimated_eval_tokens": 18272 + "soft_pair_count": 27, + "serialized_chars": 35676, + "estimated_input_tokens": 8919, + "estimated_eval_tokens": 18094 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are mostly unrelated Transformers issues that only share broad subsystem overlap (generation, processors, configs, multimodal models). None of the proposed soft edges look like true duplicates or mergeable same-bug reports.", - "confidence": 0.83, - "canonical_issue_reason": "Issue 43646 is the broadest representative of the cluster\u2019s common theme: a Transformers 5.0.0 regression that breaks custom model initialization. It\u2019s general enough to serve as the best anchor, even though the surrounding items are not duplicates.", + "summary": "These items are mostly unrelated standalone issues across training, loading, generation, multimodal, and config code paths. None of the soft pairs look like true duplicates of the same underlying bug or change.", + "confidence": 0.91, + "canonical_issue_reason": "issue:45698 is the strongest representative issue here: it is open, has active discussion, and describes a concrete, user-facing from_pretrained/save_pretrained loading regression with a clear reproduction path.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43646 is the best single issue to represent the cluster because it captures the most general, cross-cutting v5 breakage among the set. It is clearer and broader than the more model-specific reports.", + "best_issue_reason": "issue:45698 is the best issue candidate because it targets a specific core model-loading path, has ongoing activity, and is actionable for triage/repair.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:33290", + "right": "issue:36331", + "accept": false, + "reason": "Different failures: DeepSpeed/Adafactor OOM versus CustomTrainer signature mismatch." + }, { "left": "issue:44190", "right": "issue:44496", "accept": false, - "reason": "Different failures: local dataset loading in a training script vs model config missing model_type during AutoModel resolution." + "reason": "Unrelated bugs: local dataset loading in an example script versus model config/type resolution." }, { "left": "issue:44855", "right": "issue:45137", "accept": false, - "reason": "Completely different code paths: torch.jit parsing on Python 3.13 vs DeepSpeed ZeRO3 deque handling." + "reason": "Different code paths and symptoms: Python 3.13 import/parsing failure versus ZeRO3 deque error." }, { "left": "issue:43540", "right": "issue:44514", "accept": false, - "reason": "Both are multimodal, but they target different models and different bugs: video input validation vs batched chat-template padding behavior." + "reason": "Both involve video/multimodal inputs, but one is Qwen3OmniMoe video processing and the other is batched chat templating padding behavior." }, { "left": "issue:43316", "right": "issue:44877", "accept": false, - "reason": "Config API inconsistency for Gemma3TextConfig is not the same as strict config rejection for granite_speech." + "reason": "Both touch config APIs, but one is a Gemma3TextConfig discrepancy and the other is strict loading of granite_speech config; not the same bug." }, { - "left": "issue:42175", - "right": "issue:43316", + "left": "issue:43065", + "right": "issue:44617", "accept": false, - "reason": "Packaging extras/backend dependency issue is unrelated to a model config API discrepancy." + "reason": "Same model family prefix, but one is a dummy Conv2d in Sam3PixelDecoder and the other is CUDA OOM in Sam3Video." }, { - "left": "issue:43065", - "right": "issue:44617", + "left": "issue:42175", + "right": "issue:43316", "accept": false, - "reason": "A dummy Conv2d placeholder in one SAM3 component is not the same bug as CUDA OOM in Sam3Video." + "reason": "Different subsystems: backend extra installation versus model config API mismatch." }, { "left": "issue:25251", "right": "issue:28282", "accept": false, - "reason": "Pipeline top_k output nesting and a missing transformers.onnx import are unrelated." + "reason": "Pipeline output-shape change and missing PyTorch import are unrelated." }, { "left": "issue:43295", "right": "issue:43825", "accept": false, - "reason": "Processor/tokenizer regression is different from an error-message issue about translation tasks in v5." + "reason": "Both are regressions around processors/pipelines, but one is custom tokenizer/processor breakage and the other is an incorrect error message." }, { "left": "issue:43404", "right": "issue:43441", "accept": false, - "reason": "Tied lm_head weights in Mistral3ForConditionalGeneration and FlashAttention failures in Ministral-3 are distinct model bugs." + "reason": "Different bugs in nearby model families: weight tying versus FlashAttention failure." }, { "left": "issue:43673", "right": "issue:44743", "accept": false, - "reason": "Both mention cache, but one is a missing GenerationMixin cache in chunked_prefill and the other is recurrent state reset in modular_qwen3_5.py." + "reason": "Both mention cache/state, but they affect different models and different generation mechanisms." }, { "left": "issue:44162", "right": "issue:44792", "accept": false, - "reason": "ESM2 breakage and a Janus image-generation test failure are different model-specific regressions." + "reason": "ESM2 model breakage and a janus test failure are not the same issue." }, { "left": "issue:43388", "right": "issue:43688", "accept": false, - "reason": "Metric gathering dropping tuple labels is unrelated to auxiliary-loss normalization in MoE models." + "reason": "Different training/eval bugs: gather_for_metrics tuple truncation versus auxiliary-loss normalization." }, { "left": "issue:44077", "right": "issue:44534", "accept": false, - "reason": "Optional post_init enforcement for patchtsmixer is not the same as non-persistent buffers being filled with junk." + "reason": "Optional post_init enforcement and non-persistent buffer initialization are separate issues." }, { "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 question for DeBERTa-V3 is unrelated to missing transformers.onnx import." }, { "left": "issue:43550", "right": "issue:43701", "accept": false, - "reason": "torch.compile/SDPA OOM and resume_from_checkpoint key mismatch are unrelated training/runtime issues." + "reason": "torch.compile/SDPA OOM and resume_from_checkpoint key mismatch are unrelated." }, { "left": "issue:43606", @@ -9289,73 +9249,61 @@ "left": "issue:44871", "right": "issue:45584", "accept": false, - "reason": "Token ID config inconsistency for Gemma-3 is unrelated to Whisper empty-transcription generation failure." + "reason": "Both involve token/generation settings, but one is eos_token_id config inconsistency and the other is Whisper empty-transcription handling." }, { "left": "issue:41950", "right": "issue:43475", "accept": false, - "reason": "Video-classification pipeline processor selection and a missing fpn_position_embeddings attribute in Sam3VisionEncoderOutput are different bugs." + "reason": "Pipeline processor lookup bug and Sam3 vision encoder attribute error are different problems." }, { "left": "issue:43761", "right": "issue:44479", "accept": false, - "reason": "CLIPVision hidden_states regression and Qwen video-input regression are separate model-family issues." + "reason": "CLIP hidden_states regression and Qwen video-input regression are unrelated." }, { "left": "issue:43295", "right": "issue:44871", "accept": false, - "reason": "Processor/tokenizer regression is not the same as an eos_token_id configuration mismatch." + "reason": "Custom processor/tokenizer regression and Gemma eos_token_id mismatch do not share the same concrete code path." }, { "left": "issue:45127", "right": "issue:45200", "accept": false, - "reason": "LoRA merge collapse with tied embeddings is unrelated to missing mm_token_type_ids defaulting." + "reason": "LoRA merge/collapse with tied embeddings is unrelated to Gemma 4 mm_token_type_ids defaults." }, { "left": "issue:42617", "right": "issue:43575", "accept": false, - "reason": "A 3d_parallel script failure is not the same bug as tensor-parallel OOM loading a large model." + "reason": "3d_parallel script failure and tensor-parallel OOM are different issues." }, { "left": "issue:43575", "right": "issue:43742", "accept": false, - "reason": "OOM during tp load and a key error loading MobileLLM-125M are different issues." + "reason": "OOM when loading a large model with TP is not the same as a MobileLLM config KeyError." }, { "left": "issue:43452", "right": "issue:44496", "accept": false, - "reason": "gguf_file/from_pretrained breakage is not the same as an unrecognized model_type in config.json." + "reason": "GGUF tokenizer/model loading breakage and missing model_type in config.json are distinct loading errors." }, { "left": "issue:43646", "right": "issue:44913", "accept": false, - "reason": "Custom model initialization breakage and GPTNeoXConfig rotary_pct not persisting on reload are related only at a very broad config level." + "reason": "Transformer 5 custom initialization breakage and GPTNeoX rotary_pct reload persistence are different bugs." }, { "left": "issue:43504", "right": "issue:45698", "accept": false, - "reason": "Legacy field rejection for Beit preset loading is distinct from loading the wrong custom module after save_pretrained." - }, - { - "left": "issue:44734", - "right": "issue:45406", - "accept": false, - "reason": "KV-cache continuation indexing in serve and a missing _tokenizer on Gemma4Processor are different server/runtime bugs." - }, - { - "left": "issue:44617", - "right": "issue:45584", - "accept": false, - "reason": "SAM3 video OOM and Whisper empty-transcription generation failure are unrelated." + "reason": "Legacy-field preset loading for Beit semantic segmentation is not the same as loading the wrong custom module after save_pretrained." } ] }, @@ -9386,12 +9334,12 @@ "issue:43065", "issue:43335", "issue:43388", - "issue:43582", "issue:43644", "issue:43761", "issue:43827", "issue:43906", "issue:43931", + "issue:43976", "issue:43994", "issue:44060", "issue:44077", @@ -9404,6 +9352,7 @@ "issue:44521", "issue:44534", "issue:44610", + "issue:44617", "issue:44734", "issue:44737", "issue:44792", @@ -9411,7 +9360,6 @@ "issue:44871", "issue:44913", "issue:44933", - "issue:44938", "issue:44964", "issue:44977", "issue:45127", @@ -9420,7 +9368,7 @@ "issue:45357", "issue:45362", "issue:45381", - "issue:45542", + "issue:45406", "issue:45563", "issue:45584", "issue:45588", @@ -9429,6 +9377,8 @@ "issue:45907" ], "soft_pairs": [ + "issue:44734|issue:45406", + "issue:44617|issue:45584", "issue:43761|issue:44242", "issue:32090|issue:35141", "issue:43931|issue:44060", @@ -9436,217 +9386,223 @@ "issue:44464|issue:44871", "issue:44479|issue:44734", "issue:43065|issue:44610", - "issue:44792|issue:44977", + "issue:43906|issue:45290", "issue:44521|issue:45381", + "issue:44792|issue:44977", "issue:43827|issue:43994", "issue:44464|issue:45563", "issue:44534|issue:44913", - "issue:43906|issue:45290", "issue:44190|issue:44964", "issue:41720|issue:45245", "issue:44933|issue:45362", "issue:44792|issue:45600", "issue:44855|issue:45588", "issue:35141|issue:42490", - "issue:43388|issue:43582", + "issue:43388|issue:44079", "issue:45127|issue:45357", "issue:44737|issue:45803", "issue:44462|issue:45907", "issue:44077|issue:44079", "issue:43335|issue:44855", "issue:43644|issue:44913", - "issue:44938|issue:45542" + "issue:43976|issue:44855" ], - "prepared_review_unit_hash": "aa8b8d9800375b15112185a8e79a51d05671b3f3d36a2ffd4deccf9f70488998", + "prepared_review_unit_hash": "21ca06a3dd72739437b53f6c63248c4b1cbd8b2602d67cf17c14cd52af5ae3f7", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 27, - "serialized_chars": 36010, - "estimated_input_tokens": 9003, - "estimated_eval_tokens": 18262 + "soft_pair_count": 29, + "serialized_chars": 36609, + "estimated_input_tokens": 9153, + "estimated_eval_tokens": 18562 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These issues are largely unrelated across different models, configs, and runtime paths; none of the soft links look like true duplicates.", + "summary": "This cluster is mostly a set of unrelated bug reports that only share broad transformer/model-loading themes. I would anchor the cluster on the v5 non-persistent-buffer regression report, and reject all listed soft duplicate links as too semantically distant or too underspecified to merge safely.", "confidence": 0.94, - "canonical_issue_reason": null, + "canonical_issue_reason": "issue:43644 is the clearest umbrella report for a concrete v5 regression affecting non-persistent buffers; it is broad, specific, and central enough to serve as the representative issue.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 44521 is the most active/central item in the set (open and referenced), but it is still only a weak cluster anchor rather than a true duplicate target.", + "best_issue_reason": "issue:43644 is the strongest standalone representative because it describes a concrete framework-wide regression rather than a model-specific symptom or a downstream effect.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44734", + "right": "issue:45406", + "accept": false, + "reason": "Different failures: KV-cache continuation tensor indexing in serve vs a missing _tokenizer attribute in Gemma4Processor." + }, + { + "left": "issue:44617", + "right": "issue:45584", + "accept": false, + "reason": "Unrelated models and symptoms: Sam3Video CUDA OOM vs Whisper empty transcription after align_special_tokens." + }, { "left": "issue:43761", "right": "issue:44242", "accept": false, - "reason": "Different bugs: one is CLIPVisionModel hidden_states output regression, the other is MoE load-balancing loss handling when router logits are disabled." + "reason": "One is CLIPVision hidden_states being None; the other is router/load-balancing loss omission in a MoE model. Not the same bug." }, { "left": "issue:32090", "right": "issue:35141", "accept": false, - "reason": "Trainer GPU broadcast TypeError and embedding reinitialization after resize are unrelated code paths and symptoms." + "reason": "Trainer GPU broadcast TypeError is unrelated to embedding resizing/post_init reinitialization." }, { "left": "issue:43931", "right": "issue:44060", "accept": false, - "reason": "Both touch Qwen3 loading, but one is a weight-shape mismatch and the other is a tied-weights warning bug; not the same underlying issue." + "reason": "Weight-shape mismatch while loading Qwen3-VL is not the same as a tied-weights warning incorrectly linking dt_bias." }, { "left": "issue:45563", "right": "issue:45584", "accept": false, - "reason": "Both are generation-related, but stale num_return_sequences warning and Whisper empty-transcription failure are distinct bugs." + "reason": "A stale warning about num_return_sequences is not the same as Whisper failing on empty transcription." }, { "left": "issue:44464", "right": "issue:44871", "accept": false, - "reason": "Chunked generation inconsistency with compiled forward is unrelated to Gemma-3 eos_token_id config mismatch." + "reason": "Chunked generation with compiled forward and Gemma eos_token_id config mismatch are different problems." }, { "left": "issue:44479", "right": "issue:44734", "accept": false, - "reason": "Video-input regression and KV-cache continuation indexing crash are different failures in different execution paths." + "reason": "Both involve generation, but one is a video-input regression in Qwen VL models and the other is a serve KV-cache continuation crash." }, { "left": "issue:43065", "right": "issue:44610", "accept": false, - "reason": "Dummy Conv2d in Sam3PixelDecoder and OmDet-Turbo processor image-size mismatch are unrelated vision-model issues." + "reason": "A dummy Conv2d in Sam3PixelDecoder is not the same as OmDet-Turbo processor/model image-size mismatch." }, { - "left": "issue:44792", - "right": "issue:44977", + "left": "issue:43906", + "right": "issue:45290", "accept": false, - "reason": "Failed janus image-generation test and Qwen3.5 flash-attention generation bug are different model-specific problems." + "reason": "The reproduction issue lacks enough concrete overlap; apply_chat_template tool-call crashes are a different failure mode." }, { "left": "issue:44521", "right": "issue:45381", "accept": false, - "reason": "Both are multimodal, but all-zero assistant masks in chat templating and wrong vision_position_ids in Qwen2.5-VL are not the same bug." + "reason": "All-zero assistant masks in apply_chat_template are distinct from wrong vision_position_ids in Qwen2.5-VL video input." + }, + { + "left": "issue:44792", + "right": "issue:44977", + "accept": false, + "reason": "Janus image-generation test failure and Qwen3.5 flash-attention generation issues are not the same code-path bug." }, { "left": "issue:43827", "right": "issue:43994", "accept": false, - "reason": "Docs referencing pipeline() after removals and SigLIP2 producing bad outputs are unrelated." + "reason": "Docs still referencing pipeline() is unrelated to SigLIP2 producing bad outputs under AutoModel/pipeline." }, { "left": "issue:44464", "right": "issue:45563", "accept": false, - "reason": "Both involve generation behavior, but compiled-forward inconsistency and a stale warning are separate issues." + "reason": "Compiled-forward chunked generation inconsistency is not the same as a stale warning about num_return_sequences." }, { "left": "issue:44534", "right": "issue:44913", "accept": false, - "reason": "Non-persistent buffer initialization junk and GPTNeoXConfig rotary_pct reload loss are different persistence/config bugs." - }, - { - "left": "issue:43906", - "right": "issue:45290", - "accept": false, - "reason": "One is an isolated reproduction of a linked issue; the other is a crash on assistant tool-call messages. The relationship is too weak to treat as the same bug." + "reason": "A v5 non-persistent-buffer junk regression is unrelated to GPTNeoX rotary_pct not persisting on reload." }, { "left": "issue:44190", "right": "issue:44964", "accept": false, - "reason": "Local dataset loading in a training script and loading a specific multimodal model are different failure modes." + "reason": "Local dataset loading in a no-trainer script is not the same as loading a multimodal model failing in latest transformers." }, { "left": "issue:41720", "right": "issue:45245", "accept": false, - "reason": "Qwen3 auto device mapping cudaErrorAssert and category-count overflow are unrelated." + "reason": "Qwen3 auto device-mapping cudaErrorAssert is unrelated to the 2^24 category-count limit." }, { "left": "issue:44933", "right": "issue:45362", "accept": false, - "reason": "Missing import from image_utils and Qwen3.5 chat crash do not share an underlying code-path." + "reason": "A missing import from image_utils is unrelated to Qwen3.5 chat crashes." }, { "left": "issue:44792", "right": "issue:45600", "accept": false, - "reason": "Janus test failure and stale auto_mappings references are different CI breakages." + "reason": "Janus image-generation test failure and stale auto_mappings references to removed configs are different issues." }, { "left": "issue:44855", "right": "issue:45588", "accept": false, - "reason": "Python 3.13 import/parsing failure in DebertaV2 and flash_attention AttributeError on sink-less models are unrelated." + "reason": "Python 3.13 IndentationError in DebertaV2Model is unrelated to flash_attention crashing on s_aux=None." }, { "left": "issue:35141", "right": "issue:42490", "accept": false, - "reason": "Embedding reinitialization after resize and slash-suffixed local model import warning are different issues." + "reason": "Embedding reinitialization in post_init is unrelated to a local-path trailing-slash import warning." }, { "left": "issue:43388", - "right": "issue:43582", + "right": "issue:44079", "accept": false, - "reason": "Metric gathering truncation and Apple Silicon warmup TypeError are unrelated." + "reason": "A gather_for_metrics truncation bug is unrelated to ModelOutput key assignment when a field was previously None." }, { "left": "issue:45127", "right": "issue:45357", "accept": false, - "reason": "LoRA merge collapse with extended vocab and incorrect visual encoder keys on save_pretrained are different model-save bugs." + "reason": "LoRA merge collapse with extended vocab is unrelated to incorrect visual-encoder keys saved by Qwen3.5." }, { "left": "issue:44737", "right": "issue:45803", "accept": false, - "reason": "CPU arange device bug in XLNet and bare except swallowing errors in FuyuBatchFeature are unrelated." + "reason": "XLNet device placement bug and a bare-except swallowing KeyboardInterrupt are unrelated." }, { "left": "issue:44462", "right": "issue:45907", "accept": false, - "reason": "AutoTokenizer ignoring tokenizer.json and list_repo_files missing revision forwarding are different repository/API issues." + "reason": "AutoTokenizer ignoring tokenizer.json is unrelated to list_repo_files not forwarding revision." }, { "left": "issue:44077", "right": "issue:44079", "accept": false, - "reason": "Optional post_init support in patchtsmixer and ModelOutput None-key assignment are separate logic bugs." + "reason": "patchtsmixer post_init policy is unrelated to ModelOutput key assignment semantics." }, { "left": "issue:43335", "right": "issue:44855", "accept": false, - "reason": "SwitchTransformers sparse-layer config bug and DebertaV2 Python 3.13 import failure are unrelated." + "reason": "SwitchTransformers config creating a sparse layer at zero sparse layers is unrelated to the DebertaV2 Python 3.13 parsing error." }, { "left": "issue:43644", "right": "issue:44913", "accept": false, - "reason": "Non-persistent buffer junk and config reload reverting rotary_pct are different state-handling problems." - }, - { - "left": "issue:44938", - "right": "issue:45542", - "accept": false, - "reason": "Python 3.14 import/load failure and missing TensorFlow backend when only tensorboard is installed are unrelated." + "reason": "A framework-wide v5 non-persistent-buffer regression is unrelated to GPTNeoX rotary_pct persistence." } ] }, @@ -9703,29 +9659,29 @@ "issue:44561", "issue:44683", "issue:44734", - "issue:44855", "issue:44913", + "issue:44938", "issue:44964", "issue:44987", "issue:45003", "issue:45216", "issue:45245", "issue:45276", - "issue:45335", "issue:45375", "issue:45446", + "issue:45542", "issue:45584", "issue:45600", "issue:45632", "issue:45803" ], "soft_pairs": [ - "issue:43976|issue:44855", + "issue:44938|issue:45542", "issue:43550|issue:44190", "issue:43540|issue:44734", "issue:44112|issue:45600", - "issue:43650|issue:44220", "issue:43335|issue:44190", + "issue:43650|issue:44220", "issue:45276|issue:45584", "issue:44496|issue:45803", "issue:39692|issue:44190", @@ -9734,8 +9690,8 @@ "issue:41950|issue:44016", "issue:43479|issue:45375", "issue:36246|issue:41628", - "issue:44186|issue:44734", "issue:36246|issue:38617", + "issue:44186|issue:44734", "issue:38175|issue:43065", "issue:42175|issue:44279", "issue:44683|issue:45446", @@ -9751,235 +9707,228 @@ "issue:44112|issue:44964", "issue:43854|issue:44964", "issue:42907|issue:43441", - "issue:43976|issue:44987", - "issue:43295|issue:45335" + "issue:43976|issue:44987" ], - "prepared_review_unit_hash": "65d26110ffc8216a6a80d45bf030cbc435743b81f035fd6bfeac5a1a354983a9", + "prepared_review_unit_hash": "808b39d7c3fc81174cc9abbc4719d3a502b4cdf1e3f67571b30916dbae65abc9", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 33, - "serialized_chars": 37692, - "estimated_input_tokens": 9423, - "estimated_eval_tokens": 19102 + "soft_pair_count": 32, + "serialized_chars": 37304, + "estimated_input_tokens": 9326, + "estimated_eval_tokens": 18908 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are mostly a loose set of unrelated Transformers issues centered around import/load regressions, config bugs, pipeline failures, and model-specific breakages. None of the soft pairs look like the same underlying bug/change.", - "confidence": 0.76, - "canonical_issue_reason": "issue:41628 is the most generic, core-API-facing failure in the set (a direct transformers import/export regression) and is a reasonable anchor for this heterogeneous cluster.", + "summary": "This set is heterogeneous: it groups several unrelated import, config, model-loading, and runtime-regression reports, but none of the soft-linked pairs look like the same underlying bug/change.", + "confidence": 0.39, + "canonical_issue_reason": "Issue 41628 is the broadest import/export regression in the set ('Cannot import name AutoImageProcessor'), so it is the closest anchor for the import-related failures, though the overall cluster is still not a true duplicate group.", "canonical_pr_reason": null, - "best_issue_reason": "issue:41628 has the clearest, broadest user impact and the most canonical failure signature among the issues here.", + "best_issue_reason": "Issue 44964 is the strongest standalone triage candidate: it is a concrete, reproducible latest-transformers model-load failure with a specific model and clear user impact.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43976", - "right": "issue:44855", + "left": "issue:44938", + "right": "issue:45542", "accept": false, - "reason": "Both are Python/version-related failures, but one is a Transformers 5.1.0 compatibility problem and the other is a DebertaV2 import parsing error on Python 3.13; different code paths and fixes." + "reason": "Different failure modes: Python 3.14 import/load breakage vs an undefined TensorFlow backend caused by installing only tensorboard." }, { "left": "issue:43550", "right": "issue:44190", "accept": false, - "reason": "Torch.compile/SDPA model failure vs local dataset loading in an example script; unrelated subsystems and symptoms." + "reason": "One is a Bamba SDPA/torch.compile bug; the other is a local dataset loading failure in an example script." }, { "left": "issue:43540", "right": "issue:44734", "accept": false, - "reason": "One is Qwen3OmniMoe video input validation, the other is serve KV-cache continuation tensor indexing; both multimodal, but not the same bug." + "reason": "Different code paths: Qwen3OmniMoe video input validation vs transformers-serve KV-cache continuation indexing." }, { "left": "issue:44112", "right": "issue:45600", "accept": false, - "reason": "Both mention CI failures, but one is a GraniteSpeech test regression and the other is stale auto_mappings references; different root causes." + "reason": "Both mention CI, but the bugs are unrelated: a GraniteSpeech device-override test vs stale auto-mapping references to removed configs." }, { - "left": "issue:43650", - "right": "issue:44220", + "left": "issue:43335", + "right": "issue:44190", "accept": false, - "reason": "Generic placeholder issue vs a specific fbank feature extraction bug; no shared code-path." + "reason": "SwitchTransformers sparse-layer initialization is unrelated to the image-classification local dataset loading issue." }, { - "left": "issue:43335", - "right": "issue:44190", + "left": "issue:43650", + "right": "issue:44220", "accept": false, - "reason": "SwitchTransformers sparse-layer config bug vs local dataset loading failure in an example script; unrelated." + "reason": "The titles are too generic to link; one is an unspecified data request, the other is a specific fbank feature extraction bug." }, { "left": "issue:45276", "right": "issue:45584", "accept": false, - "reason": "Gemma4 embedding resizing bug vs Whisper empty-transcription generation failure; different models and mechanics." + "reason": "Different models and subsystems: Gemma4 embedding resize behavior vs Whisper generation after align_special_tokens." }, { "left": "issue:44496", "right": "issue:45803", "accept": false, - "reason": "Config/model-type loading error vs a bare-except bug in FuyuBatchFeature tensor conversion; unrelated failures." + "reason": "Unrecognized model/config loading is unrelated to the bare-except bug in FuyuBatchFeature tensor conversion." }, { "left": "issue:39692", "right": "issue:44190", "accept": false, - "reason": "SigLIP2 doc example errors vs example-script local dataset loading; only a vague documentation/example overlap." + "reason": "A documentation-example bug for SigLIP2 is not the same as a dataset-loading problem in an example script." }, { "left": "issue:45245", "right": "issue:45584", "accept": false, - "reason": "Category cardinality limit error vs Whisper generation on empty transcription; no shared underlying bug." + "reason": "Category-count runtime error and Whisper empty-transcription failure are unrelated." }, { "left": "issue:43295", "right": "issue:45276", "accept": false, - "reason": "Processor/tokenizer regression in v4.57.5 vs Gemma4 token embedding resizing; different APIs and failure modes." + "reason": "Both are regressions, but they affect different APIs and models: processor/tokenizer handling vs Gemma4 token embedding resizing." }, { "left": "issue:41950", "right": "issue:44016", "accept": false, - "reason": "Video-classification pipeline processor lookup bug vs notebook syntax error; completely different issues." + "reason": "Video-classification pipeline processor lookup is unrelated to a notebook syntax error." }, { "left": "issue:43479", "right": "issue:45375", "accept": false, - "reason": "Both involve multimodal config fields, but one is default-initialization with None and the other is a missing deepstack field being dropped by strict serialization." + "reason": "Both are config-field issues, but they involve different models and different missing/defaulted fields." }, { "left": "issue:36246", "right": "issue:41628", "accept": false, - "reason": "Both are import errors for processor-related symbols, but they concern different names, models, and likely different export points." - }, - { - "left": "issue:44186", - "right": "issue:44734", - "accept": false, - "reason": "LayoutLMv2Tokenizer NER/padding crash vs serve KV-cache continuation indexing bug; unrelated tokenization vs serving code." + "reason": "Both are import errors, but they are different missing symbols in different modules and do not indicate the same bug." }, { "left": "issue:36246", "right": "issue:38617", "accept": false, - "reason": "Two different missing imports from different modules; same symptom class only, not the same bug." + "reason": "Different missing imports in different modules; no shared code-path evidence." }, { - "left": "issue:38175", + "left": "issue:44186", + "right": "issue:44734", + "accept": false, + "reason": "Tokenizer crashes on NER padding/truncation are unrelated to serve KV-cache tensor indexing." + }, + { + "left": "issue:38175", "right": "issue:43065", "accept": false, - "reason": "Unexpected zero probabilities in SigLIP2 vs a dummy Conv2d in Sam3PixelDecoder; no concrete overlap." + "reason": "A zero-probability model-output issue is unrelated to a dummy Conv2d implementation in Sam3PixelDecoder." }, { "left": "issue:42175", "right": "issue:44279", "accept": false, - "reason": "Package/backend dependency complaint vs generic dependency issue; too broad and not the same failure." + "reason": "A packaging/backend-dependency complaint is too broad to be the same as a generic dependency issue." }, { "left": "issue:44683", "right": "issue:45446", "accept": false, - "reason": "Both touch flex_attention/PyTorch versioning, but one is compiled flex_attention failing on torch>=2.9 and the other is an import version check bug; not the same code-path problem." + "reason": "Both involve flex_attention, but one is a compiled-kernel failure on torch >=2.9 while the other is a version-check/import-gate bug; not the same concrete fix." }, { "left": "issue:44016", "right": "issue:44220", "accept": false, - "reason": "Notebook syntax error vs audio feature extraction bug; unrelated." + "reason": "Notebook syntax error and fbank feature extraction bug are unrelated." }, { "left": "issue:44466", "right": "issue:44913", "accept": false, - "reason": "Serialization of tied lm_head weights vs GPTNeoX rotary_pct reload persistence; different config/state issues." + "reason": "Serialization/tied-weights behavior is a different issue from GPTNeoXConfig rotary_pct not persisting on reload." }, { "left": "issue:43404", "right": "issue:45216", "accept": false, - "reason": "Both involve model weight handling, but one is tied lm_head weights in Mistral3 and the other is a Qwen3.5 save_pretrained regression; not mergeable as one fix." + "reason": "Different models and symptoms: untied lm_head weights in Mistral3 vs incorrect save_pretrained checkpointing for Qwen3.5." }, { "left": "issue:44246", "right": "issue:45003", "accept": false, - "reason": "Import slowdown vs unsafe sys.modules access; related only at a very high level." + "reason": "Slow import and unsafe sys.modules access may be related in spirit, but there is no evidence they are the same bug or fix." }, { "left": "issue:44734", "right": "issue:45632", "accept": false, - "reason": "Serve tensor indexing bug vs local model cache-path collision with trust_remote_code; different layers of the stack." + "reason": "Serve KV-cache indexing and trust_remote_code cache-path collisions are unrelated." }, { "left": "issue:44561", "right": "issue:45632", "accept": false, - "reason": "Removal of is_torch_fx_available breaking trust_remote_code models vs cache-path collision for local models; unrelated regressions." + "reason": "Both touch trust_remote_code, but one is a removed-helper regression and the other is a local cache-path collision; different code paths." }, { "left": "issue:43643", "right": "issue:44496", "accept": false, - "reason": "AutoConfig trust_remote_code field loss vs unrecognized model_type/config.json issue; both config-loading failures, but not the same bug." + "reason": "Missing fields in AutoConfig with trust_remote_code is not the same as an unrecognized model due to missing model_type." }, { "left": "issue:42490", "right": "issue:45632", "accept": false, - "reason": "Trailing-slash import warning vs leaf-directory cache collision; both path-related, but distinct causes and fixes." + "reason": "Path-with-trailing-slash warning and cache-path collision for same leaf names are different path-handling bugs." }, { "left": "issue:44964", "right": "issue:45600", "accept": false, - "reason": "Phi-4 multimodal loading regression vs stale Sam3LiteText auto_mappings references; different targets." + "reason": "A Phi-4 multimodal load failure is unrelated to CI breakage from removed Sam3LiteText references." }, { "left": "issue:44112", "right": "issue:44964", "accept": false, - "reason": "CI test failure in GraniteSpeech vs model-loading failure for Phi-4 multimodal; no shared code-path." + "reason": "A GraniteSpeech CI test failure is unrelated to a Phi-4 multimodal model-load failure." }, { "left": "issue:43854", "right": "issue:44964", "accept": false, - "reason": "GLM-4.7-Flash load issue vs Phi-4 multimodal load issue; same broad area, but different model-specific problems." + "reason": "Both are model-load failures, but they involve different models and likely different root causes; not mergeable as one PR/bug." }, { "left": "issue:42907", "right": "issue:43441", "accept": false, - "reason": "Dequantized save failure vs Ministral-3 FlashAttention runtime failure; unrelated model operations." + "reason": "Saving dequantized Ministral/Devstral models and FlashAttention failures in Ministral-3 are separate problems." }, { "left": "issue:43976", "right": "issue:44987", "accept": false, - "reason": "Python version compatibility issue vs loading a specific model on transformers>=5.1.0; different failure modes." - }, - { - "left": "issue:43295", - "right": "issue:45335", - "accept": false, - "reason": "Both mention embedding/tokenizer-related behavior, but one is a processor/tokenizer regression and the other is Gemma4 decoder embedding resize propagation; not the same bug." + "reason": "A Python version compatibility issue is unrelated to a specific model-loading failure on transformers>=5.1.0." } ] }, @@ -10052,6 +10001,7 @@ "issue:45823" ], "soft_pairs": [ + "issue:43295|issue:45335", "issue:44016|issue:44062", "issue:43673|issue:44336", "issue:43531|issue:45600", @@ -10067,8 +10017,8 @@ "issue:43299|issue:45357", "issue:44496|issue:45759", "issue:44521|issue:45325", - "issue:44242|issue:45200", "issue:44351|issue:45278", + "issue:44242|issue:45200", "issue:43976|issue:45405", "issue:41950|issue:45020", "issue:43479|issue:43525", @@ -10083,214 +10033,220 @@ "issue:43295|issue:44779", "issue:41950|issue:43650" ], - "prepared_review_unit_hash": "0cc9d57d95913c6f3bbafb632bafadaca5ab15b60e5c2b6c9edfd0c5ce3a3c5f", + "prepared_review_unit_hash": "dce9509f3998a3a96d1c0b22e2b0d59a33a62493286600a929a56cbf163ee1bc", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 30, - "serialized_chars": 36020, - "estimated_input_tokens": 9005, - "estimated_eval_tokens": 18266 + "soft_pair_count": 31, + "serialized_chars": 36318, + "estimated_input_tokens": 9080, + "estimated_eval_tokens": 18416 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is heterogeneous: issues span unrelated model-loading, tokenizer, config, pipeline, docs, and environment regressions across many different models/subsystems. No soft duplicate edges look safe to merge.", - "confidence": 0.96, - "canonical_issue_reason": "If forced to pick a representative issue, 44521 is the most central/active, but there is no true canonical because the cluster is not a single bug family.", + "summary": "This cluster is a loose mix of unrelated v5/v5.x regressions, model-loading/config bugs, tokenizer/processor issues, and a few docs/import problems. The titles are superficially similar, but the underlying code paths differ, so these should stay separate.", + "confidence": 0.93, + "canonical_issue_reason": "issue:45020 is the broadest and most representative regression report in the set ('recent transformers versions break models using remote_code'), so it is the best single issue to anchor the cluster even though it is not a true duplicate of the others.", "canonical_pr_reason": null, - "best_issue_reason": "44521 is the best overall issue to represent the cluster only by activity and breadth; however, the set is too mixed for a meaningful duplicate canonical.", + "best_issue_reason": "issue:45020 best captures the dominant theme of recent-version breakage and is the most reusable summary issue for this cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43295", + "right": "issue:45335", + "accept": false, + "reason": "Both are regressions, but one is about processor.tokenizer/image inputs and the other is about resize_token_embeddings not updating decoder.embed_tokens in t5gemma; different code paths." + }, { "left": "issue:44016", "right": "issue:44062", "accept": false, - "reason": "Notebook syntax error vs AddedToken keyword conflict; unrelated failures." + "reason": "A notebook syntax error and a tokenizers.AddedToken keyword-argument TypeError are unrelated bugs." }, { "left": "issue:43673", "right": "issue:44336", "accept": false, - "reason": "Generation cache bug vs ANSI coloring in loading_report; different code paths." + "reason": "Chunked-prefill cache handling and ANSI color output in loading_report are separate subsystems with different failure modes." }, { "left": "issue:43531", "right": "issue:45600", "accept": false, - "reason": "Qwen3-MoE sliding-window bug vs auto_mappings cleanup/CI breakage; not the same bug." + "reason": "Qwen3-MoE sliding_window behavior and auto_mappings referencing removed configs are not the same bug." }, { "left": "issue:43526", "right": "issue:43901", "accept": false, - "reason": "BEiT image processor label reduction bug vs docs mention return_all_scores; unrelated." + "reason": "BeitImageProcessorFast reduce_labels is a processor bug; return_all_scores is a documentation/behavior mismatch." }, { "left": "issue:44496", "right": "issue:45042", "accept": false, - "reason": "Model config loading error vs PIL/torchvision dependency issue; different problems." + "reason": "Unrecognized model/config loading and PIL image processors wrongly requiring torchvision are distinct issues." }, { "left": "issue:43824", "right": "issue:44933", "accept": false, - "reason": "Different import symbols/modules and likely different root causes." + "reason": "These are both import errors, but they involve different missing symbols in different modules." }, { "left": "issue:43299", "right": "issue:45216", "accept": false, - "reason": "Qwen3VL loading regression vs Qwen3.5 save_pretrained regression; not the same code path." + "reason": "One is a model-loading regression for Qwen3VL MoE; the other is a save_pretrained checkpoint regression for Qwen3.5." }, { "left": "issue:44075", "right": "issue:44368", "accept": false, - "reason": "SGD optimizer args ignored vs tie_word_embeddings warning; unrelated." + "reason": "SGD args being ignored and a tie_word_embeddings warning during LoRA fine-tuning are unrelated." }, { "left": "issue:45335", "right": "issue:45584", "accept": false, - "reason": "Embedding resize bug vs Whisper empty transcription failure after align_special_tokens; different issues." + "reason": "A t5gemma embedding resize bug and Whisper empty-transcription failure after align_special_tokens are different code paths." }, { "left": "issue:43976", "right": "issue:44279", "accept": false, - "reason": "Python version compatibility vs vague dependency issue; too different and underspecified." + "reason": "Python-version support problems and a generic dependency issue are not the same underlying defect." }, { "left": "issue:43381", "right": "issue:45593", "accept": false, - "reason": "Eval/gradient-checkpointing incompatibility vs auxiliary-loss omission in D-FINE; unrelated." + "reason": "Gradient checkpointing in eval mode and D-FINE aux-loss behavior when denoising is off are unrelated training behaviors." }, { "left": "issue:43901", "right": "issue:44496", "accept": false, - "reason": "Documentation mismatch vs model config loading failure; no shared bug." + "reason": "Docs wording around return_all_scores does not match the OLMo-Hybrid config/model_type loading error." }, { "left": "issue:43299", "right": "issue:45357", "accept": false, - "reason": "Loading regression vs save_pretrained visual-key regression; different failure modes." + "reason": "Both involve Qwen3.5/VL, but loading failure and incorrect visual-encoder keys in save_pretrained are different operations and fixes." }, { "left": "issue:44496", "right": "issue:45759", "accept": false, - "reason": "Unrecognized model config vs AutoModelForCausalLM text_config unwrapping; different model init paths." + "reason": "A generic unrecognized-model error and from_config not unwrapping text_config for composite multimodal configs are not the same bug." }, { "left": "issue:44521", "right": "issue:45325", "accept": false, - "reason": "All-zero assistant masks vs still-image RoPE position scaling; both multimodal but different mechanisms." + "reason": "All-zero assistant masks in apply_chat_template and Qwen2.5-VL rope position scaling are unrelated multimodal issues." }, { - "left": "issue:44242", - "right": "issue:45200", + "left": "issue:44351", + "right": "issue:45278", "accept": false, - "reason": "MoE load-balancing loss bug vs Gemma4 mm_token_type_ids defaulting; unrelated training-path issues." + "reason": "A missing HybridCache import is one concrete symbol issue; 'many import errors' is too broad and not the same bug." }, { - "left": "issue:44351", - "right": "issue:45278", + "left": "issue:44242", + "right": "issue:45200", "accept": false, - "reason": "Specific HybridCache import error vs broad import-errors-after-upgrade report; too coarse to treat as the same bug." + "reason": "Load-balancing loss gating and mm_token_type_ids defaulting to zeros are different model/training concerns." }, { "left": "issue:43976", "right": "issue:45405", "accept": false, - "reason": "Python compatibility vs PEFT minimum-version bump; different dependency problems." + "reason": "Python compatibility and an overly high MIN_PEFT_VERSION bump are separate dependency problems." }, { "left": "issue:41950", "right": "issue:45020", "accept": false, - "reason": "Video-classification processor lookup bug vs remote_code breakage across versions; unrelated." + "reason": "A video-classification pipeline selecting image processors is not the same as remote_code models breaking in recent transformers versions." }, { "left": "issue:43479", "right": "issue:43525", "accept": false, - "reason": "Multimodal config default initialization bug vs missing pad_token_id on Llama4Config; unrelated." + "reason": "Phi4MultimodalConfig defaults for None and Llama4Config missing pad_token_id are different config initialization bugs." }, { "left": "issue:44779", "right": "issue:45812", "accept": false, - "reason": "DeepSeek tokenizer regression vs Granite tokenizer regression; same broad area, but different model-specific code paths." + "reason": "Tokenizer regressions for Deepseek and Granite are both tokenization problems, but they affect different model families and likely different code paths." }, { "left": "issue:38175", "right": "issue:44617", "accept": false, - "reason": "SigLIP2 probability issue vs Sam3Video CUDA OOM; unrelated." + "reason": "SigLIP2 zero probabilities and Sam3Video CUDA OOM are completely unrelated." }, { "left": "issue:41950", "right": "issue:45823", "accept": false, - "reason": "Video pipeline processor selection bug vs Gemma4 device-map mismatch; unrelated." + "reason": "A pipeline using image processors and a Gemma4 PLE device-map mismatch are unrelated issues." }, { "left": "issue:43577", "right": "issue:45216", "accept": false, - "reason": "BLIP2 dtype propagation bug vs Qwen3.5 checkpoint-saving regression; unrelated." + "reason": "BLIP2 dtype propagation and Qwen3.5 save_pretrained checkpoint correctness are distinct regressions." }, { "left": "issue:44683", "right": "issue:45542", "accept": false, - "reason": "Torch flex_attention compile failure vs tensorboard/TensorFlow backend issue; unrelated." + "reason": "Compiled flex_attention on torch>=2.9 and tensorboard/TensorFlow backend detection are unrelated." }, { "left": "issue:39692", "right": "issue:43901", "accept": false, - "reason": "SigLIP2 docs example/model mismatch vs text-classification docs stale option; both docs-related but different topics." + "reason": "SigLIP2 docs example errors and TextClassificationPipeline docs on return_all_scores are different documentation issues with different models and failure modes." }, { "left": "issue:44279", "right": "issue:45405", "accept": false, - "reason": "Generic transformers dependency issue vs specific PEFT version pin problem; different package constraints." + "reason": "A generic dependency issue is not the same as bumping MIN_PEFT_VERSION beyond what is released." }, { "left": "issue:43299", "right": "issue:43854", "accept": false, - "reason": "Qwen3VL loading regression vs GLM-4.7-Flash unit-test load failure; different model families and code paths." + "reason": "Qwen3VL loading regressions and GLM-4.7-Flash unit-test loading failures involve different models and likely different fixes." }, { "left": "issue:43295", "right": "issue:44779", "accept": false, - "reason": "Processor/tokenizer API regression vs DeepSeek tokenizer output regression; both tokenizer-related but not the same bug." + "reason": "Processor.tokenizer/image-passing regression and Deepseek tokenizer output regression are different API surfaces." }, { "left": "issue:41950", "right": "issue:43650", "accept": false, - "reason": "Video-classification processor bug vs low-signal issue title with no clear technical overlap." + "reason": "The video-classification processor bug and the placeholder 'ADD THE DATA' issue are not related." } ] }, @@ -10365,18 +10321,18 @@ ], "soft_pairs": [ "issue:44462|issue:45081", - "issue:45081|issue:45907", "issue:44297|issue:44462", + "issue:45081|issue:45907", "issue:44016|issue:44291", "issue:33290|issue:45656", - "issue:43441|issue:45834", "issue:43450|issue:44279", + "issue:43441|issue:45834", "issue:45600|issue:45712", "issue:43065|issue:44898", "issue:44466|issue:45276", "issue:42994|issue:43317", - "issue:44261|issue:44263", "issue:43650|issue:45290", + "issue:44261|issue:44263", "issue:44623|issue:45397", "issue:44623|issue:44749", "issue:43441|issue:44683", @@ -10384,211 +10340,218 @@ "issue:43262|issue:44521", "issue:44016|issue:45823", "issue:43723|issue:44162", - "issue:44671|issue:44779", "issue:43673|issue:43761", - "issue:42175|issue:43976", "issue:44361|issue:45593", + "issue:44671|issue:44779", + "issue:42175|issue:43976", "issue:43901|issue:43994", "issue:43784|issue:45542", + "issue:43122|issue:43650", "issue:44987|issue:45542", - "issue:43122|issue:43650" + "issue:44297|issue:45397" ], - "prepared_review_unit_hash": "57062d95c9f0543a65dab0719a0f41df8e8b10a66a78b4d4b1cdeb3f43f5fe52", + "prepared_review_unit_hash": "903f9e8e03e383b199a2a8bd6fe11ae00576868b322440ed4ae92c0e0af8072d", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 28, - "serialized_chars": 36058, - "estimated_input_tokens": 9015, - "estimated_eval_tokens": 18286 + "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 of unrelated Transformers issues sharing broad language around tokenizers, loading, processors, and model regressions. I found no soft-edge pair that looks like the same underlying bug/change, so none should be merged as duplicates.", - "confidence": 0.96, - "canonical_issue_reason": "No strong canonical duplicate emerges because the items are heterogeneous. If one representative must be chosen, issue #44521 is the most active and broadly described, but it is not a true umbrella for the rest.", + "summary": "The cluster is mostly heterogeneous; the soft pairs are broad topical similarities, not the same underlying bug/change. I found no true duplicate PRs here, and none of the issue pairs are strong enough to merge as duplicates.", + "confidence": 0.94, + "canonical_issue_reason": "No single canonical issue stands out because the set spans many unrelated bugs/regressions; if one representative issue is needed, issue 44462 is the clearest self-contained user-facing bug report among the tokenizer-related items.", "canonical_pr_reason": null, - "best_issue_reason": "Issue #44521 is the best single representative only by discussion/references, not by duplicate semantics. Overall the cluster is too diverse for a meaningful canonical issue.", + "best_issue_reason": "Issue 44462 is the strongest representative bug report overall: it is concrete, reproducible, and has a clear loading-path regression description. But it is not a canonical for the whole set.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44462", "right": "issue:45081", "accept": false, - "reason": "Both involve tokenizers, but one is AutoTokenizer ignoring tokenizer.json while the other is a Mistral regex/backend_tokenizer crash; different code paths." + "reason": "Tokenizer repository loading vs Mistral regex patch crash; related area, but different code paths and failures." }, { - "left": "issue:45081", - "right": "issue:45907", + "left": "issue:44297", + "right": "issue:44462", "accept": false, - "reason": "Mistral tokenizer patch crash and list_repo_files revision forwarding are unrelated subsystems and failure modes." + "reason": "save_pretrained tokenizer_class mismatch vs AutoTokenizer ignoring tokenizer.json; same subsystem, but distinct save vs load bugs." }, { - "left": "issue:44297", - "right": "issue:44462", + "left": "issue:45081", + "right": "issue:45907", "accept": false, - "reason": "Tokenizer class mismatch on save_pretrained is a save/load serialization bug, not the same as AutoTokenizer ignoring repository tokenizer.json." + "reason": "Regex patch crash vs list_repo_files revision forwarding; unrelated functionality." }, { "left": "issue:44016", "right": "issue:44291", "accept": false, - "reason": "Notebook syntax error and init_empty_weights unexpected argument are completely different problems." + "reason": "Notebook syntax error vs init_empty_weights argument mismatch; unrelated issues." }, { "left": "issue:33290", "right": "issue:45656", "accept": false, - "reason": "Both mention DeepSpeed, but one is Adafactor OOM and the other is optimizer.step being called twice; not the same underlying bug." + "reason": "Adafactor OOM vs deepspeed optimizer.step called twice; both deepspeed-related, but different concrete bugs." }, { - "left": "issue:43441", - "right": "issue:45834", + "left": "issue:43450", + "right": "issue:44279", "accept": false, - "reason": "Ministral FlashAttention failure and Kosmos2.5 OCR index error are unrelated model-specific bugs." + "reason": "Video processor batch shape bug vs dependency/install issue; not the same bug." }, { - "left": "issue:43450", - "right": "issue:44279", + "left": "issue:43441", + "right": "issue:45834", "accept": false, - "reason": "Video processor batched-shape bug and a generic dependency issue do not describe the same defect." + "reason": "FlashAttention regression for Ministral vs Kosmos2.5 OCR index error; different models and failure modes." }, { "left": "issue:45600", "right": "issue:45712", "accept": false, - "reason": "One is stale auto_mappings references to removed configs; the other is leftover dummy classes leaking into dir(transformers). Different cleanup issues." + "reason": "Stale auto_mappings references vs leftover dummy classes; both maintenance problems, but not the same change." }, { "left": "issue:43065", "right": "issue:44898", "accept": false, - "reason": "Dummy Conv2d in Sam3PixelDecoder and Perceiver non-default-resolution failure are unrelated vision-model bugs." + "reason": "Sam3PixelDecoder dummy Conv2d vs Perceiver resolution/interpolation failure; different model paths." }, { "left": "issue:44466", "right": "issue:45276", "accept": false, - "reason": "Inconsistent lm_head.weight serialization and Gemma4 resize_token_embeddings propagation are both weight-related, but they affect different mechanisms and paths." + "reason": "lm_head.weight serialization inconsistency vs Gemma4 resize_token_embeddings propagation; same model family, but different bugs." }, { "left": "issue:42994", "right": "issue:43317", "accept": false, - "reason": "Quantized model saving failure and dequantized model loading with gpu/cpu offload are distinct quantization lifecycle bugs." + "reason": "Quantized save failure vs dequantized load/offload failure; related quantization area, but opposite operations and different symptoms." }, { - "left": "issue:44261", - "right": "issue:44263", + "left": "issue:43650", + "right": "issue:45290", "accept": false, - "reason": "MLA q_a_layernorm precision error and GlmMoeDsaIndexer torch.split return values are unrelated implementation bugs." + "reason": "Placeholder/data issue vs apply_chat_template tool-call crash; unrelated." }, { - "left": "issue:43650", - "right": "issue:45290", + "left": "issue:44261", + "right": "issue:44263", "accept": false, - "reason": "The placeholder-style 'ADD THE DATA' issue is not the same as an apply_chat_template crash on tool-call assistant messages." + "reason": "RMS norm epsilon precision mismatch vs torch.split return-value handling; unrelated internals." }, { "left": "issue:44623", "right": "issue:45397", "accept": false, - "reason": "processor.save_pretrained missing files and gemma-4 zero3 from_pretrained are different save-vs-load issues." + "reason": "processor.save_pretrained missing files vs gemma-4 zero3 from_pretrained; different save/load problems." }, { "left": "issue:44623", "right": "issue:44749", "accept": false, - "reason": "Missing processor save artifacts and a performance regression after upgrade are unrelated." + "reason": "Missing files on processor save vs performance slowdown after version upgrade; unrelated." }, { "left": "issue:43441", "right": "issue:44683", "accept": false, - "reason": "Both mention attention, but one is a model-specific FlashAttention incompatibility and the other is a compiled flex_attention torch-version failure." + "reason": "FlashAttention failure for one model vs compiled flex_attention failing on torch>=2.9; attention-related but different backends and bugs." }, { "left": "issue:43653", "right": "issue:45593", "accept": false, - "reason": "BigBirdTokenizer special-token decoding bug and D-FINE auxiliary-loss behavior are unrelated." + "reason": "BigBirdTokenizer special-token registration bug vs D-FINE auxiliary-loss omission; unrelated." }, { "left": "issue:43262", "right": "issue:44521", "accept": false, - "reason": "Both involve apply_chat_template, but one is default audio sampling rate and the other is all-zero assistant masks for multimodal inputs; different bugs." + "reason": "apply_chat_template sample-rate default vs zero assistant_masks for multimodal inputs; same API surface, but different behaviors." }, { "left": "issue:44016", "right": "issue:45823", "accept": false, - "reason": "Notebook syntax error and Gemma4 PLE device mismatch are unrelated." + "reason": "Notebook syntax error vs Gemma4 device mismatch in forward; unrelated." }, { "left": "issue:43723", "right": "issue:44162", "accept": false, - "reason": "AutoTokenizer.from_pretrained loading issue and ESM2 model breakage are not the same bug." + "reason": "AutoTokenizer loading regression vs ESM2 breakage; unrelated models and failure modes." }, { - "left": "issue:44671", - "right": "issue:44779", + "left": "issue:43673", + "right": "issue:43761", "accept": false, - "reason": "CamemBERT masked-LM regression and Deepseek tokenizer regression are separate model/tokenizer failures." + "reason": "Missing GenerationMixin cache in chunked prefill vs CLIPVisionModel hidden_states bug; unrelated." }, { - "left": "issue:43673", - "right": "issue:43761", + "left": "issue:44361", + "right": "issue:45593", "accept": false, - "reason": "GenerationMixin cache missing during chunked_prefill and CLIPVisionModel hidden_states=None are different API regressions." + "reason": "MLukeTokenizer AttributeError vs D-FINE auxiliary-loss issue; unrelated." }, { - "left": "issue:42175", - "right": "issue:43976", + "left": "issue:44671", + "right": "issue:44779", "accept": false, - "reason": "TensorFlow extra installation packaging and Python-version compatibility are different dependency/environment issues." + "reason": "CamemBERT masked-LM regression vs Deepseek tokenizer regression; both output-quality bugs, but different models and code paths." }, { - "left": "issue:44361", - "right": "issue:45593", + "left": "issue:42175", + "right": "issue:43976", "accept": false, - "reason": "MLukeTokenizer AttributeError and D-FINE auxiliary-loss behavior are unrelated." + "reason": "Missing TensorFlow backend in torch install vs Python version compatibility regression; environment-related but distinct." }, { "left": "issue:43901", "right": "issue:43994", "accept": false, - "reason": "Documentation drift for return_all_scores and SigLIP2 nonsensical outputs are not the same underlying problem." + "reason": "Docs/behavior mismatch in TextClassificationPipeline vs Siglip2 inference regression; unrelated." }, { "left": "issue:43784", "right": "issue:45542", "accept": false, - "reason": "Sentence-transformers import NameError and missing TensorFlow backend behavior are different missing-dependency problems." + "reason": "Import NameError for sentence-transformers vs tensorboard-only install without TensorFlow; unrelated packaging issues." + }, + { + "left": "issue:43122", + "right": "issue:43650", + "accept": false, + "reason": "Tokenizer behavior change across versions vs placeholder 'ADD THE DATA' report; not a duplicate." }, { "left": "issue:44987", "right": "issue:45542", "accept": false, - "reason": "Loading physical-intelligence/fast on transformers>=5.1.0 and tensorboard/TensorFlow packaging are unrelated." + "reason": "Model loading failure on transformers>=5.1.0 vs tensorboard/TensorFlow install issue; unrelated." }, { - "left": "issue:43122", - "right": "issue:43650", + "left": "issue:44297", + "right": "issue:45397", "accept": false, - "reason": "Tokenizer version drift and the malformed placeholder issue have no substantive overlap." + "reason": "Tokenizer save_pretrained metadata mismatch vs gemma-4 zero3 from_pretrained; different bugs in different subsystems." } ] }, @@ -10636,7 +10599,6 @@ "issue:44242", "issue:44246", "issue:44295", - "issue:44297", "issue:44336", "issue:44351", "issue:44373", @@ -10659,11 +10621,11 @@ "issue:45397", "issue:45561", "issue:45563", + "issue:45584", "issue:45593" ], "soft_pairs": [ "issue:43824|issue:43976", - "issue:44297|issue:45397", "issue:44188|issue:44336", "issue:43450|issue:44617", "issue:43525|issue:43881", @@ -10672,8 +10634,8 @@ "issue:41628|issue:43976", "issue:44589|issue:45362", "issue:45561|issue:45563", - "issue:42491|issue:43299", "issue:44336|issue:44373", + "issue:42491|issue:43299", "issue:44242|issue:45593", "issue:44295|issue:45397", "issue:41950|issue:44220", @@ -10682,8 +10644,8 @@ "issue:43526|issue:45003", "issue:36246|issue:42757", "issue:43352|issue:45250", - "issue:36246|issue:44351", "issue:44336|issue:44485", + "issue:36246|issue:44351", "issue:45003|issue:45561", "issue:43673|issue:44561", "issue:44246|issue:44871", @@ -10692,216 +10654,230 @@ "issue:44162|issue:45200", "issue:36246|issue:43596", "issue:44779|issue:45292", - "issue:44877|issue:45375" + "issue:44877|issue:45375", + "issue:45292|issue:45584", + "issue:43901|issue:45003" ], - "prepared_review_unit_hash": "6d0a312b87210891297844274858a7d039e598e9d56b4d0255e9e9f2f823f7d6", + "prepared_review_unit_hash": "a516a98637b51fbf71e3fb3fca77160c853015f748ecf372e6615ede0450bcc6", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 31, - "serialized_chars": 36650, - "estimated_input_tokens": 9163, - "estimated_eval_tokens": 18582 + "soft_pair_count": 32, + "serialized_chars": 36942, + "estimated_input_tokens": 9236, + "estimated_eval_tokens": 18728 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "All candidate pairs look like distinct bugs or regressions rather than duplicates. The set is broadly related to Transformers regressions, but the concrete failures span different symbols, models, configs, and subsystems, so I rejected every soft edge.", - "confidence": 0.92, - "canonical_issue_reason": "issue:43824 is the clearest concrete user-facing regression here: a specific missing model export causing an ImportError, with strong discussion and inbound-reference signal.", + "summary": "This set is mostly a mix of unrelated Transformers bugs. None of the soft pairs are strong enough to merge as duplicates; at most a few share broad themes like imports, config strictness, or model loading, but they target different concrete failures.", + "confidence": 0.84, + "canonical_issue_reason": "issue:43824 is the clearest, most specific regression report in the set: a concrete import error for Qwen2.5-VL with multiple inbound references, making it the best canonical anchor among these issues.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43824 is the strongest issue to anchor this cluster because it is specific, well-scoped, and has the most evidence of impact and reuse.", + "best_issue_reason": "issue:43824 is also the best overall issue candidate because it is narrowly scoped, user-facing, and well-documented with strong evidence of recurrence.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43824", "right": "issue:43976", "accept": false, - "reason": "Different failures: a missing model export vs a Python-version compatibility problem." - }, - { - "left": "issue:44297", - "right": "issue:45397", - "accept": false, - "reason": "Tokenizer save_pretrained metadata mismatch is unrelated to gemma-4 zero3 loading." + "reason": "Both mention import-related breakage, but one is a missing Qwen2.5-VL symbol and the other is a Python version compatibility problem; different failure modes and code paths." }, { "left": "issue:44188", "right": "issue:44336", "accept": false, - "reason": "Attention-kernel divergence on torch.compile is unrelated to ANSI codes in loading_report." + "reason": "Attention-kernel divergence under torch.compile is unrelated to ANSI escape codes in loading_report." }, { "left": "issue:43450", "right": "issue:44617", "accept": false, - "reason": "Batched video processor shape bug and CUDA OOM are different problems." + "reason": "Batched video-processor shape bugs and Sam3Video CUDA OOM are different problems." }, { "left": "issue:43525", "right": "issue:43881", "accept": false, - "reason": "Missing pad_token_id on Llama4Config is not the same as GLM-4V loading failure." + "reason": "A missing pad_token_id on Llama4Config is not the same as a glm-4v model loading failure." }, { "left": "issue:43525", "right": "issue:43643", "accept": false, - "reason": "Both are config-related, but one is a missing attribute and the other is trust_remote_code dropping fields." + "reason": "One is a config attribute error in Llama4, the other is missing fields when loading trust_remote_code configs; too broad to treat as the same bug." }, { "left": "issue:43901", "right": "issue:44246", "accept": false, - "reason": "Documentation mismatch for return_all_scores does not match a slow import regression." + "reason": "Docs/behavior mismatch for text classification pipelines is unrelated to sporadic import latency." }, { "left": "issue:41628", "right": "issue:43976", "accept": false, - "reason": "AutoImageProcessor import failure is unrelated to the Python 3.9/3.10 support issue." + "reason": "AutoImageProcessor import failure and Python version support issues are distinct regressions." }, { "left": "issue:44589", "right": "issue:45362", "accept": false, - "reason": "Float8 storage lookup failure and Qwen3.5 chat crash are different code paths." + "reason": "A Float8 storage type error is unrelated to Qwen3.5 chat crashes." }, { "left": "issue:45561", "right": "issue:45563", "accept": false, - "reason": "pytest-xdist file race is unrelated to a stale num_return_sequences warning." + "reason": "A pytest-xdist capture-file race and a stale generation warning are unrelated." }, { - "left": "issue:42491", - "right": "issue:43299", + "left": "issue:44336", + "right": "issue:44373", "accept": false, - "reason": "Both mention Qwen MoE, but one is a LoRA compatibility issue and the other is a model-loading regression." + "reason": "ANSI code emission in loading_report is unrelated to a wrong docstring for position_ids." }, { - "left": "issue:44336", - "right": "issue:44373", + "left": "issue:42491", + "right": "issue:43299", "accept": false, - "reason": "ANSI output in a report utility is unrelated to a docstring bug." + "reason": "Both mention Qwen3 MoE, but one is a LoRA compatibility issue across versions and the other is a model-loading regression for Qwen3VL MoE; not the same concrete bug." }, { "left": "issue:44242", "right": "issue:45593", "accept": false, - "reason": "Missing router loss and D-FINE auxiliary-loss behavior are different model-specific bugs." + "reason": "Missing load-balancing loss and missing auxiliary losses are different model-specific training bugs." }, { "left": "issue:44295", "right": "issue:45397", "accept": false, - "reason": "Position_ids buffer access error is unrelated to gemma-4 zero3 from_pretrained loading." + "reason": "A position_ids buffer access error and a zero3 from_pretrained issue are different failures." }, { "left": "issue:41950", "right": "issue:44220", "accept": false, - "reason": "Video-classification processor selection and fbank feature extraction are different failures." + "reason": "Video-classification pipeline processor selection and _torch_extract_fbank_features are unrelated." }, { "left": "issue:44246", - "right": "issue:44871", + "right": "issue:44496", "accept": false, - "reason": "Slow import timing is unrelated to eos_token_id configuration inconsistency." + "reason": "Import slowness and an unrecognized model/config loading error do not share the same bug." }, { "left": "issue:43531", "right": "issue:45362", "accept": false, - "reason": "Qwen3-MoE sliding_window handling and Qwen3.5 chat crashes are not the same bug." + "reason": "A sliding_window bug in Qwen3-MoE is unrelated to Qwen3.5 chat crashes." }, { "left": "issue:43526", "right": "issue:45003", "accept": false, - "reason": "BEiT label reduction logic and sys.modules access are unrelated." + "reason": "A BeitImageProcessorFast label-reduction bug and unsafe sys.modules access are different code paths." }, { "left": "issue:36246", "right": "issue:42757", "accept": false, - "reason": "Both are ImportErrors, but they affect different packages and symbols." + "reason": "Missing Qwen2.5-VL exports and missing is_offline_mode from huggingface_hub are separate import regressions." }, { "left": "issue:43352", "right": "issue:45250", "accept": false, - "reason": "Model-specific Flash Attention 2 unsupported error is not the same as a generic flash-attention issue." + "reason": "A concrete Nemotron Flash Attention 2 limitation is not the same as a generic Flash Attention 2 topic issue." }, { - "left": "issue:36246", - "right": "issue:44351", + "left": "issue:44336", + "right": "issue:44485", "accept": false, - "reason": "Missing Qwen2_5_VL export and missing HybridCache export are distinct import regressions." + "reason": "Terminal color handling and GLM-5 RoPE implementation are unrelated." }, { - "left": "issue:44336", - "right": "issue:44485", + "left": "issue:36246", + "right": "issue:44351", "accept": false, - "reason": "Loading-report ANSI formatting is unrelated to GLM-5 RoPE implementation." + "reason": "Both are import errors, but they concern different missing symbols and likely different package-export paths." }, { "left": "issue:45003", "right": "issue:45561", "accept": false, - "reason": "Unsafe sys.modules access does not match a pytest-xdist captured-info race." + "reason": "Unsafe sys.modules access and a captured_info.txt race are unrelated." }, { "left": "issue:43673", "right": "issue:44561", "accept": false, - "reason": "Chunked-prefill cache handling and trust_remote_code breakage are different mechanisms." + "reason": "Chunked prefill cache absence and removal of is_torch_fx_available affect different subsystems and failure modes." + }, + { + "left": "issue:44246", + "right": "issue:44871", + "accept": false, + "reason": "Import performance and Gemma eos_token_id inconsistency are unrelated." }, { "left": "issue:44242", "right": "issue:44857", "accept": false, - "reason": "Missing auxiliary loss and AMP CUDA crash are unrelated." + "reason": "Load-balancing-loss omission and AMP/CUDA crash in LwDetrImageLoss are different bugs." }, { "left": "issue:44295", "right": "issue:44749", "accept": false, - "reason": "Position_ids buffer reading bug is unrelated to a tokenizer/filtering slowdown regression." + "reason": "Position_ids buffer access and Chinese-language filtering slowdown are unrelated." }, { "left": "issue:44162", "right": "issue:45200", "accept": false, - "reason": "ESM2 breakage and Gemma 4 mm_token_type_ids behavior are different model bugs." + "reason": "ESM2 breakage and Gemma 4 mm_token_type_ids defaults are different model bugs." }, { "left": "issue:36246", "right": "issue:43596", "accept": false, - "reason": "Qwen2_5_VL image processor import failure is unrelated to Deepspeed/Bert index errors." + "reason": "Qwen2.5-VL import failure and a BertModel deepspeed zero3 index error are unrelated." }, { "left": "issue:44779", "right": "issue:45292", "accept": false, - "reason": "Deepseek tokenizer correctness regression is unrelated to resize_token_embeddings not updating output embeddings." + "reason": "Deepseek tokenizer corruption in v5 and resize_token_embeddings not updating output embeddings are different issues." }, { "left": "issue:44877", "right": "issue:45375", "accept": false, - "reason": "Strict granite_speech config loading and Qwen3_5MoeVisionConfig field dropping are related in theme but not the same concrete bug." + "reason": "Both involve strict config behavior, but they affect different model configs and the reported failures are not the same concrete bug." + }, + { + "left": "issue:45292", + "right": "issue:45584", + "accept": false, + "reason": "Embedding resize behavior and Whisper empty-transcription generation failure are unrelated." + }, + { + "left": "issue:43901", + "right": "issue:45003", + "accept": false, + "reason": "A pipeline doc/behavior mismatch and unsafe sys.modules access are different classes of bugs." } ] }, @@ -10936,7 +10912,6 @@ "issue:43646", "issue:43653", "issue:43673", - "issue:43901", "issue:43931", "issue:44060", "issue:44077", @@ -10960,14 +10935,15 @@ "issue:45137", "issue:45200", "issue:45229", + "issue:45278", "issue:45292", + "issue:45310", "issue:45335", "issue:45356", "issue:45375", "issue:45397", "issue:45459", "issue:45563", - "issue:45584", "issue:45588", "issue:45600", "issue:45656", @@ -10975,22 +10951,20 @@ "issue:45759" ], "soft_pairs": [ - "issue:45292|issue:45584", - "issue:43901|issue:45003", "issue:45292|issue:45701", - "issue:44496|issue:45003", "issue:43232|issue:44336", + "issue:44496|issue:45003", "issue:44987|issue:45459", "issue:43262|issue:44246", - "issue:45375|issue:45759", "issue:42886|issue:45229", + "issue:45375|issue:45759", "issue:43257|issue:45137", "issue:43065|issue:43408", "issue:44995|issue:45200", "issue:43475|issue:43540", "issue:44060|issue:44077", - "issue:44466|issue:45335", "issue:45397|issue:45656", + "issue:44466|issue:45335", "issue:44995|issue:45588", "issue:43653|issue:44162", "issue:44368|issue:44977", @@ -11003,210 +10977,226 @@ "issue:44297|issue:44623", "issue:43232|issue:43673", "issue:44361|issue:44749", - "issue:43646|issue:44162" + "issue:43646|issue:44162", + "issue:43673|issue:45278", + "issue:44623|issue:45656", + "issue:44749|issue:45656", + "issue:44749|issue:45310" ], - "prepared_review_unit_hash": "4bc168c530fc3a498564d51ffcc81c32984a630eb485323d5bd866ee5e2bd4b6", + "prepared_review_unit_hash": "5aceae6e7e93222d0f8fdf2daceba0899b61a599070156a4535ee32c3e3f7edc", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 29, - "serialized_chars": 36457, - "estimated_input_tokens": 9115, - "estimated_eval_tokens": 18486 + "soft_pair_count": 31, + "serialized_chars": 37001, + "estimated_input_tokens": 9251, + "estimated_eval_tokens": 18758 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous; the titles span unrelated bugs in tokenization, generation cache handling, model loading, Deepspeed, SAM, and docs. There are a few superficial overlaps, but no soft pair looks like the same concrete bug/change.", - "confidence": 0.91, - "canonical_issue_reason": "issue:43673 is the closest representative because it is a concrete generation/cache regression and the only item with a nearby cache-related soft match; the rest of the cluster is too broad and unrelated to share one canonical bug.", + "summary": "These items are mostly unrelated regressions across tokenization, generation, multimodal configs, serialization, and deepspeed; none of the soft pairs look like the same underlying bug or change.", + "confidence": 0.86, + "canonical_issue_reason": "issue:45278 is the broadest umbrella-style report here: a version-upgrade regression causing many import errors, unlike the more narrowly scoped model-specific bugs.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43673 is the best issue to anchor the cluster because it is specific, actionable, and sits closest to the small generation/cache subgroup, even though the overall cluster is not a true duplicate set.", + "best_issue_reason": "issue:45278 is the best representative issue because it captures a general cross-cutting failure mode after upgrade, making it the closest thing to a cluster anchor among these mostly specific reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45292", - "right": "issue:45584", - "accept": false, - "reason": "Different bugs: token embedding resize/output embeddings vs Whisper empty transcription generation failure." - }, - { - "left": "issue:43901", - "right": "issue:45003", + "right": "issue:45701", "accept": false, - "reason": "Docs mismatch in text-classification pipeline vs unsafe sys.modules access in modeling_utils." + "reason": "Embedding resize behavior is a different path from tokenizer/output changes; one is about resized embeddings, the other about tokenization differences across versions." }, { - "left": "issue:45292", - "right": "issue:45701", + "left": "issue:43232", + "right": "issue:44336", "accept": false, - "reason": "Different scope: resize_token_embeddings bug vs broad tokenization behavior changes across versions." + "reason": "Generation state updates after sync_gpus are unrelated to ANSI color handling in loading_report." }, { "left": "issue:44496", "right": "issue:45003", "accept": false, - "reason": "Unrecognized model/config loading error vs sys.modules import handling issue." - }, - { - "left": "issue:43232", - "right": "issue:44336", - "accept": false, - "reason": "Generation kwargs update after sync_gpus vs ANSI coloring in loading_report; unrelated subsystems." + "reason": "Unrecognized model/config loading is a different bug from unsafe sys.modules access in modeling_utils." }, { "left": "issue:44987", "right": "issue:45459", "accept": false, - "reason": "Both touch tokenizer errors, but one is model loading and the other is protobuf exception masking; not the same bug." + "reason": "Loading a specific model on transformers>=5.1.0 is not the same as masking tokenizer errors when protobuf is missing." }, { "left": "issue:43262", "right": "issue:44246", "accept": false, - "reason": "Audio chat-template sampling-rate bug vs import time performance regression." + "reason": "Audio sampling-rate defaults in chat templates are unrelated to intermittent import slowdowns." }, { - "left": "issue:45375", - "right": "issue:45759", + "left": "issue:42886", + "right": "issue:45229", "accept": false, - "reason": "Missing config field in strict deserialization vs from_config not unwrapping composite text_config." + "reason": "Offline cache loading for tokenizers and Gemma4 multi-GPU CUDA OOM are different failure modes." }, { - "left": "issue:42886", - "right": "issue:45229", + "left": "issue:45375", + "right": "issue:45759", "accept": false, - "reason": "Offline tokenizer cache loading vs Gemma4 multi-GPU OOM; unrelated failures." + "reason": "Missing config fields in strict serialization is not the same as unwrapping text_config for composite multimodal configs." }, { "left": "issue:43257", "right": "issue:45137", "accept": false, - "reason": "Qwen3 MOE weight conversion with accelerate/deepspeed vs ZeRO3 deque underflow; different code paths." + "reason": "MOE weight conversion with accelerate+deepspeed and a Zero3 deque pop are distinct deepspeed-related bugs." }, { "left": "issue:43065", "right": "issue:43408", "accept": false, - "reason": "Dummy Conv2d in Sam3PixelDecoder vs model-type warning for sam3_video/sam3_tracker mismatch." + "reason": "A dummy Conv2d in the pixel decoder is unrelated to a model-type warning when instantiating sam3_tracker." }, { "left": "issue:44995", "right": "issue:45200", "accept": false, - "reason": "Stale indexer cache on second forward pass vs required mm_token_type_ids for text-only fine-tuning." + "reason": "A stale cache on second forward pass is not the same as missing default mm_token_type_ids." }, { "left": "issue:43475", "right": "issue:43540", "accept": false, - "reason": "Missing vision encoder attribute vs video input ValueError in Qwen3OmniMoe; not the same failure." + "reason": "A missing fpn_position_embeddings attribute in SAM 3 Video is unrelated to Qwen3OmniMoe video input validation errors." }, { "left": "issue:44060", "right": "issue:44077", "accept": false, - "reason": "Incorrect tied-weights warning vs patchtsmixer post_init policy change; different issues." + "reason": "Incorrect tied-weights warning and a post_init allowance change are separate model/config validation issues." }, { - "left": "issue:44466", - "right": "issue:45335", + "left": "issue:45397", + "right": "issue:45656", "accept": false, - "reason": "Serialization of tied lm_head weights vs resize_token_embeddings not affecting decoder.embed_tokens." + "reason": "Gemma4 zero3 from_pretrained failure is not the same as optimizer.step being called twice under deepspeed." }, { - "left": "issue:45397", - "right": "issue:45656", + "left": "issue:44466", + "right": "issue:45335", "accept": false, - "reason": "Gemma-4 ZeRO3 loading problem vs double optimizer.step under deepspeed." + "reason": "Inconsistent lm_head serialization depending on device is different from resize_token_embeddings not propagating to decoder.embed_tokens." }, { "left": "issue:44995", "right": "issue:45588", "accept": false, - "reason": "GlmMoeDsa stale cache on second forward vs flash_attention crash on s_aux=None." + "reason": "A stale indexer cache on repeated forward passes is unrelated to flash_attention crashing on s_aux=None." }, { "left": "issue:43653", "right": "issue:44162", "accept": false, - "reason": "BigBirdTokenizer special-token registration bug vs ESM2 broken model workflow; unrelated." + "reason": "BigBirdTokenizer special-token registration is a tokenizer bug, while ESM2 being broken is a separate model regression." }, { "left": "issue:44368", "right": "issue:44977", "accept": false, - "reason": "Tied-embeddings warning during LoRA fine-tuning vs flash-attention generation failure; different concrete bugs." + "reason": "A tie_word_embeddings warning for Qwen3.5 LoRA is not the same as a flash-attention generation failure." }, { "left": "issue:44368", "right": "issue:45563", "accept": false, - "reason": "Tied-embeddings warning vs stale num_return_sequences warning in paged generate(); not duplicates." + "reason": "The tied-weights warning and the stale num_return_sequences warning are different generation warnings with different triggers." }, { "left": "issue:45137", "right": "issue:45600", "accept": false, - "reason": "DeepSpeed ZeRO3 deque error vs stale auto_mappings reference to removed Sam3LiteText configs." + "reason": "A Zero3 deque underflow is unrelated to auto_mappings references to removed SAM3 Lite configs." }, { "left": "issue:43673", "right": "issue:44734", "accept": false, - "reason": "Both involve generation/cache, but one is missing cache in chunked prefill and the other is wrong tensor indexing in serve continuation; too different to merge." + "reason": "Chunked prefill cache missing in GenerationMixin and KV-cache continuation tensor indexing are different generation/cache bugs." }, { "left": "issue:43931", "right": "issue:44368", "accept": false, - "reason": "Qwen3-VL weight-shape mismatch vs a tied-weights warning; unrelated symptoms." + "reason": "A Qwen3-VL weight-shape mismatch is unrelated to a tied-embeddings warning in Qwen3.5 LoRA." }, { "left": "issue:44869", "right": "issue:45356", "accept": false, - "reason": "Whisper trailing replacement-character decode crash vs Kimi-K2.5 codec handling regression; similar area, different bug." + "reason": "Whisper timestamp decoding crashes on a trailing replacement character, while Kimi-K2.5 has codec-handling and warning issues in its tokenizer." }, { "left": "issue:44933", "right": "issue:45459", "accept": false, - "reason": "Missing image_utils import vs tokenizer exception masking when protobuf is absent." + "reason": "A nonexistent import from image_utils is unrelated to tokenizer error handling when protobuf is absent." }, { "left": "issue:44297", "right": "issue:44623", "accept": false, - "reason": "tokenizer.save_pretrained metadata mismatch vs processor.save_pretrained missing files; different save paths and artifacts." + "reason": "Tokenizer save_pretrained writing the wrong tokenizer_class is different from processor.save_pretrained missing files." }, { "left": "issue:43232", "right": "issue:43673", "accept": false, - "reason": "Both touch generation internals, but one is sync_gpus kwargs updating and the other is cache availability; not the same bug." + "reason": "Both involve generation internals, but the concrete failures differ: sync_gpus update logic versus chunked_prefill cache handling." }, { "left": "issue:44361", "right": "issue:44749", "accept": false, - "reason": "MLukeTokenizer AttributeError on tasks vs a performance regression after upgrading transformers." + "reason": "MLukeTokenizer AttributeError on tasks is unrelated to the reported slowdown after upgrading transformers." }, { "left": "issue:43646", "right": "issue:44162", "accept": false, - "reason": "Custom model initialization breakage in v5 vs ESM2 model breakage; different root causes and surfaces." + "reason": "Custom model initialization breakage in Transformers 5.0.0 is a different class of failure from ESM2 being broken." + }, + { + "left": "issue:43673", + "right": "issue:45278", + "accept": false, + "reason": "Missing generation cache in chunked_prefill and many import errors after upgrade are not the same underlying bug." + }, + { + "left": "issue:44623", + "right": "issue:45656", + "accept": false, + "reason": "Missing files in processor.save_pretrained and duplicate optimizer steps under deepspeed are unrelated." + }, + { + "left": "issue:44749", + "right": "issue:45656", + "accept": false, + "reason": "A performance regression after version upgrade does not match optimizer.step being invoked twice." + }, + { + "left": "issue:44749", + "right": "issue:45310", + "accept": false, + "reason": "Slow filtering after upgrading to 5.3.0 is a different symptom from Qwen3.5 MOE from_pretrained failing on 5.4.0." } ] }, @@ -11244,7 +11234,9 @@ "issue:43377", "issue:43475", "issue:43482", + "issue:43504", "issue:43525", + "issue:43526", "issue:43643", "issue:43673", "issue:43704", @@ -11255,9 +11247,9 @@ "issue:44222", "issue:44242", "issue:44351", + "issue:44496", "issue:44623", "issue:44661", - "issue:44749", "issue:44792", "issue:44933", "issue:44945", @@ -11268,21 +11260,16 @@ "issue:45137", "issue:45200", "issue:45230", - "issue:45278", "issue:45310", "issue:45375", "issue:45381", "issue:45397", "issue:45478", - "issue:45656", + "issue:45561", "issue:45712", "issue:45753" ], "soft_pairs": [ - "issue:43673|issue:45278", - "issue:44623|issue:45656", - "issue:44749|issue:45656", - "issue:44749|issue:45310", "issue:42994|issue:43482", "issue:45230|issue:45397", "issue:28282|issue:30990", @@ -11295,8 +11282,8 @@ "issue:43475|issue:44661", "issue:42994|issue:43525", "issue:45375|issue:45478", - "issue:45310|issue:45375", "issue:43232|issue:44188", + "issue:45310|issue:45375", "issue:43873|issue:44945", "issue:42673|issue:43704", "issue:41628|issue:44351", @@ -11311,240 +11298,256 @@ "issue:44038|issue:45310", "issue:42222|issue:45200", "issue:43377|issue:44242", - "issue:44964|issue:45137" + "issue:44964|issue:45137", + "issue:43526|issue:45561", + "issue:43504|issue:45561", + "issue:44496|issue:45561", + "issue:42175|issue:44933" ], - "prepared_review_unit_hash": "ef0c9b42a9d76300252855e7979ebb20fac481ea73e222b732c01162d55b3c54", + "prepared_review_unit_hash": "31e129cd019502eff7a2b714d03241f04bf1b75ec41558f0575e71ded47cc29c", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 33, - "serialized_chars": 36628, - "estimated_input_tokens": 9157, - "estimated_eval_tokens": 18570 + "serialized_chars": 37264, + "estimated_input_tokens": 9316, + "estimated_eval_tokens": 18888 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly unrelated issue reports, but the packet contains at least two clear exact-duplicate pairs: 42673/43704 and 45310/45478. The soft-edge candidates are still conservatively rejected, and the earlier summary was too strong in implying the VRAM-leak pair was the only obvious duplicate.", - "confidence": 0.84, - "canonical_issue_reason": "42673 is still the best canonical pick for the VRAM-leak pair because it is the more detailed and higher-activity report, and 43704 repeats the exact same title and bug.", + "summary": "The cluster is still mostly heterogeneous, but it does contain more than one duplicate-looking pair: issue:42673/43704 is a clear same-bug duplicate pair, and issue:45310/45478 also have an identical title and look like a near-exact duplicate pair. So this is not just a single duplicate pair cluster, even though the rest of the items remain mixed.", + "confidence": 0.92, + "canonical_issue_reason": "issue:42673 is the cleanest canonical anchor for the cluster\u2019s strongest duplicate pair because it matches issue:43704 on the exact VRAM-leak title and same multi-threaded dataloader symptom. The cluster also contains another identical-title pair (issue:45310/45478), but that appears to be a separate duplicate subgroup rather than the main anchor.", "canonical_pr_reason": null, - "best_issue_reason": "42673 is a strong global issue representative because it is a concrete, clearly duplicated bug report with good detail and activity. That said, the set also includes another exact duplicate pair (45310/45478), so this cluster is not uniquely centered on only one duplicate.", + "best_issue_reason": "issue:42673 is still the best representative issue because it anchors the most clearly duplicated bug report in the packet and is the strongest evidence of an actual same-bug cluster. The other items are too mixed, and the 45310/45478 pair looks like a separate duplicate pocket rather than a better overall cluster representative.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43673", - "right": "issue:45278", - "accept": false, - "reason": "Different failure modes: generation cache regression vs broad import errors after upgrade." - }, - { - "left": "issue:44623", - "right": "issue:45656", - "accept": false, - "reason": "Processor save_pretrained file handling is unrelated to a deepspeed double optimizer step." - }, - { - "left": "issue:44749", - "right": "issue:45656", - "accept": false, - "reason": "Performance slowdown during filtering is not the same bug as duplicate optimizer stepping." - }, - { - "left": "issue:44749", - "right": "issue:45310", - "accept": false, - "reason": "Unrelated problems: filtering slowdown versus Qwen3.5 MoE from_pretrained failure." - }, { "left": "issue:42994", "right": "issue:43482", "accept": false, - "reason": "Quantized model saving failure is unrelated to GGUF loading failure in v5." + "reason": "Quantized model saving failure and GGUF loading failure are different code paths and different bug classes." }, { "left": "issue:45230", "right": "issue:45397", "accept": false, - "reason": "Generic bug report versus a specific Gemma-4 zero3 from_pretrained issue." + "reason": "A placeholder bug report is not the same issue as the Gemma-4 zero3 from_pretrained failure." }, { "left": "issue:28282", "right": "issue:30990", "accept": false, - "reason": "Torch ImportError for AutoModel is unrelated to a Sentence Transformers loading hang." + "reason": "PyTorch missing import error and Sentence Transformers loading hang are unrelated symptoms." }, { "left": "issue:44038", "right": "issue:44623", "accept": false, - "reason": "Qwen3-VL-Moe loading bug is unrelated to processor.save_pretrained missing files." + "reason": "Qwen3-VL-Moe/transformers 5 bug and processor.save_pretrained missing files are distinct failures." }, { "left": "issue:42371", "right": "issue:45092", "accept": false, - "reason": "TF32 API warning is unrelated to InternVL2 meta-init incompatibility." + "reason": "TF32 behavior settings warning is unrelated to old InternVL2 remote-code meta-init incompatibility." }, { "left": "issue:43475", "right": "issue:45381", "accept": false, - "reason": "Different model families and symptoms: SAM 3 attribute error vs Qwen2.5-VL position IDs." + "reason": "SAM 3 vision encoder missing attribute and Qwen2.5-VL video position-id bug are different models and paths." }, { "left": "issue:43475", "right": "issue:43643", "accept": false, - "reason": "A model output attribute bug is unrelated to missing AutoConfig fields with trust_remote_code." + "reason": "One is a SAM3 attribute error; the other is a generic trust_remote_code config-field omission." }, { "left": "issue:41628", "right": "issue:44933", "accept": false, - "reason": "Both are import-related, but they concern different missing symbols/modules and lack a concrete same-bug match." + "reason": "Both mention imports, but they are different missing symbols from different modules." }, { "left": "issue:45081", "right": "issue:45103", "accept": false, - "reason": "Tokenizer regex crash and auto_docstring annotation crash are unrelated code paths." + "reason": "Mistral tokenizer regex crash and auto_docstring annotation crash are unrelated." }, { "left": "issue:43475", "right": "issue:44661", "accept": false, - "reason": "A model output attribute bug is unrelated to add-new-model-like failing on TOKENIZER_MAPPING_NAMES." + "reason": "SAM3 encoder attribute failure is not the same as add-new-model-like failing inside TOKENIZER_MAPPING_NAMES." }, { "left": "issue:42994", "right": "issue:43525", "accept": false, - "reason": "Quantized saving failure and missing pad_token_id on Llama4Config are different issues." + "reason": "Quantized model saving failed and Llama4Config missing pad_token_id are different bugs." }, { "left": "issue:45375", "right": "issue:45478", "accept": false, - "reason": "Related Qwen3.5 family, but one is a missing config field and the other is a from_pretrained regression; not the same concrete bug." - }, - { - "left": "issue:45310", - "right": "issue:45375", - "accept": false, - "reason": "From_pretrained error versus silently dropped vision config field are different failures." + "reason": "They are clearly related by title, but the available evidence still points to a broader config-field omission versus a from_pretrained error, so I would not merge them without stronger same-code-path evidence." }, { "left": "issue:43232", "right": "issue:44188", "accept": false, - "reason": "Generation kwargs update after sync_gpus is unrelated to torch.compile attention kernel divergence." + "reason": "Generation kwargs update after sync_gpus and attention-kernel divergence under torch.compile are different code-path issues." + }, + { + "left": "issue:45310", + "right": "issue:45375", + "accept": false, + "reason": "A generic Qwen3.5-MoE from_pretrained error is not the same concrete bug as the missing deepstack_visual_indexes field." }, { "left": "issue:43873", "right": "issue:44945", "accept": false, - "reason": "Quantization offloading issue is unrelated to incorrect pipeline-parallel outputs." + "reason": "Quantization/offloading behavior and pipeline-parallelism output corruption are separate problems." }, { "left": "issue:42673", "right": "issue:43704", "accept": true, - "reason": "Exact same bug/title: Qwen3ForCausalLM VRAM leak across multiple dataloader threads." + "reason": "These are the same Qwen3ForCausalLM VRAM-leak report with identical titles and the same multi-threaded dataloader symptom." }, { "left": "issue:41628", "right": "issue:44351", "accept": false, - "reason": "Different missing imports: AutoImageProcessor versus HybridCache." + "reason": "AutoImageProcessor import failure and HybridCache import failure are unrelated missing-symbol bugs." }, { "left": "issue:42831", "right": "issue:44222", "accept": false, - "reason": "FP8 accuracy regression is not the same as FP8 save_pretrained failure." + "reason": "FineGrainedFP8 accuracy drift is not the same issue as FP8 save_pretrained for MoE." }, { "left": "issue:44945", "right": "issue:45753", "accept": false, - "reason": "Pipeline-parallel output mismatch is unrelated to a specific-image infinite loop in Qwen3_5." + "reason": "Pipeline-parallelism output mismatch and an infinite loop on a specific image are different behaviors." }, { "left": "issue:44222", "right": "issue:45310", "accept": false, - "reason": "FP8 save_pretrained moe and Qwen3.5 MoE from_pretrained are different bugs." + "reason": "FP8 save_pretrained for MoE and Qwen3.5-MoE from_pretrained failure do not describe the same concrete bug." }, { "left": "issue:43673", "right": "issue:44188", "accept": false, - "reason": "Generation cache regression and attention-kernel branching under torch.compile are unrelated." + "reason": "Chunked-prefill GenerationMixin cache issue and torch.compile branching divergence are different generation bugs." }, { "left": "issue:36010", "right": "issue:41628", "accept": false, - "reason": "Both are import errors, but for different symbols and different regressions." + "reason": "GenerationMixin import failure and AutoImageProcessor import failure are different missing imports." }, { "left": "issue:42994", "right": "issue:44164", "accept": false, - "reason": "Quantized model saving failure is not the same as extra_state handling in save/from_pretrained." + "reason": "Quantized model saving and extra_state save/from_pretrained handling are related only broadly, not the same bug." }, { "left": "issue:42175", "right": "issue:45712", "accept": false, - "reason": "Packaging/extra dependency behavior differs from leftover dummy class cleanup." + "reason": "Missing TensorFlow in the torch extra and dummy_pt_objects leaking into dir(transformers) are separate packaging issues." }, { "left": "issue:43232", "right": "issue:44792", "accept": false, - "reason": "Generation sync_gpus bug is unrelated to a Janus image-generation test failure." + "reason": "Generation state update and a janus image-generation test failure are not the same underlying issue." }, { - "left": "issue:44038", - "right": "issue:45310", + "left": "issue:43526", + "right": "issue:45561", "accept": false, - "reason": "Different Qwen variants and different symptoms; not enough to treat as the same bug." + "reason": "BeitImageProcessorFast label reduction and xdist captured_info.txt race are unrelated." }, { - "left": "issue:42222", - "right": "issue:45200", + "left": "issue:43504", + "right": "issue:45561", + "accept": false, + "reason": "BEiT pretrained preset legacy-field handling and the xdist race are different bugs." + }, + { + "left": "issue:44496", + "right": "issue:45137", "accept": false, - "reason": "Vitpose model breakage is unrelated to Gemma-4 token type ID defaults." + "reason": "Unrecognized model/config loading and DeepSpeed ZeRO3 deque underflow are unrelated." }, { "left": "issue:43377", "right": "issue:44242", "accept": false, - "reason": "MIMI padding-mask inconsistency is unrelated to MoE load-balancing loss omission." + "reason": "MIMI batched-vs-single output mismatch and load-balancing loss omission are different model issues." }, { "left": "issue:44964", "right": "issue:45137", "accept": false, - "reason": "Phi-4 multimodal loading failure and DeepSpeed ZeRO3 deque error are different code paths." + "reason": "Phi-4 multimodal loading failure and DeepSpeed ZeRO3 deque error are unrelated." + }, + { + "left": "issue:43526", + "right": "issue:45561", + "accept": false, + "reason": "The label-reduction bug is unrelated to the xdist captured_info.txt race." + }, + { + "left": "issue:43504", + "right": "issue:45561", + "accept": false, + "reason": "Legacy-field loading failure in BEiT is unrelated to the xdist race." + }, + { + "left": "issue:42175", + "right": "issue:44933", + "accept": false, + "reason": "TF backend packaging and image_utils import issues are different problems." + }, + { + "left": "issue:44038", + "right": "issue:45310", + "accept": false, + "reason": "Qwen3-VL-Moe transformer bug and Qwen3.5-MoE from_pretrained error are different model/version failures." + }, + { + "left": "issue:42831", + "right": "issue:44222", + "accept": false, + "reason": "FP8 accuracy regression and FP8 MoE save_pretrained failure are not the same code path." } ] }, "evaluator_result": { "accept": false, - "feedback": "The canonical duplicate call on 42673/43704 is grounded, but the summary is too strong: the packet also contains another obvious duplicate pair with the exact same title, issue 45310 and issue 45478 ('[BUG] transformers>=5.4.0, Qwen3.5 Moe from_pretrained error'). So it is not accurate to say the only clear duplicate is the Qwen3ForCausalLM VRAM leak report. The soft-edge verdicts themselves are mostly conservative, but the overall summary overstates uniqueness." + "feedback": "The cluster summary is a bit too strong: the packet contains at least two obvious duplicate-looking pairs with identical titles, not just issue:42673/43704. In particular, issue:45310 and issue:45478 share the exact same title, so saying there is \u2018only one clear duplicate pair\u2019 is not fully grounded. The conservative false soft-edge calls are mostly fine, but the overall narrative should acknowledge this additional near-exact duplicate instead of presenting the cluster as having a single clear duplicate pair." }, "evaluator_used": true, "retried": true, @@ -11565,19 +11568,21 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-404", "nodes": [ + "issue:22355", + "issue:28282", "issue:30333", "issue:30990", "issue:33290", + "issue:34689", "issue:35141", "issue:36331", - "issue:42175", + "issue:41628", + "issue:42757", "issue:42831", "issue:42898", "issue:43116", "issue:43278", "issue:43479", - "issue:43504", - "issue:43526", "issue:43644", "issue:43723", "issue:43784", @@ -11594,7 +11599,7 @@ "issue:44297", "issue:44373", "issue:44448", - "issue:44496", + "issue:44484", "issue:44568", "issue:44623", "issue:44734", @@ -11608,16 +11613,11 @@ "issue:45310", "issue:45356", "issue:45405", - "issue:45561", "issue:45656", "issue:45712", "issue:45812" ], "soft_pairs": [ - "issue:43526|issue:45561", - "issue:44496|issue:45561", - "issue:43504|issue:45561", - "issue:42175|issue:44933", "issue:42898|issue:44568", "issue:43479|issue:44869", "issue:30990|issue:35141", @@ -11625,8 +11625,8 @@ "issue:43479|issue:44079", "issue:44623|issue:45310", "issue:43784|issue:43824", - "issue:43278|issue:45137", "issue:44188|issue:44373", + "issue:43278|issue:45137", "issue:42898|issue:45812", "issue:44734|issue:45103", "issue:30333|issue:33290", @@ -11641,239 +11641,257 @@ "issue:44933|issue:45712", "issue:44222|issue:44297", "issue:44297|issue:45656", - "issue:43479|issue:43644", "issue:43116|issue:44190", + "issue:43479|issue:43644", "issue:44222|issue:44623", "issue:42831|issue:44623", "issue:44206|issue:45103", - "issue:30333|issue:36331" + "issue:41628|issue:42757", + "issue:30333|issue:36331", + "issue:22355|issue:28282", + "issue:30333|issue:34689", + "issue:44297|issue:45310", + "issue:44222|issue:44484", + "issue:28282|issue:30333" ], - "prepared_review_unit_hash": "92571dd94164bdeb7035931fcccda64b26d24c031abb770dc0ccc7ba751e7020", + "prepared_review_unit_hash": "38d473efd81445977bb8adeac31191b387dd5836491d31de943946f9a3df77c2", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 33, - "serialized_chars": 37203, - "estimated_input_tokens": 9301, - "estimated_eval_tokens": 18858 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 35, + "serialized_chars": 38240, + "estimated_input_tokens": 9560, + "estimated_eval_tokens": 19376 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is mostly a loose bag of unrelated bugs; I found no soft-edge pairs that look like the same underlying issue. The best representative issue is the broad, user-impacting Granite tokenizer regression.", - "confidence": 0.51, - "canonical_issue_reason": "issue:45812 is the most broadly scoped and severe user-facing regression in the set: a silent v4\u2192v5 tokenizer bug affecting all Granite models.", + "summary": "This is a heterogeneous cluster of mostly unrelated Transformers issues: tokenizer regressions, model-loading/import failures, training/runtime bugs, save/load serialization issues, and a few doc/config problems. None of the soft edges look like true duplicates, so all are rejected.", + "confidence": 0.88, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:45812 is the strongest anchor for the cluster because it describes a concrete, high-impact tokenizer regression with wide model coverage and clear reproduction scope.", + "best_issue_reason": "issue:45812 is the strongest standalone report in the set: it is concrete, reproducible, and describes a high-impact silent regression in AutoTokenizer for Granite models.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43526", - "right": "issue:45561", - "accept": false, - "reason": "Unrelated: BEiT label reduction bug vs pytest-xdist file-race in testing utils." - }, - { - "left": "issue:44496", - "right": "issue:45561", - "accept": false, - "reason": "Unrelated: model auto-loading/config error vs test artifact race condition." - }, - { - "left": "issue:43504", - "right": "issue:45561", - "accept": false, - "reason": "Different code paths and symptoms: BEiT pretrained loading legacy field vs testing utility race." - }, - { - "left": "issue:42175", - "right": "issue:44933", - "accept": false, - "reason": "Unrelated backend packaging issue vs missing image_utils import." - }, { "left": "issue:42898", "right": "issue:44568", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different behaviors and code paths; not the same bug." + "reason": "Both are tokenizer-related regressions, but one is about clean_up_tokenization_spaces behavior and the other about add_special_tokens not adding BOS/EOS; different code paths and symptoms." }, { "left": "issue:43479", "right": "issue:44869", "accept": false, - "reason": "Different subsystems: multimodal config default initialization vs Whisper timestamp decoding crash." + "reason": "Phi4MultimodalConfig default initialization and Whisper word-timestamp decode crashing are unrelated bugs in different subsystems." }, { "left": "issue:30990", "right": "issue:35141", "accept": false, - "reason": "No shared underlying bug: sentence-transformer loading hang vs embedding reinit on resize." + "reason": "Sentence Transformers loading hangs and embedding resize reinitialization are distinct failures with no shared code-path." }, { "left": "issue:43976", "right": "issue:44933", "accept": false, - "reason": "Unrelated: Python version compatibility vs missing import symbol." + "reason": "Python-version compatibility/import support is unrelated to the missing image_utils import symbol bug." }, { "left": "issue:43479", "right": "issue:44079", "accept": false, - "reason": "Both mention None/default handling, but in different classes and unrelated failures." + "reason": "Config initialization for Phi4Multimodal is unrelated to ModelOutput key assignment behavior." }, { "left": "issue:44623", "right": "issue:45310", "accept": false, - "reason": "Save-pretrained missing files vs Qwen3.5 MoE from_pretrained error; not the same concrete problem." + "reason": "Both mention loading/saving, but one is missing processor save artifacts and the other is a from_pretrained failure for Qwen3.5 MoE; different root causes." }, { "left": "issue:43784", "right": "issue:43824", "accept": false, - "reason": "Different import failures for different packages/models; no common code-path bug." + "reason": "Both are import errors, but they involve different packages/models and different missing symbols, so they are not the same bug." }, { - "left": "issue:43278", - "right": "issue:45137", + "left": "issue:44188", + "right": "issue:44373", "accept": false, - "reason": "Unrelated training/eval dtype regression vs DeepSpeed ZeRO3 deque error." + "reason": "A numerical/kernel divergence under torch.compile is not the same as a wrong docstring." }, { - "left": "issue:44188", - "right": "issue:44373", + "left": "issue:43278", + "right": "issue:45137", "accept": false, - "reason": "Attention-kernel divergence vs docstring typo; unrelated." + "reason": "Embedding dtype drift during eval and a DeepSpeed ZeRO3 deque IndexError are unrelated runtime bugs." }, { "left": "issue:42898", "right": "issue:45812", "accept": false, - "reason": "Both are tokenizer regressions in v5, but they are distinct failures on different APIs/models." + "reason": "Both are tokenizer regressions, but one concerns cleanup-space handling and the other wrong token IDs for Granite models; different failure mode and model scope." }, { "left": "issue:44734", "right": "issue:45103", "accept": false, - "reason": "Serve KV-cache indexing crash vs auto-docstring annotation crash; different subsystems." + "reason": "KV-cache continuation tensor indexing in transformers serve is unrelated to the auto_docstring AttributeError." }, { "left": "issue:30333", "right": "issue:33290", "accept": false, - "reason": "MLflow reporting on training failure vs DeepSpeed/Adafactor OOM; unrelated." + "reason": "MLFlow reporting on failed training and DeepSpeed OOM with Adafactor are separate training issues." }, { "left": "issue:43723", "right": "issue:44568", "accept": false, - "reason": "General tokenizer loading issue vs missing BOS/EOS addition; not the same bug." + "reason": "Tokenizer loading in v5 and add_special_tokens BOS/EOS behavior are different tokenizer problems." }, { "left": "issue:42898", "right": "issue:45356", "accept": false, - "reason": "Different tokenizer regressions with different model-specific root causes." + "reason": "These are both tokenizer regressions, but one is a generic v5 cleanup-space change and the other is Kimi-K2.5 codec handling/fix_mistral_regex behavior." }, { "left": "issue:43874", "right": "issue:45092", "accept": false, - "reason": "Different multimodal model-loading failures and different causes." + "reason": "A missing image patch count method and a remote-code checkpoint meta-init incompatibility are different multimodal model bugs." }, { "left": "issue:44060", "right": "issue:45125", "accept": false, - "reason": "Unrelated Qwen3-Next tied-weights warning vs missing tensor-parallel plan." + "reason": "Incorrect tied-weights warnings for Qwen3-Next and missing _tp_plan for Qwen3.5 MoE are separate model-specific issues." }, { "left": "issue:42898", "right": "issue:44779", "accept": false, - "reason": "Both involve tokenizers, but the reported regressions and affected models differ." + "reason": "Both are tokenizer regressions, but they affect different models and different logic paths." }, { "left": "issue:45405", "right": "issue:45712", "accept": false, - "reason": "Dependency version pin issue vs leftover dummy classes/check-repo problem." + "reason": "A dependency version bump is unrelated to cleanup of leftover dummy classes and repo-check failures." }, { - "left": "issue:44933", - "right": "issue:45712", + "left": "issue:42898", + "right": "issue:44448", "accept": false, - "reason": "Missing import vs dummy-class cleanup; unrelated." + "reason": "One is a tokenizer cleanup change; the other is a model output difference for pegasus between major versions." }, { "left": "issue:44295", "right": "issue:45656", "accept": false, - "reason": "Buffer read/write bug vs optimizer step duplication under DeepSpeed; different symptoms and layers." + "reason": "Reading a registered position_ids buffer and a double optimizer-step in DeepSpeed are unrelated training/runtime issues." + }, + { + "left": "issue:44933", + "right": "issue:45712", + "accept": false, + "reason": "Missing image_utils import and leftover dummy classes are different import/cleanup problems." }, { "left": "issue:44222", "right": "issue:44297", "accept": false, - "reason": "FP8 save_pretrained MoE issue vs tokenizer_class mismatch in tokenizer_config; unrelated save paths." + "reason": "FP8 MoE save_pretrained and tokenizer_class metadata mismatch are both serialization-related, but they affect different object types and failure modes." }, { "left": "issue:44297", "right": "issue:45656", "accept": false, - "reason": "Tokenizer metadata mismatch vs DeepSpeed optimizer stepping bug." - }, - { - "left": "issue:43479", - "right": "issue:43644", - "accept": false, - "reason": "Config defaulting bug vs non-persistent buffer junk; not the same underlying issue." + "reason": "Tokenizer save metadata mismatch and double optimizer steps are unrelated." }, { "left": "issue:43116", "right": "issue:44190", "accept": false, - "reason": "Different example scripts and failures: multi-label output handling vs local dataset loading." + "reason": "Multi-label classification example behavior and local dataset loading in a different example script are not the same bug." }, { - "left": "issue:44206", - "right": "issue:45103", + "left": "issue:43479", + "right": "issue:43644", "accept": false, - "reason": "Feature extractor argument regression vs auto-docstring annotation crash." + "reason": "Phi4MultimodalConfig default config initialization is unrelated to non-persistent buffer initialization in v5.0.0." }, { "left": "issue:44222", "right": "issue:44623", "accept": false, - "reason": "Both touch save_pretrained, but one is FP8 MoE serialization and the other processor file output." + "reason": "Both involve save_pretrained, but one is FP8 MoE serialization and the other is missing processor files; not the same code-path." }, { "left": "issue:42831", "right": "issue:44623", "accept": false, - "reason": "Accuracy regression in FineGrainedFP8 vs missing processor files; unrelated." + "reason": "FineGrainedFP8 accuracy drift is not the same as processor save_pretrained missing files." }, { "left": "issue:44206", "right": "issue:45103", "accept": false, - "reason": "Different components and failures; no shared code-path." + "reason": "An audio feature extractor center-arg regression and an auto_docstring AttributeError are unrelated." + }, + { + "left": "issue:41628", + "right": "issue:42757", + "accept": false, + "reason": "These are separate import errors from different modules/symbols, not one shared missing import bug." }, { "left": "issue:30333", "right": "issue:36331", "accept": false, - "reason": "MLflow failure reporting vs custom trainer loss-signature breakage." + "reason": "MLFlow reporting on failed jobs and the compute_loss signature change are distinct trainer issues." + }, + { + "left": "issue:22355", + "right": "issue:28282", + "accept": false, + "reason": "Missing transformers.onnx and missing PyTorch for AutoModel are different environment/import failures." + }, + { + "left": "issue:30333", + "right": "issue:34689", + "accept": false, + "reason": "MLFlow reporting on failed training is unrelated to the Llama 3.2 Vision model-loading regression." + }, + { + "left": "issue:44297", + "right": "issue:45310", + "accept": false, + "reason": "Tokenizer save metadata mismatch and Qwen3.5 from_pretrained failure are different bugs." + }, + { + "left": "issue:44222", + "right": "issue:44484", + "accept": false, + "reason": "FP8 MoE save_pretrained serialization and a question about max_shard_size documentation are not duplicates." + }, + { + "left": "issue:28282", + "right": "issue:30333", + "accept": false, + "reason": "A missing PyTorch dependency import error is unrelated to MLFlow not being reported after failed training." } ] }, @@ -11897,11 +11915,8 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-404", "nodes": [ - "issue:22355", - "issue:28282", "issue:30333", "issue:30990", - "issue:34689", "issue:35141", "issue:38175", "issue:39401", @@ -11910,25 +11925,27 @@ "issue:42831", "issue:42886", "issue:42994", + "issue:43066", + "issue:43122", "issue:43232", "issue:43479", + "issue:43576", "issue:43723", "issue:43756", + "issue:43784", "issue:43824", "issue:44038", "issue:44060", - "issue:44222", "issue:44261", "issue:44263", "issue:44295", - "issue:44297", "issue:44360", "issue:44373", "issue:44442", "issue:44483", - "issue:44484", "issue:44485", "issue:44514", + "issue:44521", "issue:44568", "issue:44623", "issue:44829", @@ -11943,24 +11960,19 @@ "issue:45357", "issue:45406", "issue:45593", + "issue:45753", "issue:45834", "issue:46002" ], "soft_pairs": [ - "issue:41628|issue:42757", - "issue:22355|issue:28282", - "issue:30333|issue:34689", - "issue:44297|issue:45310", - "issue:44222|issue:44484", - "issue:28282|issue:30333", - "issue:43232|issue:45834", "issue:43723|issue:44977", + "issue:43232|issue:45834", "issue:30333|issue:35141", "issue:44263|issue:45092", - "issue:44295|issue:44623", - "issue:44514|issue:45103", "issue:44483|issue:45406", "issue:44373|issue:44485", + "issue:44295|issue:44623", + "issue:44514|issue:45103", "issue:44263|issue:44360", "issue:44261|issue:45092", "issue:45092|issue:45357", @@ -11978,228 +11990,220 @@ "issue:42757|issue:45278", "issue:42886|issue:42994", "issue:43232|issue:45593", - "issue:43756|issue:44060" + "issue:43756|issue:44060", + "issue:42831|issue:45753", + "issue:43576|issue:43784", + "issue:42886|issue:43066", + "issue:43122|issue:44521" ], - "prepared_review_unit_hash": "e111f0bc865c656e2f204667317589284546f4e181dfb4adbbafd3146d23e65e", + "prepared_review_unit_hash": "402a7d859d8062df05f71fd5a55a4693eb695bb9fe790adf2fbf14a966fac08d", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 37049, - "estimated_input_tokens": 9263, - "estimated_eval_tokens": 18782 + "soft_pair_count": 30, + "serialized_chars": 36434, + "estimated_input_tokens": 9109, + "estimated_eval_tokens": 18474 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is dominated by superficially similar Transformers import/tokenizer/model-loading regressions, but the pairwise evidence does not show a shared underlying defect. The soft matches mostly span different code paths, different models, or different failure modes, so none should be merged as duplicates.", - "confidence": 0.88, - "canonical_issue_reason": "issue:43723 is the broadest and most central issue here: a generic v5 tokenizer-loading failure that also sits in the middle of several tokenizer-related soft matches. It is still only a loose umbrella, not a true duplicate hub for the whole cluster.", + "summary": "The cluster is dominated by Transformers v5 regression reports, but the soft-linked pairs mostly describe different code paths, models, or failure modes rather than true duplicates. No soft edge looks safe to merge.", + "confidence": 0.38, + "canonical_issue_reason": "issue:43723 is the broadest and most central report here: a generic AutoTokenizer/from_pretrained loading failure in v5, which best represents the tokenization/loading regression subgroup in this cluster.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43723 is the best representative issue because it is the most generic, recurring tokenizer-loading regression among the candidates. That said, the cluster is too heterogeneous for a strong canonical issue choice.", + "best_issue_reason": "issue:43723 is the best overall issue candidate because it is the most generic and reusable summary of the cluster\u2019s recurring tokenizer-loading regressions; other issues are either model-specific or about unrelated subsystems.