diff --git a/analysis/current/analysis-report-hybrid.json b/analysis/current/analysis-report-hybrid.json index 5b16a5c968900a04dbe33b441d7d014431b7c817..c049595e0ac61e0bf1dc9512b54f8ce83e85490d 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": "20260425T000023Z", - "generated_at": "2026-04-25T00:03:41Z", + "snapshot_id": "20260425T060019Z", + "generated_at": "2026-04-25T06:10:11Z", "evidence_quality": "full", "llm_enrichment": true, "meta_bugs": [ @@ -1150,152 +1150,73 @@ ] }, { - "cluster_id": "cluster-43656-4", - "summary": "Cluster of 1 issues and 3 PRs centered on issue #43824.", + "cluster_id": "cluster-43240-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #43240.", "status": "open", - "confidence": 0.8, - "canonical_issue_number": 43824, - "canonical_pr_number": 43656, + "confidence": 0.75, + "canonical_issue_number": 43240, + "canonical_pr_number": 43251, "issue_numbers": [ - 43824 + 43240 ], "pr_numbers": [ - 43656, - 43836, - 43842 + 43251, + 43254 ], "evidence_types": [ "closing_reference", - "shared_issue_target", - "soft_similarity" + "shared_issue_target" ], "pr_comparisons": [ { - "left_pr_number": 43656, - "right_pr_number": 43836, - "code_similarity": 0.36, - "size_similarity": 0.176, + "left_pr_number": 43251, + "right_pr_number": 43254, + "code_similarity": 0.794, + "size_similarity": 0.64, "file_overlap": 1.0, - "area_overlap": 0.051, - "patch_similarity": 0.048, + "area_overlap": 0.667, + "patch_similarity": 0.882, "shared_filenames": [ - "src/transformers/cli/serve.py" + "src/transformers/loss/loss_utils.py" ], "shared_file_areas": [ { - "filename": "src/transformers/cli/serve.py", + "filename": "src/transformers/loss/loss_utils.py", "left_ranges": [ - [ - 11, - 18 - ], [ 30, - 36 - ], - [ - 315, - 323 - ], - [ - 665, - 671 - ], - [ - 931, - 937 - ], - [ - 1843, - 1849 - ], - [ - 1868, - 1874 + 50 ] ], "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 + 43 ] ] } ] - }, + } + ] + }, + { + "cluster_id": "cluster-43824-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #43824.", + "status": "open", + "confidence": 0.55, + "canonical_issue_number": 43824, + "canonical_pr_number": 43836, + "issue_numbers": [ + 43824 + ], + "pr_numbers": [ + 43836, + 43842 + ], + "evidence_types": [ + "closing_reference" + ], + "pr_comparisons": [ { "left_pr_number": 43836, "right_pr_number": 43842, @@ -1347,56 +1268,6 @@ } ] }, - { - "cluster_id": "cluster-43240-3", - "summary": "Cluster of 1 issues and 2 PRs centered on issue #43240.", - "status": "open", - "confidence": 0.75, - "canonical_issue_number": 43240, - "canonical_pr_number": 43251, - "issue_numbers": [ - 43240 - ], - "pr_numbers": [ - 43251, - 43254 - ], - "evidence_types": [ - "closing_reference", - "shared_issue_target" - ], - "pr_comparisons": [ - { - "left_pr_number": 43251, - "right_pr_number": 43254, - "code_similarity": 0.794, - "size_similarity": 0.64, - "file_overlap": 1.0, - "area_overlap": 0.667, - "patch_similarity": 0.882, - "shared_filenames": [ - "src/transformers/loss/loss_utils.py" - ], - "shared_file_areas": [ - { - "filename": "src/transformers/loss/loss_utils.py", - "left_ranges": [ - [ - 30, - 50 - ] - ], - "right_ranges": [ - [ - 30, - 43 - ] - ] - } - ] - } - ] - }, { "cluster_id": "cluster-45081-3", "summary": "Cluster of 1 issues and 2 PRs centered on issue #45081.", @@ -1810,16 +1681,6 @@ "target_issue_number": 43240, "reason": "PRs in cluster-43240-3 are treated as duplicates because they converge on issue #43240 with closing_reference, shared_issue_target evidence." }, - { - "cluster_id": "cluster-43656-4", - "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, @@ -1829,6 +1690,15 @@ "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-43824-3", + "canonical_pr_number": 43836, + "duplicate_pr_numbers": [ + 43842 + ], + "target_issue_number": 43824, + "reason": "PRs in cluster-43824-3 are treated as duplicates because they converge on issue #43824 with closing_reference evidence." + }, { "cluster_id": "cluster-43979-11", "canonical_pr_number": 44007, diff --git a/analysis/current/analysis-report-hybrid.llm-reviews.json b/analysis/current/analysis-report-hybrid.llm-reviews.json index 978e52d90c57a0bcd551c279f09f21847d601f3e..715dc1a474709dfd6884f6651904c566f7b1f83c 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": "20260425T000023Z", - "generated_at": "2026-04-25T00:03:41Z", + "snapshot_id": "20260425T060019Z", + "generated_at": "2026-04-25T06:10:11Z", "model": "gpt-5.4-mini?service_tier=flex", "reviews": [ { @@ -13,8 +13,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-13244-3", "nodes": [ @@ -26,12 +26,12 @@ "issue:13244|issue:16618", "issue:15980|issue:16618" ], - "prepared_review_unit_hash": "3bd1aa1e5016ec9382b72cbe7c74a64e990c7c4e0b250800fedbd71ab44829a9", + "prepared_review_unit_hash": "9c0637be12a32d10044e3b0909db2a845cb6aa29108062979602cbe64c140d1f", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 2479, + "serialized_chars": 2477, "estimated_input_tokens": 620, "estimated_eval_tokens": 1496 }, @@ -39,7 +39,7 @@ "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 2479, + "serialized_chars": 2477, "estimated_input_tokens": 620, "estimated_eval_tokens": 1496 }, @@ -48,24 +48,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These issues do not appear to be duplicates of one another: one is about TAPAS tokenization, one about auth-related private model downloads, and one about a model-loading failure for bert-base-uncased. The only somewhat related pair is the last two, but they describe different failure modes and should not be merged.", - "confidence": 0.9, - "canonical_issue_reason": "Issue 16618 is the most generally framed model-loading report, so it is the best single representative of the cluster if one must be chosen.", + "summary": "These three issues are not duplicates of one another: one is about Tapas tokenization mismatch, one about a bad auth/error message for private models, and one about failing to load bert-base-uncased. The cluster looks spurious rather than a single underlying bug.", + "confidence": 0.98, + "canonical_issue_reason": "No single canonical issue fits here; the reports describe unrelated problems.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 16618 is the strongest standalone issue for triage because it is broad, clearly user-facing, and has the most discussion activity among the set.", + "best_issue_reason": "Issue 16618 is the broadest and most active report in the set, but it still does not represent the other two bugs.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:13244", "right": "issue:16618", "accept": false, - "reason": "Different topics and code paths: TAPAS tokenization vs generic model loading for bert-base-uncased." + "reason": "Tapas tokenization mismatch vs. model loading failure for bert-base-uncased are different code-path problems." }, { "left": "issue:15980", "right": "issue:16618", "accept": false, - "reason": "Both involve model loading, but one is specifically an unauthenticated private-model download error while the other is a generic inability to load bert-base-uncased; not the same underlying bug." + "reason": "Private-model login/error-message handling is unrelated to the general 'can't load bert-base-uncased' report." } ] }, @@ -84,8 +84,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -154,8 +154,8 @@ "issue:44805|issue:45161", "issue:44805|issue:44918", "issue:44661|issue:44805", - "issue:34634|issue:35545", "issue:30277|issue:43638", + "issue:34634|issue:35545", "issue:43596|issue:43638", "issue:43975|issue:44457", "issue:43039|issue:45161", @@ -203,370 +203,358 @@ "issue:41093|issue:41762", "issue:43296|issue:43366" ], - "prepared_review_unit_hash": "b09f576435fcf2d9c74b2cdf4b8a4c6a872d3e820b00fad10d5df17d08c66eec", + "prepared_review_unit_hash": "0803d07dfe8d664e5b8b7f87a0c8722b9cd99c34a2bcdf649913c3fa521b2477", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 63, - "serialized_chars": 46475, - "estimated_input_tokens": 11619, - "estimated_eval_tokens": 23494 + "serialized_chars": 46481, + "estimated_input_tokens": 11621, + "estimated_eval_tokens": 23498 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most items are unrelated, with one clear duplicate cluster around DeepSpeed ZeRO-3 / index-out-of-bounds loading failures for BERT/Gemma-style models. A smaller docs-only similarity cluster exists, but the evidence is weaker and broader than the ZeRO-3 bug cluster.", - "confidence": 0.72, - "canonical_issue_reason": "issue:44805 is the best representative of the ZeRO-3 loading failure cluster because its title is the most general form of the shared index-out-of-bounds/mask-shape error and it maps cleanly to the other Bert/ZeRO-3 variants.", + "summary": "This cluster is mostly heterogeneous, but there is a small clear duplicate family around DeepSpeed ZeRO-3 loading failures that produce \"index 0 is out of bounds for dimension 0 with size 0\". Most other soft pairs are only loosely similar and should stay separate.", + "confidence": 0.76, + "canonical_issue_reason": "Issue 43596 is the clearest representative of the ZeRO-3 index-out-of-bounds loading bug and matches the strongest near-duplicate reports in this set.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44805 is the strongest canonical issue candidate among the provided items; it is the broadest and most reusable description of the shared underlying bug.", + "best_issue_reason": "43596 is the most central issue to anchor the duplicate family because it names the concrete DeepSpeed ZeRO-3 failure mode without being overly model-specific.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44322", "right": "issue:45464", "accept": false, - "reason": "Different Qwen3.5 failures and different code paths; same model family only, not the same bug." + "reason": "Both are Qwen3.5-related, but they describe different failures in different entrypoints and are not the same concrete bug." }, { "left": "issue:43716", "right": "issue:45237", "accept": false, - "reason": "Different failure modes: dtype mismatch vs AMD GPU execution failure." + "reason": "One is a dtype mismatch in image preprocessing, the other is a GPU/AMD compatibility problem; too different." }, { "left": "issue:43828", "right": "issue:45237", "accept": false, - "reason": "Autocast dtype mismatch is unrelated to AMD GPU support issues." + "reason": "Autocast dtype mismatch is not the same underlying issue as model failing on AMD GPUs." }, { "left": "issue:45237", "right": "issue:45507", "accept": false, - "reason": "Different problem areas; one is GPU backend support, the other is an invalid method call in a model implementation." + "reason": "Different model families and different failure modes; no clear shared code-path bug." }, { "left": "issue:44292", "right": "issue:45237", "accept": false, - "reason": "Different model/runtime failures with no clear shared code-path." + "reason": "Both mention Qwen-3 variants, but the reported errors are unrelated." }, { "left": "issue:30277", "right": "issue:35545", "accept": false, - "reason": "DeepSpeed collective mismatch and ONNX export error are unrelated." + "reason": "DeepSpeed collective mismatch and ONNX export failure are unrelated." }, { "left": "issue:35545", "right": "issue:42915", "accept": false, - "reason": "Different model/export paths and different symptoms." + "reason": "Different models and different export/runtime problems; not the same bug." }, { "left": "issue:35545", "right": "issue:41720", "accept": false, - "reason": "ONNX export error and auto device-mapping cuda assert are not the same bug." + "reason": "ONNX export error and CUDA assert under auto device mapping are different issues." }, { "left": "issue:44509", "right": "issue:44512", "accept": false, - "reason": "Both are docs regressions in v5, but they reference different removed features and do not clearly share one concrete change." + "reason": "Both are docs regressions, but they cover different removed commands/tasks and do not look like one change." }, { "left": "issue:35545", "right": "issue:36010", "accept": false, - "reason": "ONNX export and GenerationMixin import breakage are unrelated." + "reason": "Export to ONNX and missing GenerationMixin import are unrelated." }, { "left": "issue:43638", "right": "issue:44805", - "accept": true, - "reason": "Same ZeRO-3 index-0-out-of-bounds failure pattern on Bert-family loading." + "accept": false, + "reason": "Both are index/shape errors, but the titles point to different concrete failure patterns." }, { "left": "issue:43596", "right": "issue:44805", - "accept": true, - "reason": "Same underlying ZeRO-3 loading/index-out-of-bounds bug, just on a different Bert variant." + "accept": false, + "reason": "Zero-3 init index-out-of-bounds and mask-shape mismatch are not the same bug." }, { "left": "issue:44805", "right": "issue:45161", "accept": false, - "reason": "One is a ZeRO-3 loading bug; the other is a GPT-OSS tensor-parallel issue." + "reason": "Different error signatures and different subsystems." }, { - "left": "issue:44661", - "right": "issue:44805", + "left": "issue:44805", + "right": "issue:44918", "accept": false, - "reason": "Tokenizer mapping generation bug is unrelated to ZeRO-3 loading errors." + "reason": "These are separate training/runtime issues with different error types." }, { - "left": "issue:34634", - "right": "issue:35545", + "left": "issue:44661", + "right": "issue:44805", "accept": false, - "reason": "BarkProcessor preset handling and ONNX export are unrelated." + "reason": "Tokenizer mapping logic and mask/tensor shape mismatch are unrelated." }, { "left": "issue:30277", "right": "issue:43638", "accept": false, - "reason": "Both are DeepSpeed-related, but the symptoms and code paths differ materially." + "reason": "Both involve DeepSpeed, but one is a collective mismatch and the other is a size-0 index error." + }, + { + "left": "issue:34634", + "right": "issue:35545", + "accept": false, + "reason": "Processor voice_preset and ONNX export problems are unrelated." }, { "left": "issue:43596", "right": "issue:43638", "accept": true, - "reason": "Near-identical ZeRO-3 index-out-of-bounds failures on BertModel loading." + "reason": "These describe the same ZeRO-3 index-0-out-of-bounds loading failure with only model-specific wording differences." }, { "left": "issue:43975", "right": "issue:44457", "accept": false, - "reason": "Different model-specific inference/detokenization bugs." + "reason": "Different language/model-specific inference behavior; no evidence they are the same bug." }, { "left": "issue:43039", "right": "issue:45161", "accept": false, - "reason": "Liger Kernel cross-entropy call and GPT-OSS tensor-parallel failure are different issues." + "reason": "Liger Kernel cross-entropy invocation and GPT-OSS tensor-parallel failure are different problems." }, { "left": "issue:44292", "right": "issue:44457", "accept": false, - "reason": "Different model/runtime problems with no shared bug." + "reason": "Different Qwen3.5/LoRA-related reports with different symptoms." }, { "left": "issue:34928", "right": "issue:44805", "accept": false, - "reason": "Activation checkpointing/FSDP mismatch is unrelated to ZeRO-3 index-out-of-bounds loading." + "reason": "Activation-checkpointing/FSDP recomputation mismatch is distinct from the mask index error." }, { "left": "issue:24643", "right": "issue:30277", "accept": false, - "reason": "Different DeepSpeed errors and different failure conditions." + "reason": "Both mention DeepSpeed, but the training weight-dimension error and collective mismatch are different bugs." }, { "left": "issue:43975", "right": "issue:45237", "accept": false, - "reason": "Different model families and different runtime failures." + "reason": "Unrelated model detokenization vs AMD GPU runtime issue." }, { "left": "issue:43638", "right": "issue:45161", "accept": false, - "reason": "Bert ZeRO-3 loading error is unrelated to GPT-OSS tensor parallelism." + "reason": "Zero-3/BERT load failure is unrelated to GPT-OSS tensor-parallel behavior." }, { "left": "issue:34928", "right": "issue:36331", "accept": false, - "reason": "Activation checkpointing/FSDP tensor mismatch is unrelated to the custom trainer loss signature issue." + "reason": "FSDP checkpointing mismatch and CustomTrainer signature change are unrelated." }, { "left": "issue:43596", "right": "issue:45161", "accept": false, - "reason": "Different model and different failure path." + "reason": "Different model-loading bug versus tensor-parallel runtime issue." }, { "left": "issue:43716", - "right": "issue:45507", + "right": "issue:44292", "accept": false, - "reason": "Dtype mismatch and invalid method invocation are unrelated." + "reason": "Different models and different error contexts; not the same underlying defect." }, { "left": "issue:39290", "right": "issue:45081", "accept": false, - "reason": "Different model/configuration issues with no shared concrete bug." + "reason": "Gemma3/vLLM attribute error and Mistral tokenizer patch crash are unrelated." }, { "left": "issue:43828", "right": "issue:45161", "accept": false, - "reason": "Autocast dtype mismatch is unrelated to GPT-OSS TP failure." + "reason": "Autocast dtype mismatch and GPT-OSS TP failure do not match." }, { - "left": "issue:43425", - "right": "issue:44292", + "left": "issue:30277", + "right": "issue:36331", "accept": false, - "reason": "Torch version incompatibility and NVFP4 model execution error are unrelated." + "reason": "Collective mismatch in DeepSpeed and trainer kwarg incompatibility are separate issues." }, { - "left": "issue:44918", - "right": "issue:45161", + "left": "issue:43039", + "right": "issue:44805", "accept": false, - "reason": "TRL embedding unpacking and GPT-OSS tensor-parallel failures are unrelated." + "reason": "Liger Kernel cross-entropy invocation is unrelated to mask/tensor indexing errors." }, { "left": "issue:34928", - "right": "issue:41720", + "right": "issue:36010", "accept": false, - "reason": "FSDP checkpointing error and Qwen3 auto-device-map cuda assert are different bugs." + "reason": "Activation checkpointing/FSDP issue does not match the import error." + }, + { + "left": "issue:41762", + "right": "issue:43596", + "accept": true, + "reason": "Both are the same DeepSpeed ZeRO-3 index-0-out-of-bounds failure appearing during model loading." }, { "left": "issue:43827", "right": "issue:44512", "accept": false, - "reason": "Both are v5 docs regressions, but they concern different removed items and are not clearly one change." + "reason": "Both are docs regressions, but they target different removed guidance and do not look like one duplicate." }, { "left": "issue:43716", "right": "issue:44560", "accept": false, - "reason": "Different model-specific dtype/video pipeline failures." + "reason": "Different model families and different symptom classes." }, { "left": "issue:43716", "right": "issue:44918", "accept": false, - "reason": "Different model/input handling failures." + "reason": "Image-preprocessor dtype mismatch and Qwen3.5 SFT embedding unpacking are unrelated." }, { "left": "issue:43039", "right": "issue:43828", "accept": false, - "reason": "Different model/runtime paths; one is kernel selection, the other is autocast mismatch." + "reason": "Both touch loss/dtype behavior, but they are not the same concrete bug." }, { "left": "issue:30277", "right": "issue:36331", "accept": false, - "reason": "DeepSpeed collective mismatch and trainer signature mismatch are unrelated." + "reason": "DeepSpeed rank collective mismatch and CustomTrainer API mismatch are unrelated." }, { "left": "issue:43039", - "right": "issue:44805", - "accept": false, - "reason": "Liger Kernel cross-entropy call does not match ZeRO-3 loading failures." - }, - { - "left": "issue:34928", - "right": "issue:36010", - "accept": false, - "reason": "Checkpointing/FSDP issue and GenerationMixin import breakage are unrelated." - }, - { - "left": "issue:41762", - "right": "issue:43638", - "accept": true, - "reason": "Same ZeRO-3 index-out-of-bounds loading bug, presented on different model names." - }, - { - "left": "issue:43901", - "right": "issue:44512", - "accept": false, - "reason": "Different stale-doc references and no clear single underlying change." - }, - { - "left": "issue:39290", - "right": "issue:41720", + "right": "issue:44560", "accept": false, - "reason": "Different Qwen/Gemma runtime issues with distinct failure signatures." + "reason": "Liger Kernel cross-entropy usage and Qwen3-vl embedding video StopIteration are different failures." }, { "left": "issue:43866", "right": "issue:44863", "accept": false, - "reason": "Checkpoint corruption and NemotronH loading failure are not the same bug." + "reason": "Corrupted checkpoint and checkpoint loading failure are too vague and not clearly the same artifact." }, { "left": "issue:43854", "right": "issue:43866", "accept": false, - "reason": "Unit-test load failure and checkpoint corruption are different problems." + "reason": "Unit test load failure and corrupted checkpoint are different problems." }, { "left": "issue:43425", "right": "issue:45070", "accept": false, - "reason": "Torch incompatibility and pydantic field breakage are unrelated." + "reason": "Torch compatibility and pydantic model field regression are unrelated." }, { "left": "issue:41093", "right": "issue:43638", "accept": false, - "reason": "Mask-shape mismatch and ZeRO-3 index-out-of-bounds are different failure modes." + "reason": "Mask shape mismatch and ZeRO-3 zero-size index error are different error signatures." }, { "left": "issue:39290", "right": "issue:43531", "accept": false, - "reason": "Different Qwen/Gemma config issues, not the same bug." + "reason": "Both involve recent Qwen/Gemma-style models, but the reported missing attribute differs and there is no clear same bug." }, { "left": "issue:43404", "right": "issue:43866", "accept": false, - "reason": "lm_head tying bug and checkpoint corruption are unrelated." + "reason": "Untied lm_head weights and corrupted checkpoint are unrelated." }, { "left": "issue:45362", "right": "issue:45464", "accept": false, - "reason": "Same model family, but different symptoms and execution paths; not the same concrete bug." + "reason": "Both are Qwen3.5, but different sizes and different runtime paths make this too weak to merge." }, { "left": "issue:43792", "right": "issue:45070", "accept": false, - "reason": "Whisper runtime failure and pydantic field regression are unrelated." + "reason": "Whisper inference failure and pydantic field regression are unrelated." }, { "left": "issue:43541", "right": "issue:43828", "accept": false, - "reason": "Grouped_mm tracing failure and autocast dtype mismatch are different bugs." + "reason": "Grouped_mm tracing error and autocast dtype mismatch are different runtime issues." }, { "left": "issue:36010", "right": "issue:39290", "accept": false, - "reason": "GenerationMixin import error and Gemma3/vLLM config issue are unrelated." + "reason": "ImportError for GenerationMixin and Gemma3 config attribute error are unrelated." }, { "left": "issue:44841", "right": "issue:45084", "accept": false, - "reason": "Processor failure and template compilation error are unrelated." + "reason": "Processor failure and template compilation error are not the same bug." }, { "left": "issue:41093", "right": "issue:41762", "accept": false, - "reason": "Both are shape/index problems, but the reported failure modes and likely code paths differ." + "reason": "Both are index-related, but one is a mask-size mismatch and the other is a zero-size load failure." }, { "left": "issue:43296", "right": "issue:43366", "accept": false, - "reason": "Different model support/load issues with no shared concrete change." + "reason": "Different model-support requests/errors; not the same concrete issue." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary stays within the packet evidence and the soft-edge calls are conservative. The ZeRO-3/BERT/Gemma cluster is the clearest duplicate set, and the docs-only cluster is appropriately treated as weaker." + "feedback": "The summary is grounded in the packet and stays conservative. The two accepted ZeRO-3 index-out-of-bounds pairs are well supported by the titles, and the rest of the soft-edge rejections do not overclaim duplication from the available evidence." }, "evaluator_used": true, "retried": false, @@ -582,8 +570,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -665,27 +653,27 @@ "issue:43950|issue:44534", "issue:43644|issue:43950", "issue:44292|issue:45507", - "issue:43828|issue:44928", "issue:39692|issue:43054", + "issue:43828|issue:44928", "issue:41553|issue:43883", "issue:43866|issue:44403", "issue:36331|issue:41762", "issue:44560|issue:45507" ], - "prepared_review_unit_hash": "2af7e5185179a934c9888ce3cec0fd72d917bd15ea784d833f44aeacd9f4115b", + "prepared_review_unit_hash": "8e4a76b427a830fbff24d9f89ece1457b2939ffc99ee7f27ae39ec5dec96a71e", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 34, - "serialized_chars": 37571, + "serialized_chars": 37569, "estimated_input_tokens": 9393, "estimated_eval_tokens": 19042 }, @@ -694,216 +682,222 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is mostly unrelated issue reports; the only clear duplicates are the non-persistent buffer regression reports in Transformers 5.x. The rest differ by model, failure mode, or subsystem and should stay separate.", + "summary": "Most pairs are not true duplicates; the only clear duplicate cluster is the non-persistent buffer corruption regression in Transformers v5 (`43644`, `43950`, `44534`).", "confidence": 0.84, - "canonical_issue_reason": "issue:43950 is the strongest canonical issue for the duplicate buffer regression because it states the exact regression (`from_pretrained()` corrupting `register_buffer(persistent=False)`) and is the most specific root-cause description.", + "canonical_issue_reason": "`issue:43950` is the clearest and most specific statement of the underlying bug (`from_pretrained` corrupting `register_buffer(persistent=False)` buffers), and it matches the other buffer-corruption reports.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43950 best represents the duplicate cluster because it precisely names the v5 non-persistent-buffer bug, making it the most useful anchor for triage and linking related reports.", + "best_issue_reason": "`issue:43950` is the most precise, actionable representative of the duplicate cluster and best captures the concrete regression.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43572", "right": "issue:45070", "accept": false, - "reason": "Both are v5/config-related regressions, but one is about `StableLmConfig.pad_token_idx` and the other is a pydantic `PretrainedConfig` field issue; different bugs." + "reason": "Both are v5 config-related, but one is a missing `pad_token_idx` field in StableLM and the other is a pydantic `PretrainedConfig` regression; not the same bug." }, { "left": "issue:44387", "right": "issue:45005", "accept": false, - "reason": "One reports int4 quantization memory/OOM behavior, the other tied-weights problems in translation models; different failure modes." + "reason": "Both mention v5/quantization or weights, but the symptoms and code paths differ: int4 memory/OOM vs tied weights in translation models." }, { "left": "issue:43541", "right": "issue:45161", "accept": false, - "reason": "Both involve MoE/generation paths, but one is a torch-dynamo `grouped_mm` tracing error and the other is a tensor-parallel support issue for GPT-OSS; not the same bug." + "reason": "Both involve MoE models, but one is a torch dynamo grouped_mm failure and the other is a TP-only support issue; not the same underlying problem." }, { "left": "issue:43828", "right": "issue:45464", "accept": false, - "reason": "Autocast dtype mismatch in Phi-tiny-MoE is unrelated to a streaming chat/completions failure on Qwen3.5-0.8B." + "reason": "Different failures and subsystems: autocast dtype mismatch vs chat/completions streaming inference." }, { "left": "issue:44291", "right": "issue:45070", "accept": false, - "reason": "`init_empty_weights` unexpectedly passing `_is_hf_initialized` is a different initialization bug than the pydantic config-field regression." + "reason": "`init_empty_weights` unexpected-arg error is a constructor/init path bug, while 45070 is a pydantic config-field regression." }, { "left": "issue:43366", "right": "issue:45464", "accept": false, - "reason": "GGUF support for gpt-oss and a streaming API failure are unrelated." + "reason": "GGUF gpt-oss support and Qwen3.5 streaming API failures are unrelated changes with different targets." }, { "left": "issue:43531", "right": "issue:45070", "accept": false, - "reason": "A Qwen3-MoE `sliding_window` issue is not the same as the generic `PretrainedConfig` field regression." + "reason": "Qwen3-MoE sliding-window handling and pydantic `PretrainedConfig` field issues are not the same bug." }, { "left": "issue:44661", "right": "issue:45464", "accept": false, - "reason": "Tokenizer-mapping failure in `add-new-model-like` is unrelated to a Qwen3.5 streaming inference bug." + "reason": "Tokenizer mapping/add-new-model-like logic is unrelated to the Qwen3.5 streaming inference failure." }, { "left": "issue:39290", "right": "issue:43742", "accept": false, - "reason": "Both are load-time errors, but one is a missing `sliding_window_pattern` attribute for Gemma3/vLLM and the other is a key error loading MobileLLM." + "reason": "Gemma3/vLLM missing attribute and MobileLLM loading key error are different model-specific loading bugs." }, { "left": "issue:43054", "right": "issue:43994", "accept": false, - "reason": "Both mention SigLIP2, but one is about worse text embeddings while the other is about nonsensical outputs in AutoModel/pipeline; too different without stronger evidence." + "reason": "Both are about SigLIP2, but one reports embedding quality degradation and the other a pipeline/AutoModel output issue; not clearly the same concrete bug." }, { "left": "issue:43541", "right": "issue:44589", "accept": false, - "reason": "Torch-dynamo grouped_mm tracing and missing Float8 storage are distinct runtime errors." + "reason": "Dynamo tracing on Mixtral float32 and missing Float8 storage are different runtime failures." }, { "left": "issue:41093", "right": "issue:43596", "accept": false, - "reason": "Both are indexing/shape errors, but one is a mask-length mismatch and the other is a DeepSpeed ZeRO-3 zero-size tensor index error." + "reason": "Both are index errors, but they arise in different paths: masking mismatch vs DeepSpeed ZeRO-3/BertModel init." }, { "left": "issue:42915", "right": "issue:43541", "accept": false, - "reason": "FineGrainedFP8Config failure for Qwen3-MoE is not the same as a Mixtral grouped_mm tracing error." + "reason": "Qwen3Moe FP8 failure and Mixtral torch dynamo grouped_mm failure are unrelated model/runtime bugs." }, { "left": "issue:40990", "right": "issue:44960", "accept": false, - "reason": "High perplexity on GPT-OSS and a GLM5 issue are unrelated model-specific reports." + "reason": "Perplexity regression on gpt-oss-20b and the generic `GLM5` issue do not appear to describe the same bug." }, { "left": "issue:39186", "right": "issue:44918", "accept": false, - "reason": "An FSDP weight-shape runtime error is unrelated to TRL SFT input-embedding unpacking on Qwen3.5." + "reason": "FSDP `'weight' must be 2-D` and Qwen3.5 embedding unpacking with TRL SFT trainer are different training/loading issues." }, { "left": "issue:43404", "right": "issue:45005", "accept": false, - "reason": "Both involve tied weights, but one is a Mistral3 image-text model bug and the other is a broader v5 translation-model tied-weights regression; not enough to merge as the same bug." + "reason": "Both mention tied weights, but one is a Mistral3 lm_head tying bug and the other is a broader translation-model v5 report; not the same concrete failure." }, { "left": "issue:43645", "right": "issue:45399", "accept": false, - "reason": "Custom-model initialization in notebooks and flash-attn2 fallback gating are different issues." + "reason": "Jupyter custom-model initialization and flash-attn2 fallback checks are unrelated regressions." }, { "left": "issue:44530", "right": "issue:45005", "accept": false, - "reason": "PagedAttentionCache group-type errors on Qwen3.5 models are unrelated to tied-weight handling." + "reason": "PagedAttentionCache/linear_attention on Qwen3.5 is unrelated to tied-weights handling in translation models." }, { "left": "issue:44863", "right": "issue:45005", "accept": false, - "reason": "NemotronH checkpoint loading is a model-compatibility problem, not the tied-weights regression." + "reason": "NemotronH checkpoint loading and tied-weights translation-model issues are different model-specific bugs." }, { "left": "issue:43873", "right": "issue:45005", "accept": false, - "reason": "Quantization/offloading behavior is unrelated to tied weights." + "reason": "Quantization/offloading behavior and tied-weights handling are different failure modes." }, { "left": "issue:44360", "right": "issue:44512", "accept": false, - "reason": "A DSA indexer activation request and a docs typo are unrelated." + "reason": "DSA indexer behavior and removed-run-command docs are unrelated." + }, + { + "left": "issue:42915", + "right": "issue:45399", + "accept": false, + "reason": "Qwen3Moe FP8 support and flash-attn2 fallback gating are different code paths." }, { "left": "issue:44918", "right": "issue:45507", "accept": false, - "reason": "TRL SFT embedding unpacking and a GraniteMoEHybrid invalid-method call are different code paths." + "reason": "TRL SFT unpacking and GraniteMoEHybrid invalid method calls are unrelated trainer/model bugs." }, { "left": "issue:44492", "right": "issue:44512", "accept": false, - "reason": "Cache-strategy typo and removed-run-command docs are unrelated documentation issues." + "reason": "Cache-strategy typo and docs about a removed command are both documentation-level but not the same issue." }, { "left": "issue:43975", "right": "issue:44292", "accept": false, - "reason": "Detokenization in v5 and a Qwen-3-8B-NVFP4 runtime error are different model/runtime problems." + "reason": "Detokenization on DeepSeek Coder and Qwen-3-8B-NVFP4 runtime errors are different model problems." }, { "left": "issue:43950", "right": "issue:44534", "accept": true, - "reason": "These describe the same v5 regression: non-persistent buffers being populated incorrectly during loading." + "reason": "Same regression: both report v5 corrupting non-persistent buffers / `register_buffer(persistent=False)` state during loading." }, { "left": "issue:43644", "right": "issue:43950", "accept": true, - "reason": "Both report the same non-persistent-buffer corruption regression in Transformers 5.x." + "reason": "Same underlying bug: Transformers v5/from_pretrained is filling or corrupting non-persistent buffers with junk." }, { "left": "issue:44292", "right": "issue:45507", "accept": false, - "reason": "Qwen-3-8B-NVFP4 runtime failure is unrelated to GraniteMoEHybrid invalid-method calls." + "reason": "Qwen-3-8B-NVFP4 runtime error and GraniteMoEHybrid invalid method are unrelated model/runtime failures." }, { - "left": "issue:43828", - "right": "issue:44928", + "left": "issue:39692", + "right": "issue:43054", "accept": false, - "reason": "Autocast dtype mismatch and Qwen3.5 RLHF NaN explosion are different failures, even though both involve MoE/precision concerns." + "reason": "Both involve SigLIP2, but one is docs-example/quantization issues and the other is a text-embedding quality complaint; not the same bug." }, { - "left": "issue:39692", - "right": "issue:43054", + "left": "issue:43828", + "right": "issue:44928", "accept": false, - "reason": "A broken SigLIP2 docs example is not the same as a separate report that SigLIP2 embeddings are worse than SigLIP." + "reason": "Autocast dtype mismatch and RLHF gradient explosion from 3D position ids are distinct failures." }, { "left": "issue:41553", "right": "issue:43883", "accept": false, - "reason": "Bad AutoTokenizer error messaging for Voxtral is unrelated to missing `all_tied_weights_keys` on MolmoForCausalLM." + "reason": "Bad AutoTokenizer error messaging for Voxtral and missing `all_tied_weights_keys` on MolmoForCausalLM are unrelated." }, { "left": "issue:43866", "right": "issue:44403", "accept": false, - "reason": "A corrupted Ovis2 checkpoint and loading noise are not the same bug." + "reason": "Corrupted Ovis2 checkpoint and generic loading noise are not the same concrete bug." }, { "left": "issue:36331", "right": "issue:41762", "accept": false, - "reason": "A custom trainer signature mismatch is unrelated to a Gemma3 DeepSpeed ZeRO-3 indexing error." + "reason": "CustomTrainer `compute_loss` signature mismatch and Gemma3 ZeRO-3 loading index error are unrelated training/loading bugs." }, { "left": "issue:44560", "right": "issue:45507", "accept": false, - "reason": "Qwen3-vl video StopIteration and GraniteMoEHybrid invalid-method are unrelated." + "reason": "Qwen3-vl-embedding video StopIteration and GraniteMoEHybrid invalid method calls are different errors." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. The two accepted soft-edge pairs are well supported by near-identical issue titles and descriptions about the same non-persistent-buffer regression in Transformers 5.x, and the other rejections avoid overclaiming duplicate status." + "feedback": "The cluster summary is grounded in the packet, and the accepted soft-edge links are conservative: both accepted pairs clearly point to the same Transformers v5 non-persistent buffer corruption regression. The rejections are also generally cautious and the reasoning does not overstate equivalence." }, "evaluator_used": true, "retried": false, @@ -919,8 +913,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -982,8 +976,8 @@ "issue:43881|issue:43883", "issue:44403|issue:44451", "issue:43975|issue:45520", - "issue:42947|issue:43856", "issue:43643|issue:45020", + "issue:42947|issue:43856", "issue:44403|issue:44488", "issue:43782|issue:44960", "issue:42491|issue:44368", @@ -1011,259 +1005,259 @@ "issue:42915|issue:43872", "issue:43257|issue:45305" ], - "prepared_review_unit_hash": "d758c0340a6177fa2fdb80123af951978fefa5db6e53f3406086f833a67499e2", + "prepared_review_unit_hash": "cd8879a6f1e6dd337c20e23631136f7b305c5486979dda1591d027ee5aeb276f", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 36, - "serialized_chars": 38600, - "estimated_input_tokens": 9650, - "estimated_eval_tokens": 19556 + "serialized_chars": 38601, + "estimated_input_tokens": 9651, + "estimated_eval_tokens": 19558 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a mix of unrelated reports with several tight duplicate clusters. The SigLIP2 group should be described more conservatively: 43493 is the best anchor for the HF-vs-JAX discrepancy, but 38175 looks only loosely related from the available evidence rather than a confirmed duplicate of 43054.", - "confidence": 0.63, - "canonical_issue_reason": "43493 is the strongest single anchor because it names the concrete SigLIP2 HF-vs-JAX implementation discrepancy; it is the cleanest root-cause-style report in the subgroup.", + "summary": "This cluster is mostly heterogeneous, but a few soft pairs are true duplicates: SigLIP2 implementation mismatch, stale v5 docs, Whisper fbank extraction, Qwen3 ZeRO-3 weight loading, and the _is_hf_initialized/init_empty_weights TypeError. Most other pairs are only broadly related by subsystem, not the same bug.", + "confidence": 0.89, + "canonical_issue_reason": "issue:45020 is the broadest representative of the cluster\u2019s main theme: recent Transformers v5 loading/tokenizer regressions, especially around remote_code and model initialization.", "canonical_pr_reason": null, - "best_issue_reason": "43493 is the best overall issue representative because it captures a specific implementation mismatch rather than just a downstream symptom, and it is the clearest anchor for the SigLIP2 cluster.", + "best_issue_reason": "issue:45020 is the best single representative because it is broad, current, and central to the model-loading regression theme spanning several items in this cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41553", "right": "issue:43881", "accept": false, - "reason": "Different models and different failures; both are loading issues but not the same bug." + "reason": "Different models and failures: Voxtral AutoTokenizer error-message issue vs GLM-4v-9b loading failure." }, { "left": "issue:43872", "right": "issue:44589", "accept": false, - "reason": "Both involve compatibility/loading problems, but the `_is_hf_initialized` kwarg bug is distinct from the Float8 storage error." + "reason": "Both are quantization/storage-related, but the root causes differ: bitsandbytes _is_hf_initialized vs missing Float8 storage type." }, { "left": "issue:38175", "right": "issue:43054", "accept": false, - "reason": "They are both SigLIP2-related quality reports, but the packet only supports broad thematic similarity, not a confirmed shared underlying implementation bug." + "reason": "Same model family, but zero probabilities vs degraded text embeddings are not clearly the same underlying bug." }, { "left": "issue:42947", "right": "issue:44387", "accept": false, - "reason": "Different problems: LoRA gradient checkpointing vs int4 quantization memory/OOM." + "reason": "Both concern training/memory behavior, but gradient checkpointing/LoRA and int4 CUDA reserved memory are distinct issues." }, { "left": "issue:43054", "right": "issue:43493", "accept": true, - "reason": "Same SigLIP2 discrepancy family; 43493 states the root HF-vs-JAX mismatch behind the degraded outputs." + "reason": "Both point to a SigLIP2 implementation mismatch versus reference behavior, with degraded embeddings as a likely symptom of the same bug." }, { "left": "issue:43881", "right": "issue:43883", "accept": false, - "reason": "Different model families and different failure modes." + "reason": "Unrelated failures: GLM-4v loading vs Molmo missing all_tied_weights_keys." }, { "left": "issue:44403", "right": "issue:44451", "accept": false, - "reason": "Generic loading noise/warnings vs a specific model load failure." + "reason": "Generic loading noise is not the same as a specific ScandiBERT loading regression." }, { "left": "issue:43975", "right": "issue:45520", "accept": false, - "reason": "Unrelated subsystems: tokenizer detokenization vs flash-attn import/key errors." + "reason": "Different areas: tokenizer detokenization regression vs flash_attn import KeyError on Python 3.13." }, { - "left": "issue:42947", - "right": "issue:43856", + "left": "issue:43643", + "right": "issue:45020", "accept": false, - "reason": "Both mention memory, but the training/checkpointing contexts and root causes differ." + "reason": "Both involve remote_code/config handling, but one is missing fields in AutoConfig and the other is broader model breakage; not the same concrete bug." }, { - "left": "issue:43643", - "right": "issue:45020", + "left": "issue:42947", + "right": "issue:43856", "accept": false, - "reason": "Both are remote_code-related, but one is a specific AutoConfig symptom and the other is a broad umbrella report; not enough to confirm the same bug." + "reason": "LoRA gradient checkpointing and Qwen3 MoE memory usage are different training bugs." }, { "left": "issue:44403", "right": "issue:44488", "accept": false, - "reason": "These are separate model-loading reports, not the same defect." + "reason": "Both are loading-related, but the underlying models and symptoms differ." }, { "left": "issue:43782", "right": "issue:44960", "accept": false, - "reason": "Different model families and incompatible error types." + "reason": "Qwen3VL weight_only loading error and GLM5 loading are unrelated model-specific problems." }, { "left": "issue:42491", "right": "issue:44368", "accept": false, - "reason": "LoRA hf4.x/hf5.x compatibility breakage is unrelated to a tie_word_embeddings warning." + "reason": "LoRA model compatibility on hf4.x/hf5.x is not the same as a tie_word_embeddings warning." }, { "left": "issue:41762", "right": "issue:42915", "accept": false, - "reason": "Different models and different DeepSpeed/config problems." + "reason": "Both mention DeepSpeed, but Gemma3 ZeRO-3 index error and Qwen3 MoE FP8 failure are different code paths." }, { "left": "issue:43975", "right": "issue:44451", "accept": false, - "reason": "Different issues: detokenization regression vs a model load failure." + "reason": "Different tokenizer/load failures for different models." }, { "left": "issue:45020", "right": "issue:45356", "accept": false, - "reason": "Both mention remote_code/tokenizer regressions, but the Kimi tokenizer codec bug is a different concrete problem." + "reason": "Broad remote_code regressions and Kimi-K2.5 tokenizer codec/fix_mistral_regex regression are distinct bugs." }, { "left": "issue:44661", "right": "issue:45020", "accept": false, - "reason": "Tokenizer-mapping template generation bug is not the same as remote_code loading regressions." + "reason": "add-new-model-like failure in tokenizer mapping is a separate codegen/tooling bug, not the same as remote_code model breakage." }, { "left": "issue:43054", "right": "issue:43278", "accept": false, - "reason": "SigLIP2 embedding quality vs a dtype change during evaluate; not the same bug." + "reason": "SigLIP2 quality discrepancy and BF16-to-FP32 dtype change are different symptoms with no clear shared root cause." }, { "left": "issue:43824", "right": "issue:43872", "accept": false, - "reason": "Importing a missing VL class is unrelated to the bitsandbytes `_is_hf_initialized` incompatibility." + "reason": "Qwen2.5-VL import error and bitsandbytes init_empty_weights incompatibility are unrelated." }, { "left": "issue:44991", "right": "issue:45399", "accept": false, - "reason": "Tokenizer loading failure and flash-attn fallback logic are different code paths." + "reason": "Tokenizer loading failure for est-roberta and flash-attn2 fallback blocking are unrelated." }, { "left": "issue:43827", "right": "issue:44509", "accept": true, - "reason": "Same docs cleanup: both still reference pipeline tasks removed in v5." + "reason": "Same documentation regression: stale docs still reference pipeline tasks removed in v5." }, { "left": "issue:43792", "right": "issue:44220", - "accept": false, - "reason": "The Whisper report is too generic to confirm it is the same fbank helper bug." + "accept": true, + "reason": "Both describe the same Whisper audio-feature path; the fbank extraction helper is the likely concrete cause of Whisper failing to run." }, { "left": "issue:45081", "right": "issue:45399", "accept": false, - "reason": "Tokenizer regex patch crash and flash-attn fallback gating are unrelated." + "reason": "Mistral regex patch crash and flash-attn2 fallback blocking are separate issues." }, { "left": "issue:43404", "right": "issue:43950", "accept": false, - "reason": "Weight-tying in Mistral3 and non-persistent buffer corruption are distinct issues." + "reason": "lm_head tying in Mistral3 and silent corruption of non-persistent buffers are unrelated model-loading bugs." }, { "left": "issue:43257", "right": "issue:43866", "accept": false, - "reason": "Qwen3 MoE deepspeed conversion bug vs an Ovis2 checkpoint corruption report." + "reason": "Qwen3 MoE weight conversion with accelerate+deepspeed and Ovis2 checkpoint corruption are different problems." }, { "left": "issue:42915", "right": "issue:45020", "accept": false, - "reason": "Qwen3 MoE FP8 training/memory issue is unrelated to remote_code loading regressions." + "reason": "Qwen3 MoE FP8 training failure is narrower and different from the broad remote_code loading regressions." }, { "left": "issue:43742", "right": "issue:45399", "accept": false, - "reason": "Different models and different failure mechanisms." + "reason": "MobileLLM loading failure and flash-attn2 fallback checks are unrelated." }, { "left": "issue:41553", "right": "issue:44991", "accept": false, - "reason": "Different model families and different tokenizer-loading errors." + "reason": "Both are tokenizer-loading issues, but they affect different models and likely different root causes." }, { "left": "issue:43742", "right": "issue:45020", "accept": false, - "reason": "A specific model key error is not the same as the broader remote_code regression." + "reason": "MobileLLM KeyError is a specific model load bug, not the same as the broad remote_code regression bucket." }, { "left": "issue:43940", "right": "issue:45313", "accept": true, - "reason": "Same DeepSpeed ZeRO-3 weight-loading failure class on Qwen3-family models." + "reason": "Same concrete failure class: Qwen3-family DeepSpeed ZeRO-3 weight loading leaves parameters missing." }, { "left": "issue:43872", "right": "issue:44291", "accept": true, - "reason": "Exact same `_is_hf_initialized` / init_empty_weights incompatibility." + "reason": "These are the same TypeError class: loading with init_empty_weights hits unexpected _is_hf_initialized handling." }, { "left": "issue:43854", "right": "issue:44863", "accept": false, - "reason": "Different model families and different checkpoint-loading failures." + "reason": "GLM-4.7-Flash unit-test loading problem and NemotronH checkpoint loading failure are distinct model-specific bugs." }, { "left": "issue:44928", "right": "issue:45588", "accept": false, - "reason": "RLHF gradient explosion from position_ids/SDPA fallback is unrelated to flash_attention s_aux handling." + "reason": "RLHF NaN from 3D position_ids/SDPA fallback and flash_attention s_aux=None crash are different failures." }, { "left": "issue:41553", "right": "issue:44488", "accept": false, - "reason": "Separate model-loading failures, not the same bug." + "reason": "Different tokenizer/model-loading failures for Voxtral and sleng-bert." }, { "left": "issue:42915", "right": "issue:43872", "accept": false, - "reason": "Qwen3 MoE FP8 training memory issue is unrelated to the bitsandbytes `_is_hf_initialized` crash." + "reason": "Qwen3 MoE FP8 training failure and bitsandbytes init_empty_weights incompatibility are unrelated." }, { "left": "issue:43257", "right": "issue:45305", "accept": false, - "reason": "Different Qwen3 training/load bugs with different symptoms and code paths." + "reason": "Qwen3 MoE weight conversion/loading and gradient averaging under DeepSpeed are different bugs." } ] }, "evaluator_result": { - "accept": false, - "feedback": "Mostly grounded, but one soft-edge verdict is too strong for the evidence available. The SigLIP2 pair (issue:38175 \u2194 issue:43054) is only supported by broad title similarity; from the packet alone it is safer to call them related symptoms, not the same underlying implementation mismatch. Because the cluster summary also presents SigLIP2 as a clear duplicate subgroup, the overall writeup is a bit overconfident and should be made more conservative." + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative on the soft edges. The accepted pairs are plausibly the same bug class, and the rejected pairs are mostly only subsystem-level matches rather than clear duplicates. No overstatement stands out." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -1276,8 +1270,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -1361,14 +1355,14 @@ "issue:42491|issue:43720", "issue:44960|issue:44977" ], - "prepared_review_unit_hash": "4a0772e4c38cbdc509a62209740a10779b54923e621469b01c702cc07f6bfb33", + "prepared_review_unit_hash": "aa7006a996d29ff284f956a94d66ee4d595c3f1c16fbf2747192deb628227d94", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 47, @@ -1383,198 +1377,198 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a loose collection of mostly unrelated Transformers issue reports, with a few clear duplicate subclusters: non-persistent buffer corruption, custom model initialization failures, Qwen video-regression reports, and repeated mask-shape IndexErrors. No PRs are present.", - "confidence": 0.83, - "canonical_issue_reason": "issue:44534 is the cleanest representative of the strongest duplicate subcluster: it is the broader, clearer restatement of the non-persistent-buffer regression and subsumes issue:43644.", + "summary": "This set is mostly a bag of unrelated issues, with only two clear duplicate pairs: the Transformers 5.0 custom model initialization regression, and the non-persistent buffer corruption regression. The rest are only superficially similar by title and should stay separate.", + "confidence": 0.86, + "canonical_issue_reason": "No single canonical issue exists for the full set because the items span many unrelated bugs; if a duplicate anchor is required, issue 43644 is the best representative of the non-persistent-buffer regression and is essentially duplicated by 44534.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44534 is the best global issue representative because it is the most general and clearly phrased duplicate target among the issues in this cluster.", + "best_issue_reason": "Issue 43644 is the strongest issue candidate because it is the clearest, earliest report of the non-persistent buffer junk regression and has a near-identical duplicate in 44534.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:36331", "right": "issue:41093", "accept": false, - "reason": "Different failures: custom trainer API mismatch vs tensor mask-shape IndexError." + "reason": "Different failures: trainer API mismatch vs mask/tensor shape IndexError." }, { "left": "issue:41093", "right": "issue:44805", - "accept": true, - "reason": "Same generic mask/tensor shape-mismatch error; titles indicate the same underlying IndexError regression." + "accept": false, + "reason": "Only the error wording overlaps; the reported shapes/models and likely root causes differ." }, { "left": "issue:43872", "right": "issue:45538", "accept": false, - "reason": "Unrelated bugs: bitsandbytes constructor incompatibility vs tokenizer max-length constant." + "reason": "Unrelated bugs: bitsandbytes initialization error vs tokenizer model_max_length behavior." }, { "left": "issue:44479", "right": "issue:44560", - "accept": true, - "reason": "Both report the same Qwen video-input regression in the 5.3.0 timeframe; same affected code path is likely." + "accept": false, + "reason": "Same broad Qwen3/VL area, but different concrete bugs and code paths (video regression vs embedding/tracks issue)." }, { "left": "issue:44530", "right": "issue:44863", "accept": false, - "reason": "Different model/path failures: PagedAttentionCache linear_attention crash vs NemotronH checkpoint loading." + "reason": "Different subsystems and failures: cache/group-type crash vs checkpoint loading for a different model family." }, { "left": "issue:43366", "right": "issue:43828", "accept": false, - "reason": "Different areas and symptoms: GGUF/gpt-oss support vs autocast dtype mismatch in Phi-tiny-MoE." + "reason": "Unrelated: GGUF gpt-oss architecture support vs autocast dtype mismatch." }, { "left": "issue:24643", "right": "issue:30064", "accept": false, - "reason": "Completely different bugs: DeepSpeed training weight shape error vs void segmentation-map processing." + "reason": "DeepSpeed training weight-shape error is not the same bug as void segmentation map processing." }, { "left": "issue:44155", "right": "issue:45584", "accept": false, - "reason": "Different audio/ASR issues with no shared code path evident." + "reason": "Different models and mechanics: batched AudioFlamingo track leakage vs Whisper empty-transcription handling." }, { "left": "issue:43366", "right": "issue:43531", "accept": false, - "reason": "Different features/models: gpt-oss GGUF support vs Qwen3-MoE sliding_window handling." + "reason": "Both mention Qwen-related topics, but one is GGUF support and the other is a sliding_window issue; not the same bug." }, { "left": "issue:43550", "right": "issue:44910", "accept": false, - "reason": "Different attention/compile failures in different models and code paths." + "reason": "Different attention/backend crashes with different models and error modes." }, { "left": "issue:43645", "right": "issue:43646", "accept": true, - "reason": "Very close duplicate wording around custom model initialization breaking in Transformers 5.0." + "reason": "Same underlying Transformers 5.0 custom model initialization regression; the notebook report is a context-specific manifestation of the general bug." }, { "left": "issue:43950", "right": "issue:45325", "accept": false, - "reason": "Different regressions: non-persistent buffer corruption vs Qwen2.5-VL rope/position-id scaling." + "reason": "Non-persistent buffer corruption regression vs Qwen2.5-VL rope scaling bug; different code paths." }, { "left": "issue:44560", "right": "issue:44918", "accept": false, - "reason": "Both are Qwen video/embedding related, but the concrete failures and code paths differ (video StopIteration vs TRL SFT unpacking)." + "reason": "Both involve Qwen models, but one is a video StopIteration regression and the other is a TRL embedding unpacking failure." }, { "left": "issue:43366", "right": "issue:44589", "accept": false, - "reason": "Unrelated: GGUF architecture support vs float8 storage lookup failure." + "reason": "No shared bug: GGUF support request vs missing Float8 storage type." }, { "left": "issue:43824", "right": "issue:43957", "accept": false, - "reason": "Different load/import regressions with different mechanisms and affected models." + "reason": "ImportError for Qwen2.5-VL is unrelated to meta-device loading regressions." }, { "left": "issue:42915", "right": "issue:43957", "accept": false, - "reason": "Different bugs: FineGrainedFP8Config failure vs meta-device loading failure." + "reason": "Qwen3Moe FP8 failure and meta-device loading problems are different issues." }, { "left": "issue:42617", "right": "issue:43366", "accept": false, - "reason": "Different subjects: 3D parallel script failure vs GGUF gpt-oss support." + "reason": "Different problems: 3d_parallel.py execution vs GGUF architecture support." }, { "left": "issue:38175", "right": "issue:43994", "accept": false, - "reason": "Same model family, but the reports are too symptom-level and not clearly the same underlying bug." + "reason": "Same model family, but one is zero-probabilities behavior and the other is an AutoModel/pipeline nonsensical-results report with different root cause." }, { "left": "issue:43646", "right": "issue:43950", "accept": false, - "reason": "Different custom-model regression vs silent non-persistent-buffer corruption." + "reason": "Custom model init regression is distinct from the non-persistent buffer corruption regression." }, { "left": "issue:44560", "right": "issue:44805", "accept": false, - "reason": "Different code paths and symptoms: Qwen video StopIteration vs mask/tensor shape mismatch." + "reason": "Different Qwen video issues; the reported failures and symptoms do not line up." }, { "left": "issue:36010", "right": "issue:42915", "accept": false, - "reason": "Different failures: missing GenerationMixin import vs Qwen3-MoE FP8 config issue." + "reason": "ImportError for GenerationMixin is unrelated to Qwen3Moe FineGrainedFP8 loading." }, { "left": "issue:43638", "right": "issue:44661", "accept": false, - "reason": "Different bugs: DeepSpeed Zero-3/Bert index error vs add-new-model-like tokenizer mapping issue." + "reason": "DeepSpeed/Bert index error vs tokenizer mapping-name constraint are unrelated." }, { "left": "issue:44877", "right": "issue:45030", "accept": false, - "reason": "Both are config-validation related, but the affected models and failure modes differ." + "reason": "Both are config-loading regressions, but the models and failure mechanisms differ." }, { "left": "issue:43596", "right": "issue:44661", "accept": false, - "reason": "Different code paths: Zero-3 Bert init error vs tokenizer-mapping registration failure." + "reason": "Different bugs: ZeRO-3 init index error vs add-new-model-like failing in TOKENIZER_MAPPING_NAMES." }, { "left": "issue:43317", "right": "issue:44756", "accept": false, - "reason": "Different problems: dequantized-model offload loading vs mmap OOM on Strix Halo." + "reason": "device_map/offload loading failure is not the same as mmap/OOM on Strix Halo." }, { "left": "issue:43644", "right": "issue:44534", "accept": true, - "reason": "Direct duplicates: same non-persistent-buffer regression, with 44534 being the broader rephrasing." + "reason": "Same underlying Transformers v5 regression: non-persistent buffers get corrupted/filled with junk." }, { "left": "issue:45313", "right": "issue:45357", "accept": false, - "reason": "Different Qwen3.5 load/save regressions affecting different components." + "reason": "Different Qwen3.5 regressions: ZeRO-3 language_model weight loading vs save_pretrained visual encoder keys." }, { "left": "issue:38175", "right": "issue:39692", "accept": false, - "reason": "SigLIP2 is the only shared context, but 39692 is a broader documentation-example issue with additional quantization/model-processor problems." + "reason": "SigLIP2 output anomaly vs docs example/model-processor mismatch and quantization failure are not the same bug." }, { "left": "issue:42491", "right": "issue:43720", "accept": false, - "reason": "Unrelated loading/compatibility problems in different quantization/accelerate paths." + "reason": "LoRA checkpoint compatibility on hf5.x is unrelated to packed-weight handling in BitNet AutoBitLinear." }, { "left": "issue:44960", "right": "issue:44977", "accept": false, - "reason": "Different model families and failures: GLM5 vs Qwen3.5 flash-attention generation." + "reason": "GLM5 and Qwen3.5 flash-attention failures are different model-specific bugs." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet: it correctly reflects that the set is mostly a collection of issue reports, and the named duplicate subclusters are supported by the issue titles. The no-PR statement is also consistent with the packet. The soft-edge acceptances are generally conservative, with the strongest evidence on the non-persistent-buffer and mask-shape pairs. " + "feedback": "The analyst output is grounded in the packet and stays conservative overall. The two accepted duplicate pairs are well supported by the titles, and the rejected soft-edge pairs mostly hinge on clearly different model/subsystem behaviors. No mergeability issues apply here because there are no PR pairs." }, "evaluator_used": true, "retried": false, @@ -1590,8 +1584,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -1653,8 +1647,8 @@ "issue:43257|issue:43854", "issue:39692|issue:43994", "issue:44488|issue:44991", - "issue:43856|issue:44387", "issue:45362|issue:45406", + "issue:43856|issue:44387", "issue:43644|issue:43645", "issue:42915|issue:44756", "issue:44661|issue:44683", @@ -1672,8 +1666,8 @@ "issue:43425|issue:43957", "issue:44387|issue:44756", "issue:43854|issue:45216", - "issue:44292|issue:44912", "issue:42947|issue:43381", + "issue:44292|issue:44912", "issue:44291|issue:44756", "issue:44743|issue:44960", "issue:45276|issue:45292", @@ -1683,268 +1677,232 @@ "issue:43646|issue:43708", "issue:45292|issue:45335" ], - "prepared_review_unit_hash": "f11c5d19c87a85f9fe938011bbc90b84a995dde28423c1230b7841abea208c04", + "prepared_review_unit_hash": "9c84d8d354c202417f84c58d5517d2e55f9e2a68aaff7647ce8319de72f6fe11", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 38, - "serialized_chars": 38399, - "estimated_input_tokens": 9600, - "estimated_eval_tokens": 19456 + "serialized_chars": 38392, + "estimated_input_tokens": 9598, + "estimated_eval_tokens": 19452 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "There is one clear duplicate cluster around `resize_token_embeddings` not propagating to output/decoder embeddings (issues 45276, 45292, 45335). The rest of the soft pairs are mostly thematically similar but appear to be different bugs/models and should stay separate.", - "confidence": 0.89, - "canonical_issue_reason": "issue:45292 is the broadest, model-agnostic report of the shared `resize_token_embeddings` bug and best covers the narrower Gemma4/T5Gemma variants.", + "summary": "The cluster is mostly unrelated. The only clear duplicate set is the `resize_token_embeddings` regression around model-specific embeddings/output heads; the rest are separate model-loading, quantization, checkpointing, or runtime bugs that only overlap at a broad subsystem level.", + "confidence": 0.92, + "canonical_issue_reason": "Issue 45276 is the best canonical issue: it is the broadest and most representative report for the `resize_token_embeddings` propagation bug, and it predates the closest follow-up reports.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45292 is the best cluster anchor: open, general, and directly describes the common failure mode that the other two issues narrow to specific model components.", + "best_issue_reason": "Issue 45276 is the strongest global issue representative because it captures the core regression most generally and matches the other accepted duplicate reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44451", "right": "issue:44488", "accept": false, - "reason": "Same general area (model loading) but different models and no clear shared code-path bug; likely related, not duplicate." + "reason": "Both are model-loading complaints, but they involve different models and no evidence of the same root cause." }, { "left": "issue:43708", "right": "issue:44075", "accept": false, - "reason": "Different trainer/optimizer problems; the titles do not describe the same underlying defect." + "reason": "Different bugs: checkpoint resume step \u0917\u0923culation vs optimizer arguments being ignored." }, { "left": "issue:43854", "right": "issue:43931", "accept": false, - "reason": "Both are model-loading failures, but for different models and distinct symptoms (unit-test load vs weight-shape mismatch)." + "reason": "Different models and different failure modes; not the same underlying defect." }, { "left": "issue:43856", "right": "issue:43873", "accept": false, - "reason": "Both involve memory/quantization, but one is Qwen3 MoE training memory use and the other is offloading under quantization; not the same bug." + "reason": "One is training memory usage, the other is quantization/offloading behavior; too different to be duplicates." }, { "left": "issue:42491", "right": "issue:45276", "accept": false, - "reason": "Different models and different failures; one is a LoRA compatibility issue, the other is resize-token-embedding propagation." + "reason": "LoRA hf4.x/hf5.x compatibility is unrelated to the `resize_token_embeddings` regression." }, { "left": "issue:41720", "right": "issue:42915", "accept": false, - "reason": "Both mention Qwen3 MoE, but one is a device-mapping crash and the other is a conversion/loading issue under accelerate+deepspeed." + "reason": "Both mention Qwen3/MoE, but the concrete failures are different code paths and symptoms." }, { "left": "issue:43257", "right": "issue:43854", "accept": false, - "reason": "Unrelated model families and failure modes; no shared concrete bug." + "reason": "Accelerate/deepspeed conversion issue vs model-load test failure; not the same bug." }, { "left": "issue:39692", "right": "issue:43994", "accept": false, - "reason": "Both are SigLIP2-related, but one is a docs/example issue and the other is a model/pipeline behavior report; not enough evidence of one underlying bug." + "reason": "Same model family, but one is a docs/example problem and the other is inference quality; insufficient to merge." }, { "left": "issue:44488", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer-loading regressions, but they concern different models and likely different configs/root causes." + "reason": "Different affected models and different tokenizer/load regressions." }, { - "left": "issue:43856", - "right": "issue:44387", + "left": "issue:45362", + "right": "issue:45406", "accept": false, - "reason": "Both are memory-related, but one is training-memory inefficiency and the other is int4 reserved-memory OOM; not the same change." + "reason": "Different products and failures: chat crash vs processor attribute error." }, { - "left": "issue:45362", - "right": "issue:45406", + "left": "issue:43856", + "right": "issue:44387", "accept": false, - "reason": "Different model classes and different processor/runtime failures; not a duplicate." + "reason": "Both are memory-related, but the reported causes are different (training footprint vs int4 OOM regression)." }, { "left": "issue:43644", "right": "issue:43645", "accept": false, - "reason": "Both are Transformers 5.0 regressions, but one is junk non-persistent buffers and the other is custom model init in notebooks; separate bugs." + "reason": "Both affect Transformers 5 custom-model scenarios, but the specific defects are different." }, { "left": "issue:42915", "right": "issue:44756", "accept": false, - "reason": "Unrelated issues: Qwen3 MoE conversion vs Strix Halo mmap/OOM." + "reason": "Unrelated: Qwen3 MoE FP8 failure vs mmap OOM mitigation on Strix Halo." }, { "left": "issue:44661", "right": "issue:44683", "accept": false, - "reason": "Completely different areas: tokenizer mapping helper vs compiled flex attention on new torch." + "reason": "Different areas entirely: tokenizer mapping validation vs compiled flex_attention on newer torch." }, { "left": "issue:42491", "right": "issue:43931", "accept": false, - "reason": "Different Qwen3/VL problems with no shared underlying code path." + "reason": "Different models and unrelated failures." }, { "left": "issue:44910", "right": "issue:45588", "accept": false, - "reason": "Both touch flash attention, but one is a Qwen3.5 packed-sequence interpretation bug and the other is an AttributeError on sink-less models; not the same defect." + "reason": "Both touch flash attention, but one is a packed-sequence/position_ids bug and the other is an s_aux None AttributeError." }, { "left": "issue:43404", "right": "issue:43454", "accept": false, - "reason": "Both are lm_head weight-tying complaints, but they target different model implementations and symptoms; too broad to merge." - }, - { - "left": "issue:42491", - "right": "issue:43856", - "accept": false, - "reason": "Different Qwen3 MoE issues: LoRA compatibility vs training memory usage." + "reason": "Both mention untied lm_head weights, but the model families and reported behavior differ enough that this is not clearly the same bug." }, { "left": "issue:45276", "right": "issue:45335", "accept": true, - "reason": "Same underlying `resize_token_embeddings` propagation bug; both report the embeddings not updating in different model wrappers." + "reason": "Same underlying bug: `resize_token_embeddings` does not propagate to model-specific embedding/output modules." }, { "left": "issue:36010", "right": "issue:39404", "accept": false, - "reason": "No shared code path: import error for GenerationMixin vs Whisper pipeline `return_language` regression." + "reason": "Unrelated import error vs Whisper pipeline regression." }, { "left": "issue:44164", "right": "issue:44291", "accept": false, - "reason": "Both involve save/load behavior, but one is `extra_state` handling and the other is `init_empty_weights`/`_is_hf_initialized`; different bugs." + "reason": "Different save/load initialization failures with different triggers." }, { "left": "issue:44402", "right": "issue:44488", "accept": false, - "reason": "Tokenizer vocab mismatch vs model loading failure for a different model; not the same issue." + "reason": "Different models and different tokenizer/vocab symptoms; not a duplicate pair." }, { "left": "issue:41720", "right": "issue:44155", "accept": false, - "reason": "Distinct subsystems and symptoms: Qwen3 device mapping crash vs AudioFlamingo3 batched inference leakage." + "reason": "Different domains: Qwen3 device-mapping crash vs AudioFlamingo3 batched inference leak." }, { "left": "issue:43873", "right": "issue:44387", "accept": false, - "reason": "Related to memory pressure, but one is quantization offloading and the other is int4 CUDA reserved memory/OOM; not duplicate enough." + "reason": "Both are quantization/memory issues, but they describe different regressions and code paths." }, { "left": "issue:39692", "right": "issue:44960", "accept": false, - "reason": "Different models and problem types; one is SigLIP2 docs/model mismatch, the other is GLM5." + "reason": "SigLIP2 docs/inference issue is unrelated to GLM5." }, { "left": "issue:43425", "right": "issue:43957", "accept": false, - "reason": "Both are version-compatibility regressions, but one is torch 2.10 and the other is `torch.device(\"meta\")` model loading; separate causes." + "reason": "Torch version incompatibility is unrelated to meta-device loading failures." }, { "left": "issue:44387", "right": "issue:44756", "accept": false, - "reason": "Both mention memory, but one is quantization OOM and the other is Strix Halo mmap behavior; unrelated." + "reason": "Different OOM scenarios with different proposed fixes; not the same defect." }, { "left": "issue:43854", "right": "issue:45216", "accept": false, - "reason": "Different models and different failures: GLM-4.7-Flash test loading vs Qwen3.5 checkpoint saving." + "reason": "Different models and different regressions: load failure in tests vs incorrect save_pretrained checkpoint." }, { - "left": "issue:44292", - "right": "issue:44912", + "left": "issue:42947", + "right": "issue:43381", "accept": false, - "reason": "Both are quantization-related, but on different models and with different failure modes; not the same change." + "reason": "Both mention gradient checkpointing, but one is inefficacy under LoRA and the other is eval-mode prohibition." }, { - "left": "issue:42947", - "right": "issue:43381", + "left": "issue:44292", + "right": "issue:44912", "accept": false, - "reason": "Both involve gradient checkpointing, but one is PEFT/LoRA effectiveness and the other is eval-mode usage; different bugs." + "reason": "Different quantization formats and different loading failures; not mergeable as one bug." }, { "left": "issue:44291", "right": "issue:44756", "accept": false, - "reason": "No meaningful overlap beyond being regressions; `_is_hf_initialized` loading error is unrelated to mmap/OOM." + "reason": "Unexpected `_is_hf_initialized` argument is unrelated to mmap/OOM on Strix Halo." }, { "left": "issue:44743", "right": "issue:44960", "accept": false, - "reason": "Different model implementations and different runtime behavior bugs." + "reason": "Recurrent-state cache bug vs GLM5 issue; unrelated." }, { "left": "issue:45276", "right": "issue:45292", "accept": true, - "reason": "Same `resize_token_embeddings` bug; 45292 is the general report and 45276 is a Gemma4-specific instance." - }, - { - "left": "issue:43381", - "right": "issue:44928", - "accept": false, - "reason": "Both concern checkpointing/sequence handling, but one is eval-mode checkpointing and the other is Qwen3.5 3D position_ids/SDPA behavior." - }, - { - "left": "issue:43957", - "right": "issue:44756", - "accept": false, - "reason": "Different loading path problems with no shared underlying defect." - }, - { - "left": "issue:44164", - "right": "issue:44756", - "accept": false, - "reason": "Save/from_pretrained extra_state handling is unrelated to mmap/OOM on Strix Halo." - }, - { - "left": "issue:43646", - "right": "issue:43708", - "accept": false, - "reason": "Different initialization/resume bugs; no shared code-path evidence." - }, - { - "left": "issue:45292", - "right": "issue:45335", - "accept": true, - "reason": "Same `resize_token_embeddings` propagation issue; 45335 is a t5gemma-specific manifestation." + "reason": "Same `resize_token_embeddings` propagation bug, with `output_embeddings` called out in both reports." } ] }, "evaluator_result": { "accept": true, - "feedback": "The cluster summary is well grounded in the packet. The resize_token_embeddings group (45276, 45292, 45335) is supported by the titles and the chosen anchor/canonical issue is conservative. The other soft-pair rejections also stay on the cautious side and do not overclaim duplication from thematic similarity alone." + "feedback": "The summary is grounded in the packet and stays conservative on the soft edges. The two accepted duplicates around `resize_token_embeddings` are well-supported by the issue titles and similarity, and the other pairwise rejections do not overclaim shared root cause. No mergeability problems for PR pairs here." }, "evaluator_used": true, "retried": false, @@ -1960,8 +1918,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -2022,8 +1980,8 @@ "issue:43425|issue:43828", "issue:43828|issue:43957", "issue:44589|issue:44756", - "issue:43906|issue:44530", "issue:43931|issue:44863", + "issue:43906|issue:44530", "issue:43611|issue:43645", "issue:39404|issue:44912", "issue:43425|issue:43716", @@ -2041,8 +1999,8 @@ "issue:43596|issue:44683", "issue:44589|issue:45084", "issue:36032|issue:43452", - "issue:43582|issue:43632", "issue:44991|issue:45081", + "issue:43582|issue:43632", "issue:32090|issue:39404", "issue:43957|issue:44164", "issue:40990|issue:43421", @@ -2052,250 +2010,244 @@ "issue:43425|issue:45161", "issue:43611|issue:43644" ], - "prepared_review_unit_hash": "6c8a8cbe78b171370d188bf37d57edaee6e99a2441cc440a6490ebe9ec7989a3", + "prepared_review_unit_hash": "5b5edd9395ff00742810924d51557e560100c9789e579dbe13df3b617f3446d0", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 36, - "serialized_chars": 38488, - "estimated_input_tokens": 9622, - "estimated_eval_tokens": 19500 + "serialized_chars": 38484, + "estimated_input_tokens": 9621, + "estimated_eval_tokens": 19498 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose collection of Transformers v5 regressions, especially model/tokenizer loading problems, but the soft pairs are not the same concrete bugs. No soft merges are warranted.", - "confidence": 0.72, - "canonical_issue_reason": "issue:43611 is the broadest and most representative loading-regression report in the set: it describes a v5 break affecting model loading via `base_model_prefix`, which aligns with the overall theme of the cluster better than the narrower model-specific issues.", + "summary": "This cluster is a loose grab-bag of Transformers v5 regressions, mostly around model loading, tokenizers, and a few unrelated runtime/docs bugs. The most representative issue is the broad v5 model-loading regression around `base_model_prefix`.", + "confidence": 0.61, + "canonical_issue_reason": "Issue 43611 is the broadest, most representative v5 regression in the set: it directly summarizes a general model-loading breakage that overlaps the cluster\u2019s dominant theme.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43611 is the best single issue to anchor this cluster because it is general, user-facing, and closest to the common failure mode across the nearby reports (loading regressions after v5).", + "best_issue_reason": "43611 is the best cluster anchor because its scope is general enough to represent the recurring Transformers v5 loading failures without being tied to one specific model family.