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:41628", - "right": "issue:42757", - "accept": false, - "reason": "Both are import errors, but one is AutoImageProcessor and the other is is_offline_mode from huggingface_hub; different symbols and likely different fixes." - }, - { - "left": "issue:22355", - "right": "issue:28282", - "accept": false, - "reason": "No module named transformers.onnx is unrelated to AutoModel requiring PyTorch; they concern different missing dependencies." - }, - { - "left": "issue:30333", - "right": "issue:34689", - "accept": false, - "reason": "MLFlow reporting on failed training jobs is unrelated to a model-loading regression for Llama 3.2 Vision." - }, - { - "left": "issue:44297", - "right": "issue:45310", - "accept": false, - "reason": "One is tokenizer_config save metadata mismatch, the other is a Qwen3.5 MoE from_pretrained load failure; different bug surfaces." - }, - { - "left": "issue:44222", - "right": "issue:44484", - "accept": false, - "reason": "FP8 save_pretrained for MoE and max_shard_size behavior are both save-related, but they are different issues with different root causes." - }, - { - "left": "issue:28282", - "right": "issue:30333", + "left": "issue:43723", + "right": "issue:44977", "accept": false, - "reason": "PyTorch-not-found import error and MLFlow job reporting are unrelated failures." + "reason": "Both mention v5-era generation/tokenizer problems, but one is a generic tokenizer load failure and the other is a Qwen3.5 flash-attention generation bug; different underlying paths." }, { "left": "issue:43232", "right": "issue:45834", "accept": false, - "reason": "Generation kwargs handling after sync_gpus is unrelated to a long-OCR index error in Kosmos2.5." - }, - { - "left": "issue:43723", - "right": "issue:44977", - "accept": false, - "reason": "Generic tokenizer loading in v5 and Qwen3.5 flash-attention generation problems are different code paths." + "reason": "Generation state update after sync_gpus is unrelated to Kosmos2.5 OCR long-input indexing; no shared bug mechanism." }, { "left": "issue:30333", "right": "issue:35141", "accept": false, - "reason": "Failed-job MLFlow reporting is unrelated to embedding reinitialization during post_init." + "reason": "MLflow job status reporting and output embedding reinitialization are unrelated." }, { "left": "issue:44263", "right": "issue:45092", "accept": false, - "reason": "torch.split return handling in a DSA indexer is unrelated to InternVL2 meta-initialization incompatibility." + "reason": "torch.split return handling in GlmMoeDsaIndexer is not the same bug as InternVL2 meta-initialization compatibility." }, { - "left": "issue:44295", - "right": "issue:44623", + "left": "issue:44483", + "right": "issue:45406", "accept": false, - "reason": "Reading position_ids after buffer registration and processor.save_pretrained missing files are distinct processor/model-save bugs." + "reason": "Chat/completions request handling and Gemma4Processor _tokenizer initialization are distinct failures." }, { - "left": "issue:44514", - "right": "issue:45103", + "left": "issue:44373", + "right": "issue:44485", "accept": false, - "reason": "A Qwen2_5_VLProcessor batched padding crash is unrelated to an auto-docstring AttributeError from future annotations." + "reason": "A wrong docstring for position_ids is not the same as the GLM-5 RoPE implementation bug." }, { - "left": "issue:44483", - "right": "issue:45406", + "left": "issue:44295", + "right": "issue:44623", "accept": false, - "reason": "Chat-completions request handling and Gemma4Processor missing _tokenizer are different failures despite both involving serving." + "reason": "Buffer/readback of position_ids and processor.save_pretrained file emission are separate serialization issues." }, { - "left": "issue:44373", - "right": "issue:44485", + "left": "issue:44514", + "right": "issue:45103", "accept": false, - "reason": "A wrong docstring for position_ids is not the same as a GLM-5 RoPE implementation bug." + "reason": "apply_chat_template batching crash and auto_docstring annotation handling are unrelated." }, { "left": "issue:44263", "right": "issue:44360", "accept": false, - "reason": "Both mention the DSA indexer, but one is about torch.split outputs and the other about a missing ReLU; not the same defect." + "reason": "Both are in the DSA/indexer area, but one is about torch.split outputs and the other about a missing ReLU; different code defects." }, { "left": "issue:44261", "right": "issue:45092", "accept": false, - "reason": "Missing rms_norm_eps in MLA q_a_layernorm is unrelated to remote-code checkpoint/meta-init incompatibility." + "reason": "MLA q_a_layernorm epsilon precision and InternVL2 meta-init incompatibility do not point to the same change." }, { "left": "issue:45092", "right": "issue:45357", "accept": false, - "reason": "Meta initialization incompatibility and incorrect visual encoder keys in save_pretrained are separate issues." + "reason": "Remote-code meta initialization incompatibility and Qwen3.5 incorrect visual encoder save keys are different issues." }, { "left": "issue:43723", "right": "issue:44871", "accept": false, - "reason": "Tokenizer loading in v5 and Gemma-3 eos_token_id inconsistency are different tokenizer/model-config problems." + "reason": "Tokenizer loading failure and eos_token_id mismatch are both tokenizer-related, but they are different concrete bugs." }, { "left": "issue:43723", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer load problems, but the affected model and failure mode differ; not the same underlying bug." + "reason": "Both are tokenizer loading regressions, but the reports are model-specific and do not clearly share the same failing path." }, { "left": "issue:30990", "right": "issue:38175", "accept": false, - "reason": "Sentence Transformers loading stalling and siglip2 zero probabilities are unrelated model behavior reports." + "reason": "Sentence Transformers loading hang and SigLIP2 zero-probabilities are unrelated." }, { "left": "issue:44829", "right": "issue:46002", "accept": false, - "reason": "Flash-attention-3 training degeneration and NaNs in classification heads after init/checkpointing are different numerical issues." + "reason": "Flash-attention-3 training degeneration is not the same as NaNs appearing after checkout/init in classification heads." }, { "left": "issue:43723", "right": "issue:44442", "accept": false, - "reason": "AutoTokenizer failing to load FastSpeech2ConformerTokenizer is a specific tokenizer registration problem, not the same as the broader v5 tokenizer-loading issue." + "reason": "Generic AutoTokenizer load failure and FastSpeech2ConformerTokenizer load failure are different tokenizer implementations and likely different causes." }, { "left": "issue:43479", "right": "issue:44913", "accept": false, - "reason": "Phi4Multimodal default config initialization and GPTNeoX rotary_pct reload behavior are both config-related but not the same bug." + "reason": "Incorrect default config initialization and rotary_pct not persisting on reload are separate config serialization bugs." }, { "left": "issue:41628", "right": "issue:43824", "accept": false, - "reason": "Different missing imports: AutoImageProcessor versus Qwen2_5_VLForConditionalGeneration." + "reason": "Missing AutoImageProcessor import and missing Qwen2_5_VLForConditionalGeneration import are different symbol/export issues." }, { "left": "issue:42831", "right": "issue:45310", "accept": false, - "reason": "FineGrainedFP8 accuracy concerns are unrelated to a Qwen3.5 MoE from_pretrained error." + "reason": "FineGrainedFP8 accuracy loss and Qwen3.5 moe from_pretrained failure are not the same bug." }, { "left": "issue:44263", "right": "issue:44373", "accept": false, - "reason": "DSA indexer torch.split behavior and a position_ids docstring issue have no shared defect." + "reason": "DSA split return values and a wrong docstring for position_ids are unrelated." }, { "left": "issue:42831", "right": "issue:44038", "accept": false, - "reason": "FineGrainedFP8 accuracy issues and Qwen3-VL-Moe bugs are separate problems." + "reason": "FP8 accuracy regression and Qwen3-VL-Moe bug are different model/runtime issues." }, { "left": "issue:39401", "right": "issue:44568", "accept": false, - "reason": "Qwen3 offset_mapping errors and mdeberta-v3 add_special_tokens/BOS-EOS behavior are different tokenizer bugs." + "reason": "Wrong offset_mapping and missing BOS/EOS on add_special_tokens are both tokenizer behavior problems, but they affect different behaviors and code paths." }, { "left": "issue:42757", "right": "issue:45278", "accept": false, - "reason": "is_offline_mode import failure is one specific breakage; many import errors after upgrade is a broader report, not the same defect." + "reason": "A specific is_offline_mode import error and a broad set of import errors after upgrade are related in theme but not clearly the same underlying defect." }, { "left": "issue:42886", "right": "issue:42994", "accept": false, - "reason": "Tokenizer cache loading offline and quantized model saving are unrelated issues." + "reason": "Offline tokenizer cache loading and quantized model saving failure are unrelated." }, { "left": "issue:43232", "right": "issue:45593", "accept": false, - "reason": "Generation kwargs after sync_gpus and missing auxiliary losses in D-FINE are unrelated." + "reason": "Generation kwargs update after sync_gpus is unrelated to D-FINE auxiliary-loss behavior." }, { "left": "issue:43756", "right": "issue:44060", "accept": false, - "reason": "Smollm3 RoPE-layer dropping and Qwen3-Next tied-weights warning are different architecture/config problems." + "reason": "RoPE layer selection in Smollm3 and tied-weights warning behavior in Qwen3-Next are different model configuration bugs." + }, + { + "left": "issue:42831", + "right": "issue:45753", + "accept": false, + "reason": "FP8 accuracy regression and Qwen3_5 infinite loop on one image are different failures." + }, + { + "left": "issue:43576", + "right": "issue:43784", + "accept": false, + "reason": "Broken transformers env command and sentence-transformers import NameError are distinct issues, even if both are environment/version related." + }, + { + "left": "issue:42886", + "right": "issue:43066", + "accept": false, + "reason": "Offline cache tokenizer loading and wrong tokenizer decoder type in v5 are different tokenizer problems." + }, + { + "left": "issue:43122", + "right": "issue:44521", + "accept": false, + "reason": "Different tokenization across versions and zero assistant masks in multimodal apply_chat_template are not the same bug." } ] }, @@ -12228,30 +12232,30 @@ "issue:29127", "issue:30990", "issue:33357", + "issue:36296", "issue:39401", "issue:41628", "issue:42371", - "issue:42831", - "issue:42886", "issue:42907", - "issue:43066", "issue:43116", "issue:43122", "issue:43232", "issue:43408", "issue:43452", "issue:43475", - "issue:43576", "issue:43646", "issue:43784", + "issue:43824", "issue:43937", "issue:43976", "issue:44038", "issue:44117", + "issue:44222", "issue:44261", "issue:44263", "issue:44295", "issue:44297", + "issue:44351", "issue:44360", "issue:44373", "issue:44485", @@ -12269,14 +12273,9 @@ "issue:45278", "issue:45290", "issue:45600", - "issue:45753", "issue:46002" ], "soft_pairs": [ - "issue:42831|issue:45753", - "issue:43576|issue:43784", - "issue:42886|issue:43066", - "issue:43122|issue:44521", "issue:22355|issue:28218", "issue:44117|issue:44704", "issue:42371|issue:44263", @@ -12290,8 +12289,8 @@ "issue:41628|issue:45278", "issue:43976|issue:44945", "issue:39401|issue:44488", - "issue:43784|issue:45600", "issue:44261|issue:44373", + "issue:43784|issue:45600", "issue:42907|issue:43937", "issue:43408|issue:43475", "issue:44360|issue:44485", @@ -12303,222 +12302,219 @@ "issue:29127|issue:30990", "issue:43122|issue:44779", "issue:44038|issue:44295", - "issue:42907|issue:43452" + "issue:36296|issue:43116", + "issue:42907|issue:43452", + "issue:43824|issue:44351", + "issue:44038|issue:44222" ], - "prepared_review_unit_hash": "76cb8dbc1b226a3d4a7160ef4873fbb92947b4ca9da1ea060f9014028f336a74", + "prepared_review_unit_hash": "2f3d75c94c77c12867acd9c6f0e800ec7457990e439c4672f6b54d53832c4d71", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 31, - "serialized_chars": 36519, - "estimated_input_tokens": 9130, - "estimated_eval_tokens": 18516 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 35654, + "estimated_input_tokens": 8914, + "estimated_eval_tokens": 18084 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is a grab bag of mostly unrelated Transformers issues spanning tokenization, multimodal/chat templates, model loading, generation, config plumbing, and a few model-specific bugs. None of the soft-similarity pairs look like the same underlying bug or change, so I would reject all soft edges.", - "confidence": 0.88, + "summary": "This cluster is a loose grab-bag of unrelated Transformers issues, not a duplicate set. The soft pairs share broad themes like imports, tokenization, multimodal/chat-template behavior, or version regressions, but each points to a different concrete bug or model path, so none should be merged.", + "confidence": 0.96, "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:43122 is the closest thing to a representative issue because it is a broad tokenizer regression report and overlaps thematically with several nearby tokenization-related complaints, even though it is not a duplicate of the others.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:42831", - "right": "issue:45753", - "accept": false, - "reason": "Different problems: FP8 accuracy regression vs Qwen3_5 image infinite loop." - }, - { - "left": "issue:43576", - "right": "issue:43784", - "accept": false, - "reason": "CLI/env command breakage is unrelated to sentence-transformers import NameError." - }, - { - "left": "issue:42886", - "right": "issue:43066", - "accept": false, - "reason": "Offline cache loading failure is separate from v5 decoder-type mismatch." - }, - { - "left": "issue:43122", - "right": "issue:44521", - "accept": false, - "reason": "Tokenizer regression vs multimodal assistant_mask generation bug; different code paths." - }, { "left": "issue:22355", "right": "issue:28218", "accept": false, - "reason": "Import error for transformers.onnx vs added-space tokenization issue." + "reason": "Different subsystems: ONNX import path vs tokenizer spacing after added token." }, { "left": "issue:44117", "right": "issue:44704", "accept": false, - "reason": "Mapping fallback bug is unrelated to AutoProcessor cached_file kwargs forwarding." + "reason": "Tokenizer mapping fallback bug vs AutoProcessor kwarg forwarding; unrelated code paths." }, { "left": "issue:42371", "right": "issue:44263", "accept": false, - "reason": "TF32 settings warning and torch.split behavior are unrelated." + "reason": "TF32 settings warning is unrelated to the torch.split return-value issue." }, { "left": "issue:43646", "right": "issue:46002", "accept": false, - "reason": "Custom model initialization breakage differs from NaNs in classification heads." + "reason": "Generic Transformers 5.0 custom-init breakage vs NaNs in classification heads; not the same failure." }, { "left": "issue:30990", "right": "issue:33357", "accept": false, - "reason": "Loading hang and Mac bus error are distinct failure modes." + "reason": "Both are loading/runtime problems, but one is SentenceTransformers stuck loading and the other is a Mac bus error on a CLIP model." }, { "left": "issue:44295", "right": "issue:44297", "accept": false, - "reason": "position_ids buffer access bug is unrelated to tokenizer_class saved in config." + "reason": "One concerns reading position_ids as a buffer, the other tokenizer_class saved in tokenizer_config.json." }, { "left": "issue:44704", "right": "issue:45290", "accept": false, - "reason": "Processor loading kwargs vs chat template crash with tool-call messages." + "reason": "Both touch chat/processor APIs only loosely; cached_file kwargs forwarding is not the same as assistant messages with tool calls crashing." }, { "left": "issue:43232", "right": "issue:44743", "accept": false, - "reason": "Generation kwargs update bug is different from recurrent-state reset in modular_qwen3_5." + "reason": "Generation kwarg update after sync_gpus is separate from recurrent state reset in modular_qwen3_5." }, { "left": "issue:42907", "right": "issue:44945", "accept": false, - "reason": "Model save failure is unrelated to incorrect pipeline-parallel output." + "reason": "Dequantized save failure for specific models vs incorrect output under pipeline parallelism." }, { "left": "issue:43122", "right": "issue:44514", "accept": false, - "reason": "General tokenization change vs batched Qwen2_5_VL chat-template crash." + "reason": "Tokenizer behavior change across versions is not the same as Qwen2_5_VLProcessor batched chat_template crash." }, { "left": "issue:41628", "right": "issue:45278", "accept": false, - "reason": "Single missing import is not the same as broad import errors after upgrade." + "reason": "Both are import-related, but one is a missing AutoImageProcessor symbol and the other is broad import regressions after upgrade." }, { "left": "issue:43976", "right": "issue:44945", "accept": false, - "reason": "Python version compatibility issue is unrelated to pipeline-parallel inference output." + "reason": "Python version compatibility issue is unrelated to pipeline parallelism output corruption." }, { "left": "issue:39401", "right": "issue:44488", "accept": false, - "reason": "Qwen3 offset_mapping bug and sleng-bert loading failure are different models/issues." + "reason": "Qwen3 offset_mapping bug and inability to load cjvt/sleng-bert are different tokenizer/model-loading issues." }, { - "left": "issue:43784", - "right": "issue:45600", + "left": "issue:44261", + "right": "issue:44373", "accept": false, - "reason": "Import-time NameError in sentence-transformers is unrelated to stale auto_mappings references." + "reason": "Precision error from missing rms_norm_eps is not the same as a wrong docstring for position_ids." }, { - "left": "issue:44261", - "right": "issue:44373", + "left": "issue:43784", + "right": "issue:45600", "accept": false, - "reason": "Layernorm precision/config bug is unrelated to a wrong docstring." + "reason": "Sentence-transformers import NameError and CI breakage from removed Sam3LiteText configs are unrelated." }, { "left": "issue:42907", "right": "issue:43937", "accept": false, - "reason": "Saving dequantized models and invalid GenerationConfig are unrelated." + "reason": "Both are model/config problems, but one is dequantized save_pretrained and the other is invalid GenerationConfig for GLM-5." }, { "left": "issue:43408", "right": "issue:43475", "accept": false, - "reason": "Model-type warning is not the same as missing fpn_position_embeddings attribute." + "reason": "Both mention SAM 3 video, but one is a wrong model-type warning and the other is a missing attribute on Sam3VisionEncoderOutput." }, { "left": "issue:44360", "right": "issue:44485", "accept": false, - "reason": "DSA indexer activation discussion is unrelated to GLM-5 RoPE implementation." + "reason": "DSA indexer ReLU concern and GLM-5 RoPE implementation are separate algorithmic issues." }, { "left": "issue:39401", "right": "issue:44521", "accept": false, - "reason": "Offset mapping bug and all-zero assistant masks are separate tokenizer/multimodal issues." + "reason": "Wrong offset_mapping for Qwen3 is different from all-zero assistant_masks in multimodal chat templating." }, { "left": "issue:43116", "right": "issue:45200", "accept": false, - "reason": "Example-script multi-label behavior is unrelated to Gemma 4 mm_token_type_ids defaults." + "reason": "Multi-label classification example bug is unrelated to Gemma 4 mm_token_type_ids defaults." }, { "left": "issue:44625", "right": "issue:44829", "accept": false, - "reason": "num_labels propagation bug does not match flash_attention_3 training degeneration." + "reason": "num_labels propagation in Qwen3.5 is unrelated to flash_attention_3 causing degenerate training." }, { "left": "issue:44038", "right": "issue:44297", "accept": false, - "reason": "Qwen3-VL-Moe model bug is unrelated to tokenizer save_pretrained class mismatch." + "reason": "Qwen3-VL-Moe bug and tokenizer_class mismatch in saved config are different components." }, { "left": "issue:42907", "right": "issue:44964", "accept": false, - "reason": "Saving dequantized Ministrals/Devstrals and loading Phi-4 multimodal are different failures." + "reason": "Dequantized save failure is unrelated to loading Phi-4 multimodal with latest transformers." }, { "left": "issue:29127", "right": "issue:30990", "accept": false, - "reason": "LayoutLMv3 error messaging and Sentence Transformers loading hang are unrelated." + "reason": "Clarity of a LayoutLMv3 error message vs SentenceTransformers loading hang are not the same bug." }, { "left": "issue:43122", "right": "issue:44779", "accept": false, - "reason": "Both concern tokenization, but they are different model/version regressions with different symptoms." + "reason": "Both are tokenizer regressions across versions, but they affect different models and failure modes." }, { "left": "issue:44038", "right": "issue:44295", "accept": false, - "reason": "Qwen3-VL-Moe bug is unrelated to position_ids buffer read error." + "reason": "Qwen3-VL-Moe bug and position_ids buffer read error are unrelated." + }, + { + "left": "issue:36296", + "right": "issue:43116", + "accept": false, + "reason": "Tensor parallel training bug and multi-label classification example bug do not share the same code path." }, { "left": "issue:42907", "right": "issue:43452", "accept": false, - "reason": "Dequantized model save failure and gguf_file load-path breakage are different code paths." + "reason": "Saving dequantized models and gguf_file breaking from_pretrained are separate loading/saving paths." + }, + { + "left": "issue:43824", + "right": "issue:44351", + "accept": false, + "reason": "Missing Qwen2_5_VLForConditionalGeneration import and missing HybridCache import are different API-export regressions." + }, + { + "left": "issue:44038", + "right": "issue:44222", + "accept": false, + "reason": "Qwen3-VL-Moe issue is not the same as FP8 save_pretrained behavior for moe models." } ] }, @@ -12544,8 +12540,9 @@ "nodes": [ "issue:22355", "issue:29127", - "issue:36296", + "issue:33453", "issue:38617", + "issue:39401", "issue:42222", "issue:42371", "issue:42673", @@ -12561,8 +12558,8 @@ "issue:43611", "issue:43673", "issue:43784", - "issue:43824", "issue:43825", + "issue:43827", "issue:43874", "issue:43937", "issue:43976", @@ -12591,9 +12588,6 @@ "issue:46002" ], "soft_pairs": [ - "issue:36296|issue:43116", - "issue:43824|issue:44351", - "issue:44038|issue:44222", "issue:45216|issue:45478", "issue:42757|issue:44351", "issue:43278|issue:44945", @@ -12619,210 +12613,213 @@ "issue:43116|issue:43825", "issue:45083|issue:45446", "issue:43937|issue:45563", - "issue:44117|issue:45588" + "issue:44117|issue:45588", + "issue:33453|issue:39401", + "issue:39401|issue:44779", + "issue:43827|issue:45446" ], - "prepared_review_unit_hash": "8c89a348e3f1981b1d845368f1b519ebb000b3b262b6e6cb8a60c759780eadf7", + "prepared_review_unit_hash": "9fddbf5d30f6cfe2871aaaa0f90cb178a98383a375d2ad3f4767451eff33c063", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 35394, - "estimated_input_tokens": 8849, - "estimated_eval_tokens": 17954 + "serialized_chars": 35924, + "estimated_input_tokens": 8981, + "estimated_eval_tokens": 18218 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: it groups many Transformers v5 regressions and model-specific bugs, but the soft edges are mostly broad thematic similarities rather than true duplicates. I do not see any pair that is clearly the same underlying bug or mergeable into one PR.", - "confidence": 0.87, - "canonical_issue_reason": "issue:45278 is the broadest umbrella-style report in the set ('Many import errors after update from 4.57.0 to 5.5.0'), so it is the most plausible representative issue for the cluster even though the cluster is not a true duplicate set.", + "summary": "This is a noisy cluster of mostly unrelated Transformers issues; the soft pairs share broad themes like v5 regressions or import/load failures, but none look like the same concrete bug. Issue 45478 is the most specific representative, but the set is not a true duplicate cluster.", + "confidence": 0.83, + "canonical_issue_reason": "issue:45478 is the clearest, most actionable single bug report in the set (a narrow Qwen3.5 Moe from_pretrained failure), so it is the best representative if a canonical issue must be chosen.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 best represents the cluster because it captures the widest version-regression/import-error pattern; the other issues are narrower, model-specific, or unrelated symptoms.", + "best_issue_reason": "issue:45478 has the strongest standalone triage value: specific model, specific failing path, and a plausible repro scope. Still, it is only a weak canonical candidate because the surrounding issues are largely unrelated.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:36296", - "right": "issue:43116", - "accept": false, - "reason": "Tensor-parallel training failure vs. multi-label example-script output bug; different code paths and symptoms." - }, - { - "left": "issue:43824", - "right": "issue:44351", - "accept": false, - "reason": "Both are import errors, but they involve different missing symbols/classes and different API surfaces." - }, - { - "left": "issue:44038", - "right": "issue:44222", - "accept": false, - "reason": "Qwen3-VL-Moe runtime bug vs. FP8 save_pretrained bug; different failure modes and likely different fixes." - }, { "left": "issue:45216", "right": "issue:45478", "accept": false, - "reason": "Both concern Qwen3.5 save/from_pretrained regressions, but one is about incorrect checkpoint saving and the other about a from_pretrained load error; not the same concrete bug." + "reason": "Both involve Qwen3.5 and save/load, but one is save_pretrained checkpoint contents and the other is from_pretrained loading failure; different concrete bugs." }, { "left": "issue:42757", "right": "issue:44351", "accept": false, - "reason": "Different missing imports from different modules; same broad class of breakage but not the same underlying issue." + "reason": "Both are import errors, but they are for different missing symbols from different packages and likely unrelated failure points." }, { "left": "issue:43278", "right": "issue:44945", "accept": false, - "reason": "Dtype change between train/evaluate vs. incorrect output under pipeline parallelism; unrelated behavior." + "reason": "Embedding dtype drift in eval is unrelated to incorrect LLM output under pipeline parallelism." }, { "left": "issue:43122", "right": "issue:44977", "accept": false, - "reason": "Tokenizer output drift across versions vs. flash-attention generation bug; too different to merge." + "reason": "Tokenizer output drift across versions is not the same as a flash-attention generation failure." }, { "left": "issue:43784", "right": "issue:45446", "accept": false, - "reason": "Import-time NameError in sentence-transformers integration vs. a PyTorch version check bug for AuxRequest; different root causes." + "reason": "Different import/version-check failures in different code paths; no shared underlying bug." }, { "left": "issue:43493", "right": "issue:45753", "accept": false, - "reason": "SigLIP2 implementation discrepancy vs. Qwen3.5 infinite loop on one image; both model-related but not the same bug." + "reason": "SigLIP2 implementation mismatch and a Qwen3.5 image infinite loop are different model-specific problems." }, { "left": "issue:42907", "right": "issue:44222", "accept": false, - "reason": "Saving dequantized Ministral/Devstral models vs. FP8 save_pretrained moe; both serialization-related but clearly different model families and mechanics." + "reason": "Both mention saving model states, but dequantized Ministral/Devstral saving and FP8 MoE save_pretrained are distinct bugs." }, { "left": "issue:43992", "right": "issue:44704", "accept": false, - "reason": "Missing embed_tokens.weight on UMT5Encoder load vs. AutoProcessor not forwarding kwargs; unrelated loader paths." + "reason": "Both are from_pretrained-related, but one misses a model weight and the other drops kwargs in AutoProcessor; different code paths." }, { "left": "issue:43066", "right": "issue:44779", "accept": false, - "reason": "Tokenizer decoder type bug vs. Deepseek tokenization regression; both tokenizer issues but not the same failure." + "reason": "Tokenizer decoder-type regression and DeepSeek tokenization corruption are separate tokenizer bugs." }, { "left": "issue:38617", "right": "issue:45278", "accept": false, - "reason": "Both are import errors, but one is a single missing symbol and the other is a broad version-upgrade report; not the same concrete bug." + "reason": "45278 is a broad bundle of import errors; 38617 is one specific missing symbol. Too generic to treat as the same bug." }, { "left": "issue:42673", "right": "issue:45072", "accept": false, - "reason": "VRAM leak in Qwen3ForCausalLM with multi-threaded dataloaders vs. dtype mismatch in CI inference; different runtime problems." + "reason": "VRAM leakage in threaded Qwen3ForCausalLM usage is unrelated to dtype mismatches in inference." }, { "left": "issue:43937", "right": "issue:44485", "accept": false, - "reason": "GenerationConfig validation error for GLM-5 vs. RoPE implementation discussion; not the same bug." + "reason": "GenerationConfig validation failure and GLM-5 RoPE discussion are not the same concrete defect." }, { "left": "issue:22355", "right": "issue:29127", "accept": false, - "reason": "Missing transformers.onnx module vs. LayoutLMv3 box-validation error message; unrelated." + "reason": "A missing module import and a LayoutLMv3 error-message clarity issue are unrelated." }, { "left": "issue:43611", "right": "issue:46002", "accept": false, - "reason": "base_model_prefix loading regression vs. NaNs in classification heads after checkout/init; different symptoms and code paths." + "reason": "Loading models with base_model_prefix and NaNs in classification heads are different failure modes." }, { "left": "issue:42371", "right": "issue:43874", "accept": false, - "reason": "TF32 behavior/settings issue vs. missing get_number_of_image_patches on an image processor; unrelated." + "reason": "TF32 settings guidance and a missing multimodal image-processor method are unrelated." }, { "left": "issue:43976", "right": "issue:45446", "accept": false, - "reason": "Python version compatibility problem vs. PyTorch version gate bug; both version-related but not the same issue." + "reason": "Python version compatibility and a PyTorch version check bug are different release/platform issues." }, { "left": "issue:42222", "right": "issue:43441", "accept": false, - "reason": "VitPose models broken vs. Ministral-3 FlashAttention failure; different models and failure mechanisms." + "reason": "Broken ViTpose models and Ministral FlashAttention failures are separate model-specific regressions." }, { "left": "issue:43064", "right": "issue:43673", "accept": false, - "reason": "Optimizer state corruption with FSDP2+PEFT+cpu_ram_efficient_loading vs. GenerationMixin cache missing in chunked_prefill; unrelated training vs. generation bugs." + "reason": "Distributed optimizer-state corruption during training is unrelated to missing generation cache in chunked prefill." }, { "left": "issue:43116", "right": "issue:43493", "accept": false, - "reason": "Example script multi-label output bug vs. SigLIP2 implementation discrepancy; no shared code path." + "reason": "Empty multi-label classification outputs and SigLIP2 JAX parity issues are unrelated." }, { "left": "issue:44075", "right": "issue:44936", "accept": false, - "reason": "SGD args ignored during optimization vs. evaluate() failing after train(); both trainer-related but different bugs." + "reason": "Unused SGD args and trainer.evaluate() failing after train() are different trainer bugs." }, { "left": "issue:45357", "right": "issue:45478", "accept": false, - "reason": "Qwen3.5 save_pretrained visual encoder key regression vs. Qwen3.5 Moe from_pretrained load error; closely related family, but not the same concrete bug." + "reason": "Both concern Qwen3.5 save/load behavior, but one is incorrect visual encoder keys on save and the other is a from_pretrained error; not the same bug." }, { "left": "issue:44038", "right": "issue:45478", "accept": false, - "reason": "Different Qwen3.5/Qwen3-VL-Moe failures with different entry points (general bug vs. load error)." + "reason": "Both involve Qwen3.x/Moe models, but the reported failures are different and not clearly the same code-path regression." }, { "left": "issue:43116", "right": "issue:43825", "accept": false, - "reason": "Example script classification output bug vs. pipeline translation-task error messaging; unrelated behavior." + "reason": "Example-script multi-label output and pipeline translation messaging are separate issues." }, { "left": "issue:45083", "right": "issue:45446", "accept": false, - "reason": "Helper output-length calculation in qwen3_omni_moe vs. PyTorch version check for AuxRequest; different layers." + "reason": "A helper length-calculation bug in qwen3_omni_moe is unrelated to a PyTorch version-check import problem." }, { "left": "issue:43937", "right": "issue:45563", "accept": false, - "reason": "GLM-5 GenerationConfig invalid vs. stale warning for paged generate(); unrelated generation issues." + "reason": "Invalid GenerationConfig and stale num_return_sequences warning are different generation-layer issues." }, { "left": "issue:44117", "right": "issue:45588", "accept": false, - "reason": "TOKENIZER_MAPPING_NAMES None handling vs. flash_attention crash on s_aux=None; different modules and failure modes." + "reason": "Tokenizer mapping None-handling and flash_attention s_aux=None crashes are unrelated." + }, + { + "left": "issue:33453", + "right": "issue:39401", + "accept": false, + "reason": "Tokenizer loading regression and Qwen3 tokenizer offset mapping bug are different tokenizer problems." + }, + { + "left": "issue:39401", + "right": "issue:44779", + "accept": false, + "reason": "Wrong offset_mapping and incorrect tokenization results are both tokenizer-related, but the concrete bugs and affected behavior differ." + }, + { + "left": "issue:43827", + "right": "issue:45446", + "accept": false, + "reason": "Docs still referencing pipeline() and an import/version-check bug are not the same underlying issue." } ] }, @@ -12846,7 +12843,6 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-404", "nodes": [ - "issue:33453", "issue:38617", "issue:39401", "issue:41628", @@ -12863,11 +12859,12 @@ "issue:43421", "issue:43450", "issue:43582", + "issue:43673", "issue:43742", - "issue:43827", "issue:43906", "issue:43937", "issue:43986", + "issue:44075", "issue:44117", "issue:44220", "issue:44261", @@ -12877,7 +12874,6 @@ "issue:44483", "issue:44484", "issue:44496", - "issue:44779", "issue:44829", "issue:44933", "issue:44977", @@ -12892,14 +12888,12 @@ "issue:45464", "issue:45542", "issue:45561", + "issue:45656", "issue:46002" ], "soft_pairs": [ - "issue:33453|issue:39401", - "issue:39401|issue:44779", - "issue:43827|issue:45446", - "issue:43937|issue:44977", "issue:44466|issue:46002", + "issue:43937|issue:44977", "issue:42673|issue:43906", "issue:45405|issue:45446", "issue:44291|issue:45071", @@ -12919,55 +12913,216 @@ "issue:42222|issue:45542", "issue:42490|issue:42673", "issue:42175|issue:42222", - "issue:38617|issue:44220", "issue:43937|issue:44496", - "issue:42371|issue:42831", + "issue:38617|issue:44220", "issue:44483|issue:45464", - "issue:43066|issue:43122" + "issue:42371|issue:42831", + "issue:43066|issue:43122", + "issue:44075|issue:45656", + "issue:43673|issue:45341" ], - "prepared_review_unit_hash": "6020e0895fcc683fdbec86cc2fbf67cb27e290e17b02c9006691d2bd787fc97f", + "prepared_review_unit_hash": "50d0b13290d23a48d6c78d322b2201c918755ec27402d802b5abd4139a3480a6", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 29, - "serialized_chars": 35620, - "estimated_input_tokens": 8905, - "estimated_eval_tokens": 18066 + "soft_pair_count": 28, + "serialized_chars": 35284, + "estimated_input_tokens": 8821, + "estimated_eval_tokens": 17898 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose set of distinct Transformers bug reports. The tokenizer-v5 issues are related at the theme level, but there isn\u2019t enough evidence here to treat issue:43066 and issue:43122 as the same concrete bug or a mergeable duplicate pair.", - "confidence": 0.47, - "canonical_issue_reason": "issue:43122 is the broadest tokenizer regression report in the cluster, framing the 4.57.3\u21925.0 behavior change generically and making it the best representative for the shared tokenizer-v5 topic.", + "summary": "This cluster is highly heterogeneous: it mixes import errors, tokenizer regressions, model-specific bugs, training/optimizer issues, and testing-utils problems. None of the soft-similarity pairs look like the same concrete underlying bug or change, so no duplicates should be merged here.", + "confidence": 0.97, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:43122 is the best cluster representative because it captures the cross-version tokenizer behavior change in the broadest terms, which is more useful than the narrower decoder-type symptom report.", + "best_issue_reason": "issue:43122 is the closest broad regression report in the set because it describes tokenizer behavior changing across versions, but the cluster is too mixed for it to serve as a true canonical issue.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44466", + "right": "issue:46002", + "accept": false, + "reason": "Different failure modes: tied-weight serialization/device-dependent save behavior vs NaNs during checkout/init." + }, + { + "left": "issue:43937", + "right": "issue:44977", + "accept": false, + "reason": "Unrelated model-specific generation issues: invalid GenerationConfig vs flash-attention generation failure." + }, + { + "left": "issue:42673", + "right": "issue:43906", + "accept": false, + "reason": "One is a VRAM leak in multithreaded dataloading; the other is just an isolated reproduction of a different issue." + }, + { + "left": "issue:45405", + "right": "issue:45446", + "accept": false, + "reason": "PEFT version pinning/release mismatch is unrelated to the flex_attention PyTorch version check bug." + }, + { + "left": "issue:44291", + "right": "issue:45071", + "accept": false, + "reason": "Different code paths: init_empty_weights argument handling vs PreTrainedConfig type-checking in v5.4." + }, + { + "left": "issue:43450", + "right": "issue:43986", + "accept": false, + "reason": "Batched video processor shape bug is not the same as the AutoProcessor crash when torchvision is missing." + }, + { + "left": "issue:45341", + "right": "issue:45561", + "accept": false, + "reason": "Both touch testing utilities, but one is a small utility bug and the other is a pytest-xdist race on a capture file." + }, + { + "left": "issue:44117", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer mapping returning None and Mistral regex patch crashing are separate tokenizer failures." + }, + { + "left": "issue:38617", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but for different symbols and modules with different root causes." + }, + { + "left": "issue:44410", + "right": "issue:45310", + "accept": false, + "reason": "Different model families and symptoms: missing QKV projections vs Qwen3.5 MoE from_pretrained failure." + }, + { + "left": "issue:39401", + "right": "issue:43421", + "accept": false, + "reason": "Wrong offset_mapping in Qwen3 tokenizer is a different bug from runtime post-processor updates for special tokens." + }, + { + "left": "issue:44829", + "right": "issue:45216", + "accept": false, + "reason": "Flash-attention training degeneration and incorrect save_pretrained serialization are distinct issues." + }, + { + "left": "issue:44484", + "right": "issue:45310", + "accept": false, + "reason": "A question about max_shard_size is unrelated to a Qwen3.5 MoE loading error." + }, + { + "left": "issue:42371", + "right": "issue:44261", + "accept": false, + "reason": "TF32 API settings and MLA rms_norm_eps precision error are different problems." + }, + { + "left": "issue:42673", + "right": "issue:43582", + "accept": false, + "reason": "VRAM leak in multi-threaded inference is unrelated to an Apple Silicon TypeError in caching_allocator_warmup." + }, + { + "left": "issue:42617", + "right": "issue:45341", + "accept": false, + "reason": "Running 3d_parallel.py and a testing_utils bug are unrelated." + }, + { + "left": "issue:43408", + "right": "issue:43742", + "accept": false, + "reason": "A model-type warning for sam3_tracker is not the same as a KeyError loading MobileLLM-125M." + }, + { + "left": "issue:44933", + "right": "issue:45278", + "accept": false, + "reason": "One is a specific nonexistent import; the other is a broad report of many import errors after upgrade." + }, + { + "left": "issue:42222", + "right": "issue:45542", + "accept": false, + "reason": "VitPose breakage and missing TensorFlow backend due to tensorboard-only install are unrelated." + }, + { + "left": "issue:42490", + "right": "issue:42673", + "accept": false, + "reason": "A path-with-slash import warning is not the same as a VRAM leak in threaded dataloading." + }, + { + "left": "issue:42175", + "right": "issue:42222", + "accept": false, + "reason": "TensorFlow backend packaging and VitPose model breakage are separate issues." + }, + { + "left": "issue:43937", + "right": "issue:44496", + "accept": false, + "reason": "Invalid GenerationConfig and unrecognized model/config.json problems are different failure classes." + }, + { + "left": "issue:38617", + "right": "issue:44220", + "accept": false, + "reason": "ImportError for layer_type_validation is unrelated to the _torch_extract_fbank_features issue." + }, + { + "left": "issue:44483", + "right": "issue:45464", + "accept": false, + "reason": "Both mention chat/completions, but one is a generic API request rejection and the other is a streaming inference failure on a specific model." + }, + { + "left": "issue:42371", + "right": "issue:42831", + "accept": false, + "reason": "TF32 behavior controls and FineGrainedFP8 accuracy problems do not share the same root cause." + }, { "left": "issue:43066", "right": "issue:43122", "accept": false, - "reason": "The titles indicate related tokenizer regressions in v5, but that is not enough to conservatively conclude the same underlying bug or one PR-fixable change without body-level evidence tying the failure modes together." + "reason": "Both are tokenizer regressions, but one is about decoder type in v5 and the other is about tokenization differences; not the same concrete bug." + }, + { + "left": "issue:44075", + "right": "issue:45656", + "accept": false, + "reason": "SGD args being ignored is unrelated to optimizer.step() being called twice under deepspeed." + }, + { + "left": "issue:43673", + "right": "issue:45341", + "accept": false, + "reason": "Generation cache missing during chunked_prefill and a testing_utils bug are unrelated." } ] }, - "evaluator_result": { - "accept": false, - "feedback": "The cluster summary is mostly grounded, but it overreaches on the lone accepted soft pair. From the packet we only have titles for issue:43066 and issue:43122, which show related tokenizer regressions in v5, but not enough evidence to conservatively mark them as the same bug/mergeable fix. The rest of the soft-edge refusals look appropriately cautious. I\u2019d keep the overall summary, but change the 43066\u219443122 verdict to a non-accept unless more body-level evidence ties them together." - }, - "evaluator_used": true, - "retried": true, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null @@ -12991,6 +13146,7 @@ "issue:42757", "issue:43012", "issue:43064", + "issue:43066", "issue:43097", "issue:43335", "issue:43408", @@ -12999,16 +13155,16 @@ "issue:43502", "issue:43531", "issue:43577", - "issue:43673", "issue:43873", "issue:43874", + "issue:43906", "issue:43931", "issue:43976", "issue:43986", "issue:43992", + "issue:44016", "issue:44038", "issue:44062", - "issue:44075", "issue:44117", "issue:44261", "issue:44263", @@ -13031,12 +13187,9 @@ "issue:45459", "issue:45478", "issue:45542", - "issue:45588", - "issue:45656" + "issue:45588" ], "soft_pairs": [ - "issue:44075|issue:45656", - "issue:43673|issue:45341", "issue:43012|issue:43577", "issue:44484|issue:45478", "issue:43874|issue:44261", @@ -13046,15 +13199,15 @@ "issue:41628|issue:45071", "issue:43012|issue:43408", "issue:42222|issue:45459", - "issue:43441|issue:44829", "issue:45071|issue:45542", + "issue:43441|issue:44829", "issue:43931|issue:44410", - "issue:43931|issue:44038", "issue:43335|issue:45071", + "issue:43931|issue:44038", "issue:45003|issue:45341", "issue:43097|issue:43452", - "issue:43986|issue:45341", "issue:44117|issue:44987", + "issue:43986|issue:45341", "issue:43986|issue:44484", "issue:43874|issue:44263", "issue:43408|issue:44368", @@ -13072,252 +13225,266 @@ "issue:44704|issue:45103", "issue:43502|issue:43992", "issue:43986|issue:44117", - "issue:43097|issue:43873" + "issue:43097|issue:43873", + "issue:43986|issue:44016", + "issue:43066|issue:43906" ], - "prepared_review_unit_hash": "b89557839026f3c5b47676b25e3d2719d2251367128402e8f240ec57d7bdc91a", + "prepared_review_unit_hash": "dffe142b16a27cfbc808589568ee4ccd6ee3ce5d591264ae90c881b749c23ec3", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 38, - "serialized_chars": 39104, - "estimated_input_tokens": 9776, - "estimated_eval_tokens": 19808 + "serialized_chars": 39163, + "estimated_input_tokens": 9791, + "estimated_eval_tokens": 19838 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a grab bag of unrelated Transformers issues: import errors, model-specific load/config bugs, optimizer/training regressions, and warning/behavior changes. None of the soft pairs look like the same underlying bug, so I would not consolidate them as duplicates.", - "confidence": 0.88, - "canonical_issue_reason": "Issue #43097 is the broadest, most umbrella-like regression report in the set (v5 breaking change, multiple references, general impact), so it is the best single representative if one must be chosen. That said, the cluster is not a true duplicate set.", + "summary": "This cluster is mostly heterogeneous: import errors, model-specific loading bugs, configuration regressions, and unrelated runtime warnings. The soft-similarity pairs all look superficially related but refer to different concrete code paths or failure modes, so they should stay separate.", + "confidence": 0.82, + "canonical_issue_reason": "issue:43097 is the closest thing to a canonical anchor because it describes a concrete, high-impact Transformers v5 API break with downstream references, and it best matches the broad version-regression theme among these otherwise mixed reports.", "canonical_pr_reason": null, - "best_issue_reason": "#43097 is the best global issue candidate because it is the most generally applicable and externally referenced report among otherwise disparate bugs. It is still only a loose representative, not a real duplicate target.", + "best_issue_reason": "There is no true single issue covering the whole set, but issue:43097 is the strongest triage anchor because it is specific, impactful, and broadly representative of the v5 breakage subgroup.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44075", - "right": "issue:45656", - "accept": false, - "reason": "Both are optimizer-related, but one is about SGD args being ignored while the other is a double optimizer.step with deepspeed; different code-paths and failure modes." - }, - { - "left": "issue:43673", - "right": "issue:45341", - "accept": false, - "reason": "Cache handling during generation/chunked prefill is unrelated to a testing_utils.py bug." - }, { "left": "issue:43012", "right": "issue:43577", "accept": false, - "reason": "A bfloat16 compilation warning and Blip2 dtype remaining float32 are different symptoms with different causes." + "reason": "Different bugs: one is a bfloat16 compilation warning, the other is a Blip2 dtype propagation issue." }, { "left": "issue:44484", "right": "issue:45478", "accept": false, - "reason": "save_pretrained max shard size discussion is unrelated to a Qwen3.5 MoE from_pretrained error." + "reason": "Unrelated: max_shard_size default behavior vs a Qwen3.5-Moe from_pretrained loading error." }, { "left": "issue:43874", "right": "issue:44261", "accept": false, - "reason": "Missing get_number_of_image_patches in an image processor and a missing rms_norm_eps config field are different model/component bugs." + "reason": "Different model/code paths: missing GLM46V image-patch API vs MLA q_a_layernorm precision mismatch." }, { "left": "issue:44410", "right": "issue:45478", "accept": false, - "reason": "These affect different Qwen variants and different failures: missing projections vs a load error in from_pretrained." + "reason": "Different Qwen families and failures: missing projections in Qwen3Next vs Qwen3.5-Moe loading error." }, { "left": "issue:43097", "right": "issue:44493", "accept": false, - "reason": "Removing tie_embeddings_and_encoder_decoder is a separate API change from unexpected position_id key warnings." + "reason": "Different regression classes: removed config flag vs unexpected position-id key handling." }, { "left": "issue:44263", "right": "issue:44485", "accept": false, - "reason": "torch.split return-value handling in GLM-MoE is unrelated to a GLM-5 RoPE implementation issue." + "reason": "Separate internals: torch.split return-value handling vs GLM-5 RoPE implementation." }, { "left": "issue:41628", "right": "issue:45071", "accept": false, - "reason": "Missing AutoImageProcessor import and PretrainedConfig type-checking breakage are different regressions." + "reason": "Different APIs: missing AutoImageProcessor import vs PretrainedConfig type checking breakage." }, { "left": "issue:43012", "right": "issue:43408", "accept": false, - "reason": "A precision warning during bfloat16 compilation is not the same as a sam3_video/sam3_tracker model-type mismatch warning." + "reason": "Both are warnings, but they come from different subsystems and model types with different causes." }, { "left": "issue:42222", "right": "issue:45459", "accept": false, - "reason": "VitPose model breakage and tokenizer protobuf error hiding are unrelated subsystems." - }, - { - "left": "issue:43441", - "right": "issue:44829", - "accept": false, - "reason": "Both mention flash attention, but one is a Ministral-3 load failure and the other is degenerate training behavior; different bugs." + "reason": "Unrelated: VitPose model breakage vs tokenizer protobuf exception masking." }, { "left": "issue:45071", "right": "issue:45542", "accept": false, - "reason": "A PretrainedConfig typing regression is unrelated to a tf-backend error caused by tensorboard-only installation." + "reason": "Config typing regression and TensorBoard/TensorFlow backend error are unrelated." }, { - "left": "issue:43931", - "right": "issue:44410", + "left": "issue:43441", + "right": "issue:44829", "accept": false, - "reason": "Qwen3-VL weight-shape mismatch and qwen3next missing projections are different model-loading bugs." + "reason": "Different FlashAttention failures: one is a model loading issue, the other is a degenerate training behavior." }, { "left": "issue:43931", - "right": "issue:44038", + "right": "issue:44410", "accept": false, - "reason": "These are different Qwen model families with different load failures, not the same defect." + "reason": "Different problems: Qwen3-VL weight shape mismatch vs missing Qwen3Next layer projections." }, { "left": "issue:43335", "right": "issue:45071", "accept": false, - "reason": "SwitchTransformers sparse-layer creation and config type-checking are unrelated regressions." + "reason": "SwitchTransformers sparse-layer construction bug and PretrainedConfig type checking breakage are unrelated." + }, + { + "left": "issue:43931", + "right": "issue:44038", + "accept": false, + "reason": "Different model families and symptoms: load-time weight mismatch vs Qwen3.5/Qwen3-VL-Moe regression." }, { "left": "issue:45003", "right": "issue:45341", "accept": false, - "reason": "Unsafe sys.modules access and a testing_utils.py bug do not share the same underlying issue." + "reason": "Unsafe sys.modules access is unrelated to the small testing_utils bug." }, { "left": "issue:43097", "right": "issue:43452", "accept": false, - "reason": "Tie-embeddings removal and gguf_file loading failures are separate API/load-path problems." + "reason": "Different load-path regressions: removed tie_embeddings flag vs gguf_file tokenizer/model loading breakage." }, { - "left": "issue:43986", - "right": "issue:45341", + "left": "issue:44117", + "right": "issue:44987", "accept": false, - "reason": "Missing torchvision support in AutoProcessor is unrelated to a testing_utils.py bug." + "reason": "Tokenizer mapping None is not the same bug as loading physical-intelligence/fast on v5.1.0." }, { - "left": "issue:44117", - "right": "issue:44987", + "left": "issue:43986", + "right": "issue:45341", "accept": false, - "reason": "A TOKENIZER_MAPPING_NAMES None-handling bug and a physical-intelligence/fast load failure are distinct tokenizer/model-loading issues." + "reason": "AutoProcessor/torchvision crash and testing_utils bug are unrelated." }, { "left": "issue:43986", "right": "issue:44484", "accept": false, - "reason": "These concern different areas entirely: video processor dependencies vs save_pretrained shard sizing." + "reason": "Different issues: missing torchvision dependency handling vs a save_pretrained shard-size question." }, { "left": "issue:43874", "right": "issue:44263", "accept": false, - "reason": "Missing image-patch counting and torch.split return values are different bugs in different GLM components." + "reason": "Both involve model internals, but one is a missing image-patch method and the other is torch.split output handling." }, { "left": "issue:43408", "right": "issue:44368", "accept": false, - "reason": "A sam3 model-type warning and a tie_word_embeddings warning in Qwen fine-tuning are unrelated." + "reason": "Different warnings from different configs: SAM3 model-type mismatch vs tie_word_embeddings warning." }, { "left": "issue:43976", "right": "issue:45071", "accept": false, - "reason": "Python version support breakage is unrelated to PretrainedConfig type checking." + "reason": "Python version compatibility and config type checking are unrelated." }, { "left": "issue:43986", "right": "issue:45372", "accept": false, - "reason": "Missing torchvision and missing ReasoningEffort imports are different dependency/import problems." + "reason": "Different dependency/import failures: no torchvision vs missing ReasoningEffort import." }, { "left": "issue:43531", "right": "issue:45341", "accept": false, - "reason": "A Qwen3-MoE sliding_window issue is unrelated to a testing_utils.py bug." + "reason": "Qwen3-MoE sliding_window bug and a testing_utils bug are unrelated." }, { "left": "issue:38617", "right": "issue:42757", "accept": false, - "reason": "Both are import errors, but they involve different missing symbols from different modules and likely different fixes." + "reason": "Both are import errors, but for different missing symbols and different modules, so not the same underlying bug." }, { "left": "issue:44261", "right": "issue:44485", "accept": false, - "reason": "A missing rms_norm_eps config field and GLM-5 RoPE implementation are not the same bug." + "reason": "Distinct model issues: rms_norm_eps precision error vs GLM-5 RoPE implementation." }, { "left": "issue:43064", "right": "issue:44315", "accept": false, - "reason": "Wrong optimizer states under FSDP2+PEFT and Liger Kernel not being applied with model_init are different training-stack problems." + "reason": "Different training bugs: wrong optimizer state partitioning in FSDP2+PEFT vs Liger Kernel not being applied with model_init." }, { "left": "issue:44410", "right": "issue:45375", "accept": false, - "reason": "Missing attn projections in qwen3next and a missing deepstack_visual_indexes config field are different model/config regressions." + "reason": "Different Qwen3.5-Moe issues: missing projections vs missing config field being dropped." }, { "left": "issue:43097", "right": "issue:45588", "accept": false, - "reason": "Tie-embeddings removal and a flash_attention crash on sink-less models are unrelated bugs." + "reason": "Removed tie_embeddings flag and a FlashAttention s_aux=None crash are unrelated." }, { "left": "issue:44062", "right": "issue:44117", "accept": false, - "reason": "AddedToken special-argument duplication and tokenizer mapping None are different tokenizer issues." + "reason": "Different tokenizer failures: AddedToken kwarg collision vs TOKENIZER_MAPPING_NAMES returning None." }, { "left": "issue:38617", "right": "issue:44351", "accept": false, - "reason": "layer_type_validation and HybridCache are different missing imports from different parts of the package." + "reason": "Both are import failures, but they concern different missing exports and different root causes." + }, + { + "left": "issue:44038", + "right": "issue:44410", + "accept": false, + "reason": "Different Qwen variants and different defects: general v5/Qwen3.5-Moe bug vs missing layer projections." + }, + { + "left": "issue:44704", + "right": "issue:45103", + "accept": false, + "reason": "AutoProcessor kwargs forwarding and auto_docstring annotation crashes are unrelated." + }, + { + "left": "issue:43502", + "right": "issue:43992", + "accept": false, + "reason": "Local-files-only network leakage and missing UMT5 encoder weights are separate loading bugs." + }, + { + "left": "issue:43986", + "right": "issue:44117", + "accept": false, + "reason": "No torchvision crash and tokenizer mapping None are unrelated failures." }, { - "left": "issue:44038", - "right": "issue:44410", + "left": "issue:43097", + "right": "issue:43873", "accept": false, - "reason": "These are different Qwen-family bugs with different symptoms; not one underlying defect." + "reason": "Different regressions: removed config option vs quantization/offloading not working as expected." }, { - "left": "issue:44704", - "right": "issue:45103", + "left": "issue:43986", + "right": "issue:44016", "accept": false, - "reason": "AutoProcessor kwargs forwarding and an auto_docstring annotations crash are unrelated." + "reason": "A runtime import/dependency crash is not the same as a notebook syntax error." }, { - "left": "issue:43502", - "right": "issue:43992", + "left": "issue:43066", + "right": "issue:43906", "accept": false, - "reason": "Unexpected network access with local_files_only and a missing embed_tokens.weight in UMT5Encoder are different issues." + "reason": "Wrong tokenizer decoder type and an isolated reproduction of another issue are not the same concrete bug." } ] }, @@ -13341,6 +13508,7 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-404", "nodes": [ + "issue:36010", "issue:36683", "issue:38617", "issue:42490", @@ -13350,7 +13518,6 @@ "issue:42994", "issue:43012", "issue:43064", - "issue:43066", "issue:43097", "issue:43208", "issue:43299", @@ -13368,12 +13535,11 @@ "issue:43825", "issue:43867", "issue:43901", - "issue:43906", + "issue:43937", "issue:43976", - "issue:43986", "issue:43992", - "issue:44016", "issue:44038", + "issue:44077", "issue:44222", "issue:44230", "issue:44368", @@ -13384,6 +13550,7 @@ "issue:44514", "issue:44655", "issue:44683", + "issue:45216", "issue:45290", "issue:45381", "issue:45478", @@ -13391,24 +13558,22 @@ "issue:45753" ], "soft_pairs": [ - "issue:43986|issue:44016", - "issue:43066|issue:43906", "issue:43097|issue:43901", "issue:43825|issue:44655", "issue:43992|issue:44484", "issue:43208|issue:43550", "issue:43012|issue:43550", - "issue:43352|issue:44380", "issue:42886|issue:43381", + "issue:43352|issue:44380", "issue:43012|issue:44368", "issue:42907|issue:44655", "issue:42907|issue:45753", "issue:44380|issue:45479", "issue:43352|issue:43976", "issue:43064|issue:44380", + "issue:44483|issue:44514", "issue:43208|issue:43526", "issue:36683|issue:43582", - "issue:44483|issue:44514", "issue:44222|issue:45478", "issue:42490|issue:43012", "issue:43618|issue:43761", @@ -13418,234 +13583,262 @@ "issue:38617|issue:43502", "issue:44038|issue:44230", "issue:44483|issue:45290", + "issue:44483|issue:45381", "issue:43299|issue:43867", "issue:43756|issue:44485", - "issue:44483|issue:45381", "issue:43901|issue:44655", "issue:43299|issue:45753", "issue:36683|issue:43475", - "issue:42994|issue:44655" + "issue:42994|issue:44655", + "issue:43352|issue:43937", + "issue:44077|issue:44380", + "issue:43867|issue:45216", + "issue:36010|issue:42757" ], - "prepared_review_unit_hash": "47bd11a5af625aaaacc642ff52e08c0dccd69ff5b573817862b6a03c4a6c61b5", + "prepared_review_unit_hash": "0e49f0a621bbc6498f82d8a4470f00678797e16c936b3ee2aa3c0bf97adb8f32", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 34, - "serialized_chars": 37648, - "estimated_input_tokens": 9412, - "estimated_eval_tokens": 19080 + "soft_pair_count": 36, + "serialized_chars": 38227, + "estimated_input_tokens": 9557, + "estimated_eval_tokens": 19370 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster looks like mostly unrelated Transformers bug reports that only share superficial regression/import-error wording. I don\u2019t see any soft pair that clearly describes the same underlying bug or change.", - "confidence": 0.84, - "canonical_issue_reason": "issue:44483 is the best single anchor if one must be chosen: it is a broad, user-facing regression with concrete API impact and clear discussion, while the other issues are narrower, model-specific, or unrelated.", + "summary": "This cluster is heterogeneous: most items are distinct Transformers regressions around imports, loading/saving, attention, or chat/template behavior. None of the soft pairs look like the same underlying bug, so all soft edges are rejected.", + "confidence": 0.24, + "canonical_issue_reason": "issue:43097 is the broadest representative here: a concrete v5 breaking-change regression with downstream references, while the other items are narrower or clearly unrelated symptoms.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44483 is the strongest representative issue in this set because it is well-scoped, high-impact, and more central than the rest; however, it is not a duplicate of the others.", + "best_issue_reason": "issue:43097 is the best umbrella issue in the set because it is a user-facing breaking regression with some cross-links, even though the cluster is too mixed to treat as a true duplicate group.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43986", - "right": "issue:44016", - "accept": false, - "reason": "Different problems: a video-model AutoProcessor crash vs a notebook syntax error." - }, - { - "left": "issue:43066", - "right": "issue:43906", - "accept": false, - "reason": "Different bugs: tokenizer decoder-type regression vs an isolated reproduction of another issue." - }, { "left": "issue:43097", "right": "issue:43901", "accept": false, - "reason": "Removed API behavior vs documentation mismatch for an unrelated pipeline flag." + "reason": "Removal of a config field in v5 is not the same bug as a docs/behavior mismatch for return_all_scores." }, { "left": "issue:43825", "right": "issue:44655", "accept": false, - "reason": "Different code paths: pipeline task messaging vs saving Pipeline objects." + "reason": "Chat-template task support is unrelated to saving Pipeline objects." }, { "left": "issue:43992", "right": "issue:44484", "accept": false, - "reason": "Model weight-loading bug vs save_pretrained shard-size default question." + "reason": "Missing embed_tokens.weight during load is unrelated to the save_pretrained shard-size default question." }, { "left": "issue:43208", "right": "issue:43550", "accept": false, - "reason": "Unrelated: xLSTM training issues vs Bamba torch.compile/SDPA failure." + "reason": "xLSTM training bugs are not the same as Bamba torch.compile/SDPA failures." }, { "left": "issue:43012", "right": "issue:43550", "accept": false, - "reason": "Precision warning during compile is not the same as the SDPA compile failure." + "reason": "A bfloat16 precision warning is a different issue from a compile-time SDPA failure." }, { - "left": "issue:43352", - "right": "issue:44380", + "left": "issue:42886", + "right": "issue:43381", "accept": false, - "reason": "Different concrete issues: unsupported Flash Attention 2 for Nemotron vs GPT2 attention scaling being ignored." + "reason": "Offline cache loading is unrelated to gradient checkpointing being disallowed in eval mode." }, { - "left": "issue:42886", - "right": "issue:43381", + "left": "issue:43352", + "right": "issue:44380", "accept": false, - "reason": "Offline tokenizer cache failure and eval-mode gradient checkpointing are unrelated." + "reason": "Nemotron FlashAttention 2 support is a different code path from GPT2 attention scaling being ignored." }, { "left": "issue:43012", "right": "issue:44368", "accept": false, - "reason": "A bfloat16 precision warning is not the same as the tie_word_embeddings warning." + "reason": "PyTorch float32 warning on bfloat16 compile is unrelated to a tie_word_embeddings config warning." }, { "left": "issue:42907", "right": "issue:44655", "accept": false, - "reason": "Saving dequantized models and saving Pipeline objects are different failure modes." + "reason": "Dequantized model save failures are not the same as Pipeline serialization failures." }, { "left": "issue:42907", "right": "issue:45753", "accept": false, - "reason": "Model saving bug vs an infinite-loop inference bug on a specific image." + "reason": "Saving dequantized models and an image-triggered infinite loop are unrelated bugs." }, { "left": "issue:44380", "right": "issue:45479", "accept": false, - "reason": "Attention backend behavior and sequence-classification loss degeneration are unrelated." + "reason": "Attention scaling being ignored is unrelated to degenerate zero loss in sequence classification." }, { "left": "issue:43352", "right": "issue:43976", "accept": false, - "reason": "Model-specific Flash Attention support issue vs Python version compatibility problem." + "reason": "FlashAttention support in a specific model is unrelated to the Python-version compatibility issue." }, { "left": "issue:43064", "right": "issue:44380", "accept": false, - "reason": "Distributed optimizer-state mismatch is unrelated to attention scaling under SDPA/FlashAttention." + "reason": "FSDP2 optimizer-state corruption is not the same as GPT2 attention scaling being ignored." + }, + { + "left": "issue:44483", + "right": "issue:44514", + "accept": false, + "reason": "Chat/completions request rejection is a server/API issue, not the same as apply_chat_template crashing on tool-call messages." }, { "left": "issue:43208", "right": "issue:43526", "accept": false, - "reason": "xLSTM training bugs vs BeitImageProcessorFast reduce_labels behavior." + "reason": "xLSTM training problems are unrelated to BeitImageProcessorFast reduce_labels behavior." }, { "left": "issue:36683", "right": "issue:43582", "accept": false, - "reason": "Gemma3Config missing vocab_size and Apple Silicon TypeError are unrelated." - }, - { - "left": "issue:44483", - "right": "issue:44514", - "accept": false, - "reason": "Chat/completions request rejection vs batched chat-template crash with tool calls." + "reason": "A missing config attribute is unrelated to an Apple Silicon allocator TypeError." }, { "left": "issue:44222", "right": "issue:45478", "accept": false, - "reason": "FP8 save_pretrained issue vs Qwen3.5 MoE from_pretrained loading error." + "reason": "FP8 save_pretrained issues and Qwen3.5 moe from_pretrained loading errors are different directions and different failures." }, { "left": "issue:42490", "right": "issue:43012", "accept": false, - "reason": "Trailing-slash local-path warning vs a compile-time bfloat16 warning." + "reason": "Trailing-slash import warnings are unrelated to bfloat16 compile warnings." }, { "left": "issue:43618", "right": "issue:43761", "accept": false, - "reason": "Both are CLIP regressions, but one is missing attentions and the other missing hidden_states; not the same bug." + "reason": "Both touch CLIP outputs, but one is missing attentions and the other is missing hidden_states; not the same bug." }, { "left": "issue:43012", "right": "issue:44683", "accept": false, - "reason": "A warning about precision is not the same as compiled flex_attention failing on torch >= 2.9." + "reason": "A warning during compile is not the same as flex_attention breaking on torch>=2.9." }, { "left": "issue:42757", "right": "issue:43824", "accept": false, - "reason": "Missing hub import and missing model-class import are different import errors." + "reason": "These are different missing imports from different modules and symbols." + }, + { + "left": "issue:43618", + "right": "issue:44380", + "accept": false, + "reason": "CLIP output fields and GPT2 attention scaling are different subsystems and symptoms." }, { "left": "issue:38617", "right": "issue:43502", "accept": false, - "reason": "Missing configuration import vs network requests despite local_files_only=True are unrelated." + "reason": "ImportError for layer_type_validation is unrelated to local_files_only still making API requests." }, { "left": "issue:44038", "right": "issue:44230", "accept": false, - "reason": "Qwen3-VL-Moe loading bug vs FP8 inference support request." + "reason": "Qwen3-VL-Moe loading regressions are not the same as FP8 inference support work." }, { "left": "issue:44483", "right": "issue:45290", "accept": false, - "reason": "Different chat-related failures: endpoint request rejection vs assistant tool-call template crash." + "reason": "Endpoint request validation and apply_chat_template crashes are separate code paths." + }, + { + "left": "issue:44483", + "right": "issue:45381", + "accept": false, + "reason": "Chat/completions request rejection is unrelated to Qwen2.5-VL video vision_position_ids being wrong." }, { "left": "issue:43299", "right": "issue:43867", "accept": false, - "reason": "Qwen3VL-Moe loading regression vs generic sorted-state_dict load error." + "reason": "Qwen3-VL-Moe loading problems are not the same as a sorted-state_dict load error." }, { "left": "issue:43756", "right": "issue:44485", "accept": false, - "reason": "Smollm3 RoPE-layer mismatch vs a GLM-5 RoPE implementation discussion; not the same issue." - }, - { - "left": "issue:44483", - "right": "issue:45381", - "accept": false, - "reason": "Chat/completions API rejection is unrelated to Qwen2.5-VL video position-id bugs." + "reason": "Smollm3 RoPE-layer omission is unrelated to a GLM-5 RoPE implementation discussion." }, { "left": "issue:43901", "right": "issue:44655", "accept": false, - "reason": "Documentation about return_all_scores vs saving Pipeline objects are different concerns." + "reason": "A docs mismatch about return_all_scores is unrelated to saving Pipeline objects." }, { "left": "issue:43299", "right": "issue:45753", "accept": false, - "reason": "Qwen3.5 MoE loading regression vs an image-specific infinite loop." + "reason": "Qwen3.5 VL/Moe loading regressions are unrelated to an image-specific infinite loop." }, { "left": "issue:36683", "right": "issue:43475", "accept": false, - "reason": "Gemma3Config missing vocab_size and SAM 3 Video missing fpn_position_embeddings are different model-specific bugs." + "reason": "Gemma3Config missing vocab_size and Sam3VisionEncoderOutput missing an attribute are unrelated model-specific bugs." + }, + { + "left": "issue:42994", + "right": "issue:44655", + "accept": false, + "reason": "Quantized model save failure is not the same as Pipeline.save_pretrained serialization." + }, + { + "left": "issue:43352", + "right": "issue:43937", + "accept": false, + "reason": "Unsupported FlashAttention in one model is unrelated to an invalid GenerationConfig in GLM-5." + }, + { + "left": "issue:44077", + "right": "issue:44380", + "accept": false, + "reason": "patchtsmixer post_init validation and GPT2 attention scaling are different issues." + }, + { + "left": "issue:43867", + "right": "issue:45216", + "accept": false, + "reason": "Both are load/save regression flavored, but one is sorted-state_dict loading while the other is a Qwen3.5 save_pretrained checkpoint regression; not the same bug." + }, + { + "left": "issue:36010", + "right": "issue:42757", + "accept": false, + "reason": "Different import errors for different symbols and packages; same surface pattern, different underlying issue." } ] }, @@ -13671,12 +13864,10 @@ "nodes": [ "issue:28282", "issue:33453", - "issue:36010", "issue:36296", "issue:36683", "issue:41669", "issue:42490", - "issue:42757", "issue:42890", "issue:42898", "issue:43010", @@ -13689,7 +13880,6 @@ "issue:43618", "issue:43867", "issue:43931", - "issue:43937", "issue:43976", "issue:43992", "issue:44075", @@ -13701,10 +13891,13 @@ "issue:44355", "issue:44360", "issue:44380", + "issue:44484", "issue:44492", "issue:44655", "issue:44855", + "issue:44857", "issue:44938", + "issue:44977", "issue:45071", "issue:45216", "issue:45278", @@ -13719,10 +13912,6 @@ "issue:45850" ], "soft_pairs": [ - "issue:43352|issue:43937", - "issue:44077|issue:44380", - "issue:43867|issue:45216", - "issue:36010|issue:42757", "issue:42898|issue:43066", "issue:36683|issue:44077", "issue:43611|issue:44655", @@ -13752,240 +13941,237 @@ "issue:42490|issue:44075", "issue:36296|issue:43575", "issue:28282|issue:33453", - "issue:44360|issue:45440" + "issue:44360|issue:45440", + "issue:43867|issue:44484", + "issue:43618|issue:44857", + "issue:43618|issue:44977" ], - "prepared_review_unit_hash": "cc79cace4c28c120ccbd3497f26ee409508297b8729ab6fa9a2726b97d229845", + "prepared_review_unit_hash": "a1d93bb1664f891ab918dc4ba022010c9f51da52c2036af41b2cde91aaa64800", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 34, - "serialized_chars": 37367, - "estimated_input_tokens": 9342, - "estimated_eval_tokens": 18940 + "soft_pair_count": 33, + "serialized_chars": 37072, + "estimated_input_tokens": 9268, + "estimated_eval_tokens": 18792 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a collection of unrelated Transformers regressions and import/load bugs rather than true duplicates. None of the soft pairs look safe to merge.", - "confidence": 0.86, - "canonical_issue_reason": "issue:45278 is the broadest umbrella-style report here, covering many import failures after a version upgrade, so it is the closest representative issue even though the cluster is heterogeneous.", + "summary": "This cluster is a loose grab-bag of Transformers regressions and performance issues, mostly model-specific loading/saving/config bugs rather than true duplicates. I found no pair that clearly describes the same underlying bug closely enough to merge as duplicates.", + "confidence": 0.31, + "canonical_issue_reason": "If a single anchor is needed, issue 45850 is the broadest umbrella report about widespread breakage in v5, but the cluster is too heterogeneous to justify a strong canonical duplicate choice.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the most general and representative issue in the set, but it should be treated as a loose umbrella rather than a duplicate target for the others.", + "best_issue_reason": "Issue 45850 is the best global triage anchor because it captures the widest user impact and can absorb related v5 breakage reports, even though most items here are not duplicates of it.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43352", - "right": "issue:43937", - "accept": false, - "reason": "Different bugs: Flash Attention support error vs invalid GenerationConfig validation." - }, - { - "left": "issue:44077", - "right": "issue:44380", - "accept": false, - "reason": "Unrelated optimizer-argument handling vs GPT-2 attention scaling bug." - }, - { - "left": "issue:43867", - "right": "issue:45216", - "accept": false, - "reason": "Different model-loading/serialization failures with no shared concrete code path." - }, - { - "left": "issue:36010", - "right": "issue:42757", - "accept": false, - "reason": "Different import-time errors for different symbols." - }, { "left": "issue:42898", "right": "issue:43066", "accept": false, - "reason": "Tokenizer cleanup behavior change vs wrong decoder type, not the same defect." + "reason": "Both are tokenizer-related, but one is a v5 behavior change and the other is a wrong decoder type; different bugs." }, { "left": "issue:36683", "right": "issue:44077", "accept": false, - "reason": "Gemma config attribute bug vs patchtsmixer post-init policy change." + "reason": "Unrelated model config attribute error vs a separate rule change for optional post_init." }, { "left": "issue:43611", "right": "issue:44655", "accept": false, - "reason": "Different serialization/loading problems in different APIs." + "reason": "Different code paths: base_model_prefix loading regression vs Pipeline save_pretrained support." }, { "left": "issue:43976", "right": "issue:44938", "accept": false, - "reason": "Python-version compatibility regressions, but on different versions and failure modes." + "reason": "Both are Python-version failures, but the reported failure modes and fixes are different." }, { "left": "issue:43352", "right": "issue:45593", "accept": false, - "reason": "Flash-Attention support error vs auxiliary-loss bug." + "reason": "Completely different models and subsystems: FlashAttention support vs auxiliary loss handling." }, { "left": "issue:36683", "right": "issue:43525", "accept": false, - "reason": "Different missing config attributes on different models." + "reason": "Same general pattern of missing config attributes, but for different models and different missing fields." }, { "left": "issue:43618", "right": "issue:45216", "accept": false, - "reason": "CLIP attentions regression vs Qwen save_pretrained regression." + "reason": "CLIP attentions regression and Qwen3.5 save_pretrained corruption are unrelated." }, { "left": "issue:44263", "right": "issue:45440", "accept": false, - "reason": "Different tensor-parallel/model-implementation bugs." + "reason": "Both involve MoE-style model code, but the concrete bugs and model implementations differ." }, { "left": "issue:36683", "right": "issue:44242", "accept": false, - "reason": "Different model/config issues, not the same load path." + "reason": "Different issues: missing Gemma3Config field vs load balancing loss not added." }, { "left": "issue:44188", "right": "issue:44380", "accept": false, - "reason": "torch.compile branching bug vs GPT-2 attention scaling bug." + "reason": "Both mention attention behavior under backend changes, but they affect different mechanisms and are not the same bug." }, { "left": "issue:43064", "right": "issue:44492", "accept": false, - "reason": "FSDP/PEFT optimizer-state bug vs cache strategy typo." + "reason": "Training/FSDP optimizer-state corruption is unrelated to a cache strategy typo." }, { "left": "issue:43352", "right": "issue:43931", "accept": false, - "reason": "Flash-Attention support error vs Qwen3-VL weight-shape mismatch." + "reason": "Both are Qwen/Nemotron model loading issues, but they are different models and different failure signatures." }, { "left": "issue:43992", "right": "issue:45310", "accept": false, - "reason": "Different from_pretrained failures on different models." + "reason": "Both are from_pretrained errors, but for different models and different missing/corrupted weights." }, { "left": "issue:43010", "right": "issue:45278", "accept": false, - "reason": "Cache no_grad bug vs broad version-wide import-breakage report." + "reason": "One is about no_grad decoration on cache updates; the other is a broad import regression after version updates." }, { "left": "issue:44355", "right": "issue:44855", "accept": false, - "reason": "Compiled-file loading issue vs Python 3.13 decorator parsing failure." + "reason": "Different root causes: compiled Python file errors vs a Python 3.13 torch.jit.script parsing problem." }, { "left": "issue:43867", "right": "issue:45375", "accept": false, - "reason": "State_dict ordering bug vs missing visual-index config field." + "reason": "State-dict loading error and a missing config field are not the same underlying bug." }, { "left": "issue:41669", "right": "issue:44246", "accept": false, - "reason": "Import-time performance regression vs xdist file race." + "reason": "Both concern import performance, but one is about import * overhead and the other about intermittent import slowness." }, { "left": "issue:36683", "right": "issue:43064", "accept": false, - "reason": "Different issues about model config attributes vs distributed optimizer state." + "reason": "Missing config attribute vs FSDP2/PEFT optimizer-state corruption; no shared code-path bug." }, { "left": "issue:42890", "right": "issue:45561", "accept": false, - "reason": "Test seeding flakiness vs xdist captured-info file race." + "reason": "Both are test flakiness, but one is seed-related nondeterminism and the other is an xdist file race." }, { "left": "issue:44938", "right": "issue:45071", "accept": false, - "reason": "Different Python-version regressions affecting different code paths." + "reason": "Different version/regression topics with different failure modes: import failure in Python 3.14 vs config type-checking breakage." }, { "left": "issue:43867", "right": "issue:45357", "accept": false, - "reason": "State_dict ordering bug vs Qwen3.5 save_pretrained key regression." + "reason": "One is a sorted state_dict load error; the other is incorrect visual encoder keys on save_pretrained." }, { "left": "issue:43010", "right": "issue:45850", "accept": false, - "reason": "Specific cache update bug vs vague broad breakage complaint." + "reason": "A cache no_grad concern is not the same bug as a broad 'many models break' regression report." }, { "left": "issue:43992", "right": "issue:45478", "accept": false, - "reason": "Different Qwen3.5/Qwen2.5 load failures, not the same issue." + "reason": "Different models and likely different weight-mapping/loading failures." }, { "left": "issue:43010", "right": "issue:43976", "accept": false, - "reason": "Cache no_grad bug vs Python 3.10 compatibility regression." + "reason": "Unrelated: cache update semantics vs Python-version compatibility." }, { "left": "issue:44938", "right": "issue:45381", "accept": false, - "reason": "Python 3.14 load failure vs Qwen2.5-VL video-position-id bug." + "reason": "Different model families and different bugs; both are release regressions only at a very high level." }, { "left": "issue:43618", "right": "issue:44077", "accept": false, - "reason": "CLIP output regression vs patchtsmixer post-init gating; unrelated." + "reason": "CLIP attentions assignment and patchtsmixer post_init policy are unrelated." }, { "left": "issue:42490", "right": "issue:44075", "accept": false, - "reason": "Trailing-slash import warning vs SGD optimizer args not being used." + "reason": "Local-path import warning vs optimizer argument handling are unrelated." }, { "left": "issue:36296", "right": "issue:43575", "accept": false, - "reason": "Both touch tensor parallelism, but one is a training bug and the other is an OOM on model load." + "reason": "Both mention tensor parallelism, but one is a training bug and the other is an OOM during model loading." }, { "left": "issue:28282", "right": "issue:33453", "accept": false, - "reason": "PyTorch-missing import error vs tokenizer loading regression." + "reason": "Completely different problems: missing PyTorch dependency vs tokenizer loading regression." }, { "left": "issue:44360", "right": "issue:45440", "accept": false, - "reason": "DSA indexer ReLU absence vs DeepSeekV3MoE implementation divergence." + "reason": "Both touch model internals, but the DSA indexer ReLU issue and DeepSeekV3 implementation divergence are different bugs." + }, + { + "left": "issue:43867", + "right": "issue:44484", + "accept": false, + "reason": "State-dict loading failure and max_shard_size behavior are unrelated." + }, + { + "left": "issue:43618", + "right": "issue:44857", + "accept": false, + "reason": "Different models and symptoms; no shared code-path problem." + }, + { + "left": "issue:43618", + "right": "issue:44977", + "accept": false, + "reason": "Flash-attention generation bug in Qwen3.5 is unrelated to CLIPOutput attentions assignment." } ] }, @@ -14027,11 +14213,10 @@ "issue:43299", "issue:43502", "issue:43576", - "issue:43618", + "issue:43673", "issue:43723", "issue:43824", "issue:43827", - "issue:43867", "issue:43874", "issue:44062", "issue:44112", @@ -14042,26 +14227,24 @@ "issue:44355", "issue:44360", "issue:44387", - "issue:44484", "issue:44485", + "issue:44492", "issue:44556", "issue:44655", "issue:44749", - "issue:44857", + "issue:44908", + "issue:44938", "issue:44961", "issue:44964", - "issue:44977", "issue:45071", "issue:45125", "issue:45341", "issue:45375", + "issue:45468", "issue:45478", "issue:45600" ], "soft_pairs": [ - "issue:43867|issue:44484", - "issue:43618|issue:44857", - "issue:43618|issue:44977", "issue:42890|issue:44964", "issue:42890|issue:44112", "issue:42890|issue:45600", @@ -14070,8 +14253,8 @@ "issue:44222|issue:45125", "issue:42890|issue:45341", "issue:44387|issue:44556", - "issue:43010|issue:43065", "issue:42548|issue:44351", + "issue:43010|issue:43065", "issue:43502|issue:44485", "issue:44355|issue:44749", "issue:45125|issue:45478", @@ -14079,9 +14262,9 @@ "issue:30990|issue:31795", "issue:45125|issue:45375", "issue:30333|issue:33453", - "issue:43299|issue:44556", - "issue:43502|issue:44222", "issue:42831|issue:43874", + "issue:43502|issue:44222", + "issue:43299|issue:44556", "issue:44315|issue:44360", "issue:43502|issue:44961", "issue:44355|issue:45071", @@ -14095,265 +14278,262 @@ "issue:43576|issue:45071", "issue:43576|issue:43824", "issue:31795|issue:38617", - "issue:41628|issue:42548" + "issue:41628|issue:42548", + "issue:44961|issue:45468", + "issue:38617|issue:43576", + "issue:43576|issue:44938", + "issue:43673|issue:44492", + "issue:44908|issue:44961", + "issue:44485|issue:44961" ], - "prepared_review_unit_hash": "1a200e183498bece84c061ee41cb00dfd23599c09c81eae439d5b79ac4ab0f4f", + "prepared_review_unit_hash": "fa21192ad3c50af8bdf97cf98b3d18bbc7252af863ef2633cdccf21edb554d5f", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 37, - "serialized_chars": 38135, - "estimated_input_tokens": 9534, - "estimated_eval_tokens": 19324 + "soft_pair_count": 40, + "serialized_chars": 39025, + "estimated_input_tokens": 9757, + "estimated_eval_tokens": 19770 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly a mix of unrelated Transformers issues that share a loose v5/regression flavor (loading, imports, tensor parallelism, FP8, tests), but they do not look like true duplicates of one underlying bug or change. No PRs are present. I selected one broad loading-regression issue as the best cluster representative, but the overall cluster is heterogeneous.", - "confidence": 0.62, - "canonical_issue_reason": "issue:44556 is the broadest representative of the recurring v5 loading/regression theme, covering a generic checkpoint reload failure after upgrading versions.", + "summary": "This cluster is mostly a grab bag of unrelated Transformers issues (imports, model loading, flaky tests, generation, FP8, tensor parallelism). There is no strong duplicate cluster overall; only one soft pair looks plausibly like the same Qwen3.5 MoE loading bug.", + "confidence": 0.91, + "canonical_issue_reason": "No single canonical issue fits the whole set: the reports span different subsystems and failure modes, so collapsing them into one tracker would be misleading.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44556 best fits the cluster header because it is a general upgrade/reload breakage and is closer to the common theme than the narrower import/test/model-specific reports.", + "best_issue_reason": "Issue 44556 is the closest broad representative of the repeated v5 loading-regression theme, but it still does not cover most of the other issues here.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43867", - "right": "issue:44484", - "accept": false, - "reason": "Different bugs: a state_dict ordering load failure versus a question about save_pretrained shard size." - }, - { - "left": "issue:43618", - "right": "issue:44857", - "accept": false, - "reason": "Unrelated symptoms: CLIPOutput attentions regression versus a float16 AMP crash in LwDetrImageLoss." - }, - { - "left": "issue:43618", - "right": "issue:44977", - "accept": false, - "reason": "Different model paths and failures: CLIP output fields versus Qwen3.5 flash-attention generation." - }, { "left": "issue:42890", "right": "issue:44964", "accept": false, - "reason": "Test flakiness from missing set_seed is unrelated to a Phi-4 multimodal model-loading failure." + "reason": "SAM HQ test flakiness from missing seeding is unrelated to a Phi-4 multimodal loading failure." }, { "left": "issue:42890", "right": "issue:44112", "accept": false, - "reason": "One is a CI test seeding issue; the other is a stale device override test in GraniteSpeech." + "reason": "Both mention CI/test failures, but they involve different models and different root causes." }, { "left": "issue:42890", "right": "issue:45600", "accept": false, - "reason": "Different test-related problems: missing set_seed versus broken auto_mappings entries for removed configs." + "reason": "Flaky SAM HQ integration tests and stale auto_mappings config references are different bugs." }, { "left": "issue:43502", "right": "issue:44062", "accept": false, - "reason": "Local-files-only network leakage is unrelated to the AddedToken special-argument collision." + "reason": "Local-files-only network leakage and AddedToken keyword collisions are unrelated code paths." }, { "left": "issue:43723", "right": "issue:44556", "accept": false, - "reason": "Both are loading regressions, but one is tokenizer loading and the other is generic checkpoint reload after version upgrade; not the same concrete bug." + "reason": "Both are loading-related, but one is tokenizer loading and the other is checkpoint reload compatibility; too different to merge." }, { "left": "issue:44222", "right": "issue:45125", "accept": false, - "reason": "Different FP8/tensor-parallel areas: save_pretrained support versus missing _tp_plan for TP inference." + "reason": "FP8 save_pretrained behavior and missing tensor-parallel plan metadata are distinct problems." }, { "left": "issue:42890", "right": "issue:45341", "accept": false, - "reason": "A SamHQ integration-test seeding problem is not the same as a general bug in testing_utils.py." + "reason": "A seed-related flaky test and a generic testing_utils bug are not clearly the same defect." }, { "left": "issue:44387", "right": "issue:44556", "accept": false, - "reason": "Int4 CUDA memory growth is a performance/OOM regression, not a checkpoint reload compatibility issue." + "reason": "int4 quantization memory growth causing OOM is not the same as checkpoint reload failure after upgrade." }, { - "left": "issue:43010", - "right": "issue:43065", + "left": "issue:42548", + "right": "issue:44351", "accept": false, - "reason": "Different code paths: no_grad decoration for cache/layer updates versus a dummy Conv2d choice in Sam3PixelDecoder." + "reason": "These are different missing top-level imports for different symbols." }, { - "left": "issue:42548", - "right": "issue:44351", + "left": "issue:43010", + "right": "issue:43065", "accept": false, - "reason": "Both are import errors, but they concern different missing exports and do not look like one concrete bug." + "reason": "Cache update no_grad semantics and a dummy Conv2d in Sam3PixelDecoder are unrelated." }, { "left": "issue:43502", "right": "issue:44485", "accept": false, - "reason": "Network access with local_files_only is unrelated to the GLM-5 RoPE implementation discussion." + "reason": "Unexpected API requests with local_files_only=True is unrelated to GLM-5 RoPE implementation." }, { "left": "issue:44355", "right": "issue:44749", "accept": false, - "reason": "Compiled-file errors are unrelated to the reported data-filtering slowdown after upgrading Transformers." + "reason": "Compiled Python file errors and a data-filtering slowdown are not the same underlying bug." }, { "left": "issue:45125", "right": "issue:45478", - "accept": false, - "reason": "Missing tensor-parallel metadata is different from the Qwen3.5 Moe from_pretrained regression." + "accept": true, + "reason": "Both point to the same Qwen3.5 MoE loading path in v5.4.0; missing _tp_plan metadata plausibly explains the from_pretrained failure." }, { "left": "issue:43010", "right": "issue:43232", "accept": false, - "reason": "Different generation bugs: no_grad on update methods versus post-sync_gpus _update_model_kwargs_for_generation behavior." + "reason": "These are different generation/cache issues with different triggers and code paths." }, { "left": "issue:30990", "right": "issue:31795", "accept": false, - "reason": "SentenceTransformers loading hang is unrelated to a documentation wording issue in model.forward." + "reason": "Sentence Transformers getting stuck loading and a documentation clarification in model.forward are unrelated." }, { "left": "issue:45125", "right": "issue:45375", "accept": false, - "reason": "Tensor-parallel support and a missing deepstack_visual_indexes config field are different issues." + "reason": "Same model family, but a missing tensor-parallel plan and a missing config field are different bugs." }, { "left": "issue:30333", "right": "issue:33453", "accept": false, - "reason": "MLflow job-status reporting failure and tokenizer loading regression are unrelated." + "reason": "MLFlow failure reporting and tokenizer loading regression are unrelated." }, { - "left": "issue:43299", - "right": "issue:44556", + "left": "issue:42831", + "right": "issue:43874", "accept": false, - "reason": "Both mention v5 loading breakage, but they affect different model families and likely different root causes." + "reason": "FP8 accuracy degradation and a missing image-patch count method are different failures." }, { "left": "issue:43502", "right": "issue:44222", "accept": false, - "reason": "Network calls with local_files_only do not match FP8 save_pretrained support for MoE models." + "reason": "Local-files-only network leakage and FP8 save_pretrained behavior do not share a concrete code path." }, { - "left": "issue:42831", - "right": "issue:43874", + "left": "issue:43299", + "right": "issue:44556", "accept": false, - "reason": "An FP8 accuracy issue is unrelated to a missing image-processor method causing token-count errors." + "reason": "Both are version-related loading regressions, but for different models and different failure modes." }, { "left": "issue:44315", "right": "issue:44360", "accept": false, - "reason": "Liger Kernel application in model_init is unrelated to the DSA indexer ReLU discussion." + "reason": "Applying Liger Kernel via model_init and a missing ReLU in the DSA indexer are unrelated." }, { "left": "issue:43502", "right": "issue:44961", "accept": false, - "reason": "A filesystem/network policy bug is unrelated to the nonspecific 'racoon' issue." - }, - { - "left": "issue:44355", - "right": "issue:45071", - "accept": false, - "reason": "Compiled-file errors and PretrainedConfig type-checking regressions are different problems." - }, - { - "left": "issue:30333", - "right": "issue:36296", - "accept": false, - "reason": "MLflow reporting on failed training jobs is unrelated to tensor parallel training behavior." - }, - { - "left": "issue:44355", - "right": "issue:44655", - "accept": false, - "reason": "Compiled Python file errors do not match the inability to save Pipeline objects." + "reason": "A network access bug and a placeholder-named issue are clearly unrelated." }, { "left": "issue:43827", "right": "issue:44961", "accept": false, - "reason": "Docs still referencing pipeline() are unrelated to the 'racoon' issue." + "reason": "Documentation references to pipeline() and the racoon issue are unrelated." }, { "left": "issue:44222", "right": "issue:44230", "accept": false, - "reason": "Both are FP8-related, but one is save_pretrained support and the other is inference support for Qwen3-VL/Qwen3.5 MoE." + "reason": "FP8 save_pretrained for MoE and FP8 inference support for Qwen3-VL/Qwen3.5 are adjacent but not the same bug." }, { "left": "issue:33453", "right": "issue:34689", "accept": false, - "reason": "Tokenizer-loading regression and Llama 3.2 multimodal model-loading failure are different concrete bugs." + "reason": "Tokenizer loading regression and Llama 3.2 vision model loading regression are different issues." }, { "left": "issue:42757", "right": "issue:43576", "accept": false, - "reason": "Missing is_offline_mode import from huggingface_hub is unrelated to the broken transformers env command." + "reason": "Missing is_offline_mode import and a broken transformers env CLI are separate problems." }, { "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 a tensor-parallel training bug are unrelated." }, { "left": "issue:43576", "right": "issue:45071", "accept": false, - "reason": "A broken CLI/env command is unrelated to PretrainedConfig type checking." + "reason": "Broken env CLI behavior and PretrainedConfig type-checking regression are different code paths." }, { "left": "issue:43576", "right": "issue:43824", "accept": false, - "reason": "The env command failure and the Qwen2_5_VL import error are different regressions." + "reason": "The env command bug and a Qwen2_5_VL import error are unrelated." }, { "left": "issue:31795", "right": "issue:38617", "accept": false, - "reason": "Documentation confusion in model.forward is unrelated to the layer_type_validation import error." + "reason": "Documentation confusion in model.forward and an ImportError for layer_type_validation are unrelated." }, { - "left": "issue:41628", - "right": "issue:42548", + "left": "issue:44961", + "right": "issue:45468", + "accept": false, + "reason": "These titles are unrelated and provide no evidence of the same underlying bug." + }, + { + "left": "issue:38617", + "right": "issue:43576", + "accept": false, + "reason": "A missing import from configuration_utils and a broken env CLI are different issues." + }, + { + "left": "issue:43576", + "right": "issue:44938", + "accept": false, + "reason": "Broken CLI behavior in v5 and Python 3.14 load failure are not the same defect." + }, + { + "left": "issue:43673", + "right": "issue:44492", + "accept": false, + "reason": "Missing cache in chunked_prefill generation and a cache-strategy typo are unrelated." + }, + { + "left": "issue:44908", + "right": "issue:44961", "accept": false, - "reason": "Both are import failures, but they concern different classes and do not appear to be the same underlying breakage." + "reason": "inverse_sqrt scheduler kwargs handling and the racoon issue are unrelated." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet and stays conservative. It correctly characterizes the set as mostly unrelated Transformers issues while acknowledging the one plausible duplicate pair. The accepted soft edge for issue:45125 and issue:45478 is supported by their titles and the shared Qwen3.5 MoE loading path. I did not see overstatement or an unsupported canonical merge." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -14370,8 +14550,9 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-404", "nodes": [ + "issue:31795", + "issue:33453", "issue:36296", - "issue:38617", "issue:41628", "issue:41669", "issue:42371", @@ -14383,9 +14564,11 @@ "issue:43572", "issue:43576", "issue:43673", + "issue:43827", "issue:43901", "issue:43976", "issue:44038", + "issue:44077", "issue:44246", "issue:44261", "issue:44273", @@ -14396,15 +14579,13 @@ "issue:44355", "issue:44360", "issue:44393", - "issue:44485", "issue:44492", "issue:44556", "issue:44623", "issue:44704", + "issue:44861", "issue:44908", - "issue:44938", "issue:44945", - "issue:44961", "issue:45003", "issue:45071", "issue:45092", @@ -14413,19 +14594,12 @@ "issue:45278", "issue:45341", "issue:45362", - "issue:45468", "issue:45478", "issue:45542", "issue:45753", "issue:45850" ], "soft_pairs": [ - "issue:44961|issue:45468", - "issue:38617|issue:43576", - "issue:43576|issue:44938", - "issue:43673|issue:44492", - "issue:44908|issue:44961", - "issue:44485|issue:44961", "issue:36296|issue:43901", "issue:41628|issue:44355", "issue:36296|issue:43526", @@ -14438,8 +14612,8 @@ "issue:44038|issue:45125", "issue:44393|issue:45071", "issue:43576|issue:44556", - "issue:43525|issue:45850", "issue:41669|issue:43673", + "issue:43525|issue:45850", "issue:44556|issue:45092", "issue:44492|issue:45850", "issue:43572|issue:45850", @@ -14459,264 +14633,262 @@ "issue:41669|issue:45341", "issue:44492|issue:44945", "issue:43976|issue:45850", - "issue:44393|issue:45478" + "issue:44393|issue:45478", + "issue:44077|issue:44861", + "issue:43827|issue:44908", + "issue:31795|issue:33453", + "issue:41669|issue:44908" ], - "prepared_review_unit_hash": "47e7b70cc6d9894855886dc6f6b620e5ab4fb43f648aa735854a1b84357c1621", + "prepared_review_unit_hash": "026da175eec397a2f34f10e82c2d078d77ddc9730be5c8405e4f0ce29f3eb640", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 40, - "serialized_chars": 38856, - "estimated_input_tokens": 9714, - "estimated_eval_tokens": 19684 + "soft_pair_count": 38, + "serialized_chars": 38367, + "estimated_input_tokens": 9592, + "estimated_eval_tokens": 19440 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is mostly a grab bag of unrelated Transformers issues. A few cluster around v5 upgrade regressions, import/loading failures, or model-specific breakages, but none of the soft pairs look like the same concrete bug or a safe duplicate merge.", - "confidence": 0.66, - "canonical_issue_reason": "Issue 45278 is the best umbrella representative: it clearly captures the broad v5 upgrade regression theme around import failures, which recurs across several nearby reports more cleanly than the highly model-specific or vague issues.", + "summary": "This cluster is mostly a grab bag of v5 migration regressions, import/load failures, and model-specific bugs. None of the soft pairs look like the same concrete underlying defect, so I would not collapse any of them as duplicates.", + "confidence": 0.76, + "canonical_issue_reason": "issue:45850 is the broadest umbrella-style report in the set, explicitly describing many model breakages after v5; it is the best cluster anchor despite being vague.", "canonical_pr_reason": null, - "best_issue_reason": "45278 is the most representative issue for triage because it describes the shared user-visible symptom at the right abstraction level and is specific enough to anchor the broader v5 compatibility/import-breakage theme.", + "best_issue_reason": "issue:45850 is the most representative issue for the overall v5-regression theme, though it is too general to be a strong canonical on its own.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44961", - "right": "issue:45468", - "accept": false, - "reason": "Unrelated: one is a nonsensical titled issue, the other is a Gemma-4 audio positional encoding bug; no shared code path." - }, - { - "left": "issue:38617", - "right": "issue:43576", - "accept": false, - "reason": "Both mention v5 breakage, but one is a specific ImportError and the other is a broken CLI command; too different to be the same bug." - }, - { - "left": "issue:43576", - "right": "issue:44938", - "accept": false, - "reason": "Different failures: broken env command vs Python 3.14 import/load failure; no evidence of one underlying defect." - }, - { - "left": "issue:43673", - "right": "issue:44492", - "accept": false, - "reason": "Cache-related wording is superficial only; one is a GenerationMixin cache regression, the other is a typo in cache strategies." - }, - { - "left": "issue:44908", - "right": "issue:44961", - "accept": false, - "reason": "Completely unrelated scheduler kwargs bug vs nonsensical titled issue." - }, - { - "left": "issue:44485", - "right": "issue:44961", - "accept": false, - "reason": "No overlap in subsystem or failure mode; RoPE implementation discussion is unrelated to the other issue." - }, { "left": "issue:36296", "right": "issue:43901", "accept": false, - "reason": "Tensor parallel training bug and pipeline docs mismatch are different areas and different bug classes." + "reason": "Tensor-parallel training bug vs. pipeline docs wording; different problem and code path." }, { "left": "issue:41628", "right": "issue:44355", "accept": false, - "reason": "Both are import/loading complaints, but one is a missing export and the other is about running compiled Python files; not the same defect." + "reason": "Specific missing import symbol vs. compiled-file execution error; unrelated failure modes." }, { "left": "issue:36296", "right": "issue:43526", "accept": false, - "reason": "Tensor parallel training and BeitImageProcessorFast label reduction are unrelated." + "reason": "Tensor-parallel training issue vs. BeitImageProcessorFast label reduction bug; not the same defect." }, { "left": "issue:45362", "right": "issue:45850", "accept": false, - "reason": "One is a specific chat crash for Qwen3.5-35B; the other is a broad complaint that many models break in v5. Not the same bug." + "reason": "A specific Qwen3.5 chat crash vs. a vague broad complaint about v5; not a shared concrete bug." }, { "left": "issue:44261", "right": "issue:44360", "accept": false, - "reason": "Different model/math bugs: RMS norm epsilon precision issue vs DSA indexer ReLU omission." + "reason": "Layernorm precision/config mismatch vs. DSA indexer ReLU omission; unrelated." }, { "left": "issue:44623", "right": "issue:45230", "accept": false, - "reason": "Too generic to match; missing processor save files is a concrete serialization bug, while the other issue has no identifying detail." + "reason": "Concrete save_pretrained missing-files bug vs. generic bug report; too little overlap." }, { "left": "issue:44273", "right": "issue:45542", "accept": false, - "reason": "Lazy loading regression and TF backend undefined error are unrelated." + "reason": "Lazy loading malfunction vs. TensorFlow backend import error; different root causes." }, { "left": "issue:44273", "right": "issue:45003", "accept": false, - "reason": "Lazy loading behavior and unsafe sys.modules access are not the same failure mode." + "reason": "Lazy loading complaint vs. sys.modules access bug; related area but not the same issue." }, { "left": "issue:44393", "right": "issue:45753", "accept": false, - "reason": "Both involve Qwen3-VL-ish behavior, but one is 2D box output errors and the other is an infinite loop on a specific image; not enough to merge." + "reason": "2D bbox hallucination/error vs. infinite-loop-on-image bug; different symptoms and likely different causes." }, { "left": "issue:44038", "right": "issue:45125", "accept": false, - "reason": "Both touch Qwen3 MoE models, but one is a broad v5 incompatibility and the other is a missing tensor-parallel plan; different concrete fixes." + "reason": "Qwen3-VL-Moe bug vs. missing _tp_plan for tensor parallelism; not the same concrete change." }, { "left": "issue:44393", "right": "issue:45071", "accept": false, - "reason": "Model output bug vs core PretrainedConfig type-checking regression; unrelated." + "reason": "Model output bug vs. PretrainedConfig type-checking regression; unrelated." }, { "left": "issue:43576", "right": "issue:44556", "accept": false, - "reason": "Both are version-upgrade regressions, but one is a broken CLI command and the other is checkpoint reload compatibility; too broad to be duplicates." + "reason": "Broken transformers env command vs. checkpoint reload incompatibility; different surfaces." }, { - "left": "issue:43525", - "right": "issue:45850", + "left": "issue:41669", + "right": "issue:43673", "accept": false, - "reason": "Specific Llama4 pad_token_id AttributeError vs a vague umbrella complaint about many models breaking." + "reason": "Import-time slowdown from wildcard imports vs. missing GenerationMixin cache in chunked prefill; unrelated." }, { - "left": "issue:41669", - "right": "issue:43673", + "left": "issue:43525", + "right": "issue:45850", "accept": false, - "reason": "Import-star slowdown and missing GenerationMixin cache are different code paths and different bugs." + "reason": "Llama4Config pad_token_id attribute error vs. broad v5 complaint; not the same bug." }, { "left": "issue:44556", "right": "issue:45092", "accept": false, - "reason": "Both concern v5 loading incompatibilities, but the former is general checkpoint reload failure and the latter is a specific remote-code/meta-init incompatibility; not the same issue." + "reason": "Both are v5 checkpoint-loading regressions, but one is generic reload failure and the other is a specific InternVL2/meta-init incompatibility; not enough evidence they are identical." }, { "left": "issue:44492", "right": "issue:45850", "accept": false, - "reason": "A typo in cache strategies is not the same as a broad report that many models break in v5." + "reason": "Cache-strategy typo vs. broad model breakage claim; unrelated." }, { "left": "issue:43572", "right": "issue:45850", "accept": false, - "reason": "Specific StableLmConfig missing pad_token_idx vs a vague multi-model regression complaint." + "reason": "Missing pad_token_idx after v5 update vs. generic v5 breakage report; different bugs." }, { "left": "issue:42371", "right": "issue:44492", "accept": false, - "reason": "TF32 settings/API behavior is unrelated to cache-strategy typo." + "reason": "TF32 behavior warning vs. cache-strategy typo; unrelated." }, { "left": "issue:44704", "right": "issue:44908", "accept": false, - "reason": "Both involve kwargs handling, but in different APIs and subsystems; no shared concrete bug." + "reason": "AutoProcessor kwargs plumbing vs. inverse_sqrt scheduler kwargs handling; different code paths." }, { "left": "issue:42890", "right": "issue:44908", "accept": false, - "reason": "Integration-test seed flakiness and scheduler kwargs bug are unrelated." + "reason": "Flaky SAM-HQ integration tests vs. scheduler kwargs bug; unrelated." }, { "left": "issue:45278", "right": "issue:45850", "accept": false, - "reason": "Both are broad v5 regressions, but one is specifically import errors after upgrade and the other is a generic complaint that many models break; not the same underlying bug." + "reason": "Import errors after upgrading to 5.5.0 vs. a vague claim that many models break; too broad to be the same concrete defect." }, { "left": "issue:44246", "right": "issue:44273", "accept": false, - "reason": "Import is slow sometimes vs lazy loading broken; related theme, but not the same defect." + "reason": "Import slowdown/performance regression vs. lazy loading not working properly; related theme but not the same bug." }, { "left": "issue:43408", "right": "issue:44315", "accept": false, - "reason": "SAM3 type mismatch warning and Liger Kernel application with model_init are unrelated." + "reason": "sam3_video/sam3_tracker class mismatch warning vs. Liger Kernel application with model_init; unrelated." }, { "left": "issue:43976", "right": "issue:44393", "accept": false, - "reason": "Python version compatibility issue vs a model output hallucination/box error; no overlap." + "reason": "Python version compatibility problem vs. Qwen3-VL bbox output bug; unrelated." }, { "left": "issue:43352", "right": "issue:44393", "accept": false, - "reason": "Flash Attention support error and Qwen3-VL box output bug are unrelated." + "reason": "Flash Attention 2 support error vs. Qwen3-VL output hallucination; different issues." }, { "left": "issue:44297", "right": "issue:45230", "accept": false, - "reason": "Tokenizer save metadata mismatch does not match the generic bug report." + "reason": "tokenizer_class mismatch on save vs. generic bug report; not the same bug." }, { "left": "issue:44295", "right": "issue:45230", "accept": false, - "reason": "Position_ids buffer access error is a specific tensor/buffer bug; the other issue is too generic." + "reason": "position_ids buffer read error vs. generic bug report; unrelated." + }, + { + "left": "issue:43576", + "right": "issue:44908", + "accept": false, + "reason": "CLI env command breakage vs. scheduler kwarg handling; unrelated." }, { "left": "issue:41669", "right": "issue:44351", "accept": false, - "reason": "Import-star performance regression and HybridCache import failure are different import-related issues, not the same bug." + "reason": "Wildcard import performance regression vs. HybridCache import failure; different symbols and causes." }, { "left": "issue:41669", "right": "issue:45341", "accept": false, - "reason": "Model import-star slowdown vs a small testing_utils bug are unrelated." + "reason": "Import performance regression vs. a testing_utils bug; unrelated." }, { "left": "issue:44492", "right": "issue:44945", "accept": false, - "reason": "Cache-strategy typo and pipeline-parallel LLM output correctness are unrelated." + "reason": "Cache-strategy typo vs. incorrect pipeline-parallel LLM output; unrelated." }, { "left": "issue:43976", "right": "issue:45850", "accept": false, - "reason": "Python 3.9/3.10 compatibility issue is not the same as a generic many-models-broken report." + "reason": "Python version support issue vs. broad v5 complaint; not the same defect." }, { "left": "issue:44393", "right": "issue:45478", "accept": false, - "reason": "Different Qwen3.5-related failures: one is 2D box output error, the other is a from_pretrained error on Qwen3.5 MoE." + "reason": "Qwen3-VL bbox output bug vs. Qwen3.5-Moe from_pretrained error; different model paths and symptoms." + }, + { + "left": "issue:44077", + "right": "issue:44861", + "accept": false, + "reason": "patchtsmixer post_init policy change vs. tied-weight-key AttributeError; unrelated." + }, + { + "left": "issue:43827", + "right": "issue:44908", + "accept": false, + "reason": "Documentation still referencing pipeline() vs. scheduler kwargs being ignored; different topics." + }, + { + "left": "issue:31795", + "right": "issue:33453", + "accept": false, + "reason": "Forward-signature documentation confusion vs. tokenizer loading regression; unrelated." + }, + { + "left": "issue:41669", + "right": "issue:44908", + "accept": false, + "reason": "Import-time regression from wildcard imports vs. scheduler kwargs bug; unrelated." } ] }, @@ -14761,8 +14933,6 @@ "issue:43606", "issue:43618", "issue:43704", - "issue:43827", - "issue:44077", "issue:44230", "issue:44261", "issue:44263", @@ -14786,10 +14956,6 @@ "issue:45850" ], "soft_pairs": [ - "issue:44077|issue:44861", - "issue:43827|issue:44908", - "issue:31795|issue:33453", - "issue:41669|issue:44908", "issue:44315|issue:44829", "issue:43576|issue:44273", "issue:44655|issue:44861", @@ -14829,94 +14995,70 @@ "issue:31515|issue:33453", "issue:43298|issue:44230" ], - "prepared_review_unit_hash": "931f54afddf7a3b4461342b5940e8d082f0ad5035a89ab32746b889e1c4842d4", + "prepared_review_unit_hash": "9de6ceded600149673dc714e35aeddfcafd77bbc6884df72d0a428caf59ab853", "component_budget": { "node_count": 404, "item_count": 404, "soft_pair_count": 1511, - "serialized_chars": 779569, - "estimated_input_tokens": 194893, - "estimated_eval_tokens": 390042 + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 }, "budget": { - "node_count": 44, - "item_count": 44, - "soft_pair_count": 42, - "serialized_chars": 37280, - "estimated_input_tokens": 9320, - "estimated_eval_tokens": 18896 + "node_count": 42, + "item_count": 42, + "soft_pair_count": 38, + "serialized_chars": 34927, + "estimated_input_tokens": 8732, + "estimated_eval_tokens": 17720 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is highly heterogeneous: most pairs are superficially similar only because they mention v5 regressions, loading, or model-specific bugs, but they target different code paths and fixes. No soft edge is strong enough to merge as a duplicate.", - "confidence": 0.95, - "canonical_issue_reason": "issue:41669 is the best canonical issue anchor: it is concrete, regression-focused, and has enough discussion/inbound linkage to serve as a representative high-impact issue.", + "summary": "These artifacts are mostly a loose collection of unrelated Transformers bug reports, with a few broad v5-regression themes. None of the soft pairs look like true duplicates of the same concrete bug or change, so they should stay separate.", + "confidence": 0.71, + "canonical_issue_reason": "issue:45850 is the broadest umbrella here (\u201cSince 5.0 version it breaks too many models\u201d) and best matches the cluster-wide v5 regression theme, even though it is very generic.", "canonical_pr_reason": null, - "best_issue_reason": "issue:41669 is the strongest standalone issue overall because it describes a specific, measurable regression with broader impact than the more vague or one-off reports.", + "best_issue_reason": "issue:45850 is the best representative issue for the cluster because it summarizes the shared high-level complaint without implying a specific code-path fix.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44077", - "right": "issue:44861", - "accept": false, - "reason": "Different bugs: optional post_init validation vs AttributeError in tied-weight key handling." - }, - { - "left": "issue:43827", - "right": "issue:44908", - "accept": false, - "reason": "Docs still referencing pipeline() is unrelated to inverse_sqrt scheduler kwargs being ignored." - }, - { - "left": "issue:31795", - "right": "issue:33453", - "accept": false, - "reason": "Model.forward docs confusion and tokenizer loading regression are different problems." - }, - { - "left": "issue:41669", - "right": "issue:44908", - "accept": false, - "reason": "Import-* performance regression is unrelated to scheduler kwargs handling." - }, { "left": "issue:44315", "right": "issue:44829", "accept": false, - "reason": "Liger Kernel not applied with model_init and flash_attention_3 degenerate training affect different code paths." + "reason": "Different symptoms and code paths: model_init/Liger Kernel application vs flash_attention_3 training degeneration." }, { "left": "issue:43576", "right": "issue:44273", "accept": false, - "reason": "Broken transformers env command and lazy loading failure are unrelated." + "reason": "One is the v5 `transformers env` command, the other is lazy loading behavior; no concrete overlap." }, { "left": "issue:44655", "right": "issue:44861", "accept": false, - "reason": "Saving Pipeline objects and _get_tied_weight_keys crash are distinct save/load path bugs." + "reason": "`save_pretrained` pipeline saving failure is unrelated to `_get_tied_weight_keys` crashing on a list." }, { "left": "issue:44360", "right": "issue:45468", "accept": false, - "reason": "DSA indexer ReLU omission and Gemma-4 audio positional encoding are unrelated model components." + "reason": "DSA indexer ReLU discussion and Gemma4Audio positional encoding are different model internals." }, { "left": "issue:45230", "right": "issue:45310", "accept": false, - "reason": "A generic bug report does not identify the same from_pretrained error in Qwen3.5 MoE." + "reason": "`Bug report` is too generic to show the same bug as the Qwen3.5 MoE from_pretrained error." }, { "left": "issue:44273", "right": "issue:44908", "accept": false, - "reason": "Lazy loading failure is unrelated to inverse_sqrt scheduler kwargs being dropped." + "reason": "Lazy loading and inverse_sqrt scheduler kwargs are unrelated features." }, { "left": "issue:43519", @@ -14928,187 +15070,187 @@ "left": "issue:43519", "right": "issue:44908", "accept": false, - "reason": "Timestamp math in a processor and scheduler kwargs handling are unrelated." + "reason": "Processor timestamp math is unrelated to scheduler kwarg handling." }, { "left": "issue:31795", "right": "issue:34689", "accept": false, - "reason": "Documentation confusion and Llama 3.2 vision model loading regression are not the same issue." + "reason": "Documentation confusion does not match a model loading regression." }, { "left": "issue:41669", "right": "issue:43704", "accept": false, - "reason": "Import-star slowdown and VRAM leak in multithreaded inference are different regressions." + "reason": "Import-star cleanup/performance work is unrelated to a VRAM leak in multi-threaded inference." }, { "left": "issue:31795", "right": "issue:37428", "accept": false, - "reason": "Docs wording issue and flash_attention_utils import error are unrelated." + "reason": "Docs wording issue vs import error in flash attention utilities are not the same bug." }, { "left": "issue:44315", "right": "issue:45092", "accept": false, - "reason": "Both touch model initialization, but one is kernel application with model_init and the other is remote-code/meta-init incompatibility; not the same bug." + "reason": "Both involve model construction/loading, but the specific failure modes and causes differ." }, { "left": "issue:44485", "right": "issue:45468", "accept": false, - "reason": "GLM-5 RoPE logic and Gemma-4 audio positional encoding are different implementations." + "reason": "GLM-5 RoPE math and Gemma-4 audio positional encoding are distinct model implementations." }, { "left": "issue:43704", "right": "issue:44485", "accept": false, - "reason": "VRAM leak in dataloader threads is unrelated to RoPE implementation." + "reason": "VRAM leakage in dataloader threads is unrelated to RoPE implementation." }, { "left": "issue:43502", "right": "issue:43519", "accept": false, - "reason": "local_files_only being ignored and timestamp calculation are separate issues." + "reason": "Accidental API requests with local_files_only and Qwen3VL timestamp errors are unrelated." }, { "left": "issue:42371", "right": "issue:43704", "accept": false, - "reason": "TF32 settings API guidance is unrelated to the VRAM leak report." + "reason": "TF32 settings behavior and VRAM leakage are different subsystems." }, { "left": "issue:43519", "right": "issue:44623", "accept": false, - "reason": "Processor timestamp math and missing save_pretrained files are different bugs." + "reason": "Processor timestamp calculation and missing save_pretrained files are separate issues." }, { "left": "issue:43519", "right": "issue:43704", "accept": false, - "reason": "Timestamp calculation bug does not match the VRAM leak in multiple threads." + "reason": "No shared code path between Qwen3VL timestamp handling and VRAM leaks." }, { "left": "issue:44263", "right": "issue:45468", "accept": false, - "reason": "torch.split return values in the DSA indexer and Gemma-4 audio positional encoding are unrelated." + "reason": "torch.split return values in a DSA indexer are unrelated to Gemma4Audio positional encoding." }, { "left": "issue:43606", "right": "issue:45468", "accept": false, - "reason": "CPU offload device mismatch and audio positional encoding are different code paths." + "reason": "CPU offload device mismatch for bark-small is not the same bug as Gemma4Audio positional encoding." }, { "left": "issue:44261", "right": "issue:45468", "accept": false, - "reason": "Missing rms_norm_eps in MLA q_a_layernorm is unrelated to Gemma-4 audio positional encoding." + "reason": "MLA q_a_layernorm precision/config issue is unrelated to Gemma4Audio positional encoding." }, { "left": "issue:42913", "right": "issue:43066", "accept": false, - "reason": "Both concern tokenizer changes in v5, but one is broad behavior drift and the other is a specific decoder-type bug; not clearly the same fix." + "reason": "Both mention tokenizer v5 behavior, but one is a broad behavioral regression while the other is a specific decoder-type problem; not enough to merge." }, { "left": "issue:44908", "right": "issue:44998", "accept": false, - "reason": "Scheduler kwargs loss and an unrelated low-signal issue titled unemployment are not the same." + "reason": "Scheduler kwarg handling and an unrelated placeholder issue are not duplicates." }, { "left": "issue:44492", "right": "issue:44998", "accept": false, - "reason": "Cache-strategy typo and the unrelated unemployment issue do not match." + "reason": "Typo in cache strategies is unrelated to the placeholder issue." }, { "left": "issue:44861", "right": "issue:45290", "accept": false, - "reason": "Tied-weight-key AttributeError and chat-template crashes on tool-call messages are different failures." + "reason": "A tied-weights helper crash is unrelated to chat template crashes on tool-call assistant messages." }, { "left": "issue:44998", "right": "issue:45850", "accept": false, - "reason": "A nonsense unemployment issue is not the same as a broad v5 regression complaint." + "reason": "'Unemployment' is not a meaningful bug report and does not describe the same problem as the broad v5 regression complaint." }, { "left": "issue:43618", "right": "issue:44998", "accept": false, - "reason": "CLIPOutput attentions regression is unrelated to the unemployment issue." + "reason": "CLIPOutput attentions assignment regression is unrelated to the placeholder issue." }, { "left": "issue:44998", "right": "issue:45341", "accept": false, - "reason": "Unemployment and a testing_utils bug are unrelated." + "reason": "Placeholder/empty issue does not match the testing_utils bug." }, { "left": "issue:43525", "right": "issue:44861", "accept": false, - "reason": "Missing pad_token_id in Llama4Config and tied-weight-key handling are different config/path bugs." + "reason": "Different AttributeErrors in different components; no shared underlying bug." }, { "left": "issue:43572", "right": "issue:44861", "accept": false, - "reason": "StableLmConfig pad_token_idx removal and tied-weight-key crash are unrelated." + "reason": "Missing pad_token_idx in a config is unrelated to tied-weight key handling." }, { "left": "issue:31515", "right": "issue:38617", "accept": false, - "reason": "Slow checkpoint loading and import error for layer_type_validation are unrelated." + "reason": "Slow checkpoint loading and an import error in configuration_utils are unrelated." }, { "left": "issue:44230", "right": "issue:44623", "accept": false, - "reason": "FP8 Qwen3-VL/MoE support and processor.save_pretrained file omission are different bugs." + "reason": "FP8 model support/inference and processor.save_pretrained file emission are different issues." }, { "left": "issue:30333", "right": "issue:31515", "accept": false, - "reason": "MLflow reporting on training failure and slow from_pretrained checkpoint loading have no shared code path." + "reason": "MLflow reporting on failed training jobs and slow from_pretrained checkpoint loading are unrelated." }, { "left": "issue:31515", "right": "issue:36296", "accept": false, - "reason": "Checkpoint loading slowness and tensor parallel training bug are distinct problems." + "reason": "Checkpoint load performance and tensor-parallel training bugs do not share the same concrete failure." }, { "left": "issue:31515", "right": "issue:31795", "accept": false, - "reason": "Pretrained loading performance and forward-doc confusion are unrelated." + "reason": "Loading slowdown and documentation confusion are not duplicates." }, { "left": "issue:44230", "right": "issue:45310", "accept": false, - "reason": "Qwen3.5 MoE FP8 support request and from_pretrained loading error are not the same concrete bug." + "reason": "Both involve Qwen3.5 MoE, but one is FP8 support and the other is a from_pretrained error; too different to merge." }, { "left": "issue:31515", "right": "issue:33453", "accept": false, - "reason": "Checkpoint loading slowdown and tokenizer loading regression are different failure modes." + "reason": "Slow checkpoint loading is not the same as a tokenizer-loading regression." }, { "left": "issue:43298", "right": "issue:44230", "accept": false, - "reason": "Continuous batching not supporting audio models and FP8 Qwen support are unrelated." + "reason": "Continuous batching for audio models and Qwen3-VL/Qwen3.5 FP8 support are different features and failures." } ] }, @@ -15140,7 +15282,7 @@ "issue:13244|issue:16618", "issue:15980|issue:16618" ], - "prepared_review_unit_hash": "d81e87497b382c778b6b53d98246672fd0dff0489e51094430927e0852d68115", + "prepared_review_unit_hash": "f3c2fc74fc672ee7b42181d41b58e92f7eeb2ff72069f4764f07a170a591060a", "component_budget": { "node_count": 3, "item_count": 3, @@ -15162,24 +15304,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These issues do not appear to be duplicates of one another. Two are about model-loading/authentication behavior, while one is about Tapas tokenization differences; the overlap is too superficial to treat them as the same underlying bug.", - "confidence": 0.88, - "canonical_issue_reason": "No clear canonical duplicate exists because the cluster mixes unrelated reports. If one issue must anchor the cluster, #16618 is the broadest model-loading report.", + "summary": "These issues are not duplicates: one is about Tapas tokenization diverging from TensorFlow code, one is about a misleading private-model download/login error, and one is a generic model-loading failure for bert-base-uncased.", + "confidence": 0.95, + "canonical_issue_reason": "#16618 is the broadest, most central user-facing failure in the set, but it still describes a different problem from the other two issues.", "canonical_pr_reason": null, - "best_issue_reason": "#16618 is the best representative only in the sense that it is the most general and central-sounding of the three, but it still does not subsume the others.", + "best_issue_reason": "#16618 is the best cluster anchor because its title is the most general and it has the strongest discussion activity, even though it is not the same underlying bug as the others.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:13244", "right": "issue:16618", "accept": false, - "reason": "Different bugs: Tapas tokenization mismatch vs. inability to load bert-base-uncased. They involve different code paths and symptoms." + "reason": "Tapas tokenization mismatch is a different code path and symptom from a failure to load bert-base-uncased; they are not the same underlying bug." }, { "left": "issue:15980", "right": "issue:16618", "accept": false, - "reason": "Both mention loading models, but one is specifically a bad auth/error message for private-model downloads while the other reports a general failure loading a public model. Not the same underlying issue." + "reason": "A bad unauthenticated private-model download message is an authentication/download UX bug, not the same as a model-loading failure for bert-base-uncased." } ] }, @@ -15209,12 +15351,12 @@ "soft_pairs": [ "issue:26350|issue:33260" ], - "prepared_review_unit_hash": "ce45d12a1b19f95ec399fab591896c5d576aaed59fadf569a7feb183ad25837a", + "prepared_review_unit_hash": "74bf751df619f32360334a6a972358487190ce061986e1f41300c39d495f69d5", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1590, + "serialized_chars": 1589, "estimated_input_tokens": 398, "estimated_eval_tokens": 1052 }, @@ -15222,7 +15364,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1590, + "serialized_chars": 1589, "estimated_input_tokens": 398, "estimated_eval_tokens": 1052 }, @@ -15231,18 +15373,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are not duplicates: one tracks Flash Attention 2 support for more architectures, the other tracks GGUF support for more architectures. They share a community-contribution framing but not the same feature work.", + "summary": "These two issues are about different features: Flash Attention 2 support vs GGUF support. They share a \u201ccommunity contribution\u201d framing, but not the same underlying bug or change, so they should not be deduplicated.", "confidence": 0.98, - "canonical_issue_reason": "No true canonical duplicate here; the issues describe different feature requests and should not be merged into one tracker.", + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Neither issue is a better duplicate representative of the other. If forced to pick a cluster representative, 26350 is older and more discussed, but it is still a distinct request.", + "best_issue_reason": "Issue #26350 is the more established representative of its own feature request, with higher discussion activity, but it is not a duplicate of #33260.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:26350", "right": "issue:33260", "accept": false, - "reason": "Different targets: Flash Attention 2 support vs GGUF support. Same broad 'community contribution' framing, but not the same underlying bug or change." + "reason": "Different requested capabilities and code paths: Flash Attention 2 vs GGUF support. Similar wording about community contributions is too broad to indicate the same issue." } ] }, @@ -15284,7 +15426,7 @@ "issue:43834|issue:44365", "issue:44365|issue:44370" ], - "prepared_review_unit_hash": "6708a0640a08936e4c9eeb7e42546312cf15800676fe534ea1671ccfeb4a8108", + "prepared_review_unit_hash": "078481a518474127f496a6c37c7920be3b93d6cbe278327a0c8cb42c7cd56ad8", "component_budget": { "node_count": 5, "item_count": 5, @@ -15306,78 +15448,78 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster contains two distinct language-specific docs translation requests (Turkish and Spanish) plus three near-identical generic template issues. The generic placeholder issues look like true duplicates of the same request, but the Turkish and Spanish issues are separate language targets and should stay distinct.", - "confidence": 0.93, - "canonical_issue_reason": "Issue 43834 is the best representative of the duplicate generic-template subgroup: it matches the other two almost exactly, is already closed, and predates the other near-identical copies.", + "summary": "The cluster is mostly a set of separate docs-translation requests by language, plus three nearly identical placeholder/template issues. Only the placeholder trio looks duplicate; the Turkish and Spanish issues are distinct tasks and should not be merged with them.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 43834 is the best representative of the duplicate subgroup because it is one of the near-identical placeholder translation issues and the earliest of the trio. It is not a good canonical for the full cluster, since the Turkish and Spanish requests are separate language-specific tasks.", "canonical_pr_reason": null, - "best_issue_reason": "43834 is the strongest global issue representative because it is part of the only clear duplicate set in the cluster and is the earliest/most established instance of that identical template issue.", + "best_issue_reason": "43834 is the clearest canonical candidate among the items because it matches the other placeholder issues almost exactly and is more representative than the language-specific requests.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:27088", "right": "issue:44370", "accept": false, - "reason": "Different target language request; Turkish docs translation is not the same underlying issue as the generic template entry." + "reason": "Different concrete request: Turkish translation vs a generic placeholder translation template. Similar category, but not the same underlying issue." }, { "left": "issue:27088", "right": "issue:44365", "accept": false, - "reason": "Different target language request; Turkish and generic placeholder translation issues are not duplicates." + "reason": "Turkish-specific translation request is not the same as the generic placeholder template issue." }, { "left": "issue:27088", "right": "issue:43834", "accept": false, - "reason": "Different target language request; Turkish docs translation is separate from the generic template issue." + "reason": "Distinct language-specific request versus generic placeholder template; not duplicates." }, { "left": "issue:27088", "right": "issue:28936", "accept": false, - "reason": "Turkish vs Spanish are different translation requests, so they are not the same underlying issue." + "reason": "Different target language and therefore a different translation task." }, { "left": "issue:28936", "right": "issue:43834", "accept": false, - "reason": "Spanish docs translation is distinct from the generic template issue and should not be merged as a duplicate." + "reason": "Spanish translation request is separate from the generic placeholder template issue." }, { "left": "issue:28936", "right": "issue:44365", "accept": false, - "reason": "Spanish docs translation is a different language-specific request than the generic template issue." + "reason": "Different concrete language request versus generic placeholder template." }, { "left": "issue:28936", "right": "issue:44370", "accept": false, - "reason": "Spanish docs translation is not the same underlying issue as the generic template entry." + "reason": "Spanish translation request is not the same underlying issue as the placeholder template." }, { "left": "issue:43834", "right": "issue:44370", "accept": true, - "reason": "These two issues are near-identical generic template entries with almost the same content, indicating the same underlying request." + "reason": "These two placeholder translation issues are effectively the same text and appear to be duplicate template filings." }, { "left": "issue:43834", "right": "issue:44365", "accept": true, - "reason": "Near-perfect similarity and identical template-style content make these clear duplicates of the same issue." + "reason": "Near-identical placeholder issue title/body; same underlying template request." }, { "left": "issue:44365", "right": "issue:44370", "accept": true, - "reason": "These are effectively the same generic template issue with nearly identical bodies and titles." + "reason": "Near-exact duplicates of the same placeholder translation issue, differing only trivially." } ] }, "evaluator_result": { "accept": true, - "feedback": "" + "feedback": "The summary is well grounded in the packet. The three generic placeholder issues are near-identical by title and similarity scores, while the Turkish and Spanish issues are clearly separate language-specific requests. The canonical choice is conservative, and the pairwise verdicts align with the evidence." }, "evaluator_used": true, "retried": false, @@ -15451,244 +15593,244 @@ "issue:37015|issue:45027", "issue:43837|issue:45177" ], - "prepared_review_unit_hash": "77f6da0403a04d7e6bdf6bfa76887ed87126bea1cd87f723dbcf9df66c4521df", + "prepared_review_unit_hash": "f5c42ecfea37824ab8a889fbc82ce06be998568a6593f3e41e78539daeec12d1", "component_budget": { "node_count": 15, "item_count": 15, "soft_pair_count": 35, - "serialized_chars": 20703, - "estimated_input_tokens": 5176, - "estimated_eval_tokens": 10608 + "serialized_chars": 20708, + "estimated_input_tokens": 5177, + "estimated_eval_tokens": 10610 }, "budget": { "node_count": 15, "item_count": 15, "soft_pair_count": 35, - "serialized_chars": 20703, - "estimated_input_tokens": 5176, - "estimated_eval_tokens": 10608 + "serialized_chars": 20708, + "estimated_input_tokens": 5177, + "estimated_eval_tokens": 10610 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is not a duplicate set; it contains many independent model-addition requests across different architectures, modalities, and vendor/model families. The shared wording is generic (\u201cadd/support model\u201d), but the underlying change requests are distinct.", - "confidence": 0.95, - "canonical_issue_reason": null, + "summary": "This is a broad thematic cluster of model-addition requests, not a true duplicate set. The issues mention different models, architectures, and modalities (vision, text, ASR/TTS, audio, OCR), so none of the soft edges should be merged as the same underlying change.", + "confidence": 0.96, + "canonical_issue_reason": "Closest anchor in this non-duplicate theme cluster: issue 43671 is a detailed, active support request and the best representative of the set, though it is not a duplicate of the others.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "Issue 43671 is the strongest standalone representative here because it is specific, actively discussed, and clearly scoped, but it still should not subsume the other model requests.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:36817", "right": "issue:41211", "accept": false, - "reason": "EuroBert vs DEIMv2 are unrelated model requests; same broad feature area, not the same underlying addition." + "reason": "Different requested models and changes: EuroBert config vs DEIMv2 support." }, { "left": "issue:36817", "right": "issue:43671", "accept": false, - "reason": "EuroBert config support and Qwen3-TTS support target different model families and modalities." + "reason": "EuroBert config and Qwen3-TTS are unrelated model additions with different code paths." }, { "left": "issue:36817", "right": "issue:36879", "accept": false, - "reason": "EuroBert config support and RF-DETR are distinct model additions with different code paths." + "reason": "Different model families and tasks: EuroBert vs RF-DETR." }, { "left": "issue:36879", "right": "issue:41211", "accept": false, - "reason": "RF-DETR and DEIMv2 are separate vision-model requests, not a duplicate bug/change." + "reason": "RF-DETR and DEIMv2 are distinct detection models; not the same underlying change." }, { "left": "issue:36879", "right": "issue:43909", "accept": false, - "reason": "RF-DETR and LFM2.5 Audio 1.5B are different model families and tasks." + "reason": "RF-DETR vs LFM2.5 Audio are different domains and implementations." }, { "left": "issue:36879", "right": "issue:44121", "accept": false, - "reason": "RF-DETR and the weight-sparse transformer request are unrelated model additions." + "reason": "RF-DETR support and weight-sparse transformer support are unrelated code changes." }, { "left": "issue:36879", "right": "issue:43837", "accept": false, - "reason": "RF-DETR and Qwen3-ASR are different models and different modalities." + "reason": "RF-DETR and Qwen3-ASR are different model types and subsystems." }, { "left": "issue:36879", "right": "issue:43671", "accept": false, - "reason": "RF-DETR and Qwen3-TTS are distinct model support requests." + "reason": "RF-DETR and Qwen3-TTS are unrelated model additions; same broad repo area only." }, { "left": "issue:37015", "right": "issue:43837", "accept": false, - "reason": "NeoBERT and Qwen3-ASR are unrelated requests; only the generic add-support wording overlaps." + "reason": "NeoBERT and Qwen3-ASR are different models with different integration paths." }, { "left": "issue:37015", "right": "issue:44121", "accept": false, - "reason": "NeoBERT support and the sparse transformer request are different model additions." + "reason": "NeoBERT support and sparse-transformer support are not the same bug/change." }, { "left": "issue:37015", "right": "issue:43909", "accept": false, - "reason": "NeoBERT and LFM2.5 Audio 1.5B are separate model families with different functionality." + "reason": "NeoBERT is a text model; LFM2.5 Audio is an audio model." }, { "left": "issue:36879", "right": "issue:37015", "accept": false, - "reason": "RF-DETR and NeoBERT are unrelated model requests, not one concrete code-path fix." + "reason": "RF-DETR and NeoBERT are unrelated model requests." }, { "left": "issue:37015", "right": "issue:42503", "accept": false, - "reason": "NeoBERT and ModernVBERT are different BERT-like model requests with different targets." + "reason": "NeoBERT and ModernVBERT are distinct model requests, not one mergeable fix." }, { "left": "issue:41115", "right": "issue:43837", "accept": false, - "reason": "MiniCPM3 architecture support and Qwen3-ASR are distinct model requests." + "reason": "MiniCPM3 architecture and Qwen3-ASR support are different concrete additions." }, { "left": "issue:41115", "right": "issue:44121", "accept": false, - "reason": "MiniCPM3 architecture support and the sparse transformer request are not the same underlying change." + "reason": "MiniCPM3 and the sparse-transformer request target different code paths." }, { "left": "issue:41115", "right": "issue:42503", "accept": false, - "reason": "MiniCPM3 architecture support and ModernVBERT models are separate additions." + "reason": "MiniCPM3 and ModernVBERT are separate model support requests." }, { "left": "issue:41115", "right": "issue:45027", "accept": false, - "reason": "MiniCPM3 architecture support and Voxtral TTS support are different modalities and code paths." + "reason": "MiniCPM3 architecture support and Voxtral TTS support are unrelated changes." }, { "left": "issue:41211", "right": "issue:43671", "accept": false, - "reason": "DEIMv2 and Qwen3-TTS are different model requests with no shared concrete fix." + "reason": "DEIMv2 and Qwen3-TTS are different model integrations." }, { "left": "issue:42503", "right": "issue:43837", "accept": false, - "reason": "ModernVBERT and Qwen3-ASR are unrelated model additions." + "reason": "ModernVBERT and Qwen3-ASR are different models and tasks." }, { "left": "issue:42503", "right": "issue:43671", "accept": false, - "reason": "ModernVBERT models and Qwen3-TTS support are different requests." + "reason": "ModernVBERT and Qwen3-TTS are unrelated model support requests." }, { "left": "issue:42503", "right": "issue:42738", "accept": false, - "reason": "ModernVBERT and BERT-like models with RoPE overlap only at a broad theme, not the same model implementation." + "reason": "One is a specific model request, the other is a broad RoPE architectural proposal; not the same concrete change." }, { "left": "issue:42738", "right": "issue:45027", "accept": false, - "reason": "BERT-like models with RoPE and Voxtral TTS support are unrelated beyond being model requests." + "reason": "BERT-like RoPE support and Voxtral TTS are unrelated." }, { "left": "issue:43671", "right": "issue:43837", "accept": false, - "reason": "Qwen3-TTS and Qwen3-ASR are separate sibling models, but they are different support requests." + "reason": "Qwen3-TTS and Qwen3-ASR share a family name but target different modalities and integration work." }, { "left": "issue:43671", "right": "issue:45027", "accept": false, - "reason": "Qwen3-TTS and Voxtral-4B-TTS-2603 are distinct TTS-model support requests." + "reason": "Both are TTS-related requests, but for different models and code paths." }, { "left": "issue:43671", "right": "issue:45177", "accept": false, - "reason": "Qwen3-TTS and DeepSeek-OCR2 are different modalities and different code paths." + "reason": "Qwen3-TTS and DeepSeek-OCR2 are unrelated model requests." }, { "left": "issue:43671", "right": "issue:44121", "accept": false, - "reason": "Qwen3-TTS support and the sparse transformer request are unrelated model additions." + "reason": "TTS support and sparse-transformer support are different underlying changes." }, { "left": "issue:43671", "right": "issue:43909", "accept": false, - "reason": "Qwen3-TTS and LFM2.5 Audio 1.5B are separate audio-model requests." + "reason": "Qwen3-TTS and LFM2.5 Audio are different audio-model integrations." }, { "left": "issue:43837", "right": "issue:43909", "accept": false, - "reason": "Qwen3-ASR and LFM2.5 Audio 1.5B are different audio model support requests." + "reason": "Qwen3-ASR and LFM2.5 Audio are distinct audio-related model requests." }, { "left": "issue:43837", "right": "issue:45322", "accept": false, - "reason": "Qwen3-ASR and EUPE are unrelated; one is speech recognition, the other is a perception encoder." + "reason": "Qwen3-ASR and EUPE are unrelated model additions." }, { "left": "issue:43837", "right": "issue:44121", "accept": false, - "reason": "Qwen3-ASR support and the sparse transformer request are not the same concrete model change." + "reason": "ASR support and weight-sparse transformer support are not the same code-path problem." }, { "left": "issue:43909", "right": "issue:44205", "accept": false, - "reason": "LFM2.5 Audio 1.5B and SAM3-LiteText are different model families and tasks." + "reason": "LFM2.5 Audio and SAM3-LiteText are different model families and tasks." }, { "left": "issue:44121", "right": "issue:45322", "accept": false, - "reason": "The weight-sparse transformer request and EUPE are distinct model additions." + "reason": "Sparse-transformer support and EUPE support are unrelated changes." }, { "left": "issue:43837", "right": "issue:45027", "accept": false, - "reason": "Qwen3-ASR and Voxtral TTS are different speech-model support requests." + "reason": "Qwen3-ASR and Voxtral TTS are different modalities and model integrations." }, { "left": "issue:37015", "right": "issue:45027", "accept": false, - "reason": "NeoBERT and Voxtral TTS are unrelated model additions." + "reason": "NeoBERT and Voxtral TTS are unrelated model support requests." }, { "left": "issue:43837", "right": "issue:45177", "accept": false, - "reason": "Qwen3-ASR and DeepSeek-OCR2 are different modalities and implementation areas." + "reason": "Qwen3-ASR and DeepSeek-OCR2 are different model additions." } ] }, @@ -15731,12 +15873,12 @@ "issue:43668|issue:44031", "issue:43668|issue:43813" ], - "prepared_review_unit_hash": "d3c4100e5d9165a8b735249fe995a001f962604ed6dc5607e34c7220f1a77830", + "prepared_review_unit_hash": "e6f6010e9df87c15bb65015914185133f4d91a24aefadcfc6298a457f8674f1e", "component_budget": { "node_count": 7, "item_count": 7, "soft_pair_count": 9, - "serialized_chars": 7188, + "serialized_chars": 7185, "estimated_input_tokens": 1797, "estimated_eval_tokens": 3850 }, @@ -15744,7 +15886,7 @@ "node_count": 7, "item_count": 7, "soft_pair_count": 9, - "serialized_chars": 7188, + "serialized_chars": 7185, "estimated_input_tokens": 1797, "estimated_eval_tokens": 3850 }, @@ -15753,66 +15895,66 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "No duplicate cluster here: the issues cover unrelated models, configs, tokenizer warnings, and a typo, with no pair describing the same concrete bug.", - "confidence": 0.98, - "canonical_issue_reason": "No clear canonical issue: the items are about different subsystems and different failure modes, so none can safely represent the whole cluster as a duplicate target.", + "summary": "These issues are only superficially similar: they span unrelated bugs in model config typing, masking, tokenization warnings, quantization typos, loss plumbing, and a gating logic question. No pair looks like the same underlying defect.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 43240 is the broadest and most infrastructure-level report in the set, with the highest inbound references; if one issue must anchor this noisy cluster, it is the least specific and most central title.", "canonical_pr_reason": null, - "best_issue_reason": "#42936 is the most concrete standalone bug report (specific model, specific behavior, plausible fix), but it is not a good duplicate anchor for the rest of the set.", + "best_issue_reason": "Issue 43240 is the best representative only in a loose triage sense because it describes a generic framework bug rather than a model-specific typo or one-off parameter issue.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:37934", "right": "issue:43668", "accept": false, - "reason": "RMS-norm question for Llama4TextL2Norm and a ModernBERT type-hint issue are unrelated model/config problems." + "reason": "Unrelated model behavior question vs. config type-hint issue; no shared code path." }, { "left": "issue:42936", "right": "issue:43240", "accept": false, - "reason": "Mask2former ignore_value handling and missing kwargs in loss calculation are different bugs in different code paths." + "reason": "Mask2former ignore_value handling and loss kwargs plumbing are different bugs in different components." }, { "left": "issue:42936", "right": "issue:46018", "accept": false, - "reason": "Mask2former loss-argument behavior and DeepSeek-V4 gating are unrelated model-specific issues." + "reason": "Masking loss usage and DeepSeek expert gating are separate model behaviors." }, { "left": "issue:42936", "right": "issue:44031", "accept": false, - "reason": "A model ignore_value bug is not the same as a tokenizer regex warning regression." + "reason": "Mask2former loss masking and tokenizer regex warnings are unrelated subsystems." }, { "left": "issue:42936", "right": "issue:43813", "accept": false, - "reason": "Mask2former loss handling and a quantization-typo issue do not share the same underlying defect." + "reason": "Loss ignore_value usage and a quantization typo are not the same change." }, { "left": "issue:42936", "right": "issue:43668", "accept": false, - "reason": "A runtime model bug and a config type-hint mistake are distinct issues." + "reason": "Different model areas: segmentation loss handling vs. ModernBERT config typing." }, { "left": "issue:43240", "right": "issue:43668", "accept": false, - "reason": "Loss kwargs propagation and ModernBERT norm_eps typing concern different code and different symptoms." + "reason": "Generic kwargs propagation bug does not match a type-hint correction." }, { "left": "issue:43668", "right": "issue:44031", "accept": false, - "reason": "Incorrect type hint in ModernBERTConfig is unrelated to tokenizer regex warnings after 4.57.3." + "reason": "A config annotation bug is unrelated to tokenizer regex warning behavior." }, { "left": "issue:43668", "right": "issue:43813", "accept": false, - "reason": "ModernBERT config typing and a typo in orig_conversion.quantization_operation are unrelated changes." + "reason": "ModernBERT norm_eps typing and quantization typo are distinct documentation/code issues." } ] }, @@ -15842,7 +15984,7 @@ "soft_pairs": [ "issue:41084|issue:43037" ], - "prepared_review_unit_hash": "697a8ef0af6fe70431794fb8ea963d80c387767583b4a03dcafd9fcab1bb5ff4", + "prepared_review_unit_hash": "8ab2a7c022e229ddfa5e8a81aa624b87117a481543133d8165cadac182c8e368", "component_budget": { "node_count": 2, "item_count": 2, @@ -15864,18 +16006,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues do not appear to describe the same underlying problem: one is about block decoding, the other is a request for DeepSeek v3.2 support. The soft similarity is too weak to treat them as duplicates.", - "confidence": 0.93, - "canonical_issue_reason": "No clear canonical issue: the items are about different features/problems and should not be merged as duplicates.", + "summary": "These two issues are not duplicates: one is about block decoding behavior, the other is a request for DeepSeek v3.2 support. They describe different feature areas and different underlying changes.", + "confidence": 0.98, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue #43037 is the more specific and actionable request, but it is not a duplicate of #41084.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41084", "right": "issue:43037", "accept": false, - "reason": "Different topics and likely different code paths: block decoding vs model support. The shared similarity is only superficial." + "reason": "Different requests and code paths: block decoding vs model support. No evidence they describe the same underlying bug or change." } ] }, @@ -15905,7 +16047,7 @@ "soft_pairs": [ "issue:43598|issue:43600" ], - "prepared_review_unit_hash": "b5c487d7e8c60f860577c99a135c1dd976963e434b66877f3abf2f91d2f7e053", + "prepared_review_unit_hash": "1769909196a533af01e6fdce040f2d38b9bbc139ee253e96083b6a7e68c879d3", "component_budget": { "node_count": 2, "item_count": 2, @@ -15927,18 +16069,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are thematically related to custom trainers and Trainer internals, but they describe different problems: one is about `remove_unused_column` customizability, the other about FLOPS estimation for custom trainers. I do not see evidence that they are the same underlying bug.", - "confidence": 0.93, + "summary": "These are related to Trainer customizability, but they describe different problems: one is about `remove_unused_column` behavior, the other about FLOPS estimation for custom trainers. They should not be merged as duplicates.", + "confidence": 0.96, "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue 43598 is the broader Trainer customization request and is the better cluster representative, but it should not be treated as a duplicate of 43600.", + "best_issue_reason": "No single issue is a clear duplicate canonical for the other; if forced to pick a representative, #43598 is the more concrete Trainer-behavior customization request.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43598", "right": "issue:43600", "accept": false, - "reason": "Related subsystem, but different concrete concerns and no shared bug path; one is about `remove_unused_column`, the other about FLOPS accounting for custom trainers." + "reason": "Different underlying concerns: one changes `remove_unused_column` handling, the other changes FLOPS estimation for custom trainers. Shared Trainer/custom-trainer context is too broad to treat as the same bug." } ] }, @@ -15968,12 +16110,12 @@ "soft_pairs": [ "issue:43684|issue:44405" ], - "prepared_review_unit_hash": "4a769f8b2979ea000db9ee21021fdc9f16adb7ca9e358e6b6a7f9592959732c9", + "prepared_review_unit_hash": "78c883e5ed914504c1e8c7663b8215f96574a617fc17b7d758a80191483c5129", "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 }, @@ -15981,7 +16123,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1611, + "serialized_chars": 1612, "estimated_input_tokens": 403, "estimated_eval_tokens": 1062 }, @@ -15990,18 +16132,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These issues both concern AutoModel support for Qwen variants, but they request different registrations for different model classes and tasks. They are not the same underlying bug or change request.", - "confidence": 0.95, - "canonical_issue_reason": "Issue #43684 is the broader model-registration request and is a better representative of this cluster than the more specific sequence-classification support request in #44405.", + "summary": "These are not duplicates: one asks for Qwen3-Omni registration in AutoModel/AutoModelForConditionalGeneration, while the other asks for AutoModelForSequenceClassification support for Qwen3.5. They target different model families and different task heads.", + "confidence": 0.96, + "canonical_issue_reason": "Issue 43684 is the broader model-registration request and is the closer representative of the cluster topic, but it is still separate from the Qwen3.5 sequence-classification request.", "canonical_pr_reason": null, - "best_issue_reason": "#43684 is the better canonical issue because it covers a general AutoModel/AutoModelForConditionalGeneration registration gap, whereas #44405 is a separate task-specific support request for AutoModelForSequenceClassification on Qwen3.5.", + "best_issue_reason": "Issue 43684 is the better single representative because it concerns the core registration gap in AutoModel mappings; issue 44405 is a distinct task-specific feature request.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43684", "right": "issue:44405", "accept": false, - "reason": "Different target models and APIs: Qwen3-Omni registration for AutoModel/AutoModelForConditionalGeneration vs Qwen3.5 sequence-classification support for AutoModelForSequenceClassification. They are related only at a broad subsystem level, not the same change." + "reason": "Different model families (Qwen3-Omni vs Qwen3.5) and different APIs (AutoModel/AutoModelForConditionalGeneration vs AutoModelForSequenceClassification); not the same underlying bug or change." } ] }, @@ -16038,12 +16180,12 @@ "issue:43845|issue:43859", "issue:43846|issue:43859" ], - "prepared_review_unit_hash": "766a5057ca8c34151fbd60c49da1feda0776f49de5216859517edefb71d03114", + "prepared_review_unit_hash": "b160aa6500515d4778cfb0d74988123d0af1fa56116c5d4c9d5b2b5e59117e27", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4267, + "serialized_chars": 4266, "estimated_input_tokens": 1067, "estimated_eval_tokens": 2390 }, @@ -16051,7 +16193,7 @@ "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4267, + "serialized_chars": 4266, "estimated_input_tokens": 1067, "estimated_eval_tokens": 2390 }, @@ -16060,57 +16202,57 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Issues 43845, 43846, and 43859 have the same title, so they are likely duplicates of one another; 43835 (\u201cFett\u201d) looks separate. No PRs are present.", - "confidence": 0.79, - "canonical_issue_reason": "Issue 43845 is the earliest of the identical-title \u201chuggingface \u201d reports and is a reasonable representative of that cluster.", + "summary": "Issues 43845, 43846, and 43859 look like the same vague duplicate report cluster with the same title; 43835 is a different title and does not fit the cluster well.", + "confidence": 0.66, + "canonical_issue_reason": "Issue 43845 is the earliest of the near-identical 'huggingface' reports and is the best representative of that duplicate cluster.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43845 is the best cluster anchor because it is the earliest matching-title report and has the clearest tie to the duplicate set.", + "best_issue_reason": "43845 is the strongest issue candidate because it is earliest and most central to the repeated same-title reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43835", "right": "issue:43845", "accept": false, - "reason": "Different title and no shared concrete details; this does not look like the same report." + "reason": "Different title ('Fett' vs 'huggingface') and no concrete evidence they describe the same underlying problem." }, { "left": "issue:43835", "right": "issue:43846", "accept": false, - "reason": "Different title and no shared concrete details; this does not look like the same report." + "reason": "Different title and no shared issue-specific details to justify treating them as duplicates." }, { "left": "issue:43835", "right": "issue:43859", "accept": false, - "reason": "Different title and no shared concrete details; this does not look like the same report." + "reason": "Different title with no substantive overlap beyond weak generic similarity." }, { "left": "issue:43845", "right": "issue:43846", "accept": true, - "reason": "Same title and closely matching metadata, so these are likely duplicate reports." + "reason": "Same title, nearly simultaneous creation, and very similar metadata; likely duplicate reports of the same issue." }, { "left": "issue:43845", "right": "issue:43859", "accept": true, - "reason": "Same title and broadly similar issue metadata, so these are likely duplicate reports." + "reason": "Same title and close enough pattern to plausibly be the same vague underlying report cluster." }, { "left": "issue:43846", "right": "issue:43859", "accept": true, - "reason": "Same title and broadly similar issue metadata, so these are likely duplicate reports." + "reason": "Same title and strongly overlapping duplicate-style reporting; likely the same issue." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The analysis is mostly grounded, but it overstates the evidence in a few places. The packet supports that 43845/43846/43859 share the exact title and have soft similarity, so they are plausibly related, but calling them duplicate reports of the same \"underlying placeholder issue\" goes beyond what is shown. Similarly, \"very similar metadata\" is a bit strong for 43859 given the differing body length and activity. 43835 being separate is a reasonable conservative call, but the cluster summary should stay closer to \"likely duplicate by identical title\" rather than asserting a specific placeholder/root-cause interpretation." + "accept": true, + "feedback": "The summary is grounded in the packet: 43845/43846/43859 share the same title, while 43835 has a different title. The duplicate-cluster wording is appropriately cautious given the lack of body/detail evidence, and the soft-edge verdicts stay conservative by not merging 43835 with the others." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -16145,7 +16287,7 @@ "issue:44572|issue:44937", "issue:44573|issue:44937" ], - "prepared_review_unit_hash": "75b98ae52c45c0c5061ad4b5e5ffdba295788261a31b23c1c176020ef525e98e", + "prepared_review_unit_hash": "204ef58bf658e85fe1e47c9c7f90c918b6507dcb591ecfffd3ac546147e0e705", "component_budget": { "node_count": 5, "item_count": 5, @@ -16167,66 +16309,66 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These look like unrelated low-signal issues (mostly spam/generic titles) rather than a true duplicate set. The only somewhat substantive item is the documentation enhancement, but it does not match the spam-like reports.", - "confidence": 0.14, + "summary": "This cluster does not look like a true duplicate set. It is mostly spam/placeholder issues, plus one unrelated documentation suggestion and one unrelated \u201cGoogle LLC\u201d post.", + "confidence": 0.94, "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue #44418 is the only non-spam, potentially actionable report; the others are generic/spam titles with no concrete bug details.", + "best_issue_reason": "Issue 44418 is the only non-spam, semi-descriptive item, so it is the best anchor if a representative must be chosen.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44371", "right": "issue:44572", "accept": false, - "reason": "Both are just '' with no concrete bug/change description; no evidence they refer to the same underlying issue beyond generic similarity." + "reason": "Both are titled spam placeholders, but there is no concrete underlying bug or change to unify." }, { "left": "issue:44371", "right": "issue:44573", "accept": false, - "reason": "Both are spam-like placeholders and lack any shared technical problem statement." + "reason": "Both are spam placeholders only; identical low-quality titles are not enough to treat them as the same issue." }, { "left": "issue:44371", "right": "issue:44418", "accept": false, - "reason": "A spam placeholder and a documentation enhancement are not the same underlying issue." + "reason": "Spam placeholder versus a documentation enhancement request; clearly different subjects." }, { "left": "issue:44371", "right": "issue:44937", "accept": false, - "reason": "Spam placeholder vs. an unrelated titled issue; no shared bug or change." + "reason": "Spam placeholder versus an unrelated \u201cGoogle LLC\u201d post; no shared bug or change." }, { "left": "issue:44418", "right": "issue:44572", "accept": false, - "reason": "Documentation enhancement suggestion does not match a spam placeholder." + "reason": "Documentation enhancement request versus spam placeholder; not the same issue." }, { "left": "issue:44418", "right": "issue:44573", "accept": false, - "reason": "Documentation enhancement vs. spam placeholder; no duplicate signal." + "reason": "Documentation enhancement request versus spam placeholder; not a duplicate pair." }, { "left": "issue:44572", "right": "issue:44573", "accept": false, - "reason": "Both are spam titles, but there is no concrete bug or request to confirm they are the same underlying report." + "reason": "Identical spam titles, but they do not indicate the same actionable underlying problem or change." }, { "left": "issue:44572", "right": "issue:44937", "accept": false, - "reason": "Spam placeholder and unrelated issue title; insufficient evidence of a duplicate." + "reason": "Spam placeholder versus unrelated \u201cGoogle LLC\u201d post; no concrete overlap." }, { "left": "issue:44573", "right": "issue:44937", "accept": false, - "reason": "Spam placeholder vs. unrelated issue title; not the same problem." + "reason": "Spam placeholder versus unrelated \u201cGoogle LLC\u201d post; not the same issue." } ] }, @@ -16256,7 +16398,7 @@ "soft_pairs": [ "issue:45145|issue:45151" ], - "prepared_review_unit_hash": "5595e119a7401bd51faa31f6e8c9339ba51eb947465168ac094319dae6b4883e", + "prepared_review_unit_hash": "59ca5d4b25b78e40611cf66b74bf046f7b0690ca9ebd1ce6bca3a6d6dba8dd96", "component_budget": { "node_count": 2, "item_count": 2, @@ -16278,27 +16420,27 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two closed issues share the exact same title and have moderately high soft-similarity evidence, so they are likely duplicates of the same energy-optimization request. The evidence supports grouping them, but the body-text similarity should be described conservatively since only lengths and similarity metrics are available.", - "confidence": 0.89, - "canonical_issue_reason": "issue #45145 is the better canonical issue because it was created first and already has an inbound reference, making it the more established thread for this duplicate pair.", + "summary": "These two issues appear to be duplicates of the same energy-optimization proposal/report: identical title, very similar body size, and strong textual overlap. Keep the earlier issue as canonical.", + "confidence": 0.99, + "canonical_issue_reason": "issue:45145 is the earlier-created issue and already has an inbound reference, so it is the best canonical representative for this duplicate pair.", "canonical_pr_reason": null, - "best_issue_reason": "issue #45145 is the strongest representative for the cluster: it is earlier, has the same title as the other issue, and has slightly more discussion context.", + "best_issue_reason": "issue:45145 is the strongest issue to keep because it predates 45151 and appears to be the original thread for the same content.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45145", "right": "issue:45151", "accept": true, - "reason": "Exact same title plus moderately high soft-similarity and no conflicting targets or filenames suggests the same underlying request; this is a reasonable duplicate edge, though the evidence is softer than text-level comparison." + "reason": "Same title and highly similar body suggest the same underlying issue/report, so this is a duplicate pair." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The cluster evidence supports that these issues share the exact same title and are likely related, but the analyst overstates the packet by claiming the bodies are \"highly similar\" and \"near-identical\"\u2014the packet only provides body lengths, not body text. The soft-edge accept is a bit too strong for soft-similarity-only evidence, so the verdict should be more conservative." + "accept": true, + "feedback": "The conclusion is grounded in the packet: the issues have identical titles, similar body lengths, and a soft-similarity link with Jaccard 0.63. Choosing the earlier issue as canonical is conservative and supported by the timestamps and inbound reference. Confidence is a bit high, but not clearly overstated given the available evidence." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -16376,40 +16518,40 @@ "soft_pairs": [ "issue:45672|issue:45674" ], - "prepared_review_unit_hash": "29ddf7fdfb94af5e5ad37c6aa9bd3b6e94b1001887906be121a7eef83c860e07", + "prepared_review_unit_hash": "d3e169eca8fd35257fd7be825b3ae56a6c534f7f194e8249ad0e42d99d8a1f8a", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1669, - "estimated_input_tokens": 418, - "estimated_eval_tokens": 1092 + "serialized_chars": 1668, + "estimated_input_tokens": 417, + "estimated_eval_tokens": 1090 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1669, - "estimated_input_tokens": 418, - "estimated_eval_tokens": 1092 + "serialized_chars": 1668, + "estimated_input_tokens": 417, + "estimated_eval_tokens": 1090 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These look like two separate 4-bit inference bugs, not duplicates: one is a Gemma4/audio-module uint8-weight `torch.finfo()` TypeError, the other is a BitsAndBytesConfig regression where `llm_int8_skip_modules` drops the default `lm_head` exclusion and triggers an AssertionError.", - "confidence": 0.94, - "canonical_issue_reason": "No single canonical duplicate here; if a representative is needed, issue #45674 is slightly cleaner because it describes a specific BitsAndBytesConfig regression with a clearer configuration-level fix path.", + "summary": "These are two distinct 4-bit inference bugs in different code paths: one in Gemma4 audio modules hitting uint8/torch.finfo, the other in BitsAndBytesConfig skip-module handling clearing lm_head exclusion. Not duplicates.", + "confidence": 0.97, + "canonical_issue_reason": "issue #45674 is the more general, framework-level 4-bit inference bug and has broader applicability than the Gemma4-specific audio-module failure.", "canonical_pr_reason": null, - "best_issue_reason": "Issue #45674 is the better cluster representative only in the sense that it is more narrowly scoped and actionable, but it is not the same bug as #45672.", + "best_issue_reason": "issue #45674 is the best representative because it describes a reusable configuration regression in BitsAndBytesConfig rather than a model-specific symptom.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45672", "right": "issue:45674", "accept": false, - "reason": "Different failure modes and code paths: Gemma4 audio modules hit a `torch.finfo()` TypeError on uint8 weights, while BitsAndBytesConfig loses the default `lm_head` exclusion and asserts in 4-bit inference. Same broad area, but not the same underlying bug." + "reason": "Different failure modes and code paths: Gemma4 audio-module torch.finfo on uint8 weights vs. BitsAndBytesConfig skip-module logic removing default lm_head exclusion. They are not the same underlying bug." } ] }, @@ -16439,7 +16581,7 @@ "soft_pairs": [ "issue:45762|issue:45763" ], - "prepared_review_unit_hash": "4183987cdaeb3358fbe2eaedba375396d59c8db057bb6f85b20fe1ff29f28fb8", + "prepared_review_unit_hash": "860048ddeafcdd98ccdb1af7e67d6136bff062433cb6264ba274278034db581f", "component_budget": { "node_count": 2, "item_count": 2, @@ -16461,18 +16603,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both artifacts look like unrelated personal-name posts with no bug report content, so they should not be clustered as duplicates.", - "confidence": 0.98, - "canonical_issue_reason": "If one must be chosen, issue 45762 is marginally earlier, but neither issue is a meaningful canonical bug report; both are unrelated personal-data-style posts.", + "summary": "The two closed issues appear to be unrelated spam/PII-style entries with different personal names and dates, not duplicate bug reports.", + "confidence": 0.97, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Neither issue is suitable as a canonical issue for triage because neither describes a reproducible problem or actionable feature request.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45762", "right": "issue:45763", "accept": false, - "reason": "Different names and birth dates; no shared bug, feature, or code context. This looks like unrelated spam/personal posts rather than duplicate issues." + "reason": "They only share a vague naming pattern; the titles contain different people and dates, with no evidence of the same underlying bug or request." } ] }, @@ -16502,7 +16644,7 @@ "soft_pairs": [ "issue:45764|issue:45765" ], - "prepared_review_unit_hash": "b9885288196eb89a5cc38a3f1a9d9a399c378602564dcf9c97ba386466193352", + "prepared_review_unit_hash": "fd14d972b6cc1fdbead2cf2f2fdb401625a9f012aa28e865a66fc2ff9aecfbd8", "component_budget": { "node_count": 2, "item_count": 2, @@ -16524,18 +16666,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both items look like unrelated low-information spam/personal-name issues, not duplicates of the same bug report.", - "confidence": 0.93, + "summary": "The two issues are unrelated title-only entries that look like personal-name/date spam, not the same bug or change.", + "confidence": 0.98, "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Neither issue has substantive bug content; if forced to keep one, either would be equally non-canonical because both are just name/date placeholders.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45764", "right": "issue:45765", "accept": false, - "reason": "The titles are different personal-name/date strings with no shared problem description, code path, or issue context; this does not look like the same underlying issue." + "reason": "No shared technical problem or code change; the titles are different personal-name/date strings and do not indicate duplicate bug reports." } ] }, @@ -16575,91 +16717,91 @@ "issue:45871|issue:45873", "issue:45872|issue:45873" ], - "prepared_review_unit_hash": "4d08cb7dc8da087a78757f572f0d903a71ecea0ba0cddefdc64ce15b31703fab", + "prepared_review_unit_hash": "eea9b421b4abed37db54a9b5e6fad5589dd9a2ccd59ba6824522fe695f55da64", "component_budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 8, - "serialized_chars": 5720, - "estimated_input_tokens": 1430, - "estimated_eval_tokens": 3116 + "serialized_chars": 5724, + "estimated_input_tokens": 1431, + "estimated_eval_tokens": 3118 }, "budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 8, - "serialized_chars": 5720, - "estimated_input_tokens": 1430, - "estimated_eval_tokens": 3116 + "serialized_chars": 5724, + "estimated_input_tokens": 1431, + "estimated_eval_tokens": 3118 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The cluster is not one single duplicate set: it contains two exact duplicate issue pairs (45870/45871 and 45872/45873) plus one unrelated financial dataset promo issue (45869). Only the identical-title pairs should be merged; 45869 should stand alone.", - "confidence": 0.89, - "canonical_issue_reason": "45870 is the clearest representative of the legitimate duplicate set 45870/45871 and is the earliest of that pair; 45872/45873 is a separate request/topic, and 45869 is unrelated.", + "summary": "The cluster appears to contain two likely duplicate issue pairs: 45870/45871 and 45872/45873. Their identical titles and identical body lengths make them strong near-duplicates, though the full bodies are not visible here. Issue 45869 is a separate finance-related request and should remain separate.", + "confidence": 0.94, + "canonical_issue_reason": "45872 is a good representative open issue because it is concrete and has a very strong likely-duplicate twin in 45873. The 45870/45871 pair also looks duplicated, but 45872/45873 is the more specific request within the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "45870 is the best overall issue to keep as a representative because it is an exact-duplicate pair member, cleanly worded, and the unrelated promo issue should not influence canonical selection.", + "best_issue_reason": "45872 is the best overall issue to keep as the representative item: it is specific, actionable, and its paired twin 45873 appears to be a likely duplicate. 45869 is different, and 45870/45871 are another likely duplicate pair but less distinctive.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45869", "right": "issue:45870", "accept": false, - "reason": "Different topics: a financial dataset promo vs. a DataForge example request. Shared finance wording is too broad to indicate the same issue." + "reason": "Same broad finance context, but this is a dataset bundle versus an example request; different underlying issue." }, { "left": "issue:45869", "right": "issue:45871", "accept": false, - "reason": "Different topics: a financial dataset promo vs. a DataForge example request." + "reason": "Related domain wording, but the request type is different; not the same issue." }, { "left": "issue:45870", "right": "issue:45871", "accept": true, - "reason": "Exact title match and same DataForge financial time-series example request; these look like duplicates of the same issue." + "reason": "Identical title and identical body length strongly suggest near-identical duplicate issues, even though the full bodies are not shown." }, { "left": "issue:45870", "right": "issue:45872", "accept": false, - "reason": "Both are DataForge/financial-time-series themed, but one is an example request and the other is a scraping-service request; not the same underlying issue." + "reason": "Both are DataForge-related, but one is an example request and the other is a scraping-service request; different underlying work." }, { "left": "issue:45870", "right": "issue:45873", "accept": false, - "reason": "Same as 45870/45872: related wording, but different concrete request and code path/topic." + "reason": "Shared finance/time-series theme only; the requests appear different and are not the same issue." }, { "left": "issue:45871", "right": "issue:45872", "accept": false, - "reason": "Different concrete requests: example prediction vs. scraping service." + "reason": "The pair shares only a broad subject area, not the same concrete request." }, { "left": "issue:45871", "right": "issue:45873", "accept": false, - "reason": "Different concrete requests: example prediction vs. scraping service." + "reason": "Different request types despite related wording; not a duplicate pair." }, { "left": "issue:45872", "right": "issue:45873", "accept": true, - "reason": "Exact title match and same DataForge financial time-series + scraping service request; these are duplicates." + "reason": "Identical title and identical body length make these look like near-identical duplicates, though the full body text is not available here." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet: 45870/45871 and 45872/45873 have identical titles and matching body lengths, while 45869 is only soft-similar and clearly a different topic. The canonical choice of 45870 as the representative of the earlier duplicate pair is conservative, and the soft-edge verdicts do not overstate the weaker matches." + "accept": false, + "feedback": "The clustering is plausible, but the writeup is a bit too strong for the packet. The visible evidence supports very high similarity from identical titles and identical body lengths for 45870/45871 and 45872/45873, but it does not show the full issue bodies, so phrases like \"exact duplicate\" and \"matching issue content\" overstate what is directly evidenced. Please soften to \"likely duplicates\" or \"near-identical duplicates\" unless fuller content is available. The 45869 separation looks grounded." }, "evaluator_used": true, - "retried": false, + "retried": true, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -16667,7 +16809,7 @@ { "label": "pull_request", "component_index": 1, - "component_count": 27, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -16710,148 +16852,148 @@ "pull_request:43823|pull_request:45213", "pull_request:43251|pull_request:43254" ], - "prepared_review_unit_hash": "74eb2728e044b6185e9c8fb4cd9ba105080c315bd84207d4b24aa467ab49b1b9", + "prepared_review_unit_hash": "35fb03b62c7d2aaa2ab84d6264f1424a245df4507f225aa27cc620ec25e6e9b3", "component_budget": { "node_count": 12, "item_count": 12, "soft_pair_count": 18, - "serialized_chars": 32876, - "estimated_input_tokens": 8219, - "estimated_eval_tokens": 16694 + "serialized_chars": 32877, + "estimated_input_tokens": 8220, + "estimated_eval_tokens": 16696 }, "budget": { "node_count": 12, "item_count": 12, "soft_pair_count": 18, - "serialized_chars": 32876, - "estimated_input_tokens": 8219, - "estimated_eval_tokens": 16694 + "serialized_chars": 32877, + "estimated_input_tokens": 8220, + "estimated_eval_tokens": 16696 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is mostly heterogeneous PRs spanning unrelated model additions, docs, tests, refactors, and bug fixes. The only clear duplicate-like pair is the two `fixed_cross_entropy` patches (`#43251` and `#43254`), which target the same code path and same tracked issue.", - "confidence": 0.9, + "summary": "This cluster is mostly heterogeneous: model additions, processor/tokenizer fixes, tests, and unrelated refactors. The only true duplicate-like pair is the two `fixed_cross_entropy` PRs targeting the same `loss_utils.py` bug.", + "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": "`pull_request:43251` is the best canonical PR for the only true duplicate pair: it is issue-linked, touches the same `fixed_cross_entropy` path as `#43254`, and is the more concrete/complete fix. The rest of the cluster is not one cohesive duplicate set.", + "canonical_pr_reason": "No single cluster-wide canonical PR fits because most items are unrelated. If forced to pick the nearest duplicate anchor, PR 43251 is the better canonical for the loss-utils fix since it directly targets issue 43240 and implements the broader cross-entropy kwargs pass-through.", "best_issue_reason": null, - "best_pr_reason": "`pull_request:43251` is the strongest representative artifact in the cluster because it has an explicit issue target and matches the only accepted duplicate-like change; most other PRs are unrelated and should not be merged into a single canonical change.", + "best_pr_reason": "PR 43251 is the strongest representative of the only real duplicate cluster: it is tied to an explicit issue, touches the shared bug in `fixed_cross_entropy`, and is the broader implementation of that fix.", "soft_edge_verdicts": [ { "left": "pull_request:39895", "right": "pull_request:43424", "accept": false, - "reason": "Unrelated PRs: VideoPrism model addition vs Executorch dynamic-shape test. Different subsystems and no shared bug/change." + "reason": "Videoprism model integration and an executorch dynamic-shape test are unrelated changes." }, { "left": "pull_request:41855", "right": "pull_request:43823", "accept": false, - "reason": "Tokenizer missing-methods patch vs adding MobileLLM model support. Same broad area of model code, but not the same underlying change." + "reason": "A Mistral tokenizer method addition and a new MobileLLM model PR do not address the same bug or change." }, { "left": "pull_request:41855", "right": "pull_request:43424", "accept": false, - "reason": "Mistral tokenizer helper methods vs Executorch export test; no shared code path or fix." + "reason": "Tokenizer utility work and executorch export tests are unrelated." }, { "left": "pull_request:42668", "right": "pull_request:43823", "accept": false, - "reason": "Processor robustness changes vs MobileLLM model addition. Different concrete changes and no same bug." + "reason": "Processor auto-loading robustness and adding MobileLLM are different changes." }, { "left": "pull_request:43251", "right": "pull_request:43823", "accept": false, - "reason": "Cross-entropy utility fix vs MobileLLM model addition. Completely different functionality." + "reason": "Cross-entropy kwargs support and MobileLLM model addition are unrelated code paths." }, { "left": "pull_request:43251", "right": "pull_request:44827", "accept": false, - "reason": "`fixed_cross_entropy` kwargs fix vs Mistral4 test/model changes. Not the same code path or issue." + "reason": "The loss-utils fix and Mistral4 test updates do not look like the same bug or change." }, { "left": "pull_request:43424", "right": "pull_request:44827", "accept": false, - "reason": "Executorch dynamic-shape test vs Mistral4 test/model fix. Both are tests-heavy, but they are not the same underlying bug/change." + "reason": "Executorch dynamic-shape export testing and Mistral4 test fixes are unrelated." }, { "left": "pull_request:43424", "right": "pull_request:43823", "accept": false, - "reason": "Executorch export test vs MobileLLM model addition. No concrete overlap beyond general framework work." + "reason": "Executorch export test coverage and MobileLLM model scaffolding are unrelated." }, { "left": "pull_request:43823", "right": "pull_request:45655", "accept": false, - "reason": "Model addition PR vs config resolution-order fix. Different layers and unrelated changes." + "reason": "Adding MobileLLM and changing config-class resolution order are separate changes." }, { "left": "pull_request:43823", "right": "pull_request:43995", "accept": false, - "reason": "MobileLLM model addition vs Falcon output-interface refactor. No same bug or mergeable combined change." + "reason": "MobileLLM model addition and Falcon output-collection refactoring are unrelated." }, { "left": "pull_request:43995", "right": "pull_request:44827", "accept": false, - "reason": "Falcon interface refactor vs Mistral4 test/model changes. Same broad modeling area, but not the same concrete problem." + "reason": "Falcon refactoring and Mistral4 test fixes address different code paths." }, { "left": "pull_request:43995", "right": "pull_request:45213", "accept": false, - "reason": "Falcon refactor vs model-creation skill scaffolding. Unrelated artifacts." + "reason": "A Falcon model refactor and the model-creation skill PR are not the same underlying change." }, { "left": "pull_request:44827", "right": "pull_request:45170", "accept": false, - "reason": "Mistral4 fixes vs `layrnorm`\u2192`layernorm` renaming across CLIP-like models. Different bugs and different model families." + "reason": "Mistral4 test fixes and the `layrnorm` typo normalization patch are different problems." }, { "left": "pull_request:44827", "right": "pull_request:45655", "accept": false, - "reason": "Mistral4 work vs config resolution-order fix. No shared code-path issue." + "reason": "Mistral4 tests and config resolution-order fixes are unrelated." }, { "left": "pull_request:43823", "right": "pull_request:44827", "accept": false, - "reason": "MobileLLM model addition vs Mistral4 fixes. Both are model PRs, but not the same underlying change." + "reason": "MobileLLM addition and Mistral4 test changes are unrelated." }, { "left": "pull_request:44827", "right": "pull_request:45213", "accept": false, - "reason": "Mistral4 fixes vs model-creation skill repository changes. Not the same artifact type or bug." + "reason": "Mistral4 tests and the do-not-merge model-creation skill are unrelated." }, { "left": "pull_request:43823", "right": "pull_request:45213", "accept": false, - "reason": "MobileLLM model addition vs model-creation skill. Unrelated." + "reason": "MobileLLM model addition and the model-creation skill PR are not the same change." }, { "left": "pull_request:43251", "right": "pull_request:43254", "accept": true, - "reason": "Both PRs patch the same `fixed_cross_entropy` function for the same issue (`#43240`) and add support for the same kwargs (`weight`/`label_smoothing`). They are direct alternate implementations of the same fix." + "reason": "Both PRs fix the same `loss_utils.fixed_cross_entropy` issue in `nn.functional.cross_entropy`, share issue target 43240, and could plausibly be merged into one PR." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet: the cluster is heterogeneous, and the only explicitly supported duplicate-like pair is pull_request:43251 / pull_request:43254, which share the same file and issue target (43240). The verdicts are conservative, and no unsupported mergeability claims were made for unrelated PRs." + "feedback": "The assessment is grounded in the packet and stays conservative. The only accepted pair, PR 43251 and PR 43254, is legitimately mergeable as the same `fixed_cross_entropy`/`nn.functional.cross_entropy` bugfix with the shared issue target 43240. The remaining soft-edge rejections are appropriately cautious given the clearly unrelated file sets and PR intents. No evidence is overstated." }, "evaluator_used": true, "retried": false, @@ -16862,7 +17004,7 @@ { "label": "pull_request", "component_index": 2, - "component_count": 27, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -16878,12 +17020,12 @@ "soft_pairs": [ "pull_request:40546|pull_request:42781" ], - "prepared_review_unit_hash": "940ae0013b39d8bffa99e4bd7d20e09d1ef22c0c35fed567cda66bc32fd3d051", + "prepared_review_unit_hash": "2a4f971bef35e9e29e22f5684893aa54c314ffabd62d5f7a31ecf85db72f43c2", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 7226, + "serialized_chars": 7225, "estimated_input_tokens": 1807, "estimated_eval_tokens": 3870 }, @@ -16891,7 +17033,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 7226, + "serialized_chars": 7225, "estimated_input_tokens": 1807, "estimated_eval_tokens": 3870 }, @@ -16900,18 +17042,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open PRs for VibeVoice, but they are different changes: one adds the base VibeVoice integration, the other adds a realtime/acoustic-tokenizer extension. They overlap on shared model-family plumbing, but they are not the same concrete fix/change.", - "confidence": 0.91, + "summary": "These are related VibeVoice PRs, but they are not duplicates: one is the initial VibeVoice integration, the other adds realtime/VibeVoice acoustic-tokenizer support and related generation changes.", + "confidence": 0.92, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #40546 is the foundational VibeVoice implementation and the broader base integration for the model family.", + "canonical_pr_reason": "PR 40546 is the broader base implementation of VibeVoice and establishes the core integration that the later realtime work builds on.", "best_issue_reason": null, - "best_pr_reason": "PR #40546 is the best representative of the cluster because it is the core implementation PR rather than a follow-on realtime extension.", + "best_pr_reason": "PR 40546 is the best standalone PR to represent the cluster because it introduces the primary VibeVoice model support; PR 42781 is a separate follow-up feature set for realtime/tokenizer support.", "soft_edge_verdicts": [ { "left": "pull_request:40546", "right": "pull_request:42781", "accept": false, - "reason": "Reject: #40546 implements baseline VibeVoice support, while #42781 adds realtime support plus an acoustic tokenizer and generation changes. They touch related files but solve different underlying feature sets and would not plausibly be merged as one PR." + "reason": "Do not merge: 40546 implements the base VibeVoice model, while 42781 adds realtime-specific functionality plus acoustic-tokenizer and generation updates. They touch the same family of files but address different concrete features rather than the same underlying change." } ] }, @@ -16925,7 +17067,7 @@ { "label": "pull_request", "component_index": 3, - "component_count": 27, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -16941,7 +17083,7 @@ "soft_pairs": [ "pull_request:40755|pull_request:41224" ], - "prepared_review_unit_hash": "ddafc6f6cd107079cd47312bd5e28854b2bac91ef7951a57c0919b51d6bd845c", + "prepared_review_unit_hash": "bb6639e9c7fdec288ce81d466e59e744ae68874d98fb7d566e7b2f995dba8903", "component_budget": { "node_count": 2, "item_count": 2, @@ -16963,18 +17105,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two PRs are unrelated model-feature additions in different areas: TimesFM covariate forecasting vs. DINOv3 image-classification support. They do not look like the same underlying change and should not be merged as duplicates.", - "confidence": 0.97, + "summary": "These two pull requests are unrelated: one adds covariate-based forecasting support for TimesFM, while the other adds image-classification support for DINOv3ViT. They do not describe the same bug or change and should not be deduplicated.", + "confidence": 0.99, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR as a duplicate cluster is justified here; the PRs target different models and different code paths. If one must be chosen, PR 41224 is the more self-contained and narrowly scoped change.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR 41224 is the better standalone PR candidate because it is tightly focused on one concrete capability (DINOv3ViT image classification) with a smaller, cleaner scope. PR 40755 is broader and touches more files for a different feature area.", + "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "pull_request:40755", "right": "pull_request:41224", "accept": false, - "reason": "Different model families and different user-facing features: TimesFM covariate forecasting versus DINOv3 image classification. They do not fix the same bug or implement the same change, so this is not a duplicate or mergeable soft match." + "reason": "Different models and code paths (TimesFM forecasting with covariates vs DINOv3 image classification support); no shared target issue or filenames, and they could not plausibly be merged into one PR." } ] }, @@ -16988,7 +17130,7 @@ { "label": "pull_request", "component_index": 4, - "component_count": 27, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -17051,7 +17193,7 @@ { "label": "pull_request", "component_index": 5, - "component_count": 27, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -17069,7 +17211,7 @@ "pull_request:41797|pull_request:45186", "pull_request:45186|pull_request:45630" ], - "prepared_review_unit_hash": "ea0e68b68a54c8c6f5085bc5891d6ed73ca3b142e34038aeb96674340886fe12", + "prepared_review_unit_hash": "456cf687b93956a56d097e9651a6ec0e562d3c10a8f51d169fdf941f380bf54a", "component_budget": { "node_count": 3, "item_count": 3, @@ -17091,24 +17233,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are three unrelated model-addition PRs that only overlap on boilerplate registration/docs changes. They do not describe the same underlying bug or the same concrete code-path fix, so none should be deduplicated together.", - "confidence": 0.97, + "summary": "All three items are separate new-model addition PRs, not duplicates. They share scaffold files and auto-registry touchpoints, but each introduces a different model family and codepath.", + "confidence": 0.92, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "PR 45630 is the best cluster anchor: it has the highest review activity and the broadest integration surface, but it is still not a duplicate of the others.", "best_issue_reason": null, - "best_pr_reason": "PR 45630 is the strongest standalone anchor only because it has the broadest model-integration surface and highest review activity, but it is still not a duplicate of the others.", + "best_pr_reason": "PR 45630 is the most substantial and best-documented representative of the cluster, though the cluster members are functionally unrelated.", "soft_edge_verdicts": [ { "left": "pull_request:41797", "right": "pull_request:45186", "accept": false, - "reason": "Different models and distinct integrations: DeepSeek OCR vs Isaac. Shared auto-registration/docs scaffolding is generic template overlap, not the same change." + "reason": "DeepSeek OCR vs Isaac are different model additions with different docs/tests and implementation codepaths; shared boilerplate files are not enough to treat them as the same change." }, { "left": "pull_request:45186", "right": "pull_request:45630", "accept": false, - "reason": "Isaac and Kimi2-6 are different model additions with different files and code paths. Similar repository plumbing is expected and does not indicate the same fix." + "reason": "Isaac and Kimi2-6 add different models and different supporting modules; they only overlap on standard registry/docs scaffolding, so they are not duplicates." } ] }, @@ -17122,7 +17264,7 @@ { "label": "pull_request", "component_index": 6, - "component_count": 27, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -17138,7 +17280,7 @@ "soft_pairs": [ "pull_request:42975|pull_request:42976" ], - "prepared_review_unit_hash": "393ee4b59c15f17ccf3e954bba5e0dbe4c18416e974d88ccb50ccc459ffa791d", + "prepared_review_unit_hash": "b57585c1a91e6554c1f095fb10763e606d44856dba816765ef42584aa799786b", "component_budget": { "node_count": 2, "item_count": 2, @@ -17160,18 +17302,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two independent workflow-maintenance PRs: one upgrades core GitHub Actions for Node 24 compatibility, the other bumps Docker build actions and a few related workflow actions. They overlap in CI workflows but do not fix the same underlying change.", - "confidence": 0.92, + "summary": "Two open PRs both update GitHub Actions versions in workflow files, but they address different upgrade tracks: one is specifically for Node 24 compatibility, the other is a broader \u201clatest versions\u201d refresh centered on docker/build-push-action. They overlap in files but not in the underlying change intent.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 42975 is the clearer representative because its scope and intent are explicit: updating major GitHub Actions versions to keep workflows Node 24 compatible across many files.", + "canonical_pr_reason": "PR 42975 is the more specific and representative workflow-maintenance change: it targets Node 24 compatibility and updates several core GitHub Actions pins across many workflows.", "best_issue_reason": null, - "best_pr_reason": "PR 42975 is the best cluster representative since it captures the broader, more clearly stated workflow runtime-compatibility upgrade rather than a narrower action-version bump.", + "best_pr_reason": "PR 42975 is the best representative of the cluster because its title and edits align cleanly around one concrete maintenance goal, whereas PR 42976 is a separate, broader action-version bump.", "soft_edge_verdicts": [ { "left": "pull_request:42975", "right": "pull_request:42976", "accept": false, - "reason": "Both are workflow dependency upgrades, but they change different actions and different operational paths (checkout/cache/upload-artifact vs docker/build-push-action). They are similar maintenance, not the same concrete fix and not mergeable as one single PR without conflating distinct upgrades." + "reason": "Reject: these are both GitHub Actions upgrade PRs, but they fix different things. 42975 is about Node 24 compatibility for checkout/cache/upload-artifact, while 42976 updates docker/build-push-action and a few related workflow actions. They share workflow files but not the same concrete bug/change." } ] }, @@ -17185,7 +17327,7 @@ { "label": "pull_request", "component_index": 7, - "component_count": 27, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -17214,12 +17356,12 @@ "pull_request:44601|pull_request:45838", "pull_request:45490|pull_request:45838" ], - "prepared_review_unit_hash": "1fef83008e319e38db4e846806164ac4714ced274232cf677a28c3c411ac25b6", + "prepared_review_unit_hash": "4a63e5a20df5f4a80837f5a2f06f80d488cade6308509f691f136d018d780c60", "component_budget": { "node_count": 7, "item_count": 7, "soft_pair_count": 9, - "serialized_chars": 23541, + "serialized_chars": 23542, "estimated_input_tokens": 5886, "estimated_eval_tokens": 12028 }, @@ -17227,7 +17369,7 @@ "node_count": 7, "item_count": 7, "soft_pair_count": 9, - "serialized_chars": 23541, + "serialized_chars": 23542, "estimated_input_tokens": 5886, "estimated_eval_tokens": 12028 }, @@ -17236,66 +17378,66 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are mostly unrelated model-add or infrastructure PRs that only overlap on generic Transformers plumbing (docs, auto mappings, model registration). No pair looks like the same underlying bug/change, so none of the soft edges should be merged.", - "confidence": 0.97, + "summary": "This cluster does not contain duplicate PRs; it mixes unrelated model-addition PRs and one distributed-loading PR. The shared file patterns are mostly boilerplate (docs/toctree, auto registries), not evidence of the same underlying change.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "Add Molmo2 is the strongest standalone PR in this set: it has the broadest model/processor/video coverage and the highest review activity, so it is the best representative if one PR must be chosen.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "Add Molmo2 is the most complete and actively reviewed of the unrelated PRs, making it the best overall PR candidate in this cluster even though it is not a duplicate of the others.", + "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "pull_request:42978", "right": "pull_request:43448", "accept": false, - "reason": "ViT NEPA and Molmo are different models with different code paths; shared docs/auto-registration files are generic model-add boilerplate." + "reason": "Different models: ViT NEPA vs Molmo. Overlap is only generic model-registration/docs files, not the same code-path or feature." }, { "left": "pull_request:42978", "right": "pull_request:43451", "accept": false, - "reason": "ViT NEPA vs Molmo2 is a different model family and a different implementation, not the same change." + "reason": "Different models: ViT NEPA vs Molmo2. Shared auto/docs files are boilerplate for adding new architectures, not a duplicate change." }, { "left": "pull_request:42978", "right": "pull_request:45490", "accept": false, - "reason": "ViT NEPA and CTSM are unrelated model additions; overlap is only in common Transformers registration/docs." + "reason": "Different models and purposes: ViT NEPA vs CTSM. Same broad pattern of adding a model, but not the same underlying bug or implementation." }, { "left": "pull_request:43448", "right": "pull_request:43451", "accept": false, - "reason": "Molmo and Molmo2 are distinct model additions with different files, processors, and tests; similar naming is not enough to treat them as duplicates." + "reason": "Molmo and Molmo2 are distinct model families/releases with different files and implementations; not mergeable as one PR." }, { "left": "pull_request:43448", "right": "pull_request:45490", "accept": false, - "reason": "Molmo and CTSM target different model architectures/tasks and do not fix the same concrete code-path problem." + "reason": "Molmo vs CTSM are unrelated model additions; shared registry/docs edits are incidental." }, { "left": "pull_request:43451", "right": "pull_request:45490", "accept": false, - "reason": "Molmo2 and CTSM are unrelated model integrations; they only share generic auto/model registration changes." + "reason": "Molmo2 and CTSM are unrelated architectures; no concrete shared bug or code path." }, { "left": "pull_request:43973", "right": "pull_request:45838", "accept": false, - "reason": "Lfm2 Audio and PP-OCRv6 text recognition are different model domains and code paths; no same-bug/same-change evidence." + "reason": "Lfm2 Audio and PP-OCRv6 text recognition are completely different modalities and model stacks; the overlap is just new-model scaffolding." }, { "left": "pull_request:44601", "right": "pull_request:45838", "accept": false, - "reason": "Native pipeline-parallel loading support is infrastructure work, while PP-OCRv6 is a model-support PR; they are not the same concrete change." + "reason": "One is distributed pipeline-parallel loading infrastructure, the other is a model-support PR. They do not fix the same concrete problem." }, { "left": "pull_request:45490", "right": "pull_request:45838", "accept": false, - "reason": "CTSM and PP-OCRv6 are different model additions with different targets and implementations, so they should not be merged as duplicates." + "reason": "CTSM and PP-OCRv6 are unrelated model additions; any similarity is only in shared auto/docs plumbing." } ] }, @@ -17309,7 +17451,7 @@ { "label": "pull_request", "component_index": 8, - "component_count": 27, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -17328,7 +17470,7 @@ "pull_request:43291|pull_request:44815", "pull_request:43665|pull_request:44815" ], - "prepared_review_unit_hash": "c2376e88fa5d99d5f6c880c873529fbac0f69a4448383fb84d21d81ad6f4d549", + "prepared_review_unit_hash": "d26ab0ef44374118fa7e3068487b43f0936f87b2457c4d9f3c9390736acbfc0a", "component_budget": { "node_count": 3, "item_count": 3, @@ -17350,30 +17492,30 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These three PRs are not duplicates: one is Whisper tokenizer/test work, one is unrelated test-only noise in CLIP/ViT, and one is an FP8 dequant/loading fix. There is no single underlying bug or change tying them together.", - "confidence": 0.96, + "summary": "These pull requests do not look like duplicates of the same change. PR 44815 is a substantive FP8 dequant/loading fix, PR 43291 is a Whisper tokenizer/test fix, and PR 43665 appears to be test-only noise for CLIP/ViT rather than a real bug fix.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "44815 is the most substantive standalone fix and the clearest anchor, but the cluster is heterogeneous rather than a true duplicate set.", + "canonical_pr_reason": "PR 44815 is the strongest canonical candidate because it contains the clearest concrete code-path fix with implementation changes across loading/FP8 logic and tests.", "best_issue_reason": null, - "best_pr_reason": "44815 is the best representative because it contains the real implementation change; 43291 is narrow Whisper test/decoder work and 43665 is only test assertions.", + "best_pr_reason": "PR 44815 is the best representative of a real fix in the cluster; the other PRs are either unrelated or mostly test-focused.", "soft_edge_verdicts": [ { "left": "pull_request:43291", "right": "pull_request:43665", "accept": false, - "reason": "Whisper tokenizer/decoder changes with tests vs. unrelated CLIP/ViT test assertions; no shared concrete bug or code path." + "reason": "Whisper tokenizer decoding/tests versus CLIP/ViT test assertions; different models and different bugs, so not the same underlying change." }, { "left": "pull_request:43291", "right": "pull_request:44815", "accept": false, - "reason": "Whisper decoding/test fix vs. FP8 dequant/model-loading changes; different subsystems and unrelated fixes." + "reason": "Whisper decode behavior fix and FP8/dequant loading fix are unrelated code paths with no shared concrete bug." }, { "left": "pull_request:43665", "right": "pull_request:44815", "accept": false, - "reason": "Test-only CLIP/ViT PR vs. dequantization/loading implementation fix; not the same underlying issue." + "reason": "One is test-only CLIP/ViT edits, the other is a broad FP8/dequant implementation fix; they are not mergeable as one duplicate fix." } ] }, @@ -17387,7 +17529,7 @@ { "label": "pull_request", "component_index": 9, - "component_count": 27, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -17403,12 +17545,12 @@ "soft_pairs": [ "pull_request:43488|pull_request:43532" ], - "prepared_review_unit_hash": "799f70126c7021e3fb9a7cb5da26b2cc50cfa1aec2c9c2117e5dedec456c7918", + "prepared_review_unit_hash": "e207655d40f016347a1892651e611f647ab65beb4d13eda55b00055b9e21c9af", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4563, + "serialized_chars": 4562, "estimated_input_tokens": 1141, "estimated_eval_tokens": 2538 }, @@ -17416,7 +17558,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4563, + "serialized_chars": 4562, "estimated_input_tokens": 1141, "estimated_eval_tokens": 2538 }, @@ -17425,18 +17567,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two pull requests are not duplicates: they share only a Vit test file, but one is a small setup/version/dependency change while the other rewires CI/workflow and testing utility code.", - "confidence": 0.97, + "summary": "These two pull requests are not duplicates: one is a repo-bot/setup/version-format test PR, while the other rewires CI/comment workflows and touches testing/notification code. The only shared file is a test file, which is not enough to treat them as the same change.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43488 is the more self-contained change, centered on setup/version and dependency-table edits; PR 43532 is a broader CI/testing refactor and not the same concrete fix.", + "canonical_pr_reason": "No canonical PR: the PRs address different code paths and purposes, so there is no single representative duplicate target here.", "best_issue_reason": null, - "best_pr_reason": "PR 43488 is the closest thing to a canonical representative because it has a single, focused code-path change; PR 43532 spans workflows, utilities, and tests and is not a good duplicate anchor.", + "best_pr_reason": "PR #43532 is the better standalone representative of a substantive code change because it spans CI/workflow and related utility updates, whereas #43488 is mainly a bot/check-oriented test PR.", "soft_edge_verdicts": [ { "left": "pull_request:43488", "right": "pull_request:43532", "accept": false, - "reason": "Shared Vit test filename alone is not enough; 43488 changes setup.py/version and dependency metadata, while 43532 changes CI workflows, testing utilities, and reporting code. They do not fix the same bug or belong in one merged PR." + "reason": "Different underlying changes: #43488 adjusts setup/dependency/version metadata and a vit test; #43532 changes workflow triggers plus testing and notification utilities. Shared test filename is incidental." } ] }, @@ -17450,7 +17592,7 @@ { "label": "pull_request", "component_index": 10, - "component_count": 27, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -17466,12 +17608,12 @@ "soft_pairs": [ "pull_request:43612|pull_request:43613" ], - "prepared_review_unit_hash": "97f8eb91b7fd34be160335d2606f7572c567fcb55000d5f1ffa3cac392fb585e", + "prepared_review_unit_hash": "4f3bd470d5da9da6dac3297bfe7c7ff4cad64e0d30d05f0a5aa125c5bd661219", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6083, + "serialized_chars": 6082, "estimated_input_tokens": 1521, "estimated_eval_tokens": 3298 }, @@ -17479,7 +17621,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6083, + "serialized_chars": 6082, "estimated_input_tokens": 1521, "estimated_eval_tokens": 3298 }, @@ -17488,18 +17630,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open PRs add different new pipelines: one for promptable concept segmentation and one for promptable visual segmentation. They share documentation/registry plumbing, but they target different models, files, and task semantics, so they are not duplicates.", - "confidence": 0.97, + "summary": "These are two separate feature PRs for different pipelines: promptable concept segmentation vs promptable visual segmentation. They share a common implementation pattern and some documentation/registration files, but they do not fix the same underlying bug or change the same concrete code path.", + "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "No single canonical PR for deduplication: the PRs are distinct feature additions targeting different pipelines and model sets.", "best_issue_reason": null, - "best_pr_reason": "PR 43613 is the broader feature addition and touches more of the segmentation/model integration surface, so it is the better representative of this cluster, but it is not a duplicate of 43612.", + "best_pr_reason": "PR 43613 is the broader representative if one must be chosen, since it covers more related model integrations and tests, but it is still not a duplicate of 43612.", "soft_edge_verdicts": [ { "left": "pull_request:43612", "right": "pull_request:43613", "accept": false, - "reason": "Different features: concept segmentation vs visual segmentation. Shared docs/registry changes are infrastructure overlap, not the same underlying code-path fix, and they would not plausibly be merged as one PR." + "reason": "Both add different pipeline features with different names, docs, tests, and model integrations; overlap is broad infrastructure, not the same concrete change, so they should not be merged as duplicates." } ] }, @@ -17513,7 +17655,7 @@ { "label": "pull_request", "component_index": 11, - "component_count": 27, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -17529,7 +17671,7 @@ "soft_pairs": [ "pull_request:43651|pull_request:43663" ], - "prepared_review_unit_hash": "e1a98475f0b3741056a828e3c7352e84c7290122103281b2860cf76e28295d44", + "prepared_review_unit_hash": "d239e1c3b0d5174477fa3aae9cba35c07573f35e8b1b57342999a94ea5c77370", "component_budget": { "node_count": 2, "item_count": 2, @@ -17551,18 +17693,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two separate Trainer extension-point PRs: one changes gradient-accumulation loss scaling, the other changes signature-column filtering for custom trainers. They share a file but not the same bug or mergeable change.", + "summary": "The two PRs both edit `trainer.py`, but they implement different trainer extensibility changes: one alters gradient-accumulation loss scaling, the other factors out signature-column filtering. Shared file/subsystem alone is not enough to treat them as duplicates.", "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "No clear canonical duplicate: PR #43651 changes loss scaling behavior in `training_step`, while PR #43663 adds an overridable helper for signature-column filtering. They target different code paths and behaviors.", "best_issue_reason": null, - "best_pr_reason": "Neither PR is a good canonical representative for a duplicate cluster here; they address different code paths and should be triaged independently.", + "best_pr_reason": "PR #43651 is the better representative only in a loose triage sense because it is a focused, concrete trainer behavior change; however, it is not a duplicate of PR #43663.", "soft_edge_verdicts": [ { "left": "pull_request:43651", "right": "pull_request:43663", "accept": false, - "reason": "Different underlying changes: loss scaling in training_step vs. signature-column extraction for data filtering. Same file only, not the same fix." + "reason": "Different changes in the same file: #43651 introduces `_loss_is_scaled_for_ga` and modifies gradient-accumulation loss scaling, while #43663 adds `_get_signature_columns` for column filtering. They do not fix the same underlying bug or implement the same change." } ] }, @@ -17576,7 +17718,7 @@ { "label": "pull_request", "component_index": 12, - "component_count": 27, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -17594,52 +17736,52 @@ "pull_request:43656|pull_request:43842", "pull_request:43836|pull_request:43842" ], - "prepared_review_unit_hash": "7e5c74b169175b4502aaebc13b6aafb2e4a5b243ef991dfc0ac656cc7a6a7017", + "prepared_review_unit_hash": "78c95989db646c7fbb6fe2b928b0b2f8d1ea38de6924156900794a5cc5d20890", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 5991, - "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": 5991, - "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 target the same CLI serve.py TypeAdapter annotation failure, with 43842 being the most focused fix and the best representative of the cluster.", - "confidence": 0.82, + "summary": "All three PRs touch `src/transformers/cli/serve.py` and aim at the same runtime annotation/`TypeAdapter` import-time NameError family. PR #43842 is the clearest, most targeted fix; #43836 is a close duplicate variant; #43656 appears to address the same underlying CLI annotation NameError more broadly.", + "confidence": 0.84, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43842 most directly addresses the concrete import/type-annotation failure in `src/transformers/cli/serve.py` with the smallest, most focused change.", + "canonical_pr_reason": "PR #43842 is the most explicit and focused fix: it directly addresses `TypeAdapter` being unavailable at runtime when `pydantic` is not installed, with a minimal change set and a clear title.", "best_issue_reason": null, - "best_pr_reason": "PR 43842 is the clearest standalone fix for the `TypeAdapter` NameError in the CLI serve code path and is the best canonical representative for this cluster.", + "best_pr_reason": "PR #43842 is the best representative because it is the most direct, low-risk, and narrowly scoped solution to the same `TypeAdapter` NameError in the CLI.", "soft_edge_verdicts": [ { "left": "pull_request:43656", "right": "pull_request:43842", "accept": true, - "reason": "Both are fixes for the same `TypeAdapter` NameError in `src/transformers/cli/serve.py`; 43656 is broader, but the core bug and code path overlap enough that they could plausibly be merged into one PR." + "reason": "Both PRs target the same `serve.py` annotation/import-time NameError in the CLI and are alternative fixes for the same underlying runtime typing problem." }, { "left": "pull_request:43836", "right": "pull_request:43842", "accept": true, - "reason": "Both PRs address the same `TypeAdapter` annotation/runtime-name-resolution problem in the CLI serve path, with 43836 being a broader variant of the same fix rather than a separate issue." + "reason": "Both PRs explicitly target the same `TypeAdapter` NameError in `src/transformers/cli/serve.py`; #43836 is just a broader/rougher variant of the same fix." } ] }, "evaluator_result": { "accept": true, - "feedback": "The cluster is grounded in the packet: all three PRs modify `src/transformers/cli/serve.py` around the same annotation/runtime-name-resolution area, and 43842 is the most focused representative. The pairwise accepts are plausible as mergeable fixes for the same bug family, with 43842 being the narrowest change." + "feedback": "The grouping is grounded in the packet: all three PRs modify `src/transformers/cli/serve.py` and their titles/diffs point to the same TypeAdapter-related NameError class. The 43836\u201343842 link is especially strong because they share issue target 43824. The 43656 link is looser, but still plausible based on the title and diff preview, so the soft-edge verdicts stay within a conservative interpretation." }, "evaluator_used": true, "retried": false, @@ -17650,7 +17792,7 @@ { "label": "pull_request", "component_index": 13, - "component_count": 27, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -17716,7 +17858,7 @@ { "label": "pull_request", "component_index": 14, - "component_count": 27, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 2, "status": "reviewed", @@ -17772,12 +17914,12 @@ "pull_request:44019|pull_request:44085", "pull_request:44018|pull_request:44019", "pull_request:44019|pull_request:44071", - "pull_request:44068|pull_request:44722", "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:44085", "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", @@ -17828,12 +17970,12 @@ "pull_request:44044|pull_request:44129", "pull_request:44044|pull_request:44154" ], - "prepared_review_unit_hash": "41df2495cb655a9977f77b2cb255b21ccc87a4b4d57b5726481832639eb9ed7f", + "prepared_review_unit_hash": "55e6b5a1a451cf1c254e32e2870c6486cfb51cba7adc86d49a79ae7aa7ebde7e", "component_budget": { "node_count": 28, "item_count": 28, "soft_pair_count": 78, - "serialized_chars": 84897, + "serialized_chars": 84898, "estimated_input_tokens": 21225, "estimated_eval_tokens": 42706 }, @@ -17841,7 +17983,7 @@ "node_count": 27, "item_count": 27, "soft_pair_count": 72, - "serialized_chars": 75095, + "serialized_chars": 75093, "estimated_input_tokens": 18774, "estimated_eval_tokens": 37804 }, @@ -17850,423 +17992,339 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a broad family of model-specific standardized output-tracing refactors. The clearest duplicate slices are the GPT-J/CodeGen edits and the GPT-Neo edits; the other PRs are mostly same-pattern changes in different model implementations and should stay separate unless they touch the exact same code path.", - "confidence": 0.84, - "canonical_issue_reason": "#43979 is the shared tracking issue for this output-tracing refactor family and is explicitly linked by many of the PRs across the cluster.", - "canonical_pr_reason": "PR 44722 is the best single representative PR for the strongest duplicate slice: it covers the GPT-J/CodeGen output-tracing path, links to #43979, and is broad enough to subsume the narrower GPT-J variants.", - "best_issue_reason": "Issue #43979 is the best cluster anchor because it cleanly unifies the repeated standardized-output-tracing work across many model-specific PRs, not just the GPT-J/CodeGen and GPT-Neo slices.", - "best_pr_reason": "PR 44722 is the best global PR representative because it captures the most complete GPT-J/CodeGen refactor variant while staying on the same concrete code path as the other strong duplicates.", + "summary": "This is a topical cluster of output-tracing refactors across many unrelated model families. Most pairs share only the same tracking issue or a generic refactor theme; only same-model variants like GPT-J/CodeGen and GPT-Neo look like true duplicates.", + "confidence": 0.87, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single PR cleanly canonicalizes the whole cluster because it spans many different models. If forced to pick a representative, 44722 is the strongest canonical PR for the GPT-J/CodeGen subgroup: it is the latest, most consolidated version of that concrete output-tracing refactor.", + "best_issue_reason": null, + "best_pr_reason": "44722 is the best standalone PR representative among the near-duplicates because it covers the same concrete GPT-J/CodeGen output-tracing change with the broadest, most polished implementation. It is not a global canonical for the entire heterogeneous cluster.", "soft_edge_verdicts": [ { "left": "pull_request:43996", "right": "pull_request:44085", "accept": false, - "reason": "Same refactor theme, but CVT/FNet and GPT-J are different model code paths." + "reason": "Same generic output-tracing topic and issue target, but different model families (CVT/FNet vs GPT-J) and unrelated code paths." }, { "left": "pull_request:43996", "right": "pull_request:44044", "accept": false, - "reason": "Different model implementations; shared issue link alone is not enough." + "reason": "Both are output-tracing refactors, but they touch different models and implementations; not the same concrete change." }, { "left": "pull_request:44066", "right": "pull_request:44085", "accept": true, - "reason": "Both are GPT-J output-tracing refactors on the same underlying code path; the scope differs, but the change family is the same." + "reason": "Both edit the same GPT-J output-tracing code path in modeling_gptj.py; the change is substantively the same refactor, with 44066 just broader." }, { "left": "pull_request:44007", "right": "pull_request:44072", "accept": false, - "reason": "RegNet and EfficientNet are separate model implementations, not one duplicate fix." + "reason": "Same umbrella issue, but different architectures and files; these are separate model-specific refactors." }, { "left": "pull_request:44072", "right": "pull_request:44722", "accept": false, - "reason": "EfficientNet vs GPT-J/CodeGen: same pattern, different concrete model path." + "reason": "Both are standardized output-tracing PRs, but EfficientNet and GPT-J are unrelated code paths and cannot be merged as one fix." }, { "left": "pull_request:44066", "right": "pull_request:44072", "accept": false, - "reason": "GPT-J/CodeGen and EfficientNet are unrelated code paths." + "reason": "Different model families and different implementation details; the shared issue target is too broad to make them duplicates." }, { "left": "pull_request:44013", "right": "pull_request:44044", "accept": false, - "reason": "MobileNetV2 and DeBERTa are different model-specific changes." + "reason": "MobileNetV2 and DeBERTa are unrelated code paths; they only share the broad output-tracing theme." }, { "left": "pull_request:44066", "right": "pull_request:44086", "accept": false, - "reason": "GPT-J/CodeGen refactor vs MGP-STR refactor; different model families." + "reason": "GPT-J and MGP-STR are different models; the PRs are thematically similar but not the same change." }, { "left": "pull_request:44066", "right": "pull_request:44071", "accept": false, - "reason": "Same general output-tracing theme, but GPT-J/CodeGen and MPT are different implementations." + "reason": "GPT-J vs MPT; same refactor pattern, but not the same concrete bug/change." }, { "left": "pull_request:44018", "right": "pull_request:44068", "accept": true, - "reason": "Both target GPT-Neo output tracing on the same model code path, so they are plausible duplicate variants." + "reason": "Same GPT-Neo files and same decorator-based output-tracing refactor; these are clearly the same underlying change." }, { "left": "pull_request:44066", "right": "pull_request:44068", "accept": false, - "reason": "GPT-J/CodeGen and GPT-Neo are different model paths." + "reason": "GPT-J/CodeGen and GPT-Neo are different model implementations, so this is only thematic similarity." }, { "left": "pull_request:44018", "right": "pull_request:44066", "accept": false, - "reason": "GPT-Neo and GPT-J/CodeGen are separate model refactors." + "reason": "Different models and code paths; they share a refactor style, not the same fix." }, { "left": "pull_request:44019", "right": "pull_request:44722", "accept": false, - "reason": "ResNet and GPT-J/CodeGen are unrelated implementations." + "reason": "ResNet output tracing and GPT-J output tracing are unrelated model changes." }, { "left": "pull_request:44086", "right": "pull_request:44722", "accept": false, - "reason": "MGP-STR and GPT-J/CodeGen are different model-specific changes." + "reason": "MGP-STR and GPT-J are different code paths; same theme, different fixes." }, { "left": "pull_request:44019", "right": "pull_request:44085", "accept": false, - "reason": "ResNet vs GPT-J: same pattern, different code path." + "reason": "ResNet and GPT-J are unrelated implementations, so this is not a duplicate." }, { "left": "pull_request:44018", "right": "pull_request:44019", "accept": false, - "reason": "GPT-Neo and ResNet are different model implementations." + "reason": "GPT-Neo and ResNet are different models with different output-shaping logic." }, { "left": "pull_request:44019", "right": "pull_request:44071", "accept": false, - "reason": "ResNet and MPT are different code paths." - }, - { - "left": "pull_request:44068", - "right": "pull_request:44722", - "accept": false, - "reason": "GPT-Neo vs GPT-J/CodeGen; similar theme but not the same fix." + "reason": "ResNet and MPT are separate refactors, only loosely related by the tracking theme." }, { "left": "pull_request:44068", "right": "pull_request:44116", "accept": false, - "reason": "GPT-Neo and Flaubert/XLM are different model-specific refactors." + "reason": "GPT-Neo and Flaubert/XLM are unrelated model code paths." }, { - "left": "pull_request:44071", + "left": "pull_request:44068", "right": "pull_request:44722", "accept": false, - "reason": "MPT and GPT-J/CodeGen are separate code paths." + "reason": "GPT-Neo and GPT-J are different model families; same refactor pattern but not the same PR." }, { "left": "pull_request:44018", - "right": "pull_request:44116", - "accept": false, - "reason": "GPT-Neo vs Flaubert/XLM; not the same underlying bug/change." - }, - { - "left": "pull_request:44085", - "right": "pull_request:44086", - "accept": false, - "reason": "GPT-J and MGP-STR are different model implementations." - }, - { - "left": "pull_request:44059", "right": "pull_request:44086", "accept": false, - "reason": "GPT2 and MGP-STR are unrelated concrete changes." - }, - { - "left": "pull_request:44059", - "right": "pull_request:44722", - "accept": false, - "reason": "GPT2 vs GPT-J/CodeGen; same theme only." - }, - { - "left": "pull_request:44098", - "right": "pull_request:44101", - "accept": false, - "reason": "ViLT and Flaubert/XLM are different model-specific refactors." + "reason": "GPT-Neo and MGP-STR are unrelated implementations." }, { "left": "pull_request:44018", - "right": "pull_request:44071", - "accept": false, - "reason": "GPT-Neo and MPT are different code paths." - }, - { - "left": "pull_request:44017", - "right": "pull_request:44722", - "accept": false, - "reason": "SegFormer vs GPT-J/CodeGen; different model implementations." - }, - { - "left": "pull_request:44017", - "right": "pull_request:44161", - "accept": false, - "reason": "SegFormer and LongT5 are different concrete changes." - }, - { - "left": "pull_request:44017", - "right": "pull_request:44073", - "accept": false, - "reason": "SegFormer and VisualBert are unrelated model code paths." - }, - { - "left": "pull_request:44068", - "right": "pull_request:44071", - "accept": false, - "reason": "GPT-Neo and MPT are not the same fix." - }, - { - "left": "pull_request:44071", - "right": "pull_request:44076", - "accept": false, - "reason": "MPT and ImageGPT are different model-specific refactors." - }, - { - "left": "pull_request:43997", - "right": "pull_request:44073", + "right": "pull_request:44085", "accept": false, - "reason": "RegNet and VisualBert are different model implementations." + "reason": "GPT-Neo vs GPT-J; same issue theme, but not the same underlying change." }, { - "left": "pull_request:43997", - "right": "pull_request:44056", + "left": "pull_request:44071", + "right": "pull_request:44085", "accept": false, - "reason": "RegNet and MPNet are separate code paths." + "reason": "MPT and GPT-J are different model code paths." }, { - "left": "pull_request:44073", - "right": "pull_request:44101", + "left": "pull_request:44068", + "right": "pull_request:44086", "accept": false, - "reason": "VisualBert and Flaubert/XLM are different model-specific changes." + "reason": "GPT-Neo and MGP-STR are unrelated refactors." }, { "left": "pull_request:44059", - "right": "pull_request:44071", + "right": "pull_request:44086", "accept": false, - "reason": "GPT2 and MPT are different code paths." + "reason": "GPT2 and MGP-STR are different models with different output-tracing implementations." }, { - "left": "pull_request:44056", + "left": "pull_request:44098", "right": "pull_request:44101", "accept": false, - "reason": "MPNet and Flaubert/XLM are unrelated implementations." - }, - { - "left": "pull_request:44073", - "right": "pull_request:44161", - "accept": false, - "reason": "VisualBert and LongT5 are different model-specific refactors." + "reason": "ViLT and Flaubert/XLM are different code paths; only the generic theme overlaps." }, { - "left": "pull_request:44059", - "right": "pull_request:44068", + "left": "pull_request:44018", + "right": "pull_request:44116", "accept": false, - "reason": "GPT2 and GPT-Neo are different concrete model changes." + "reason": "GPT-Neo and Flaubert/XLM are separate model implementations." }, { - "left": "pull_request:44018", - "right": "pull_request:44059", + "left": "pull_request:44017", + "right": "pull_request:44722", "accept": false, - "reason": "GPT-Neo and GPT2 are separate model code paths." + "reason": "Segformer and GPT-J are unrelated model-specific refactors." }, { - "left": "pull_request:44101", + "left": "pull_request:44017", "right": "pull_request:44161", "accept": false, - "reason": "Flaubert/XLM and LongT5 are different implementations." + "reason": "Segformer and LongT5 are different code paths; shared issue target is too broad." }, { - "left": "pull_request:43997", - "right": "pull_request:44017", + "left": "pull_request:44073", + "right": "pull_request:44101", "accept": false, - "reason": "RegNet and SegFormer are unrelated model refactors." + "reason": "VisualBert and Flaubert/XLM are unrelated model implementations." }, { - "left": "pull_request:44071", - "right": "pull_request:44129", + "left": "pull_request:44059", + "right": "pull_request:44068", "accept": false, - "reason": "MPT and SpeechT5 are different code paths." + "reason": "GPT2 and GPT-Neo have different blocks and output handling; not the same change." }, { "left": "pull_request:44074", "right": "pull_request:44101", "accept": false, - "reason": "TextNet and Flaubert/XLM are different model-specific changes." + "reason": "TextNet and Flaubert/XLM are separate model-specific refactors." }, { "left": "pull_request:44056", "right": "pull_request:44161", "accept": false, - "reason": "MPNet and LongT5 are separate refactors." + "reason": "MPNet and LongT5 are different models; same generic feature, different fixes." }, { "left": "pull_request:43997", - "right": "pull_request:44018", + "right": "pull_request:44017", "accept": false, - "reason": "RegNet and GPT-Neo are not the same fix." + "reason": "RegNet and Segformer are unrelated code paths despite the shared tracking issue." }, { "left": "pull_request:43997", - "right": "pull_request:44161", + "right": "pull_request:44074", "accept": false, - "reason": "RegNet and LongT5 are different model implementations." + "reason": "RegNet and TextNet are distinct model implementations." }, { - "left": "pull_request:44074", - "right": "pull_request:44161", + "left": "pull_request:43997", + "right": "pull_request:44018", "accept": false, - "reason": "TextNet and LongT5 are different concrete changes." + "reason": "RegNet and GPT-Neo are unrelated model-specific changes." }, { - "left": "pull_request:44018", - "right": "pull_request:44161", + "left": "pull_request:44071", + "right": "pull_request:44722", "accept": false, - "reason": "GPT-Neo and LongT5 are unrelated model paths." + "reason": "MPT and GPT-J are different architectures; not a duplicate change." }, { - "left": "pull_request:44010", + "left": "pull_request:44059", "right": "pull_request:44071", "accept": false, - "reason": "SqueezeBert and MPT are different model implementations." + "reason": "GPT2 and MPT are different code paths." }, { - "left": "pull_request:44010", - "right": "pull_request:44073", + "left": "pull_request:44056", + "right": "pull_request:44101", "accept": false, - "reason": "SqueezeBert and VisualBert are separate code paths." + "reason": "MPNet and Flaubert/XLM are different refactors in different files." }, { - "left": "pull_request:43997", - "right": "pull_request:44074", + "left": "pull_request:44074", + "right": "pull_request:44161", "accept": false, - "reason": "RegNet and TextNet are different model-specific refactors." + "reason": "TextNet and LongT5 are separate model implementations." }, { "left": "pull_request:43997", - "right": "pull_request:44129", + "right": "pull_request:44056", "accept": false, - "reason": "RegNet and SpeechT5 are unrelated implementations." + "reason": "RegNet and MPNet are unrelated model families." }, { "left": "pull_request:44071", - "right": "pull_request:44129", - "accept": false, - "reason": "MPT and SpeechT5 are different model code paths." - }, - { - "left": "pull_request:44010", - "right": "pull_request:44018", + "right": "pull_request:44076", "accept": false, - "reason": "SqueezeBert and GPT-Neo are different models." + "reason": "MPT and ImageGPT are different model code paths." }, { - "left": "pull_request:43996", + "left": "pull_request:43997", "right": "pull_request:44161", "accept": false, - "reason": "CVT/FNet and LongT5 are different concrete changes." + "reason": "RegNet and LongT5 are unrelated changes." }, { "left": "pull_request:43996", "right": "pull_request:44017", "accept": false, - "reason": "CVT/FNet and SegFormer are separate model-specific refactors." + "reason": "CVT/FNet and Segformer are different model-specific refactors." }, { "left": "pull_request:44010", - "right": "pull_request:44044", + "right": "pull_request:44073", "accept": false, - "reason": "SqueezeBert and DeBERTa are different model implementations." + "reason": "SqueezeBert and VisualBert are unrelated model implementations." }, { - "left": "pull_request:44044", - "right": "pull_request:44074", + "left": "pull_request:44010", + "right": "pull_request:44018", "accept": false, - "reason": "DeBERTa and TextNet are different code paths." + "reason": "SqueezeBert and GPT-Neo only share the broad output-tracing theme." }, { - "left": "pull_request:44044", - "right": "pull_request:44129", + "left": "pull_request:44010", + "right": "pull_request:44044", "accept": false, - "reason": "DeBERTa and SpeechT5 are not the same underlying fix." + "reason": "SqueezeBert and DeBERTa are different code paths." }, { "left": "pull_request:44044", - "right": "pull_request:44154", + "right": "pull_request:44074", "accept": false, - "reason": "DeBERTa and ViLT are separate model-specific changes." + "reason": "DeBERTa and TextNet are unrelated refactors." }, { "left": "pull_request:44066", "right": "pull_request:44722", "accept": true, - "reason": "Both are GPT-J/CodeGen output-tracing variants on the same underlying code path." + "reason": "Same GPT-J output-tracing refactor, and both touch the same concrete file; 44722 is effectively a broader variant of the same change." }, { "left": "pull_request:44085", "right": "pull_request:44722", "accept": true, - "reason": "Both touch the GPT-J standardized output-tracing path; the scope differs, but the fix is the same family." + "reason": "Same GPT-J output-tracing change in the same file; the broader 44722 PR subsumes the GPT-J-specific part." }, { "left": "pull_request:44085", "right": "pull_request:44129", "accept": false, - "reason": "GPT-J and SpeechT5 are different model implementations." + "reason": "GPT-J and SpeechT5 are unrelated model code paths." }, { "left": "pull_request:43996", "right": "pull_request:44007", "accept": false, - "reason": "CVT/FNet and RegNet are different model-specific refactors." - }, - { - "left": "pull_request:43996", - "right": "pull_request:44129", - "accept": false, - "reason": "CVT/FNet and SpeechT5 are unrelated code paths." + "reason": "Different model families and different concrete output-tracing implementations." }, { "left": "pull_request:44044", "right": "pull_request:44129", "accept": false, - "reason": "DeBERTa and SpeechT5 are not the same concrete change." + "reason": "DeBERTa and SpeechT5 are separate model-specific changes." }, { "left": "pull_request:44044", "right": "pull_request:44154", "accept": false, - "reason": "DeBERTa and ViLT are different underlying fixes." + "reason": "DeBERTa and Vits are unrelated implementations." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The anchor/PR choices are mostly grounded, but the summary is too narrow for the packet and a couple of soft-edge accepts are a bit too assertive. In particular, the packet contains other strong shared-issue/soft-similarity pairs beyond GPT-J/CodeGen and GPT-Neo (e.g. CVT/FNet with RegNet/ResNet/SpeechT5/DeBERTa variants), so saying those two are the only strong duplicate groups is not fully supported. Also, the accepted duplicate verdicts rely on model-family similarity and shared issue links more than clearly identical bugfix scope, so they read a little less conservative than requested." + "accept": true, + "feedback": "The summary is grounded in the packet: it correctly treats the cluster as mostly heterogeneous, and it only elevates the GPT-J/CodeGen and GPT-Neo same-file variants as true duplicates. The accepted soft-edge pairs are conservative and look mergeable into single bugfix PRs, while the rejected pairs stay appropriately scoped to different model/code paths." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -18274,7 +18332,7 @@ { "label": "pull_request", "component_index": 14, - "component_count": 27, + "component_count": 25, "review_unit_index": 2, "review_unit_count": 2, "status": "reviewed", @@ -18301,12 +18359,12 @@ "pull_request:43997|pull_request:44010", "pull_request:44010|pull_request:44129" ], - "prepared_review_unit_hash": "cab1ebbe6b520533bcb6c9c050df375b52d5c53e3c5ff33c86be827f1ee526ea", + "prepared_review_unit_hash": "56a66d818554db9853016539f1807ddb9ccb678b23d44998b340f68b032588b9", "component_budget": { "node_count": 28, "item_count": 28, "soft_pair_count": 78, - "serialized_chars": 84897, + "serialized_chars": 84898, "estimated_input_tokens": 21225, "estimated_eval_tokens": 42706 }, @@ -18314,57 +18372,57 @@ "node_count": 8, "item_count": 8, "soft_pair_count": 6, - "serialized_chars": 17123, - "estimated_input_tokens": 4281, - "estimated_eval_tokens": 8818 + "serialized_chars": 17126, + "estimated_input_tokens": 4282, + "estimated_eval_tokens": 8820 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These PRs are all model-specific refactors to standardized output tracing/capture_outputs across different architectures, not the same underlying bug or change. They should not be merged as duplicates just because they share the same refactor theme.", - "confidence": 0.94, + "summary": "A cluster of model-specific PRs migrating different Transformers models to the standardized output-capture/tracing helpers. They share a common refactor theme, but each targets a different model code path, so they are not duplicates of one another.", + "confidence": 0.88, "canonical_issue_reason": null, - "canonical_pr_reason": "pull_request:44129 is the best representative because it has an explicit issue target, a clear output-tracing refactor, and appears more complete than the others.", + "canonical_pr_reason": "PR 44073 is the clearest representative: it explicitly updates VisualBert to the new capture_outputs/can_return_tuple pattern and its title matches the cluster theme well.", "best_issue_reason": null, - "best_pr_reason": "pull_request:44129 is the strongest standalone PR for the cluster: it is explicitly linked to an issue and cleanly implements the standardized output-capture migration for SpeechT5.", + "best_pr_reason": "PR 44073 is the best cluster representative because it most directly reflects the shared standardized output-tracing migration, with a clear model-specific implementation and concise scope.", "soft_edge_verdicts": [ { "left": "pull_request:44024", "right": "pull_request:44076", "accept": false, - "reason": "Different models and different code paths (FocalNet vs ImageGPT); same refactor style, but not the same concrete change and not mergeable as one PR." + "reason": "Both are output-tracing refactors, but FocalNet and ImageGPT are different model implementations with different hook targets and forward paths; this is not the same underlying change." }, { "left": "pull_request:44073", "right": "pull_request:44074", "accept": false, - "reason": "VisualBert and TextNet are separate model implementations with distinct forward/output logic; this is a shared pattern, not a duplicate fix." + "reason": "VisualBert and TextNet are separate model migrations. The refactor pattern is similar, but they touch different code paths and are not mergeable as one duplicate PR." }, { "left": "pull_request:44056", "right": "pull_request:44073", "accept": false, - "reason": "MPNet output tracing and VisualBert output tracing are unrelated model-specific migrations; no evidence of the same underlying bug." + "reason": "MPNet and VisualBert each add capture_outputs/can_return_tuple plumbing in different model files; same broad theme, but not the same concrete bug or change." }, { "left": "pull_request:44056", "right": "pull_request:44074", "accept": false, - "reason": "MPNet and TextNet touch different architectures and code paths, so this is not the same change despite similar decorator usage." + "reason": "These are distinct model-specific refactors in MPNet and TextNet. Similar tooling, but different forward implementations and output plumbing." }, { "left": "pull_request:43997", "right": "pull_request:44010", "accept": false, - "reason": "RegNet and SqueezeBert are different models with different forward signatures and output handling; too broad to treat as duplicates." + "reason": "RegNet and SqueezeBert are unrelated model code paths. The diffs are both output-tracing related, but they are independent migrations, not duplicates." }, { "left": "pull_request:44010", "right": "pull_request:44129", "accept": false, - "reason": "SqueezeBert and SpeechT5 are distinct model refactors in different modules; they share a library-wide pattern but not the same concrete PR." + "reason": "SqueezeBert and SpeechT5 both adjust output capture, but SpeechT5 has a different explicit issue target and a different attention/cache signature; these are separate changes." } ] }, @@ -18378,7 +18436,7 @@ { "label": "pull_request", "component_index": 15, - "component_count": 27, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18444,262 +18502,274 @@ "pull_request:44028|pull_request:44030", "pull_request:44029|pull_request:44030" ], - "prepared_review_unit_hash": "46a54b6bf120c340aaa3c4bd6685b77a4e066fff240cc9c69936b43a19c60984", + "prepared_review_unit_hash": "aef848aede654c5267e382442b25f64735280b444087b1e473ebc57aa1d4c80c", "component_budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 40, - "serialized_chars": 40175, - "estimated_input_tokens": 10044, - "estimated_eval_tokens": 20344 + "serialized_chars": 40171, + "estimated_input_tokens": 10043, + "estimated_eval_tokens": 20342 }, "budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 40, - "serialized_chars": 40175, - "estimated_input_tokens": 10044, - "estimated_eval_tokens": 20344 + "serialized_chars": 40171, + "estimated_input_tokens": 10043, + "estimated_eval_tokens": 20342 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "A set of similarly titled but mostly independent model-specific PRs that apply the output-tracing refactor pattern to different architectures. They share theme and implementation style, but not the same concrete bug or change.", - "confidence": 0.94, + "summary": "This cluster is a set of similar-looking but separate model-specific output-tracing refactors. They share the same tooling pattern, but not the same underlying bug or code path, so I would not collapse them as duplicates.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "No true canonical PR: these are separate model-specific refactors, not duplicates of one code change.", + "canonical_pr_reason": "PR 44025 is a good canonical representative because it is a cohesive multi-file refactor in one related model family and clearly shows the output-tracing migration pattern.", "best_issue_reason": null, - "best_pr_reason": "PR 44003 is the strongest representative because it spans two related model files and has the broadest concrete output-tracing update among the set.", + "best_pr_reason": "PR 44025 is the best single PR anchor for the cluster: it is self-contained, touches closely related variants, and most clearly exemplifies the common refactor theme without being overly specialized.", "soft_edge_verdicts": [ { "left": "pull_request:43998", "right": "pull_request:44002", "accept": false, - "reason": "Same refactor theme, but different model code paths (`timm_backbone` vs `upernet`)." + "reason": "Different model families and different forward/output paths; same refactor theme, not the same change." }, { "left": "pull_request:43998", "right": "pull_request:44000", "accept": false, - "reason": "Different models and outputs (`timm_backbone` vs `vision_text_dual_encoder`); only the refactor pattern matches." + "reason": "Both refactor output handling, but they target unrelated models and code paths." }, { "left": "pull_request:43998", "right": "pull_request:44001", "accept": false, - "reason": "Unrelated model implementations (`timm_backbone` vs `univnet`), so not the same change." + "reason": "Separate model implementations with no shared bug or mergeable code-path change." }, { "left": "pull_request:43998", "right": "pull_request:44003", "accept": false, - "reason": "Different architectures and paths (`timm_backbone` vs `mamba/falcon_mamba`); not mergeable as one PR." - }, - { - "left": "pull_request:43998", - "right": "pull_request:44004", - "accept": false, - "reason": "Both are output-tracing refactors, but for different model internals (`timm_backbone` vs `codegen`)." + "reason": "Different subsystems/models; one is timm_backbone, the other is mamba/falcon_mamba." }, { "left": "pull_request:43998", "right": "pull_request:44029", "accept": false, - "reason": "Different model families (`timm_backbone` vs `rwkv`); no shared concrete bug." + "reason": "Unrelated model code paths despite similar output-tracing refactor intent." }, { "left": "pull_request:43999", "right": "pull_request:44001", "accept": false, - "reason": "Different models (`mobilenet_v1` vs `univnet`) and different output plumbing." + "reason": "Mobilenet_v1 and univnet are distinct models with separate output APIs." }, { "left": "pull_request:43999", "right": "pull_request:44002", "accept": false, - "reason": "Different model code paths; same refactor style is not enough to duplicate." + "reason": "Different models and behavior; only the refactor pattern matches." }, { "left": "pull_request:43999", "right": "pull_request:44025", "accept": false, - "reason": "`mobilenet_v1` and `depth_anything` are separate implementations with no single shared fix." + "reason": "Different model families and different tracing mechanics; not one underlying fix." }, { "left": "pull_request:43999", "right": "pull_request:44026", "accept": false, - "reason": "Different wrapper/model code (`mobilenet_v1` vs `vision_encoder_decoder`)." + "reason": "No shared concrete bug; these are separate encoder/decoder vs vision model refactors." }, { "left": "pull_request:43999", "right": "pull_request:44027", "accept": false, - "reason": "Different model families (`mobilenet_v1` vs `speech_encoder_decoder`)." + "reason": "Distinct model families with unrelated forward implementations." }, { "left": "pull_request:44000", "right": "pull_request:44001", "accept": false, - "reason": "Different tasks and architectures (`vision_text_dual_encoder` vs `univnet`)." + "reason": "Similar API cleanup, but they affect different model code paths and outputs." }, { "left": "pull_request:44000", "right": "pull_request:44002", "accept": false, - "reason": "Separate model code paths; only the tracing refactor motif is shared." + "reason": "Different models; same broad refactor theme is not enough to deduplicate." }, { "left": "pull_request:44000", "right": "pull_request:44003", "accept": false, - "reason": "Distinct subsystems (`vision_text_dual_encoder` vs `mamba/falcon_mamba`) and unrelated fixes." + "reason": "These touch unrelated model stacks and would not plausibly merge into one PR." }, { "left": "pull_request:44000", - "right": "pull_request:44004", + "right": "pull_request:44025", "accept": false, - "reason": "Different model implementations (`vision_text_dual_encoder` vs `codegen`)." + "reason": "Different model families and different output-tracing implementations." }, { - "left": "pull_request:44000", - "right": "pull_request:44025", + "left": "pull_request:44001", + "right": "pull_request:44002", "accept": false, - "reason": "Different model families; not the same concrete behavior change." + "reason": "Separate model-specific refactors, not the same underlying change." }, { "left": "pull_request:44001", - "right": "pull_request:44002", + "right": "pull_request:44003", "accept": false, - "reason": "`univnet` and `upernet` are unrelated model paths." + "reason": "Mamba/falcon_mamba is unrelated to univnet beyond the shared refactor motif." }, { "left": "pull_request:44001", "right": "pull_request:44025", "accept": false, - "reason": "Different models and output structures; not a duplicate fix." + "reason": "Different models and no evidence of a shared defect or shared code path." }, { - "left": "pull_request:44001", - "right": "pull_request:44030", + "left": "pull_request:44002", + "right": "pull_request:44029", "accept": false, - "reason": "Separate model implementations (`univnet` vs `dpr`) with different call signatures." + "reason": "UperNet and RWKV are unrelated implementations; same theme only." }, { "left": "pull_request:44002", "right": "pull_request:44003", "accept": false, - "reason": "Different model families (`upernet` vs `mamba/falcon_mamba`) and no shared code-path bug." + "reason": "Distinct model stacks and distinct output-capture changes." }, { "left": "pull_request:44003", "right": "pull_request:44028", "accept": false, - "reason": "Both are refactors, but `falcon_mamba/mamba` and `superpoint` are unrelated code paths." + "reason": "Mamba/falcon_mamba and SuperPoint are unrelated; no shared bug or merged PR candidate." }, { "left": "pull_request:44003", "right": "pull_request:44004", "accept": false, - "reason": "Different model internals (`mamba/falcon_mamba` vs `codegen`) and not the same change." + "reason": "These are different code paths and different levels of API refactoring." }, { "left": "pull_request:43998", "right": "pull_request:44004", "accept": false, - "reason": "Different model code paths; same output-tracing pattern only." + "reason": "timm_backbone and codegen are unrelated models; only the naming pattern is similar." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44004", + "accept": false, + "reason": "UnivNet and CodeGen do not share a concrete code-path problem." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44004", + "accept": false, + "reason": "Different model architectures and output contracts." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44004", + "accept": false, + "reason": "Vision-text dual encoder vs CodeGen are unrelated changes." }, { "left": "pull_request:44025", "right": "pull_request:44026", "accept": false, - "reason": "Different encoder-decoder wrappers and model stacks; not one fix." + "reason": "Different model families; same refactor style but not the same fix." }, { "left": "pull_request:44025", "right": "pull_request:44027", "accept": false, - "reason": "`depth_anything` and `speech_encoder_decoder` are separate implementations." + "reason": "Depth Anything and Speech Encoder-Decoder are unrelated model paths." }, { "left": "pull_request:44025", "right": "pull_request:44028", "accept": false, - "reason": "Different model families (`depth_anything` vs `superpoint`)." + "reason": "Different models and different output-recording mechanics; not duplicate PRs." }, { "left": "pull_request:44025", "right": "pull_request:44029", "accept": false, - "reason": "Different architectures (`depth_anything` vs `rwkv`); not mergeable as a single PR." + "reason": "Depth Anything and RWKV are separate model stacks with no shared bug." }, { "left": "pull_request:44025", "right": "pull_request:44030", "accept": false, - "reason": "Separate model code paths (`depth_anything` vs `dpr`)." + "reason": "Different models and different concrete forward methods." }, { "left": "pull_request:44026", - "right": "pull_request:44027", + "right": "pull_request:44029", "accept": false, - "reason": "Both are encoder-decoder wrappers, but they fix different model files and behaviors." + "reason": "Vision encoder-decoder and RWKV are unrelated; same pattern only." }, { "left": "pull_request:44026", - "right": "pull_request:44028", + "right": "pull_request:44027", "accept": false, - "reason": "Different model families (`vision_encoder_decoder` vs `superpoint`)." + "reason": "Vision encoder-decoder and speech encoder-decoder are separate code paths with distinct implementations." }, { "left": "pull_request:44026", - "right": "pull_request:44029", + "right": "pull_request:44028", "accept": false, - "reason": "Unrelated architectures (`vision_encoder_decoder` vs `rwkv`)." + "reason": "Different models; not the same underlying bug or fix." }, { "left": "pull_request:44026", "right": "pull_request:44030", "accept": false, - "reason": "Different model implementations (`vision_encoder_decoder` vs `dpr`)." + "reason": "Vision encoder-decoder and DPR do not share a concrete code-path problem." }, { "left": "pull_request:44027", "right": "pull_request:44028", "accept": false, - "reason": "Different model code paths; only the refactor pattern overlaps." + "reason": "Speech encoder-decoder and SuperPoint are unrelated model changes." }, { "left": "pull_request:44027", "right": "pull_request:44029", "accept": false, - "reason": "Separate architectures (`speech_encoder_decoder` vs `rwkv`)." + "reason": "Different architectures and output APIs; not duplicate changes." }, { "left": "pull_request:44027", "right": "pull_request:44030", "accept": false, - "reason": "Different model stacks (`speech_encoder_decoder` vs `dpr`)." + "reason": "Speech encoder-decoder and DPR are separate model implementations." }, { "left": "pull_request:44028", "right": "pull_request:44029", "accept": false, - "reason": "Different models (`superpoint` vs `rwkv`) and no shared concrete bug." + "reason": "SuperPoint and RWKV are unrelated code paths." }, { "left": "pull_request:44028", "right": "pull_request:44030", "accept": false, - "reason": "Different model implementations (`superpoint` vs `dpr`)." + "reason": "SuperPoint and DPR are different models with different output contracts." }, { "left": "pull_request:44029", "right": "pull_request:44030", "accept": false, - "reason": "Separate model families (`rwkv` vs `dpr`); not the same fix." + "reason": "RWKV and DPR are different model families; no shared concrete fix." } ] }, @@ -18713,7 +18783,7 @@ { "label": "pull_request", "component_index": 16, - "component_count": 27, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18732,52 +18802,52 @@ "pull_request:44662|pull_request:45075", "pull_request:45075|pull_request:45401" ], - "prepared_review_unit_hash": "7b641df278a9a025a2c431d73819bd5fff31f35239997232e8bbfa3631ca6abc", + "prepared_review_unit_hash": "320ad4651be130179970fde0ae4b07844d8d151fc1a409226a04ab19e85192e4", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 10914, - "estimated_input_tokens": 2729, - "estimated_eval_tokens": 5714 + "serialized_chars": 10912, + "estimated_input_tokens": 2728, + "estimated_eval_tokens": 5712 }, "budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 10914, - "estimated_input_tokens": 2729, - "estimated_eval_tokens": 5714 + "serialized_chars": 10912, + "estimated_input_tokens": 2728, + "estimated_eval_tokens": 5712 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "All three PRs are separate model integrations (PenguinVL, DeepSeek-OCR-2, and Voxtral TTS). They share only common Transformers boilerplate around auto mappings, docs, and package wiring, not the same underlying change.", + "summary": "All three pull requests are separate model-addition implementations (PenguinVL, DeepSeek-OCR-2, and Voxtral TTS). They share Transformers scaffolding files and docs registration, but they do not fix the same concrete model or code-path issue, so none should be deduplicated together.", "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45075 is the strongest representative: it\u2019s a full model integration with the most extensive implementation/test surface and higher review activity, making it the best cluster anchor.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR #45075 is the best overall PR candidate in this set because it is a substantive end-to-end model addition, appears more complete than the draft Voxtral PR, and is more clearly packaged as a standalone integration than the others.", + "best_pr_reason": "No PR is a good canonical representative for the cluster because they target different models. If one must be singled out, 45075 is the most mature standalone implementation, but it should not absorb the others.", "soft_edge_verdicts": [ { "left": "pull_request:44662", "right": "pull_request:45401", "accept": false, - "reason": "Different model additions for different products/domains (PenguinVL vs Voxtral TTS). Overlap is limited to generic model-registration/docs plumbing, not the same code-path fix." + "reason": "Different model families and capabilities: PenguinVL vs Voxtral TTS. Shared auto-registration/docs files are generic model-add boilerplate, not the same underlying change." }, { "left": "pull_request:44662", "right": "pull_request:45075", "accept": false, - "reason": "These add unrelated models (PenguinVL vs DeepSeek-OCR-2). Shared files are generic framework wiring, but the concrete implementations, processors, and tests are for different models." + "reason": "Different model additions: PenguinVL vs DeepSeek-OCR-2. Overlap is limited to common integration scaffolding, not a shared bug or fix." }, { "left": "pull_request:45075", "right": "pull_request:45401", "accept": false, - "reason": "DeepSeek-OCR-2 and Voxtral TTS are distinct model integrations with different architectures and targets; shared auto-mapping/docs changes are boilerplate, not a common underlying change." + "reason": "Different concrete models and tasks: DeepSeek-OCR-2 vs Voxtral TTS. Similar file patterns are expected for new model support and do not indicate duplicate PRs." } ] }, @@ -18791,7 +18861,7 @@ { "label": "pull_request", "component_index": 17, - "component_count": 27, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18807,12 +18877,12 @@ "soft_pairs": [ "pull_request:44979|pull_request:45363" ], - "prepared_review_unit_hash": "31c75df0207bae5e38ce1d541bb779ed71431a429c3564baa377644d6337e10d", + "prepared_review_unit_hash": "e3beb93d5b4e2de35e77549f1b63049d82decbb041cefb92bc2eb9bddf43353d", "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 }, @@ -18820,7 +18890,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4407, + "serialized_chars": 4408, "estimated_input_tokens": 1102, "estimated_eval_tokens": 2460 }, @@ -18829,18 +18899,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are related to model/kernel fusion infrastructure, but they are not the same change. #44979 introduces a generic Module Fusion API, while #45363 implements n-to-1 kernel fusion via KernelConfig and integrates it through model loading paths.", - "confidence": 0.94, + "summary": "These PRs are related by the broader theme of fusion, but they implement different changes: one introduces a general module fusion API, while the other adds n-to-1 kernel fusion support via KernelConfig and related integration plumbing. They are not duplicates and should stay separate.", + "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "No issue is present. As a PR representative, #45363 is the broader, more end-to-end fusion feature because it wires KernelConfig through the core loading/integration path, whereas #44979 is the lower-level module-fusion API scaffold.", + "canonical_pr_reason": "PR 44979 is the more foundational change: it adds the new Module Fusion API and test coverage, which is a core library feature rather than a usage example or a narrower integration path.", "best_issue_reason": null, - "best_pr_reason": "#45363 is the best cluster representative because it is the more complete user-facing fusion implementation and touches the main integration points, while #44979 is a narrower API foundation.", + "best_pr_reason": "PR 44979 is the best representative for the cluster because it defines the core abstraction, whereas PR 45363 is a separate kernel-fusion implementation built on adjacent infrastructure.", "soft_edge_verdicts": [ { "left": "pull_request:44979", "right": "pull_request:45363", "accept": false, - "reason": "Different underlying changes: #44979 adds a generic module fusion API, while #45363 adds KernelConfig-based n-to-1 kernel fusion integration. They share the fusion theme but do not appear mergeable as one PR." + "reason": "Different concrete changes and code paths: 44979 adds a general module fusion API, while 45363 adds KernelConfig-based n-to-1 kernel fusion and example/integration updates. They are related conceptually but not the same underlying fix or mergeable into one PR." } ] }, @@ -18854,7 +18924,7 @@ { "label": "pull_request", "component_index": 18, - "component_count": 27, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18870,7 +18940,7 @@ "soft_pairs": [ "pull_request:45086|pull_request:45317" ], - "prepared_review_unit_hash": "43a487442a835d0c12a3c8dbcd5c49dc7ee0283551025700d3e6bf30defd843d", + "prepared_review_unit_hash": "9b2a5bb108debe3bd590002048ab5183d2d89db1c6810d2fb3368f4b4c1f4954", "component_budget": { "node_count": 2, "item_count": 2, @@ -18892,18 +18962,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both pull requests address the same concrete bug in `_patch_mistral_regex`: accessing `tokenizer.backend_tokenizer.pre_tokenizer` causes an AttributeError, and both switch to `tokenizer.pre_tokenizer` on the same code path. The later PR also adds test coverage, making it the better representative of the fix.", - "confidence": 0.96, + "summary": "Both PRs target the same Mistral regex patch bug in the same file with the same code change; 45317 is the more complete version because it adds tests and clearer context.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45317 is the clearest canonical representative because it fixes the same AttributeError on the same Mistral pre-tokenizer path as PR 45086, but also includes tests and a more explicit title/scope.", + "canonical_pr_reason": "45317 is the stronger canonical PR because it covers the same fix as 45086 and adds test coverage, making it the more complete representative of the change.", "best_issue_reason": null, - "best_pr_reason": "PR 45317 is the best overall PR in the cluster because it contains the same code fix as 45086 plus validation in tests, making it more complete and merge-ready.", + "best_pr_reason": "45317 is the best PR for the cluster: same concrete bug fix, same code path, and it includes validation/tests rather than only the minimal code edit.", "soft_edge_verdicts": [ { "left": "pull_request:45086", "right": "pull_request:45317", "accept": true, - "reason": "Same underlying bug and same fix in `_patch_mistral_regex`; 45317 is essentially a superset of 45086 with added tests." + "reason": "Accept: both PRs fix the same AttributeError in _patch_mistral_regex by changing the same tokenizer pre-tokenizer accessors; 45317 is essentially a superset with tests, so they could plausibly be merged into one PR." } ] }, @@ -18917,7 +18987,7 @@ { "label": "pull_request", "component_index": 19, - "component_count": 27, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18933,7 +19003,7 @@ "soft_pairs": [ "pull_request:45115|pull_request:45569" ], - "prepared_review_unit_hash": "1fe33424b2526bead0494ec0ffce82eca6bb05d14808567ce7455d34efa49115", + "prepared_review_unit_hash": "3041fe4fb12a25a8834292750ce2931ab6aece11078260b5540af3e99aeb76b4", "component_budget": { "node_count": 2, "item_count": 2, @@ -18955,18 +19025,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two NemotronH PRs overlap on shared model files, but they address different changes: one is a conversion/refactor toward GraniteMoeHybrid-compatible naming, the other is a broader model split into NemotronH Dense/Sparse with docs and tests. They are related, but not the same underlying change.", - "confidence": 0.87, + "summary": "These are related to NemotronH work, but they are not the same PR: one is a conversion/refactor to GraniteMoeHybrid-compatible mappings, while the other is a broader model split/addition for NemotronH dense/sparse variants and docs. The overlap is mostly shared files in the NemotronH package, not the same concrete code change.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45569 is the better canonical PR because it is the more complete, user-facing model change: it introduces the dense/sparse variants, updates auto/config/model plumbing, docs, and tests. PR #45115 is narrower and mostly about conversion/mapping refactors.", + "canonical_pr_reason": "PR #45569 is the better canonical representative for this cluster because it is the broader, more user-facing model-structure change introducing the dedicated NemotronH dense/sparse variants and associated docs/tests, whereas #45115 is a narrower conversion-mapping/refactor PR.", "best_issue_reason": null, - "best_pr_reason": "PR #45569 is the strongest representative of the cluster due to its broader scope and end-to-end implementation. PR #45115 is better treated as a separate, smaller refactor/conversion patch.", + "best_pr_reason": "PR #45569 is the strongest representative PR overall: it covers the main architecture/package rework and is more complete than the smaller conversion-mapping-only PR.", "soft_edge_verdicts": [ { "left": "pull_request:45115", "right": "pull_request:45569", "accept": false, - "reason": "Shared NemotronH internals do not make these the same fix. #45115 focuses on checkpoint conversion and structural renaming for GraniteMoeHybrid compatibility, while #45569 is a larger architectural split into NemotronHDense/Sparse with docs and tests. They could not realistically be merged as one identical PR." + "reason": "Reject: both touch NemotronH code, but 45115 is about conversion mapping and inheritance/refactoring, while 45569 adds dedicated NemotronH dense/sparse model variants, docs, and tests. They are adjacent work, not the same underlying change and would not naturally merge into one PR." } ] }, @@ -18980,7 +19050,7 @@ { "label": "pull_request", "component_index": 20, - "component_count": 27, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18996,7 +19066,7 @@ "soft_pairs": [ "pull_request:45332|pull_request:45333" ], - "prepared_review_unit_hash": "148cc396aa7eb1c307a600fc1b8dad5947b04e62ee1b11e66492e79d2e68c29b", + "prepared_review_unit_hash": "b2673f3349e5b14a6009543295a38a98f11e53a6e0dbbeb995a939619b41c377", "component_budget": { "node_count": 2, "item_count": 2, @@ -19018,18 +19088,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs are part of the same heterogeneous-model feature area, but they are not duplicates: one adds per-layer config support while the other adds broader modeling/runtime support on top of that.", - "confidence": 0.89, + "summary": "These are related heterogeneous-config workstreams, but they are not duplicate PRs: one adds per-layer configuration plumbing, while the other adds broader heterogeneous model/runtime support on top of it.", + "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45332 is the broader end-to-end change: it extends cache, masking, modeling, and gpt_oss paths for heterogeneous per-layer behavior, so it is the better canonical representative of the feature work.", + "canonical_pr_reason": "No duplicate canonical PR: PR 45332 is the broader heterogeneous model/runtime implementation, while PR 45333 is the narrower config-plumbing foundation. They are related but not the same change.", "best_issue_reason": null, - "best_pr_reason": "PR #45332 is the strongest standalone PR because it covers the concrete runtime/modeling code paths rather than only the config layer, making it the most complete representative of the feature.", + "best_pr_reason": "PR 45332 is the best representative of the cluster because it covers the larger end-to-end heterogeneous support surface and subsumes the config-only groundwork in 45333.", "soft_edge_verdicts": [ { "left": "pull_request:45332", "right": "pull_request:45333", "accept": false, - "reason": "These are related but distinct feature slices. #45333 focuses on heterogeneous config plumbing, while #45332 implements additional modeling/runtime changes in cache, masking, and model code. They do not look like the same concrete fix and would not naturally collapse into one duplicate PR." + "reason": "Reject: 45333 is only per-layer config support, whereas 45332 adds heterogeneous modeling/cache/runtime behavior. They share infrastructure files, but they are different concrete changes and not the same PR-level bug fix or feature." } ] }, @@ -19043,7 +19113,7 @@ { "label": "pull_request", "component_index": 21, - "component_count": 27, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -19059,7 +19129,7 @@ "soft_pairs": [ "pull_request:45415|pull_request:45425" ], - "prepared_review_unit_hash": "6f4956e52d7b5d71adbbe4bf27448c0db4bcf22af28d0088a04f5894c4260f1c", + "prepared_review_unit_hash": "468747db5fb4157075d83107b7a4b3e00bfaf65a45921168c7d6ba6122864b41", "component_budget": { "node_count": 2, "item_count": 2, @@ -19081,18 +19151,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs are typing-focused, but they are not duplicates: #45415 is a broad project-wide type-checking sweep, while #45425 adds specific typing protocols/helpers for model-utils and PEFT-related code paths.", - "confidence": 0.91, + "summary": "Both PRs are typing-related, but they are not duplicates: #45415 is a broad repo-wide type-checking sweep, while #45425 adds focused typing protocols/helpers for specific modules. They overlap in a few files, but not in the same concrete change.", + "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "#45425 is the more focused, concrete change; #45415 is a wide cleanup across many files and a different scope of work.", + "canonical_pr_reason": "PR #45415 is the broadest representative of the cluster, with a repo-wide type-checking pass and pyproject typing-ignore policy changes across many files.", "best_issue_reason": null, - "best_pr_reason": "#45425 best represents a single underlying change because it targets specific typing contracts and integrations, whereas #45415 is a broad typing pass spanning many unrelated files.", + "best_pr_reason": "PR #45415 is the strongest cluster anchor because it captures the main typing-enable effort at a larger, more central scope than the narrower follow-up in #45425.", "soft_edge_verdicts": [ { "left": "pull_request:45415", "right": "pull_request:45425", "accept": false, - "reason": "They share the general theme of type checking and a few filenames, but they fix different problems: one is a broad typing sweep, the other adds specific typing protocol support. They would not plausibly merge into one PR as the same concrete change." + "reason": "Related typing work, but not the same underlying change: #45415 is a wide mechanical type-checking sweep, while #45425 introduces specific typing aliases/protocols for modeling_utils and related code paths. They would not plausibly merge as one duplicate PR." } ] }, @@ -19106,109 +19176,7 @@ { "label": "pull_request", "component_index": 22, - "component_count": 27, - "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-45487-6", - "nodes": [ - "pull_request:45487", - "pull_request:45817", - "pull_request:45843", - "pull_request:45858", - "pull_request:45890", - "pull_request:45926" - ], - "soft_pairs": [ - "pull_request:45487|pull_request:45843", - "pull_request:45487|pull_request:45926", - "pull_request:45817|pull_request:45843", - "pull_request:45858|pull_request:45890", - "pull_request:45858|pull_request:45926", - "pull_request:45890|pull_request:45926" - ], - "prepared_review_unit_hash": "07ab05a92ef35ac278b873972fee4dd13698e2452aada3b69761baadcc6dd335", - "component_budget": { - "node_count": 6, - "item_count": 6, - "soft_pair_count": 6, - "serialized_chars": 14718, - "estimated_input_tokens": 3680, - "estimated_eval_tokens": 7616 - }, - "budget": { - "node_count": 6, - "item_count": 6, - "soft_pair_count": 6, - "serialized_chars": 14718, - "estimated_input_tokens": 3680, - "estimated_eval_tokens": 7616 - }, - "overflow_policy": "truncate_then_skip", - "trimmed": false, - "aggressively_trimmed": false, - "split": false, - "analyst_result": { - "summary": "This cluster is a loose topical grouping around model-parallel fixes and XPU test expectations, but the PRs target different models and different concrete code paths. I do not see any true duplicates among them.", - "confidence": 0.89, - "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45817 is the clearest standalone fix: it directly addresses a concrete model-parallel device-placement bug in Gemma4 code and includes an accompanying test update.", - "best_issue_reason": null, - "best_pr_reason": "PR 45817 is the strongest representative of the cluster because it fixes an actual runtime path in model code rather than only updating expectations, and the change is specific and well-scoped.", - "soft_edge_verdicts": [ - { - "left": "pull_request:45487", - "right": "pull_request:45843", - "accept": false, - "reason": "Both mention model-parallel problems, but they affect different models (AltCLIP/ChineseCLIP vs Deimv2) and the fixes are not the same code path or mergeable as one PR." - }, - { - "left": "pull_request:45487", - "right": "pull_request:45926", - "accept": false, - "reason": "Different models and different changes: one fixes device placement in model code, the other adds XPU test expectations for HyperCLOVAX." - }, - { - "left": "pull_request:45817", - "right": "pull_request:45843", - "accept": false, - "reason": "Both are model-parallel PRs, but they target different architectures and likely different failure modes; not the same underlying bug." - }, - { - "left": "pull_request:45858", - "right": "pull_request:45890", - "accept": false, - "reason": "These are test-only expectation updates for different models, not the same bug or change." - }, - { - "left": "pull_request:45858", - "right": "pull_request:45926", - "accept": false, - "reason": "Different model test expectations; same broad XPU theme is too weak to treat as a duplicate." - }, - { - "left": "pull_request:45890", - "right": "pull_request:45926", - "accept": false, - "reason": "Both are XPU expectation updates, but for unrelated models and separate test files, so they are not 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": 27, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -19224,7 +19192,7 @@ "soft_pairs": [ "pull_request:45548|pull_request:45662" ], - "prepared_review_unit_hash": "cefd351ae40b9ade0812c3ef7ea6945c5a474c414349d749070d2df09281e9db", + "prepared_review_unit_hash": "f34ddd3191d7791e835bc636a416ecd9b251226be8fa7b19bbbd813bd260b810", "component_budget": { "node_count": 2, "item_count": 2, @@ -19246,18 +19214,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs are about expert parallelism (EP) loading, but they fix different failure modes: PR #45548 preserves `distributed_config` and adds an EP-detection property for accelerate/ZeRO-3 loading, while PR #45662 fixes expert parameters being overwritten by rank-0 broadcast in FSDP2 loading. They share a subsystem, not the same underlying bug.", - "confidence": 0.94, + "summary": "Two EP-related pull requests, but they fix different bugs in different loading paths: one is about DeepSpeed ZeRO-3 / accelerate launch serialization-loading behavior, the other about FSDP2 rank-0 broadcast overwriting experts. They are related by theme, not the same underlying change.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45662 is the more concrete and self-contained code-path fix for a loading corruption bug, with changes centered on `core_model_loading.py` and EP tensor handling.", + "canonical_pr_reason": "No canonical duplicate PR here; the two patches address distinct code-path failures. If forced to pick an anchor, 45662 is the more concrete bug fix in core loading behavior.", "best_issue_reason": null, - "best_pr_reason": "PR #45662 is the best representative of the cluster because it addresses a specific EP loading defect and has the strongest direct code-path alignment; PR #45548 is a separate, narrower loading-compatibility fix.", + "best_pr_reason": "45662 is the stronger representative of the cluster only because it has the more explicit, self-contained failure mode and broader implementation changes; 45548 targets a separate loading/serialization issue.", "soft_edge_verdicts": [ { "left": "pull_request:45548", "right": "pull_request:45662", "accept": false, - "reason": "Different bugs and merge scopes: #45548 fixes missing `distributed_config` / EP detection for accelerate launch with ZeRO-3, while #45662 fixes expert parameters being overwritten during FSDP2 loading. Shared EP wording is not enough to treat them as duplicates." + "reason": "Both are EP-related, but 45548 fixes distributed_config serialization / has_ep handling for DeepSpeed ZeRO-3 loading, while 45662 fixes experts being overwritten by rank-0 broadcast in a different FSDP2 loading path. Different underlying bugs, not mergeable as one PR." } ] }, @@ -19270,8 +19238,8 @@ }, { "label": "pull_request", - "component_index": 24, - "component_count": 27, + "component_index": 23, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -19287,7 +19255,7 @@ "soft_pairs": [ "pull_request:45639|pull_request:45645" ], - "prepared_review_unit_hash": "774b33fd03a90f79fc33ed227a246fe304a796834664abbe85e5004759f97673", + "prepared_review_unit_hash": "b09064fae7efcc344d454e480eaeebdbe7a4162126d35105b319fa1d8ab8e48d", "component_budget": { "node_count": 2, "item_count": 2, @@ -19309,18 +19277,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs target the same xdist collision in captured_info handling; #45645 is the broader superset and appears to subsume #45639.", - "confidence": 0.94, - "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45645 is the better canonical PR because it includes the core testing_utils xdist-safe fix from #45639 and also updates CI/workflow log collection to consume the new per-worker artifacts.", - "best_issue_reason": null, - "best_pr_reason": "PR #45645 is the strongest representative of the cluster: it addresses the same captured_info collision problem end-to-end, with the additional workflow changes making it more complete than #45639.", + "summary": "Both PRs address the same xdist-related captured_info collision problem tied to issue #45561; #45645 is a broader superset of #45639\u2019s testing_utils fix and adds CI/log handling for the same artifact path.", + "confidence": 0.84, + "canonical_issue_reason": "No issue artifacts were provided, so there is no issue canonical to choose here.", + "canonical_pr_reason": "PR #45645 is the better canonical PR because it subsumes the core testing_utils fix from #45639 and extends the same change to CI/workflow artifact handling and notification paths.", + "best_issue_reason": "No issue nodes were supplied, so there is no best issue to rank.", + "best_pr_reason": "PR #45645 is the best PR because it appears to be the more complete end-to-end fix for the captured_info/xdist collision problem, including the workflow changes needed to consume multiple captured_info files.", "soft_edge_verdicts": [ { "left": "pull_request:45639", "right": "pull_request:45645", "accept": true, - "reason": "Accept: both are about the same concrete bug in captured_info artifact handling under xdist, and #45645 clearly builds on the same fix rather than a different change." + "reason": "Same underlying bug: xdist-safe handling of captured_info artifacts. #45645 contains the core fix from #45639 and adds related CI/log collection changes, so these look like duplicate/overlapping implementations of the same change." } ] }, @@ -19333,8 +19301,8 @@ }, { "label": "pull_request", - "component_index": 25, - "component_count": 27, + "component_index": 24, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -19396,8 +19364,8 @@ }, { "label": "pull_request", - "component_index": 26, - "component_count": 27, + "component_index": 25, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -19459,80 +19427,6 @@ "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null - }, - { - "label": "pull_request", - "component_index": 27, - "component_count": 27, - "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-45997-3", - "nodes": [ - "pull_request:45997", - "pull_request:46005", - "pull_request:46010" - ], - "soft_pairs": [ - "pull_request:45997|pull_request:46005", - "pull_request:46005|pull_request:46010" - ], - "prepared_review_unit_hash": "e73c2260ac249c73f1b4daabeda5d4097ca1adb436be7dbd33e0205e9dfc5175", - "component_budget": { - "node_count": 3, - "item_count": 3, - "soft_pair_count": 2, - "serialized_chars": 6650, - "estimated_input_tokens": 1663, - "estimated_eval_tokens": 3582 - }, - "budget": { - "node_count": 3, - "item_count": 3, - "soft_pair_count": 2, - "serialized_chars": 6650, - "estimated_input_tokens": 1663, - "estimated_eval_tokens": 3582 - }, - "overflow_policy": "truncate_then_skip", - "trimmed": false, - "aggressively_trimmed": false, - "split": false, - "analyst_result": { - "summary": "All three PRs revolve around the same cache sequence-length type bug in `StaticLayer/StaticCache.get_seq_length()`. Two are narrow variants of the same fix, and one is a broader patch with downstream adjustments and tests.", - "confidence": 0.91, - "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46010 is the most complete representative: it includes the core `get_seq_length()` int-return fix plus related downstream changes and tests, so it subsumes the narrower variants.", - "best_issue_reason": null, - "best_pr_reason": "PR 46010 is the strongest standalone fix for the cluster because it addresses the same bug with broader coverage and validation, making it the best merge candidate.", - "soft_edge_verdicts": [ - { - "left": "pull_request:45997", - "right": "pull_request:46005", - "accept": true, - "reason": "Same underlying cache bug on the exact same code path; 45997 is a narrower subset of 46005's fix (scalar tensor + `.item()` return)." - }, - { - "left": "pull_request:46005", - "right": "pull_request:46010", - "accept": true, - "reason": "Both target the same `StaticCache.get_seq_length()` tensor-vs-int bug; 46010 appears to be the broader, more complete version with downstream adjustments and tests, so they are the same change family." - } - ] - }, - "evaluator_result": { - "accept": true, - "feedback": "Grounded and conservative overall. The packet supports that 45997 is a narrower subset of the cache-length fix, and that 46005 and 46010 address the same bug with 46010 adding broader downstream changes/tests. The summary does not overstate beyond the diff evidence, and the paired PRs look mergeable as the same bugfix family." - }, - "evaluator_used": true, - "retried": false, - "accepted_nontrivial_soft_edge": true, - "error_kind": null, - "error_message": null } ] } diff --git a/analysis/current/manifest.json b/analysis/current/manifest.json index 4261b295e22c4718c26aa67ddc12a1e024ac778c..65f617ebca54608eebd2a8475409c099c15508e0 100644 --- a/analysis/current/manifest.json +++ b/analysis/current/manifest.json @@ -1,8 +1,8 @@ { - "analysis_id": "hybrid-model-20260518t060213z", + "analysis_id": "hybrid-model-20260518t120024z", "archived_artifacts": { - "hybrid": "snapshots/20260518T060213Z/analysis-runs/hybrid-model-20260518t060213z/analysis-report-hybrid.json", - "hybrid_reviews": "snapshots/20260518T060213Z/analysis-runs/hybrid-model-20260518t060213z/analysis-report-hybrid.llm-reviews.json" + "hybrid": "snapshots/20260518T120024Z/analysis-runs/hybrid-model-20260518t120024z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260518T120024Z/analysis-runs/hybrid-model-20260518t120024z/analysis-report-hybrid.llm-reviews.json" }, "artifacts": { "hybrid": "analysis/current/analysis-report-hybrid.json", @@ -10,9 +10,9 @@ }, "channel": "canonical", "model": null, - "published_at": "2026-05-18T06:15:35Z", + "published_at": "2026-05-18T12:15:51Z", "repo": "huggingface/transformers", "schema_version": 1, - "snapshot_id": "20260518T060213Z", + "snapshot_id": "20260518T120024Z", "variant": "hybrid" } diff --git a/snapshots/20260518T120024Z/analysis-runs/hybrid-model-20260518t120024z/analysis-report-hybrid.json b/snapshots/20260518T120024Z/analysis-runs/hybrid-model-20260518t120024z/analysis-report-hybrid.json new file mode 100644 index 0000000000000000000000000000000000000000..413e13b0d61eecd8b47dacfcc442f43bfb879abf --- /dev/null +++ b/snapshots/20260518T120024Z/analysis-runs/hybrid-model-20260518t120024z/analysis-report-hybrid.json @@ -0,0 +1,2205 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260518T120024Z", + "generated_at": "2026-05-18T12:15:50Z", + "evidence_quality": "full", + "llm_enrichment": true, + "meta_bugs": [ + { + "cluster_id": "cluster-43979-11", + "summary": "Cluster of 1 issues and 10 PRs centered on issue #43979.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 43979, + "canonical_pr_number": 44007, + "issue_numbers": [ + 43979 + ], + "pr_numbers": [ + 43996, + 44007, + 44013, + 44044, + 44066, + 44072, + 44085, + 44129, + 44154, + 44722 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 43996, + "right_pr_number": 44007, + "code_similarity": 0.179, + "size_similarity": 0.576, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.429, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44013, + "code_similarity": 0.122, + "size_similarity": 0.318, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.392, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44044, + "code_similarity": 0.245, + "size_similarity": 0.864, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.479, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44066, + "code_similarity": 0.225, + "size_similarity": 0.818, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.408, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44072, + "code_similarity": 0.14, + "size_similarity": 0.303, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.528, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44085, + "code_similarity": 0.216, + "size_similarity": 0.783, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.398, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44129, + "code_similarity": 0.163, + "size_similarity": 0.643, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.229, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44154, + "code_similarity": 0.153, + "size_similarity": 0.535, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.31, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44722, + "code_similarity": 0.225, + "size_similarity": 0.848, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.368, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44013, + "code_similarity": 0.19, + "size_similarity": 0.553, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.531, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44044, + "code_similarity": 0.186, + "size_similarity": 0.667, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.354, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44066, + "code_similarity": 0.188, + "size_similarity": 0.704, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.315, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44072, + "code_similarity": 0.212, + "size_similarity": 0.526, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.708, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44085, + "code_similarity": 0.195, + "size_similarity": 0.735, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.318, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44129, + "code_similarity": 0.103, + "size_similarity": 0.37, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.191, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44154, + "code_similarity": 0.238, + "size_similarity": 0.93, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.344, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44722, + "code_similarity": 0.178, + "size_similarity": 0.679, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.28, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44044, + "code_similarity": 0.126, + "size_similarity": 0.368, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.351, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44066, + "code_similarity": 0.127, + "size_similarity": 0.389, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.325, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44072, + "code_similarity": 0.29, + "size_similarity": 0.952, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.667, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44085, + "code_similarity": 0.131, + "size_similarity": 0.406, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.329, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44129, + "code_similarity": 0.07, + "size_similarity": 0.205, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.192, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44154, + "code_similarity": 0.177, + "size_similarity": 0.594, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.389, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44722, + "code_similarity": 0.118, + "size_similarity": 0.375, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.287, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44066, + "code_similarity": 0.25, + "size_similarity": 0.947, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.404, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44072, + "code_similarity": 0.136, + "size_similarity": 0.351, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.442, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44085, + "code_similarity": 0.24, + "size_similarity": 0.906, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.394, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44129, + "code_similarity": 0.147, + "size_similarity": 0.555, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.243, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44154, + "code_similarity": 0.17, + "size_similarity": 0.62, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.306, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44722, + "code_similarity": 0.257, + "size_similarity": 0.982, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.402, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44072, + "code_similarity": 0.133, + "size_similarity": 0.37, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.393, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44085, + "code_similarity": 0.763, + "size_similarity": 0.957, + "file_overlap": 0.5, + "area_overlap": 0.825, + "patch_similarity": 0.887, + "shared_filenames": [ + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 41 + ], + [ + 174, + 181 + ], + [ + 250, + 257 + ], + [ + 398, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 466 + ], + [ + 471, + 488 + ], + [ + 518, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 606 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ], + "right_ranges": [ + [ + 33, + 40 + ], + [ + 173, + 180 + ], + [ + 249, + 256 + ], + [ + 397, + 405 + ], + [ + 408, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 465 + ], + [ + 470, + 489 + ], + [ + 517, + 539 + ], + [ + 553, + 559 + ], + [ + 565, + 574 + ], + [ + 579, + 597 + ], + [ + 600, + 611 + ], + [ + 633, + 639 + ], + [ + 645, + 652 + ], + [ + 657, + 673 + ], + [ + 716, + 728 + ], + [ + 737, + 743 + ], + [ + 748, + 771 + ], + [ + 789, + 794 + ] + ] + } + ] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44129, + "code_similarity": 0.145, + "size_similarity": 0.526, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.263, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44154, + "code_similarity": 0.174, + "size_similarity": 0.654, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.286, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44722, + "code_similarity": 0.907, + "size_similarity": 0.964, + "file_overlap": 1.0, + "area_overlap": 0.808, + "patch_similarity": 0.874, + "shared_filenames": [ + "src/transformers/models/codegen/modeling_codegen.py", + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/codegen/modeling_codegen.py", + "left_ranges": [ + [ + 245, + 266 + ] + ], + "right_ranges": [ + [ + 228, + 234 + ] + ] + }, + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 41 + ], + [ + 174, + 181 + ], + [ + 250, + 257 + ], + [ + 398, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 466 + ], + [ + 471, + 488 + ], + [ + 518, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 606 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ], + "right_ranges": [ + [ + 33, + 42 + ], + [ + 175, + 182 + ], + [ + 251, + 258 + ], + [ + 399, + 420 + ], + [ + 426, + 435 + ], + [ + 460, + 467 + ], + [ + 472, + 489 + ], + [ + 519, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 612 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 663 + ], + [ + 666, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ] + } + ] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44085, + "code_similarity": 0.137, + "size_similarity": 0.387, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.398, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44129, + "code_similarity": 0.074, + "size_similarity": 0.195, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.231, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44154, + "code_similarity": 0.175, + "size_similarity": 0.566, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.414, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44722, + "code_similarity": 0.124, + "size_similarity": 0.357, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.347, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44129, + "code_similarity": 0.141, + "size_similarity": 0.503, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.272, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44154, + "code_similarity": 0.18, + "size_similarity": 0.684, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.289, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44722, + "code_similarity": 0.728, + "size_similarity": 0.923, + "file_overlap": 0.5, + "area_overlap": 0.791, + "patch_similarity": 0.78, + "shared_filenames": [ + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 40 + ], + [ + 173, + 180 + ], + [ + 249, + 256 + ], + [ + 397, + 405 + ], + [ + 408, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 465 + ], + [ + 470, + 489 + ], + [ + 517, + 539 + ], + [ + 553, + 559 + ], + [ + 565, + 574 + ], + [ + 579, + 597 + ], + [ + 600, + 611 + ], + [ + 633, + 639 + ], + [ + 645, + 652 + ], + [ + 657, + 673 + ], + [ + 716, + 728 + ], + [ + 737, + 743 + ], + [ + 748, + 771 + ], + [ + 789, + 794 + ] + ], + "right_ranges": [ + [ + 33, + 42 + ], + [ + 175, + 182 + ], + [ + 251, + 258 + ], + [ + 399, + 420 + ], + [ + 426, + 435 + ], + [ + 460, + 467 + ], + [ + 472, + 489 + ], + [ + 519, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 612 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 663 + ], + [ + 666, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ] + } + ] + }, + { + "left_pr_number": 44129, + "right_pr_number": 44154, + "code_similarity": 0.099, + "size_similarity": 0.344, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.199, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44129, + "right_pr_number": 44722, + "code_similarity": 0.146, + "size_similarity": 0.545, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.247, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44154, + "right_pr_number": 44722, + "code_similarity": 0.164, + "size_similarity": 0.631, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.255, + "shared_filenames": [], + "shared_file_areas": [] + } + ] + }, + { + "cluster_id": "cluster-41211-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #41211.", + "status": "open", + "confidence": 0.55, + "canonical_issue_number": 41211, + "canonical_pr_number": 44339, + "issue_numbers": [ + 41211 + ], + "pr_numbers": [ + 41356, + 44339 + ], + "evidence_types": [ + "closing_reference" + ], + "pr_comparisons": [ + { + "left_pr_number": 41356, + "right_pr_number": 44339, + "code_similarity": 0.155, + "size_similarity": 0.094, + "file_overlap": 0.273, + "area_overlap": 0.078, + "patch_similarity": 0.183, + "shared_filenames": [ + "docs/source/en/model_doc/deimv2.md", + "src/transformers/models/deimv2/__init__.py", + "src/transformers/models/deimv2/configuration_deimv2.py", + "src/transformers/models/deimv2/modeling_deimv2.py", + "tests/models/deimv2/__init__.py", + "tests/models/deimv2/test_modeling_deimv2.py" + ], + "shared_file_areas": [ + { + "filename": "docs/source/en/model_doc/deimv2.md", + "left_ranges": [ + [ + 1, + 132 + ] + ], + "right_ranges": [ + [ + 1, + 65 + ] + ] + }, + { + "filename": "src/transformers/models/deimv2/__init__.py", + "left_ranges": [ + [ + 1, + 15 + ] + ], + "right_ranges": [ + [ + 1, + 29 + ] + ] + }, + { + "filename": "src/transformers/models/deimv2/configuration_deimv2.py", + "left_ranges": [ + [ + 1, + 74 + ] + ], + "right_ranges": [ + [ + 1, + 266 + ] + ] + }, + { + "filename": "tests/models/deimv2/test_modeling_deimv2.py", + "left_ranges": [ + [ + 1, + 15 + ] + ], + "right_ranges": [ + [ + 1, + 1734 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-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-43240-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #43240.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 43240, + "canonical_pr_number": 43251, + "issue_numbers": [ + 43240 + ], + "pr_numbers": [ + 43251, + 43254 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 43251, + "right_pr_number": 43254, + "code_similarity": 0.794, + "size_similarity": 0.64, + "file_overlap": 1.0, + "area_overlap": 0.667, + "patch_similarity": 0.882, + "shared_filenames": [ + "src/transformers/loss/loss_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/loss/loss_utils.py", + "left_ranges": [ + [ + 30, + 50 + ] + ], + "right_ranges": [ + [ + 30, + 43 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-41115-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #41115.", + "status": "open", + "confidence": 0.55, + "canonical_issue_number": 41115, + "canonical_pr_number": 45613, + "issue_numbers": [ + 41115 + ], + "pr_numbers": [ + 41116, + 45613 + ], + "evidence_types": [ + "closing_reference" + ], + "pr_comparisons": [ + { + "left_pr_number": 41116, + "right_pr_number": 45613, + "code_similarity": 0.805, + "size_similarity": 0.885, + "file_overlap": 0.75, + "area_overlap": 0.854, + "patch_similarity": 0.693, + "shared_filenames": [ + "docs/source/en/_toctree.yml", + "docs/source/en/model_doc/minicpm3.md", + "src/transformers/models/auto/modeling_auto.py", + "src/transformers/models/minicpm3/__init__.py", + "src/transformers/models/minicpm3/configuration_minicpm3.py", + "src/transformers/models/minicpm3/modeling_minicpm3.py", + "src/transformers/models/minicpm3/modular_minicpm3.py", + "tests/models/minicpm3/__init__.py", + "tests/models/minicpm3/test_modeling_minicpm3.py" + ], + "shared_file_areas": [ + { + "filename": "docs/source/en/_toctree.yml", + "left_ranges": [ + [ + 713, + 720 + ] + ], + "right_ranges": [ + [ + 713, + 720 + ] + ] + }, + { + "filename": "docs/source/en/model_doc/minicpm3.md", + "left_ranges": [ + [ + 1, + 67 + ] + ], + "right_ranges": [ + [ + 1, + 45 + ] + ] + }, + { + "filename": "src/transformers/models/auto/modeling_auto.py", + "left_ranges": [ + [ + 285, + 291 + ], + [ + 700, + 706 + ], + [ + 1298, + 1304 + ] + ], + "right_ranges": [ + [ + 699, + 705 + ], + [ + 1301, + 1307 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/__init__.py", + "left_ranges": [ + [ + 1, + 29 + ] + ], + "right_ranges": [ + [ + 1, + 29 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/configuration_minicpm3.py", + "left_ranges": [ + [ + 1, + 141 + ] + ], + "right_ranges": [ + [ + 1, + 126 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/modeling_minicpm3.py", + "left_ranges": [ + [ + 1, + 544 + ] + ], + "right_ranges": [ + [ + 1, + 522 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/modular_minicpm3.py", + "left_ranges": [ + [ + 1, + 444 + ] + ], + "right_ranges": [ + [ + 1, + 342 + ] + ] + }, + { + "filename": "tests/models/minicpm3/test_modeling_minicpm3.py", + "left_ranges": [ + [ + 1, + 109 + ] + ], + "right_ranges": [ + [ + 1, + 136 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-45647-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45647.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 45647, + "canonical_pr_number": 45738, + "issue_numbers": [ + 45647 + ], + "pr_numbers": [ + 45737, + 45738 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 45737, + "right_pr_number": 45738, + "code_similarity": 0.526, + "size_similarity": 0.11, + "file_overlap": 0.5, + "area_overlap": 1.0, + "patch_similarity": 0.024, + "shared_filenames": [ + "src/transformers/models/musicgen_melody/modeling_musicgen_melody.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/musicgen_melody/modeling_musicgen_melody.py", + "left_ranges": [ + [ + 491, + 497 + ] + ], + "right_ranges": [ + [ + 491, + 497 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-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-45561-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45561.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 45561, + "canonical_pr_number": 45645, + "issue_numbers": [ + 45561 + ], + "pr_numbers": [ + 45639, + 45645 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 45639, + "right_pr_number": 45645, + "code_similarity": 0.532, + "size_similarity": 0.619, + "file_overlap": 0.5, + "area_overlap": 0.543, + "patch_similarity": 0.451, + "shared_filenames": [ + "src/transformers/testing_utils.py", + "tests/utils/test_testing_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/testing_utils.py", + "left_ranges": [ + [ + 3529, + 3536 + ], + [ + 3753, + 3787 + ] + ], + "right_ranges": [ + [ + 3525, + 3558 + ], + [ + 3782, + 3788 + ] + ] + }, + { + "filename": "tests/utils/test_testing_utils.py", + "left_ranges": [ + [ + 1, + 86 + ] + ], + "right_ranges": [ + [ + 1, + 114 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-43698-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #43698.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 43698, + "canonical_pr_number": 43779, + "issue_numbers": [ + 43698 + ], + "pr_numbers": [ + 43779, + 43816 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 43779, + "right_pr_number": 43816, + "code_similarity": 0.418, + "size_similarity": 0.538, + "file_overlap": 1.0, + "area_overlap": 0.02, + "patch_similarity": 0.02, + "shared_filenames": [ + "src/transformers/integrations/integration_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/integrations/integration_utils.py", + "left_ranges": [ + [ + 2227, + 2241 + ], + [ + 2303, + 2309 + ] + ], + "right_ranges": [ + [ + 2278, + 2291 + ], + [ + 2309, + 2322 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-44018-2", + "summary": "Cluster of 2 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 44068, + "issue_numbers": [], + "pr_numbers": [ + 44018, + 44068 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 44018, + "right_pr_number": 44068, + "code_similarity": 0.766, + "size_similarity": 0.939, + "file_overlap": 1.0, + "area_overlap": 0.425, + "patch_similarity": 0.866, + "shared_filenames": [ + "src/transformers/models/gpt_neo/modeling_gpt_neo.py", + "tests/models/gpt_neo/test_modeling_gpt_neo.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/gpt_neo/modeling_gpt_neo.py", + "left_ranges": [ + [ + 26, + 46 + ], + [ + 138, + 143 + ], + [ + 182, + 187 + ], + [ + 283, + 295 + ], + [ + 327, + 341 + ], + [ + 348, + 354 + ], + [ + 360, + 369 + ], + [ + 399, + 406 + ], + [ + 411, + 419 + ], + [ + 428, + 433 + ], + [ + 472, + 492 + ], + [ + 507, + 513 + ], + [ + 519, + 528 + ], + [ + 541, + 559 + ], + [ + 562, + 567 + ], + [ + 595, + 601 + ], + [ + 607, + 614 + ], + [ + 627, + 643 + ], + [ + 685, + 690 + ], + [ + 708, + 714 + ], + [ + 720, + 727 + ], + [ + 740, + 757 + ], + [ + 761, + 766 + ], + [ + 780, + 786 + ], + [ + 791, + 798 + ], + [ + 807, + 822 + ], + [ + 840, + 845 + ] + ], + "right_ranges": [ + [ + 26, + 31 + ], + [ + 34, + 44 + ], + [ + 136, + 141 + ], + [ + 180, + 185 + ], + [ + 281, + 293 + ], + [ + 325, + 339 + ], + [ + 346, + 352 + ], + [ + 358, + 367 + ], + [ + 397, + 404 + ], + [ + 409, + 417 + ], + [ + 426, + 434 + ], + [ + 463, + 483 + ], + [ + 498, + 504 + ], + [ + 510, + 519 + ], + [ + 532, + 550 + ], + [ + 553, + 564 + ], + [ + 586, + 592 + ], + [ + 598, + 605 + ], + [ + 618, + 634 + ], + [ + 676, + 688 + ], + [ + 699, + 705 + ], + [ + 711, + 718 + ], + [ + 731, + 748 + ], + [ + 752, + 762 + ], + [ + 771, + 777 + ], + [ + 782, + 789 + ], + [ + 798, + 813 + ], + [ + 831, + 836 + ] + ] + }, + { + "filename": "tests/models/gpt_neo/test_modeling_gpt_neo.py", + "left_ranges": [ + [ + 458, + 464 + ] + ], + "right_ranges": [ + [ + 458, + 464 + ] + ] + } + ] + } + ] + } + ], + "duplicate_issues": [], + "duplicate_prs": [ + { + "cluster_id": "cluster-41115-3", + "canonical_pr_number": 45613, + "duplicate_pr_numbers": [ + 41116 + ], + "target_issue_number": 41115, + "reason": "PRs in cluster-41115-3 are treated as duplicates because they converge on issue #41115 with closing_reference evidence." + }, + { + "cluster_id": "cluster-41211-3", + "canonical_pr_number": 44339, + "duplicate_pr_numbers": [ + 41356 + ], + "target_issue_number": 41211, + "reason": "PRs in cluster-41211-3 are treated as duplicates because they converge on issue #41211 with closing_reference evidence." + }, + { + "cluster_id": "cluster-43240-3", + "canonical_pr_number": 43251, + "duplicate_pr_numbers": [ + 43254 + ], + "target_issue_number": 43240, + "reason": "PRs in cluster-43240-3 are treated as duplicates because they converge on issue #43240 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-43656-4", + "canonical_pr_number": 43656, + "duplicate_pr_numbers": [ + 43836, + 43842 + ], + "target_issue_number": 43824, + "reason": "PRs in cluster-43656-4 are treated as duplicates because they converge on issue #43824 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-43698-3", + "canonical_pr_number": 43779, + "duplicate_pr_numbers": [ + 43816 + ], + "target_issue_number": 43698, + "reason": "PRs in cluster-43698-3 are treated as duplicates because they converge on issue #43698 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-43979-11", + "canonical_pr_number": 44007, + "duplicate_pr_numbers": [ + 43996, + 44013, + 44044, + 44066, + 44072, + 44085, + 44129, + 44154, + 44722 + ], + "target_issue_number": 43979, + "reason": "PRs in cluster-43979-11 are treated as duplicates because they converge on issue #43979 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-44018-2", + "canonical_pr_number": 44068, + "duplicate_pr_numbers": [ + 44018 + ], + "target_issue_number": null, + "reason": "PRs in cluster-44018-2 are treated as duplicates because they share soft_similarity evidence." + }, + { + "cluster_id": "cluster-45081-3", + "canonical_pr_number": 45317, + "duplicate_pr_numbers": [ + 45086 + ], + "target_issue_number": 45081, + "reason": "PRs in cluster-45081-3 are treated as duplicates because they converge on issue #45081 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-45561-3", + "canonical_pr_number": 45645, + "duplicate_pr_numbers": [ + 45639 + ], + "target_issue_number": 45561, + "reason": "PRs in cluster-45561-3 are treated as duplicates because they converge on issue #45561 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-45647-3", + "canonical_pr_number": 45738, + "duplicate_pr_numbers": [ + 45737 + ], + "target_issue_number": 45647, + "reason": "PRs in cluster-45647-3 are treated as duplicates because they converge on issue #45647 with closing_reference, shared_issue_target evidence." + } + ], + "best_issue": { + "cluster_id": "cluster-45647-3", + "issue_number": 45647, + "reason": "Issue #45647 is the strongest global issue candidate because it is open, belongs to a cluster with 3 artifacts, and carries 12 discussion comments plus 8 inbound references.", + "score": 162.0 + }, + "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/20260518T120024Z/analysis-runs/hybrid-model-20260518t120024z/analysis-report-hybrid.llm-reviews.json b/snapshots/20260518T120024Z/analysis-runs/hybrid-model-20260518t120024z/analysis-report-hybrid.llm-reviews.json new file mode 100644 index 0000000000000000000000000000000000000000..135dcc9ce0e39263c393d8877c40bd5ce3f1c3d5 --- /dev/null +++ b/snapshots/20260518T120024Z/analysis-runs/hybrid-model-20260518t120024z/analysis-report-hybrid.llm-reviews.json @@ -0,0 +1,19432 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260518T120024Z", + "generated_at": "2026-05-18T12:15:50Z", + "model": "gpt-5.4-mini?service_tier=flex", + "reviews": [ + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 1, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "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:43425", + "issue:43541", + "issue:43596", + "issue:43638", + "issue:43716", + "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:44805", + "issue:44841", + "issue:44863", + "issue:44918", + "issue:45081", + "issue:45161", + "issue:45237", + "issue:45464", + "issue:45507", + "issue:45685", + "issue:45812", + "issue:45823", + "issue:45859", + "issue:45874", + "issue:45901", + "issue:45902", + "issue:45920", + "issue:46015" + ], + "soft_pairs": [ + "issue:44322|issue:45464", + "issue:45237|issue:45685", + "issue:43716|issue:45237", + "issue:43828|issue:45237", + "issue:45237|issue:45507", + "issue:44292|issue:45237", + "issue:30277|issue:35545", + "issue:35545|issue:42915", + "issue:35545|issue:41720", + "issue:45901|issue:46015", + "issue:44509|issue:44512", + "issue:35545|issue:36010", + "issue:44805|issue:45823", + "issue:43638|issue:44805", + "issue:43596|issue:44805", + "issue:44805|issue:45161", + "issue:44805|issue:44918", + "issue:34634|issue:35545", + "issue:30277|issue:43638", + "issue:43596|issue:43638", + "issue:43039|issue:45161", + "issue:43975|issue:44457", + "issue:44292|issue:44457", + "issue:34928|issue:44805", + "issue:24643|issue:30277", + "issue:45859|issue:45902", + "issue:43638|issue:45161", + "issue:34928|issue:36331", + "issue:43596|issue:45161", + "issue:43716|issue:45507", + "issue:45823|issue:45874", + "issue:39290|issue:45081", + "issue:43425|issue:44292", + "issue:43828|issue:45161", + "issue:34928|issue:41720", + "issue:44918|issue:45161", + "issue:43716|issue:45685", + "issue:43541|issue:43716", + "issue:43828|issue:44292", + "issue:45507|issue:45685", + "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:43828|issue:45685", + "issue:44560|issue:45685", + "issue:43039|issue:43828", + "issue:30277|issue:36331", + "issue:45237|issue:46015", + "issue:34928|issue:36010", + "issue:43039|issue:44805", + "issue:41762|issue:43596", + "issue:43901|issue:44512", + "issue:39290|issue:41720", + "issue:45237|issue:45901", + "issue:44841|issue:45823", + "issue:43039|issue:44560", + "issue:43975|issue:46015", + "issue:45812|issue:45920", + "issue:43866|issue:44863", + "issue:44863|issue:46015", + "issue:44292|issue:46015", + "issue:43854|issue:43866", + "issue:41093|issue:43638" + ], + "prepared_review_unit_hash": "79ed1f2392eac2d8c62cced47b776ec6fac3df1e51400e86f8dff6456431c81f", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 66, + "serialized_chars": 47560, + "estimated_input_tokens": 11890, + "estimated_eval_tokens": 24036 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly a loose topical cluster, not one real duplicate set. The only strong duplicate-looking groups are the Zero-3/BERT IndexError pair and the v5 AutoTokenizer token-ID regression pair; the rest are distinct model/back-end/doc issues.", + "confidence": 0.71, + "canonical_issue_reason": "issue:45920 is the broadest, clearest representative of the shared v5 AutoTokenizer wrong-token-ID regression, and it subsumes the narrower Granite-only report in issue:45812.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45920 is the strongest standalone issue in this cluster: broad impact, clear regression framing, and it captures the common tokenizer code path better than the narrower instance.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44322", + "right": "issue:45464", + "accept": false, + "reason": "Same model family, but different failures: missing config field vs streaming chat/completions crash." + }, + { + "left": "issue:45237", + "right": "issue:45685", + "accept": false, + "reason": "Different backends and symptoms: AMD GPU failure vs MPS histogram unsupported for Int." + }, + { + "left": "issue:43716", + "right": "issue:45237", + "accept": false, + "reason": "Different model and error path: image preprocessor/model dtype mismatch vs AMD GPU runtime failure." + }, + { + "left": "issue:43828", + "right": "issue:45237", + "accept": false, + "reason": "Both are runtime errors, but the underlying bugs are unrelated." + }, + { + "left": "issue:45237", + "right": "issue:45507", + "accept": false, + "reason": "Different model-specific issues; no shared concrete code path." + }, + { + "left": "issue:44292", + "right": "issue:45237", + "accept": false, + "reason": "Different model families and failure modes; not the same bug." + }, + { + "left": "issue:30277", + "right": "issue:35545", + "accept": false, + "reason": "DeepSpeed collective mismatch is unrelated to ONNX export failure." + }, + { + "left": "issue:35545", + "right": "issue:42915", + "accept": false, + "reason": "Both involve model support, but one is ONNX export and the other is FineGrainedFP8 loading." + }, + { + "left": "issue:35545", + "right": "issue:41720", + "accept": false, + "reason": "Different issues: export-to-ONNX vs device-map CUDA assert." + }, + { + "left": "issue:45901", + "right": "issue:46015", + "accept": false, + "reason": "Unrelated: table-question-answering crash vs outdated RAG examples." + }, + { + "left": "issue:44509", + "right": "issue:44512", + "accept": false, + "reason": "Both are stale docs, but they target different removed features and pages." + }, + { + "left": "issue:35545", + "right": "issue:36010", + "accept": false, + "reason": "Different problems: export error vs GenerationMixin import breakage." + }, + { + "left": "issue:44805", + "right": "issue:45823", + "accept": false, + "reason": "Similar-looking runtime errors, but different models and failure modes." + }, + { + "left": "issue:43638", + "right": "issue:44805", + "accept": false, + "reason": "Both are IndexError reports, but one is zero3/BERT init and the other is a mask/tensor shape mismatch." + }, + { + "left": "issue:43596", + "right": "issue:44805", + "accept": false, + "reason": "Different underlying bugs: zero3/BERT out-of-bounds vs mask/tensor shape mismatch." + }, + { + "left": "issue:44805", + "right": "issue:45161", + "accept": false, + "reason": "Different model paths: masking/indexing error vs GPT-OSS tensor-parallel issue." + }, + { + "left": "issue:44805", + "right": "issue:44918", + "accept": false, + "reason": "Different subsystems and errors; not the same bug." + }, + { + "left": "issue:34634", + "right": "issue:35545", + "accept": false, + "reason": "Processor voice_preset bug is unrelated to ModernBERT ONNX export." + }, + { + "left": "issue:30277", + "right": "issue:43638", + "accept": false, + "reason": "Both involve DeepSpeed, but the actual failures are different." + }, + { + "left": "issue:43596", + "right": "issue:43638", + "accept": true, + "reason": "Same Zero-3/BERT out-of-bounds IndexError on model loading/training; looks like one bug with the same code path." + }, + { + "left": "issue:43039", + "right": "issue:45161", + "accept": false, + "reason": "Liger Kernel cross-entropy routing is unrelated to GPT-OSS tensor-parallel failure." + }, + { + "left": "issue:43975", + "right": "issue:44457", + "accept": false, + "reason": "Different issues: detokenization regression vs LoRA merge/save mismatch." + }, + { + "left": "issue:44292", + "right": "issue:44457", + "accept": false, + "reason": "Different models and symptoms; not a duplicate." + }, + { + "left": "issue:34928", + "right": "issue:44805", + "accept": false, + "reason": "Activation-checkpointing/FSDP tensor recomputation issue is unrelated to the mask-shape IndexError." + }, + { + "left": "issue:24643", + "right": "issue:30277", + "accept": false, + "reason": "Both mention DeepSpeed, but the runtime errors come from different code paths." + }, + { + "left": "issue:45859", + "right": "issue:45902", + "accept": false, + "reason": "Same class name, but different bugs: CPU offload incompatibility vs uninitialized-memory after materialization." + }, + { + "left": "issue:43638", + "right": "issue:45161", + "accept": false, + "reason": "Different models and failure modes; not the same bug." + }, + { + "left": "issue:44509", + "right": "issue:44512", + "accept": false, + "reason": "Both are doc cleanup topics, but they concern different removed CLI/docs features." + }, + { + "left": "issue:44841", + "right": "issue:45823", + "accept": false, + "reason": "Processor failure and device-mapping mismatch are unrelated." + }, + { + "left": "issue:43425", + "right": "issue:44292", + "accept": false, + "reason": "Torch version compatibility is unrelated to Qwen-3-8B-NVFP4 runtime failure." + }, + { + "left": "issue:43828", + "right": "issue:45685", + "accept": false, + "reason": "Autocast dtype mismatch vs MPS histogram unsupported are unrelated." + }, + { + "left": "issue:44560", + "right": "issue:45685", + "accept": false, + "reason": "Video StopIteration bug is unrelated to the MPS Int histogram error." + }, + { + "left": "issue:43039", + "right": "issue:43828", + "accept": false, + "reason": "Different model/integration bugs; no shared concrete code path." + }, + { + "left": "issue:30277", + "right": "issue:36331", + "accept": false, + "reason": "DeepSpeed collective mismatch is unrelated to custom trainer signature breakage." + }, + { + "left": "issue:45237", + "right": "issue:46015", + "accept": false, + "reason": "Model-on-AMD-GPU failure is unrelated to outdated RAG docs." + }, + { + "left": "issue:34928", + "right": "issue:36010", + "accept": false, + "reason": "Training/checkpointing issue is unrelated to GenerationMixin import error." + }, + { + "left": "issue:43039", + "right": "issue:44805", + "accept": false, + "reason": "Different subsystems and errors; not the same bug." + }, + { + "left": "issue:41762", + "right": "issue:43638", + "accept": false, + "reason": "Both are loading/runtime errors, but one is Qwen3 auto device mapping and the other is BERT zero3 init/training." + }, + { + "left": "issue:43901", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs issues, but they target different outdated API references." + }, + { + "left": "issue:39290", + "right": "issue:41720", + "accept": false, + "reason": "Different Qwen3/Gemma3-vs-Qwen3-vl style failures; not the same issue." + }, + { + "left": "issue:45237", + "right": "issue:45901", + "accept": false, + "reason": "Different models/features: AMD GPU support vs table-question-answering crash." + }, + { + "left": "issue:44841", + "right": "issue:45823", + "accept": false, + "reason": "Processor failure and device-map forward mismatch are unrelated." + }, + { + "left": "issue:43039", + "right": "issue:44560", + "accept": false, + "reason": "Liger Kernel cross-entropy issue is unrelated to Qwen3-vl video StopIteration." + }, + { + "left": "issue:43975", + "right": "issue:46015", + "accept": false, + "reason": "Detokenization regression is unrelated to stale RAG docs." + }, + { + "left": "issue:45812", + "right": "issue:45920", + "accept": true, + "reason": "Same underlying v5 AutoTokenizer wrong-token-ID regression; 45920 is the broader umbrella and 45812 is a narrower instance." + }, + { + "left": "issue:43866", + "right": "issue:44863", + "accept": false, + "reason": "Both are checkpoint/model-loading reports, but they concern different models and different failures." + }, + { + "left": "issue:44863", + "right": "issue:46015", + "accept": false, + "reason": "NemotronH checkpoint loading bug is unrelated to docs cleanup." + }, + { + "left": "issue:44292", + "right": "issue:46015", + "accept": false, + "reason": "Qwen-3-8B-NVFP4 runtime failure is unrelated to outdated RAG docs." + }, + { + "left": "issue:43854", + "right": "issue:43866", + "accept": false, + "reason": "Unit-test model-loading issue and corrupted checkpoint report are different problems." + }, + { + "left": "issue:41093", + "right": "issue:43638", + "accept": false, + "reason": "Similar-looking IndexErrors, but the mask/tensor mismatch and zero3/BERT out-of-bounds are different bugs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Summary is grounded in the packet and stays conservative. The two accepted duplicate pairs are plausible from the titles/body framing, and the broader cluster characterization as mostly loose topical similarity is supported. Canonical issue choice for the v5 AutoTokenizer regression is reasonable. No mergeability concerns apply here since there is no PR pair." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 2, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:36010", + "issue:39186", + "issue:39290", + "issue:40990", + "issue:41093", + "issue:41762", + "issue:42915", + "issue:43054", + "issue:43296", + "issue:43366", + "issue:43404", + "issue:43425", + "issue:43531", + "issue:43541", + "issue:43572", + "issue:43596", + "issue:43645", + "issue:43716", + "issue:43792", + "issue:43828", + "issue:43866", + "issue:43975", + "issue:43994", + "issue:44291", + "issue:44292", + "issue:44387", + "issue:44530", + "issue:44589", + "issue:44841", + "issue:44918", + "issue:44960", + "issue:45005", + "issue:45020", + "issue:45070", + "issue:45084", + "issue:45161", + "issue:45362", + "issue:45399", + "issue:45464", + "issue:45542", + "issue:45632", + "issue:45685", + "issue:45698", + "issue:45759", + "issue:45800", + "issue:45823", + "issue:45901", + "issue:45902" + ], + "soft_pairs": [ + "issue:39290|issue:43531", + "issue:43792|issue:45901", + "issue:43425|issue:45070", + "issue:43596|issue:45800", + "issue:43975|issue:45901", + "issue:43792|issue:45070", + "issue:45362|issue:45464", + "issue:44292|issue:45901", + "issue:39290|issue:45823", + "issue:43404|issue:43866", + "issue:45464|issue:45685", + "issue:43541|issue:45685", + "issue:43541|issue:43828", + "issue:44841|issue:45084", + "issue:36010|issue:39290", + "issue:41093|issue:41762", + "issue:45800|issue:45823", + "issue:43296|issue:43366", + "issue:43572|issue:45070", + "issue:44387|issue:45005", + "issue:43541|issue:45161", + "issue:43828|issue:45464", + "issue:44291|issue:45070", + "issue:43366|issue:45464", + "issue:45020|issue:45800", + "issue:43531|issue:45070", + "issue:41093|issue:43596", + "issue:43541|issue:44589", + "issue:43054|issue:43994", + "issue:45542|issue:45800", + "issue:45759|issue:45902", + "issue:40990|issue:44960", + "issue:39186|issue:44918", + "issue:43716|issue:44918", + "issue:45632|issue:45698", + "issue:43404|issue:45005", + "issue:43645|issue:45399", + "issue:44530|issue:45005", + "issue:42915|issue:45399" + ], + "prepared_review_unit_hash": "b4d759ca2c2da0c69a39553ab66b552df56b66b52501b8432dadf66c223f2ece", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 39, + "serialized_chars": 39064, + "estimated_input_tokens": 9766, + "estimated_eval_tokens": 19788 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a heterogeneous mix of unrelated Transformers regressions across different models, loaders, precision modes, and runtime backends. The only strong duplicate-like pair is the SigLIP2 output-quality reports.", + "confidence": 0.76, + "canonical_issue_reason": "Use issue 43994 as the canonical issue for the SigLIP2 regression: it states the broader user-visible failure more directly, while 43054 reads as a narrower comparative symptom report.", + "canonical_pr_reason": null, + "best_issue_reason": "If a single issue must represent the one clear duplicate subgroup here, 43994 is the clearest and most general report. There is no strong cluster-wide canonical because the rest of the items are not the same bug.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:39290", + "right": "issue:43531", + "accept": false, + "reason": "Both mention sliding_window, but they involve different model families and different failure contexts; not the same concrete bug." + }, + { + "left": "issue:43792", + "right": "issue:45901", + "accept": false, + "reason": "Whisper loading failure and table QA crash are unrelated code paths and symptoms." + }, + { + "left": "issue:43425", + "right": "issue:45070", + "accept": false, + "reason": "Different regressions: one is a torch version incompatibility, the other is a PretrainedConfig/pydantic breakage." + }, + { + "left": "issue:43596", + "right": "issue:45800", + "accept": false, + "reason": "Both are version/runtime errors, but the repros and causes differ substantially (Deepspeed/Bert vs torch-transformers incompatibility)." + }, + { + "left": "issue:43975", + "right": "issue:45901", + "accept": false, + "reason": "Detokenization bug and table-question-answering crash are unrelated." + }, + { + "left": "issue:45362", + "right": "issue:45464", + "accept": false, + "reason": "Same broad model family, but one is a transformers chat crash and the other is a streaming chat/completions API failure; not enough evidence of one bug." + }, + { + "left": "issue:44292", + "right": "issue:45901", + "accept": false, + "reason": "Qwen-3 NVFP4 runtime failure is unrelated to the table QA crash." + }, + { + "left": "issue:39290", + "right": "issue:45823", + "accept": false, + "reason": "Different models and different symptoms: sliding_window attribute error vs device mismatch in forward." + }, + { + "left": "issue:43404", + "right": "issue:43866", + "accept": false, + "reason": "lm_head weight tying in Mistral3 and a corrupted Ovis2 checkpoint are different problems." + }, + { + "left": "issue:45464", + "right": "issue:45685", + "accept": false, + "reason": "Streaming inference API failure and MPS histogram error are unrelated." + }, + { + "left": "issue:43541", + "right": "issue:45685", + "accept": false, + "reason": "Different backends and ops; no shared underlying code-path bug is apparent." + }, + { + "left": "issue:43541", + "right": "issue:43828", + "accept": false, + "reason": "Both are precision/runtime issues, but they affect different models and failure mechanisms." + }, + { + "left": "issue:44841", + "right": "issue:45084", + "accept": false, + "reason": "Processor loading failure and template compilation error are not the same issue." + }, + { + "left": "issue:36010", + "right": "issue:39290", + "accept": false, + "reason": "ImportError for GenerationMixin and a sliding_window config error are unrelated." + }, + { + "left": "issue:41093", + "right": "issue:41762", + "accept": false, + "reason": "Both are index/shape errors during loading, but the repro paths and underlying tensors differ." + }, + { + "left": "issue:45800", + "right": "issue:45823", + "accept": false, + "reason": "Torch version incompatibility and a Gemma4 device mismatch are different bugs." + }, + { + "left": "issue:43296", + "right": "issue:43366", + "accept": false, + "reason": "PaddleOCR-VL loading in vLLM and GPT-OSS GGUF support are unrelated requests/failures." + }, + { + "left": "issue:43572", + "right": "issue:45070", + "accept": false, + "reason": "Both are v5 regressions, but they hit different config fields and different failure modes." + }, + { + "left": "issue:44387", + "right": "issue:45005", + "accept": false, + "reason": "Quantization OOM and tied-weights behavior are unrelated." + }, + { + "left": "issue:43541", + "right": "issue:45161", + "accept": false, + "reason": "Both mention MoE/parallelism, but the concrete failures differ too much to be duplicates." + }, + { + "left": "issue:43828", + "right": "issue:45464", + "accept": false, + "reason": "Autocast dtype mismatch and chat/completions streaming failure are not the same bug." + }, + { + "left": "issue:44291", + "right": "issue:45070", + "accept": false, + "reason": "Unexpected _is_hf_initialized argument and a pydantic PretrainedConfig breakage are different regressions." + }, + { + "left": "issue:43366", + "right": "issue:45464", + "accept": false, + "reason": "GPT-OSS GGUF architecture support and Qwen streaming inference are unrelated." + }, + { + "left": "issue:45020", + "right": "issue:45800", + "accept": false, + "reason": "Generic remote_code breakage and torch-transformers incompatibility are different root causes." + }, + { + "left": "issue:43531", + "right": "issue:45070", + "accept": false, + "reason": "Sliding_window handling and PretrainedConfig field regression do not look like the same bug." + }, + { + "left": "issue:41093", + "right": "issue:43596", + "accept": false, + "reason": "Both are index errors with Deepspeed-related loading, but the affected models and failure shapes differ." + }, + { + "left": "issue:43541", + "right": "issue:44589", + "accept": false, + "reason": "Grouped_mm tracing failure and missing Float8 storage are different low-level errors." + }, + { + "left": "issue:43054", + "right": "issue:43994", + "accept": true, + "reason": "Both reports describe SigLIP2 producing bad/nonsensical text-embedding results across interfaces, which looks like the same underlying model bug." + }, + { + "left": "issue:43645", + "right": "issue:45399", + "accept": false, + "reason": "Custom model initialization in notebooks and flash-attn fallback blocking are unrelated." + }, + { + "left": "issue:42915", + "right": "issue:45399", + "accept": false, + "reason": "FineGrainedFP8Config failure and flash-attn fallback logic are different issues." + }, + { + "left": "issue:45632", + "right": "issue:45698", + "accept": false, + "reason": "Both concern custom-module loading, but one is a leaf-directory cache collision and the other is wrong module loading after save_pretrained; not enough evidence they are the same bug." + }, + { + "left": "issue:40990", + "right": "issue:44960", + "accept": false, + "reason": "High perplexity on gpt-oss and GLM5 are unrelated model-specific regressions." + }, + { + "left": "issue:39186", + "right": "issue:44918", + "accept": false, + "reason": "FSDP weight shape error and Qwen3.5 embedding unpacking are different failures." + }, + { + "left": "issue:43716", + "right": "issue:44918", + "accept": false, + "reason": "Gemma-3 image/text dtype mismatch and Qwen3.5 input embedding unpacking do not share a concrete code path." + }, + { + "left": "issue:45632", + "right": "issue:45698", + "accept": false, + "reason": "Both are about custom code loading, but the titles point to different mechanisms and different failure conditions." + }, + { + "left": "issue:43404", + "right": "issue:45005", + "accept": false, + "reason": "Tied lm_head in Mistral3 and translation-model tied-weights issues are not the same bug." + }, + { + "left": "issue:43645", + "right": "issue:45399", + "accept": false, + "reason": "Notebook custom-model initialization and flash-attn fallback behavior are unrelated." + }, + { + "left": "issue:44530", + "right": "issue:45005", + "accept": false, + "reason": "PagedAttentionCache linear_attention failure and tied-weights translation issues are unrelated." + }, + { + "left": "issue:42915", + "right": "issue:45399", + "accept": false, + "reason": "Qwen3Moe FP8 failure and flash-attn fallback gating are different issues." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: the cluster is largely heterogeneous by model and failure mode, and the only accepted duplicate pair (43054/43994) is plausibly the same SigLIP2 regression based on the titles. The soft-edge rejections are generally conservative and well-justified." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 3, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:36331", + "issue:38175", + "issue:39692", + "issue:41553", + "issue:41628", + "issue:41762", + "issue:42915", + "issue:42947", + "issue:43054", + "issue:43493", + "issue:43643", + "issue:43644", + "issue:43782", + "issue:43828", + "issue:43856", + "issue:43866", + "issue:43872", + "issue:43873", + "issue:43881", + "issue:43883", + "issue:43950", + "issue:43975", + "issue:44292", + "issue:44360", + "issue:44387", + "issue:44403", + "issue:44451", + "issue:44488", + "issue:44492", + "issue:44512", + "issue:44534", + "issue:44589", + "issue:44863", + "issue:44910", + "issue:44918", + "issue:44928", + "issue:44960", + "issue:45005", + "issue:45020", + "issue:45507", + "issue:45520", + "issue:45663", + "issue:45685", + "issue:45701", + "issue:45800", + "issue:45823", + "issue:45859", + "issue:45902" + ], + "soft_pairs": [ + "issue:43873|issue:45005", + "issue:44360|issue:44512", + "issue:44863|issue:45005", + "issue:44492|issue:44512", + "issue:44918|issue:45507", + "issue:43975|issue:44292", + "issue:43950|issue:44534", + "issue:44292|issue:45507", + "issue:39692|issue:43054", + "issue:43644|issue:43950", + "issue:43828|issue:44928", + "issue:36331|issue:41762", + "issue:41628|issue:45800", + "issue:43950|issue:45902", + "issue:43866|issue:44403", + "issue:43872|issue:44589", + "issue:43054|issue:45701", + "issue:38175|issue:43054", + "issue:41553|issue:43883", + "issue:44910|issue:45859", + "issue:45663|issue:45902", + "issue:41553|issue:43881", + "issue:42947|issue:44387", + "issue:43054|issue:43493", + "issue:44403|issue:44451", + "issue:43643|issue:45020", + "issue:43975|issue:45520", + "issue:44403|issue:44488", + "issue:42947|issue:43856", + "issue:41762|issue:42915", + "issue:43782|issue:44960", + "issue:43881|issue:43883", + "issue:45685|issue:45823" + ], + "prepared_review_unit_hash": "4168fccd8f77e3ede6f9154eb56cdbf01b1d47310a77866f4557ff249063b56e", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37200, + "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 mostly heterogeneous. The only clear duplicate subgroup is the Transformers v5 non-persistent-buffer corruption regression; most other pairs are related by subsystem or model family only, not the same underlying bug.", + "confidence": 0.86, + "canonical_issue_reason": "Issue 43950 is the best canonical issue for the only true duplicate set here: it clearly describes the non-persistent-buffer corruption regression in v5, which 43644 and 44534 restate.", + "canonical_pr_reason": null, + "best_issue_reason": "43950 is the strongest standalone issue in the cluster because it is specific about the failure mode and regression, making it the best anchor for deduping the matching buffer-corruption reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43873", + "right": "issue:45005", + "accept": false, + "reason": "Different problems: quantization/offloading behavior vs tied-weights issues in translation models." + }, + { + "left": "issue:44360", + "right": "issue:44512", + "accept": false, + "reason": "Unrelated: a DSA indexer bug versus a docs typo about a removed run command." + }, + { + "left": "issue:44863", + "right": "issue:45005", + "accept": false, + "reason": "Different model-loading failure than the tied-weights v5 issue." + }, + { + "left": "issue:44492", + "right": "issue:44512", + "accept": false, + "reason": "Both are documentation-related, but they are distinct doc issues with different content." + }, + { + "left": "issue:44918", + "right": "issue:45507", + "accept": false, + "reason": "Different code paths and failures: TRL embedding unpacking versus an invalid method call in GraniteMoEHybrid." + }, + { + "left": "issue:43975", + "right": "issue:44292", + "accept": false, + "reason": "Different runtime failures affecting different models; no clear shared root cause." + }, + { + "left": "issue:43950", + "right": "issue:44534", + "accept": true, + "reason": "Same underlying regression: v5/from_pretrained mishandles non-persistent buffers and leaves them as junk/corrupted." + }, + { + "left": "issue:44292", + "right": "issue:45507", + "accept": false, + "reason": "Distinct model-specific loading/runtime errors, not the same bug." + }, + { + "left": "issue:39692", + "right": "issue:43054", + "accept": false, + "reason": "Both mention SigLIP2, but one is a docs/example problem and the other is an embedding-quality complaint." + }, + { + "left": "issue:43644", + "right": "issue:43950", + "accept": true, + "reason": "Same bug: non-persistent buffers are being filled with junk/corrupted during v5 loading." + }, + { + "left": "issue:43828", + "right": "issue:44928", + "accept": false, + "reason": "Different issues: autocast dtype mismatch versus Qwen3.5 position-id/SDPA behavior." + }, + { + "left": "issue:36331", + "right": "issue:41762", + "accept": false, + "reason": "Completely different failures: trainer API mismatch versus ZeRO-3 loading index error." + }, + { + "left": "issue:41628", + "right": "issue:45800", + "accept": false, + "reason": "Different bugs: missing import/export versus a torch/transformers compatibility problem." + }, + { + "left": "issue:43950", + "right": "issue:45902", + "accept": false, + "reason": "Related buffer/memory symptoms, but not the same concrete bug or code path." + }, + { + "left": "issue:43866", + "right": "issue:44403", + "accept": false, + "reason": "Checkpoint corruption and noisy loading warnings are not the same underlying issue." + }, + { + "left": "issue:43872", + "right": "issue:44589", + "accept": false, + "reason": "Different quantization/storage incompatibilities with different error surfaces." + }, + { + "left": "issue:43054", + "right": "issue:45701", + "accept": false, + "reason": "SigLIP2 embedding quality versus a general tokenization-version change; no direct duplicate evidence." + }, + { + "left": "issue:38175", + "right": "issue:43054", + "accept": false, + "reason": "Same model family, but the reported symptoms are too different to treat as one bug." + }, + { + "left": "issue:41553", + "right": "issue:43883", + "accept": false, + "reason": "Unrelated loading/error-reporting issues for different models." + }, + { + "left": "issue:44910", + "right": "issue:45859", + "accept": false, + "reason": "Different Qwen3.5-related failures: flash-attn crash versus CPU-offload incompatibility." + }, + { + "left": "issue:45663", + "right": "issue:45902", + "accept": false, + "reason": "Different Gemma4 vs Qwen3.5 code paths and failure modes." + }, + { + "left": "issue:41553", + "right": "issue:43881", + "accept": false, + "reason": "Distinct model-loading problems with different causes." + }, + { + "left": "issue:42947", + "right": "issue:44387", + "accept": false, + "reason": "Both are memory-related, but one is LoRA checkpointing behavior and the other is int4 quantization OOM." + }, + { + "left": "issue:43054", + "right": "issue:43493", + "accept": false, + "reason": "Closely related SigLIP2 complaints, but not clearly the same concrete bug." + }, + { + "left": "issue:44403", + "right": "issue:44451", + "accept": false, + "reason": "Generic loading noise versus a specific model-loading failure; not the same issue." + }, + { + "left": "issue:43643", + "right": "issue:45020", + "accept": false, + "reason": "Remote-code regressions are related, but one is missing fields in AutoConfig and the other is a broader compatibility complaint." + }, + { + "left": "issue:43975", + "right": "issue:45520", + "accept": false, + "reason": "Different failure classes: detokenization regression versus flash_attn import error." + }, + { + "left": "issue:44403", + "right": "issue:44488", + "accept": false, + "reason": "Both are model-loading problems, but they concern different models and likely different causes." + }, + { + "left": "issue:42947", + "right": "issue:43856", + "accept": false, + "reason": "Both mention memory, but one is gradient-checkpointing/LoRA and the other is MoE training efficiency." + }, + { + "left": "issue:41762", + "right": "issue:42915", + "accept": false, + "reason": "Different loading failures for different models/configurations." + }, + { + "left": "issue:43782", + "right": "issue:44960", + "accept": false, + "reason": "Different models and unrelated errors." + }, + { + "left": "issue:43881", + "right": "issue:43883", + "accept": false, + "reason": "Separate model-loading bugs with different error signatures." + }, + { + "left": "issue:45685", + "right": "issue:45823", + "accept": false, + "reason": "Different issues: MPS histogram support versus Gemma4 device placement mismatch." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded and conservative overall. The only clear duplicate subgroup is the Transformers v5 non-persistent-buffer corruption regression centered on issue 43950, and the other soft-edge calls are phrased cautiously and match the packet titles/body themes without overclaiming. " + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 4, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:24643", + "issue:30064", + "issue:36331", + "issue:41093", + "issue:41553", + "issue:42491", + "issue:42915", + "issue:42947", + "issue:43257", + "issue:43366", + "issue:43404", + "issue:43638", + "issue:43742", + "issue:43792", + "issue:43824", + "issue:43827", + "issue:43828", + "issue:43854", + "issue:43866", + "issue:43872", + "issue:43940", + "issue:43950", + "issue:44220", + "issue:44291", + "issue:44368", + "issue:44479", + "issue:44488", + "issue:44509", + "issue:44530", + "issue:44560", + "issue:44661", + "issue:44805", + "issue:44863", + "issue:44928", + "issue:44991", + "issue:45020", + "issue:45081", + "issue:45305", + "issue:45313", + "issue:45325", + "issue:45356", + "issue:45399", + "issue:45538", + "issue:45588", + "issue:45663", + "issue:45823", + "issue:45902", + "issue:45923" + ], + "soft_pairs": [ + "issue:42491|issue:44368", + "issue:43824|issue:43872", + "issue:45020|issue:45823", + "issue:44991|issue:45399", + "issue:44661|issue:45020", + "issue:43828|issue:45823", + "issue:45081|issue:45399", + "issue:43827|issue:44509", + "issue:45020|issue:45356", + "issue:43404|issue:43950", + "issue:43742|issue:45399", + "issue:42915|issue:45020", + "issue:43638|issue:45823", + "issue:43257|issue:43866", + "issue:43792|issue:44220", + "issue:43872|issue:44291", + "issue:42915|issue:43872", + "issue:43940|issue:45313", + "issue:41553|issue:44991", + "issue:43854|issue:44863", + "issue:36331|issue:41093", + "issue:43257|issue:45305", + "issue:42947|issue:45923", + "issue:44928|issue:45588", + "issue:43872|issue:45538", + "issue:41093|issue:44805", + "issue:41553|issue:44488", + "issue:44479|issue:44560", + "issue:44530|issue:44863", + "issue:24643|issue:30064", + "issue:45325|issue:45902", + "issue:43366|issue:43828", + "issue:45663|issue:45823" + ], + "prepared_review_unit_hash": "ee69c055fddf0a4a383ec368cd53ce1436ce65428fcad8bb753988d6a409af46", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 38120, + "estimated_input_tokens": 9530, + "estimated_eval_tokens": 19316 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly a loose family of model-loading/regression reports, not one tight duplicate set. A few soft edges are solid duplicates, especially the docs pipeline-removal pair, the `_is_hf_initialized` / `init_empty_weights` compatibility pair, the Qwen VL video regression pair, and the mask-shape IndexError pair for 41093/44805. The 36331/41093 link should be rejected: those are different bugs.", + "confidence": 0.72, + "canonical_issue_reason": "issue:45020 is the broadest representative issue in the cluster: it captures the recent-transformers/model-loading regression theme that many of the narrower reports orbit, even though the cluster is not a single exact-duplicate family.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45020 is the best anchor because it is the most umbrella-like and representative of the cluster\u2019s shared regression pattern, rather than one narrow symptom report.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42491", + "right": "issue:44368", + "accept": false, + "reason": "Related Qwen3/LoRA themes, but one is a cross-version LoRA compatibility break and the other is a tie-word-embeddings warning; different failure modes." + }, + { + "left": "issue:43824", + "right": "issue:43872", + "accept": false, + "reason": "Missing Qwen2_5_VL import vs bitsandbytes `Int8Params.__new__` incompatibility; unrelated code paths." + }, + { + "left": "issue:45020", + "right": "issue:45823", + "accept": false, + "reason": "Broad remote_code/model-loading regression vs Gemma4 PLE device mismatch during forward; not the same bug." + }, + { + "left": "issue:44991", + "right": "issue:45399", + "accept": false, + "reason": "Tokenizer load failure for one model vs flash-attention fallback gating; same broad area, different root cause." + }, + { + "left": "issue:44661", + "right": "issue:45020", + "accept": false, + "reason": "Tokenizer-mapping edge case in `add-new-model-like` vs broad remote_code regression; different defects." + }, + { + "left": "issue:43828", + "right": "issue:45823", + "accept": false, + "reason": "Autocast dtype mismatch in Phi MoE vs Gemma4 PLE device placement mismatch; distinct bugs." + }, + { + "left": "issue:45081", + "right": "issue:45399", + "accept": false, + "reason": "Tokenizer regex patch crash vs flash-attn fallback blocked by other checks; different subsystems." + }, + { + "left": "issue:43827", + "right": "issue:44509", + "accept": true, + "reason": "Same docs regression after v5 pipeline removals; both report stale references to removed summarization/translation pipeline tasks." + }, + { + "left": "issue:45020", + "right": "issue:45356", + "accept": false, + "reason": "General remote_code breakage vs Kimi-K2.5 tokenizer codec / regex-warning regression; different failure paths." + }, + { + "left": "issue:43404", + "right": "issue:43950", + "accept": false, + "reason": "Weight-tying bug in Mistral3 vs non-persistent buffer corruption on `from_pretrained()`; different loading mechanisms." + }, + { + "left": "issue:43742", + "right": "issue:45399", + "accept": false, + "reason": "MobileLLM key error vs flash-attention fallback gating; both load-time issues, but not the same defect." + }, + { + "left": "issue:42915", + "right": "issue:45020", + "accept": false, + "reason": "Specific Qwen3 MoE FP8 config failure vs broad remote_code/model-loading regression; too different in scope and cause." + }, + { + "left": "issue:43638", + "right": "issue:45823", + "accept": false, + "reason": "ZeRO-3 index error on a Bert model vs Gemma4 PLE device mismatch; unrelated root causes." + }, + { + "left": "issue:43257", + "right": "issue:43866", + "accept": false, + "reason": "Qwen3 MoE weight conversion under accelerate+deepspeed vs corrupted Ovis2 checkpoint; not the same bug." + }, + { + "left": "issue:43792", + "right": "issue:44220", + "accept": false, + "reason": "Whisper runtime failure vs `_torch_extract_fbank_features()` issue; both audio-related but distinct." + }, + { + "left": "issue:43872", + "right": "issue:44291", + "accept": true, + "reason": "Same compatibility regression around `init_empty_weights` / `_is_hf_initialized` being passed into `Int8Params` and breaking model loading." + }, + { + "left": "issue:42915", + "right": "issue:43872", + "accept": false, + "reason": "Qwen3 MoE FP8 failure and bitsandbytes parameter-construction incompatibility are different loading/configuration bugs." + }, + { + "left": "issue:43940", + "right": "issue:45313", + "accept": false, + "reason": "Both are DeepSpeed ZeRO-3 loading issues, but they describe different Qwen3-family failures and different missing-weight scenarios." + }, + { + "left": "issue:41553", + "right": "issue:44991", + "accept": false, + "reason": "Bad tokenizer error for Voxtral vs tokenizer load failure for EMBEDDIA/est-roberta; same area, different bug." + }, + { + "left": "issue:43854", + "right": "issue:44863", + "accept": false, + "reason": "GLM-4.7-Flash unit-test loading issue vs NemotronH checkpoint-loading failure; different model and code path." + }, + { + "left": "issue:36331", + "right": "issue:41093", + "accept": false, + "reason": "`compute_loss()` got unexpected `num_items_in_batch` TypeError vs mask-shape IndexError; the titles and symptoms do not support the same bug." + }, + { + "left": "issue:43257", + "right": "issue:45305", + "accept": false, + "reason": "Missing weight conversion under DeepSpeed vs gradient averaging not happening with GAS and `model_accepts_loss_kwargs=True`; different behaviors." + }, + { + "left": "issue:42947", + "right": "issue:45923", + "accept": false, + "reason": "Both mention gradient checkpointing, but one is ineffective PEFT LoRA checkpointing and the other is a missing capability flag on a trust_remote_code model." + }, + { + "left": "issue:44928", + "right": "issue:45588", + "accept": false, + "reason": "RLHF gradient explosion from 3D position_ids / SDPA fallback vs flash_attention crash on `s_aux=None`; different bugs." + }, + { + "left": "issue:43872", + "right": "issue:45538", + "accept": false, + "reason": "bitsandbytes parameter-construction regression vs CLIPTokenizer max-length sentinel; unrelated." + }, + { + "left": "issue:41093", + "right": "issue:44805", + "accept": true, + "reason": "Same mask-shape IndexError, just with slightly different tensor sizes and wording." + }, + { + "left": "issue:41553", + "right": "issue:44488", + "accept": false, + "reason": "Both are loading-related, but one is a Voxtral tokenizer error and the other is a separate failure loading `cjvt/sleng-bert`." + }, + { + "left": "issue:44479", + "right": "issue:44560", + "accept": true, + "reason": "Both describe the same v5.3.0 video-input regression in the Qwen VL family; `44560` looks like a concrete manifestation of the broader bug." + }, + { + "left": "issue:44530", + "right": "issue:44863", + "accept": false, + "reason": "PagedAttentionCache `linear_attention` crash vs NemotronH checkpoint-loading failure; unrelated code paths." + }, + { + "left": "issue:24643", + "right": "issue:30064", + "accept": false, + "reason": "DeepSpeed training weight-shape RuntimeError vs image-processor void-segmentation-map failure; no shared root cause." + }, + { + "left": "issue:45325", + "right": "issue:45902", + "accept": false, + "reason": "Qwen2.5-VL rope-index scaling bug vs uninitialized-memory after `to_empty(cuda)` materialization; different internals." + }, + { + "left": "issue:43366", + "right": "issue:43828", + "accept": false, + "reason": "GGUF gpt-oss support request vs Phi MoE autocast dtype mismatch; not the same bug." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "Reject: one accepted soft-edge verdict is not grounded in the packet. `issue:36331` is a `compute_loss()` unexpected `num_items_in_batch` TypeError, while `issue:41093` is a mask-shape IndexError; the titles do not support treating them as the same bug. Because of that, the summary also overstates the cluster\u2019s duplicate strength by including this pair among the strongest matches. The other acceptances (`43827`/`44509`, `43872`/`44291`, `44479`/`44560`, `41093`/`44805`) are more conservative and plausible." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 5, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:36010", + "issue:38175", + "issue:39692", + "issue:41720", + "issue:42491", + "issue:42617", + "issue:42915", + "issue:43317", + "issue:43366", + "issue:43531", + "issue:43550", + "issue:43596", + "issue:43632", + "issue:43638", + "issue:43644", + "issue:43645", + "issue:43646", + "issue:43708", + "issue:43720", + "issue:43824", + "issue:43854", + "issue:43931", + "issue:43950", + "issue:43957", + "issue:43994", + "issue:44075", + "issue:44155", + "issue:44451", + "issue:44488", + "issue:44534", + "issue:44560", + "issue:44589", + "issue:44661", + "issue:44756", + "issue:44805", + "issue:44877", + "issue:44910", + "issue:44918", + "issue:44960", + "issue:44977", + "issue:45030", + "issue:45276", + "issue:45313", + "issue:45357", + "issue:45584", + "issue:45823", + "issue:45834" + ], + "soft_pairs": [ + "issue:44155|issue:45584", + "issue:43366|issue:43531", + "issue:43950|issue:45834", + "issue:43550|issue:44910", + "issue:43824|issue:43957", + "issue:43645|issue:43646", + "issue:42915|issue:43957", + "issue:43366|issue:44589", + "issue:44560|issue:44918", + "issue:42617|issue:43366", + "issue:36010|issue:42915", + "issue:44661|issue:45823", + "issue:38175|issue:43994", + "issue:43632|issue:45834", + "issue:44560|issue:44805", + "issue:43638|issue:44661", + "issue:43596|issue:44661", + "issue:44661|issue:44805", + "issue:44877|issue:45030", + "issue:43317|issue:44756", + "issue:38175|issue:39692", + "issue:43644|issue:44534", + "issue:44960|issue:44977", + "issue:42491|issue:43720", + "issue:45313|issue:45357", + "issue:41720|issue:42915", + "issue:44451|issue:44488", + "issue:42491|issue:45276", + "issue:42915|issue:44756", + "issue:43708|issue:44075", + "issue:43854|issue:43931" + ], + "prepared_review_unit_hash": "98cae0867a07c7d3aa5eae99f04da9e45b7e5fa2653a3b1d5efedc3f3229c6c5", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36262, + "estimated_input_tokens": 9066, + "estimated_eval_tokens": 18388 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly heterogeneous, but a few soft pairs look like true duplicates: the non-persistent buffer regression, the custom model initialization regression, and the SigLIP2 output regression. The rest are too model- or symptom-specific to merge safely.", + "confidence": 0.83, + "canonical_issue_reason": "Issue 43644 is the clearest canonical issue among the near-duplicates: 44534 reports the same Transformers 5.x non-persistent-buffer corruption regression with essentially the same symptom.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43644 is the strongest issue to keep as the representative tracker because it is concrete, broadly applicable, and has an exact near-duplicate.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44155", + "right": "issue:45584", + "accept": false, + "reason": "Different models and failure modes: batched AudioFlamingo track leakage vs Whisper empty-transcription generation." + }, + { + "left": "issue:43366", + "right": "issue:43531", + "accept": false, + "reason": "Both mention Qwen/GGUF-style model issues, but they describe different concrete bugs and code paths." + }, + { + "left": "issue:43950", + "right": "issue:45834", + "accept": false, + "reason": "Non-persistent buffer corruption is unrelated to Kosmos2.5 OCR index errors." + }, + { + "left": "issue:43550", + "right": "issue:44910", + "accept": false, + "reason": "Different regressions: Bamba torch.compile/SDPA vs Qwen3.5 flash-attention position_ids handling." + }, + { + "left": "issue:43824", + "right": "issue:43957", + "accept": false, + "reason": "Importing a missing class and meta-device loading failures are distinct problems." + }, + { + "left": "issue:43645", + "right": "issue:43646", + "accept": true, + "reason": "These read like the same custom-model initialization regression in Transformers 5.0.0, with the notebook mention just a narrower manifestation." + }, + { + "left": "issue:42915", + "right": "issue:43957", + "accept": false, + "reason": "Qwen3Moe FP8 failure and generic meta-device loading issues are different bugs." + }, + { + "left": "issue:43366", + "right": "issue:44589", + "accept": false, + "reason": "GGUF support request and Float8 storage lookup failure are unrelated." + }, + { + "left": "issue:44560", + "right": "issue:44918", + "accept": false, + "reason": "Both involve Qwen3.x, but one is a video StopIteration error and the other is an input-embedding unpacking failure in TRL." + }, + { + "left": "issue:42617", + "right": "issue:43366", + "accept": false, + "reason": "3d_parallel.py execution failure is not the same as GGUF architecture support." + }, + { + "left": "issue:36010", + "right": "issue:42915", + "accept": false, + "reason": "GenerationMixin import error and Qwen3Moe FP8 failure are unrelated." + }, + { + "left": "issue:44661", + "right": "issue:45823", + "accept": false, + "reason": "Tokenizer mapping/model registration logic is unrelated to Gemma4 PLE device placement during forward." + }, + { + "left": "issue:38175", + "right": "issue:43994", + "accept": true, + "reason": "Same underlying SigLIP2 regression: both report incorrect/degenerate outputs for google/siglip2-base-patch16-224 with AutoModel/pipeline usage." + }, + { + "left": "issue:43632", + "right": "issue:45834", + "accept": false, + "reason": "HF initialization flag regression and Kosmos2.5 OCR index errors are not the same bug." + }, + { + "left": "issue:44560", + "right": "issue:44805", + "accept": false, + "reason": "Different errors and model paths: video StopIteration vs mask/tensor shape mismatch." + }, + { + "left": "issue:43638", + "right": "issue:44661", + "accept": false, + "reason": "Zero3/Bert indexing failure is unrelated to tokenizer mapping registration constraints." + }, + { + "left": "issue:43596", + "right": "issue:44661", + "accept": false, + "reason": "A BertModel Zero3 init IndexError is not the same as the TOKENIZER_MAPPING_NAMES add-new-model-like failure." + }, + { + "left": "issue:44661", + "right": "issue:44805", + "accept": false, + "reason": "Tokenizer mapping validation and mask-shape IndexErrors are different code paths." + }, + { + "left": "issue:44877", + "right": "issue:45030", + "accept": false, + "reason": "Both mention config validation, but they are distinct model/config failures with different triggers." + }, + { + "left": "issue:43317", + "right": "issue:44756", + "accept": false, + "reason": "GPU/CPU offload dequantization failure and Strix Halo mmap OOM are different loading problems." + }, + { + "left": "issue:38175", + "right": "issue:39692", + "accept": false, + "reason": "Same model family, but one is a degenerate output bug and the other is a documentation example with model/processor mismatch and quantization failure." + }, + { + "left": "issue:43644", + "right": "issue:44534", + "accept": true, + "reason": "Near-exact duplicate: both report Transformers 5.x corrupting non-persistent buffers; 44534 is essentially a reworded repeat." + }, + { + "left": "issue:44960", + "right": "issue:44977", + "accept": false, + "reason": "GLM5 issue and Qwen3.5 flash-attention generation failure are unrelated." + }, + { + "left": "issue:42491", + "right": "issue:43720", + "accept": false, + "reason": "LoRA/Qwen3_moe hf4.x compatibility and BitNet packed-weight loading are separate regressions." + }, + { + "left": "issue:45313", + "right": "issue:45357", + "accept": false, + "reason": "ZeRO-3 language-model weight loading and save_pretrained visual-encoder key serialization are different problems." + }, + { + "left": "issue:41720", + "right": "issue:42915", + "accept": false, + "reason": "Qwen3 auto device mapping cuda assert and Qwen3Moe FP8 failure are not the same bug." + }, + { + "left": "issue:44451", + "right": "issue:44488", + "accept": false, + "reason": "Both are BERT model-loading complaints, but the titles point to different models and no shared concrete failure is evident." + }, + { + "left": "issue:43854", + "right": "issue:43931", + "accept": false, + "reason": "GLM-4.7-Flash unit-test loading and Qwen3-VL-30B-A3B weight-shape mismatch are different model-specific issues." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The judgments are mostly conservative and well grounded in the packet. The accepted soft pairs are plausible same-bug matches based on closely aligned titles and symptoms, especially 43644/44534, 43645/43646, and 38175/43994. The rejected pairs generally stay on the safe side by requiring more than superficial model-family overlap. The canonical issue choice and summary are supported by the evidence provided." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 6, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:36010", + "issue:39404", + "issue:39692", + "issue:41720", + "issue:42491", + "issue:42947", + "issue:43257", + "issue:43335", + "issue:43404", + "issue:43425", + "issue:43454", + "issue:43644", + "issue:43645", + "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:44488", + "issue:44743", + "issue:44756", + "issue:44910", + "issue:44912", + "issue:44960", + "issue:44991", + "issue:45216", + "issue:45276", + "issue:45292", + "issue:45335", + "issue:45362", + "issue:45406", + "issue:45488", + "issue:45588", + "issue:45663", + "issue:45684", + "issue:45698", + "issue:45704", + "issue:45759", + "issue:45812", + "issue:45823", + "issue:45859" + ], + "soft_pairs": [ + "issue:45488|issue:45812", + "issue:39692|issue:43994", + "issue:43257|issue:43854", + "issue:44488|issue:44991", + "issue:42491|issue:43931", + "issue:43644|issue:43645", + "issue:43856|issue:43873", + "issue:45362|issue:45406", + "issue:43856|issue:44387", + "issue:44910|issue:45588", + "issue:42491|issue:45292", + "issue:45276|issue:45335", + "issue:44164|issue:44291", + "issue:36010|issue:39404", + "issue:43404|issue:43454", + "issue:41720|issue:44155", + "issue:39692|issue:44960", + "issue:43425|issue:43957", + "issue:44402|issue:44488", + "issue:42947|issue:45704", + "issue:45663|issue:45759", + "issue:43873|issue:44387", + "issue:43335|issue:45663", + "issue:45663|issue:45859", + "issue:44291|issue:44756", + "issue:44155|issue:45823", + "issue:44292|issue:44912", + "issue:44387|issue:44756", + "issue:43854|issue:45216", + "issue:44743|issue:44960", + "issue:43957|issue:44756", + "issue:45276|issue:45292", + "issue:45684|issue:45698" + ], + "prepared_review_unit_hash": "813abd841e93cbee47f5e7109ee5c44197c7a16f99b69cbfc67ba7fdb9494fc3", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37775, + "estimated_input_tokens": 9444, + "estimated_eval_tokens": 19144 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is mostly a loose collection of unrelated Transformers issues. The only clear duplicate-like cluster is around `resize_token_embeddings` not propagating to nested/output embeddings across related model variants.", + "confidence": 0.69, + "canonical_issue_reason": "`issue:45292` is the most generic and reusable statement of the shared `resize_token_embeddings` regression, so it best serves as the representative issue for the small embedding-resize cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "`issue:45292` is the broadest, least model-specific report among the candidates, and it cleanly subsumes the more specific Gemma4/T5Gemma variants.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45488", + "right": "issue:45812", + "accept": false, + "reason": "Different tokenizer regressions in different model families; same broad area, not the same bug." + }, + { + "left": "issue:39692", + "right": "issue:43994", + "accept": false, + "reason": "Both involve SigLIP2, but one is a docs/example problem and the other is a model loading/output issue." + }, + { + "left": "issue:43257", + "right": "issue:43854", + "accept": false, + "reason": "Different models and failure modes during loading; no concrete shared code-path is evident." + }, + { + "left": "issue:44488", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer-loading complaints, but for different models and likely different root causes." + }, + { + "left": "issue:42491", + "right": "issue:43931", + "accept": false, + "reason": "Both mention Qwen3 loading issues, but one is LoRA/version compatibility and the other is a shape mismatch in a different model." + }, + { + "left": "issue:43644", + "right": "issue:43645", + "accept": false, + "reason": "Distinct regressions: junk-filled buffers vs Jupyter custom-model initialization." + }, + { + "left": "issue:43856", + "right": "issue:43873", + "accept": false, + "reason": "Both are memory-related, but one is training memory efficiency and the other is offload behavior under quantization." + }, + { + "left": "issue:45362", + "right": "issue:45406", + "accept": false, + "reason": "Different models and different crash sites; not the same bug." + }, + { + "left": "issue:43856", + "right": "issue:44387", + "accept": false, + "reason": "Both mention memory, but the reported conditions and failure modes differ materially." + }, + { + "left": "issue:44910", + "right": "issue:45588", + "accept": false, + "reason": "Both touch flash attention, but the failures are unrelated: packed-sequence misread vs missing `s_aux` handling." + }, + { + "left": "issue:42491", + "right": "issue:45292", + "accept": false, + "reason": "A Qwen3 MoE LoRA compatibility issue is unrelated to embedding resizing behavior." + }, + { + "left": "issue:45276", + "right": "issue:45335", + "accept": true, + "reason": "Both report `resize_token_embeddings` failing to update nested model embeddings; these look like the same underlying regression across closely related model variants." + }, + { + "left": "issue:44164", + "right": "issue:44291", + "accept": false, + "reason": "Separate serialization/loading bugs with different triggers and error types." + }, + { + "left": "issue:36010", + "right": "issue:39404", + "accept": false, + "reason": "Completely different import vs Whisper pipeline behavior issues." + }, + { + "left": "issue:43404", + "right": "issue:43454", + "accept": false, + "reason": "Both are untied `lm_head` reports, but they target different model families and there isn't enough evidence of one shared code-path bug." + }, + { + "left": "issue:41720", + "right": "issue:44155", + "accept": false, + "reason": "Different subsystems and symptoms: device-mapping crash vs batched inference token leak." + }, + { + "left": "issue:39692", + "right": "issue:44960", + "accept": false, + "reason": "Unrelated model-specific issue vs generic GLM5 loading report." + }, + { + "left": "issue:43425", + "right": "issue:43957", + "accept": false, + "reason": "Torch version compatibility and `meta`-device loading are different problems." + }, + { + "left": "issue:44402", + "right": "issue:44488", + "accept": false, + "reason": "Tokenizer vocab mismatch is not the same as a model loading failure." + }, + { + "left": "issue:42947", + "right": "issue:45704", + "accept": false, + "reason": "Gradient checkpointing/LoRA and an apex RMSNorm leak are unrelated." + }, + { + "left": "issue:45663", + "right": "issue:45759", + "accept": false, + "reason": "Different model architecture bugs: FSDP/shared KV state vs composite config unwrapping." + }, + { + "left": "issue:43873", + "right": "issue:44387", + "accept": false, + "reason": "Both concern memory, but one is offloading behavior and the other is int4 reserved-memory growth." + }, + { + "left": "issue:43335", + "right": "issue:45663", + "accept": false, + "reason": "Sparse-layer construction bug is unrelated to Gemma4 attention state rebuilding." + }, + { + "left": "issue:45663", + "right": "issue:45859", + "accept": false, + "reason": "Different Qwen3/Gemma4 issues with different execution paths." + }, + { + "left": "issue:44291", + "right": "issue:44756", + "accept": false, + "reason": "`init_empty_weights` argument handling and Strix Halo mmap OOM are unrelated." + }, + { + "left": "issue:44155", + "right": "issue:45823", + "accept": false, + "reason": "Different model families and different failure modes." + }, + { + "left": "issue:44292", + "right": "issue:44912", + "accept": false, + "reason": "Different quantized-model loading failures with no clear shared root cause." + }, + { + "left": "issue:44387", + "right": "issue:44756", + "accept": false, + "reason": "Both are OOM-related, but one is quantization memory growth and the other is platform-specific mmap behavior." + }, + { + "left": "issue:43854", + "right": "issue:45216", + "accept": false, + "reason": "Different regressions: GLM-4.7 unit-test loading vs Qwen3.5 checkpoint save correctness." + }, + { + "left": "issue:44743", + "right": "issue:44960", + "accept": false, + "reason": "Recurrent-state cache behavior and a GLM5 loading issue are unrelated." + }, + { + "left": "issue:43957", + "right": "issue:44756", + "accept": false, + "reason": "`meta`-device model loading and Strix Halo OOM mitigation are different issues." + }, + { + "left": "issue:45276", + "right": "issue:45292", + "accept": true, + "reason": "Same `resize_token_embeddings` regression; one is a model-specific report and the other is the generic version of the same bug." + }, + { + "left": "issue:45684", + "right": "issue:45698", + "accept": false, + "reason": "One is file permission propagation in `save_pretrained`, the other is wrong custom-module resolution on reload." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The representative issue choice for the resize_token_embeddings cluster is supported by the titles, and the accepted 45276/45335 pair looks like the same bugfix area and mergeable into one PR. The remaining soft-edge rejections are phrased cautiously and do not overstate shared root cause beyond what the packet shows." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 7, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:36032", + "issue:39404", + "issue:42947", + "issue:43257", + "issue:43299", + "issue:43381", + "issue:43425", + "issue:43452", + "issue:43596", + "issue:43611", + "issue:43645", + "issue:43646", + "issue:43708", + "issue:43716", + "issue:43828", + "issue:43901", + "issue:43906", + "issue:43931", + "issue:43957", + "issue:43994", + "issue:44079", + "issue:44164", + "issue:44265", + "issue:44402", + "issue:44451", + "issue:44479", + "issue:44509", + "issue:44530", + "issue:44589", + "issue:44671", + "issue:44683", + "issue:44756", + "issue:44805", + "issue:44863", + "issue:44877", + "issue:44898", + "issue:44912", + "issue:44918", + "issue:44928", + "issue:44991", + "issue:45084", + "issue:45292", + "issue:45335", + "issue:45588", + "issue:45663", + "issue:45701", + "issue:45759", + "issue:45823" + ], + "soft_pairs": [ + "issue:43381|issue:44928", + "issue:44164|issue:44756", + "issue:42947|issue:43381", + "issue:44479|issue:44877", + "issue:45588|issue:45663", + "issue:43646|issue:43708", + "issue:45292|issue:45335", + "issue:43645|issue:44164", + "issue:43257|issue:44863", + "issue:44402|issue:44451", + "issue:43828|issue:43957", + "issue:43425|issue:43828", + "issue:44589|issue:44756", + "issue:44683|issue:45823", + "issue:39404|issue:45701", + "issue:43906|issue:44530", + "issue:43425|issue:45823", + "issue:43611|issue:43645", + "issue:43425|issue:43716", + "issue:44530|issue:45084", + "issue:39404|issue:44912", + "issue:43901|issue:44509", + "issue:44265|issue:44898", + "issue:44479|issue:44805", + "issue:43994|issue:44079", + "issue:43299|issue:43931", + "issue:39404|issue:44671", + "issue:43425|issue:44918", + "issue:44589|issue:45084", + "issue:44164|issue:45759", + "issue:36032|issue:43452", + "issue:43596|issue:44683", + "issue:44671|issue:44991" + ], + "prepared_review_unit_hash": "02858d2dff42c29ba3926b4ac434b4cefbf62e53887ff642c9a7e90832b2d150", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37674, + "estimated_input_tokens": 9419, + "estimated_eval_tokens": 19094 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a loose collection of unrelated Transformers regressions, largely around model loading and v5 compatibility. Only the token-embedding resize reports (45292/45335) look like the same underlying bug; the rest are better kept separate.", + "confidence": 0.88, + "canonical_issue_reason": "Issue 43611 is the best anchor: it is a broad, concrete Transformers 5.0 model-loading regression and is representative of the cluster\u2019s main theme without being overly model-specific.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43611 is the strongest standalone issue to keep: clear symptom, broad impact, and a central regression category that matches the surrounding cluster better than the narrow model-specific reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43381", + "right": "issue:44928", + "accept": false, + "reason": "Both mention gradient/checkpointing, but the failures are different: eval-mode checkpointing vs RLHF NaN/SDPA fallback on Qwen3.5." + }, + { + "left": "issue:44164", + "right": "issue:44756", + "accept": false, + "reason": "Unrelated bugs: serialization of extra_state vs disabling mmap on Strix Halo to avoid OOM." + }, + { + "left": "issue:42947", + "right": "issue:43381", + "accept": false, + "reason": "Both involve gradient checkpointing, but one is about LoRA effectiveness and the other about eval-mode incompatibility." + }, + { + "left": "issue:44479", + "right": "issue:44877", + "accept": false, + "reason": "Different subsystems and failures: Qwen VL video-input regression vs strict config blocking granite_speech loading." + }, + { + "left": "issue:45588", + "right": "issue:45663", + "accept": false, + "reason": "Different code paths: flash_attention sink-less crash vs Gemma4 FSDP2 shared_kv_states rebuild bug." + }, + { + "left": "issue:43646", + "right": "issue:43708", + "accept": false, + "reason": "Custom model init in notebooks is unrelated to resume_from_checkpoint max_steps calculation." + }, + { + "left": "issue:45292", + "right": "issue:45335", + "accept": true, + "reason": "Same underlying resize_token_embeddings bug; one report names output_embeddings and the other decoder.embed_tokens, both saying resizing does not propagate." + }, + { + "left": "issue:43645", + "right": "issue:44164", + "accept": false, + "reason": "Notebook custom-model initialization and extra_state serialization are different issues." + }, + { + "left": "issue:43257", + "right": "issue:44863", + "accept": false, + "reason": "Both are model-loading problems, but they affect different architectures and failure modes; not the same bug." + }, + { + "left": "issue:44402", + "right": "issue:44451", + "accept": false, + "reason": "Tokenizer vocab-size mismatch is not enough to equate it with a separate model-loading failure for ScandiBERT." + }, + { + "left": "issue:43828", + "right": "issue:43957", + "accept": false, + "reason": "Different failures: autocast dtype mismatch vs meta-device loading breakage." + }, + { + "left": "issue:43425", + "right": "issue:43828", + "accept": false, + "reason": "Torch version incompatibility and a model-specific autocast dtype mismatch are not the same issue." + }, + { + "left": "issue:44589", + "right": "issue:44756", + "accept": false, + "reason": "Float8 storage lookup failure is unrelated to the Strix Halo mmap OOM report." + }, + { + "left": "issue:44683", + "right": "issue:45823", + "accept": false, + "reason": "Compiled flex_attention on torch>=2.9 and Gemma4 device-mapping mismatch are separate bugs." + }, + { + "left": "issue:39404", + "right": "issue:45701", + "accept": false, + "reason": "Whisper pipeline return_language regression is unrelated to a generic tokenization-change report." + }, + { + "left": "issue:43906", + "right": "issue:44530", + "accept": false, + "reason": "The isolated reproduction issue does not describe the same failure as the Qwen3.5 PagedAttentionCache crash." + }, + { + "left": "issue:43425", + "right": "issue:45823", + "accept": false, + "reason": "Torch 2.10 incompatibility and Gemma4 PLE device mismatch are different problems." + }, + { + "left": "issue:43611", + "right": "issue:43645", + "accept": false, + "reason": "Base_model_prefix loading regression is not the same as custom model initialization in notebooks." + }, + { + "left": "issue:43425", + "right": "issue:43716", + "accept": false, + "reason": "A Torch version incompatibility is unrelated to a Mistral-3 image preprocessor/model dtype mismatch." + }, + { + "left": "issue:44530", + "right": "issue:45084", + "accept": false, + "reason": "Linear-attention group crash and non-template-node compilation error are different code paths." + }, + { + "left": "issue:39404", + "right": "issue:44912", + "accept": false, + "reason": "Whisper pipeline language return and MXFP4 quantization fallback are unrelated." + }, + { + "left": "issue:43901", + "right": "issue:44509", + "accept": false, + "reason": "Both are docs-related, but they cover different removed/changed pipeline docs and should not be merged as duplicates." + }, + { + "left": "issue:44265", + "right": "issue:44898", + "accept": false, + "reason": "torch.export/torch_compilable_check failure and Perceiver non-default-resolution failure are separate issues." + }, + { + "left": "issue:44479", + "right": "issue:44805", + "accept": false, + "reason": "Qwen video-input regression and a mask-shape IndexError are not the same bug." + }, + { + "left": "issue:43994", + "right": "issue:44079", + "accept": false, + "reason": "Siglip2 incorrect outputs and ModelOutput key assignment are different behaviors." + }, + { + "left": "issue:43299", + "right": "issue:43931", + "accept": false, + "reason": "Both involve Qwen3-VL loading, but one is a version-specific MoE loading regression and the other is a concrete weight-shape mismatch; too different to merge." + }, + { + "left": "issue:39404", + "right": "issue:44671", + "accept": false, + "reason": "Whisper pipeline behavior and CamemBERT masked-LM predictions are unrelated." + }, + { + "left": "issue:43425", + "right": "issue:44918", + "accept": false, + "reason": "Torch incompatibility is unrelated to Qwen3.5 embedding unpacking in TRL SFT." + }, + { + "left": "issue:44589", + "right": "issue:45084", + "accept": false, + "reason": "Float8 storage lookup and non-template-node compilation are unrelated." + }, + { + "left": "issue:44164", + "right": "issue:45759", + "accept": false, + "reason": "extra_state serialization failure is unrelated to unwrapping text_config in composite Qwen multimodal configs." + }, + { + "left": "issue:36032", + "right": "issue:43452", + "accept": false, + "reason": "Tokenizer special-token method conflict and gguf_file loading breakage are different bugs." + }, + { + "left": "issue:43596", + "right": "issue:44683", + "accept": false, + "reason": "Zero3 init/BertModel index error is unrelated to compiled flex_attention failing on torch>=2.9." + }, + { + "left": "issue:44671", + "right": "issue:44991", + "accept": false, + "reason": "CamemBERT prediction regression and tokenizer loading failure for est-roberta are not the same issue." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet and stays conservative: it treats most pairs as separate regressions, and the one accepted pair (45292/45335) is a plausible same-bug resize_token_embeddings issue. The chosen anchor issue and cluster characterization are subjective but supported by the issue titles." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 8, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:32090", + "issue:34567", + "issue:36032", + "issue:39404", + "issue:40444", + "issue:40990", + "issue:42915", + "issue:43089", + "issue:43317", + "issue:43404", + "issue:43421", + "issue:43452", + "issue:43482", + "issue:43541", + "issue:43582", + "issue:43611", + "issue:43632", + "issue:43643", + "issue:43644", + "issue:43645", + "issue:43646", + "issue:43716", + "issue:43856", + "issue:43883", + "issue:44164", + "issue:44206", + "issue:44220", + "issue:44451", + "issue:44464", + "issue:44466", + "issue:44479", + "issue:44488", + "issue:44530", + "issue:44661", + "issue:44683", + "issue:44737", + "issue:44756", + "issue:44811", + "issue:44849", + "issue:44987", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45325", + "issue:45584", + "issue:45859", + "issue:45987" + ], + "soft_pairs": [ + "issue:32090|issue:39404", + "issue:43856|issue:44756", + "issue:44991|issue:45081", + "issue:44737|issue:45859", + "issue:44164|issue:44479", + "issue:43582|issue:43632", + "issue:44488|issue:45081", + "issue:44661|issue:44683", + "issue:40990|issue:43421", + "issue:44488|issue:44987", + "issue:43611|issue:43644", + "issue:44206|issue:44220", + "issue:43716|issue:45084", + "issue:43643|issue:44530", + "issue:36032|issue:42915", + "issue:34567|issue:40444", + "issue:43883|issue:44488", + "issue:44811|issue:45584", + "issue:44464|issue:45987", + "issue:44451|issue:44987", + "issue:43644|issue:43646", + "issue:43452|issue:43482", + "issue:42915|issue:44530", + "issue:43541|issue:45084", + "issue:43317|issue:43482", + "issue:44849|issue:45325", + "issue:36032|issue:43645", + "issue:44987|issue:44991", + "issue:43404|issue:44466", + "issue:43089|issue:45987" + ], + "prepared_review_unit_hash": "369bd14dd10ded0fc164eb90b068d049ea45d385c908e597db72bdaa60ac13f0", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36126, + "estimated_input_tokens": 9032, + "estimated_eval_tokens": 18320 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The candidate set is highly heterogeneous; most pairs are only loosely similar by subsystem or regression theme, not the same underlying bug/change. No PRs are present. I reject all soft-edge pairs.", + "confidence": 0.9, + "canonical_issue_reason": "No single canonical issue: the items span unrelated bugs and regressions, with only broad thematic overlap in a few loading/tokenizer/model cases.", + "canonical_pr_reason": null, + "best_issue_reason": "No issue is a strong global canonical for this cluster; if forced, issue:43452 is the broadest loading-regression report, but it still does not represent the whole set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:32090", + "right": "issue:39404", + "accept": false, + "reason": "Different components and failures: Trainer GPU broadcast TypeError vs Whisper pipeline return_language regression." + }, + { + "left": "issue:43856", + "right": "issue:44756", + "accept": false, + "reason": "Both mention memory/OOM, but one is Qwen3 MoE training memory use and the other is mmap on Strix Halo; not the same bug." + }, + { + "left": "issue:44991", + "right": "issue:45081", + "accept": false, + "reason": "Both are tokenizer-loading regressions, but for different models and different code paths (EST-RoBERTa vs Mistral regex patching)." + }, + { + "left": "issue:44737", + "right": "issue:45859", + "accept": false, + "reason": "Both involve device placement, but XLNet positional encoding CPU allocation and Qwen3.5 MoE CPU-offload incompatibility are distinct bugs." + }, + { + "left": "issue:44164", + "right": "issue:44479", + "accept": false, + "reason": "Save/from_pretrained extra_state handling and a Qwen VL video-input regression are unrelated." + }, + { + "left": "issue:43582", + "right": "issue:43632", + "accept": false, + "reason": "Apple Silicon allocator warmup TypeError is unrelated to the _is_hf_initialized v5 regression." + }, + { + "left": "issue:44488", + "right": "issue:45081", + "accept": false, + "reason": "Different failures: loading cjvt/sleng-bert vs a Mistral tokenizer regex crash." + }, + { + "left": "issue:44661", + "right": "issue:44683", + "accept": false, + "reason": "Tokenizer mapping/model registration bug vs compiled flex_attention on torch>=2.9; unrelated." + }, + { + "left": "issue:40990", + "right": "issue:43421", + "accept": false, + "reason": "Perplexity issue on gpt-oss-20b is unrelated to runtime special-token post-processor updates." + }, + { + "left": "issue:44488", + "right": "issue:44987", + "accept": false, + "reason": "Both are model-loading failures, but for different repos and different reported breakages." + }, + { + "left": "issue:43611", + "right": "issue:43644", + "accept": false, + "reason": "Both are v5 regressions, but base_model_prefix loading and non-persistent buffer junk are separate issues." + }, + { + "left": "issue:44206", + "right": "issue:44220", + "accept": false, + "reason": "Both are audio-feature related, but the unsupported center arg and _torch_extract_fbank_features crash are not clearly the same bug." + }, + { + "left": "issue:43716", + "right": "issue:45084", + "accept": false, + "reason": "Mistral-3 dtype mismatch in image preprocessing is unrelated to a template-node compilation TypeError." + }, + { + "left": "issue:43643", + "right": "issue:44530", + "accept": false, + "reason": "AutoConfig trust_remote_code missing fields is unrelated to PagedAttentionCache linear_attention crash." + }, + { + "left": "issue:36032", + "right": "issue:42915", + "accept": false, + "reason": "Tokenizer add_special_tokens conflict and Qwen3 MoE FP8 config failure are unrelated." + }, + { + "left": "issue:34567", + "right": "issue:40444", + "accept": false, + "reason": "TrainerState token counting bug is unrelated to iterable multi-image Qwen2.5-VL finetuning failure." + }, + { + "left": "issue:43883", + "right": "issue:44488", + "accept": false, + "reason": "Molmo tied-weights attribute error is unrelated to loading cjvt/sleng-bert." + }, + { + "left": "issue:44811", + "right": "issue:45584", + "accept": false, + "reason": "Both are Whisper-related, but batch_decode skip_special_tokens handling and empty-transcription generation after align_special_tokens are different code paths." + }, + { + "left": "issue:44464", + "right": "issue:45987", + "accept": false, + "reason": "Chunked generation with compiled forward is not the same as StaticCache.get_seq_length returning the wrong type." + }, + { + "left": "issue:44451", + "right": "issue:44987", + "accept": false, + "reason": "Separate model-loading failures for different models; no shared underlying bug is evident." + }, + { + "left": "issue:43644", + "right": "issue:43646", + "accept": false, + "reason": "Non-persistent buffer initialization junk vs custom model initialization breakage are distinct regressions." + }, + { + "left": "issue:43452", + "right": "issue:43482", + "accept": false, + "reason": "Both involve GGUF loading, but one is a broad gguf_file/from_pretrained regression and the other is a specific model failure; not enough evidence they are the same bug." + }, + { + "left": "issue:42915", + "right": "issue:44530", + "accept": false, + "reason": "Qwen3 MoE FP8 config failure is unrelated to PagedAttentionCache linear_attention crash." + }, + { + "left": "issue:43541", + "right": "issue:45084", + "accept": false, + "reason": "Torch dynamo grouped_mm tracing failure is unrelated to template-node compilation errors." + }, + { + "left": "issue:43317", + "right": "issue:43482", + "accept": false, + "reason": "Device-map dequantized-model offload failure is unrelated to Qwen2.5-GGUF loading." + }, + { + "left": "issue:44849", + "right": "issue:45325", + "accept": false, + "reason": "Qwen3.5 output_hidden_states bug and Qwen2.5-VL rope-index scaling bug are different issues." + }, + { + "left": "issue:36032", + "right": "issue:43645", + "accept": false, + "reason": "Tokenizer special-token conflict is unrelated to Jupyter/custom-model initialization breakage." + }, + { + "left": "issue:44987", + "right": "issue:44991", + "accept": false, + "reason": "Both are loading regressions, but they affect different artifacts and likely different failure causes." + }, + { + "left": "issue:43404", + "right": "issue:44466", + "accept": false, + "reason": "lm_head tying in Mistral3 and inconsistent lm_head serialization by device are related in theme but not the same concrete bug." + }, + { + "left": "issue:43089", + "right": "issue:45987", + "accept": false, + "reason": "Generation GPU-sync overhead is unrelated to StaticCache.get_seq_length returning the wrong type." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 9, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:36032", + "issue:40444", + "issue:40990", + "issue:41720", + "issue:42175", + "issue:42915", + "issue:43262", + "issue:43296", + "issue:43322", + "issue:43377", + "issue:43404", + "issue:43425", + "issue:43531", + "issue:43540", + "issue:43611", + "issue:43632", + "issue:43645", + "issue:43646", + "issue:43742", + "issue:43819", + "issue:43883", + "issue:43994", + "issue:44062", + "issue:44079", + "issue:44220", + "issue:44265", + "issue:44291", + "issue:44403", + "issue:44451", + "issue:44479", + "issue:44534", + "issue:44589", + "issue:44610", + "issue:44683", + "issue:44756", + "issue:44811", + "issue:44877", + "issue:44987", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45127", + "issue:45216", + "issue:45325", + "issue:45357", + "issue:45507", + "issue:45542", + "issue:45663" + ], + "soft_pairs": [ + "issue:43262|issue:43377", + "issue:45084|issue:45507", + "issue:43883|issue:44991", + "issue:43994|issue:44610", + "issue:43646|issue:44534", + "issue:43322|issue:43540", + "issue:43645|issue:44534", + "issue:44403|issue:45081", + "issue:43377|issue:43819", + "issue:44265|issue:44610", + "issue:42915|issue:45084", + "issue:45216|issue:45357", + "issue:44451|issue:45081", + "issue:44987|issue:45081", + "issue:44220|issue:44683", + "issue:42915|issue:44291", + "issue:43404|issue:45127", + "issue:43632|issue:44756", + "issue:44062|issue:44987", + "issue:44479|issue:45325", + "issue:36032|issue:43742", + "issue:42915|issue:44589", + "issue:40444|issue:45325", + "issue:43296|issue:43531", + "issue:43611|issue:44534", + "issue:36032|issue:41720", + "issue:43531|issue:44877", + "issue:43425|issue:44403", + "issue:44079|issue:45663", + "issue:40990|issue:44811", + "issue:43632|issue:44534", + "issue:42175|issue:45542" + ], + "prepared_review_unit_hash": "8670bd5d7d4997cd76018ed0431ecf55d8eb95506d557c048c5ef63ca36bae9e", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37278, + "estimated_input_tokens": 9320, + "estimated_eval_tokens": 18896 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous, with one loose subtheme around Transformers v5 loading/initialization regressions, but the soft-edge candidates are not true duplicates and span different code paths, models, or modalities.", + "confidence": 0.74, + "canonical_issue_reason": "Issue 43632 is the broadest and most central Transformers v5 regression in this set, touching model initialization/loading behavior that overlaps conceptually with several other issues here.", + "canonical_pr_reason": null, + "best_issue_reason": "43632 is the best representative of the cluster\u2019s main theme: a core Transformers v5 initialization regression with broad downstream impact, unlike the more model-specific or modality-specific reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43262", + "right": "issue:43377", + "accept": false, + "reason": "Audio processor chat-template sampling-rate default vs MIMI encoder batch-padding-mask mismatch; different components and bugs." + }, + { + "left": "issue:45084", + "right": "issue:45507", + "accept": false, + "reason": "Template compilation error vs GraniteMoEHybrid calling an invalid method; unrelated failure modes." + }, + { + "left": "issue:43883", + "right": "issue:44991", + "accept": false, + "reason": "Missing `all_tied_weights_keys` on MolmoForCausalLM vs tokenizer loading failure for EMBEDDIA/est-roberta; different subsystems." + }, + { + "left": "issue:43994", + "right": "issue:44610", + "accept": false, + "reason": "SigLIP2 bad outputs via AutoModel/pipeline vs OmDet-Turbo processor/model input-size mismatch; not the same bug." + }, + { + "left": "issue:43646", + "right": "issue:44534", + "accept": false, + "reason": "Custom model initialization breakage vs non-persistent buffers being filled with junk; both v5 regressions, but different code paths." + }, + { + "left": "issue:43322", + "right": "issue:43540", + "accept": false, + "reason": "Llava Next segfault vs Qwen3OmniMoe video ValueError; both multimodal, but distinct crashes and models." + }, + { + "left": "issue:43645", + "right": "issue:44534", + "accept": false, + "reason": "Jupyter/custom model init regression vs buffer initialization regression; not the same underlying issue." + }, + { + "left": "issue:44403", + "right": "issue:45081", + "accept": false, + "reason": "Generic loading noise vs Mistral regex patch crash on tokenizer backend access; different behaviors and fixes." + }, + { + "left": "issue:43377", + "right": "issue:43819", + "accept": false, + "reason": "MIMI encoder padding-mask discrepancy vs DAC STE mismatch in `from_latents`; different model internals and symptoms." + }, + { + "left": "issue:44265", + "right": "issue:44610", + "accept": false, + "reason": "`torch.export.export` failure with `torch_compilable_check` vs processor output-size mismatch; unrelated export vs preprocessing bugs." + }, + { + "left": "issue:42915", + "right": "issue:45084", + "accept": false, + "reason": "Qwen3Moe FP8 config failure vs non-template compilation error; different model/config and failure mechanism." + }, + { + "left": "issue:45216", + "right": "issue:45357", + "accept": false, + "reason": "Both are Qwen3.5 save_pretrained regressions, but one concerns checkpoint correctness and the other visual encoder key naming; related area but not the same concrete bug." + }, + { + "left": "issue:44451", + "right": "issue:45081", + "accept": false, + "reason": "Model load failure for vesteinn/ScandiBERT vs tokenizer backend attribute crash in Mistral regex patch; different root causes." + }, + { + "left": "issue:44987", + "right": "issue:45081", + "accept": false, + "reason": "Model loading failure for physical-intelligence/fast vs Mistral tokenizer regex patch crash; different models and code paths." + }, + { + "left": "issue:44220", + "right": "issue:44683", + "accept": false, + "reason": "`_torch_extract_fbank_features()` issue vs compiled flex_attention failure on newer torch; unrelated audio feature extraction vs attention compilation." + }, + { + "left": "issue:42915", + "right": "issue:44291", + "accept": false, + "reason": "Qwen3Moe FP8 config issue vs `init_empty_weights`/_is_hf_initialized argument handling; distinct initialization/configuration problems." + }, + { + "left": "issue:43404", + "right": "issue:45127", + "accept": false, + "reason": "Mistral3 untied lm_head weights vs LoRA merge collapse with extended vocab; both involve weights, but not the same bug." + }, + { + "left": "issue:43632", + "right": "issue:44756", + "accept": false, + "reason": "`_is_hf_initialized` regression vs mmap/OOM workaround on Strix Halo; unrelated loading/runtime concerns." + }, + { + "left": "issue:44062", + "right": "issue:44987", + "accept": false, + "reason": "Duplicate `special` keyword in AddedToken construction vs tokenizer/model loading failure for a specific checkpoint; different tokenizer-layer bugs." + }, + { + "left": "issue:44479", + "right": "issue:45325", + "accept": false, + "reason": "Video input regression across Qwen variants vs Qwen2.5-VL rope index scaling bug; both visual models, but different preprocessing/math errors." + }, + { + "left": "issue:36032", + "right": "issue:43742", + "accept": false, + "reason": "T5 tokenizer special-token method conflict vs MobileLLM key error on load; unrelated tokenizer vs model-loading issues." + }, + { + "left": "issue:42915", + "right": "issue:44589", + "accept": false, + "reason": "Qwen3Moe FP8 config failure vs missing Float8 storage type on load; both FP8-adjacent, but different error surfaces and fixes." + }, + { + "left": "issue:40444", + "right": "issue:45325", + "accept": false, + "reason": "IterableDataset multi-image finetuning failure vs Qwen2.5-VL rope index scaling bug; different multimodal training vs inference issues." + }, + { + "left": "issue:43296", + "right": "issue:43531", + "accept": false, + "reason": "PaddleOCR-VL load failure in vLLM vs Qwen3-MoE sliding_window behavior; different models and code paths." + }, + { + "left": "issue:43611", + "right": "issue:44534", + "accept": false, + "reason": "`base_model_prefix` loading regression vs non-persistent buffer junk values; both v5-related, but not the same bug." + }, + { + "left": "issue:36032", + "right": "issue:41720", + "accept": false, + "reason": "T5 tokenizer load error vs Qwen3 auto device mapping cuda assert; different model families and failure layers." + }, + { + "left": "issue:43531", + "right": "issue:44877", + "accept": false, + "reason": "Qwen3-MoE sliding_window issue vs strict config blocking granite_speech load; unrelated model behavior vs config validation." + }, + { + "left": "issue:43425", + "right": "issue:44403", + "accept": false, + "reason": "Torch 2.10 incompatibility vs generic transformer-loading noise; one is dependency compatibility, the other a runtime loading symptom." + }, + { + "left": "issue:44079", + "right": "issue:45663", + "accept": false, + "reason": "ModelOutput key assignment bug vs Gemma4 FSDP2 shared_kv_states KeyError; different abstractions and failure sites." + }, + { + "left": "issue:40990", + "right": "issue:44811", + "accept": false, + "reason": "High perplexity on GPT-OSS/WikiText-2 vs Whisper batch_decode skip_special_tokens regression; unrelated evaluation vs decoding behavior." + }, + { + "left": "issue:43632", + "right": "issue:44534", + "accept": false, + "reason": "_is_hf_initialized flag regression vs non-persistent buffer initialization bug; both Transformers v5 issues, but distinct code paths and symptoms." + }, + { + "left": "issue:42175", + "right": "issue:45542", + "accept": false, + "reason": "Packaging/install backend inclusion issue vs tensorboard-only install causing undefined tf backend; different dependency resolution problems." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 10, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:33666", + "issue:35707", + "issue:40444", + "issue:41553", + "issue:41720", + "issue:42175", + "issue:43296", + "issue:43317", + "issue:43377", + "issue:43381", + "issue:43421", + "issue:43482", + "issue:43525", + "issue:43531", + "issue:43582", + "issue:43611", + "issue:43632", + "issue:43644", + "issue:43646", + "issue:43749", + "issue:43792", + "issue:43818", + "issue:43856", + "issue:43994", + "issue:44079", + "issue:44265", + "issue:44291", + "issue:44403", + "issue:44462", + "issue:44534", + "issue:44568", + "issue:44589", + "issue:44683", + "issue:44829", + "issue:44898", + "issue:44991", + "issue:45020", + "issue:45072", + "issue:45081", + "issue:45325", + "issue:45357", + "issue:45381", + "issue:45479", + "issue:45491", + "issue:45538", + "issue:45812", + "issue:45923" + ], + "soft_pairs": [ + "issue:41553|issue:43582", + "issue:43421|issue:44568", + "issue:43792|issue:44589", + "issue:43818|issue:45357", + "issue:44683|issue:45020", + "issue:33666|issue:40444", + "issue:44403|issue:44991", + "issue:44462|issue:45812", + "issue:43317|issue:43856", + "issue:43482|issue:45538", + "issue:44829|issue:45479", + "issue:43381|issue:45923", + "issue:35707|issue:40444", + "issue:44079|issue:44534", + "issue:44291|issue:44589", + "issue:43482|issue:43525", + "issue:43377|issue:45491", + "issue:43749|issue:44079", + "issue:40444|issue:45381", + "issue:41720|issue:45081", + "issue:43644|issue:44079", + "issue:44265|issue:45072", + "issue:43531|issue:44291", + "issue:45325|issue:45381", + "issue:43632|issue:45072", + "issue:43994|issue:44898", + "issue:43611|issue:43646", + "issue:42175|issue:43296" + ], + "prepared_review_unit_hash": "a20bb808612ae457568959432161281640b615db802f178ae684aaafb3d0eb56", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35640, + "estimated_input_tokens": 8910, + "estimated_eval_tokens": 18076 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is not a duplicate set; it\u2019s a loose collection of mostly unrelated transformer regression reports, many centered on Transformers v5 compatibility but covering different code paths and model families.", + "confidence": 0.83, + "canonical_issue_reason": "issue:43646 is the best representative anchor: it is a broad, concrete Transformers v5 compatibility regression about custom model initialization, which matches the dominant theme of the cluster better than the narrower tokenizer/model-specific reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43646 is the strongest standalone issue to keep: it\u2019s specific enough to be actionable yet broad enough to represent the main v5 regression pattern seen here.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41553", + "right": "issue:43582", + "accept": false, + "reason": "Different subsystems and failures: tokenizer error messaging vs Apple Silicon allocator warmup TypeError." + }, + { + "left": "issue:43421", + "right": "issue:44568", + "accept": false, + "reason": "Both touch tokenizers, but one is runtime special-token post-processing and the other is a specific BOS/EOS regression for one tokenizer; not the same bug." + }, + { + "left": "issue:43792", + "right": "issue:44589", + "accept": false, + "reason": "Completely different failure modes: Whisper model loading/runtime vs Float8 storage lookup." + }, + { + "left": "issue:43818", + "right": "issue:45357", + "accept": false, + "reason": "Video-LLaVA weight/temporal-attention issue is unrelated to Qwen3.5 save_pretrained key names." + }, + { + "left": "issue:44683", + "right": "issue:45020", + "accept": false, + "reason": "One is torch >=2.9 flex_attention compilation, the other is remote_code model loading in recent Transformers." + }, + { + "left": "issue:33666", + "right": "issue:40444", + "accept": false, + "reason": "Both are Qwen VL training issues, but one is multi-GPU training and the other is IterableDataset multi-image prompt failures; not the same bug." + }, + { + "left": "issue:44403", + "right": "issue:44991", + "accept": false, + "reason": "Generic loading noise vs tokenizer loading failure for a specific model; different symptoms and code paths." + }, + { + "left": "issue:44462", + "right": "issue:45812", + "accept": false, + "reason": "Tokenizer.json discovery bug and Granite token-id regression are both tokenizer-related but not the same concrete failure." + }, + { + "left": "issue:43317", + "right": "issue:43856", + "accept": false, + "reason": "device_map/offload loading failure is unrelated to Qwen3 MoE memory efficiency during training." + }, + { + "left": "issue:43482", + "right": "issue:45538", + "accept": false, + "reason": "GGUF loading under Transformers v5 and CLIPTokenizer max-length behavior are unrelated." + }, + { + "left": "issue:44829", + "right": "issue:45479", + "accept": false, + "reason": "Flash-attention-3 training degeneration is not the same as zero-loss classification with num_labels=1." + }, + { + "left": "issue:43381", + "right": "issue:45923", + "accept": false, + "reason": "Eval-mode gradient checkpointing misuse vs missing supports_gradient_checkpointing flag on a trust_remote_code model." + }, + { + "left": "issue:35707", + "right": "issue:40444", + "accept": false, + "reason": "Progressive generation with inputs_embeds/past_key_values is unrelated to VL finetuning with multiple images per prompt." + }, + { + "left": "issue:44079", + "right": "issue:44534", + "accept": false, + "reason": "ModelOutput key assignment bug and non-persistent buffer junk are different internals with different effects." + }, + { + "left": "issue:44291", + "right": "issue:44589", + "accept": false, + "reason": "Empty-weights initialization arg mismatch is unrelated to Float8 storage resolution." + }, + { + "left": "issue:43482", + "right": "issue:43525", + "accept": false, + "reason": "GGUF loading failure and missing pad_token_id on Llama4Config are different model/config regressions." + }, + { + "left": "issue:43377", + "right": "issue:45491", + "accept": false, + "reason": "MIMI encoder padding-mask inconsistency is unrelated to Gemma3 sliding-window NaN embeddings." + }, + { + "left": "issue:43749", + "right": "issue:44079", + "accept": false, + "reason": "FSDP CPU RAM efficient loading is unrelated to ModelOutput key bookkeeping." + }, + { + "left": "issue:40444", + "right": "issue:45381", + "accept": false, + "reason": "Multi-image prompt finetuning failure is not the same as Qwen2.5-VL video position-id regression." + }, + { + "left": "issue:41720", + "right": "issue:45081", + "accept": false, + "reason": "A800 cudaErrorAssert with auto device mapping is unrelated to Mistral tokenizer regex patch crashes." + }, + { + "left": "issue:43644", + "right": "issue:44079", + "accept": false, + "reason": "Non-persistent buffer corruption and ModelOutput key handling are separate regressions." + }, + { + "left": "issue:44265", + "right": "issue:45072", + "accept": false, + "reason": "torch.export/torch_compilable_check failure is unrelated to dtype mismatches in inference." + }, + { + "left": "issue:43531", + "right": "issue:44291", + "accept": false, + "reason": "Qwen3-MoE sliding-window behavior does not match the init_empty_weights/_is_hf_initialized issue." + }, + { + "left": "issue:45325", + "right": "issue:45381", + "accept": false, + "reason": "Both are Qwen2.5-VL positional-id bugs, but one is still-image temporal scaling and the other is video vision_position_ids; too different to merge." + }, + { + "left": "issue:43632", + "right": "issue:45072", + "accept": false, + "reason": "_is_hf_initialized breakage is a model initialization regression, not a dtype mismatch issue." + }, + { + "left": "issue:43994", + "right": "issue:44898", + "accept": false, + "reason": "SigLIP2 AutoModel/pipeline nonsense results and Perceiver non-default resolution failure are unrelated vision-model bugs." + }, + { + "left": "issue:43611", + "right": "issue:43646", + "accept": false, + "reason": "Both are Transformers v5 compatibility regressions, but one is base_model_prefix loading and the other is custom model initialization; different concrete bugs." + }, + { + "left": "issue:42175", + "right": "issue:43296", + "accept": false, + "reason": "Pip extras backend packaging is unrelated to loading PaddleOCR-VL in vLLM." + }, + { + "left": "issue:44534", + "right": "issue:43644", + "accept": false, + "reason": "Both mention v5-era regressions, but one concerns buffer contents and the other concerns non-persistent buffer junk specifically; not the same artifact." + }, + { + "left": "issue:44568", + "right": "issue:43421", + "accept": false, + "reason": "A specific add_special_tokens regression on one tokenizer is not the same as runtime post-processor update behavior." + }, + { + "left": "issue:45020", + "right": "issue:43632", + "accept": false, + "reason": "Remote_code model breakage and _is_hf_initialized handling are different compatibility regressions." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 11, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:34634", + "issue:35707", + "issue:39692", + "issue:40444", + "issue:40990", + "issue:41720", + "issue:42915", + "issue:42947", + "issue:43257", + "issue:43262", + "issue:43296", + "issue:43317", + "issue:43377", + "issue:43454", + "issue:43482", + "issue:43550", + "issue:43572", + "issue:43606", + "issue:43701", + "issue:43746", + "issue:43792", + "issue:43819", + "issue:43873", + "issue:43881", + "issue:43957", + "issue:44186", + "issue:44265", + "issue:44291", + "issue:44387", + "issue:44466", + "issue:44479", + "issue:44488", + "issue:44568", + "issue:44610", + "issue:44617", + "issue:44877", + "issue:44898", + "issue:45072", + "issue:45325", + "issue:45357", + "issue:45584", + "issue:45701", + "issue:45759", + "issue:45800", + "issue:45823", + "issue:45834", + "issue:45920", + "issue:45923" + ], + "soft_pairs": [ + "issue:43454|issue:44466", + "issue:45325|issue:45834", + "issue:43792|issue:43957", + "issue:42915|issue:43482", + "issue:43296|issue:45800", + "issue:44387|issue:44617", + "issue:43701|issue:45325", + "issue:44479|issue:45823", + "issue:43606|issue:44265", + "issue:43262|issue:43746", + "issue:43881|issue:44488", + "issue:44291|issue:44877", + "issue:40990|issue:45584", + "issue:44568|issue:45920", + "issue:34634|issue:35707", + "issue:43873|issue:45701", + "issue:44186|issue:44265", + "issue:43606|issue:43819", + "issue:35707|issue:41720", + "issue:41720|issue:45823", + "issue:43454|issue:43550", + "issue:43377|issue:44610", + "issue:39692|issue:40444", + "issue:43317|issue:43873", + "issue:43377|issue:44898", + "issue:43257|issue:45923", + "issue:43454|issue:45072", + "issue:43572|issue:45759", + "issue:40990|issue:42947", + "issue:44466|issue:45357" + ], + "prepared_review_unit_hash": "0effb5cb84a1b0f44a53ebd87a83ecd5bad739755840caa06027a72e3b2f7766", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36787, + "estimated_input_tokens": 9197, + "estimated_eval_tokens": 18650 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is a grab-bag of mostly unrelated bug reports; the soft pairs share broad themes like loading, offloading, tokenization, or multimodal regressions, but not the same concrete root cause. No soft pair looks safe to merge as a duplicate or same-change match.", + "confidence": 0.89, + "canonical_issue_reason": "If a single issue must represent the set, issue 45920 is the broadest systemic regression: it affects multiple model families, has clear impact, and is the most generally reusable reference point among these otherwise unrelated reports.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45920 is the strongest overall issue candidate because it describes a wide tokenizer regression across many models, is open, and has the best chance of being a useful umbrella reference compared with the narrower, model-specific reports here.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43454", + "right": "issue:44466", + "accept": false, + "reason": "Both mention lm_head/tied weights, but one is a generation bug in AyaVision while the other is a serialization inconsistency tied to device state; different concrete failure modes." + }, + { + "left": "issue:45325", + "right": "issue:45834", + "accept": false, + "reason": "Both are vision/multimodal indexing problems, but they affect different models and different code paths; no evidence they stem from the same bug." + }, + { + "left": "issue:43792", + "right": "issue:43957", + "accept": false, + "reason": "Whisper runtime failure and meta-device loading breakage are different model-loading problems with different triggers and symptoms." + }, + { + "left": "issue:42915", + "right": "issue:43482", + "accept": false, + "reason": "FineGrainedFP8Config failure and GGUF loading failure are unrelated quantization/loading issues." + }, + { + "left": "issue:43296", + "right": "issue:45800", + "accept": false, + "reason": "These are different compatibility problems: one is a vLLM/PaddleOCR-VL load failure, the other is a torch/transformers version incompatibility." + }, + { + "left": "issue:44387", + "right": "issue:44617", + "accept": false, + "reason": "Both mention OOM/memory, but one is a general int4 quantization reserved-memory regression and the other is a specific model running out of memory." + }, + { + "left": "issue:43701", + "right": "issue:45325", + "accept": false, + "reason": "Checkpoint resume key mismatch is unrelated to the Qwen2.5-VL RoPE position-id bug." + }, + { + "left": "issue:44479", + "right": "issue:45823", + "accept": false, + "reason": "Video-input regression for Qwen VL models and Gemma4 PLE device mismatch are different forward-pass bugs in different models." + }, + { + "left": "issue:43606", + "right": "issue:44265", + "accept": false, + "reason": "CPU offload device mismatch for Bark is unrelated to torch.export failing on torch_compilable_check." + }, + { + "left": "issue:43262", + "right": "issue:43746", + "accept": false, + "reason": "Audio chat-template sampling-rate defaulting and GraniteSpeech PEFT checkpoint loading are separate audio/model-loading issues." + }, + { + "left": "issue:43881", + "right": "issue:44488", + "accept": false, + "reason": "Different models, different load failures, and no shared concrete failure mechanism." + }, + { + "left": "issue:44291", + "right": "issue:44877", + "accept": false, + "reason": "Unexpected _is_hf_initialized argument with init_empty_weights is a constructor/load API issue, while strict config rejection is a config-schema problem." + }, + { + "left": "issue:40990", + "right": "issue:45584", + "accept": false, + "reason": "Perplexity on GPT-OSS and Whisper empty-transcription generation failure are unrelated inference bugs." + }, + { + "left": "issue:44568", + "right": "issue:45920", + "accept": false, + "reason": "One is a missing BOS/EOS addition for a specific tokenizer; the other is a broader wrong-token-IDs regression across multiple tokenizers." + }, + { + "left": "issue:34634", + "right": "issue:35707", + "accept": false, + "reason": "Bark voice preset handling and progressive generation with inputs_embeds/past_key_values are unrelated features." + }, + { + "left": "issue:43873", + "right": "issue:45701", + "accept": false, + "reason": "Offloading/quantization behavior and version-dependent tokenization are different subsystems and failure modes." + }, + { + "left": "issue:44186", + "right": "issue:44265", + "accept": false, + "reason": "Tokenizer crashes on NER/batched padding is unrelated to export failures involving torch_compilable_check." + }, + { + "left": "issue:43606", + "right": "issue:43819", + "accept": false, + "reason": "CPU offload device mismatch for Bark and DAC latent/forward mismatch are different models and different bugs." + }, + { + "left": "issue:35707", + "right": "issue:41720", + "accept": false, + "reason": "Inputs_embeds generation and Qwen3 auto device-map CUDA asserts are not the same code-path problem." + }, + { + "left": "issue:41720", + "right": "issue:45823", + "accept": false, + "reason": "Both involve device placement, but one is a CUDA assert with Qwen3 and the other is a Gemma4 PLE device mismatch; not the same bug." + }, + { + "left": "issue:43454", + "right": "issue:43550", + "accept": false, + "reason": "lm_head weight tying/serialization and torch.compile+SDPA failure are unrelated issues." + }, + { + "left": "issue:43377", + "right": "issue:44610", + "accept": false, + "reason": "Missing padding-mask support in MIMI encoder is unrelated to OmDet-Turbo producing the wrong input resolution." + }, + { + "left": "issue:39692", + "right": "issue:40444", + "accept": false, + "reason": "Doc-example model/processor mismatch and Qwen2.5-VL finetuning with multiple images are different problems." + }, + { + "left": "issue:43317", + "right": "issue:43873", + "accept": false, + "reason": "Both mention offload/quantization, but one is a dequantized model load failure on gpu+cpu offload and the other is a broader offloading regression; insufficient evidence of the same root cause." + }, + { + "left": "issue:43377", + "right": "issue:44898", + "accept": false, + "reason": "MIMI padding-mask batching bug and Perceiver non-default resolution failure are different model-specific shape/vision issues." + }, + { + "left": "issue:43257", + "right": "issue:45923", + "accept": false, + "reason": "Qwen3 MOE weight conversion under accelerate+deepspeed and missing supports_gradient_checkpointing on a trust_remote_code model are unrelated load/config issues." + }, + { + "left": "issue:43454", + "right": "issue:45072", + "accept": false, + "reason": "AyaVision tied-weight generation issue and dtype mismatches in SwitchTransformers/TimmWrapperModel are different bugs." + }, + { + "left": "issue:43572", + "right": "issue:45759", + "accept": false, + "reason": "Missing pad_token_idx in StableLmConfig and failure to unwrap text_config for composite Qwen multimodal configs are different configuration issues." + }, + { + "left": "issue:40990", + "right": "issue:42947", + "accept": false, + "reason": "High perplexity on a model evaluation task and gradient checkpointing ineffectiveness with LoRA are unrelated." + }, + { + "left": "issue:44466", + "right": "issue:45357", + "accept": false, + "reason": "Both concern serialization, but one is lm_head weight inconsistency by device and the other is incorrect visual encoder keys for Qwen3.5; not the same fix." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 12, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:36331", + "issue:40990", + "issue:41720", + "issue:42617", + "issue:42915", + "issue:43295", + "issue:43296", + "issue:43316", + "issue:43317", + "issue:43322", + "issue:43404", + "issue:43450", + "issue:43454", + "issue:43526", + "issue:43550", + "issue:43577", + "issue:43606", + "issue:43708", + "issue:43720", + "issue:43749", + "issue:43873", + "issue:43881", + "issue:43883", + "issue:44008", + "issue:44186", + "issue:44387", + "issue:44423", + "issue:44479", + "issue:44534", + "issue:44610", + "issue:44671", + "issue:44743", + "issue:44756", + "issue:44877", + "issue:44898", + "issue:44913", + "issue:44918", + "issue:44977", + "issue:44991", + "issue:45072", + "issue:45161", + "issue:45357", + "issue:45406", + "issue:45632", + "issue:45684", + "issue:45701", + "issue:45704", + "issue:45823" + ], + "soft_pairs": [ + "issue:43526|issue:43883", + "issue:43454|issue:43720", + "issue:43749|issue:44534", + "issue:41720|issue:43322", + "issue:43881|issue:44991", + "issue:43296|issue:44877", + "issue:41720|issue:44977", + "issue:44008|issue:45823", + "issue:42617|issue:42915", + "issue:40990|issue:45701", + "issue:43404|issue:45357", + "issue:44186|issue:45072", + "issue:45632|issue:45684", + "issue:43577|issue:43708", + "issue:44387|issue:45704", + "issue:44423|issue:45406", + "issue:43873|issue:44756", + "issue:36331|issue:42617", + "issue:44877|issue:44991", + "issue:43550|issue:43720", + "issue:36331|issue:42915", + "issue:43450|issue:45701", + "issue:44610|issue:44898", + "issue:44186|issue:44610", + "issue:44479|issue:44918", + "issue:41720|issue:42617", + "issue:43550|issue:43606", + "issue:43316|issue:44913", + "issue:42915|issue:43317", + "issue:43295|issue:45701", + "issue:41720|issue:45161", + "issue:43526|issue:43881", + "issue:44671|issue:44743" + ], + "prepared_review_unit_hash": "dc525243b0ba7a56c2544a691fc2f3adb2c68022740f65cd72cae25995e216a5", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37608, + "estimated_input_tokens": 9402, + "estimated_eval_tokens": 19060 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a heterogeneous cluster of Transformers bug reports, mostly around model loading, offloading, quantization, multimodal processors, and a few trainer/config regressions. None of the soft pairs look like the same underlying bug or change, so I would not merge any of them as duplicates.", + "confidence": 0.89, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "issue:43873 is the broadest and most central report in the cluster: it is open, has inbound references, and covers the shared offloading/quantization failure mode that several other issues only loosely resemble.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43526", + "right": "issue:43883", + "accept": false, + "reason": "Different bugs: one is a BeitImageProcessorFast label-reduction bug, the other is a Molmo tied-weights attribute error." + }, + { + "left": "issue:43454", + "right": "issue:43720", + "accept": false, + "reason": "Different model paths and failures: AyaVision lm_head tying vs BitNet packed-weight unpacking during accelerate loading." + }, + { + "left": "issue:43749", + "right": "issue:44534", + "accept": false, + "reason": "Both mention loading/runtime stability, but the concrete problems differ: FSDP CPU RAM-efficient loading vs non-persistent buffers being filled with junk." + }, + { + "left": "issue:41720", + "right": "issue:43322", + "accept": false, + "reason": "Unrelated symptoms and models: Qwen3 auto device mapping CUDA assert vs Llava Next segmentation fault." + }, + { + "left": "issue:43881", + "right": "issue:44991", + "accept": false, + "reason": "Different scope: a specific glm-4v-9b load failure vs a tokenizer-loading regression for EMBEDDIA/est-roberta." + }, + { + "left": "issue:43296", + "right": "issue:44877", + "accept": false, + "reason": "Different root causes: PaddleOCR-VL load failure in vLLM vs strict config blocking granite_speech config loading." + }, + { + "left": "issue:41720", + "right": "issue:44977", + "accept": false, + "reason": "Both involve Qwen3.x, but one is an auto device mapping CUDA assert and the other is a flash-attention generation bug." + }, + { + "left": "issue:44008", + "right": "issue:45823", + "accept": false, + "reason": "Both are Gemma-related multimodal issues, but one is a variable-name collision in Gemma3nModel.forward and the other is a PLE device mismatch in Gemma4." + }, + { + "left": "issue:42617", + "right": "issue:42915", + "accept": false, + "reason": "Different failure modes and models: 3d_parallel.py execution problem vs Qwen3Moe failing with FineGrainedFP8Config." + }, + { + "left": "issue:40990", + "right": "issue:45701", + "accept": false, + "reason": "The titles are too general to match: high perplexity on GPT-OSS/WikiText-2 vs a broad tokenization-version complaint." + }, + { + "left": "issue:43404", + "right": "issue:45357", + "accept": false, + "reason": "Both concern multimodal weight handling, but one is untied lm_head weights in Mistral3ForConditionalGeneration and the other is incorrect visual encoder keys on save_pretrained for Qwen3.5." + }, + { + "left": "issue:44186", + "right": "issue:45072", + "accept": false, + "reason": "Different classes of bugs: tokenizer crashes on NER/batched padding vs dtype mismatches in SwitchTransformers and TimmWrapperModel under bf16 inference." + }, + { + "left": "issue:45632", + "right": "issue:45684", + "accept": false, + "reason": "One is a cache-path collision for trust_remote_code; the other is save_pretrained copying read-only permissions." + }, + { + "left": "issue:43577", + "right": "issue:43708", + "accept": false, + "reason": "Different problems: Blip2 dtype not propagating to submodules vs resume_from_checkpoint miscomputing max_steps after batch-size changes." + }, + { + "left": "issue:44387", + "right": "issue:45704", + "accept": false, + "reason": "Both mention memory, but one is int4 quantization OOM in Transformers 5.x and the other is a T5 RMSNorm memory leak from apex." + }, + { + "left": "issue:44423", + "right": "issue:45406", + "accept": false, + "reason": "Both are serve-related multimodal crashes, but the concrete errors differ: AttributeError on str.to vs Gemma4Processor missing _tokenizer." + }, + { + "left": "issue:43873", + "right": "issue:44756", + "accept": false, + "reason": "Related offloading/memory theme, but not the same bug: general quantization offloading behavior vs a Strix Halo mmap OOM workaround." + }, + { + "left": "issue:36331", + "right": "issue:42617", + "accept": false, + "reason": "Completely different areas: Trainer.compute_loss signature regression vs a 3d_parallel.py runtime failure." + }, + { + "left": "issue:44877", + "right": "issue:44991", + "accept": false, + "reason": "Different kinds of loading regressions: strict config rejection vs tokenizer loading failure for a specific model." + }, + { + "left": "issue:43550", + "right": "issue:43720", + "accept": false, + "reason": "Different concrete failures: Bamba torch.compile with SDPA vs BitNet packed-weight unpacking during accelerate loading." + }, + { + "left": "issue:36331", + "right": "issue:42915", + "accept": false, + "reason": "Unrelated trainer API breakage vs Qwen3Moe FP8 loading failure." + }, + { + "left": "issue:43450", + "right": "issue:45701", + "accept": false, + "reason": "Video processor batched-shape bug does not match a broad tokenization version-change report." + }, + { + "left": "issue:44610", + "right": "issue:44898", + "accept": false, + "reason": "Both are image-size/preprocessing issues, but the models and root causes differ: OmDet-Turbo processor output size mismatch vs Perceiver interpolation failure." + }, + { + "left": "issue:44186", + "right": "issue:44610", + "accept": false, + "reason": "Different domains: LayoutLMv2 tokenizer crash on NER/batched padding vs OmDet-Turbo processor producing the wrong image resolution." + }, + { + "left": "issue:44479", + "right": "issue:44918", + "accept": false, + "reason": "Both involve Qwen3.5-family issues, but one is a video-input regression and the other is unpacking Qwen3.5 input embeddings with trl SFT trainer." + }, + { + "left": "issue:41720", + "right": "issue:42617", + "accept": false, + "reason": "Different models and failure surfaces: Qwen3 device-mapping CUDA assert vs inability to run 3d_parallel.py." + }, + { + "left": "issue:43550", + "right": "issue:43606", + "accept": false, + "reason": "Both are device/compile-related, but one is Bamba SDPA under torch.compile and the other is bark-small CPU offload device mismatch." + }, + { + "left": "issue:43316", + "right": "issue:44913", + "accept": false, + "reason": "Different API issues: Gemma3TextConfig discrepancy vs GPTNeoXConfig rotary_pct not persisting on reload." + }, + { + "left": "issue:42915", + "right": "issue:43317", + "accept": false, + "reason": "Different loading failures: Qwen3Moe FineGrainedFP8Config vs dequantized model failing with device_map=auto on GPU+CPU offload." + }, + { + "left": "issue:43295", + "right": "issue:45701", + "accept": false, + "reason": "A regression in processor.tokenizer/image handling is unrelated to a generic tokenization-version complaint." + }, + { + "left": "issue:41720", + "right": "issue:45161", + "accept": false, + "reason": "Both touch Qwen/GPT-OSS-style MoE topics, but one is a CUDA assert on auto device mapping and the other is a TP-only problem with GPT-OSS MoE." + }, + { + "left": "issue:43526", + "right": "issue:43881", + "accept": false, + "reason": "Unrelated bugs: BeitImageProcessorFast label reduction vs glm-4v-9b model loading failure." + }, + { + "left": "issue:44671", + "right": "issue:44743", + "accept": false, + "reason": "Both are Qwen-related language-model bugs, but one is CamemBERT masked-LM predictions and the other is recurrent-state reset behavior in modular_qwen3_5.py." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 13, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:30064", + "issue:33357", + "issue:34567", + "issue:34634", + "issue:35707", + "issue:36010", + "issue:41720", + "issue:42490", + "issue:42617", + "issue:42915", + "issue:43322", + "issue:43329", + "issue:43388", + "issue:43421", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43572", + "issue:43606", + "issue:43653", + "issue:43720", + "issue:43746", + "issue:43782", + "issue:43873", + "issue:43927", + "issue:44186", + "issue:44361", + "issue:44368", + "issue:44451", + "issue:44492", + "issue:44509", + "issue:44514", + "issue:44545", + "issue:44625", + "issue:44898", + "issue:44991", + "issue:45059", + "issue:45072", + "issue:45081", + "issue:45357", + "issue:45381", + "issue:45406", + "issue:45464", + "issue:45479", + "issue:45701", + "issue:45704", + "issue:45759", + "issue:45812" + ], + "soft_pairs": [ + "issue:45081|issue:45812", + "issue:43531|issue:43572", + "issue:43540|issue:43782", + "issue:36010|issue:42617", + "issue:43421|issue:45812", + "issue:44186|issue:44361", + "issue:42915|issue:43322", + "issue:43388|issue:45059", + "issue:45406|issue:45464", + "issue:33357|issue:35707", + "issue:43421|issue:45701", + "issue:43606|issue:43720", + "issue:41720|issue:43746", + "issue:30064|issue:34634", + "issue:43550|issue:43653", + "issue:44492|issue:44509", + "issue:43873|issue:45704", + "issue:44514|issue:44545", + "issue:43606|issue:43653", + "issue:42490|issue:43329", + "issue:43531|issue:45759", + "issue:34634|issue:41720", + "issue:34567|issue:35707", + "issue:43421|issue:43927", + "issue:44625|issue:45479", + "issue:44368|issue:45357", + "issue:44898|issue:45072", + "issue:43782|issue:45381", + "issue:44451|issue:44991" + ], + "prepared_review_unit_hash": "f06099e90e732cb1aedd024ea9552bf398c0abb5bd60ea234476addc5ae4c003", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36658, + "estimated_input_tokens": 9165, + "estimated_eval_tokens": 18586 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most pairs are unrelated. The only likely duplicate in this packet is issue:44514 / issue:44545: they have identical titles and very similar timing, so 44514 is the best canonical issue, but the packet only supports a conservative title-level duplicate call.", + "confidence": 0.92, + "canonical_issue_reason": "issue:44514 is the earlier of the two near-identical reports, and both items share the same crash title, making it the best canonical issue candidate on the evidence shown.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44514 is the strongest issue candidate because it is the earlier report in the likely duplicate pair and the titles match exactly; the packet does not provide stronger body-level proof.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45081", + "right": "issue:45812", + "accept": false, + "reason": "Different tokenizer bugs in different models: Mistral regex patch crash vs Granite token ID regression." + }, + { + "left": "issue:43531", + "right": "issue:43572", + "accept": false, + "reason": "Unrelated config issues: Qwen3-MoE sliding_window behavior vs StableLmConfig missing pad_token_idx." + }, + { + "left": "issue:43540", + "right": "issue:43782", + "accept": false, + "reason": "Different Qwen3 multimodal failures with different symptoms and code paths." + }, + { + "left": "issue:36010", + "right": "issue:42617", + "accept": false, + "reason": "ImportError for GenerationMixin is unrelated to a 3d_parallel.py runtime failure." + }, + { + "left": "issue:43421", + "right": "issue:45812", + "accept": false, + "reason": "Runtime tokenizer post-processor updates are not the same bug as Granite token ID regression." + }, + { + "left": "issue:44186", + "right": "issue:44361", + "accept": false, + "reason": "Different tokenizer crashes affecting different models and tasks." + }, + { + "left": "issue:42915", + "right": "issue:43322", + "accept": false, + "reason": "Qwen3Moe FP8 config failure and Llava Next segmentation fault are distinct issues." + }, + { + "left": "issue:43388", + "right": "issue:45059", + "accept": false, + "reason": "Different bugs: metric gathering label truncation vs SAM3 text/bounding-box behavior." + }, + { + "left": "issue:45406", + "right": "issue:45464", + "accept": false, + "reason": "Gemma4Processor missing _tokenizer in serve is unrelated to Qwen3.5 streaming inference failure." + }, + { + "left": "issue:33357", + "right": "issue:35707", + "accept": false, + "reason": "MacOS bus error with a CLIP model is unrelated to progressive generation with inputs_embeds and past_key_values." + }, + { + "left": "issue:43421", + "right": "issue:45701", + "accept": false, + "reason": "Broad tokenization-version behavior is not the same as a specific post-processor update bug." + }, + { + "left": "issue:43606", + "right": "issue:43720", + "accept": false, + "reason": "CPU offload device mismatch and BitNet packed-weight unpacking are different failure mechanisms." + }, + { + "left": "issue:41720", + "right": "issue:43746", + "accept": false, + "reason": "Qwen3 auto-device-map cuda assert is unrelated to GraniteSpeech PEFT checkpoint loading." + }, + { + "left": "issue:30064", + "right": "issue:34634", + "accept": false, + "reason": "Void segmentation map processing and Bark voice_preset handling are unrelated." + }, + { + "left": "issue:43550", + "right": "issue:43653", + "accept": false, + "reason": "torch.compile SDPA failure is unrelated to BigBirdTokenizer special-token registration." + }, + { + "left": "issue:44492", + "right": "issue:44509", + "accept": false, + "reason": "A cache-strategy typo is not the same as stale docs for removed pipeline tasks." + }, + { + "left": "issue:43873", + "right": "issue:45704", + "accept": false, + "reason": "Quantization/offloading behavior and an apex RMSNorm memory leak are unrelated." + }, + { + "left": "issue:44514", + "right": "issue:44545", + "accept": true, + "reason": "The titles are identical and the reports are adjacent, so they very likely describe the same crash; the packet only supports a conservative duplicate call." + }, + { + "left": "issue:43606", + "right": "issue:43653", + "accept": false, + "reason": "Different subsystems and failure modes: CPU offload device mismatch vs tokenizer special-token registration." + }, + { + "left": "issue:42490", + "right": "issue:43329", + "accept": false, + "reason": "Trailing-slash import warning is unrelated to an undefined helper in multimodal token counting." + }, + { + "left": "issue:43531", + "right": "issue:45759", + "accept": false, + "reason": "Sliding-window behavior in Qwen3-MoE is not the same as from_config failing to unwrap text_config." + }, + { + "left": "issue:34634", + "right": "issue:41720", + "accept": false, + "reason": "BarkProcessor voice_preset and Qwen3 auto device mapping errors are unrelated." + }, + { + "left": "issue:34567", + "right": "issue:35707", + "accept": false, + "reason": "TrainerState token counter issues are unrelated to generation with inputs_embeds and past_key_values." + }, + { + "left": "issue:43421", + "right": "issue:43927", + "accept": false, + "reason": "Tokenizer backend post-processor updates and DiaConfig save/load token IDs are different bugs." + }, + { + "left": "issue:44625", + "right": "issue:45479", + "accept": false, + "reason": "Config propagation of num_labels is not the same as degenerate zero loss with num_labels=1." + }, + { + "left": "issue:44368", + "right": "issue:45357", + "accept": false, + "reason": "A tie_word_embeddings warning is unrelated to incorrect visual encoder keys during save_pretrained." + }, + { + "left": "issue:44898", + "right": "issue:45072", + "accept": false, + "reason": "Perceiver interpolation at non-default resolution and bfloat16 dtype mismatches are distinct failures." + }, + { + "left": "issue:43782", + "right": "issue:45381", + "accept": false, + "reason": "Qwen3VL weight_only loading error and qwen2.5-vl video vision_position_ids bug are different code paths." + }, + { + "left": "issue:44451", + "right": "issue:44991", + "accept": false, + "reason": "Different tokenizer-loading regressions affecting different models." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The duplicate-pair conclusion for issue:44514 / issue:44545 is plausible from the titles, but the writeup overstates what is supported by the packet. Phrases like \"precise repro,\" \"narrow code path,\" and \"exact duplicate\" imply body-level evidence that is not shown here. Keep the conclusion more conservative: the titles are nearly identical and the pair looks very likely to be the same bug, but the packet alone does not prove the full repro details." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 14, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:30064", + "issue:36010", + "issue:39692", + "issue:40444", + "issue:41720", + "issue:42886", + "issue:42915", + "issue:43381", + "issue:43421", + "issue:43525", + "issue:43531", + "issue:43540", + "issue:43596", + "issue:43618", + "issue:43653", + "issue:43697", + "issue:43720", + "issue:43746", + "issue:43819", + "issue:43844", + "issue:43873", + "issue:43994", + "issue:44008", + "issue:44062", + "issue:44077", + "issue:44186", + "issue:44220", + "issue:44451", + "issue:44589", + "issue:44610", + "issue:44625", + "issue:44661", + "issue:44671", + "issue:44843", + "issue:44918", + "issue:44936", + "issue:44987", + "issue:45042", + "issue:45081", + "issue:45356", + "issue:45588", + "issue:45663", + "issue:45704", + "issue:45759", + "issue:45859", + "issue:45907", + "issue:45923" + ], + "soft_pairs": [ + "issue:44077|issue:44625", + "issue:44987|issue:45042", + "issue:45759|issue:45859", + "issue:43531|issue:44589", + "issue:39692|issue:44610", + "issue:44062|issue:44451", + "issue:45704|issue:45923", + "issue:43421|issue:43653", + "issue:45588|issue:45759", + "issue:45663|issue:45923", + "issue:44918|issue:44936", + "issue:42886|issue:44843", + "issue:43381|issue:43844", + "issue:43421|issue:44186", + "issue:43994|issue:44671", + "issue:45704|issue:45859", + "issue:43525|issue:43531", + "issue:45081|issue:45356", + "issue:41720|issue:45704", + "issue:43746|issue:43873", + "issue:39692|issue:43873", + "issue:40444|issue:43746", + "issue:45356|issue:45907", + "issue:43596|issue:44220", + "issue:42915|issue:43746", + "issue:43720|issue:43819", + "issue:43540|issue:44008", + "issue:44220|issue:44661", + "issue:30064|issue:36010", + "issue:43653|issue:43720", + "issue:43618|issue:43697" + ], + "prepared_review_unit_hash": "a40f43c9a5ba1917d9ca968e9bb5a156e672bfe1cbef09e806c3fa04ddfb9df9", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36639, + "estimated_input_tokens": 9160, + "estimated_eval_tokens": 18576 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a heterogeneous mix of unrelated Transformer bugs and regressions; the soft pairs mostly share only broad subsystem or keyword similarity, not the same concrete failing code path. I did not find any pair that should be merged as duplicates.", + "confidence": 0.92, + "canonical_issue_reason": "issue:44843 is the clearest, most actionable bug report here: it names a specific failing path (_patch_mistral_regex), an explicit offline-mode regression, and a reproducible symptom.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44843 is the strongest standalone issue in the set because it is narrowly scoped, reproducible, and points to a concrete code path with a likely fix surface.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44077", + "right": "issue:44625", + "accept": false, + "reason": "Different model families and failures: PatchTSMixer post-init handling vs Qwen3.5 config propagation." + }, + { + "left": "issue:44987", + "right": "issue:45042", + "accept": false, + "reason": "One is a model-loading regression; the other is an image-processor torchvision dependency issue." + }, + { + "left": "issue:45759", + "right": "issue:45859", + "accept": false, + "reason": "Both involve Qwen 3.x, but one is config unwrapping for from_config and the other is CPU-offload incompatibility in rotary embedding." + }, + { + "left": "issue:43531", + "right": "issue:44589", + "accept": false, + "reason": "Unrelated symptoms: sliding-window behavior in Qwen3-MoE vs a Float8 storage loading error." + }, + { + "left": "issue:39692", + "right": "issue:44610", + "accept": false, + "reason": "Both are multimodal/image-adjacent, but one is SigLIP2 docs/example errors and the other is an OmDet-Turbo processor/model size mismatch." + }, + { + "left": "issue:44062", + "right": "issue:44451", + "accept": false, + "reason": "Tokenizer construction error vs a specific model-load failure for ScandiBERT; no shared failing path is evident." + }, + { + "left": "issue:45704", + "right": "issue:45923", + "accept": false, + "reason": "T5 RMSNorm/apex leak is unrelated to a trust_remote_code gradient-checkpointing flag omission." + }, + { + "left": "issue:43421", + "right": "issue:43653", + "accept": false, + "reason": "Both touch tokenizer special tokens, but one is a backend post-processor update bug and the other is a BigBirdTokenizer registration/decode issue." + }, + { + "left": "issue:45588", + "right": "issue:45759", + "accept": false, + "reason": "Flash-attention s_aux handling is unrelated to Qwen config unwrapping." + }, + { + "left": "issue:45663", + "right": "issue:45923", + "accept": false, + "reason": "Gemma-4 FSDP2 shared_kv_states bug is a different training-path failure than the Nemotron gradient-checkpointing flag issue." + }, + { + "left": "issue:44918", + "right": "issue:44936", + "accept": false, + "reason": "Both involve trainer workflows, but one is SFT embedding unpacking and the other is evaluate-after-train state handling." + }, + { + "left": "issue:42886", + "right": "issue:44843", + "accept": false, + "reason": "Both mention offline/tokenizer loading, but one is cache loading in offline mode and the other is an unconditional model_info call in _patch_mistral_regex." + }, + { + "left": "issue:43381", + "right": "issue:43844", + "accept": false, + "reason": "Gradient checkpointing eval-mode restriction is unrelated to DeepSpeed ZeRO-3 gradient explosion." + }, + { + "left": "issue:43421", + "right": "issue:44186", + "accept": false, + "reason": "Shared tokenizer theme only; the concrete bugs are different (runtime special-token updates vs LayoutLMv2 NER/padding crashes)." + }, + { + "left": "issue:43994", + "right": "issue:44671", + "accept": false, + "reason": "Different models and outcomes: SigLIP2 nonsensical outputs vs CamemBERT masked-LM prediction regression." + }, + { + "left": "issue:45704", + "right": "issue:45859", + "accept": false, + "reason": "RMSNorm memory leak in T5 is unrelated to Qwen rotary embedding CPU-offload behavior." + }, + { + "left": "issue:43525", + "right": "issue:43531", + "accept": false, + "reason": "Both are Qwen-related, but one is a missing pad_token_id attribute and the other is a sliding_window issue." + }, + { + "left": "issue:45081", + "right": "issue:45356", + "accept": false, + "reason": "Both involve fix_mistral_regex, but one is an AttributeError crash while the other is a codec-handling regression with a warning issue." + }, + { + "left": "issue:41720", + "right": "issue:45704", + "accept": false, + "reason": "Qwen3 auto device mapping CUDA assert is unrelated to T5/apex RMSNorm memory leak." + }, + { + "left": "issue:43746", + "right": "issue:43873", + "accept": false, + "reason": "PEFT local-checkpoint loading failure is not the same bug as quantization offloading behavior." + }, + { + "left": "issue:39692", + "right": "issue:43873", + "accept": false, + "reason": "SigLIP2 doc/example errors and quantization offloading problems are different issue classes." + }, + { + "left": "issue:40444", + "right": "issue:43746", + "accept": false, + "reason": "Multi-image IterableDataset finetuning failure is unrelated to PEFT-adapter local checkpoint loading." + }, + { + "left": "issue:45356", + "right": "issue:45907", + "accept": false, + "reason": "Kimi tokenizer codec/fix_mistral_regex regression is unrelated to list_repo_files revision forwarding." + }, + { + "left": "issue:43596", + "right": "issue:44220", + "accept": false, + "reason": "DeepSpeed zero3/BertModel indexing error is unrelated to the audio fbank feature extraction issue." + }, + { + "left": "issue:42915", + "right": "issue:43746", + "accept": false, + "reason": "Qwen3Moe FP8 config failure and GraniteSpeech PEFT checkpoint loading are different model-path bugs." + }, + { + "left": "issue:43720", + "right": "issue:43819", + "accept": false, + "reason": "Packed-weight loading in BitNet is unrelated to DAC.from_latents / missing STE mismatch." + }, + { + "left": "issue:43540", + "right": "issue:44008", + "accept": false, + "reason": "Video input processing error in Qwen3OmniMoe is unrelated to Gemma 3n audio_mel_mask variable collision." + }, + { + "left": "issue:44220", + "right": "issue:44661", + "accept": false, + "reason": "Audio feature extraction bug and TOKENIZER_MAPPING_NAMES add-new-model-like failure are unrelated." + }, + { + "left": "issue:30064", + "right": "issue:36010", + "accept": false, + "reason": "Image segmentation map processing failure and a GenerationMixin import error do not share the same code path." + }, + { + "left": "issue:43653", + "right": "issue:43720", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and BitNet packed-weight loading are unrelated issues." + }, + { + "left": "issue:43618", + "right": "issue:43697", + "accept": false, + "reason": "CLIPOutput attentions assignment regression is unrelated to RTDetrV2 output differences in v5." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 15, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:30064", + "issue:36331", + "issue:39692", + "issue:41720", + "issue:43295", + "issue:43322", + "issue:43335", + "issue:43421", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43577", + "issue:43643", + "issue:43653", + "issue:43716", + "issue:43756", + "issue:43782", + "issue:43818", + "issue:43819", + "issue:43825", + "issue:43950", + "issue:43994", + "issue:44008", + "issue:44062", + "issue:44186", + "issue:44448", + "issue:44488", + "issue:44561", + "issue:44671", + "issue:44811", + "issue:44821", + "issue:44843", + "issue:44877", + "issue:44898", + "issue:44977", + "issue:44991", + "issue:45020", + "issue:45072", + "issue:45081", + "issue:45084", + "issue:45183", + "issue:45381", + "issue:45479", + "issue:45584", + "issue:45685", + "issue:45834", + "issue:45987" + ], + "soft_pairs": [ + "issue:45183|issue:45479", + "issue:43295|issue:45834", + "issue:43335|issue:43756", + "issue:43716|issue:44008", + "issue:43550|issue:45072", + "issue:43550|issue:43819", + "issue:43950|issue:45987", + "issue:44062|issue:44488", + "issue:43322|issue:45685", + "issue:44186|issue:44898", + "issue:44843|issue:45081", + "issue:43653|issue:43819", + "issue:43643|issue:44877", + "issue:43421|issue:44811", + "issue:43577|issue:43818", + "issue:43540|issue:45381", + "issue:43782|issue:44977", + "issue:43531|issue:43643", + "issue:44448|issue:44671", + "issue:45584|issue:45834", + "issue:39692|issue:45834", + "issue:43825|issue:43994", + "issue:44561|issue:45020", + "issue:30064|issue:36331", + "issue:41720|issue:44991", + "issue:44821|issue:45084" + ], + "prepared_review_unit_hash": "ff8cb69162c4eec612d7a438b59867e19c8dfc4e7bfdc671cdc1ebfeeba13f1c", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 26, + "serialized_chars": 35290, + "estimated_input_tokens": 8823, + "estimated_eval_tokens": 17902 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These issues are mostly heterogeneous model- or subsystem-specific regressions, not true duplicates. I would reject every soft pair; the closest umbrella issue in the set is the broad remote_code regression report, but even that is only a loose representative, not a duplicate of the others.", + "confidence": 0.42, + "canonical_issue_reason": "issue:45020 is the broadest umbrella in the set: it covers recent-version regressions affecting remote_code models, which is the closest common thread among a few tokenizer/config-loading reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45020 is the best representative issue because it is the most general and likely to gather related remote_code breakages, even though the rest of the cluster is too diverse to be a single duplicate set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45183", + "right": "issue:45479", + "accept": false, + "reason": "Different bugs: input validation error messaging vs a sequence-classification zero-loss regression." + }, + { + "left": "issue:43295", + "right": "issue:45834", + "accept": false, + "reason": "Processor/tokenizer regression is unrelated to Kosmos2.5 long-OCR index errors." + }, + { + "left": "issue:43335", + "right": "issue:43756", + "accept": false, + "reason": "SwitchTransformers sparse-layer construction and Smollm3 RoPE-layer dropping are unrelated model-specific issues." + }, + { + "left": "issue:43716", + "right": "issue:44008", + "accept": false, + "reason": "Mistral-3 dtype mismatch and Gemma3n variable-name collision are different failure modes in different code paths." + }, + { + "left": "issue:43550", + "right": "issue:45072", + "accept": false, + "reason": "Both are model/runtime regressions, but they affect different models and different mechanisms (SDPA compile vs dtype mismatch)." + }, + { + "left": "issue:43550", + "right": "issue:43819", + "accept": false, + "reason": "Bamba torch.compile/SDPA failure is unrelated to DAC.from_latents missing STE parity." + }, + { + "left": "issue:43950", + "right": "issue:45987", + "accept": false, + "reason": "from_pretrained buffer corruption and StaticCache.get_seq_length returning a tensor are separate API regressions." + }, + { + "left": "issue:44062", + "right": "issue:44488", + "accept": false, + "reason": "AddedToken special-argument duplication is unrelated to loading cjvt/sleng-bert." + }, + { + "left": "issue:43322", + "right": "issue:45685", + "accept": false, + "reason": "Llava Next loading segfault and MPS histogram-on-int error are unrelated." + }, + { + "left": "issue:44186", + "right": "issue:44898", + "accept": false, + "reason": "LayoutLMv2 tokenizer padding/truncation crashes and Perceiver image-size failures are different model-specific bugs." + }, + { + "left": "issue:44843", + "right": "issue:45081", + "accept": false, + "reason": "Both touch _patch_mistral_regex, but one is an offline hub call bug and the other is a backend_tokenizer attribute crash; not the same underlying defect." + }, + { + "left": "issue:43653", + "right": "issue:43819", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and DAC STE mismatch are unrelated." + }, + { + "left": "issue:43643", + "right": "issue:44877", + "accept": false, + "reason": "Missing fields from AutoConfig with trust_remote_code and granite_speech strict-config loading are not the same concrete bug." + }, + { + "left": "issue:43421", + "right": "issue:44811", + "accept": false, + "reason": "Runtime post-processor updates for special tokens and Whisper batch_decode skip_special_tokens behavior are distinct tokenizer issues." + }, + { + "left": "issue:43577", + "right": "issue:43818", + "accept": false, + "reason": "Blip2 dtype preservation and Video-LLaVA temporal-attention/shared-weight problems are unrelated." + }, + { + "left": "issue:43540", + "right": "issue:45381", + "accept": false, + "reason": "Both mention video inputs, but Qwen3OmniMoe ValueError and qwen2.5-vl vision_position_ids are different bugs." + }, + { + "left": "issue:43782", + "right": "issue:44977", + "accept": false, + "reason": "Qwen3VL weight_only loading error and Qwen3.5 flash-attention generation issue are separate failures." + }, + { + "left": "issue:43531", + "right": "issue:43643", + "accept": false, + "reason": "Sliding-window handling in Qwen3-MoE is unrelated to AutoConfig trust_remote_code field loss." + }, + { + "left": "issue:44448", + "right": "issue:44671", + "accept": false, + "reason": "Version-specific output changes for different models do not indicate the same underlying bug." + }, + { + "left": "issue:45584", + "right": "issue:45834", + "accept": false, + "reason": "Whisper empty-transcription generation and Kosmos2.5 OCR index errors are unrelated." + }, + { + "left": "issue:39692", + "right": "issue:45834", + "accept": false, + "reason": "Documentation/example errors for SigLIP2 do not match Kosmos2.5 long-input indexing failures." + }, + { + "left": "issue:43825", + "right": "issue:43994", + "accept": false, + "reason": "Pipeline error-message wording and SigLIP2 nonsensical outputs are unrelated." + }, + { + "left": "issue:44561", + "right": "issue:45020", + "accept": false, + "reason": "44561 is a specific trust_remote_code regression caused by a removed helper; 45020 is a broader umbrella report, but not the same concrete bug." + }, + { + "left": "issue:30064", + "right": "issue:36331", + "accept": false, + "reason": "ImageProcessor void-segmentation-map failure and CustomTrainer compute_loss signature breakage are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:44991", + "accept": false, + "reason": "Qwen3 auto device mapping cuda assert and tokenizer loading for EMBEDDIA/est-roberta are unrelated." + }, + { + "left": "issue:44821", + "right": "issue:45084", + "accept": false, + "reason": "AutoImageProcessor URL loading and non-template-node compilation errors are different issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 16, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:34634", + "issue:36010", + "issue:36331", + "issue:42175", + "issue:43262", + "issue:43329", + "issue:43526", + "issue:43577", + "issue:43582", + "issue:43606", + "issue:43644", + "issue:43650", + "issue:43653", + "issue:43697", + "issue:43720", + "issue:43742", + "issue:43749", + "issue:43782", + "issue:43819", + "issue:43825", + "issue:43827", + "issue:43906", + "issue:44060", + "issue:44186", + "issue:44291", + "issue:44368", + "issue:44442", + "issue:44479", + "issue:44485", + "issue:44509", + "issue:44554", + "issue:44610", + "issue:44671", + "issue:44805", + "issue:44857", + "issue:44912", + "issue:44936", + "issue:45020", + "issue:45072", + "issue:45081", + "issue:45325", + "issue:45406", + "issue:45459", + "issue:45600", + "issue:45701", + "issue:45803", + "issue:45834" + ], + "soft_pairs": [ + "issue:43644|issue:43749", + "issue:44186|issue:44442", + "issue:44485|issue:44509", + "issue:44479|issue:45834", + "issue:43606|issue:45072", + "issue:43577|issue:43582", + "issue:43825|issue:43827", + "issue:44805|issue:44936", + "issue:34634|issue:36010", + "issue:44671|issue:45701", + "issue:44610|issue:45072", + "issue:43782|issue:45325", + "issue:43720|issue:45072", + "issue:42175|issue:45081", + "issue:43749|issue:44291", + "issue:43650|issue:43906", + "issue:45081|issue:45406", + "issue:43582|issue:44857", + "issue:44060|issue:44368", + "issue:43262|issue:43819", + "issue:43526|issue:43697", + "issue:43526|issue:44554", + "issue:34634|issue:36331", + "issue:45459|issue:45803", + "issue:43329|issue:45600", + "issue:43742|issue:45020", + "issue:43653|issue:44912" + ], + "prepared_review_unit_hash": "0f311b30d34efcc0c6b393942136dda291b94bca6c6023720d4170d57cf6dae1", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35616, + "estimated_input_tokens": 8904, + "estimated_eval_tokens": 18064 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly a grab bag of unrelated Transformers issues. No soft edge looks like a true duplicate with the same concrete bug path; the closest pair is the v5 pipeline-removal messaging/docs pair, but I\u2019d still keep it split.", + "confidence": 0.84, + "canonical_issue_reason": "Issue 43825 is the most self-contained, actionable bug report in the set: a concrete runtime regression in the error message around removed pipeline tasks.", + "canonical_pr_reason": null, + "best_issue_reason": "43825 is the best representative issue because it is specific and reproducible; most other items are broader environment/model regressions or unrelated subsystems.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43644", + "right": "issue:43749", + "accept": false, + "reason": "Both concern loading/runtime behavior, but they are different failures: junk non-persistent buffers in v5 vs broken FSDP CPU RAM efficient loading." + }, + { + "left": "issue:44186", + "right": "issue:44442", + "accept": false, + "reason": "Different tokenizer paths and failures: LayoutLMv2 NER/padding crash vs AutoTokenizer loading FastSpeech2ConformerTokenizer." + }, + { + "left": "issue:44485", + "right": "issue:44509", + "accept": false, + "reason": "One is a GLM-5 RoPE implementation discussion, the other is docs still mentioning removed pipeline tasks; not the same bug/change." + }, + { + "left": "issue:44479", + "right": "issue:45834", + "accept": false, + "reason": "Different models and symptoms: video-input regression for Qwen VL models vs Kosmos2.5 OCR index error." + }, + { + "left": "issue:43606", + "right": "issue:45072", + "accept": false, + "reason": "Both are inference/load-time issues, but device mismatch on Bark CPU offload is unrelated to dtype mismatch in SwitchTransformers/TimmWrapperModel." + }, + { + "left": "issue:43577", + "right": "issue:43582", + "accept": false, + "reason": "Blip2 dtype propagation and Apple Silicon warmup TypeError are separate bugs in different code paths." + }, + { + "left": "issue:43825", + "right": "issue:43827", + "accept": false, + "reason": "They are related by the v5 pipeline removals, but one is an error-message bug and the other is a docs cleanup; not the same concrete artifact." + }, + { + "left": "issue:44805", + "right": "issue:44936", + "accept": false, + "reason": "Different failures: mask/index shape mismatch vs trainer.evaluate() breaking after training." + }, + { + "left": "issue:34634", + "right": "issue:36010", + "accept": false, + "reason": "BarkProcessor voice_preset failure is unrelated to the GenerationMixin import error." + }, + { + "left": "issue:44671", + "right": "issue:45701", + "accept": false, + "reason": "CamemBERT masked-LM regression and a general tokenization-change report are not the same underlying bug." + }, + { + "left": "issue:44610", + "right": "issue:45072", + "accept": false, + "reason": "OmDet-Turbo processor image-size mismatch is unrelated to SwitchTransformers/TimmWrapperModel dtype mismatches." + }, + { + "left": "issue:43782", + "right": "issue:45325", + "accept": false, + "reason": "Qwen3VL weight_only loading error and Qwen2.5-VL rope indexing bug affect different code paths." + }, + { + "left": "issue:43720", + "right": "issue:45072", + "accept": false, + "reason": "Packed-weight unpacking during accelerate loading is a different issue than bfloat16 dtype mismatches." + }, + { + "left": "issue:42175", + "right": "issue:45081", + "accept": false, + "reason": "Backend-install dependency problems are unrelated to a tokenizer regex patch crash." + }, + { + "left": "issue:43749", + "right": "issue:44291", + "accept": false, + "reason": "Both involve model loading, but FSDP CPU RAM efficient loading and init_empty_weights/_is_hf_initialized are distinct failures." + }, + { + "left": "issue:43650", + "right": "issue:43906", + "accept": false, + "reason": "The titles do not describe the same bug; one is a vague request, the other is a reproduction of another issue." + }, + { + "left": "issue:45081", + "right": "issue:45406", + "accept": false, + "reason": "Tokenizer regex patch crash and Gemma4Processor missing _tokenizer in serve are separate processor/tokenizer bugs." + }, + { + "left": "issue:43582", + "right": "issue:44857", + "accept": false, + "reason": "Apple Silicon caching allocator warmup TypeError is unrelated to LwDetrImageLoss float16 AMP crash on CUDA." + }, + { + "left": "issue:44060", + "right": "issue:44368", + "accept": false, + "reason": "Both mention tied-weights warnings in Qwen models, but they describe different warning surfaces and likely different fixes; not enough to treat as the same bug." + }, + { + "left": "issue:43262", + "right": "issue:43819", + "accept": false, + "reason": "Audio chat-template sampling-rate defaulting and DAC.from_latents missing STE are unrelated audio-processing issues." + }, + { + "left": "issue:43526", + "right": "issue:43697", + "accept": false, + "reason": "BeitImageProcessorFast label reduction and RTDetrV2 output differences are different model/processor bugs." + }, + { + "left": "issue:43526", + "right": "issue:44554", + "accept": false, + "reason": "No shared failure mode: label reduction bug vs MPS attention correctness issue." + }, + { + "left": "issue:34634", + "right": "issue:36331", + "accept": false, + "reason": "BarkProcessor voice_preset and CustomTrainer.compute_loss signature mismatch are unrelated." + }, + { + "left": "issue:45459", + "right": "issue:45803", + "accept": false, + "reason": "Both involve exception handling, but they are different code paths and different bugs; not the same underlying issue." + }, + { + "left": "issue:43329", + "right": "issue:45600", + "accept": false, + "reason": "A multimodal token-count video-branch bug is unrelated to stale auto-mapping references for removed configs." + }, + { + "left": "issue:43742", + "right": "issue:45020", + "accept": false, + "reason": "KeyError loading MobileLLM-125M and generic remote_code breakage are too broad to be the same concrete issue." + }, + { + "left": "issue:43653", + "right": "issue:44912", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and MXFP4 quantization fallback are unrelated bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 17, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:34567", + "issue:36010", + "issue:36331", + "issue:38175", + "issue:39692", + "issue:40990", + "issue:42490", + "issue:43257", + "issue:43262", + "issue:43329", + "issue:43450", + "issue:43452", + "issue:43454", + "issue:43531", + "issue:43550", + "issue:43577", + "issue:43606", + "issue:43638", + "issue:43643", + "issue:43645", + "issue:43696", + "issue:43720", + "issue:43746", + "issue:43749", + "issue:43819", + "issue:43854", + "issue:43927", + "issue:43931", + "issue:44479", + "issue:44568", + "issue:44589", + "issue:44610", + "issue:44756", + "issue:44811", + "issue:44849", + "issue:44912", + "issue:44936", + "issue:44977", + "issue:44991", + "issue:45042", + "issue:45084", + "issue:45356", + "issue:45488", + "issue:45538", + "issue:45701", + "issue:45759", + "issue:45859" + ], + "soft_pairs": [ + "issue:45084|issue:45538", + "issue:43638|issue:44936", + "issue:43452|issue:44991", + "issue:43746|issue:44811", + "issue:38175|issue:44610", + "issue:44589|issue:45042", + "issue:43854|issue:44977", + "issue:43819|issue:43927", + "issue:39692|issue:44912", + "issue:36010|issue:43645", + "issue:43257|issue:43329", + "issue:43720|issue:44610", + "issue:43550|issue:44912", + "issue:43454|issue:43577", + "issue:43257|issue:43696", + "issue:43262|issue:44811", + "issue:43329|issue:44479", + "issue:42490|issue:43257", + "issue:43531|issue:45042", + "issue:36010|issue:36331", + "issue:43749|issue:44756", + "issue:44568|issue:45356", + "issue:34567|issue:40990", + "issue:43643|issue:45759", + "issue:43329|issue:43450", + "issue:43257|issue:43931", + "issue:43606|issue:45859", + "issue:44568|issue:45488", + "issue:44849|issue:45701" + ], + "prepared_review_unit_hash": "5a80a8dab22d785c987dba14b20ac4ddef28b362d3ffb306dbb4e10b8e91f194", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 36008, + "estimated_input_tokens": 9002, + "estimated_eval_tokens": 18260 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a loose lexical-similarity cluster, not a duplicate set: the issues span training state bugs, import/load failures, tokenizer regressions, multimodal processor problems, quantization/offload issues, and model-specific generation errors.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue stands out; the cluster is too heterogeneous and the title similarities are superficial rather than describing one underlying bug.", + "canonical_pr_reason": null, + "best_issue_reason": "If forced to pick a representative, issue:45701 is the broadest/open-ended title, but it is still too generic and not a reliable duplicate anchor for the rest of the cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45084", + "right": "issue:45538", + "accept": false, + "reason": "Different problems: one is a compile-time template-node error, the other is a tokenizer max-length constant issue." + }, + { + "left": "issue:43638", + "right": "issue:44936", + "accept": false, + "reason": "Both mention training/evaluation failures, but one is a DeepSpeed ZeRO-3 indexing crash and the other is a post-train evaluate() failure; different code paths." + }, + { + "left": "issue:43452", + "right": "issue:44991", + "accept": false, + "reason": "Both are model-loading/tokenizer-related, but one concerns gguf_file handling in model/tokenizer loaders and the other is a tokenizer regression for a specific checkpoint." + }, + { + "left": "issue:43746", + "right": "issue:44811", + "accept": false, + "reason": "One is about PEFT adapter local checkpoint loading; the other is Whisper processor batch_decode skipping special tokens." + }, + { + "left": "issue:38175", + "right": "issue:44610", + "accept": false, + "reason": "Different modalities and symptoms: unexpected zero probabilities in a vision model versus an image processor size mismatch." + }, + { + "left": "issue:44589", + "right": "issue:45042", + "accept": false, + "reason": "These are unrelated loading/backend issues: a missing Float8 storage type versus PIL image processors unnecessarily requiring torchvision." + }, + { + "left": "issue:43854", + "right": "issue:44977", + "accept": false, + "reason": "Different bugs: a unit-test model load problem for GLM-4.7-Flash versus Qwen3.5 flash-attention generation behavior." + }, + { + "left": "issue:43819", + "right": "issue:43927", + "accept": false, + "reason": "Both involve model math during generation, but DAC from_latents/STE mismatch is unrelated to DiaConfig losing custom token IDs on save/load." + }, + { + "left": "issue:39692", + "right": "issue:44912", + "accept": false, + "reason": "Both mention model-specific regressions, but one is a SigLIP2 docs example error and the other is MXFP4 quantization falling back to bf16." + }, + { + "left": "issue:36010", + "right": "issue:43645", + "accept": false, + "reason": "Both are import/configuration related, but one is a GenerationMixin import error and the other is a Jupyter custom-model initialization regression." + }, + { + "left": "issue:43257", + "right": "issue:43329", + "accept": false, + "reason": "Different Qwen/MoE-related failures: weight conversion with accelerate+deepspeed versus undefined video token counting helpers." + }, + { + "left": "issue:43720", + "right": "issue:44610", + "accept": false, + "reason": "BitNet packed-weight unpacking during accelerate loading is a different failure mode than an OmDet processor/model input-size mismatch." + }, + { + "left": "issue:43550", + "right": "issue:44912", + "accept": false, + "reason": "One is a torch.compile + SDPA failure in Bamba, the other is a quantization-format regression for gpt-oss." + }, + { + "left": "issue:43454", + "right": "issue:43577", + "accept": false, + "reason": "AyaVision lm_head tying bug is unrelated to BLIP2 dtype propagation during loading." + }, + { + "left": "issue:43257", + "right": "issue:43696", + "accept": false, + "reason": "Different issues: Qwen3 MOE weight conversion/loading versus GPU OOM while loading gpt-oss-20b." + }, + { + "left": "issue:43262", + "right": "issue:44811", + "accept": false, + "reason": "Audio processor sample-rate defaulting is unrelated to Whisper batch_decode skip_special_tokens handling." + }, + { + "left": "issue:43329", + "right": "issue:44479", + "accept": false, + "reason": "The first is a missing video token-counting implementation detail; the second is a v5.3.0 video input regression across several Qwen variants." + }, + { + "left": "issue:42490", + "right": "issue:43257", + "accept": false, + "reason": "Import warning for local paths with trailing slashes is unrelated to Qwen3 MOE loading/conversion under accelerate+deepspeed." + }, + { + "left": "issue:43531", + "right": "issue:45042", + "accept": false, + "reason": "Sliding_window behavior in Qwen3-MoE is unrelated to a PIL backend dependency regression." + }, + { + "left": "issue:36010", + "right": "issue:36331", + "accept": false, + "reason": "An import error for GenerationMixin is not the same underlying bug as CustomTrainer.compute_loss receiving an unexpected num_items_in_batch kwarg." + }, + { + "left": "issue:43749", + "right": "issue:44756", + "accept": false, + "reason": "Both concern loading efficiency, but FSDP CPU RAM efficient loading and Strix Halo mmap OOM are distinct mechanisms." + }, + { + "left": "issue:44568", + "right": "issue:45356", + "accept": false, + "reason": "Tokenizer special-token handling for mdeberta-v3-base differs from Kimi-K2.5 codec handling and warning behavior." + }, + { + "left": "issue:34567", + "right": "issue:40990", + "accept": false, + "reason": "TrainerState token accounting is unrelated to GPT-oss-20b high perplexity on WikiText-2." + }, + { + "left": "issue:43643", + "right": "issue:45759", + "accept": false, + "reason": "Missing fields from trust_remote_code AutoConfig and unwrapping text_config for composite Qwen multimodal configs are different config-loading bugs." + }, + { + "left": "issue:43329", + "right": "issue:43450", + "accept": false, + "reason": "Undefined multimodal token-counting helpers are a different bug from batched video processor output shape." + }, + { + "left": "issue:43257", + "right": "issue:43931", + "accept": false, + "reason": "Qwen3 MOE loading with accelerate+deepspeed is not the same as Qwen3-VL-30B-A3B weight-shape mismatch on load." + }, + { + "left": "issue:43606", + "right": "issue:45859", + "accept": false, + "reason": "CPU-offload device mismatch in bark-small is unrelated to Qwen3_5MoeTextRotaryEmbedding incompatibility with CPU offload." + }, + { + "left": "issue:44568", + "right": "issue:45488", + "accept": false, + "reason": "Tokenizer BOS/EOS insertion regression is different from LlamaTokenizer overriding tokenizer.json pre-tokenizer settings." + }, + { + "left": "issue:44849", + "right": "issue:45701", + "accept": false, + "reason": "Qwen3.5 hidden-state bug is a specific model output issue, while 45701 is a broad tokenization-version complaint." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 18, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:28282", + "issue:29127", + "issue:33357", + "issue:34567", + "issue:34689", + "issue:38175", + "issue:41720", + "issue:42175", + "issue:42915", + "issue:43257", + "issue:43317", + "issue:43404", + "issue:43452", + "issue:43526", + "issue:43540", + "issue:43577", + "issue:43582", + "issue:43697", + "issue:43742", + "issue:43749", + "issue:43756", + "issue:43824", + "issue:43906", + "issue:43927", + "issue:44062", + "issue:44291", + "issue:44479", + "issue:44488", + "issue:44560", + "issue:44737", + "issue:44821", + "issue:44898", + "issue:44912", + "issue:44938", + "issue:44987", + "issue:44991", + "issue:45042", + "issue:45072", + "issue:45137", + "issue:45357", + "issue:45381", + "issue:45405", + "issue:45479", + "issue:45538", + "issue:45588", + "issue:45823", + "issue:45907", + "issue:45987" + ], + "soft_pairs": [ + "issue:43697|issue:45357", + "issue:42915|issue:43742", + "issue:44291|issue:45042", + "issue:33357|issue:34567", + "issue:44479|issue:45381", + "issue:43452|issue:43582", + "issue:44987|issue:45538", + "issue:41720|issue:43540", + "issue:42915|issue:44821", + "issue:44987|issue:45907", + "issue:43257|issue:43756", + "issue:42915|issue:45538", + "issue:43927|issue:44912", + "issue:43540|issue:44479", + "issue:42175|issue:43824", + "issue:43540|issue:44560", + "issue:43317|issue:45538", + "issue:43577|issue:45072", + "issue:44062|issue:44991", + "issue:42175|issue:44938", + "issue:38175|issue:44898", + "issue:43404|issue:45987", + "issue:45137|issue:45405", + "issue:43906|issue:45381", + "issue:28282|issue:34689", + "issue:42175|issue:44488", + "issue:42915|issue:43749", + "issue:43742|issue:44291", + "issue:44737|issue:45588", + "issue:43540|issue:45823", + "issue:28282|issue:29127", + "issue:43526|issue:45479" + ], + "prepared_review_unit_hash": "280637db8ddf86aeff5d973ac98243de00a77d703a9fb85131b73c1a4ee33932", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37408, + "estimated_input_tokens": 9352, + "estimated_eval_tokens": 18960 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is heterogeneous: it spans unrelated Transformers bugs across model loading, tokenizers, video inputs, dtype/device-map issues, training state, and backend/install problems. None of the soft pairs look like the same concrete bug or a mergeable duplicate PR case.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43697", + "right": "issue:45357", + "accept": false, + "reason": "Different bugs and code paths: RTDetrV2 output drift vs Qwen3.5 save_pretrained key serialization." + }, + { + "left": "issue:42915", + "right": "issue:43742", + "accept": false, + "reason": "Different model-loading failures: FineGrainedFP8Config handling vs MobileLLM-125M key error." + }, + { + "left": "issue:44291", + "right": "issue:45042", + "accept": false, + "reason": "Unrelated issues: init_empty_weights argument handling vs PIL backend requiring torchvision." + }, + { + "left": "issue:33357", + "right": "issue:34567", + "accept": false, + "reason": "Different subsystems and symptoms: MacOS bus error on CLIP vs TrainerState counter not updating." + }, + { + "left": "issue:44479", + "right": "issue:45381", + "accept": false, + "reason": "Both involve video input, but one is a broad multi-model regression and the other is a specific qwen2.5-vl vision_position_ids bug." + }, + { + "left": "issue:43452", + "right": "issue:43582", + "accept": false, + "reason": "Completely different failures: gguf_file tokenizer/model loading vs AppleSilicon warmup TypeError." + }, + { + "left": "issue:44987", + "right": "issue:45538", + "accept": false, + "reason": "Different causes and targets: one model load regression vs CLIPTokenizer model_max_length behavior." + }, + { + "left": "issue:41720", + "right": "issue:43540", + "accept": false, + "reason": "Different Qwen3 problems: auto device-map cuda assert vs video-input ValueError." + }, + { + "left": "issue:42915", + "right": "issue:44821", + "accept": false, + "reason": "Different failure modes: FP8 config issue vs loading AutoImageProcessor from URL." + }, + { + "left": "issue:44987", + "right": "issue:45907", + "accept": false, + "reason": "Different issues: model loading regression vs list_repo_files not forwarding revision." + }, + { + "left": "issue:43257", + "right": "issue:43756", + "accept": false, + "reason": "Different model families and bugs: Qwen3 MOE weight conversion vs Smollm3 RoPE-layer handling." + }, + { + "left": "issue:42915", + "right": "issue:45538", + "accept": false, + "reason": "Different topics: quantization/config loading vs tokenizer max_length default." + }, + { + "left": "issue:43927", + "right": "issue:44912", + "accept": false, + "reason": "Different layers of the stack: DiaConfig save/load token IDs vs MXFP4 quantization fallback." + }, + { + "left": "issue:43540", + "right": "issue:44479", + "accept": false, + "reason": "Related domain, but not the same bug: Qwen3OmniMoe video ValueError vs broader v5.3.0 video regression across several Qwen variants." + }, + { + "left": "issue:42175", + "right": "issue:43824", + "accept": false, + "reason": "Different problems: pip backend packaging vs missing Qwen2.5-VL import symbol." + }, + { + "left": "issue:43540", + "right": "issue:44560", + "accept": false, + "reason": "Different video-input bugs in different Qwen models and code paths." + }, + { + "left": "issue:43317", + "right": "issue:45538", + "accept": false, + "reason": "Device-map/offload loading bug vs tokenizer max_length default issue." + }, + { + "left": "issue:43577", + "right": "issue:45072", + "accept": false, + "reason": "Both mention dtype, but they affect different architectures and likely different fixes." + }, + { + "left": "issue:44062", + "right": "issue:44991", + "accept": false, + "reason": "Tokenizer-related, but one is an AddedToken constructor error and the other is a specific model tokenizer load failure." + }, + { + "left": "issue:42175", + "right": "issue:44938", + "accept": false, + "reason": "Unrelated: TensorFlow backend packaging vs Python 3.14 import/load failure." + }, + { + "left": "issue:38175", + "right": "issue:44898", + "accept": false, + "reason": "Different vision-model bugs: zero probabilities in SigLIP2 vs Perceiver non-default resolution failure." + }, + { + "left": "issue:43404", + "right": "issue:45987", + "accept": false, + "reason": "Different functionality: Mistral3 weight tying bug vs StaticCache seq length return type bug." + }, + { + "left": "issue:45137", + "right": "issue:45405", + "accept": false, + "reason": "Different concerns entirely: DeepSpeed ZeRO3 deque error vs PEFT version bump/release mismatch." + }, + { + "left": "issue:43906", + "right": "issue:45381", + "accept": false, + "reason": "No concrete overlap: isolated reproduction of another issue vs qwen2.5-vl video position-id regression." + }, + { + "left": "issue:28282", + "right": "issue:34689", + "accept": false, + "reason": "Both are loading errors, but one is missing PyTorch dependency and the other is a model-loading regression for Llama 3.2 Vision." + }, + { + "left": "issue:42175", + "right": "issue:44488", + "accept": false, + "reason": "Different issues: packaging/backends vs a specific model load failure for cjvt/sleng-bert." + }, + { + "left": "issue:42915", + "right": "issue:43749", + "accept": false, + "reason": "Different problem areas: FineGrainedFP8Config vs FSDP CPU RAM efficient loading." + }, + { + "left": "issue:43742", + "right": "issue:44291", + "accept": false, + "reason": "Different load-time exceptions affecting different APIs and model families." + }, + { + "left": "issue:44737", + "right": "issue:45588", + "accept": false, + "reason": "Different bugs: XLNet CPU tensor allocation vs flash_attention s_aux=None crash." + }, + { + "left": "issue:43540", + "right": "issue:45823", + "accept": false, + "reason": "Different model families and symptoms: Qwen3OmniMoe video parsing vs Gemma4 PLE device mismatch." + }, + { + "left": "issue:28282", + "right": "issue:29127", + "accept": false, + "reason": "Different issues: missing PyTorch dependency vs LayoutLMv3 error-message clarity for insufficient boxes." + }, + { + "left": "issue:43526", + "right": "issue:45479", + "accept": false, + "reason": "Different bug classes: BeitImageProcessorFast label reduction vs sequence-classification zero-loss behavior." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 19, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:24643", + "issue:29127", + "issue:34567", + "issue:36331", + "issue:41762", + "issue:43452", + "issue:43493", + "issue:43531", + "issue:43577", + "issue:43582", + "issue:43643", + "issue:43688", + "issue:43697", + "issue:43720", + "issue:43742", + "issue:43749", + "issue:43756", + "issue:43782", + "issue:43994", + "issue:44062", + "issue:44291", + "issue:44466", + "issue:44484", + "issue:44488", + "issue:44514", + "issue:44589", + "issue:44625", + "issue:44704", + "issue:44792", + "issue:44821", + "issue:44843", + "issue:44877", + "issue:44898", + "issue:44936", + "issue:44964", + "issue:45005", + "issue:45042", + "issue:45081", + "issue:45216", + "issue:45290", + "issue:45325", + "issue:45357", + "issue:45375", + "issue:45381", + "issue:45405", + "issue:45600", + "issue:45920" + ], + "soft_pairs": [ + "issue:43582|issue:43782", + "issue:43493|issue:43994", + "issue:43756|issue:45325", + "issue:34567|issue:36331", + "issue:44062|issue:45081", + "issue:44291|issue:44821", + "issue:45357|issue:45920", + "issue:43643|issue:45375", + "issue:44792|issue:44964", + "issue:43452|issue:44488", + "issue:45042|issue:45081", + "issue:43742|issue:44589", + "issue:45405|issue:45600", + "issue:43531|issue:44821", + "issue:43697|issue:45216", + "issue:44514|issue:45381", + "issue:43688|issue:43749", + "issue:43720|issue:44898", + "issue:44484|issue:44704", + "issue:24643|issue:29127", + "issue:41762|issue:44936", + "issue:44625|issue:44877", + "issue:44514|issue:45290", + "issue:43452|issue:44843", + "issue:43452|issue:43577", + "issue:44466|issue:45005" + ], + "prepared_review_unit_hash": "5ceb14cfb80aecb32f852d2f57ee5911b2f56cc7fe17526b025730de1e55cd79", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 26, + "serialized_chars": 35288, + "estimated_input_tokens": 8822, + "estimated_eval_tokens": 17900 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly heterogeneous issue set with one clear duplicate mini-cluster around tied-weights serialization in v5. I found one strong duplicate pair; the other soft matches are too broad or model-specific to merge confidently.", + "confidence": 0.41, + "canonical_issue_reason": "Issue 45005 is the best umbrella for the tied-weights regression: it states the broader v5 problem, while 44466 looks like a specific device-dependent manifestation of the same save/serialization bug.", + "canonical_pr_reason": null, + "best_issue_reason": "45005 is the most representative issue to anchor triage because it captures the broader underlying regression rather than one symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43582", + "right": "issue:43782", + "accept": false, + "reason": "Different subsystems and failures: Apple Silicon allocator warmup TypeError vs Qwen3VL weight_only loading error." + }, + { + "left": "issue:43493", + "right": "issue:43994", + "accept": false, + "reason": "Both involve SigLIP2, but one is a HF-vs-JAX discrepancy and the other is a bad AutoModel/pipeline result; not clearly the same concrete bug." + }, + { + "left": "issue:43756", + "right": "issue:45325", + "accept": false, + "reason": "Different models and different RoPE/position-id bugs; no shared code-path evidence." + }, + { + "left": "issue:34567", + "right": "issue:36331", + "accept": false, + "reason": "Trainer-state token accounting bug vs custom compute_loss signature mismatch; related area, not the same defect." + }, + { + "left": "issue:44062", + "right": "issue:45081", + "accept": false, + "reason": "Different tokenizer failures: duplicate AddedToken kwargs vs mistral regex patch crashing on backend_tokenizer access." + }, + { + "left": "issue:44291", + "right": "issue:44821", + "accept": false, + "reason": "init_empty_weights/_is_hf_initialized loading bug is unrelated to AutoImageProcessor URL loading." + }, + { + "left": "issue:45357", + "right": "issue:45920", + "accept": false, + "reason": "Different regressions: Qwen3.5 save_pretrained visual-key serialization vs generic AutoTokenizer token-id errors." + }, + { + "left": "issue:43643", + "right": "issue:45375", + "accept": false, + "reason": "Both are config-related, but one is remote-code config field loss and the other is a specific strict-field drop in Qwen3_5MoeVisionConfig." + }, + { + "left": "issue:44792", + "right": "issue:44964", + "accept": false, + "reason": "Both are multimodal/model-load failures, but they affect different models and likely different code paths." + }, + { + "left": "issue:43452", + "right": "issue:44488", + "accept": false, + "reason": "Generic gguf_file/from_pretrained breakage vs one model failing to load; not enough evidence they are the same bug." + }, + { + "left": "issue:45042", + "right": "issue:45081", + "accept": false, + "reason": "PIL/torchvision dependency regression is unrelated to Mistral regex patching." + }, + { + "left": "issue:43742", + "right": "issue:44589", + "accept": false, + "reason": "KeyError on MobileLLM loading and missing Float8 storage are different loader/runtime issues." + }, + { + "left": "issue:45405", + "right": "issue:45600", + "accept": false, + "reason": "PEFT version pin mismatch vs stale auto_mappings reference; both CI/package issues, but not the same change." + }, + { + "left": "issue:43531", + "right": "issue:44821", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and AutoImageProcessor URL loading are unrelated." + }, + { + "left": "issue:43697", + "right": "issue:45216", + "accept": false, + "reason": "RTDetrV2 output regression and Qwen3.5 checkpoint save regression affect different paths and symptoms." + }, + { + "left": "issue:44514", + "right": "issue:45381", + "accept": false, + "reason": "Both are Qwen2.5-VL, but one is chat-template batching and the other is video position-id calculation; not the same bug." + }, + { + "left": "issue:43688", + "right": "issue:43749", + "accept": false, + "reason": "Aux-loss normalization bug vs FSDP CPU RAM-efficient loading breakage are unrelated." + }, + { + "left": "issue:43720", + "right": "issue:44898", + "accept": false, + "reason": "BitNet packed-weight loading bug and Perceiver image-classification failure are different model-specific problems." + }, + { + "left": "issue:44484", + "right": "issue:44704", + "accept": false, + "reason": "save_pretrained shard-size behavior vs AutoProcessor cached_file kwargs forwarding are unrelated." + }, + { + "left": "issue:24643", + "right": "issue:29127", + "accept": false, + "reason": "DeepSpeed training weight-dimension runtime error vs LayoutLMv3 box validation message clarity are different issues." + }, + { + "left": "issue:41762", + "right": "issue:44936", + "accept": false, + "reason": "Gemma3 ZeRO-3 load crash and trainer evaluate-after-train failure do not look like the same underlying bug." + }, + { + "left": "issue:44625", + "right": "issue:44877", + "accept": false, + "reason": "Qwen3.5 num_labels propagation and granite_speech strict-config loading are both config bugs, but not the same concrete issue." + }, + { + "left": "issue:44514", + "right": "issue:45290", + "accept": false, + "reason": "Both touch apply_chat_template, but one is batched padding handling and the other is tool-call assistant message handling." + }, + { + "left": "issue:43452", + "right": "issue:44843", + "accept": false, + "reason": "gguf_file loading breakage is unrelated to the mistral-regex offline model_info call bug." + }, + { + "left": "issue:43452", + "right": "issue:43577", + "accept": false, + "reason": "GGUF/from_pretrained loading issue and BLIP2 dtype propagation issue are different code paths." + }, + { + "left": "issue:44466", + "right": "issue:45005", + "accept": true, + "reason": "Both describe the same tied-weights serialization regression in v5/save_pretrained; 45005 is the broader umbrella and 44466 is a concrete manifestation." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet and stays conservative overall. The only accepted pair, 44466/45005, is plausibly the same tied-weights save/serialization regression in v5, with 45005 as the broader umbrella. The other soft matches are mostly model- or code-path-specific and were reasonably rejected." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 20, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:28282", + "issue:30064", + "issue:32090", + "issue:33357", + "issue:33666", + "issue:34567", + "issue:35141", + "issue:38175", + "issue:41720", + "issue:43122", + "issue:43295", + "issue:43504", + "issue:43531", + "issue:43550", + "issue:43577", + "issue:43643", + "issue:43646", + "issue:43653", + "issue:43688", + "issue:43697", + "issue:43720", + "issue:43742", + "issue:43749", + "issue:43819", + "issue:43906", + "issue:43927", + "issue:44361", + "issue:44442", + "issue:44464", + "issue:44479", + "issue:44554", + "issue:44561", + "issue:44568", + "issue:44625", + "issue:44743", + "issue:44811", + "issue:44843", + "issue:44936", + "issue:44977", + "issue:44991", + "issue:45356", + "issue:45440", + "issue:45491", + "issue:45588", + "issue:45701", + "issue:45704", + "issue:45823", + "issue:46002" + ], + "soft_pairs": [ + "issue:43295|issue:45356", + "issue:43819|issue:44811", + "issue:44936|issue:45823", + "issue:38175|issue:43295", + "issue:44361|issue:44442", + "issue:43653|issue:44361", + "issue:30064|issue:33666", + "issue:44464|issue:44743", + "issue:41720|issue:43906", + "issue:43653|issue:44568", + "issue:28282|issue:32090", + "issue:28282|issue:33357", + "issue:43688|issue:45704", + "issue:43688|issue:45440", + "issue:45491|issue:46002", + "issue:43577|issue:43697", + "issue:28282|issue:35141", + "issue:43643|issue:44561", + "issue:44843|issue:44991", + "issue:43646|issue:43749", + "issue:43688|issue:45701", + "issue:43653|issue:43927", + "issue:43531|issue:43742", + "issue:43550|issue:43927", + "issue:44464|issue:44977", + "issue:44479|issue:44625", + "issue:43550|issue:44361", + "issue:43122|issue:43906", + "issue:43577|issue:44554", + "issue:43577|issue:45588", + "issue:43504|issue:43550", + "issue:43504|issue:43720", + "issue:32090|issue:34567", + "issue:41720|issue:44464" + ], + "prepared_review_unit_hash": "c8b48e762319516cc90deb7c21b36f281b94ad69389672bd6ca69e153f5a773b", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37924, + "estimated_input_tokens": 9481, + "estimated_eval_tokens": 19218 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a noisy mix of unrelated bugs. The strongest recurring theme is tokenizer/version-regression reports, but the soft pairs still describe different concrete failures, so none should be merged.", + "confidence": 0.9, + "canonical_issue_reason": "issue:45701 is the best umbrella representative: it explicitly frames cross-version tokenization changes and has inbound references, which fits the main recurring theme better than the more specific symptom reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45701 is the most suitable single issue to anchor this cluster because it is the broadest tokenizer-regression report and most likely to subsume the nearby version-dependent tokenization complaints.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43295", + "right": "issue:45356", + "accept": false, + "reason": "Both are tokenizer regressions, but one is about processor.tokenizer/image handling and the other is a Kimi-K2.5 codec/fix_mistral_regex regression; different failure modes." + }, + { + "left": "issue:43819", + "right": "issue:44811", + "accept": false, + "reason": "Different code paths: DAC.from_latents vs Whisper batch_decode skip_special_tokens." + }, + { + "left": "issue:44936", + "right": "issue:45823", + "accept": false, + "reason": "Trainer evaluate-after-train behavior is unrelated to Gemma4 PLE device_map forward mismatch." + }, + { + "left": "issue:38175", + "right": "issue:43295", + "accept": false, + "reason": "siglip2 zero probabilities is a model/output issue, not the tokenizer regression described by 43295." + }, + { + "left": "issue:44361", + "right": "issue:44442", + "accept": false, + "reason": "Both involve tokenizers, but they are different tokenizer classes and distinct loading failures." + }, + { + "left": "issue:43653", + "right": "issue:44361", + "accept": false, + "reason": "BigBird special-token registration/empty decode is unrelated to MLukeTokenizer task AttributeError." + }, + { + "left": "issue:30064", + "right": "issue:33666", + "accept": false, + "reason": "Image processor void segmentation maps and Qwen2-VL multi-GPU training are unrelated subsystems." + }, + { + "left": "issue:44464", + "right": "issue:44743", + "accept": false, + "reason": "Both concern generation/state handling, but one is compiled-forward inconsistency and the other is recurrent-state reset with cache; not the same bug." + }, + { + "left": "issue:41720", + "right": "issue:43906", + "accept": false, + "reason": "Qwen3 auto-device-map cuda assert is not the same as the isolated reproduction of a different upstream issue." + }, + { + "left": "issue:43653", + "right": "issue:44568", + "accept": false, + "reason": "Both are tokenizer special-token issues, but on different tokenizers with different symptoms and code paths." + }, + { + "left": "issue:28282", + "right": "issue:32090", + "accept": false, + "reason": "PyTorch missing ImportError is unrelated to Trainer _gpu_broadcast_one rejecting NoneType." + }, + { + "left": "issue:28282", + "right": "issue:33357", + "accept": false, + "reason": "Missing PyTorch import is unrelated to the MacOS bus error on a community CLIP model." + }, + { + "left": "issue:43688", + "right": "issue:45704", + "accept": false, + "reason": "Aux-loss normalization in MoE models is unrelated to the T5 apex.FusedRMSNorm memory leak." + }, + { + "left": "issue:43688", + "right": "issue:45440", + "accept": false, + "reason": "Both touch MoE models, but one is auxiliary-loss normalization and the other is parity divergence from a remote implementation." + }, + { + "left": "issue:45491", + "right": "issue:46002", + "accept": false, + "reason": "Both mention NaNs, but the triggers are different: sliding-window all-padding embeddings vs checkout/init classification heads." + }, + { + "left": "issue:43577", + "right": "issue:43697", + "accept": false, + "reason": "BLIP2 dtype preservation and RTDetrV2 output mismatch are separate model-specific bugs." + }, + { + "left": "issue:28282", + "right": "issue:35141", + "accept": false, + "reason": "Missing PyTorch is unrelated to output embeddings being reinitialized during resize." + }, + { + "left": "issue:43643", + "right": "issue:44561", + "accept": false, + "reason": "AutoConfig missing fields under trust_remote_code and removal of is_torch_fx_available are different regressions." + }, + { + "left": "issue:44843", + "right": "issue:44991", + "accept": false, + "reason": "Offline model_info() call in tokenizer patching is a different bug from failing to load EMBEDDIA/est-roberta's tokenizer." + }, + { + "left": "issue:43646", + "right": "issue:43749", + "accept": false, + "reason": "Custom model initialization breakage is unrelated to FSDP CPU RAM efficient loading." + }, + { + "left": "issue:43688", + "right": "issue:45701", + "accept": false, + "reason": "MoE auxiliary-loss normalization is unrelated to the tokenizer version-change reports." + }, + { + "left": "issue:43653", + "right": "issue:43927", + "accept": false, + "reason": "BigBird mask-token special-token registration and DiaConfig custom token IDs loss are different token-management bugs." + }, + { + "left": "issue:43531", + "right": "issue:43742", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior is unrelated to the MobileLLM-125M key error on load." + }, + { + "left": "issue:43550", + "right": "issue:43927", + "accept": false, + "reason": "torch.compile SDPA failure is unrelated to config token IDs being lost after save/load." + }, + { + "left": "issue:44464", + "right": "issue:44977", + "accept": false, + "reason": "Compiled-forward chunked generation inconsistency and flash-attention generation failure are different concrete code paths." + }, + { + "left": "issue:44479", + "right": "issue:44625", + "accept": false, + "reason": "Video-input regression across Qwen VL variants is unrelated to num_labels propagation in Qwen3.5." + }, + { + "left": "issue:43550", + "right": "issue:44361", + "accept": false, + "reason": "Bamba torch.compile/SDPA failure is unrelated to MLukeTokenizer AttributeError." + }, + { + "left": "issue:43122", + "right": "issue:43906", + "accept": false, + "reason": "Different tokenization across versions and an isolated reproduction of another issue are not the same bug." + }, + { + "left": "issue:43577", + "right": "issue:44554", + "accept": false, + "reason": "BLIP2 dtype loading and MPS attention correctness are unrelated." + }, + { + "left": "issue:43577", + "right": "issue:45588", + "accept": false, + "reason": "dtype preservation on load is unrelated to flash_attention.py crashing on s_aux=None." + }, + { + "left": "issue:43504", + "right": "issue:43550", + "accept": false, + "reason": "Beit legacy-field loading failure is unrelated to Bamba torch.compile SDPA." + }, + { + "left": "issue:43504", + "right": "issue:43720", + "accept": false, + "reason": "Legacy preset loading for Beit is unrelated to BitNet packed-weight unpacking during accelerate loading." + }, + { + "left": "issue:32090", + "right": "issue:34567", + "accept": false, + "reason": "Trainer _gpu_broadcast_one NoneType and TrainerState num_input_tokens_seen not updating are separate trainer bugs." + }, + { + "left": "issue:41720", + "right": "issue:44464", + "accept": false, + "reason": "Qwen3 auto-device-map cuda assert is unrelated to chunked generation with compiled forward." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 21, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:28218", + "issue:32090", + "issue:33357", + "issue:33666", + "issue:36010", + "issue:36331", + "issue:39692", + "issue:43257", + "issue:43262", + "issue:43278", + "issue:43295", + "issue:43299", + "issue:43334", + "issue:43452", + "issue:43493", + "issue:43504", + "issue:43525", + "issue:43540", + "issue:43572", + "issue:43606", + "issue:43688", + "issue:43720", + "issue:43742", + "issue:43756", + "issue:43825", + "issue:43844", + "issue:43927", + "issue:44361", + "issue:44451", + "issue:44488", + "issue:44568", + "issue:44610", + "issue:44625", + "issue:44821", + "issue:44843", + "issue:44849", + "issue:44913", + "issue:44964", + "issue:44977", + "issue:45083", + "issue:45200", + "issue:45229", + "issue:45357", + "issue:45405", + "issue:45803", + "issue:45812", + "issue:45987", + "issue:46002" + ], + "soft_pairs": [ + "issue:43504|issue:44610", + "issue:32090|issue:33666", + "issue:43452|issue:44451", + "issue:43742|issue:44821", + "issue:33666|issue:36010", + "issue:43278|issue:43688", + "issue:43606|issue:43927", + "issue:43493|issue:43825", + "issue:43262|issue:43295", + "issue:43720|issue:43927", + "issue:44849|issue:45357", + "issue:43756|issue:44913", + "issue:44625|issue:46002", + "issue:44568|issue:45812", + "issue:43525|issue:43572", + "issue:43720|issue:44361", + "issue:28218|issue:33357", + "issue:33666|issue:36331", + "issue:39692|issue:43493", + "issue:43540|issue:44977", + "issue:43493|issue:43844", + "issue:44964|issue:45405", + "issue:45803|issue:45987", + "issue:43278|issue:43844", + "issue:43756|issue:45083", + "issue:43257|issue:43299", + "issue:45200|issue:45229", + "issue:43334|issue:43572", + "issue:44488|issue:44843" + ], + "prepared_review_unit_hash": "c02a4a28804214324e434c2b3f5e8f667b1fc435a933db691dcd71e481ac3687", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36497, + "estimated_input_tokens": 9125, + "estimated_eval_tokens": 18506 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly a title-similarity bucket, not a duplicate set; the soft pairs are related only at a very broad subsystem level and should be rejected. The strongest standalone issue is the Granite tokenizer regression, but there is no single canonical duplicate across the whole cluster.", + "confidence": 0.87, + "canonical_issue_reason": "issue:45812 is the most central, concrete regression in the set: it has broad impact, clear tokenizer misbehavior, and the strongest discussion/reference signal.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45812 is the best representative issue to anchor triage because it is a clearly described, high-impact, user-facing regression with multiple inbound references.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43504", + "right": "issue:44610", + "accept": false, + "reason": "Different models and failures: legacy-field load error vs processor/model input-size mismatch." + }, + { + "left": "issue:32090", + "right": "issue:33666", + "accept": false, + "reason": "Trainer GPU broadcast TypeError is unrelated to Qwen2-VL multi-GPU training behavior." + }, + { + "left": "issue:43452", + "right": "issue:44451", + "accept": false, + "reason": "Both are loading failures, but one is gguf handling and the other is a specific model load regression." + }, + { + "left": "issue:43742", + "right": "issue:44821", + "accept": false, + "reason": "KeyError on MobileLLM loading is not the same as AutoImageProcessor URL loading." + }, + { + "left": "issue:33666", + "right": "issue:36010", + "accept": false, + "reason": "Multi-GPU training for Qwen2-VL is unrelated to a GenerationMixin import regression." + }, + { + "left": "issue:43278", + "right": "issue:43688", + "accept": false, + "reason": "Embedding dtype drift during eval is unrelated to auxiliary-loss normalization in MoE models." + }, + { + "left": "issue:43606", + "right": "issue:43927", + "accept": false, + "reason": "CPU offload device mismatch and DiaConfig token-id persistence are different code paths and symptoms." + }, + { + "left": "issue:43493", + "right": "issue:43825", + "accept": false, + "reason": "SigLIP2 implementation mismatch is unrelated to a pipeline error-message regression." + }, + { + "left": "issue:43262", + "right": "issue:43295", + "accept": false, + "reason": "Audio chat-template sample-rate defaulting is unrelated to processor.tokenizer/image-passing regression." + }, + { + "left": "issue:43720", + "right": "issue:43927", + "accept": false, + "reason": "BitNet packed-weight unpacking and DiaConfig token IDs are separate loading bugs." + }, + { + "left": "issue:44849", + "right": "issue:45357", + "accept": false, + "reason": "Qwen3.5 hidden-state behavior and visual-encoder save_pretrained key corruption are different regressions." + }, + { + "left": "issue:43756", + "right": "issue:44913", + "accept": false, + "reason": "Smollm3 RoPE-layer omission is not the same as GPTNeoX rotary_pct reload loss." + }, + { + "left": "issue:44625", + "right": "issue:46002", + "accept": false, + "reason": "num_labels propagation and NaNs in classification heads are distinct model init/config issues." + }, + { + "left": "issue:44568", + "right": "issue:45812", + "accept": false, + "reason": "mDeBERTa add_special_tokens tokenization bug is unrelated to Granite token-ID corruption." + }, + { + "left": "issue:43525", + "right": "issue:43572", + "accept": false, + "reason": "Both involve missing pad-token fields, but they affect different configs and failure modes." + }, + { + "left": "issue:43720", + "right": "issue:44361", + "accept": false, + "reason": "BitNet accelerate loading and MLukeTokenizer AttributeError are not the same bug." + }, + { + "left": "issue:28218", + "right": "issue:33357", + "accept": false, + "reason": "Tokenizer extra-space bug is unrelated to a MacOS bus error on a CLIP model." + }, + { + "left": "issue:33666", + "right": "issue:36331", + "accept": false, + "reason": "Qwen2-VL multi-GPU training is unrelated to CustomTrainer.compute_loss signature breakage." + }, + { + "left": "issue:39692", + "right": "issue:43493", + "accept": false, + "reason": "SigLIP2 doc example errors are documentation/runtime issues, not the same as core implementation discrepancy." + }, + { + "left": "issue:43540", + "right": "issue:44977", + "accept": false, + "reason": "Qwen3OmniMoe video processing error and Qwen3.5 flash-attention generation bug are different problems." + }, + { + "left": "issue:43493", + "right": "issue:43844", + "accept": false, + "reason": "SigLIP2 implementation mismatch is unrelated to DeepSpeed ZeRO-3 gradient growth." + }, + { + "left": "issue:44964", + "right": "issue:45405", + "accept": false, + "reason": "Phi-4 multimodal load failure is unrelated to a dependency version bump issue." + }, + { + "left": "issue:45803", + "right": "issue:45987", + "accept": false, + "reason": "Bare except in Fuyu tensor conversion and StaticCache seq-length return type are unrelated." + }, + { + "left": "issue:43278", + "right": "issue:43844", + "accept": false, + "reason": "Embedding dtype drift is unrelated to ZeRO-3 gradient instability." + }, + { + "left": "issue:43756", + "right": "issue:45083", + "accept": false, + "reason": "Smollm3 RoPE-layer dropping and Qwen3 Omni helper length behavior are separate implementation bugs." + }, + { + "left": "issue:43257", + "right": "issue:43299", + "accept": false, + "reason": "Qwen3 MOE accelerate/deepspeed conversion and Qwen3 VL loading regression are different model families and failures." + }, + { + "left": "issue:45200", + "right": "issue:45229", + "accept": false, + "reason": "Missing mm_token_type_ids default and Gemma 4 multi-GPU OOM are not the same bug." + }, + { + "left": "issue:43334", + "right": "issue:43572", + "accept": false, + "reason": "Missing pad_token_id in Qwen3-VL config and missing pad_token_idx in StableLmConfig are similar in shape but different defects." + }, + { + "left": "issue:44488", + "right": "issue:44843", + "accept": false, + "reason": "Two separate loading issues: model-specific tokenizer load failure vs offline-mode hub lookup bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 22, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:28218", + "issue:29942", + "issue:30333", + "issue:33357", + "issue:33666", + "issue:34567", + "issue:39692", + "issue:42490", + "issue:43278", + "issue:43295", + "issue:43334", + "issue:43504", + "issue:43540", + "issue:43572", + "issue:43575", + "issue:43582", + "issue:43606", + "issue:43646", + "issue:43653", + "issue:43696", + "issue:43701", + "issue:43824", + "issue:43844", + "issue:43927", + "issue:43994", + "issue:44112", + "issue:44291", + "issue:44361", + "issue:44423", + "issue:44464", + "issue:44466", + "issue:44493", + "issue:44521", + "issue:44554", + "issue:44589", + "issue:44734", + "issue:44792", + "issue:44964", + "issue:44977", + "issue:45200", + "issue:45229", + "issue:45290", + "issue:45405", + "issue:45446", + "issue:45520", + "issue:45584", + "issue:45759", + "issue:45823" + ], + "soft_pairs": [ + "issue:43334|issue:45759", + "issue:39692|issue:43295", + "issue:43646|issue:44466", + "issue:45229|issue:45823", + "issue:43824|issue:44589", + "issue:43504|issue:43606", + "issue:33357|issue:33666", + "issue:43572|issue:44291", + "issue:43927|issue:44361", + "issue:44589|issue:45520", + "issue:43504|issue:43653", + "issue:44464|issue:45584", + "issue:43606|issue:44112", + "issue:44423|issue:44734", + "issue:43575|issue:45405", + "issue:43994|issue:44493", + "issue:43278|issue:43701", + "issue:44792|issue:45405", + "issue:44521|issue:45290", + "issue:29942|issue:30333", + "issue:42490|issue:43696", + "issue:43844|issue:44964", + "issue:28218|issue:34567", + "issue:44554|issue:45446", + "issue:43582|issue:44977", + "issue:43540|issue:45200" + ], + "prepared_review_unit_hash": "bc374d2c648e817aa69515b62092c75ae34463a00206c2c4570e3e0f06ec70dc", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 26, + "serialized_chars": 35513, + "estimated_input_tokens": 8879, + "estimated_eval_tokens": 18014 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a grab bag of unrelated issue reports across tokenizers, multimodal loading, training/runtime errors, and CI; none of the soft pairs look like true duplicates.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue: the items describe different bugs in different code paths and subsystems, so there is no clear duplicate anchor for the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "No strong best issue representative; none of the issues subsume the others or appear to be the same underlying defect.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43334", + "right": "issue:45759", + "accept": false, + "reason": "Both involve Qwen multimodal configs, but one is a load-time pad_token_id AttributeError and the other is AutoModelForCausalLM.from_config not unwrapping text_config; different failure modes." + }, + { + "left": "issue:39692", + "right": "issue:43295", + "accept": false, + "reason": "Both are multimodal regression reports, but one is a SigLIP2 documentation example problem and the other is processor.tokenizer/image-passing custom model breakage." + }, + { + "left": "issue:43646", + "right": "issue:44466", + "accept": false, + "reason": "Distinct custom-model issues: init_empty_weights initialization failure versus inconsistent lm_head.weight serialization tied to device placement." + }, + { + "left": "issue:45229", + "right": "issue:45823", + "accept": false, + "reason": "Both concern Gemma4 and multi-device behavior, but one is CUDA OOM during inference and the other is a PLE device mismatch during forward." + }, + { + "left": "issue:43824", + "right": "issue:44589", + "accept": false, + "reason": "Different import/runtime problems: Qwen2_5_VL import failure versus missing Float8 storage type." + }, + { + "left": "issue:43504", + "right": "issue:43606", + "accept": false, + "reason": "Beit legacy-field loading bug and suno/bark-small CPU-offload device mismatch are unrelated code paths." + }, + { + "left": "issue:33357", + "right": "issue:33666", + "accept": false, + "reason": "Different models and failures: MacOS bus error for community CLIP versus Qwen2-VL multi-GPU training." + }, + { + "left": "issue:43572", + "right": "issue:44291", + "accept": false, + "reason": "Both mention transformers 5/init changes, but one is a missing pad_token_idx config field and the other is an unexpected _is_hf_initialized argument TypeError." + }, + { + "left": "issue:43927", + "right": "issue:44361", + "accept": false, + "reason": "Tokenizer/config token-ID persistence bug versus MLukeTokenizer task AttributeError; not the same defect." + }, + { + "left": "issue:44589", + "right": "issue:45520", + "accept": false, + "reason": "Both are import-related, but one is a storage object error and the other is a flash_attn KeyError on Python 3.13." + }, + { + "left": "issue:43504", + "right": "issue:43653", + "accept": false, + "reason": "Different tokenizer/model-loading bugs: Beit legacy field preset loading versus BigBirdTokenizer special-token registration/empty decode." + }, + { + "left": "issue:44464", + "right": "issue:45584", + "accept": false, + "reason": "Chunked generation with compiled forward is unrelated to empty transcription after align_special_tokens." + }, + { + "left": "issue:43606", + "right": "issue:44112", + "accept": false, + "reason": "One is a real CPU-offload device mismatch for bark-small; the other is a stale GraniteSpeech CI test issue." + }, + { + "left": "issue:44423", + "right": "issue:44734", + "accept": false, + "reason": "Both affect serving, but one crashes on multimodal continuous batching and the other on KV-cache continuation indexing in /v1/responses." + }, + { + "left": "issue:43575", + "right": "issue:45405", + "accept": false, + "reason": "Qwen2 model OOM with tensor parallelism is unrelated to a PEFT version pin bump." + }, + { + "left": "issue:43994", + "right": "issue:44493", + "accept": false, + "reason": "SigLIP2 nonsensical outputs and unexpected position-id keys are different regression symptoms with different causes." + }, + { + "left": "issue:43278", + "right": "issue:43701", + "accept": false, + "reason": "BF16-to-FP32 embedding dtype drift in evaluate is unrelated to resume_from_checkpoint key mismatch." + }, + { + "left": "issue:44792", + "right": "issue:45405", + "accept": false, + "reason": "Failed janus generation test and PEFT minimum-version bump do not describe the same bug." + }, + { + "left": "issue:44521", + "right": "issue:45290", + "accept": false, + "reason": "Both involve chat templates, but all-zero assistant masks for multimodal inputs is a different bug from crashing on assistant tool-call messages with no content." + }, + { + "left": "issue:29942", + "right": "issue:30333", + "accept": false, + "reason": "Flash Attention 2 test failures and MLFlow reporting after training failure are unrelated." + }, + { + "left": "issue:42490", + "right": "issue:43696", + "accept": false, + "reason": "Trailing-slash local-model import warning is unrelated to GPT-oss-20b CUDA OOM." + }, + { + "left": "issue:43844", + "right": "issue:44964", + "accept": false, + "reason": "ZeRO-3 gradient anomaly and Phi-4 multimodal loading failure are different issues." + }, + { + "left": "issue:28218", + "right": "issue:34567", + "accept": false, + "reason": "Tokenizer extra-space bug and TrainerState token-count tracking bug are unrelated." + }, + { + "left": "issue:44554", + "right": "issue:45446", + "accept": false, + "reason": "MPS attention correctness issue and a PyTorch version check bug in flex_attention are different code paths." + }, + { + "left": "issue:43582", + "right": "issue:44977", + "accept": false, + "reason": "Apple Silicon caching_allocator_warmup TypeError and Qwen3.5 flash-attention generation problems are unrelated." + }, + { + "left": "issue:43540", + "right": "issue:45200", + "accept": false, + "reason": "Qwen3OmniMoe video-processing ValueError and Gemma4 mm_token_type_ids defaulting are separate 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": 1, + "component_count": 17, + "review_unit_index": 23, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:15354", + "issue:28218", + "issue:35141", + "issue:37428", + "issue:38175", + "issue:42490", + "issue:42915", + "issue:43232", + "issue:43278", + "issue:43295", + "issue:43334", + "issue:43388", + "issue:43450", + "issue:43493", + "issue:43525", + "issue:43526", + "issue:43572", + "issue:43701", + "issue:43824", + "issue:43844", + "issue:43856", + "issue:43873", + "issue:43957", + "issue:43975", + "issue:44315", + "issue:44442", + "issue:44466", + "issue:44514", + "issue:44521", + "issue:44554", + "issue:44610", + "issue:44792", + "issue:44849", + "issue:44898", + "issue:44936", + "issue:44987", + "issue:45072", + "issue:45084", + "issue:45127", + "issue:45198", + "issue:45200", + "issue:45325", + "issue:45356", + "issue:45412", + "issue:45520", + "issue:45563", + "issue:45823" + ], + "soft_pairs": [ + "issue:15354|issue:38175", + "issue:45084|issue:45520", + "issue:45072|issue:45198", + "issue:45200|issue:45823", + "issue:37428|issue:43957", + "issue:43232|issue:44849", + "issue:44315|issue:44849", + "issue:43493|issue:43701", + "issue:43295|issue:44610", + "issue:44554|issue:45356", + "issue:44898|issue:45412", + "issue:44466|issue:45127", + "issue:43572|issue:44987", + "issue:37428|issue:43824", + "issue:45200|issue:45325", + "issue:43572|issue:43957", + "issue:43295|issue:43450", + "issue:43844|issue:44936", + "issue:43295|issue:43873", + "issue:43334|issue:43525", + "issue:44792|issue:45563", + "issue:43824|issue:43975", + "issue:43388|issue:43526", + "issue:42915|issue:43824", + "issue:38175|issue:45412", + "issue:38175|issue:43493", + "issue:44514|issue:44521", + "issue:44442|issue:45198", + "issue:28218|issue:35141", + "issue:43278|issue:43856", + "issue:42490|issue:43493" + ], + "prepared_review_unit_hash": "9c78cb0a48db868ea81eaa4ceaf810b1e3378992e3a013b81a07670bde75e844", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36676, + "estimated_input_tokens": 9169, + "estimated_eval_tokens": 18594 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These items are a heterogeneous mix of unrelated Transformers bugs; none of the soft pairs look like true duplicates, though a few share broad subsystem or model-family overlap.", + "confidence": 0.75, + "canonical_issue_reason": "I\u2019d use issue:37428 as the canonical issue because it is a clear, broadly relevant core-runtime regression and the most representative of the mixed cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:37428 is the best representative issue here: concrete, actionable, and library-level, whereas most others are narrower model-specific reports or orthogonal bugs.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:15354", + "right": "issue:38175", + "accept": false, + "reason": "Different failures: torch.jit.script generator-expression support vs SigLIP2 zero-probability output." + }, + { + "left": "issue:45084", + "right": "issue:45520", + "accept": false, + "reason": "Unrelated code paths: template-node compilation error vs flash_attn import KeyError on Python 3.13." + }, + { + "left": "issue:45072", + "right": "issue:45198", + "accept": false, + "reason": "Different components and symptoms: bfloat16 dtype mismatch vs Wav2Vec2 save/tokenization failure." + }, + { + "left": "issue:45200", + "right": "issue:45823", + "accept": false, + "reason": "Both Gemma4, but one is missing mm_token_type_ids defaulting and the other is a device_map forward mismatch." + }, + { + "left": "issue:37428", + "right": "issue:43957", + "accept": false, + "reason": "Both are loading/import regressions, but one is a missing flash attention helper import and the other is meta-device model loading." + }, + { + "left": "issue:43232", + "right": "issue:44849", + "accept": false, + "reason": "Generation kwarg update bug vs Qwen3.5 hidden-state output bug; not the same code path." + }, + { + "left": "issue:44315", + "right": "issue:44849", + "accept": false, + "reason": "Liger Kernel application in model_init is unrelated to the Qwen3.5 output_hidden_states regression." + }, + { + "left": "issue:43493", + "right": "issue:43701", + "accept": false, + "reason": "SigLIP2 implementation discrepancy and resume_from_checkpoint key mismatch are distinct problems." + }, + { + "left": "issue:43295", + "right": "issue:44610", + "accept": false, + "reason": "Processor API regression vs OmDet image-size mismatch; different failures in different processors/models." + }, + { + "left": "issue:44554", + "right": "issue:45356", + "accept": false, + "reason": "MPS attention correctness issue and Kimi tokenizer regression are unrelated." + }, + { + "left": "issue:44898", + "right": "issue:45412", + "accept": false, + "reason": "Perceiver resize/interpolation bug vs RT-DETR memory leak; no shared underlying defect." + }, + { + "left": "issue:44466", + "right": "issue:45127", + "accept": false, + "reason": "Both touch tied-weights behavior, but one is serialization inconsistency and the other is LoRA merge collapse after vocab extension." + }, + { + "left": "issue:43572", + "right": "issue:44987", + "accept": false, + "reason": "Missing pad_token_idx in one config is not the same as loading failure for physical-intelligence/fast on 5.1.0." + }, + { + "left": "issue:37428", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but they involve different missing symbols/classes and different modules." + }, + { + "left": "issue:45200", + "right": "issue:45325", + "accept": false, + "reason": "Both are Gemma/Qwen2.5-VL multimodal issues? No\u2014these are different models and different bugs." + }, + { + "left": "issue:43572", + "right": "issue:43957", + "accept": false, + "reason": "StableLm config field regression and meta-device loading breakage are separate issues." + }, + { + "left": "issue:43295", + "right": "issue:43450", + "accept": false, + "reason": "Processor.tokenizer regression is unrelated to batched video processor output shape." + }, + { + "left": "issue:43844", + "right": "issue:44936", + "accept": false, + "reason": "DeepSpeed/ZeRO gradient growth and trainer evaluate-after-train failure are different training issues." + }, + { + "left": "issue:43295", + "right": "issue:43873", + "accept": false, + "reason": "Custom processor API regression and quantization/offloading behavior are not the same bug." + }, + { + "left": "issue:43334", + "right": "issue:43525", + "accept": false, + "reason": "Missing pad_token_id on different configs/models does not indicate the same underlying defect." + }, + { + "left": "issue:44792", + "right": "issue:45563", + "accept": false, + "reason": "A failing Janus test and a stale generation warning are not the same change or bug." + }, + { + "left": "issue:43824", + "right": "issue:43975", + "accept": false, + "reason": "Importing a missing class and incorrect detokenization are unrelated." + }, + { + "left": "issue:43388", + "right": "issue:43526", + "accept": false, + "reason": "Both involve labels, but one is metric-gathering on tuple labels and the other is image-processor label reduction." + }, + { + "left": "issue:42915", + "right": "issue:43824", + "accept": false, + "reason": "FineGrainedFP8Config failure and Qwen2.5-VL import error are entirely different problems." + }, + { + "left": "issue:38175", + "right": "issue:45412", + "accept": false, + "reason": "SigLIP2 zero probabilities and RT-DETR memory not being released are unrelated." + }, + { + "left": "issue:38175", + "right": "issue:43493", + "accept": false, + "reason": "Same model family, but one reports zero probabilities and the other a JAX/HF implementation discrepancy; not enough to merge." + }, + { + "left": "issue:44514", + "right": "issue:44521", + "accept": false, + "reason": "Both concern apply_chat_template, but one is a batching crash and the other is all-zero assistant masks; different failure modes." + }, + { + "left": "issue:44442", + "right": "issue:45198", + "accept": false, + "reason": "FastSpeech2ConformerTokenizer loading failure is unrelated to Wav2Vec2 save_pretrained/tokenization failure." + }, + { + "left": "issue:28218", + "right": "issue:35141", + "accept": false, + "reason": "Tokenizer spacing after added tokens is unrelated to lm_head reinitialization after resizing embeddings." + }, + { + "left": "issue:43278", + "right": "issue:43856", + "accept": false, + "reason": "Embedding dtype drift in eval and Qwen3 MoE training memory inefficiency are different bugs." + }, + { + "left": "issue:42490", + "right": "issue:43493", + "accept": false, + "reason": "Local-path import warning and SigLIP2 JAX discrepancy are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 24, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:29127", + "issue:29942", + "issue:32090", + "issue:37428", + "issue:38175", + "issue:39692", + "issue:43262", + "issue:43278", + "issue:43335", + "issue:43425", + "issue:43450", + "issue:43504", + "issue:43526", + "issue:43550", + "issue:43575", + "issue:43606", + "issue:43611", + "issue:43653", + "issue:43819", + "issue:43825", + "issue:43827", + "issue:43901", + "issue:43931", + "issue:44112", + "issue:44206", + "issue:44351", + "issue:44360", + "issue:44448", + "issue:44493", + "issue:44561", + "issue:44610", + "issue:44743", + "issue:44779", + "issue:44855", + "issue:44857", + "issue:45083", + "issue:45127", + "issue:45200", + "issue:45245", + "issue:45290", + "issue:45310", + "issue:45381", + "issue:45478", + "issue:45584", + "issue:45588", + "issue:45593", + "issue:45823", + "issue:45859" + ], + "soft_pairs": [ + "issue:43653|issue:44112", + "issue:43450|issue:43575", + "issue:43827|issue:43901", + "issue:44206|issue:45584", + "issue:45127|issue:45245", + "issue:43262|issue:43450", + "issue:43819|issue:45083", + "issue:43819|issue:44112", + "issue:44610|issue:44857", + "issue:37428|issue:43425", + "issue:45290|issue:45381", + "issue:43335|issue:43550", + "issue:43606|issue:45083", + "issue:44448|issue:44743", + "issue:44857|issue:45593", + "issue:45200|issue:45588", + "issue:43278|issue:43825", + "issue:29942|issue:32090", + "issue:37428|issue:44351", + "issue:43611|issue:43931", + "issue:44206|issue:45823", + "issue:43335|issue:43653", + "issue:44360|issue:44561", + "issue:39692|issue:43526", + "issue:43819|issue:44493", + "issue:43550|issue:44112", + "issue:43335|issue:43504", + "issue:44855|issue:45083", + "issue:45310|issue:45478", + "issue:43335|issue:45083", + "issue:29127|issue:32090", + "issue:45588|issue:45859", + "issue:39692|issue:44493", + "issue:44448|issue:44779", + "issue:38175|issue:44493" + ], + "prepared_review_unit_hash": "fc7b3bd01d4dee7b7bf81726766eb50391f31ab17ace2788f06ce5ed616ebde0", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 35, + "serialized_chars": 38281, + "estimated_input_tokens": 9571, + "estimated_eval_tokens": 19398 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous; nearly all soft pairs are adjacent-but-distinct issues across different models, processors, or regressions. The only clear duplicate is the Qwen3.5-Moe `from_pretrained` error pair (45310/45478).", + "confidence": 0.95, + "canonical_issue_reason": "issue:45310 is the clearest canonical issue because it is the earlier, more complete report of the same Qwen3.5-Moe `from_pretrained` failure also described in issue:45478.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45310 is the best representative of the cluster: it is the only near-certain duplicate target and has the most direct, actionable bug description.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43653", + "right": "issue:44112", + "accept": false, + "reason": "Both are CI-related, but one is a tokenizer special-token decode bug and the other is a GraniteSpeech device-override test failure." + }, + { + "left": "issue:43450", + "right": "issue:43575", + "accept": false, + "reason": "Video processor batch-shape bug vs Qwen2-57B tensor-parallel OOM; different subsystems and failure modes." + }, + { + "left": "issue:43827", + "right": "issue:43901", + "accept": false, + "reason": "One is about v5 pipeline removal in docs for summarization/translation, the other is TextClassificationPipeline docs mentioning `return_all_scores`." + }, + { + "left": "issue:44206", + "right": "issue:45584", + "accept": false, + "reason": "Unsupported `center` arg in LasrFeatureExtractor is unrelated to Whisper empty-transcription failure after `align_special_tokens`." + }, + { + "left": "issue:45127", + "right": "issue:45245", + "accept": false, + "reason": "LoRA merge / tied embeddings collapse is unrelated to the category-count limit error." + }, + { + "left": "issue:43262", + "right": "issue:43450", + "accept": false, + "reason": "Audio chat-template sampling-rate default bug vs video processor batched-shape bug." + }, + { + "left": "issue:43819", + "right": "issue:45083", + "accept": false, + "reason": "DAC latent/forward mismatch is a different code path from the qwen3_omni_moe feature-length helper behavior." + }, + { + "left": "issue:43819", + "right": "issue:44112", + "accept": false, + "reason": "Different models and failures: DAC missing-STE mismatch vs GraniteSpeech CI device override." + }, + { + "left": "issue:44610", + "right": "issue:44857", + "accept": false, + "reason": "OmDet-Turbo input-size mismatch is unrelated to the LwDetrImageLoss float16/CUDA crash." + }, + { + "left": "issue:37428", + "right": "issue:43425", + "accept": false, + "reason": "ImportError for `_flash_supports_window_size` is not the same bug as Torch 2.10 incompatibility." + }, + { + "left": "issue:45290", + "right": "issue:45381", + "accept": false, + "reason": "Chat-template tool-call crash and Qwen2.5-VL video `vision_position_ids` bug affect different paths." + }, + { + "left": "issue:43335", + "right": "issue:43550", + "accept": false, + "reason": "SwitchTransformers sparse-layer construction bug vs Bamba torch.compile/SDPA failure." + }, + { + "left": "issue:43606", + "right": "issue:45083", + "accept": false, + "reason": "CPU offload device mismatch in `suno/bark-small` is unrelated to qwen3_omni_moe feature-length helper behavior." + }, + { + "left": "issue:44448", + "right": "issue:44743", + "accept": false, + "reason": "Pegasus v4/v5 output regression is unrelated to Qwen modular recurrent-state reset under cache." + }, + { + "left": "issue:44857", + "right": "issue:45593", + "accept": false, + "reason": "AMP/CUDA crash in LwDetrImageLoss is not the same as D-FINE skipping auxiliary losses when denoising is off." + }, + { + "left": "issue:45200", + "right": "issue:45588", + "accept": false, + "reason": "Gemma 4 `mm_token_type_ids` defaulting issue is distinct from the flash-attention `s_aux=None` crash." + }, + { + "left": "issue:43278", + "right": "issue:43825", + "accept": false, + "reason": "Training/evaluate embedding dtype drift is unrelated to the pipeline() v5 error-message regression." + }, + { + "left": "issue:29942", + "right": "issue:32090", + "accept": false, + "reason": "Flash Attention 2 test failures and `_gpu_broadcast_one` TypeError are separate failures in different areas." + }, + { + "left": "issue:37428", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they miss different symbols (`_flash_supports_window_size` vs `HybridCache`) and imply different fixes." + }, + { + "left": "issue:43611", + "right": "issue:43931", + "accept": false, + "reason": "`base_model_prefix` loading breakage and Qwen3-VL weight-shape mismatch are different model-loading bugs." + }, + { + "left": "issue:44206", + "right": "issue:45823", + "accept": false, + "reason": "LasrFeatureExtractor `center`-arg regression is unrelated to Gemma4 PLE device-mismatch during forward." + }, + { + "left": "issue:43335", + "right": "issue:43653", + "accept": false, + "reason": "Sparse-layer config bug and tokenizer special-token registration bug do not share the same underlying code path." + }, + { + "left": "issue:44360", + "right": "issue:44561", + "accept": false, + "reason": "DSA indexer missing ReLU is unrelated to removal of `is_torch_fx_available` breaking `trust_remote_code` models." + }, + { + "left": "issue:39692", + "right": "issue:43526", + "accept": false, + "reason": "SigLIP2 docs/model mismatch and BEiT fast processor `reduce_labels` bug are different issues." + }, + { + "left": "issue:43819", + "right": "issue:44493", + "accept": false, + "reason": "DAC forward mismatch is not the same as the widespread unexpected position-id key issue." + }, + { + "left": "issue:43550", + "right": "issue:44112", + "accept": false, + "reason": "Bamba compile/SDPA failure is unrelated to the stale GraniteSpeech device-override test." + }, + { + "left": "issue:43335", + "right": "issue:43504", + "accept": false, + "reason": "SwitchTransformers sparse-layer creation bug and BEiT semantic-segmentation legacy-field load failure are unrelated." + }, + { + "left": "issue:44855", + "right": "issue:45083", + "accept": false, + "reason": "Python 3.13 `IndentationError` in DebertaV2Model is unrelated to qwen3_omni_moe feature-length helper behavior." + }, + { + "left": "issue:45310", + "right": "issue:45478", + "accept": true, + "reason": "Same Qwen3.5-Moe `from_pretrained` failure; titles and symptoms match closely enough to be the same bug." + }, + { + "left": "issue:43335", + "right": "issue:45083", + "accept": false, + "reason": "Sparse-layer configuration and qwen3_omni_moe feature-length helper are unrelated." + }, + { + "left": "issue:29127", + "right": "issue:32090", + "accept": false, + "reason": "LayoutLMv3 error-message clarity issue is unrelated to the Trainer `_gpu_broadcast_one` TypeError." + }, + { + "left": "issue:45588", + "right": "issue:45859", + "accept": false, + "reason": "Flash-attention `s_aux=None` crash and Qwen3_5Moe CPU-offload incompatibility are different bugs." + }, + { + "left": "issue:39692", + "right": "issue:44493", + "accept": false, + "reason": "SigLIP2 docs/model mismatch is unrelated to the unexpected position-id key issue affecting many models." + }, + { + "left": "issue:44448", + "right": "issue:44779", + "accept": false, + "reason": "Pegasus output differences and Deepseek tokenizer regressions are distinct issues." + }, + { + "left": "issue:38175", + "right": "issue:44493", + "accept": false, + "reason": "SigLIP2 zero-probabilities bug is not the same as the generic unexpected position-id key regression." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The cluster summary is well grounded in the packet: the only clearly duplicated pair is issue:45310/issue:45478, which have identical titles and closely matching Qwen3.5-Moe from_pretrained reports. The rest of the soft-pair rejections are conservative and consistent with the issue titles/subsystems; no pair is overstated as a duplicate." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 25, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:28218", + "issue:29942", + "issue:32090", + "issue:33357", + "issue:34689", + "issue:35141", + "issue:37428", + "issue:40444", + "issue:42490", + "issue:43278", + "issue:43329", + "issue:43335", + "issue:43412", + "issue:43526", + "issue:43550", + "issue:43606", + "issue:43701", + "issue:43720", + "issue:43723", + "issue:43854", + "issue:43873", + "issue:43927", + "issue:44112", + "issue:44292", + "issue:44448", + "issue:44462", + "issue:44493", + "issue:44554", + "issue:44561", + "issue:44568", + "issue:44779", + "issue:44857", + "issue:44977", + "issue:44987", + "issue:45072", + "issue:45083", + "issue:45198", + "issue:45229", + "issue:45372", + "issue:45459", + "issue:45491", + "issue:45520", + "issue:45563", + "issue:45588", + "issue:45593", + "issue:45698", + "issue:45712", + "issue:45803" + ], + "soft_pairs": [ + "issue:43335|issue:45072", + "issue:43720|issue:44112", + "issue:43335|issue:45593", + "issue:44292|issue:45520", + "issue:44977|issue:45229", + "issue:42490|issue:43854", + "issue:44462|issue:44561", + "issue:44561|issue:44779", + "issue:40444|issue:43526", + "issue:34689|issue:35141", + "issue:43606|issue:44857", + "issue:44448|issue:44977", + "issue:44857|issue:45072", + "issue:43723|issue:44554", + "issue:28218|issue:32090", + "issue:43701|issue:45083", + "issue:43927|issue:45698", + "issue:44568|issue:45491", + "issue:43550|issue:45083", + "issue:45198|issue:45803", + "issue:29942|issue:35141", + "issue:37428|issue:44987", + "issue:43873|issue:44493", + "issue:32090|issue:33357", + "issue:45372|issue:45588", + "issue:43873|issue:45229", + "issue:45459|issue:45712", + "issue:43412|issue:45588", + "issue:44977|issue:45563", + "issue:43278|issue:43329" + ], + "prepared_review_unit_hash": "194161df7b29dacd7a9bcd748d24aabe08b1c993c426f57912a933f7763c536a", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36905, + "estimated_input_tokens": 9227, + "estimated_eval_tokens": 18710 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous: most items are unrelated tokenizer, model-loading, CI, attention, and serialization bugs. I found no soft pair that clearly describes the same underlying issue tightly enough to merge.", + "confidence": 0.91, + "canonical_issue_reason": "No single canonical issue stands out because the cluster is noisy and spans many unrelated bug classes; the items do not form one cohesive duplicate set.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43873 is the strongest standalone issue by activity/inbound references and has a clear, reproducible bug report, but it is still not a duplicate hub for the rest of the cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43335", + "right": "issue:45072", + "accept": false, + "reason": "Different bugs: SwitchTransformers sparse-layer config vs dtype mismatch in a different model path." + }, + { + "left": "issue:43720", + "right": "issue:44112", + "accept": false, + "reason": "Different failures: packed-weight loading vs a stale CI device-override test." + }, + { + "left": "issue:43335", + "right": "issue:45593", + "accept": false, + "reason": "Unrelated model bugs: SwitchTransformers config creation vs D-FINE auxiliary-loss handling." + }, + { + "left": "issue:44292", + "right": "issue:45520", + "accept": false, + "reason": "Both touch import/runtime environment issues, but one is NVFP4 model execution and the other is flash_attn import lookup on Python 3.13." + }, + { + "left": "issue:44977", + "right": "issue:45229", + "accept": false, + "reason": "Different concrete problems: flash-attention generation issue vs multi-GPU CUDA OOM." + }, + { + "left": "issue:42490", + "right": "issue:43854", + "accept": false, + "reason": "Local-path warning vs GLM-4.7-Flash model-loading test failure are not the same bug." + }, + { + "left": "issue:44462", + "right": "issue:44561", + "accept": false, + "reason": "Tokenizer repository loading vs v5 removal of is_torch_fx_available affect different code paths." + }, + { + "left": "issue:44561", + "right": "issue:44779", + "accept": false, + "reason": "Different regressions: trust_remote_code import breakage vs DeepSeek tokenizer output corruption." + }, + { + "left": "issue:40444", + "right": "issue:43526", + "accept": false, + "reason": "Both involve vision preprocessing, but one is iterable multi-image finetuning and the other is BEiT fast label reduction." + }, + { + "left": "issue:34689", + "right": "issue:35141", + "accept": false, + "reason": "Model loading regression vs embedding reinitialization after resize are separate issues." + }, + { + "left": "issue:43606", + "right": "issue:44857", + "accept": false, + "reason": "Device-mismatch offload bug and AMP/CUDA loss crash are different failure modes." + }, + { + "left": "issue:44448", + "right": "issue:44977", + "accept": false, + "reason": "Pegasus v4/v5 output drift is unrelated to the Qwen flash-attention generation bug." + }, + { + "left": "issue:44857", + "right": "issue:45072", + "accept": false, + "reason": "Different runtime paths: image loss AMP crash vs bfloat16 inference dtype mismatch." + }, + { + "left": "issue:43723", + "right": "issue:44554", + "accept": false, + "reason": "Tokenizer loading regression is unrelated to the MPS attention correctness issue." + }, + { + "left": "issue:28218", + "right": "issue:32090", + "accept": false, + "reason": "Tokenizer spacing bug vs Trainer broadcast TypeError are unrelated." + }, + { + "left": "issue:43701", + "right": "issue:45083", + "accept": false, + "reason": "Checkpoint key mismatch is not the same as the qwen3_omni_moe feature-length helper bug." + }, + { + "left": "issue:43927", + "right": "issue:45698", + "accept": false, + "reason": "Both involve save/load, but custom token IDs being lost is different from loading the wrong custom module." + }, + { + "left": "issue:44568", + "right": "issue:45491", + "accept": false, + "reason": "Special-token addition regression and Gemma3 NaN embeddings have different causes and code paths." + }, + { + "left": "issue:43550", + "right": "issue:45083", + "accept": false, + "reason": "torch.compile/SDPA failure is unrelated to the feature-length helper behavior bug." + }, + { + "left": "issue:45198", + "right": "issue:45803", + "accept": false, + "reason": "Wav2Vec2 save/tokenization failure vs bare-except swallowing errors in FuyuBatchFeature are not the same issue." + }, + { + "left": "issue:29942", + "right": "issue:35141", + "accept": false, + "reason": "Flash Attention 2 test failures are unrelated to embedding reinitialization after resize." + }, + { + "left": "issue:37428", + "right": "issue:44987", + "accept": false, + "reason": "Importing a removed flash-attn helper is not the same as a model loading failure for physical-intelligence/fast." + }, + { + "left": "issue:43873", + "right": "issue:44493", + "accept": false, + "reason": "Quantization/offloading behavior is unrelated to unexpected position-id keys." + }, + { + "left": "issue:32090", + "right": "issue:33357", + "accept": false, + "reason": "Trainer NoneType broadcast error and MacOS bus error with CLIP are different bugs." + }, + { + "left": "issue:45372", + "right": "issue:45588", + "accept": false, + "reason": "Gemma processor import breakage is unrelated to flash_attention.py crashing on s_aux=None." + }, + { + "left": "issue:43873", + "right": "issue:45229", + "accept": false, + "reason": "Quantization/offloading issue is not the same as Gemma4 multi-GPU OOM." + }, + { + "left": "issue:45459", + "right": "issue:45712", + "accept": false, + "reason": "Tokenizer error masking from a bare except is unrelated to leftover dummy class cleanup." + }, + { + "left": "issue:43412", + "right": "issue:45588", + "accept": false, + "reason": "Executorch export missing training guard/erfinv support is unrelated to flash_attention.py s_aux=None handling." + }, + { + "left": "issue:44977", + "right": "issue:45563", + "accept": false, + "reason": "Flash-attention generation correctness and a stale paged-generate warning are not the same bug." + }, + { + "left": "issue:43278", + "right": "issue:43329", + "accept": false, + "reason": "Training/evaluate dtype drift is unrelated to the undefined video-branch variables in _get_num_multimodal_tokens." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 26, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:10000", + "issue:15354", + "issue:25251", + "issue:28218", + "issue:29127", + "issue:29942", + "issue:30064", + "issue:32090", + "issue:41720", + "issue:42490", + "issue:42915", + "issue:42994", + "issue:43065", + "issue:43066", + "issue:43278", + "issue:43381", + "issue:43421", + "issue:43575", + "issue:43653", + "issue:43756", + "issue:43976", + "issue:43994", + "issue:44186", + "issue:44206", + "issue:44265", + "issue:44361", + "issue:44448", + "issue:44451", + "issue:44462", + "issue:44466", + "issue:44479", + "issue:44743", + "issue:44857", + "issue:44871", + "issue:44977", + "issue:45198", + "issue:45229", + "issue:45245", + "issue:45372", + "issue:45406", + "issue:45412", + "issue:45440", + "issue:45459", + "issue:45491", + "issue:45520", + "issue:45563", + "issue:45588", + "issue:45823" + ], + "soft_pairs": [ + "issue:41720|issue:45229", + "issue:10000|issue:15354", + "issue:43976|issue:45440", + "issue:42490|issue:43575", + "issue:45372|issue:45406", + "issue:28218|issue:29127", + "issue:43065|issue:43994", + "issue:45372|issue:45459", + "issue:44857|issue:45588", + "issue:43065|issue:45412", + "issue:44448|issue:44466", + "issue:45245|issue:45563", + "issue:45588|issue:45823", + "issue:43653|issue:44743", + "issue:43278|issue:43381", + "issue:25251|issue:32090", + "issue:44186|issue:45198", + "issue:44206|issue:44479", + "issue:43994|issue:45412", + "issue:43066|issue:44462", + "issue:44361|issue:45198", + "issue:44871|issue:44977", + "issue:44451|issue:45520", + "issue:44265|issue:45198", + "issue:42915|issue:42994", + "issue:43421|issue:45491", + "issue:43756|issue:45440", + "issue:29942|issue:30064" + ], + "prepared_review_unit_hash": "6f233c0a52c90b5595e33a8b04601fd6ae19d2f578eb1cefa5bf0ddc7fee0bf1", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36052, + "estimated_input_tokens": 9013, + "estimated_eval_tokens": 18282 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is highly heterogeneous; the soft pairs mostly share only broad subsystem overlap, not the same underlying bug or change. I would not merge any of the candidate pairs as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue stands out: the items span unrelated failures across tokenization, generation, quantization, vision, and model-loading paths.", + "canonical_pr_reason": null, + "best_issue_reason": "No strong global best issue for duplication canonicalization because there is no common root cause tying the set together.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41720", + "right": "issue:45229", + "accept": false, + "reason": "Different models and failure modes: device-mapping CUDA assert vs multi-GPU OOM." + }, + { + "left": "issue:10000", + "right": "issue:15354", + "accept": false, + "reason": "Unrelated model/export problems: DistilBERT issue vs torch.jit GeneratorExp scripting limitation." + }, + { + "left": "issue:43976", + "right": "issue:45440", + "accept": false, + "reason": "Different bug classes: Python-version compatibility vs DeepSeek implementation divergence." + }, + { + "left": "issue:42490", + "right": "issue:43575", + "accept": false, + "reason": "Local-path import warning and tensor-parallel OOM are unrelated." + }, + { + "left": "issue:45372", + "right": "issue:45406", + "accept": false, + "reason": "Both involve Gemma 4 loading, but one is a missing dependency symbol and the other is a missing processor attribute; not the same bug." + }, + { + "left": "issue:28218", + "right": "issue:29127", + "accept": false, + "reason": "Tokenizer spacing behavior vs LayoutLMv3 error-message clarity are different issues." + }, + { + "left": "issue:43065", + "right": "issue:43994", + "accept": false, + "reason": "Different subsystems and symptoms: dummy conv layer in Sam3PixelDecoder vs bad SigLIP2 outputs." + }, + { + "left": "issue:45372", + "right": "issue:45459", + "accept": false, + "reason": "One is an import dependency failure; the other is tokenizer error handling around protobuf." + }, + { + "left": "issue:44857", + "right": "issue:45588", + "accept": false, + "reason": "Different code paths: LW-DETR AMP crash vs flash-attention integration crash on sink-less models." + }, + { + "left": "issue:43065", + "right": "issue:45412", + "accept": false, + "reason": "Dummy layer definition and RT-DETR memory release are unrelated." + }, + { + "left": "issue:44448", + "right": "issue:44466", + "accept": false, + "reason": "Both are v5 regressions, but one is generation output mismatch and the other is serialization behavior; not the same defect." + }, + { + "left": "issue:45245", + "right": "issue:45563", + "accept": false, + "reason": "Category-limit runtime error vs stale warning in paged generate are unrelated." + }, + { + "left": "issue:45588", + "right": "issue:45823", + "accept": false, + "reason": "Different Gemma4 issues: flash-attention s_aux handling vs device mismatch in forward." + }, + { + "left": "issue:43653", + "right": "issue:44743", + "accept": false, + "reason": "Tokenizer special-token registration vs Qwen3.5 recurrent-state cache handling are unrelated." + }, + { + "left": "issue:43278", + "right": "issue:43381", + "accept": false, + "reason": "Eval-mode dtype change and checkpointing-in-eval are separate problems." + }, + { + "left": "issue:25251", + "right": "issue:32090", + "accept": false, + "reason": "Pipeline output nesting and Trainer broadcast NoneType error are different failures." + }, + { + "left": "issue:44186", + "right": "issue:45198", + "accept": false, + "reason": "Different tokenizers/models and different failure surfaces: NER padding/truncation crash vs save_pretrained/tokenization failure." + }, + { + "left": "issue:44206", + "right": "issue:44479", + "accept": false, + "reason": "Different components and regressions: feature extractor center arg vs video input regression in Qwen VL models." + }, + { + "left": "issue:43994", + "right": "issue:45412", + "accept": false, + "reason": "SigLIP2 inference corruption and RT-DETR memory leak are unrelated." + }, + { + "left": "issue:43066", + "right": "issue:44462", + "accept": false, + "reason": "Wrong decoder type and ignoring tokenizer.json are both tokenizer-related but not the same underlying bug." + }, + { + "left": "issue:44361", + "right": "issue:45198", + "accept": false, + "reason": "Different tokenizer/model paths and different errors; no shared root cause." + }, + { + "left": "issue:44871", + "right": "issue:44977", + "accept": false, + "reason": "EOS-token config inconsistency and flash-attention generation regression are unrelated." + }, + { + "left": "issue:44451", + "right": "issue:45520", + "accept": false, + "reason": "Model loading failure and Python 3.13 flash_attn import KeyError are different problems." + }, + { + "left": "issue:44265", + "right": "issue:45198", + "accept": false, + "reason": "torch.export check failure is unrelated to Wav2Vec2 save/tokenization failure." + }, + { + "left": "issue:42915", + "right": "issue:42994", + "accept": false, + "reason": "Both involve quantization, but one is FP8 config failure and the other is quantized-model saving." + }, + { + "left": "issue:43421", + "right": "issue:45491", + "accept": false, + "reason": "Runtime post-processor update and Gemma3 NaN embeddings are unrelated." + }, + { + "left": "issue:43756", + "right": "issue:45440", + "accept": false, + "reason": "Different model families and defects: RoPE-layer dropping vs DeepSeekV3MoE divergence." + }, + { + "left": "issue:29942", + "right": "issue:30064", + "accept": false, + "reason": "FlashAttention 2 test failures and void segmentation-map processing are different issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 27, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:29127", + "issue:29942", + "issue:32090", + "issue:33290", + "issue:33357", + "issue:34567", + "issue:35141", + "issue:36010", + "issue:39692", + "issue:42490", + "issue:42915", + "issue:43257", + "issue:43262", + "issue:43278", + "issue:43334", + "issue:43381", + "issue:43526", + "issue:43531", + "issue:43575", + "issue:43653", + "issue:43701", + "issue:43761", + "issue:43824", + "issue:43873", + "issue:43931", + "issue:44186", + "issue:44206", + "issue:44279", + "issue:44442", + "issue:44464", + "issue:44610", + "issue:44625", + "issue:44743", + "issue:44811", + "issue:44857", + "issue:44977", + "issue:45072", + "issue:45103", + "issue:45245", + "issue:45372", + "issue:45412", + "issue:45440", + "issue:45491", + "issue:45584", + "issue:45593", + "issue:45834", + "issue:45859" + ], + "soft_pairs": [ + "issue:33357|issue:39692", + "issue:43257|issue:43278", + "issue:29127|issue:34567", + "issue:42490|issue:43873", + "issue:32090|issue:33290", + "issue:44610|issue:45412", + "issue:45440|issue:45593", + "issue:43824|issue:44279", + "issue:43653|issue:44442", + "issue:43381|issue:43701", + "issue:43761|issue:45859", + "issue:43526|issue:45834", + "issue:45103|issue:45372", + "issue:43526|issue:44186", + "issue:29942|issue:36010", + "issue:29127|issue:33357", + "issue:44625|issue:45584", + "issue:29127|issue:35141", + "issue:43334|issue:43531", + "issue:44464|issue:45491", + "issue:44977|issue:45245", + "issue:44442|issue:45072", + "issue:42915|issue:43931", + "issue:43262|issue:44206", + "issue:43526|issue:44857", + "issue:42490|issue:43381", + "issue:43575|issue:43701", + "issue:44743|issue:44811" + ], + "prepared_review_unit_hash": "e40b762a77f3344e947354ce1e0b3fa67456a0570c3557061e8a8a2dedd556fe", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35482, + "estimated_input_tokens": 8871, + "estimated_eval_tokens": 17998 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is heterogeneous: the soft pairs share only broad themes (Qwen, processors, loading, training) but not the same concrete bug or change. I would not merge any of these pairs as duplicates.", + "confidence": 0.98, + "canonical_issue_reason": "No canonical issue: the items are mostly unrelated bugs across different models, processors, and trainer/runtime paths, so there is no clear duplicate hub.", + "canonical_pr_reason": null, + "best_issue_reason": "No single issue is a good global representative; issue 43824 is well-scoped and concrete, but the cluster is too heterogeneous for it to serve as a canonical anchor.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:33357", + "right": "issue:39692", + "accept": false, + "reason": "Both involve multimodal/model-loading problems, but one is a MacOS bus error for a community CLIP model and the other is a documentation example/model-processor mismatch with quantization failure; not the same bug." + }, + { + "left": "issue:43257", + "right": "issue:43278", + "accept": false, + "reason": "Different failure modes: MOE weights not converted under accelerate+deepspeed versus embedding dtype drifting from BF16 to FP32 during evaluation." + }, + { + "left": "issue:29127", + "right": "issue:34567", + "accept": false, + "reason": "LayoutLMv3 box-info error messaging and TrainerState token counting are unrelated code paths." + }, + { + "left": "issue:42490", + "right": "issue:43873", + "accept": false, + "reason": "Trailing-slash import warning for local models is unrelated to quantization/offloading behavior." + }, + { + "left": "issue:32090", + "right": "issue:33290", + "accept": false, + "reason": "Trainer NoneType broadcast failure and AdaFactor/deepspeed OOM are distinct runtime issues." + }, + { + "left": "issue:44610", + "right": "issue:45412", + "accept": false, + "reason": "Processor/model input-size mismatch for OmDet-Turbo is unrelated to RT-DETR memory release on deletion." + }, + { + "left": "issue:45440", + "right": "issue:45593", + "accept": false, + "reason": "DeepseekV3MoE implementation divergence and D-FINE auxiliary-loss behavior are different model bugs." + }, + { + "left": "issue:43824", + "right": "issue:44279", + "accept": false, + "reason": "Specific missing Qwen2_5_VLForConditionalGeneration import is not the same as a generic dependency issue." + }, + { + "left": "issue:43653", + "right": "issue:44442", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and FastSpeech2ConformerTokenizer load failure are different tokenizer bugs." + }, + { + "left": "issue:43381", + "right": "issue:43701", + "accept": false, + "reason": "Gradient checkpointing in eval mode and resume_from_checkpoint key mismatch are unrelated trainer issues." + }, + { + "left": "issue:43761", + "right": "issue:45859", + "accept": false, + "reason": "CLIPVisionModel hidden_states regression and Qwen3_5MoeTextRotaryEmbedding CPU offload incompatibility do not share the same code path." + }, + { + "left": "issue:43526", + "right": "issue:45834", + "accept": false, + "reason": "BeitImageProcessorFast label reduction bug and Kosmos2.5 OCR long-input index error are unrelated." + }, + { + "left": "issue:45103", + "right": "issue:45372", + "accept": false, + "reason": "Auto-docstring annotation crash and Gemma 4 processor import failure are unrelated." + }, + { + "left": "issue:43526", + "right": "issue:44186", + "accept": false, + "reason": "Both are processor/tokenization bugs, but one is BeitImageProcessorFast label reduction and the other is LayoutLMv2Tokenizer NER/padding crashes; different implementations and symptoms." + }, + { + "left": "issue:29942", + "right": "issue:36010", + "accept": false, + "reason": "Flash Attention 2 test failures and GenerationMixin import errors are unrelated." + }, + { + "left": "issue:29127", + "right": "issue:33357", + "accept": false, + "reason": "LayoutLMv3 error-message clarity and CLIP MacOS bus error are unrelated." + }, + { + "left": "issue:44625", + "right": "issue:45584", + "accept": false, + "reason": "Qwen3.5 num_labels propagation and Whisper empty-transcription generation failure are different bugs despite both touching processing/output behavior." + }, + { + "left": "issue:29127", + "right": "issue:35141", + "accept": false, + "reason": "A diagnostic message issue in LayoutLMv3 is unrelated to embedding reinitialization after resize in post_init." + }, + { + "left": "issue:43334", + "right": "issue:43531", + "accept": false, + "reason": "Missing pad_token_id in Qwen3-VL text config and sliding_window behavior in Qwen3-MoE are different configuration bugs." + }, + { + "left": "issue:44464", + "right": "issue:45491", + "accept": false, + "reason": "Compiled-forward chunked generation inconsistency and Gemma3 NaN embeddings in mixed-length batches are distinct numerical/attention issues." + }, + { + "left": "issue:44977", + "right": "issue:45245", + "accept": false, + "reason": "Qwen3.5 flash-attention generation problems and the 2^24 category limit runtime error are unrelated." + }, + { + "left": "issue:44442", + "right": "issue:45072", + "accept": false, + "reason": "Tokenizer loading failure and bfloat16 dtype mismatch in inference are different problems." + }, + { + "left": "issue:42915", + "right": "issue:43931", + "accept": false, + "reason": "Both are Qwen loading issues, but one is a FineGrainedFP8Config failure and the other is a weight-shape mismatch for Qwen3-VL; not the same concrete bug." + }, + { + "left": "issue:43262", + "right": "issue:44206", + "accept": false, + "reason": "Audio chat-template sample-rate defaulting and LasrFeatureExtractor center-arg crash are separate audio-processing bugs." + }, + { + "left": "issue:43526", + "right": "issue:44857", + "accept": false, + "reason": "BeitImageProcessorFast reduce_labels and LwDetrImageLoss AMP/CUDA crash do not overlap." + }, + { + "left": "issue:42490", + "right": "issue:43381", + "accept": false, + "reason": "Import warning for local paths and gradient checkpointing eval-mode restrictions are unrelated." + }, + { + "left": "issue:43575", + "right": "issue:43701", + "accept": false, + "reason": "Tensor-parallel OOM for a Qwen2-57B model and checkpoint key mismatch are unrelated." + }, + { + "left": "issue:44743", + "right": "issue:44811", + "accept": false, + "reason": "Qwen3_5 recurrent-state reset logic and Whisper batch_decode skip_special_tokens handling are different model-specific bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 28, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:33290", + "issue:34567", + "issue:34689", + "issue:35141", + "issue:36010", + "issue:41628", + "issue:42490", + "issue:42907", + "issue:43232", + "issue:43316", + "issue:43381", + "issue:43388", + "issue:43441", + "issue:43493", + "issue:43525", + "issue:43575", + "issue:43611", + "issue:43643", + "issue:43646", + "issue:43720", + "issue:43723", + "issue:43749", + "issue:43761", + "issue:43827", + "issue:43873", + "issue:43931", + "issue:44062", + "issue:44077", + "issue:44079", + "issue:44188", + "issue:44387", + "issue:44442", + "issue:44479", + "issue:44743", + "issue:44871", + "issue:44898", + "issue:44913", + "issue:45003", + "issue:45216", + "issue:45290", + "issue:45372", + "issue:45440", + "issue:45479", + "issue:45588", + "issue:45698", + "issue:45759", + "issue:45803", + "issue:45907" + ], + "soft_pairs": [ + "issue:44062|issue:45907", + "issue:43232|issue:43388", + "issue:43575|issue:44387", + "issue:43646|issue:44077", + "issue:43761|issue:44079", + "issue:43441|issue:45588", + "issue:45003|issue:45440", + "issue:44442|issue:45803", + "issue:43723|issue:45479", + "issue:44479|issue:45290", + "issue:43761|issue:44188", + "issue:43316|issue:43525", + "issue:44442|issue:44898", + "issue:33290|issue:34567", + "issue:42490|issue:44743", + "issue:42907|issue:45698", + "issue:34689|issue:43611", + "issue:33290|issue:35141", + "issue:33290|issue:34689", + "issue:43381|issue:43873", + "issue:43931|issue:45216", + "issue:43493|issue:43827", + "issue:44077|issue:44479", + "issue:43720|issue:44442", + "issue:43316|issue:45759", + "issue:41628|issue:45372", + "issue:43643|issue:44913", + "issue:43749|issue:43931", + "issue:34689|issue:36010", + "issue:43873|issue:44871" + ], + "prepared_review_unit_hash": "bb46484e2bdeaf964227c26dfcef48cda6a3910fcd4345ed298691ab852d27d6", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36916, + "estimated_input_tokens": 9229, + "estimated_eval_tokens": 18714 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: the items cover unrelated regressions, API breaks, loading/saving bugs, and docs issues. No soft pair looks like the same underlying bug/change, so all candidate merges are rejected.", + "confidence": 0.87, + "canonical_issue_reason": "No true duplicate cluster is evident; if a representative issue is needed, issue 43873 is the strongest signal item because it is open, broadly user-facing, and has the most inbound references/discussion.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43873 is the best standalone representative: active, high-impact, and clearly a real regression with multiple references.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44062", + "right": "issue:45907", + "accept": false, + "reason": "Different bugs: tokenizer AddedToken argument collision vs a repo file listing API regression." + }, + { + "left": "issue:43232", + "right": "issue:43388", + "accept": false, + "reason": "Different code paths: generation kwarg sync issue vs metric gathering dropping tuple labels." + }, + { + "left": "issue:43575", + "right": "issue:44387", + "accept": false, + "reason": "Both mention OOM, but one is tensor-parallel model loading and the other is int4 quantization reserved-memory growth." + }, + { + "left": "issue:43646", + "right": "issue:44077", + "accept": false, + "reason": "Related to initialization, but one is a broad v5 custom model init break and the other is a specific `patchtsmixer` post_init policy issue." + }, + { + "left": "issue:43761", + "right": "issue:44079", + "accept": false, + "reason": "One is CLIPVision hidden_states handling; the other is `ModelOutput` key assignment semantics." + }, + { + "left": "issue:43441", + "right": "issue:45588", + "accept": false, + "reason": "Both touch flash attention, but they fail in different models and different branches of the integration code." + }, + { + "left": "issue:45003", + "right": "issue:45440", + "accept": false, + "reason": "Unrelated: unsafe `sys.modules` access vs a DeepSeekV3 implementation divergence." + }, + { + "left": "issue:44442", + "right": "issue:45803", + "accept": false, + "reason": "Tokenizer loading failure is unrelated to a bare `except` in Fuyu tensor conversion." + }, + { + "left": "issue:43723", + "right": "issue:45479", + "accept": false, + "reason": "Tokenizer loading in v5 is unrelated to a sequence-classification zero-loss bug." + }, + { + "left": "issue:44479", + "right": "issue:45290", + "accept": false, + "reason": "Different regressions: multimodal video input handling vs chat template tool-call tokenization." + }, + { + "left": "issue:43761", + "right": "issue:44188", + "accept": false, + "reason": "CLIPVision hidden_states bug is unrelated to attention-kernel divergence under `torch.compile`." + }, + { + "left": "issue:43316", + "right": "issue:43525", + "accept": false, + "reason": "Both are config/API discrepancies, but they concern different models and different missing fields." + }, + { + "left": "issue:44442", + "right": "issue:44898", + "accept": false, + "reason": "Tokenizer auto-loading failure is unrelated to Perceiver image-resolution interpolation failure." + }, + { + "left": "issue:33290", + "right": "issue:34567", + "accept": false, + "reason": "OOM in DeepSpeed Adafactor is unrelated to `num_input_tokens_seen` not updating." + }, + { + "left": "issue:42490", + "right": "issue:44743", + "accept": false, + "reason": "Local-path import warning is unrelated to recurrent state reset in Qwen3.5 cache usage." + }, + { + "left": "issue:42907", + "right": "issue:45698", + "accept": false, + "reason": "Saving dequantized models is different from loading the wrong custom module after save/load." + }, + { + "left": "issue:34689", + "right": "issue:43611", + "accept": false, + "reason": "Both are model-loading regressions, but they are version- and cause-specific and not the same bug." + }, + { + "left": "issue:33290", + "right": "issue:35141", + "accept": false, + "reason": "OOM from optimizer usage is unrelated to embedding reinitialization during `post_init`." + }, + { + "left": "issue:33290", + "right": "issue:34689", + "accept": false, + "reason": "Different failure modes: training OOM vs model-loading breakage for a specific Llama vision model." + }, + { + "left": "issue:43381", + "right": "issue:43873", + "accept": false, + "reason": "Gradient checkpointing eval-mode behavior is unrelated to quantization offloading." + }, + { + "left": "issue:43931", + "right": "issue:45216", + "accept": false, + "reason": "Weight-shape mismatch on load is different from a bad checkpoint produced by `save_pretrained`." + }, + { + "left": "issue:43493", + "right": "issue:43827", + "accept": false, + "reason": "Model implementation discrepancy is unrelated to docs still referencing removed pipeline APIs." + }, + { + "left": "issue:44077", + "right": "issue:44479", + "accept": false, + "reason": "A `post_init` policy issue is unrelated to the video-input regression in Qwen VL models." + }, + { + "left": "issue:43720", + "right": "issue:44442", + "accept": false, + "reason": "BitNet accelerate-loading unpacking bug is unrelated to FastSpeech2ConformerTokenizer auto-loading." + }, + { + "left": "issue:43316", + "right": "issue:45759", + "accept": false, + "reason": "Gemma3TextConfig API mismatch is unrelated to `from_config` not unwrapping `text_config` for Qwen multimodal configs." + }, + { + "left": "issue:41628", + "right": "issue:45372", + "accept": false, + "reason": "Both are ImportErrors, but they involve different imports, packages, and loading paths." + }, + { + "left": "issue:43643", + "right": "issue:44913", + "accept": false, + "reason": "Missing fields with `trust_remote_code` is distinct from `rotary_pct` not persisting on reload." + }, + { + "left": "issue:43749", + "right": "issue:43931", + "accept": false, + "reason": "FSDP CPU RAM efficient loading failure is unrelated to a Qwen3-VL shape mismatch on model load." + }, + { + "left": "issue:34689", + "right": "issue:36010", + "accept": false, + "reason": "Different import/load regressions: a specific Llama vision loading break vs missing `GenerationMixin` import." + }, + { + "left": "issue:43873", + "right": "issue:44871", + "accept": false, + "reason": "Quantization/offloading behavior is unrelated to an EOS token config inconsistency in Gemma-3." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 29, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:16998", + "issue:22355", + "issue:25251", + "issue:28282", + "issue:33290", + "issue:36331", + "issue:41950", + "issue:42175", + "issue:42617", + "issue:43065", + "issue:43295", + "issue:43316", + "issue:43388", + "issue:43404", + "issue:43441", + "issue:43452", + "issue:43475", + "issue:43504", + "issue:43540", + "issue:43550", + "issue:43575", + "issue:43606", + "issue:43646", + "issue:43673", + "issue:43688", + "issue:43701", + "issue:43742", + "issue:43761", + "issue:43825", + "issue:44077", + "issue:44162", + "issue:44190", + "issue:44479", + "issue:44496", + "issue:44514", + "issue:44534", + "issue:44617", + "issue:44743", + "issue:44792", + "issue:44855", + "issue:44871", + "issue:44877", + "issue:44913", + "issue:45127", + "issue:45137", + "issue:45200", + "issue:45584", + "issue:45698" + ], + "soft_pairs": [ + "issue:33290|issue:36331", + "issue:44190|issue:44496", + "issue:44855|issue:45137", + "issue:43540|issue:44514", + "issue:43316|issue:44877", + "issue:43065|issue:44617", + "issue:42175|issue:43316", + "issue:25251|issue:28282", + "issue:43295|issue:43825", + "issue:43404|issue:43441", + "issue:43673|issue:44743", + "issue:44162|issue:44792", + "issue:43388|issue:43688", + "issue:44077|issue:44534", + "issue:16998|issue:22355", + "issue:43550|issue:43701", + "issue:43606|issue:43701", + "issue:44871|issue:45584", + "issue:41950|issue:43475", + "issue:43761|issue:44479", + "issue:43295|issue:44871", + "issue:45127|issue:45200", + "issue:42617|issue:43575", + "issue:43575|issue:43742", + "issue:43452|issue:44496", + "issue:43646|issue:44913", + "issue:43504|issue:45698" + ], + "prepared_review_unit_hash": "fc2ffce13f45d2c39d624aa0b57ecbff5bc9b677a417635fa5daaec6ea81a59f", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 35676, + "estimated_input_tokens": 8919, + "estimated_eval_tokens": 18094 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These items are mostly unrelated standalone issues across training, loading, generation, multimodal, and config code paths. None of the soft pairs look like true duplicates of the same underlying bug or change.", + "confidence": 0.91, + "canonical_issue_reason": "issue:45698 is the strongest representative issue here: it is open, has active discussion, and describes a concrete, user-facing from_pretrained/save_pretrained loading regression with a clear reproduction path.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45698 is the best issue candidate because it targets a specific core model-loading path, has ongoing activity, and is actionable for triage/repair.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:33290", + "right": "issue:36331", + "accept": false, + "reason": "Different failures: DeepSpeed/Adafactor OOM versus CustomTrainer signature mismatch." + }, + { + "left": "issue:44190", + "right": "issue:44496", + "accept": false, + "reason": "Unrelated bugs: local dataset loading in an example script versus model config/type resolution." + }, + { + "left": "issue:44855", + "right": "issue:45137", + "accept": false, + "reason": "Different code paths and symptoms: Python 3.13 import/parsing failure versus ZeRO3 deque error." + }, + { + "left": "issue:43540", + "right": "issue:44514", + "accept": false, + "reason": "Both involve video/multimodal inputs, but one is Qwen3OmniMoe video processing and the other is batched chat templating padding behavior." + }, + { + "left": "issue:43316", + "right": "issue:44877", + "accept": false, + "reason": "Both touch config APIs, but one is a Gemma3TextConfig discrepancy and the other is strict loading of granite_speech config; not the same bug." + }, + { + "left": "issue:43065", + "right": "issue:44617", + "accept": false, + "reason": "Same model family prefix, but one is a dummy Conv2d in Sam3PixelDecoder and the other is CUDA OOM in Sam3Video." + }, + { + "left": "issue:42175", + "right": "issue:43316", + "accept": false, + "reason": "Different subsystems: backend extra installation versus model config API mismatch." + }, + { + "left": "issue:25251", + "right": "issue:28282", + "accept": false, + "reason": "Pipeline output-shape change and missing PyTorch import are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:43825", + "accept": false, + "reason": "Both are regressions around processors/pipelines, but one is custom tokenizer/processor breakage and the other is an incorrect error message." + }, + { + "left": "issue:43404", + "right": "issue:43441", + "accept": false, + "reason": "Different bugs in nearby model families: weight tying versus FlashAttention failure." + }, + { + "left": "issue:43673", + "right": "issue:44743", + "accept": false, + "reason": "Both mention cache/state, but they affect different models and different generation mechanisms." + }, + { + "left": "issue:44162", + "right": "issue:44792", + "accept": false, + "reason": "ESM2 model breakage and a janus test failure are not the same issue." + }, + { + "left": "issue:43388", + "right": "issue:43688", + "accept": false, + "reason": "Different training/eval bugs: gather_for_metrics tuple truncation versus auxiliary-loss normalization." + }, + { + "left": "issue:44077", + "right": "issue:44534", + "accept": false, + "reason": "Optional post_init enforcement and non-persistent buffer initialization are separate issues." + }, + { + "left": "issue:16998", + "right": "issue:22355", + "accept": false, + "reason": "Model max length question for DeBERTa-V3 is unrelated to missing transformers.onnx import." + }, + { + "left": "issue:43550", + "right": "issue:43701", + "accept": false, + "reason": "torch.compile/SDPA OOM and resume_from_checkpoint key mismatch are unrelated." + }, + { + "left": "issue:43606", + "right": "issue:43701", + "accept": false, + "reason": "CPU offload device mismatch is a different failure mode from checkpoint key mismatch." + }, + { + "left": "issue:44871", + "right": "issue:45584", + "accept": false, + "reason": "Both involve token/generation settings, but one is eos_token_id config inconsistency and the other is Whisper empty-transcription handling." + }, + { + "left": "issue:41950", + "right": "issue:43475", + "accept": false, + "reason": "Pipeline processor lookup bug and Sam3 vision encoder attribute error are different problems." + }, + { + "left": "issue:43761", + "right": "issue:44479", + "accept": false, + "reason": "CLIP hidden_states regression and Qwen video-input regression are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:44871", + "accept": false, + "reason": "Custom processor/tokenizer regression and Gemma eos_token_id mismatch do not share the same concrete code path." + }, + { + "left": "issue:45127", + "right": "issue:45200", + "accept": false, + "reason": "LoRA merge/collapse with tied embeddings is unrelated to Gemma 4 mm_token_type_ids defaults." + }, + { + "left": "issue:42617", + "right": "issue:43575", + "accept": false, + "reason": "3d_parallel script failure and tensor-parallel OOM are different issues." + }, + { + "left": "issue:43575", + "right": "issue:43742", + "accept": false, + "reason": "OOM when loading a large model with TP is not the same as a MobileLLM config KeyError." + }, + { + "left": "issue:43452", + "right": "issue:44496", + "accept": false, + "reason": "GGUF tokenizer/model loading breakage and missing model_type in config.json are distinct loading errors." + }, + { + "left": "issue:43646", + "right": "issue:44913", + "accept": false, + "reason": "Transformer 5 custom initialization breakage and GPTNeoX rotary_pct reload persistence are different bugs." + }, + { + "left": "issue:43504", + "right": "issue:45698", + "accept": false, + "reason": "Legacy-field preset loading for Beit semantic segmentation is not the same as loading the wrong custom module after save_pretrained." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 30, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:32090", + "issue:35141", + "issue:41720", + "issue:42490", + "issue:43065", + "issue:43335", + "issue:43388", + "issue:43644", + "issue:43761", + "issue:43827", + "issue:43906", + "issue:43931", + "issue:43976", + "issue:43994", + "issue:44060", + "issue:44077", + "issue:44079", + "issue:44190", + "issue:44242", + "issue:44462", + "issue:44464", + "issue:44479", + "issue:44521", + "issue:44534", + "issue:44610", + "issue:44617", + "issue:44734", + "issue:44737", + "issue:44792", + "issue:44855", + "issue:44871", + "issue:44913", + "issue:44933", + "issue:44964", + "issue:44977", + "issue:45127", + "issue:45245", + "issue:45290", + "issue:45357", + "issue:45362", + "issue:45381", + "issue:45406", + "issue:45563", + "issue:45584", + "issue:45588", + "issue:45600", + "issue:45803", + "issue:45907" + ], + "soft_pairs": [ + "issue:44734|issue:45406", + "issue:44617|issue:45584", + "issue:43761|issue:44242", + "issue:32090|issue:35141", + "issue:43931|issue:44060", + "issue:45563|issue:45584", + "issue:44464|issue:44871", + "issue:44479|issue:44734", + "issue:43065|issue:44610", + "issue:43906|issue:45290", + "issue:44521|issue:45381", + "issue:44792|issue:44977", + "issue:43827|issue:43994", + "issue:44464|issue:45563", + "issue:44534|issue:44913", + "issue:44190|issue:44964", + "issue:41720|issue:45245", + "issue:44933|issue:45362", + "issue:44792|issue:45600", + "issue:44855|issue:45588", + "issue:35141|issue:42490", + "issue:43388|issue:44079", + "issue:45127|issue:45357", + "issue:44737|issue:45803", + "issue:44462|issue:45907", + "issue:44077|issue:44079", + "issue:43335|issue:44855", + "issue:43644|issue:44913", + "issue:43976|issue:44855" + ], + "prepared_review_unit_hash": "21ca06a3dd72739437b53f6c63248c4b1cbd8b2602d67cf17c14cd52af5ae3f7", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36609, + "estimated_input_tokens": 9153, + "estimated_eval_tokens": 18562 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a set of unrelated bug reports that only share broad transformer/model-loading themes. I would anchor the cluster on the v5 non-persistent-buffer regression report, and reject all listed soft duplicate links as too semantically distant or too underspecified to merge safely.", + "confidence": 0.94, + "canonical_issue_reason": "issue:43644 is the clearest umbrella report for a concrete v5 regression affecting non-persistent buffers; it is broad, specific, and central enough to serve as the representative issue.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43644 is the strongest standalone representative because it describes a concrete framework-wide regression rather than a model-specific symptom or a downstream effect.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44734", + "right": "issue:45406", + "accept": false, + "reason": "Different failures: KV-cache continuation tensor indexing in serve vs a missing _tokenizer attribute in Gemma4Processor." + }, + { + "left": "issue:44617", + "right": "issue:45584", + "accept": false, + "reason": "Unrelated models and symptoms: Sam3Video CUDA OOM vs Whisper empty transcription after align_special_tokens." + }, + { + "left": "issue:43761", + "right": "issue:44242", + "accept": false, + "reason": "One is CLIPVision hidden_states being None; the other is router/load-balancing loss omission in a MoE model. Not the same bug." + }, + { + "left": "issue:32090", + "right": "issue:35141", + "accept": false, + "reason": "Trainer GPU broadcast TypeError is unrelated to embedding resizing/post_init reinitialization." + }, + { + "left": "issue:43931", + "right": "issue:44060", + "accept": false, + "reason": "Weight-shape mismatch while loading Qwen3-VL is not the same as a tied-weights warning incorrectly linking dt_bias." + }, + { + "left": "issue:45563", + "right": "issue:45584", + "accept": false, + "reason": "A stale warning about num_return_sequences is not the same as Whisper failing on empty transcription." + }, + { + "left": "issue:44464", + "right": "issue:44871", + "accept": false, + "reason": "Chunked generation with compiled forward and Gemma eos_token_id config mismatch are different problems." + }, + { + "left": "issue:44479", + "right": "issue:44734", + "accept": false, + "reason": "Both involve generation, but one is a video-input regression in Qwen VL models and the other is a serve KV-cache continuation crash." + }, + { + "left": "issue:43065", + "right": "issue:44610", + "accept": false, + "reason": "A dummy Conv2d in Sam3PixelDecoder is not the same as OmDet-Turbo processor/model image-size mismatch." + }, + { + "left": "issue:43906", + "right": "issue:45290", + "accept": false, + "reason": "The reproduction issue lacks enough concrete overlap; apply_chat_template tool-call crashes are a different failure mode." + }, + { + "left": "issue:44521", + "right": "issue:45381", + "accept": false, + "reason": "All-zero assistant masks in apply_chat_template are distinct from wrong vision_position_ids in Qwen2.5-VL video input." + }, + { + "left": "issue:44792", + "right": "issue:44977", + "accept": false, + "reason": "Janus image-generation test failure and Qwen3.5 flash-attention generation issues are not the same code-path bug." + }, + { + "left": "issue:43827", + "right": "issue:43994", + "accept": false, + "reason": "Docs still referencing pipeline() is unrelated to SigLIP2 producing bad outputs under AutoModel/pipeline." + }, + { + "left": "issue:44464", + "right": "issue:45563", + "accept": false, + "reason": "Compiled-forward chunked generation inconsistency is not the same as a stale warning about num_return_sequences." + }, + { + "left": "issue:44534", + "right": "issue:44913", + "accept": false, + "reason": "A v5 non-persistent-buffer junk regression is unrelated to GPTNeoX rotary_pct not persisting on reload." + }, + { + "left": "issue:44190", + "right": "issue:44964", + "accept": false, + "reason": "Local dataset loading in a no-trainer script is not the same as loading a multimodal model failing in latest transformers." + }, + { + "left": "issue:41720", + "right": "issue:45245", + "accept": false, + "reason": "Qwen3 auto device-mapping cudaErrorAssert is unrelated to the 2^24 category-count limit." + }, + { + "left": "issue:44933", + "right": "issue:45362", + "accept": false, + "reason": "A missing import from image_utils is unrelated to Qwen3.5 chat crashes." + }, + { + "left": "issue:44792", + "right": "issue:45600", + "accept": false, + "reason": "Janus image-generation test failure and stale auto_mappings references to removed configs are different issues." + }, + { + "left": "issue:44855", + "right": "issue:45588", + "accept": false, + "reason": "Python 3.13 IndentationError in DebertaV2Model is unrelated to flash_attention crashing on s_aux=None." + }, + { + "left": "issue:35141", + "right": "issue:42490", + "accept": false, + "reason": "Embedding reinitialization in post_init is unrelated to a local-path trailing-slash import warning." + }, + { + "left": "issue:43388", + "right": "issue:44079", + "accept": false, + "reason": "A gather_for_metrics truncation bug is unrelated to ModelOutput key assignment when a field was previously None." + }, + { + "left": "issue:45127", + "right": "issue:45357", + "accept": false, + "reason": "LoRA merge collapse with extended vocab is unrelated to incorrect visual-encoder keys saved by Qwen3.5." + }, + { + "left": "issue:44737", + "right": "issue:45803", + "accept": false, + "reason": "XLNet device placement bug and a bare-except swallowing KeyboardInterrupt are unrelated." + }, + { + "left": "issue:44462", + "right": "issue:45907", + "accept": false, + "reason": "AutoTokenizer ignoring tokenizer.json is unrelated to list_repo_files not forwarding revision." + }, + { + "left": "issue:44077", + "right": "issue:44079", + "accept": false, + "reason": "patchtsmixer post_init policy is unrelated to ModelOutput key assignment semantics." + }, + { + "left": "issue:43335", + "right": "issue:44855", + "accept": false, + "reason": "SwitchTransformers config creating a sparse layer at zero sparse layers is unrelated to the DebertaV2 Python 3.13 parsing error." + }, + { + "left": "issue:43644", + "right": "issue:44913", + "accept": false, + "reason": "A framework-wide v5 non-persistent-buffer regression is unrelated to GPTNeoX rotary_pct persistence." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 31, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:36246", + "issue:38175", + "issue:38617", + "issue:39692", + "issue:41628", + "issue:41950", + "issue:42175", + "issue:42490", + "issue:42907", + "issue:43065", + "issue:43295", + "issue:43335", + "issue:43404", + "issue:43441", + "issue:43479", + "issue:43540", + "issue:43550", + "issue:43643", + "issue:43650", + "issue:43854", + "issue:43976", + "issue:44016", + "issue:44112", + "issue:44186", + "issue:44190", + "issue:44220", + "issue:44246", + "issue:44279", + "issue:44466", + "issue:44496", + "issue:44561", + "issue:44683", + "issue:44734", + "issue:44913", + "issue:44938", + "issue:44964", + "issue:44987", + "issue:45003", + "issue:45216", + "issue:45245", + "issue:45276", + "issue:45375", + "issue:45446", + "issue:45542", + "issue:45584", + "issue:45600", + "issue:45632", + "issue:45803" + ], + "soft_pairs": [ + "issue:44938|issue:45542", + "issue:43550|issue:44190", + "issue:43540|issue:44734", + "issue:44112|issue:45600", + "issue:43335|issue:44190", + "issue:43650|issue:44220", + "issue:45276|issue:45584", + "issue:44496|issue:45803", + "issue:39692|issue:44190", + "issue:45245|issue:45584", + "issue:43295|issue:45276", + "issue:41950|issue:44016", + "issue:43479|issue:45375", + "issue:36246|issue:41628", + "issue:36246|issue:38617", + "issue:44186|issue:44734", + "issue:38175|issue:43065", + "issue:42175|issue:44279", + "issue:44683|issue:45446", + "issue:44016|issue:44220", + "issue:44466|issue:44913", + "issue:43404|issue:45216", + "issue:44246|issue:45003", + "issue:44734|issue:45632", + "issue:44561|issue:45632", + "issue:43643|issue:44496", + "issue:42490|issue:45632", + "issue:44964|issue:45600", + "issue:44112|issue:44964", + "issue:43854|issue:44964", + "issue:42907|issue:43441", + "issue:43976|issue:44987" + ], + "prepared_review_unit_hash": "808b39d7c3fc81174cc9abbc4719d3a502b4cdf1e3f67571b30916dbae65abc9", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37304, + "estimated_input_tokens": 9326, + "estimated_eval_tokens": 18908 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is heterogeneous: it groups several unrelated import, config, model-loading, and runtime-regression reports, but none of the soft-linked pairs look like the same underlying bug/change.", + "confidence": 0.39, + "canonical_issue_reason": "Issue 41628 is the broadest import/export regression in the set ('Cannot import name AutoImageProcessor'), so it is the closest anchor for the import-related failures, though the overall cluster is still not a true duplicate group.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44964 is the strongest standalone triage candidate: it is a concrete, reproducible latest-transformers model-load failure with a specific model and clear user impact.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44938", + "right": "issue:45542", + "accept": false, + "reason": "Different failure modes: Python 3.14 import/load breakage vs an undefined TensorFlow backend caused by installing only tensorboard." + }, + { + "left": "issue:43550", + "right": "issue:44190", + "accept": false, + "reason": "One is a Bamba SDPA/torch.compile bug; the other is a local dataset loading failure in an example script." + }, + { + "left": "issue:43540", + "right": "issue:44734", + "accept": false, + "reason": "Different code paths: Qwen3OmniMoe video input validation vs transformers-serve KV-cache continuation indexing." + }, + { + "left": "issue:44112", + "right": "issue:45600", + "accept": false, + "reason": "Both mention CI, but the bugs are unrelated: a GraniteSpeech device-override test vs stale auto-mapping references to removed configs." + }, + { + "left": "issue:43335", + "right": "issue:44190", + "accept": false, + "reason": "SwitchTransformers sparse-layer initialization is unrelated to the image-classification local dataset loading issue." + }, + { + "left": "issue:43650", + "right": "issue:44220", + "accept": false, + "reason": "The titles are too generic to link; one is an unspecified data request, the other is a specific fbank feature extraction bug." + }, + { + "left": "issue:45276", + "right": "issue:45584", + "accept": false, + "reason": "Different models and subsystems: Gemma4 embedding resize behavior vs Whisper generation after align_special_tokens." + }, + { + "left": "issue:44496", + "right": "issue:45803", + "accept": false, + "reason": "Unrecognized model/config loading is unrelated to the bare-except bug in FuyuBatchFeature tensor conversion." + }, + { + "left": "issue:39692", + "right": "issue:44190", + "accept": false, + "reason": "A documentation-example bug for SigLIP2 is not the same as a dataset-loading problem in an example script." + }, + { + "left": "issue:45245", + "right": "issue:45584", + "accept": false, + "reason": "Category-count runtime error and Whisper empty-transcription failure are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:45276", + "accept": false, + "reason": "Both are regressions, but they affect different APIs and models: processor/tokenizer handling vs Gemma4 token embedding resizing." + }, + { + "left": "issue:41950", + "right": "issue:44016", + "accept": false, + "reason": "Video-classification pipeline processor lookup is unrelated to a notebook syntax error." + }, + { + "left": "issue:43479", + "right": "issue:45375", + "accept": false, + "reason": "Both are config-field issues, but they involve different models and different missing/defaulted fields." + }, + { + "left": "issue:36246", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but they are different missing symbols in different modules and do not indicate the same bug." + }, + { + "left": "issue:36246", + "right": "issue:38617", + "accept": false, + "reason": "Different missing imports in different modules; no shared code-path evidence." + }, + { + "left": "issue:44186", + "right": "issue:44734", + "accept": false, + "reason": "Tokenizer crashes on NER padding/truncation are unrelated to serve KV-cache tensor indexing." + }, + { + "left": "issue:38175", + "right": "issue:43065", + "accept": false, + "reason": "A zero-probability model-output issue is unrelated to a dummy Conv2d implementation in Sam3PixelDecoder." + }, + { + "left": "issue:42175", + "right": "issue:44279", + "accept": false, + "reason": "A packaging/backend-dependency complaint is too broad to be the same as a generic dependency issue." + }, + { + "left": "issue:44683", + "right": "issue:45446", + "accept": false, + "reason": "Both involve flex_attention, but one is a compiled-kernel failure on torch >=2.9 while the other is a version-check/import-gate bug; not the same concrete fix." + }, + { + "left": "issue:44016", + "right": "issue:44220", + "accept": false, + "reason": "Notebook syntax error and fbank feature extraction bug are unrelated." + }, + { + "left": "issue:44466", + "right": "issue:44913", + "accept": false, + "reason": "Serialization/tied-weights behavior is a different issue from GPTNeoXConfig rotary_pct not persisting on reload." + }, + { + "left": "issue:43404", + "right": "issue:45216", + "accept": false, + "reason": "Different models and symptoms: untied lm_head weights in Mistral3 vs incorrect save_pretrained checkpointing for Qwen3.5." + }, + { + "left": "issue:44246", + "right": "issue:45003", + "accept": false, + "reason": "Slow import and unsafe sys.modules access may be related in spirit, but there is no evidence they are the same bug or fix." + }, + { + "left": "issue:44734", + "right": "issue:45632", + "accept": false, + "reason": "Serve KV-cache indexing and trust_remote_code cache-path collisions are unrelated." + }, + { + "left": "issue:44561", + "right": "issue:45632", + "accept": false, + "reason": "Both touch trust_remote_code, but one is a removed-helper regression and the other is a local cache-path collision; different code paths." + }, + { + "left": "issue:43643", + "right": "issue:44496", + "accept": false, + "reason": "Missing fields in AutoConfig with trust_remote_code is not the same as an unrecognized model due to missing model_type." + }, + { + "left": "issue:42490", + "right": "issue:45632", + "accept": false, + "reason": "Path-with-trailing-slash warning and cache-path collision for same leaf names are different path-handling bugs." + }, + { + "left": "issue:44964", + "right": "issue:45600", + "accept": false, + "reason": "A Phi-4 multimodal load failure is unrelated to CI breakage from removed Sam3LiteText references." + }, + { + "left": "issue:44112", + "right": "issue:44964", + "accept": false, + "reason": "A GraniteSpeech CI test failure is unrelated to a Phi-4 multimodal model-load failure." + }, + { + "left": "issue:43854", + "right": "issue:44964", + "accept": false, + "reason": "Both are model-load failures, but they involve different models and likely different root causes; not mergeable as one PR/bug." + }, + { + "left": "issue:42907", + "right": "issue:43441", + "accept": false, + "reason": "Saving dequantized Ministral/Devstral models and FlashAttention failures in Ministral-3 are separate problems." + }, + { + "left": "issue:43976", + "right": "issue:44987", + "accept": false, + "reason": "A Python version compatibility issue is unrelated to a specific model-loading failure on transformers>=5.1.0." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 32, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:38175", + "issue:39692", + "issue:41950", + "issue:43295", + "issue:43299", + "issue:43381", + "issue:43479", + "issue:43525", + "issue:43526", + "issue:43531", + "issue:43577", + "issue:43650", + "issue:43673", + "issue:43824", + "issue:43854", + "issue:43901", + "issue:43976", + "issue:44016", + "issue:44062", + "issue:44075", + "issue:44242", + "issue:44279", + "issue:44336", + "issue:44351", + "issue:44368", + "issue:44496", + "issue:44521", + "issue:44617", + "issue:44683", + "issue:44779", + "issue:44933", + "issue:45020", + "issue:45042", + "issue:45200", + "issue:45216", + "issue:45278", + "issue:45325", + "issue:45335", + "issue:45357", + "issue:45405", + "issue:45542", + "issue:45584", + "issue:45593", + "issue:45600", + "issue:45759", + "issue:45812", + "issue:45823" + ], + "soft_pairs": [ + "issue:43295|issue:45335", + "issue:44016|issue:44062", + "issue:43673|issue:44336", + "issue:43531|issue:45600", + "issue:43526|issue:43901", + "issue:44496|issue:45042", + "issue:43824|issue:44933", + "issue:43299|issue:45216", + "issue:44075|issue:44368", + "issue:45335|issue:45584", + "issue:43976|issue:44279", + "issue:43381|issue:45593", + "issue:43901|issue:44496", + "issue:43299|issue:45357", + "issue:44496|issue:45759", + "issue:44521|issue:45325", + "issue:44351|issue:45278", + "issue:44242|issue:45200", + "issue:43976|issue:45405", + "issue:41950|issue:45020", + "issue:43479|issue:43525", + "issue:44779|issue:45812", + "issue:38175|issue:44617", + "issue:41950|issue:45823", + "issue:43577|issue:45216", + "issue:44683|issue:45542", + "issue:39692|issue:43901", + "issue:44279|issue:45405", + "issue:43299|issue:43854", + "issue:43295|issue:44779", + "issue:41950|issue:43650" + ], + "prepared_review_unit_hash": "dce9509f3998a3a96d1c0b22e2b0d59a33a62493286600a929a56cbf163ee1bc", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36318, + "estimated_input_tokens": 9080, + "estimated_eval_tokens": 18416 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose mix of unrelated v5/v5.x regressions, model-loading/config bugs, tokenizer/processor issues, and a few docs/import problems. The titles are superficially similar, but the underlying code paths differ, so these should stay separate.", + "confidence": 0.93, + "canonical_issue_reason": "issue:45020 is the broadest and most representative regression report in the set ('recent transformers versions break models using remote_code'), so it is the best single issue to anchor the cluster even though it is not a true duplicate of the others.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45020 best captures the dominant theme of recent-version breakage and is the most reusable summary issue for this cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43295", + "right": "issue:45335", + "accept": false, + "reason": "Both are regressions, but one is about processor.tokenizer/image inputs and the other is about resize_token_embeddings not updating decoder.embed_tokens in t5gemma; different code paths." + }, + { + "left": "issue:44016", + "right": "issue:44062", + "accept": false, + "reason": "A notebook syntax error and a tokenizers.AddedToken keyword-argument TypeError are unrelated bugs." + }, + { + "left": "issue:43673", + "right": "issue:44336", + "accept": false, + "reason": "Chunked-prefill cache handling and ANSI color output in loading_report are separate subsystems with different failure modes." + }, + { + "left": "issue:43531", + "right": "issue:45600", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and auto_mappings referencing removed configs are not the same bug." + }, + { + "left": "issue:43526", + "right": "issue:43901", + "accept": false, + "reason": "BeitImageProcessorFast reduce_labels is a processor bug; return_all_scores is a documentation/behavior mismatch." + }, + { + "left": "issue:44496", + "right": "issue:45042", + "accept": false, + "reason": "Unrecognized model/config loading and PIL image processors wrongly requiring torchvision are distinct issues." + }, + { + "left": "issue:43824", + "right": "issue:44933", + "accept": false, + "reason": "These are both import errors, but they involve different missing symbols in different modules." + }, + { + "left": "issue:43299", + "right": "issue:45216", + "accept": false, + "reason": "One is a model-loading regression for Qwen3VL MoE; the other is a save_pretrained checkpoint regression for Qwen3.5." + }, + { + "left": "issue:44075", + "right": "issue:44368", + "accept": false, + "reason": "SGD args being ignored and a tie_word_embeddings warning during LoRA fine-tuning are unrelated." + }, + { + "left": "issue:45335", + "right": "issue:45584", + "accept": false, + "reason": "A t5gemma embedding resize bug and Whisper empty-transcription failure after align_special_tokens are different code paths." + }, + { + "left": "issue:43976", + "right": "issue:44279", + "accept": false, + "reason": "Python-version support problems and a generic dependency issue are not the same underlying defect." + }, + { + "left": "issue:43381", + "right": "issue:45593", + "accept": false, + "reason": "Gradient checkpointing in eval mode and D-FINE aux-loss behavior when denoising is off are unrelated training behaviors." + }, + { + "left": "issue:43901", + "right": "issue:44496", + "accept": false, + "reason": "Docs wording around return_all_scores does not match the OLMo-Hybrid config/model_type loading error." + }, + { + "left": "issue:43299", + "right": "issue:45357", + "accept": false, + "reason": "Both involve Qwen3.5/VL, but loading failure and incorrect visual-encoder keys in save_pretrained are different operations and fixes." + }, + { + "left": "issue:44496", + "right": "issue:45759", + "accept": false, + "reason": "A generic unrecognized-model error and from_config not unwrapping text_config for composite multimodal configs are not the same bug." + }, + { + "left": "issue:44521", + "right": "issue:45325", + "accept": false, + "reason": "All-zero assistant masks in apply_chat_template and Qwen2.5-VL rope position scaling are unrelated multimodal issues." + }, + { + "left": "issue:44351", + "right": "issue:45278", + "accept": false, + "reason": "A missing HybridCache import is one concrete symbol issue; 'many import errors' is too broad and not the same bug." + }, + { + "left": "issue:44242", + "right": "issue:45200", + "accept": false, + "reason": "Load-balancing loss gating and mm_token_type_ids defaulting to zeros are different model/training concerns." + }, + { + "left": "issue:43976", + "right": "issue:45405", + "accept": false, + "reason": "Python compatibility and an overly high MIN_PEFT_VERSION bump are separate dependency problems." + }, + { + "left": "issue:41950", + "right": "issue:45020", + "accept": false, + "reason": "A video-classification pipeline selecting image processors is not the same as remote_code models breaking in recent transformers versions." + }, + { + "left": "issue:43479", + "right": "issue:43525", + "accept": false, + "reason": "Phi4MultimodalConfig defaults for None and Llama4Config missing pad_token_id are different config initialization bugs." + }, + { + "left": "issue:44779", + "right": "issue:45812", + "accept": false, + "reason": "Tokenizer regressions for Deepseek and Granite are both tokenization problems, but they affect different model families and likely different code paths." + }, + { + "left": "issue:38175", + "right": "issue:44617", + "accept": false, + "reason": "SigLIP2 zero probabilities and Sam3Video CUDA OOM are completely unrelated." + }, + { + "left": "issue:41950", + "right": "issue:45823", + "accept": false, + "reason": "A pipeline using image processors and a Gemma4 PLE device-map mismatch are unrelated issues." + }, + { + "left": "issue:43577", + "right": "issue:45216", + "accept": false, + "reason": "BLIP2 dtype propagation and Qwen3.5 save_pretrained checkpoint correctness are distinct regressions." + }, + { + "left": "issue:44683", + "right": "issue:45542", + "accept": false, + "reason": "Compiled flex_attention on torch>=2.9 and tensorboard/TensorFlow backend detection are unrelated." + }, + { + "left": "issue:39692", + "right": "issue:43901", + "accept": false, + "reason": "SigLIP2 docs example errors and TextClassificationPipeline docs on return_all_scores are different documentation issues with different models and failure modes." + }, + { + "left": "issue:44279", + "right": "issue:45405", + "accept": false, + "reason": "A generic dependency issue is not the same as bumping MIN_PEFT_VERSION beyond what is released." + }, + { + "left": "issue:43299", + "right": "issue:43854", + "accept": false, + "reason": "Qwen3VL loading regressions and GLM-4.7-Flash unit-test loading failures involve different models and likely different fixes." + }, + { + "left": "issue:43295", + "right": "issue:44779", + "accept": false, + "reason": "Processor.tokenizer/image-passing regression and Deepseek tokenizer output regression are different API surfaces." + }, + { + "left": "issue:41950", + "right": "issue:43650", + "accept": false, + "reason": "The video-classification processor bug and the placeholder 'ADD THE DATA' issue are not related." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 33, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:33290", + "issue:42175", + "issue:42994", + "issue:43065", + "issue:43122", + "issue:43262", + "issue:43317", + "issue:43441", + "issue:43450", + "issue:43650", + "issue:43653", + "issue:43673", + "issue:43723", + "issue:43761", + "issue:43784", + "issue:43901", + "issue:43976", + "issue:43994", + "issue:44016", + "issue:44162", + "issue:44261", + "issue:44263", + "issue:44279", + "issue:44291", + "issue:44297", + "issue:44361", + "issue:44462", + "issue:44466", + "issue:44521", + "issue:44623", + "issue:44671", + "issue:44683", + "issue:44749", + "issue:44779", + "issue:44898", + "issue:44987", + "issue:45081", + "issue:45276", + "issue:45290", + "issue:45397", + "issue:45542", + "issue:45593", + "issue:45600", + "issue:45656", + "issue:45712", + "issue:45823", + "issue:45834", + "issue:45907" + ], + "soft_pairs": [ + "issue:44462|issue:45081", + "issue:44297|issue:44462", + "issue:45081|issue:45907", + "issue:44016|issue:44291", + "issue:33290|issue:45656", + "issue:43450|issue:44279", + "issue:43441|issue:45834", + "issue:45600|issue:45712", + "issue:43065|issue:44898", + "issue:44466|issue:45276", + "issue:42994|issue:43317", + "issue:43650|issue:45290", + "issue:44261|issue:44263", + "issue:44623|issue:45397", + "issue:44623|issue:44749", + "issue:43441|issue:44683", + "issue:43653|issue:45593", + "issue:43262|issue:44521", + "issue:44016|issue:45823", + "issue:43723|issue:44162", + "issue:43673|issue:43761", + "issue:44361|issue:45593", + "issue:44671|issue:44779", + "issue:42175|issue:43976", + "issue:43901|issue:43994", + "issue:43784|issue:45542", + "issue:43122|issue:43650", + "issue:44987|issue:45542", + "issue:44297|issue:45397" + ], + "prepared_review_unit_hash": "903f9e8e03e383b199a2a8bd6fe11ae00576868b322440ed4ae92c0e0af8072d", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "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": "The cluster is mostly heterogeneous; the soft pairs are broad topical similarities, not the same underlying bug/change. I found no true duplicate PRs here, and none of the issue pairs are strong enough to merge as duplicates.", + "confidence": 0.94, + "canonical_issue_reason": "No single canonical issue stands out because the set spans many unrelated bugs/regressions; if one representative issue is needed, issue 44462 is the clearest self-contained user-facing bug report among the tokenizer-related items.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44462 is the strongest representative bug report overall: it is concrete, reproducible, and has a clear loading-path regression description. But it is not a canonical for the whole set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44462", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer repository loading vs Mistral regex patch crash; related area, but different code paths and failures." + }, + { + "left": "issue:44297", + "right": "issue:44462", + "accept": false, + "reason": "save_pretrained tokenizer_class mismatch vs AutoTokenizer ignoring tokenizer.json; same subsystem, but distinct save vs load bugs." + }, + { + "left": "issue:45081", + "right": "issue:45907", + "accept": false, + "reason": "Regex patch crash vs list_repo_files revision forwarding; unrelated functionality." + }, + { + "left": "issue:44016", + "right": "issue:44291", + "accept": false, + "reason": "Notebook syntax error vs init_empty_weights argument mismatch; unrelated issues." + }, + { + "left": "issue:33290", + "right": "issue:45656", + "accept": false, + "reason": "Adafactor OOM vs deepspeed optimizer.step called twice; both deepspeed-related, but different concrete bugs." + }, + { + "left": "issue:43450", + "right": "issue:44279", + "accept": false, + "reason": "Video processor batch shape bug vs dependency/install issue; not the same bug." + }, + { + "left": "issue:43441", + "right": "issue:45834", + "accept": false, + "reason": "FlashAttention regression for Ministral vs Kosmos2.5 OCR index error; different models and failure modes." + }, + { + "left": "issue:45600", + "right": "issue:45712", + "accept": false, + "reason": "Stale auto_mappings references vs leftover dummy classes; both maintenance problems, but not the same change." + }, + { + "left": "issue:43065", + "right": "issue:44898", + "accept": false, + "reason": "Sam3PixelDecoder dummy Conv2d vs Perceiver resolution/interpolation failure; different model paths." + }, + { + "left": "issue:44466", + "right": "issue:45276", + "accept": false, + "reason": "lm_head.weight serialization inconsistency vs Gemma4 resize_token_embeddings propagation; same model family, but different bugs." + }, + { + "left": "issue:42994", + "right": "issue:43317", + "accept": false, + "reason": "Quantized save failure vs dequantized load/offload failure; related quantization area, but opposite operations and different symptoms." + }, + { + "left": "issue:43650", + "right": "issue:45290", + "accept": false, + "reason": "Placeholder/data issue vs apply_chat_template tool-call crash; unrelated." + }, + { + "left": "issue:44261", + "right": "issue:44263", + "accept": false, + "reason": "RMS norm epsilon precision mismatch vs torch.split return-value handling; unrelated internals." + }, + { + "left": "issue:44623", + "right": "issue:45397", + "accept": false, + "reason": "processor.save_pretrained missing files vs gemma-4 zero3 from_pretrained; different save/load problems." + }, + { + "left": "issue:44623", + "right": "issue:44749", + "accept": false, + "reason": "Missing files on processor save vs performance slowdown after version upgrade; unrelated." + }, + { + "left": "issue:43441", + "right": "issue:44683", + "accept": false, + "reason": "FlashAttention failure for one model vs compiled flex_attention failing on torch>=2.9; attention-related but different backends and bugs." + }, + { + "left": "issue:43653", + "right": "issue:45593", + "accept": false, + "reason": "BigBirdTokenizer special-token registration bug vs D-FINE auxiliary-loss omission; unrelated." + }, + { + "left": "issue:43262", + "right": "issue:44521", + "accept": false, + "reason": "apply_chat_template sample-rate default vs zero assistant_masks for multimodal inputs; same API surface, but different behaviors." + }, + { + "left": "issue:44016", + "right": "issue:45823", + "accept": false, + "reason": "Notebook syntax error vs Gemma4 device mismatch in forward; unrelated." + }, + { + "left": "issue:43723", + "right": "issue:44162", + "accept": false, + "reason": "AutoTokenizer loading regression vs ESM2 breakage; unrelated models and failure modes." + }, + { + "left": "issue:43673", + "right": "issue:43761", + "accept": false, + "reason": "Missing GenerationMixin cache in chunked prefill vs CLIPVisionModel hidden_states bug; unrelated." + }, + { + "left": "issue:44361", + "right": "issue:45593", + "accept": false, + "reason": "MLukeTokenizer AttributeError vs D-FINE auxiliary-loss issue; unrelated." + }, + { + "left": "issue:44671", + "right": "issue:44779", + "accept": false, + "reason": "CamemBERT masked-LM regression vs Deepseek tokenizer regression; both output-quality bugs, but different models and code paths." + }, + { + "left": "issue:42175", + "right": "issue:43976", + "accept": false, + "reason": "Missing TensorFlow backend in torch install vs Python version compatibility regression; environment-related but distinct." + }, + { + "left": "issue:43901", + "right": "issue:43994", + "accept": false, + "reason": "Docs/behavior mismatch in TextClassificationPipeline vs Siglip2 inference regression; unrelated." + }, + { + "left": "issue:43784", + "right": "issue:45542", + "accept": false, + "reason": "Import NameError for sentence-transformers vs tensorboard-only install without TensorFlow; unrelated packaging issues." + }, + { + "left": "issue:43122", + "right": "issue:43650", + "accept": false, + "reason": "Tokenizer behavior change across versions vs placeholder 'ADD THE DATA' report; not a duplicate." + }, + { + "left": "issue:44987", + "right": "issue:45542", + "accept": false, + "reason": "Model loading failure on transformers>=5.1.0 vs tensorboard/TensorFlow install issue; unrelated." + }, + { + "left": "issue:44297", + "right": "issue:45397", + "accept": false, + "reason": "Tokenizer save_pretrained metadata mismatch vs gemma-4 zero3 from_pretrained; different bugs in different subsystems." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 34, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:36246", + "issue:41628", + "issue:41950", + "issue:42491", + "issue:42757", + "issue:43299", + "issue:43352", + "issue:43450", + "issue:43525", + "issue:43526", + "issue:43531", + "issue:43596", + "issue:43643", + "issue:43673", + "issue:43824", + "issue:43881", + "issue:43901", + "issue:43976", + "issue:44162", + "issue:44188", + "issue:44220", + "issue:44242", + "issue:44246", + "issue:44295", + "issue:44336", + "issue:44351", + "issue:44373", + "issue:44485", + "issue:44496", + "issue:44561", + "issue:44589", + "issue:44617", + "issue:44749", + "issue:44779", + "issue:44857", + "issue:44871", + "issue:44877", + "issue:45003", + "issue:45200", + "issue:45250", + "issue:45292", + "issue:45362", + "issue:45375", + "issue:45397", + "issue:45561", + "issue:45563", + "issue:45584", + "issue:45593" + ], + "soft_pairs": [ + "issue:43824|issue:43976", + "issue:44188|issue:44336", + "issue:43450|issue:44617", + "issue:43525|issue:43881", + "issue:43525|issue:43643", + "issue:43901|issue:44246", + "issue:41628|issue:43976", + "issue:44589|issue:45362", + "issue:45561|issue:45563", + "issue:44336|issue:44373", + "issue:42491|issue:43299", + "issue:44242|issue:45593", + "issue:44295|issue:45397", + "issue:41950|issue:44220", + "issue:44246|issue:44496", + "issue:43531|issue:45362", + "issue:43526|issue:45003", + "issue:36246|issue:42757", + "issue:43352|issue:45250", + "issue:44336|issue:44485", + "issue:36246|issue:44351", + "issue:45003|issue:45561", + "issue:43673|issue:44561", + "issue:44246|issue:44871", + "issue:44242|issue:44857", + "issue:44295|issue:44749", + "issue:44162|issue:45200", + "issue:36246|issue:43596", + "issue:44779|issue:45292", + "issue:44877|issue:45375", + "issue:45292|issue:45584", + "issue:43901|issue:45003" + ], + "prepared_review_unit_hash": "a516a98637b51fbf71e3fb3fca77160c853015f748ecf372e6615ede0450bcc6", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 36942, + "estimated_input_tokens": 9236, + "estimated_eval_tokens": 18728 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly a mix of unrelated Transformers bugs. None of the soft pairs are strong enough to merge as duplicates; at most a few share broad themes like imports, config strictness, or model loading, but they target different concrete failures.", + "confidence": 0.84, + "canonical_issue_reason": "issue:43824 is the clearest, most specific regression report in the set: a concrete import error for Qwen2.5-VL with multiple inbound references, making it the best canonical anchor among these issues.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43824 is also the best overall issue candidate because it is narrowly scoped, user-facing, and well-documented with strong evidence of recurrence.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43824", + "right": "issue:43976", + "accept": false, + "reason": "Both mention import-related breakage, but one is a missing Qwen2.5-VL symbol and the other is a Python version compatibility problem; different failure modes and code paths." + }, + { + "left": "issue:44188", + "right": "issue:44336", + "accept": false, + "reason": "Attention-kernel divergence under torch.compile is unrelated to ANSI escape codes in loading_report." + }, + { + "left": "issue:43450", + "right": "issue:44617", + "accept": false, + "reason": "Batched video-processor shape bugs and Sam3Video CUDA OOM are different problems." + }, + { + "left": "issue:43525", + "right": "issue:43881", + "accept": false, + "reason": "A missing pad_token_id on Llama4Config is not the same as a glm-4v model loading failure." + }, + { + "left": "issue:43525", + "right": "issue:43643", + "accept": false, + "reason": "One is a config attribute error in Llama4, the other is missing fields when loading trust_remote_code configs; too broad to treat as the same bug." + }, + { + "left": "issue:43901", + "right": "issue:44246", + "accept": false, + "reason": "Docs/behavior mismatch for text classification pipelines is unrelated to sporadic import latency." + }, + { + "left": "issue:41628", + "right": "issue:43976", + "accept": false, + "reason": "AutoImageProcessor import failure and Python version support issues are distinct regressions." + }, + { + "left": "issue:44589", + "right": "issue:45362", + "accept": false, + "reason": "A Float8 storage type error is unrelated to Qwen3.5 chat crashes." + }, + { + "left": "issue:45561", + "right": "issue:45563", + "accept": false, + "reason": "A pytest-xdist capture-file race and a stale generation warning are unrelated." + }, + { + "left": "issue:44336", + "right": "issue:44373", + "accept": false, + "reason": "ANSI code emission in loading_report is unrelated to a wrong docstring for position_ids." + }, + { + "left": "issue:42491", + "right": "issue:43299", + "accept": false, + "reason": "Both mention Qwen3 MoE, but one is a LoRA compatibility issue across versions and the other is a model-loading regression for Qwen3VL MoE; not the same concrete bug." + }, + { + "left": "issue:44242", + "right": "issue:45593", + "accept": false, + "reason": "Missing load-balancing loss and missing auxiliary losses are different model-specific training bugs." + }, + { + "left": "issue:44295", + "right": "issue:45397", + "accept": false, + "reason": "A position_ids buffer access error and a zero3 from_pretrained issue are different failures." + }, + { + "left": "issue:41950", + "right": "issue:44220", + "accept": false, + "reason": "Video-classification pipeline processor selection and _torch_extract_fbank_features are unrelated." + }, + { + "left": "issue:44246", + "right": "issue:44496", + "accept": false, + "reason": "Import slowness and an unrecognized model/config loading error do not share the same bug." + }, + { + "left": "issue:43531", + "right": "issue:45362", + "accept": false, + "reason": "A sliding_window bug in Qwen3-MoE is unrelated to Qwen3.5 chat crashes." + }, + { + "left": "issue:43526", + "right": "issue:45003", + "accept": false, + "reason": "A BeitImageProcessorFast label-reduction bug and unsafe sys.modules access are different code paths." + }, + { + "left": "issue:36246", + "right": "issue:42757", + "accept": false, + "reason": "Missing Qwen2.5-VL exports and missing is_offline_mode from huggingface_hub are separate import regressions." + }, + { + "left": "issue:43352", + "right": "issue:45250", + "accept": false, + "reason": "A concrete Nemotron Flash Attention 2 limitation is not the same as a generic Flash Attention 2 topic issue." + }, + { + "left": "issue:44336", + "right": "issue:44485", + "accept": false, + "reason": "Terminal color handling and GLM-5 RoPE implementation are unrelated." + }, + { + "left": "issue:36246", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they concern different missing symbols and likely different package-export paths." + }, + { + "left": "issue:45003", + "right": "issue:45561", + "accept": false, + "reason": "Unsafe sys.modules access and a captured_info.txt race are unrelated." + }, + { + "left": "issue:43673", + "right": "issue:44561", + "accept": false, + "reason": "Chunked prefill cache absence and removal of is_torch_fx_available affect different subsystems and failure modes." + }, + { + "left": "issue:44246", + "right": "issue:44871", + "accept": false, + "reason": "Import performance and Gemma eos_token_id inconsistency are unrelated." + }, + { + "left": "issue:44242", + "right": "issue:44857", + "accept": false, + "reason": "Load-balancing-loss omission and AMP/CUDA crash in LwDetrImageLoss are different bugs." + }, + { + "left": "issue:44295", + "right": "issue:44749", + "accept": false, + "reason": "Position_ids buffer access and Chinese-language filtering slowdown are unrelated." + }, + { + "left": "issue:44162", + "right": "issue:45200", + "accept": false, + "reason": "ESM2 breakage and Gemma 4 mm_token_type_ids defaults are different model bugs." + }, + { + "left": "issue:36246", + "right": "issue:43596", + "accept": false, + "reason": "Qwen2.5-VL import failure and a BertModel deepspeed zero3 index error are unrelated." + }, + { + "left": "issue:44779", + "right": "issue:45292", + "accept": false, + "reason": "Deepseek tokenizer corruption in v5 and resize_token_embeddings not updating output embeddings are different issues." + }, + { + "left": "issue:44877", + "right": "issue:45375", + "accept": false, + "reason": "Both involve strict config behavior, but they affect different model configs and the reported failures are not the same concrete bug." + }, + { + "left": "issue:45292", + "right": "issue:45584", + "accept": false, + "reason": "Embedding resize behavior and Whisper empty-transcription generation failure are unrelated." + }, + { + "left": "issue:43901", + "right": "issue:45003", + "accept": false, + "reason": "A pipeline doc/behavior mismatch and unsafe sys.modules access are different classes of bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 35, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:42886", + "issue:43065", + "issue:43232", + "issue:43257", + "issue:43262", + "issue:43408", + "issue:43475", + "issue:43540", + "issue:43646", + "issue:43653", + "issue:43673", + "issue:43931", + "issue:44060", + "issue:44077", + "issue:44162", + "issue:44246", + "issue:44297", + "issue:44336", + "issue:44361", + "issue:44368", + "issue:44466", + "issue:44496", + "issue:44623", + "issue:44734", + "issue:44749", + "issue:44869", + "issue:44933", + "issue:44977", + "issue:44987", + "issue:44995", + "issue:45003", + "issue:45137", + "issue:45200", + "issue:45229", + "issue:45278", + "issue:45292", + "issue:45310", + "issue:45335", + "issue:45356", + "issue:45375", + "issue:45397", + "issue:45459", + "issue:45563", + "issue:45588", + "issue:45600", + "issue:45656", + "issue:45701", + "issue:45759" + ], + "soft_pairs": [ + "issue:45292|issue:45701", + "issue:43232|issue:44336", + "issue:44496|issue:45003", + "issue:44987|issue:45459", + "issue:43262|issue:44246", + "issue:42886|issue:45229", + "issue:45375|issue:45759", + "issue:43257|issue:45137", + "issue:43065|issue:43408", + "issue:44995|issue:45200", + "issue:43475|issue:43540", + "issue:44060|issue:44077", + "issue:45397|issue:45656", + "issue:44466|issue:45335", + "issue:44995|issue:45588", + "issue:43653|issue:44162", + "issue:44368|issue:44977", + "issue:44368|issue:45563", + "issue:45137|issue:45600", + "issue:43673|issue:44734", + "issue:43931|issue:44368", + "issue:44869|issue:45356", + "issue:44933|issue:45459", + "issue:44297|issue:44623", + "issue:43232|issue:43673", + "issue:44361|issue:44749", + "issue:43646|issue:44162", + "issue:43673|issue:45278", + "issue:44623|issue:45656", + "issue:44749|issue:45656", + "issue:44749|issue:45310" + ], + "prepared_review_unit_hash": "5aceae6e7e93222d0f8fdf2daceba0899b61a599070156a4535ee32c3e3f7edc", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37001, + "estimated_input_tokens": 9251, + "estimated_eval_tokens": 18758 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These items are mostly unrelated regressions across tokenization, generation, multimodal configs, serialization, and deepspeed; none of the soft pairs look like the same underlying bug or change.", + "confidence": 0.86, + "canonical_issue_reason": "issue:45278 is the broadest umbrella-style report here: a version-upgrade regression causing many import errors, unlike the more narrowly scoped model-specific bugs.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 is the best representative issue because it captures a general cross-cutting failure mode after upgrade, making it the closest thing to a cluster anchor among these mostly specific reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45292", + "right": "issue:45701", + "accept": false, + "reason": "Embedding resize behavior is a different path from tokenizer/output changes; one is about resized embeddings, the other about tokenization differences across versions." + }, + { + "left": "issue:43232", + "right": "issue:44336", + "accept": false, + "reason": "Generation state updates after sync_gpus are unrelated to ANSI color handling in loading_report." + }, + { + "left": "issue:44496", + "right": "issue:45003", + "accept": false, + "reason": "Unrecognized model/config loading is a different bug from unsafe sys.modules access in modeling_utils." + }, + { + "left": "issue:44987", + "right": "issue:45459", + "accept": false, + "reason": "Loading a specific model on transformers>=5.1.0 is not the same as masking tokenizer errors when protobuf is missing." + }, + { + "left": "issue:43262", + "right": "issue:44246", + "accept": false, + "reason": "Audio sampling-rate defaults in chat templates are unrelated to intermittent import slowdowns." + }, + { + "left": "issue:42886", + "right": "issue:45229", + "accept": false, + "reason": "Offline cache loading for tokenizers and Gemma4 multi-GPU CUDA OOM are different failure modes." + }, + { + "left": "issue:45375", + "right": "issue:45759", + "accept": false, + "reason": "Missing config fields in strict serialization is not the same as unwrapping text_config for composite multimodal configs." + }, + { + "left": "issue:43257", + "right": "issue:45137", + "accept": false, + "reason": "MOE weight conversion with accelerate+deepspeed and a Zero3 deque pop are distinct deepspeed-related bugs." + }, + { + "left": "issue:43065", + "right": "issue:43408", + "accept": false, + "reason": "A dummy Conv2d in the pixel decoder is unrelated to a model-type warning when instantiating sam3_tracker." + }, + { + "left": "issue:44995", + "right": "issue:45200", + "accept": false, + "reason": "A stale cache on second forward pass is not the same as missing default mm_token_type_ids." + }, + { + "left": "issue:43475", + "right": "issue:43540", + "accept": false, + "reason": "A missing fpn_position_embeddings attribute in SAM 3 Video is unrelated to Qwen3OmniMoe video input validation errors." + }, + { + "left": "issue:44060", + "right": "issue:44077", + "accept": false, + "reason": "Incorrect tied-weights warning and a post_init allowance change are separate model/config validation issues." + }, + { + "left": "issue:45397", + "right": "issue:45656", + "accept": false, + "reason": "Gemma4 zero3 from_pretrained failure is not the same as optimizer.step being called twice under deepspeed." + }, + { + "left": "issue:44466", + "right": "issue:45335", + "accept": false, + "reason": "Inconsistent lm_head serialization depending on device is different from resize_token_embeddings not propagating to decoder.embed_tokens." + }, + { + "left": "issue:44995", + "right": "issue:45588", + "accept": false, + "reason": "A stale indexer cache on repeated forward passes is unrelated to flash_attention crashing on s_aux=None." + }, + { + "left": "issue:43653", + "right": "issue:44162", + "accept": false, + "reason": "BigBirdTokenizer special-token registration is a tokenizer bug, while ESM2 being broken is a separate model regression." + }, + { + "left": "issue:44368", + "right": "issue:44977", + "accept": false, + "reason": "A tie_word_embeddings warning for Qwen3.5 LoRA is not the same as a flash-attention generation failure." + }, + { + "left": "issue:44368", + "right": "issue:45563", + "accept": false, + "reason": "The tied-weights warning and the stale num_return_sequences warning are different generation warnings with different triggers." + }, + { + "left": "issue:45137", + "right": "issue:45600", + "accept": false, + "reason": "A Zero3 deque underflow is unrelated to auto_mappings references to removed SAM3 Lite configs." + }, + { + "left": "issue:43673", + "right": "issue:44734", + "accept": false, + "reason": "Chunked prefill cache missing in GenerationMixin and KV-cache continuation tensor indexing are different generation/cache bugs." + }, + { + "left": "issue:43931", + "right": "issue:44368", + "accept": false, + "reason": "A Qwen3-VL weight-shape mismatch is unrelated to a tied-embeddings warning in Qwen3.5 LoRA." + }, + { + "left": "issue:44869", + "right": "issue:45356", + "accept": false, + "reason": "Whisper timestamp decoding crashes on a trailing replacement character, while Kimi-K2.5 has codec-handling and warning issues in its tokenizer." + }, + { + "left": "issue:44933", + "right": "issue:45459", + "accept": false, + "reason": "A nonexistent import from image_utils is unrelated to tokenizer error handling when protobuf is absent." + }, + { + "left": "issue:44297", + "right": "issue:44623", + "accept": false, + "reason": "Tokenizer save_pretrained writing the wrong tokenizer_class is different from processor.save_pretrained missing files." + }, + { + "left": "issue:43232", + "right": "issue:43673", + "accept": false, + "reason": "Both involve generation internals, but the concrete failures differ: sync_gpus update logic versus chunked_prefill cache handling." + }, + { + "left": "issue:44361", + "right": "issue:44749", + "accept": false, + "reason": "MLukeTokenizer AttributeError on tasks is unrelated to the reported slowdown after upgrading transformers." + }, + { + "left": "issue:43646", + "right": "issue:44162", + "accept": false, + "reason": "Custom model initialization breakage in Transformers 5.0.0 is a different class of failure from ESM2 being broken." + }, + { + "left": "issue:43673", + "right": "issue:45278", + "accept": false, + "reason": "Missing generation cache in chunked_prefill and many import errors after upgrade are not the same underlying bug." + }, + { + "left": "issue:44623", + "right": "issue:45656", + "accept": false, + "reason": "Missing files in processor.save_pretrained and duplicate optimizer steps under deepspeed are unrelated." + }, + { + "left": "issue:44749", + "right": "issue:45656", + "accept": false, + "reason": "A performance regression after version upgrade does not match optimizer.step being invoked twice." + }, + { + "left": "issue:44749", + "right": "issue:45310", + "accept": false, + "reason": "Slow filtering after upgrading to 5.3.0 is a different symptom from Qwen3.5 MOE from_pretrained failing on 5.4.0." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 36, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:28282", + "issue:30990", + "issue:36010", + "issue:41628", + "issue:42175", + "issue:42222", + "issue:42371", + "issue:42673", + "issue:42831", + "issue:42994", + "issue:43232", + "issue:43377", + "issue:43475", + "issue:43482", + "issue:43504", + "issue:43525", + "issue:43526", + "issue:43643", + "issue:43673", + "issue:43704", + "issue:43873", + "issue:44038", + "issue:44164", + "issue:44188", + "issue:44222", + "issue:44242", + "issue:44351", + "issue:44496", + "issue:44623", + "issue:44661", + "issue:44792", + "issue:44933", + "issue:44945", + "issue:44964", + "issue:45081", + "issue:45092", + "issue:45103", + "issue:45137", + "issue:45200", + "issue:45230", + "issue:45310", + "issue:45375", + "issue:45381", + "issue:45397", + "issue:45478", + "issue:45561", + "issue:45712", + "issue:45753" + ], + "soft_pairs": [ + "issue:42994|issue:43482", + "issue:45230|issue:45397", + "issue:28282|issue:30990", + "issue:44038|issue:44623", + "issue:42371|issue:45092", + "issue:43475|issue:45381", + "issue:43475|issue:43643", + "issue:41628|issue:44933", + "issue:45081|issue:45103", + "issue:43475|issue:44661", + "issue:42994|issue:43525", + "issue:45375|issue:45478", + "issue:43232|issue:44188", + "issue:45310|issue:45375", + "issue:43873|issue:44945", + "issue:42673|issue:43704", + "issue:41628|issue:44351", + "issue:42831|issue:44222", + "issue:44945|issue:45753", + "issue:44222|issue:45310", + "issue:43673|issue:44188", + "issue:36010|issue:41628", + "issue:42994|issue:44164", + "issue:42175|issue:45712", + "issue:43232|issue:44792", + "issue:44038|issue:45310", + "issue:42222|issue:45200", + "issue:43377|issue:44242", + "issue:44964|issue:45137", + "issue:43526|issue:45561", + "issue:43504|issue:45561", + "issue:44496|issue:45561", + "issue:42175|issue:44933" + ], + "prepared_review_unit_hash": "31e129cd019502eff7a2b714d03241f04bf1b75ec41558f0575e71ded47cc29c", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37264, + "estimated_input_tokens": 9316, + "estimated_eval_tokens": 18888 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is still mostly heterogeneous, but it does contain more than one duplicate-looking pair: issue:42673/43704 is a clear same-bug duplicate pair, and issue:45310/45478 also have an identical title and look like a near-exact duplicate pair. So this is not just a single duplicate pair cluster, even though the rest of the items remain mixed.", + "confidence": 0.92, + "canonical_issue_reason": "issue:42673 is the cleanest canonical anchor for the cluster\u2019s strongest duplicate pair because it matches issue:43704 on the exact VRAM-leak title and same multi-threaded dataloader symptom. The cluster also contains another identical-title pair (issue:45310/45478), but that appears to be a separate duplicate subgroup rather than the main anchor.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:42673 is still the best representative issue because it anchors the most clearly duplicated bug report in the packet and is the strongest evidence of an actual same-bug cluster. The other items are too mixed, and the 45310/45478 pair looks like a separate duplicate pocket rather than a better overall cluster representative.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42994", + "right": "issue:43482", + "accept": false, + "reason": "Quantized model saving failure and GGUF loading failure are different code paths and different bug classes." + }, + { + "left": "issue:45230", + "right": "issue:45397", + "accept": false, + "reason": "A placeholder bug report is not the same issue as the Gemma-4 zero3 from_pretrained failure." + }, + { + "left": "issue:28282", + "right": "issue:30990", + "accept": false, + "reason": "PyTorch missing import error and Sentence Transformers loading hang are unrelated symptoms." + }, + { + "left": "issue:44038", + "right": "issue:44623", + "accept": false, + "reason": "Qwen3-VL-Moe/transformers 5 bug and processor.save_pretrained missing files are distinct failures." + }, + { + "left": "issue:42371", + "right": "issue:45092", + "accept": false, + "reason": "TF32 behavior settings warning is unrelated to old InternVL2 remote-code meta-init incompatibility." + }, + { + "left": "issue:43475", + "right": "issue:45381", + "accept": false, + "reason": "SAM 3 vision encoder missing attribute and Qwen2.5-VL video position-id bug are different models and paths." + }, + { + "left": "issue:43475", + "right": "issue:43643", + "accept": false, + "reason": "One is a SAM3 attribute error; the other is a generic trust_remote_code config-field omission." + }, + { + "left": "issue:41628", + "right": "issue:44933", + "accept": false, + "reason": "Both mention imports, but they are different missing symbols from different modules." + }, + { + "left": "issue:45081", + "right": "issue:45103", + "accept": false, + "reason": "Mistral tokenizer regex crash and auto_docstring annotation crash are unrelated." + }, + { + "left": "issue:43475", + "right": "issue:44661", + "accept": false, + "reason": "SAM3 encoder attribute failure is not the same as add-new-model-like failing inside TOKENIZER_MAPPING_NAMES." + }, + { + "left": "issue:42994", + "right": "issue:43525", + "accept": false, + "reason": "Quantized model saving failed and Llama4Config missing pad_token_id are different bugs." + }, + { + "left": "issue:45375", + "right": "issue:45478", + "accept": false, + "reason": "They are clearly related by title, but the available evidence still points to a broader config-field omission versus a from_pretrained error, so I would not merge them without stronger same-code-path evidence." + }, + { + "left": "issue:43232", + "right": "issue:44188", + "accept": false, + "reason": "Generation kwargs update after sync_gpus and attention-kernel divergence under torch.compile are different code-path issues." + }, + { + "left": "issue:45310", + "right": "issue:45375", + "accept": false, + "reason": "A generic Qwen3.5-MoE from_pretrained error is not the same concrete bug as the missing deepstack_visual_indexes field." + }, + { + "left": "issue:43873", + "right": "issue:44945", + "accept": false, + "reason": "Quantization/offloading behavior and pipeline-parallelism output corruption are separate problems." + }, + { + "left": "issue:42673", + "right": "issue:43704", + "accept": true, + "reason": "These are the same Qwen3ForCausalLM VRAM-leak report with identical titles and the same multi-threaded dataloader symptom." + }, + { + "left": "issue:41628", + "right": "issue:44351", + "accept": false, + "reason": "AutoImageProcessor import failure and HybridCache import failure are unrelated missing-symbol bugs." + }, + { + "left": "issue:42831", + "right": "issue:44222", + "accept": false, + "reason": "FineGrainedFP8 accuracy drift is not the same issue as FP8 save_pretrained for MoE." + }, + { + "left": "issue:44945", + "right": "issue:45753", + "accept": false, + "reason": "Pipeline-parallelism output mismatch and an infinite loop on a specific image are different behaviors." + }, + { + "left": "issue:44222", + "right": "issue:45310", + "accept": false, + "reason": "FP8 save_pretrained for MoE and Qwen3.5-MoE from_pretrained failure do not describe the same concrete bug." + }, + { + "left": "issue:43673", + "right": "issue:44188", + "accept": false, + "reason": "Chunked-prefill GenerationMixin cache issue and torch.compile branching divergence are different generation bugs." + }, + { + "left": "issue:36010", + "right": "issue:41628", + "accept": false, + "reason": "GenerationMixin import failure and AutoImageProcessor import failure are different missing imports." + }, + { + "left": "issue:42994", + "right": "issue:44164", + "accept": false, + "reason": "Quantized model saving and extra_state save/from_pretrained handling are related only broadly, not the same bug." + }, + { + "left": "issue:42175", + "right": "issue:45712", + "accept": false, + "reason": "Missing TensorFlow in the torch extra and dummy_pt_objects leaking into dir(transformers) are separate packaging issues." + }, + { + "left": "issue:43232", + "right": "issue:44792", + "accept": false, + "reason": "Generation state update and a janus image-generation test failure are not the same underlying issue." + }, + { + "left": "issue:43526", + "right": "issue:45561", + "accept": false, + "reason": "BeitImageProcessorFast label reduction and xdist captured_info.txt race are unrelated." + }, + { + "left": "issue:43504", + "right": "issue:45561", + "accept": false, + "reason": "BEiT pretrained preset legacy-field handling and the xdist race are different bugs." + }, + { + "left": "issue:44496", + "right": "issue:45137", + "accept": false, + "reason": "Unrecognized model/config loading and DeepSpeed ZeRO3 deque underflow are unrelated." + }, + { + "left": "issue:43377", + "right": "issue:44242", + "accept": false, + "reason": "MIMI batched-vs-single output mismatch and load-balancing loss omission are different model issues." + }, + { + "left": "issue:44964", + "right": "issue:45137", + "accept": false, + "reason": "Phi-4 multimodal loading failure and DeepSpeed ZeRO3 deque error are unrelated." + }, + { + "left": "issue:43526", + "right": "issue:45561", + "accept": false, + "reason": "The label-reduction bug is unrelated to the xdist captured_info.txt race." + }, + { + "left": "issue:43504", + "right": "issue:45561", + "accept": false, + "reason": "Legacy-field loading failure in BEiT is unrelated to the xdist race." + }, + { + "left": "issue:42175", + "right": "issue:44933", + "accept": false, + "reason": "TF backend packaging and image_utils import issues are different problems." + }, + { + "left": "issue:44038", + "right": "issue:45310", + "accept": false, + "reason": "Qwen3-VL-Moe transformer bug and Qwen3.5-MoE from_pretrained error are different model/version failures." + }, + { + "left": "issue:42831", + "right": "issue:44222", + "accept": false, + "reason": "FP8 accuracy regression and FP8 MoE save_pretrained failure are not the same code path." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The cluster summary is a bit too strong: the packet contains at least two obvious duplicate-looking pairs with identical titles, not just issue:42673/43704. In particular, issue:45310 and issue:45478 share the exact same title, so saying there is \u2018only one clear duplicate pair\u2019 is not fully grounded. The conservative false soft-edge calls are mostly fine, but the overall narrative should acknowledge this additional near-exact duplicate instead of presenting the cluster as having a single clear duplicate pair." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 37, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:22355", + "issue:28282", + "issue:30333", + "issue:30990", + "issue:33290", + "issue:34689", + "issue:35141", + "issue:36331", + "issue:41628", + "issue:42757", + "issue:42831", + "issue:42898", + "issue:43116", + "issue:43278", + "issue:43479", + "issue:43644", + "issue:43723", + "issue:43784", + "issue:43824", + "issue:43874", + "issue:43976", + "issue:44060", + "issue:44079", + "issue:44188", + "issue:44190", + "issue:44206", + "issue:44222", + "issue:44295", + "issue:44297", + "issue:44373", + "issue:44448", + "issue:44484", + "issue:44568", + "issue:44623", + "issue:44734", + "issue:44779", + "issue:44869", + "issue:44933", + "issue:45092", + "issue:45103", + "issue:45125", + "issue:45137", + "issue:45310", + "issue:45356", + "issue:45405", + "issue:45656", + "issue:45712", + "issue:45812" + ], + "soft_pairs": [ + "issue:42898|issue:44568", + "issue:43479|issue:44869", + "issue:30990|issue:35141", + "issue:43976|issue:44933", + "issue:43479|issue:44079", + "issue:44623|issue:45310", + "issue:43784|issue:43824", + "issue:44188|issue:44373", + "issue:43278|issue:45137", + "issue:42898|issue:45812", + "issue:44734|issue:45103", + "issue:30333|issue:33290", + "issue:43723|issue:44568", + "issue:42898|issue:45356", + "issue:43874|issue:45092", + "issue:44060|issue:45125", + "issue:42898|issue:44779", + "issue:45405|issue:45712", + "issue:42898|issue:44448", + "issue:44295|issue:45656", + "issue:44933|issue:45712", + "issue:44222|issue:44297", + "issue:44297|issue:45656", + "issue:43116|issue:44190", + "issue:43479|issue:43644", + "issue:44222|issue:44623", + "issue:42831|issue:44623", + "issue:44206|issue:45103", + "issue:41628|issue:42757", + "issue:30333|issue:36331", + "issue:22355|issue:28282", + "issue:30333|issue:34689", + "issue:44297|issue:45310", + "issue:44222|issue:44484", + "issue:28282|issue:30333" + ], + "prepared_review_unit_hash": "38d473efd81445977bb8adeac31191b387dd5836491d31de943946f9a3df77c2", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 35, + "serialized_chars": 38240, + "estimated_input_tokens": 9560, + "estimated_eval_tokens": 19376 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a heterogeneous cluster of mostly unrelated Transformers issues: tokenizer regressions, model-loading/import failures, training/runtime bugs, save/load serialization issues, and a few doc/config problems. None of the soft edges look like true duplicates, so all are rejected.", + "confidence": 0.88, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "issue:45812 is the strongest standalone report in the set: it is concrete, reproducible, and describes a high-impact silent regression in AutoTokenizer for Granite models.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42898", + "right": "issue:44568", + "accept": false, + "reason": "Both are tokenizer-related regressions, but one is about clean_up_tokenization_spaces behavior and the other about add_special_tokens not adding BOS/EOS; different code paths and symptoms." + }, + { + "left": "issue:43479", + "right": "issue:44869", + "accept": false, + "reason": "Phi4MultimodalConfig default initialization and Whisper word-timestamp decode crashing are unrelated bugs in different subsystems." + }, + { + "left": "issue:30990", + "right": "issue:35141", + "accept": false, + "reason": "Sentence Transformers loading hangs and embedding resize reinitialization are distinct failures with no shared code-path." + }, + { + "left": "issue:43976", + "right": "issue:44933", + "accept": false, + "reason": "Python-version compatibility/import support is unrelated to the missing image_utils import symbol bug." + }, + { + "left": "issue:43479", + "right": "issue:44079", + "accept": false, + "reason": "Config initialization for Phi4Multimodal is unrelated to ModelOutput key assignment behavior." + }, + { + "left": "issue:44623", + "right": "issue:45310", + "accept": false, + "reason": "Both mention loading/saving, but one is missing processor save artifacts and the other is a from_pretrained failure for Qwen3.5 MoE; different root causes." + }, + { + "left": "issue:43784", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but they involve different packages/models and different missing symbols, so they are not the same bug." + }, + { + "left": "issue:44188", + "right": "issue:44373", + "accept": false, + "reason": "A numerical/kernel divergence under torch.compile is not the same as a wrong docstring." + }, + { + "left": "issue:43278", + "right": "issue:45137", + "accept": false, + "reason": "Embedding dtype drift during eval and a DeepSpeed ZeRO3 deque IndexError are unrelated runtime bugs." + }, + { + "left": "issue:42898", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer regressions, but one concerns cleanup-space handling and the other wrong token IDs for Granite models; different failure mode and model scope." + }, + { + "left": "issue:44734", + "right": "issue:45103", + "accept": false, + "reason": "KV-cache continuation tensor indexing in transformers serve is unrelated to the auto_docstring AttributeError." + }, + { + "left": "issue:30333", + "right": "issue:33290", + "accept": false, + "reason": "MLFlow reporting on failed training and DeepSpeed OOM with Adafactor are separate training issues." + }, + { + "left": "issue:43723", + "right": "issue:44568", + "accept": false, + "reason": "Tokenizer loading in v5 and add_special_tokens BOS/EOS behavior are different tokenizer problems." + }, + { + "left": "issue:42898", + "right": "issue:45356", + "accept": false, + "reason": "These are both tokenizer regressions, but one is a generic v5 cleanup-space change and the other is Kimi-K2.5 codec handling/fix_mistral_regex behavior." + }, + { + "left": "issue:43874", + "right": "issue:45092", + "accept": false, + "reason": "A missing image patch count method and a remote-code checkpoint meta-init incompatibility are different multimodal model bugs." + }, + { + "left": "issue:44060", + "right": "issue:45125", + "accept": false, + "reason": "Incorrect tied-weights warnings for Qwen3-Next and missing _tp_plan for Qwen3.5 MoE are separate model-specific issues." + }, + { + "left": "issue:42898", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different models and different logic paths." + }, + { + "left": "issue:45405", + "right": "issue:45712", + "accept": false, + "reason": "A dependency version bump is unrelated to cleanup of leftover dummy classes and repo-check failures." + }, + { + "left": "issue:42898", + "right": "issue:44448", + "accept": false, + "reason": "One is a tokenizer cleanup change; the other is a model output difference for pegasus between major versions." + }, + { + "left": "issue:44295", + "right": "issue:45656", + "accept": false, + "reason": "Reading a registered position_ids buffer and a double optimizer-step in DeepSpeed are unrelated training/runtime issues." + }, + { + "left": "issue:44933", + "right": "issue:45712", + "accept": false, + "reason": "Missing image_utils import and leftover dummy classes are different import/cleanup problems." + }, + { + "left": "issue:44222", + "right": "issue:44297", + "accept": false, + "reason": "FP8 MoE save_pretrained and tokenizer_class metadata mismatch are both serialization-related, but they affect different object types and failure modes." + }, + { + "left": "issue:44297", + "right": "issue:45656", + "accept": false, + "reason": "Tokenizer save metadata mismatch and double optimizer steps are unrelated." + }, + { + "left": "issue:43116", + "right": "issue:44190", + "accept": false, + "reason": "Multi-label classification example behavior and local dataset loading in a different example script are not the same bug." + }, + { + "left": "issue:43479", + "right": "issue:43644", + "accept": false, + "reason": "Phi4MultimodalConfig default config initialization is unrelated to non-persistent buffer initialization in v5.0.0." + }, + { + "left": "issue:44222", + "right": "issue:44623", + "accept": false, + "reason": "Both involve save_pretrained, but one is FP8 MoE serialization and the other is missing processor files; not the same code-path." + }, + { + "left": "issue:42831", + "right": "issue:44623", + "accept": false, + "reason": "FineGrainedFP8 accuracy drift is not the same as processor save_pretrained missing files." + }, + { + "left": "issue:44206", + "right": "issue:45103", + "accept": false, + "reason": "An audio feature extractor center-arg regression and an auto_docstring AttributeError are unrelated." + }, + { + "left": "issue:41628", + "right": "issue:42757", + "accept": false, + "reason": "These are separate import errors from different modules/symbols, not one shared missing import bug." + }, + { + "left": "issue:30333", + "right": "issue:36331", + "accept": false, + "reason": "MLFlow reporting on failed jobs and the compute_loss signature change are distinct trainer issues." + }, + { + "left": "issue:22355", + "right": "issue:28282", + "accept": false, + "reason": "Missing transformers.onnx and missing PyTorch for AutoModel are different environment/import failures." + }, + { + "left": "issue:30333", + "right": "issue:34689", + "accept": false, + "reason": "MLFlow reporting on failed training is unrelated to the Llama 3.2 Vision model-loading regression." + }, + { + "left": "issue:44297", + "right": "issue:45310", + "accept": false, + "reason": "Tokenizer save metadata mismatch and Qwen3.5 from_pretrained failure are different bugs." + }, + { + "left": "issue:44222", + "right": "issue:44484", + "accept": false, + "reason": "FP8 MoE save_pretrained serialization and a question about max_shard_size documentation are not duplicates." + }, + { + "left": "issue:28282", + "right": "issue:30333", + "accept": false, + "reason": "A missing PyTorch dependency import error is unrelated to MLFlow not being reported after failed training." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 38, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:30333", + "issue:30990", + "issue:35141", + "issue:38175", + "issue:39401", + "issue:41628", + "issue:42757", + "issue:42831", + "issue:42886", + "issue:42994", + "issue:43066", + "issue:43122", + "issue:43232", + "issue:43479", + "issue:43576", + "issue:43723", + "issue:43756", + "issue:43784", + "issue:43824", + "issue:44038", + "issue:44060", + "issue:44261", + "issue:44263", + "issue:44295", + "issue:44360", + "issue:44373", + "issue:44442", + "issue:44483", + "issue:44485", + "issue:44514", + "issue:44521", + "issue:44568", + "issue:44623", + "issue:44829", + "issue:44871", + "issue:44913", + "issue:44977", + "issue:44991", + "issue:45092", + "issue:45103", + "issue:45278", + "issue:45310", + "issue:45357", + "issue:45406", + "issue:45593", + "issue:45753", + "issue:45834", + "issue:46002" + ], + "soft_pairs": [ + "issue:43723|issue:44977", + "issue:43232|issue:45834", + "issue:30333|issue:35141", + "issue:44263|issue:45092", + "issue:44483|issue:45406", + "issue:44373|issue:44485", + "issue:44295|issue:44623", + "issue:44514|issue:45103", + "issue:44263|issue:44360", + "issue:44261|issue:45092", + "issue:45092|issue:45357", + "issue:43723|issue:44871", + "issue:43723|issue:44991", + "issue:30990|issue:38175", + "issue:44829|issue:46002", + "issue:43723|issue:44442", + "issue:43479|issue:44913", + "issue:41628|issue:43824", + "issue:42831|issue:45310", + "issue:44263|issue:44373", + "issue:42831|issue:44038", + "issue:39401|issue:44568", + "issue:42757|issue:45278", + "issue:42886|issue:42994", + "issue:43232|issue:45593", + "issue:43756|issue:44060", + "issue:42831|issue:45753", + "issue:43576|issue:43784", + "issue:42886|issue:43066", + "issue:43122|issue:44521" + ], + "prepared_review_unit_hash": "402a7d859d8062df05f71fd5a55a4693eb695bb9fe790adf2fbf14a966fac08d", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36434, + "estimated_input_tokens": 9109, + "estimated_eval_tokens": 18474 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is dominated by Transformers v5 regression reports, but the soft-linked pairs mostly describe different code paths, models, or failure modes rather than true duplicates. No soft edge looks safe to merge.", + "confidence": 0.38, + "canonical_issue_reason": "issue:43723 is the broadest and most central report here: a generic AutoTokenizer/from_pretrained loading failure in v5, which best represents the tokenization/loading regression subgroup in this cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43723 is the best overall issue candidate because it is the most generic and reusable summary of the cluster\u2019s recurring tokenizer-loading regressions; other issues are either model-specific or about unrelated subsystems.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43723", + "right": "issue:44977", + "accept": false, + "reason": "Both mention v5-era generation/tokenizer problems, but one is a generic tokenizer load failure and the other is a Qwen3.5 flash-attention generation bug; different underlying paths." + }, + { + "left": "issue:43232", + "right": "issue:45834", + "accept": false, + "reason": "Generation state update after sync_gpus is unrelated to Kosmos2.5 OCR long-input indexing; no shared bug mechanism." + }, + { + "left": "issue:30333", + "right": "issue:35141", + "accept": false, + "reason": "MLflow job status reporting and output embedding reinitialization are unrelated." + }, + { + "left": "issue:44263", + "right": "issue:45092", + "accept": false, + "reason": "torch.split return handling in GlmMoeDsaIndexer is not the same bug as InternVL2 meta-initialization compatibility." + }, + { + "left": "issue:44483", + "right": "issue:45406", + "accept": false, + "reason": "Chat/completions request handling and Gemma4Processor _tokenizer initialization are distinct failures." + }, + { + "left": "issue:44373", + "right": "issue:44485", + "accept": false, + "reason": "A wrong docstring for position_ids is not the same as the GLM-5 RoPE implementation bug." + }, + { + "left": "issue:44295", + "right": "issue:44623", + "accept": false, + "reason": "Buffer/readback of position_ids and processor.save_pretrained file emission are separate serialization issues." + }, + { + "left": "issue:44514", + "right": "issue:45103", + "accept": false, + "reason": "apply_chat_template batching crash and auto_docstring annotation handling are unrelated." + }, + { + "left": "issue:44263", + "right": "issue:44360", + "accept": false, + "reason": "Both are in the DSA/indexer area, but one is about torch.split outputs and the other about a missing ReLU; different code defects." + }, + { + "left": "issue:44261", + "right": "issue:45092", + "accept": false, + "reason": "MLA q_a_layernorm epsilon precision and InternVL2 meta-init incompatibility do not point to the same change." + }, + { + "left": "issue:45092", + "right": "issue:45357", + "accept": false, + "reason": "Remote-code meta initialization incompatibility and Qwen3.5 incorrect visual encoder save keys are different issues." + }, + { + "left": "issue:43723", + "right": "issue:44871", + "accept": false, + "reason": "Tokenizer loading failure and eos_token_id mismatch are both tokenizer-related, but they are different concrete bugs." + }, + { + "left": "issue:43723", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer loading regressions, but the reports are model-specific and do not clearly share the same failing path." + }, + { + "left": "issue:30990", + "right": "issue:38175", + "accept": false, + "reason": "Sentence Transformers loading hang and SigLIP2 zero-probabilities are unrelated." + }, + { + "left": "issue:44829", + "right": "issue:46002", + "accept": false, + "reason": "Flash-attention-3 training degeneration is not the same as NaNs appearing after checkout/init in classification heads." + }, + { + "left": "issue:43723", + "right": "issue:44442", + "accept": false, + "reason": "Generic AutoTokenizer load failure and FastSpeech2ConformerTokenizer load failure are different tokenizer implementations and likely different causes." + }, + { + "left": "issue:43479", + "right": "issue:44913", + "accept": false, + "reason": "Incorrect default config initialization and rotary_pct not persisting on reload are separate config serialization bugs." + }, + { + "left": "issue:41628", + "right": "issue:43824", + "accept": false, + "reason": "Missing AutoImageProcessor import and missing Qwen2_5_VLForConditionalGeneration import are different symbol/export issues." + }, + { + "left": "issue:42831", + "right": "issue:45310", + "accept": false, + "reason": "FineGrainedFP8 accuracy loss and Qwen3.5 moe from_pretrained failure are not the same bug." + }, + { + "left": "issue:44263", + "right": "issue:44373", + "accept": false, + "reason": "DSA split return values and a wrong docstring for position_ids are unrelated." + }, + { + "left": "issue:42831", + "right": "issue:44038", + "accept": false, + "reason": "FP8 accuracy regression and Qwen3-VL-Moe bug are different model/runtime issues." + }, + { + "left": "issue:39401", + "right": "issue:44568", + "accept": false, + "reason": "Wrong offset_mapping and missing BOS/EOS on add_special_tokens are both tokenizer behavior problems, but they affect different behaviors and code paths." + }, + { + "left": "issue:42757", + "right": "issue:45278", + "accept": false, + "reason": "A specific is_offline_mode import error and a broad set of import errors after upgrade are related in theme but not clearly the same underlying defect." + }, + { + "left": "issue:42886", + "right": "issue:42994", + "accept": false, + "reason": "Offline tokenizer cache loading and quantized model saving failure are unrelated." + }, + { + "left": "issue:43232", + "right": "issue:45593", + "accept": false, + "reason": "Generation kwargs update after sync_gpus is unrelated to D-FINE auxiliary-loss behavior." + }, + { + "left": "issue:43756", + "right": "issue:44060", + "accept": false, + "reason": "RoPE layer selection in Smollm3 and tied-weights warning behavior in Qwen3-Next are different model configuration bugs." + }, + { + "left": "issue:42831", + "right": "issue:45753", + "accept": false, + "reason": "FP8 accuracy regression and Qwen3_5 infinite loop on one image are different failures." + }, + { + "left": "issue:43576", + "right": "issue:43784", + "accept": false, + "reason": "Broken transformers env command and sentence-transformers import NameError are distinct issues, even if both are environment/version related." + }, + { + "left": "issue:42886", + "right": "issue:43066", + "accept": false, + "reason": "Offline cache tokenizer loading and wrong tokenizer decoder type in v5 are different tokenizer problems." + }, + { + "left": "issue:43122", + "right": "issue:44521", + "accept": false, + "reason": "Different tokenization across versions and zero assistant masks in multimodal apply_chat_template are not the same bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 39, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:22355", + "issue:28218", + "issue:29127", + "issue:30990", + "issue:33357", + "issue:36296", + "issue:39401", + "issue:41628", + "issue:42371", + "issue:42907", + "issue:43116", + "issue:43122", + "issue:43232", + "issue:43408", + "issue:43452", + "issue:43475", + "issue:43646", + "issue:43784", + "issue:43824", + "issue:43937", + "issue:43976", + "issue:44038", + "issue:44117", + "issue:44222", + "issue:44261", + "issue:44263", + "issue:44295", + "issue:44297", + "issue:44351", + "issue:44360", + "issue:44373", + "issue:44485", + "issue:44488", + "issue:44514", + "issue:44521", + "issue:44625", + "issue:44704", + "issue:44743", + "issue:44779", + "issue:44829", + "issue:44945", + "issue:44964", + "issue:45200", + "issue:45278", + "issue:45290", + "issue:45600", + "issue:46002" + ], + "soft_pairs": [ + "issue:22355|issue:28218", + "issue:44117|issue:44704", + "issue:42371|issue:44263", + "issue:43646|issue:46002", + "issue:30990|issue:33357", + "issue:44295|issue:44297", + "issue:44704|issue:45290", + "issue:43232|issue:44743", + "issue:42907|issue:44945", + "issue:43122|issue:44514", + "issue:41628|issue:45278", + "issue:43976|issue:44945", + "issue:39401|issue:44488", + "issue:44261|issue:44373", + "issue:43784|issue:45600", + "issue:42907|issue:43937", + "issue:43408|issue:43475", + "issue:44360|issue:44485", + "issue:39401|issue:44521", + "issue:43116|issue:45200", + "issue:44625|issue:44829", + "issue:44038|issue:44297", + "issue:42907|issue:44964", + "issue:29127|issue:30990", + "issue:43122|issue:44779", + "issue:44038|issue:44295", + "issue:36296|issue:43116", + "issue:42907|issue:43452", + "issue:43824|issue:44351", + "issue:44038|issue:44222" + ], + "prepared_review_unit_hash": "2f3d75c94c77c12867acd9c6f0e800ec7457990e439c4672f6b54d53832c4d71", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 35654, + "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 a loose grab-bag of unrelated Transformers issues, not a duplicate set. The soft pairs share broad themes like imports, tokenization, multimodal/chat-template behavior, or version regressions, but each points to a different concrete bug or model path, so none should be merged.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:22355", + "right": "issue:28218", + "accept": false, + "reason": "Different subsystems: ONNX import path vs tokenizer spacing after added token." + }, + { + "left": "issue:44117", + "right": "issue:44704", + "accept": false, + "reason": "Tokenizer mapping fallback bug vs AutoProcessor kwarg forwarding; unrelated code paths." + }, + { + "left": "issue:42371", + "right": "issue:44263", + "accept": false, + "reason": "TF32 settings warning is unrelated to the torch.split return-value issue." + }, + { + "left": "issue:43646", + "right": "issue:46002", + "accept": false, + "reason": "Generic Transformers 5.0 custom-init breakage vs NaNs in classification heads; not the same failure." + }, + { + "left": "issue:30990", + "right": "issue:33357", + "accept": false, + "reason": "Both are loading/runtime problems, but one is SentenceTransformers stuck loading and the other is a Mac bus error on a CLIP model." + }, + { + "left": "issue:44295", + "right": "issue:44297", + "accept": false, + "reason": "One concerns reading position_ids as a buffer, the other tokenizer_class saved in tokenizer_config.json." + }, + { + "left": "issue:44704", + "right": "issue:45290", + "accept": false, + "reason": "Both touch chat/processor APIs only loosely; cached_file kwargs forwarding is not the same as assistant messages with tool calls crashing." + }, + { + "left": "issue:43232", + "right": "issue:44743", + "accept": false, + "reason": "Generation kwarg update after sync_gpus is separate from recurrent state reset in modular_qwen3_5." + }, + { + "left": "issue:42907", + "right": "issue:44945", + "accept": false, + "reason": "Dequantized save failure for specific models vs incorrect output under pipeline parallelism." + }, + { + "left": "issue:43122", + "right": "issue:44514", + "accept": false, + "reason": "Tokenizer behavior change across versions is not the same as Qwen2_5_VLProcessor batched chat_template crash." + }, + { + "left": "issue:41628", + "right": "issue:45278", + "accept": false, + "reason": "Both are import-related, but one is a missing AutoImageProcessor symbol and the other is broad import regressions after upgrade." + }, + { + "left": "issue:43976", + "right": "issue:44945", + "accept": false, + "reason": "Python version compatibility issue is unrelated to pipeline parallelism output corruption." + }, + { + "left": "issue:39401", + "right": "issue:44488", + "accept": false, + "reason": "Qwen3 offset_mapping bug and inability to load cjvt/sleng-bert are different tokenizer/model-loading issues." + }, + { + "left": "issue:44261", + "right": "issue:44373", + "accept": false, + "reason": "Precision error from missing rms_norm_eps is not the same as a wrong docstring for position_ids." + }, + { + "left": "issue:43784", + "right": "issue:45600", + "accept": false, + "reason": "Sentence-transformers import NameError and CI breakage from removed Sam3LiteText configs are unrelated." + }, + { + "left": "issue:42907", + "right": "issue:43937", + "accept": false, + "reason": "Both are model/config problems, but one is dequantized save_pretrained and the other is invalid GenerationConfig for GLM-5." + }, + { + "left": "issue:43408", + "right": "issue:43475", + "accept": false, + "reason": "Both mention SAM 3 video, but one is a wrong model-type warning and the other is a missing attribute on Sam3VisionEncoderOutput." + }, + { + "left": "issue:44360", + "right": "issue:44485", + "accept": false, + "reason": "DSA indexer ReLU concern and GLM-5 RoPE implementation are separate algorithmic issues." + }, + { + "left": "issue:39401", + "right": "issue:44521", + "accept": false, + "reason": "Wrong offset_mapping for Qwen3 is different from all-zero assistant_masks in multimodal chat templating." + }, + { + "left": "issue:43116", + "right": "issue:45200", + "accept": false, + "reason": "Multi-label classification example bug is unrelated to Gemma 4 mm_token_type_ids defaults." + }, + { + "left": "issue:44625", + "right": "issue:44829", + "accept": false, + "reason": "num_labels propagation in Qwen3.5 is unrelated to flash_attention_3 causing degenerate training." + }, + { + "left": "issue:44038", + "right": "issue:44297", + "accept": false, + "reason": "Qwen3-VL-Moe bug and tokenizer_class mismatch in saved config are different components." + }, + { + "left": "issue:42907", + "right": "issue:44964", + "accept": false, + "reason": "Dequantized save failure is unrelated to loading Phi-4 multimodal with latest transformers." + }, + { + "left": "issue:29127", + "right": "issue:30990", + "accept": false, + "reason": "Clarity of a LayoutLMv3 error message vs SentenceTransformers loading hang are not the same bug." + }, + { + "left": "issue:43122", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions across versions, but they affect different models and failure modes." + }, + { + "left": "issue:44038", + "right": "issue:44295", + "accept": false, + "reason": "Qwen3-VL-Moe bug and position_ids buffer read error are unrelated." + }, + { + "left": "issue:36296", + "right": "issue:43116", + "accept": false, + "reason": "Tensor parallel training bug and multi-label classification example bug do not share the same code path." + }, + { + "left": "issue:42907", + "right": "issue:43452", + "accept": false, + "reason": "Saving dequantized models and gguf_file breaking from_pretrained are separate loading/saving paths." + }, + { + "left": "issue:43824", + "right": "issue:44351", + "accept": false, + "reason": "Missing Qwen2_5_VLForConditionalGeneration import and missing HybridCache import are different API-export regressions." + }, + { + "left": "issue:44038", + "right": "issue:44222", + "accept": false, + "reason": "Qwen3-VL-Moe issue is not the same as FP8 save_pretrained behavior for moe models." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 40, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:22355", + "issue:29127", + "issue:33453", + "issue:38617", + "issue:39401", + "issue:42222", + "issue:42371", + "issue:42673", + "issue:42757", + "issue:42907", + "issue:43064", + "issue:43066", + "issue:43116", + "issue:43122", + "issue:43278", + "issue:43441", + "issue:43493", + "issue:43611", + "issue:43673", + "issue:43784", + "issue:43825", + "issue:43827", + "issue:43874", + "issue:43937", + "issue:43976", + "issue:43992", + "issue:44038", + "issue:44075", + "issue:44117", + "issue:44222", + "issue:44351", + "issue:44485", + "issue:44704", + "issue:44779", + "issue:44936", + "issue:44945", + "issue:44977", + "issue:45072", + "issue:45083", + "issue:45216", + "issue:45278", + "issue:45357", + "issue:45446", + "issue:45478", + "issue:45563", + "issue:45588", + "issue:45753", + "issue:46002" + ], + "soft_pairs": [ + "issue:45216|issue:45478", + "issue:42757|issue:44351", + "issue:43278|issue:44945", + "issue:43122|issue:44977", + "issue:43784|issue:45446", + "issue:43493|issue:45753", + "issue:42907|issue:44222", + "issue:43992|issue:44704", + "issue:43066|issue:44779", + "issue:38617|issue:45278", + "issue:42673|issue:45072", + "issue:43937|issue:44485", + "issue:22355|issue:29127", + "issue:43611|issue:46002", + "issue:42371|issue:43874", + "issue:43976|issue:45446", + "issue:42222|issue:43441", + "issue:43064|issue:43673", + "issue:43116|issue:43493", + "issue:44075|issue:44936", + "issue:45357|issue:45478", + "issue:44038|issue:45478", + "issue:43116|issue:43825", + "issue:45083|issue:45446", + "issue:43937|issue:45563", + "issue:44117|issue:45588", + "issue:33453|issue:39401", + "issue:39401|issue:44779", + "issue:43827|issue:45446" + ], + "prepared_review_unit_hash": "9fddbf5d30f6cfe2871aaaa0f90cb178a98383a375d2ad3f4767451eff33c063", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 35924, + "estimated_input_tokens": 8981, + "estimated_eval_tokens": 18218 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a noisy cluster of mostly unrelated Transformers issues; the soft pairs share broad themes like v5 regressions or import/load failures, but none look like the same concrete bug. Issue 45478 is the most specific representative, but the set is not a true duplicate cluster.", + "confidence": 0.83, + "canonical_issue_reason": "issue:45478 is the clearest, most actionable single bug report in the set (a narrow Qwen3.5 Moe from_pretrained failure), so it is the best representative if a canonical issue must be chosen.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45478 has the strongest standalone triage value: specific model, specific failing path, and a plausible repro scope. Still, it is only a weak canonical candidate because the surrounding issues are largely unrelated.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45216", + "right": "issue:45478", + "accept": false, + "reason": "Both involve Qwen3.5 and save/load, but one is save_pretrained checkpoint contents and the other is from_pretrained loading failure; different concrete bugs." + }, + { + "left": "issue:42757", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they are for different missing symbols from different packages and likely unrelated failure points." + }, + { + "left": "issue:43278", + "right": "issue:44945", + "accept": false, + "reason": "Embedding dtype drift in eval is unrelated to incorrect LLM output under pipeline parallelism." + }, + { + "left": "issue:43122", + "right": "issue:44977", + "accept": false, + "reason": "Tokenizer output drift across versions is not the same as a flash-attention generation failure." + }, + { + "left": "issue:43784", + "right": "issue:45446", + "accept": false, + "reason": "Different import/version-check failures in different code paths; no shared underlying bug." + }, + { + "left": "issue:43493", + "right": "issue:45753", + "accept": false, + "reason": "SigLIP2 implementation mismatch and a Qwen3.5 image infinite loop are different model-specific problems." + }, + { + "left": "issue:42907", + "right": "issue:44222", + "accept": false, + "reason": "Both mention saving model states, but dequantized Ministral/Devstral saving and FP8 MoE save_pretrained are distinct bugs." + }, + { + "left": "issue:43992", + "right": "issue:44704", + "accept": false, + "reason": "Both are from_pretrained-related, but one misses a model weight and the other drops kwargs in AutoProcessor; different code paths." + }, + { + "left": "issue:43066", + "right": "issue:44779", + "accept": false, + "reason": "Tokenizer decoder-type regression and DeepSeek tokenization corruption are separate tokenizer bugs." + }, + { + "left": "issue:38617", + "right": "issue:45278", + "accept": false, + "reason": "45278 is a broad bundle of import errors; 38617 is one specific missing symbol. Too generic to treat as the same bug." + }, + { + "left": "issue:42673", + "right": "issue:45072", + "accept": false, + "reason": "VRAM leakage in threaded Qwen3ForCausalLM usage is unrelated to dtype mismatches in inference." + }, + { + "left": "issue:43937", + "right": "issue:44485", + "accept": false, + "reason": "GenerationConfig validation failure and GLM-5 RoPE discussion are not the same concrete defect." + }, + { + "left": "issue:22355", + "right": "issue:29127", + "accept": false, + "reason": "A missing module import and a LayoutLMv3 error-message clarity issue are unrelated." + }, + { + "left": "issue:43611", + "right": "issue:46002", + "accept": false, + "reason": "Loading models with base_model_prefix and NaNs in classification heads are different failure modes." + }, + { + "left": "issue:42371", + "right": "issue:43874", + "accept": false, + "reason": "TF32 settings guidance and a missing multimodal image-processor method are unrelated." + }, + { + "left": "issue:43976", + "right": "issue:45446", + "accept": false, + "reason": "Python version compatibility and a PyTorch version check bug are different release/platform issues." + }, + { + "left": "issue:42222", + "right": "issue:43441", + "accept": false, + "reason": "Broken ViTpose models and Ministral FlashAttention failures are separate model-specific regressions." + }, + { + "left": "issue:43064", + "right": "issue:43673", + "accept": false, + "reason": "Distributed optimizer-state corruption during training is unrelated to missing generation cache in chunked prefill." + }, + { + "left": "issue:43116", + "right": "issue:43493", + "accept": false, + "reason": "Empty multi-label classification outputs and SigLIP2 JAX parity issues are unrelated." + }, + { + "left": "issue:44075", + "right": "issue:44936", + "accept": false, + "reason": "Unused SGD args and trainer.evaluate() failing after train() are different trainer bugs." + }, + { + "left": "issue:45357", + "right": "issue:45478", + "accept": false, + "reason": "Both concern Qwen3.5 save/load behavior, but one is incorrect visual encoder keys on save and the other is a from_pretrained error; not the same bug." + }, + { + "left": "issue:44038", + "right": "issue:45478", + "accept": false, + "reason": "Both involve Qwen3.x/Moe models, but the reported failures are different and not clearly the same code-path regression." + }, + { + "left": "issue:43116", + "right": "issue:43825", + "accept": false, + "reason": "Example-script multi-label output and pipeline translation messaging are separate issues." + }, + { + "left": "issue:45083", + "right": "issue:45446", + "accept": false, + "reason": "A helper length-calculation bug in qwen3_omni_moe is unrelated to a PyTorch version-check import problem." + }, + { + "left": "issue:43937", + "right": "issue:45563", + "accept": false, + "reason": "Invalid GenerationConfig and stale num_return_sequences warning are different generation-layer issues." + }, + { + "left": "issue:44117", + "right": "issue:45588", + "accept": false, + "reason": "Tokenizer mapping None-handling and flash_attention s_aux=None crashes are unrelated." + }, + { + "left": "issue:33453", + "right": "issue:39401", + "accept": false, + "reason": "Tokenizer loading regression and Qwen3 tokenizer offset mapping bug are different tokenizer problems." + }, + { + "left": "issue:39401", + "right": "issue:44779", + "accept": false, + "reason": "Wrong offset_mapping and incorrect tokenization results are both tokenizer-related, but the concrete bugs and affected behavior differ." + }, + { + "left": "issue:43827", + "right": "issue:45446", + "accept": false, + "reason": "Docs still referencing pipeline() and an import/version-check bug are not the same underlying issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 41, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:38617", + "issue:39401", + "issue:41628", + "issue:42175", + "issue:42222", + "issue:42371", + "issue:42490", + "issue:42617", + "issue:42673", + "issue:42831", + "issue:43066", + "issue:43122", + "issue:43408", + "issue:43421", + "issue:43450", + "issue:43582", + "issue:43673", + "issue:43742", + "issue:43906", + "issue:43937", + "issue:43986", + "issue:44075", + "issue:44117", + "issue:44220", + "issue:44261", + "issue:44291", + "issue:44410", + "issue:44466", + "issue:44483", + "issue:44484", + "issue:44496", + "issue:44829", + "issue:44933", + "issue:44977", + "issue:45071", + "issue:45081", + "issue:45216", + "issue:45278", + "issue:45310", + "issue:45341", + "issue:45405", + "issue:45446", + "issue:45464", + "issue:45542", + "issue:45561", + "issue:45656", + "issue:46002" + ], + "soft_pairs": [ + "issue:44466|issue:46002", + "issue:43937|issue:44977", + "issue:42673|issue:43906", + "issue:45405|issue:45446", + "issue:44291|issue:45071", + "issue:43450|issue:43986", + "issue:45341|issue:45561", + "issue:44117|issue:45081", + "issue:38617|issue:41628", + "issue:44410|issue:45310", + "issue:39401|issue:43421", + "issue:44829|issue:45216", + "issue:44484|issue:45310", + "issue:42371|issue:44261", + "issue:42673|issue:43582", + "issue:42617|issue:45341", + "issue:43408|issue:43742", + "issue:44933|issue:45278", + "issue:42222|issue:45542", + "issue:42490|issue:42673", + "issue:42175|issue:42222", + "issue:43937|issue:44496", + "issue:38617|issue:44220", + "issue:44483|issue:45464", + "issue:42371|issue:42831", + "issue:43066|issue:43122", + "issue:44075|issue:45656", + "issue:43673|issue:45341" + ], + "prepared_review_unit_hash": "50d0b13290d23a48d6c78d322b2201c918755ec27402d802b5abd4139a3480a6", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35284, + "estimated_input_tokens": 8821, + "estimated_eval_tokens": 17898 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous: it mixes import errors, tokenizer regressions, model-specific bugs, training/optimizer issues, and testing-utils problems. None of the soft-similarity pairs look like the same concrete underlying bug or change, so no duplicates should be merged here.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "issue:43122 is the closest broad regression report in the set because it describes tokenizer behavior changing across versions, but the cluster is too mixed for it to serve as a true canonical issue.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44466", + "right": "issue:46002", + "accept": false, + "reason": "Different failure modes: tied-weight serialization/device-dependent save behavior vs NaNs during checkout/init." + }, + { + "left": "issue:43937", + "right": "issue:44977", + "accept": false, + "reason": "Unrelated model-specific generation issues: invalid GenerationConfig vs flash-attention generation failure." + }, + { + "left": "issue:42673", + "right": "issue:43906", + "accept": false, + "reason": "One is a VRAM leak in multithreaded dataloading; the other is just an isolated reproduction of a different issue." + }, + { + "left": "issue:45405", + "right": "issue:45446", + "accept": false, + "reason": "PEFT version pinning/release mismatch is unrelated to the flex_attention PyTorch version check bug." + }, + { + "left": "issue:44291", + "right": "issue:45071", + "accept": false, + "reason": "Different code paths: init_empty_weights argument handling vs PreTrainedConfig type-checking in v5.4." + }, + { + "left": "issue:43450", + "right": "issue:43986", + "accept": false, + "reason": "Batched video processor shape bug is not the same as the AutoProcessor crash when torchvision is missing." + }, + { + "left": "issue:45341", + "right": "issue:45561", + "accept": false, + "reason": "Both touch testing utilities, but one is a small utility bug and the other is a pytest-xdist race on a capture file." + }, + { + "left": "issue:44117", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer mapping returning None and Mistral regex patch crashing are separate tokenizer failures." + }, + { + "left": "issue:38617", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but for different symbols and modules with different root causes." + }, + { + "left": "issue:44410", + "right": "issue:45310", + "accept": false, + "reason": "Different model families and symptoms: missing QKV projections vs Qwen3.5 MoE from_pretrained failure." + }, + { + "left": "issue:39401", + "right": "issue:43421", + "accept": false, + "reason": "Wrong offset_mapping in Qwen3 tokenizer is a different bug from runtime post-processor updates for special tokens." + }, + { + "left": "issue:44829", + "right": "issue:45216", + "accept": false, + "reason": "Flash-attention training degeneration and incorrect save_pretrained serialization are distinct issues." + }, + { + "left": "issue:44484", + "right": "issue:45310", + "accept": false, + "reason": "A question about max_shard_size is unrelated to a Qwen3.5 MoE loading error." + }, + { + "left": "issue:42371", + "right": "issue:44261", + "accept": false, + "reason": "TF32 API settings and MLA rms_norm_eps precision error are different problems." + }, + { + "left": "issue:42673", + "right": "issue:43582", + "accept": false, + "reason": "VRAM leak in multi-threaded inference is unrelated to an Apple Silicon TypeError in caching_allocator_warmup." + }, + { + "left": "issue:42617", + "right": "issue:45341", + "accept": false, + "reason": "Running 3d_parallel.py and a testing_utils bug are unrelated." + }, + { + "left": "issue:43408", + "right": "issue:43742", + "accept": false, + "reason": "A model-type warning for sam3_tracker is not the same as a KeyError loading MobileLLM-125M." + }, + { + "left": "issue:44933", + "right": "issue:45278", + "accept": false, + "reason": "One is a specific nonexistent import; the other is a broad report of many import errors after upgrade." + }, + { + "left": "issue:42222", + "right": "issue:45542", + "accept": false, + "reason": "VitPose breakage and missing TensorFlow backend due to tensorboard-only install are unrelated." + }, + { + "left": "issue:42490", + "right": "issue:42673", + "accept": false, + "reason": "A path-with-slash import warning is not the same as a VRAM leak in threaded dataloading." + }, + { + "left": "issue:42175", + "right": "issue:42222", + "accept": false, + "reason": "TensorFlow backend packaging and VitPose model breakage are separate issues." + }, + { + "left": "issue:43937", + "right": "issue:44496", + "accept": false, + "reason": "Invalid GenerationConfig and unrecognized model/config.json problems are different failure classes." + }, + { + "left": "issue:38617", + "right": "issue:44220", + "accept": false, + "reason": "ImportError for layer_type_validation is unrelated to the _torch_extract_fbank_features issue." + }, + { + "left": "issue:44483", + "right": "issue:45464", + "accept": false, + "reason": "Both mention chat/completions, but one is a generic API request rejection and the other is a streaming inference failure on a specific model." + }, + { + "left": "issue:42371", + "right": "issue:42831", + "accept": false, + "reason": "TF32 behavior controls and FineGrainedFP8 accuracy problems do not share the same root cause." + }, + { + "left": "issue:43066", + "right": "issue:43122", + "accept": false, + "reason": "Both are tokenizer regressions, but one is about decoder type in v5 and the other is about tokenization differences; not the same concrete bug." + }, + { + "left": "issue:44075", + "right": "issue:45656", + "accept": false, + "reason": "SGD args being ignored is unrelated to optimizer.step() being called twice under deepspeed." + }, + { + "left": "issue:43673", + "right": "issue:45341", + "accept": false, + "reason": "Generation cache missing during chunked_prefill and a testing_utils bug are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 42, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:38617", + "issue:41628", + "issue:42222", + "issue:42757", + "issue:43012", + "issue:43064", + "issue:43066", + "issue:43097", + "issue:43335", + "issue:43408", + "issue:43441", + "issue:43452", + "issue:43502", + "issue:43531", + "issue:43577", + "issue:43873", + "issue:43874", + "issue:43906", + "issue:43931", + "issue:43976", + "issue:43986", + "issue:43992", + "issue:44016", + "issue:44038", + "issue:44062", + "issue:44117", + "issue:44261", + "issue:44263", + "issue:44315", + "issue:44351", + "issue:44368", + "issue:44410", + "issue:44484", + "issue:44485", + "issue:44493", + "issue:44704", + "issue:44829", + "issue:44987", + "issue:45003", + "issue:45071", + "issue:45103", + "issue:45341", + "issue:45372", + "issue:45375", + "issue:45459", + "issue:45478", + "issue:45542", + "issue:45588" + ], + "soft_pairs": [ + "issue:43012|issue:43577", + "issue:44484|issue:45478", + "issue:43874|issue:44261", + "issue:44410|issue:45478", + "issue:43097|issue:44493", + "issue:44263|issue:44485", + "issue:41628|issue:45071", + "issue:43012|issue:43408", + "issue:42222|issue:45459", + "issue:45071|issue:45542", + "issue:43441|issue:44829", + "issue:43931|issue:44410", + "issue:43335|issue:45071", + "issue:43931|issue:44038", + "issue:45003|issue:45341", + "issue:43097|issue:43452", + "issue:44117|issue:44987", + "issue:43986|issue:45341", + "issue:43986|issue:44484", + "issue:43874|issue:44263", + "issue:43408|issue:44368", + "issue:43976|issue:45071", + "issue:43986|issue:45372", + "issue:43531|issue:45341", + "issue:38617|issue:42757", + "issue:44261|issue:44485", + "issue:43064|issue:44315", + "issue:44410|issue:45375", + "issue:43097|issue:45588", + "issue:44062|issue:44117", + "issue:38617|issue:44351", + "issue:44038|issue:44410", + "issue:44704|issue:45103", + "issue:43502|issue:43992", + "issue:43986|issue:44117", + "issue:43097|issue:43873", + "issue:43986|issue:44016", + "issue:43066|issue:43906" + ], + "prepared_review_unit_hash": "dffe142b16a27cfbc808589568ee4ccd6ee3ce5d591264ae90c881b749c23ec3", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 38, + "serialized_chars": 39163, + "estimated_input_tokens": 9791, + "estimated_eval_tokens": 19838 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: import errors, model-specific loading bugs, configuration regressions, and unrelated runtime warnings. The soft-similarity pairs all look superficially related but refer to different concrete code paths or failure modes, so they should stay separate.", + "confidence": 0.82, + "canonical_issue_reason": "issue:43097 is the closest thing to a canonical anchor because it describes a concrete, high-impact Transformers v5 API break with downstream references, and it best matches the broad version-regression theme among these otherwise mixed reports.", + "canonical_pr_reason": null, + "best_issue_reason": "There is no true single issue covering the whole set, but issue:43097 is the strongest triage anchor because it is specific, impactful, and broadly representative of the v5 breakage subgroup.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43012", + "right": "issue:43577", + "accept": false, + "reason": "Different bugs: one is a bfloat16 compilation warning, the other is a Blip2 dtype propagation issue." + }, + { + "left": "issue:44484", + "right": "issue:45478", + "accept": false, + "reason": "Unrelated: max_shard_size default behavior vs a Qwen3.5-Moe from_pretrained loading error." + }, + { + "left": "issue:43874", + "right": "issue:44261", + "accept": false, + "reason": "Different model/code paths: missing GLM46V image-patch API vs MLA q_a_layernorm precision mismatch." + }, + { + "left": "issue:44410", + "right": "issue:45478", + "accept": false, + "reason": "Different Qwen families and failures: missing projections in Qwen3Next vs Qwen3.5-Moe loading error." + }, + { + "left": "issue:43097", + "right": "issue:44493", + "accept": false, + "reason": "Different regression classes: removed config flag vs unexpected position-id key handling." + }, + { + "left": "issue:44263", + "right": "issue:44485", + "accept": false, + "reason": "Separate internals: torch.split return-value handling vs GLM-5 RoPE implementation." + }, + { + "left": "issue:41628", + "right": "issue:45071", + "accept": false, + "reason": "Different APIs: missing AutoImageProcessor import vs PretrainedConfig type checking breakage." + }, + { + "left": "issue:43012", + "right": "issue:43408", + "accept": false, + "reason": "Both are warnings, but they come from different subsystems and model types with different causes." + }, + { + "left": "issue:42222", + "right": "issue:45459", + "accept": false, + "reason": "Unrelated: VitPose model breakage vs tokenizer protobuf exception masking." + }, + { + "left": "issue:45071", + "right": "issue:45542", + "accept": false, + "reason": "Config typing regression and TensorBoard/TensorFlow backend error are unrelated." + }, + { + "left": "issue:43441", + "right": "issue:44829", + "accept": false, + "reason": "Different FlashAttention failures: one is a model loading issue, the other is a degenerate training behavior." + }, + { + "left": "issue:43931", + "right": "issue:44410", + "accept": false, + "reason": "Different problems: Qwen3-VL weight shape mismatch vs missing Qwen3Next layer projections." + }, + { + "left": "issue:43335", + "right": "issue:45071", + "accept": false, + "reason": "SwitchTransformers sparse-layer construction bug and PretrainedConfig type checking breakage are unrelated." + }, + { + "left": "issue:43931", + "right": "issue:44038", + "accept": false, + "reason": "Different model families and symptoms: load-time weight mismatch vs Qwen3.5/Qwen3-VL-Moe regression." + }, + { + "left": "issue:45003", + "right": "issue:45341", + "accept": false, + "reason": "Unsafe sys.modules access is unrelated to the small testing_utils bug." + }, + { + "left": "issue:43097", + "right": "issue:43452", + "accept": false, + "reason": "Different load-path regressions: removed tie_embeddings flag vs gguf_file tokenizer/model loading breakage." + }, + { + "left": "issue:44117", + "right": "issue:44987", + "accept": false, + "reason": "Tokenizer mapping None is not the same bug as loading physical-intelligence/fast on v5.1.0." + }, + { + "left": "issue:43986", + "right": "issue:45341", + "accept": false, + "reason": "AutoProcessor/torchvision crash and testing_utils bug are unrelated." + }, + { + "left": "issue:43986", + "right": "issue:44484", + "accept": false, + "reason": "Different issues: missing torchvision dependency handling vs a save_pretrained shard-size question." + }, + { + "left": "issue:43874", + "right": "issue:44263", + "accept": false, + "reason": "Both involve model internals, but one is a missing image-patch method and the other is torch.split output handling." + }, + { + "left": "issue:43408", + "right": "issue:44368", + "accept": false, + "reason": "Different warnings from different configs: SAM3 model-type mismatch vs tie_word_embeddings warning." + }, + { + "left": "issue:43976", + "right": "issue:45071", + "accept": false, + "reason": "Python version compatibility and config type checking are unrelated." + }, + { + "left": "issue:43986", + "right": "issue:45372", + "accept": false, + "reason": "Different dependency/import failures: no torchvision vs missing ReasoningEffort import." + }, + { + "left": "issue:43531", + "right": "issue:45341", + "accept": false, + "reason": "Qwen3-MoE sliding_window bug and a testing_utils bug are unrelated." + }, + { + "left": "issue:38617", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but for different missing symbols and different modules, so not the same underlying bug." + }, + { + "left": "issue:44261", + "right": "issue:44485", + "accept": false, + "reason": "Distinct model issues: rms_norm_eps precision error vs GLM-5 RoPE implementation." + }, + { + "left": "issue:43064", + "right": "issue:44315", + "accept": false, + "reason": "Different training bugs: wrong optimizer state partitioning in FSDP2+PEFT vs Liger Kernel not being applied with model_init." + }, + { + "left": "issue:44410", + "right": "issue:45375", + "accept": false, + "reason": "Different Qwen3.5-Moe issues: missing projections vs missing config field being dropped." + }, + { + "left": "issue:43097", + "right": "issue:45588", + "accept": false, + "reason": "Removed tie_embeddings flag and a FlashAttention s_aux=None crash are unrelated." + }, + { + "left": "issue:44062", + "right": "issue:44117", + "accept": false, + "reason": "Different tokenizer failures: AddedToken kwarg collision vs TOKENIZER_MAPPING_NAMES returning None." + }, + { + "left": "issue:38617", + "right": "issue:44351", + "accept": false, + "reason": "Both are import failures, but they concern different missing exports and different root causes." + }, + { + "left": "issue:44038", + "right": "issue:44410", + "accept": false, + "reason": "Different Qwen variants and different defects: general v5/Qwen3.5-Moe bug vs missing layer projections." + }, + { + "left": "issue:44704", + "right": "issue:45103", + "accept": false, + "reason": "AutoProcessor kwargs forwarding and auto_docstring annotation crashes are unrelated." + }, + { + "left": "issue:43502", + "right": "issue:43992", + "accept": false, + "reason": "Local-files-only network leakage and missing UMT5 encoder weights are separate loading bugs." + }, + { + "left": "issue:43986", + "right": "issue:44117", + "accept": false, + "reason": "No torchvision crash and tokenizer mapping None are unrelated failures." + }, + { + "left": "issue:43097", + "right": "issue:43873", + "accept": false, + "reason": "Different regressions: removed config option vs quantization/offloading not working as expected." + }, + { + "left": "issue:43986", + "right": "issue:44016", + "accept": false, + "reason": "A runtime import/dependency crash is not the same as a notebook syntax error." + }, + { + "left": "issue:43066", + "right": "issue:43906", + "accept": false, + "reason": "Wrong tokenizer decoder type and an isolated reproduction of another issue are not the same concrete bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 43, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:36010", + "issue:36683", + "issue:38617", + "issue:42490", + "issue:42757", + "issue:42886", + "issue:42907", + "issue:42994", + "issue:43012", + "issue:43064", + "issue:43097", + "issue:43208", + "issue:43299", + "issue:43352", + "issue:43381", + "issue:43475", + "issue:43502", + "issue:43526", + "issue:43550", + "issue:43582", + "issue:43618", + "issue:43756", + "issue:43761", + "issue:43824", + "issue:43825", + "issue:43867", + "issue:43901", + "issue:43937", + "issue:43976", + "issue:43992", + "issue:44038", + "issue:44077", + "issue:44222", + "issue:44230", + "issue:44368", + "issue:44380", + "issue:44483", + "issue:44484", + "issue:44485", + "issue:44514", + "issue:44655", + "issue:44683", + "issue:45216", + "issue:45290", + "issue:45381", + "issue:45478", + "issue:45479", + "issue:45753" + ], + "soft_pairs": [ + "issue:43097|issue:43901", + "issue:43825|issue:44655", + "issue:43992|issue:44484", + "issue:43208|issue:43550", + "issue:43012|issue:43550", + "issue:42886|issue:43381", + "issue:43352|issue:44380", + "issue:43012|issue:44368", + "issue:42907|issue:44655", + "issue:42907|issue:45753", + "issue:44380|issue:45479", + "issue:43352|issue:43976", + "issue:43064|issue:44380", + "issue:44483|issue:44514", + "issue:43208|issue:43526", + "issue:36683|issue:43582", + "issue:44222|issue:45478", + "issue:42490|issue:43012", + "issue:43618|issue:43761", + "issue:43012|issue:44683", + "issue:42757|issue:43824", + "issue:43618|issue:44380", + "issue:38617|issue:43502", + "issue:44038|issue:44230", + "issue:44483|issue:45290", + "issue:44483|issue:45381", + "issue:43299|issue:43867", + "issue:43756|issue:44485", + "issue:43901|issue:44655", + "issue:43299|issue:45753", + "issue:36683|issue:43475", + "issue:42994|issue:44655", + "issue:43352|issue:43937", + "issue:44077|issue:44380", + "issue:43867|issue:45216", + "issue:36010|issue:42757" + ], + "prepared_review_unit_hash": "0e49f0a621bbc6498f82d8a4470f00678797e16c936b3ee2aa3c0bf97adb8f32", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 36, + "serialized_chars": 38227, + "estimated_input_tokens": 9557, + "estimated_eval_tokens": 19370 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: most items are distinct Transformers regressions around imports, loading/saving, attention, or chat/template behavior. None of the soft pairs look like the same underlying bug, so all soft edges are rejected.", + "confidence": 0.24, + "canonical_issue_reason": "issue:43097 is the broadest representative here: a concrete v5 breaking-change regression with downstream references, while the other items are narrower or clearly unrelated symptoms.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43097 is the best umbrella issue in the set because it is a user-facing breaking regression with some cross-links, even though the cluster is too mixed to treat as a true duplicate group.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43097", + "right": "issue:43901", + "accept": false, + "reason": "Removal of a config field in v5 is not the same bug as a docs/behavior mismatch for return_all_scores." + }, + { + "left": "issue:43825", + "right": "issue:44655", + "accept": false, + "reason": "Chat-template task support is unrelated to saving Pipeline objects." + }, + { + "left": "issue:43992", + "right": "issue:44484", + "accept": false, + "reason": "Missing embed_tokens.weight during load is unrelated to the save_pretrained shard-size default question." + }, + { + "left": "issue:43208", + "right": "issue:43550", + "accept": false, + "reason": "xLSTM training bugs are not the same as Bamba torch.compile/SDPA failures." + }, + { + "left": "issue:43012", + "right": "issue:43550", + "accept": false, + "reason": "A bfloat16 precision warning is a different issue from a compile-time SDPA failure." + }, + { + "left": "issue:42886", + "right": "issue:43381", + "accept": false, + "reason": "Offline cache loading is unrelated to gradient checkpointing being disallowed in eval mode." + }, + { + "left": "issue:43352", + "right": "issue:44380", + "accept": false, + "reason": "Nemotron FlashAttention 2 support is a different code path from GPT2 attention scaling being ignored." + }, + { + "left": "issue:43012", + "right": "issue:44368", + "accept": false, + "reason": "PyTorch float32 warning on bfloat16 compile is unrelated to a tie_word_embeddings config warning." + }, + { + "left": "issue:42907", + "right": "issue:44655", + "accept": false, + "reason": "Dequantized model save failures are not the same as Pipeline serialization failures." + }, + { + "left": "issue:42907", + "right": "issue:45753", + "accept": false, + "reason": "Saving dequantized models and an image-triggered infinite loop are unrelated bugs." + }, + { + "left": "issue:44380", + "right": "issue:45479", + "accept": false, + "reason": "Attention scaling being ignored is unrelated to degenerate zero loss in sequence classification." + }, + { + "left": "issue:43352", + "right": "issue:43976", + "accept": false, + "reason": "FlashAttention support in a specific model is unrelated to the Python-version compatibility issue." + }, + { + "left": "issue:43064", + "right": "issue:44380", + "accept": false, + "reason": "FSDP2 optimizer-state corruption is not the same as GPT2 attention scaling being ignored." + }, + { + "left": "issue:44483", + "right": "issue:44514", + "accept": false, + "reason": "Chat/completions request rejection is a server/API issue, not the same as apply_chat_template crashing on tool-call messages." + }, + { + "left": "issue:43208", + "right": "issue:43526", + "accept": false, + "reason": "xLSTM training problems are unrelated to BeitImageProcessorFast reduce_labels behavior." + }, + { + "left": "issue:36683", + "right": "issue:43582", + "accept": false, + "reason": "A missing config attribute is unrelated to an Apple Silicon allocator TypeError." + }, + { + "left": "issue:44222", + "right": "issue:45478", + "accept": false, + "reason": "FP8 save_pretrained issues and Qwen3.5 moe from_pretrained loading errors are different directions and different failures." + }, + { + "left": "issue:42490", + "right": "issue:43012", + "accept": false, + "reason": "Trailing-slash import warnings are unrelated to bfloat16 compile warnings." + }, + { + "left": "issue:43618", + "right": "issue:43761", + "accept": false, + "reason": "Both touch CLIP outputs, but one is missing attentions and the other is missing hidden_states; not the same bug." + }, + { + "left": "issue:43012", + "right": "issue:44683", + "accept": false, + "reason": "A warning during compile is not the same as flex_attention breaking on torch>=2.9." + }, + { + "left": "issue:42757", + "right": "issue:43824", + "accept": false, + "reason": "These are different missing imports from different modules and symbols." + }, + { + "left": "issue:43618", + "right": "issue:44380", + "accept": false, + "reason": "CLIP output fields and GPT2 attention scaling are different subsystems and symptoms." + }, + { + "left": "issue:38617", + "right": "issue:43502", + "accept": false, + "reason": "ImportError for layer_type_validation is unrelated to local_files_only still making API requests." + }, + { + "left": "issue:44038", + "right": "issue:44230", + "accept": false, + "reason": "Qwen3-VL-Moe loading regressions are not the same as FP8 inference support work." + }, + { + "left": "issue:44483", + "right": "issue:45290", + "accept": false, + "reason": "Endpoint request validation and apply_chat_template crashes are separate code paths." + }, + { + "left": "issue:44483", + "right": "issue:45381", + "accept": false, + "reason": "Chat/completions request rejection is unrelated to Qwen2.5-VL video vision_position_ids being wrong." + }, + { + "left": "issue:43299", + "right": "issue:43867", + "accept": false, + "reason": "Qwen3-VL-Moe loading problems are not the same as a sorted-state_dict load error." + }, + { + "left": "issue:43756", + "right": "issue:44485", + "accept": false, + "reason": "Smollm3 RoPE-layer omission is unrelated to a GLM-5 RoPE implementation discussion." + }, + { + "left": "issue:43901", + "right": "issue:44655", + "accept": false, + "reason": "A docs mismatch about return_all_scores is unrelated to saving Pipeline objects." + }, + { + "left": "issue:43299", + "right": "issue:45753", + "accept": false, + "reason": "Qwen3.5 VL/Moe loading regressions are unrelated to an image-specific infinite loop." + }, + { + "left": "issue:36683", + "right": "issue:43475", + "accept": false, + "reason": "Gemma3Config missing vocab_size and Sam3VisionEncoderOutput missing an attribute are unrelated model-specific bugs." + }, + { + "left": "issue:42994", + "right": "issue:44655", + "accept": false, + "reason": "Quantized model save failure is not the same as Pipeline.save_pretrained serialization." + }, + { + "left": "issue:43352", + "right": "issue:43937", + "accept": false, + "reason": "Unsupported FlashAttention in one model is unrelated to an invalid GenerationConfig in GLM-5." + }, + { + "left": "issue:44077", + "right": "issue:44380", + "accept": false, + "reason": "patchtsmixer post_init validation and GPT2 attention scaling are different issues." + }, + { + "left": "issue:43867", + "right": "issue:45216", + "accept": false, + "reason": "Both are load/save regression flavored, but one is sorted-state_dict loading while the other is a Qwen3.5 save_pretrained checkpoint regression; not the same bug." + }, + { + "left": "issue:36010", + "right": "issue:42757", + "accept": false, + "reason": "Different import errors for different symbols and packages; same surface pattern, different underlying issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 44, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:28282", + "issue:33453", + "issue:36296", + "issue:36683", + "issue:41669", + "issue:42490", + "issue:42890", + "issue:42898", + "issue:43010", + "issue:43064", + "issue:43066", + "issue:43352", + "issue:43525", + "issue:43575", + "issue:43611", + "issue:43618", + "issue:43867", + "issue:43931", + "issue:43976", + "issue:43992", + "issue:44075", + "issue:44077", + "issue:44188", + "issue:44242", + "issue:44246", + "issue:44263", + "issue:44355", + "issue:44360", + "issue:44380", + "issue:44484", + "issue:44492", + "issue:44655", + "issue:44855", + "issue:44857", + "issue:44938", + "issue:44977", + "issue:45071", + "issue:45216", + "issue:45278", + "issue:45310", + "issue:45357", + "issue:45375", + "issue:45381", + "issue:45440", + "issue:45478", + "issue:45561", + "issue:45593", + "issue:45850" + ], + "soft_pairs": [ + "issue:42898|issue:43066", + "issue:36683|issue:44077", + "issue:43611|issue:44655", + "issue:43976|issue:44938", + "issue:43352|issue:45593", + "issue:36683|issue:43525", + "issue:43618|issue:45216", + "issue:44263|issue:45440", + "issue:36683|issue:44242", + "issue:44188|issue:44380", + "issue:43064|issue:44492", + "issue:43352|issue:43931", + "issue:43992|issue:45310", + "issue:43010|issue:45278", + "issue:44355|issue:44855", + "issue:43867|issue:45375", + "issue:41669|issue:44246", + "issue:36683|issue:43064", + "issue:42890|issue:45561", + "issue:44938|issue:45071", + "issue:43867|issue:45357", + "issue:43010|issue:45850", + "issue:43992|issue:45478", + "issue:43010|issue:43976", + "issue:44938|issue:45381", + "issue:43618|issue:44077", + "issue:42490|issue:44075", + "issue:36296|issue:43575", + "issue:28282|issue:33453", + "issue:44360|issue:45440", + "issue:43867|issue:44484", + "issue:43618|issue:44857", + "issue:43618|issue:44977" + ], + "prepared_review_unit_hash": "a1d93bb1664f891ab918dc4ba022010c9f51da52c2036af41b2cde91aaa64800", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37072, + "estimated_input_tokens": 9268, + "estimated_eval_tokens": 18792 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose grab-bag of Transformers regressions and performance issues, mostly model-specific loading/saving/config bugs rather than true duplicates. I found no pair that clearly describes the same underlying bug closely enough to merge as duplicates.", + "confidence": 0.31, + "canonical_issue_reason": "If a single anchor is needed, issue 45850 is the broadest umbrella report about widespread breakage in v5, but the cluster is too heterogeneous to justify a strong canonical duplicate choice.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45850 is the best global triage anchor because it captures the widest user impact and can absorb related v5 breakage reports, even though most items here are not duplicates of it.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42898", + "right": "issue:43066", + "accept": false, + "reason": "Both are tokenizer-related, but one is a v5 behavior change and the other is a wrong decoder type; different bugs." + }, + { + "left": "issue:36683", + "right": "issue:44077", + "accept": false, + "reason": "Unrelated model config attribute error vs a separate rule change for optional post_init." + }, + { + "left": "issue:43611", + "right": "issue:44655", + "accept": false, + "reason": "Different code paths: base_model_prefix loading regression vs Pipeline save_pretrained support." + }, + { + "left": "issue:43976", + "right": "issue:44938", + "accept": false, + "reason": "Both are Python-version failures, but the reported failure modes and fixes are different." + }, + { + "left": "issue:43352", + "right": "issue:45593", + "accept": false, + "reason": "Completely different models and subsystems: FlashAttention support vs auxiliary loss handling." + }, + { + "left": "issue:36683", + "right": "issue:43525", + "accept": false, + "reason": "Same general pattern of missing config attributes, but for different models and different missing fields." + }, + { + "left": "issue:43618", + "right": "issue:45216", + "accept": false, + "reason": "CLIP attentions regression and Qwen3.5 save_pretrained corruption are unrelated." + }, + { + "left": "issue:44263", + "right": "issue:45440", + "accept": false, + "reason": "Both involve MoE-style model code, but the concrete bugs and model implementations differ." + }, + { + "left": "issue:36683", + "right": "issue:44242", + "accept": false, + "reason": "Different issues: missing Gemma3Config field vs load balancing loss not added." + }, + { + "left": "issue:44188", + "right": "issue:44380", + "accept": false, + "reason": "Both mention attention behavior under backend changes, but they affect different mechanisms and are not the same bug." + }, + { + "left": "issue:43064", + "right": "issue:44492", + "accept": false, + "reason": "Training/FSDP optimizer-state corruption is unrelated to a cache strategy typo." + }, + { + "left": "issue:43352", + "right": "issue:43931", + "accept": false, + "reason": "Both are Qwen/Nemotron model loading issues, but they are different models and different failure signatures." + }, + { + "left": "issue:43992", + "right": "issue:45310", + "accept": false, + "reason": "Both are from_pretrained errors, but for different models and different missing/corrupted weights." + }, + { + "left": "issue:43010", + "right": "issue:45278", + "accept": false, + "reason": "One is about no_grad decoration on cache updates; the other is a broad import regression after version updates." + }, + { + "left": "issue:44355", + "right": "issue:44855", + "accept": false, + "reason": "Different root causes: compiled Python file errors vs a Python 3.13 torch.jit.script parsing problem." + }, + { + "left": "issue:43867", + "right": "issue:45375", + "accept": false, + "reason": "State-dict loading error and a missing config field are not the same underlying bug." + }, + { + "left": "issue:41669", + "right": "issue:44246", + "accept": false, + "reason": "Both concern import performance, but one is about import * overhead and the other about intermittent import slowness." + }, + { + "left": "issue:36683", + "right": "issue:43064", + "accept": false, + "reason": "Missing config attribute vs FSDP2/PEFT optimizer-state corruption; no shared code-path bug." + }, + { + "left": "issue:42890", + "right": "issue:45561", + "accept": false, + "reason": "Both are test flakiness, but one is seed-related nondeterminism and the other is an xdist file race." + }, + { + "left": "issue:44938", + "right": "issue:45071", + "accept": false, + "reason": "Different version/regression topics with different failure modes: import failure in Python 3.14 vs config type-checking breakage." + }, + { + "left": "issue:43867", + "right": "issue:45357", + "accept": false, + "reason": "One is a sorted state_dict load error; the other is incorrect visual encoder keys on save_pretrained." + }, + { + "left": "issue:43010", + "right": "issue:45850", + "accept": false, + "reason": "A cache no_grad concern is not the same bug as a broad 'many models break' regression report." + }, + { + "left": "issue:43992", + "right": "issue:45478", + "accept": false, + "reason": "Different models and likely different weight-mapping/loading failures." + }, + { + "left": "issue:43010", + "right": "issue:43976", + "accept": false, + "reason": "Unrelated: cache update semantics vs Python-version compatibility." + }, + { + "left": "issue:44938", + "right": "issue:45381", + "accept": false, + "reason": "Different model families and different bugs; both are release regressions only at a very high level." + }, + { + "left": "issue:43618", + "right": "issue:44077", + "accept": false, + "reason": "CLIP attentions assignment and patchtsmixer post_init policy are unrelated." + }, + { + "left": "issue:42490", + "right": "issue:44075", + "accept": false, + "reason": "Local-path import warning vs optimizer argument handling are unrelated." + }, + { + "left": "issue:36296", + "right": "issue:43575", + "accept": false, + "reason": "Both mention tensor parallelism, but one is a training bug and the other is an OOM during model loading." + }, + { + "left": "issue:28282", + "right": "issue:33453", + "accept": false, + "reason": "Completely different problems: missing PyTorch dependency vs tokenizer loading regression." + }, + { + "left": "issue:44360", + "right": "issue:45440", + "accept": false, + "reason": "Both touch model internals, but the DSA indexer ReLU issue and DeepSeekV3 implementation divergence are different bugs." + }, + { + "left": "issue:43867", + "right": "issue:44484", + "accept": false, + "reason": "State-dict loading failure and max_shard_size behavior are unrelated." + }, + { + "left": "issue:43618", + "right": "issue:44857", + "accept": false, + "reason": "Different models and symptoms; no shared code-path problem." + }, + { + "left": "issue:43618", + "right": "issue:44977", + "accept": false, + "reason": "Flash-attention generation bug in Qwen3.5 is unrelated to CLIPOutput attentions assignment." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 45, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:30333", + "issue:30990", + "issue:31795", + "issue:33453", + "issue:34689", + "issue:36296", + "issue:38617", + "issue:41628", + "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:44112", + "issue:44222", + "issue:44230", + "issue:44315", + "issue:44351", + "issue:44355", + "issue:44360", + "issue:44387", + "issue:44485", + "issue:44492", + "issue:44556", + "issue:44655", + "issue:44749", + "issue:44908", + "issue:44938", + "issue:44961", + "issue:44964", + "issue:45071", + "issue:45125", + "issue:45341", + "issue:45375", + "issue:45468", + "issue:45478", + "issue:45600" + ], + "soft_pairs": [ + "issue:42890|issue:44964", + "issue:42890|issue:44112", + "issue:42890|issue:45600", + "issue:43502|issue:44062", + "issue:43723|issue:44556", + "issue:44222|issue:45125", + "issue:42890|issue:45341", + "issue:44387|issue:44556", + "issue:42548|issue:44351", + "issue:43010|issue:43065", + "issue:43502|issue:44485", + "issue:44355|issue:44749", + "issue:45125|issue:45478", + "issue:43010|issue:43232", + "issue:30990|issue:31795", + "issue:45125|issue:45375", + "issue:30333|issue:33453", + "issue:42831|issue:43874", + "issue:43502|issue:44222", + "issue:43299|issue:44556", + "issue:44315|issue:44360", + "issue:43502|issue:44961", + "issue:44355|issue:45071", + "issue:30333|issue:36296", + "issue:44355|issue:44655", + "issue:43827|issue:44961", + "issue:44222|issue:44230", + "issue:33453|issue:34689", + "issue:42757|issue:43576", + "issue:33453|issue:36296", + "issue:43576|issue:45071", + "issue:43576|issue:43824", + "issue:31795|issue:38617", + "issue:41628|issue:42548", + "issue:44961|issue:45468", + "issue:38617|issue:43576", + "issue:43576|issue:44938", + "issue:43673|issue:44492", + "issue:44908|issue:44961", + "issue:44485|issue:44961" + ], + "prepared_review_unit_hash": "fa21192ad3c50af8bdf97cf98b3d18bbc7252af863ef2633cdccf21edb554d5f", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 40, + "serialized_chars": 39025, + "estimated_input_tokens": 9757, + "estimated_eval_tokens": 19770 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab bag of unrelated Transformers issues (imports, model loading, flaky tests, generation, FP8, tensor parallelism). There is no strong duplicate cluster overall; only one soft pair looks plausibly like the same Qwen3.5 MoE loading bug.", + "confidence": 0.91, + "canonical_issue_reason": "No single canonical issue fits the whole set: the reports span different subsystems and failure modes, so collapsing them into one tracker would be misleading.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44556 is the closest broad representative of the repeated v5 loading-regression theme, but it still does not cover most of the other issues here.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42890", + "right": "issue:44964", + "accept": false, + "reason": "SAM HQ test flakiness from missing seeding is unrelated to a Phi-4 multimodal loading failure." + }, + { + "left": "issue:42890", + "right": "issue:44112", + "accept": false, + "reason": "Both mention CI/test failures, but they involve different models and different root causes." + }, + { + "left": "issue:42890", + "right": "issue:45600", + "accept": false, + "reason": "Flaky SAM HQ integration tests and stale auto_mappings config references are different bugs." + }, + { + "left": "issue:43502", + "right": "issue:44062", + "accept": false, + "reason": "Local-files-only network leakage and AddedToken keyword collisions are unrelated code paths." + }, + { + "left": "issue:43723", + "right": "issue:44556", + "accept": false, + "reason": "Both are loading-related, but one is tokenizer loading and the other is checkpoint reload compatibility; too different to merge." + }, + { + "left": "issue:44222", + "right": "issue:45125", + "accept": false, + "reason": "FP8 save_pretrained behavior and missing tensor-parallel plan metadata are distinct problems." + }, + { + "left": "issue:42890", + "right": "issue:45341", + "accept": false, + "reason": "A seed-related flaky test and a generic testing_utils bug are not clearly the same defect." + }, + { + "left": "issue:44387", + "right": "issue:44556", + "accept": false, + "reason": "int4 quantization memory growth causing OOM is not the same as checkpoint reload failure after upgrade." + }, + { + "left": "issue:42548", + "right": "issue:44351", + "accept": false, + "reason": "These are different missing top-level imports for different symbols." + }, + { + "left": "issue:43010", + "right": "issue:43065", + "accept": false, + "reason": "Cache update no_grad semantics and a dummy Conv2d in Sam3PixelDecoder are unrelated." + }, + { + "left": "issue:43502", + "right": "issue:44485", + "accept": false, + "reason": "Unexpected API requests with local_files_only=True is unrelated to GLM-5 RoPE implementation." + }, + { + "left": "issue:44355", + "right": "issue:44749", + "accept": false, + "reason": "Compiled Python file errors and a data-filtering slowdown are not the same underlying bug." + }, + { + "left": "issue:45125", + "right": "issue:45478", + "accept": true, + "reason": "Both point to the same Qwen3.5 MoE loading path in v5.4.0; missing _tp_plan metadata plausibly explains the from_pretrained failure." + }, + { + "left": "issue:43010", + "right": "issue:43232", + "accept": false, + "reason": "These are different generation/cache issues with different triggers and code paths." + }, + { + "left": "issue:30990", + "right": "issue:31795", + "accept": false, + "reason": "Sentence Transformers getting stuck loading and a documentation clarification in model.forward are unrelated." + }, + { + "left": "issue:45125", + "right": "issue:45375", + "accept": false, + "reason": "Same model family, but a missing tensor-parallel plan and a missing config field are different bugs." + }, + { + "left": "issue:30333", + "right": "issue:33453", + "accept": false, + "reason": "MLFlow failure reporting and tokenizer loading regression are unrelated." + }, + { + "left": "issue:42831", + "right": "issue:43874", + "accept": false, + "reason": "FP8 accuracy degradation and a missing image-patch count method are different failures." + }, + { + "left": "issue:43502", + "right": "issue:44222", + "accept": false, + "reason": "Local-files-only network leakage and FP8 save_pretrained behavior do not share a concrete code path." + }, + { + "left": "issue:43299", + "right": "issue:44556", + "accept": false, + "reason": "Both are version-related loading regressions, but for different models and different failure modes." + }, + { + "left": "issue:44315", + "right": "issue:44360", + "accept": false, + "reason": "Applying Liger Kernel via model_init and a missing ReLU in the DSA indexer are unrelated." + }, + { + "left": "issue:43502", + "right": "issue:44961", + "accept": false, + "reason": "A network access bug and a placeholder-named issue are clearly unrelated." + }, + { + "left": "issue:43827", + "right": "issue:44961", + "accept": false, + "reason": "Documentation references to pipeline() and the racoon issue are unrelated." + }, + { + "left": "issue:44222", + "right": "issue:44230", + "accept": false, + "reason": "FP8 save_pretrained for MoE and FP8 inference support for Qwen3-VL/Qwen3.5 are adjacent but not the same bug." + }, + { + "left": "issue:33453", + "right": "issue:34689", + "accept": false, + "reason": "Tokenizer loading regression and Llama 3.2 vision model loading regression are different issues." + }, + { + "left": "issue:42757", + "right": "issue:43576", + "accept": false, + "reason": "Missing is_offline_mode import and a broken transformers env CLI are separate problems." + }, + { + "left": "issue:33453", + "right": "issue:36296", + "accept": false, + "reason": "Tokenizer loading regression and a tensor-parallel training bug are unrelated." + }, + { + "left": "issue:43576", + "right": "issue:45071", + "accept": false, + "reason": "Broken env CLI behavior and PretrainedConfig type-checking regression are different code paths." + }, + { + "left": "issue:43576", + "right": "issue:43824", + "accept": false, + "reason": "The env command bug and a Qwen2_5_VL import error are unrelated." + }, + { + "left": "issue:31795", + "right": "issue:38617", + "accept": false, + "reason": "Documentation confusion in model.forward and an ImportError for layer_type_validation are unrelated." + }, + { + "left": "issue:44961", + "right": "issue:45468", + "accept": false, + "reason": "These titles are unrelated and provide no evidence of the same underlying bug." + }, + { + "left": "issue:38617", + "right": "issue:43576", + "accept": false, + "reason": "A missing import from configuration_utils and a broken env CLI are different issues." + }, + { + "left": "issue:43576", + "right": "issue:44938", + "accept": false, + "reason": "Broken CLI behavior in v5 and Python 3.14 load failure are not the same defect." + }, + { + "left": "issue:43673", + "right": "issue:44492", + "accept": false, + "reason": "Missing cache in chunked_prefill generation and a cache-strategy typo are unrelated." + }, + { + "left": "issue:44908", + "right": "issue:44961", + "accept": false, + "reason": "inverse_sqrt scheduler kwargs handling and the racoon issue are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet and stays conservative. It correctly characterizes the set as mostly unrelated Transformers issues while acknowledging the one plausible duplicate pair. The accepted soft edge for issue:45125 and issue:45478 is supported by their titles and the shared Qwen3.5 MoE loading path. I did not see overstatement or an unsupported canonical merge." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 46, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:31795", + "issue:33453", + "issue:36296", + "issue:41628", + "issue:41669", + "issue:42371", + "issue:42890", + "issue:43352", + "issue:43408", + "issue:43525", + "issue:43526", + "issue:43572", + "issue:43576", + "issue:43673", + "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:44492", + "issue:44556", + "issue:44623", + "issue:44704", + "issue:44861", + "issue:44908", + "issue:44945", + "issue:45003", + "issue:45071", + "issue:45092", + "issue:45125", + "issue:45230", + "issue:45278", + "issue:45341", + "issue:45362", + "issue:45478", + "issue:45542", + "issue:45753", + "issue:45850" + ], + "soft_pairs": [ + "issue:36296|issue:43901", + "issue:41628|issue:44355", + "issue:36296|issue:43526", + "issue:45362|issue:45850", + "issue:44261|issue:44360", + "issue:44623|issue:45230", + "issue:44273|issue:45542", + "issue:44273|issue:45003", + "issue:44393|issue:45753", + "issue:44038|issue:45125", + "issue:44393|issue:45071", + "issue:43576|issue:44556", + "issue:41669|issue:43673", + "issue:43525|issue:45850", + "issue:44556|issue:45092", + "issue:44492|issue:45850", + "issue:43572|issue:45850", + "issue:42371|issue:44492", + "issue:44704|issue:44908", + "issue:42890|issue:44908", + "issue:45278|issue:45850", + "issue:44246|issue:44273", + "issue:43408|issue:44315", + "issue:43976|issue:44393", + "issue:43352|issue:44393", + "issue:44297|issue:45230", + "issue:44295|issue:45230", + "issue:43576|issue:44908", + "issue:44908|issue:45071", + "issue:41669|issue:44351", + "issue:41669|issue:45341", + "issue:44492|issue:44945", + "issue:43976|issue:45850", + "issue:44393|issue:45478", + "issue:44077|issue:44861", + "issue:43827|issue:44908", + "issue:31795|issue:33453", + "issue:41669|issue:44908" + ], + "prepared_review_unit_hash": "026da175eec397a2f34f10e82c2d078d77ddc9730be5c8405e4f0ce29f3eb640", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 38, + "serialized_chars": 38367, + "estimated_input_tokens": 9592, + "estimated_eval_tokens": 19440 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab bag of v5 migration regressions, import/load failures, and model-specific bugs. None of the soft pairs look like the same concrete underlying defect, so I would not collapse any of them as duplicates.", + "confidence": 0.76, + "canonical_issue_reason": "issue:45850 is the broadest umbrella-style report in the set, explicitly describing many model breakages after v5; it is the best cluster anchor despite being vague.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45850 is the most representative issue for the overall v5-regression theme, though it is too general to be a strong canonical on its own.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36296", + "right": "issue:43901", + "accept": false, + "reason": "Tensor-parallel training bug vs. pipeline docs wording; different problem and code path." + }, + { + "left": "issue:41628", + "right": "issue:44355", + "accept": false, + "reason": "Specific missing import symbol vs. compiled-file execution error; unrelated failure modes." + }, + { + "left": "issue:36296", + "right": "issue:43526", + "accept": false, + "reason": "Tensor-parallel training issue vs. BeitImageProcessorFast label reduction bug; not the same defect." + }, + { + "left": "issue:45362", + "right": "issue:45850", + "accept": false, + "reason": "A specific Qwen3.5 chat crash vs. a vague broad complaint about v5; not a shared concrete bug." + }, + { + "left": "issue:44261", + "right": "issue:44360", + "accept": false, + "reason": "Layernorm precision/config mismatch vs. DSA indexer ReLU omission; unrelated." + }, + { + "left": "issue:44623", + "right": "issue:45230", + "accept": false, + "reason": "Concrete save_pretrained missing-files bug vs. generic bug report; too little overlap." + }, + { + "left": "issue:44273", + "right": "issue:45542", + "accept": false, + "reason": "Lazy loading malfunction vs. TensorFlow backend import error; different root causes." + }, + { + "left": "issue:44273", + "right": "issue:45003", + "accept": false, + "reason": "Lazy loading complaint vs. sys.modules access bug; related area but not the same issue." + }, + { + "left": "issue:44393", + "right": "issue:45753", + "accept": false, + "reason": "2D bbox hallucination/error vs. infinite-loop-on-image bug; different symptoms and likely different causes." + }, + { + "left": "issue:44038", + "right": "issue:45125", + "accept": false, + "reason": "Qwen3-VL-Moe bug vs. missing _tp_plan for tensor parallelism; not the same concrete change." + }, + { + "left": "issue:44393", + "right": "issue:45071", + "accept": false, + "reason": "Model output bug vs. PretrainedConfig type-checking regression; unrelated." + }, + { + "left": "issue:43576", + "right": "issue:44556", + "accept": false, + "reason": "Broken transformers env command vs. checkpoint reload incompatibility; different surfaces." + }, + { + "left": "issue:41669", + "right": "issue:43673", + "accept": false, + "reason": "Import-time slowdown from wildcard imports vs. missing GenerationMixin cache in chunked prefill; unrelated." + }, + { + "left": "issue:43525", + "right": "issue:45850", + "accept": false, + "reason": "Llama4Config pad_token_id attribute error vs. broad v5 complaint; not the same bug." + }, + { + "left": "issue:44556", + "right": "issue:45092", + "accept": false, + "reason": "Both are v5 checkpoint-loading regressions, but one is generic reload failure and the other is a specific InternVL2/meta-init incompatibility; not enough evidence they are identical." + }, + { + "left": "issue:44492", + "right": "issue:45850", + "accept": false, + "reason": "Cache-strategy typo vs. broad model breakage claim; unrelated." + }, + { + "left": "issue:43572", + "right": "issue:45850", + "accept": false, + "reason": "Missing pad_token_idx after v5 update vs. generic v5 breakage report; different bugs." + }, + { + "left": "issue:42371", + "right": "issue:44492", + "accept": false, + "reason": "TF32 behavior warning vs. cache-strategy typo; unrelated." + }, + { + "left": "issue:44704", + "right": "issue:44908", + "accept": false, + "reason": "AutoProcessor kwargs plumbing vs. inverse_sqrt scheduler kwargs handling; different code paths." + }, + { + "left": "issue:42890", + "right": "issue:44908", + "accept": false, + "reason": "Flaky SAM-HQ integration tests vs. scheduler kwargs bug; unrelated." + }, + { + "left": "issue:45278", + "right": "issue:45850", + "accept": false, + "reason": "Import errors after upgrading to 5.5.0 vs. a vague claim that many models break; too broad to be the same concrete defect." + }, + { + "left": "issue:44246", + "right": "issue:44273", + "accept": false, + "reason": "Import slowdown/performance regression vs. lazy loading not working properly; related theme but not the same bug." + }, + { + "left": "issue:43408", + "right": "issue:44315", + "accept": false, + "reason": "sam3_video/sam3_tracker class mismatch warning vs. Liger Kernel application with model_init; unrelated." + }, + { + "left": "issue:43976", + "right": "issue:44393", + "accept": false, + "reason": "Python version compatibility problem vs. Qwen3-VL bbox output bug; unrelated." + }, + { + "left": "issue:43352", + "right": "issue:44393", + "accept": false, + "reason": "Flash Attention 2 support error vs. Qwen3-VL output hallucination; different issues." + }, + { + "left": "issue:44297", + "right": "issue:45230", + "accept": false, + "reason": "tokenizer_class mismatch on save vs. generic bug report; not the same bug." + }, + { + "left": "issue:44295", + "right": "issue:45230", + "accept": false, + "reason": "position_ids buffer read error vs. generic bug report; unrelated." + }, + { + "left": "issue:43576", + "right": "issue:44908", + "accept": false, + "reason": "CLI env command breakage vs. scheduler kwarg handling; unrelated." + }, + { + "left": "issue:41669", + "right": "issue:44351", + "accept": false, + "reason": "Wildcard import performance regression vs. HybridCache import failure; different symbols and causes." + }, + { + "left": "issue:41669", + "right": "issue:45341", + "accept": false, + "reason": "Import performance regression vs. a testing_utils bug; unrelated." + }, + { + "left": "issue:44492", + "right": "issue:44945", + "accept": false, + "reason": "Cache-strategy typo vs. incorrect pipeline-parallel LLM output; unrelated." + }, + { + "left": "issue:43976", + "right": "issue:45850", + "accept": false, + "reason": "Python version support issue vs. broad v5 complaint; not the same defect." + }, + { + "left": "issue:44393", + "right": "issue:45478", + "accept": false, + "reason": "Qwen3-VL bbox output bug vs. Qwen3.5-Moe from_pretrained error; different model paths and symptoms." + }, + { + "left": "issue:44077", + "right": "issue:44861", + "accept": false, + "reason": "patchtsmixer post_init policy change vs. tied-weight-key AttributeError; unrelated." + }, + { + "left": "issue:43827", + "right": "issue:44908", + "accept": false, + "reason": "Documentation still referencing pipeline() vs. scheduler kwargs being ignored; different topics." + }, + { + "left": "issue:31795", + "right": "issue:33453", + "accept": false, + "reason": "Forward-signature documentation confusion vs. tokenizer loading regression; unrelated." + }, + { + "left": "issue:41669", + "right": "issue:44908", + "accept": false, + "reason": "Import-time regression from wildcard imports vs. scheduler kwargs bug; unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 17, + "review_unit_index": 47, + "review_unit_count": 47, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-404", + "nodes": [ + "issue:30333", + "issue:31515", + "issue:31795", + "issue:33453", + "issue:34689", + "issue:36296", + "issue:37428", + "issue:38617", + "issue:41669", + "issue:42371", + "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:44623", + "issue:44655", + "issue:44829", + "issue:44861", + "issue:44908", + "issue:44998", + "issue:45092", + "issue:45230", + "issue:45290", + "issue:45310", + "issue:45341", + "issue:45468", + "issue:45850" + ], + "soft_pairs": [ + "issue:44315|issue:44829", + "issue:43576|issue:44273", + "issue:44655|issue:44861", + "issue:44360|issue:45468", + "issue:45230|issue:45310", + "issue:44273|issue:44908", + "issue:43519|issue:44485", + "issue:43519|issue:44908", + "issue:31795|issue:34689", + "issue:41669|issue:43704", + "issue:31795|issue:37428", + "issue:44315|issue:45092", + "issue:44485|issue:45468", + "issue:43704|issue:44485", + "issue:43502|issue:43519", + "issue:42371|issue:43704", + "issue:43519|issue:44623", + "issue:43519|issue:43704", + "issue:44263|issue:45468", + "issue:43606|issue:45468", + "issue:44261|issue:45468", + "issue:42913|issue:43066", + "issue:44908|issue:44998", + "issue:44492|issue:44998", + "issue:44861|issue:45290", + "issue:44998|issue:45850", + "issue:43618|issue:44998", + "issue:44998|issue:45341", + "issue:43525|issue:44861", + "issue:43572|issue:44861", + "issue:31515|issue:38617", + "issue:44230|issue:44623", + "issue:30333|issue:31515", + "issue:31515|issue:36296", + "issue:31515|issue:31795", + "issue:44230|issue:45310", + "issue:31515|issue:33453", + "issue:43298|issue:44230" + ], + "prepared_review_unit_hash": "9de6ceded600149673dc714e35aeddfcafd77bbc6884df72d0a428caf59ab853", + "component_budget": { + "node_count": 404, + "item_count": 404, + "soft_pair_count": 1511, + "serialized_chars": 779549, + "estimated_input_tokens": 194888, + "estimated_eval_tokens": 390032 + }, + "budget": { + "node_count": 42, + "item_count": 42, + "soft_pair_count": 38, + "serialized_chars": 34927, + "estimated_input_tokens": 8732, + "estimated_eval_tokens": 17720 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These artifacts are mostly a loose collection of unrelated Transformers bug reports, with a few broad v5-regression themes. None of the soft pairs look like true duplicates of the same concrete bug or change, so they should stay separate.", + "confidence": 0.71, + "canonical_issue_reason": "issue:45850 is the broadest umbrella here (\u201cSince 5.0 version it breaks too many models\u201d) and best matches the cluster-wide v5 regression theme, even though it is very generic.