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44479", "right": "issue:44877", "accept": false, - "reason": "Both are v5 regressions, but one is a video-input pipeline issue and the other is strict config handling; different code paths and failure modes." + "reason": "Both are v5 regressions, but one is a video-input bug for Qwen VL models and the other is a config-loading issue for granite_speech; different code paths." }, { "left": "issue:44402", "right": "issue:44451", "accept": false, - "reason": "One is tokenizer vocab boundary metadata, the other is a model-load failure for a specific checkpoint; not the same bug." + "reason": "Both involve tokenizer/model loading failures, but the symptoms and likely causes differ: vocab-size mismatch vs inability to load a specific tokenizer." }, { "left": "issue:43645", "right": "issue:44164", "accept": false, - "reason": "Custom model initialization in notebooks and `extra_state` save/load handling are unrelated." + "reason": "Custom-model notebook initialization and save/from_pretrained extra_state handling are different loading/saving problems." }, { "left": "issue:43257", "right": "issue:44863", "accept": false, - "reason": "Both are loading failures, but for different model families and different mechanisms (`accelerate+deepspeed` conversion vs NemotronH checkpoint compatibility)." + "reason": "Different model families and failure modes: Qwen3 MoE conversion with accelerate/deepspeed vs NemotronH checkpoint loading." }, { "left": "issue:43425", "right": "issue:43828", "accept": false, - "reason": "Torch version incompatibility and autocast dtype mismatch are distinct problems." + "reason": "Torch-version incompatibility is not the same as an autocast dtype mismatch in a specific model." }, { "left": "issue:43828", "right": "issue:43957", "accept": false, - "reason": "Autocast dtype mismatch is not the same as meta-device loading breakage." + "reason": "One is an autocast dtype mismatch; the other is meta-device loading breakage. Related only at a high level." }, { "left": "issue:44589", "right": "issue:44756", "accept": false, - "reason": "Float8 storage lookup error and Apple Silicon mmap OOM are unrelated." + "reason": "Storage lookup failure and mmap/OOM behavior are unrelated bugs." }, { - "left": "issue:43906", - "right": "issue:44530", + "left": "issue:43931", + "right": "issue:44863", "accept": false, - "reason": "An isolated reproduction of an old issue is not the same as the Qwen3.5 PagedAttentionCache crash." + "reason": "Both are model-loading errors, but they concern different architectures and likely different conversion/shape issues." }, { - "left": "issue:43931", - "right": "issue:44863", + "left": "issue:43906", + "right": "issue:44530", "accept": false, - "reason": "Different models and different failure signatures: Qwen3-VL shape mismatch vs NemotronH checkpoint loading." + "reason": "A reproduction of an older issue and a PagedAttentionCache crash are not the same underlying bug." }, { "left": "issue:43611", "right": "issue:43645", "accept": false, - "reason": "Both involve model loading, but one is a `base_model_prefix` regression and the other is notebook-defined custom model initialization; not one concrete bug." + "reason": "Both are v5 regressions, but one is about `base_model_prefix` loading and the other about notebook-defined custom models." }, { "left": "issue:39404", "right": "issue:44912", "accept": false, - "reason": "Whisper `return_language` pipeline regression and GPT-OSS MXFP4 quantization fallback are unrelated." + "reason": "Whisper pipeline language return and GPT-OSS MXFP4 quantization fallback are unrelated." }, { "left": "issue:43425", "right": "issue:43716", "accept": false, - "reason": "General torch compatibility and Mistral image dtype mismatch are different bugs." + "reason": "Torch compatibility and image-preprocessor/model dtype mismatch are distinct issues." }, { "left": "issue:44530", "right": "issue:45084", "accept": false, - "reason": "`PagedAttentionCache` group-type crash and template compilation failure are separate code paths." + "reason": "PagedAttentionCache group-type handling and template-node compilation errors are different failures." }, { "left": "issue:43901", "right": "issue:44509", "accept": false, - "reason": "Both are docs issues, but they cover different pipeline/task deprecations and are not duplicates." + "reason": "Both are docs-related, but they target different APIs and removals." }, { "left": "issue:43299", "right": "issue:43931", "accept": false, - "reason": "Both mention Qwen3-VL-related loading, but one is a general MoE loading breakage and the other is a checkpoint shape mismatch; too different to merge." + "reason": "Both mention Qwen3-VL loading, but one is about MoE weight conversion and the other about weight-shape mismatch; not enough to treat as the same bug." }, { "left": "issue:44265", "right": "issue:44898", "accept": false, - "reason": "Torch export with `torch_compilable_check` and Perceiver interpolation failure are unrelated." + "reason": "Torch export with `torch_compilable_check` and Perceiver non-default resolution loading are different code paths." }, { "left": "issue:44479", "right": "issue:44805", "accept": false, - "reason": "Video-input regression and mask shape mismatch are not the same underlying bug." + "reason": "A video-input regression and a mask shape IndexError are unrelated." }, { "left": "issue:43994", "right": "issue:44079", "accept": false, - "reason": "Nonsensical model outputs and `ModelOutput` key assignment with `None` are unrelated." + "reason": "Incorrect SigLIP2 outputs and `ModelOutput` key assignment are different problems." }, { "left": "issue:44671", "right": "issue:44991", "accept": false, - "reason": "CamemBERT masked-LM prediction errors and tokenizer loading failure are different issues." + "reason": "CamemBERT masked-LM prediction errors and tokenizer loading failures for a different model are not the same bug." }, { "left": "issue:43856", "right": "issue:44756", "accept": false, - "reason": "Training memory inefficiency and Strix Halo mmap OOM are separate performance/memory problems." + "reason": "Memory inefficiency in Qwen3 MoE training is unrelated to Strix Halo mmap OOM avoidance." }, { "left": "issue:43425", "right": "issue:44918", "accept": false, - "reason": "Torch version compatibility and Qwen3.5 embedding unpacking are different failure modes." + "reason": "Torch version incompatibility and TRL SFT embedding unpacking are different failures." }, { "left": "issue:39404", "right": "issue:44671", "accept": false, - "reason": "Whisper pipeline behavior and CamemBERT MLM outputs are unrelated." + "reason": "Whisper return_language and CamemBERT masked-LM regressions are unrelated model-specific bugs." }, { "left": "issue:40990", "right": "issue:42947", "accept": false, - "reason": "High perplexity on a model and gradient checkpointing not taking effect are distinct issues." + "reason": "High perplexity on GPT-OSS and ineffective gradient checkpointing with LoRA are not the same issue." }, { "left": "issue:43596", "right": "issue:44683", "accept": false, - "reason": "DeepSpeed zero3 init index error and compiled flex_attention failure on newer torch versions are unrelated." + "reason": "DeepSpeed ZeRO-3 initialization IndexError and compiled flex_attention on torch>=2.9 are unrelated." }, { - "left": "issue:44589", - "right": "issue:45084", + "left": "issue:43883", + "right": "issue:44488", "accept": false, - "reason": "Missing Float8 storage and template-node compilation errors are different bugs." + "reason": "Missing `all_tied_weights_keys` on Molmo and loading `cjvt/sleng-bert` are different model-loading problems." }, { - "left": "issue:36032", - "right": "issue:43452", + "left": "issue:44488", + "right": "issue:45081", "accept": false, - "reason": "Tokenizer special-token method conflict and `gguf_file` loading breakage are not the same problem." + "reason": "Tokenizer load failure for `sleng-bert` and the Mistral regex patch crash are distinct tokenizer bugs." }, { - "left": "issue:43582", - "right": "issue:43632", + "left": "issue:36032", + "right": "issue:43452", "accept": false, - "reason": "Apple Silicon warmup TypeError and `_is_hf_initialized` flag breakage are different runtime issues." + "reason": "T5 tokenizer special-token naming conflict and `gguf_file` loading breakage are unrelated." }, { "left": "issue:44991", "right": "issue:45081", "accept": false, - "reason": "Tokenizer loading failure and `_patch_mistral_regex` backend-tokenizer crash are related to tokenizers, but they are different concrete failures." + "reason": "Both are tokenizer regressions, but they affect different tokenizers and likely different fix paths." + }, + { + "left": "issue:43582", + "right": "issue:43632", + "accept": false, + "reason": "Apple Silicon allocator warmup TypeError and `_is_hf_initialized` breakage are different runtime issues." }, { "left": "issue:32090", "right": "issue:39404", "accept": false, - "reason": "Trainer GPU broadcast `NoneType` error and Whisper `return_language` regression are unrelated." + "reason": "Trainer GPU broadcast TypeError and Whisper pipeline language return are unrelated." }, { "left": "issue:43957", "right": "issue:44164", "accept": false, - "reason": "Meta-device loading and `extra_state` save/load are separate mechanisms." + "reason": "Meta-device loading issues and extra_state save/load handling are distinct code paths." }, { "left": "issue:40990", "right": "issue:43421", "accept": false, - "reason": "Perplexity regression and runtime special-token post-processor update are unrelated." + "reason": "Perplexity anomaly and runtime special-token post-processor updates are unrelated." }, { "left": "issue:43883", "right": "issue:44488", "accept": false, - "reason": "Missing `all_tied_weights_keys` and inability to load a specific BERT checkpoint are different issues." - }, - { - "left": "issue:44164", - "right": "issue:44479", - "accept": false, - "reason": "`extra_state` serialization and video-input regression are unrelated." - }, - { - "left": "issue:44488", - "right": "issue:45081", - "accept": false, - "reason": "Both involve loading/tokenizer behavior, but one is a model-loading failure and the other is a tokenizer regex patch crash; not the same bug." + "reason": "These are separate tokenizer/model-loading failures with no clear shared root cause." }, { - "left": "issue:43425", - "right": "issue:45161", + "left": "issue:44589", + "right": "issue:45084", "accept": false, - "reason": "Torch compatibility and GPT-OSS tensor-parallel loading are different problems." + "reason": "Float8 storage lookup and template-node compilation errors are unrelated." }, { "left": "issue:43611", "right": "issue:43644", "accept": false, - "reason": "Both are v5 model-loading regressions, but they target different failures: `base_model_prefix` loading vs non-persistent buffer initialization." + "reason": "Both are broad v5 regressions, but one is about `base_model_prefix` and the other about non-persistent buffer initialization; not the same bug." } ] }, @@ -2314,8 +2266,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -2372,14 +2324,14 @@ "issue:44811|issue:45584", "issue:44206|issue:44220", "issue:44488|issue:44987", - "issue:42947|issue:44756", "issue:43716|issue:45084", "issue:34567|issue:40444", "issue:43883|issue:44991", + "issue:42947|issue:44756", "issue:43644|issue:43646", "issue:43643|issue:44530", - "issue:43452|issue:43482", "issue:44451|issue:44987", + "issue:43452|issue:43482", "issue:43317|issue:43482", "issue:36032|issue:42915", "issue:43541|issue:45084", @@ -2390,8 +2342,8 @@ "issue:43262|issue:43377", "issue:36032|issue:43645", "issue:44987|issue:44991", - "issue:43646|issue:44534", "issue:45084|issue:45507", + "issue:43646|issue:44534", "issue:43994|issue:44610", "issue:43645|issue:44534", "issue:45216|issue:45357", @@ -2402,232 +2354,232 @@ "issue:41553|issue:43582", "issue:44220|issue:44683" ], - "prepared_review_unit_hash": "f862802ba83fc25a8eb496eedd17212f4cf94f26fb57cef28ecbbfbc2e1ae789", + "prepared_review_unit_hash": "a5b18d7e3de6320ee75afcbb76ef1a71fb1ecb2dd09f23756acea42d923f91e2", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 37202, - "estimated_input_tokens": 9301, - "estimated_eval_tokens": 18858 + "serialized_chars": 37198, + "estimated_input_tokens": 9300, + "estimated_eval_tokens": 18856 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a heterogeneous cluster of unrelated Transformers regressions; only a couple of pairs look like true duplicates, centered on GGUF loading and Qwen3.5 save_pretrained serialization.", - "confidence": 0.67, - "canonical_issue_reason": "Issue 45216 is the cleanest anchor: it states the Qwen3.5 save_pretrained regression directly, and 45357 reads like a follow-up/variant of the same checkpoint-serialization bug.", + "summary": "This set is mostly heterogeneous, with many unrelated Transformer bugs. The only clear duplicate-like pair is the Qwen3.5 save_pretrained regression (#45216 / #45357); the rest are too model- or symptom-specific to merge together.", + "confidence": 0.82, + "canonical_issue_reason": "No single canonical issue across the full set: most items describe unrelated bugs. If a duplicate anchor is needed, #45216 is the best one because #45357 is a later, narrower follow-up of the same Qwen3.5 save_pretrained regression.", "canonical_pr_reason": null, - "best_issue_reason": "45216 is the best representative issue because it is concrete, specific to the underlying regression, and has the clearest near-duplicate in 45357.", + "best_issue_reason": "#45216 is the best representative issue for the only strong duplicate pair, since it is earlier and describes the broader regression; #45357 is a narrower persistence of the same problem.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44811", "right": "issue:45584", "accept": false, - "reason": "Both are Whisper-related, but one is about batch_decode skip_special_tokens handling and the other about empty-transcription generation after align_special_tokens; different failure modes." + "reason": "Both are Whisper-related, but one is batch_decode skip-special-token behavior and the other is empty-transcription generation after align_special_tokens; different failure modes." }, { "left": "issue:44206", "right": "issue:44220", "accept": false, - "reason": "Both touch audio feature extraction, but the titles point to different bugs: an unsupported center argument vs a separate _torch_extract_fbank_features() issue." + "reason": "Both touch audio feature extraction, but one is an unsupported `center` arg in `LasrFeatureExtractor` and the other is a low-level `_torch_extract_fbank_features()` issue; not the same bug." }, { "left": "issue:44488", "right": "issue:44987", "accept": false, - "reason": "Both are model-loading failures, but they affect different models and describe different regressions; too broad to be the same bug." - }, - { - "left": "issue:42947", - "right": "issue:44756", - "accept": false, - "reason": "Different areas entirely: LoRA/gradient checkpointing behavior vs Apple Silicon mmap/OOM handling." + "reason": "Both are loading regressions, but they involve different models and likely different code paths (`cjvt/sleng-bert` vs `physical-intelligence/fast`)." }, { "left": "issue:43716", "right": "issue:45084", "accept": false, - "reason": "Mistral image dtype mismatch is unrelated to a template compilation TypeError." + "reason": "Completely different problems: Mistral-3 image preprocessor dtype mismatch vs a template compilation TypeError." }, { "left": "issue:34567", "right": "issue:40444", "accept": false, - "reason": "TrainerState token counting and Qwen2.5-VL multi-image IterableDataset failures are unrelated." + "reason": "Unrelated: Trainer token accounting vs Qwen2.5-VL iterable dataset multi-image loading failure." }, { "left": "issue:43883", "right": "issue:44991", "accept": false, - "reason": "Different model/tokenizer-loading regressions with different root symptoms; not the same concrete bug." + "reason": "Both are loading-related, but one is a missing `all_tied_weights_keys` attribute and the other is a tokenizer load failure for a different model; not the same underlying bug." + }, + { + "left": "issue:42947", + "right": "issue:44756", + "accept": false, + "reason": "Different subsystems and symptoms: LoRA gradient checkpointing vs Apple Silicon mmap/OOM handling." }, { "left": "issue:43644", "right": "issue:43646", "accept": false, - "reason": "Both are v5 regression reports, but one is about non-persistent buffers and the other about custom model initialization; too generic to merge." + "reason": "Both are Transformers 5 breakages, but one is non-persistent buffer initialization and the other is custom model initialization; too broad and not the same defect." }, { "left": "issue:43643", "right": "issue:44530", "accept": false, - "reason": "AutoConfig trust_remote_code field loss is unrelated to Qwen3.5 PagedAttentionCache crashes." - }, - { - "left": "issue:43452", - "right": "issue:43482", - "accept": true, - "reason": "Both point to the same GGUF/from_pretrained loading regression in the v5 loader path; one is the generic API break, the other a specific Qwen2.5-GGUF manifestation." + "reason": "Different issues: missing fields from `AutoConfig.from_pretrained(trust_remote_code=True)` vs PagedAttentionCache group-type handling on Qwen3.5." }, { "left": "issue:44451", "right": "issue:44987", "accept": false, - "reason": "Both are loading failures, but they are for different repositories/models and do not indicate the same code-path bug." + "reason": "Both fail on model loading, but they are different models with no evidence of a shared code-path bug." + }, + { + "left": "issue:43452", + "right": "issue:43482", + "accept": false, + "reason": "Both mention GGUF, but one is a broad `gguf_file` API break and the other is a model-specific v5 loading failure; not enough evidence they are the same bug." }, { "left": "issue:43317", "right": "issue:43482", "accept": false, - "reason": "Device-map offload/dequantized-model loading is a different failure mode from GGUF loading in v5." + "reason": "Different loading paths: `device_map=auto` offload/dequantization vs GGUF model loading." }, { "left": "issue:36032", "right": "issue:42915", "accept": false, - "reason": "T5 tokenizer special-token method conflict and Qwen3Moe FineGrainedFP8Config failure are unrelated." + "reason": "Unrelated: T5 tokenizer method conflict vs Qwen3Moe FP8 configuration failure." }, { "left": "issue:43541", "right": "issue:45084", "accept": false, - "reason": "Torch Dynamo grouped_mm tracing failure is unrelated to the template-node compilation error." + "reason": "Different compilation failures with no shared code path." }, { "left": "issue:39404", "right": "issue:43825", "accept": false, - "reason": "Both involve pipeline behavior, but one is a Whisper return_language regression and the other is an error-message/docs issue about translation support." + "reason": "Both involve pipeline/Whisper wording, but one is a functional `return_language` regression and the other is an error-message issue about unsupported translation tasks." }, { "left": "issue:44849", "right": "issue:45325", "accept": false, - "reason": "Different Qwen model families and different bugs: hidden-states output vs rope index scaling." + "reason": "Different Qwen variants and different bugs: hidden-states behavior vs image temporal position-id scaling." }, { "left": "issue:42915", "right": "issue:44530", "accept": false, - "reason": "Qwen3Moe FP8 configuration failure and Qwen3.5 PagedAttentionCache linear_attention crash are different issues." + "reason": "Both mention Qwen, but one is FP8 config failure and the other is an invalid cache group type; not the same concrete bug." }, { "left": "issue:43404", "right": "issue:44466", "accept": false, - "reason": "Both concern tied weights, but one is an untied lm_head bug in Mistral3ForConditionalGeneration and the other is a device-dependent save_pretrained serialization issue." + "reason": "Related to tied weights, but one is a missing tie in `Mistral3ForConditionalGeneration` and the other is inconsistent serialization by device; separate issues." }, { "left": "issue:43262", "right": "issue:43377", "accept": false, - "reason": "Audio chat-template sampling-rate defaults and MIMI padding-mask batching differences are separate bugs." + "reason": "Both are audio-processing issues, but one is chat-template sampling-rate defaults and the other is missing padding-mask support in MIMI batching." }, { "left": "issue:36032", "right": "issue:43645", "accept": false, - "reason": "Tokenizer method conflict and custom model initialization in notebooks are not the same bug." + "reason": "Different Transformers 5 regressions: tokenizer method name collision vs custom model initialization in notebooks." }, { "left": "issue:44987", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer/model loading regressions, but they affect different models and likely different loader paths." + "reason": "Both are v5 loading regressions, but they concern different models and different failure modes; no shared underlying bug is evident." }, { - "left": "issue:43646", - "right": "issue:44534", + "left": "issue:45084", + "right": "issue:45507", "accept": false, - "reason": "Custom model initialization and non-persistent buffer junk are related by theme, but the titles describe different concrete regressions." + "reason": "Different template/model-method failures with no concrete overlap." }, { - "left": "issue:45084", - "right": "issue:45507", + "left": "issue:43646", + "right": "issue:44534", "accept": false, - "reason": "Template compilation failure and GraniteMoEHybrid invalid-method calls are unrelated." + "reason": "Both mention custom model/init behavior, but one is generic custom model initialization and the other is non-persistent buffer serialization junk." }, { "left": "issue:43994", "right": "issue:44610", "accept": false, - "reason": "Both involve model/processor mismatch, but they are for different models and different preprocessing expectations." + "reason": "Both are vision-related, but one is bad predictions from SigLIP2 and the other is a processor size mismatch for OmDet-Turbo." }, { "left": "issue:43645", "right": "issue:44534", "accept": false, - "reason": "Custom notebook model initialization and v5 non-persistent buffer junk are not the same concrete bug." + "reason": "Different regressions: notebook custom-model initialization vs junk-filled non-persistent buffers." }, { "left": "issue:45216", "right": "issue:45357", "accept": true, - "reason": "Same Qwen3.5 save_pretrained regression: both report incorrect checkpoint serialization, with 45357 narrowing it to visual encoder keys." + "reason": "Same underlying Qwen3.5 `save_pretrained` regression; #45357 is a later, narrower report that still saves incorrect visual encoder keys." }, { "left": "issue:43322", "right": "issue:43540", "accept": false, - "reason": "Llava Next loading segfault and Qwen3OmniMoe video-input ValueError are different model-specific failures." + "reason": "Both affect multimodal loading/processing, but one is a Llava Next segfault and the other is a Qwen3OmniMoe video-input ValueError." }, { "left": "issue:43377", "right": "issue:43819", "accept": false, - "reason": "MIMI batch-vs-single padding-mask discrepancy is unrelated to DAC.from_latents missing STE behavior." + "reason": "Different audio-model bugs: missing padding-mask support in MIMI encoder vs DAC latent/forward mismatch without STE." }, { "left": "issue:44403", "right": "issue:45081", "accept": false, - "reason": "Generic loading noise and a Mistral regex patch AttributeError are different issues." + "reason": "One is generic noisy model-loading output, the other is a specific Mistral tokenizer regex patch crash." }, { "left": "issue:44451", "right": "issue:45081", "accept": false, - "reason": "Different model-loading failures with no shared concrete bug path." + "reason": "Both are tokenizer/model-loading failures, but for different models and different root causes." }, { "left": "issue:41553", "right": "issue:43582", "accept": false, - "reason": "Bad AutoTokenizer error message for Voxtral and Apple Silicon caching_allocator_warmup TypeError are unrelated." + "reason": "Unrelated: bad AutoTokenizer error message for Voxtral vs Apple Silicon caching_allocator_warmup TypeError." }, { "left": "issue:44220", "right": "issue:44683", "accept": false, - "reason": "Audio fbank feature extraction and compiled flex_attention on torch>=2.9 are unrelated." + "reason": "Different low-level failures: audio fbank extraction vs compiled flex_attention on newer torch." } ] }, "evaluator_result": { "accept": true, - "feedback": "The cluster summary is grounded in the packet and stays conservative. The two accepted pairs are well supported by their titles alone: 43452/43482 both point to the same GGUF/v5 loading regression, and 45216/45357 both describe the same Qwen3.5 save_pretrained serialization bug. The remaining pairwise rejections avoid overclaiming and generally distinguish different models or failure modes appropriately." + "feedback": "The summary is grounded in the packet and stays appropriately conservative. The only accepted duplicate-like pair, #45216/#45357, is plausibly the same Qwen3.5 save_pretrained regression and looks mergeable into one bugfix PR. The other soft-edge pairs are rejected with cautious, title-based reasoning rather than overclaiming evidence." }, "evaluator_used": true, "retried": false, @@ -2643,8 +2595,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -2702,24 +2654,24 @@ "issue:43404|issue:45127", "issue:42915|issue:45084", "issue:44987|issue:45081", - "issue:43296|issue:43531", "issue:44479|issue:45325", + "issue:43296|issue:43531", "issue:44062|issue:44987", "issue:42915|issue:44291", "issue:40990|issue:44811", - "issue:42175|issue:45542", - "issue:43317|issue:43856", "issue:43611|issue:44534", + "issue:43317|issue:43856", + "issue:42175|issue:45542", "issue:36032|issue:43742", "issue:40444|issue:45325", "issue:43632|issue:44534", "issue:43421|issue:44568", "issue:43818|issue:45357", - "issue:42915|issue:44589", "issue:43531|issue:44877", + "issue:42915|issue:44589", "issue:36032|issue:41720", - "issue:44683|issue:45020", "issue:43425|issue:44403", + "issue:44683|issue:45020", "issue:43482|issue:45538", "issue:44829|issue:45479", "issue:33666|issue:40444", @@ -2728,214 +2680,214 @@ "issue:44079|issue:44534", "issue:35707|issue:40444" ], - "prepared_review_unit_hash": "d7ffc5c6c082c02c1c6e3fd7228f2baa7b9c0bb8dfe5b48c0d0dbe0f87802dba", + "prepared_review_unit_hash": "a28d5f7018859e92f8f51d63cf5adfe56345c1daa2ed96da09f5699a5c13ee10", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 30, - "serialized_chars": 36158, - "estimated_input_tokens": 9040, - "estimated_eval_tokens": 18336 + "serialized_chars": 36162, + "estimated_input_tokens": 9041, + "estimated_eval_tokens": 18338 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is highly heterogeneous: the suggested pairs share at most broad subsystem vocabulary (tokenizers, Qwen, v5, FP8, etc.) but not the same concrete bug or code path. I don\u2019t see a true duplicate cluster here, so no canonical issue/PR stands out.", - "confidence": 0.96, - "canonical_issue_reason": null, + "summary": "This cluster is heterogeneous: the soft-pair candidates mostly share only broad subsystem or version context, not the same concrete bug or fix path. I reject all soft edges and see no clear canonical duplicate representative.", + "confidence": 0.97, + "canonical_issue_reason": "No single issue clearly represents the cluster; the items span unrelated bugs across tokenizers, Qwen/VL, packaging, loading, and attention behavior.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "No issue is a strong global representative. The cluster does not appear to center on one underlying defect, so a canonical issue choice would be arbitrary.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43632", "right": "issue:44756", "accept": false, - "reason": "Different problems: HF init flag regression vs Strix Halo mmap/OOM handling." + "reason": "Unrelated: HF init flag regression vs mmap/OOM on Strix Halo." }, { "left": "issue:44265", "right": "issue:44610", "accept": false, - "reason": "torch.export/torch_compilable_check is unrelated to OmDet-Turbo input-size mismatch." + "reason": "Different bugs: torch export/check utility failure vs OmDet-Turbo processor size mismatch." }, { "left": "issue:43404", "right": "issue:45127", "accept": false, - "reason": "Both mention weight tying, but one is missing tie in Mistral3 and the other is LoRA merge collapse; different bugs." + "reason": "Distinct causes: lm_head weight tying bug vs LoRA merge with extended vocabulary collapse." }, { "left": "issue:42915", "right": "issue:45084", "accept": false, - "reason": "Qwen3Moe FP8 config failure and template-node compilation error are unrelated failure modes." + "reason": "FP8 config failure and template-node compilation error are unrelated code paths." }, { "left": "issue:44987", "right": "issue:45081", "accept": false, - "reason": "Different models and code paths: physical-intelligence/fast loading vs Mistral regex patch tokenizer crash." + "reason": "Both are loading-related, but one is remote-code/model loading and the other is a Mistral regex patch crash." }, { - "left": "issue:43296", - "right": "issue:43531", + "left": "issue:44479", + "right": "issue:45325", "accept": false, - "reason": "PaddleOCR-VL load failure and Qwen3-MoE sliding_window bug are unrelated." + "reason": "Same model family, but video-input regression and still-image rope-index scaling are different symptoms and likely different fixes." }, { - "left": "issue:44479", - "right": "issue:45325", + "left": "issue:43296", + "right": "issue:43531", "accept": false, - "reason": "Both are Qwen VL regressions, but one is a broad video-input regression and the other is a specific still-image RoPE scaling bug." + "reason": "Different model families and failure modes: PaddleOCR-VL loading vs Qwen3-MoE sliding_window behavior." }, { "left": "issue:44062", "right": "issue:44987", "accept": false, - "reason": "AddedToken special-argument conflict is a tokenizer API bug, not the same as the physical-intelligence/fast loading failure." + "reason": "Tokenizer AddedToken argument conflict is unrelated to loading a remote-code model." }, { "left": "issue:42915", "right": "issue:44291", "accept": false, - "reason": "Both involve loading in v5, but FP8 config failure and _is_hf_initialized/init_empty_weights TypeError are distinct issues." + "reason": "Both touch transformers v5-era loading, but FP8 config failure and _is_hf_initialized/init_empty_weights mismatch are different bugs." }, { "left": "issue:40990", "right": "issue:44811", "accept": false, - "reason": "Perplexity evaluation issue is unrelated to Whisper batch_decode skip_special_tokens behavior." + "reason": "Perplexity/eval issue vs Whisper batch_decode special-token handling; no shared bug." }, { - "left": "issue:42175", - "right": "issue:45542", + "left": "issue:43611", + "right": "issue:44534", "accept": false, - "reason": "Both concern TensorFlow packaging, but one is missing tf backend in torch extra and the other is a tensorboard-only backend error." + "reason": "base_model_prefix loading breakage is separate from non-persistent buffer initialization junk." }, { "left": "issue:43317", "right": "issue:43856", "accept": false, - "reason": "GPU/CPU offload loading failure and Qwen3 MoE memory inefficiency are different problems." + "reason": "Loading with gpu+cpu offload is unrelated to training-time memory inefficiency." }, { - "left": "issue:43611", - "right": "issue:44534", + "left": "issue:42175", + "right": "issue:45542", "accept": false, - "reason": "Both are v5 regressions, but base_model_prefix loading and non-persistent buffer corruption are separate code paths." + "reason": "Both are TensorFlow packaging/backend issues, but one is missing TF in the torch extra and the other is an undefined backend error with tensorboard-only install." }, { "left": "issue:36032", "right": "issue:43742", "accept": false, - "reason": "Tokenizer special-token conflict and MobileLLM-125M key error are unrelated load failures." + "reason": "Tokenizer special-token conflict vs model key error; different underlying failures." }, { "left": "issue:40444", "right": "issue:45325", "accept": false, - "reason": "Different Qwen2.5-VL bugs: multi-image iterable finetuning vs RoPE temporal-position scaling." + "reason": "Both involve Qwen VL, but iterable-dataset multi-image finetuning and rope-index scaling for still images are not the same bug." }, { "left": "issue:43632", "right": "issue:44534", "accept": false, - "reason": "HF initialization flag regression is not the same as junk-filled non-persistent buffers." + "reason": "Unrelated: _is_hf_initialized flag regression vs non-persistent buffer junk in v5." }, { "left": "issue:43421", "right": "issue:44568", "accept": false, - "reason": "Special-token runtime update bug and mdeberta-v3 BOS/EOS insertion bug are related only superficially." + "reason": "Shared special-token theme, but runtime post-processor updates and add_special_tokens=True regression are different surfaces and not clearly one fix." }, { "left": "issue:43818", "right": "issue:45357", "accept": false, - "reason": "Video-LLaVA weight/temporal-attention issue is unrelated to Qwen3.5 save_pretrained key mismatch." + "reason": "Video-LLaVA tower/temporal-attention issue is unrelated to Qwen3.5 save_pretrained visual-key regression." }, { - "left": "issue:42915", - "right": "issue:44589", + "left": "issue:43531", + "right": "issue:44877", "accept": false, - "reason": "FP8 config failure and missing Float8 storage type are different FP8-related bugs, not the same defect." + "reason": "Qwen3-MoE sliding_window behavior and granite_speech strict config loading are unrelated." }, { - "left": "issue:43531", - "right": "issue:44877", + "left": "issue:42915", + "right": "issue:44589", "accept": false, - "reason": "Qwen3-MoE sliding_window bug and granite_speech strict-config loading are unrelated." + "reason": "FineGrainedFP8Config failure is unrelated to missing Float8 storage type lookup." }, { "left": "issue:36032", "right": "issue:41720", "accept": false, - "reason": "Tokenizer load conflict and Qwen3 auto device-map cuda assert are unrelated." + "reason": "Tokenizer load conflict vs CUDA assert during auto device mapping are different problem classes." }, { - "left": "issue:44683", - "right": "issue:45020", + "left": "issue:43425", + "right": "issue:44403", "accept": false, - "reason": "Compiled flex_attention on torch>=2.9 and remote_code loading regressions are different areas." + "reason": "Torch compatibility issue is unrelated to noisy transformer loading logs." }, { - "left": "issue:43425", - "right": "issue:44403", + "left": "issue:44683", + "right": "issue:45020", "accept": false, - "reason": "Torch version incompatibility is not the same as noisy transformer loading." + "reason": "Compiled flex_attention on new torch versions is unrelated to remote_code model breakage." }, { "left": "issue:43482", "right": "issue:45538", "accept": false, - "reason": "Qwen2.5-GGUF/v5 loading failure and CLIPTokenizer max_length bug are unrelated." + "reason": "GGUF loading with transformers v5 and CLIPTokenizer model_max_length are different defects." }, { "left": "issue:44829", "right": "issue:45479", "accept": false, - "reason": "Both involve bad training behavior, but flash_attention_3 degeneration and single-label/num_labels=1 zero-loss are distinct causes." + "reason": "Both affect training loss, but flash_attention_3 degeneration and num_labels=1 zero-loss behavior have different root causes." }, { "left": "issue:33666", "right": "issue:40444", "accept": false, - "reason": "Same broad family (Qwen VL training), but multi-GPU training and multi-image iterable finetuning are not the same bug." + "reason": "Both mention Qwen VL training, but multi-GPU training and multi-image iterable dataset failures are not the same bug." }, { "left": "issue:44403", "right": "issue:44991", "accept": false, - "reason": "Unnecessary load noise and tokenizer load failure for est-roberta are different issues." + "reason": "Generic loading noise is unrelated to tokenizer loading failure for EMBEDDIA/est-roberta." }, { "left": "issue:43792", "right": "issue:44589", "accept": false, - "reason": "Whisper runtime failure and missing Float8 storage object are unrelated." + "reason": "Whisper run failure and Float8 storage lookup error are unrelated." }, { "left": "issue:44079", "right": "issue:44534", "accept": false, - "reason": "ModelOutput key assignment and non-persistent buffer corruption are different internal bugs." + "reason": "ModelOutput key assignment and non-persistent buffer junk are separate internal behaviors." }, { "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 finetuning." + "reason": "Progressive generation with inputs_embeds/past_key_values is unrelated to Qwen2.5-VL multi-image finetuning." } ] }, @@ -2954,8 +2906,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -3015,8 +2967,8 @@ "issue:43749|issue:44079", "issue:44079|issue:45072", "issue:43644|issue:44079", - "issue:43632|issue:45072", "issue:40444|issue:45381", + "issue:43632|issue:45072", "issue:44265|issue:45072", "issue:43881|issue:44488", "issue:45325|issue:45381", @@ -3039,208 +2991,208 @@ "issue:40990|issue:45584", "issue:34634|issue:35707" ], - "prepared_review_unit_hash": "c22d20c078767bf7e4f2a00609c0508fbdd893e92cef0217ed7a6050c50b69b5", + "prepared_review_unit_hash": "8036223205b231423116aa0b4e79a355b114956e81d139dea9a4dd8db108e853", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36393, - "estimated_input_tokens": 9099, - "estimated_eval_tokens": 18454 + "serialized_chars": 36388, + "estimated_input_tokens": 9097, + "estimated_eval_tokens": 18450 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose thematic group of transformer-regression reports, not a set of true duplicates. The strongest center is around Transformers v5 loading/initialization regressions, but the paired items usually hit different code paths or model families.", - "confidence": 0.74, - "canonical_issue_reason": "Issue 43632 is the best representative because it captures a broad, core v5 initialization regression that can plausibly underlie several nearby loading bugs, while still being concrete and well-scoped.", - "canonical_pr_reason": null, - "best_issue_reason": "Issue 43632 is the best issue to keep as the cluster anchor: it is broad enough to represent the v5 loading/init breakage theme, but specific enough to remain actionable.", - "best_pr_reason": null, + "summary": "This is a heterogeneous set of mostly unrelated issues, with only a few superficial overlaps around Transformers v5 regressions, quantization/offloading, and a couple of model-family-specific bugs. None of the soft edges look like true duplicates.", + "confidence": 0.91, + "canonical_issue_reason": "No single canonical issue is safe here: the items span many distinct bugs and model families, so picking one representative would be misleading.", + "canonical_pr_reason": "No PRs are present in this cluster.", + "best_issue_reason": "There is no strong global best issue; the closest broad items are still too narrow or too different from the rest of the set to serve as a reliable representative.", + "best_pr_reason": "No PRs are present in this cluster.", "soft_edge_verdicts": [ { "left": "issue:43377", "right": "issue:45491", "accept": false, - "reason": "Both mention sequence/batching behavior, but one is MIMI encoder padding-mask inconsistency and the other is Gemma3 sliding-window NaNs; different models and different failure mechanisms." + "reason": "Different models and failure modes: batched-vs-single padding-mask mismatch vs mixed-length sliding-window NaNs." }, { "left": "issue:43482", "right": "issue:43525", "accept": false, - "reason": "Both are Transformers v5 load regressions, but one is GGUF loading and the other is a missing pad_token_id attribute on Llama4Config; distinct bugs." + "reason": "Both are load failures, but one is a Qwen2.5-GGUF/v5 loading issue and the other is a Llama4Config missing `pad_token_id`." }, { "left": "issue:44291", "right": "issue:44589", "accept": false, - "reason": "Both involve v5/init-loading breakage, but one is an unexpected _is_hf_initialized argument with init_empty_weights and the other is a missing Float8 storage type; not the same code path." + "reason": "Different code paths: unexpected `_is_hf_initialized` argument vs missing Float8 storage object." }, { "left": "issue:43749", "right": "issue:44079", "accept": false, - "reason": "FSDP CPU RAM efficient loading and ModelOutput key assignment are unrelated behaviors; only a broad v5 context overlaps." + "reason": "FSDP CPU RAM efficient loading is unrelated to `ModelOutput` key assignment behavior." }, { "left": "issue:44079", "right": "issue:45072", "accept": false, - "reason": "ModelOutput key handling and CI dtype mismatches in SwitchTransformers/TimmWrapperModel are different bugs with different symptoms." + "reason": "`ModelOutput` key handling and bf16 dtype mismatches are unrelated bugs." }, { "left": "issue:43644", "right": "issue:44079", "accept": false, - "reason": "Non-persistent buffer serialization corruption is a separate state-dict/load issue from ModelOutput key assignment." + "reason": "Non-persistent buffer serialization/junk is not the same bug as `ModelOutput` key assignment." }, { - "left": "issue:43632", - "right": "issue:45072", + "left": "issue:40444", + "right": "issue:45381", "accept": false, - "reason": "The _is_hf_initialized regression and bfloat16 dtype mismatches are different failures; they may both surface in v5 but not the same underlying bug." + "reason": "Different modalities and symptoms: multi-image finetuning failure vs Qwen2.5-VL video position-id bug." }, { - "left": "issue:40444", - "right": "issue:45381", + "left": "issue:43632", + "right": "issue:45072", "accept": false, - "reason": "One is multi-image iterable dataset finetuning for Qwen2.5-VL, the other is video vision_position_ids in Qwen2.5-VL; related model family but different input modality bug." + "reason": "`_is_hf_initialized` regression and bfloat16 inference dtype mismatch are separate issues." }, { "left": "issue:44265", "right": "issue:45072", "accept": false, - "reason": "torch.export.export with torch_compilable_check and inference dtype mismatches are unrelated issues." + "reason": "`torch.export`/`torch_compilable_check` failure is not the same as a bfloat16 dtype mismatch." }, { "left": "issue:43881", "right": "issue:44488", "accept": false, - "reason": "GLM-4V load failure and cjvt/sleng-bert tokenizer/model load failure are unrelated model-specific loading problems." + "reason": "Different models and likely different loader failures; no shared concrete bug." }, { "left": "issue:45325", "right": "issue:45381", "accept": false, - "reason": "Both are Qwen2.5-VL position-id bugs, but one is still-image temporal scaling and the other is video input vision_position_ids; related but distinct paths." + "reason": "Same model family, but still-image temporal scaling and video vision-position-id errors are different code paths." }, { "left": "issue:43531", "right": "issue:44291", "accept": false, - "reason": "Sliding_window behavior in Qwen3-MoE and init_empty_weights/_is_hf_initialized are different regression classes." + "reason": "Qwen3-MoE sliding-window behavior is unrelated to the `init_empty_weights` argument regression." }, { "left": "issue:43611", "right": "issue:43646", "accept": false, - "reason": "Base_model_prefix loading and custom model initialization are both v5 load/init regressions, but they are different concrete breakages." + "reason": "Both are Transformers v5 regressions, but one is about `base_model_prefix` loading and the other about custom model initialization; not the same concrete bug." }, { "left": "issue:41720", "right": "issue:45081", "accept": false, - "reason": "Qwen3 auto device mapping cudaErrorAssert and Mistral tokenizer regex patch crash are unrelated." + "reason": "Qwen3 auto device-map CUDA assert and Mistral tokenizer regex crash are unrelated." }, { "left": "issue:42175", "right": "issue:43296", "accept": false, - "reason": "TensorFlow backend packaging and PaddleOCR-VL model loading are different issues with no shared failure path." + "reason": "Packaging/backend selection issue is unrelated to PaddleOCR-VL loading failure." }, { "left": "issue:43454", "right": "issue:44466", "accept": false, - "reason": "AyaVision lm_head weight tying and inconsistent lm_head.weight serialization are both weight-tying related, but one is a missing tie and the other is device-dependent serialization behavior." + "reason": "Both mention tied `lm_head` weights, but one is a generation-quality/init issue and the other is device-dependent serialization; not the same fix." }, { "left": "issue:43994", "right": "issue:44898", "accept": false, - "reason": "SigLIP2 nonsensical pipeline output and Perceiver non-default resolution failure are different model-specific inference issues." + "reason": "Different vision-model failures with different symptoms and likely different root causes." }, { "left": "issue:44387", "right": "issue:44617", "accept": false, - "reason": "Int4 quantization memory regression and Sam3Video OOM are both memory-related, but the causes and affected code paths differ." + "reason": "Both mention OOM, but one is an int4 quantization memory regression and the other is a model-specific video OOM." }, { "left": "issue:43828", "right": "issue:44683", "accept": false, - "reason": "torch.autocast dtype mismatch in Phi-tiny-MoE is unrelated to compiled flex_attention failures on torch >= 2.9." + "reason": "Autocast dtype mismatch and compiled flex_attention failure on torch >= 2.9 are different problems." }, { "left": "issue:42915", "right": "issue:43482", "accept": false, - "reason": "Qwen3Moe FP8 config failure and Qwen2.5-GGUF loading failure are different model/loading paths." + "reason": "Qwen3Moe FP8 config failure and Qwen2.5-GGUF loading failure are unrelated." }, { "left": "issue:43526", "right": "issue:43883", "accept": false, - "reason": "BeitImageProcessorFast label reduction and Molmo all_tied_weights_keys are unrelated issues." + "reason": "Image processor `reduce_labels` bug is unrelated to missing `all_tied_weights_keys` on a model class." }, { "left": "issue:43792", "right": "issue:43957", "accept": false, - "reason": "Whisper large-v2 runtime failure and meta-device loading regressions on transformers 5 are separate problems." + "reason": "Whisper runtime failure and meta-device model loading regression do not share the same root cause." }, { "left": "issue:43701", "right": "issue:45325", "accept": false, - "reason": "Checkpoint key mismatch during resume and Qwen2.5-VL rope-index scaling are unrelated bugs." + "reason": "Checkpoint resume key mismatch is unrelated to Qwen2.5-VL rope/position-id scaling." }, { "left": "issue:43262", "right": "issue:43746", "accept": false, - "reason": "Audio chat-template sample-rate defaulting and GraniteSpeech PEFT local checkpoint loading are different audio/model-loading issues." + "reason": "Audio chat-template sampling-rate default and PEFT local-checkpoint loading are different bugs." }, { "left": "issue:43632", "right": "issue:45563", "accept": false, - "reason": "_is_hf_initialized regression and stale paged generate warning are not the same bug." + "reason": "An initialization flag regression is unrelated to a stale generate warning." }, { "left": "issue:43881", "right": "issue:44991", "accept": false, - "reason": "GLM-4V loading failure and tokenizer loading failure for EMBEDDIA/est-roberta are distinct model/tokenizer regressions." + "reason": "Different model/tokenizer loading failures with no shared code path." }, { "left": "issue:43317", "right": "issue:43873", "accept": false, - "reason": "gpu+cpu offload failing for dequantized models and offloading with quantization not working are related only at a high level; the concrete failure modes differ." + "reason": "Both concern offloading/quantization, but one is a dequantized-model load failure with gpu+cpu offload and the other is a broader offloading complaint; not the same concrete bug." }, { "left": "issue:40990", "right": "issue:45584", "accept": false, - "reason": "High perplexity on gpt-oss-20b and Whisper empty transcription failure are unrelated generation/evaluation issues." + "reason": "Perplexity/evaluation anomaly and Whisper empty-transcription generation failure are unrelated." }, { "left": "issue:34634", "right": "issue:35707", "accept": false, - "reason": "BarkProcessor voice_preset and progressive generation with inputs_embeds/past_key_values are completely unrelated." + "reason": "Bark voice_preset support and progressive generation with inputs_embeds/past_key_values are unrelated." } ] }, @@ -3259,8 +3211,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -3292,7 +3244,6 @@ "issue:44186", "issue:44265", "issue:44291", - "issue:44292", "issue:44423", "issue:44466", "issue:44479", @@ -3319,17 +3270,17 @@ "issue:44186|issue:44265", "issue:43606|issue:43819", "issue:44610|issue:45325", - "issue:43377|issue:44898", "issue:35707|issue:41720", + "issue:43377|issue:44898", "issue:43454|issue:43550", - "issue:43454|issue:45072", "issue:43377|issue:44610", - "issue:44466|issue:45357", + "issue:43454|issue:45072", "issue:39692|issue:40444", + "issue:44466|issue:45357", "issue:43749|issue:44534", "issue:43296|issue:44991", - "issue:43296|issue:44877", "issue:43792|issue:44683", + "issue:43296|issue:44877", "issue:41720|issue:43322", "issue:43454|issue:43720", "issue:43404|issue:45357", @@ -3337,11 +3288,11 @@ "issue:44423|issue:45406", "issue:43526|issue:43881", "issue:44186|issue:45072", - "issue:43577|issue:43708", "issue:43873|issue:44756", + "issue:43577|issue:43708", "issue:42617|issue:42915", - "issue:44610|issue:44898", "issue:44877|issue:44991", + "issue:44610|issue:44898", "issue:43550|issue:43720", "issue:44479|issue:44918", "issue:44186|issue:44610", @@ -3351,271 +3302,264 @@ "issue:44671|issue:44743", "issue:42915|issue:43317", "issue:43550|issue:43606", - "issue:41720|issue:42617", - "issue:44292|issue:44479" + "issue:41720|issue:42617" ], - "prepared_review_unit_hash": "af40a5c109f29ff6676b5131695b1388f8c0cea334d0bbd463cecab90012a688", + "prepared_review_unit_hash": "7852e5b2837c4dbfd48907421944cee58f7022d1c59336beb3d907bd0f0dae02", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 39, - "serialized_chars": 39466, - "estimated_input_tokens": 9867, - "estimated_eval_tokens": 19990 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 38, + "serialized_chars": 38640, + "estimated_input_tokens": 9660, + "estimated_eval_tokens": 19576 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is highly heterogeneous: the titles point to many different bugs across loading, multimodal, tokenizer, and training paths. I do not see any soft pair that is clearly the same underlying issue, so I would keep them separate.", - "confidence": 0.16, - "canonical_issue_reason": "issue:43749 is the broadest, most infrastructure-level report here: a concrete loading regression with cross-references, making it the best representative issue in this otherwise mixed set.", + "summary": "This cluster is mostly heterogeneous: the soft links point to broad thematic overlap around loading/offloading, quantization, serialization, and multimodal preprocessing, but the paired issues do not appear to describe the same underlying bug. I would not merge any of the soft-edge candidates.", + "confidence": 0.76, + "canonical_issue_reason": "issue:43873 is the broadest representative of the recurring loading/offloading theme, but the cluster is not actually a duplicate set and most items have distinct root causes.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43749 is the strongest single issue to anchor on because it describes a clear, central failure mode in model loading rather than a narrow model-specific symptom.", + "best_issue_reason": "issue:43873 is the closest umbrella issue and has the strongest cross-cutting theme among the set, but it still does not subsume the model-specific and API-specific reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43606", "right": "issue:44265", "accept": false, - "reason": "Different failures: CPU offload device mismatch vs torch.export failing on torch_compilable_check." + "reason": "CPU-offload device mismatch vs torch.export failure on torch_compilable_check; different bugs." }, { "left": "issue:44291", "right": "issue:44877", "accept": false, - "reason": "Unexpected init argument vs strict config loading; different code paths and fixes." + "reason": "Unexpected _is_hf_initialized arg during init_empty_weights vs strict config loading; unrelated root causes." }, { "left": "issue:44186", "right": "issue:44265", "accept": false, - "reason": "Tokenizer crash on NER/padding vs export failure in compiled-check models." + "reason": "Tokenizer padding/NER crash vs export-time check failure; not the same issue." }, { "left": "issue:43606", "right": "issue:43819", "accept": false, - "reason": "Offload device mismatch is unrelated to DAC.from_latents missing STE behavior." + "reason": "CPU offload mismatch vs DAC STE/latents mismatch; different code paths." }, { "left": "issue:44610", "right": "issue:45325", "accept": false, - "reason": "Processor/model image-size mismatch vs Qwen2.5-VL temporal position scaling bug." - }, - { - "left": "issue:43377", - "right": "issue:44898", - "accept": false, - "reason": "Batched-vs-single padding mask bug vs Perceiver non-default resolution failure." + "reason": "Processor/model input-size mismatch vs rope-index temporal scaling regression; different defects." }, { "left": "issue:35707", "right": "issue:41720", "accept": false, - "reason": "Generation with inputs_embeds/past_key_values is unrelated to Qwen3 auto device map cudaErrorAssert." + "reason": "Progressive generation with inputs_embeds/past_key_values vs auto device-map cuda assert; unrelated." }, { - "left": "issue:43454", - "right": "issue:43550", + "left": "issue:43377", + "right": "issue:44898", "accept": false, - "reason": "lm_head weight tying bug vs Bamba torch.compile/SDPA failure." + "reason": "MIMI padding-mask batching bug vs Perceiver resolution handling bug; different model paths." }, { "left": "issue:43454", - "right": "issue:45072", + "right": "issue:43550", "accept": false, - "reason": "Weight-tying serialization issue vs dtype mismatch in SwitchTransformers/TimmWrapperModel." + "reason": "lm_head tying/serialization bug vs torch.compile SDPA failure; not the same change." }, { "left": "issue:43377", "right": "issue:44610", "accept": false, - "reason": "Different multimodal issues: MIMI padding masks vs OmDet-Turbo processor resolution mismatch." + "reason": "Padding-mask batching issue vs OmDet image-size mismatch; different preprocessing bugs." }, { - "left": "issue:44466", - "right": "issue:45357", + "left": "issue:43454", + "right": "issue:45072", "accept": false, - "reason": "Device-dependent lm_head serialization vs Qwen3.5 visual encoder key saving regression." + "reason": "Weight-tying serialization issue vs dtype mismatch in CI; different symptoms and fixes." }, { "left": "issue:39692", "right": "issue:40444", "accept": false, - "reason": "SigLIP2 doc example errors are unrelated to Qwen2.5-VL IterableDataset finetuning failure." + "reason": "Doc example errors vs finetuning failure with multiple images; different reports." + }, + { + "left": "issue:44466", + "right": "issue:45357", + "accept": false, + "reason": "Generic tied-weight serialization inconsistency vs Qwen3.5 visual-key save_pretrained regression; not the same bug." }, { "left": "issue:43749", "right": "issue:44534", "accept": false, - "reason": "FSDP CPU-efficient loading broken is different from v5 non-persistent buffer junk filling." + "reason": "FSDP RAM-efficient loading failure vs non-persistent buffer junk-initialization; different mechanisms." }, { "left": "issue:43296", "right": "issue:44991", "accept": false, - "reason": "PaddleOCR-VL loading in vLLM differs from tokenizer loading failure for est-roberta." + "reason": "PaddleOCR-VL load failure vs tokenizer load regression for est-roberta; unrelated." }, { - "left": "issue:43296", - "right": "issue:44877", + "left": "issue:43792", + "right": "issue:44683", "accept": false, - "reason": "PaddleOCR-VL/vLLM loading issue is not the same as granite_speech strict-config rejection." + "reason": "Whisper load/run issue vs compiled flex_attention incompatibility; different subsystems." }, { - "left": "issue:43792", - "right": "issue:44683", + "left": "issue:43296", + "right": "issue:44877", "accept": false, - "reason": "Whisper model load failure is unrelated to compiled flex_attention on torch >= 2.9." + "reason": "Model loading failure vs strict config rejection; no shared underlying bug." }, { "left": "issue:41720", "right": "issue:43322", "accept": false, - "reason": "Qwen3 auto device-map assert differs from Llava Next segmentation fault on load." + "reason": "Qwen3 device-map cuda assert vs Llava Next segfault; different models and failures." }, { "left": "issue:43454", "right": "issue:43720", "accept": false, - "reason": "Different loading/serialization bugs: lm_head tying vs unpacking packed weights during accelerate load." + "reason": "lm_head tying bug vs packed-weight unpacking during accelerate loading; different code paths." }, { "left": "issue:43404", "right": "issue:45357", "accept": false, - "reason": "Mistral3 lm_head tying bug is different from Qwen3.5 visual encoder key regression." + "reason": "Mistral3 lm_head tying bug vs Qwen3.5 visual-encoder key serialization bug; not duplicates." }, { "left": "issue:41720", "right": "issue:44977", "accept": false, - "reason": "cudaErrorAssert on A800 is not the same as Qwen3.5 flash-attention generation regression." + "reason": "Device-mapping crash vs flash-attention generation regression; different Qwen3.5 issues." }, { "left": "issue:44423", "right": "issue:45406", "accept": false, - "reason": "Continuous-batching multimodal crash vs Gemma4Processor missing _tokenizer; different processors and failures." + "reason": "Continuous-batching multimodal crash vs Gemma4Processor missing _tokenizer; similar area, different bug." }, { "left": "issue:43526", "right": "issue:43881", "accept": false, - "reason": "BeitImageProcessorFast label reduction bug is unrelated to glm-4v-9b loading failure." + "reason": "BeitImageProcessorFast label reduction bug vs glm-4v loading failure; unrelated." }, { "left": "issue:44186", "right": "issue:45072", "accept": false, - "reason": "LayoutLMv2 tokenizer crash differs from dtype mismatch in SwitchTransformers/TimmWrapperModel." + "reason": "Tokenizer crash vs dtype mismatch in inference; no common root cause." }, { - "left": "issue:43577", - "right": "issue:43708", + "left": "issue:43873", + "right": "issue:44756", "accept": false, - "reason": "Blip2 dtype not propagated is unrelated to Trainer resume/max_steps recalculation." + "reason": "Quantization/offloading issue vs mmap OOM workaround; different failure modes." }, { - "left": "issue:43873", - "right": "issue:44756", + "left": "issue:43577", + "right": "issue:43708", "accept": false, - "reason": "Quantization/offloading behavior issue vs mmap OOM on Strix Halo; not the same bug." + "reason": "dtype preservation bug vs resume_from_checkpoint max_steps miscalculation; unrelated." }, { "left": "issue:42617", "right": "issue:42915", "accept": false, - "reason": "3d_parallel.py failure and Qwen3Moe FineGrainedFP8Config failure are different runtime issues." + "reason": "3d_parallel startup failure vs Qwen3Moe FineGrainedFP8Config failure; different models and errors." }, { - "left": "issue:44610", - "right": "issue:44898", + "left": "issue:44877", + "right": "issue:44991", "accept": false, - "reason": "OmDet-Turbo processor output size mismatch is not the same as Perceiver interpolation failure." + "reason": "Config-loading strictness vs tokenizer-loading regression; distinct issues." }, { - "left": "issue:44877", - "right": "issue:44991", + "left": "issue:44610", + "right": "issue:44898", "accept": false, - "reason": "Strict config loading for granite_speech is unrelated to tokenizer loading for est-roberta." + "reason": "Processor output size mismatch vs Perceiver non-default resolution failure; not the same bug." }, { "left": "issue:43550", "right": "issue:43720", "accept": false, - "reason": "torch.compile SDPA failure and packed-weight unpacking during load are different code paths." + "reason": "torch.compile SDPA failure vs accelerate packed-weight loading; different code paths." }, { "left": "issue:44479", "right": "issue:44918", "accept": false, - "reason": "Video-input regression in v5.3.0 is distinct from input-embedding unpacking in trl SFT trainer." + "reason": "Video-input regression vs SFT input-embedding unpacking failure; unrelated." }, { "left": "issue:44186", "right": "issue:44610", "accept": false, - "reason": "Tokenizer crash on NER inputs differs from OmDet-Turbo processor/model resolution mismatch." + "reason": "Tokenizer batching crash vs image processor resolution mismatch; different defects." }, { "left": "issue:36331", "right": "issue:42915", "accept": false, - "reason": "Trainer custom compute_loss signature error is unrelated to Qwen3Moe FP8 loading." + "reason": "Trainer compute_loss signature breakage vs Qwen3Moe FP8 failure; no overlap." }, { "left": "issue:36331", "right": "issue:42617", "accept": false, - "reason": "Unexpected Trainer keyword arg and 3d_parallel.py failure are different training/runtime problems." + "reason": "CustomTrainer API break vs 3d_parallel startup failure; unrelated." }, { "left": "issue:43316", "right": "issue:44913", "accept": false, - "reason": "Gemma3TextConfig API discrepancy is unrelated to GPTNeoX rotary_pct reload behavior." + "reason": "Gemma3TextConfig API discrepancy vs GPTNeoX rotary_pct not persisting on reload; different config bugs." }, { "left": "issue:44671", "right": "issue:44743", "accept": false, - "reason": "CamemBERT MLM prediction regression and Qwen3.5 recurrent-state reset are different model bugs." + "reason": "CamemBERT masked-LM prediction regression vs qwen3_5 recurrent-state reset; different models and logic." }, { "left": "issue:42915", "right": "issue:43317", "accept": false, - "reason": "Qwen3Moe FineGrainedFP8Config failure is not the same as dequantized model offload loading failure." + "reason": "Qwen3Moe FP8 failure vs device_map=auto dequantized-load offload failure; different bugs." }, { "left": "issue:43550", "right": "issue:43606", "accept": false, - "reason": "Bamba SDPA compile bug and Bark CPU offload mismatch are unrelated." + "reason": "Bamba torch.compile/SDPA bug vs CPU offload device mismatch; not the same issue." }, { "left": "issue:41720", "right": "issue:42617", "accept": false, - "reason": "Qwen3 device-map assert differs from 3d_parallel.py execution failure." - }, - { - "left": "issue:44292", - "right": "issue:44479", - "accept": false, - "reason": "Qwen-3-8B-NVFP4 running error is not the same as Qwen v5.3.0 video-input regression." + "reason": "Qwen3 auto device-map assert vs 3d_parallel startup failure; unrelated." } ] }, @@ -3634,8 +3578,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -3651,7 +3595,6 @@ "issue:42886", "issue:42915", "issue:43322", - "issue:43381", "issue:43388", "issue:43421", "issue:43531", @@ -3663,14 +3606,15 @@ "issue:43720", "issue:43746", "issue:43782", - "issue:43844", "issue:43873", "issue:43927", "issue:44062", "issue:44186", + "issue:44292", "issue:44361", "issue:44368", "issue:44451", + "issue:44479", "issue:44492", "issue:44509", "issue:44514", @@ -3690,20 +3634,21 @@ ], "soft_pairs": [ "issue:43531|issue:43572", + "issue:44292|issue:44479", "issue:41720|issue:45161", "issue:43540|issue:43782", "issue:43388|issue:45059", "issue:44186|issue:44361", - "issue:41720|issue:43873", "issue:42915|issue:43322", + "issue:41720|issue:43873", "issue:36010|issue:42617", "issue:45406|issue:45464", "issue:44514|issue:44545", "issue:36331|issue:41720", "issue:33357|issue:35707", "issue:41720|issue:43746", - "issue:43606|issue:43720", "issue:30064|issue:34634", + "issue:43606|issue:43720", "issue:44492|issue:44509", "issue:43550|issue:43653", "issue:44368|issue:45357", @@ -3717,223 +3662,222 @@ "issue:42886|issue:44843", "issue:43421|issue:43653", "issue:44451|issue:44991", - "issue:44062|issue:44451", - "issue:43381|issue:43844" + "issue:44062|issue:44451" ], - "prepared_review_unit_hash": "3e9eeddf8eb4d697a4ed643f0ba42b83f879698a66e2594a3f67e1fc118f1fc9", + "prepared_review_unit_hash": "23ca7a7f42bc48e7ebde864b67fa218895228b3b0500112b06b5c0abd3128e50", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36765, - "estimated_input_tokens": 9192, - "estimated_eval_tokens": 18640 + "serialized_chars": 36737, + "estimated_input_tokens": 9185, + "estimated_eval_tokens": 18626 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most pairs are unrelated superficial similarities. Two soft duplicate candidates stand out: the Qwen2_5_VLProcessor batched padding=False crash, and the HF_HUB_OFFLINE tokenizer-loading regression.", - "confidence": 0.78, - "canonical_issue_reason": "Issue 44514 is the best canonical issue: it has the fuller repro for the `Qwen2_5_VLProcessor.apply_chat_template` crash with `padding=False`, and 44545 appears to be the same bug with less detail.", + "summary": "Only two soft edges look like true duplicates: the Qwen2_5_VLProcessor batched-padding crash and the HF_HUB_OFFLINE tokenizer-cache failure. The remaining pairs share broad areas but not the same concrete bug/change.", + "confidence": 0.96, + "canonical_issue_reason": "issue:44514 is the cleanest canonical issue for its pair: it is the older report and has an exact title match with 44545, so it best serves as the duplicate anchor.", "canonical_pr_reason": null, - "best_issue_reason": "44514 is the strongest issue to anchor this cluster because it is a precise, reproducible report and the clearest representative of an actual duplicate pair.", + "best_issue_reason": "issue:44843 is the strongest standalone issue overall because it states the offline-mode failure clearly and points to the likely root cause (`_patch_mistral_regex` calling `model_info()` unconditionally), making it the most actionable anchor.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43531", "right": "issue:43572", "accept": false, - "reason": "Different models and failure modes: Qwen3-MoE sliding-window behavior vs StableLmConfig pad token metadata." + "reason": "Different models and failure modes: Qwen3-MoE sliding-window behavior vs StableLmConfig pad_token_idx missing after v5 update." + }, + { + "left": "issue:44292", + "right": "issue:44479", + "accept": false, + "reason": "Both involve Qwen models, but one is an NVFP4 loading error and the other is a video-input regression in Qwen VL/MoE paths." }, { "left": "issue:41720", "right": "issue:45161", "accept": false, - "reason": "Both involve model execution, but the reported problems are different: A800 cudaErrorAssert vs GPT-OSS MoE tensor-parallel behavior." + "reason": "Different concrete problems: auto device mapping CUDA assert on A800 vs GPT-OSS MoE tensor-parallelism failure." }, { "left": "issue:43540", "right": "issue:43782", "accept": false, - "reason": "Different code paths: video input processing in Qwen3OmniMoe vs Qwen3VL `from_pretrained(weight_only=True)` loading error." + "reason": "One is a Qwen3OmniMoe video-input processing ValueError; the other is a Qwen3VL from_pretrained weight_only loading error." }, { "left": "issue:43388", "right": "issue:45059", "accept": false, - "reason": "Unrelated bugs in evaluation metric gathering vs SAM3 text/bounding-box prompting behavior." + "reason": "Metric gathering label truncation in Trainer/Mask2Former is not the same bug as SAM3 text/bbox weird behavior." }, { "left": "issue:44186", "right": "issue:44361", "accept": false, - "reason": "Both are tokenizer/task failures, but they affect different tokenizer families and different underlying errors." + "reason": "Both are tokenizer crashes, but they affect different models and appear to stem from different code paths and inputs." }, { - "left": "issue:41720", - "right": "issue:43873", + "left": "issue:42915", + "right": "issue:43322", "accept": false, - "reason": "Different issues: device-map assertion on Qwen3 vs offloading/quantization behavior." + "reason": "Qwen3Moe FineGrainedFP8Config failure and Llava Next segfault are unrelated concrete bugs." }, { - "left": "issue:42915", - "right": "issue:43322", + "left": "issue:41720", + "right": "issue:43873", "accept": false, - "reason": "Different failures in different model stacks: FineGrainedFP8 on Qwen3Moe vs Llava Next load-time segfault." + "reason": "Both relate loosely to loading/offloading, but the reported failures and triggers are different enough to keep separate." }, { "left": "issue:36010", "right": "issue:42617", "accept": false, - "reason": "No shared underlying bug; import-path regression is unrelated to running `3d_parallel.py`." + "reason": "GenerationMixin import error and a 3d_parallel script failure are unrelated." }, { "left": "issue:45406", "right": "issue:45464", "accept": false, - "reason": "Different surfaces and symptoms: transformers-serve processor init crash vs chat/completions streaming failure." + "reason": "Gemma4Processor missing _tokenizer in serve is a different issue from Qwen3.5 streaming inference failure." }, { "left": "issue:44514", "right": "issue:44545", "accept": true, - "reason": "Same exact `Qwen2_5_VLProcessor.apply_chat_template` crash on batched input with `padding=False`; 44545 is a duplicate report." + "reason": "Exact same title and same crash scenario; 44514 is the older canonical report." }, { "left": "issue:36331", "right": "issue:41720", "accept": false, - "reason": "Different bugs: custom trainer signature mismatch vs model device-mapping assertion." + "reason": "Trainer compute_loss signature mismatch is unrelated to Qwen3 auto device-mapping CUDA asserts." }, { "left": "issue:33357", "right": "issue:35707", "accept": false, - "reason": "No common code path; MacOS bus error in CLIP loading is unrelated to progressive generation with `inputs_embeds` and `past_key_values`." + "reason": "MacOS bus error with a CLIP model is not the same underlying bug as progressive generation with inputs_embeds/past_key_values." }, { "left": "issue:41720", "right": "issue:43746", "accept": false, - "reason": "Different problems: Qwen3 device-map failure vs PEFT-adapter checkpoint loading for GraniteSpeech." + "reason": "Different failures: device-mapping CUDA assert vs PEFT-adapter local checkpoint loading for GraniteSpeech." }, { - "left": "issue:43606", - "right": "issue:43720", + "left": "issue:30064", + "right": "issue:34634", "accept": false, - "reason": "Both touch loading/offload, but one is a CPU-offload device mismatch and the other is BitNet packed-weight unpacking during accelerate loading." + "reason": "Image Processor void segmentation maps and BarkProcessor voice_preset are entirely different subsystems and bugs." }, { - "left": "issue:30064", - "right": "issue:34634", + "left": "issue:43606", + "right": "issue:43720", "accept": false, - "reason": "Unrelated processor bugs: segmentation-map handling vs Bark voice preset parsing." + "reason": "Both touch accelerate/offloading/loading, but one is a CPU-offload device mismatch and the other is unpacking packed weights in BitNet." }, { "left": "issue:44492", "right": "issue:44509", "accept": false, - "reason": "A documentation typo is not the same as stale pipeline-task docs in v5; different scope and fix." + "reason": "A cache-strategy typo and docs mentioning removed pipeline tasks are unrelated documentation/config issues." }, { "left": "issue:43550", "right": "issue:43653", "accept": false, - "reason": "Different tokenizer/model issues: torch.compile + SDPA vs mask-token registration and empty decode output." + "reason": "torch.compile/SDPA failure in Bamba-9B-v2 is unrelated to BigBirdTokenizer special-token registration." }, { "left": "issue:44368", "right": "issue:45357", "accept": false, - "reason": "Different regressions affecting different parts of model config/save behavior." + "reason": "A tie_word_embeddings warning during LoRA fine-tuning is not the same as incorrect visual-encoder keys being saved." }, { "left": "issue:44625", "right": "issue:45479", "accept": false, - "reason": "Related to classification config, but not the same bug: one is `num_labels` propagation in Qwen3.5, the other is degenerate zero loss for single-label classification across models." + "reason": "num_labels propagation in Qwen3.5 config is different from the cross-model zero-loss classification bug." }, { "left": "issue:43421", "right": "issue:43927", "accept": false, - "reason": "Different tokenizer-state bugs: runtime post-processor updates vs saved config losing custom token IDs." + "reason": "Runtime post-processor updates for special tokens and save/load loss of custom token IDs are related broadly, but not the same concrete bug." }, { "left": "issue:43606", "right": "issue:43653", "accept": false, - "reason": "Different bugs in different models: device mismatch during offload vs tokenizer special-token registration." + "reason": "Device mismatch during CPU offload and mask-token registration in a tokenizer are unrelated." }, { "left": "issue:43782", "right": "issue:45381", "accept": false, - "reason": "Different Qwen vision-language failures: `weight_only=True` load error vs incorrect video `vision_position_ids`." + "reason": "from_pretrained weight_only loading error is distinct from wrong vision_position_ids in Qwen2.5-VL video input." }, { "left": "issue:34567", "right": "issue:35707", "accept": false, - "reason": "No shared underlying issue; trainer token counting is unrelated to generation with cached past key values." + "reason": "TrainerState token counting and progressive generation with past_key_values are different issues." }, { "left": "issue:34634", "right": "issue:41720", "accept": false, - "reason": "Bark processor voice preset handling is unrelated to Qwen3 device-mapping crashes." + "reason": "BarkProcessor voice_preset and Qwen3 auto device mapping are unrelated." }, { "left": "issue:44898", "right": "issue:45072", "accept": false, - "reason": "Different models and failure modes: Perceiver image resolution handling vs dtype mismatches in inference." + "reason": "Perceiver interpolation at non-default resolution and dtype mismatches in bfloat16 inference do not describe the same bug." }, { "left": "issue:42886", "right": "issue:44843", "accept": true, - "reason": "Both describe tokenizer loading breaking in HF_HUB_OFFLINE mode due to an unconditional hub metadata call in the tokenizer-loading path." + "reason": "Both describe the same HF_HUB_OFFLINE tokenizer-loading failure; 44843 supplies the likely root cause via an unconditional `model_info()` call." }, { "left": "issue:43421", "right": "issue:43653", "accept": false, - "reason": "Different tokenizer bugs: post-processor not refreshed at runtime vs mask token not registered as special." + "reason": "TokenizersBackend post-processor refresh and BigBirdTokenizer special-token registration are different tokenization bugs." }, { "left": "issue:44451", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer load failures after newer versions, but they hit different models and the root cause is not established as the same." + "reason": "Both are tokenizer-loading regressions, but for different models and not enough evidence of a shared underlying cause." }, { "left": "issue:44062", "right": "issue:44451", "accept": false, - "reason": "The AddedToken keyword-argument error is a specific tokenizer construction bug, while ScandiBERT loading failure is a separate model-specific report." - }, - { - "left": "issue:43381", - "right": "issue:43844", - "accept": false, - "reason": "Unrelated training/gradient issues: eval-mode gradient checkpointing vs ZeRO-3 gradient explosion." + "reason": "AddedToken constructor argument handling is unrelated to ScandiBERT tokenizer loading." } ] }, "evaluator_result": { "accept": true, - "feedback": "The cluster summary is grounded in the packet, and the two accepted soft edges are conservative enough: 44514/44545 are near-identical reports of the same Qwen2_5_VLProcessor padding=False crash, and 42886/44843 both point to the same HF_HUB_OFFLINE tokenizer-loading regression. The rejected pairs are also framed cautiously and do not overclaim duplicates." + "feedback": "Grounded and conservative overall. The two accepted duplicate pairs are well supported by the packet: 44514/44545 have the same title and scenario, and 42886/44843 describe the same HF_HUB_OFFLINE tokenizer-loading failure with a plausible shared root cause. The remaining soft edges are rejected cautiously without overclaiming similarity." }, "evaluator_used": true, "retried": false, @@ -3949,8 +3893,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -3959,12 +3903,12 @@ "issue:39692", "issue:40444", "issue:43335", + "issue:43381", "issue:43421", "issue:43525", "issue:43531", "issue:43540", "issue:43550", - "issue:43577", "issue:43596", "issue:43618", "issue:43653", @@ -3973,10 +3917,9 @@ "issue:43720", "issue:43746", "issue:43756", - "issue:43818", "issue:43819", + "issue:43844", "issue:43873", - "issue:43927", "issue:43994", "issue:44008", "issue:44062", @@ -4004,6 +3947,7 @@ "issue:45479" ], "soft_pairs": [ + "issue:43381|issue:43844", "issue:44918|issue:44936", "issue:44077|issue:44625", "issue:44987|issue:45042", @@ -4014,242 +3958,234 @@ "issue:45081|issue:45356", "issue:43525|issue:43531", "issue:43596|issue:44220", - "issue:45183|issue:45479", "issue:43746|issue:45381", + "issue:45183|issue:45479", "issue:39692|issue:43873", "issue:43618|issue:43697", "issue:43540|issue:44008", - "issue:43746|issue:43873", "issue:40444|issue:43746", + "issue:43746|issue:43873", "issue:44220|issue:44661", "issue:43716|issue:44008", "issue:43335|issue:43756", "issue:43720|issue:43819", "issue:43653|issue:43720", - "issue:44843|issue:45081", "issue:43421|issue:44811", "issue:30064|issue:36010", + "issue:44843|issue:45081", "issue:44062|issue:44488", "issue:44186|issue:44898", "issue:43540|issue:45381", - "issue:43550|issue:45072", - "issue:43577|issue:43818", - "issue:43550|issue:43927" + "issue:43550|issue:45072" ], - "prepared_review_unit_hash": "68eda53b31bae9b20c6a5ed8e508c5f6db13830f46cc2ced5cff3abba4766f85", + "prepared_review_unit_hash": "f309753de1960f2da45be8c380ea47234b7a5dcd7928486e16379c9f0cb9bae2", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 31, - "serialized_chars": 37390, - "estimated_input_tokens": 9348, - "estimated_eval_tokens": 18952 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36405, + "estimated_input_tokens": 9102, + "estimated_eval_tokens": 18460 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is not a duplicate cluster: the items span many unrelated bugs across tokenizers, vision/video processors, quantization, training, and model config. All soft edges should be rejected.", - "confidence": 0.97, - "canonical_issue_reason": "No canonical issue: the collection is heterogeneous and does not center on one underlying bug or change.", + "summary": "This cluster is highly heterogeneous; the soft pairs are mostly broad-topic similarities rather than true duplicates, so I rejected all of them. The closest-looking pair is the Mistral tokenizer regex set, but even there the failures are different enough to keep separate.", + "confidence": 0.95, + "canonical_issue_reason": "issue:45081 is the most concrete, code-path-specific bug in the set (a direct tokenizer loading crash with a clear trigger and traceback), so it is the best representative issue if one must be chosen.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43873 is the closest loose representative because it is open and has the most active discussion/inbound references, but it is still not a duplicate target for the rest of the set.", + "best_issue_reason": "issue:45081 has the strongest actionable reproduction and the most precisely scoped failure mode among these unrelated issues.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43381", + "right": "issue:43844", + "accept": false, + "reason": "Different subsystems and failure modes: eval-mode gradient checkpointing vs ZeRO-3 gradient explosion during training." + }, { "left": "issue:44918", "right": "issue:44936", "accept": false, - "reason": "Different failures: Qwen3.5 embedding unpacking in TRL SFT vs trainer.evaluate() after train()." + "reason": "One is an embeddings/SFT trainer failure, the other is evaluate() breaking after train(); not the same bug." }, { "left": "issue:44077", "right": "issue:44625", "accept": false, - "reason": "Unrelated: patchtsmixer post_init policy vs Qwen3.5 num_labels propagation." + "reason": "Unrelated model/config problems: patchtsmixer post_init policy vs Qwen3.5 num_labels propagation." }, { "left": "issue:44987", "right": "issue:45042", "accept": false, - "reason": "Different subsystems: loading a physical-intelligence model vs PIL image processors wrongly requiring torchvision." + "reason": "Different bugs and surfaces: model loading regression vs PIL image processor torchvision dependency issue." }, { "left": "issue:43531", "right": "issue:44589", "accept": false, - "reason": "Different bugs: Qwen3-MoE sliding_window vs Float8 storage lookup failure." + "reason": "Qwen3-MoE sliding_window bug is unrelated to the Float8 storage lookup error." }, { "left": "issue:43994", "right": "issue:44671", "accept": false, - "reason": "Different models and symptoms: SigLIP2 inference issues vs CamemBERT masked LM regression." + "reason": "Both are model-output regressions, but they involve different models and distinct code paths." }, { "left": "issue:39692", "right": "issue:44610", "accept": false, - "reason": "Doc/example mistakes for SigLIP2 are unrelated to OmDet-Turbo processor size mismatch." + "reason": "Doc/example errors for SigLIP2 do not match an OmDet-Turbo processor/model input-size mismatch." }, { "left": "issue:43421", "right": "issue:44186", "accept": false, - "reason": "Both are tokenizer-related, but they are distinct code paths and failures." + "reason": "Tokenizer post-processor update behavior is not the same as LayoutLMv2 NER padding/truncation crashes." }, { "left": "issue:45081", "right": "issue:45356", "accept": false, - "reason": "Both touch mistral regex handling, but one is a crash and the other is a tokenizer regression/codec warning; not the same concrete bug." + "reason": "Both mention fix_mistral_regex, but one is a loading crash and the other is a tokenizer regression/warning issue; not the same underlying defect." }, { "left": "issue:43525", "right": "issue:43531", "accept": false, - "reason": "Llama4 pad_token_id attribute error is unrelated to Qwen3-MoE sliding_window behavior." + "reason": "Llama4Config pad_token_id absence is unrelated to Qwen3-MoE sliding_window handling." }, { "left": "issue:43596", "right": "issue:44220", "accept": false, - "reason": "Deepspeed zero3/BertModel initialization error is unrelated to fbank feature extraction." + "reason": "BertModel/ZeRO-3 initialization failure is unrelated to audio feature extraction." }, { - "left": "issue:45183", - "right": "issue:45479", + "left": "issue:43746", + "right": "issue:45381", "accept": false, - "reason": "Input validation error messaging is unrelated to the zero-loss sequence-classification bug." + "reason": "PEFT checkpoint loading for GraniteSpeech is unrelated to Qwen2.5-VL video vision_position_ids." }, { - "left": "issue:43746", - "right": "issue:45381", + "left": "issue:45183", + "right": "issue:45479", "accept": false, - "reason": "GraniteSpeech PEFT checkpoint loading is unrelated to Qwen2.5-VL video position ids." + "reason": "Input validation error messages are not the same bug as the degenerate zero-loss classification behavior." }, { "left": "issue:39692", "right": "issue:43873", "accept": false, - "reason": "SigLIP2 documentation/example errors are unrelated to quantization offloading behavior." + "reason": "A documentation example bug is not the same as an offloading/quantization runtime bug." }, { "left": "issue:43618", "right": "issue:43697", "accept": false, - "reason": "CLIPOutput attentions regression is unrelated to RTDetrV2 output differences." + "reason": "Missing CLIP attentions assignment and RTDetrV2 output differences are different regressions." }, { "left": "issue:43540", "right": "issue:44008", "accept": false, - "reason": "Qwen3OmniMoe video processing failure is unrelated to Gemma3n audio mask name collision." + "reason": "Different models and failure types: Qwen3OmniMoe video processing vs Gemma 3n variable collision." }, { - "left": "issue:43746", - "right": "issue:43873", + "left": "issue:40444", + "right": "issue:43746", "accept": false, - "reason": "PEFT adapter loading and quantized offloading are separate bugs." + "reason": "Multi-image Qwen2.5-VL finetuning failure is unrelated to GraniteSpeech PEFT checkpoint loading." }, { - "left": "issue:40444", - "right": "issue:43746", + "left": "issue:43746", + "right": "issue:43873", "accept": false, - "reason": "Qwen2.5-VL iterable dataset multi-image training failure is unrelated to GraniteSpeech local checkpoint loading." + "reason": "Checkpoint loading with PEFT adapters is unrelated to quantization offloading behavior." }, { "left": "issue:44220", "right": "issue:44661", "accept": false, - "reason": "Audio feature extraction bug is unrelated to add-new-model-like failing in tokenizer mapping." + "reason": "Audio feature extraction and add-new-model-like/tokenizer mapping failure are unrelated." }, { "left": "issue:43716", "right": "issue:44008", "accept": false, - "reason": "Mistral image preprocessor dtype mismatch is unrelated to Gemma3n variable collision." + "reason": "Mistral-3 image dtype mismatch is unrelated to Gemma 3n audio tensor name collision." }, { "left": "issue:43335", "right": "issue:43756", "accept": false, - "reason": "SwitchTransformers sparse-layer config bug is unrelated to Smollm3 RoPE layer dropping." + "reason": "Sparse-layer creation bug and RoPE-layer dropping are different model/config regressions." }, { "left": "issue:43720", "right": "issue:43819", "accept": false, - "reason": "BitNet loading/unpacking issue is unrelated to DAC forward/from_latents mismatch." + "reason": "Packed-weight loading in BitNet is unrelated to DAC forward/from_latents mismatch." }, { "left": "issue:43653", "right": "issue:43720", "accept": false, - "reason": "BigBirdTokenizer special-token decode bug is unrelated to BitNet accelerate-loading bug." - }, - { - "left": "issue:44843", - "right": "issue:45081", - "accept": false, - "reason": "Both involve mistral-regex code, but one breaks offline model_info calls while the other crashes on a missing backend_tokenizer attribute." + "reason": "BigBirdTokenizer special-token decoding bug is unrelated to BitNet packed-weight loading." }, { "left": "issue:43421", "right": "issue:44811", "accept": false, - "reason": "Runtime special-token post-processing and Whisper batch_decode skip_special_tokens are different behaviors." + "reason": "Both involve tokenization/decoding, but one is runtime post-processor updates and the other is Whisper batch_decode skip_special_tokens behavior." }, { "left": "issue:30064", "right": "issue:36010", "accept": false, - "reason": "Void segmentation map processing is unrelated to GenerationMixin import failure." + "reason": "Image processor segmentation-map handling is unrelated to a GenerationMixin import error." + }, + { + "left": "issue:44843", + "right": "issue:45081", + "accept": false, + "reason": "Both touch _patch_mistral_regex, but one is unconditional hub metadata access/offline failure and the other is an attribute error on tokenizer loading." }, { "left": "issue:44062", "right": "issue:44488", "accept": false, - "reason": "AddedToken special-argument TypeError is not the same bug as loading cjvt/sleng-bert." + "reason": "AddedToken special-argument duplication is unrelated to the model-loading problem for cjvt/sleng-bert." }, { "left": "issue:44186", "right": "issue:44898", "accept": false, - "reason": "LayoutLMv2 tokenizer crashes are unrelated to Perceiver image classification at non-default resolution." + "reason": "Tokenizer padding/truncation crashes are not the same as Perceiver image resolution interpolation failures." }, { "left": "issue:43540", "right": "issue:45381", "accept": false, - "reason": "Both mention video inputs, but they are different model families and different failure modes." + "reason": "Both are video-related, but they affect different models and different internal bugs." }, { "left": "issue:43550", "right": "issue:45072", "accept": false, - "reason": "torch.compile/SDPA failure in Bamba is unrelated to bf16 dtype mismatches in SwitchTransformers/TimmWrapperModel." - }, - { - "left": "issue:43577", - "right": "issue:43818", - "accept": false, - "reason": "Blip2 dtype handling is unrelated to Video-LLaVA temporal-attention/weight-sharing problems." - }, - { - "left": "issue:43550", - "right": "issue:43927", - "accept": false, - "reason": "Bamba torch.compile SDPA regression is unrelated to DiaConfig custom token ID loss." + "reason": "Bamba torch.compile SDPA failure is unrelated to dtype mismatches in SwitchTransformers/TimmWrapperModel inference." } ] }, @@ -4268,8 +4204,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -4292,14 +4228,15 @@ "issue:43720", "issue:43749", "issue:43782", + "issue:43818", "issue:43819", "issue:43825", "issue:43827", "issue:43906", + "issue:43927", "issue:43994", "issue:44060", "issue:44186", - "issue:44291", "issue:44368", "issue:44442", "issue:44448", @@ -4310,7 +4247,6 @@ "issue:44671", "issue:44805", "issue:44821", - "issue:44857", "issue:44877", "issue:44936", "issue:44977", @@ -4323,10 +4259,12 @@ "issue:45406" ], "soft_pairs": [ + "issue:43577|issue:43818", "issue:43643|issue:44877", - "issue:43577|issue:43582", + "issue:43550|issue:43927", "issue:43782|issue:44977", "issue:43531|issue:43643", + "issue:43577|issue:43582", "issue:44448|issue:44671", "issue:43644|issue:43749", "issue:43653|issue:43819", @@ -4340,8 +4278,8 @@ "issue:41720|issue:44991", "issue:44805|issue:44936", "issue:30064|issue:36331", - "issue:43606|issue:45072", "issue:44485|issue:44509", + "issue:43606|issue:45072", "issue:42175|issue:45081", "issue:43782|issue:45325", "issue:44610|issue:45072", @@ -4350,170 +4288,174 @@ "issue:45081|issue:45406", "issue:43720|issue:45072", "issue:43262|issue:43819", - "issue:43653|issue:44671", - "issue:43582|issue:44857", - "issue:43749|issue:44291" + "issue:43653|issue:44671" ], - "prepared_review_unit_hash": "b36fa89192bc3b82a60d484e384712abcbc234fbcf89edf0b1a4a0e2dd7cc3ec", + "prepared_review_unit_hash": "a40f6d534b3accdb894f805fe61485b9afab9d7fbb8d4c540563e4fd505556c9", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36840, - "estimated_input_tokens": 9210, - "estimated_eval_tokens": 18676 + "serialized_chars": 36866, + "estimated_input_tokens": 9217, + "estimated_eval_tokens": 18690 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a heterogeneous set of unrelated transformer issues with only a few loose theme overlaps (v5 regressions, remote_code, tokenizer/processor loading). No PRs are present. I would not merge any of the soft pairs as duplicates.", - "confidence": 0.95, - "canonical_issue_reason": "No single canonical duplicate stands out; the cluster is too heterogeneous. If an umbrella is needed, 45020 is the broadest remote_code-regression report.", + "summary": "The cluster is mostly a set of unrelated regression reports that share broad Transformers-version or model-family vocabulary rather than the same underlying bug. I found no soft-edge pair that is clearly the same concrete issue/change, so all candidate links are rejected.", + "confidence": 0.71, + "canonical_issue_reason": "issue:45020 is the broadest active report in the set and best fits a loose umbrella role for recent-version remote_code regressions, though the cluster overall is too heterogeneous for a strong canonical choice.", "canonical_pr_reason": null, - "best_issue_reason": "45020 is the closest thing to an umbrella issue because it captures the broad 'recent versions break remote_code' theme seen in a few nearby reports, though it still does not unify the cluster.", + "best_issue_reason": "issue:45020 has the widest scope, is still open, and can plausibly serve as the representative tracking issue for the cluster\u2019s loose theme of recent-version breakages.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43577", + "right": "issue:43818", + "accept": false, + "reason": "Different problems: one is a dtype-loading issue in BLIP2, the other is a Video-LLaVA model-structure/weight-sharing bug." + }, { "left": "issue:43643", "right": "issue:44877", "accept": false, - "reason": "Both involve config/model loading, but one is missing fields with trust_remote_code and the other is strict config rejection for granite_speech; different bugs." + "reason": "Both involve config loading, but one is about missing fields from trust_remote_code and the other is a strict-config failure for granite_speech; not the same bug." }, { - "left": "issue:43577", - "right": "issue:43582", + "left": "issue:43550", + "right": "issue:43927", "accept": false, - "reason": "BLIP2 dtype propagation and Apple Silicon allocator warmup TypeError are unrelated code paths." + "reason": "Unrelated: torch.compile/SDPA failure for Bamba versus DiaConfig losing custom token IDs on save/load." }, { "left": "issue:43782", "right": "issue:44977", "accept": false, - "reason": "Different models and failures: Qwen3VL weight_only loading error vs Qwen3.5 flash-attention generation issue." + "reason": "Different model and failure mode: Qwen3VL weight_only loading error versus Qwen3.5 flash-attention generation issue." }, { "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": "No concrete overlap beyond being Transformer model/config issues; sliding_window behavior and missing config fields are distinct." + }, + { + "left": "issue:43577", + "right": "issue:43582", + "accept": false, + "reason": "BLIP2 dtype mismatch and AppleSilicon caching_allocator_warmup TypeError are unrelated." }, { "left": "issue:44448", "right": "issue:44671", "accept": false, - "reason": "Both are v5 regressions, but Pegasus generation output changes and CamemBERT masked LM predictions are distinct model/code paths." + "reason": "Both are v5 output regressions, but for different models and different code paths; too broad to merge." }, { "left": "issue:43644", "right": "issue:43749", "accept": false, - "reason": "Non-persistent buffer junk and FSDP RAM-efficient loading are separate loading/initialization problems." + "reason": "Non-persistent buffer initialization bug and FSDP CPU RAM efficient loading breakage are separate loading-path problems." }, { "left": "issue:43653", "right": "issue:43819", "accept": false, - "reason": "BigBirdTokenizer special-token registration and DAC from_latents/STE mismatch are unrelated." + "reason": "Tokenizer special-token registration and DAC latent reconstruction are unrelated." }, { "left": "issue:44561", "right": "issue:45020", "accept": false, - "reason": "45020 is a broad umbrella for remote_code breakage, while 44561 is a specific is_torch_fx_available removal regression; not the same concrete bug." + "reason": "45020 is a broad umbrella report, while 44561 is a specific v5 regression from removing is_torch_fx_available; not clearly the same concrete issue." }, { "left": "issue:43825", "right": "issue:43994", "accept": false, - "reason": "Pipeline error-message text and SigLIP nonsensical outputs are unrelated." - }, - { - "left": "issue:43550", - "right": "issue:43819", - "accept": false, - "reason": "torch.compile + SDPA failure in Bamba is unrelated to DAC latent decoding mismatch." + "reason": "Pipeline error-message regression and SigLIP2 nonsensical outputs are different issues." }, { "left": "issue:43825", "right": "issue:43827", "accept": false, - "reason": "One is a runtime error message, the other is stale docs; related topic but not the same bug." + "reason": "Related only at the documentation/pipeline-removal theme; one is an error message bug, the other is stale docs." }, { "left": "issue:44186", "right": "issue:44442", "accept": false, - "reason": "LayoutLMv2 NER padding/truncation crash and FastSpeech2Conformer tokenizer loading are different tokenizer issues." + "reason": "LayoutLMv2Tokenizer NER/padding crash and FastSpeech2ConformerTokenizer loading failure are unrelated tokenizer issues." }, { "left": "issue:44821", "right": "issue:45084", "accept": false, - "reason": "AutoImageProcessor URL loading and template compilation errors are unrelated." + "reason": "AutoImageProcessor URL loading and template-node compilation errors do not share the same failure mode." }, { "left": "issue:43650", "right": "issue:43906", "accept": false, - "reason": "Too little concrete overlap; one is a vague 'ADD THE DATA' issue and the other is a reproduction of a different linked report." + "reason": "A low-content placeholder issue and an isolated reproduction of another issue are not duplicates." }, { "left": "issue:41720", "right": "issue:44991", "accept": false, - "reason": "Qwen3 auto device mapping CUDA assert and est-roberta tokenizer loading are unrelated." + "reason": "Different symptoms and subsystems: Qwen3 device-mapping CUDA assert versus tokenizer loading for EMBEDDIA/est-roberta." }, { "left": "issue:44805", "right": "issue:44936", "accept": false, - "reason": "Mask-shape IndexError and trainer.evaluate-after-train failure are distinct issues." + "reason": "Mask-shape IndexError and trainer.evaluate after train are not the same bug." }, { "left": "issue:30064", "right": "issue:36331", "accept": false, - "reason": "Void segmentation map processing and CustomTrainer compute_loss kwargs are unrelated." + "reason": "Image processor void segmentation map handling is unrelated to CustomTrainer.compute_loss keyword-arg incompatibility." }, { - "left": "issue:43606", - "right": "issue:45072", + "left": "issue:44485", + "right": "issue:44509", "accept": false, - "reason": "CPU offload device mismatch in Bark-small and bf16 dtype mismatches in wrappers are different bugs." + "reason": "GLM-5 RoPE discussion and docs mentioning removed pipeline tasks are different topics." }, { - "left": "issue:44485", - "right": "issue:44509", + "left": "issue:43606", + "right": "issue:45072", "accept": false, - "reason": "GLM-5 RoPE discussion and docs still mentioning removed pipeline tasks are unrelated." + "reason": "CPU offload device mismatch and bfloat16 dtype mismatch are separate inference bugs." }, { "left": "issue:42175", "right": "issue:45081", "accept": false, - "reason": "TensorFlow backend installation and Mistral regex patch crash are unrelated." + "reason": "TensorFlow extra installation issue and Mistral regex patch crash are unrelated." }, { "left": "issue:43782", "right": "issue:45325", "accept": false, - "reason": "Different Qwen model families and different bugs: weight_only loading vs still-image RoPE scaling." + "reason": "Both are Qwen VL issues, but one is a weight_only loading error and the other is a rope-index scaling bug; not the same change." }, { "left": "issue:44610", "right": "issue:45072", "accept": false, - "reason": "Processor image-size mismatch and bf16 dtype mismatch are not the same code-path problem." + "reason": "Processor image-size mismatch and dtype mismatch in inference are different defects." }, { "left": "issue:34634", @@ -4525,43 +4467,31 @@ "left": "issue:44060", "right": "issue:44368", "accept": false, - "reason": "Both mention tied-weights warnings, but they involve different Qwen variants and different triggering conditions; not clearly the same concrete bug." + "reason": "Both mention tied-weights warnings around Qwen models, but one is about an incorrect internal weight tie and the other is a user-visible warning during LoRA fine-tuning." }, { "left": "issue:45081", "right": "issue:45406", "accept": false, - "reason": "Both are processor/tokenizer attribute errors, but on different processors and different missing attributes; not mergeable as one fix." + "reason": "Mistral regex patching crash and Gemma4Processor missing _tokenizer in transformers serve are different processor bugs." }, { "left": "issue:43720", "right": "issue:45072", "accept": false, - "reason": "Packed-weight unpacking in BitNet and dtype mismatches in other models are unrelated." + "reason": "BitNet packed-weight unpacking and bfloat16 dtype mismatches are not the same underlying problem." }, { "left": "issue:43262", "right": "issue:43819", "accept": false, - "reason": "Audio chat-template sampling-rate defaulting and DAC latent mismatch are unrelated." + "reason": "Audio sampling-rate defaulting and DAC latent reconstruction mismatch are unrelated." }, { "left": "issue:43653", "right": "issue:44671", "accept": false, - "reason": "BigBirdTokenizer special-token registration and CamemBERT masked LM output regression are different model/tokenizer bugs." - }, - { - "left": "issue:43582", - "right": "issue:44857", - "accept": false, - "reason": "Apple Silicon caching_allocator_warmup TypeError and LwDetrImageLoss AMP crash are unrelated." - }, - { - "left": "issue:43749", - "right": "issue:44291", - "accept": false, - "reason": "Both concern loading, but FSDP CPU RAM efficient loading and init_empty_weights with unexpected _is_hf_initialized are distinct failures." + "reason": "Tokenizer special-token registration and CamemBERT masked-LM prediction regression are different issues." } ] }, @@ -4580,18 +4510,16 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ - "issue:34567", "issue:34634", "issue:36010", "issue:36331", "issue:38175", "issue:39692", - "issue:40990", "issue:43257", "issue:43262", "issue:43329", @@ -4601,6 +4529,7 @@ "issue:43531", "issue:43550", "issue:43577", + "issue:43582", "issue:43638", "issue:43645", "issue:43653", @@ -4614,6 +4543,7 @@ "issue:43854", "issue:43927", "issue:43931", + "issue:44291", "issue:44479", "issue:44554", "issue:44568", @@ -4621,24 +4551,25 @@ "issue:44610", "issue:44756", "issue:44811", + "issue:44857", "issue:44912", "issue:44936", "issue:44977", "issue:44991", "issue:45042", - "issue:45072", "issue:45084", "issue:45356", - "issue:45357", "issue:45488", "issue:45538", "issue:45600" ], "soft_pairs": [ + "issue:43749|issue:44291", + "issue:43582|issue:44857", "issue:43526|issue:43697", "issue:43526|issue:44554", - "issue:43329|issue:45600", "issue:45084|issue:45538", + "issue:43329|issue:45600", "issue:43653|issue:44912", "issue:43454|issue:43577", "issue:44568|issue:45488", @@ -4647,14 +4578,14 @@ "issue:43452|issue:44991", "issue:43257|issue:43329", "issue:43746|issue:44811", - "issue:43854|issue:44977", "issue:38175|issue:44610", + "issue:43854|issue:44977", "issue:44589|issue:45042", "issue:43257|issue:43696", "issue:39692|issue:44912", "issue:44568|issue:45356", - "issue:43819|issue:43927", "issue:36010|issue:43645", + "issue:43819|issue:43927", "issue:43720|issue:44610", "issue:43262|issue:44811", "issue:43550|issue:44912", @@ -4663,237 +4594,228 @@ "issue:43329|issue:44479", "issue:43257|issue:43931", "issue:43257|issue:43756", - "issue:36010|issue:36331", - "issue:34567|issue:40990", - "issue:43697|issue:45357", - "issue:43577|issue:45072" + "issue:36010|issue:36331" ], - "prepared_review_unit_hash": "ef7779cfb1cd1971437635f94a1811ef8055961ae7b6f7c8bb412c2ebf9f55eb", + "prepared_review_unit_hash": "6b59f64fc4b00194160d74c2c96dc296206bbeb3c9a663b8a7af88ab77831f0e", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 37525, - "estimated_input_tokens": 9382, - "estimated_eval_tokens": 19020 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36684, + "estimated_input_tokens": 9171, + "estimated_eval_tokens": 18598 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a broad grab-bag of mostly unrelated closed issues across training, tokenizers, multimodal processing, loading, and model-specific regressions. I only see one plausible duplicate-style soft edge: the two video-regression reports around Qwen/VLM multimodal token counting. Everything else looks like distinct bugs.", - "confidence": 0.79, - "canonical_issue_reason": "No single canonical issue stands out because the cluster is heterogeneous; only issue:43329 has a clear duplicate-like counterpart in the soft pairs.", + "summary": "This set is mostly a grab bag of unrelated Transformers regressions across loading, tokenization, processors, and training. I\u2019d only merge one soft edge: the Qwen video-input regression pair (43329, 44479). Everything else looks like broad thematic similarity rather than the same underlying bug.", + "confidence": 0.37, + "canonical_issue_reason": "issue:43329 is the most concrete root-cause report here: it points to the broken video branch in `_get_num_multimodal_tokens` with undefined helpers/variables, which is the clearest diagnostic match for the only plausible duplicate pair.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43329 is the strongest representative because it describes a concrete, reproducible multimodal video-path bug and is the only item that plausibly matches another report in the cluster.", + "best_issue_reason": "issue:43329 is the best representative of the cluster\u2019s only real duplicate-like subcase because it is specific, actionable, and pinpoints the failing code path rather than just describing user-visible symptoms.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43749", + "right": "issue:44291", + "accept": false, + "reason": "Both involve model loading, but one is FSDP CPU RAM efficient loading and the other is an `init_empty_weights` argument mismatch; different failure points and fixes." + }, + { + "left": "issue:43582", + "right": "issue:44857", + "accept": false, + "reason": "Apple Silicon allocator warmup TypeError vs CUDA float16 AMP loss crash; unrelated platform and subsystem." + }, { "left": "issue:43526", "right": "issue:43697", "accept": false, - "reason": "Different subsystems and failure modes: Beit image label reduction vs RTDetrV2 output drift." + "reason": "BEiT `reduce_labels` fast-processor bug and RTDetrV2 output drift are both vision-related, but they are different code paths and symptoms." }, { "left": "issue:43526", "right": "issue:44554", "accept": false, - "reason": "Unrelated bugs: image processor label handling vs an attention correctness issue on MPS." + "reason": "BEiT image-label handling vs an MPS attention correctness issue; no shared bug." }, { - "left": "issue:43329", - "right": "issue:45600", + "left": "issue:45084", + "right": "issue:45538", "accept": false, - "reason": "Different code paths: multimodal video token counting vs auto-mapping references to removed configs." + "reason": "Template-node compile failure and CLIP tokenizer max-length handling are unrelated." }, { - "left": "issue:45084", - "right": "issue:45538", + "left": "issue:43329", + "right": "issue:45600", "accept": false, - "reason": "No overlap in underlying bug: compilation error vs tokenizer max_length default behavior." + "reason": "Video multimodal token counting bug vs stale `auto_mappings.py` config references; different parts of the library." }, { "left": "issue:43653", "right": "issue:44912", "accept": false, - "reason": "Tokenizer special-token registration is unrelated to MXFP4 quantization loading." + "reason": "BigBird special-token registration and GPT-oss MXFP4 loading behavior are unrelated." }, { "left": "issue:43454", "right": "issue:43577", "accept": false, - "reason": "Different model families and issues: lm_head tying vs dtype propagation in BLIP2." + "reason": "AyaVision lm_head tying and BLIP-2 dtype propagation are different model-initialization bugs." }, { "left": "issue:44568", "right": "issue:45488", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different models and different mechanisms." + "reason": "Both are tokenizer regressions, but one is special-token/BOS-EOS handling and the other is pre-tokenizer override behavior." }, { "left": "issue:34634", "right": "issue:36331", "accept": false, - "reason": "BarkProcessor voice_preset failure is unrelated to CustomTrainer compute_loss keyword handling." + "reason": "BarkProcessor voice_preset and CustomTrainer `compute_loss` signature are unrelated." }, { "left": "issue:43638", "right": "issue:44936", "accept": false, - "reason": "Different trainer failures: ZeRO-3/non-pretrained BERT indexing vs evaluate-after-train behavior." + "reason": "Zero3 training index error vs evaluate-after-train failure; different trainer issues." }, { "left": "issue:43452", "right": "issue:44991", "accept": false, - "reason": "Both involve loading/tokenizers, but one is gguf_file + auto classes and the other is a specific tokenizer regression; not the same bug." + "reason": "GGUF loading breaks for auto classes, while the other is a tokenizer-loading regression for a specific checkpoint; same broad area only." }, { "left": "issue:43257", "right": "issue:43329", "accept": false, - "reason": "Qwen3 MOE weight conversion/load issue is unrelated to multimodal video token counting." + "reason": "Qwen3 MoE weight conversion under accelerate/deepspeed is unrelated to the video multimodal token-count bug." }, { "left": "issue:43746", "right": "issue:44811", "accept": false, - "reason": "PEFT checkpoint loading for GraniteSpeech is unrelated to Whisper batch_decode special-token handling." + "reason": "PEFT checkpoint loading and Whisper batch decode skip-special-tokens are different audio/model-loading bugs." }, { - "left": "issue:43854", - "right": "issue:44977", + "left": "issue:38175", + "right": "issue:44610", "accept": false, - "reason": "Different model-specific inference/load problems: GLM-4.7-Flash loading vs Qwen3.5 flash-attention generation." + "reason": "Unexpected zero probabilities in SigLIP2 vs OmDet-Turbo processor size mismatch; both vision issues, but not the same bug." }, { - "left": "issue:38175", - "right": "issue:44610", + "left": "issue:43854", + "right": "issue:44977", "accept": false, - "reason": "SigLIP2 zero probabilities and OmDet-Turbo input-size mismatch are distinct vision bugs." + "reason": "GLM-4.7-Flash load test failure vs Qwen3.5 flash-attention generation regression; different models and code paths." }, { "left": "issue:44589", "right": "issue:45042", "accept": false, - "reason": "Float8 storage lookup error is unrelated to PIL backend requiring torchvision." + "reason": "Float8 storage lookup during loading vs PIL backend requiring torchvision; unrelated." }, { "left": "issue:43257", "right": "issue:43696", "accept": false, - "reason": "Qwen3 MOE conversion/loading differs from a CUDA OOM on GPT-oss-20b." + "reason": "Qwen3 MoE weight conversion and GPT-oss OOM are not the same failure." }, { "left": "issue:39692", "right": "issue:44912", "accept": false, - "reason": "Documentation example errors are unrelated to MXFP4 quantization loading." + "reason": "SigLIP2 doc-example/model-processor mismatch and GPT-oss MXFP4 fallback are different quantization/model-loading problems." }, { "left": "issue:44568", "right": "issue:45356", "accept": false, - "reason": "Both are tokenizer-related, but they concern different tokenizers and different regression mechanisms." + "reason": "Special-token addition regression vs Kimi-K2.5 codec-handling regression; distinct tokenizer bugs." }, { - "left": "issue:43819", - "right": "issue:43927", + "left": "issue:36010", + "right": "issue:43645", "accept": false, - "reason": "DAC forward/from_latents mismatch is unrelated to DiaConfig save/load token ID loss." + "reason": "GenerationMixin import breakage and custom-model notebook initialization failure are related to v5 API changes but not the same bug." }, { - "left": "issue:36010", - "right": "issue:43645", + "left": "issue:43819", + "right": "issue:43927", "accept": false, - "reason": "GenerationMixin import failure and Jupyter custom-model init issues are distinct API breakages." + "reason": "DAC `from_latents` mismatch and DiaConfig save/load token-ID loss are unrelated." }, { "left": "issue:43720", "right": "issue:44610", "accept": false, - "reason": "BitNet packed-weight loading is unrelated to OmDet-Turbo processor size mismatch." + "reason": "BitNet packed-weight loading and OmDet processor image-size mismatch are different problems." }, { "left": "issue:43262", "right": "issue:44811", "accept": false, - "reason": "Audio chat-template sampling-rate default is unrelated to Whisper batch_decode skipping special tokens." + "reason": "Audio chat-template sample-rate default and Whisper batch decode skip-special-tokens are distinct processor bugs." }, { "left": "issue:43550", "right": "issue:44912", "accept": false, - "reason": "torch.compile/SDPA failure is unrelated to MXFP4 quantization fallback." + "reason": "Bamba torch.compile/SDPA failure vs GPT-oss quantization fallback are unrelated." }, { "left": "issue:43749", "right": "issue:44756", "accept": false, - "reason": "FSDP CPU RAM-efficient loading and Strix Halo mmap/OOM are separate memory-loading issues." + "reason": "Both mention loading/OOM, but one is FSDP CPU RAM efficient loading and the other is mmap on Strix Halo; different root causes." }, { "left": "issue:43531", "right": "issue:45042", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior is unrelated to PIL backend torchvision requirements." + "reason": "Qwen3-MoE sliding window behavior and PIL torchvision dependency handling are unrelated." }, { "left": "issue:43329", "right": "issue:44479", "accept": true, - "reason": "These both point to the same video-input/multimodal regression path in Qwen/VLM processing, with 43329 naming the broken video branch directly." + "reason": "These look like the same video-input regression in Qwen VLM processors; 43329 identifies the concrete broken video token-counting path, while 44479 reports the broader symptom across affected models." }, { "left": "issue:43257", "right": "issue:43931", "accept": false, - "reason": "Qwen3 MOE loading/weight conversion is unrelated to Qwen3-VL shape mismatch." + "reason": "Qwen3 MoE weight conversion and Qwen3-VL shape mismatch are different loading bugs." }, { "left": "issue:43257", "right": "issue:43756", "accept": false, - "reason": "Different model families and bugs: MOE loading vs Smollm3 RoPE-layer configuration." + "reason": "Qwen3 MoE weight conversion is unrelated to Smollm3 RoPE-layer dropping." }, { "left": "issue:36010", "right": "issue:36331", "accept": false, - "reason": "Importing GenerationMixin and CustomTrainer.compute_loss keyword mismatch are unrelated failures." - }, - { - "left": "issue:34567", - "right": "issue:40990", - "accept": false, - "reason": "Trainer input-token counter bug is unrelated to GPT-oss perplexity behavior." - }, - { - "left": "issue:43697", - "right": "issue:45357", - "accept": false, - "reason": "RTDetrV2 output drift and Qwen3.5 visual-encoder save_pretrained regression are separate model-specific bugs." - }, - { - "left": "issue:43577", - "right": "issue:45072", - "accept": false, - "reason": "BLIP2 dtype propagation and SwitchTransformers/TimmWrapperModel bfloat16 mismatches are different loading/inference bugs." + "reason": "GenerationMixin import failure and `compute_loss` signature mismatch are unrelated." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet: it stays conservative about the cluster being heterogeneous, and it only endorses one soft pair that is plausibly the same multimodal video-path regression. No soft-edge verdict overstates the evidence." + "feedback": "The summary is grounded in the packet and stays conservative. The single accepted soft edge (43329, 44479) is plausibly the same video-input regression, and the rejected pairs are reasonably distinguished by different failure modes or code paths. No clear overstatement of evidence." }, "evaluator_used": true, "retried": false, @@ -4909,8 +4831,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -4920,6 +4842,7 @@ "issue:34689", "issue:36010", "issue:38175", + "issue:40990", "issue:41720", "issue:42175", "issue:42915", @@ -4927,20 +4850,18 @@ "issue:43329", "issue:43450", "issue:43452", - "issue:43493", "issue:43526", "issue:43540", + "issue:43577", "issue:43582", - "issue:43643", "issue:43646", + "issue:43697", "issue:43742", "issue:43749", - "issue:43756", "issue:43782", "issue:43824", "issue:43906", "issue:43927", - "issue:43994", "issue:44062", "issue:44291", "issue:44479", @@ -4954,9 +4875,9 @@ "issue:44987", "issue:44991", "issue:45042", + "issue:45072", "issue:45137", - "issue:45325", - "issue:45375", + "issue:45357", "issue:45381", "issue:45405", "issue:45479", @@ -4964,261 +4885,254 @@ "issue:45588" ], "soft_pairs": [ + "issue:34567|issue:40990", + "issue:43697|issue:45357", "issue:43329|issue:43450", - "issue:43452|issue:43582", - "issue:41720|issue:43540", + "issue:43577|issue:45072", "issue:44479|issue:45381", - "issue:42175|issue:44938", + "issue:41720|issue:43540", + "issue:43452|issue:43582", "issue:44987|issue:45538", + "issue:42175|issue:44938", "issue:43317|issue:45538", "issue:43927|issue:44912", - "issue:43540|issue:44560", "issue:44291|issue:45042", - "issue:33357|issue:34567", "issue:43906|issue:45381", + "issue:43540|issue:44560", + "issue:33357|issue:34567", "issue:43540|issue:44479", "issue:42915|issue:45538", "issue:42915|issue:43742", "issue:42175|issue:43824", - "issue:44737|issue:45588", "issue:43526|issue:45479", "issue:42915|issue:44821", - "issue:42175|issue:44488", "issue:44062|issue:44991", + "issue:44737|issue:45588", + "issue:42175|issue:44488", "issue:45137|issue:45405", "issue:38175|issue:44898", - "issue:43582|issue:43782", "issue:43646|issue:43749", + "issue:43582|issue:43782", "issue:28282|issue:34689", - "issue:36010|issue:42175", - "issue:43643|issue:45375", - "issue:43756|issue:45325", - "issue:43742|issue:44291", - "issue:43493|issue:43994" + "issue:36010|issue:42175" ], - "prepared_review_unit_hash": "0df656a80f63ba86672f02b6dfe8e0bfbe247e97fcac64ac5c0af9ed87387132", + "prepared_review_unit_hash": "f39cab3844e68597ebb04904a026187ed22343b431e2f29d68f78eb151c6ea55", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 31, - "serialized_chars": 37106, - "estimated_input_tokens": 9277, - "estimated_eval_tokens": 18810 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36212, + "estimated_input_tokens": 9053, + "estimated_eval_tokens": 18362 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The packet is mostly heterogeneous, but there are two solid duplicate subclusters: the Qwen video-input regression and the SigLIP2 discrepancy reports. The tokenizer pair `issue:44062` / `issue:44991` is not strong enough to merge on the evidence shown and is rejected conservatively.", - "confidence": 0.64, - "canonical_issue_reason": "Issue 44479 is the broadest, most explicit representative of the Qwen video-input regression subcluster, covering multiple affected Qwen VL variants and serving as the best anchor among the candidate issues.", + "summary": "Mostly unrelated issue reports with one plausible duplicate pair around the Qwen video-input regression; no PRs are present.", + "confidence": 0.44, + "canonical_issue_reason": "Issue 44479 is the broadest and clearest report: it describes a v5.3.0 video-input regression across multiple Qwen model families, which likely subsumes the more specific qwen2.5-vl manifestation.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 44479 is the strongest global issue candidate because it is the most central and broadly scoped duplicate target in the packet; the other items are either narrower repros or unrelated bugs.", + "best_issue_reason": "Issue 44479 is the best representative issue because it captures the shared regression surface and is broad enough to anchor the related Qwen video-input discussion.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43329", - "right": "issue:43450", + "left": "issue:34567", + "right": "issue:40990", "accept": false, - "reason": "Both are video-related, but one is an undefined-variable bug in multimodal token counting and the other is a batched-shape bug in video processors." + "reason": "Unrelated: training-state metric update vs perplexity evaluation behavior." }, { - "left": "issue:43452", - "right": "issue:43582", + "left": "issue:43697", + "right": "issue:45357", "accept": false, - "reason": "Completely different areas: gguf/tokenizer loading vs Apple Silicon allocator warmup." + "reason": "Different bugs and model paths: RTDetr output drift vs Qwen3.5 save_pretrained key regression." }, { - "left": "issue:41720", - "right": "issue:43540", + "left": "issue:43329", + "right": "issue:43450", + "accept": false, + "reason": "Both involve video/multimodal code, but one is an undefined-variable token-counting bug and the other is a batched processor shape issue." + }, + { + "left": "issue:43577", + "right": "issue:45072", "accept": false, - "reason": "Different failure modes and code paths: auto device-mapping CUDA assert vs Qwen3OmniMoe video-input validation." + "reason": "Both mention dtype, but they affect different models and different failure modes." }, { "left": "issue:44479", "right": "issue:45381", "accept": true, - "reason": "Same underlying Qwen VL video-position-id regression; 45381 is a narrower repro of the broader 44479 video-input breakage." + "reason": "Likely same underlying Qwen video-input regression: 44479 is the umbrella report, while 45381 is a specific qwen2.5-vl symptom around video position IDs." }, { - "left": "issue:42175", - "right": "issue:44938", + "left": "issue:41720", + "right": "issue:43540", "accept": false, - "reason": "Unrelated packaging/backend issue vs Python 3.14 load failure." + "reason": "Different concrete failures: auto device mapping CUDA assert vs video-processing ValueError in a different Qwen model." + }, + { + "left": "issue:43452", + "right": "issue:43582", + "accept": false, + "reason": "No shared code path: gguf loading regression vs Apple Silicon allocator warmup type error." }, { "left": "issue:44987", "right": "issue:45538", "accept": false, - "reason": "Tokenizer/model loading failure and CLIP max-length constant issue are different bugs." + "reason": "Unrelated: model-loading failure for a specific repo vs CLIP tokenizer max-length semantics." + }, + { + "left": "issue:42175", + "right": "issue:44938", + "accept": false, + "reason": "Packaging/backend availability issue vs a Python 3.14 loading failure; too generic to treat as duplicates." }, { "left": "issue:43317", "right": "issue:45538", "accept": false, - "reason": "Unrelated device-map/offload bug vs tokenizer configuration bug." + "reason": "Different areas: device_map/offload loading bug vs tokenizer max-length constant." }, { "left": "issue:43927", "right": "issue:44912", "accept": false, - "reason": "Different code paths: config save/load losing token IDs vs quantization fallback to bf16." + "reason": "Different model/config bugs: DiaConfig persistence vs MXFP4 quantization fallback." }, { - "left": "issue:43540", - "right": "issue:44560", + "left": "issue:44291", + "right": "issue:45042", "accept": false, - "reason": "Both concern Qwen video inputs, but the model variants and concrete errors differ too much to treat as the same bug." + "reason": "Different load-time API mismatch vs image-processor torchvision dependency regression." }, { - "left": "issue:44291", - "right": "issue:45042", + "left": "issue:43906", + "right": "issue:45381", "accept": false, - "reason": "Unrelated issues: init_empty_weights argument handling vs PIL backend torchvision dependency." + "reason": "A reproduction of another issue, but not the same concrete bug as the qwen2.5-vl video-position-ID regression." }, { - "left": "issue:33357", - "right": "issue:34567", + "left": "issue:43540", + "right": "issue:44560", "accept": false, - "reason": "MacOS bus error on CLIP vs TrainerState token-count tracking are unrelated." + "reason": "Both are video-related, but they describe different models and distinct observed failures." }, { - "left": "issue:43906", - "right": "issue:45381", + "left": "issue:33357", + "right": "issue:34567", "accept": false, - "reason": "No clear evidence these are the same bug; one is an isolated reproduction of a different issue." + "reason": "Unrelated Mac bus error in CLIP vs TrainerState token-count tracking." }, { "left": "issue:43540", "right": "issue:44479", "accept": false, - "reason": "Related theme, but not the same concrete bug: different Qwen video code paths and different symptoms." + "reason": "Related theme, but not enough to confirm the same code-path bug; one is a specific model ValueError and the other is a broader regression report." }, { "left": "issue:42915", "right": "issue:45538", "accept": false, - "reason": "FP8 config failure and CLIP tokenizer max-length bug are unrelated." + "reason": "Different categories: FP8 config loading vs tokenizer max-length behavior." }, { "left": "issue:42915", "right": "issue:43742", "accept": false, - "reason": "Different problems: FineGrainedFP8Config loading vs MobileLLM key error." + "reason": "Different failures in different model-loading paths." }, { "left": "issue:42175", "right": "issue:43824", "accept": false, - "reason": "Packaging/backend installation issue vs missing Qwen class import." - }, - { - "left": "issue:44737", - "right": "issue:45588", - "accept": false, - "reason": "XLNet CPU allocation bug and flash-attention null-aux crash are unrelated." + "reason": "Packaging/backend issue vs missing Qwen2.5-VL class import." }, { "left": "issue:43526", "right": "issue:45479", "accept": false, - "reason": "Different model families and failure modes: BEiT label reduction vs degenerate zero loss in sequence classification." + "reason": "Different model families and different functional bugs." }, { "left": "issue:42915", "right": "issue:44821", "accept": false, - "reason": "FP8 config bug and AutoImageProcessor URL loading are unrelated." + "reason": "Unrelated: FP8 Qwen3Moe failure vs AutoImageProcessor URL loading." }, { - "left": "issue:42175", - "right": "issue:44488", + "left": "issue:44062", + "right": "issue:44991", "accept": false, - "reason": "Backend installation issue vs model/tokenizer loading failure." + "reason": "Tokenizer-added-token argument bug vs tokenizer-loading regression for a specific model." }, { - "left": "issue:44062", - "right": "issue:44991", + "left": "issue:44737", + "right": "issue:45588", + "accept": false, + "reason": "XLNet CPU placement bug vs flash-attention sinkless-model AttributeError." + }, + { + "left": "issue:42175", + "right": "issue:44488", "accept": false, - "reason": "One side is a specific AddedToken TypeError and the other is a generic tokenizer-loading failure; the available evidence does not show the same underlying regression." + "reason": "Backend/install packaging issue vs model-loading failure for cjvt/sleng-bert." }, { "left": "issue:45137", "right": "issue:45405", "accept": false, - "reason": "DeepSpeed ZeRO3 deque error and PEFT version pin issue are unrelated." + "reason": "DeepSpeed ZeRO3 runtime error vs dependency version bump; unrelated." }, { "left": "issue:38175", "right": "issue:44898", "accept": false, - "reason": "Different model/feature areas: SigLIP2 output anomaly vs Perceiver positional-encoding interpolation." + "reason": "Different vision-model bugs with different symptoms and code paths." }, { - "left": "issue:43582", - "right": "issue:43782", + "left": "issue:43646", + "right": "issue:43749", "accept": false, - "reason": "Apple Silicon TypeError and Qwen3VL weight_only loading error are different bugs." + "reason": "Transformers 5 custom init breakage vs FSDP CPU RAM efficient loading regression." }, { - "left": "issue:43646", - "right": "issue:43749", + "left": "issue:43582", + "right": "issue:43782", "accept": false, - "reason": "General custom model init breakage vs FSDP CPU RAM efficient loading are not the same concrete failure." + "reason": "Apple Silicon allocator warmup type error vs Qwen3VL weight-only load error." }, { "left": "issue:28282", "right": "issue:34689", "accept": false, - "reason": "Missing PyTorch import and Llama 3.2 Vision loading regression are unrelated." + "reason": "Missing PyTorch import vs Llama 3.2 Vision loading breakage; not the same issue." }, { "left": "issue:36010", "right": "issue:42175", "accept": false, - "reason": "GenerationMixin import failure and TensorFlow backend packaging are unrelated." - }, - { - "left": "issue:43643", - "right": "issue:45375", - "accept": false, - "reason": "Both involve config fields, but the titles indicate different mechanisms and no clear same-bug evidence." - }, - { - "left": "issue:43756", - "right": "issue:45325", - "accept": false, - "reason": "RoPE-layer omission and temporal position-id scaling are different bugs." - }, - { - "left": "issue:43742", - "right": "issue:44291", - "accept": false, - "reason": "MobileLLM key error and init_empty_weights argument mismatch are unrelated." - }, - { - "left": "issue:43493", - "right": "issue:43994", - "accept": true, - "reason": "Same SigLIP2 model family and same underlying incorrect-output discrepancy; 43994 is a user-facing repro of the HF-vs-JAX mismatch." + "reason": "Import path regression for GenerationMixin vs missing TensorFlow backend in an install extra." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The cluster summary is mostly grounded, but one soft-edge accept is too strong for the packet: `issue:44062` / `issue:44991` only shows a specific AddedToken TypeError on one side and a generic tokenizer-loading failure on the other, so calling them the same underlying regression goes beyond the evidence available here. The Qwen video and SigLIP2 overlaps are plausible, but that pair makes the verdict non-conservative." + "accept": true, + "feedback": "Grounded overall: the canonical issue choice is supported by the packet, no PRs are present, and the one accepted soft-edge pair (44479/45381) is stated cautiously. The rejected pairs are generally conservative and do not overclaim duplication." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -5231,8 +5145,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -5241,8 +5155,8 @@ "issue:34567", "issue:36331", "issue:41762", - "issue:43295", "issue:43452", + "issue:43493", "issue:43531", "issue:43577", "issue:43643", @@ -5252,11 +5166,11 @@ "issue:43720", "issue:43742", "issue:43749", + "issue:43756", "issue:43819", + "issue:43994", "issue:44062", "issue:44291", - "issue:44464", - "issue:44466", "issue:44484", "issue:44488", "issue:44514", @@ -5265,7 +5179,6 @@ "issue:44589", "issue:44625", "issue:44704", - "issue:44743", "issue:44792", "issue:44811", "issue:44821", @@ -5274,20 +5187,24 @@ "issue:44898", "issue:44936", "issue:44964", - "issue:45005", "issue:45042", "issue:45081", "issue:45216", "issue:45290", - "issue:45356", + "issue:45325", + "issue:45375", "issue:45381", "issue:45405", "issue:45440", "issue:45600" ], "soft_pairs": [ - "issue:28282|issue:29127", + "issue:43643|issue:45375", + "issue:43756|issue:45325", + "issue:43742|issue:44291", "issue:44484|issue:44704", + "issue:43493|issue:43994", + "issue:28282|issue:29127", "issue:34567|issue:36331", "issue:44062|issue:45081", "issue:44291|issue:44821", @@ -5295,13 +5212,13 @@ "issue:44792|issue:44964", "issue:44514|issue:45290", "issue:45405|issue:45600", - "issue:43688|issue:43749", "issue:43643|issue:44561", + "issue:43688|issue:43749", "issue:44514|issue:45381", - "issue:43452|issue:43577", "issue:43697|issue:45216", - "issue:43742|issue:44589", + "issue:43452|issue:43577", "issue:45042|issue:45081", + "issue:43742|issue:44589", "issue:43531|issue:44821", "issue:43688|issue:45440", "issue:43452|issue:44843", @@ -5309,211 +5226,211 @@ "issue:43653|issue:44568", "issue:43720|issue:44898", "issue:43819|issue:44811", - "issue:41762|issue:44936", - "issue:43295|issue:45356", - "issue:44466|issue:45005", - "issue:44464|issue:44743" + "issue:41762|issue:44936" ], - "prepared_review_unit_hash": "1900b5e6f3f8b4004d3f6d8d01e5d6234391e7e6efaff901060e88c7b1c4381a", + "prepared_review_unit_hash": "497a926599b4d3624f159cbaefa2e132eacd11606be0e0cf46edecd1917a48a5", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 27, - "serialized_chars": 36193, - "estimated_input_tokens": 9049, - "estimated_eval_tokens": 18354 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35845, + "estimated_input_tokens": 8962, + "estimated_eval_tokens": 18180 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most pairs are unrelated bug reports touching different code paths. The only plausible duplicate is the v5 tied-weights serialization/saving regression pair.", + "summary": "The cluster is mostly a loose set of unrelated Transformers bug reports that only share high-level loading/API vocabulary. I rejected all soft edges because the paired items do not describe the same concrete underlying bug or change.", "confidence": 0.62, - "canonical_issue_reason": "issue:45005 is the broader, more canonical description of the v5 tied-weights save/serialization regression; issue:44466 reads like a specific manifestation depending on model device.", + "canonical_issue_reason": "issue:43452 is the most broadly representative loading-related report in the set, spanning core `from_pretrained` behavior for multiple auto classes and fitting the loose thematic center of the cluster better than the more specialized reports.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45005 is the best global issue candidate because it captures the underlying tied-weights problem more generally and is easier to map to related symptoms.", + "best_issue_reason": "It is a clear, user-facing regression in a core loading path and is broad enough to act as the best single representative issue for this heterogeneous cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:28282", - "right": "issue:29127", + "left": "issue:43643", + "right": "issue:45375", + "accept": false, + "reason": "Both involve missing config fields, but they are different mechanisms on different models: `trust_remote_code` returning incomplete config vs a strict config dropping `deepstack_visual_indexes`." + }, + { + "left": "issue:43756", + "right": "issue:45325", + "accept": false, + "reason": "Both mention RoPE/position IDs, but the bugs are model-specific and distinct: Smollm3 dropping RoPE layers vs Qwen2.5-VL temporal position scaling." + }, + { + "left": "issue:43742", + "right": "issue:44291", "accept": false, - "reason": "Different failures: missing PyTorch dependency vs clearer LayoutLMv3 error messaging." + "reason": "One is a key error when loading MobileLLM, the other is an `init_empty_weights`/`_is_hf_initialized` TypeError. Different failure modes and code paths." }, { "left": "issue:44484", "right": "issue:44704", "accept": false, - "reason": "One is about max shard size defaults; the other is AutoProcessor failing to forward kwargs to cached_file." + "reason": "A save-pretrained sharding-size question is unrelated to `AutoProcessor` failing to forward kwargs to `cached_file`." + }, + { + "left": "issue:43493", + "right": "issue:43994", + "accept": false, + "reason": "Both are about SigLIP2, but one is a HF-vs-JAX implementation discrepancy and the other is bad outputs with AutoModel/pipeline; not the same concrete bug." + }, + { + "left": "issue:28282", + "right": "issue:29127", + "accept": false, + "reason": "Torch import failure for AutoModel is unrelated to unclear error messaging for insufficient boxes in LayoutLMv3." }, { "left": "issue:34567", "right": "issue:36331", "accept": false, - "reason": "TrainerState token counting and CustomTrainer.compute_loss keyword incompatibility are separate trainer issues." + "reason": "Trainer state tracking and a custom `compute_loss` signature break are separate trainer issues with different symptoms and fixes." }, { "left": "issue:44062", "right": "issue:45081", "accept": false, - "reason": "AddedToken special-argument duplication is unrelated to the Mistral regex backend_tokenizer crash." + "reason": "One is an `AddedToken` constructor conflict, the other is a Mistral regex patch crashing on tokenizer internals; they are different tokenizer bugs." }, { "left": "issue:44291", "right": "issue:44821", "accept": false, - "reason": "init_empty_weights/_is_hf_initialized and loading AutoImageProcessor from URL are different loading failures." + "reason": "`init_empty_weights` loading breakage and URL loading for `AutoImageProcessor` are unrelated." }, { "left": "issue:43452", "right": "issue:44488", "accept": false, - "reason": "Both are model loading problems, but the affected APIs and failure modes are not the same." + "reason": "Both are loading-related, but one is a GGUF-specific breakage in auto classes and the other is a model-specific inability to load `cjvt/sleng-bert`." }, { "left": "issue:44792", "right": "issue:44964", "accept": false, - "reason": "A failed Janus image-generation test is not the same bug as failing to load Phi-4 multimodal." + "reason": "Different multimodal models and different reports; no evidence they share the same underlying defect." }, { "left": "issue:44514", "right": "issue:45290", "accept": false, - "reason": "Both involve apply_chat_template, but one is batched padding behavior and the other is tool-call assistant message handling." + "reason": "Both crash in `apply_chat_template`, but the triggering conditions differ materially: batched input with `padding=False` vs assistant messages with tool calls and no content." }, { "left": "issue:45405", "right": "issue:45600", "accept": false, - "reason": "Unreleased PEFT minimum version and removed Sam3LiteText config references are unrelated CI/versioning issues." - }, - { - "left": "issue:43688", - "right": "issue:43749", - "accept": false, - "reason": "Auxiliary-loss normalization in MoE models is unrelated to broken FSDP CPU RAM-efficient loading." + "reason": "A PEFT version bump issue is unrelated to stale `auto_mappings` references to removed configs." }, { "left": "issue:43643", "right": "issue:44561", "accept": false, - "reason": "Missing fields from trust_remote_code AutoConfig and removal of is_torch_fx_available are different regressions." + "reason": "Both involve config/remote-code compatibility, but one is missing fields from `AutoConfig`, while the other is removal of `is_torch_fx_available` breaking remote-code models." }, { - "left": "issue:44514", - "right": "issue:45381", + "left": "issue:43688", + "right": "issue:43749", "accept": false, - "reason": "Different Qwen2.5-VL bugs: chat templating on batches vs wrong video vision_position_ids." + "reason": "Auxiliary-loss normalization and FSDP CPU RAM efficient loading are unrelated subsystems and failure modes." }, { - "left": "issue:43452", - "right": "issue:43577", + "left": "issue:44514", + "right": "issue:45381", "accept": false, - "reason": "gguf_file loading breakage is unrelated to BLIP2 dtype propagation staying float32." + "reason": "Both concern Qwen2.5-VL chat/template behavior, but one is a template crash on batched padding and the other is incorrect `vision_position_ids`; not the same bug." }, { "left": "issue:43697", "right": "issue:45216", "accept": false, - "reason": "RTDetrV2 output divergence and Qwen3.5 save_pretrained checkpoint correctness are different problems." + "reason": "RTDetrV2 output mismatch and Qwen3.5 checkpoint save corruption are different regression classes with different code paths." }, { - "left": "issue:43742", - "right": "issue:44589", + "left": "issue:43452", + "right": "issue:43577", "accept": false, - "reason": "MobileLLM key error and missing Float8 storage type are separate loading/storage issues." + "reason": "`gguf_file`/auto-class loading issues are unrelated to BLIP2 dtype propagation for `qformer`." }, { "left": "issue:45042", "right": "issue:45081", "accept": false, - "reason": "PIL/torchvision dependency regression is not the same as the Mistral regex tokenizer crash." + "reason": "Torchvision being incorrectly required for PIL backend image processors is unrelated to the Mistral regex/tokenizer crash." + }, + { + "left": "issue:43742", + "right": "issue:44589", + "accept": false, + "reason": "MobileLLM key error and a missing `Float8_e4m3fnStorage` type are separate loading/runtime failures." }, { "left": "issue:43531", "right": "issue:44821", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and AutoImageProcessor URL loading are unrelated." + "reason": "Qwen3-MoE sliding-window behavior and image processor loading from URL are unrelated." }, { "left": "issue:43688", "right": "issue:45440", "accept": false, - "reason": "Both are MoE-related, but auxiliary-loss normalization and DeepSeekV3 implementation divergence are different bugs." + "reason": "Both involve MoE-style models, but the reported defects are distinct: auxiliary-loss normalization vs implementation divergence from a remote model." }, { "left": "issue:43452", "right": "issue:44843", "accept": false, - "reason": "gguf_file loading failure is unrelated to model_info being called in offline mode." + "reason": "GGUF/auto-class loading and offline `model_info()` calls in `_patch_mistral_regex` are different loading regressions." }, { "left": "issue:44625", "right": "issue:44877", "accept": false, - "reason": "num_labels propagation and strict granite_speech config loading are separate config issues." + "reason": "Qwen3.5 `num_labels` propagation and strict config preventing `granite_speech` loading are separate config bugs." }, { "left": "issue:43653", "right": "issue:44568", "accept": false, - "reason": "BigBird mask-token special registration and mdeberta BOS/EOS insertion are different tokenizer bugs." + "reason": "Special-token registration causing empty decode output is different from `add_special_tokens=True` failing to add BOS/EOS." }, { "left": "issue:43720", "right": "issue:44898", "accept": false, - "reason": "BitNet packed-weight loading and Perceiver non-default resolution handling are unrelated." + "reason": "BitNet packed-weight loading and Perceiver non-default-resolution failure are unrelated." }, { "left": "issue:43819", "right": "issue:44811", "accept": false, - "reason": "DAC latent reconstruction mismatch and Whisper batch_decode skip_special_tokens behavior are different code paths." + "reason": "DAC latent decoding mismatch and Whisper `batch_decode` ignoring `skip_special_tokens` are different bugs in different components." }, { "left": "issue:41762", "right": "issue:44936", "accept": false, - "reason": "Gemma3 ZeRO-3 load failure and trainer.evaluate after train are not the same issue." - }, - { - "left": "issue:43295", - "right": "issue:45356", - "accept": false, - "reason": "Custom processor/tokenizer regression and Kimi-K2.5 codec/fix_mistral_regex regression are different tokenizer issues." - }, - { - "left": "issue:44466", - "right": "issue:45005", - "accept": true, - "reason": "Both describe the same v5 tied-weights/save_pretrained serialization regression; 45005 is the broader translation-model manifestation and 44466 is a device-dependent symptom." - }, - { - "left": "issue:44464", - "right": "issue:44743", - "accept": false, - "reason": "Chunked generation with compiled forward and Qwen3.5 recurrent state reset are distinct generation/runtime bugs." + "reason": "DeepSpeed ZeRO-3 loading for Gemma3 and `trainer.evaluate()` after `train()` are unrelated trainer/model-loading issues." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is well grounded in the packet and stays mostly conservative. The soft-edge verdicts generally separate clearly different issue titles/code paths, and the lone accepted pair is a plausible duplicate based on the shared tied-weights/v5 serialization theme. No obvious overstatement beyond the evidence in the titles." - }, - "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 }, @@ -5525,8 +5442,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -5543,7 +5460,6 @@ "issue:41720", "issue:43122", "issue:43262", - "issue:43278", "issue:43295", "issue:43329", "issue:43452", @@ -5554,7 +5470,6 @@ "issue:43577", "issue:43643", "issue:43653", - "issue:43688", "issue:43697", "issue:43720", "issue:43742", @@ -5564,15 +5479,17 @@ "issue:44442", "issue:44451", "issue:44464", + "issue:44466", "issue:44479", "issue:44554", - "issue:44610", "issue:44625", + "issue:44743", "issue:44843", "issue:44849", "issue:44936", "issue:44977", "issue:44991", + "issue:45005", "issue:45084", "issue:45325", "issue:45356", @@ -5580,6 +5497,9 @@ "issue:45588" ], "soft_pairs": [ + "issue:43295|issue:45356", + "issue:44466|issue:45005", + "issue:44464|issue:44743", "issue:41720|issue:43906", "issue:24643|issue:29127", "issue:43525|issue:43643", @@ -5597,249 +5517,253 @@ "issue:43577|issue:45588", "issue:30064|issue:33666", "issue:43577|issue:44554", - "issue:44464|issue:44977", "issue:43653|issue:43927", + "issue:44464|issue:44977", "issue:43531|issue:43742", "issue:43329|issue:45325", "issue:44479|issue:44625", "issue:43550|issue:44361", - "issue:41720|issue:44464", "issue:43452|issue:44451", "issue:43504|issue:43720", + "issue:41720|issue:44464", "issue:43504|issue:43550", - "issue:44849|issue:45357", "issue:43262|issue:45084", + "issue:44849|issue:45357", "issue:32090|issue:34567", - "issue:32090|issue:33666", - "issue:43504|issue:44610", - "issue:43278|issue:43688" + "issue:32090|issue:33666" ], - "prepared_review_unit_hash": "c7aadc8dc6bb4a8d4f8e3ad1b7007cb8775c8bdd11fefbdfe90ec8604c976f36", + "prepared_review_unit_hash": "2fd41c04be8e77d6954e741c61678b11eb23dbf55502cc321b32cf5f9dd76f89", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 33, - "serialized_chars": 37812, - "estimated_input_tokens": 9453, - "estimated_eval_tokens": 19162 + "soft_pair_count": 34, + "serialized_chars": 38189, + "estimated_input_tokens": 9548, + "estimated_eval_tokens": 19352 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a heterogeneous mix of unrelated Transformers issues: model loading, tokenizers, multimodal processors, training/TrainerState, dtype drift, and a few Qwen/BLIP/BEiT-specific regressions. The soft edges are mostly superficial text similarity, not true duplicates.", - "confidence": 0.96, - "canonical_issue_reason": "Issue 43504 is the closest thing to a hub because it is a concrete model-loading failure and sits near several other load-time regressions by wording, but the cluster is still too mixed to have a real duplicate canonical issue.", + "summary": "This set is heterogeneous: the soft links are mostly broad subsystem similarity (tokenizers, Qwen, tied weights, generation) rather than the same underlying bug. I would not merge any of these pairs as duplicates.", + "confidence": 0.95, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue 43504 is the least bad representative of the cluster\u2019s broad loading-regression theme, but it is not a safe global canonical because the linked issues span different models and code paths.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43295", + "right": "issue:45356", + "accept": false, + "reason": "Both are tokenizer/processor regressions, but they affect different code paths: processor.tokenizer/images handling vs Kimi-K2.5 codec and warning behavior." + }, + { + "left": "issue:44466", + "right": "issue:45005", + "accept": false, + "reason": "Both mention tied weights, but one is a device-dependent lm_head serialization bug and the other is a broader translation-model tied-weights issue." + }, + { + "left": "issue:44464", + "right": "issue:44743", + "accept": false, + "reason": "Different generation bugs: compiled-forward inconsistency vs recurrent state reset with cache and seq_len>1." + }, { "left": "issue:41720", "right": "issue:43906", "accept": false, - "reason": "Different bugs: Qwen3 auto device mapping CUDA assert vs an isolated reproduction of another issue." + "reason": "No clear match; one is a Qwen3 auto-device-map cuda assert, the other is an isolated reproduction of a different issue." }, { "left": "issue:24643", "right": "issue:29127", "accept": false, - "reason": "Unrelated symptoms and code paths: DeepSpeed weight shape error vs LayoutLMv3 box validation/message clarity." + "reason": "Completely different failures: DeepSpeed training weight-shape error vs LayoutLMv3 box-validation error message." }, { "left": "issue:43525", "right": "issue:43643", "accept": false, - "reason": "Both are config-related, but one is a missing pad_token_id on Llama4 and the other is trust_remote_code dropping fields; not the same bug." + "reason": "Both are config/field problems, but they concern different models and different missing-field behavior." }, { "left": "issue:43577", "right": "issue:43697", "accept": false, - "reason": "Different models and failures: BLIP2 dtype preservation vs RTDetrV2 output drift." + "reason": "Unrelated bugs: BLIP2 dtype not propagated vs RTDetrV2 output mismatch in v5." }, { "left": "issue:38175", "right": "issue:43295", "accept": false, - "reason": "SigLIP2 zero-probability issue and processor.tokenizer/images regression are different problems." + "reason": "Different models and symptoms: zero probabilities in SigLIP2 vs processor.tokenizer/images regression." }, { "left": "issue:44361", "right": "issue:44442", "accept": false, - "reason": "Different tokenizer failures: MLukeTokenizer task error vs FastSpeech2ConformerTokenizer load failure." + "reason": "Both are tokenizer-loading failures, but they are different models and different error modes." }, { "left": "issue:44843", "right": "issue:44991", "accept": false, - "reason": "Offline model_info() call in tokenizer patch vs EMBEDDIA/est-roberta loading regression; different causes." + "reason": "Both involve tokenizer loading, but one is an offline model_info call bug and the other is a model-specific v5 tokenizer load failure." }, { "left": "issue:43653", "right": "issue:44361", "accept": false, - "reason": "BigBird special-token decode bug is not the same as MLukeTokenizer task AttributeError." + "reason": "Different tokenizers and failures: BigBird special-token registration vs MLukeTokenizer attribute error." }, { "left": "issue:44479", "right": "issue:44936", "accept": false, - "reason": "Qwen video-input regression vs trainer.evaluate() failing after train(); unrelated." + "reason": "Different areas entirely: Qwen3.5 video regression vs Trainer evaluate-after-train behavior." }, { "left": "issue:43122", "right": "issue:43906", "accept": false, - "reason": "Version-to-version tokenization change is not the same as the isolated reproduction of another bug." + "reason": "Only loosely related through tokenization/versioning; the reported failures and targets are not the same bug." }, { "left": "issue:44977", "right": "issue:45356", "accept": false, - "reason": "Different surfaces: flash-attention generation failure in Qwen3.5 vs tokenizer codec handling regression in Kimi-K2.5." + "reason": "Different regressions: Qwen3.5 flash-attention generation vs Kimi-K2.5 tokenizer codec handling." }, { "left": "issue:28282", "right": "issue:32090", "accept": false, - "reason": "Missing PyTorch import error vs NoneType passed to _gpu_broadcast_one; unrelated." + "reason": "ImportError for missing PyTorch is unrelated to Trainer NoneType broadcast during distributed loading." }, { "left": "issue:28282", "right": "issue:33357", "accept": false, - "reason": "Environment import failure vs MacOS bus error in CLIP model loading; not the same issue." + "reason": "Different problems: missing PyTorch dependency vs MacOS bus error with a CLIP model." }, { "left": "issue:28282", "right": "issue:35141", "accept": false, - "reason": "No shared bug path: PyTorch missing import vs embedding reinitialization after resize." + "reason": "No shared bug; dependency import failure is unrelated to embedding resize/reinit behavior." }, { "left": "issue:43577", "right": "issue:45588", "accept": false, - "reason": "BLIP2 dtype mismatch and flash_attention.py crashing on s_aux=None are different code paths." + "reason": "Unrelated: BLIP2 dtype propagation vs flash_attention crash on s_aux=None." }, { "left": "issue:30064", "right": "issue:33666", "accept": false, - "reason": "Void segmentation-map processing and Qwen2-VL multi-GPU training are unrelated." + "reason": "Different subsystems and failures: image processor void segmentation maps vs Qwen2-VL multi-GPU training." }, { "left": "issue:43577", "right": "issue:44554", "accept": false, - "reason": "BLIP2 dtype bug is not the same as the MPS attention correctness issue." + "reason": "Different code paths: BLIP2 dtype issue vs MPS attention correctness issue when value head dim differs." }, { - "left": "issue:44464", - "right": "issue:44977", + "left": "issue:43653", + "right": "issue:43927", "accept": false, - "reason": "Chunked generation compiled-forward inconsistency vs Qwen3.5 flash-attention generation failure; different bugs." + "reason": "Both are tokenizer/config correctness issues, but the underlying bugs are different and not mergeable as one fix." }, { - "left": "issue:43653", - "right": "issue:43927", + "left": "issue:44464", + "right": "issue:44977", "accept": false, - "reason": "BigBird mask-token registration issue and DiaConfig custom-token-ID loss are different tokenizer/config bugs." + "reason": "Different generation regressions: compiled forward chunking vs flash-attention generation on Qwen3.5." }, { "left": "issue:43531", "right": "issue:43742", "accept": false, - "reason": "Qwen3-MoE sliding_window bug vs MobileLLM loading key error; unrelated." + "reason": "No overlap beyond being model-loading/runtime issues; sliding_window bug and MobileLLM key error are distinct." }, { "left": "issue:43329", "right": "issue:45325", "accept": false, - "reason": "Both touch multimodal/video logic, but one is undefined helpers in token counting and the other is rope-index scaling for still images." + "reason": "Both are Qwen multimodal/video-related, but one is an undefined-symbol video branch bug and the other is rope/position-id scaling." }, { "left": "issue:44479", "right": "issue:44625", "accept": false, - "reason": "Video-input regression vs num_labels propagation bug; different model parts." + "reason": "Different bugs in Qwen3.5: video input regression vs num_labels propagation." }, { "left": "issue:43550", "right": "issue:44361", "accept": false, - "reason": "torch.compile/SDPA failure in Bamba is unrelated to MLukeTokenizer AttributeError." - }, - { - "left": "issue:41720", - "right": "issue:44464", - "accept": false, - "reason": "CUDA assert in Qwen3 auto device mapping and chunked generation inconsistency are separate failures." + "reason": "Different components and failures: torch.compile/SDPA on Bamba vs tokenizer attribute error on MLuke." }, { "left": "issue:43452", "right": "issue:44451", "accept": false, - "reason": "gguf_file breaks tokenizer/model loading APIs, but ScandiBERT load failure is a different model-specific issue." + "reason": "Both are loading issues, but one is gguf_file handling in Auto* loaders and the other is a model-specific tokenizer load failure." }, { "left": "issue:43504", "right": "issue:43720", "accept": false, - "reason": "BEiT legacy-field load bug and BitNet packed-weight unpacking bug are both loading-related but not the same code path." + "reason": "Unrelated model-loading bugs: legacy field in BEiT vs packed-weight unpacking in BitNet/Aircelerate loading." }, { - "left": "issue:43504", - "right": "issue:43550", + "left": "issue:41720", + "right": "issue:44464", "accept": false, - "reason": "BEiT pretrained-load legacy field issue vs Bamba torch.compile SDPA bug; unrelated." + "reason": "Different failures and execution paths: auto device mapping cuda assert vs chunked generation with compiled forward." }, { - "left": "issue:44849", - "right": "issue:45357", + "left": "issue:43504", + "right": "issue:43550", "accept": false, - "reason": "Qwen3.5 output_hidden_states bug and save_pretrained visual-key regression are different regressions." + "reason": "Different models and issues; shared only at a very broad \u201cruntime bug\u201d level." }, { "left": "issue:43262", "right": "issue:45084", "accept": false, - "reason": "Audio processor sampling-rate default bug vs chat template compilation error; different subsystems." + "reason": "Audio chat-template sampling-rate defaulting is unrelated to compilation failing on non-template nodes." }, { - "left": "issue:32090", - "right": "issue:34567", + "left": "issue:44849", + "right": "issue:45357", "accept": false, - "reason": "NoneType broadcast error in trainer vs TrainerState num_input_tokens_seen not updating; unrelated." + "reason": "Different regressions in Qwen3.5: hidden-states behavior vs incorrect visual-encoder keys on save_pretrained." }, { "left": "issue:32090", - "right": "issue:33666", - "accept": false, - "reason": "Trainer broadcast failure and Qwen2-VL multi-GPU training issue are not the same bug." - }, - { - "left": "issue:43504", - "right": "issue:44610", + "right": "issue:34567", "accept": false, - "reason": "BEiT legacy-field loading issue vs OmDet-Turbo processor/model input-size mismatch; different failures." + "reason": "Trainer distributed broadcast failure is unrelated to TrainerState.num_input_tokens_seen not updating." }, { - "left": "issue:43278", - "right": "issue:43688", + "left": "issue:32090", + "right": "issue:33666", "accept": false, - "reason": "Embedding dtype drift during eval vs auxiliary-loss normalization bug; unrelated." + "reason": "Different training-time problems: NoneType broadcast vs multi-GPU training for Qwen2-VL." } ] }, @@ -5863,7 +5787,6 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ - "issue:33357", "issue:33666", "issue:36010", "issue:36331", @@ -5874,6 +5797,7 @@ "issue:43299", "issue:43334", "issue:43493", + "issue:43504", "issue:43525", "issue:43540", "issue:43572", @@ -5891,13 +5815,13 @@ "issue:43873", "issue:43927", "issue:44060", - "issue:44291", "issue:44361", "issue:44423", "issue:44466", "issue:44488", "issue:44521", "issue:44589", + "issue:44610", "issue:44625", "issue:44734", "issue:44821", @@ -5913,21 +5837,23 @@ "issue:45584" ], "soft_pairs": [ + "issue:43504|issue:44610", + "issue:43278|issue:43688", "issue:43334|issue:43572", + "issue:43525|issue:43572", "issue:43756|issue:44913", "issue:43540|issue:43716", - "issue:43525|issue:43572", "issue:43493|issue:43825", "issue:43742|issue:44821", "issue:45200|issue:45229", "issue:33666|issue:36010", - "issue:43688|issue:43873", "issue:43720|issue:43927", "issue:43606|issue:43927", + "issue:43688|issue:43873", "issue:43257|issue:43299", "issue:43278|issue:43844", - "issue:43540|issue:44977", "issue:44488|issue:44843", + "issue:43540|issue:44977", "issue:43720|issue:44361", "issue:39692|issue:43493", "issue:44964|issue:45405", @@ -5941,224 +5867,222 @@ "issue:44423|issue:44734", "issue:43824|issue:44589", "issue:43688|issue:43746", - "issue:39692|issue:43295", - "issue:33357|issue:33666", - "issue:43572|issue:44291" + "issue:39692|issue:43295" ], - "prepared_review_unit_hash": "e265358a354d294653517b540cb753ff3377f748df229c2bcc92d3579e5bc800", + "prepared_review_unit_hash": "0a8f87fa71a06c139253751805aa1355af77bb1c9044c8e53a7a11b1ec9f1f3b", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 37175, - "estimated_input_tokens": 9294, - "estimated_eval_tokens": 18844 + "serialized_chars": 37135, + "estimated_input_tokens": 9284, + "estimated_eval_tokens": 18824 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These issues are mostly a loose cluster of recent Transformers regressions around loading/config/init paths, but they are not duplicates of one another. I found no safe PR cluster here.", - "confidence": 0.66, - "canonical_issue_reason": "issue:43646 is the broadest umbrella regression report in the set: a generic v5 custom model initialization breakage that best represents the cluster\u2019s repeated loading/init failures.", + "summary": "The cluster is mostly a grab-bag of distinct model-loading, config, tokenizer, multimodal, and training regressions that only look similar at the title level. I would not merge any of the soft-edge pairs; they describe different concrete bugs. The broadest representative issue is the generic v5/custom-initialization regression.", + "confidence": 0.56, + "canonical_issue_reason": "issue:43646 is the broadest umbrella item in the set: a generic Transformers 5.0 custom model initialization regression that can represent the wider class of load/init failures here.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43646 is the most representative top-level issue for this cluster, though it is still not a duplicate of the other reports.", + "best_issue_reason": "issue:43646 is the most representative anchor for this cluster because it is broader than the model-specific loading/config bugs and best matches the recurring v5 regression theme.