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45850 is the best representative issue for the cluster because it summarizes the shared high-level complaint without implying a specific code-path fix.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44315", + "right": "issue:44829", + "accept": false, + "reason": "Different symptoms and code paths: model_init/Liger Kernel application vs flash_attention_3 training degeneration." + }, + { + "left": "issue:43576", + "right": "issue:44273", + "accept": false, + "reason": "One is the v5 `transformers env` command, the other is lazy loading behavior; no concrete overlap." + }, + { + "left": "issue:44655", + "right": "issue:44861", + "accept": false, + "reason": "`save_pretrained` pipeline saving failure is unrelated to `_get_tied_weight_keys` crashing on a list." + }, + { + "left": "issue:44360", + "right": "issue:45468", + "accept": false, + "reason": "DSA indexer ReLU discussion and Gemma4Audio positional encoding are different model internals." + }, + { + "left": "issue:45230", + "right": "issue:45310", + "accept": false, + "reason": "`Bug report` is too generic to show the same bug as the Qwen3.5 MoE from_pretrained error." + }, + { + "left": "issue:44273", + "right": "issue:44908", + "accept": false, + "reason": "Lazy loading and inverse_sqrt scheduler kwargs are unrelated features." + }, + { + "left": "issue:43519", + "right": "issue:44485", + "accept": false, + "reason": "Qwen3VL timestamp calculation and GLM-5 RoPE implementation are different model-specific bugs." + }, + { + "left": "issue:43519", + "right": "issue:44908", + "accept": false, + "reason": "Processor timestamp math is unrelated to scheduler kwarg handling." + }, + { + "left": "issue:31795", + "right": "issue:34689", + "accept": false, + "reason": "Documentation confusion does not match a model loading regression." + }, + { + "left": "issue:41669", + "right": "issue:43704", + "accept": false, + "reason": "Import-star cleanup/performance work is unrelated to a VRAM leak in multi-threaded inference." + }, + { + "left": "issue:31795", + "right": "issue:37428", + "accept": false, + "reason": "Docs wording issue vs import error in flash attention utilities are not the same bug." + }, + { + "left": "issue:44315", + "right": "issue:45092", + "accept": false, + "reason": "Both involve model construction/loading, but the specific failure modes and causes differ." + }, + { + "left": "issue:44485", + "right": "issue:45468", + "accept": false, + "reason": "GLM-5 RoPE math and Gemma-4 audio positional encoding are distinct model implementations." + }, + { + "left": "issue:43704", + "right": "issue:44485", + "accept": false, + "reason": "VRAM leakage in dataloader threads is unrelated to RoPE implementation." + }, + { + "left": "issue:43502", + "right": "issue:43519", + "accept": false, + "reason": "Accidental API requests with local_files_only and Qwen3VL timestamp errors are unrelated." + }, + { + "left": "issue:42371", + "right": "issue:43704", + "accept": false, + "reason": "TF32 settings behavior and VRAM leakage are different subsystems." + }, + { + "left": "issue:43519", + "right": "issue:44623", + "accept": false, + "reason": "Processor timestamp calculation and missing save_pretrained files are separate issues." + }, + { + "left": "issue:43519", + "right": "issue:43704", + "accept": false, + "reason": "No shared code path between Qwen3VL timestamp handling and VRAM leaks." + }, + { + "left": "issue:44263", + "right": "issue:45468", + "accept": false, + "reason": "torch.split return values in a DSA indexer are unrelated to Gemma4Audio positional encoding." + }, + { + "left": "issue:43606", + "right": "issue:45468", + "accept": false, + "reason": "CPU offload device mismatch for bark-small is not the same bug as Gemma4Audio positional encoding." + }, + { + "left": "issue:44261", + "right": "issue:45468", + "accept": false, + "reason": "MLA q_a_layernorm precision/config issue is unrelated to Gemma4Audio positional encoding." + }, + { + "left": "issue:42913", + "right": "issue:43066", + "accept": false, + "reason": "Both mention tokenizer v5 behavior, but one is a broad behavioral regression while the other is a specific decoder-type problem; not enough to merge." + }, + { + "left": "issue:44908", + "right": "issue:44998", + "accept": false, + "reason": "Scheduler kwarg handling and an unrelated placeholder issue are not duplicates." + }, + { + "left": "issue:44492", + "right": "issue:44998", + "accept": false, + "reason": "Typo in cache strategies is unrelated to the placeholder issue." + }, + { + "left": "issue:44861", + "right": "issue:45290", + "accept": false, + "reason": "A tied-weights helper crash is unrelated to chat template crashes on tool-call assistant messages." + }, + { + "left": "issue:44998", + "right": "issue:45850", + "accept": false, + "reason": "'Unemployment' is not a meaningful bug report and does not describe the same problem as the broad v5 regression complaint." + }, + { + "left": "issue:43618", + "right": "issue:44998", + "accept": false, + "reason": "CLIPOutput attentions assignment regression is unrelated to the placeholder issue." + }, + { + "left": "issue:44998", + "right": "issue:45341", + "accept": false, + "reason": "Placeholder/empty issue does not match the testing_utils bug." + }, + { + "left": "issue:43525", + "right": "issue:44861", + "accept": false, + "reason": "Different AttributeErrors in different components; no shared underlying bug." + }, + { + "left": "issue:43572", + "right": "issue:44861", + "accept": false, + "reason": "Missing pad_token_idx in a config is unrelated to tied-weight key handling." + }, + { + "left": "issue:31515", + "right": "issue:38617", + "accept": false, + "reason": "Slow checkpoint loading and an import error in configuration_utils are unrelated." + }, + { + "left": "issue:44230", + "right": "issue:44623", + "accept": false, + "reason": "FP8 model support/inference and processor.save_pretrained file emission are different issues." + }, + { + "left": "issue:30333", + "right": "issue:31515", + "accept": false, + "reason": "MLflow reporting on failed training jobs and slow from_pretrained checkpoint loading are unrelated." + }, + { + "left": "issue:31515", + "right": "issue:36296", + "accept": false, + "reason": "Checkpoint load performance and tensor-parallel training bugs do not share the same concrete failure." + }, + { + "left": "issue:31515", + "right": "issue:31795", + "accept": false, + "reason": "Loading slowdown and documentation confusion are not duplicates." + }, + { + "left": "issue:44230", + "right": "issue:45310", + "accept": false, + "reason": "Both involve Qwen3.5 MoE, but one is FP8 support and the other is a from_pretrained error; too different to merge." + }, + { + "left": "issue:31515", + "right": "issue:33453", + "accept": false, + "reason": "Slow checkpoint loading is not the same as a tokenizer-loading regression." + }, + { + "left": "issue:43298", + "right": "issue:44230", + "accept": false, + "reason": "Continuous batching for audio models and Qwen3-VL/Qwen3.5 FP8 support are different features and failures." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 17, + "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": "f3c2fc74fc672ee7b42181d41b58e92f7eeb2ff72069f4764f07a170a591060a", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 2478, + "estimated_input_tokens": 620, + "estimated_eval_tokens": 1496 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 2478, + "estimated_input_tokens": 620, + "estimated_eval_tokens": 1496 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These issues are not duplicates: one is about Tapas tokenization diverging from TensorFlow code, one is about a misleading private-model download/login error, and one is a generic model-loading failure for bert-base-uncased.", + "confidence": 0.95, + "canonical_issue_reason": "#16618 is the broadest, most central user-facing failure in the set, but it still describes a different problem from the other two issues.", + "canonical_pr_reason": null, + "best_issue_reason": "#16618 is the best cluster anchor because its title is the most general and it has the strongest discussion activity, even though it is not the same underlying bug as the others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:13244", + "right": "issue:16618", + "accept": false, + "reason": "Tapas tokenization mismatch is a different code path and symptom from a failure to load bert-base-uncased; they are not the same underlying bug." + }, + { + "left": "issue:15980", + "right": "issue:16618", + "accept": false, + "reason": "A bad unauthenticated private-model download message is an authentication/download UX bug, not the same as a model-loading failure for bert-base-uncased." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 3, + "component_count": 17, + "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": "74bf751df619f32360334a6a972358487190ce061986e1f41300c39d495f69d5", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1589, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1589, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues are about different features: Flash Attention 2 support vs GGUF support. They share a \u201ccommunity contribution\u201d framing, but not the same underlying bug or change, so they should not be deduplicated.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "Issue #26350 is the more established representative of its own feature request, with higher discussion activity, but it is not a duplicate of #33260.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:26350", + "right": "issue:33260", + "accept": false, + "reason": "Different requested capabilities and code paths: Flash Attention 2 vs GGUF support. Similar wording about community contributions is too broad to indicate 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": 4, + "component_count": 17, + "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-27088-5", + "nodes": [ + "issue:27088", + "issue:28936", + "issue:43834", + "issue:44365", + "issue:44370" + ], + "soft_pairs": [ + "issue:27088|issue:44370", + "issue:27088|issue:44365", + "issue:27088|issue:43834", + "issue:27088|issue:28936", + "issue:28936|issue:43834", + "issue:28936|issue:44365", + "issue:28936|issue:44370", + "issue:43834|issue:44370", + "issue:43834|issue:44365", + "issue:44365|issue:44370" + ], + "prepared_review_unit_hash": "078481a518474127f496a6c37c7920be3b93d6cbe278327a0c8cb42c7cd56ad8", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 10, + "serialized_chars": 6443, + "estimated_input_tokens": 1611, + "estimated_eval_tokens": 3478 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 10, + "serialized_chars": 6443, + "estimated_input_tokens": 1611, + "estimated_eval_tokens": 3478 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The cluster is mostly a set of separate docs-translation requests by language, plus three nearly identical placeholder/template issues. Only the placeholder trio looks duplicate; the Turkish and Spanish issues are distinct tasks and should not be merged with them.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 43834 is the best representative of the duplicate subgroup because it is one of the near-identical placeholder translation issues and the earliest of the trio. It is not a good canonical for the full cluster, since the Turkish and Spanish requests are separate language-specific tasks.", + "canonical_pr_reason": null, + "best_issue_reason": "43834 is the clearest canonical candidate among the items because it matches the other placeholder issues almost exactly and is more representative than the language-specific requests.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:27088", + "right": "issue:44370", + "accept": false, + "reason": "Different concrete request: Turkish translation vs a generic placeholder translation template. Similar category, but not the same underlying issue." + }, + { + "left": "issue:27088", + "right": "issue:44365", + "accept": false, + "reason": "Turkish-specific translation request is not the same as the generic placeholder template issue." + }, + { + "left": "issue:27088", + "right": "issue:43834", + "accept": false, + "reason": "Distinct language-specific request versus generic placeholder template; not duplicates." + }, + { + "left": "issue:27088", + "right": "issue:28936", + "accept": false, + "reason": "Different target language and therefore a different translation task." + }, + { + "left": "issue:28936", + "right": "issue:43834", + "accept": false, + "reason": "Spanish translation request is separate from the generic placeholder template issue." + }, + { + "left": "issue:28936", + "right": "issue:44365", + "accept": false, + "reason": "Different concrete language request versus generic placeholder template." + }, + { + "left": "issue:28936", + "right": "issue:44370", + "accept": false, + "reason": "Spanish translation request is not the same underlying issue as the placeholder template." + }, + { + "left": "issue:43834", + "right": "issue:44370", + "accept": true, + "reason": "These two placeholder translation issues are effectively the same text and appear to be duplicate template filings." + }, + { + "left": "issue:43834", + "right": "issue:44365", + "accept": true, + "reason": "Near-identical placeholder issue title/body; same underlying template request." + }, + { + "left": "issue:44365", + "right": "issue:44370", + "accept": true, + "reason": "Near-exact duplicates of the same placeholder translation issue, differing only trivially." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet. The three generic placeholder issues are near-identical by title and similarity scores, while the Turkish and Spanish issues are clearly separate language-specific requests. The canonical choice is conservative, and the pairwise verdicts align with the evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 5, + "component_count": 17, + "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-36817-15", + "nodes": [ + "issue:36817", + "issue:36879", + "issue:37015", + "issue:41115", + "issue:41211", + "issue:42503", + "issue:42738", + "issue:43671", + "issue:43837", + "issue:43909", + "issue:44121", + "issue:44205", + "issue:45027", + "issue:45177", + "issue:45322" + ], + "soft_pairs": [ + "issue:36817|issue:41211", + "issue:36817|issue:43671", + "issue:36817|issue:36879", + "issue:36879|issue:41211", + "issue:36879|issue:43909", + "issue:36879|issue:44121", + "issue:36879|issue:43837", + "issue:36879|issue:43671", + "issue:37015|issue:43837", + "issue:37015|issue:44121", + "issue:37015|issue:43909", + "issue:36879|issue:37015", + "issue:37015|issue:42503", + "issue:41115|issue:43837", + "issue:41115|issue:44121", + "issue:41115|issue:42503", + "issue:41115|issue:45027", + "issue:41211|issue:43671", + "issue:42503|issue:43837", + "issue:42503|issue:43671", + "issue:42503|issue:42738", + "issue:42738|issue:45027", + "issue:43671|issue:43837", + "issue:43671|issue:45027", + "issue:43671|issue:45177", + "issue:43671|issue:44121", + "issue:43671|issue:43909", + "issue:43837|issue:43909", + "issue:43837|issue:45322", + "issue:43837|issue:44121", + "issue:43909|issue:44205", + "issue:44121|issue:45322", + "issue:43837|issue:45027", + "issue:37015|issue:45027", + "issue:43837|issue:45177" + ], + "prepared_review_unit_hash": "f5c42ecfea37824ab8a889fbc82ce06be998568a6593f3e41e78539daeec12d1", + "component_budget": { + "node_count": 15, + "item_count": 15, + "soft_pair_count": 35, + "serialized_chars": 20708, + "estimated_input_tokens": 5177, + "estimated_eval_tokens": 10610 + }, + "budget": { + "node_count": 15, + "item_count": 15, + "soft_pair_count": 35, + "serialized_chars": 20708, + "estimated_input_tokens": 5177, + "estimated_eval_tokens": 10610 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This is a broad thematic cluster of model-addition requests, not a true duplicate set. The issues mention different models, architectures, and modalities (vision, text, ASR/TTS, audio, OCR), so none of the soft edges should be merged as the same underlying change.", + "confidence": 0.96, + "canonical_issue_reason": "Closest anchor in this non-duplicate theme cluster: issue 43671 is a detailed, active support request and the best representative of the set, though it is not a duplicate of the others.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43671 is the strongest standalone representative here because it is specific, actively discussed, and clearly scoped, but it still should not subsume the other model requests.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36817", + "right": "issue:41211", + "accept": false, + "reason": "Different requested models and changes: EuroBert config vs DEIMv2 support." + }, + { + "left": "issue:36817", + "right": "issue:43671", + "accept": false, + "reason": "EuroBert config and Qwen3-TTS are unrelated model additions with different code paths." + }, + { + "left": "issue:36817", + "right": "issue:36879", + "accept": false, + "reason": "Different model families and tasks: EuroBert vs RF-DETR." + }, + { + "left": "issue:36879", + "right": "issue:41211", + "accept": false, + "reason": "RF-DETR and DEIMv2 are distinct detection models; not the same underlying change." + }, + { + "left": "issue:36879", + "right": "issue:43909", + "accept": false, + "reason": "RF-DETR vs LFM2.5 Audio are different domains and implementations." + }, + { + "left": "issue:36879", + "right": "issue:44121", + "accept": false, + "reason": "RF-DETR support and weight-sparse transformer support are unrelated code changes." + }, + { + "left": "issue:36879", + "right": "issue:43837", + "accept": false, + "reason": "RF-DETR and Qwen3-ASR are different model types and subsystems." + }, + { + "left": "issue:36879", + "right": "issue:43671", + "accept": false, + "reason": "RF-DETR and Qwen3-TTS are unrelated model additions; same broad repo area only." + }, + { + "left": "issue:37015", + "right": "issue:43837", + "accept": false, + "reason": "NeoBERT and Qwen3-ASR are different models with different integration paths." + }, + { + "left": "issue:37015", + "right": "issue:44121", + "accept": false, + "reason": "NeoBERT support and sparse-transformer support are not the same bug/change." + }, + { + "left": "issue:37015", + "right": "issue:43909", + "accept": false, + "reason": "NeoBERT is a text model; LFM2.5 Audio is an audio model." + }, + { + "left": "issue:36879", + "right": "issue:37015", + "accept": false, + "reason": "RF-DETR and NeoBERT are unrelated model requests." + }, + { + "left": "issue:37015", + "right": "issue:42503", + "accept": false, + "reason": "NeoBERT and ModernVBERT are distinct model requests, not one mergeable fix." + }, + { + "left": "issue:41115", + "right": "issue:43837", + "accept": false, + "reason": "MiniCPM3 architecture and Qwen3-ASR support are different concrete additions." + }, + { + "left": "issue:41115", + "right": "issue:44121", + "accept": false, + "reason": "MiniCPM3 and the sparse-transformer request target different code paths." + }, + { + "left": "issue:41115", + "right": "issue:42503", + "accept": false, + "reason": "MiniCPM3 and ModernVBERT are separate model support requests." + }, + { + "left": "issue:41115", + "right": "issue:45027", + "accept": false, + "reason": "MiniCPM3 architecture support and Voxtral TTS support are unrelated changes." + }, + { + "left": "issue:41211", + "right": "issue:43671", + "accept": false, + "reason": "DEIMv2 and Qwen3-TTS are different model integrations." + }, + { + "left": "issue:42503", + "right": "issue:43837", + "accept": false, + "reason": "ModernVBERT and Qwen3-ASR are different models and tasks." + }, + { + "left": "issue:42503", + "right": "issue:43671", + "accept": false, + "reason": "ModernVBERT and Qwen3-TTS are unrelated model support requests." + }, + { + "left": "issue:42503", + "right": "issue:42738", + "accept": false, + "reason": "One is a specific model request, the other is a broad RoPE architectural proposal; not the same concrete change." + }, + { + "left": "issue:42738", + "right": "issue:45027", + "accept": false, + "reason": "BERT-like RoPE support and Voxtral TTS are unrelated." + }, + { + "left": "issue:43671", + "right": "issue:43837", + "accept": false, + "reason": "Qwen3-TTS and Qwen3-ASR share a family name but target different modalities and integration work." + }, + { + "left": "issue:43671", + "right": "issue:45027", + "accept": false, + "reason": "Both are TTS-related requests, but for different models and code paths." + }, + { + "left": "issue:43671", + "right": "issue:45177", + "accept": false, + "reason": "Qwen3-TTS and DeepSeek-OCR2 are unrelated model requests." + }, + { + "left": "issue:43671", + "right": "issue:44121", + "accept": false, + "reason": "TTS support and sparse-transformer support are different underlying changes." + }, + { + "left": "issue:43671", + "right": "issue:43909", + "accept": false, + "reason": "Qwen3-TTS and LFM2.5 Audio are different audio-model integrations." + }, + { + "left": "issue:43837", + "right": "issue:43909", + "accept": false, + "reason": "Qwen3-ASR and LFM2.5 Audio are distinct audio-related model requests." + }, + { + "left": "issue:43837", + "right": "issue:45322", + "accept": false, + "reason": "Qwen3-ASR and EUPE are unrelated model additions." + }, + { + "left": "issue:43837", + "right": "issue:44121", + "accept": false, + "reason": "ASR support and weight-sparse transformer support are not the same code-path problem." + }, + { + "left": "issue:43909", + "right": "issue:44205", + "accept": false, + "reason": "LFM2.5 Audio and SAM3-LiteText are different model families and tasks." + }, + { + "left": "issue:44121", + "right": "issue:45322", + "accept": false, + "reason": "Sparse-transformer support and EUPE support are unrelated changes." + }, + { + "left": "issue:43837", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-ASR and Voxtral TTS are different modalities and model integrations." + }, + { + "left": "issue:37015", + "right": "issue:45027", + "accept": false, + "reason": "NeoBERT and Voxtral TTS are unrelated model support requests." + }, + { + "left": "issue:43837", + "right": "issue:45177", + "accept": false, + "reason": "Qwen3-ASR and DeepSeek-OCR2 are different model additions." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 6, + "component_count": 17, + "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-37934-7", + "nodes": [ + "issue:37934", + "issue:42936", + "issue:43240", + "issue:43668", + "issue:43813", + "issue:44031", + "issue:46018" + ], + "soft_pairs": [ + "issue:37934|issue:43668", + "issue:42936|issue:43240", + "issue:42936|issue:46018", + "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": "e6f6010e9df87c15bb65015914185133f4d91a24aefadcfc6298a457f8674f1e", + "component_budget": { + "node_count": 7, + "item_count": 7, + "soft_pair_count": 9, + "serialized_chars": 7185, + "estimated_input_tokens": 1797, + "estimated_eval_tokens": 3850 + }, + "budget": { + "node_count": 7, + "item_count": 7, + "soft_pair_count": 9, + "serialized_chars": 7185, + "estimated_input_tokens": 1797, + "estimated_eval_tokens": 3850 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These issues are only superficially similar: they span unrelated bugs in model config typing, masking, tokenization warnings, quantization typos, loss plumbing, and a gating logic question. No pair looks like the same underlying defect.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 43240 is the broadest and most infrastructure-level report in the set, with the highest inbound references; if one issue must anchor this noisy cluster, it is the least specific and most central title.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43240 is the best representative only in a loose triage sense because it describes a generic framework bug rather than a model-specific typo or one-off parameter issue.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:37934", + "right": "issue:43668", + "accept": false, + "reason": "Unrelated model behavior question vs. config type-hint issue; no shared code path." + }, + { + "left": "issue:42936", + "right": "issue:43240", + "accept": false, + "reason": "Mask2former ignore_value handling and loss kwargs plumbing are different bugs in different components." + }, + { + "left": "issue:42936", + "right": "issue:46018", + "accept": false, + "reason": "Masking loss usage and DeepSeek expert gating are separate model behaviors." + }, + { + "left": "issue:42936", + "right": "issue:44031", + "accept": false, + "reason": "Mask2former loss masking and tokenizer regex warnings are unrelated subsystems." + }, + { + "left": "issue:42936", + "right": "issue:43813", + "accept": false, + "reason": "Loss ignore_value usage and a quantization typo are not the same change." + }, + { + "left": "issue:42936", + "right": "issue:43668", + "accept": false, + "reason": "Different model areas: segmentation loss handling vs. ModernBERT config typing." + }, + { + "left": "issue:43240", + "right": "issue:43668", + "accept": false, + "reason": "Generic kwargs propagation bug does not match a type-hint correction." + }, + { + "left": "issue:43668", + "right": "issue:44031", + "accept": false, + "reason": "A config annotation bug is unrelated to tokenizer regex warning behavior." + }, + { + "left": "issue:43668", + "right": "issue:43813", + "accept": false, + "reason": "ModernBERT norm_eps typing and quantization typo are distinct documentation/code issues." + } + ] + }, + "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": 17, + "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": "8ab2a7c022e229ddfa5e8a81aa624b87117a481543133d8165cadac182c8e368", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1480, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1480, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues are not duplicates: one is about block decoding behavior, the other is a request for DeepSeek v3.2 support. They describe different feature areas and different underlying changes.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41084", + "right": "issue:43037", + "accept": false, + "reason": "Different requests and code paths: block decoding vs model support. No evidence they describe the same underlying bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 8, + "component_count": 17, + "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": "1769909196a533af01e6fdce040f2d38b9bbc139ee253e96083b6a7e68c879d3", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1584, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1584, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related to Trainer customizability, but they describe different problems: one is about `remove_unused_column` behavior, the other about FLOPS estimation for custom trainers. They should not be merged as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "No single issue is a clear duplicate canonical for the other; if forced to pick a representative, #43598 is the more concrete Trainer-behavior customization request.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43598", + "right": "issue:43600", + "accept": false, + "reason": "Different underlying concerns: one changes `remove_unused_column` handling, the other changes FLOPS estimation for custom trainers. Shared Trainer/custom-trainer context is too broad to treat as 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": 9, + "component_count": 17, + "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": "78c883e5ed914504c1e8c7663b8215f96574a617fc17b7d758a80191483c5129", + "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 not duplicates: one asks for Qwen3-Omni registration in AutoModel/AutoModelForConditionalGeneration, while the other asks for AutoModelForSequenceClassification support for Qwen3.5. They target different model families and different task heads.", + "confidence": 0.96, + "canonical_issue_reason": "Issue 43684 is the broader model-registration request and is the closer representative of the cluster topic, but it is still separate from the Qwen3.5 sequence-classification request.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43684 is the better single representative because it concerns the core registration gap in AutoModel mappings; issue 44405 is a distinct task-specific feature request.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43684", + "right": "issue:44405", + "accept": false, + "reason": "Different model families (Qwen3-Omni vs Qwen3.5) and different APIs (AutoModel/AutoModelForConditionalGeneration vs AutoModelForSequenceClassification); not the same underlying bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 10, + "component_count": 17, + "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": "b160aa6500515d4778cfb0d74988123d0af1fa56116c5d4c9d5b2b5e59117e27", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4266, + "estimated_input_tokens": 1067, + "estimated_eval_tokens": 2390 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4266, + "estimated_input_tokens": 1067, + "estimated_eval_tokens": 2390 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Issues 43845, 43846, and 43859 look like the same vague duplicate report cluster with the same title; 43835 is a different title and does not fit the cluster well.", + "confidence": 0.66, + "canonical_issue_reason": "Issue 43845 is the earliest of the near-identical 'huggingface' reports and is the best representative of that duplicate cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "43845 is the strongest issue candidate because it is earliest and most central to the repeated same-title reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43835", + "right": "issue:43845", + "accept": false, + "reason": "Different title ('Fett' vs 'huggingface') and no concrete evidence they describe the same underlying problem." + }, + { + "left": "issue:43835", + "right": "issue:43846", + "accept": false, + "reason": "Different title and no shared issue-specific details to justify treating them as duplicates." + }, + { + "left": "issue:43835", + "right": "issue:43859", + "accept": false, + "reason": "Different title with no substantive overlap beyond weak generic similarity." + }, + { + "left": "issue:43845", + "right": "issue:43846", + "accept": true, + "reason": "Same title, nearly simultaneous creation, and very similar metadata; likely duplicate reports of the same issue." + }, + { + "left": "issue:43845", + "right": "issue:43859", + "accept": true, + "reason": "Same title and close enough pattern to plausibly be the same vague underlying report cluster." + }, + { + "left": "issue:43846", + "right": "issue:43859", + "accept": true, + "reason": "Same title and strongly overlapping duplicate-style reporting; likely the same issue." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: 43845/43846/43859 share the same title, while 43835 has a different title. The duplicate-cluster wording is appropriately cautious given the lack of body/detail evidence, and the soft-edge verdicts stay conservative by not merging 43835 with the others." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 11, + "component_count": 17, + "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": "204ef58bf658e85fe1e47c9c7f90c918b6507dcb591ecfffd3ac546147e0e705", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 9, + "serialized_chars": 5791, + "estimated_input_tokens": 1448, + "estimated_eval_tokens": 3152 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 9, + "serialized_chars": 5791, + "estimated_input_tokens": 1448, + "estimated_eval_tokens": 3152 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster does not look like a true duplicate set. It is mostly spam/placeholder issues, plus one unrelated documentation suggestion and one unrelated \u201cGoogle LLC\u201d post.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44418 is the only non-spam, semi-descriptive item, so it is the best anchor if a representative must be chosen.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44371", + "right": "issue:44572", + "accept": false, + "reason": "Both are titled spam placeholders, but there is no concrete underlying bug or change to unify." + }, + { + "left": "issue:44371", + "right": "issue:44573", + "accept": false, + "reason": "Both are spam placeholders only; identical low-quality titles are not enough to treat them as the same issue." + }, + { + "left": "issue:44371", + "right": "issue:44418", + "accept": false, + "reason": "Spam placeholder versus a documentation enhancement request; clearly different subjects." + }, + { + "left": "issue:44371", + "right": "issue:44937", + "accept": false, + "reason": "Spam placeholder versus an unrelated \u201cGoogle LLC\u201d post; no shared bug or change." + }, + { + "left": "issue:44418", + "right": "issue:44572", + "accept": false, + "reason": "Documentation enhancement request versus spam placeholder; not the same issue." + }, + { + "left": "issue:44418", + "right": "issue:44573", + "accept": false, + "reason": "Documentation enhancement request versus spam placeholder; not a duplicate pair." + }, + { + "left": "issue:44572", + "right": "issue:44573", + "accept": false, + "reason": "Identical spam titles, but they do not indicate the same actionable underlying problem or change." + }, + { + "left": "issue:44572", + "right": "issue:44937", + "accept": false, + "reason": "Spam placeholder versus unrelated \u201cGoogle LLC\u201d post; no concrete overlap." + }, + { + "left": "issue:44573", + "right": "issue:44937", + "accept": false, + "reason": "Spam placeholder versus unrelated \u201cGoogle LLC\u201d post; 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": 12, + "component_count": 17, + "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": "59ca5d4b25b78e40611cf66b74bf046f7b0690ca9ebd1ce6bca3a6d6dba8dd96", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1645, + "estimated_input_tokens": 412, + "estimated_eval_tokens": 1080 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1645, + "estimated_input_tokens": 412, + "estimated_eval_tokens": 1080 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues appear to be duplicates of the same energy-optimization proposal/report: identical title, very similar body size, and strong textual overlap. Keep the earlier issue as canonical.", + "confidence": 0.99, + "canonical_issue_reason": "issue:45145 is the earlier-created issue and already has an inbound reference, so it is the best canonical representative for this duplicate pair.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45145 is the strongest issue to keep because it predates 45151 and appears to be the original thread for the same content.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45145", + "right": "issue:45151", + "accept": true, + "reason": "Same title and highly similar body suggest the same underlying issue/report, so this is a duplicate pair." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The conclusion is grounded in the packet: the issues have identical titles, similar body lengths, and a soft-similarity link with Jaccard 0.63. Choosing the earlier issue as canonical is conservative and supported by the timestamps and inbound reference. Confidence is a bit high, but not clearly overstated given the available evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 13, + "component_count": 17, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45295-2", + "nodes": [ + "issue:45295", + "issue:45373" + ], + "soft_pairs": [], + "prepared_review_unit_hash": "f5fd95fada2f530ab0999e343c17c16266cde0c7182bb22509972ca44827cd49", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 0, + "serialized_chars": 1286, + "estimated_input_tokens": 322, + "estimated_eval_tokens": 900 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 0, + "serialized_chars": 1286, + "estimated_input_tokens": 322, + "estimated_eval_tokens": 900 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues describe the same missing Gemma 4 sequence-classification support; the later, more specific open issue is the better cluster representative.", + "confidence": 0.94, + "canonical_issue_reason": "issue #45373 is the clearest canonical tracker: it is open, explicitly names the missing class `Gemma4ForSequenceClassification`, and states the module-level gap relative to Gemma 2/3.", + "canonical_pr_reason": null, + "best_issue_reason": "issue #45373 best captures the concrete request and likely remaining work, while #45295 is a shorter earlier duplicate with less detail and is already closed.", + "best_pr_reason": null, + "soft_edge_verdicts": [] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 14, + "component_count": 17, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45672-2", + "nodes": [ + "issue:45672", + "issue:45674" + ], + "soft_pairs": [ + "issue:45672|issue:45674" + ], + "prepared_review_unit_hash": "d3e169eca8fd35257fd7be825b3ae56a6c534f7f194e8249ad0e42d99d8a1f8a", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1668, + "estimated_input_tokens": 417, + "estimated_eval_tokens": 1090 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1668, + "estimated_input_tokens": 417, + "estimated_eval_tokens": 1090 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two distinct 4-bit inference bugs in different code paths: one in Gemma4 audio modules hitting uint8/torch.finfo, the other in BitsAndBytesConfig skip-module handling clearing lm_head exclusion. Not duplicates.", + "confidence": 0.97, + "canonical_issue_reason": "issue #45674 is the more general, framework-level 4-bit inference bug and has broader applicability than the Gemma4-specific audio-module failure.", + "canonical_pr_reason": null, + "best_issue_reason": "issue #45674 is the best representative because it describes a reusable configuration regression in BitsAndBytesConfig rather than a model-specific symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45672", + "right": "issue:45674", + "accept": false, + "reason": "Different failure modes and code paths: Gemma4 audio-module torch.finfo on uint8 weights vs. BitsAndBytesConfig skip-module logic removing default lm_head exclusion. They are not the same underlying bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 15, + "component_count": 17, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45762-2", + "nodes": [ + "issue:45762", + "issue:45763" + ], + "soft_pairs": [ + "issue:45762|issue:45763" + ], + "prepared_review_unit_hash": "860048ddeafcdd98ccdb1af7e67d6136bff062433cb6264ba274278034db581f", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1477, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1477, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two closed issues appear to be unrelated spam/PII-style entries with different personal names and dates, not duplicate bug reports.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45762", + "right": "issue:45763", + "accept": false, + "reason": "They only share a vague naming pattern; the titles contain different people and dates, with no evidence of the same underlying bug or request." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 16, + "component_count": 17, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45764-2", + "nodes": [ + "issue:45764", + "issue:45765" + ], + "soft_pairs": [ + "issue:45764|issue:45765" + ], + "prepared_review_unit_hash": "fd14d972b6cc1fdbead2cf2f2fdb401625a9f012aa28e865a66fc2ff9aecfbd8", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1481, + "estimated_input_tokens": 371, + "estimated_eval_tokens": 998 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1481, + "estimated_input_tokens": 371, + "estimated_eval_tokens": 998 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues are unrelated title-only entries that look like personal-name/date spam, not the same bug or change.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45764", + "right": "issue:45765", + "accept": false, + "reason": "No shared technical problem or code change; the titles are different personal-name/date strings and do not indicate duplicate bug reports." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 17, + "component_count": 17, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45869-5", + "nodes": [ + "issue:45869", + "issue:45870", + "issue:45871", + "issue:45872", + "issue:45873" + ], + "soft_pairs": [ + "issue:45869|issue:45870", + "issue:45869|issue:45871", + "issue:45870|issue:45871", + "issue:45870|issue:45872", + "issue:45870|issue:45873", + "issue:45871|issue:45872", + "issue:45871|issue:45873", + "issue:45872|issue:45873" + ], + "prepared_review_unit_hash": "eea9b421b4abed37db54a9b5e6fad5589dd9a2ccd59ba6824522fe695f55da64", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 8, + "serialized_chars": 5724, + "estimated_input_tokens": 1431, + "estimated_eval_tokens": 3118 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 8, + "serialized_chars": 5724, + "estimated_input_tokens": 1431, + "estimated_eval_tokens": 3118 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The cluster appears to contain two likely duplicate issue pairs: 45870/45871 and 45872/45873. Their identical titles and identical body lengths make them strong near-duplicates, though the full bodies are not visible here. Issue 45869 is a separate finance-related request and should remain separate.", + "confidence": 0.94, + "canonical_issue_reason": "45872 is a good representative open issue because it is concrete and has a very strong likely-duplicate twin in 45873. The 45870/45871 pair also looks duplicated, but 45872/45873 is the more specific request within the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "45872 is the best overall issue to keep as the representative item: it is specific, actionable, and its paired twin 45873 appears to be a likely duplicate. 45869 is different, and 45870/45871 are another likely duplicate pair but less distinctive.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45869", + "right": "issue:45870", + "accept": false, + "reason": "Same broad finance context, but this is a dataset bundle versus an example request; different underlying issue." + }, + { + "left": "issue:45869", + "right": "issue:45871", + "accept": false, + "reason": "Related domain wording, but the request type is different; not the same issue." + }, + { + "left": "issue:45870", + "right": "issue:45871", + "accept": true, + "reason": "Identical title and identical body length strongly suggest near-identical duplicate issues, even though the full bodies are not shown." + }, + { + "left": "issue:45870", + "right": "issue:45872", + "accept": false, + "reason": "Both are DataForge-related, but one is an example request and the other is a scraping-service request; different underlying work." + }, + { + "left": "issue:45870", + "right": "issue:45873", + "accept": false, + "reason": "Shared finance/time-series theme only; the requests appear different and are not the same issue." + }, + { + "left": "issue:45871", + "right": "issue:45872", + "accept": false, + "reason": "The pair shares only a broad subject area, not the same concrete request." + }, + { + "left": "issue:45871", + "right": "issue:45873", + "accept": false, + "reason": "Different request types despite related wording; not a duplicate pair." + }, + { + "left": "issue:45872", + "right": "issue:45873", + "accept": true, + "reason": "Identical title and identical body length make these look like near-identical duplicates, though the full body text is not available here." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The clustering is plausible, but the writeup is a bit too strong for the packet. The visible evidence supports very high similarity from identical titles and identical body lengths for 45870/45871 and 45872/45873, but it does not show the full issue bodies, so phrases like \"exact duplicate\" and \"matching issue content\" overstate what is directly evidenced. Please soften to \"likely duplicates\" or \"near-identical duplicates\" unless fuller content is available. The 45869 separation looks grounded." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 1, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-39895-12", + "nodes": [ + "pull_request:39895", + "pull_request:41855", + "pull_request:42668", + "pull_request:43251", + "pull_request:43254", + "pull_request:43424", + "pull_request:43823", + "pull_request:43995", + "pull_request:44827", + "pull_request:45170", + "pull_request:45213", + "pull_request:45655" + ], + "soft_pairs": [ + "pull_request:39895|pull_request:43424", + "pull_request:41855|pull_request:43823", + "pull_request:41855|pull_request:43424", + "pull_request:42668|pull_request:43823", + "pull_request:43251|pull_request:43823", + "pull_request:43251|pull_request:44827", + "pull_request:43424|pull_request:44827", + "pull_request:43424|pull_request:43823", + "pull_request:43823|pull_request:45655", + "pull_request:43823|pull_request:43995", + "pull_request:43995|pull_request:44827", + "pull_request:43995|pull_request:45213", + "pull_request:44827|pull_request:45170", + "pull_request:44827|pull_request:45655", + "pull_request:43823|pull_request:44827", + "pull_request:44827|pull_request:45213", + "pull_request:43823|pull_request:45213", + "pull_request:43251|pull_request:43254" + ], + "prepared_review_unit_hash": "35fb03b62c7d2aaa2ab84d6264f1424a245df4507f225aa27cc620ec25e6e9b3", + "component_budget": { + "node_count": 12, + "item_count": 12, + "soft_pair_count": 18, + "serialized_chars": 32877, + "estimated_input_tokens": 8220, + "estimated_eval_tokens": 16696 + }, + "budget": { + "node_count": 12, + "item_count": 12, + "soft_pair_count": 18, + "serialized_chars": 32877, + "estimated_input_tokens": 8220, + "estimated_eval_tokens": 16696 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: model additions, processor/tokenizer fixes, tests, and unrelated refactors. The only true duplicate-like pair is the two `fixed_cross_entropy` PRs targeting the same `loss_utils.py` bug.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single cluster-wide canonical PR fits because most items are unrelated. If forced to pick the nearest duplicate anchor, PR 43251 is the better canonical for the loss-utils fix since it directly targets issue 43240 and implements the broader cross-entropy kwargs pass-through.", + "best_issue_reason": null, + "best_pr_reason": "PR 43251 is the strongest representative of the only real duplicate cluster: it is tied to an explicit issue, touches the shared bug in `fixed_cross_entropy`, and is the broader implementation of that fix.", + "soft_edge_verdicts": [ + { + "left": "pull_request:39895", + "right": "pull_request:43424", + "accept": false, + "reason": "Videoprism model integration and an executorch dynamic-shape test are unrelated changes." + }, + { + "left": "pull_request:41855", + "right": "pull_request:43823", + "accept": false, + "reason": "A Mistral tokenizer method addition and a new MobileLLM model PR do not address the same bug or change." + }, + { + "left": "pull_request:41855", + "right": "pull_request:43424", + "accept": false, + "reason": "Tokenizer utility work and executorch export tests are unrelated." + }, + { + "left": "pull_request:42668", + "right": "pull_request:43823", + "accept": false, + "reason": "Processor auto-loading robustness and adding MobileLLM are different changes." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43823", + "accept": false, + "reason": "Cross-entropy kwargs support and MobileLLM model addition are unrelated code paths." + }, + { + "left": "pull_request:43251", + "right": "pull_request:44827", + "accept": false, + "reason": "The loss-utils fix and Mistral4 test updates do not look like the same bug or change." + }, + { + "left": "pull_request:43424", + "right": "pull_request:44827", + "accept": false, + "reason": "Executorch dynamic-shape export testing and Mistral4 test fixes are unrelated." + }, + { + "left": "pull_request:43424", + "right": "pull_request:43823", + "accept": false, + "reason": "Executorch export test coverage and MobileLLM model scaffolding are unrelated." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45655", + "accept": false, + "reason": "Adding MobileLLM and changing config-class resolution order are separate changes." + }, + { + "left": "pull_request:43823", + "right": "pull_request:43995", + "accept": false, + "reason": "MobileLLM model addition and Falcon output-collection refactoring are unrelated." + }, + { + "left": "pull_request:43995", + "right": "pull_request:44827", + "accept": false, + "reason": "Falcon refactoring and Mistral4 test fixes address different code paths." + }, + { + "left": "pull_request:43995", + "right": "pull_request:45213", + "accept": false, + "reason": "A Falcon model refactor and the model-creation skill PR are not the same underlying change." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45170", + "accept": false, + "reason": "Mistral4 test fixes and the `layrnorm` typo normalization patch are different problems." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45655", + "accept": false, + "reason": "Mistral4 tests and config resolution-order fixes are unrelated." + }, + { + "left": "pull_request:43823", + "right": "pull_request:44827", + "accept": false, + "reason": "MobileLLM addition and Mistral4 test changes are unrelated." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45213", + "accept": false, + "reason": "Mistral4 tests and the do-not-merge model-creation skill are unrelated." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45213", + "accept": false, + "reason": "MobileLLM model addition and the model-creation skill PR are not the same change." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43254", + "accept": true, + "reason": "Both PRs fix the same `loss_utils.fixed_cross_entropy` issue in `nn.functional.cross_entropy`, share issue target 43240, and could plausibly be merged into one PR." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The assessment is grounded in the packet and stays conservative. The only accepted pair, PR 43251 and PR 43254, is legitimately mergeable as the same `fixed_cross_entropy`/`nn.functional.cross_entropy` bugfix with the shared issue target 43240. The remaining soft-edge rejections are appropriately cautious given the clearly unrelated file sets and PR intents. No evidence is overstated." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 2, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-40546-2", + "nodes": [ + "pull_request:40546", + "pull_request:42781" + ], + "soft_pairs": [ + "pull_request:40546|pull_request:42781" + ], + "prepared_review_unit_hash": "2a4f971bef35e9e29e22f5684893aa54c314ffabd62d5f7a31ecf85db72f43c2", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7225, + "estimated_input_tokens": 1807, + "estimated_eval_tokens": 3870 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7225, + "estimated_input_tokens": 1807, + "estimated_eval_tokens": 3870 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related VibeVoice PRs, but they are not duplicates: one is the initial VibeVoice integration, the other adds realtime/VibeVoice acoustic-tokenizer support and related generation changes.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 40546 is the broader base implementation of VibeVoice and establishes the core integration that the later realtime work builds on.", + "best_issue_reason": null, + "best_pr_reason": "PR 40546 is the best standalone PR to represent the cluster because it introduces the primary VibeVoice model support; PR 42781 is a separate follow-up feature set for realtime/tokenizer support.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40546", + "right": "pull_request:42781", + "accept": false, + "reason": "Do not merge: 40546 implements the base VibeVoice model, while 42781 adds realtime-specific functionality plus acoustic-tokenizer and generation updates. They touch the same family of files but address different concrete features rather than the same underlying change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 3, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-40755-2", + "nodes": [ + "pull_request:40755", + "pull_request:41224" + ], + "soft_pairs": [ + "pull_request:40755|pull_request:41224" + ], + "prepared_review_unit_hash": "bb6639e9c7fdec288ce81d466e59e744ae68874d98fb7d566e7b2f995dba8903", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4675, + "estimated_input_tokens": 1169, + "estimated_eval_tokens": 2594 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4675, + "estimated_input_tokens": 1169, + "estimated_eval_tokens": 2594 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two pull requests are unrelated: one adds covariate-based forecasting support for TimesFM, while the other adds image-classification support for DINOv3ViT. They do not describe the same bug or change and should not be deduplicated.", + "confidence": 0.99, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "pull_request:40755", + "right": "pull_request:41224", + "accept": false, + "reason": "Different models and code paths (TimesFM forecasting with covariates vs DINOv3 image classification support); no shared target issue or filenames, and they 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": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41356-2", + "nodes": [ + "pull_request:41356", + "pull_request:44339" + ], + "soft_pairs": [ + "pull_request:41356|pull_request:44339" + ], + "prepared_review_unit_hash": "aead1308c167b0bae6ff11bf9358340baa5e76f816618383418b505a10a9cd01", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6050, + "estimated_input_tokens": 1513, + "estimated_eval_tokens": 3282 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6050, + "estimated_input_tokens": 1513, + "estimated_eval_tokens": 3282 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two DEIMv2 pull requests target the same model family, but they are not the same change: PR 41356 is an early/basic model + image processor add, while PR 44339 is a much broader follow-up that adds loss code, auto-mappings, conversion tooling, modular files, and more integration work.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44339 is the more complete and representative DEIMv2 implementation; PR 41356 looks like an earlier partial version of the same model bring-up.", + "best_issue_reason": null, + "best_pr_reason": "PR 44339 is the strongest single artifact for the cluster because it subsumes the basic model work and includes the broader integration needed for a full Transformers addition.", + "soft_edge_verdicts": [ + { + "left": "pull_request:41356", + "right": "pull_request:44339", + "accept": false, + "reason": "Same model area, but not the same concrete change: 44339 expands the implementation substantially and adds new code paths beyond the basic PR 41356. They look like related iterations, not mergeable duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 5, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41797-3", + "nodes": [ + "pull_request:41797", + "pull_request:45186", + "pull_request:45630" + ], + "soft_pairs": [ + "pull_request:41797|pull_request:45186", + "pull_request:45186|pull_request:45630" + ], + "prepared_review_unit_hash": "456cf687b93956a56d097e9651a6ec0e562d3c10a8f51d169fdf941f380bf54a", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 10409, + "estimated_input_tokens": 2603, + "estimated_eval_tokens": 5462 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 10409, + "estimated_input_tokens": 2603, + "estimated_eval_tokens": 5462 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All three items are separate new-model addition PRs, not duplicates. They share scaffold files and auto-registry touchpoints, but each introduces a different model family and codepath.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45630 is the best cluster anchor: it has the highest review activity and the broadest integration surface, but it is still not a duplicate of the others.", + "best_issue_reason": null, + "best_pr_reason": "PR 45630 is the most substantial and best-documented representative of the cluster, though the cluster members are functionally unrelated.", + "soft_edge_verdicts": [ + { + "left": "pull_request:41797", + "right": "pull_request:45186", + "accept": false, + "reason": "DeepSeek OCR vs Isaac are different model additions with different docs/tests and implementation codepaths; shared boilerplate files are not enough to treat them as the same change." + }, + { + "left": "pull_request:45186", + "right": "pull_request:45630", + "accept": false, + "reason": "Isaac and Kimi2-6 add different models and different supporting modules; they only overlap on standard registry/docs scaffolding, so they are not 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": 6, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-42975-2", + "nodes": [ + "pull_request:42975", + "pull_request:42976" + ], + "soft_pairs": [ + "pull_request:42975|pull_request:42976" + ], + "prepared_review_unit_hash": "b57585c1a91e6554c1f095fb10763e606d44856dba816765ef42584aa799786b", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5935, + "estimated_input_tokens": 1484, + "estimated_eval_tokens": 3224 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5935, + "estimated_input_tokens": 1484, + "estimated_eval_tokens": 3224 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two open PRs both update GitHub Actions versions in workflow files, but they address different upgrade tracks: one is specifically for Node 24 compatibility, the other is a broader \u201clatest versions\u201d refresh centered on docker/build-push-action. They overlap in files but not in the underlying change intent.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 42975 is the more specific and representative workflow-maintenance change: it targets Node 24 compatibility and updates several core GitHub Actions pins across many workflows.", + "best_issue_reason": null, + "best_pr_reason": "PR 42975 is the best representative of the cluster because its title and edits align cleanly around one concrete maintenance goal, whereas PR 42976 is a separate, broader action-version bump.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42975", + "right": "pull_request:42976", + "accept": false, + "reason": "Reject: these are both GitHub Actions upgrade PRs, but they fix different things. 42975 is about Node 24 compatibility for checkout/cache/upload-artifact, while 42976 updates docker/build-push-action and a few related workflow actions. They share workflow files but not the same concrete bug/change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 7, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-42978-7", + "nodes": [ + "pull_request:42978", + "pull_request:43448", + "pull_request:43451", + "pull_request:43973", + "pull_request:44601", + "pull_request:45490", + "pull_request:45838" + ], + "soft_pairs": [ + "pull_request:42978|pull_request:43448", + "pull_request:42978|pull_request:43451", + "pull_request:42978|pull_request:45490", + "pull_request:43448|pull_request:43451", + "pull_request:43448|pull_request:45490", + "pull_request:43451|pull_request:45490", + "pull_request:43973|pull_request:45838", + "pull_request:44601|pull_request:45838", + "pull_request:45490|pull_request:45838" + ], + "prepared_review_unit_hash": "4a63e5a20df5f4a80837f5a2f06f80d488cade6308509f691f136d018d780c60", + "component_budget": { + "node_count": 7, + "item_count": 7, + "soft_pair_count": 9, + "serialized_chars": 23542, + "estimated_input_tokens": 5886, + "estimated_eval_tokens": 12028 + }, + "budget": { + "node_count": 7, + "item_count": 7, + "soft_pair_count": 9, + "serialized_chars": 23542, + "estimated_input_tokens": 5886, + "estimated_eval_tokens": 12028 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster does not contain duplicate PRs; it mixes unrelated model-addition PRs and one distributed-loading PR. The shared file patterns are mostly boilerplate (docs/toctree, auto registries), not evidence of the same underlying change.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "pull_request:42978", + "right": "pull_request:43448", + "accept": false, + "reason": "Different models: ViT NEPA vs Molmo. Overlap is only generic model-registration/docs files, not the same code-path or feature." + }, + { + "left": "pull_request:42978", + "right": "pull_request:43451", + "accept": false, + "reason": "Different models: ViT NEPA vs Molmo2. Shared auto/docs files are boilerplate for adding new architectures, not a duplicate change." + }, + { + "left": "pull_request:42978", + "right": "pull_request:45490", + "accept": false, + "reason": "Different models and purposes: ViT NEPA vs CTSM. Same broad pattern of adding a model, but not the same underlying bug or implementation." + }, + { + "left": "pull_request:43448", + "right": "pull_request:43451", + "accept": false, + "reason": "Molmo and Molmo2 are distinct model families/releases with different files and implementations; not mergeable as one PR." + }, + { + "left": "pull_request:43448", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo vs CTSM are unrelated model additions; shared registry/docs edits are incidental." + }, + { + "left": "pull_request:43451", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo2 and CTSM are unrelated architectures; no concrete shared bug or code path." + }, + { + "left": "pull_request:43973", + "right": "pull_request:45838", + "accept": false, + "reason": "Lfm2 Audio and PP-OCRv6 text recognition are completely different modalities and model stacks; the overlap is just new-model scaffolding." + }, + { + "left": "pull_request:44601", + "right": "pull_request:45838", + "accept": false, + "reason": "One is distributed pipeline-parallel loading infrastructure, the other is a model-support PR. They do not fix the same concrete problem." + }, + { + "left": "pull_request:45490", + "right": "pull_request:45838", + "accept": false, + "reason": "CTSM and PP-OCRv6 are unrelated model additions; any similarity is only in shared auto/docs plumbing." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 8, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43291-3", + "nodes": [ + "pull_request:43291", + "pull_request:43665", + "pull_request:44815" + ], + "soft_pairs": [ + "pull_request:43291|pull_request:43665", + "pull_request:43291|pull_request:44815", + "pull_request:43665|pull_request:44815" + ], + "prepared_review_unit_hash": "d26ab0ef44374118fa7e3068487b43f0936f87b2457c4d9f3c9390736acbfc0a", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 7093, + "estimated_input_tokens": 1774, + "estimated_eval_tokens": 3804 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 7093, + "estimated_input_tokens": 1774, + "estimated_eval_tokens": 3804 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These pull requests do not look like duplicates of the same change. PR 44815 is a substantive FP8 dequant/loading fix, PR 43291 is a Whisper tokenizer/test fix, and PR 43665 appears to be test-only noise for CLIP/ViT rather than a real bug fix.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44815 is the strongest canonical candidate because it contains the clearest concrete code-path fix with implementation changes across loading/FP8 logic and tests.", + "best_issue_reason": null, + "best_pr_reason": "PR 44815 is the best representative of a real fix in the cluster; the other PRs are either unrelated or mostly test-focused.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43291", + "right": "pull_request:43665", + "accept": false, + "reason": "Whisper tokenizer decoding/tests versus CLIP/ViT test assertions; different models and different bugs, so not the same underlying change." + }, + { + "left": "pull_request:43291", + "right": "pull_request:44815", + "accept": false, + "reason": "Whisper decode behavior fix and FP8/dequant loading fix are unrelated code paths with no shared concrete bug." + }, + { + "left": "pull_request:43665", + "right": "pull_request:44815", + "accept": false, + "reason": "One is test-only CLIP/ViT edits, the other is a broad FP8/dequant implementation fix; they are not mergeable as one duplicate fix." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 9, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43488-2", + "nodes": [ + "pull_request:43488", + "pull_request:43532" + ], + "soft_pairs": [ + "pull_request:43488|pull_request:43532" + ], + "prepared_review_unit_hash": "e207655d40f016347a1892651e611f647ab65beb4d13eda55b00055b9e21c9af", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4562, + "estimated_input_tokens": 1141, + "estimated_eval_tokens": 2538 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4562, + "estimated_input_tokens": 1141, + "estimated_eval_tokens": 2538 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two pull requests are not duplicates: one is a repo-bot/setup/version-format test PR, while the other rewires CI/comment workflows and touches testing/notification code. The only shared file is a test file, which is not enough to treat them as the same change.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: the PRs address different code paths and purposes, so there is no single representative duplicate target here.", + "best_issue_reason": null, + "best_pr_reason": "PR #43532 is the better standalone representative of a substantive code change because it spans CI/workflow and related utility updates, whereas #43488 is mainly a bot/check-oriented test PR.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43488", + "right": "pull_request:43532", + "accept": false, + "reason": "Different underlying changes: #43488 adjusts setup/dependency/version metadata and a vit test; #43532 changes workflow triggers plus testing and notification utilities. Shared test filename is incidental." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 10, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43612-2", + "nodes": [ + "pull_request:43612", + "pull_request:43613" + ], + "soft_pairs": [ + "pull_request:43612|pull_request:43613" + ], + "prepared_review_unit_hash": "4f3bd470d5da9da6dac3297bfe7c7ff4cad64e0d30d05f0a5aa125c5bd661219", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6082, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6082, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two separate feature PRs for different pipelines: promptable concept segmentation vs promptable visual segmentation. They share a common implementation pattern and some documentation/registration files, but they do not fix the same underlying bug or change the same concrete code path.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR for deduplication: the PRs are distinct feature additions targeting different pipelines and model sets.", + "best_issue_reason": null, + "best_pr_reason": "PR 43613 is the broader representative if one must be chosen, since it covers more related model integrations and tests, but it is still not a duplicate of 43612.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43612", + "right": "pull_request:43613", + "accept": false, + "reason": "Both add different pipeline features with different names, docs, tests, and model integrations; overlap is broad infrastructure, not the same concrete change, so they should not be merged as duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 11, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43651-2", + "nodes": [ + "pull_request:43651", + "pull_request:43663" + ], + "soft_pairs": [ + "pull_request:43651|pull_request:43663" + ], + "prepared_review_unit_hash": "d239e1c3b0d5174477fa3aae9cba35c07573f35e8b1b57342999a94ea5c77370", + "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": "The two PRs both edit `trainer.py`, but they implement different trainer extensibility changes: one alters gradient-accumulation loss scaling, the other factors out signature-column filtering. Shared file/subsystem alone is not enough to treat them as duplicates.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "No clear canonical duplicate: PR #43651 changes loss scaling behavior in `training_step`, while PR #43663 adds an overridable helper for signature-column filtering. They target different code paths and behaviors.", + "best_issue_reason": null, + "best_pr_reason": "PR #43651 is the better representative only in a loose triage sense because it is a focused, concrete trainer behavior change; however, it is not a duplicate of PR #43663.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43651", + "right": "pull_request:43663", + "accept": false, + "reason": "Different changes in the same file: #43651 introduces `_loss_is_scaled_for_ga` and modifies gradient-accumulation loss scaling, while #43663 adds `_get_signature_columns` for column filtering. They do not fix the same underlying bug or implement the same change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 12, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43656-3", + "nodes": [ + "pull_request:43656", + "pull_request:43836", + "pull_request:43842" + ], + "soft_pairs": [ + "pull_request:43656|pull_request:43842", + "pull_request:43836|pull_request:43842" + ], + "prepared_review_unit_hash": "78c95989db646c7fbb6fe2b928b0b2f8d1ea38de6924156900794a5cc5d20890", + "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 touch `src/transformers/cli/serve.py` and aim at the same runtime annotation/`TypeAdapter` import-time NameError family. PR #43842 is the clearest, most targeted fix; #43836 is a close duplicate variant; #43656 appears to address the same underlying CLI annotation NameError more broadly.", + "confidence": 0.84, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43842 is the most explicit and focused fix: it directly addresses `TypeAdapter` being unavailable at runtime when `pydantic` is not installed, with a minimal change set and a clear title.", + "best_issue_reason": null, + "best_pr_reason": "PR #43842 is the best representative because it is the most direct, low-risk, and narrowly scoped solution to the same `TypeAdapter` NameError in the CLI.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43656", + "right": "pull_request:43842", + "accept": true, + "reason": "Both PRs target the same `serve.py` annotation/import-time NameError in the CLI and are alternative fixes for the same underlying runtime typing problem." + }, + { + "left": "pull_request:43836", + "right": "pull_request:43842", + "accept": true, + "reason": "Both PRs explicitly target the same `TypeAdapter` NameError in `src/transformers/cli/serve.py`; #43836 is just a broader/rougher variant of the same fix." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The grouping is grounded in the packet: all three PRs modify `src/transformers/cli/serve.py` and their titles/diffs point to the same TypeAdapter-related NameError class. The 43836\u201343842 link is especially strong because they share issue target 43824. The 43656 link is looser, but still plausible based on the title and diff preview, so the soft-edge verdicts stay within a conservative interpretation." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 13, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43779-2", + "nodes": [ + "pull_request:43779", + "pull_request:43816" + ], + "soft_pairs": [ + "pull_request:43779|pull_request:43816" + ], + "prepared_review_unit_hash": "e7c674adcc031aa8883109a5dfc678b4407fb20b10764302092e384d031ba83a", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4362, + "estimated_input_tokens": 1091, + "estimated_eval_tokens": 2438 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4362, + "estimated_input_tokens": 1091, + "estimated_eval_tokens": 2438 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both pull requests appear to implement the same SwanLab integration change: adding run id/resume support in `integration_utils.py` for issue #43698, with one variant also expanding the docs. They are strong duplicate PR candidates.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43816 is the cleaner representative: it carries the same SwanLab id/resume support plus the accompanying documentation updates, making it the more complete version of the shared change.", + "best_issue_reason": null, + "best_pr_reason": "PR #43816 is the best global representative because it appears to include the full feature change and docs for SwanLab run id/resume support, while #43779 is the narrower implementation-focused variant.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43779", + "right": "pull_request:43816", + "accept": true, + "reason": "Same underlying SwanLab callback enhancement for id/resume support in the same file and same issue target; the doc differences do not change the core code-path change, so these are duplicate PRs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: both PRs target issue #43698, touch the same file, and the previews indicate the same SwanLab id/resume support with #43816 adding docs. The duplicate/mergeable verdict is conservative enough given the shared target and overlapping code change." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 14, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 2, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43996-28", + "nodes": [ + "pull_request:43996", + "pull_request:43997", + "pull_request:44007", + "pull_request:44010", + "pull_request:44013", + "pull_request:44017", + "pull_request:44018", + "pull_request:44019", + "pull_request:44044", + "pull_request:44056", + "pull_request:44059", + "pull_request:44066", + "pull_request:44068", + "pull_request:44071", + "pull_request:44072", + "pull_request:44073", + "pull_request:44074", + "pull_request:44076", + "pull_request:44085", + "pull_request:44086", + "pull_request:44098", + "pull_request:44101", + "pull_request:44116", + "pull_request:44129", + "pull_request:44154", + "pull_request:44161", + "pull_request:44722" + ], + "soft_pairs": [ + "pull_request:43996|pull_request:44085", + "pull_request:43996|pull_request:44044", + "pull_request:44066|pull_request:44085", + "pull_request:44007|pull_request:44072", + "pull_request:44072|pull_request:44722", + "pull_request:44066|pull_request:44072", + "pull_request:44013|pull_request:44044", + "pull_request:44066|pull_request:44086", + "pull_request: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:44073|pull_request:44101", + "pull_request:44059|pull_request:44071", + "pull_request:44056|pull_request:44101", + "pull_request:44073|pull_request:44161", + "pull_request:44059|pull_request:44068", + "pull_request:44018|pull_request:44059", + "pull_request:44101|pull_request:44161", + "pull_request:43997|pull_request:44017", + "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:44018|pull_request:44161", + "pull_request:43997|pull_request:44071", + "pull_request:44010|pull_request:44071", + "pull_request:44018|pull_request:44129", + "pull_request:44010|pull_request:44073", + "pull_request:43997|pull_request:44074", + "pull_request:43997|pull_request:44129", + "pull_request:44071|pull_request:44129", + "pull_request:44010|pull_request:44018", + "pull_request:43996|pull_request:44161", + "pull_request:43996|pull_request:44017", + "pull_request:44010|pull_request:44044", + "pull_request:44044|pull_request:44074", + "pull_request:44066|pull_request:44722", + "pull_request:44085|pull_request:44722", + "pull_request:44085|pull_request:44129", + "pull_request:43996|pull_request:44007", + "pull_request:43996|pull_request:44129", + "pull_request:44044|pull_request:44129", + "pull_request:44044|pull_request:44154" + ], + "prepared_review_unit_hash": "55e6b5a1a451cf1c254e32e2870c6486cfb51cba7adc86d49a79ae7aa7ebde7e", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 78, + "serialized_chars": 84898, + "estimated_input_tokens": 21225, + "estimated_eval_tokens": 42706 + }, + "budget": { + "node_count": 27, + "item_count": 27, + "soft_pair_count": 72, + "serialized_chars": 75093, + "estimated_input_tokens": 18774, + "estimated_eval_tokens": 37804 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a topical cluster of output-tracing refactors across many unrelated model families. Most pairs share only the same tracking issue or a generic refactor theme; only same-model variants like GPT-J/CodeGen and GPT-Neo look like true duplicates.", + "confidence": 0.87, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single PR cleanly canonicalizes the whole cluster because it spans many different models. If forced to pick a representative, 44722 is the strongest canonical PR for the GPT-J/CodeGen subgroup: it is the latest, most consolidated version of that concrete output-tracing refactor.", + "best_issue_reason": null, + "best_pr_reason": "44722 is the best standalone PR representative among the near-duplicates because it covers the same concrete GPT-J/CodeGen output-tracing change with the broadest, most polished implementation. It is not a global canonical for the entire heterogeneous cluster.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43996", + "right": "pull_request:44085", + "accept": false, + "reason": "Same generic output-tracing topic and issue target, but different model families (CVT/FNet vs GPT-J) and unrelated code paths." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44044", + "accept": false, + "reason": "Both are output-tracing refactors, but they touch different models and implementations; not the same concrete change." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44085", + "accept": true, + "reason": "Both edit the same GPT-J output-tracing code path in modeling_gptj.py; the change is substantively the same refactor, with 44066 just broader." + }, + { + "left": "pull_request:44007", + "right": "pull_request:44072", + "accept": false, + "reason": "Same umbrella issue, but different architectures and files; these are separate model-specific refactors." + }, + { + "left": "pull_request:44072", + "right": "pull_request:44722", + "accept": false, + "reason": "Both are standardized output-tracing PRs, but EfficientNet and GPT-J are unrelated code paths and cannot be merged as one fix." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44072", + "accept": false, + "reason": "Different model families and different implementation details; the shared issue target is too broad to make them duplicates." + }, + { + "left": "pull_request:44013", + "right": "pull_request:44044", + "accept": false, + "reason": "MobileNetV2 and DeBERTa are unrelated code paths; they only share the broad output-tracing theme." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J and MGP-STR are different models; the PRs are thematically similar but not the same change." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-J vs MPT; same refactor pattern, but not the same concrete bug/change." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44068", + "accept": true, + "reason": "Same GPT-Neo files and same decorator-based output-tracing refactor; these are clearly the same underlying change." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44068", + "accept": false, + "reason": "GPT-J/CodeGen and GPT-Neo are different model implementations, so this is only thematic similarity." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44066", + "accept": false, + "reason": "Different models and code paths; they share a refactor style, not the same fix." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44722", + "accept": false, + "reason": "ResNet output tracing and GPT-J output tracing are unrelated model changes." + }, + { + "left": "pull_request:44086", + "right": "pull_request:44722", + "accept": false, + "reason": "MGP-STR and GPT-J are different code paths; same theme, different fixes." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44085", + "accept": false, + "reason": "ResNet and GPT-J are unrelated implementations, so this is not a duplicate." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44019", + "accept": false, + "reason": "GPT-Neo and ResNet are different models with different output-shaping logic." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44071", + "accept": false, + "reason": "ResNet and MPT are separate refactors, only loosely related by the tracking theme." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo and Flaubert/XLM are unrelated model code paths." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-Neo and GPT-J are different model families; same refactor pattern but not the same PR." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo and MGP-STR are unrelated implementations." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44085", + "accept": false, + "reason": "GPT-Neo vs GPT-J; same issue theme, but not the same underlying change." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44085", + "accept": false, + "reason": "MPT and GPT-J are different model code paths." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo and MGP-STR are unrelated refactors." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT2 and MGP-STR are different models with different output-tracing implementations." + }, + { + "left": "pull_request:44098", + "right": "pull_request:44101", + "accept": false, + "reason": "ViLT and Flaubert/XLM are different code paths; only the generic theme overlaps." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo and Flaubert/XLM are separate model implementations." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44722", + "accept": false, + "reason": "Segformer and GPT-J are unrelated model-specific refactors." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44161", + "accept": false, + "reason": "Segformer and LongT5 are different code paths; shared issue target is too broad." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44101", + "accept": false, + "reason": "VisualBert and Flaubert/XLM are unrelated model implementations." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44068", + "accept": false, + "reason": "GPT2 and GPT-Neo have different blocks and output handling; not the same change." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44101", + "accept": false, + "reason": "TextNet and Flaubert/XLM are separate model-specific refactors." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44161", + "accept": false, + "reason": "MPNet and LongT5 are different models; same generic feature, different fixes." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44017", + "accept": false, + "reason": "RegNet and Segformer are unrelated code paths despite the shared tracking issue." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44074", + "accept": false, + "reason": "RegNet and TextNet are distinct model implementations." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44018", + "accept": false, + "reason": "RegNet and GPT-Neo are unrelated model-specific changes." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44722", + "accept": false, + "reason": "MPT and GPT-J are different architectures; not a duplicate change." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT2 and MPT are different code paths." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44101", + "accept": false, + "reason": "MPNet and Flaubert/XLM are different refactors in different files." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44161", + "accept": false, + "reason": "TextNet and LongT5 are separate model implementations." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44056", + "accept": false, + "reason": "RegNet and MPNet are unrelated model families." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44076", + "accept": false, + "reason": "MPT and ImageGPT are different model code paths." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44161", + "accept": false, + "reason": "RegNet and LongT5 are unrelated changes." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44017", + "accept": false, + "reason": "CVT/FNet and Segformer are different model-specific refactors." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44073", + "accept": false, + "reason": "SqueezeBert and VisualBert are unrelated model implementations." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44018", + "accept": false, + "reason": "SqueezeBert and GPT-Neo only share the broad output-tracing theme." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44044", + "accept": false, + "reason": "SqueezeBert and DeBERTa are different code paths." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44074", + "accept": false, + "reason": "DeBERTa and TextNet are unrelated refactors." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44722", + "accept": true, + "reason": "Same GPT-J output-tracing refactor, and both touch the same concrete file; 44722 is effectively a broader variant of the same change." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44722", + "accept": true, + "reason": "Same GPT-J output-tracing change in the same file; the broader 44722 PR subsumes the GPT-J-specific part." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44129", + "accept": false, + "reason": "GPT-J and SpeechT5 are unrelated model code paths." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44007", + "accept": false, + "reason": "Different model families and different concrete output-tracing implementations." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44129", + "accept": false, + "reason": "DeBERTa and SpeechT5 are separate model-specific changes." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44154", + "accept": false, + "reason": "DeBERTa and Vits are unrelated implementations." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: it correctly treats the cluster as mostly heterogeneous, and it only elevates the GPT-J/CodeGen and GPT-Neo same-file variants as true duplicates. The accepted soft-edge pairs are conservative and look mergeable into single bugfix PRs, while the rejected pairs stay appropriately scoped to different model/code paths." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 14, + "component_count": 25, + "review_unit_index": 2, + "review_unit_count": 2, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43996-28", + "nodes": [ + "pull_request:43997", + "pull_request:44010", + "pull_request:44024", + "pull_request:44056", + "pull_request:44073", + "pull_request:44074", + "pull_request:44076", + "pull_request:44129" + ], + "soft_pairs": [ + "pull_request:44024|pull_request:44076", + "pull_request:44073|pull_request:44074", + "pull_request:44056|pull_request:44073", + "pull_request:44056|pull_request:44074", + "pull_request:43997|pull_request:44010", + "pull_request:44010|pull_request:44129" + ], + "prepared_review_unit_hash": "56a66d818554db9853016539f1807ddb9ccb678b23d44998b340f68b032588b9", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 78, + "serialized_chars": 84898, + "estimated_input_tokens": 21225, + "estimated_eval_tokens": 42706 + }, + "budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 6, + "serialized_chars": 17126, + "estimated_input_tokens": 4282, + "estimated_eval_tokens": 8820 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "A cluster of model-specific PRs migrating different Transformers models to the standardized output-capture/tracing helpers. They share a common refactor theme, but each targets a different model code path, so they are not duplicates of one another.", + "confidence": 0.88, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44073 is the clearest representative: it explicitly updates VisualBert to the new capture_outputs/can_return_tuple pattern and its title matches the cluster theme well.", + "best_issue_reason": null, + "best_pr_reason": "PR 44073 is the best cluster representative because it most directly reflects the shared standardized output-tracing migration, with a clear model-specific implementation and concise scope.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44024", + "right": "pull_request:44076", + "accept": false, + "reason": "Both are output-tracing refactors, but FocalNet and ImageGPT are different model implementations with different hook targets and forward paths; this is not the same underlying change." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44074", + "accept": false, + "reason": "VisualBert and TextNet are separate model migrations. The refactor pattern is similar, but they touch different code paths and are not mergeable as one duplicate PR." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44073", + "accept": false, + "reason": "MPNet and VisualBert each add capture_outputs/can_return_tuple plumbing in different model files; same broad theme, but not the same concrete bug or change." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44074", + "accept": false, + "reason": "These are distinct model-specific refactors in MPNet and TextNet. Similar tooling, but different forward implementations and output plumbing." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44010", + "accept": false, + "reason": "RegNet and SqueezeBert are unrelated model code paths. The diffs are both output-tracing related, but they are independent migrations, not duplicates." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44129", + "accept": false, + "reason": "SqueezeBert and SpeechT5 both adjust output capture, but SpeechT5 has a different explicit issue target and a different attention/cache signature; these are separate changes." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 15, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "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:44028", + "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": "aef848aede654c5267e382442b25f64735280b444087b1e473ebc57aa1d4c80c", + "component_budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 40, + "serialized_chars": 40171, + "estimated_input_tokens": 10043, + "estimated_eval_tokens": 20342 + }, + "budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 40, + "serialized_chars": 40171, + "estimated_input_tokens": 10043, + "estimated_eval_tokens": 20342 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is a set of similar-looking but separate model-specific output-tracing refactors. They share the same tooling pattern, but not the same underlying bug or code path, so I would not collapse them as duplicates.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44025 is a good canonical representative because it is a cohesive multi-file refactor in one related model family and clearly shows the output-tracing migration pattern.", + "best_issue_reason": null, + "best_pr_reason": "PR 44025 is the best single PR anchor for the cluster: it is self-contained, touches closely related variants, and most clearly exemplifies the common refactor theme without being overly specialized.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43998", + "right": "pull_request:44002", + "accept": false, + "reason": "Different model families and different forward/output paths; same refactor theme, not the same change." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44000", + "accept": false, + "reason": "Both refactor output handling, but they target unrelated models and code paths." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44001", + "accept": false, + "reason": "Separate model implementations with no shared bug or mergeable code-path change." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44003", + "accept": false, + "reason": "Different subsystems/models; one is timm_backbone, the other is mamba/falcon_mamba." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44029", + "accept": false, + "reason": "Unrelated model code paths despite similar output-tracing refactor intent." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44001", + "accept": false, + "reason": "Mobilenet_v1 and univnet are distinct models with separate output APIs." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44002", + "accept": false, + "reason": "Different models and behavior; only the refactor pattern matches." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44025", + "accept": false, + "reason": "Different model families and different tracing mechanics; not one underlying fix." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44026", + "accept": false, + "reason": "No shared concrete bug; these are separate encoder/decoder vs vision model refactors." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44027", + "accept": false, + "reason": "Distinct model families with unrelated forward implementations." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44001", + "accept": false, + "reason": "Similar API cleanup, but they affect different model code paths and outputs." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44002", + "accept": false, + "reason": "Different models; same broad refactor theme is not enough to deduplicate." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44003", + "accept": false, + "reason": "These touch unrelated model stacks and would not plausibly merge into one PR." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44025", + "accept": false, + "reason": "Different model families and different output-tracing implementations." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44002", + "accept": false, + "reason": "Separate model-specific refactors, not the same underlying change." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44003", + "accept": false, + "reason": "Mamba/falcon_mamba is unrelated to univnet beyond the shared refactor motif." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44025", + "accept": false, + "reason": "Different models and no evidence of a shared defect or shared code path." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44029", + "accept": false, + "reason": "UperNet and RWKV are unrelated implementations; same theme only." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44003", + "accept": false, + "reason": "Distinct model stacks and distinct output-capture changes." + }, + { + "left": "pull_request:44003", + "right": "pull_request:44028", + "accept": false, + "reason": "Mamba/falcon_mamba and SuperPoint are unrelated; no shared bug or merged PR candidate." + }, + { + "left": "pull_request:44003", + "right": "pull_request:44004", + "accept": false, + "reason": "These are different code paths and different levels of API refactoring." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44004", + "accept": false, + "reason": "timm_backbone and codegen are unrelated models; only the naming pattern is similar." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44004", + "accept": false, + "reason": "UnivNet and CodeGen do not share a concrete code-path problem." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44004", + "accept": false, + "reason": "Different model architectures and output contracts." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44004", + "accept": false, + "reason": "Vision-text dual encoder vs CodeGen are unrelated changes." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44026", + "accept": false, + "reason": "Different model families; same refactor style but not the same fix." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44027", + "accept": false, + "reason": "Depth Anything and Speech Encoder-Decoder are unrelated model paths." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44028", + "accept": false, + "reason": "Different models and different output-recording mechanics; not duplicate PRs." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44029", + "accept": false, + "reason": "Depth Anything and RWKV are separate model stacks with no shared bug." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44030", + "accept": false, + "reason": "Different models and different concrete forward methods." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44029", + "accept": false, + "reason": "Vision encoder-decoder and RWKV are unrelated; same pattern only." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44027", + "accept": false, + "reason": "Vision encoder-decoder and speech encoder-decoder are separate code paths with distinct implementations." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44028", + "accept": false, + "reason": "Different models; not the same underlying bug or fix." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44030", + "accept": false, + "reason": "Vision encoder-decoder and DPR do not share a concrete code-path problem." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44028", + "accept": false, + "reason": "Speech encoder-decoder and SuperPoint are unrelated model changes." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44029", + "accept": false, + "reason": "Different architectures and output APIs; not duplicate changes." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44030", + "accept": false, + "reason": "Speech encoder-decoder and DPR are separate model implementations." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44029", + "accept": false, + "reason": "SuperPoint and RWKV are unrelated code paths." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44030", + "accept": false, + "reason": "SuperPoint and DPR are different models with different output contracts." + }, + { + "left": "pull_request:44029", + "right": "pull_request:44030", + "accept": false, + "reason": "RWKV and DPR are different model families; no shared concrete fix." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 16, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44662-3", + "nodes": [ + "pull_request:44662", + "pull_request:45075", + "pull_request:45401" + ], + "soft_pairs": [ + "pull_request:44662|pull_request:45401", + "pull_request:44662|pull_request:45075", + "pull_request:45075|pull_request:45401" + ], + "prepared_review_unit_hash": "320ad4651be130179970fde0ae4b07844d8d151fc1a409226a04ab19e85192e4", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 10912, + "estimated_input_tokens": 2728, + "estimated_eval_tokens": 5712 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 10912, + "estimated_input_tokens": 2728, + "estimated_eval_tokens": 5712 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All three pull requests are separate model-addition implementations (PenguinVL, DeepSeek-OCR-2, and Voxtral TTS). They share Transformers scaffolding files and docs registration, but they do not fix the same concrete model or code-path issue, so none should be deduplicated together.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "No PR is a good canonical representative for the cluster because they target different models. If one must be singled out, 45075 is the most mature standalone implementation, but it should not absorb the others.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44662", + "right": "pull_request:45401", + "accept": false, + "reason": "Different model families and capabilities: PenguinVL vs Voxtral TTS. Shared auto-registration/docs files are generic model-add boilerplate, not the same underlying change." + }, + { + "left": "pull_request:44662", + "right": "pull_request:45075", + "accept": false, + "reason": "Different model additions: PenguinVL vs DeepSeek-OCR-2. Overlap is limited to common integration scaffolding, not a shared bug or fix." + }, + { + "left": "pull_request:45075", + "right": "pull_request:45401", + "accept": false, + "reason": "Different concrete models and tasks: DeepSeek-OCR-2 vs Voxtral TTS. Similar file patterns are expected for new model support and do not indicate 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": 17, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44979-2", + "nodes": [ + "pull_request:44979", + "pull_request:45363" + ], + "soft_pairs": [ + "pull_request:44979|pull_request:45363" + ], + "prepared_review_unit_hash": "e3beb93d5b4e2de35e77549f1b63049d82decbb041cefb92bc2eb9bddf43353d", + "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 by the broader theme of fusion, but they implement different changes: one introduces a general module fusion API, while the other adds n-to-1 kernel fusion support via KernelConfig and related integration plumbing. They are not duplicates and should stay separate.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44979 is the more foundational change: it adds the new Module Fusion API and test coverage, which is a core library feature rather than a usage example or a narrower integration path.", + "best_issue_reason": null, + "best_pr_reason": "PR 44979 is the best representative for the cluster because it defines the core abstraction, whereas PR 45363 is a separate kernel-fusion implementation built on adjacent infrastructure.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44979", + "right": "pull_request:45363", + "accept": false, + "reason": "Different concrete changes and code paths: 44979 adds a general module fusion API, while 45363 adds KernelConfig-based n-to-1 kernel fusion and example/integration updates. They are related conceptually but not the same underlying fix or mergeable into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 18, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45086-2", + "nodes": [ + "pull_request:45086", + "pull_request:45317" + ], + "soft_pairs": [ + "pull_request:45086|pull_request:45317" + ], + "prepared_review_unit_hash": "9b2a5bb108debe3bd590002048ab5183d2d89db1c6810d2fb3368f4b4c1f4954", + "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 target the same Mistral regex patch bug in the same file with the same code change; 45317 is the more complete version because it adds tests and clearer context.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "45317 is the stronger canonical PR because it covers the same fix as 45086 and adds test coverage, making it the more complete representative of the change.", + "best_issue_reason": null, + "best_pr_reason": "45317 is the best PR for the cluster: same concrete bug fix, same code path, and it includes validation/tests rather than only the minimal code edit.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45086", + "right": "pull_request:45317", + "accept": true, + "reason": "Accept: both PRs fix the same AttributeError in _patch_mistral_regex by changing the same tokenizer pre-tokenizer accessors; 45317 is essentially a superset with tests, so they could 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": 19, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45115-2", + "nodes": [ + "pull_request:45115", + "pull_request:45569" + ], + "soft_pairs": [ + "pull_request:45115|pull_request:45569" + ], + "prepared_review_unit_hash": "3041fe4fb12a25a8834292750ce2931ab6aece11078260b5540af3e99aeb76b4", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6066, + "estimated_input_tokens": 1517, + "estimated_eval_tokens": 3290 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6066, + "estimated_input_tokens": 1517, + "estimated_eval_tokens": 3290 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related to NemotronH work, but they are not the same PR: one is a conversion/refactor to GraniteMoeHybrid-compatible mappings, while the other is a broader model split/addition for NemotronH dense/sparse variants and docs. The overlap is mostly shared files in the NemotronH package, not the same concrete code change.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45569 is the better canonical representative for this cluster because it is the broader, more user-facing model-structure change introducing the dedicated NemotronH dense/sparse variants and associated docs/tests, whereas #45115 is a narrower conversion-mapping/refactor PR.", + "best_issue_reason": null, + "best_pr_reason": "PR #45569 is the strongest representative PR overall: it covers the main architecture/package rework and is more complete than the smaller conversion-mapping-only PR.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45115", + "right": "pull_request:45569", + "accept": false, + "reason": "Reject: both touch NemotronH code, but 45115 is about conversion mapping and inheritance/refactoring, while 45569 adds dedicated NemotronH dense/sparse model variants, docs, and tests. They are adjacent work, not the same underlying change and would not naturally 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": 20, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45332-2", + "nodes": [ + "pull_request:45332", + "pull_request:45333" + ], + "soft_pairs": [ + "pull_request:45332|pull_request:45333" + ], + "prepared_review_unit_hash": "b2673f3349e5b14a6009543295a38a98f11e53a6e0dbbeb995a939619b41c377", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5398, + "estimated_input_tokens": 1350, + "estimated_eval_tokens": 2956 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5398, + "estimated_input_tokens": 1350, + "estimated_eval_tokens": 2956 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related heterogeneous-config workstreams, but they are not duplicate PRs: one adds per-layer configuration plumbing, while the other adds broader heterogeneous model/runtime support on top of it.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "No duplicate canonical PR: PR 45332 is the broader heterogeneous model/runtime implementation, while PR 45333 is the narrower config-plumbing foundation. They are related but not the same change.", + "best_issue_reason": null, + "best_pr_reason": "PR 45332 is the best representative of the cluster because it covers the larger end-to-end heterogeneous support surface and subsumes the config-only groundwork in 45333.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45332", + "right": "pull_request:45333", + "accept": false, + "reason": "Reject: 45333 is only per-layer config support, whereas 45332 adds heterogeneous modeling/cache/runtime behavior. They share infrastructure files, but they are different concrete changes and not the same PR-level bug fix or feature." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 21, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45415-2", + "nodes": [ + "pull_request:45415", + "pull_request:45425" + ], + "soft_pairs": [ + "pull_request:45415|pull_request:45425" + ], + "prepared_review_unit_hash": "468747db5fb4157075d83107b7a4b3e00bfaf65a45921168c7d6ba6122864b41", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5513, + "estimated_input_tokens": 1379, + "estimated_eval_tokens": 3014 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5513, + "estimated_input_tokens": 1379, + "estimated_eval_tokens": 3014 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs are typing-related, but they are not duplicates: #45415 is a broad repo-wide type-checking sweep, while #45425 adds focused typing protocols/helpers for specific modules. They overlap in a few files, but not in the same concrete change.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45415 is the broadest representative of the cluster, with a repo-wide type-checking pass and pyproject typing-ignore policy changes across many files.", + "best_issue_reason": null, + "best_pr_reason": "PR #45415 is the strongest cluster anchor because it captures the main typing-enable effort at a larger, more central scope than the narrower follow-up in #45425.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45415", + "right": "pull_request:45425", + "accept": false, + "reason": "Related typing work, but not the same underlying change: #45415 is a wide mechanical type-checking sweep, while #45425 introduces specific typing aliases/protocols for modeling_utils and related code paths. They would not plausibly merge 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": 22, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45548-2", + "nodes": [ + "pull_request:45548", + "pull_request:45662" + ], + "soft_pairs": [ + "pull_request:45548|pull_request:45662" + ], + "prepared_review_unit_hash": "f34ddd3191d7791e835bc636a416ecd9b251226be8fa7b19bbbd813bd260b810", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4508, + "estimated_input_tokens": 1127, + "estimated_eval_tokens": 2510 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4508, + "estimated_input_tokens": 1127, + "estimated_eval_tokens": 2510 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two EP-related pull requests, but they fix different bugs in different loading paths: one is about DeepSpeed ZeRO-3 / accelerate launch serialization-loading behavior, the other about FSDP2 rank-0 broadcast overwriting experts. They are related by theme, not the same underlying change.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical duplicate PR here; the two patches address distinct code-path failures. If forced to pick an anchor, 45662 is the more concrete bug fix in core loading behavior.", + "best_issue_reason": null, + "best_pr_reason": "45662 is the stronger representative of the cluster only because it has the more explicit, self-contained failure mode and broader implementation changes; 45548 targets a separate loading/serialization issue.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45548", + "right": "pull_request:45662", + "accept": false, + "reason": "Both are EP-related, but 45548 fixes distributed_config serialization / has_ep handling for DeepSpeed ZeRO-3 loading, while 45662 fixes experts being overwritten by rank-0 broadcast in a different FSDP2 loading path. Different underlying bugs, not mergeable as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 23, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45639-2", + "nodes": [ + "pull_request:45639", + "pull_request:45645" + ], + "soft_pairs": [ + "pull_request:45639|pull_request:45645" + ], + "prepared_review_unit_hash": "b09064fae7efcc344d454e480eaeebdbe7a4162126d35105b319fa1d8ab8e48d", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4589, + "estimated_input_tokens": 1148, + "estimated_eval_tokens": 2552 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4589, + "estimated_input_tokens": 1148, + "estimated_eval_tokens": 2552 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs address the same xdist-related captured_info collision problem tied to issue #45561; #45645 is a broader superset of #45639\u2019s testing_utils fix and adds CI/log handling for the same artifact path.", + "confidence": 0.84, + "canonical_issue_reason": "No issue artifacts were provided, so there is no issue canonical to choose here.", + "canonical_pr_reason": "PR #45645 is the better canonical PR because it subsumes the core testing_utils fix from #45639 and extends the same change to CI/workflow artifact handling and notification paths.", + "best_issue_reason": "No issue nodes were supplied, so there is no best issue to rank.", + "best_pr_reason": "PR #45645 is the best PR because it appears to be the more complete end-to-end fix for the captured_info/xdist collision problem, including the workflow changes needed to consume multiple captured_info files.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45639", + "right": "pull_request:45645", + "accept": true, + "reason": "Same underlying bug: xdist-safe handling of captured_info artifacts. #45645 contains the core fix from #45639 and adds related CI/log collection changes, so these look like duplicate/overlapping implementations of the same change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 24, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45683-2", + "nodes": [ + "pull_request:45683", + "pull_request:45734" + ], + "soft_pairs": [ + "pull_request:45683|pull_request:45734" + ], + "prepared_review_unit_hash": "eba4ca0f26b2d3be0c00e316270a9341d31af2208b44dbab31b1d3a8e81e7538", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4469, + "estimated_input_tokens": 1118, + "estimated_eval_tokens": 2492 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4469, + "estimated_input_tokens": 1118, + "estimated_eval_tokens": 2492 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two open PRs touch the same quantization helper, but they fix different bugs: one special-cases audio submodules, the other changes how default skip lists are combined with user-supplied skips. They should not be clustered as duplicates.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "45734 is the better representative PR because it fixes the core skip-module composition logic in `get_modules_to_not_convert`, which is broader and more central than the narrow audio-module special case in 45683.", + "best_issue_reason": null, + "best_pr_reason": "45734 is the strongest standalone PR in the set: it addresses a concrete quantization failure mode in the main code path and its fix is directly about the helper\u2019s behavior rather than a model-specific workaround.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45683", + "right": "pull_request:45734", + "accept": false, + "reason": "Both touch quantization skip logic, but 45683 adds audio-tower/embed_audio exclusions while 45734 changes additive handling of `skip_modules`; these are different underlying bugs and would not naturally merge into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 25, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45737-2", + "nodes": [ + "pull_request:45737", + "pull_request:45738" + ], + "soft_pairs": [ + "pull_request:45737|pull_request:45738" + ], + "prepared_review_unit_hash": "cc2ade73a49ec6a8883a7a24c433988645588bcd049e8aad2bce56b1b24daa1c", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4131, + "estimated_input_tokens": 1033, + "estimated_eval_tokens": 2322 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4131, + "estimated_input_tokens": 1033, + "estimated_eval_tokens": 2322 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs address the same MusicgenMelody cache initialization bug in `forward()`: replacing `EncoderDecoderCache(DynamicCache, DynamicCache)` with a single `DynamicCache`. PR #45737 is the more complete duplicate because it bundles the fix with a regression test, while #45738 is the fix-only variant.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45737 is the best canonical PR because it contains the same code fix as #45738 plus a regression test, making it the most complete representation of the bug fix.", + "best_issue_reason": null, + "best_pr_reason": "PR #45737 is the strongest overall candidate: it fixes the concrete cache bug and adds a targeted regression test, whereas #45738 only changes the implementation.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45737", + "right": "pull_request:45738", + "accept": true, + "reason": "Same underlying code-path bug in `musicgen_melody` cache setup; the diffs match on the core fix and the test-only addition in #45737 could plausibly live in the same PR as #45738." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The assessment is grounded in the packet: both PRs make the same `musicgen_melody` cache initialization change, and #45737 only adds a regression test on top of that fix. The soft-edge accept is conservative enough because the diffs are mergeable into one PR for the same bugfix." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + } + ] +} diff --git a/snapshots/20260518T120024Z/analysis-runs/hybrid-model-20260518t120024z/manifest.json b/snapshots/20260518T120024Z/analysis-runs/hybrid-model-20260518t120024z/manifest.json new file mode 100644 index 0000000000000000000000000000000000000000..e8cf7b0d7de6a1e80c98d2742b02a4e9b3624654 --- /dev/null +++ b/snapshots/20260518T120024Z/analysis-runs/hybrid-model-20260518t120024z/manifest.json @@ -0,0 +1,14 @@ +{ + "analysis_id": "hybrid-model-20260518t120024z", + "artifacts": { + "hybrid": "snapshots/20260518T120024Z/analysis-runs/hybrid-model-20260518t120024z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260518T120024Z/analysis-runs/hybrid-model-20260518t120024z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "model": null, + "published_at": "2026-05-18T12:15:51Z", + "repo": "huggingface/transformers", + "schema_version": 1, + "snapshot_id": "20260518T120024Z", + "variant": "hybrid" +} diff --git a/snapshots/20260518T120024Z/manifest.json b/snapshots/20260518T120024Z/manifest.json index 7971b8ba3ccc8a6650899320ea5f425cfbd9b56d..422ccbabbae15a8af869d29e6f018037b6fcc8bc 100644 --- a/snapshots/20260518T120024Z/manifest.json +++ b/snapshots/20260518T120024Z/manifest.json @@ -32,6 +32,24 @@ "timeline_events": 1560 }, "extracted_at": "2026-05-18T12:00:24Z", + "published_analysis": { + "canonical_analysis_id": "hybrid-model-20260518t120024z", + "runs": { + "hybrid-model-20260518t120024z": { + "analysis_id": "hybrid-model-20260518t120024z", + "artifacts": { + "hybrid": "snapshots/20260518T120024Z/analysis-runs/hybrid-model-20260518t120024z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260518T120024Z/analysis-runs/hybrid-model-20260518t120024z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "manifest_path": "snapshots/20260518T120024Z/analysis-runs/hybrid-model-20260518t120024z/manifest.json", + "model": null, + "published_at": "2026-05-18T12:15:51Z", + "variant": "hybrid" + } + }, + "schema_version": 1 + }, "repo": "huggingface/transformers", "snapshot_id": "20260518T120024Z", "watermark": {