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43504", + "right": "issue:44610", + "accept": false, + "reason": "Different failures: BEiT legacy-field loading vs OmDet processor/model image-size mismatch." + }, + { + "left": "issue:43278", + "right": "issue:43688", + "accept": false, + "reason": "Both mention dtype/normalization issues, but one is an eval dtype regression and the other is auxiliary-loss normalization in MoE models." + }, { "left": "issue:43334", "right": "issue:43572", "accept": false, - "reason": "Both are config-related token-id errors, but they hit different models, different missing fields (pad_token_id vs pad_token_idx), and likely different code paths." + "reason": "Different config fields and models: Qwen3-VL pad_token_id vs StableLm pad_token_idx." + }, + { + "left": "issue:43525", + "right": "issue:43572", + "accept": false, + "reason": "Both are missing config attributes, but they affect different model families and different fields." }, { "left": "issue:43756", "right": "issue:44913", "accept": false, - "reason": "Both concern config behavior, but one is about Smollm3 RoPE layer retention and the other about GPTNeoX rotary_pct reload persistence; not the same bug." + "reason": "Smollm3 RoPE-layer dropping is unrelated to GPTNeoX rotary_pct reload behavior." }, { "left": "issue:43540", "right": "issue:43716", "accept": false, - "reason": "Both are multimodal-input problems, but one is a Qwen3OmniMoe video-processing ValueError and the other is a Mistral-3 image dtype mismatch." - }, - { - "left": "issue:43525", - "right": "issue:43572", - "accept": false, - "reason": "Similar missing-pad-token symptom, but different config classes and different missing attributes; not enough to treat as one duplicate." + "reason": "Qwen3OmniMoe video-input processing error is not the same as Mistral-3 image-preprocessor dtype mismatch." }, { "left": "issue:43493", "right": "issue:43825", "accept": false, - "reason": "SigLIP2 model discrepancy vs a pipeline error message regression are unrelated bugs." + "reason": "SigLIP2 implementation mismatch is unrelated to pipeline translation-task messaging in v5." }, { "left": "issue:43742", "right": "issue:44821", "accept": false, - "reason": "Different loading paths and failures: MobileLLM key error versus AutoImageProcessor-from-URL failure." + "reason": "MobileLLM key error and AutoImageProcessor URL loading failure are different code paths and model types." }, { "left": "issue:45200", "right": "issue:45229", "accept": false, - "reason": "Gemma4 text-finetuning token-type-id defaulting is unrelated to a multi-GPU inference OOM." + "reason": "Gemma 4 text-only fine-tuning defaults are unrelated to Gemma 4 multi-GPU inference OOM." }, { "left": "issue:33666", "right": "issue:36010", "accept": false, - "reason": "Multi-GPU Qwen2-VL training and a GenerationMixin import error are completely different issues." - }, - { - "left": "issue:43688", - "right": "issue:43873", - "accept": false, - "reason": "Aux-loss normalization in MoE models and quantization/offload behavior are separate problems." + "reason": "Multi-GPU training for Qwen2-VL is unrelated to an import error for GenerationMixin." }, { "left": "issue:43720", "right": "issue:43927", "accept": false, - "reason": "Packed-weight unpacking during accelerate load and DiaConfig token-id persistence are unrelated." + "reason": "Packed-weight unpacking during accelerate loading is not the same bug as DiaConfig losing custom token IDs on save/load." }, { "left": "issue:43606", "right": "issue:43927", "accept": false, - "reason": "CPU offload device mismatch and config token-id loss are different failure modes." + "reason": "CPU offload device mismatch and config token-ID loss are separate failures." + }, + { + "left": "issue:43688", + "right": "issue:43873", + "accept": false, + "reason": "Aux-loss normalization in MoE models is unrelated to quantization/offloading behavior." }, { "left": "issue:43257", "right": "issue:43299", "accept": false, - "reason": "Both mention Qwen3 MoE loading, but one is accelerate/deepspeed weight conversion and the other is a transformers dev load regression for Qwen3VL MoE." + "reason": "Both involve Qwen3 MoE loading, but one is accelerate+deepspeed conversion and the other is a v5 loading regression; not the same concrete bug." }, { "left": "issue:43278", "right": "issue:43844", "accept": false, - "reason": "Embedding dtype drift during evaluate and DeepSpeed ZeRO-3 gradient inflation are distinct training/runtime bugs." + "reason": "Embedding dtype drift in eval is unrelated to ZeRO-3 gradient inflation." }, { - "left": "issue:43540", - "right": "issue:44977", + "left": "issue:44488", + "right": "issue:44843", "accept": false, - "reason": "Video-input processing failure and flash-attention generation issues in Qwen3.5 are not the same defect." + "reason": "Model-load failure for cjvt/sleng-bert is unrelated to offline tokenizer hub access in _patch_mistral_regex." }, { - "left": "issue:44488", - "right": "issue:44843", + "left": "issue:43540", + "right": "issue:44977", "accept": false, - "reason": "Model load failure for cjvt/sleng-bert and an offline-mode regression in Mistral regex patching are unrelated." + "reason": "Video-input processing error in Qwen3OmniMoe is not the same as flash-attention generation issues in Qwen3.5." }, { "left": "issue:43720", "right": "issue:44361", "accept": false, - "reason": "BitNet packed-weight loading and MLukeTokenizer task handling are separate code paths." + "reason": "Both are loader-related, but packed-weight unpacking and MLukeTokenizer task AttributeError are distinct bugs." }, { "left": "issue:39692", "right": "issue:43493", "accept": false, - "reason": "SigLIP2 docs-example mistakes are not the same as the underlying HF-vs-JAX implementation discrepancy." + "reason": "SigLIP2 documentation issues are not the same as SigLIP2 implementation-vs-JAX discrepancies." }, { "left": "issue:44964", "right": "issue:45405", "accept": false, - "reason": "Phi-4 multimodal loading and the unreleased PEFT version bump are unrelated." + "reason": "Phi-4 multimodal load failure and PEFT minimum-version bump are unrelated." }, { "left": "issue:43493", "right": "issue:43844", "accept": false, - "reason": "SigLIP2 implementation mismatch and DeepSpeed gradient inflation are different bugs." + "reason": "SigLIP2 model mismatch and ZeRO-3 gradient instability are different problem classes." }, { "left": "issue:33666", "right": "issue:36331", "accept": false, - "reason": "Qwen2-VL multi-GPU training and compute_loss signature incompatibility are not duplicates." + "reason": "Multi-GPU training setup for Qwen2-VL is unrelated to the CustomTrainer.compute_loss signature break." }, { "left": "issue:43756", "right": "issue:44060", "accept": false, - "reason": "Smollm3 RoPE-layer loss and Qwen3-Next tied-weights warning are different model/config problems." + "reason": "Smollm3 RoPE-layer count bug is unrelated to Qwen3-Next tied-weights warnings." }, { "left": "issue:44589", "right": "issue:45520", "accept": false, - "reason": "A missing storage backend for Float8 and a flash_attn import_utils KeyError are unrelated." + "reason": "Float8 storage lookup failure and flash_attn KeyError on Python 3.13 are different import/runtime issues." }, { "left": "issue:44625", "right": "issue:45584", "accept": false, - "reason": "Qwen3.5 num_labels propagation and Whisper empty-transcription generation are separate issues." + "reason": "Qwen3.5 num_labels propagation and Whisper empty-transcription generation are unrelated." }, { "left": "issue:44521", "right": "issue:45290", "accept": false, - "reason": "Both involve apply_chat_template, but one is all-zero assistant masks for multimodal inputs and the other is a crash on tool-call messages with no content." + "reason": "All-zero assistant masks for multimodal inputs is not the same as tool-call assistant-message crash." }, { "left": "issue:43646", "right": "issue:44466", "accept": false, - "reason": "Both are v5/tied-weight-adjacent regressions, but one is generic custom model initialization and the other is device-dependent lm_head serialization; not the same concrete bug." + "reason": "Generic custom model initialization regression is broader, but lm_head.weight serialization/device-dependent tied-weights behavior is a distinct bug." }, { "left": "issue:44423", "right": "issue:44734", "accept": false, - "reason": "Both are serve crashes, but the failing code paths differ: multimodal continuous batching versus KV-cache continuation indexing." + "reason": "Continuous-batching multimodal crash and /v1/responses KV-cache indexing bug are different serving-path failures." }, { "left": "issue:43824", "right": "issue:44589", "accept": false, - "reason": "Qwen2.5-VL import failure and Float8 storage lookup failure are unrelated." + "reason": "Qwen2_5_VL import failure and Float8 storage lookup failure are unrelated." }, { "left": "issue:43688", "right": "issue:43746", "accept": false, - "reason": "MoE auxiliary-loss normalization and GraniteSpeech PEFT checkpoint loading are different problems." + "reason": "Aux-loss normalization in MoE models is unrelated to GraniteSpeech PEFT local-checkpoint loading." }, { "left": "issue:39692", "right": "issue:43295", "accept": false, - "reason": "SigLIP2 documentation errors and processor.tokenizer/image-tokenizer regression are not the same issue." - }, - { - "left": "issue:33357", - "right": "issue:33666", - "accept": false, - "reason": "MacOS bus error with a community CLIP model and Qwen2-VL multi-GPU training are unrelated." - }, - { - "left": "issue:43572", - "right": "issue:44291", - "accept": false, - "reason": "Both are v5 loading regressions, but one is a missing pad_token_idx config field and the other is an unexpected _is_hf_initialized argument in init_empty_weights." + "reason": "SigLIP2 doc-example errors are unrelated to the v4.57.5 processor.tokenizer regression." } ] }, @@ -6177,14 +6101,16 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ "issue:15354", "issue:29942", "issue:30333", + "issue:33357", + "issue:33666", "issue:37428", "issue:38175", "issue:39401", @@ -6195,6 +6121,7 @@ "issue:43504", "issue:43525", "issue:43540", + "issue:43572", "issue:43575", "issue:43582", "issue:43606", @@ -6205,48 +6132,47 @@ "issue:43957", "issue:43994", "issue:44112", + "issue:44291", "issue:44315", "issue:44361", "issue:44464", - "issue:44466", "issue:44493", "issue:44554", "issue:44568", "issue:44610", "issue:44792", "issue:44849", - "issue:44898", "issue:44964", "issue:44977", "issue:45072", "issue:45084", - "issue:45127", "issue:45198", "issue:45200", "issue:45325", "issue:45356", "issue:45405", - "issue:45412", "issue:45446", "issue:45520", "issue:45584" ], "soft_pairs": [ + "issue:33357|issue:33666", + "issue:43572|issue:44291", "issue:43994|issue:44493", "issue:43504|issue:43606", "issue:44464|issue:45584", "issue:43504|issue:43653", - "issue:43927|issue:44361", "issue:43575|issue:45405", + "issue:43927|issue:44361", "issue:43278|issue:43701", "issue:44792|issue:45405", - "issue:29942|issue:30333", "issue:43606|issue:44112", + "issue:29942|issue:30333", "issue:15354|issue:38175", "issue:44554|issue:45446", "issue:43540|issue:45200", - "issue:45084|issue:45520", "issue:43232|issue:44849", + "issue:45084|issue:45520", "issue:43844|issue:44964", "issue:39401|issue:44568", "issue:43582|issue:44977", @@ -6256,200 +6182,198 @@ "issue:44554|issue:45356", "issue:43334|issue:43525", "issue:43295|issue:44610", - "issue:45200|issue:45325", - "issue:44466|issue:45127", - "issue:44898|issue:45412" + "issue:45200|issue:45325" ], - "prepared_review_unit_hash": "4ceb043023b5768f38f363880e5d1d6f174d9929acc5d68404d323b678a6fce8", + "prepared_review_unit_hash": "af3b007793388969c289cd0c34340e224109487c74a3dc4865f23cb3ae9e06cf", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 27, - "serialized_chars": 35999, - "estimated_input_tokens": 9000, - "estimated_eval_tokens": 18256 + "serialized_chars": 35862, + "estimated_input_tokens": 8966, + "estimated_eval_tokens": 18188 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: the items span unrelated bugs in tokenizers, generation, model loading, dtype/device behavior, CI, and versioning. The soft links are mostly superficial keyword overlap, so none of the candidate pairs look like true duplicates.", - "confidence": 0.93, - "canonical_issue_reason": null, + "summary": "No soft pair looks like a true duplicate: the items mostly share only broad keywords such as model loading, tokenizer, generation, or dtype issues, but they describe different failure modes and code paths.", + "confidence": 0.62, + "canonical_issue_reason": "Issue 43957 is the broadest, most representative report in this set: it describes a general model-loading regression affecting multiple models under transformers 5.x.", "canonical_pr_reason": null, - "best_issue_reason": "No single issue is a safe global representative; the cluster mixes unrelated failure modes across different models and subsystems, so there is no clear canonical issue to anchor on.", + "best_issue_reason": "Issue 43957 is the best standalone representative because it captures a wider cross-model regression than the more model-specific reports.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:33357", + "right": "issue:33666", + "accept": false, + "reason": "Both mention model usage on specialized hardware/training, but one is a MacOS bus error for CLIP and the other is a Qwen2-VL multi-GPU training request; different bugs and workflows." + }, + { + "left": "issue:43572", + "right": "issue:44291", + "accept": false, + "reason": "One is about missing pad_token_idx after a 5.0 update; the other is an init_empty_weights/_is_hf_initialized argument error. Related release context, but not the same bug." + }, { "left": "issue:43994", "right": "issue:44493", "accept": false, - "reason": "Different bugs: SigLIP2 bad outputs vs a position-id warning affecting many models; not the same failing code path." + "reason": "Both involve wrong outputs, but one is SigLIP2 nonsense results and the other is unexpected position-id keys across many models; different root causes." }, { "left": "issue:43504", "right": "issue:43606", "accept": false, - "reason": "Legacy-field model loading bug vs CPU offload device mismatch; unrelated symptoms and fixes." + "reason": "Beit pretrained loading with a legacy field is unrelated to a Bark CPU offload device mismatch." }, { "left": "issue:44464", "right": "issue:45584", "accept": false, - "reason": "Compiled-forward chunked generation issue vs empty-transcription token alignment bug; distinct generation paths." + "reason": "Chunked generation with compiled forward is a generation-path consistency issue; Whisper empty transcription after align_special_tokens is a tokenizer/special-token edge case." }, { "left": "issue:43504", "right": "issue:43653", "accept": false, - "reason": "Pretrained model preset loading vs BigBirdTokenizer special-token decode behavior; different subsystems." + "reason": "A legacy-field load failure in BEiT is unrelated to BigBirdTokenizer mask-token registration and decode behavior." }, { - "left": "issue:43927", - "right": "issue:44361", + "left": "issue:43575", + "right": "issue:45405", "accept": false, - "reason": "Save/load custom token IDs causing IndexError vs tokenizer AttributeError on tasks; not the same bug." + "reason": "Qwen2-57B-A14B-Instruct TP OOM and MIN_PEFT_VERSION being bumped to an unreleased version are unrelated." }, { - "left": "issue:43575", - "right": "issue:45405", + "left": "issue:43927", + "right": "issue:44361", "accept": false, - "reason": "TP-related OOM vs a release/version-constant problem; no shared underlying change." + "reason": "DiaConfig losing custom token IDs after save/load is a config persistence bug; MLukeTokenizer failing on tasks is a tokenizer/task-specific AttributeError." }, { "left": "issue:43278", "right": "issue:43701", "accept": false, - "reason": "BF16-to-FP32 eval dtype drift vs resume checkpoint key mismatch; unrelated failure modes." + "reason": "BF16-to-FP32 embedding dtype drift and resume_from_checkpoint key mismatch are different training-state issues." }, { "left": "issue:44792", "right": "issue:45405", "accept": false, - "reason": "Janus image-generation test failure vs PEFT minimum-version bump; no common code path." + "reason": "A janus image-generation test failure is unrelated to a dependency version bump issue." }, { - "left": "issue:29942", - "right": "issue:30333", + "left": "issue:43606", + "right": "issue:44112", "accept": false, - "reason": "Flash Attention 2 test failures vs MLFlow reporting after failed training; unrelated concerns." + "reason": "Both are CI/device-related, but Bark CPU offload mismatch and GraniteSpeech stale device override are distinct bugs." }, { - "left": "issue:43606", - "right": "issue:44112", + "left": "issue:29942", + "right": "issue:30333", "accept": false, - "reason": "Both mention devices, but one is a real CPU-offload mismatch and the other is a stale CI test; not the same bug." + "reason": "Failing Flash Attention 2 tests and MLFlow not reporting failed jobs are unrelated." }, { "left": "issue:15354", "right": "issue:38175", "accept": false, - "reason": "TorchScript GeneratorExp limitation vs zero probabilities in SigLIP2; completely different issues." + "reason": "GeneratorExp/JIT scripting incompatibility is unrelated to SigLIP2 zero-probability outputs." }, { "left": "issue:44554", "right": "issue:45446", "accept": false, - "reason": "MPS attention correctness on mismatched heads vs flex_attention import version check; different code paths." + "reason": "An MPS attention correctness bug with mismatched head dims is unrelated to a PyTorch version check in flex_attention." }, { "left": "issue:43540", "right": "issue:45200", "accept": false, - "reason": "Qwen3OmniMoe video-input processing vs Gemma 4 mm_token_type_ids handling; both multimodal, but not the same defect." + "reason": "Qwen3OmniMoe video-input validation and Gemma 4 mm_token_type_ids defaults are separate multimodal input bugs." }, { - "left": "issue:45084", - "right": "issue:45520", + "left": "issue:43232", + "right": "issue:44849", "accept": false, - "reason": "Compiler template-node error vs flash_attn import failure on Python 3.13; unrelated." + "reason": "A generation-state update bug after sync_gpus is not the same as Qwen3.5 output_hidden_states behavior." }, { - "left": "issue:43232", - "right": "issue:44849", + "left": "issue:45084", + "right": "issue:45520", "accept": false, - "reason": "Generation kwargs update after sync_gpus vs Qwen3.5 hidden-states bug; same broad area, different bug." + "reason": "Template-node compilation failure and flash_attn import KeyError on Python 3.13 are unrelated failures." }, { "left": "issue:43844", "right": "issue:44964", "accept": false, - "reason": "ZeRO-3 gradient inflation vs Phi-4 multimodal loading failure; no shared underlying problem." + "reason": "ZeRO-3 gradient growth with DeepSpeed is unrelated to Phi-4 multimodal model loading failures." }, { "left": "issue:39401", "right": "issue:44568", "accept": false, - "reason": "Qwen3 offset_mapping regression vs mdeberta special-token insertion; both tokenizer bugs, but distinct." + "reason": "Qwen3 tokenizer offset mapping is a tokenizer alignment bug; mdeberta add_special_tokens not adding BOS/EOS is a different special-token handling issue." }, { "left": "issue:43582", "right": "issue:44977", "accept": false, - "reason": "Apple Silicon allocator TypeError vs Qwen3.5 flash-attention generation bug; unrelated." + "reason": "Apple Silicon caching_allocator_warmup TypeError and Qwen3.5 flash-attention generation are different runtime bugs." }, { "left": "issue:45072", "right": "issue:45198", "accept": false, - "reason": "bfloat16 dtype mismatches in inference vs Wav2Vec2 save/tokenization failure; different subsystems." + "reason": "Both are inference issues, but dtype mismatches in bfloat16 models and Wav2Vec2 save_pretrained/tokenization failures are not the same code path." }, { "left": "issue:44315", "right": "issue:44849", "accept": false, - "reason": "Liger Kernel not applied with model_init vs Qwen3.5 hidden-states issue; not the same fix." + "reason": "Liger Kernel not applied with model_init is about trainer/model initialization hooks, not Qwen3.5 hidden-state outputs." }, { "left": "issue:37428", "right": "issue:43957", "accept": false, - "reason": "Flash-support import failure vs meta-device model loading breakage; separate problems." + "reason": "Flash-support import failure and meta-device model-loading regressions are both loading-related but clearly separate errors." }, { "left": "issue:44554", "right": "issue:45356", "accept": false, - "reason": "MPS attention correctness vs Kimi-K2.5 tokenizer codec regression; unrelated." + "reason": "An MPS attention correctness issue is unrelated to Kimi-K2.5 tokenizer codec/reg-ex warnings." }, { "left": "issue:43334", "right": "issue:43525", "accept": false, - "reason": "Both mention missing pad_token_id, but they affect different model/config classes and load paths." + "reason": "Qwen3-VL missing pad_token_id and Llama4Config missing pad_token_id are similar symptomatically, but they affect different model families and likely different config paths." }, { "left": "issue:43295", "right": "issue:44610", "accept": false, - "reason": "Processor.tokenizer regression vs OmDet-Turbo input-size mismatch; different processor/model contracts." + "reason": "Processor.tokenizer regression with images is unrelated to OmDet-Turbo processor image-size mismatch." }, { "left": "issue:45200", "right": "issue:45325", "accept": false, - "reason": "Gemma 4 token-type handling vs Qwen2.5-VL rope-index scaling; distinct multimodal bugs." - }, - { - "left": "issue:44466", - "right": "issue:45127", - "accept": false, - "reason": "lm_head.weight serialization inconsistency vs LoRA merge collapse with extended vocab; related theme but different concrete failures." - }, - { - "left": "issue:44898", - "right": "issue:45412", - "accept": false, - "reason": "Perceiver resize/interpolate failure vs RT-DETR memory leak on deletion; no common underlying issue." + "reason": "Gemma 4 token-type-id defaults and Qwen2.5-VL rope-index scaling are different multimodal bookkeeping bugs." } ] }, @@ -6468,8 +6392,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -6499,16 +6423,16 @@ "issue:44206", "issue:44360", "issue:44442", - "issue:44462", "issue:44464", + "issue:44466", "issue:44514", "issue:44521", "issue:44561", "issue:44568", "issue:44589", - "issue:44779", "issue:44792", "issue:44849", + "issue:44898", "issue:44936", "issue:44987", "issue:44991", @@ -6523,6 +6447,8 @@ "issue:45584" ], "soft_pairs": [ + "issue:44466|issue:45127", + "issue:44898|issue:45412", "issue:43493|issue:43701", "issue:43873|issue:44464", "issue:44792|issue:45563", @@ -6535,9 +6461,9 @@ "issue:43824|issue:43975", "issue:43295|issue:43450", "issue:43295|issue:43873", + "issue:44442|issue:45198", "issue:44360|issue:44561", "issue:38175|issue:45412", - "issue:44442|issue:45198", "issue:43295|issue:43994", "issue:43122|issue:44849", "issue:44206|issue:45584", @@ -6546,222 +6472,220 @@ "issue:38175|issue:43493", "issue:43827|issue:43901", "issue:44568|issue:44991", - "issue:43653|issue:44112", "issue:43450|issue:43493", + "issue:43653|issue:44112", "issue:45290|issue:45381", "issue:45127|issue:45245", - "issue:43819|issue:45083", - "issue:44462|issue:44561", - "issue:44561|issue:44779" + "issue:43819|issue:45083" ], - "prepared_review_unit_hash": "2e2c5cadd241920eda5e9d6ea375f9ab1e604685fa0abc24204edfc0d42d2627", + "prepared_review_unit_hash": "b47851ceb17d1d6c19f013525e704552c6d07a8d46652b1373a1677b13051bab", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36650, - "estimated_input_tokens": 9163, - "estimated_eval_tokens": 18582 + "serialized_chars": 36759, + "estimated_input_tokens": 9190, + "estimated_eval_tokens": 18636 }, "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 regressions/docs issues. The closest overlap is around multimodal chat-template handling, but none of the soft pairs are strong enough to treat as duplicates.", - "confidence": 0.66, - "canonical_issue_reason": "issue:44521 is the most active, well-referenced open issue and the best umbrella candidate among the multimodal/chat-template regressions here, though the overall cluster is heterogeneous.", + "summary": "This set is a loose collection of unrelated Transformers issues rather than one duplicate cluster. The soft pairs mostly share broad subsystems or model families, but not the same concrete bug/change, so all soft edges should be rejected.", + "confidence": 0.97, + "canonical_issue_reason": "issue:43824 is the best canonical anchor here: it is a concrete, reproducible import regression with a clear failure mode and strong user-facing impact, and it has more discussion/reference signal than most others.", "canonical_pr_reason": null, - "best_issue_reason": "If a single issue must anchor the set, issue:44521 is the strongest choice because it has the most ongoing discussion and inbound references while describing a concrete, user-facing bug.", + "best_issue_reason": "issue:43824 is the strongest representative issue in this mixed set because it is self-contained, clearly actionable, and the most clearly described regression.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44466", + "right": "issue:45127", + "accept": false, + "reason": "Different bugs: tied-weight serialization inconsistency vs LoRA merge collapse after vocab extension." + }, + { + "left": "issue:44898", + "right": "issue:45412", + "accept": false, + "reason": "Different model paths and symptoms: Perceiver resolution failure vs RT-DETR memory leak." + }, { "left": "issue:43493", "right": "issue:43701", "accept": false, - "reason": "Different bugs: SigLIP2 model discrepancy vs checkpoint resume key mismatch." + "reason": "Same repo area but unrelated failures: SigLIP2 output mismatch vs resume checkpoint key mismatch." }, { "left": "issue:43873", "right": "issue:44464", "accept": false, - "reason": "Unrelated areas: quantization/offloading vs chunked generation with compiled forward." + "reason": "Offloading/quantization behavior is unrelated to compiled-forward chunked generation inconsistency." }, { "left": "issue:44792", "right": "issue:45563", "accept": false, - "reason": "Different failures: janus image-generation test vs stale generation warning." + "reason": "Janus test failure and stale generate warning are different issues and likely different code paths." }, { "left": "issue:44514", "right": "issue:44521", "accept": false, - "reason": "Same API family, but different concrete defects: batched padding crash vs zero assistant masks in multimodal inputs." + "reason": "Both touch apply_chat_template, but one crashes on batched padding=False while the other returns all-zero assistant masks." }, { "left": "issue:37428", "right": "issue:43824", "accept": false, - "reason": "Both are ImportErrors, but they involve different missing symbols and different model families." + "reason": "Both are ImportErrors, but for different missing symbols and different code paths." }, { "left": "issue:43388", "right": "issue:43526", "accept": false, - "reason": "Both involve labels, but one drops tuple elements in gather_for_metrics and the other collapses labels in a fast image processor." + "reason": "Trainer metric gathering bug and BeitImageProcessorFast label reduction bug are distinct." }, { "left": "issue:43572", "right": "issue:44987", "accept": false, - "reason": "Different problems: missing StableLmConfig field vs model loading failure for a specific checkpoint." + "reason": "Missing config field after v5 update is not the same as a model-loading failure for a remote model." }, { "left": "issue:43844", "right": "issue:44936", "accept": false, - "reason": "Training instability with ZeRO-3 is unrelated to trainer.evaluate() failing after train()." + "reason": "ZeRO-3 gradient growth and trainer.evaluate() failing after train are different training behaviors." }, { "left": "issue:43572", "right": "issue:44589", "accept": false, - "reason": "No overlap: config field regression vs Float8 storage lookup error." + "reason": "Config API regression vs Float8 storage lookup error; no shared concrete bug." }, { "left": "issue:43824", "right": "issue:43975", "accept": false, - "reason": "ImportError for a missing class is not the same as DeepSeek detokenization corruption." + "reason": "Both involve v5 regressions, but one is an import failure and the other is incorrect detokenization." }, { "left": "issue:43295", "right": "issue:43450", "accept": false, - "reason": "Processor/tokenizer regression and batched video shape bug are different code paths and symptoms." + "reason": "Custom processor/tokenizer crash and batched video shape bug are different multimodal regressions." }, { "left": "issue:43295", "right": "issue:43873", "accept": false, - "reason": "Custom processor regression is unrelated to quantization/offloading behavior." + "reason": "Processor API regression is unrelated to quantization/offloading behavior." + }, + { + "left": "issue:44442", + "right": "issue:45198", + "accept": false, + "reason": "Tokenizer load failure and save/tokenization failure are separate tokenizer issues." }, { "left": "issue:44360", "right": "issue:44561", "accept": false, - "reason": "DSA indexer ReLU discussion is unrelated to removing is_torch_fx_available." + "reason": "DSA indexing discussion is unrelated to removal of is_torch_fx_available breaking trust_remote_code." }, { "left": "issue:38175", "right": "issue:45412", "accept": false, - "reason": "SigLIP2 zero-probabilities bug is unrelated to RT-DETR memory release." - }, - { - "left": "issue:44442", - "right": "issue:45198", - "accept": false, - "reason": "Different tokenizer/model-loading failures in different architectures." + "reason": "SigLIP2 inference anomaly and RT-DETR memory reclamation bug are unrelated." }, { "left": "issue:43295", "right": "issue:43994", "accept": false, - "reason": "Processor/tokenizer regression and SigLIP2 inference nonsense are not the same bug." + "reason": "Different models and failures: custom multimodal processor regression vs SigLIP2 nonsensical outputs." }, { "left": "issue:43122", "right": "issue:44849", "accept": false, - "reason": "Tokenization drift across versions is unrelated to Qwen3.5 hidden-state output behavior." + "reason": "Tokenizer version drift and Qwen3.5 hidden-state bug are unrelated." }, { "left": "issue:44206", "right": "issue:45584", "accept": false, - "reason": "Feature-extractor center-arg crash and Whisper empty-transcription failure are different issues." + "reason": "Unsupported feature-extractor arg crash and Whisper empty-transcription failure are different bugs." }, { "left": "issue:42915", "right": "issue:43824", "accept": false, - "reason": "Qwen3Moe FP8 config failure and Qwen2.5-VL import error are unrelated." + "reason": "Qwen3Moe FP8 config failure is unrelated to a missing exported VL class." }, { "left": "issue:43278", "right": "issue:43856", "accept": false, - "reason": "Embedding dtype mismatch and MoE memory usage are different training issues." + "reason": "Eval dtype change and MoE memory usage are different training-time issues." }, { "left": "issue:38175", "right": "issue:43493", "accept": false, - "reason": "Same general SigLIP2 area, but one is zero probabilities and the other is HF-vs-JAX discrepancy." + "reason": "Same model family, but zero probabilities and JAX parity discrepancy are not the same concrete failure." }, { "left": "issue:43827", "right": "issue:43901", "accept": false, - "reason": "Both are docs issues, but they refer to different pages and different deprecated pipeline APIs." + "reason": "Both are docs-only regressions, but they refer to different pages and different outdated API mentions." }, { "left": "issue:44568", "right": "issue:44991", "accept": false, - "reason": "Different tokenizer regressions affecting different models and failure modes." + "reason": "Special-token handling bug and tokenizer loading failure are distinct." }, { - "left": "issue:43653", - "right": "issue:44112", + "left": "issue:43450", + "right": "issue:43493", "accept": false, - "reason": "Tokenizer special-token bug and GraniteSpeech CI test flake are unrelated." + "reason": "Batched video shape bug is unrelated to SigLIP2 implementation discrepancy." }, { - "left": "issue:43450", - "right": "issue:43493", + "left": "issue:43653", + "right": "issue:44112", "accept": false, - "reason": "Video batching shape bug and SigLIP2 implementation discrepancy do not share the same code path." + "reason": "BigBirdTokenizer special-token registration and GraniteSpeech CI device override are unrelated." }, { "left": "issue:45290", "right": "issue:45381", "accept": false, - "reason": "Assistant-message/tool-call chat-template crash is not the same as Qwen2.5-VL video position-id issue." + "reason": "Chat-template crash with tool-call assistant messages is unrelated to Qwen2.5-VL video vision_position_ids." }, { "left": "issue:45127", "right": "issue:45245", "accept": false, - "reason": "LoRA/tied-embedding collapse and category-cardinality RuntimeError are unrelated." + "reason": "LoRA/tied-weights collapse is unrelated to the 2^24 category limit runtime error." }, { "left": "issue:43819", "right": "issue:45083", "accept": false, - "reason": "DAC latent decoding mismatch and qwen3_omni_moe feature-length helper behavior are different bugs." - }, - { - "left": "issue:44462", - "right": "issue:44561", - "accept": false, - "reason": "Tokenizer JSON loading and removal of is_torch_fx_available affect different layers of the stack." - }, - { - "left": "issue:44561", - "right": "issue:44779", - "accept": false, - "reason": "Remote-code compatibility break and DeepSeek detokenization regression are different issues." + "reason": "DAC latent/forward mismatch and qwen3_omni_moe feature-length helper behavior are different code paths." } ] }, @@ -6780,8 +6704,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -6791,7 +6715,6 @@ "issue:37428", "issue:38175", "issue:39692", - "issue:40444", "issue:43262", "issue:43278", "issue:43335", @@ -6804,7 +6727,6 @@ "issue:43606", "issue:43611", "issue:43653", - "issue:43720", "issue:43723", "issue:43819", "issue:43825", @@ -6813,8 +6735,10 @@ "issue:44292", "issue:44351", "issue:44448", + "issue:44462", "issue:44493", "issue:44554", + "issue:44561", "issue:44568", "issue:44610", "issue:44743", @@ -6835,14 +6759,16 @@ "issue:45593" ], "soft_pairs": [ + "issue:44462|issue:44561", + "issue:44561|issue:44779", "issue:45200|issue:45588", "issue:43450|issue:43575", "issue:44610|issue:44857", "issue:44857|issue:45593", "issue:37428|issue:43425", "issue:43262|issue:43450", - "issue:43819|issue:44112", "issue:44448|issue:44743", + "issue:43819|issue:44112", "issue:45310|issue:45478", "issue:43606|issue:45083", "issue:43278|issue:43825", @@ -6850,285 +6776,263 @@ "issue:37428|issue:44351", "issue:39692|issue:43526", "issue:43611|issue:43931", - "issue:43819|issue:44493", "issue:43335|issue:43550", + "issue:43819|issue:44493", "issue:45200|issue:45381", "issue:43450|issue:43526", "issue:39692|issue:44493", "issue:44855|issue:45083", - "issue:29127|issue:32090", "issue:44448|issue:44779", - "issue:38175|issue:44493", + "issue:29127|issue:32090", "issue:43335|issue:43653", - "issue:44292|issue:45520", + "issue:38175|issue:44493", "issue:44977|issue:45229", + "issue:44292|issue:45520", "issue:43335|issue:45083", "issue:43335|issue:43504", "issue:43550|issue:44112", "issue:44568|issue:45491", "issue:43335|issue:45072", - "issue:43723|issue:44554", - "issue:40444|issue:43526", - "issue:44448|issue:44977", - "issue:43335|issue:45593", - "issue:43720|issue:44112" + "issue:43723|issue:44554" ], - "prepared_review_unit_hash": "7d408e6d01071dc1cada7bdb13614bae5b33d50bcab3099ae5721b76b010fae6", + "prepared_review_unit_hash": "135b3d7fbdc3571120a3e7aefe95004d0a76a1e1ff2c72366e962c43fe381b34", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 37, - "serialized_chars": 38895, - "estimated_input_tokens": 9724, - "estimated_eval_tokens": 19704 + "soft_pair_count": 35, + "serialized_chars": 38239, + "estimated_input_tokens": 9560, + "estimated_eval_tokens": 19376 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a set of unrelated transformer bug reports. The only clear duplicate pair is the repeated Qwen3.5 MoE `from_pretrained` error; the rest are similar only at a broad subsystem level and should stay separate.", - "confidence": 0.98, - "canonical_issue_reason": "issue:45310 is the earlier report with the same title and the same Qwen3.5 MoE `from_pretrained` failure; it is the better canonical target for the duplicate pair with issue:45478.", + "summary": "The cluster is mostly heterogeneous v5/model-specific regressions that should not be merged. I found one clear duplicate pair: the two identical Qwen3.5 MoE `from_pretrained` error reports. No PRs are present in the cluster.", + "confidence": 0.78, + "canonical_issue_reason": "issue:45310 is the clearest canonical issue because it is a concrete, well-scoped bug report and has an exact duplicate in the cluster (`issue:45478`) with the same title/failure mode.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45310 is the clearest representative issue in the cluster because it is a precise, reproducible bug report and has a direct duplicate.", + "best_issue_reason": "issue:45310 is the best representative issue in this cluster: it is specific, reproducible, and duplicated verbatim by another issue, whereas most other items are unrelated regressions or generic v5 breakages.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44462", + "right": "issue:44561", + "accept": false, + "reason": "Both are v5-era failures, but one is tokenizer file selection and the other is a removed utility breaking trust_remote_code model loading; different root causes." + }, + { + "left": "issue:44561", + "right": "issue:44779", + "accept": false, + "reason": "Remote-code loading breakage and DeepSeek tokenizer output regression are separate tokenizer/model-loading bugs." + }, { "left": "issue:45200", "right": "issue:45588", "accept": false, - "reason": "Both involve model/integration behavior, but one is Gemma 4 multimodal token ids and the other is a flash-attention sink-less crash; different code paths and bugs." + "reason": "Gemma4 text-only token-id defaults and flash-attention sink-less crash are unrelated code paths." }, { "left": "issue:43450", "right": "issue:43575", "accept": false, - "reason": "Video processor shape handling and Qwen2.5-VL tensor-parallel OOM are unrelated failures." + "reason": "Video processor batching shape bug and Qwen TP OOM are different subsystems and failure modes." }, { "left": "issue:44610", "right": "issue:44857", "accept": false, - "reason": "Processor input-size mismatch and AMP/CUDA loss crash are different bugs in different components." + "reason": "OmDet preprocessing size mismatch and LwDetr AMP/CUDA loss crash are different vision bugs." }, { "left": "issue:44857", "right": "issue:45593", "accept": false, - "reason": "Both are training-time issues, but one is a float16/CUDA crash and the other is missing auxiliary losses; not the same defect." + "reason": "Float16 loss crash and missing auxiliary losses when denoising is off are unrelated training issues." }, { "left": "issue:37428", "right": "issue:43425", "accept": false, - "reason": "An import error for a missing flash-attention symbol and a Torch 2.10 incompatibility are separate compatibility problems." + "reason": "A missing flash-attention import symbol is not the same concrete bug as broad Torch 2.10 incompatibility." }, { "left": "issue:43262", "right": "issue:43450", "accept": false, - "reason": "Audio chat-template sampling rate default and batched video processor shape are unrelated multimodal bugs." + "reason": "Audio chat-template sampling-rate defaults and video batch-shape handling are different processor bugs." }, { - "left": "issue:43819", - "right": "issue:44112", + "left": "issue:44448", + "right": "issue:44743", "accept": false, - "reason": "DAC forward/from_latents mismatch and GraniteSpeech stale device-override CI failure are different issues." + "reason": "Pegasus v4/v5 output drift and Qwen3.5 recurrent-state reset are unrelated model-specific regressions." }, { - "left": "issue:44448", - "right": "issue:44743", + "left": "issue:43819", + "right": "issue:44112", "accept": false, - "reason": "Pegasus v4/v5 output differences and Qwen3.5 recurrent-state reset are different model behaviors and code paths." + "reason": "DAC forward/from_latents mismatch and GraniteSpeech CI device-override test failure are not the same bug." }, { "left": "issue:45310", "right": "issue:45478", "accept": true, - "reason": "Exact same Qwen3.5 MoE `from_pretrained` error title and same underlying loading bug; clearly duplicates." + "reason": "Same title, same model family, same `from_pretrained` failure; this is a clear duplicate report." }, { "left": "issue:43606", "right": "issue:45083", "accept": false, - "reason": "CPU offload device mismatch for bark-small and `_get_feat_extract_output_lengths` behavior in qwen3_omni_moe are unrelated." + "reason": "CPU offload device mismatch in Bark and `_get_feat_extract_output_lengths` behavior in qwen3_omni_moe are unrelated." }, { "left": "issue:43278", "right": "issue:43825", "accept": false, - "reason": "BF16/FP32 embedding dtype drift and pipeline translation-task messaging are different concerns." + "reason": "BF16-to-FP32 eval dtype drift and `pipeline()` translation-task messaging are different issues." }, { "left": "issue:29942", "right": "issue:32090", "accept": false, - "reason": "Flash Attention 2 test failures and `_gpu_broadcast_one` rejecting NoneType are distinct failures." + "reason": "Flash Attention 2 test failures and `_gpu_broadcast_one` NoneType TypeError are different failures." }, { "left": "issue:37428", "right": "issue:44351", "accept": false, - "reason": "Missing `_flash_supports_window_size` and missing `HybridCache` imports are both import errors but for different symbols and fixes." + "reason": "These are both import errors, but for different missing symbols and different breakages." }, { "left": "issue:39692", "right": "issue:43526", "accept": false, - "reason": "SigLIP2 doc example issues and BeitImageProcessorFast `reduce_labels` behavior are unrelated." + "reason": "Documentation/example errors for SigLIP2 are unrelated to BeitImageProcessorFast `reduce_labels` behavior." }, { "left": "issue:43611", "right": "issue:43931", "accept": false, - "reason": "`base_model_prefix` loading breakage and Qwen3-VL weight-shape mismatch are different model-loading problems." + "reason": "`base_model_prefix` loading regression and Qwen3-VL weight-shape mismatch are distinct model-loading problems." }, { - "left": "issue:43819", - "right": "issue:44493", + "left": "issue:43335", + "right": "issue:43550", "accept": false, - "reason": "DAC latent conversion mismatch and unexpected position-id keys are unrelated bugs." + "reason": "SwitchTransformers sparse-layer config bug and Bamba `torch.compile`/SDPA failure are different code paths." }, { - "left": "issue:43335", - "right": "issue:43550", + "left": "issue:43819", + "right": "issue:44493", "accept": false, - "reason": "SwitchTransformers sparse-layer config and Bamba torch.compile/SDPA failure are different model-specific problems." + "reason": "DAC latent conversion mismatch and widespread unexpected position-id keys are unrelated." }, { "left": "issue:45200", "right": "issue:45381", "accept": false, - "reason": "Gemma 4 text-only fine-tuning defaults and Qwen2.5-VL video position ids are different multimodal issues." + "reason": "Both concern multimodal ids, but Gemma4 `mm_token_type_ids` defaults and Qwen2.5-VL video position ids are separate bugs." }, { "left": "issue:43450", "right": "issue:43526", "accept": false, - "reason": "Both touch Beit-related processors, but one is batched video shape handling and the other is image-label reduction; not the same bug." + "reason": "Video batching shape handling and Beit fast processor `reduce_labels` are different processor defects." }, { "left": "issue:39692", "right": "issue:44493", "accept": false, - "reason": "Documentation example errors and position-id warnings in many models are not the same underlying defect." + "reason": "A SigLIP2 docs/example report is not the same as a generic position-id key regression across many models." }, { "left": "issue:44855", "right": "issue:45083", "accept": false, - "reason": "Python 3.13 `IndentationError` in DebertaV2 import and qwen3_omni_moe feature-length helper behavior are unrelated." - }, - { - "left": "issue:29127", - "right": "issue:32090", - "accept": false, - "reason": "LayoutLMv3 box-clarity error messaging and trainer broadcast NoneType are unrelated failures." + "reason": "Python 3.13 `@torch.jit.script` parsing/indentation failure and qwen3_omni_moe helper behavior are unrelated." }, { "left": "issue:44448", "right": "issue:44779", "accept": false, - "reason": "Pegasus output differences and DeepSeek tokenizer regressions are separate model/tokenizer issues." + "reason": "Both mention v4/v5 output changes, but Pegasus generation drift and DeepSeek tokenizer regression are different underlying bugs." }, { - "left": "issue:38175", - "right": "issue:44493", + "left": "issue:29127", + "right": "issue:32090", "accept": false, - "reason": "SigLIP2 zero probabilities and position-id key warnings are not the same bug." + "reason": "LayoutLMv3 error-message clarity and Trainer `_gpu_broadcast_one` NoneType crash are unrelated." }, { "left": "issue:43335", "right": "issue:43653", "accept": false, - "reason": "SwitchTransformers sparse-layer config and BigBirdTokenizer special-token decode issue are unrelated." + "reason": "SwitchTransformers config and BigBirdTokenizer special-token registration are different model/tokenizer bugs." }, { - "left": "issue:44292", - "right": "issue:45520", + "left": "issue:38175", + "right": "issue:44493", "accept": false, - "reason": "Qwen-3-8B-NVFP4 runtime error and Python 3.13 flash_attn import KeyError are different failures." + "reason": "SigLIP2 zero-probabilities and position-id key warnings are different symptoms with different likely causes." }, { "left": "issue:44977", "right": "issue:45229", "accept": false, - "reason": "Qwen3.5 flash-attention generation bug and Gemma4 multi-GPU CUDA OOM are unrelated." + "reason": "Qwen3.5 flash-attention generation bug and Gemma4 multi-GPU OOM are unrelated." + }, + { + "left": "issue:44292", + "right": "issue:45520", + "accept": false, + "reason": "Qwen-3-8B-NVFP4 runtime failure and Python 3.13 flash_attn import key error are different problems." }, { "left": "issue:43335", "right": "issue:45083", "accept": false, - "reason": "SwitchTransformers config bug and qwen3_omni_moe feature-length helper behavior are unrelated." + "reason": "SwitchTransformers sparse-layer config and qwen3_omni_moe feature-length helper behavior are not the same bug." }, { "left": "issue:43335", "right": "issue:43504", "accept": false, - "reason": "SwitchTransformers sparse-layer creation and Beit semantic-segmentation legacy-field loading are different issues." + "reason": "Sparse-layer creation in SwitchTransformers and BEiT legacy-field loading failure are unrelated." }, { "left": "issue:43550", "right": "issue:44112", "accept": false, - "reason": "Bamba SDPA/torch.compile failure and GraniteSpeech CI device override failure are unrelated." + "reason": "Bamba SDPA/compile failure and GraniteSpeech stale CI test are not the same underlying issue." }, { "left": "issue:44568", "right": "issue:45491", "accept": false, - "reason": "mDeBERTa special-token addition regression and Gemma3 NaN embeddings are different tokenizer/model problems." - }, - { - "left": "issue:43335", - "right": "issue:45072", - "accept": false, - "reason": "SwitchTransformers config bug and bfloat16 dtype mismatches in inference are not the same underlying issue." + "reason": "Tokenizer special-token insertion and Gemma3 NaN embeddings under sliding-window attention are different issues." }, { "left": "issue:43723", "right": "issue:44554", "accept": false, "reason": "AutoTokenizer loading regression and MPS attention correctness issue are unrelated." - }, - { - "left": "issue:40444", - "right": "issue:43526", - "accept": false, - "reason": "Qwen2.5-VL iterable dataset multi-image finetuning failure and BeitImageProcessorFast label reduction are different areas." - }, - { - "left": "issue:44448", - "right": "issue:44977", - "accept": false, - "reason": "Pegasus output drift and Qwen3.5 flash-attention generation failure are separate model-specific bugs." - }, - { - "left": "issue:43335", - "right": "issue:45593", - "accept": false, - "reason": "SwitchTransformers sparse-layer config and D-FINE auxiliary-loss behavior are unrelated training/model issues." - }, - { - "left": "issue:43720", - "right": "issue:44112", - "accept": false, - "reason": "BitNet packed-weight unpacking during accelerate loading and GraniteSpeech CI stale device override are different failures." } ] }, "evaluator_result": { "accept": true, - "feedback": "Accepted. The canonical pair is well grounded by the packet: issue:45310 and issue:45478 have identical titles and clearly match. The remaining soft-edge rejections stay conservative at the title/body-summary level, and the cluster summary does not overstate the evidence." + "feedback": "The summary is grounded in the packet: the cluster is indeed issue-only, and the one duplicate pair (issue:45310 / issue:45478) is a clear match by title and failure mode. The soft-edge rejections are conservative and do not overclaim equivalence, so I don\u2019t see evidence of an overstated mergeability verdict." }, "evaluator_used": true, "retried": false, @@ -7144,8 +7048,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -7156,23 +7060,24 @@ "issue:34689", "issue:35141", "issue:37428", + "issue:40444", "issue:41720", "issue:43065", "issue:43066", "issue:43278", "issue:43329", - "issue:43381", + "issue:43335", "issue:43412", - "issue:43421", + "issue:43526", "issue:43550", "issue:43606", "issue:43653", "issue:43701", + "issue:43720", "issue:43873", "issue:43976", "issue:43994", - "issue:44186", - "issue:44265", + "issue:44112", "issue:44448", "issue:44451", "issue:44462", @@ -7184,7 +7089,6 @@ "issue:44987", "issue:45072", "issue:45083", - "issue:45198", "issue:45229", "issue:45245", "issue:45372", @@ -7192,26 +7096,30 @@ "issue:45412", "issue:45440", "issue:45459", - "issue:45491", "issue:45520", "issue:45563", - "issue:45588" + "issue:45588", + "issue:45593" ], "soft_pairs": [ - "issue:45372|issue:45588", + "issue:40444|issue:43526", + "issue:44448|issue:44977", + "issue:43335|issue:45593", + "issue:43720|issue:44112", "issue:43066|issue:44462", + "issue:45372|issue:45588", "issue:34689|issue:35141", "issue:44977|issue:45563", "issue:44857|issue:45072", "issue:43606|issue:44857", "issue:43412|issue:45588", "issue:43873|issue:45229", - "issue:45372|issue:45406", - "issue:43873|issue:44493", "issue:37428|issue:44987", + "issue:43873|issue:44493", + "issue:45372|issue:45406", "issue:43701|issue:45083", - "issue:29942|issue:35141", "issue:43976|issue:45440", + "issue:29942|issue:35141", "issue:45245|issue:45563", "issue:43550|issue:45083", "issue:41720|issue:45229", @@ -7224,233 +7132,222 @@ "issue:43065|issue:45412", "issue:25251|issue:32090", "issue:43653|issue:44743", - "issue:44857|issue:45588", - "issue:44186|issue:45198", - "issue:43421|issue:45491", - "issue:43278|issue:43381", - "issue:43994|issue:45412", - "issue:44265|issue:45198" + "issue:44857|issue:45588" ], - "prepared_review_unit_hash": "b3e62fc6f83b2d19584561cf0791db84f72df75daaca74d11c982c2c96df9382", + "prepared_review_unit_hash": "784d466bd7b168a4f8d2a829a66036c8814c72bd2546f9974e025738aeea3615", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 32, - "serialized_chars": 36797, - "estimated_input_tokens": 9200, - "estimated_eval_tokens": 18656 + "soft_pair_count": 31, + "serialized_chars": 36492, + "estimated_input_tokens": 9123, + "estimated_eval_tokens": 18502 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set looks like a loose collection of Transformers issues with superficial thematic overlap (tokenizers, Gemma/Qwen, flash-attn, offloading, dtype, export), but the soft pairs mostly describe different root causes and should not be deduplicated. No PRs are present. Issue 45372 is the best cluster anchor only because it is the most connected item in the candidate graph, not because it matches the others exactly.", - "confidence": 0.56, - "canonical_issue_reason": "Issue 45372 is the most central node among the soft links, with multiple nearby tokenizer/processor-loading references, so it is the best available anchor for this set.", + "summary": "This cluster is highly heterogeneous: the soft pairs span unrelated bugs across loading, tokenization, multimodal processing, flash-attention, offloading, and CI. I rejected every soft edge as a duplicate false positive. With no PRs present, I selected the broadest active issue as the best issue representative.", + "confidence": 0.69, + "canonical_issue_reason": "Issue 43873 is the broadest and most active open report here, with multiple inbound references and discussion around offloading/quantization behavior, making it the least specific umbrella representative among these unrelated issues.", "canonical_pr_reason": null, - "best_issue_reason": "45372 has the broadest apparent overlap with other candidates in the provided similarity graph; it is the best representative issue, though still not a clear duplicate hub.", + "best_issue_reason": "Issue 43873 is the best available issue candidate because it is open, actively discussed, and broadly scoped compared with the other more niche or closed reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:45372", - "right": "issue:45588", + "left": "issue:40444", + "right": "issue:43526", + "accept": false, + "reason": "Different subsystems and failures: Qwen2.5-VL iterable dataset finetuning vs BEiT fast image processor label reduction." + }, + { + "left": "issue:44448", + "right": "issue:44977", "accept": false, - "reason": "Both involve tokenizer/processor-related failures, but one is a missing import for Gemma 4 loading and the other is a flash-attention runtime crash on `s_aux=None`; different code paths." + "reason": "Both mention output differences, but they concern different models and code paths: Pegasus v4/v5 mismatch vs Qwen3.5 flash-attention generation." + }, + { + "left": "issue:43335", + "right": "issue:45593", + "accept": false, + "reason": "Unrelated bugs in different models: SwitchTransformers sparse-layer config vs D-FINE auxiliary losses." + }, + { + "left": "issue:43720", + "right": "issue:44112", + "accept": false, + "reason": "Packed-weight loading in BitNet is unrelated to a stale GraniteSpeech CI device-override test." }, { "left": "issue:43066", "right": "issue:44462", "accept": false, - "reason": "Both touch tokenizer loading/serialization, but one is a wrong decoder type in v5 and the other is AutoTokenizer ignoring `tokenizer.json`; not the same bug." + "reason": "Both are tokenizer-related, but one is a decoder-type regression and the other is AutoTokenizer ignoring tokenizer.json; different underlying bugs." + }, + { + "left": "issue:45372", + "right": "issue:45588", + "accept": false, + "reason": "Gemma 4 processor import failure and flash-attention s_aux=None crash are different failure modes in different code paths." }, { "left": "issue:34689", "right": "issue:35141", "accept": false, - "reason": "Model loading failure for Llama 3.2 Vision is unrelated to the embedding reinitialization bug in `post_init`." + "reason": "Model loading failure for Llama 3.2 Vision is unrelated to post_init reinitializing embeddings after resize." }, { "left": "issue:44977", "right": "issue:45563", "accept": false, - "reason": "Flash-attention generation failure and a stale `num_return_sequences` warning are unrelated behavior." + "reason": "Qwen3.5 flash-attention generation bug is unrelated to a stale warning about num_return_sequences." }, { "left": "issue:44857", "right": "issue:45072", "accept": false, - "reason": "Both are dtype/device-related inference issues, but they affect different models and failure modes." + "reason": "Different numeric/runtime issues in different models: LwDetrImageLoss AMP crash vs dtype mismatches in SwitchTransformers/TimmWrapperModel." }, { "left": "issue:43606", "right": "issue:44857", "accept": false, - "reason": "CPU offload device mismatch and float16 AMP CUDA crash are separate bugs." + "reason": "CPU offload device mismatch for bark-small is not the same bug as the LwDetrImageLoss float16 CUDA crash." }, { "left": "issue:43412", "right": "issue:45588", "accept": false, - "reason": "Executorch export / unsupported op failure is unrelated to the flash_attention `s_aux=None` crash." + "reason": "Gemma3n ExecuTorch export problems are unrelated to an s_aux=None flash-attention crash." }, { "left": "issue:43873", "right": "issue:45229", "accept": false, - "reason": "Both mention offload/memory behavior, but one is quantization offloading and the other is multi-GPU OOM on Gemma4." + "reason": "Quantization/offloading behavior and Gemma4 multi-GPU OOM are different issues, not one concrete code-path bug." }, { - "left": "issue:45372", - "right": "issue:45406", + "left": "issue:37428", + "right": "issue:44987", "accept": false, - "reason": "Both are Gemma 4 loading/serving issues, but one is a missing `ReasoningEffort` import and the other is a missing `_tokenizer` attribute." + "reason": "The missing internal flash-attention import and loading physical-intelligence/fast may be adjacent, but the reports do not describe the same concrete bug." }, { "left": "issue:43873", "right": "issue:44493", "accept": false, - "reason": "Quantized offloading problems and unexpected position-id keys are different issues." + "reason": "Offloading/quantization behavior is unrelated to unexpected position-id keys during model loading." }, { - "left": "issue:37428", - "right": "issue:44987", + "left": "issue:45372", + "right": "issue:45406", "accept": false, - "reason": "Import error for `_flash_supports_window_size` and Qwen3.5 flash-attention generation failure are distinct failures." + "reason": "Both involve Gemma 4, but one is an external import failure in processor loading and the other is a missing _tokenizer attribute in serve; different root causes." }, { "left": "issue:43701", "right": "issue:45083", "accept": false, - "reason": "Checkpoint resume key mismatch is unrelated to the qwen3_omni_moe feature-extraction helper behavior." + "reason": "Checkpoint key mismatch is unrelated to the qwen3_omni_moe feature-extraction length helper bug." }, { - "left": "issue:29942", - "right": "issue:35141", + "left": "issue:43976", + "right": "issue:45440", "accept": false, - "reason": "Flash Attention 2 test failures do not match the embedding reinitialization bug." + "reason": "Python version compatibility regression is unrelated to DeepseekV3MoE divergence from the remote implementation." }, { - "left": "issue:43976", - "right": "issue:45440", + "left": "issue:29942", + "right": "issue:35141", "accept": false, - "reason": "Python version incompatibility and DeepSeekV3MoE divergence are unrelated." + "reason": "Flash Attention 2 test failures are unrelated to embedding reinitialization after token resize." }, { "left": "issue:45245", "right": "issue:45563", "accept": false, - "reason": "Category-count limit error is unrelated to the stale generation warning." + "reason": "Category-count runtime limit and a stale warning in paged generate() are not the same underlying problem." }, { "left": "issue:43550", "right": "issue:45083", "accept": false, - "reason": "torch.compile/SDPA failure and feature-length helper behavior are different code paths." + "reason": "Bamba SDPA torch.compile failure is unrelated to the qwen3_omni_moe feature-length helper behavior." }, { "left": "issue:41720", "right": "issue:45229", "accept": false, - "reason": "Qwen3 auto device mapping CUDA assert and Gemma4 multi-GPU OOM are different model/runtime problems." + "reason": "Qwen3 auto device-mapping cudaErrorAssert and Gemma4 multi-GPU OOM are both large-model runtime issues, but not the same bug." }, { "left": "issue:43278", "right": "issue:43329", "accept": false, - "reason": "Embedding dtype drift between train/eval is unrelated to the undefined video-branch helper variables." + "reason": "Embedding dtype drift during evaluation is unrelated to undefined variables in multimodal token counting." }, { "left": "issue:32090", "right": "issue:33357", "accept": false, - "reason": "Trainer `_gpu_broadcast_one` NoneType error and MacOS bus error on CLIP are unrelated." + "reason": "Trainer broadcast TypeError and MacOS bus error on CLIP are unrelated failures." }, { "left": "issue:45372", "right": "issue:45459", "accept": false, - "reason": "Tokenizer import failure and protobuf-related error masking are different tokenizer-loading bugs." + "reason": "Gemma 4 processor import failure and tokenizer error masking from protobuf absence are different issues." }, { "left": "issue:44451", "right": "issue:45520", "accept": false, - "reason": "ScandiBERT load failure and Python 3.13 `flash_attn` import-key error are unrelated." + "reason": "Loading a specific model and a Python 3.13 flash_attn KeyError are unrelated." }, { "left": "issue:44448", "right": "issue:44466", "accept": false, - "reason": "Both involve v4/v5 behavior changes, but one is output divergence and the other is serialization/tied-weight inconsistency; not the same concrete bug." + "reason": "Both involve version-dependent behavior, but Pegasus output differences and lm_head serialization inconsistency are different bugs." }, { "left": "issue:43065", "right": "issue:43994", "accept": false, - "reason": "Dummy Conv2d in Sam3PixelDecoder and SigLIP2 nonsensical AutoModel/pipeline outputs are different model issues." + "reason": "Dummy Conv2d in Sam3PixelDecoder is unrelated to SigLIP2 nonsensical AutoModel/pipeline outputs." }, { "left": "issue:43065", "right": "issue:45412", "accept": false, - "reason": "Dummy layer in Sam3PixelDecoder and RT-DETR memory release issue are unrelated." + "reason": "A placeholder Conv2d in Sam3PixelDecoder is unrelated to RT-DETR memory not being released." }, { "left": "issue:25251", "right": "issue:32090", "accept": false, - "reason": "Pipeline `top_k` output shape change and `_gpu_broadcast_one` NoneType error are unrelated." + "reason": "Pipeline top_k nesting behavior and a Trainer broadcast TypeError have no common underlying bug." }, { "left": "issue:43653", "right": "issue:44743", "accept": false, - "reason": "BigBirdTokenizer special-token decode bug and Qwen3.5 recurrent-state reset bug are unrelated." + "reason": "BigBirdTokenizer special-token decode bug is unrelated to recurrent state reset in qwen3_5 modular code." }, { "left": "issue:44857", "right": "issue:45588", "accept": false, - "reason": "AMP/CUDA loss crash and flash_attention `s_aux=None` crash are separate inference failures." - }, - { - "left": "issue:44186", - "right": "issue:45198", - "accept": false, - "reason": "LayoutLMv2 NER/padding crash and Wav2Vec2 save/tokenization failure are different tokenizer bugs." - }, - { - "left": "issue:43421", - "right": "issue:45491", - "accept": false, - "reason": "Runtime special-token post-processor update feature and Gemma3 NaN embeddings are unrelated." - }, - { - "left": "issue:43278", - "right": "issue:43381", - "accept": false, - "reason": "Train/eval embedding dtype mismatch is unrelated to gradient checkpointing in eval mode." - }, - { - "left": "issue:43994", - "right": "issue:45412", - "accept": false, - "reason": "SigLIP2 output issue and RT-DETR memory release problem are unrelated." - }, - { - "left": "issue:44265", - "right": "issue:45198", - "accept": false, - "reason": "torch.export failure with `torch_compilable_check` and Wav2Vec2 save/tokenization failure are unrelated." + "reason": "A CUDA AMP loss crash and an s_aux=None flash-attention crash are different model/runtime failures." } ] }, @@ -7469,8 +7366,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -7491,7 +7388,7 @@ "issue:43278", "issue:43334", "issue:43381", - "issue:43441", + "issue:43421", "issue:43526", "issue:43531", "issue:43653", @@ -7499,10 +7396,11 @@ "issue:43756", "issue:43761", "issue:43824", - "issue:43931", + "issue:43994", "issue:44079", "issue:44186", "issue:44206", + "issue:44265", "issue:44279", "issue:44361", "issue:44442", @@ -7520,18 +7418,22 @@ "issue:45412", "issue:45440", "issue:45491", - "issue:45588", "issue:45593" ], "soft_pairs": [ + "issue:44186|issue:45198", + "issue:43421|issue:45491", + "issue:44265|issue:45198", + "issue:43278|issue:43381", + "issue:43994|issue:45412", "issue:42915|issue:42994", "issue:44361|issue:45198", "issue:44206|issue:44479", "issue:44871|issue:44977", - "issue:43257|issue:43278", "issue:43756|issue:45440", - "issue:29942|issue:30064", + "issue:43257|issue:43278", "issue:44610|issue:45412", + "issue:29942|issue:30064", "issue:45103|issue:45372", "issue:43824|issue:44279", "issue:33357|issue:39692", @@ -7542,213 +7444,223 @@ "issue:43653|issue:44442", "issue:43381|issue:43701", "issue:43526|issue:44186", - "issue:43334|issue:43531", "issue:44464|issue:45491", + "issue:43334|issue:43531", "issue:44977|issue:45245", - "issue:29942|issue:36010", "issue:43262|issue:44206", + "issue:29942|issue:36010", "issue:29127|issue:33357", "issue:43761|issue:44079", - "issue:29127|issue:35141", - "issue:42915|issue:43931", - "issue:43441|issue:45588" + "issue:29127|issue:35141" ], - "prepared_review_unit_hash": "4ed09d18c624dafd247c92e9d6e459a5b0bb6bf2f002b8c520c1b00fac56b5db", + "prepared_review_unit_hash": "77edb973a1351d1e6c94b2b5aba3a2b1fac90dbb895e76241359e584b61db6d1", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 28, - "serialized_chars": 36120, - "estimated_input_tokens": 9030, - "estimated_eval_tokens": 18316 + "soft_pair_count": 31, + "serialized_chars": 37064, + "estimated_input_tokens": 9266, + "estimated_eval_tokens": 18788 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: the soft-linked pairs mostly share a broad subsystem or surface-level symptom, but not the same concrete bug/change. I would not merge any of the soft edges. The most representative issue is the CLIPVisionModel hidden-states regression because it is a precise, reproducible code-path bug.", - "confidence": 0.84, - "canonical_issue_reason": "No clear duplicate theme across the cluster; issue 43761 is the most concrete, reproducible regression among the set, so it is the best representative issue.", + "summary": "These issues are largely unrelated Hugging Face Transformers bug reports across different subsystems; I don\u2019t see any convincing duplicate pairs, so no soft edges should be merged.", + "confidence": 0.99, + "canonical_issue_reason": "issue:43824 has the strongest impact signal in this set (6 inbound references) and is a clear, self-contained reproducible import failure.", "canonical_pr_reason": null, - "best_issue_reason": "43761 has a narrow failure mode, a specific model path, and a straightforward regression signature, making it the most actionable issue in this mixed cluster.", + "best_issue_reason": "issue:43824 is the best cluster anchor because it is concrete, externally referenced, and easy to identify as a distinct bug report.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44186", + "right": "issue:45198", + "accept": false, + "reason": "Different models and failure modes: LayoutLMv2 tokenizer padding/NER crash vs Wav2Vec2 save/tokenization failure." + }, + { + "left": "issue:43421", + "right": "issue:45491", + "accept": false, + "reason": "Tokenizer post-processor update behavior is unrelated to Gemma3 NaNs from sliding-window/all-padding attention." + }, + { + "left": "issue:44265", + "right": "issue:45198", + "accept": false, + "reason": "torch.export / torch_compilable_check failure is a different code path from Wav2Vec2 tokenization/save_pretrained issues." + }, + { + "left": "issue:43278", + "right": "issue:43381", + "accept": false, + "reason": "Embedding dtype drift in eval is unrelated to the gradient-checkpointing eval-mode limitation." + }, + { + "left": "issue:43994", + "right": "issue:45412", + "accept": false, + "reason": "SigLIP2 bad outputs and RT-DETR memory release are different bugs in different subsystems." + }, { "left": "issue:42915", "right": "issue:42994", "accept": false, - "reason": "Both involve Qwen3-MoE/quantization, but one is an FP8 config failure and the other is a quantized saving problem; different stages and likely different bugs." + "reason": "Qwen3Moe FP8 config failure is not the same as generic quantized model saving failure." }, { "left": "issue:44361", "right": "issue:45198", "accept": false, - "reason": "Different models and different failure surfaces: MLukeTokenizer task errors vs Wav2Vec2 save/tokenization failure." + "reason": "MLukeTokenizer AttributeError is unrelated to Wav2Vec2 save/tokenization behavior." }, { "left": "issue:44206", "right": "issue:44479", "accept": false, - "reason": "Unsupported feature-extractor arg vs video-input regression; same broad area, not the same code-path bug." + "reason": "Unsupported feature arg in LasrFeatureExtractor is a different regression from Qwen video-input behavior." }, { "left": "issue:44871", "right": "issue:44977", "accept": false, - "reason": "Tokenizer/model EOS config mismatch is unrelated to the flash-attention generation failure." - }, - { - "left": "issue:43257", - "right": "issue:43278", - "accept": false, - "reason": "Loading/conversion of Qwen3-MoE weights is distinct from BF16-to-FP32 dtype drift during evaluate." + "reason": "Gemma eos-token config mismatch is unrelated to Qwen3.5 flash-attention generation bugs." }, { "left": "issue:43756", "right": "issue:45440", "accept": false, - "reason": "RoPE layer dropping and DeepSeek implementation divergence are different model discrepancies." + "reason": "Smollm3 RoPE-layer mismatch and DeepseekV3MoE divergence are separate model-implementation issues." }, { - "left": "issue:29942", - "right": "issue:30064", + "left": "issue:43257", + "right": "issue:43278", "accept": false, - "reason": "Failing Flash Attention 2 tests and void segmentation map processing are unrelated issues." + "reason": "Qwen3 MoE weight conversion under accelerate+deepspeed is unrelated to training/eval embedding dtype changes." }, { "left": "issue:44610", "right": "issue:45412", "accept": false, - "reason": "Processor output-size mismatch and RT-DETR memory leaks are not the same underlying bug." + "reason": "Processor image-size mismatch is not the same as RT-DETR memory not being released on delete." + }, + { + "left": "issue:29942", + "right": "issue:30064", + "accept": false, + "reason": "Flash Attention 2 test failures and void segmentation map processing are different failures in different areas." }, { "left": "issue:45103", "right": "issue:45372", "accept": false, - "reason": "auto_docstring annotation parsing crash is unrelated to the mistral_common import dependency break." + "reason": "auto_docstring future-annotations crash is unrelated to the mistral_common ReasoningEffort import error." }, { "left": "issue:43824", "right": "issue:44279", "accept": false, - "reason": "A specific missing import and a generic dependency issue do not indicate the same fix." + "reason": "Missing Qwen2_5_VLForConditionalGeneration import is a specific API/export bug, not a generic dependency issue." }, { "left": "issue:33357", "right": "issue:39692", "accept": false, - "reason": "MacOS CLIP bus error and SigLIP2 doc example errors are different problems." + "reason": "MacOS CLIP bus error and SigLIP2 doc-example errors are different models and failure modes." }, { "left": "issue:45440", "right": "issue:45593", "accept": false, - "reason": "Native model divergence and missing auxiliary losses are separate model-behavior bugs." + "reason": "DeepseekV3MoE implementation divergence is unrelated to D-FINE auxiliary-loss behavior." }, { "left": "issue:29127", "right": "issue:34567", "accept": false, - "reason": "Error-message clarity for LayoutLMv3 is unrelated to TrainerState token-count tracking." + "reason": "LayoutLMv3 error-message clarity and TrainerState token-count tracking are unrelated bugs." }, { "left": "issue:32090", "right": "issue:33290", "accept": false, - "reason": "NoneType broadcast error and Adafactor/deepspeed OOM are different Trainer/deepspeed failures." + "reason": "Trainer NoneType broadcast failure is unrelated to DeepSpeed AdaFactor OOM." }, { "left": "issue:45072", "right": "issue:45229", "accept": false, - "reason": "Bfloat16 dtype mismatches are not the same as multi-GPU CUDA OOM." + "reason": "bfloat16 dtype mismatch in inference is not the same as Gemma4 multi-GPU OOM." }, { "left": "issue:43653", "right": "issue:44442", "accept": false, - "reason": "BigBird special-token registration and AutoTokenizer loading a FastSpeech2Conformer tokenizer are unrelated." + "reason": "BigBirdTokenizer special-token registration and AutoTokenizer loading FastSpeech2ConformerTokenizer are different tokenizer issues." }, { "left": "issue:43381", "right": "issue:43701", "accept": false, - "reason": "Eval-mode gradient-checkpointing restriction and resume checkpoint key mismatch are separate training-state bugs." + "reason": "Gradient checkpointing in eval mode is unrelated to resume-from-checkpoint key mismatch." }, - { - "left": "issue:43526", - "right": "issue:44186", - "accept": false, - "reason": "reduce_labels returning one label and LayoutLMv2Tokenizer padding/truncation crashes are different tokenizer issues." - }, - { - "left": "issue:43334", - "right": "issue:43531", + { + "left": "issue:43526", + "right": "issue:44186", "accept": false, - "reason": "Missing pad_token_id in Qwen3-VL and a Qwen3-MoE sliding_window issue are distinct model/config regressions." + "reason": "BeitImageProcessorFast reduce_labels bug is unrelated to LayoutLMv2Tokenizer crashes on padding/truncation." }, { "left": "issue:44464", "right": "issue:45491", "accept": false, - "reason": "Compiled-forward chunked generation inconsistency and Gemma3 NaNs from all-padding windows are not the same bug." + "reason": "Chunked generation with compiled forward is a different path from Gemma3 NaNs in mixed-length batching." }, { - "left": "issue:44977", - "right": "issue:45245", + "left": "issue:43334", + "right": "issue:43531", "accept": false, - "reason": "Flash-attention generation failure and category-count runtime limit are unrelated." + "reason": "Qwen3-VL pad_token_id loading error and Qwen3-MoE sliding_window behavior are separate issues." }, { - "left": "issue:29942", - "right": "issue:36010", + "left": "issue:44977", + "right": "issue:45245", "accept": false, - "reason": "Flash Attention 2 test failures and GenerationMixin import error are different failure modes." + "reason": "Flash-attention generation regression is unrelated to the category-count limit runtime error." }, { "left": "issue:43262", "right": "issue:44206", "accept": false, - "reason": "Audio chat-template sampling-rate defaulting and unsupported center arg in a feature extractor are unrelated." + "reason": "Audio chat-template sampling-rate default is unrelated to LasrFeatureExtractor's unsupported center arg." }, { - "left": "issue:29127", - "right": "issue:33357", + "left": "issue:29942", + "right": "issue:36010", "accept": false, - "reason": "LayoutLMv3 error messaging and a CLIP MacOS bus error do not share a concrete fix." + "reason": "Flash Attention test failures are unrelated to the GenerationMixin import error." }, { - "left": "issue:43761", - "right": "issue:44079", + "left": "issue:29127", + "right": "issue:33357", "accept": false, - "reason": "The CLIPVisionModel hidden-states regression may involve ModelOutput handling, but this is not clearly the same concrete bug as generic key assignment when a key was previously None." + "reason": "LayoutLMv3 error messaging and CLIP MacOS bus errors are not the same underlying bug." }, { "left": "issue:29127", "right": "issue:35141", "accept": false, - "reason": "LayoutLMv3 error handling and embedding reinitialization in post_init are unrelated." - }, - { - "left": "issue:42915", - "right": "issue:43931", - "accept": false, - "reason": "Qwen3-MoE FP8 failure and Qwen3-VL weight-shape mismatch are different loading issues." - }, - { - "left": "issue:43441", - "right": "issue:45588", - "accept": false, - "reason": "Ministral FlashAttention failure and sink-less flash_attention.py s_aux=None crash are different attention-path bugs." + "reason": "LayoutLMv3 error clarity and output-embedding reinitialization are different bugs in different code paths." } ] }, @@ -7767,21 +7679,21 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ "issue:33290", "issue:34567", "issue:34689", - "issue:35141", "issue:39692", "issue:41628", + "issue:42915", "issue:43232", "issue:43316", - "issue:43381", "issue:43388", + "issue:43441", "issue:43452", "issue:43493", "issue:43525", @@ -7822,22 +7734,24 @@ "issue:45600" ], "soft_pairs": [ + "issue:43441|issue:45588", + "issue:42915|issue:43931", "issue:43232|issue:43388", "issue:43526|issue:44857", "issue:44442|issue:45072", "issue:43723|issue:45479", "issue:44479|issue:45290", - "issue:43526|issue:44610", "issue:45003|issue:45440", + "issue:43526|issue:44610", "issue:44743|issue:44811", "issue:43452|issue:45588", "issue:43931|issue:45216", - "issue:45588|issue:45600", "issue:44977|issue:45584", - "issue:43761|issue:44188", + "issue:45588|issue:45600", "issue:43316|issue:43525", - "issue:43575|issue:43701", + "issue:43761|issue:44188", "issue:43646|issue:44077", + "issue:43575|issue:43701", "issue:43994|issue:44206", "issue:39692|issue:45491", "issue:43575|issue:44387", @@ -7847,220 +7761,204 @@ "issue:44442|issue:44898", "issue:34689|issue:43611", "issue:41628|issue:45372", - "issue:43493|issue:43827", - "issue:43381|issue:43873", - "issue:33290|issue:34689", - "issue:33290|issue:35141", - "issue:44077|issue:44479" + "issue:43493|issue:43827" ], - "prepared_review_unit_hash": "bf2ed62e00b36489b5e3fa679536890ee45cfc5f7d3ad9909c0c4f7f50d6ac3d", + "prepared_review_unit_hash": "d7b00247e999fcfa3808e320b6e13dcfc832c8fa7e45cc8eafbd386f301db5f3", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 36854, - "estimated_input_tokens": 9214, - "estimated_eval_tokens": 18684 + "soft_pair_count": 28, + "serialized_chars": 36198, + "estimated_input_tokens": 9050, + "estimated_eval_tokens": 18356 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous: it mixes unrelated tokenizer/loading, generation/cache, vision processor, quantization/offloading, documentation, and model-specific regression issues. None of the proposed soft pairs look like the same underlying bug or change, so all soft edges should be rejected.", - "confidence": 0.88, - "canonical_issue_reason": "Issue 43873 is the broadest and most active item here (open, with multiple inbound references) and is the closest thing to a cluster anchor, but the surrounding issues are too unrelated for it to be a true duplicate hub.", + "summary": "The soft pairs mostly span unrelated regressions in different subsystems or models, so I rejected all of them; the only plausible hub is the flash-attention-related issue around `s_aux=None`, but even that is not the same concrete bug as the paired reports.", + "confidence": 0.77, + "canonical_issue_reason": "issue:45588 is the best central anchor in this set: it is open and has multiple flash-attention-related soft links, making it the clearest hub among the provided issues.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43873 is the best administrative representative because it has the most activity and references, but it still does not meaningfully cover the rest of the set.", + "best_issue_reason": "issue:45588 is the strongest issue candidate overall because it is the most connected and most general among the soft-linked flash-attention reports.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43441", + "right": "issue:45588", + "accept": false, + "reason": "Both involve FlashAttention, but they describe different failure modes and likely different fixes: Ministral-3 RC regression vs `s_aux=None` on sink-less models." + }, + { + "left": "issue:42915", + "right": "issue:43931", + "accept": false, + "reason": "Different models and different errors: FineGrainedFP8Config failure vs weight-shape mismatch on Qwen3-VL loading." + }, { "left": "issue:43232", "right": "issue:43388", "accept": false, - "reason": "Generation kwargs update logic and metric label gathering are unrelated code paths." + "reason": "Unrelated code paths: generation kwargs after `sync_gpus` vs metric gathering dropping tuple labels." }, { "left": "issue:43526", "right": "issue:44857", "accept": false, - "reason": "ImageProcessor label reduction and AMP/CUDA loss crashes are different bugs." + "reason": "Different components and symptoms: BeitImageProcessorFast label reduction bug vs LwDetrImageLoss AMP crash." }, { "left": "issue:44442", "right": "issue:45072", "accept": false, - "reason": "Tokenizer loading failure and bfloat16 dtype mismatch are unrelated issues." + "reason": "Tokenizer loading failure and dtype mismatch are unrelated bugs." }, { "left": "issue:43723", "right": "issue:45479", "accept": false, - "reason": "Tokenizer loading regression and sequence-classification zero-loss bug do not share the same fix." + "reason": "Tokenizer loading in v5 and zero-loss classification behavior are not the same issue." }, { "left": "issue:44479", "right": "issue:45290", "accept": false, - "reason": "Video-input regression and chat template tool-call crash affect different subsystems." + "reason": "Video-input regression for Qwen VL models is separate from chat-template tool-call handling." }, { - "left": "issue:43526", - "right": "issue:44610", + "left": "issue:45003", + "right": "issue:45440", "accept": false, - "reason": "Label reduction in an image processor is unrelated to OmDet input-size mismatch." + "reason": "`sys.modules` access in modeling utils is unrelated to DeepSeekV3MoE divergence from remote implementation." }, { - "left": "issue:45003", - "right": "issue:45440", + "left": "issue:43526", + "right": "issue:44610", "accept": false, - "reason": "Unsafe sys.modules access and DeepSeekV3MoE divergence are distinct problems." + "reason": "Both are model/processor-related, but one is label reduction and the other is input-size mismatch." }, { "left": "issue:44743", "right": "issue:44811", "accept": false, - "reason": "Qwen recurrent-state reset and Whisper batch_decode token handling are not the same bug." + "reason": "Cache/reset behavior in recurrent states is unrelated to Whisper batch decode special-token handling." }, { "left": "issue:43452", "right": "issue:45588", "accept": false, - "reason": "GGUF loading breakage and flash-attention sink handling are different code paths." + "reason": "Tokenizer/model loading with `gguf_file` is a different failure mode than FlashAttention `s_aux=None`." }, { "left": "issue:43931", "right": "issue:45216", "accept": false, - "reason": "Shape-mismatch loading and save_pretrained checkpoint corruption are related to the model family but not the same concrete defect." - }, - { - "left": "issue:45588", - "right": "issue:45600", - "accept": false, - "reason": "Flash-attention AttributeError and removed Sam3LiteText auto-mapping references are unrelated." + "reason": "Both concern Qwen models, but one is a loading shape mismatch and the other is a broken `save_pretrained` checkpoint." }, { "left": "issue:44977", "right": "issue:45584", "accept": false, - "reason": "Qwen flash-attention generation and Whisper empty-transcription generation failures are different bugs." + "reason": "Qwen3.5 FlashAttention generation regression and Whisper empty-transcription generation failure are unrelated." }, { - "left": "issue:43761", - "right": "issue:44188", + "left": "issue:45588", + "right": "issue:45600", "accept": false, - "reason": "CLIP hidden_states regression and attention-kernel divergence are separate failures." + "reason": "FlashAttention `s_aux=None` crash is unrelated to removed `Sam3LiteText` references in auto mappings." }, { "left": "issue:43316", "right": "issue:43525", "accept": false, - "reason": "Gemma3TextConfig API mismatch and missing Llama4 pad_token_id are not the same issue." + "reason": "Different config API mismatches: Gemma3TextConfig discrepancy vs missing `pad_token_id` on Llama4Config." }, { - "left": "issue:43575", - "right": "issue:43701", + "left": "issue:43761", + "right": "issue:44188", "accept": false, - "reason": "OOM during large-model TP loading and resume_from_checkpoint key mismatch are unrelated." + "reason": "CLIPVision hidden-state regression and attention-kernel divergence under `torch.compile` are separate bugs." }, { "left": "issue:43646", "right": "issue:44077", "accept": false, - "reason": "Generic custom model initialization breakage and patchtsmixer post_init validation are not the same concrete bug." + "reason": "Generic custom-model init breakage is not the same as the `patchtsmixer` `post_init` constraint." + }, + { + "left": "issue:43575", + "right": "issue:43701", + "accept": false, + "reason": "Tensor-parallel OOM and checkpoint key mismatch are unrelated." }, { "left": "issue:43994", "right": "issue:44206", "accept": false, - "reason": "SigLIP2 bad outputs and LasrFeatureExtractor center-arg crash are different problems." + "reason": "SigLIP2 wrong outputs and LastFeatureExtractor center-arg crash are different problems." }, { "left": "issue:39692", "right": "issue:45491", "accept": false, - "reason": "Documentation example errors and Gemma3 NaN embeddings are unrelated." + "reason": "SigLIP2 docs/example errors are unrelated to Gemma3 mixed-length NaN embeddings." }, { "left": "issue:43575", "right": "issue:44387", "accept": false, - "reason": "Tensor-parallel load OOM and int4 reserved-memory OOM have different causes and fixes." + "reason": "Both mention memory, but one is model loading OOM and the other is int4 reserved-memory growth." }, { "left": "issue:43873", "right": "issue:45588", "accept": false, - "reason": "Quantization offloading issues and flash-attention sink-less crashes are unrelated." + "reason": "Quantization/offloading behavior is unrelated to the FlashAttention sink-less crash." }, { "left": "issue:43673", "right": "issue:44743", "accept": false, - "reason": "Missing generation cache in v5.0.0 and Qwen recurrent-state reset are not the same bug." + "reason": "Both involve cache/generation, but they affect different mechanisms and code paths." }, { "left": "issue:33290", "right": "issue:34567", "accept": false, - "reason": "Adafactor/deepspeed OOM and TrainerState token-count tracking are unrelated." + "reason": "Deepspeed Adafactor OOM and `num_input_tokens_seen` not updating are unrelated trainer issues." }, { "left": "issue:44442", "right": "issue:44898", "accept": false, - "reason": "Tokenizer loading failure and Perceiver image-resolution failure are different areas." + "reason": "Tokenizer loading failure is unrelated to Perceiver interpolation/non-default resolution failure." }, { "left": "issue:34689", "right": "issue:43611", "accept": false, - "reason": "Both involve loading, but they target different model families and different root causes." + "reason": "Different release regressions and different model-loading assumptions; not the same concrete bug." }, { "left": "issue:41628", "right": "issue:45372", "accept": false, - "reason": "Different missing imports in different processor-loading paths; not the same defect." + "reason": "Both are import errors, but they involve different missing symbols and different loading flows." }, { "left": "issue:43493", "right": "issue:43827", "accept": false, - "reason": "SigLIP2 implementation mismatch and stale docs after pipeline removal are unrelated." - }, - { - "left": "issue:43381", - "right": "issue:43873", - "accept": false, - "reason": "Gradient-checkpointing eval-mode restriction and quantization offloading are separate issues." - }, - { - "left": "issue:33290", - "right": "issue:34689", - "accept": false, - "reason": "Optimizer OOM and model-loading regression do not share a code fix." - }, - { - "left": "issue:33290", - "right": "issue:35141", - "accept": false, - "reason": "OOM in training and embedding reinitialization during post_init are unrelated." - }, - { - "left": "issue:44077", - "right": "issue:44479", - "accept": false, - "reason": "PatchTSMixer post_init policy and Qwen video-input regression are different bugs." + "reason": "SigLIP2 implementation mismatch and docs still referencing removed `pipeline()` are unrelated." } ] }, @@ -8079,8 +7977,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -8090,6 +7988,7 @@ "issue:28282", "issue:33290", "issue:34689", + "issue:35141", "issue:36010", "issue:36331", "issue:41950", @@ -8113,6 +8012,7 @@ "issue:43856", "issue:43873", "issue:43931", + "issue:44077", "issue:44162", "issue:44190", "issue:44442", @@ -8120,7 +8020,6 @@ "issue:44496", "issue:44514", "issue:44617", - "issue:44734", "issue:44792", "issue:44855", "issue:44871", @@ -8129,14 +8028,17 @@ "issue:45127", "issue:45137", "issue:45200", - "issue:45406", "issue:45584" ], "soft_pairs": [ + "issue:43381|issue:43873", + "issue:33290|issue:34689", + "issue:33290|issue:35141", + "issue:44077|issue:44479", "issue:43720|issue:44442", "issue:43643|issue:44913", - "issue:43749|issue:43931", "issue:44190|issue:44496", + "issue:43749|issue:43931", "issue:43873|issue:44871", "issue:43540|issue:44514", "issue:44162|issue:44792", @@ -8150,193 +8052,210 @@ "issue:43388|issue:43688", "issue:43065|issue:44617", "issue:16998|issue:22355", - "issue:41950|issue:43475", "issue:43316|issue:43716", + "issue:41950|issue:43475", "issue:25251|issue:28282", "issue:45127|issue:45200", "issue:43761|issue:44479", "issue:43295|issue:43825", - "issue:44871|issue:45584", - "issue:44734|issue:45406" + "issue:44871|issue:45584" ], - "prepared_review_unit_hash": "f3cd8ef27bb3f3e6858aeccfe903a6e10359cc207d03e568c2a9644e47d0a4aa", + "prepared_review_unit_hash": "cd53b91dc9e07fd57b93d3e9e41a20c335c06cb91c1d332ce2067861b633f736", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 25, - "serialized_chars": 34627, - "estimated_input_tokens": 8657, - "estimated_eval_tokens": 17570 + "soft_pair_count": 28, + "serialized_chars": 35511, + "estimated_input_tokens": 8878, + "estimated_eval_tokens": 18012 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: the items cover unrelated bugs across loading, configs, pipelines, training, and multimodal models. None of the soft-similarity pairs look like the same underlying issue, so all should remain separate.", - "confidence": 0.94, - "canonical_issue_reason": "No single canonical issue stands out; the issues span distinct code paths and models, and the pairwise similarities are mostly superficial title overlap.", + "summary": "This cluster is not a true duplicate set: the paired items span unrelated bugs across loading, training, tokenization, multimodal processors, and config semantics. I would not merge any soft edge.", + "confidence": 0.14, + "canonical_issue_reason": "No single issue cleanly represents the cluster; the items are too heterogeneous to treat one as canonical.", "canonical_pr_reason": null, - "best_issue_reason": "If one issue must represent the cluster, issue 43761 is the clearest self-contained regression report, but it is not a true duplicate anchor for the rest of the set.", + "best_issue_reason": "issue:43749 has the strongest discussion/inbound activity, but it still only represents one of many unrelated failure modes.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43381", + "right": "issue:43873", + "accept": false, + "reason": "Gradient checkpointing in eval mode is unrelated to quantization/offloading behavior." + }, + { + "left": "issue:33290", + "right": "issue:34689", + "accept": false, + "reason": "Different failure modes: Deepspeed/Adafactor OOM vs Llama 3.2 Vision model-loading regression." + }, + { + "left": "issue:33290", + "right": "issue:35141", + "accept": false, + "reason": "Training-memory OOM and token-embedding reinitialization are distinct bugs." + }, + { + "left": "issue:44077", + "right": "issue:44479", + "accept": false, + "reason": "Unrelated model-specific config validation vs video-input regression in Qwen VL models." + }, { "left": "issue:43720", "right": "issue:44442", "accept": false, - "reason": "BitNet packed-weight loading vs AutoTokenizer/FastSpeech2 tokenizer loading; unrelated failure modes." + "reason": "BitNet packed-weight loading and tokenizer loading are different code paths." }, { "left": "issue:43643", "right": "issue:44913", "accept": false, - "reason": "trust_remote_code config field loss vs GPTNeoX rotary_pct reset on reload; different config bugs." + "reason": "Missing fields from trust_remote_code config loading is unrelated to GPTNeoX rotary_pct persistence." }, { - "left": "issue:43749", - "right": "issue:43931", + "left": "issue:44190", + "right": "issue:44496", "accept": false, - "reason": "FSDP CPU RAM-efficient loading breakage vs Qwen3-VL weight-shape mismatch; different loading paths." + "reason": "Local dataset loading in a training script is unrelated to unrecognized model/config loading." }, { - "left": "issue:44190", - "right": "issue:44496", + "left": "issue:43749", + "right": "issue:43931", "accept": false, - "reason": "local dataset script issue vs unrecognized model/config.json error; unrelated subsystems." + "reason": "FSDP CPU RAM-efficient loading and Qwen3-VL shape mismatch are separate load-time issues." }, { "left": "issue:43873", "right": "issue:44871", "accept": false, - "reason": "quantization/offloading behavior vs Gemma eos_token_id mismatch; not the same bug." + "reason": "Quantization/offloading behavior is unrelated to Gemma-3 eos_token_id inconsistency." }, { "left": "issue:43540", "right": "issue:44514", "accept": false, - "reason": "Qwen3OmniMoe video-input processing vs Qwen2_5_VL batched chat-template padding crash; distinct code paths." + "reason": "Both are video-related, but they affect different Qwen processors and different error conditions." }, { "left": "issue:44162", "right": "issue:44792", "accept": false, - "reason": "ESM2 model breakage vs janus image-generation test failure; unrelated models and symptoms." + "reason": "ESM2 breakage and janus image-generation test failure do not share the same underlying bug." }, { "left": "issue:43316", "right": "issue:44877", "accept": false, - "reason": "Gemma3TextConfig API discrepancy vs strict granite_speech config loading; different config issues." + "reason": "Gemma3TextConfig API mismatch is unrelated to strict config loading for granite_speech." }, { "left": "issue:43381", "right": "issue:43856", "accept": false, - "reason": "gradient checkpointing in eval mode vs Qwen3 MoE training memory usage; different problems." + "reason": "Eval-mode gradient checkpointing is not the same as MoE training memory inefficiency." }, { "left": "issue:33290", "right": "issue:36331", "accept": false, - "reason": "Adafactor OOM in deepspeed vs CustomTrainer signature mismatch; unrelated training bugs." + "reason": "Deepspeed OOM with Adafactor is unrelated to CustomTrainer.compute_loss signature breakage." }, { "left": "issue:34689", "right": "issue:36010", "accept": false, - "reason": "Llama 3.2 vision model loading regression vs GenerationMixin import error; different code paths." + "reason": "Llama 3.2 Vision model loading and GenerationMixin import failure are different regressions." }, { "left": "issue:44855", "right": "issue:45137", "accept": false, - "reason": "Python 3.13 IndentationError in torch.jit parsing vs DeepSpeed ZeRO3 deque underflow; unrelated." + "reason": "Python 3.13 import parsing error and DeepSpeed ZeRO3 deque error are unrelated." }, { "left": "issue:42175", "right": "issue:43316", "accept": false, - "reason": "packaging/backends installation issue vs Gemma3TextConfig API discrepancy; not a duplicate." + "reason": "Backend packaging omission has no overlap with Gemma3TextConfig API discrepancy." }, { "left": "issue:43404", "right": "issue:43441", "accept": false, - "reason": "Mistral3 lm_head tying bug vs Ministral-3 FlashAttention failure; same family but different concrete problems." + "reason": "Same model family, but weight tying and FlashAttention failures are distinct code-path bugs." }, { "left": "issue:43388", "right": "issue:43688", "accept": false, - "reason": "metrics batch-label truncation vs auxiliary-loss normalization; distinct training/eval logic." + "reason": "Metric gathering label truncation and auxiliary-loss normalization are unrelated training issues." }, { "left": "issue:43065", "right": "issue:44617", "accept": false, - "reason": "dummy Conv2d implementation question vs Sam3Video CUDA OOM; different issues." + "reason": "Dummy Conv2d in Sam3PixelDecoder is not the same as Sam3Video OOM behavior." }, { "left": "issue:16998", "right": "issue:22355", "accept": false, - "reason": "model_max_length question vs missing transformers.onnx module; unrelated topics." + "reason": "Model_max_length behavior and missing transformers.onnx module are unrelated." }, { - "left": "issue:41950", - "right": "issue:43475", + "left": "issue:43316", + "right": "issue:43716", "accept": false, - "reason": "video-classification pipeline processor lookup vs Sam3VisionEncoderOutput missing attribute; different bugs." + "reason": "Gemma3TextConfig API differences do not match Mistral-3 image preprocessor dtype mismatch." }, { - "left": "issue:43316", - "right": "issue:43716", + "left": "issue:41950", + "right": "issue:43475", "accept": false, - "reason": "Gemma3TextConfig API mismatch vs Mistral-3 image-preprocessor dtype mismatch; unrelated models." + "reason": "Video-classification processor lookup and missing Sam3VisionEncoderOutput attribute are separate issues." }, { "left": "issue:25251", "right": "issue:28282", "accept": false, - "reason": "pipeline top_k output-shape change vs AutoModel PyTorch import error; not the same bug." + "reason": "Pipeline output-shape behavior is unrelated to missing PyTorch import errors." }, { "left": "issue:45127", "right": "issue:45200", "accept": false, - "reason": "LoRA merge with extended vocabulary/tied embeddings vs Gemma 4 mm_token_type_ids defaulting; different mechanisms." + "reason": "LoRA merge collapse with tied embeddings is not the same as missing mm_token_type_ids defaults." }, { "left": "issue:43761", "right": "issue:44479", "accept": false, - "reason": "CLIPVisionModel hidden_states regression vs broad v5 video-input regression across Qwen models; different subsystems." + "reason": "CLIPVisionModel hidden_states regression and Qwen video-input regression are different models and failures." }, { "left": "issue:43295", "right": "issue:43825", "accept": false, - "reason": "processor.tokenizer/images-to-tokenizer regression vs translation-task error message; one API break, one messaging bug." + "reason": "Processor.tokenizer regression is unrelated to an inaccurate pipeline error message." }, { "left": "issue:44871", "right": "issue:45584", "accept": false, - "reason": "Gemma eos_token_id config mismatch vs Whisper empty transcription after align_special_tokens; different models and code paths." - }, - { - "left": "issue:44734", - "right": "issue:45406", - "accept": false, - "reason": "serve KV-cache continuation indexing bug vs Gemma4Processor missing _tokenizer; unrelated failures." + "reason": "Gemma-3 eos_token_id mismatch and Whisper empty-transcription handling are different generation bugs." } ] }, @@ -8355,8 +8274,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -8403,12 +8322,14 @@ "issue:45357", "issue:45362", "issue:45381", + "issue:45406", "issue:45542", "issue:45563", "issue:45584", "issue:45600" ], "soft_pairs": [ + "issue:44734|issue:45406", "issue:44077|issue:44534", "issue:43295|issue:44871", "issue:43550|issue:43701", @@ -8417,8 +8338,8 @@ "issue:43906|issue:45290", "issue:44617|issue:45584", "issue:43761|issue:44242", - "issue:43452|issue:44496", "issue:43931|issue:44060", + "issue:43452|issue:44496", "issue:45563|issue:45584", "issue:44521|issue:45381", "issue:43575|issue:43742", @@ -8434,165 +8355,171 @@ "issue:43827|issue:43994", "issue:44534|issue:44913", "issue:41720|issue:45245", - "issue:44938|issue:45542", - "issue:44792|issue:45600", "issue:44190|issue:44964", - "issue:45127|issue:45357" + "issue:44938|issue:45542", + "issue:45127|issue:45357", + "issue:44792|issue:45600" ], - "prepared_review_unit_hash": "6a3f814b39c43276d7228d2d59fbfe6b1bec610785dd555d4b5c848dd1911857", + "prepared_review_unit_hash": "f4c2c006e106aa45c9e0e87196e9f7f005ee4e470613b9f418ff392c8509ddb6", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 29, - "serialized_chars": 35677, - "estimated_input_tokens": 8920, - "estimated_eval_tokens": 18096 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36583, + "estimated_input_tokens": 9146, + "estimated_eval_tokens": 18548 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is a loose, heterogeneous cluster of Transformers issues spanning unrelated regressions, loading failures, generation bugs, and model-specific problems. None of the soft pairs look like true duplicates; they usually share only a broad subsystem or version context. Issue 44521 is the best cluster anchor because it is the most active and broadly applicable multimodal bug in the set.", - "confidence": 0.74, - "canonical_issue_reason": "Issue 44521 is the most central/open and has the strongest discussion/inbound reference activity, so it is the best representative anchor for this otherwise mixed cluster.", + "summary": "This cluster is heterogeneous: the soft-similarity pairs mostly share broad areas (serving, multimodal, generation, config loading) but do not describe the same concrete bug/change. I would reject all proposed merges; no PRs are present. If one issue must anchor the set, issue 44521 is the strongest open representative because it has the most inbound references and broad activity.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 44521 is the best representative by current activity and inbound references, but the cluster has no single true duplicate center.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 44521 is broad, currently active, and has the most external references, making it the most useful issue to track as the cluster's representative.", + "best_issue_reason": "Open, actively discussed, and the most referenced issue in the set, so it is the best practical anchor.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44734", + "right": "issue:45406", + "accept": false, + "reason": "Both involve transformers serve, but one is KV-cache tensor indexing and the other is a missing processor attribute; different failures." + }, { "left": "issue:44077", "right": "issue:44534", "accept": false, - "reason": "Different failure classes: optional post_init behavior in patchtsmixer vs junk-filled non-persistent buffers in v5." + "reason": "Unrelated bugs: patchtsmixer post_init validation vs non-persistent buffer corruption in v5." }, { "left": "issue:43295", "right": "issue:44871", "accept": false, - "reason": "Both involve model/processor config, but one is a tokenizer/processor regression and the other is an eos_token_id mismatch." + "reason": "Different regressions: processor/tokenizer access breakage vs eos_token_id config mismatch." }, { "left": "issue:43550", "right": "issue:43701", "accept": false, - "reason": "Unrelated bugs: torch.compile SDPA failure for Bamba vs checkpoint key mismatch on resume." + "reason": "torch.compile/SDPA model failure is unrelated to resume_from_checkpoint key mismatches." }, { "left": "issue:43606", "right": "issue:43701", "accept": false, - "reason": "CPU offload device mismatch is a different code path from resume_from_checkpoint key mismatch." + "reason": "CPU offload device mismatch is a different code path from checkpoint key handling." }, { "left": "issue:43646", "right": "issue:44913", "accept": false, - "reason": "Custom model init breakage is not the same as GPTNeoX config reload losing rotary_pct." + "reason": "Custom model init breakage is not the same as rotary_pct not persisting on reload." }, { "left": "issue:43906", "right": "issue:45290", "accept": false, - "reason": "Both touch chat/template behavior, but one is an isolated reproduction of an older issue and the other is a tokenizer crash on tool-call assistant messages." + "reason": "The reproduction target and failure mode differ: isolated prior issue reproduction vs chat template tool-call crash." }, { "left": "issue:44617", "right": "issue:45584", "accept": false, - "reason": "Completely different models and symptoms: Sam3Video CUDA OOM vs Whisper empty transcription handling." + "reason": "CUDA OOM in Sam3Video is unrelated to Whisper empty-transcription failure after align_special_tokens." }, { "left": "issue:43761", "right": "issue:44242", "accept": false, - "reason": "Different bugs in different models: CLIPVision hidden_states omission vs MoE load-balancing loss not added." + "reason": "CLIPVision hidden_states bug is different from MoE load-balancing loss omission." }, { - "left": "issue:43452", - "right": "issue:44496", + "left": "issue:43931", + "right": "issue:44060", "accept": false, - "reason": "Both are loading failures, but one is gguf_file handling and the other is an unrecognized config/model_type issue." + "reason": "Model weight shape mismatch is not the same as the incorrect tied-weights warning." }, { - "left": "issue:43931", - "right": "issue:44060", + "left": "issue:43452", + "right": "issue:44496", "accept": false, - "reason": "Shape-mismatch loading for Qwen3-VL is unrelated to the tied-weights warning in Qwen3-Next." + "reason": "GGUF tokenizer/model loading breakage is a different issue from missing model_type in config.json." }, { "left": "issue:45563", "right": "issue:45584", "accept": false, - "reason": "Paged generate warning bug is unrelated to Whisper empty transcription after align_special_tokens." + "reason": "Stale num_return_sequences warning is unrelated to Whisper generation crashing on empty output." }, { "left": "issue:44521", "right": "issue:45381", "accept": false, - "reason": "Both are multimodal, but assistant_masks being all zero is not the same bug as wrong vision_position_ids in Qwen2.5-VL video input." + "reason": "Both are multimodal/VL-related, but assistant_mask generation and vision_position_ids are distinct bugs." }, { "left": "issue:43575", "right": "issue:43742", "accept": false, - "reason": "Tensor-parallel OOM for Qwen2-57B is a different problem from a MobileLLM config key error." + "reason": "TP OOM on Qwen2-57B and key errors when loading MobileLLM are different load-time failures." }, { "left": "issue:42617", "right": "issue:43575", "accept": false, - "reason": "3d_parallel.py runtime failure is not the same as model-loading OOM with tensor parallelism." + "reason": "3d_parallel.py runtime failure does not match Qwen2 tensor-parallel OOM." }, { "left": "issue:44479", "right": "issue:44734", "accept": false, - "reason": "Video-input regression in VL models is a different code path from KV-cache continuation indexing in serve." + "reason": "Video-input regression across Qwen VL models is unrelated to serve-side KV-cache indexing." }, { "left": "issue:44933", "right": "issue:45362", "accept": false, - "reason": "Missing import from image_utils is a packaging/API issue, not the same as Qwen3.5 chat generation crashes." + "reason": "Missing image_utils import and Qwen3.5 chat crashes are different defects." }, { "left": "issue:41720", "right": "issue:43575", "accept": false, - "reason": "cudaErrorAssert on auto device mapping is distinct from a tensor-parallel out-of-memory failure." + "reason": "Auto device mapping cudaErrorAssert is not the same as tensor-parallel OOM." }, { "left": "issue:32090", "right": "issue:35141", "accept": false, - "reason": "Trainer GPU broadcast NoneType error is unrelated to embedding resize/post_init reinitialization." + "reason": "Trainer GPU broadcast NoneType error is unrelated to embedding reinitialization in post_init." }, { "left": "issue:44464", "right": "issue:45563", "accept": false, - "reason": "Compiled-forward chunked generation inconsistency is a different bug from a stale warning in paged generate()." + "reason": "Compiled-forward generation inconsistency is not the same as a stale warning." }, { "left": "issue:43065", "right": "issue:44610", "accept": false, - "reason": "Sam3PixelDecoder dummy Conv2d and OmDet-Turbo processor size mismatch are separate vision-model issues." + "reason": "Dummy Conv2d in Sam3PixelDecoder and OmDet-Turbo input-size mismatch are separate model issues." }, { "left": "issue:44464", "right": "issue:44871", "accept": false, - "reason": "Compiled generation inconsistency and eos_token_id config mismatch are unrelated." + "reason": "Chunked generation inconsistency is unrelated to eos_token_id configuration drift." }, { "left": "issue:44792", @@ -8604,37 +8531,55 @@ "left": "issue:43827", "right": "issue:43994", "accept": false, - "reason": "Docs still referencing pipeline() is a documentation break, not the same as SigLIP2 producing bad outputs." + "reason": "Documentation references to removed pipeline() are not the same as SigLIP2 producing bad outputs." }, { "left": "issue:44534", "right": "issue:44913", "accept": false, - "reason": "Non-persistent buffer junk in v5 and GPTNeoX rotary_pct reload regression affect different mechanisms." + "reason": "Non-persistent buffer junk in v5 is unrelated to GPTNeoX rotary_pct not persisting." }, { "left": "issue:41720", "right": "issue:45245", "accept": false, - "reason": "Qwen3 device-mapping assert and category-count limit error are unrelated." + "reason": "Qwen3 auto-device-map CUDA assert is unrelated to the 2^24 categories limit error." }, { - "left": "issue:44938", - "right": "issue:45542", + "left": "issue:44190", + "right": "issue:44964", "accept": false, - "reason": "Python 3.14 load failure is an environment/runtime compatibility issue, not the TensorBoard-without-TensorFlow backend error." + "reason": "Local dataset loading in an example script is a different problem from Phi-4 multimodal model loading." }, { - "left": "issue:44190", - "right": "issue:44964", + "left": "issue:44938", + "right": "issue:45542", "accept": false, - "reason": "Local dataset loading failure and Phi-4-multimodal model loading failure are different problems." + "reason": "Python 3.14 import/load failure is unrelated to the TensorBoard-without-TensorFlow backend error." }, { "left": "issue:45127", "right": "issue:45357", "accept": false, - "reason": "LoRA merge collapse with extended vocab and incorrect visual encoder save keys are separate regression areas." + "reason": "LoRA merge collapse with extended vocab is a different bug from incorrect visual encoder keys being saved." + }, + { + "left": "issue:44242", + "right": "issue:44521", + "accept": false, + "reason": "MoE load-balancing loss omission is unrelated to assistant mask generation for multimodal chat templates." + }, + { + "left": "issue:44617", + "right": "issue:45584", + "accept": false, + "reason": "Sam3Video OOM and Whisper empty-transcription crash are unrelated generation failures." + }, + { + "left": "issue:44792", + "right": "issue:45600", + "accept": false, + "reason": "Janus image-generation test failure is unrelated to CI breakage from removed Sam3LiteText configs." } ] }, @@ -8717,37 +8662,37 @@ "issue:43644|issue:44913", "issue:45276|issue:45584", "issue:43976|issue:44855", - "issue:44077|issue:44079", "issue:43404|issue:45216", + "issue:44077|issue:44079", "issue:43550|issue:44190", - "issue:44112|issue:45600", "issue:43116|issue:44190", + "issue:44112|issue:45600", "issue:43295|issue:45276", "issue:45245|issue:45584", "issue:41950|issue:44016", "issue:39692|issue:44190", - "issue:44734|issue:45632", "issue:36246|issue:41628", + "issue:44734|issue:45632", "issue:43479|issue:45375", "issue:45406|issue:45632", - "issue:44561|issue:45632", "issue:36246|issue:38617", + "issue:44561|issue:45632", "issue:44514|issue:44734", - "issue:38175|issue:43065", "issue:44016|issue:44220", + "issue:38175|issue:43065", "issue:42175|issue:44279", "issue:44683|issue:45446", "issue:44466|issue:44913", "issue:44246|issue:45003" ], - "prepared_review_unit_hash": "e83f5375faa75e6db9b24826b3397f4c797310f12b8c78d9e87cfdefaf67a7e6", + "prepared_review_unit_hash": "35db967a3dfbef3fd5e89fa1b92ba62e26932da9cdb718fe8ab845ad2ee5b754", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 48, @@ -8762,198 +8707,198 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is highly heterogeneous: the soft pairs span unrelated import errors, model-specific regressions, docs/notebook issues, and pipeline bugs. I found no convincing duplicate pairs, so the set should not be merged into a single canonical bug/PR cluster.", - "confidence": 0.96, - "canonical_issue_reason": "No canonical issue: the items cover many distinct subsystems and failure modes, with no shared concrete code path or single underlying defect tying them together.", + "summary": "This cluster is heterogeneous; the soft edges mostly connect unrelated bugs that only share broad subsystem vocabulary (imports, configs, cache, multimodal, CI, etc.). I would not merge any of the proposed pairs as duplicates.", + "confidence": 0.72, + "canonical_issue_reason": "issue:44561 is the best representative: it describes a clear, high-impact framework regression in `trust_remote_code` handling caused by a removed utility, with broad applicability beyond one model family.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43295 is the strongest standalone report: it\u2019s a clear regression with a specific version, reproducible custom-model breakage, and a concrete API interaction to fix.", + "best_issue_reason": "issue:44561 has the strongest combination of clarity, scope, and downstream impact, so it is the best single issue anchor for this cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43388", "right": "issue:43582", "accept": false, - "reason": "Different failures: metric gathering label truncation vs Apple Silicon allocator warmup TypeError." + "reason": "Different failures: metric gathering last-batch label handling vs Apple Silicon allocator warmup TypeError." }, { "left": "issue:43761", "right": "issue:44743", "accept": false, - "reason": "Unrelated models and bugs: CLIP hidden_states regression vs Qwen3 recurrent-state reset with cache." + "reason": "Unrelated model bugs: CLIP vision hidden states regression vs Qwen3 recurrent state reset with cache." }, { "left": "issue:44855", "right": "issue:45588", "accept": false, - "reason": "Different code paths: Python 3.13 import/parse error in DebertaV2 vs flash-attention crash on s_aux=None." + "reason": "Different code paths and symptoms: Python 3.13 JIT parsing/import indentation issue vs flash-attention `s_aux=None` crash." }, { "left": "issue:43650", "right": "issue:44220", "accept": false, - "reason": "No shared defect: a vague data issue vs an audio feature-extraction failure." + "reason": "No shared underlying bug; one is an unclear data-related issue, the other is a specific `_torch_extract_fbank_features()` problem." }, { "left": "issue:44355", "right": "issue:44855", "accept": false, - "reason": "Both are import/compile-related, but one is about compiled Python file errors and the other is a Python 3.13 JIT parsing issue." + "reason": "Both involve import/compile-time errors, but the concrete root causes and affected paths are different." }, { "left": "issue:43540", "right": "issue:44734", "accept": false, - "reason": "Different components: Qwen3Omni video input validation vs serve KV-cache continuation indexing." + "reason": "Different multimodal/runtime bugs: Qwen3OmniMoe video processing ValueError vs serve KV-cache tensor indexing crash." }, { "left": "issue:43644", "right": "issue:44913", "accept": false, - "reason": "Different bugs: non-persistent buffer junk values vs GPTNeoX config reload losing rotary_pct." + "reason": "One is about non-persistent buffer serialization junk; the other is GPTNeoX config reload losing `rotary_pct`." }, { "left": "issue:45276", "right": "issue:45584", "accept": false, - "reason": "Separate model behaviors: Gemma4 embedding resize bug vs Whisper empty-transcription generation failure." + "reason": "Gemma4 embedding resize mismatch is unrelated to Whisper empty-transcription generation after `align_special_tokens`." }, { "left": "issue:43976", "right": "issue:44855", "accept": false, - "reason": "Different regressions: package/Python-version compatibility vs DebertaV2 import parsing on 3.13." + "reason": "Different regressions: package Python-version compatibility vs a Python 3.13 TorchScript parsing/indentation failure." }, { - "left": "issue:44077", - "right": "issue:44079", + "left": "issue:43404", + "right": "issue:45216", "accept": false, - "reason": "Distinct core issues: optional post_init handling vs ModelOutput key assignment semantics." + "reason": "Both mention weights/checkpointing, but they affect different models and concrete failure modes; not the same bug." }, { - "left": "issue:43404", - "right": "issue:45216", + "left": "issue:44077", + "right": "issue:44079", "accept": false, - "reason": "Both mention tied weights, but they affect different models and different save/serialization paths." + "reason": "Unrelated issues: `post_init` validation vs `ModelOutput` key assignment logic." }, { "left": "issue:43550", "right": "issue:44190", "accept": false, - "reason": "Unrelated: Bamba torch.compile SDPA failure vs local dataset loading in an example script." + "reason": "Different domains: `torch.compile` SDPA failure in Bamba vs local dataset loading in an image-classification example." }, { - "left": "issue:44112", - "right": "issue:45600", + "left": "issue:43116", + "right": "issue:44190", "accept": false, - "reason": "Both touch CI, but they describe different breakages in different files and subsystems." + "reason": "Separate example-script bugs with different symptoms and likely different causes." }, { - "left": "issue:43116", - "right": "issue:44190", + "left": "issue:44112", + "right": "issue:45600", "accept": false, - "reason": "Different example-script failures: empty multi-label results vs inability to load a local dataset." + "reason": "Both are CI-related, but one is a stale device-override test and the other is a broken auto-mapping reference." }, { "left": "issue:43295", "right": "issue:45276", "accept": false, - "reason": "Different model APIs and symptoms: processor/tokenizer regression vs embedding resize not propagating." + "reason": "Different processor/tokenizer regression vs Gemma4 embedding resize behavior; no shared concrete code-path." }, { "left": "issue:45245", "right": "issue:45584", "accept": false, - "reason": "Unrelated errors: category-cardinality runtime limit vs Whisper empty transcription after token alignment." + "reason": "Unrelated errors: category-cardinality limit vs Whisper empty transcription handling." }, { "left": "issue:41950", "right": "issue:44016", "accept": false, - "reason": "Different problem classes: pipeline selecting image processors vs a notebook syntax error." + "reason": "A video-classification pipeline bug and a notebook syntax error are not the same underlying issue." }, { "left": "issue:39692", "right": "issue:44190", "accept": false, - "reason": "Docs/example mistakes are not the same as a local-dataset loading bug." - }, - { - "left": "issue:44734", - "right": "issue:45632", - "accept": false, - "reason": "Different areas: server KV-cache indexing vs trust_remote_code cache-path collision." + "reason": "Documentation/example errors versus local dataset loading failure; only broad area overlaps." }, { "left": "issue:36246", "right": "issue:41628", "accept": false, - "reason": "Both are import errors, but they involve different symbols, modules, and likely different missing exports." + "reason": "Both are import errors, but they concern different symbols and different API surfaces; not the same bug." + }, + { + "left": "issue:44734", + "right": "issue:45632", + "accept": false, + "reason": "Serve KV-cache continuation crash and trust-remote-code cache-path collision are unrelated." }, { "left": "issue:43479", "right": "issue:45375", "accept": false, - "reason": "Different config bugs in different multimodal models; no shared underlying defect." + "reason": "Different configuration bugs in different model families; one is default initialization, the other is a missing strict-serialization field." }, { "left": "issue:45406", "right": "issue:45632", "accept": false, - "reason": "Serve processor attribute crash vs remote-code cache collision are unrelated." + "reason": "Gemma4Processor missing `_tokenizer` in serve is unrelated to trust_remote_code cache collisions." }, { - "left": "issue:44561", - "right": "issue:45632", + "left": "issue:36246", + "right": "issue:38617", "accept": false, - "reason": "Both mention trust_remote_code, but one is an API removal regression and the other is a cache-path bug." + "reason": "Different import failures affecting different names/modules; no shared root cause is evident." }, { - "left": "issue:36246", - "right": "issue:38617", + "left": "issue:44561", + "right": "issue:45632", "accept": false, - "reason": "Both are import failures, but for unrelated names/modules and different root causes." + "reason": "One is a removed-API regression breaking trust_remote_code models; the other is a local cache-path collision bug." }, { "left": "issue:44514", "right": "issue:44734", "accept": false, - "reason": "Different Qwen serve/processor crashes: batched chat-template padding vs KV-cache continuation indexing." + "reason": "Different Qwen2.5-VL chat-template batching issue vs serve KV-cache continuation indexing bug." }, { - "left": "issue:38175", - "right": "issue:43065", + "left": "issue:44016", + "right": "issue:44220", "accept": false, - "reason": "No overlap: SigLIP2 probability issue vs Sam3PixelDecoder dummy layer behavior." + "reason": "Notebook syntax error and audio feature-extraction issue are unrelated." }, { - "left": "issue:44016", - "right": "issue:44220", + "left": "issue:38175", + "right": "issue:43065", "accept": false, - "reason": "Notebook syntax error and audio feature extraction are unrelated." + "reason": "Different model-specific behavior: SigLIP2 zero probabilities vs Sam3PixelDecoder dummy Conv2d." }, { "left": "issue:42175", "right": "issue:44279", "accept": false, - "reason": "Different dependency problems: missing tensorflow in extras vs a generic dependency issue." + "reason": "Backend/dependency installation issue vs generic dependency issue; no concrete shared bug." }, { "left": "issue:44683", "right": "issue:45446", "accept": false, - "reason": "Related file area, but not the same bug: torch>=2.9 flex_attention failure vs wrong version gating for AuxRequest import." + "reason": "Both touch flex_attention/PyTorch versioning, but the failures are distinct and not mergeable as one fix." }, { "left": "issue:44466", "right": "issue:44913", "accept": false, - "reason": "Both involve serialization/reload, but one is tied-weight saving and the other is config field persistence." + "reason": "Different state/serialization bugs: tied lm_head weight serialization vs GPTNeoX config reload behavior." }, { "left": "issue:44246", "right": "issue:45003", "accept": false, - "reason": "Import slowness and unsafe sys.modules access are different issues, even if both concern import-time behavior." + "reason": "Import performance complaint and unsafe `sys.modules[]` access are not the same underlying issue." } ] }, @@ -9030,14 +8975,14 @@ "issue:43299|issue:45216", "issue:42907|issue:43441", "issue:43643|issue:44496", - "issue:43295|issue:45335", "issue:44016|issue:44062", + "issue:43295|issue:45335", "issue:44964|issue:45600", - "issue:44112|issue:44964", "issue:43976|issue:44987", + "issue:43673|issue:44336", + "issue:44112|issue:44964", "issue:45561|issue:45632", "issue:43531|issue:45600", - "issue:43673|issue:44336", "issue:45335|issue:45584", "issue:43824|issue:44933", "issue:43976|issue:44279", @@ -9050,222 +8995,216 @@ "issue:44521|issue:45325", "issue:44242|issue:45200", "issue:43381|issue:45593", + "issue:44351|issue:45278", "issue:43927|issue:45632", "issue:43901|issue:44496", - "issue:44351|issue:45278", - "issue:44683|issue:45542", "issue:43976|issue:45405", + "issue:44683|issue:45542", "issue:43479|issue:43525", "issue:44462|issue:45081" ], - "prepared_review_unit_hash": "808e95c27082892363fcb40a0e77dc4fb33d2e8ef03280ff6d73e178e7bc7ee2", + "prepared_review_unit_hash": "018d0195f305cc6745ec7b4e6ef6e3dbc8633d39a3e09947bdb40dcb132ce65c", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36768, - "estimated_input_tokens": 9192, - "estimated_eval_tokens": 18640 + "serialized_chars": 36769, + "estimated_input_tokens": 9193, + "estimated_eval_tokens": 18642 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These pairs mostly share only broad subsystem keywords (loading, saving, configs, tokenizers, multimodal) and do not appear to be the same underlying bug/change. I would not merge any of the soft pairs. As a standalone representative issue, 44521 is the strongest high-signal report here, but it is not a duplicate of the others.", - "confidence": 0.82, - "canonical_issue_reason": "No clear duplicate cluster emerges from the provided issues; if a single issue must represent the set, 44521 is the strongest high-signal bug report (open, reproducible, multiple inbound refs).", + "summary": "This cluster is mostly heterogeneous; the soft links are driven by superficial similarity (same model family, same subsystem, or generic import/load regressions) rather than the same concrete bug. I would keep the cluster centered on the broad upgrade/import-regression report, but reject the proposed duplicate edges.", + "confidence": 0.8, + "canonical_issue_reason": "issue:45278 is the broadest and most representative report here: a user-facing breakage after upgrading to 5.5.0 that can plausibly absorb several related import-regression reports.", "canonical_pr_reason": null, - "best_issue_reason": "44521 is the most actionable and broadly impactful issue in the set, with strong discussion/inbound attention and a concrete multimodal regression.", + "best_issue_reason": "It is the best global issue representative because it is broad, central to the cluster\u2019s common theme (5.x upgrade/import failures), and more likely than the others to serve as an umbrella canonical issue.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43299", "right": "issue:45216", "accept": false, - "reason": "Different model families and different failure modes: Qwen3VL MoE loading vs Qwen3.5 save_pretrained checkpoint key corruption." + "reason": "Both are Qwen3-family regressions, but one is a loading failure and the other is a save_pretrained checkpoint regression; different code paths and different symptoms." }, { "left": "issue:42907", "right": "issue:43441", "accept": false, - "reason": "Both mention Ministral, but one is about saving dequantized models and the other about FlashAttention loading." + "reason": "Dequantized save failure vs FlashAttention load failure for Ministral-3 are distinct problems." }, { "left": "issue:43643", "right": "issue:44496", "accept": false, - "reason": "Missing fields from trust_remote_code config loading is unrelated to unrecognized model_type in config.json." + "reason": "Both involve config/model loading, but one is missing fields in AutoConfig with trust_remote_code and the other is an unrecognized model due to missing model_type; not the same bug." }, { - "left": "issue:43295", - "right": "issue:45335", + "left": "issue:44016", + "right": "issue:44062", "accept": false, - "reason": "Processor/tokenizer image handling regression is unrelated to t5gemma embedding resize not reaching decoder.embed_tokens." + "reason": "A notebook syntax error and an AddedToken argument conflict are unrelated." }, { - "left": "issue:44016", - "right": "issue:44062", + "left": "issue:43295", + "right": "issue:45335", "accept": false, - "reason": "Notebook syntax error and AddedToken special-arg TypeError are different bugs in different parts of the stack." + "reason": "processor.tokenizer/image passing regression and t5gemma embedding resize behavior are separate issues." }, { "left": "issue:44964", "right": "issue:45600", "accept": false, - "reason": "Phi-4 multimodal loading and auto_mappings referencing removed Sam3LiteText configs are distinct issues." + "reason": "Phi-4 multimodal loading and stale auto_mappings references to Sam3LiteText configs are different failures." }, { - "left": "issue:44112", - "right": "issue:44964", + "left": "issue:43976", + "right": "issue:44987", "accept": false, - "reason": "CI stale device-override test failure is unrelated to a model loading regression." + "reason": "Python version compatibility and loading physical-intelligence/fast are unrelated." }, { - "left": "issue:43976", - "right": "issue:44987", + "left": "issue:43673", + "right": "issue:44336", + "accept": false, + "reason": "GenerationMixin cache missing during chunked_prefill and ANSI codes in loading_report are different subsystems and bugs." + }, + { + "left": "issue:44112", + "right": "issue:44964", "accept": false, - "reason": "Python version compatibility bug and physical-intelligence/fast loading failure are not the same issue." + "reason": "CI test flakiness in GraniteSpeech is unrelated to Phi-4 multimodal model loading." }, { "left": "issue:45561", "right": "issue:45632", "accept": false, - "reason": "Both involve filesystem behavior, but xdist captured_info.txt races and trust_remote_code cache path collisions are different problems." + "reason": "A pytest-xdist file race and a trust_remote_code cache-path collision are different concurrency/path issues." }, { "left": "issue:43531", "right": "issue:45600", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and removed-config references in auto_mappings are unrelated." - }, - { - "left": "issue:43673", - "right": "issue:44336", - "accept": false, - "reason": "GenerationMixin cache missing in chunked_prefill is unrelated to ANSI codes in loading_report." + "reason": "Qwen3-MoE sliding_window behavior and removed Sam3LiteText references in auto_mappings are unrelated." }, { "left": "issue:45335", "right": "issue:45584", "accept": false, - "reason": "t5gemma embedding resize bug and Whisper empty-transcription generation failure are different code paths." + "reason": "Embedding resize propagation and Whisper empty-transcription handling are distinct bugs." }, { "left": "issue:43824", "right": "issue:44933", "accept": false, - "reason": "Both are import errors, but for different symbols and causes; not the same underlying defect." + "reason": "Both are import-related, but one is a missing Qwen2.5-VL export and the other is a non-existent image_utils import; not the same symbol or code path." }, { "left": "issue:43976", "right": "issue:44279", "accept": false, - "reason": "Python support mismatch and a generic dependency issue are not duplicates." + "reason": "Python support/version mismatch and a generic dependency issue are not the same problem." }, { "left": "issue:44496", "right": "issue:45042", "accept": false, - "reason": "Model config recognition failure is unrelated to PIL backend unexpectedly requiring torchvision." + "reason": "Missing model_type config loading and PIL image processors requiring torchvision are unrelated." }, { "left": "issue:43854", "right": "issue:44964", "accept": false, - "reason": "Different model targets and different load failures; both are load bugs, not the same one." + "reason": "Two different model-specific loading failures; same broad area, but not the same concrete bug." }, { "left": "issue:43577", "right": "issue:45216", "accept": false, - "reason": "Blip2 dtype propagation and Qwen3.5 save_pretrained are unrelated." + "reason": "BLIP2 dtype propagation and Qwen3.5 save_pretrained checkpoint correctness are unrelated." }, { "left": "issue:43299", "right": "issue:45357", "accept": false, - "reason": "Both are Qwen-related regressions, but one is MoE loading and the other is incorrect visual-encoder keys on save_pretrained." + "reason": "Qwen3VL loading and Qwen3.5 visual-encoder save_pretrained key layout are different regressions." }, { "left": "issue:43526", "right": "issue:43901", "accept": false, - "reason": "BeitImageProcessorFast label reduction bug is unrelated to pipeline docs mentioning return_all_scores." + "reason": "A BeitImageProcessorFast bug and a docs/behavior mismatch in TextClassificationPipeline are unrelated." }, { "left": "issue:44075", "right": "issue:44368", "accept": false, - "reason": "SGD arguments not used and tie_word_embeddings warnings concern different subsystems." + "reason": "Ignored SGD optimizer args and a Qwen3.5 tie_word_embeddings warning are not the same bug." }, { "left": "issue:44521", "right": "issue:45325", "accept": false, - "reason": "Multimodal assistant_masks all-zero and Qwen2.5-VL rope-index scaling are separate multimodal bugs." + "reason": "All-zero assistant masks in multimodal chat templates and Qwen2.5-VL rope-index scaling are separate multimodal issues." }, { "left": "issue:44242", "right": "issue:45200", "accept": false, - "reason": "Load-balancing loss gating and mm_token_type_ids defaults are different model-training issues." + "reason": "Missing load-balancing loss and Gemma 4 mm_token_type_ids defaults are different model-training bugs." }, { "left": "issue:43381", "right": "issue:45593", "accept": false, - "reason": "Gradient checkpointing eval-mode behavior is unrelated to D-FINE auxiliary-loss handling." + "reason": "Gradient checkpointing in eval mode and D-FINE auxiliary-loss behavior are unrelated." }, { - "left": "issue:43927", - "right": "issue:45632", + "left": "issue:44351", + "right": "issue:45278", "accept": false, - "reason": "DiaConfig save/load token-id loss and trust_remote_code cache-path collisions are different save/load problems." + "reason": "44351 is one specific missing import (HybridCache); 45278 is a broad complaint about many import errors after upgrade, not the same concrete failure." }, { "left": "issue:43901", "right": "issue:44496", "accept": false, - "reason": "Documentation mismatch for return_all_scores is unrelated to model-type resolution failure." + "reason": "A pipeline docs mismatch and an unrecognized model/config loading error are unrelated." }, { - "left": "issue:44351", - "right": "issue:45278", + "left": "issue:43976", + "right": "issue:45405", "accept": false, - "reason": "A specific HybridCache import error and a broad import-error report after upgrade are not the same concrete bug." + "reason": "Python compatibility and a PEFT minimum-version bump are different dependency/version issues." }, { "left": "issue:44683", "right": "issue:45542", "accept": false, - "reason": "Compiled flex_attention on torch>=2.9 is unrelated to tensorboard-without-TensorFlow backend handling." - }, - { - "left": "issue:43976", - "right": "issue:45405", - "accept": false, - "reason": "Python compatibility and the unreleased MIN_PEFT_VERSION bump are separate release issues." + "reason": "Compiled flex_attention on torch>=2.9 and a TensorFlow backend error from tensorboard installation are unrelated." }, { "left": "issue:43479", "right": "issue:43525", "accept": false, - "reason": "Phi4Multimodal default config initialization and Llama4Config missing pad_token_id are different config bugs." + "reason": "Different config attributes on different models; both are config bugs, but not the same one." }, { "left": "issue:44462", "right": "issue:45081", "accept": false, - "reason": "Tokenizer JSON selection and Mistral regex patching fail in different tokenizer-loading code paths." + "reason": "AutoTokenizer ignoring tokenizer.json and _patch_mistral_regex crashing on backend_tokenizer access are distinct tokenizer-loading bugs." } ] }, @@ -9284,8 +9223,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -9344,8 +9283,8 @@ "issue:41950|issue:43650", "issue:44496|issue:44877", "issue:43299|issue:43854", - "issue:38175|issue:44617", "issue:44462|issue:44779", + "issue:38175|issue:44617", "issue:43295|issue:44779", "issue:39692|issue:43901", "issue:43650|issue:45290", @@ -9354,8 +9293,8 @@ "issue:44016|issue:44291", "issue:43262|issue:44521", "issue:42994|issue:43317", - "issue:43441|issue:44683", "issue:44261|issue:44263", + "issue:43441|issue:44683", "issue:43065|issue:44898", "issue:44466|issue:45276", "issue:44623|issue:45397", @@ -9367,202 +9306,202 @@ "issue:44987|issue:45542", "issue:43653|issue:45593" ], - "prepared_review_unit_hash": "76eccc8f15cb146bf651200c5a2cf40472cf6d8770d8f4341cb7b747e8a4a8cb", + "prepared_review_unit_hash": "3445cd7f5940f976dde3c5008537684e22c9204532e3ff04c3c50c2944243ef3", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 35388, - "estimated_input_tokens": 8847, - "estimated_eval_tokens": 17950 + "serialized_chars": 35392, + "estimated_input_tokens": 8848, + "estimated_eval_tokens": 17952 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab bag of distinct Transformers regressions and docs issues, not a single duplicate set. A few items are thematically adjacent (tokenizer/loading, processor/chat-template, v5 regressions), but the concrete failure modes differ enough that none of the soft edges should be merged.", - "confidence": 0.93, - "canonical_issue_reason": "No single canonical issue fits this cluster: it mixes unrelated bugs across tokenizers, processors, model loading, attention, quantization, docs, and model-specific failures.", + "summary": "This set is heterogeneous: most items are distinct regressions around tokenizers, processors, model loading, quantization, or docs, not one duplicate cluster.", + "confidence": 0.34, + "canonical_issue_reason": "No single true duplicate stands out from the titles alone. If forced to pick one representative issue, 45020 is the broadest umbrella report about recent transformers versions breaking remote_code models.", "canonical_pr_reason": null, - "best_issue_reason": "If forced to pick a representative issue, 45020 is the broadest v5/remote_code regression umbrella, but it is still too generic to serve as a true duplicate target for the rest.", + "best_issue_reason": "45020 is the most general, user-facing regression report in the set and best represents the cluster\u2019s overall theme better than the narrower subsystem-specific issues.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41950", "right": "issue:45020", "accept": false, - "reason": "Different bugs: video-classification pipeline processor lookup vs broad remote_code model-loading breakage." + "reason": "Different bugs: video-classification processor selection vs. remote_code model-loading regressions." }, { "left": "issue:44297", "right": "issue:44462", "accept": false, - "reason": "Both tokenizer-related, but one is save_pretrained metadata/class mismatch and the other is AutoTokenizer ignoring tokenizer.json; different code paths." + "reason": "Both concern tokenizers, but one is save_pretrained metadata and the other is repo loading behavior; not the same concrete bug." }, { "left": "issue:44279", "right": "issue:45405", "accept": false, - "reason": "Unrelated: generic dependency problem vs a PEFT minimum-version bump issue." + "reason": "Unrelated: a generic dependency issue vs. a version pin bump to an unreleased PEFT release." }, { "left": "issue:41950", "right": "issue:43650", "accept": false, - "reason": "No overlap beyond being noisy issues; video pipeline processor lookup is unrelated to the placeholder 'ADD THE DATA' report." + "reason": "Completely unrelated: video pipeline processor lookup vs. placeholder issue text." }, { "left": "issue:44496", "right": "issue:44877", "accept": false, - "reason": "Different subsystems: strict config loading for granite_speech vs Perceiver resize/interpolate behavior." + "reason": "Both are config-loading failures, but for different models and different strictness checks; not one bug." }, { "left": "issue:43299", "right": "issue:43854", "accept": false, - "reason": "Both involve model loading, but for different model families and likely different failure causes." + "reason": "Different model-loading failures for different architectures; no evidence of a shared code-path bug." }, { - "left": "issue:38175", - "right": "issue:44617", + "left": "issue:44462", + "right": "issue:44779", "accept": false, - "reason": "SigLIP2 zero-probabilities inference bug is unrelated to Sam3Video CUDA OOM." + "reason": "Tokenizer loading from repo vs. incorrect Deepseek tokenization are different regressions." }, { - "left": "issue:44462", - "right": "issue:44779", + "left": "issue:38175", + "right": "issue:44617", "accept": false, - "reason": "Tokenizer loading from repository vs DeepSeek tokenization correctness are distinct tokenizer bugs." + "reason": "Unrelated: SigLIP2 probability output vs. Sam3Video CUDA memory usage." }, { "left": "issue:43295", "right": "issue:44779", "accept": false, - "reason": "Processor.tokenizer regression with images passed to tokenizer is different from DeepSeek tokenizer output drift." + "reason": "Custom processor.tokenizer access regression vs. Deepseek tokenizer output regression; distinct failures." }, { "left": "issue:39692", "right": "issue:43901", "accept": false, - "reason": "Both docs-related, but different docs and different errors: SigLIP2 example failures vs return_all_scores documentation drift." + "reason": "Both are docs-related, but one is a model/processor mismatch and quantization example issue; the other is stale pipeline docs." }, { "left": "issue:43650", "right": "issue:45290", "accept": false, - "reason": "Placeholder issue is not the same as the specific apply_chat_template tool-call crash." + "reason": "Placeholder issue text vs. a specific apply_chat_template crash on tool-call assistant messages." }, { "left": "issue:43723", "right": "issue:44162", "accept": false, - "reason": "AutoTokenizer.from_pretrained regression is not the same as ESM2 model breakage." + "reason": "Tokenizer loading regression vs. an ESM2 model breakage; unrelated." }, { "left": "issue:43450", "right": "issue:44279", "accept": false, - "reason": "Video processor batched-shape bug is unrelated to a dependency issue." + "reason": "Video processor shape bug vs. dependency/version issue; not the same underlying change." }, { "left": "issue:44016", "right": "issue:44291", "accept": false, - "reason": "Notebook syntax error vs init_empty_weights unexpected argument TypeError are unrelated." + "reason": "Notebook syntax error vs. init_empty_weights loading TypeError; unrelated." }, { "left": "issue:43262", "right": "issue:44521", "accept": false, - "reason": "Both involve apply_chat_template, but one is audio sampling-rate defaulting and the other is multimodal assistant mask generation." + "reason": "Both mention apply_chat_template, but one is audio sampling-rate defaulting and the other is multimodal assistant mask generation." }, { "left": "issue:42994", "right": "issue:43317", "accept": false, - "reason": "Quantized model save failure and dequantized offload/load failure are different stages and failure modes." + "reason": "Different quantization-related failures: saving quantized models vs. loading dequantized models with device_map=auto." }, { - "left": "issue:43441", - "right": "issue:44683", + "left": "issue:44261", + "right": "issue:44263", "accept": false, - "reason": "FlashAttention failure on a specific model is not the same as compiled flex_attention failing on newer torch." + "reason": "Precision/config issue in MLA layernorm vs. a torch.split return-value issue in a different model component." }, { - "left": "issue:44261", - "right": "issue:44263", + "left": "issue:43441", + "right": "issue:44683", "accept": false, - "reason": "Layernorm epsilon precision discussion is unrelated to torch.split return values in GlmMoeDsaIndexer." + "reason": "Different attention-backend problems: FlashAttention failure for a model vs. compiled flex_attention failing on newer torch." }, { "left": "issue:43065", "right": "issue:44898", "accept": false, - "reason": "Dummy Conv2d in Sam3PixelDecoder is unrelated to Perceiver image-classification resolution handling." + "reason": "Unrelated model/component bugs: dummy Conv2d in Sam3PixelDecoder vs. Perceiver image classification resize handling." }, { "left": "issue:44466", "right": "issue:45276", "accept": false, - "reason": "Both touch serialization/weights, but one is device-dependent lm_head serialization and the other is resize_token_embeddings not propagating to tied embeddings." + "reason": "Serialization of lm_head.weight vs. resize_token_embeddings not propagating; related area but different concrete bugs." }, { "left": "issue:44623", "right": "issue:45397", "accept": false, - "reason": "processor.save_pretrained missing files is a different bug from gemma-4 zero3 from_pretrained loading." + "reason": "processor.save_pretrained missing files vs. gemma-4 zero3 from_pretrained; different failure modes." }, { "left": "issue:43122", "right": "issue:43650", "accept": false, - "reason": "Tokenizer version drift is unrelated to the placeholder issue." + "reason": "Tokenizer behavior regression vs. placeholder issue text; no overlap." }, { "left": "issue:44623", "right": "issue:44749", "accept": false, - "reason": "Missing processor files on save is unrelated to the reported slowdown after upgrading Transformers." + "reason": "Processor saving bug vs. performance regression during data filtering; unrelated." }, { "left": "issue:43784", "right": "issue:45542", "accept": false, - "reason": "Import-time NameError in sentence-transformers is unrelated to the TensorBoard/TensorFlow backend installation issue." + "reason": "Import-time NameError from transformers integration vs. tensorboard/TensorFlow backend detection issue." }, { "left": "issue:44671", "right": "issue:44779", "accept": false, - "reason": "CamemBERT masked-LM prediction regression is a different model-specific bug from DeepSeek tokenization." + "reason": "Different regressions: CamemBERT masked LM predictions vs. Deepseek tokenizer output." }, { "left": "issue:43673", "right": "issue:43761", "accept": false, - "reason": "Missing GenerationMixin cache in chunked_prefill is unrelated to CLIPVisionModel not returning hidden_states." + "reason": "Cache-missing generation regression vs. CLIP vision hidden_states regression; unrelated." }, { "left": "issue:44987", "right": "issue:45542", "accept": false, - "reason": "Model loading failure for physical-intelligence/fast is unrelated to the TensorBoard-only backend issue." + "reason": "Model loading under transformers>=5.1.0 vs. tensorboard without TensorFlow backend handling; different areas." }, { "left": "issue:43653", "right": "issue:45593", "accept": false, - "reason": "BigBirdTokenizer special-token registration bug is unrelated to D-FINE auxiliary-loss behavior." + "reason": "Tokenizer special-token registration bug vs. D-FINE auxiliary-loss behavior; unrelated." } ] }, @@ -9581,8 +9520,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -9640,47 +9579,47 @@ "issue:43901|issue:43994", "issue:44297|issue:45397", "issue:44361|issue:45593", - "issue:45561|issue:45563", "issue:43475|issue:44016", + "issue:45561|issue:45563", "issue:43824|issue:43976", "issue:43295|issue:44617", + "issue:43525|issue:43881", "issue:43352|issue:45250", "issue:44589|issue:45362", - "issue:43525|issue:43881", - "issue:41628|issue:43976", "issue:44188|issue:44336", - "issue:43901|issue:44246", + "issue:41628|issue:43976", "issue:44295|issue:45397", + "issue:43901|issue:44246", "issue:42491|issue:43299", "issue:41950|issue:44220", - "issue:44242|issue:45593", "issue:44336|issue:44373", + "issue:44242|issue:45593", "issue:44462|issue:45356", "issue:44661|issue:45542", - "issue:45003|issue:45561", "issue:44246|issue:44496", + "issue:45003|issue:45561", "issue:36246|issue:42757", - "issue:43531|issue:45362", "issue:36246|issue:44351", + "issue:43531|issue:45362", "issue:43526|issue:45003", - "issue:43673|issue:44561", "issue:44336|issue:44485", + "issue:43673|issue:44561", "issue:44162|issue:45200" ], - "prepared_review_unit_hash": "03d63de90f6ab2089f76b8d997e24cddb9570b00ffe1aff808534e853ad4a001", + "prepared_review_unit_hash": "927a8378370efb0e1783a633764ddf08e8dce0b6a876aef02ed14a6fc2ddf28f", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36335, + "serialized_chars": 36334, "estimated_input_tokens": 9084, "estimated_eval_tokens": 18424 }, @@ -9689,192 +9628,186 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a loose mix of Transformers issues around import failures, model-loading regressions, tokenizer/pipeline bugs, and unrelated docs/test/model-specific problems. There are no PRs here, and none of the soft pairs look like true duplicates of the same underlying bug.", - "confidence": 0.64, - "canonical_issue_reason": "issue:43824 is the strongest anchor: it has the most inbound references and discussion, and it represents the broad import/load-breakage theme shared by several nearby reports.", + "summary": "This cluster is not a true duplicate set; it mixes many unrelated Transformers issues (imports, model loading, generation, tokenizer behavior, docs, tests, and backend/runtime problems). I would not merge any of the soft pairs.", + "confidence": 0.91, + "canonical_issue_reason": "issue:43824 is the strongest representative issue here: it has the highest discussion/inbound signal and a concrete, user-facing import regression. That said, the overall set is heterogeneous rather than one coherent bug family.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43824 is the best single representative for the cluster because it is the most referenced/discussed and sits closest to the recurring import/loading regressions, even though the overall cluster is not tight.", + "best_issue_reason": "issue:43824 is the best single issue to anchor the cluster because it is specific, high-impact, and clearly described. It is still only a representative, not a duplicate target for the rest.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:42175", "right": "issue:43976", "accept": false, - "reason": "TensorFlow backend packaging under `.[torch]` is unrelated to the Python 3.9+/5.1.0 compatibility problem." + "reason": "Packaging/backend dependency complaint vs a Python-version compatibility bug; different failure modes and fixes." }, { "left": "issue:43901", "right": "issue:43994", "accept": false, - "reason": "Docs about `return_all_scores` do not match the SigLIP2 bad-output model behavior." + "reason": "Docs mismatch in TextClassificationPipeline vs a model inference correctness issue; not the same underlying bug." }, { "left": "issue:44297", "right": "issue:45397", "accept": false, - "reason": "Tokenizer class mismatch on save is a different failure mode from Gemma-4 `from_pretrained` breakage." + "reason": "Tokenizer metadata saved wrong vs a zero3 from_pretrained loading failure; unrelated code paths." }, { "left": "issue:44361", "right": "issue:45593", "accept": false, - "reason": "MLukeTokenizer task AttributeError is unrelated to D-FINE auxiliary-loss behavior." + "reason": "Tokenizer/task AttributeError vs missing auxiliary losses in D-FINE; different models and mechanisms." }, { - "left": "issue:45561", - "right": "issue:45563", + "left": "issue:43475", + "right": "issue:44016", "accept": false, - "reason": "A pytest-xdist file race is not the same bug as the stale `num_return_sequences` warning in paged generate." + "reason": "SAM 3 video model attribute bug vs a notebook syntax error; clearly unrelated." }, { - "left": "issue:43475", - "right": "issue:44016", + "left": "issue:45561", + "right": "issue:45563", "accept": false, - "reason": "SAM 3 Video missing attribute and a notebook syntax error are entirely different problems." + "reason": "pytest-xdist file race vs stale generation warning; testing infrastructure vs generation logic." }, { "left": "issue:43824", "right": "issue:43976", "accept": false, - "reason": "Qwen2_5_VL import failure is unrelated to Transformers 5.1.0/Python version incompatibility." + "reason": "Import/export regression for a Qwen2.5 VL symbol vs a Python support/version issue; not the same bug." }, { "left": "issue:43295", "right": "issue:44617", "accept": false, - "reason": "Processor/tokenizer regression is a different code path from the SAM3 Video OOM report." + "reason": "Processor/tokenizer API regression vs CUDA OOM in SAM3Video; different symptoms and fixes." + }, + { + "left": "issue:43525", + "right": "issue:43881", + "accept": false, + "reason": "Missing config attribute in Llama4 vs a separate model-loading failure for GLM-4V; not one code path." }, { "left": "issue:43352", "right": "issue:45250", "accept": false, - "reason": "Unsupported Flash Attention 2.0 for Nemotron is not the same as a generic Flash Attention 2.0 issue." + "reason": "Model does not support Flash Attention 2.0 vs a broad Flash Attention topic issue; not the same concrete problem." }, { "left": "issue:44589", "right": "issue:45362", "accept": false, - "reason": "Float8 storage lookup failure and Qwen3.5 chat crash are different model/runtime issues." + "reason": "Storage serialization/type error vs a Qwen3.5 chat crash; unrelated runtime failures." }, { - "left": "issue:43525", - "right": "issue:43881", + "left": "issue:44188", + "right": "issue:44336", "accept": false, - "reason": "Missing `pad_token_id` in Llama4Config is unrelated to glm-4v-9b loading failure." + "reason": "torch.compile attention-kernel divergence vs ANSI code output formatting; no overlap in underlying bug." }, { "left": "issue:41628", "right": "issue:43976", "accept": false, - "reason": "`AutoImageProcessor` import failure is not the same as the Transformers 5.1.0/Python compatibility bug." + "reason": "AutoImageProcessor import failure vs Python-version incompatibility; different root causes." }, { - "left": "issue:44188", - "right": "issue:44336", + "left": "issue:44295", + "right": "issue:45397", "accept": false, - "reason": "torch.compile attention divergence and ANSI codes in loading_report are unrelated." + "reason": "Position_ids buffer access bug vs distributed loading issue; unrelated behavior." }, { "left": "issue:43901", "right": "issue:44246", "accept": false, - "reason": "A docs mismatch is not the same underlying issue as slow `import transformers`." - }, - { - "left": "issue:44295", - "right": "issue:45397", - "accept": false, - "reason": "Registered-buffer `position_ids` access differs from the Gemma-4 zero3 loading regression." + "reason": "Pipeline documentation mismatch vs slow import performance; not the same issue." }, { "left": "issue:42491", "right": "issue:43299", "accept": false, - "reason": "Both involve Qwen3 MoE, but one is a LoRA hf4.x\u2192hf5.x incompatibility and the other is a model-loading regression on v5 dev, so they are not the same bug." + "reason": "Both involve Qwen MoE/version regressions, but they target different model families and failure points, so they are not safely mergeable." }, { "left": "issue:41950", "right": "issue:44220", "accept": false, - "reason": "Video-classification pipeline image-processor lookup and `_torch_extract_fbank_features()` are different pipelines." + "reason": "Video pipeline selecting image processors vs audio fbank extraction; different subsystems." }, { - "left": "issue:44242", - "right": "issue:45593", + "left": "issue:44336", + "right": "issue:44373", "accept": false, - "reason": "MoE load-balancing loss and D-FINE auxiliary losses are unrelated training objectives." + "reason": "Terminal ANSI formatting bug vs wrong docstring; docs and runtime output are separate concerns." }, { - "left": "issue:44336", - "right": "issue:44373", + "left": "issue:44242", + "right": "issue:45593", "accept": false, - "reason": "ANSI output in loading_report and a wrong docstring for `position_ids` are not duplicates." + "reason": "Missing load-balancing loss in MoE routing vs missing auxiliary losses in D-FINE; both are loss issues but not the same code path." }, { "left": "issue:44462", "right": "issue:45356", "accept": false, - "reason": "AutoTokenizer ignoring `tokenizer.json` is different from the Kimi-K2.5 codec/regression report." + "reason": "AutoTokenizer ignoring tokenizer.json vs tokenizer codec/regex regression; different tokenizer loading bugs." }, { "left": "issue:44661", "right": "issue:45542", "accept": false, - "reason": "TOKENIZER_MAPPING_NAMES model-addition failure is unrelated to the missing TensorFlow backend when only tensorboard is installed." - }, - { - "left": "issue:45003", - "right": "issue:45561", - "accept": false, - "reason": "Unsafe `sys.modules` access and the xdist captured-info race are different failure mechanisms." + "reason": "Model-registration template failure vs missing TensorFlow backend dependency; unrelated setup problems." }, { "left": "issue:44246", "right": "issue:44496", "accept": false, - "reason": "Slow import timing is not the same as an unrecognized-model config error." + "reason": "Slow import vs unrecognized model/config loading; different failure modes." }, { "left": "issue:36246", "right": "issue:42757", "accept": false, - "reason": "Both are import errors, but they concern different modules and different missing symbols." + "reason": "Import error in transformers Qwen2.5 VL export vs missing huggingface_hub symbol; same symptom class, different package and symbol." }, { - "left": "issue:43531", - "right": "issue:45362", + "left": "issue:36246", + "right": "issue:44351", "accept": false, - "reason": "Qwen3-MoE `sliding_window` behavior is a different issue from Qwen3.5 chat crashes." + "reason": "Different missing exports from transformers; superficial similarity only, not one underlying bug." }, { - "left": "issue:36246", - "right": "issue:44351", + "left": "issue:43531", + "right": "issue:45362", "accept": false, - "reason": "Qwen2_5_VL image-processor import failure is unrelated to missing `HybridCache` import." + "reason": "Qwen3-MoE sliding_window behavior vs a Qwen3.5 chat crash; not the same concrete issue." }, { "left": "issue:43526", "right": "issue:45003", "accept": false, - "reason": "BeitImageProcessorFast label reduction and unsafe `sys.modules` access are unrelated bugs." + "reason": "BEiT label reduction logic bug vs unsafe sys.modules access in modeling_utils; unrelated code paths." }, { - "left": "issue:43673", - "right": "issue:44561", + "left": "issue:44336", + "right": "issue:44485", "accept": false, - "reason": "Chunked-prefill cache behavior in v5 is a different code path from removal of `is_torch_fx_available` breaking trust-remote-code models." + "reason": "ANSI terminal output bug vs GLM-5 RoPE implementation discussion; no shared failure." }, { - "left": "issue:44336", - "right": "issue:44485", + "left": "issue:43673", + "right": "issue:44561", "accept": false, - "reason": "Loading-report ANSI codes are unrelated to GLM-5 RoPE implementation." + "reason": "Chunked-prefill cache regression vs removal of is_torch_fx_available affecting trust_remote_code; both are version regressions, but not the same fix." }, { "left": "issue:44162", "right": "issue:45200", "accept": false, - "reason": "ESM2 breakage and Gemma 4 mm token defaults are different model-specific regressions." + "reason": "ESM2 regression impacting workflows vs Gemma 4 mm_token_type_ids defaulting; different models and failures." } ] }, @@ -9893,8 +9826,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -9947,231 +9880,231 @@ "issue:45588" ], "soft_pairs": [ - "issue:44242|issue:44857", "issue:44295|issue:44749", + "issue:44242|issue:44857", "issue:44246|issue:44871", "issue:44877|issue:45375", "issue:36246|issue:43596", "issue:43066|issue:44779", "issue:43441|issue:43540", "issue:45292|issue:45584", - "issue:44779|issue:45292", "issue:44987|issue:45459", - "issue:42886|issue:45229", + "issue:44779|issue:45292", "issue:41628|issue:41950", - "issue:44995|issue:45200", + "issue:42886|issue:45229", "issue:43065|issue:43408", + "issue:44995|issue:45200", "issue:43901|issue:45003", "issue:43232|issue:44336", "issue:43257|issue:45137", "issue:44496|issue:45003", + "issue:43295|issue:45292", "issue:44995|issue:45588", "issue:44060|issue:44077", - "issue:43295|issue:45292", - "issue:43262|issue:44246", "issue:43653|issue:44162", + "issue:43262|issue:44246", "issue:43475|issue:43540", "issue:44368|issue:45563", - "issue:44368|issue:44977", "issue:44466|issue:45335", + "issue:44368|issue:44977", "issue:44933|issue:45459" ], - "prepared_review_unit_hash": "716a855b2940d0aa5ee0652ffaf51f7fdca8a503e51f40a0253cb8720ec94aa1", + "prepared_review_unit_hash": "753cb1695756858fa3cc9a12133a8900da23d25a7a536affddd195d520694a54", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 35650, - "estimated_input_tokens": 8913, - "estimated_eval_tokens": 18082 + "serialized_chars": 35635, + "estimated_input_tokens": 8909, + "estimated_eval_tokens": 18074 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is broad and mostly heterogeneous: import/load regressions, tokenizer/config bugs, generation issues, and model-specific SAM/Qwen/DeepSpeed problems. I did not find any pair that clearly describes the same underlying bug/change closely enough to merge as duplicates.", - "confidence": 0.91, - "canonical_issue_reason": "If forced to pick one, issue 45003 is the broadest infrastructure-level bug in the set, but the cluster is not a tight duplicate group.", + "summary": "This cluster is mostly heterogeneous: the soft pairs share surface vocabulary (tokenizers, embeddings, generation, import/load paths) but usually describe different concrete bugs. I would not merge any of the suggested soft edges.", + "confidence": 0.89, + "canonical_issue_reason": "If a single issue must anchor the set, issue 45292 is the best fit because it is a concrete, reproducible API regression with broader applicability and external references; however, the overall cluster is not a true duplicate group.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45003 is the best global representative because it is a general core-library bug rather than a model-specific symptom, though it still does not unify the cluster.", + "best_issue_reason": "Issue 45292 is the strongest general-purpose bug report among the items and the most plausible long-term canonical reference for the resize_token_embeddings-related threads.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44242", - "right": "issue:44857", + "left": "issue:44295", + "right": "issue:44749", "accept": false, - "reason": "Different failures in different subsystems: MoE load-balancing loss vs LW-DETR AMP/CUDA crash." + "reason": "Both mention position/processing behavior, but one is a buffer read error and the other is a tokenizer performance regression; different code paths and symptoms." }, { - "left": "issue:44295", - "right": "issue:44749", + "left": "issue:44242", + "right": "issue:44857", "accept": false, - "reason": "Position_ids buffer access and post-upgrade data slowdown are unrelated." + "reason": "Load-balancing loss gating in MOE is unrelated to a CUDA float16 crash in a detection loss; no shared underlying bug." }, { "left": "issue:44246", "right": "issue:44871", "accept": false, - "reason": "Import latency bug vs Gemma eos_token_id config mismatch; no shared code path." + "reason": "Import latency and eos_token_id mismatch are unrelated issues despite both being configuration/runtime complaints." }, { "left": "issue:44877", "right": "issue:45375", "accept": false, - "reason": "Both involve config strictness, but they affect different model configs and different missing fields." + "reason": "Both involve config parsing/strictness, but they target different model families and different missing-field behavior." }, { "left": "issue:36246", "right": "issue:43596", "accept": false, - "reason": "ImportError for Qwen2.5 VL image processor vs DeepSpeed/BertModel index error are unrelated." + "reason": "One is a missing export/import error, the other is a DeepSpeed ZeRO3 index error in BertModel; no common fix path." }, { "left": "issue:43066", "right": "issue:44779", "accept": false, - "reason": "Tokenizer decoder-type warning in v5 vs DeepSeek tokenizer incorrect output; same broad area, different bug." + "reason": "Both are tokenizer-related, but one is a wrong decoder-type warning and the other is incorrect Deepseek tokenization output; not the same defect." }, { "left": "issue:43441", "right": "issue:43540", "accept": false, - "reason": "FlashAttention failure for Ministral-3 and video-processing ValueError for Qwen3OmniMoe are distinct." + "reason": "FlashAttention failure in Ministral and video-input validation in Qwen3OmniMoe are different model paths with different failure modes." }, { "left": "issue:45292", "right": "issue:45584", "accept": false, - "reason": "Embedding resize propagation bug vs Whisper empty-transcription generation failure are different code paths." - }, - { - "left": "issue:44779", - "right": "issue:45292", - "accept": false, - "reason": "DeepSeek tokenizer regression is not the same as resize_token_embeddings not updating output embeddings." + "reason": "resize_token_embeddings behavior and Whisper empty-transcription generation after align_special_tokens are separate regressions." }, { "left": "issue:44987", "right": "issue:45459", "accept": false, - "reason": "Model loading failure and protobuf-related tokenizer error masking are unrelated." + "reason": "Model-loading failure for a specific checkpoint is unrelated to protobuf exception handling in tokenizer decoding." }, { - "left": "issue:42886", - "right": "issue:45229", + "left": "issue:44779", + "right": "issue:45292", "accept": false, - "reason": "Offline cache tokenizer load bug vs multi-GPU CUDA OOM are not duplicates." + "reason": "Tokenizer incorrect results and embedding resize not propagating affect different APIs and likely different fixes." }, { "left": "issue:41628", "right": "issue:41950", "accept": false, - "reason": "General AutoImageProcessor import failure vs video-classification pipeline picking image processors; related area, not same bug." + "reason": "AutoImageProcessor import failure and video-classification pipeline misrouting are both image/video related but not the same bug." }, { - "left": "issue:44995", - "right": "issue:45200", + "left": "issue:42886", + "right": "issue:45229", "accept": false, - "reason": "Stale cache crash in GlmMoeDsa vs mm_token_type_ids defaulting issue are unrelated." + "reason": "Offline cache loading for tokenizers is unrelated to Gemma4 multi-GPU OOM." }, { "left": "issue:43065", "right": "issue:43408", "accept": false, - "reason": "SAM3 pixel decoder dummy layer and sam3_video/sam3_tracker type warning are different defects." + "reason": "A dummy Conv2d in SAM3 pixel decoder is not the same as a model-type mismatch warning for sam3_video vs sam3_tracker." + }, + { + "left": "issue:44995", + "right": "issue:45200", + "accept": false, + "reason": "A stale cache crash on second forward pass is unrelated to mm_token_type_ids defaulting for text-only fine-tuning." }, { "left": "issue:43901", "right": "issue:45003", "accept": false, - "reason": "Docs mismatch for return_all_scores is unrelated to sys.modules access in modeling_utils." + "reason": "Documentation drift for return_all_scores and unsafe sys.modules access are unrelated." }, { "left": "issue:43232", "right": "issue:44336", "accept": false, - "reason": "Generation kwargs/sync_gpus bug and ANSI color output in loading_report are unrelated." + "reason": "Generation kwargs after sync_gpus and stray ANSI output in loading_report do not share a code path." }, { "left": "issue:43257", "right": "issue:45137", "accept": false, - "reason": "Qwen3 MoE weight conversion problem and ZeRO3 deque pop error are different DeepSpeed failures." + "reason": "Qwen3 MOE weight conversion under accelerate/deepspeed and a ZeRO3 deque underflow are separate distributed-training bugs." }, { "left": "issue:44496", "right": "issue:45003", "accept": false, - "reason": "Unrecognized model config and unsafe sys.modules access are separate loading issues." + "reason": "Unrecognized model/config loading and unsafe sys.modules access are different model-loading failures." + }, + { + "left": "issue:43295", + "right": "issue:45292", + "accept": false, + "reason": "Processor.tokenizer regression is about multimodal processor API behavior, while 45292 is about token embedding resizing." }, { "left": "issue:44995", "right": "issue:45588", "accept": false, - "reason": "Different crashes: stale indexer cache vs flash_attention s_aux=None handling." + "reason": "A stale indexer cache in GLM MOE is unrelated to flash_attention.py crashing on s_aux=None." }, { "left": "issue:44060", "right": "issue:44077", "accept": false, - "reason": "Tied-weights warning bug and patchtsmixer post_init restriction are unrelated." + "reason": "Tied-weights warning behavior and patchtsmixer post_init policy are not the same underlying issue." }, { - "left": "issue:43295", - "right": "issue:45292", + "left": "issue:43653", + "right": "issue:44162", "accept": false, - "reason": "Processor.tokenizer/image handling regression is not the same as resize_token_embeddings not updating output embeddings." + "reason": "BigBirdTokenizer special-token registration and ESM2 breakage are different tokenizer/model regressions." }, { "left": "issue:43262", "right": "issue:44246", "accept": false, - "reason": "Audio chat-template sampling-rate default vs import slowness are unrelated." - }, - { - "left": "issue:43653", - "right": "issue:44162", - "accept": false, - "reason": "BigBirdTokenizer special-token registration bug and ESM2 regression are different model-specific issues." + "reason": "Audio chat-template sampling-rate defaulting and import latency are unrelated." }, { "left": "issue:43475", "right": "issue:43540", "accept": false, - "reason": "Sam3VisionEncoderOutput missing field and Qwen3OmniMoe video ValueError are distinct SAM/Qwen bugs." + "reason": "Missing fpn_position_embeddings in SAM3 video output is a different bug from Qwen3OmniMoe video input processing." }, { "left": "issue:44368", "right": "issue:45563", "accept": false, - "reason": "Word-embedding tie warning and stale num_return_sequences warning are different warning bugs." + "reason": "A tied-embeddings warning in Qwen3.5 and a stale warning for num_return_sequences are different warning-regression reports." }, { - "left": "issue:44368", - "right": "issue:44977", + "left": "issue:44466", + "right": "issue:45335", "accept": false, - "reason": "Qwen3.5 tie_word_embeddings warning and DeepSeek tokenizer regression are unrelated." + "reason": "Both concern embedding/tied-weight behavior, but one is serialization inconsistency and the other is resize_token_embeddings not updating decoder.embed_tokens; close but not the same bug." }, { - "left": "issue:44466", - "right": "issue:45335", + "left": "issue:44368", + "right": "issue:44977", "accept": false, - "reason": "Serialization inconsistency across devices and t5gemma resize propagation bug are separate weight-handling issues." + "reason": "A tied-weights warning and a FlashAttention generation failure in Qwen3.5 are separate regressions." }, { "left": "issue:44933", "right": "issue:45459", "accept": false, - "reason": "Missing import from image_utils and protobuf decode-error masking are unrelated." + "reason": "Missing import from image_utils and protobuf-related tokenizer exception masking are unrelated import/error-handling bugs." } ] }, @@ -10190,8 +10123,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -10216,9 +10149,9 @@ "issue:43931", "issue:44038", "issue:44162", - "issue:44188", "issue:44222", "issue:44297", + "issue:44351", "issue:44361", "issue:44368", "issue:44623", @@ -10261,8 +10194,8 @@ "issue:44038|issue:44623", "issue:42994|issue:43482", "issue:43701|issue:44749", - "issue:43475|issue:45381", "issue:42371|issue:45092", + "issue:43475|issue:45381", "issue:45375|issue:45478", "issue:43475|issue:43643", "issue:41628|issue:44933", @@ -10275,231 +10208,237 @@ "issue:42831|issue:44222", "issue:45310|issue:45397", "issue:45362|issue:45375", - "issue:43232|issue:44188" + "issue:41628|issue:44351" ], - "prepared_review_unit_hash": "9f3295fb9ab485166bb024ca16c141eee772daa4385d74d7c8c13cd9cf0f2c39", + "prepared_review_unit_hash": "a5a66186dc29a33786f6f247e33ebae4fd65af52400a556c82617e046374a327", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 36705, - "estimated_input_tokens": 9177, - "estimated_eval_tokens": 18610 + "serialized_chars": 36663, + "estimated_input_tokens": 9166, + "estimated_eval_tokens": 18588 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This packet is still mostly heterogeneous, with many unrelated model-specific regressions and API breakages. There are at least two clear duplicate-like issue pairs by title/symptom: the confirmed Qwen3ForCausalLM VRAM leak pair (42673/43704) and another identical-title pair (45310/45478) about the Qwen3.5 Moe from_pretrained error, even though only the former was surfaced in soft_pairs. The rest of the soft candidates remain conservative non-matches.", - "confidence": 0.94, - "canonical_issue_reason": "issue:42673 is the clearest canonical report for the Qwen3ForCausalLM VRAM leak: it is earlier, more detailed, and matches issue:43704 exactly on title and symptom.", + "summary": "This set is mostly a grab-bag of unrelated Transformers issues. Only one soft pair is a true duplicate: the Qwen3ForCausalLM VRAM leak reports. The rest are same-subsystem or same-version regressions at best, but not the same underlying bug.", + "confidence": 0.95, + "canonical_issue_reason": "issue:42673 is the better canonical because issue:43704 is an exact duplicate of the same Qwen3ForCausalLM VRAM leak in multiple dataloader threads, and 42673 has the fuller report and more discussion/references.", "canonical_pr_reason": null, - "best_issue_reason": "issue:42673 is the best representative issue because it is the earliest and fuller report in the only confirmed duplicate pair, and it anchors the strongest duplicate signal in the set.", + "best_issue_reason": "Given the cluster is heterogeneous, issue:42673 is the strongest representative only because it is backed by a true duplicate pair; none of the other issues are close enough to serve as a broader cluster canonical.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43673", "right": "issue:44734", "accept": false, - "reason": "Both are generation/cache-related, but they describe different failures in different code paths: chunked_prefill cache missing vs KV-cache continuation tensor indexing." + "reason": "Both mention generation/cache behavior, but one is chunked_prefill cache handling and the other is /v1/responses KV-cache continuation indexing; different code paths." }, { "left": "issue:44869", "right": "issue:45356", "accept": false, - "reason": "Both involve tokenizer/decoding regressions, but one is Whisper timestamp decoding and the other is Kimi tokenizer codec handling; not the same bug." + "reason": "Both involve tokenizer/codec edge cases, but they are different models and different failures; no shared underlying bug is evident." }, { "left": "issue:45137", "right": "issue:45600", "accept": false, - "reason": "DeepSpeed ZeRO3 deque error and Sam3LiteText auto-mapping CI breakage are unrelated." + "reason": "DeepSpeed ZeRO3 deque failure and auto_mappings CI breakage are unrelated bugs." }, { "left": "issue:43931", "right": "issue:44368", "accept": false, - "reason": "Qwen3-VL weight-shape mismatch and a tie_word_embeddings warning are different model/config issues." + "reason": "A weight-shape mismatch and a warning about tie_word_embeddings during LoRA fine-tuning are not the same issue." }, { "left": "issue:43232", "right": "issue:43673", "accept": false, - "reason": "These are separate generation bugs: sync_gpus kwargs handling versus missing GenerationMixin cache during chunked prefill." + "reason": "Both are generation-related, but one is sync_gpus/_update_model_kwargs_for_generation and the other is chunked_prefill cache missing; not the same defect." }, { "left": "issue:44297", "right": "issue:44623", "accept": false, - "reason": "Both are save_pretrained problems, but one is tokenizer metadata mismatch and the other is missing processor files; different components and failures." + "reason": "Tokenizer save_pretrained metadata mismatch versus processor.save_pretrained missing files are distinct save-path bugs." }, { "left": "issue:43673", "right": "issue:45278", "accept": false, - "reason": "The v5 import-error report is broad and unrelated to the specific generation cache bug in 43673." + "reason": "A specific generation cache bug is not the same as a broad set of import errors after upgrading." }, { "left": "issue:43646", "right": "issue:44162", "accept": false, - "reason": "Custom model initialization breakage and ESM2-specific breakage are too different to treat as the same underlying bug." + "reason": "Generic custom model initialization breakage and the ESM2-specific regression are not the same underlying problem." }, { "left": "issue:44361", "right": "issue:44749", "accept": false, - "reason": "MLukeTokenizer AttributeError and a large slowdown after upgrade are unrelated." + "reason": "MLukeTokenizer task AttributeError and a performance slowdown after upgrade are unrelated." }, { "left": "issue:42673", "right": "issue:43704", "accept": true, - "reason": "Same title, same model, same symptom: VRAM leak when Qwen3ForCausalLM is used in multiple dataloader threads." + "reason": "Exact same Qwen3ForCausalLM VRAM leak in multiple dataloader threads; 43704 is a duplicate of 42673." }, { "left": "issue:28282", "right": "issue:30990", "accept": false, - "reason": "PyTorch missing ImportError and Sentence Transformers hanging on load are not the same underlying issue." + "reason": "Missing PyTorch import error and a Sentence Transformers loading hang are different failure modes." }, { "left": "issue:44749", "right": "issue:45310", "accept": false, - "reason": "Performance regression in data filtering and Qwen3.5 MoE from_pretrained failure are different bugs." + "reason": "Performance regression after upgrading is unrelated to a Qwen3.5 Moe from_pretrained error." }, { "left": "issue:45230", "right": "issue:45397", "accept": false, - "reason": "Generic bug report and Gemma-4 Zero3 from_pretrained failure are not the same issue." + "reason": "Generic placeholder bug report and gemma-4 zero3 loading failure are not the same bug." }, { "left": "issue:44038", "right": "issue:44623", "accept": false, - "reason": "Qwen3-VL-Moe v5 bug and missing processor.save_pretrained files are unrelated." + "reason": "Qwen3-VL-Moe loading failure and processor.save_pretrained missing files touch different code paths." }, { "left": "issue:42994", "right": "issue:43482", "accept": false, - "reason": "Quantized model saving failure and Qwen2.5-GGUF loading failure are different operations and code paths." + "reason": "Quantized model saving failure is distinct from GGUF loading failure." }, { "left": "issue:43701", "right": "issue:44749", "accept": false, - "reason": "Checkpoint resume key mismatch is unrelated to the reported training slowdown." + "reason": "Checkpoint resume key mismatch and a throughput regression are unrelated." }, { - "left": "issue:43475", - "right": "issue:45381", + "left": "issue:42371", + "right": "issue:45092", "accept": false, - "reason": "SAM 3 Video missing attribute and qwen2.5-vl video position-id bug are different model-specific regressions." + "reason": "TF32 API-setting guidance and an old remote-code checkpoint/meta-init incompatibility are different issues." }, { - "left": "issue:42371", - "right": "issue:45092", + "left": "issue:43475", + "right": "issue:45381", "accept": false, - "reason": "TF32 API guidance and InternVL2 meta-initialization incompatibility are unrelated." + "reason": "SAM 3 Video missing attribute and qwen2.5-vl video position-id bug are different model-specific regressions." }, { "left": "issue:45375", "right": "issue:45478", "accept": false, - "reason": "Missing deepstack_visual_indexes field in config and Qwen3.5 Moe from_pretrained error are different problems." + "reason": "A missing Qwen3_5MoeVisionConfig field dropped by strict handling is not the same as a generic Qwen3.5 Moe from_pretrained error." }, { "left": "issue:43475", "right": "issue:43643", "accept": false, - "reason": "SAM 3 Video output attribute error and AutoConfig missing fields with trust_remote_code are not the same bug." + "reason": "SAM 3 Video attribute access failure and trust_remote_code missing-fields behavior are unrelated." }, { "left": "issue:41628", "right": "issue:44933", "accept": false, - "reason": "These are both import-related, but they concern different symbols and likely different import paths; not enough to unify." + "reason": "Missing AutoImageProcessor import and a different nonexistent import from image_utils are not clearly the same bug." }, { "left": "issue:45081", "right": "issue:45103", "accept": false, - "reason": "Tokenizer regex patch crash and auto_docstring future-annotations crash are unrelated." + "reason": "Tokenizer regex patch crash and a future-annotations docstring helper crash are separate AttributeError sites." }, { "left": "issue:45310", "right": "issue:45375", "accept": false, - "reason": "Both involve Qwen-related v5 regressions, but one is a from_pretrained error and the other is a config-field drop; different underlying failures." + "reason": "Qwen3.5 Moe from_pretrained error and a missing vision-config field are different failures, even if both involve Qwen configs." }, { "left": "issue:42994", "right": "issue:43525", "accept": false, - "reason": "Quantized model saving failure and Llama4Config missing pad_token_id are different issues." + "reason": "Quantized save failure and missing pad_token_id on Llama4Config are unrelated." }, { "left": "issue:42222", "right": "issue:45200", "accept": false, - "reason": "VitPose breakage and Gemma-4 mm_token_type_ids defaulting are unrelated model-specific bugs." + "reason": "VitPose model breakage and Gemma 4 text-only fine-tuning token-type defaults are different model issues." }, { "left": "issue:43873", "right": "issue:44945", "accept": false, - "reason": "Quantization offloading behavior and pipeline parallelism output correctness are different problems." + "reason": "Quantization/offloading behavior and pipeline-parallelism output corruption are different problems." }, { "left": "issue:43475", "right": "issue:44661", "accept": false, - "reason": "SAM 3 video encoder attribute error and add-new-model-like tokenizer mapping failure are unrelated." + "reason": "SAM 3 Video attribute failure and add-new-model-like tokenizer-mapping failure are not the same code path." }, { "left": "issue:42831", "right": "issue:44222", "accept": false, - "reason": "FineGrainedFP8 accuracy regression and FP8 save_pretrained failure are distinct FP8-related issues, not the same bug." + "reason": "FineGrainedFP8 accuracy degradation and FP8 save_pretrained bugs are different FP8-related issues." }, { "left": "issue:45310", "right": "issue:45397", "accept": false, - "reason": "Qwen3.5 Moe from_pretrained failure and Gemma-4 Zero3 from_pretrained failure are different model paths." + "reason": "Qwen3.5 Moe loading error and gemma-4 zero3 loading error are separate model regressions." }, { "left": "issue:45362", "right": "issue:45375", "accept": false, - "reason": "Qwen3.5-35B chat crash and missing deepstack_visual_indexes config field are unrelated." + "reason": "Qwen3.5-35B chat crash and missing deepstack_visual_indexes in a vision config are unrelated." }, { - "left": "issue:43232", - "right": "issue:44188", + "left": "issue:41628", + "right": "issue:44351", + "accept": false, + "reason": "AutoImageProcessor import failure and HybridCache import failure are different missing-symbol regressions." + }, + { + "left": "issue:45081", + "right": "issue:45356", "accept": false, - "reason": "Generation kwargs handling after sync_gpus and attention-kernel divergence under torch.compile are different code paths and symptoms." + "reason": "Mistral regex patch crash and Kimi-K2.5 tokenizer regression are distinct tokenizer bugs." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The summary overstates uniqueness. Besides the accepted Qwen3ForCausalLM duplicate pair (42673/43704), the packet also contains another identical-title duplicate candidate: issue:45310 and issue:45478 both read \"[BUG] transformers>=5.4.0, Qwen3.5 Moe from_pretrained error\". Even if that pair was not surfaced in soft_pairs, the writeup should not claim there is only one duplicate-like pair in the set. The rest of the conservative pairwise reasoning is mostly grounded, but this omission makes the overall summary too strong." + "accept": true, + "feedback": "The summary is grounded in the packet, and the duplicate call on issue:42673/issue:43704 is supported by the identical title and report pattern. The remaining soft-edge rejections stay conservative and do not overclaim shared root causes." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -10512,8 +10451,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -10545,7 +10484,7 @@ "issue:44188", "issue:44222", "issue:44242", - "issue:44351", + "issue:44297", "issue:44373", "issue:44448", "issue:44496", @@ -10567,15 +10506,15 @@ "issue:45561" ], "soft_pairs": [ - "issue:41628|issue:44351", + "issue:43232|issue:44188", "issue:44222|issue:45310", "issue:42994|issue:44164", "issue:44038|issue:45310", + "issue:43232|issue:44792", "issue:44496|issue:45561", "issue:43526|issue:45561", - "issue:43232|issue:44792", - "issue:43504|issue:45561", "issue:43377|issue:44242", + "issue:43504|issue:45561", "issue:36010|issue:41628", "issue:43673|issue:44188", "issue:30990|issue:35141", @@ -10591,216 +10530,235 @@ "issue:44060|issue:45125", "issue:43723|issue:44568", "issue:42898|issue:45356", - "issue:44734|issue:45103", "issue:43278|issue:45137", + "issue:44734|issue:45103", "issue:42898|issue:44779", - "issue:30333|issue:33290", "issue:43874|issue:45092", + "issue:30333|issue:33290", "issue:42898|issue:44448", - "issue:41628|issue:45459" + "issue:41628|issue:45459", + "issue:44222|issue:44297" ], - "prepared_review_unit_hash": "7235562a98130ca40aa000b24342bf404212115f1bf79b220407fe280ab08e12", + "prepared_review_unit_hash": "dc37dc9fd1922b47509c9e4ff77430137fea80f305545fbfe8adb410c1514b9e", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 31, - "serialized_chars": 36970, - "estimated_input_tokens": 9243, - "estimated_eval_tokens": 18742 + "soft_pair_count": 32, + "serialized_chars": 37314, + "estimated_input_tokens": 9329, + "estimated_eval_tokens": 18914 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is a loose collection of mostly unrelated Transformers v5-era bug reports (imports, tokenizer regressions, multimodal configs, save/load, generation, and backend issues). None of the soft pairs looks like the same concrete bug/change, so all are rejected.", - "confidence": 0.89, - "canonical_issue_reason": "issue:42898 is the broadest, most generic user-facing regression in the set and the best placeholder representative for this mostly v5-regression cluster, though it does not duplicate the others.", + "summary": "The cluster is a set of mostly unrelated transformer bugs and regressions; none of the soft pairs look like the same underlying issue, so no duplicate merges are justified.", + "confidence": 0.93, + "canonical_issue_reason": "Issue 42898 is the best single representative because it is a clear, user-facing v5 tokenizer regression with a broad and well-scoped description; however, it does not subsume the other reports.", "canonical_pr_reason": null, - "best_issue_reason": "issue:42898 is the most reusable canonical issue because it describes a common v5 behavior break rather than a narrow model-specific or infrastructure-specific failure.", + "best_issue_reason": "Issue 42898 is the strongest anchor for this cluster if one must be chosen: it is concrete, reproducible, and broadly representative of the v5 behavior-change reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:41628", - "right": "issue:44351", + "left": "issue:43232", + "right": "issue:44188", "accept": false, - "reason": "Both are import errors, but for different symbols and different code paths; this is shared symptom, not the same bug." + "reason": "Both mention generation internals, but one is about sync_gpus/update kwargs and the other about torch.compile branching in attention kernels; different code paths and symptoms." }, { "left": "issue:44222", "right": "issue:45310", "accept": false, - "reason": "Both involve Qwen3.5/MoE, but one is FP8 save_pretrained and the other is a from_pretrained load failure; not the same fix." + "reason": "Both involve MoE/FP8 wording, but one is about save_pretrained for FP8 and the other is a from_pretrained loading error for Qwen3.5 MoE; not the same bug." }, { "left": "issue:42994", "right": "issue:44164", "accept": false, - "reason": "Both touch persistence, but quantized model saving and extra_state handling are distinct serialization bugs." + "reason": "One is about quantized model saving failure, the other about save/from_pretrained handling extra_state; related persistence area, but distinct failure modes and fixes." }, { "left": "issue:44038", "right": "issue:45310", "accept": false, - "reason": "Both mention Qwen VL/MoE, but they are different model/version failures and not the same code-path problem." + "reason": "Qwen3-VL-Moe transformers v5 compatibility is a different model-loading issue from Qwen3.5 MoE from_pretrained errors." + }, + { + "left": "issue:43232", + "right": "issue:44792", + "accept": false, + "reason": "Generation cache/update logic is unrelated to a failed janus image-generation test." }, { "left": "issue:44496", "right": "issue:45561", "accept": false, - "reason": "Completely unrelated: model config loading versus a pytest-xdist race in testing utils." + "reason": "Model config auto-discovery and pytest-xdist file races are unrelated." }, { "left": "issue:43526", "right": "issue:45561", "accept": false, - "reason": "Different subsystems and failure modes; image processor label reduction is unrelated to the test file race." + "reason": "A BeitImageProcessorFast label-reduction bug is unrelated to a testing-utils race on captured_info.txt." }, { - "left": "issue:43232", - "right": "issue:44792", + "left": "issue:43377", + "right": "issue:44242", "accept": false, - "reason": "One is a generation kwargs bug after sync_gpus; the other is a Janus image generation test failure. Too different." + "reason": "Missing padding-mask support in MIMI encoder is unrelated to MoE load-balancing loss gating." }, { "left": "issue:43504", "right": "issue:45561", "accept": false, - "reason": "Semantic-segmentation preset loading and a testing-utils race are unrelated." - }, - { - "left": "issue:43377", - "right": "issue:44242", - "accept": false, - "reason": "Different bugs: padding mask batched-vs-single mismatch versus MoE load-balancing loss gating." + "reason": "A legacy-field model preset loading bug is unrelated to a test-worker file race." }, { "left": "issue:36010", "right": "issue:41628", "accept": false, - "reason": "Both are import failures, but they are different missing exports with no shared concrete fix indicated." + "reason": "Both are import errors, but they concern different symbols and modules: GenerationMixin vs AutoImageProcessor." }, { "left": "issue:43673", "right": "issue:44188", "accept": false, - "reason": "Generation cache support and attention-kernel divergence under torch.compile are separate problems." + "reason": "Chunked-prefill cache handling and compile-time attention-kernel branching are different generation bugs." }, { "left": "issue:30990", "right": "issue:35141", "accept": false, - "reason": "Stuck loading sentence-transformers is unrelated to token embedding reinitialization during resize." + "reason": "Sentence Transformers loading hangs and token-embedding resizing reinitialization are unrelated." }, { "left": "issue:42898", "right": "issue:44568", "accept": false, - "reason": "Both are tokenizer regressions in v5, but they affect different behaviors and different tokenizers." + "reason": "Both are tokenizer v5 regressions, but one is about clean_up_tokenization_spaces and the other about add_special_tokens/BOS-EOS behavior; different semantics and code paths." }, { "left": "issue:43479", "right": "issue:44079", "accept": false, - "reason": "Both involve None-handling, but in unrelated classes and code paths; not mergeable as one bug." + "reason": "Default multimodal config initialization and ModelOutput key assignment are unrelated." }, { "left": "issue:43479", "right": "issue:44869", "accept": false, - "reason": "Multimodal config defaults and Whisper timestamp decoding are unrelated." + "reason": "Default multimodal config initialization is unrelated to Whisper word-timestamp decoding on trailing replacement characters." }, { "left": "issue:43976", "right": "issue:44933", "accept": false, - "reason": "Python-version compatibility and a missing import are unrelated issues." + "reason": "Python version compatibility and a missing image_utils import are different issues with no shared underlying bug." }, { "left": "issue:42175", "right": "issue:44933", "accept": false, - "reason": "Backend dependency packaging and image_utils import exposure are unrelated." + "reason": "Backend dependency installation and a nonexistent image_utils import are unrelated." }, { "left": "issue:44964", "right": "issue:45137", "accept": false, - "reason": "Model loading failure and DeepSpeed ZeRO3 deque handling are unrelated." + "reason": "Phi-4 multimodal loading and DeepSpeed ZeRO3 deque underflow are different failures." }, { "left": "issue:43784", "right": "issue:43824", "accept": false, - "reason": "Different import errors affecting different APIs/models; no same underlying bug." + "reason": "Both are import failures, but one is a NameError during sentence-transformers import and the other is a missing Qwen2.5-VL class export." }, { "left": "issue:44623", "right": "issue:45310", "accept": false, - "reason": "Processor save_pretrained missing files is a different path from Qwen3.5 MoE from_pretrained loading." + "reason": "Processor save_pretrained missing files is unrelated to Qwen3.5 MoE from_pretrained loading errors." }, { "left": "issue:44188", "right": "issue:44373", "accept": false, - "reason": "Generation runtime behavior and a docstring typo are not the same bug." + "reason": "A generation-kernel divergence bug is unrelated to a wrong docstring for position_ids." }, { "left": "issue:44060", "right": "issue:45125", "accept": false, - "reason": "Both are Qwen3.5 MoE internals, but tied-weights warning and tensor-parallel plan support are different issues." + "reason": "Both mention Qwen3 MoE/tensor parallelism, but one is a tied-weights warning bug and the other is a missing _tp_plan attribute; different concrete problems." }, { "left": "issue:43723", "right": "issue:44568", "accept": false, - "reason": "Tokenizer loading in v5 and BOS/EOS add_special_tokens behavior are different regressions." + "reason": "Tokenizer loading in v5 and special-token addition behavior are distinct tokenizer issues." }, { "left": "issue:42898", "right": "issue:45356", "accept": false, - "reason": "Both are tokenizer-related regressions, but the specific failure modes and affected code paths differ." + "reason": "Both are tokenizer regressions, but one is clean_up_tokenization_spaces and the other is Kimi-K2.5 codec handling with a misleading warning; not the same defect." }, { - "left": "issue:44734", - "right": "issue:45103", + "left": "issue:43278", + "right": "issue:45137", "accept": false, - "reason": "Serving KV-cache indexing and auto-docstring annotation handling are unrelated." + "reason": "Embedding dtype drift in evaluate and a ZeRO3 deque pop error are unrelated." }, { - "left": "issue:43278", - "right": "issue:45137", + "left": "issue:44734", + "right": "issue:45103", "accept": false, - "reason": "Embedding dtype drift in evaluate and a ZeRO3 deque IndexError are unrelated." + "reason": "KV-cache continuation indexing in serving and an auto_docstring AttributeError from future annotations are unrelated." }, { "left": "issue:42898", "right": "issue:44779", "accept": false, - "reason": "Both are tokenizer regressions in v5, but they are model-specific and not the same concrete bug." + "reason": "Both are tokenizer regressions in v5, but they affect different models and different mechanisms: cleanup spaces versus DeepSeek codec output." + }, + { + "left": "issue:43874", + "right": "issue:45092", + "accept": false, + "reason": "A missing image-patch counting method and an old InternVL2 meta-init incompatibility are different multimodal-loading issues." }, { "left": "issue:30333", "right": "issue:33290", "accept": false, - "reason": "MLflow reporting on failed jobs and DeepSpeed OOM are unrelated." + "reason": "MLflow job status reporting and Deepspeed/Adafactor OOM are unrelated training issues." }, { - "left": "issue:43874", - "right": "issue:45092", + "left": "issue:42898", + "right": "issue:44448", + "accept": false, + "reason": "A tokenizer cleanup regression and a Pegasus output mismatch across v4/v5 are different behavior changes." + }, + { + "left": "issue:41628", + "right": "issue:45459", + "accept": false, + "reason": "AutoImageProcessor import failure and protobuf-related tokenizer error masking are different import/runtime issues." + }, + { + "left": "issue:44222", + "right": "issue:44297", "accept": false, - "reason": "Different multimodal issues: a missing image-processor method versus legacy remote-code checkpoint meta-init incompatibility." + "reason": "FP8 MoE save_pretrained behavior and tokenizer_class metadata mismatch are unrelated save-path bugs." } ] }, @@ -10819,8 +10777,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -10873,7 +10831,6 @@ "issue:45459" ], "soft_pairs": [ - "issue:44222|issue:44297", "issue:42898|issue:43295", "issue:43479|issue:43644", "issue:44442|issue:45459", @@ -10894,253 +10851,253 @@ "issue:28282|issue:30333", "issue:44295|issue:44623", "issue:44514|issue:45103", - "issue:43479|issue:44913", "issue:43723|issue:44871", + "issue:43479|issue:44913", "issue:44261|issue:45092", "issue:43723|issue:44991", - "issue:43723|issue:44442", "issue:44373|issue:44485", + "issue:43723|issue:44442", "issue:45092|issue:45357", "issue:30990|issue:38175", "issue:42886|issue:42994", "issue:42831|issue:45310", "issue:41628|issue:43824", "issue:42831|issue:44038", - "issue:43576|issue:43784", - "issue:42757|issue:45278" + "issue:42757|issue:45278", + "issue:43576|issue:43784" ], - "prepared_review_unit_hash": "46c1584c60316f39b14eaebc1ef462ba6065f7a143c54d82795402c5bcf3cbff", + "prepared_review_unit_hash": "de99eca2efe21f3403b8e850e71dda177ea26d57ce219d326a0714d3373d2da0", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 35, - "serialized_chars": 37352, - "estimated_input_tokens": 9338, - "estimated_eval_tokens": 18932 + "soft_pair_count": 34, + "serialized_chars": 37052, + "estimated_input_tokens": 9263, + "estimated_eval_tokens": 18782 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is mostly heterogeneous. Most soft pairs share a subsystem or symptom class, but not the same concrete bug, so they should not be merged as duplicates. The broadest issue is the v5 import-regression report, but it still doesn\u2019t cleanly cover the rest.", - "confidence": 0.43, - "canonical_issue_reason": "issue:45278 is the broadest umbrella here: it describes many import errors after upgrading to v5.5.0, which makes it the best candidate to subsume related import/load regressions.", + "summary": "This cluster is heterogeneous: most items are unrelated regressions or import/load failures in different subsystems, so the soft duplicate links are not strong enough to merge. No PRs are present.", + "confidence": 0.95, + "canonical_issue_reason": "No single canonical issue cleanly subsumes the cluster. The closest thematic anchor is issue:43723 (tokenizer loading in v5), but the surrounding items span unrelated bugs and are not duplicate-equivalent.", "canonical_pr_reason": null, - "best_issue_reason": "It is the most representative high-level issue in the cluster, centered on the recurring post-v5 import breakage theme rather than a narrow model-specific symptom.", + "best_issue_reason": "If one issue must represent the set, issue:43723 is the closest representative because it captures the broad tokenizer-loading regression theme seen in a few nearby reports, but it still should not be treated as a duplicate hub.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44222", - "right": "issue:44297", - "accept": false, - "reason": "Both touch save_pretrained/tokenizer persistence, but one is about FP8 Moe model saving and the other about tokenizer_class metadata; different bugs." - }, { "left": "issue:42898", "right": "issue:43295", "accept": false, - "reason": "One is about clean_up_tokenization_spaces behavior, the other about processor.tokenizer access and image passing; unrelated regressions." + "reason": "Both are v5 regression-style issues, but one is about `clean_up_tokenization_spaces` behavior and the other about `processor.tokenizer` access; different code paths and fixes." }, { "left": "issue:43479", "right": "issue:43644", "accept": false, - "reason": "Different failure modes: default multimodal config initialization vs non-persistent buffers being filled with junk." + "reason": "Different problems: config default initialization for Phi4Multimodal vs non-persistent buffer garbage in Transformers 5.0." }, { "left": "issue:44442", "right": "issue:45459", "accept": false, - "reason": "Tokenizer load failure for a specific tokenizer is not the same as protobuf-related exception masking." + "reason": "Both involve tokenizer-related failures, but one is AutoTokenizer loading a specific tokenizer class and the other is protobuf error masking; not the same bug." }, { "left": "issue:44222", "right": "issue:44623", "accept": false, - "reason": "save_pretrained is shared broadly, but these are different artifacts and bugs: FP8 Moe saving vs missing processor files." + "reason": "Both mention saving, but one is FP8/MoE `save_pretrained` and the other is missing processor save files; different save paths." }, { "left": "issue:42831", "right": "issue:44623", "accept": false, - "reason": "An FP8 accuracy issue is not the same as processor.save_pretrained missing files." + "reason": "Accuracy loss in FineGrainedFP8 is unrelated to processor save file omissions." }, { "left": "issue:44206", "right": "issue:45103", "accept": false, - "reason": "One is a feature-extractor argument regression; the other is an auto-docstring AttributeError from future annotations." + "reason": "One is a feature extractor regression with unsupported `center`; the other is an auto-docstring annotation crash." }, { "left": "issue:41628", "right": "issue:42757", "accept": false, - "reason": "Both are import errors, but for different missing symbols from different packages and likely different causes." + "reason": "Both are import errors, but they concern different missing symbols from different packages and likely different dependency/version issues." }, { "left": "issue:44297", "right": "issue:45310", "accept": false, - "reason": "Tokenizer metadata mismatch is unrelated to the Qwen3.5 MoE from_pretrained error." + "reason": "Tokenizer config mismatch on save is not the same as Qwen3.5 MoE `from_pretrained` loading failure." }, { "left": "issue:22355", "right": "issue:28282", "accept": false, - "reason": "Both are import problems, but one is missing transformers.onnx and the other is a missing PyTorch dependency; not the same bug." + "reason": "Different import failures: missing `transformers.onnx` vs missing PyTorch for `AutoModel`." }, { "left": "issue:44222", "right": "issue:44484", "accept": false, - "reason": "save_pretrained shard-size behavior is unrelated to FP8 Moe model saving failure." + "reason": "One is a concrete FP8 MoE save failure; the other is a question about `max_shard_size` defaults." }, { "left": "issue:44483", "right": "issue:45406", "accept": false, - "reason": "An API request rejection and a Gemma4Processor _tokenizer AttributeError are different code paths." + "reason": "Chat-completions request rejection is unrelated to Gemma4Processor missing `_tokenizer` in serve." }, { "left": "issue:43723", "right": "issue:44977", "accept": false, - "reason": "Both involve loading/generation, but one is generic tokenizer loading and the other is Qwen3.5 flash-attention generation." + "reason": "Both are tokenizer/model-loading regressions in v5, but one is generic `AutoTokenizer` loading and the other is Qwen3.5 flash-attention generation; not the same code-path bug." }, { "left": "issue:44263", "right": "issue:44360", "accept": false, - "reason": "Both concern the DSA indexer, but one is about torch.split outputs and the other about a missing ReLU; different defects." + "reason": "Both mention GLM DSA internals, but one is a `torch.split` return-value issue and the other is a missing ReLU in the indexer." }, { "left": "issue:30333", "right": "issue:36331", "accept": false, - "reason": "MLFlow job status reporting and CustomTrainer compute_loss signature handling are unrelated." + "reason": "MLflow reporting on failed jobs is unrelated to the trainer `compute_loss(num_items_in_batch)` signature break." }, { "left": "issue:44263", "right": "issue:45092", "accept": false, - "reason": "Different models and failures: DSA indexer split behavior vs remote-code checkpoint incompatibility with meta init." + "reason": "GLM indexer behavior and InternVL2 meta-initialization incompatibility are different bugs." }, { "left": "issue:30333", "right": "issue:34689", "accept": false, - "reason": "Training-job MLFlow reporting is unrelated to Llama 3.2 Vision model-loading regression." + "reason": "Training-job reporting with MLflow is unrelated to Llama 3.2 Vision model-loading regression." }, { "left": "issue:30333", "right": "issue:35141", "accept": false, - "reason": "Job reporting and embedding reinitialization after resize are unrelated." + "reason": "Failed-job reporting has nothing to do with output embeddings being reinitialized after resizing token embeddings." }, { "left": "issue:28282", "right": "issue:30333", "accept": false, - "reason": "A missing PyTorch dependency import error is unrelated to MLFlow reporting on failed training jobs." + "reason": "PyTorch-missing import error and MLflow failed-job reporting are unrelated." }, { "left": "issue:44295", "right": "issue:44623", "accept": false, - "reason": "Reading position_ids after buffer registration is not the same as missing files in processor.save_pretrained." + "reason": "Registering `position_ids` as a buffer and processor save-file omissions are unrelated." }, { "left": "issue:44514", "right": "issue:45103", "accept": false, - "reason": "A batched apply_chat_template crash is unrelated to the auto_docstring AttributeError." + "reason": "Qwen2_5_VL batched chat-template crash and auto-docstring annotation crash are unrelated." + }, + { + "left": "issue:43723", + "right": "issue:44871", + "accept": false, + "reason": "Tokenizer loading in v5 and Gemma-3 EOS token config inconsistency are different issues." }, { "left": "issue:43479", "right": "issue:44913", "accept": false, - "reason": "Default multimodal config initialization is unrelated to GPTNeoXConfig rotary_pct reload behavior." + "reason": "Phi4Multimodal config initialization and GPTNeoX rotary_pct reloading are unrelated config bugs." }, { - "left": "issue:43723", - "right": "issue:44871", + "left": "issue:44261", + "right": "issue:45092", "accept": false, - "reason": "Generic tokenizer loading failure and Gemma-3 eos_token_id inconsistency are different issues." + "reason": "MLA q_a_layernorm precision error is not the same as old InternVL2 checkpoint meta-init incompatibility." }, { "left": "issue:43723", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer-related regressions, but they affect different models and appear to have different root causes." + "reason": "Both concern tokenizer loading in v5, but the evidence is too weak to say they are the same concrete bug or fixable by one PR." }, { - "left": "issue:43723", - "right": "issue:44442", + "left": "issue:44373", + "right": "issue:44485", "accept": false, - "reason": "Tokenizer loading failure and FastSpeech2ConformerTokenizer loading are both load issues, but not the same concrete bug." + "reason": "A wrong docstring for `position_ids` is not the same as the GLM-5 RoPE implementation discussion." }, { - "left": "issue:44373", - "right": "issue:44485", + "left": "issue:43723", + "right": "issue:44442", "accept": false, - "reason": "A docstring mistake and a GLM-5 RoPE implementation discussion are not duplicates." + "reason": "Generic tokenizer loading regression and FastSpeech2ConformerTokenizer import/load failure are different failures." }, { "left": "issue:45092", "right": "issue:45357", "accept": false, - "reason": "Meta-init incompatibility for old InternVL2 checkpoints is unrelated to incorrect visual encoder keys in save_pretrained." + "reason": "Both involve Qwen3.5-family models, but one is an old remote-code meta-init incompatibility and the other is incorrect visual encoder keys on save." }, { "left": "issue:30990", "right": "issue:38175", "accept": false, - "reason": "Sentence-Transformers getting stuck loading and SigLIP2 zero probabilities are different problems." + "reason": "Sentence Transformers getting stuck loading and SigLIP2 zero-probabilities are unrelated model behaviors." }, { "left": "issue:42886", "right": "issue:42994", "accept": false, - "reason": "Offline cache tokenizer loading and quantized model saving failure are separate bugs." + "reason": "Offline tokenizer cache loading and quantized model saving are different code paths." }, { "left": "issue:42831", "right": "issue:45310", "accept": false, - "reason": "FP8 accuracy issues are unrelated to Qwen3.5 MoE from_pretrained errors." + "reason": "FP8 accuracy issues are unrelated to Qwen3.5 MoE `from_pretrained` loading errors." }, { "left": "issue:41628", "right": "issue:43824", "accept": false, - "reason": "Different missing imports: AutoImageProcessor versus Qwen2_5_VLForConditionalGeneration." + "reason": "Both are import errors from transformers, but they involve different missing exported classes and likely different fixes." }, { "left": "issue:42831", "right": "issue:44038", "accept": false, - "reason": "FP8 accuracy and Qwen3.5-v5 bug reports are not the same code-path problem." + "reason": "FineGrainedFP8 accuracy and Qwen3.5-VL-Moe bug reports are unrelated." }, { - "left": "issue:43576", - "right": "issue:43784", + "left": "issue:42757", + "right": "issue:45278", "accept": false, - "reason": "Broken `transformers env` command and sentence-transformers import NameError have different causes and effects." + "reason": "A specific missing `huggingface_hub.is_offline_mode` import is not the same as a broad report of many import errors after upgrading to v5.5.0." }, { - "left": "issue:42757", - "right": "issue:45278", + "left": "issue:43576", + "right": "issue:43784", "accept": false, - "reason": "The broad import-regression umbrella may include this symptom, but the specific missing is_offline_mode import is not itself the same bug as the whole cluster." + "reason": "Broken `transformers env` command and `sentence-transformers` import NameError are different regressions." } ] }, @@ -11159,8 +11116,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -11216,17 +11173,17 @@ "issue:43122|issue:44521", "issue:44263|issue:44373", "issue:43232|issue:45593", - "issue:42886|issue:43066", "issue:44117|issue:44704", + "issue:42886|issue:43066", "issue:43122|issue:44514", - "issue:30990|issue:33357", "issue:42371|issue:44263", - "issue:44295|issue:44297", + "issue:30990|issue:33357", "issue:44704|issue:45290", + "issue:44295|issue:44297", "issue:42907|issue:44945", "issue:43976|issue:44945", - "issue:43232|issue:44743", "issue:39401|issue:44488", + "issue:43232|issue:44743", "issue:43232|issue:45245", "issue:42907|issue:43937", "issue:43408|issue:43475", @@ -11234,228 +11191,252 @@ "issue:43784|issue:45600", "issue:44625|issue:44829", "issue:44360|issue:44485", - "issue:43116|issue:45200", "issue:39401|issue:44521", + "issue:43116|issue:45200", "issue:43122|issue:44779", - "issue:29127|issue:30990", "issue:44038|issue:44297", + "issue:29127|issue:30990", "issue:44261|issue:44373", "issue:44038|issue:44295", "issue:42371|issue:42831", "issue:42907|issue:44964", - "issue:36296|issue:43116", "issue:44038|issue:44222", + "issue:36296|issue:43116", "issue:42907|issue:43452", "issue:43992|issue:44704" ], - "prepared_review_unit_hash": "54128892aaae869e6a6ccd91519a8c90862d6ed7f75e655bce0fcf3457464c93", + "prepared_review_unit_hash": "81d4856a5ac29d80c275ab3660bd5364e9a2220e904c968b21ab9310e1afec43", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 34, - "serialized_chars": 36859, - "estimated_input_tokens": 9215, - "estimated_eval_tokens": 18686 + "serialized_chars": 36855, + "estimated_input_tokens": 9214, + "estimated_eval_tokens": 18684 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous: the items cover unrelated tokenizer, generation, import, multimodal, training, and model-specific bugs. None of the soft-edge pairs look like the same underlying issue closely enough to merge as duplicates.", - "confidence": 0.97, - "canonical_issue_reason": null, + "summary": "Mostly false-positive similarity edges across unrelated Transformers bugs; the closest cluster is around tokenizer/regression behavior in v5, but none of the candidate pairs look like true duplicates of the same concrete bug.", + "confidence": 0.23, + "canonical_issue_reason": "issue:43122 is the broadest tokenizer-regression report here and sits near several of the tokenizer/chat-template soft links, so it is the best representative issue for triage.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "Same as canonical: it is the most general issue in the nearest tokenizer-related subset, but the overall cluster is too mixed to treat as one duplicate family.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43122", "right": "issue:44521", "accept": false, - "reason": "Both mention tokenization/chat behavior, but one is a tokenizer-v4/v5 discrepancy and the other is zero assistant masks for multimodal inputs; different code paths and symptoms." + "reason": "Both involve tokenizer/template output, but one is a version-wide tokenization change and the other is zero assistant masks for multimodal inputs." }, { "left": "issue:44263", "right": "issue:44373", "accept": false, - "reason": "One is a torch.split indexing bug in a model component, the other is a docstring issue for position_ids; not the same bug." + "reason": "One is a torch.split behavior bug and the other is only a wrong docstring for position_ids; not the same defect." }, { "left": "issue:43232", "right": "issue:45593", "accept": false, - "reason": "Generation sync/update logic vs D-FINE auxiliary loss handling are unrelated model subsystems." - }, - { - "left": "issue:42886", - "right": "issue:43066", - "accept": false, - "reason": "Offline cache loading failure and wrong tokenizer decoder type are distinct tokenizer issues." + "reason": "Generation kwargs after sync_gpus and missing auxiliary losses in D-FINE are unrelated code paths." }, { "left": "issue:44117", "right": "issue:44704", "accept": false, - "reason": "TOKENIZER_MAPPING_NAMES returning None is a registry bug; AutoProcessor not forwarding kwargs to cached_file is an API plumbing bug." + "reason": "Both touch from_pretrained internals, but one is TOKENIZER_MAPPING_NAMES returning None and the other is cached_file kwargs passthrough." }, { - "left": "issue:43122", - "right": "issue:44514", + "left": "issue:42886", + "right": "issue:43066", "accept": false, - "reason": "Both involve tokenization, but one is differing tokenization outputs across versions and the other is a batched apply_chat_template crash with padding=False." + "reason": "Offline cache loading failure and wrong tokenizer decoder type are different tokenizer bugs." }, { - "left": "issue:30990", - "right": "issue:33357", + "left": "issue:43122", + "right": "issue:44514", "accept": false, - "reason": "Sentence Transformers loading hang and a MacOS bus error with a CLIP model are unrelated runtime failures." + "reason": "Both are tokenizer-related, but this is a batched chat-template crash, not the same regression as version-to-version tokenization differences." }, { "left": "issue:42371", "right": "issue:44263", "accept": false, - "reason": "TF32 configuration guidance and a torch.split return-value bug are unrelated." + "reason": "TF32 behavior settings and a torch.split return-value issue are unrelated." }, { - "left": "issue:44295", - "right": "issue:44297", + "left": "issue:30990", + "right": "issue:33357", "accept": false, - "reason": "Reading position_ids after buffer registration and tokenizer_class mismatch on save_pretrained are separate save/load issues." + "reason": "Sentence-transformers loading hang and a MacOS bus error with a CLIP model are different failure modes." }, { "left": "issue:44704", "right": "issue:45290", "accept": false, - "reason": "Both touch chat/template flows, but one is cached_file kwargs forwarding and the other is a crash on assistant tool-call messages; different failure points." + "reason": "AutoProcessor kwargs forwarding and chat-template crashes with tool-call messages are separate bugs." + }, + { + "left": "issue:44295", + "right": "issue:44297", + "accept": false, + "reason": "Reading a buffer-backed position_ids tensor and save_pretrained writing the wrong tokenizer_class are different issues." }, { "left": "issue:42907", "right": "issue:44945", "accept": false, - "reason": "Saving dequantized Ministral/Devstral models and incorrect output under pipeline parallelism are unrelated." + "reason": "Saving dequantized Ministrals/Devstrals and incorrect output under pipeline parallelism do not share the same root cause." }, { "left": "issue:43976", "right": "issue:44945", "accept": false, - "reason": "Python-version compatibility failure and pipeline-parallel output corruption are not the same bug." + "reason": "Python version compatibility and pipeline-parallel output corruption are unrelated." }, { - "left": "issue:43232", - "right": "issue:44743", + "left": "issue:39401", + "right": "issue:44488", "accept": false, - "reason": "Both involve generation/cache internals, but one is sync_gpus update handling and the other is recurrent-state reset in modular_qwen3_5.py; different concrete code paths." + "reason": "Qwen3 offset_mapping errors and failure to load a specific tokenizer/model are not the same bug." }, { - "left": "issue:39401", - "right": "issue:44488", + "left": "issue:43232", + "right": "issue:44743", "accept": false, - "reason": "Wrong offset_mapping for Qwen3 tokenizer and failure to load cjvt/sleng-bert are distinct tokenizer problems." + "reason": "These are both generation/cache related, but one is sync_gpus kwargs handling and the other is recurrent state reset with cache." }, { "left": "issue:43232", "right": "issue:45245", "accept": false, - "reason": "Generation kwargs update after sync_gpus is unrelated to the 2^24 categories runtime error." + "reason": "Generation kwargs handling and a categories-limit runtime error are unrelated." }, { "left": "issue:42907", "right": "issue:43937", "accept": false, - "reason": "Dequantized model save failure and invalid GenerationConfig for GLM-5 are unrelated." + "reason": "Dequantized save failures and invalid GenerationConfig for GLM-5 are different issues." }, { "left": "issue:43408", "right": "issue:43475", "accept": false, - "reason": "Both are SAM 3 Video related, but one is a model-type warning and the other is a missing attribute on Sam3VisionEncoderOutput; not the same bug." + "reason": "A sam3_video/sam3_tracker type warning is not the same as the missing fpn_position_embeddings attribute." }, { "left": "issue:41628", "right": "issue:45278", "accept": false, - "reason": "AutoImageProcessor import failure is a specific symbol-export problem; the other reports many import errors after an upgrade and is much broader." + "reason": "A specific missing AutoImageProcessor import is not enough to equate with broad import failures after upgrade." }, { "left": "issue:43784", "right": "issue:45600", "accept": false, - "reason": "NameError importing sentence-transformers and auto_mappings references to removed Sam3LiteText configs are unrelated import/compatibility failures." + "reason": "Sentence-transformers import NameError and Sam3LiteText auto_mappings breakage are unrelated." }, { "left": "issue:44625", "right": "issue:44829", "accept": false, - "reason": "num_labels propagation in Qwen3.5 and flash_attention_3 degenerate training are different model/configuration bugs." + "reason": "Qwen3.5 num_labels propagation and flash_attention_3 training degeneration are different problems." }, { "left": "issue:44360", "right": "issue:44485", "accept": false, - "reason": "Missing ReLU in a DSA indexer and GLM-5 RoPE implementation are unrelated architectural issues." + "reason": "DSA indexer missing ReLU and GLM-5 RoPE implementation discussion are unrelated." }, { - "left": "issue:43116", - "right": "issue:45200", + "left": "issue:39401", + "right": "issue:44521", "accept": false, - "reason": "Empty multi-label results in an example script and Gemma 4 mm_token_type_ids defaulting are unrelated application bugs." + "reason": "Offset_mapping correctness and all-zero assistant_masks are both tokenizer-related, but they are distinct bugs." }, { - "left": "issue:39401", - "right": "issue:44521", + "left": "issue:43116", + "right": "issue:45200", "accept": false, - "reason": "Tokenizer offset mapping bug and multimodal assistant_mask zeroing are different failures." + "reason": "Multi-label example-script behavior and Gemma 4 mm_token_type_ids defaults are separate issues." }, { "left": "issue:43122", "right": "issue:44779", "accept": false, - "reason": "Both are tokenizer regressions in v5, but they affect different models and behaviors; not enough to treat as one duplicate bug." + "reason": "Both are tokenization regressions in v5, but they affect different models and symptoms, so they are not the same bug." + }, + { + "left": "issue:44038", + "right": "issue:44297", + "accept": false, + "reason": "A Qwen3-VL-Moe bug and tokenizer_class mismatch on save_pretrained are unrelated." }, { "left": "issue:29127", "right": "issue:30990", "accept": false, - "reason": "LayoutLMv3 error-message clarity and Sentence Transformers loading hang have no substantive overlap." + "reason": "LayoutLMv3 error-message clarity and Sentence Transformers loading hang are unrelated." + }, + { + "left": "issue:44261", + "right": "issue:44373", + "accept": false, + "reason": "MLA q_a_layernorm precision/config issue and wrong position_ids docstring are not the same defect." }, { "left": "issue:44038", - "right": "issue:44297", + "right": "issue:44295", "accept": false, - "reason": "Qwen3-VL-Moe bug and tokenizer_class mismatch on save_pretrained are unrelated." + "reason": "Qwen3-VL-Moe and position_ids buffer read failure are different bugs." }, { - "left": "issue:44261", - "right": "issue:44373", + "left": "issue:42371", + "right": "issue:42831", + "accept": false, + "reason": "TF32 API-setting guidance and FineGrainedFP8 accuracy issues are not the same root cause." + }, + { + "left": "issue:42907", + "right": "issue:44964", "accept": false, - "reason": "A numerical precision/config issue in MLA q_a_layernorm is not the same as a wrong docstring for position_ids." + "reason": "Dequantized save failure and inability to load Phi-4 multimodal are unrelated." }, { "left": "issue:44038", "right": "issue:44222", "accept": false, - "reason": "Both mention MOE/FP8-adjacent topics, but one is a Qwen3-VL-Moe load bug and the other is FP8 save_pretrained behavior; different concrete problems." + "reason": "Qwen3-VL-Moe regression and FP8 save_pretrained moe are only loosely related by model family, not the same bug." + }, + { + "left": "issue:36296", + "right": "issue:43116", + "accept": false, + "reason": "Tensor parallel training bug and multi-label classification example failure are unrelated." }, { "left": "issue:42907", "right": "issue:43452", "accept": false, - "reason": "Saving dequantized models and gguf_file breaking from_pretrained are distinct model-loading/saving bugs." + "reason": "Dequantized Ministrals/Devstrals save failure and gguf_file load failure are different code paths." }, { "left": "issue:43992", "right": "issue:44704", "accept": false, - "reason": "Missing embed_tokens.weight for UMT5Encoder and AutoProcessor kwargs forwarding are unrelated from_pretrained issues." + "reason": "Missing embed_tokens.weight in UMT5Encoder and AutoProcessor cached_file kwargs handling are unrelated." } ] }, @@ -11474,8 +11455,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -11509,18 +11490,19 @@ "issue:44075", "issue:44117", "issue:44222", + "issue:44291", "issue:44351", "issue:44485", "issue:44779", "issue:44936", "issue:44945", "issue:44977", + "issue:45071", "issue:45072", "issue:45083", "issue:45216", "issue:45278", "issue:45341", - "issue:45405", "issue:45446", "issue:45478", "issue:45561", @@ -11558,222 +11540,222 @@ "issue:43937|issue:44977", "issue:45341|issue:45561", "issue:42673|issue:43906", - "issue:45405|issue:45446" + "issue:44291|issue:45071" ], - "prepared_review_unit_hash": "787b2d7c324517c2da68777872efb5e3993813c915429ba1b2f7bc6a3fc49af9", + "prepared_review_unit_hash": "7d59db835bee6fee1d619ce0b55f7178f9d391d16471c5515d84703acded473f", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 36001, - "estimated_input_tokens": 9001, - "estimated_eval_tokens": 18258 + "serialized_chars": 36612, + "estimated_input_tokens": 9153, + "estimated_eval_tokens": 18562 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These are mostly superficially similar bug reports, but they do not describe the same concrete defect or mergeable fix. I found no true duplicate pair among the soft edges, so there is no safe canonical issue/PR for the cluster.", - "confidence": 0.93, - "canonical_issue_reason": null, + "summary": "The set is highly heterogeneous: most pairs share only broad symptoms (imports, tokenization, save/load, training) or a common subsystem, but not the same concrete bug. I reject all soft edges and see no true duplicate cluster here.", + "confidence": 0.84, + "canonical_issue_reason": "No single issue cleanly represents a duplicate cluster. If forced, issue:45278 is the broadest umbrella, but the items are too diverse for a real canonical duplicate.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "issue:45278 is the closest thing to a cluster anchor because it describes a broad import-regression symptom, but it is still too generic to serve as a strong canonical for the whole set.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45216", "right": "issue:45478", "accept": false, - "reason": "Same model family and version range, but one is a save_pretrained checkpoint regression and the other is a from_pretrained load error; different code paths and symptoms." + "reason": "Same model family and version range, but one is save_pretrained output corruption and the other is from_pretrained loading failure; not the same bug." }, { "left": "issue:43824", "right": "issue:44351", "accept": false, - "reason": "Both are import errors, but they concern different missing symbols with different causes." + "reason": "Both are import errors, but for different missing symbols and likely different breakages." }, { "left": "issue:43122", "right": "issue:44977", "accept": false, - "reason": "Tokenizer regression vs flash-attention generation failure; unrelated underlying bugs." + "reason": "Both concern tokenizer behavior, but one is a generic version-to-version change and the other is a DeepSeek-specific regression." }, { "left": "issue:42907", "right": "issue:44222", "accept": false, - "reason": "Both involve saving quantized models, but they target different quantization formats and model families." + "reason": "Both mention saving/quantized models, but they are different model families and different failure modes." }, { "left": "issue:42757", "right": "issue:44351", "accept": false, - "reason": "Different missing imports; no shared failure mechanism beyond generic ImportError." + "reason": "Both are import failures, but they involve different packages/symbols and different root causes." }, { "left": "issue:43278", "right": "issue:44945", "accept": false, - "reason": "Training/eval dtype mismatch is not the same defect as wrong outputs under pipeline parallelism." + "reason": "Different areas: dtype change in evaluation versus incorrect outputs under pipeline parallelism." }, { "left": "issue:43784", "right": "issue:45446", "accept": false, - "reason": "Sentence-transformers import breakage vs a flex_attention version check bug; unrelated." + "reason": "One is a missing internal name during import; the other is a bad version gate for a flex_attention import." }, { "left": "issue:42222", "right": "issue:43441", "accept": false, - "reason": "Different model families and different failure modes." + "reason": "Different model families and different subsystems: VitPose breakage versus FlashAttention failure." }, { "left": "issue:42673", "right": "issue:45072", "accept": false, - "reason": "VRAM leak in threaded dataloading is unrelated to bfloat16 dtype mismatches." + "reason": "VRAM leak in Qwen3ForCausalLM is unrelated to dtype mismatches in SwitchTransformers/TimmWrapperModel." }, { "left": "issue:38617", "right": "issue:45278", "accept": false, - "reason": "One is a specific missing import, the other is a broad report of many import errors; too coarse to treat as the same bug." + "reason": "Both are import-related regressions, but one is a single missing symbol while the other is a broad update report; not the same concrete bug." }, { "left": "issue:43064", "right": "issue:43673", "accept": false, - "reason": "Wrong optimizer state distribution and missing generation cache are distinct bugs." + "reason": "Training optimizer-state corruption is unrelated to a missing generation cache in chunked prefill." }, { "left": "issue:43937", "right": "issue:44485", "accept": false, - "reason": "GenerationConfig validation failure vs RoPE implementation discussion; not the same concrete issue." + "reason": "GenerationConfig validation failure is unrelated to a RoPE implementation discussion." }, { "left": "issue:22355", "right": "issue:29127", "accept": false, - "reason": "Module import failure and a LayoutLMv3 error-message clarity issue are unrelated." + "reason": "Completely different failures: missing transformers.onnx module versus unclear LayoutLMv3 error message." }, { "left": "issue:42371", "right": "issue:43874", "accept": false, - "reason": "TF32 settings guidance vs a missing image-patch helper method; different subsystems and bugs." + "reason": "TF32 settings advice and missing multimodal-token helper are unrelated." }, { "left": "issue:43116", "right": "issue:43493", "accept": false, - "reason": "Example script multi-label output bug vs SigLIP2 implementation discrepancy; not the same defect." + "reason": "Example-script classification bug versus SigLIP2 implementation discrepancy; different code paths." }, { "left": "issue:44075", "right": "issue:44936", "accept": false, - "reason": "Unused SGD args and train-then-evaluate failure are separate training bugs." + "reason": "Optimizer argument handling and trainer.evaluate after train are different bugs." }, { "left": "issue:44222", "right": "issue:45478", "accept": false, - "reason": "Both mention Qwen3.5/MoE saving/loading, but one is FP8 save_pretrained and the other is a from_pretrained error." + "reason": "Both involve Qwen3.5/MoE save/load flows, but one is FP8 save_pretrained and the other is from_pretrained loading; not the same defect." }, { "left": "issue:44038", "right": "issue:45478", "accept": false, - "reason": "Same broad model area, but one is a vague Qwen3-VL-Moe bug and the other is a specific Qwen3.5 MoE load error." + "reason": "Same broad Qwen3.5 area, but different symptoms and likely different failing code paths." }, { "left": "issue:43116", "right": "issue:43825", "accept": false, - "reason": "Example script behavior vs pipeline API error messaging; not the same bug." + "reason": "Classification example failure is unrelated to a pipeline() error message regression." }, { "left": "issue:43976", "right": "issue:45446", "accept": false, - "reason": "Python version compatibility issue vs a PyTorch version-guard bug." + "reason": "Python-version compatibility issue versus a flex_attention import guard bug." }, { "left": "issue:43937", "right": "issue:45563", "accept": false, - "reason": "Invalid GenerationConfig and a stale warning on paged generate are unrelated." + "reason": "GenerationConfig invalidity and a stale warning about num_return_sequences are different issues." }, { "left": "issue:39401", "right": "issue:44779", "accept": false, - "reason": "Both tokenizer regressions, but they affect different models and different behaviors (offset mapping vs tokenization output)." + "reason": "Both tokenizer-related, but one is offset_mapping and the other is generic incorrect tokenization in DeepSeek." }, { "left": "issue:33453", "right": "issue:39401", "accept": false, - "reason": "Tokenizer loading regression and offset_mapping bug are not the same concrete defect." + "reason": "Tokenizer-loading regression and Qwen3 offset_mapping bug are not the same concrete failure." }, { "left": "issue:43931", "right": "issue:44945", "accept": false, - "reason": "Weight-shape mismatch on model load is unrelated to incorrect outputs under pipeline parallelism." + "reason": "Model weight-shape mismatch and pipeline-parallelism incorrect output are unrelated." }, { "left": "issue:45083", "right": "issue:45446", "accept": false, - "reason": "Multimodal feature helper behavior and a flex_attention import/version check are unrelated." + "reason": "A helper-function behavior change in qwen3_omni_moe is unrelated to a PyTorch version check bug." }, { "left": "issue:44117", "right": "issue:45588", "accept": false, - "reason": "Tokenizer mapping assumption bug vs flash_attention crash on sink-less models; different code paths." + "reason": "Tokenizer mapping fallback bug and flash_attention s_aux=None crash are different subsystems and failures." }, { "left": "issue:43827", "right": "issue:45446", "accept": false, - "reason": "Documentation still referencing pipeline() is unrelated to the flex_attention import guard." + "reason": "Documentation mentioning pipeline() is unrelated to a runtime version-check import bug." }, { "left": "issue:43937", "right": "issue:44977", "accept": false, - "reason": "Different GLM-5 generation-config bug vs Qwen3.5 flash-attention generation bug." + "reason": "Different generation/tokenization problems with different model contexts." }, { "left": "issue:45341", "right": "issue:45561", "accept": false, - "reason": "A generic testing_utils bug is not the same as a pytest-xdist race on captured_info.txt." + "reason": "A vague testing_utils bug is not enough to merge with a specific xdist race on captured_info.txt." }, { "left": "issue:42673", "right": "issue:43906", "accept": false, - "reason": "An isolated reproduction of another issue is not itself the same bug as a VRAM leak." + "reason": "VRAM leak in Qwen3ForCausalLM is unrelated to a reproduction of another issue." }, { - "left": "issue:45405", - "right": "issue:45446", + "left": "issue:44291", + "right": "issue:45071", "accept": false, - "reason": "A dependency minimum-version bump issue is unrelated to the AuxRequest flex_attention import guard." + "reason": "Unexpected _is_hf_initialized argument during init_empty_weights is unrelated to a PretrainedConfig type-checking breakage." } ] }, @@ -11792,8 +11774,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -11825,7 +11807,6 @@ "issue:44220", "issue:44261", "issue:44263", - "issue:44291", "issue:44410", "issue:44483", "issue:44484", @@ -11834,20 +11815,21 @@ "issue:44496", "issue:44829", "issue:44933", - "issue:45003", "issue:45071", "issue:45081", "issue:45216", "issue:45278", "issue:45310", "issue:45341", + "issue:45405", + "issue:45446", "issue:45459", "issue:45464", "issue:45478", "issue:45542" ], "soft_pairs": [ - "issue:44291|issue:45071", + "issue:45405|issue:45446", "issue:39401|issue:43421", "issue:43450|issue:43986", "issue:44484|issue:45310", @@ -11855,278 +11837,250 @@ "issue:38617|issue:41628", "issue:44410|issue:45310", "issue:44117|issue:45081", - "issue:42673|issue:43582", "issue:43408|issue:43742", + "issue:42673|issue:43582", "issue:42222|issue:45542", - "issue:42673|issue:43577", "issue:42371|issue:44261", + "issue:42673|issue:43577", "issue:43937|issue:44496", - "issue:42175|issue:42222", "issue:38617|issue:44220", "issue:43012|issue:43577", + "issue:42175|issue:42222", "issue:44483|issue:45464", "issue:44933|issue:45278", - "issue:42617|issue:45341", "issue:44484|issue:45478", + "issue:42617|issue:45341", "issue:43066|issue:43122", "issue:43673|issue:45341", "issue:44410|issue:45478", "issue:43097|issue:44493", - "issue:42222|issue:45459", "issue:43874|issue:44261", + "issue:42222|issue:45459", "issue:43441|issue:44829", - "issue:43012|issue:43408", "issue:43986|issue:44484", + "issue:43012|issue:43408", "issue:44263|issue:44485", "issue:41628|issue:45071", "issue:45071|issue:45542", - "issue:43931|issue:44410", - "issue:45003|issue:45341" + "issue:43931|issue:44410" ], - "prepared_review_unit_hash": "b13c14c8e13dac9fb9bd15434edc9d166d64339caf3f7fa3213104e35e601f71", + "prepared_review_unit_hash": "ac0bbe99f108a1aa6ec2458e234c56532bc2602ef102b24057fc7106ad4908c8", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 35, - "serialized_chars": 37973, - "estimated_input_tokens": 9494, - "estimated_eval_tokens": 19244 + "soft_pair_count": 34, + "serialized_chars": 37633, + "estimated_input_tokens": 9409, + "estimated_eval_tokens": 19074 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous: most items are distinct bug reports spanning tokenizers, model loading, multimodal processors, packaging/import regressions, and API behavior. I don\u2019t see a single underlying defect that cleanly subsumes the whole set.", - "confidence": 0.95, - "canonical_issue_reason": "No single issue cleanly represents the whole cluster; the items are mostly unrelated. If forced to pick a broad umbrella, #45278 is the closest because it describes many import errors after a version upgrade, but it does not actually cover most of the other reports.", + "summary": "This cluster is mostly a grab bag of unrelated issues. The only clearly supported duplicate set is the Qwen3.5 MoE `from_pretrained` regression between 45310 and 45478. Issue 44484 is about `save_pretrained()` / `max_shard_size` and is not, by title alone, enough to treat it as the same bug.", + "confidence": 0.82, + "canonical_issue_reason": "issue:45478 is the clearest representative of the Qwen3.5 MoE `from_pretrained` regression; it matches issue:45310 closely and has slightly stronger discussion signal. Issue:44484 does not have enough evidence to join that duplicate set.", "canonical_pr_reason": null, - "best_issue_reason": "#45278 is the broadest and most recent open issue in the set, so it is the best umbrella candidate for generic import-regression triage, though it is still not a true duplicate target for the rest of the cluster.", + "best_issue_reason": "issue:45478 is the best issue to keep as the representative of the only clearly duplicated bug in the cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44291", - "right": "issue:45071", + "left": "issue:45405", + "right": "issue:45446", "accept": false, - "reason": "Both are version-regression loader failures, but one is an init_empty_weights/_is_hf_initialized TypeError and the other is a PretrainedConfig type-checking break; different code paths." + "reason": "Both are version/dependency-related, but one is about PEFT pinning and the other is about an import/version check for AuxRequest; different bugs." }, { "left": "issue:39401", "right": "issue:43421", "accept": false, - "reason": "Both involve tokenizers, but one is wrong offset mapping and the other is runtime special-token/post-processor updates; not the same bug." + "reason": "Both concern tokenizer behavior, but one is wrong `offset_mapping` and the other is runtime special-token post-processor updates; not the same issue." }, { "left": "issue:43450", "right": "issue:43986", "accept": false, - "reason": "One is a batched video-processor shape bug; the other is an AutoProcessor crash without torchvision. Different failure modes." + "reason": "One is a video-processor batch shape bug; the other is a crash when `torchvision` is missing. Different code paths and symptoms." }, { "left": "issue:44484", "right": "issue:45310", "accept": false, - "reason": "Unrelated issues: max_shard_size behavior versus a Qwen3.5 MoE from_pretrained error." + "reason": "`44484` is about `save_pretrained()` / `max_shard_size`, while `45310` is a Qwen3.5 MoE `from_pretrained` error. The titles do not establish the same underlying bug." }, { "left": "issue:44829", "right": "issue:45216", "accept": false, - "reason": "One is FA3 training degeneration; the other is incorrect checkpoint saving for Qwen3.5. Different concrete problems." + "reason": "FlashAttention-3 training degeneration and Qwen3.5 checkpoint save corruption are distinct problems." }, { "left": "issue:38617", "right": "issue:41628", "accept": false, - "reason": "Both are import errors, but they concern different missing symbols and modules." + "reason": "Both are import errors, but for different symbols (`layer_type_validation` vs `AutoImageProcessor`)." }, { "left": "issue:44410", "right": "issue:45310", "accept": false, - "reason": "Different model bugs: missing Qwen3Next projections versus Qwen3.5 MoE loading failure." + "reason": "One is missing projections in qwen3next layer 0; the other is a Qwen3.5 MoE loading regression. Different bugs." }, { "left": "issue:44117", - "right": "issue:45081", - "accept": false, - "reason": "Tokenizer internals are involved in both, but one is TOKENIZER_MAPPING_NAMES returning None and the other is a Mistral regex patch crash." - }, - { - "left": "issue:42673", - "right": "issue:43582", + "right": "issue:45081", "accept": false, - "reason": "VRAM leak in Qwen3ForCausalLM vs Apple Silicon caching_allocator_warmup TypeError; unrelated." + "reason": "`TOKENIZER_MAPPING_NAMES` returning `None` and the Mistral regex patch crash are unrelated tokenizer internals." }, { "left": "issue:43408", "right": "issue:43742", "accept": false, - "reason": "Different model-loading issues: sam3 type mismatch versus MobileLLM key error." + "reason": "A model-type warning for sam3_tracker is not the same as a key error when loading MobileLLM." }, { - "left": "issue:42222", - "right": "issue:45542", + "left": "issue:42673", + "right": "issue:43582", "accept": false, - "reason": "Vitpose model breakage and a TensorFlow backend-install issue are unrelated." + "reason": "A VRAM leak in Qwen3ForCausalLM multi-threaded loading is unrelated to an Apple Silicon allocator `TypeError`." }, { - "left": "issue:42673", - "right": "issue:43577", + "left": "issue:42222", + "right": "issue:45542", "accept": false, - "reason": "Both mention memory/dtype concerns, but the VRAM leak and Blip2 dtype loading bug are different code paths." + "reason": "VitPose breakage and undefined TF backend errors are different dependency/model-loading problems." }, { "left": "issue:42371", "right": "issue:44261", "accept": false, - "reason": "TF32 API warning and MLA rms_norm_eps precision issue are unrelated." + "reason": "TF32 behavior/settings guidance is unrelated to the MLA `rms_norm_eps` precision bug." }, { - "left": "issue:43937", - "right": "issue:44496", + "left": "issue:42673", + "right": "issue:43577", "accept": false, - "reason": "GenerationConfig validation for GLM-5 versus missing model_type for an unrecognized model are different loader/config problems." + "reason": "The Qwen3 VRAM leak and BLIP2 dtype staying float32 are both about memory/precision, but not the same bug." }, { - "left": "issue:42175", - "right": "issue:42222", + "left": "issue:43937", + "right": "issue:44496", "accept": false, - "reason": "Package extra/backend dependency issue versus a vitpose model regression; no shared underlying bug." + "reason": "GLM-5 invalid `GenerationConfig` and OLMo-Hybrid unrecognized model/config are distinct loading/configuration failures." }, { "left": "issue:38617", "right": "issue:44220", "accept": false, - "reason": "ImportError for a missing config symbol is not the same as an audio feature-extraction failure." + "reason": "An import error from `configuration_utils` is not the same as the `_torch_extract_fbank_features()` issue." }, { "left": "issue:43012", "right": "issue:43577", "accept": false, - "reason": "Both touch precision/dtype, but one is a PyTorch warning on compilation and the other is incorrect dtype propagation while loading Blip2." + "reason": "Both involve dtype/precision, but one is a compile-time warning and the other is incorrect loading dtypes in BLIP2." + }, + { + "left": "issue:42175", + "right": "issue:42222", + "accept": false, + "reason": "Missing TensorFlow in the `[torch]` install and broken VitPose models are unrelated." }, { "left": "issue:44483", "right": "issue:45464", "accept": false, - "reason": "Both are chat/completions-related, but one is a general request rejection and the other is a streaming inference failure on a specific model." + "reason": "Both mention chat/completions, but one is a generic endpoint request failure and the other is a model-specific streaming inference problem." }, { "left": "issue:44933", "right": "issue:45278", "accept": false, - "reason": "The latter is a broad import-regression report, but this pair does not establish the same missing symbol or exact failure." + "reason": "A specific missing import from `image_utils` is not enough to equate it with the broader import-error report after upgrading to 5.5.0." }, { - "left": "issue:42617", - "right": "issue:45341", + "left": "issue:44484", + "right": "issue:45478", "accept": false, - "reason": "A 3d_parallel runtime issue and a testing_utils bug are unrelated." + "reason": "`44484` is about `save_pretrained()` / `max_shard_size`, while `45478` is a Qwen3.5 MoE `from_pretrained` error. The evidence does not support treating them as the same bug." }, { - "left": "issue:44484", - "right": "issue:45478", + "left": "issue:42617", + "right": "issue:45341", "accept": false, - "reason": "Different issues: shard-size behavior versus another Qwen3.5 MoE from_pretrained error report." + "reason": "`3d_parallel.py` execution failure and a `testing_utils.py` bug are unrelated." }, { "left": "issue:43066", "right": "issue:43122", "accept": false, - "reason": "Both concern tokenizer behavior across v5, but one is decoder type and the other is tokenization output changes; not the same defect." + "reason": "Both are tokenizer regression reports, but one targets decoder-type metadata and the other reports changed tokenization output; not clearly the same concrete bug." }, { "left": "issue:43673", "right": "issue:45341", "accept": false, - "reason": "Generation cache missing in chunked_prefill is unrelated to a testing utility bug." + "reason": "Missing generation cache in chunked prefill and a testing utility bug do not share the same code path." }, { "left": "issue:44410", "right": "issue:45478", "accept": false, - "reason": "Different Qwen3-related loader/model bugs, not the same concrete failure." + "reason": "qwen3next missing attention projections is unrelated to the Qwen3.5 MoE loading regression." }, { "left": "issue:43097", "right": "issue:44493", "accept": false, - "reason": "One is a removed config flag in v5; the other is unexpected position-id keys. Related theme, but not the same bug." - }, - { - "left": "issue:42222", - "right": "issue:45459", - "accept": false, - "reason": "Vitpose model breakage and tokenizer error hiding when protobuf is absent are unrelated." + "reason": "Removal of `tie_embeddings_and_encoder_decoder` and unexpected `position_id` keys are different regressions." }, { "left": "issue:43874", "right": "issue:44261", "accept": false, - "reason": "Missing image-patch helper in GLM46V versus MLA layernorm epsilon precision issue; different subsystems." - }, - { - "left": "issue:43441", - "right": "issue:44829", - "accept": false, - "reason": "Ministral FlashAttention failure and FA3 degenerate training are different model/attention problems." + "reason": "A missing image-patch helper and an `rms_norm_eps` precision mismatch are unrelated bugs." }, { - "left": "issue:43012", - "right": "issue:43408", - "accept": false, - "reason": "Precision-warning behavior and a sam3_video/sam3_tracker type warning are unrelated." - }, - { - "left": "issue:43986", - "right": "issue:44484", + "left": "issue:41628", + "right": "issue:45071", "accept": false, - "reason": "Video processor import/runtime crash versus save_pretrained shard sizing; not the same issue." + "reason": "A missing top-level import and a `PreTrainedConfig` type-checking breakage are different failures." }, { "left": "issue:44263", "right": "issue:44485", "accept": false, - "reason": "torch.split return handling in GlmMoeDsaIndexer is unrelated to GLM-5 RoPE implementation." - }, - { - "left": "issue:41628", - "right": "issue:45071", - "accept": false, - "reason": "AutoImageProcessor import failure and PretrainedConfig type-checking break are different regressions." + "reason": "Both are GLM-related, but one concerns `torch.split` handling and the other RoPE implementation; not the same issue." }, { "left": "issue:45071", "right": "issue:45542", "accept": false, - "reason": "PretrainedConfig type checking and TensorFlow backend installation are unrelated." + "reason": "`PreTrainedConfig` type checking and undefined TF backend errors are unrelated." }, { "left": "issue:43931", "right": "issue:44410", "accept": false, - "reason": "Qwen3-VL weight-shape mismatch and qwen3next missing projections are different model-loading problems." - }, - { - "left": "issue:45003", - "right": "issue:45341", - "accept": false, - "reason": "Unsafe sys.modules access in modeling_utils is not the same as a testing_utils bug." + "reason": "Qwen3-VL weight-shape mismatch and qwen3next missing projections are different model-loading bugs." } ] }, - "evaluator_result": null, - "evaluator_used": false, - "retried": false, + "evaluator_result": { + "accept": false, + "feedback": "The cluster summary is too strong about the Qwen3.5 MoE duplicate set. In the packet, 45310 and 45478 clearly match on `Qwen3.5 Moe from_pretrained error`, but 44484\u2019s title is about `save_pretrained()` / `max_shard_size` and does not, by itself, ground the claim that it is the same bug. The accepted soft-edge verdicts involving 44484 therefore overstate the evidence. The rest of the rejected pairs look conservative, but I can\u2019t accept the cluster with that unsupported equivalence." + }, + "evaluator_used": true, + "retried": true, "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null @@ -12139,15 +12093,14 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ "issue:38617", "issue:42757", "issue:42886", - "issue:42907", "issue:43012", "issue:43064", "issue:43066", @@ -12185,6 +12138,7 @@ "issue:44704", "issue:44829", "issue:44987", + "issue:45003", "issue:45071", "issue:45103", "issue:45245", @@ -12194,6 +12148,7 @@ "issue:45588" ], "soft_pairs": [ + "issue:45003|issue:45341", "issue:44117|issue:44987", "issue:43986|issue:45341", "issue:43097|issue:43452", @@ -12209,251 +12164,256 @@ "issue:38617|issue:42757", "issue:44261|issue:44485", "issue:44410|issue:45375", - "issue:44038|issue:44410", "issue:38617|issue:44351", + "issue:44038|issue:44410", "issue:44062|issue:44117", "issue:43097|issue:45588", "issue:43986|issue:44117", "issue:44704|issue:45103", "issue:43502|issue:43992", - "issue:43097|issue:43873", "issue:43992|issue:44484", "issue:43824|issue:43986", + "issue:43097|issue:43873", "issue:43066|issue:43906", "issue:43352|issue:44380", "issue:43825|issue:44655", "issue:38617|issue:43097", "issue:43012|issue:44368", - "issue:43012|issue:43550", "issue:42886|issue:43381", - "issue:42907|issue:44655" + "issue:43012|issue:43550" ], - "prepared_review_unit_hash": "4200dc2148cffb64509de27e2126b29db432f7f4a2d0bcec79102b0fa49b932d", + "prepared_review_unit_hash": "617e895a6f98e5fe2713f014fb05f78b1ef8463d1823f84152e4298749265309", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 33, - "serialized_chars": 37583, - "estimated_input_tokens": 9396, - "estimated_eval_tokens": 19048 + "serialized_chars": 37573, + "estimated_input_tokens": 9394, + "estimated_eval_tokens": 19044 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: it mixes unrelated Transformers v5 regressions across imports, tokenizer/model loading, multimodal processors, training, and attention backends. None of the proposed soft pairs look like the same underlying bug/change, so no soft edge should be merged.", - "confidence": 0.93, - "canonical_issue_reason": "No single canonical duplicate: the items span many distinct failures and API regressions, so there is no clear one-bug anchor for the whole cluster.", + "summary": "The cluster is mostly a noisy mix of unrelated Transformers v5 regressions and model-specific bugs. I do not see any soft pair that is the same underlying bug/change, so all candidate edges are rejected. The closest representative issue is the v5 breaking-change report about `tie_embeddings_and_encoder_decoder` being removed.", + "confidence": 0.66, + "canonical_issue_reason": "Issue 43097 is the most cluster-like representative: it describes a concrete, cross-cutting v5 API/config break with inbound references, while most other issues are narrower model- or dependency-specific failures.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43097 is the strongest representative because it captures a broad v5 breaking-change regression with clear API impact, but it is still not a duplicate target for the rest of the cluster.", + "best_issue_reason": "43097 is the best anchor for this set because it captures a broad Transformers v5 regression that plausibly sits near the center of the cluster\u2019s general theme of breaking changes.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:45003", + "right": "issue:45341", + "accept": false, + "reason": "Different bug classes: unsafe `sys.modules` access vs a testing utility issue." + }, { "left": "issue:44117", "right": "issue:44987", "accept": false, - "reason": "Tokenizer mapping None vs a specific model-loading failure for physical-intelligence/fast; different code paths and symptoms." + "reason": "Tokenizer mapping fallback bug vs a model-loading failure for `physical-intelligence/fast`." }, { "left": "issue:43986", "right": "issue:45341", "accept": false, - "reason": "Video-model loading without torchvision vs a small testing_utils bug; unrelated components." + "reason": "AutoProcessor/video-model crash without torchvision vs unrelated testing utility bug." }, { "left": "issue:43097", "right": "issue:43452", "accept": false, - "reason": "Removed config flag in v5 vs gguf_file handling for AutoTokenizer/AutoModel loading; not the same bug." + "reason": "Removed config flag warning vs `gguf_file` loading breakage; different code paths." }, { "left": "issue:43931", "right": "issue:44038", "accept": false, - "reason": "Different Qwen3-VL model failures: weight-shape mismatch vs a Qwen3-VL-Moe bug." + "reason": "Weight-shape mismatch for one model vs a separate Qwen3-VL-Moe loading problem." }, { "left": "issue:43335", "right": "issue:45071", "accept": false, - "reason": "SwitchTransformers sparse-layer construction bug vs PretrainedConfig type-checking regression; unrelated." + "reason": "SwitchTransformers layer creation bug vs PretrainedConfig type-checking regression." }, { "left": "issue:43408", "right": "issue:44368", "accept": false, - "reason": "SAM3 video-type warning vs a Qwen3.5 tie_word_embeddings warning; different models and failure modes." + "reason": "Model-type warning mismatch vs tie_word_embeddings warning; not the same failure." }, { "left": "issue:43874", "right": "issue:44263", "accept": false, - "reason": "Missing image-processor method in GLM46V vs torch.split return handling in GlmMoeDsaIndexer; unrelated." + "reason": "Missing image-patch method vs incorrect `torch.split()` return handling." }, { "left": "issue:43976", "right": "issue:45071", "accept": false, - "reason": "Python-version compatibility problem vs config type checking; not the same underlying issue." + "reason": "Python version compatibility issue vs config type-checking failure." }, { "left": "issue:44829", "right": "issue:45245", "accept": false, - "reason": "Degenerate training with flash_attention_3 vs categorical cardinality limit error; completely different problems." + "reason": "FA3 training degeneration vs category-cardinality runtime limit; unrelated." }, { "left": "issue:43986", "right": "issue:45372", "accept": false, - "reason": "Missing torchvision for video processor loading vs mistral_common ReasoningEffort import failure; unrelated." + "reason": "Missing torchvision dependency vs `mistral_common` import failure; different models/dependencies." }, { "left": "issue:43064", "right": "issue:44315", "accept": false, - "reason": "FSDP2/PEFT optimizer-state corruption vs Liger Kernel not applied with model_init; different training paths." + "reason": "FSDP2 optimizer-state corruption vs Liger Kernel not being applied." }, { "left": "issue:43531", "right": "issue:45341", "accept": false, - "reason": "Qwen3-MoE sliding_window bug vs a testing_utils issue; no shared bug." + "reason": "Qwen3-MoE `sliding_window` behavior vs a testing utility bug." }, { "left": "issue:38617", "right": "issue:42757", "accept": false, - "reason": "Both are import errors, but for different symbols/modules and unrelated dependency paths." + "reason": "Both are import errors, but for different missing names and different modules." }, { "left": "issue:44261", "right": "issue:44485", "accept": false, - "reason": "MLA q_a_layernorm epsilon/config precision issue vs GLM-5 RoPE implementation; distinct model internals." + "reason": "RMS norm epsilon precision issue vs GLM-5 RoPE implementation." }, { "left": "issue:44410", "right": "issue:45375", "accept": false, - "reason": "Missing projections in qwen3next vs missing deepstack_visual_indexes in Qwen3_5MoeVisionConfig; different configs and symptoms." + "reason": "Missing attention projections vs missing config field in a different model family." }, { - "left": "issue:44038", - "right": "issue:44410", + "left": "issue:38617", + "right": "issue:44351", "accept": false, - "reason": "Qwen3-VL-Moe issue vs qwen3next projection bug; different architectures and failures." + "reason": "Different import failures: `layer_type_validation` vs `HybridCache`." }, { - "left": "issue:38617", - "right": "issue:44351", + "left": "issue:44038", + "right": "issue:44410", "accept": false, - "reason": "layer_type_validation import error vs HybridCache import error; separate missing exports." + "reason": "Qwen3-VL-Moe loading bug vs qwen3next missing projections; not the same underlying defect." }, { "left": "issue:44062", "right": "issue:44117", "accept": false, - "reason": "AddedToken special-argument conflict vs TOKENIZER_MAPPING_NAMES returning None; different tokenizer code paths." + "reason": "`AddedToken` argument duplication vs tokenizer mapping returning None." }, { "left": "issue:43097", "right": "issue:45588", "accept": false, - "reason": "v5 config removal vs flash_attention sink-less crash; unrelated subsystems." + "reason": "Config-removal warning vs flash_attention crash on `s_aux=None`; different bugs." }, { "left": "issue:43986", "right": "issue:44117", "accept": false, - "reason": "Video processor loading without torchvision vs tokenizer mapping None; unrelated." + "reason": "Video AutoProcessor dependency crash vs tokenizer mapping fallback bug." }, { "left": "issue:44704", "right": "issue:45103", "accept": false, - "reason": "AutoProcessor kwargs not forwarded to cached_file vs auto_docstring crash with future annotations; unrelated utility layers." + "reason": "Kwargs not forwarded to `cached_file` vs docstring/annotation-related AttributeError." }, { "left": "issue:43502", "right": "issue:43992", "accept": false, - "reason": "local_files_only still causing network requests vs missing embed_tokens.weight in UMT5Encoder; different loading bugs." - }, - { - "left": "issue:43097", - "right": "issue:43873", - "accept": false, - "reason": "API removal in v5 vs offloading/quantization behavior; not the same issue." + "reason": "Offline/network leakage bug vs missing `embed_tokens.weight` during loading." }, { "left": "issue:43992", "right": "issue:44484", "accept": false, - "reason": "Missing checkpoint weight in UMT5Encoder vs max_shard_size question; unrelated save/load behavior." + "reason": "Missing state-dict key vs `max_shard_size` default question." }, { "left": "issue:43824", "right": "issue:43986", "accept": false, - "reason": "Qwen2_5_VL import error vs video-model loading crash without torchvision; different models and root causes." + "reason": "Missing model symbol import vs video processor crash without torchvision." + }, + { + "left": "issue:43097", + "right": "issue:43873", + "accept": false, + "reason": "Removed config attribute warning vs quantization offloading behavior." }, { "left": "issue:43066", "right": "issue:43906", "accept": false, - "reason": "Wrong tokenizer decoder type vs an isolated reproduction of a separate issue; no concrete shared bug." + "reason": "Wrong tokenizer decoder type vs a reproduction of a different issue." }, { "left": "issue:43352", "right": "issue:44380", "accept": false, - "reason": "Flash Attention 2 unsupported gate vs GPT2 attention-scaling ignored under SDPA/FlashAttention; different backend problems." + "reason": "Flash Attention support gating vs GPT2 attention scaling being ignored." }, { "left": "issue:43825", "right": "issue:44655", "accept": false, - "reason": "Pipeline translation-task error vs inability to save Pipeline objects; different pipeline concerns." + "reason": "Pipeline translation-task messaging vs inability to save Pipeline objects." }, { "left": "issue:38617", "right": "issue:43097", "accept": false, - "reason": "ImportError for layer_type_validation vs removed tie_embeddings_and_encoder_decoder; unrelated regressions." + "reason": "ImportError for a missing symbol vs a removed config field warning." }, { - "left": "issue:43012", - "right": "issue:44368", + "left": "issue:44261", + "right": "issue:44485", "accept": false, - "reason": "bfloat16 compile warning vs tie_word_embeddings warning during Qwen3.5 LoRA; same general warning category, but not the same bug." + "reason": "Precision mismatch in MLA q_a_layernorm vs GLM-5 RoPE implementation." }, { "left": "issue:43012", - "right": "issue:43550", + "right": "issue:44368", "accept": false, - "reason": "PyTorch precision warning during compile vs Bamba SDPA compile failure; different failure modes." + "reason": "bfloat16 compilation warning vs tie_word_embeddings warning." }, { "left": "issue:42886", "right": "issue:43381", "accept": false, - "reason": "Offline tokenizer cache loading vs gradient checkpointing in eval mode; unrelated." + "reason": "Offline cache-loading failure vs gradient checkpointing being disallowed in eval mode." }, { - "left": "issue:42907", - "right": "issue:44655", + "left": "issue:43012", + "right": "issue:43550", "accept": false, - "reason": "Failed save of dequantized Ministrals vs saving Pipeline objects; both involve save_pretrained, but distinct object types and causes." + "reason": "PyTorch precision warning vs Bamba-9B-v2 compile/SDPA failure." } ] }, @@ -12472,8 +12432,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -12481,6 +12441,7 @@ "issue:36683", "issue:38617", "issue:42757", + "issue:42907", "issue:42994", "issue:43012", "issue:43064", @@ -12506,8 +12467,8 @@ "issue:44038", "issue:44077", "issue:44188", + "issue:44220", "issue:44230", - "issue:44242", "issue:44263", "issue:44380", "issue:44483", @@ -12522,35 +12483,35 @@ "issue:45310", "issue:45381", "issue:45440", - "issue:45478", "issue:45479", "issue:45593" ], "soft_pairs": [ + "issue:42907|issue:44655", "issue:43208|issue:43550", "issue:44380|issue:45479", "issue:43352|issue:43976", "issue:43064|issue:44380", - "issue:43208|issue:43526", "issue:44483|issue:44514", + "issue:43208|issue:43526", "issue:43618|issue:44380", "issue:36683|issue:43582", "issue:43618|issue:43761", - "issue:43012|issue:44683", "issue:44483|issue:45290", - "issue:44038|issue:44230", + "issue:43012|issue:44683", "issue:42757|issue:43824", + "issue:44038|issue:44230", "issue:44483|issue:45381", "issue:43299|issue:43867", "issue:38617|issue:43502", - "issue:43901|issue:44655", "issue:43756|issue:44485", + "issue:43901|issue:44655", "issue:42994|issue:44655", "issue:43867|issue:45216", "issue:36683|issue:43475", - "issue:44655|issue:45478", - "issue:43352|issue:43937", + "issue:44220|issue:44655", "issue:43066|issue:45381", + "issue:43352|issue:43937", "issue:44077|issue:44380", "issue:43064|issue:44492", "issue:44263|issue:45440", @@ -12561,247 +12522,252 @@ "issue:44188|issue:44380", "issue:43618|issue:45216", "issue:43352|issue:45593", - "issue:43992|issue:45310", - "issue:36683|issue:44242" + "issue:43992|issue:45310" ], - "prepared_review_unit_hash": "868e9c0ac85c00d2176448438500caf85a0fa5280a7fb9b3d4a3ae0bfa1c11a8", + "prepared_review_unit_hash": "189834f88c865a0bd70f83c6075a362629939539de30f3be8e7505f5a68fc38c", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 36, - "serialized_chars": 38123, - "estimated_input_tokens": 9531, - "estimated_eval_tokens": 19318 + "serialized_chars": 38094, + "estimated_input_tokens": 9524, + "estimated_eval_tokens": 19304 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly unrelated Transformers issues; the only clear duplicate pair is the Qwen3.5 MoE from_pretrained regression. No pull requests were present.", - "confidence": 0.89, - "canonical_issue_reason": "issue:45310 is the earlier, more specific report of the shared Qwen3.5 MoE from_pretrained failure, and issue:45478 reads as a later duplicate of the same version-regression.", + "summary": "This set is mostly heterogeneous: many title-level similarities (imports, saving, Qwen/CLIP/model regressions) but the underlying bugs and code paths differ. I would not merge any soft pair as a duplicate.", + "confidence": 0.88, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:45310 is the best cluster representative because it captures the concrete bug shape and version scope most clearly, with a near-identical follow-up report duplicating it.", + "best_issue_reason": "issue:43502 is the strongest standalone representative here: concrete, reproducible API regression, broad user impact, and more discussion/inbound references than most others.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:42907", + "right": "issue:44655", + "accept": false, + "reason": "Both involve saving, but one is dequantized model saving and the other is Pipeline.save_pretrained; different failure paths." + }, { "left": "issue:43208", "right": "issue:43550", "accept": false, - "reason": "Different models and failure modes: xLSTM training bugs vs Bamba torch.compile/SDPA issue." + "reason": "xLSTM training bugs vs Bamba torch.compile/SDPA regression; different models and code paths." }, { "left": "issue:44380", "right": "issue:45479", "accept": false, - "reason": "Both are model bugs, but one is GPT-2 attention scaling and the other is a sequence-classification zero-loss bug; different code paths." + "reason": "Attention scaling under SDPA/FlashAttention is unrelated to single-label classification zero-loss behavior." }, { "left": "issue:43352", "right": "issue:43976", "accept": false, - "reason": "Unrelated problems: Flash Attention 2 support for Nemotron vs Python-version compatibility." + "reason": "Nemotron Flash Attention support and Python/version compatibility are unrelated issues." }, { "left": "issue:43064", "right": "issue:44380", "accept": false, - "reason": "FSDP2 optimizer-state corruption is unrelated to GPT-2 attention scaling under SDPA/FlashAttention." + "reason": "FSDP/PEFT optimizer-state bug is unrelated to GPT2 attention scaling under backend kernels." }, { - "left": "issue:43208", - "right": "issue:43526", + "left": "issue:44483", + "right": "issue:44514", "accept": false, - "reason": "xLSTM training bugs and BeitImageProcessorFast label reduction are different subsystems and symptoms." + "reason": "Chat-completions request rejection and batched chat template padding crash are different layers of the stack." }, { - "left": "issue:44483", - "right": "issue:44514", + "left": "issue:43208", + "right": "issue:43526", "accept": false, - "reason": "Both touch chat-related APIs, but one is request acceptance and the other is batched chat template padding behavior." + "reason": "xLSTM training bugs vs BeitImageProcessorFast label reduction bug; no shared underlying defect." }, { "left": "issue:43618", "right": "issue:44380", "accept": false, - "reason": "CLIP output attentions regression is not the same bug as GPT-2 attention scaling being ignored." + "reason": "CLIP attentions assignment regression is separate from GPT2 attention scaling config being ignored." }, { "left": "issue:36683", "right": "issue:43582", "accept": false, - "reason": "Gemma3Config missing vocab_size and Apple Silicon allocator TypeError are unrelated." + "reason": "Gemma3 vocab_size attribute error and AppleSilicon TypeError in caching_allocator_warmup are unrelated." }, { "left": "issue:43618", "right": "issue:43761", "accept": false, - "reason": "Both involve CLIP regressions, but one is missing attentions assignment and the other is hidden_states=None; not the same defect." - }, - { - "left": "issue:43012", - "right": "issue:44683", - "accept": false, - "reason": "A warning about bf16 precision is distinct from a compiled flex_attention failure on torch>=2.9." + "reason": "Both are CLIP regressions, but one drops attentions and the other drops hidden_states; not the same bug." }, { "left": "issue:44483", "right": "issue:45290", "accept": false, - "reason": "Different chat-template/API failures: request validation vs assistant tool-call content handling." + "reason": "Server-side request validation issue vs apply_chat_template crash on tool-call assistant messages." }, { - "left": "issue:44038", - "right": "issue:44230", + "left": "issue:43012", + "right": "issue:44683", "accept": false, - "reason": "Both mention Qwen VL/MoE, but one is broad v5 loading support and the other is fp8 inference support; different requests." + "reason": "PyTorch warning on bfloat16 compile and flex_attention failure on torch>=2.9 are distinct compiler/backend problems." }, { "left": "issue:42757", "right": "issue:43824", "accept": false, - "reason": "Both are import errors, but for different missing symbols and likely different fixes." + "reason": "Different missing imports from different modules; same symptom, different symbols and root causes." + }, + { + "left": "issue:44038", + "right": "issue:44230", + "accept": false, + "reason": "Qwen3-VL-Moe loading regression vs fp8 inference support request; not the same change or bug." }, { "left": "issue:44483", "right": "issue:45381", "accept": false, - "reason": "Chat/completions request handling is unrelated to Qwen2.5-VL vision_position_ids generation." + "reason": "Chat/completions request handling and Qwen2.5-VL video vision_position_ids are unrelated." }, { "left": "issue:43299", "right": "issue:43867", "accept": false, - "reason": "Qwen3VL-Moe loading breakage and sorted-state_dict load error are not the same underlying issue." + "reason": "Qwen3VL-Moe loading regression and state_dict-sorted load error differ in model path and failure mode." }, { "left": "issue:38617", "right": "issue:43502", "accept": false, - "reason": "ImportError from configuration_utils and local_files_only network leakage are different bug classes." + "reason": "Missing configuration_utils import and local_files_only still making network calls are unrelated." }, { - "left": "issue:43901", - "right": "issue:44655", + "left": "issue:43756", + "right": "issue:44485", "accept": false, - "reason": "Documentation mismatch for return_all_scores is unrelated to saving Pipeline objects." + "reason": "Smollm3 RoPE-layer mismatch is unrelated to GLM-5 RoPE implementation discussion." }, { - "left": "issue:43756", - "right": "issue:44485", + "left": "issue:43901", + "right": "issue:44655", "accept": false, - "reason": "Smollm3 RoPE layer dropout and GLM-5 RoPE implementation discussion are different model-specific issues." + "reason": "Documentation mismatch for return_all_scores and Pipeline save_pretrained failure are different concerns." }, { "left": "issue:42994", "right": "issue:44655", "accept": false, - "reason": "Quantized model saving failure is not the same as Pipeline save_pretrained serialization failure." + "reason": "Quantized model save failure is not the same as saving Pipeline objects." }, { "left": "issue:43867", "right": "issue:45216", "accept": false, - "reason": "Both concern loading/saving, but one is sorted state_dict loading and the other is Qwen3.5 checkpoint correctness on save_pretrained." + "reason": "State_dict ordering load error and Qwen3.5 checkpoint corruption after save_pretrained are different bugs." }, { "left": "issue:36683", "right": "issue:43475", "accept": false, - "reason": "Gemma3Config missing vocab_size and Sam3VisionEncoderOutput missing fpn_position_embeddings are different attribute errors." + "reason": "Gemma3 vocab_size attribute error and SAM3 video missing fpn_position_embeddings are unrelated model config issues." }, { - "left": "issue:44655", - "right": "issue:45478", + "left": "issue:44220", + "right": "issue:44655", "accept": false, - "reason": "Pipeline serialization and Qwen3.5 MoE from_pretrained are unrelated." + "reason": "_torch_extract_fbank_features bug and Pipeline serialization bug are different subsystems." }, { - "left": "issue:43352", - "right": "issue:43937", + "left": "issue:43066", + "right": "issue:45381", "accept": false, - "reason": "Flash Attention support for Nemotron and GLM-5 GenerationConfig validation are different issues." + "reason": "Tokenizer decoder-type regression and Qwen2.5-VL video position-id bug are unrelated." }, { - "left": "issue:43066", - "right": "issue:45381", + "left": "issue:43352", + "right": "issue:43937", "accept": false, - "reason": "Tokenizer decoder type mismatch is unrelated to Qwen2.5-VL video position IDs." + "reason": "FlashAttention support gating for Nemotron and GLM-5 GenerationConfig validation are different issues." }, { "left": "issue:44077", "right": "issue:44380", "accept": false, - "reason": "A patchtsmixer post_init policy issue is unrelated to GPT-2 attention scaling under SDPA/FlashAttention." + "reason": "patchtsmixer post_init allowance and GPT2 attention scaling under SDPA/FlashAttention are unrelated." }, { "left": "issue:43064", "right": "issue:44492", "accept": false, - "reason": "Optimizer-state corruption and a typo in cache strategies are not the same bug." + "reason": "Distributed optimizer-state bug and a cache-strategy typo are not the same underlying problem." }, { "left": "issue:44263", "right": "issue:45440", "accept": false, - "reason": "Different MoE-related code paths and models; not the same concrete defect." + "reason": "torch.split return handling in GLM-Moe indexer is unrelated to DeepseekV3MoE implementation divergence." }, { "left": "issue:43976", "right": "issue:44938", "accept": false, - "reason": "Both are environment compatibility reports, but for different Python versions and failure causes." + "reason": "Both are environment-compatibility reports, but Python 3.9/3.10 and Python 3.14 failures have different causes." }, { - "left": "issue:43618", - "right": "issue:45216", + "left": "issue:36683", + "right": "issue:44077", "accept": false, - "reason": "CLIP attentions regression and Qwen3.5 save_pretrained corruption are unrelated." + "reason": "Gemma3 config attribute error is unrelated to patchtsmixer post_init policy." }, { - "left": "issue:43352", - "right": "issue:45593", + "left": "issue:36683", + "right": "issue:43525", "accept": false, - "reason": "Flash Attention support for Nemotron and D-FINE auxiliary-loss logic are different subsystems." + "reason": "Gemma3 vocab_size missing and Llama4 pad_token_id missing are different config attribute regressions." }, { - "left": "issue:43992", - "right": "issue:45310", + "left": "issue:36010", + "right": "issue:42757", "accept": false, - "reason": "UMT5 missing embed_tokens.weight and Qwen3.5 MoE from_pretrained are unrelated model-loading bugs." + "reason": "Different import failures from different modules and packages; not the same missing symbol." }, { - "left": "issue:36683", - "right": "issue:44242", + "left": "issue:44188", + "right": "issue:44380", "accept": false, - "reason": "Gemma3Config attribute error and missing load-balancing loss are different problems." + "reason": "Torch.compile branching divergence in attention kernels is not the same as GPT2 attention scaling config being ignored." }, { - "left": "issue:45310", - "right": "issue:45478", - "accept": true, - "reason": "Same concrete Qwen3.5 MoE from_pretrained failure with the same version regression; these look like duplicates that one fix could resolve together." + "left": "issue:43618", + "right": "issue:45216", + "accept": false, + "reason": "CLIP attentions regression and Qwen3.5 checkpoint save corruption are unrelated." }, { - "left": "issue:44188", - "right": "issue:44380", + "left": "issue:43352", + "right": "issue:45593", "accept": false, - "reason": "Both involve attention/compile behavior, but one is divergent kernels from branching under torch.compile and the other is scaling config being ignored in backend selection." + "reason": "FlashAttention support for Nemotron and D-FINE auxiliary-loss behavior are different model bugs." }, { - "left": "issue:36010", - "right": "issue:42757", + "left": "issue:43992", + "right": "issue:45310", "accept": false, - "reason": "Both are import errors, but for different exports from different modules and no shared concrete bug path." + "reason": "Missing embed_tokens.weight in UMT5Encoder and Qwen3.5 Moe from_pretrained error are different load paths and failures." } ] }, @@ -12820,8 +12786,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -12849,13 +12815,13 @@ "issue:44077", "issue:44112", "issue:44222", + "issue:44242", "issue:44246", "issue:44351", "issue:44355", "issue:44360", "issue:44387", "issue:44484", - "issue:44485", "issue:44556", "issue:44749", "issue:44857", @@ -12875,13 +12841,14 @@ "issue:45600" ], "soft_pairs": [ + "issue:36683|issue:44242", "issue:43010|issue:45278", "issue:36683|issue:43352", "issue:41669|issue:44246", "issue:43867|issue:45375", "issue:36683|issue:43064", - "issue:42890|issue:45561", "issue:43992|issue:45478", + "issue:42890|issue:45561", "issue:44938|issue:45071", "issue:43867|issue:45357", "issue:43010|issue:43976", @@ -12899,255 +12866,254 @@ "issue:43010|issue:44075", "issue:43502|issue:44062", "issue:43723|issue:44556", - "issue:42890|issue:45341", "issue:44222|issue:45125", + "issue:42890|issue:45341", "issue:44387|issue:44556", "issue:43010|issue:43065", "issue:43010|issue:43232", "issue:42548|issue:44351", "issue:45125|issue:45478", "issue:44355|issue:44749", - "issue:45125|issue:45375", - "issue:43502|issue:44485" + "issue:45125|issue:45375" ], - "prepared_review_unit_hash": "bc45c234b685a155e7aad7879932f23317729e6cea89d949a042066af2ea63f3", + "prepared_review_unit_hash": "680a4d8cfe5f60e8baff9bffaea25f07d7d0f8dc3eb86667a4eb383ca4f1736f", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 34, - "serialized_chars": 37341, - "estimated_input_tokens": 9336, - "estimated_eval_tokens": 18928 + "serialized_chars": 37356, + "estimated_input_tokens": 9339, + "estimated_eval_tokens": 18934 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is mostly a grab bag of unrelated Transformers issues. The only clear soft duplicate is the import-time slowdown pair; most other pairs share a subsystem or release version but not the same concrete bug.", - "confidence": 0.6, - "canonical_issue_reason": "issue:41669 is the most concrete root-cause report for the import slowdown; issue:44246 is the vaguer symptom report and fits as the duplicate.", + "summary": "This cluster is mostly heterogeneous and centered on unrelated loading/import/model-specific regressions. Only the import-time slowdown pair looks like a true duplicate; the rest are better treated as distinct bugs.", + "confidence": 0.71, + "canonical_issue_reason": "issue:45278 is the best umbrella issue: it captures the broad post-upgrade import-breakage theme that most closely represents this mixed cluster, whereas the other items are narrower model- or test-specific regressions.", "canonical_pr_reason": null, - "best_issue_reason": "issue:41669 is the best representative because it identifies the specific cause and is actionable, unlike the broader symptom-only report.", + "best_issue_reason": "issue:45278 is the strongest global representative because it is broad, user-facing, and aligned with the dominant import/load regression theme across the set.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:36683", + "right": "issue:44242", + "accept": false, + "reason": "Gemma3Config vocab_size missing is a model-config bug; load-balancing loss omission is a different MoE training issue." + }, { "left": "issue:43010", "right": "issue:45278", "accept": false, - "reason": "Both touch import/generation ergonomics, but one is a cache no_grad concern and the other is a broad import-regression report; different bugs." + "reason": "Cache update/no_grad behavior is unrelated to broad import errors after upgrading." }, { "left": "issue:36683", "right": "issue:43352", "accept": false, - "reason": "Gemma3 config vocab_size and Nemotron flash-attn support are unrelated model-specific failures." + "reason": "Different subsystems and symptoms: Gemma3Config attribute error vs Flash Attention 2.0 support gate." }, { "left": "issue:41669", "right": "issue:44246", "accept": true, - "reason": "Both describe the same import-time slowdown; 41669 gives the concrete cause (model import *) and 44246 reports the symptom." + "reason": "Both describe the same import-time slowdown regression; 41669 identifies the `import *` cause and 44246 reports the user-visible import slowness." }, { "left": "issue:43867", "right": "issue:45375", "accept": false, - "reason": "One is a state_dict loading error, the other is a Qwen3.5 vision config field being dropped; not the same code path." + "reason": "State-dict loading failure and a missing config field are different bugs in different code paths." }, { "left": "issue:36683", "right": "issue:43064", "accept": false, - "reason": "Different failures: missing Gemma3Config attribute vs FSDP2/PEFT optimizer-state mismatch." + "reason": "Gemma3Config missing vocab_size is unrelated to the FSDP2/PEFT optimizer-state issue." }, { - "left": "issue:42890", - "right": "issue:45561", + "left": "issue:43992", + "right": "issue:45478", "accept": false, - "reason": "Both are test-related, but one is missing set_seed coverage and the other is an xdist file race; distinct bugs." + "reason": "Both concern model loading, but they affect different models and likely different root causes." }, { - "left": "issue:43992", - "right": "issue:45478", + "left": "issue:42890", + "right": "issue:45561", "accept": false, - "reason": "Different Qwen/UMT5 load failures with different symptoms and likely different causes." + "reason": "One is flaky test seeding, the other is an xdist file-race in testing utils; not the same bug." }, { "left": "issue:44938", "right": "issue:45071", "accept": false, - "reason": "Both are version-related regressions, but one is Python 3.14 loading and the other is PretrainedConfig type checking." + "reason": "Python 3.14 import failure and PretrainedConfig type-checking regression are distinct compatibility issues." }, { "left": "issue:43867", "right": "issue:45357", "accept": false, - "reason": "A state_dict sorting/load issue is not the same as incorrect visual encoder keys in save_pretrained." + "reason": "A state_dict sort/load issue is not the same as incorrect visual encoder keys in save_pretrained." }, { "left": "issue:43010", "right": "issue:43976", "accept": false, - "reason": "No-no_grad cache update and Python-version support failure are unrelated." + "reason": "Cache update decoration behavior is unrelated to Python-version support in Transformers 5.1.0." }, { "left": "issue:44938", "right": "issue:45381", "accept": false, - "reason": "Python 3.14 import/load failure and Qwen2.5-VL video position ids are different problems." + "reason": "Python 3.14 load failure and Qwen2.5-VL vision_position_ids are unrelated." }, { "left": "issue:44360", "right": "issue:45440", "accept": false, - "reason": "DSA indexer/ReLU discussion and DeepSeekV3MoE divergence are unrelated model implementation issues." + "reason": "DSA indexer activation bug and DeepSeekV3MoE divergence are different implementation issues." }, { "left": "issue:43618", "right": "issue:44077", "accept": false, - "reason": "CLIPOutput attentions and patchtsmixer post_init policy do not match the same bug." + "reason": "CLIPOutput attentions regression and patchtsmixer post_init policy are unrelated." }, { "left": "issue:36296", "right": "issue:43575", "accept": false, - "reason": "Tensor-parallel training bug vs TP OOM on loading a large model: related area, but not the same underlying issue." + "reason": "Tensor-parallel training bug and TP OOM on a specific large model are not the same concrete failure." }, { "left": "issue:43867", "right": "issue:44484", "accept": false, - "reason": "Loading error from sorted state_dict is unrelated to max_shard_size behavior in save_pretrained." + "reason": "State_dict loading error is not the same as the default max_shard_size question." }, { "left": "issue:43618", "right": "issue:45588", "accept": false, - "reason": "Different crash sites and causes: CLIP attentions vs flash_attention on sink-less models." + "reason": "Different failures: CLIPOutput attentions missing vs flash_attention.py crashing on s_aux=None." }, { "left": "issue:43618", "right": "issue:44857", "accept": false, - "reason": "Different issues in different code paths; not duplicates." + "reason": "CLIPOutput attentions regression and AMP/CUDA crash in LwDetrImageLoss are unrelated." }, { "left": "issue:28282", "right": "issue:33453", "accept": false, - "reason": "AutoModel PyTorch import error and tokenizer loading regression are unrelated." + "reason": "AutoModel missing PyTorch dependency and tokenizer loading regression are different import/load problems." }, { "left": "issue:42890", "right": "issue:44964", "accept": false, - "reason": "Test seeding failures and Phi-4 multimodal model loading are unrelated." + "reason": "Seed-related test flakiness is unrelated to Phi-4 multimodal model loading." }, { "left": "issue:42890", "right": "issue:45600", "accept": false, - "reason": "A test seeding problem is not the same as stale auto_mappings references breaking CI." + "reason": "Missing set_seed in a test is unrelated to removed Sam3LiteText config references in auto_mappings." }, { "left": "issue:42890", "right": "issue:44112", "accept": false, - "reason": "Both are CI/test failures, but the specific causes differ." + "reason": "Flaky integration test seeding and GraniteSpeech device-override CI failure are different test bugs." }, { "left": "issue:43010", "right": "issue:44075", "accept": false, - "reason": "Cache update no_grad and SGD args ignored are different behavior bugs." + "reason": "Cache update decoration is unrelated to SGD args not being applied." }, { "left": "issue:43502", "right": "issue:44062", "accept": false, - "reason": "Offline download enforcement and AddedToken argument handling are unrelated tokenizer/loading issues." + "reason": "Ignoring local_files_only and AddedToken special-arg duplication are different tokenizer/file-fetch bugs." }, { "left": "issue:43723", "right": "issue:44556", "accept": false, - "reason": "Tokenizer loading in v5 and checkpoint reload incompatibility after upgrade are not the same concrete bug." + "reason": "Tokenizer loading regression and checkpoint reload incompatibility are related only loosely; the concrete failures differ." }, { - "left": "issue:42890", - "right": "issue:45341", + "left": "issue:44222", + "right": "issue:45125", "accept": false, - "reason": "Missing set_seed in tests and a testing_utils bug are too different without clearer evidence." + "reason": "FP8 save_pretrained behavior and missing `_tp_plan` for tensor parallelism are different MoE issues." }, { - "left": "issue:44222", - "right": "issue:45125", + "left": "issue:42890", + "right": "issue:45341", "accept": false, - "reason": "FP8 save_pretrained for MoE and missing _tp_plan for Qwen3_5Moe are different features/failures." + "reason": "Missing test seeding is not the same as the generic testing_utils bug referenced here." }, { "left": "issue:44387", "right": "issue:44556", "accept": false, - "reason": "Quantization memory OOM and checkpoint reload incompatibility are unrelated." + "reason": "Int4 quantization OOM and checkpoint reload failure are distinct regressions." }, { "left": "issue:43010", "right": "issue:43065", "accept": false, - "reason": "Cache/layer no_grad and Sam3PixelDecoder dummy Conv2d are unrelated implementation issues." + "reason": "Cache update no_grad behavior and the Sam3PixelDecoder dummy Conv2d issue are unrelated." }, { "left": "issue:43010", "right": "issue:43232", "accept": false, - "reason": "Generic generation kwargs handling and cache update no_sync behavior are different bugs." + "reason": "Cache update no_grad behavior and generation kwarg synchronization are different code paths." }, { "left": "issue:42548", "right": "issue:44351", "accept": false, - "reason": "Cannot import PreTrainedModel and cannot import HybridCache are distinct import/export regressions." - }, - { - "left": "issue:45125", - "right": "issue:45478", - "accept": false, - "reason": "Both involve Qwen3.5 MoE, but one is tensor-parallel metadata and the other is a from_pretrained loading error." + "reason": "Missing PreTrainedModel export and missing HybridCache export are both import errors, but not the same missing symbol or root cause." }, { "left": "issue:44355", "right": "issue:44749", "accept": false, - "reason": "Compiled Python file errors and a performance regression after upgrade are not the same issue." + "reason": "Compiled Python-file errors and slow filtering after upgrade are unrelated." }, { "left": "issue:45125", - "right": "issue:45375", + "right": "issue:45478", "accept": false, - "reason": "Missing _tp_plan and missing deepstack_visual_indexes are different Qwen3.5 model/config problems." + "reason": "Missing `_tp_plan` and Qwen3.5 Moe from_pretrained error are related only at a high level; not the same concrete bug." }, { - "left": "issue:43502", - "right": "issue:44485", + "left": "issue:44222", + "right": "issue:45375", "accept": false, - "reason": "Offline API blocking and GLM-5 RoPE implementation are unrelated." + "reason": "FP8 MoE save_pretrained and missing deepstack_visual_indexes are different Qwen3.5 vision/config issues." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. The only accepted soft edge, issue:41669 vs issue:44246, is well supported by the titles and the import-slowdown theme. The remaining rejections read as cautious and do not overclaim duplicates. No mergeability concerns apply since these are issue pairs, not PR pairs." + "feedback": "The summary is grounded in the packet: issue:45278 is a reasonable umbrella for the import/load-related regressions, and the rest of the cluster is mostly heterogeneous. The only accepted soft edge, issue:41669 \u2194 issue:44246, looks like a plausible duplicate import-slowdown pair. The other pairwise rejections are conservative and consistent with the issue titles." }, "evaluator_used": true, "retried": false, @@ -13162,9 +13128,9 @@ "review_unit_index": 41, "review_unit_count": 43, "status": "reviewed", - "reason": null, - "source": "cache", - "cache_hit": true, + "reason": null, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -13218,6 +13184,7 @@ "issue:45542" ], "soft_pairs": [ + "issue:43502|issue:44485", "issue:30990|issue:31795", "issue:44075|issue:44871", "issue:44315|issue:44360", @@ -13257,262 +13224,268 @@ "issue:44393|issue:45071", "issue:41669|issue:43673" ], - "prepared_review_unit_hash": "43c46936dabf8bac77d0fe6b71c11ebd2f6c5deb41751d62bc61df348f5c09ed", + "prepared_review_unit_hash": "99194bfbff11ba0d515efd1661cab690c54cadf7855fb5ed9774db2c75d3382a", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 38, - "serialized_chars": 38352, - "estimated_input_tokens": 9588, - "estimated_eval_tokens": 19432 + "soft_pair_count": 39, + "serialized_chars": 38649, + "estimated_input_tokens": 9663, + "estimated_eval_tokens": 19582 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous: it mixes unrelated import errors, model-loading regressions, docs issues, cache/FP8/tensor-parallel bugs, and CLI/runtime problems. I don\u2019t see any safe duplicate merges, and there are no PRs to evaluate.", - "confidence": 0.19, - "canonical_issue_reason": "If forced to pick one, issue 43576 is the broadest v5 regression-style report and the closest thing to a generic umbrella issue, but it does not \u05d1\u05d0\u05de\u05ea subsume the rest of the cluster.", + "summary": "This cluster is mostly a loose set of Transformers v5 regression reports, especially import/loading failures, but the soft pairs do not look like true duplicates of the same concrete bug. I would not merge any of the proposed soft edges.", + "confidence": 0.64, + "canonical_issue_reason": "issue:43576 is the most central and broadly connected report in the set, with multiple soft similarities to other v5 breakage/import-loading issues. It best represents the cluster theme, even though it is not a duplicate of the others.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43576 is the best single representative only because it is broad and version-gated; however, the overall set is too mixed for a strong canonical issue.", + "best_issue_reason": "issue:43576 is the best single issue to anchor this cluster because it is the most connected to the other v5-related reports and captures the shared breakage theme more broadly than the narrower model-specific issues.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43502", + "right": "issue:44485", + "accept": false, + "reason": "Different problems: offline/local-files-only networking behavior vs GLM-5 RoPE implementation." + }, { "left": "issue:30990", "right": "issue:31795", "accept": false, - "reason": "Unrelated symptoms: a loading hang vs documentation confusion about model.forward arguments." + "reason": "Unrelated topics: Sentence Transformers load hang vs documentation confusion about forward arguments." }, { "left": "issue:44075", "right": "issue:44871", "accept": false, - "reason": "Different bugs in different areas: SGD optimizer arguments vs Gemma EOS token configuration." + "reason": "Different bugs: SGD optimizer args ignored vs Gemma-3 EOS token config mismatch." }, { "left": "issue:44315", "right": "issue:44360", "accept": false, - "reason": "Liger Kernel application bug and DSA indexer ReLU complaint are unrelated." + "reason": "No shared code-path or symptom; model_init/Liger Kernel issue is unrelated to the DSA indexer ReLU discussion." }, { "left": "issue:30333", "right": "issue:33453", "accept": false, - "reason": "MLflow job-status reporting and tokenizer loading regression are different code paths." + "reason": "Different failures: MLFlow reporting on training failure vs tokenizer loading regression." }, { "left": "issue:43010", "right": "issue:44368", "accept": false, - "reason": "Cache update no_grad issue is unrelated to tie_word_embeddings warnings." + "reason": "Different areas: cache/layer no_grad decoration vs tie_word_embeddings warning during Qwen fine-tuning." }, { "left": "issue:43502", "right": "issue:44222", "accept": false, - "reason": "local_files_only network leakage and FP8 save_pretrained are separate failures." + "reason": "Different issues: local_files_only network requests vs FP8 moe save_pretrained behavior." }, { "left": "issue:43299", "right": "issue:44556", "accept": false, - "reason": "Both are loading regressions, but they involve different models and failure modes; not the same concrete bug." + "reason": "Both are v5-related loading regressions, but they affect different models and scenarios; not the same concrete bug." }, { "left": "issue:43502", "right": "issue:44961", "accept": false, - "reason": "API requests despite local_files_only is unrelated to the racoon issue." + "reason": "No meaningful overlap; local_files_only networking bug vs unrelated 'racoon' issue." }, { "left": "issue:44222", "right": "issue:44230", "accept": false, - "reason": "Same FP8/MoE theme, but one is save_pretrained behavior and the other is inference support; too different to merge." + "reason": "Related FP8/moe area, but one is save_pretrained and the other is inference support; not the same bug." }, { "left": "issue:44355", "right": "issue:44655", "accept": false, - "reason": "Compiled Python file errors and Pipeline save_pretrained are unrelated." + "reason": "Different functionality: compiled Python file errors vs saving Pipeline objects." }, { "left": "issue:30333", "right": "issue:36296", "accept": false, - "reason": "MLflow reporting failure and tensor parallel training bug do not share the same underlying cause." + "reason": "Different subsystems and symptoms: MLFlow reporting vs tensor parallel training." }, { "left": "issue:44355", "right": "issue:45071", "accept": false, - "reason": "Compiled-file runtime errors are unrelated to PretrainedConfig type checking regressions." + "reason": "Different issues: compiled file execution error vs PretrainedConfig type-checking regression." }, { "left": "issue:42831", "right": "issue:43874", "accept": false, - "reason": "FineGrainedFP8 accuracy issue and missing image-processor method are different bugs." + "reason": "Different model-specific bugs: FineGrainedFP8 accuracy vs GLM46V image processor token counting." }, { "left": "issue:33453", "right": "issue:34689", "accept": false, - "reason": "Both are loading-related, but one is a tokenizer regression and the other is a specific Llama 3.2 vision model-loading breakage." + "reason": "Both are loading regressions, but for different models and likely different root causes." }, { "left": "issue:42757", "right": "issue:43576", "accept": false, - "reason": "ImportError from huggingface_hub and broken transformers env command are separate issues." + "reason": "Both are v5 compatibility complaints, but one is a missing import and the other is the env command being broken; not the same bug." }, { "left": "issue:33453", "right": "issue:36296", "accept": false, - "reason": "Tokenizer loading regression and tensor-parallel training bug are unrelated." + "reason": "Unrelated: tokenizer loading regression vs tensor parallel training bug." }, { "left": "issue:43827", "right": "issue:44961", "accept": false, - "reason": "Docs still referencing pipeline() is unrelated to racoon." + "reason": "Docs reference to pipeline() is unrelated to the 'racoon' issue." }, { "left": "issue:43576", "right": "issue:45071", "accept": false, - "reason": "Both are v5 regressions, but one is a broken CLI command and the other is a type-checking issue; not the same code path." + "reason": "Both involve v5 breakage, but one is the env command and the other is PretrainedConfig typing; different concrete failures." }, { "left": "issue:43576", "right": "issue:43824", "accept": false, - "reason": "Broken env command and missing Qwen2_5_VLForConditionalGeneration import are different failures." + "reason": "Both are import/loading regressions, but the missing symbol and the env command bug are not the same change or code-path." }, { "left": "issue:38617", "right": "issue:43576", "accept": false, - "reason": "Missing layer_type_validation import and broken env command are unrelated import/CLI problems." + "reason": "Different specific failures: missing layer_type_validation import vs broken env command." }, { "left": "issue:43673", "right": "issue:44492", "accept": false, - "reason": "Chunked_prefill cache missing and a typo in cache strategies are not the same bug." + "reason": "Cache-missing during chunked prefill is not the same as a typo in cache strategies." }, { "left": "issue:44961", "right": "issue:45468", "accept": false, - "reason": "A placeholder racoon issue is unrelated to Gemma-4 audio positional encoding." + "reason": "No overlap: 'racoon' vs Gemma-4 audio positional encoding." }, { "left": "issue:31795", "right": "issue:38617", "accept": false, - "reason": "Documentation confusion and an ImportError are different issue types with different fixes." + "reason": "Documentation confusion vs import error; not duplicate bugs." }, { "left": "issue:43576", "right": "issue:44351", "accept": false, - "reason": "Broken env command and missing HybridCache import are both v5-related but not the same concrete defect." + "reason": "Both are v5 import-related, but they concern different missing symbols and likely different fixes." }, { "left": "issue:41628", "right": "issue:42548", "accept": false, - "reason": "Both are import errors, but they concern different exported symbols and likely different loader paths." + "reason": "Both are import errors, but for different classes and missing symbols; not enough evidence of the same underlying bug." }, { "left": "issue:44908", "right": "issue:44961", "accept": false, - "reason": "Scheduler kwargs bug and racoon are unrelated." + "reason": "Scheduler kwargs bug is unrelated to the 'racoon' issue." }, { "left": "issue:44485", "right": "issue:44961", "accept": false, - "reason": "GLM-5 RoPE implementation discussion is unrelated to racoon." + "reason": "RoPE implementation discussion is unrelated to 'racoon'." }, { "left": "issue:36296", "right": "issue:43901", "accept": false, - "reason": "Tensor parallel training bug and docs mentioning return_all_scores are unrelated." + "reason": "Tensor parallel training bug is unrelated to a docs mismatch in TextClassificationPipeline." }, { "left": "issue:36296", "right": "issue:43526", "accept": false, - "reason": "Tensor parallel training and BeitImageProcessorFast reduce_labels are unrelated." + "reason": "Different model/training bug vs BeitImageProcessorFast reduce_labels issue." }, { "left": "issue:41628", "right": "issue:44355", "accept": false, - "reason": "ImportError for AutoImageProcessor and compiled-file errors are not the same issue." + "reason": "Import error for AutoImageProcessor is unrelated to compiled Python file execution errors." }, { "left": "issue:44273", "right": "issue:45003", "accept": false, - "reason": "Lazy loading malfunction and unsafe sys.modules access are different bug reports." + "reason": "Both mention loading behavior, but lazy loading problems and unsafe sys.modules access are not clearly the same bug." }, { "left": "issue:44261", "right": "issue:44360", "accept": false, - "reason": "MLA q_a_layernorm epsilon mismatch and missing ReLU in DSA indexer are unrelated." + "reason": "MLA q_a_layernorm precision/config issue is unrelated to the DSA indexer ReLU topic." }, { "left": "issue:44623", "right": "issue:45230", "accept": false, - "reason": "Processor save_pretrained missing files and a generic bug report do not establish a shared underlying bug." + "reason": "Generic 'Bug report' is too vague to treat as the same issue as missing processor.save_pretrained files." }, { "left": "issue:44273", "right": "issue:45542", "accept": false, - "reason": "Lazy loading and undefined tf backend error are different problems." + "reason": "Lazy loading behavior is unrelated to the TensorFlow backend error when only tensorboard is installed." }, { "left": "issue:43576", "right": "issue:44556", "accept": false, - "reason": "Broken env command and checkpoint reload failure after upgrading to v5.2/v5.3 are distinct regressions." + "reason": "Both are v5 breakages, but one is the env command and the other is checkpoint reload compatibility; not the same concrete bug." }, { "left": "issue:44038", "right": "issue:45125", "accept": false, - "reason": "Qwen3-VL-Moe loading bug and missing _tp_plan for tensor parallelism are different code paths." + "reason": "Different Qwen/moe and tensor-parallel problems; not the same underlying change." }, { "left": "issue:44393", "right": "issue:45071", "accept": false, - "reason": "Qwen3-VL 2D box output error and PretrainedConfig type-checking regression are unrelated." + "reason": "Qwen3-VL bounding-box output issue is unrelated to PretrainedConfig type checking." }, { "left": "issue:41669", "right": "issue:43673", "accept": false, - "reason": "Import-star cleanup in models and missing GenerationMixin cache during chunked_prefill are different changes." + "reason": "Import-star cleanup/perf issue is unrelated to a missing cache during chunked prefill." } ] }, @@ -13531,8 +13504,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -13629,323 +13602,320 @@ "issue:44263|issue:45468", "issue:44261|issue:45468", "issue:43606|issue:45468", - "issue:42913|issue:43066", "issue:42890|issue:43704", + "issue:42913|issue:43066", "issue:44908|issue:44998" ], - "prepared_review_unit_hash": "2f1a6c3146f9f6aef6f43e633bba86536aa2493ab5902d72058799d42db9201e", + "prepared_review_unit_hash": "49fde7466f69de5fd53f79114bc3d20afb46bc84605228a528fecfbe42ffe323", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 46, - "serialized_chars": 40776, - "estimated_input_tokens": 10194, - "estimated_eval_tokens": 20644 + "serialized_chars": 40783, + "estimated_input_tokens": 10196, + "estimated_eval_tokens": 20648 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This packet is mostly a set of unrelated issue reports. I\u2019m not accepting the two title-similar pairs as duplicates: 44556/45092 and 44246/44273 only have thematic overlap, with no body/diff evidence that they share the same root cause or would be fixed cleanly by one change.", - "confidence": 0.9, - "canonical_issue_reason": "issue:44556 is the clearest standalone report of the v5 checkpoint-reload regression theme and the strongest representative issue in this set.", + "summary": "This is a mixed set of mostly unrelated Transformers issues, with several loose clusters around tokenizer/model loading in v5, but no soft pair is strong enough to treat as the same underlying bug under strict duplicate triage.", + "confidence": 0.76, + "canonical_issue_reason": "issue:43723 is the clearest broad anchor in the set because it directly describes an AutoTokenizer.from_pretrained loading regression in v5, which matches the repeated tokenizer-loading theme better than the narrower or more model-specific reports.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44556 is the best overall anchor here because it is specific, high-signal, and describes a concrete loading failure rather than a vague symptom.", + "best_issue_reason": "issue:43723 is also the best overall issue to keep as the representative reference for this cluster because it is the most central and reusable description of the shared v5 tokenizer-loading problem area.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44556", "right": "issue:45092", "accept": false, - "reason": "Only title-level similarity about v5 checkpoint/model loading; no body or diff evidence that the same code path or root cause is involved." + "reason": "Both are v5 checkpoint-loading complaints, but one is a generic reload failure and the other is a specific InternVL2 remote-code/meta-init incompatibility; not the same concrete bug." }, { "left": "issue:42371", "right": "issue:44492", "accept": false, - "reason": "TF32 configuration guidance and a cache-strategy typo are different bugs in different areas." + "reason": "TF32 behavior settings and a cache-strategy typo are unrelated problems." }, { "left": "issue:42890", "right": "issue:44908", "accept": false, - "reason": "Test flakiness from missing seeds is unrelated to inverse_sqrt scheduler kwargs handling." + "reason": "A flaky integration test missing set_seed is unrelated to the inverse_sqrt scheduler ignoring kwargs." }, { "left": "issue:44704", "right": "issue:44908", "accept": false, - "reason": "AutoProcessor kwargs forwarding and scheduler kwargs are separate code paths with different failures." + "reason": "AutoProcessor kwargs forwarding and scheduler kwargs handling are different code paths and different bugs." }, { "left": "issue:44246", "right": "issue:44273", "accept": false, - "reason": "Only thematic overlap around loading/import performance; no evidence they are the same underlying bug or one fix would cover both." + "reason": "Import-time slowness and lazy-loading correctness are related only at a high level, not the same bug." }, { "left": "issue:43408", "right": "issue:44315", "accept": false, - "reason": "Wrong model-type warning and Liger Kernel not applied with model_init are unrelated symptoms." + "reason": "A model-type warning for sam3_video/sam3_tracker is unrelated to Liger Kernel application with model_init." }, { "left": "issue:44297", "right": "issue:45230", "accept": false, - "reason": "Tokenizer_class mismatch on save_pretrained is not the same as a generic bug report." + "reason": "A tokenizer_class mismatch in save_pretrained is not the same as an unspecified generic bug report." }, { "left": "issue:43352", "right": "issue:44393", "accept": false, - "reason": "Flash Attention 2 support warning and Qwen3-VL bbox output issues are different model problems." + "reason": "Flash Attention 2 support for Nemotron and Qwen3-VL bounding-box output issues are unrelated." }, { "left": "issue:43576", "right": "issue:44908", "accept": false, - "reason": "Broken transformers env command is unrelated to scheduler kwargs handling." + "reason": "A broken transformers env command is unrelated to scheduler kwargs handling." }, { "left": "issue:43976", "right": "issue:44393", "accept": false, - "reason": "Python version support issue and Qwen3-VL output bug do not share a concrete code path." + "reason": "Python version compatibility and Qwen3-VL output errors are unrelated." }, { "left": "issue:44908", "right": "issue:45071", "accept": false, - "reason": "Scheduler kwargs handling and PretrainedConfig type checking are unrelated." + "reason": "Scheduler kwargs handling and PretrainedConfig type checking are different regressions." }, { "left": "issue:44295", "right": "issue:45230", "accept": false, - "reason": "position_ids buffer access error is unrelated to a vague generic bug report." + "reason": "A position_ids buffer access error is not the same as a generic bug report." }, { "left": "issue:41669", "right": "issue:44351", "accept": false, - "reason": "Import-star performance slowdown and missing HybridCache import are different problems." + "reason": "Removing import * for performance is not the same as failing to import HybridCache." }, { "left": "issue:41669", "right": "issue:45341", "accept": false, - "reason": "Import-time slowdown and a testing_utils bug are unrelated." + "reason": "Import cleanup/performance and a testing_utils bug are unrelated." }, { "left": "issue:44492", "right": "issue:44945", "accept": false, - "reason": "Cache-strategy typo and pipeline-parallelism output corruption are different bugs." + "reason": "A cache-strategy typo does not match an incorrect pipeline-parallelism output bug." }, { "left": "issue:41669", "right": "issue:44492", "accept": false, - "reason": "Model import performance regression and cache-strategy typo are unrelated." + "reason": "Import cleanup and cache-strategy wording are unrelated." }, { "left": "issue:44315", "right": "issue:44829", "accept": false, - "reason": "Liger Kernel application issue and flash_attention_3 training degeneration are distinct problems." + "reason": "Liger Kernel not being applied with model_init and flash_attention_3 causing degenerate training are different issues." }, { "left": "issue:44393", "right": "issue:45478", "accept": false, - "reason": "Qwen3-VL bounding-box hallucination/error and Qwen3.5 MoE from_pretrained failure are different model issues." + "reason": "Qwen3-VL 2D bounding-box hallucinations and Qwen3.5 MoE from_pretrained loading failures are unrelated." }, { "left": "issue:31795", "right": "issue:33453", "accept": false, - "reason": "Documentation confusion about model.forward is not the same as a tokenizer-loading regression." + "reason": "Documentation confusion about forward arguments is not the same as a tokenizer-loading regression." }, { "left": "issue:43827", "right": "issue:44908", "accept": false, - "reason": "Docs still referencing pipeline() and scheduler kwargs handling are unrelated." + "reason": "Docs still referencing pipeline() after removals is unrelated to inverse_sqrt scheduler kwargs." }, { "left": "issue:44077", "right": "issue:44861", "accept": false, - "reason": "post_init policy and _tied_weights_keys AttributeError are separate model-internal issues." + "reason": "An optional post_init restriction and a tied-weights AttributeError are different bugs." }, { "left": "issue:41669", "right": "issue:44908", "accept": false, - "reason": "Import-time performance slowdown and scheduler config handling have no overlap in code path." + "reason": "Import-time cleanup and scheduler behavior are unrelated." }, { "left": "issue:43723", "right": "issue:44393", "accept": false, - "reason": "Tokenizer loading regression and Qwen3-VL bbox output bug are unrelated." + "reason": "Tokenizer loading in v5 and Qwen3-VL bounding-box output errors are unrelated." }, { "left": "issue:44655", "right": "issue:44861", "accept": false, - "reason": "Saving pipeline objects and tied-weight key handling are different failure modes." + "reason": "Saving pipeline objects and _get_tied_weight_keys crashing on a list are different failure modes." }, { "left": "issue:43576", "right": "issue:44273", "accept": false, - "reason": "Broken env command and lazy-loading/import slowdown are different features and bugs." + "reason": "The env command issue is unrelated to lazy loading not working properly." }, { "left": "issue:44360", "right": "issue:45468", "accept": false, - "reason": "DSA ReLU concern and Gemma4 audio positional encoding are unrelated." + "reason": "A missing ReLU in the DSA indexer and Gemma-4 audio positional encoding are unrelated." }, { "left": "issue:45230", "right": "issue:45310", "accept": false, - "reason": "Generic bug report is too vague to be the same as the specific Qwen3.5 MoE from_pretrained error." + "reason": "A generic bug report is not the same as a specific Qwen3.5 MoE from_pretrained error." }, { "left": "issue:44273", "right": "issue:44908", "accept": false, - "reason": "Lazy loading is not working properly and scheduler kwargs handling are separate subsystems." + "reason": "Lazy loading failures and scheduler kwargs handling are unrelated." }, { "left": "issue:41669", "right": "issue:43704", "accept": false, - "reason": "Import-time slowdown and VRAM leak in dataloader threads are unrelated." + "reason": "Import * cleanup/performance and VRAM leakage in dataloader threads are different issues." }, { "left": "issue:43519", "right": "issue:44485", "accept": false, - "reason": "Timestamp calculation bug and GLM-5 RoPE implementation are different model-specific problems." + "reason": "Qwen3VL timestamp calculation and GLM-5 RoPE implementation are separate model bugs." }, { "left": "issue:43519", "right": "issue:44908", "accept": false, - "reason": "Processor timestamp math and scheduler kwargs handling are unrelated." + "reason": "Processor timestamp calculation and scheduler kwargs handling are unrelated." }, { "left": "issue:31795", "right": "issue:34689", "accept": false, - "reason": "Documentation confusion and Llama 3.2 Vision loading regression are not the same bug." + "reason": "Documentation confusion and a Llama 3.2 model-loading regression are not the same bug." }, { "left": "issue:31515", "right": "issue:31795", "accept": false, - "reason": "Checkpoint loading slowdown and forward documentation confusion are unrelated." + "reason": "Slow checkpoint loading and forward-argument documentation confusion are unrelated." }, { "left": "issue:44315", "right": "issue:45092", "accept": false, - "reason": "Liger Kernel application and v5 checkpoint/meta-init loading failure are different code paths." + "reason": "Liger Kernel application with model_init and old checkpoint/meta-init incompatibility are different problems." }, { "left": "issue:43704", "right": "issue:44485", "accept": false, - "reason": "VRAM leak in dataloader threads is unrelated to RoPE implementation issues." + "reason": "VRAM leakage across dataloader threads is unrelated to GLM-5 RoPE implementation." }, { "left": "issue:44485", "right": "issue:45468", "accept": false, - "reason": "GLM-5 RoPE discussion and Gemma4 audio positional encoding are different model implementations." + "reason": "GLM-5 RoPE and Gemma-4 audio positional encoding are unrelated model-specific bugs." }, { "left": "issue:42371", "right": "issue:43704", "accept": false, - "reason": "TF32 settings and VRAM leak in dataloader threads are unrelated." + "reason": "TF32 settings and VRAM leakage in dataloader threads are unrelated." }, { "left": "issue:43502", "right": "issue:43519", "accept": false, - "reason": "local_files_only network requests and Qwen3VL timestamp calculation are separate bugs." + "reason": "Unexpected API requests with local_files_only and Qwen3VL timestamp calculation are unrelated." }, { "left": "issue:43519", "right": "issue:44623", "accept": false, - "reason": "Timestamp computation bug and processor.save_pretrained missing files are different issues." + "reason": "Processor timestamp math and missing files in processor.save_pretrained are different bugs." }, { "left": "issue:43519", "right": "issue:43704", "accept": false, - "reason": "Timestamp computation bug and VRAM leak are unrelated." + "reason": "Timestamp calculation and VRAM leakage across threads are unrelated." }, { "left": "issue:44263", "right": "issue:45468", "accept": false, - "reason": "torch.split return semantics and Gemma4 audio positional encoding are different problems." + "reason": "A torch.split return-value issue and Gemma-4 audio positional encoding are different bugs." }, { "left": "issue:44261", "right": "issue:45468", "accept": false, - "reason": "MLA q_a_layernorm precision/config issue is unrelated to Gemma4 audio positional encoding." + "reason": "An MLA q_a_layernorm precision issue is unrelated to Gemma-4 audio positional encoding." }, { "left": "issue:43606", "right": "issue:45468", "accept": false, - "reason": "CPU offload device mismatch is unrelated to Gemma4 audio positional encoding." + "reason": "CPU offload device mismatch and Gemma-4 audio positional encoding are unrelated." }, { - "left": "issue:42913", - "right": "issue:43066", + "left": "issue:42890", + "right": "issue:43704", "accept": false, - "reason": "Tokenizer behavior differences and wrong decoder type in v5 are related thematically but not the same concrete bug." + "reason": "A flaky integration test and VRAM leakage in dataloader threads are unrelated." }, { - "left": "issue:42890", - "right": "issue:43704", + "left": "issue:42913", + "right": "issue:43066", "accept": false, - "reason": "Test seeding flakiness is unrelated to VRAM leakage." + "reason": "Both mention tokenizer changes in v5, but one is a broad behavior-difference report and the other is a specific decoder-type mismatch; not enough to call them the same bug." }, { "left": "issue:44908", "right": "issue:44998", "accept": false, - "reason": "Scheduler configuration bug and an unrelated placeholder issue are not a match." + "reason": "A scheduler kwargs bug and an unrelated placeholder issue are not the same thing." } ] }, - "evaluator_result": { - "accept": false, - "feedback": "The writeup is mostly conservative, but it still overstates the evidence for the two accepted soft edges. In the packet, 44556/45092 and 44246/44273 are only supported by title-level thematic similarity; there is no body/diff evidence showing the same root cause or that a single PR would cleanly fix both. I\u2019d keep those pairs as tentative rather than accepted duplicate clusters." - }, - "evaluator_used": true, - "retried": true, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null @@ -13958,8 +13928,8 @@ "review_unit_count": 43, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-15354-374", "nodes": [ @@ -13998,118 +13968,118 @@ "issue:31515|issue:33453", "issue:43298|issue:44230" ], - "prepared_review_unit_hash": "55065f5b7a5682e02817892a07b80cd0bb68e1231277967451f0fb63768bfbd6", + "prepared_review_unit_hash": "04281dd5a6716af263c30c1d3922adecaa5126f2841630e3c8d4107beac9bffc", "component_budget": { "node_count": 374, "item_count": 374, "soft_pair_count": 1395, - "serialized_chars": 720210, - "estimated_input_tokens": 180053, - "estimated_eval_tokens": 360362 + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 }, "budget": { "node_count": 18, "item_count": 18, "soft_pair_count": 14, - "serialized_chars": 14257, - "estimated_input_tokens": 3565, - "estimated_eval_tokens": 7386 + "serialized_chars": 14255, + "estimated_input_tokens": 3564, + "estimated_eval_tokens": 7384 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is highly heterogeneous: most items are unrelated bugs, and every soft-similarity pair appears to be a false positive. I would not merge any of these issues as duplicates.", + "summary": "This cluster is highly mixed: the issues share superficial wording around crashes, loading, or config names, but they do not describe the same underlying bug. None of the soft pairs look safe to merge as duplicates.", "confidence": 0.95, - "canonical_issue_reason": null, + "canonical_issue_reason": "No single canonical issue stands out as a true duplicate hub; the items span unrelated failures across loading, configs, templates, training, and misc. bugs.", "canonical_pr_reason": null, - "best_issue_reason": "No single issue is a safe canonical representative because the connected items span unrelated subsystems and failure modes.", + "best_issue_reason": "issue:31515 is the closest thing to a representative because it is a concrete from_pretrained/loading report and has the most superficial overlap with other load-related items, but it still is not a true duplicate center.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44861", "right": "issue:45290", "accept": false, - "reason": "Different bugs in different paths: tied-weight key handling vs chat template/tool-call serialization." + "reason": "Both are crashes, but one is about tied weights internals and the other about chat template tool-call handling; different code paths and failure modes." }, { "left": "issue:44998", "right": "issue:45341", "accept": false, - "reason": "One is a nonsense-title issue and the other is a testing_utils bug; no shared symptom or code path." + "reason": "'Unemployment' is unrelated to a small testing_utils bug; no shared underlying behavior." }, { "left": "issue:44492", "right": "issue:44998", "accept": false, - "reason": "A typo in cache strategy docs/config is unrelated to the unrelated 'Unemployment' issue." + "reason": "A cache-strategy typo and an unrelated issue title do not describe the same bug or change." }, { "left": "issue:43618", "right": "issue:44998", "accept": false, - "reason": "CLIPOutput attentions regression is unrelated to the other issue." + "reason": "CLIPOutput attentions assignment is a specific model-output regression, unrelated to the other item." }, { "left": "issue:44556", "right": "issue:44998", "accept": false, - "reason": "Checkpoint reload incompatibility is not the same as the unrelated 'Unemployment' issue." + "reason": "Checkpoint reload breakage after version upgrade is unrelated to the other issue; no concrete overlap." }, { "left": "issue:43525", "right": "issue:44861", "accept": false, - "reason": "Llama4Config pad_token_id attribute missing is unrelated to tied-weight key list handling." + "reason": "Missing pad_token_id on Llama4Config and tied-weight-key AttributeError are different config/model bugs." }, { "left": "issue:43572", "right": "issue:44861", "accept": false, - "reason": "StableLmConfig pad_token_idx regression is a different config attribute bug than the tied-weight crash." + "reason": "StableLmConfig pad_token_idx regression is unrelated to the tied-weight-key list crash." }, { "left": "issue:31515", "right": "issue:38617", "accept": false, - "reason": "Slow from_pretrained checkpoint loading is unrelated to an ImportError for layer_type_validation." + "reason": "Slow checkpoint loading and an ImportError for layer_type_validation are distinct problems in different code paths." }, { "left": "issue:44230", "right": "issue:44623", "accept": false, - "reason": "FP8 Qwen model support and missing processor save_pretrained files are different issues." + "reason": "FP8 support for a model family and missing processor.save_pretrained files are not the same bug." }, { "left": "issue:31515", "right": "issue:36296", "accept": false, - "reason": "Tokenizer/checkpoint loading slowness is not the same as a tensor-parallel training bug." + "reason": "from_pretrained loading slowness is unrelated to a tensor-parallel training bug." }, { "left": "issue:30333", "right": "issue:31515", "accept": false, - "reason": "MLflow reporting on failed training jobs and slow checkpoint loading are unrelated problems." + "reason": "MLflow reporting on failed training jobs and checkpoint loading speed are unrelated concerns." }, { "left": "issue:44230", "right": "issue:45310", "accept": false, - "reason": "Qwen3-VL/Qwen3.5 FP8 support request and a from_pretrained error on Qwen3.5 MoE are not the same concrete bug." + "reason": "FP8 inference support for Qwen3-VL/MoE and a Qwen3.5 MoE from_pretrained error are different concrete issues." }, { "left": "issue:31515", "right": "issue:33453", "accept": false, - "reason": "Both involve loading, but one is checkpoint load slowness and the other is a tokenizer loading regression; not the same bug." + "reason": "Tokenizer loading regression and checkpoint loading slowness are different failure modes, not the same duplicate." }, { "left": "issue:43298", "right": "issue:44230", "accept": false, - "reason": "Continuous batching audio-model support and FP8 Qwen inference support are distinct feature gaps." + "reason": "Continuous batching support for audio models and FP8 support for Qwen models are separate feature/bug reports." } ] }, @@ -14128,8 +14098,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-26350-2", "nodes": [ @@ -14139,7 +14109,7 @@ "soft_pairs": [ "issue:26350|issue:33260" ], - "prepared_review_unit_hash": "ad8d1b620509c568a6a760a93a5e059838298c90968b34d1ad4ed61ae5ebe562", + "prepared_review_unit_hash": "e0ac32febd7dbe0a2803a61f6aec3c590807364ff4bccf125285591426951abd", "component_budget": { "node_count": 2, "item_count": 2, @@ -14161,18 +14131,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are not duplicates: one requests Flash Attention 2 support for more architectures, the other requests GGUF support for more architectures. They share a broad \u201ccommunity contribution\u201d theme but describe different feature work.", - "confidence": 0.97, + "summary": "These are two distinct community-contribution enhancement issues: one for Flash Attention 2 support on more architectures, and one for GGUF support on more architectures. They share only a broad theme and are not duplicates.", + "confidence": 0.98, "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue 26350 is the stronger representative thread in this cluster because it has substantially more discussion and is the older, more established request; however, it should not be treated as a duplicate of 33260.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:26350", "right": "issue:33260", "accept": false, - "reason": "Different requested features and code paths: Flash Attention 2 architecture support vs GGUF architecture support. Similar framing, but not the same underlying change." + "reason": "Different feature requests for different backends/code paths (Flash Attention 2 vs GGUF); same broad \u2018more architectures\u2019 framing is too generic to treat as the same issue." } ] }, @@ -14191,8 +14161,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-27088-4", "nodes": [ @@ -14209,7 +14179,7 @@ "issue:43834|issue:44365", "issue:44365|issue:44370" ], - "prepared_review_unit_hash": "2148375c0b5f798b8c6e75714948c6cc19bfc19dc443baaf8853195fcd1247fe", + "prepared_review_unit_hash": "ebff596b7c9a49a465b1291325e51513294e09c9226da383fa7f0bc2a055c984", "component_budget": { "node_count": 4, "item_count": 4, @@ -14231,48 +14201,48 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "All four issues are language-translation requests using the same template. The Turkish-specific issue is the most concrete, but the other three appear to be separate template instances rather than the same underlying request; high text similarity alone is not enough to merge them as duplicates.", - "confidence": 0.71, - "canonical_issue_reason": "issue:27088 is the only concrete, specific request (Turkish). The others are generic placeholder-style variants, so this is the clearest representative of the actual problem being reported.", + "summary": "All items are about translating docs, but the placeholder-titled issues likely refer to different languages rather than the same request. The tightest near-duplicate group is the generic template issues (43834/44365/44370), while 27088 is a concrete Turkish-specific outlier.", + "confidence": 0.78, + "canonical_issue_reason": "43834 is the earliest of the near-identical template-style issues and the best representative of the generic i18n translation-request pattern.", "canonical_pr_reason": null, - "best_issue_reason": "issue:27088 is the best standalone issue because it names a specific language and has the least ambiguity; the other issues look like templated duplicates or separate language requests.", + "best_issue_reason": "43834 is the strongest cluster representative because it sits in the tightest similarity group and is the oldest item among the template-matching issues.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:27088", "right": "issue:44370", "accept": false, - "reason": "Both are translation-request templates, but one is specifically Turkish while the other is a generic placeholder-style issue. That is not enough to prove the same underlying request." + "reason": "Turkish-specific request vs. a generic placeholder language request; same topic area, but not enough evidence they are the same underlying issue." }, { "left": "issue:27088", "right": "issue:44365", "accept": false, - "reason": "High similarity likely comes from the shared i18n template; the language target is not the same concrete request, so this should not be merged as a duplicate." + "reason": "Different target language context; these look like separate translation requests, not duplicates." }, { "left": "issue:27088", "right": "issue:43834", "accept": false, - "reason": "The titles/body patterns are similar, but this looks like a different translation request instance rather than the same Turkish issue." + "reason": "The Turkish issue is concrete, while 43834 is a generic template-style request; they do not appear to be the same issue." }, { "left": "issue:43834", "right": "issue:44370", "accept": false, - "reason": "These are near-identical template issues, but there is no evidence they target the exact same language/request; template similarity alone is insufficient." + "reason": "Very similar template text, but likely different language targets, so they are not clearly the same underlying request." }, { "left": "issue:43834", "right": "issue:44365", "accept": false, - "reason": "Both look like generic translation-request variants, but not enough evidence that they are the same underlying bug/change." + "reason": "High textual similarity alone is not enough here; the placeholder language fields suggest separate translation requests." }, { "left": "issue:44365", "right": "issue:44370", "accept": false, - "reason": "Near-duplicate wording is likely due to the shared template, but they still appear to be separate issue instances without proof of the same concrete request." + "reason": "These may share the same template, but there is no evidence they target the same language or same request." } ] }, @@ -14291,8 +14261,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-36817-14", "nodes": [ @@ -14341,202 +14311,202 @@ "issue:43837|issue:45027", "issue:43837|issue:45177" ], - "prepared_review_unit_hash": "65b828b3117a9d9d6cfbb4f9604d3489c66ca1d7a13e12b4647103307d97f1ee", + "prepared_review_unit_hash": "6bfbcd9cc0fffd721ccbbf98669bbd8dddcc1f3c86c5c206ed0d3c05d2935e36", "component_budget": { "node_count": 14, "item_count": 14, "soft_pair_count": 28, - "serialized_chars": 17644, - "estimated_input_tokens": 4411, - "estimated_eval_tokens": 9078 + "serialized_chars": 17646, + "estimated_input_tokens": 4412, + "estimated_eval_tokens": 9080 }, "budget": { "node_count": 14, "item_count": 14, "soft_pair_count": 28, - "serialized_chars": 17644, - "estimated_input_tokens": 4411, - "estimated_eval_tokens": 9078 + "serialized_chars": 17646, + "estimated_input_tokens": 4412, + "estimated_eval_tokens": 9080 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These items are thematically similar model-request issues, but they are not duplicates: each asks for a different model, task, or modality (e.g. BERT variants, audio ASR/TTS, OCR, sparse transformers).", - "confidence": 0.98, - "canonical_issue_reason": "No single canonical issue: the cluster spans distinct model requests rather than the same underlying bug/change.", + "summary": "This cluster is mostly a set of separate model-addition requests, not duplicates. The similarity signal is driven by generic wording like \u201cadd model\u201d rather than the same underlying bug/change.", + "confidence": 0.95, + "canonical_issue_reason": "No single canonical issue: each item requests a different model or modality (BERT variants, audio/TTS/ASR, OCR, vision/text), so they are not duplicates.", "canonical_pr_reason": null, - "best_issue_reason": "No clear global representative for deduplication because the requests target different models and code paths.", + "best_issue_reason": "No strong duplicate candidate exists. If a representative issue were needed, issue 43671 is one of the more specific and active requests, but it is still not a duplicate of the others.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:36817", "right": "issue:41211", "accept": false, - "reason": "EuroBERT config vs DEIMv2 are different model additions with different implementations." + "reason": "Different model families (EuroBert vs DEIMv2); generic \u201cadd model\u201d similarity only." }, { "left": "issue:36817", "right": "issue:43671", "accept": false, - "reason": "EuroBERT vs Qwen3-TTS are unrelated model families and modalities." + "reason": "EuroBert request is unrelated to Qwen3-TTS support; different model and task." }, { "left": "issue:37015", "right": "issue:43837", "accept": false, - "reason": "NeoBERT and Qwen3-ASR are different models; ASR support is not the same change as adding a BERT variant." + "reason": "NeoBERT and Qwen3-ASR are different models with different modalities/tasks." }, { "left": "issue:37015", "right": "issue:44121", "accept": false, - "reason": "NeoBERT vs a weight-sparse transformer request are different architectures and code paths." + "reason": "NeoBERT request is unrelated to OpenAI weight-sparse transformer support." }, { "left": "issue:37015", "right": "issue:43909", "accept": false, - "reason": "NeoBERT is unrelated to LFM2.5 Audio; different family and modality." + "reason": "NeoBERT and LFM2.5 Audio are different model requests, not the same change." }, { "left": "issue:37015", "right": "issue:42503", "accept": false, - "reason": "NeoBERT and ModernVBERT are different model requests, not the same bug/change." + "reason": "NeoBERT vs ModernVBERT: different model families; no shared concrete code path." }, { "left": "issue:37015", "right": "issue:45027", "accept": false, - "reason": "NeoBERT vs Voxtral TTS are unrelated model/support requests." + "reason": "NeoBERT and Voxtral-4B-TTS are separate model additions, not duplicates." }, { "left": "issue:41115", "right": "issue:43837", "accept": false, - "reason": "MiniCPM3 architecture support and Qwen3-ASR support are distinct model integrations." + "reason": "MiniCPM3 architecture support and Qwen3-ASR support are different models/tasks." }, { "left": "issue:41115", "right": "issue:44121", "accept": false, - "reason": "MiniCPM3 vs weight-sparse transformer are different requested changes." + "reason": "MiniCPM3 and circuit-sparsity/circuitgpt are unrelated model requests." }, { "left": "issue:41115", "right": "issue:42503", "accept": false, - "reason": "MiniCPM3 and ModernVBERT are separate model additions." + "reason": "MiniCPM3 vs ModernVBERT: different concrete model additions." }, { "left": "issue:41115", "right": "issue:45027", "accept": false, - "reason": "MiniCPM3 and Voxtral TTS target different models and modalities." + "reason": "MiniCPM3 and Voxtral-TTS are distinct model families and modalities." }, { "left": "issue:41211", "right": "issue:43671", "accept": false, - "reason": "DEIMv2 and Qwen3-TTS are unrelated model requests." + "reason": "DEIMv2 and Qwen3-TTS are different models with different target behaviors." }, { "left": "issue:42503", "right": "issue:43837", "accept": false, - "reason": "ModernVBERT and Qwen3-ASR are different model/task integrations." + "reason": "ModernVBERT models and Qwen3-ASR support are unrelated model additions." }, { "left": "issue:42503", "right": "issue:43671", "accept": false, - "reason": "ModernVBERT and Qwen3-TTS are unrelated model requests." + "reason": "ModernVBERT and Qwen3-TTS are different model families and tasks." }, { "left": "issue:42503", "right": "issue:42738", "accept": false, - "reason": "ModernVBERT is a concrete model request; BERT-like models with RoPE is a broader architecture discussion, not the same change." + "reason": "ModernVBERT and generic BERT-like RoPE support are not the same underlying change." }, { "left": "issue:42738", "right": "issue:45027", "accept": false, - "reason": "A generic BERT-with-RoPE request is unrelated to Voxtral TTS support." + "reason": "BERT-like RoPE support and Voxtral TTS are different problem domains." }, { "left": "issue:43671", "right": "issue:43837", "accept": false, - "reason": "Qwen3-TTS and Qwen3-ASR share a prefix but are different concrete support changes (TTS vs ASR)." + "reason": "Qwen3-TTS and Qwen3-ASR are related by name but are distinct models and separate support requests." }, { "left": "issue:43671", "right": "issue:45027", "accept": false, - "reason": "Qwen3-TTS and Voxtral TTS are different model integrations." + "reason": "Both are audio-related, but they target different models and integrations." }, { "left": "issue:43671", "right": "issue:45177", "accept": false, - "reason": "Qwen3-TTS and DeepSeek-OCR2 are different modalities and implementations." + "reason": "Qwen3-TTS support is unrelated to DeepSeek-OCR2." }, { "left": "issue:43671", "right": "issue:44121", "accept": false, - "reason": "Qwen3-TTS support is unrelated to a weight-sparse transformer request." + "reason": "Qwen3-TTS and weight-sparse transformer support are different model changes." }, { "left": "issue:43671", "right": "issue:43909", "accept": false, - "reason": "Qwen3-TTS and LFM2.5 Audio are different audio-model requests, not the same underlying change." + "reason": "Qwen3-TTS and LFM2.5 Audio 1.5B are separate audio-model requests." }, { "left": "issue:43837", "right": "issue:43909", "accept": false, - "reason": "Qwen3-ASR and LFM2.5 Audio are distinct audio model requests with different code paths." + "reason": "Qwen3-ASR and LFM2.5 Audio are distinct audio-model support requests." }, { "left": "issue:43837", "right": "issue:45322", "accept": false, - "reason": "Qwen3-ASR support and EUPE are unrelated model requests." + "reason": "Qwen3-ASR support and EUPE are different models with no shared concrete bug/change." }, { "left": "issue:43837", "right": "issue:44121", "accept": false, - "reason": "Qwen3-ASR and a weight-sparse transformer request are different changes." + "reason": "ASR model support is unrelated to the weight-sparse transformer request." }, { "left": "issue:43909", "right": "issue:44205", "accept": false, - "reason": "LFM2.5 Audio support and SAM3-LiteText are different model additions." + "reason": "LFM2.5 Audio 1.5B and SAM3-LiteText are different model additions." }, { "left": "issue:44121", "right": "issue:45322", "accept": false, - "reason": "Weight-sparse transformer and EUPE are unrelated research-model requests." + "reason": "Weight-sparse transformer/circuitgpt and EUPE are unrelated model requests." }, { "left": "issue:43837", "right": "issue:45027", "accept": false, - "reason": "Qwen3-ASR and Voxtral TTS differ by model and task; not a duplicate." + "reason": "Qwen3-ASR and Voxtral-4B-TTS are different speech-model support requests." }, { "left": "issue:43837", "right": "issue:45177", "accept": false, - "reason": "Qwen3-ASR and DeepSeek-OCR2 are different model requests." + "reason": "Qwen3-ASR support is unrelated to DeepSeek-OCR2." } ] }, @@ -14555,8 +14525,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-37934-6", "nodes": [ @@ -14577,82 +14547,82 @@ "issue:43668|issue:44031", "issue:43668|issue:43813" ], - "prepared_review_unit_hash": "bb2cc08e747d7e77db1527f9f7b77d434c7014832549efc89efb6073efcf1bd3", + "prepared_review_unit_hash": "5b1f6386955b5ba32a0377b1a492ffb28f6b886be3f8b15dd0673e606ef09c45", "component_budget": { "node_count": 6, "item_count": 6, "soft_pair_count": 8, - "serialized_chars": 6291, - "estimated_input_tokens": 1573, - "estimated_eval_tokens": 3402 + "serialized_chars": 6288, + "estimated_input_tokens": 1572, + "estimated_eval_tokens": 3400 }, "budget": { "node_count": 6, "item_count": 6, "soft_pair_count": 8, - "serialized_chars": 6291, - "estimated_input_tokens": 1573, - "estimated_eval_tokens": 3402 + "serialized_chars": 6288, + "estimated_input_tokens": 1572, + "estimated_eval_tokens": 3400 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These items are only loosely text-similar; they describe different bugs in different areas (model norms, loss handling, config typing, a typo, and tokenizer warnings), so they should not be deduplicated together.", + "summary": "These items are a heterogeneous set of unrelated issues (model norm behavior, Mask2Former loss args, config typing, a typo, and tokenizer regex warnings). The soft-similarity signals are too weak to treat any pair as duplicates.", "confidence": 0.98, - "canonical_issue_reason": "No single canonical issue: the cluster spans unrelated problems across different components and bug types.", + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "No issue here is a good global representative for the whole set, because they are not the same underlying defect.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:37934", "right": "issue:43668", "accept": false, - "reason": "Different models and defects: Llama4 norm semantics vs ModernBERTConfig type-hint correctness." + "reason": "Different concerns: Llama4TextL2Norm implementation semantics vs ModernBERTConfig type-hint correctness." }, { "left": "issue:42936", "right": "issue:43240", "accept": false, - "reason": "Mask2Former ignore_value usage is a separate loss-parameter bug from kwargs not being forwarded to the loss function." + "reason": "Mask2Former ignore_value handling is a model-specific bug; kwargs loss forwarding is a broader API wiring issue." }, { "left": "issue:42936", "right": "issue:44031", "accept": false, - "reason": "Unrelated subsystems: a model loss argument issue vs tokenizer regex warning behavior." + "reason": "Unrelated subsystems: Mask2Former loss parameter usage vs tokenizer regex warning regression." }, { "left": "issue:42936", "right": "issue:43813", "accept": false, - "reason": "Different underlying problem: unused ignore_value vs a typo in a quantization attribute name." + "reason": "Different bug class and area: model loss config usage vs a typo in a quantization field name." }, { "left": "issue:42936", "right": "issue:43668", "accept": false, - "reason": "These concern different models and bug classes, not the same code-path failure." + "reason": "No shared underlying defect; one is runtime behavior in a model, the other is a config annotation issue." }, { "left": "issue:43240", "right": "issue:43668", "accept": false, - "reason": "Loss kwargs forwarding and a config type-hint error are unrelated." + "reason": "Loss kwargs forwarding and incorrect type hints are separate issues with no common code-path." }, { "left": "issue:43668", "right": "issue:44031", "accept": false, - "reason": "A ModernBERT config annotation issue is not the same as the tokenizer regex warning bug." + "reason": "ModernBERT config typing is unrelated to tokenizer regex warnings." }, { "left": "issue:43668", "right": "issue:43813", "accept": false, - "reason": "Type-hint mistake vs literal typo in a different quantization-related string." + "reason": "A type-hint mismatch and a typo in a field path are distinct, non-duplicate issues." } ] }, @@ -14671,8 +14641,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-41084-2", "nodes": [ @@ -14682,7 +14652,7 @@ "soft_pairs": [ "issue:41084|issue:43037" ], - "prepared_review_unit_hash": "ff882ecfdd7846ea64e48964f8f053805d271ca58e6f3ba487a29533b675ca33", + "prepared_review_unit_hash": "ee99f254c5eef4357c55fdf3b8a5b18278fe1d7880966fd926d5d5b65a7f5df4", "component_budget": { "node_count": 2, "item_count": 2, @@ -14704,18 +14674,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are unrelated: one is about Block Decoding and the other is a request for DeepSeek v3.2 model support, so they should not be deduplicated.", - "confidence": 0.96, + "summary": "The two issues are unrelated: one requests block decoding behavior, the other requests DeepSeek v3.2 model support. They should not be clustered as duplicates.", + "confidence": 0.98, "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "If one had to be highlighted, #43037 is the more specific, 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 for DeepSeek v3.2. Similarity is too weak to treat as the same underlying issue." + "reason": "Different topics and likely different workstreams: block decoding vs DeepSeek v3.2 support; no evidence they describe the same underlying bug or feature." } ] }, @@ -14734,8 +14704,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-43598-2", "nodes": [ @@ -14745,7 +14715,7 @@ "soft_pairs": [ "issue:43598|issue:43600" ], - "prepared_review_unit_hash": "838b2c5d64b6819652f7fe8558f40fe979826318d318a7e5340d3e597809f1d9", + "prepared_review_unit_hash": "4fc6b68ff6bdc51c8e2bf39fd4be5dacd45bd1756e920b0db6f2790f88bf47b1", "component_budget": { "node_count": 2, "item_count": 2, @@ -14767,18 +14737,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are related only at a high level to Trainer customization, but they describe different problems: one is about `remove_unused_column`, the other about FLOPS estimation for custom trainers. They should not be clustered as duplicates.", - "confidence": 0.91, - "canonical_issue_reason": "Issue #43598 is the more specific customization concern around `remove_unused_column` in `Trainer`, and it is better suited as the representative issue for that narrow topic.", + "summary": "The two closed issues are about different Trainer-related customization problems: one concerns `remove_unused_column`, the other concerns FLOPS estimation for custom trainers. They share a broad area but not the same underlying bug or fix.", + "confidence": 0.95, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "#43598 is the clearer, more actionable report with a concrete Trainer behavior to revisit; #43600 is a separate concern about FLOPS accounting and should remain distinct.", + "best_issue_reason": "Neither issue is a duplicate of the other. If one must be picked as the representative, issue #43598 is the more concrete Trainer customization problem, but the cluster does not justify a single canonical issue.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43598", "right": "issue:43600", "accept": false, - "reason": "They mention Trainer customization, but one is about column removal behavior and the other about FLOPS estimation. These are different code paths and different fixes." + "reason": "Both mention Trainer customization, but #43598 is about `remove_unused_column` behavior while #43600 is about estimating FLOPS for custom trainers. They describe different code paths and different fixes." } ] }, @@ -14797,8 +14767,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-43684-2", "nodes": [ @@ -14808,7 +14778,7 @@ "soft_pairs": [ "issue:43684|issue:44405" ], - "prepared_review_unit_hash": "37218c8cb1bcd3773dedee66f9130321e3a0ad936562cfc73088da017d90561e", + "prepared_review_unit_hash": "581542f42486834982c1808003961f2e5b9891a7119ddd0536d4a8a1afa2cc0b", "component_budget": { "node_count": 2, "item_count": 2, @@ -14830,18 +14800,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two different feature requests for different Qwen model support paths: one for Qwen3-Omni auto-registration, the other for Qwen3.5 sequence-classification support. They are not duplicates.", + "summary": "These are two separate model-support requests: one for Qwen3-Omni auto-class registration, and one for Qwen3.5 sequence-classification support. They share the general theme of adding AutoModel support, but not the same concrete bug/change.", "confidence": 0.97, - "canonical_issue_reason": "Issue 43684 is the broader/cleaner anchor because it asks for AutoModel and AutoModelForConditionalGeneration registration support, while 44405 is a separate, task-specific AutoModelForSequenceClassification request for a different Qwen variant.", + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "43684 is the best representative if one issue must be chosen, since it targets general auto-class registration; 44405 is a distinct head-specific request and should not be merged into it.", + "best_issue_reason": "Issue 43684 is the broader auto-registration request and is the clearer representative of the Qwen3-Omni gap; issue 44405 is a distinct Qwen3.5 sequence-classification support request, not a duplicate.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43684", "right": "issue:44405", "accept": false, - "reason": "Different underlying changes: Qwen3-Omni registration for auto classes vs Qwen3.5 sequence-classification support. Different model family and different task head, so not the same bug or fix." + "reason": "Different model families and different auto classes: Qwen3-Omni registration for AutoModel/AutoModelForConditionalGeneration vs Qwen3.5 sequence-classification support. They are related only at a high level, not the same underlying change." } ] }, @@ -14860,8 +14830,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-43835-4", "nodes": [ @@ -14878,80 +14848,77 @@ "issue:43845|issue:43859", "issue:43846|issue:43859" ], - "prepared_review_unit_hash": "a92bcc515c8921ceed06ab44a9dcec731b333292f66b00fea325c9d5f709fc1e", + "prepared_review_unit_hash": "4aceadcf55ba70db28212dd61d0de279a6ca17591306cd06ab01a0a2f3431209", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4264, - "estimated_input_tokens": 1066, - "estimated_eval_tokens": 2388 + "serialized_chars": 4265, + "estimated_input_tokens": 1067, + "estimated_eval_tokens": 2390 }, "budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4264, - "estimated_input_tokens": 1066, - "estimated_eval_tokens": 2388 + "serialized_chars": 4265, + "estimated_input_tokens": 1067, + "estimated_eval_tokens": 2390 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Three issues are near-identical duplicate reports with the title \"huggingface \"; the \"Fett\" issue is unrelated on the available evidence. I\u2019d cluster the three huggingface issues together and keep them separate from 43835.", - "confidence": 0.71, - "canonical_issue_reason": "Issue 43859 is the best canonical representative among the duplicate huggingface reports because it matches the repeated title and has the most discussion activity, while the others appear to be the same report.", + "summary": "All four issues are extremely low-signal and mostly consist of vague or near-empty titles. Three share the same title ('huggingface '), while 'Fett' looks unrelated. There is not enough concrete content to confidently establish true duplicates, but 43845 is the best placeholder representative among the repeated-title issues.", + "confidence": 0.41, + "canonical_issue_reason": "Issue 43845 is the earliest of the repeated 'huggingface ' reports and is the cleanest placeholder representative for that cluster, though the bodies are too sparse to prove a real underlying bug.", "canonical_pr_reason": null, - "best_issue_reason": "43859 is the strongest issue to keep as the cluster representative: same duplicated title as the other matching issues and slightly more activity, so it is the most informative anchor for triage.", + "best_issue_reason": "43845 is the most suitable global issue representative because it is earliest among the near-identical 'huggingface ' issues and has no evidence of being different from 43846 or 43859.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43835", "right": "issue:43845", "accept": false, - "reason": "Different titles and no shared body or target evidence; not enough to treat \"Fett\" and \"huggingface\" as the same underlying issue." + "reason": "Titles are generic and unrelated ('Fett' vs 'huggingface '); no shared body details or targets to support a duplicate determination." }, { "left": "issue:43835", "right": "issue:43846", "accept": false, - "reason": "Different titles and no concrete evidence of the same bug/change." + "reason": "Same as above: no concrete evidence that 'Fett' and 'huggingface ' describe the same underlying issue." }, { "left": "issue:43835", "right": "issue:43859", "accept": false, - "reason": "Different titles and no shared issue target or descriptive evidence; should not be merged." + "reason": "No meaningful overlap beyond both being vague; insufficient evidence of the same bug." }, { "left": "issue:43845", "right": "issue:43846", - "accept": true, - "reason": "Same title, very high textual similarity, and no contrary evidence; these look like duplicate reports of the same issue." + "accept": false, + "reason": "The titles match, but the issue bodies are too thin to verify they are the same underlying problem; identical generic titles alone are not enough." }, { "left": "issue:43845", "right": "issue:43859", - "accept": true, - "reason": "Same title and strong similarity; these appear to be the same underlying report." + "accept": false, + "reason": "All we have is the same generic title; there are no concrete symptoms, targets, or code references to confirm duplication." }, { "left": "issue:43846", "right": "issue:43859", - "accept": true, - "reason": "Same title and strong similarity with no distinguishing details; likely duplicates." + "accept": false, + "reason": "Identical vague titles are suggestive, but the lack of substantive description makes a strict duplicate match unjustified." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The clustering is grounded in the packet: 43845/43846/43859 share the same title and strong soft-similarity evidence, while 43835 has a different title and only weak soft similarity. Choosing 43859 as the canonical representative is also conservative because it has the only discussion activity. " - }, - "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 }, @@ -14963,8 +14930,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-44371-5", "nodes": [ @@ -14985,88 +14952,88 @@ "issue:44572|issue:44937", "issue:44573|issue:44937" ], - "prepared_review_unit_hash": "b921ee5f6beb117d409d2218412d4ee66a9d39e33afbd400bd292145d865e555", + "prepared_review_unit_hash": "4d32738a40266427712efb1b54036146fd7e4aa3c6838bca7f06614f037eb622", "component_budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 9, - "serialized_chars": 5794, - "estimated_input_tokens": 1449, - "estimated_eval_tokens": 3154 + "serialized_chars": 5791, + "estimated_input_tokens": 1448, + "estimated_eval_tokens": 3152 }, "budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 9, - "serialized_chars": 5794, - "estimated_input_tokens": 1449, - "estimated_eval_tokens": 3154 + "serialized_chars": 5791, + "estimated_input_tokens": 1448, + "estimated_eval_tokens": 3152 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "All items look like unrelated low-signal or spam issues; there is no substantive evidence that any two describe the same underlying bug or request.", - "confidence": 0.98, - "canonical_issue_reason": "issue:44418 is the only item with a non-placeholder topical title, so it is the least bad representative if one must be chosen; however, it is still too generic to be a strong canonical issue.", + "summary": "All items look like unrelated low-signal/spam issues rather than duplicates of a real bug report. No soft edge should be merged.", + "confidence": 0.88, + "canonical_issue_reason": "issue:44371 is the simplest generic spam placeholder and the earliest item, so it is the least misleading representative for the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44418 is the most usable anchor only because it has an actual topic label rather than a spam placeholder; none of the issues are strong enough to serve as a meaningful canonical report.", + "best_issue_reason": "issue:44371 is the best available issue representative because it is generic, minimal, and there is no substantive technical content to anchor a more specific choice.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44371", "right": "issue:44572", "accept": false, - "reason": "Both titles are identical spam placeholders with no body/detail, so this is not evidence of the same underlying issue." + "reason": "Both are spam-like placeholders, but there is no evidence they describe the same underlying report or change." }, { "left": "issue:44371", "right": "issue:44573", "accept": false, - "reason": "Both are generic spam placeholders; similarity is not enough to infer a shared bug or request." + "reason": "Same generic spam style, but no shared bug, feature, or reproduction details to justify duplication." }, { "left": "issue:44371", "right": "issue:44418", "accept": false, - "reason": "One is a spam placeholder and the other is a vague documentation request; no concrete shared problem is shown." + "reason": "One is a generic spam title and the other is a documentation suggestion; too little evidence they are the same issue." }, { "left": "issue:44371", "right": "issue:44937", "accept": false, - "reason": "A spam placeholder and an unrelated branded title do not indicate the same underlying issue." + "reason": "Completely different titles and no technical details tying them to the same underlying problem." }, { "left": "issue:44418", "right": "issue:44572", "accept": false, - "reason": "A generic documentation suggestion and a spam placeholder are not the same substantive report." + "reason": "Documentation suggestion versus generic spam; not the same underlying issue." }, { "left": "issue:44418", "right": "issue:44573", "accept": false, - "reason": "The two titles are too generic and one is clear spam; there is no shared concrete issue." + "reason": "Different apparent intent and no shared concrete content beyond weak similarity signals." }, { "left": "issue:44572", "right": "issue:44573", "accept": false, - "reason": "Identical spam placeholders are not sufficient evidence of duplicate substantive content." + "reason": "Both are spam placeholders, but that is not enough to conclude they are duplicate reports of the same thing." }, { "left": "issue:44572", "right": "issue:44937", "accept": false, - "reason": "A spam placeholder and an unrelated brand-name title do not describe the same bug or change." + "reason": "Generic spam title versus a specific quoted name; no evidence of the same underlying issue." }, { "left": "issue:44573", "right": "issue:44937", "accept": false, - "reason": "No concrete overlap beyond generic spammy wording; not a duplicate pair." + "reason": "No concrete overlap in problem statement, targets, or artifacts; reject as duplicate." } ] }, @@ -15085,8 +15052,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "cache", - "cache_hit": true, + "source": "llm", + "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-45145-2", "nodes": [ @@ -15096,7 +15063,7 @@ "soft_pairs": [ "issue:45145|issue:45151" ], - "prepared_review_unit_hash": "b1517103c377d88d4edaf8b204c2f1f235d0c8fd3e57a6f9d9d62bc121f0178b", + "prepared_review_unit_hash": "350f17c51c7a1e5c256a646be320d3d7a061b73b25c701bc88548180a0457d1f", "component_budget": { "node_count": 2, "item_count": 2, @@ -15118,28 +15085,28 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two closed issues share an identical title and close body lengths, and 45145 is earlier with one inbound reference. That makes 45145 the better consolidation target, but the packet only provides soft-similarity evidence, so I would not treat the pair as conclusively duplicate with high confidence.", - "confidence": 0.78, - "canonical_issue_reason": "Issue 45145 is the earlier report and has an inbound reference, so it is the more established thread to keep as the canonical issue. The evidence here is limited to identical titles and soft similarity, so the choice is pragmatic rather than definitive.", + "summary": "The two closed issues appear to be likely duplicates or near-duplicate reports: they have the exact same title, similar body sizes, and a soft-similarity signal, with no distinguishing links, files, or targets shown. #45145 is the better representative because it was created earlier and has an inbound reference.", + "confidence": 0.88, + "canonical_issue_reason": "#45145 is the better canonical pick because it was created earlier and has an inbound reference, while #45151 appears to be a near-duplicate report with the same title.", "canonical_pr_reason": null, - "best_issue_reason": "45145 is the best issue representative because it predates 45151 and already has inbound linkage, while the available evidence does not strongly distinguish the two beyond soft similarity.", + "best_issue_reason": "#45145 is the best issue to keep as the representative item because it is earlier and already has an inbound reference, making it the more established anchor for this duplicate pair.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45145", "right": "issue:45151", - "accept": false, - "reason": "Reject: the packet only marks this as soft similarity and explicitly says deterministic acceptance is false. Identical titles and similar body lengths suggest a possible duplicate, but there is not enough direct content evidence to confidently merge them as the same report." + "accept": true, + "reason": "Exact same title plus similar body length and a soft-similarity signal make these look like duplicate reports of the same underlying issue." } ] }, "evaluator_result": { "accept": false, - "feedback": "The canonical choice is plausible, but the soft-edge verdict is too strong for the packet: the pair is only linked by soft similarity, and the packet explicitly marks the pair as not deterministically acceptable. The summary also overstates the evidence by saying the bodies are \"very similar\" / \"same report duplicated\" without direct content comparison in the packet. A more conservative stance would note identical titles, similar body lengths, and 45145\u2019s earlier timestamp plus inbound reference, but stop short of asserting duplication with high confidence." + "feedback": "The cluster call is a bit too strong for the evidence shown. The packet supports that these issues have the same title, similar sizes, and a soft-similarity signal, but it does not expose the bodies, so statements like \"highly similar content\" / \"substantially similar body content\" are not directly grounded. Also, the best-issue rationale overstates linkage/activity a bit: #45145 has the inbound reference, but it does not have more discussion activity than #45151. A more conservative summary would say they appear likely duplicate or near-duplicate reports, with #45145 as the better representative because it is earlier and has an inbound reference." }, "evaluator_used": true, "retried": true, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -15244,148 +15211,148 @@ "pull_request:43823|pull_request:45213", "pull_request:43251|pull_request:43254" ], - "prepared_review_unit_hash": "e2e9016b2ee1c7bbfea9c61a2473b7d80b8d1b9c856fa308ad187d5f07789db0", + "prepared_review_unit_hash": "4a04e67c3f56402c39cfd7388722cbeb74f085903a1eaee1b16a634240919561", "component_budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 18, - "serialized_chars": 35177, - "estimated_input_tokens": 8795, - "estimated_eval_tokens": 17846 + "serialized_chars": 35176, + "estimated_input_tokens": 8794, + "estimated_eval_tokens": 17844 }, "budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 18, - "serialized_chars": 35177, - "estimated_input_tokens": 8795, - "estimated_eval_tokens": 17846 + "serialized_chars": 35176, + "estimated_input_tokens": 8794, + "estimated_eval_tokens": 17844 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Most pairs are false positives from broad topical overlap. The only true duplicate pair is the two loss-utils patches (43251/43254) that both extend fixed_cross_entropy in the same code path. The Mistral4/PEFT pair should be rejected.", - "confidence": 0.77, + "summary": "Only one pair looks like a true duplicate: the two near-identical loss-utils PRs (43251/43254). The rest are unrelated same-subsystem changes, broad model additions, or docs/test-only updates and should not be merged as duplicates.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43251 is the strongest representative: it has the explicit issue link, makes the concrete fixed_cross_entropy change, and effectively covers the same kwargs-support fix as 43254.", + "canonical_pr_reason": "PR 43251 is the best canonical representative of the duplicate loss-utils fix: it has the explicit issue link, directly patches fixed_cross_entropy, and is the clearer of the two near-identical implementations.", "best_issue_reason": null, - "best_pr_reason": "PR 43251 is the best standalone representative of the actual duplicate cluster because it is issue-linked and more clearly frames the loss-function fix; the other accepted-looking pair is not a true duplicate.", + "best_pr_reason": "PR 43251 is the strongest overall PR in the cluster for duplicate-triage purposes because it most cleanly captures the shared bug in fixed_cross_entropy and has the most concrete issue alignment.", "soft_edge_verdicts": [ { "left": "pull_request:39895", "right": "pull_request:43424", "accept": false, - "reason": "VideoPrism model/docs addition vs Executorch dynamic-shape export test; unrelated changes." + "reason": "Videoprism model addition vs executorch dynamic-shape test; different features and code paths." }, { "left": "pull_request:41855", "right": "pull_request:43823", "accept": false, - "reason": "Mistral tokenizer helper methods vs adding a new MobileLLM model; different bugs/changes." + "reason": "Mistral tokenizer methods vs adding MobileLLM; unrelated changes with only superficial tokenizer/model overlap." }, { "left": "pull_request:41855", "right": "pull_request:43424", "accept": false, - "reason": "Tokenizer utility patch vs export test addition; no shared concrete code-path issue." + "reason": "Tokenizer API additions vs executorch export test; no shared bug or implementation change." }, { "left": "pull_request:42668", "right": "pull_request:43823", "accept": false, - "reason": "Processor auto-mapping robustness vs new model implementation; unrelated." + "reason": "Processor auto-loading robustness vs adding a new model; same broad area of model support, but not the same fix." }, { "left": "pull_request:43251", "right": "pull_request:43823", "accept": false, - "reason": "Loss utility kwargs fix vs new MobileLLM model code; different areas." + "reason": "Loss-function kwargs forwarding vs new MobileLLM model files; completely different code paths." }, { "left": "pull_request:43251", "right": "pull_request:44827", "accept": false, - "reason": "Both are code fixes, but one is cross-entropy kwargs and the other is Mistral4 test/model work; not the same underlying bug." + "reason": "Cross-entropy kwargs fix vs Mistral4 test/model refactor; unrelated bugs and changes." }, { "left": "pull_request:43424", "right": "pull_request:44827", "accept": false, - "reason": "Dynamic-shape export test vs Mistral4 test/model fix; unrelated." + "reason": "Executorch dynamic-shape test vs Mistral4 test/model refactor; different functionality and no common fix." }, { "left": "pull_request:43424", "right": "pull_request:43823", "accept": false, - "reason": "Executorch export test vs new model addition; not the same change." + "reason": "Executorch export testing vs MobileLLM model addition; unrelated changes." }, { "left": "pull_request:43823", "right": "pull_request:43995", "accept": false, - "reason": "New MobileLLM model addition vs Falcon output-collection refactor; different code paths." + "reason": "New MobileLLM model implementation vs Falcon output-collection refactor; different model families and problems." }, { "left": "pull_request:43823", "right": "pull_request:45622", "accept": false, - "reason": "MobileLLM model addition vs PEFT constructor fix; unrelated." + "reason": "MobileLLM model addition vs PEFT constructor fix; unrelated code paths." }, { "left": "pull_request:43995", "right": "pull_request:45213", "accept": false, - "reason": "Falcon refactor vs model-creation skill docs; no duplicate bug/change relationship." + "reason": "Falcon refactor vs model-creation skill workflow; not the same bug or change." }, { "left": "pull_request:43995", "right": "pull_request:44827", "accept": false, - "reason": "Falcon output-interface refactor vs Mistral4 tests/model work; different functionality." + "reason": "Falcon output interface refactor vs Mistral4 tests/compatibility fix; different models and goals." }, { "left": "pull_request:44827", "right": "pull_request:45622", "accept": false, - "reason": "Mistral4 tests/modeling changes vs PEFT constructor mapping fix; broad auto/modeling overlap only, not the same concrete bug." + "reason": "Mistral4 test/model fixes vs PEFT constructor fix; unrelated changes despite both touching framework internals." }, { "left": "pull_request:44827", "right": "pull_request:45587", "accept": false, - "reason": "Mistral4 fix vs docs-only memory-management article; unrelated." + "reason": "Mistral4-related fix vs continuous batching docs; one is code/tests, the other is documentation." }, { "left": "pull_request:44827", "right": "pull_request:45170", "accept": false, - "reason": "Mistral4 test/model fix vs layrnorm->layernorm renaming across CLIP-like models; different issues." + "reason": "Mistral4 test/model work vs CLIP-like weight-renaming fix; different subsystems and concrete issues." }, { "left": "pull_request:44827", "right": "pull_request:45213", "accept": false, - "reason": "Mistral4 fix vs model-creation skill docs/dataset; unrelated." + "reason": "Mistral4 fix vs a draft model-creation skill PR; not the same underlying change." }, { "left": "pull_request:43823", "right": "pull_request:45213", "accept": false, - "reason": "New MobileLLM model addition vs skill/docs scaffold for model creation; not the same underlying change." + "reason": "MobileLLM model addition vs model-creation skill scaffolding; one creates a model, the other adds workflow tooling." }, { "left": "pull_request:43251", "right": "pull_request:43254", "accept": true, - "reason": "Both patches modify src/transformers/loss/loss_utils.py to forward the same cross_entropy kwargs (weight/label_smoothing); same underlying loss-function fix." + "reason": "Both PRs patch the same fixed_cross_entropy code path to forward supported cross_entropy kwargs; the diffs are near-identical and could be merged into one fix." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. The only accepted pair, 43251/43254, is strongly supported by the shared issue target and identical loss_utils.py / fixed_cross_entropy change. The rejected pairs are reasonably treated as topical overlap rather than duplicates, including the Mistral4/PEFT case." + "feedback": "Grounded and conservative. The only accepted duplicate pair, 43251/43254, is supported by the shared loss_utils file, the shared issue target, and near-identical fixed_cross_entropy patches. The other pairs are reasonably rejected as different bugfixes, features, or docs/tests rather than mergeable duplicates." }, "evaluator_used": true, "retried": false, @@ -15412,7 +15379,7 @@ "soft_pairs": [ "pull_request:40546|pull_request:42781" ], - "prepared_review_unit_hash": "df47438e1fbe94cdcddf3aebd1e9314ff773653cd243068b0f9749e349daf614", + "prepared_review_unit_hash": "1584b7ac07cc91c523316348d74a8bea0ef16322c7d9b77f3ae17705ba132ace", "component_budget": { "node_count": 2, "item_count": 2, @@ -15434,18 +15401,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two related VibeVoice PRs, but they are not duplicates: one adds the initial VibeVoice integration, while the other adds a separate realtime variant with additional acoustic-tokenizer and generation changes.", - "confidence": 0.87, + "summary": "These are related VibeVoice PRs, but they are not duplicates: one is the core VibeVoice implementation, the other is a draft for VibeVoice Realtime that adds extra realtime-specific files and changes. The shared filenames mostly reflect the common model family, not the same exact change.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #40546 is the earlier, broader base VibeVoice integration and looks like the main canonical PR for the model bring-up; #42781 is a later, narrower realtime extension.", + "canonical_pr_reason": "PR #40546 is the better canonical representative because it is the fuller, reviewed core VibeVoice implementation with broad model and docs integration.", "best_issue_reason": null, - "best_pr_reason": "#40546 is the best single PR to represent the cluster because it is the foundational implementation and appears closer to a complete standalone addition.", + "best_pr_reason": "PR #40546 is the best overall representative of the cluster: it is the main implementation PR, while #42781 is a draft follow-up for a more specific realtime variant.", "soft_edge_verdicts": [ { "left": "pull_request:40546", "right": "pull_request:42781", "accept": false, - "reason": "They share VibeVoice scaffolding, but #42781 adds a realtime-specific path and extra acoustic-tokenizer work rather than fixing the same concrete code-path problem. They are related follow-on changes, not one mergeable duplicate PR." + "reason": "Reject: they share the VibeVoice codebase, but #42781 is a separate draft for VibeVoice Realtime with additional realtime-specific files and changes, not the same concrete code-path fix/change as #40546." } ] }, @@ -15475,7 +15442,7 @@ "soft_pairs": [ "pull_request:40755|pull_request:41224" ], - "prepared_review_unit_hash": "6b9611ca6fe2cd8e3685f2f0bf3ebcf3cce03d23f360f783e2bfc5f1be6f1961", + "prepared_review_unit_hash": "c78a4e744c187bc6677f105e311280c04222ff551f2e406baf665f4f8b2f3f0e", "component_budget": { "node_count": 2, "item_count": 2, @@ -15497,18 +15464,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two PRs are not duplicates: one adds TimesFM covariate forecasting support across modeling code and tests, while the other adds DINOv3 image classification support and docs. They touch different models, code paths, and feature areas.", + "summary": "The pair is not a duplicate cluster: one PR adds TimesFM covariate forecasting, the other adds DINOv3 image classification support. They target different models and different code paths.", "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: the PRs are unrelated feature additions for different models (TimesFM vs DINOv3) and cannot be merged into one change.", + "canonical_pr_reason": "PR 41224 is the better canonical representative because it is a focused, self-contained model feature with substantial review activity and a clear code/test surface.", "best_issue_reason": null, - "best_pr_reason": "No best PR for a duplicate cluster; neither PR is a representative duplicate of the other since they address different concrete tasks.", + "best_pr_reason": "PR 41224 is the strongest standalone PR in the cluster: it is narrowly scoped, implementation-backed, and has more review evidence than the broader TimesFM feature PR.", "soft_edge_verdicts": [ { "left": "pull_request:40755", "right": "pull_request:41224", "accept": false, - "reason": "Different models and implementation targets: TimesFM covariate forecasting vs DINOv3 image classification; no shared bug or mergeable code-path fix." + "reason": "Different features for different models (TimesFM covariates vs DINOv3 image classification); no shared bug or change to merge into one PR." } ] }, @@ -15604,12 +15571,12 @@ "soft_pairs": [ "pull_request:41797|pull_request:45186" ], - "prepared_review_unit_hash": "7b87d376f074f7376005cdd281a6a13b13603cf15b42f260e2b542f64e3213b4", + "prepared_review_unit_hash": "61c25d0acdfc8da32f5afbba697ca7ba09864588c591c5a3f93f2131a9c1dbff", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6723, + "serialized_chars": 6722, "estimated_input_tokens": 1681, "estimated_eval_tokens": 3618 }, @@ -15617,7 +15584,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6723, + "serialized_chars": 6722, "estimated_input_tokens": 1681, "estimated_eval_tokens": 3618 }, @@ -15626,18 +15593,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two pull requests are both model-addition PRs with similar scaffolding, but they implement different models and distinct code paths (DeepSeekOCR vs Isaac), so they are not duplicates.", - "confidence": 0.97, + "summary": "These are two distinct model-addition pull requests: one for DeepSeek OCR and one for Isaac. They share only common Transformers integration boilerplate, not the same underlying code change.", + "confidence": 0.98, "canonical_issue_reason": null, "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "No single canonical PR: 41797 adds DeepSeekOCR while 45186 adds Isaac, so they do not represent the same underlying change.", + "best_pr_reason": "PR 41797 is the stronger standalone representative of the cluster because it has broader implementation scope and more review activity, but it is not a duplicate of PR 45186.", "soft_edge_verdicts": [ { "left": "pull_request:41797", "right": "pull_request:45186", "accept": false, - "reason": "Different models and implementation files; shared registry/docs updates are generic boilerplate, not the same fix or change." + "reason": "Different models and different codepaths: DeepSeek OCR vs Isaac. Shared files are generic registry/docs plumbing, not evidence of the same fix or mergeable change." } ] }, @@ -15667,12 +15634,12 @@ "soft_pairs": [ "pull_request:42166|pull_request:45426" ], - "prepared_review_unit_hash": "ccf8e9a9042ed6027ffa6b08a388077669a8076fb7b8e7e8851b7f3cf973f5c9", + "prepared_review_unit_hash": "42e6a3ab10d2a44c0bd004cfe31c9983aa8b6a61bfd6c988d5a1f98e4a44d7f5", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5408, + "serialized_chars": 5407, "estimated_input_tokens": 1352, "estimated_eval_tokens": 2960 }, @@ -15680,7 +15647,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5408, + "serialized_chars": 5407, "estimated_input_tokens": 1352, "estimated_eval_tokens": 2960 }, @@ -15689,18 +15656,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both items are independent model-integration PRs that share Transformers boilerplate but add different models (InternVLFlash vs AXK1). They are not duplicates and should not be clustered as the same change.", - "confidence": 0.97, + "summary": "The two PRs are separate model-addition workstreams (InternVLFlash vs AXK1). They overlap only in shared registry/boilerplate files, not in the underlying change or code path.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 42166 is a broader, earlier model-addition PR with more discussion/review activity; it\u2019s the better canonical representative only in a bookkeeping sense, not because it duplicates 45426.", + "canonical_pr_reason": "No canonical PR: these are different model integrations with different implementation files and different goals, so they are not duplicates.", "best_issue_reason": null, - "best_pr_reason": "PR 42166 is the strongest representative of the cluster since it is the larger, more established model-addition PR. 45426 is separate work for a different model family.", + "best_pr_reason": "No single best PR for deduplication. If forced to pick a representative, 42166 is more mature because it includes docs and tests, but it is still unrelated to 45426.", "soft_edge_verdicts": [ { "left": "pull_request:42166", "right": "pull_request:45426", "accept": false, - "reason": "They add different models and different code paths. Shared auto-registration and package boilerplate is not enough to treat them as the same underlying fix/change." + "reason": "Shared auto-registry and package plumbing only; one adds InternVLFlash and the other adds AXK1, so they do not fix or implement the same underlying code-path change." } ] }, @@ -15737,12 +15704,12 @@ "pull_request:43291|pull_request:44815", "pull_request:43665|pull_request:44815" ], - "prepared_review_unit_hash": "8359a59d22c1c115b6f55cf0ca32c53a2d7d580d432eafae6a2d4ff2f0ef6ad1", + "prepared_review_unit_hash": "5f5636f86a017ef44cd3ec8295f538556a3761fb687246323de9490bc94a8e1a", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 10170, + "serialized_chars": 10169, "estimated_input_tokens": 2543, "estimated_eval_tokens": 5342 }, @@ -15750,7 +15717,7 @@ "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 10170, + "serialized_chars": 10169, "estimated_input_tokens": 2543, "estimated_eval_tokens": 5342 }, @@ -15759,48 +15726,48 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are not duplicates of one another: one fixes Whisper ASR return_language plumbing, two are unrelated Whisper/test-only edits, and one is a separate FP8 dequant/loading fix. The cluster appears to be a loose soft-similarity bucket rather than a single change set.", - "confidence": 0.92, + "summary": "These PRs do not look like duplicates of the same underlying change. Two are Whisper-related, one is generic test noise, and one is an unrelated FP8/dequantization fix. Only the Whisper return-language PR appears to be a real bug fix; the others are too different in code path and intent.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #42227 is the clearest canonical representative because it has a specific bug fix, an explicit linked issue, and focused code changes in the Whisper ASR pipeline.", + "canonical_pr_reason": "PR 42227 is the clearest substantive fix: it targets a specific Whisper ASR bug, has an explicit issue link, and changes the actual pipeline code path rather than only tests or unrelated subsystems.", "best_issue_reason": null, - "best_pr_reason": "PR #42227 is the best single PR to represent the cluster: it is the most concrete user-facing bug fix and is substantially more well-scoped than the other test-heavy or unrelated changes.", + "best_pr_reason": "PR 42227 is the best representative artifact in this cluster because it is a concrete production fix with issue linkage and clear behavior change; the others are either test-only/junk-like or unrelated to the same bug.", "soft_edge_verdicts": [ { "left": "pull_request:42227", "right": "pull_request:43291", "accept": false, - "reason": "Both touch Whisper, but #42227 fixes ASR return_language propagation while #43291 changes tokenizer decode/test behavior; they are different code paths and not the same underlying change." + "reason": "Both are Whisper-related, but they fix different code paths: 42227 propagates return_language through the ASR pipeline, while 43291 changes tokenizer decode/timestamp handling and tests. They do not look mergeable as one duplicate PR." }, { "left": "pull_request:42227", "right": "pull_request:43665", "accept": false, - "reason": "#43665 is just vague test edits across CLIP/ViT and does not address Whisper return_language or the same bug." + "reason": "42227 is a real Whisper pipeline bug fix; 43665 is just placeholder test edits across CLIP/ViT and does not address the same underlying change." }, { "left": "pull_request:42227", "right": "pull_request:44815", "accept": false, - "reason": "These are unrelated fixes: Whisper ASR return_language versus FP8 dequant/core loading behavior." + "reason": "These are unrelated subsystems and problems: Whisper ASR parameter propagation versus FP8/dequant/model-loading changes. No shared concrete bug." }, { "left": "pull_request:43291", "right": "pull_request:43665", "accept": false, - "reason": "Different models and different purposes; one is Whisper tokenizer/test work, the other is generic model test edits." + "reason": "43291 modifies Whisper tokenizer behavior; 43665 is test noise in vision model tests. Different intent and no shared underlying bug." }, { "left": "pull_request:43291", "right": "pull_request:44815", "accept": false, - "reason": "No shared concrete bug or code path; Whisper decoding changes are unrelated to FP8 dequant/loading." + "reason": "Whisper tokenizer decode/timestamp logic is unrelated to FP8 dequantization/model loading. They do not fix the same issue." }, { "left": "pull_request:43665", "right": "pull_request:44815", "accept": false, - "reason": "Generic model test changes are not the same underlying fix as the FP8 dequant/loading PR." + "reason": "One is placeholder test edits, the other is a substantive quantization/dequant fix. Not the same bug or change." } ] }, @@ -15830,7 +15797,7 @@ "soft_pairs": [ "pull_request:42975|pull_request:42976" ], - "prepared_review_unit_hash": "36f6a00b81b4e03833c6378e53ce19875beb54a8acc791f25f7e3b926fcf2ef6", + "prepared_review_unit_hash": "92ef2a134638e3629d7417b08d4d00d2953b55d7b40a09cfb4cf2a4f1ad25b20", "component_budget": { "node_count": 2, "item_count": 2, @@ -15852,18 +15819,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs are CI/workflow dependency bumps, but they upgrade different GitHub Actions and solve different maintenance tasks. The overlap in workflow files is incidental, not evidence of the same underlying change.", - "confidence": 0.97, + "summary": "These are related workflow-maintenance PRs, but they change different GitHub Actions dependencies and solve different upgrade concerns. They are similar in theme, not the same underlying change.", + "confidence": 0.89, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: #42975 upgrades checkout/cache/upload-artifact for Node 24 compatibility, while #42976 bumps docker/build-push-action and related actions; they are distinct changes.", + "canonical_pr_reason": "PR #42975 is the broader, more specific workflow-runtime upgrade: it updates several core actions for Node 24 compatibility and touches more workflow entry points. It is a better umbrella representative of this maintenance cluster than the narrower Docker-action bump in #42976.", "best_issue_reason": null, - "best_pr_reason": "If a single representative were needed, #42975 is the broader workflow-action compatibility upgrade, but it still should not be merged with #42976 as a duplicate.", + "best_pr_reason": "PR #42975 is the best single representative because its title and diff clearly describe the main compatibility-driven Actions upgrade, while #42976 is a smaller, separate dependency refresh focused on docker/build-push-action.", "soft_edge_verdicts": [ { "left": "pull_request:42975", "right": "pull_request:42976", "accept": false, - "reason": "Different action families and different upgrade goals: Node 24 compatibility for general workflow actions vs. Docker build action version bumps. Not the same concrete fix and not plausibly one PR." + "reason": "Both are GitHub Actions version bumps, but they update different actions in different workflows (core checkout/cache/upload-artifact vs docker/build-push-action). They are related maintenance, not the same concrete bug or change, so I would not merge them as duplicates." } ] }, @@ -15885,7 +15852,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-42978-10", + "cluster_id": "cluster-42978-9", "nodes": [ "pull_request:42978", "pull_request:43448", @@ -15894,7 +15861,6 @@ "pull_request:44601", "pull_request:45283", "pull_request:45490", - "pull_request:45616", "pull_request:45626", "pull_request:45630" ], @@ -15904,9 +15870,7 @@ "pull_request:42978|pull_request:45490", "pull_request:43448|pull_request:43451", "pull_request:43448|pull_request:45490", - "pull_request:43448|pull_request:45616", "pull_request:43451|pull_request:45490", - "pull_request:43451|pull_request:45616", "pull_request:43973|pull_request:45283", "pull_request:44601|pull_request:45626", "pull_request:45283|pull_request:45626", @@ -15914,118 +15878,106 @@ "pull_request:45490|pull_request:45630", "pull_request:45626|pull_request:45630" ], - "prepared_review_unit_hash": "eddad9e46fa11dfe279d595e60e93121adc1de3cd9310aa8d4e89d0c9e8b4468", + "prepared_review_unit_hash": "f1e612125d6dfcc47d82456fe0eeb0b1ba89e20ee1de622adcdaabf7fb26103a", "component_budget": { - "node_count": 10, - "item_count": 10, - "soft_pair_count": 14, - "serialized_chars": 33117, - "estimated_input_tokens": 8280, - "estimated_eval_tokens": 16816 + "node_count": 9, + "item_count": 9, + "soft_pair_count": 12, + "serialized_chars": 29350, + "estimated_input_tokens": 7338, + "estimated_eval_tokens": 14932 }, "budget": { - "node_count": 10, - "item_count": 10, - "soft_pair_count": 14, - "serialized_chars": 33117, - "estimated_input_tokens": 8280, - "estimated_eval_tokens": 16816 + "node_count": 9, + "item_count": 9, + "soft_pair_count": 12, + "serialized_chars": 29350, + "estimated_input_tokens": 7338, + "estimated_eval_tokens": 14932 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is not a true duplicate set: it mixes several unrelated model-addition PRs with one distributed loading change, and the only overlap is standard Transformers boilerplate (docs, auto-registry, tests).", - "confidence": 0.95, + "summary": "These PRs are not duplicates: they cover different model additions and one distributed-loading feature, with only shared boilerplate in docs/auto registration. All soft edges should be rejected.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "No single PR is a valid canonical representative for this cluster; the items target different models or different infrastructure code paths, so shared file patterns are superficial.", + "canonical_pr_reason": "No single canonical PR fits this cluster; the items are distinct feature/model additions rather than the same underlying change.", "best_issue_reason": null, - "best_pr_reason": "No PR is globally suitable as the best representative duplicate: the PRs span distinct model integrations and a separate distributed-parallel feature, so there is no one change that subsumes the rest.", + "best_pr_reason": "No PR is a good global representative because the similarities are superficial (docs, auto-mapping, tests) and the concrete code paths differ.", "soft_edge_verdicts": [ { "left": "pull_request:42978", "right": "pull_request:43448", "accept": false, - "reason": "ViT NEPA vs Molmo are different model additions; shared auto/docs files are routine scaffolding, not the same underlying change." + "reason": "Different models (ViT NEPA vs Molmo); overlap is only generic model-registration/docs boilerplate." }, { "left": "pull_request:42978", "right": "pull_request:43451", "accept": false, - "reason": "ViT NEPA and Molmo2 are unrelated model integrations with different architectures and filenames beyond boilerplate registry/docs edits." + "reason": "Different models (ViT NEPA vs Molmo2) with distinct files and code paths." }, { "left": "pull_request:42978", "right": "pull_request:45490", "accept": false, - "reason": "ViT NEPA model support is unrelated to CTSM time-series model support; overlap is only generic model-registration plumbing." + "reason": "ViT NEPA model addition vs CTSM model addition; not the same bug/change." }, { "left": "pull_request:43448", "right": "pull_request:43451", "accept": false, - "reason": "Molmo and Molmo2 are separate model families and separate codebases; this is not one concrete change that could be merged into a single PR." + "reason": "Molmo and Molmo2 are separate model support PRs, not one fix split across two changes." }, { "left": "pull_request:43448", "right": "pull_request:45490", "accept": false, - "reason": "Molmo vs CTSM are different model additions with different implementation files and purposes." - }, - { - "left": "pull_request:43448", - "right": "pull_request:45616", - "accept": false, - "reason": "Molmo and DeepSeek V4 are distinct model additions; shared registry/docs edits do not indicate the same bug or feature." + "reason": "Molmo vs CTSM are unrelated model additions; shared auto/model-doc edits are boilerplate." }, { "left": "pull_request:43451", "right": "pull_request:45490", "accept": false, - "reason": "Molmo2 and CTSM are unrelated model additions; similar auto-mapping changes are standard boilerplate across model PRs." - }, - { - "left": "pull_request:43451", - "right": "pull_request:45616", - "accept": false, - "reason": "Molmo2 and DeepSeek V4 are separate model integrations, not one shared code-path fix." + "reason": "Molmo2 and CTSM are different models with different implementation files and targets." }, { "left": "pull_request:43973", "right": "pull_request:45283", "accept": false, - "reason": "Lfm2 Audio model support and Qwen3.5 GGUF loading support address different subsystems and cannot plausibly be merged as one PR." + "reason": "Lfm2Audio model support and Qwen3.5 GGUF loading are different feature areas and code paths." }, { "left": "pull_request:44601", "right": "pull_request:45626", "accept": false, - "reason": "Native pipeline-parallel loading support is an infrastructure change, while PP-FormulaNet is a model addition; same prefix/boilerplate only." + "reason": "Pipeline-parallel infrastructure work is unrelated to PP-FormulaNet model support." }, { "left": "pull_request:45283", "right": "pull_request:45626", "accept": false, - "reason": "Qwen3.5 GGUF loading support and PP-FormulaNet model support are unrelated changes in different code paths." + "reason": "GGUF loading support for Qwen3.5 and PP-FormulaNet model integration are distinct changes." }, { "left": "pull_request:45490", "right": "pull_request:45626", "accept": false, - "reason": "CTSM model support and PP-FormulaNet model support are distinct model PRs with no shared underlying bug or feature." + "reason": "CTSM model support and PP-FormulaNet model support are separate model additions." }, { "left": "pull_request:45490", "right": "pull_request:45630", "accept": false, - "reason": "CTSM and Kimi2-6 are separate model additions; overlap in docs/auto registry is not enough to treat them as duplicates." + "reason": "CTSM and Kimi2-6 are different models; similarity is limited to standard auto/docs scaffolding." }, { "left": "pull_request:45626", "right": "pull_request:45630", "accept": false, - "reason": "PP-FormulaNet and Kimi2-6 are different model integrations with different implementation and test files." + "reason": "PP-FormulaNet and Kimi2-6 are unrelated model support PRs with different implementation files." } ] }, @@ -16058,52 +16010,52 @@ "pull_request:43488|pull_request:45476", "pull_request:43532|pull_request:45476" ], - "prepared_review_unit_hash": "c50657b4c0635e36df3f01206a49b3389b301562359b34f557e787f7819e933a", + "prepared_review_unit_hash": "2d6920781c3cc68e5504d58e25f0d34fe74baf8a6f33939e3ddfa6159af59186", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 7469, - "estimated_input_tokens": 1868, - "estimated_eval_tokens": 3992 + "serialized_chars": 7468, + "estimated_input_tokens": 1867, + "estimated_eval_tokens": 3990 }, "budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 7469, - "estimated_input_tokens": 1868, - "estimated_eval_tokens": 3992 + "serialized_chars": 7468, + "estimated_input_tokens": 1867, + "estimated_eval_tokens": 3990 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "All three items are unrelated \"don't merge\" test PRs with different fabricated changes; the only overlap is a shared test filename, which is not enough to treat them as duplicates.", - "confidence": 0.92, + "summary": "All three are unrelated synthetic \"don't merge\" PRs that only overlap on broad test/workflow areas and a shared vit test file; none look like the same underlying change or bug fix.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43532 is the most central of the three because it touches multiple workflow/testing files and has the strongest overlap with another item, but it is still not a duplicate of the others.", + "canonical_pr_reason": "No canonical PR: these are separate, non-duplicative bot/demo changes with different purposes (setup/versioning, workflow display tweaks, and CI/workflow invocation plus unrelated code noise).", "best_issue_reason": null, - "best_pr_reason": "PR 43532 is the best representative PR for the cluster because it is the most broadly connected artifact and shares the clearest superficial overlap with 45476, though the underlying changes are still distinct.", + "best_pr_reason": "No single PR is a good cluster representative for deduplication because the overlap is superficial and the concrete code paths differ.", "soft_edge_verdicts": [ { "left": "pull_request:43488", "right": "pull_request:43532", "accept": false, - "reason": "Different fake changes: 43488 edits packaging/versioning and dependency tables, while 43532 edits CI workflows and testing utilities. Shared vit test file is incidental." + "reason": "Shared vit test filename is not enough: 43488 is a setup/dependency/version-formatting PR, while 43532 changes workflow automation and testing utilities. Different underlying changes." }, { "left": "pull_request:43488", "right": "pull_request:45476", "accept": false, - "reason": "No same underlying bug or change: 43488 is about setup/version/deps, whereas 45476 adds a PR CI workflow and unrelated dummy code in CLIP/Chinese-CLIP files." + "reason": "43488 only tweaks setup metadata and dependency listing; 45476 adds a workflow caller and unrelated tokenization/test noise. They do not fix the same concrete problem." }, { "left": "pull_request:43532", "right": "pull_request:45476", "accept": false, - "reason": "Even though both touch tests/workflows, they do not fix the same concrete code-path problem and could not plausibly be merged as one PR; the overlap is only superficial." + "reason": "Both are synthetic PRs with some overlapping test/workflow files, but they touch different code paths and goals; no evidence they could be merged into one fix." } ] }, @@ -16133,12 +16085,12 @@ "soft_pairs": [ "pull_request:43612|pull_request:43613" ], - "prepared_review_unit_hash": "fdba583e59f3fa8ab5a5c919ba748fc9b17a4268d67291becc98c258df0af3ef", + "prepared_review_unit_hash": "04e50e9bc2e5702c735c83fcbae1eb6f68ec7ee18f74340263f330d25072ba0d", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6082, + "serialized_chars": 6083, "estimated_input_tokens": 1521, "estimated_eval_tokens": 3298 }, @@ -16146,7 +16098,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6082, + "serialized_chars": 6083, "estimated_input_tokens": 1521, "estimated_eval_tokens": 3298 }, @@ -16155,18 +16107,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two related but distinct PRs add new segmentation pipelines: one for promptable concept segmentation and one for promptable visual segmentation. They share documentation and pipeline-registration scaffolding, but they target different tasks and model integrations, so they should not be deduplicated.", - "confidence": 0.98, + "summary": "These are two closely named but distinct pipeline additions: one for Promptable Concept Segmentation and one for Promptable Visual Segmentation. They share documentation and registry plumbing, but the actual pipeline implementations, model integrations, and tests target different tasks, so they should not be deduplicated.", + "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "No true duplicate PR pair here; if selecting a single representative, PR 43613 is the stronger/broader one because it adds the visual-segmentation pipeline with multiple model integrations and dedicated tests/docs.", "best_issue_reason": null, - "best_pr_reason": "PR #43613 is the broader, more representative pipeline addition, with wider model coverage and more supporting tests/docs, but it is still a separate feature from #43612 rather than a duplicate.", + "best_pr_reason": "PR 43613 is the best representative of the cluster because it is the more complete feature branch and has clearer task scope with broader supporting changes.", "soft_edge_verdicts": [ { "left": "pull_request:43612", "right": "pull_request:43613", "accept": false, - "reason": "Different underlying features and codepaths: concept segmentation vs visual segmentation. Shared docs/registry updates are not enough to treat them as one mergeable PR." + "reason": "Different underlying features and code paths: concept segmentation vs visual segmentation. Shared docs/registry updates and similar naming are not enough to treat them as the same change, and they would not plausibly be merged as one PR." } ] }, @@ -16196,7 +16148,7 @@ "soft_pairs": [ "pull_request:43651|pull_request:43663" ], - "prepared_review_unit_hash": "f5095a79bd1f3c88eb0dd740145ee5d95837afbf44f3bd496a992d75cce17024", + "prepared_review_unit_hash": "e3384a568fbb81b6b0f7327a98b63eb89495861b768e696e2a60b7b6abd621a2", "component_budget": { "node_count": 2, "item_count": 2, @@ -16218,18 +16170,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both pull requests modify `src/transformers/trainer.py`, but they address different trainer extension points: one changes gradient-accumulation loss scaling, the other adds a customizable signature-column helper for column filtering. They are not the same underlying bug/change.", - "confidence": 0.96, + "summary": "Two open PRs in the same trainer file, but they address unrelated Trainer extensibility points: gradient-accumulation loss scaling vs. signature-column filtering. The soft similarity is file-level only, not a shared underlying bug/change.", + "confidence": 0.98, "canonical_issue_reason": null, "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "No single PR is a clear canonical representative because the two changes are unrelated and should remain separate.", + "best_pr_reason": "PR #43651 is the more concrete change: it fixes a specific training-step loss-scaling behavior and introduces a narrowly scoped hook. PR #43663 is also valid, but it targets a different Trainer concern.", "soft_edge_verdicts": [ { "left": "pull_request:43651", "right": "pull_request:43663", "accept": false, - "reason": "Same file, but different code paths and different intents: loss scaling vs. signature-column filtering. They could not plausibly be merged as one PR without mixing unrelated changes." + "reason": "Different code paths and behaviors: #43651 changes gradient accumulation loss normalization, while #43663 changes how Trainer derives signature columns for input filtering. They are not the same bug or mergeable into one fix." } ] }, @@ -16261,56 +16213,56 @@ "pull_request:43656|pull_request:43842", "pull_request:43836|pull_request:43842" ], - "prepared_review_unit_hash": "7ae6a917abf72373cdd61f66d0d5a4f1a863dd854e8afa54d1365a6c2a82e3b9", + "prepared_review_unit_hash": "466e6bdca13e43bf071fdca69e92afa07d3be700cbbd2cdb5a0544abdf7fc0d3", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 5992, - "estimated_input_tokens": 1498, - "estimated_eval_tokens": 3252 + "serialized_chars": 5993, + "estimated_input_tokens": 1499, + "estimated_eval_tokens": 3254 }, "budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 5992, - "estimated_input_tokens": 1498, - "estimated_eval_tokens": 3252 + "serialized_chars": 5993, + "estimated_input_tokens": 1499, + "estimated_eval_tokens": 3254 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "All three PRs appear to be alternative fixes for the same TypeAdapter-related NameError in `src/transformers/cli/serve.py`, with 43842 the most direct and focused patch.", - "confidence": 0.79, + "summary": "PR #43842 is the strongest canonical choice: it makes the most targeted fix for the `TypeAdapter` NameError in `serve.py`. PRs #43656 and #43836 are related in the same area, but the packet does not conservatively establish that they are the same underlying fix, so they should be treated as possibly overlapping rather than definitively duplicate.", + "confidence": 0.87, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43842 is the cleanest, most targeted fix: it explicitly handles `TypeAdapter` under `TYPE_CHECKING` and avoids runtime NameError with minimal scope.", + "canonical_pr_reason": "PR #43842 is the cleanest minimal fix: it directly addresses the `TypeAdapter` runtime/import issue with a small, explicit change in `src/transformers/cli/serve.py` and avoids the broader refactoring present in the other PRs.", "best_issue_reason": null, - "best_pr_reason": "43842 is the best standalone PR to keep because it directly addresses the runtime typing import problem with the smallest, clearest change set.", + "best_pr_reason": "PR #43842 is the best representative of the cluster because it most narrowly and directly fixes the concrete `TypeAdapter` NameError case.", "soft_edge_verdicts": [ { "left": "pull_request:43656", "right": "pull_request:43842", - "accept": true, - "reason": "Both target the same TypeAdapter NameError in the CLI serve module and use alternative typing-only remedies (`from __future__ import annotations` vs `TYPE_CHECKING` import/string annotation)." + "accept": false, + "reason": "The packet shows only shared filename and broad thematic similarity; the different explicit issue target and limited diff evidence do not establish the same concrete failure mode." }, { "left": "pull_request:43836", "right": "pull_request:43842", - "accept": true, - "reason": "Both are about the same TypeAdapter NameError in `serve.py`; the titles and diff context indicate overlapping fixes for the same runtime annotation issue, even if 43836 is broader." + "accept": false, + "reason": "Although both target the same issue, #43836 appears to be a much broader `serve.py` rewrite, so the evidence is insufficient to say it is the same specific bugfix or that the PRs could cleanly merge as one change." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The packet supports the main conclusion: all three PRs are in `src/transformers/cli/serve.py` and their titles/diffs point to the same `TypeAdapter` runtime-annotation bug family, with 43842 the smallest targeted fix. The soft-edge acceptances are also conservative enough: 43836 clearly overlaps on the same issue target and 43656 appears to be another typing/annotation-based workaround in the same module." + "accept": false, + "feedback": "The main conclusion that PR #43842 is the cleanest/minimal fix is grounded. However, the soft-edge verdicts are too strong. For 43656 vs 43842, the packet only shows soft filename similarity plus a different explicit issue target; it does not establish that both PRs fix the same TypeAdapter failure mode. For 43836 vs 43842, they do share the same issue target, but the preview shows 43836 is a much broader serve.py rewrite, so calling it simply another variant of the same bugfix is not sufficiently conservative. The summary should frame these as related or possibly overlapping, not definitively the same underlying fix." }, "evaluator_used": true, - "retried": false, - "accepted_nontrivial_soft_edge": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -16496,12 +16448,12 @@ "pull_request:44044|pull_request:44154", "pull_request:44024|pull_request:44076" ], - "prepared_review_unit_hash": "73994f39096e66f4536d9a2f504c809f7feabfaf48304f314ed11e424fc25813", + "prepared_review_unit_hash": "709cb79e419cbe2e4b361692350a70867de96495b094ce811b93d9ea4b8ea596", "component_budget": { "node_count": 28, "item_count": 28, "soft_pair_count": 77, - "serialized_chars": 84501, + "serialized_chars": 84502, "estimated_input_tokens": 21126, "estimated_eval_tokens": 42508 }, @@ -16509,7 +16461,7 @@ "node_count": 28, "item_count": 28, "soft_pair_count": 72, - "serialized_chars": 76934, + "serialized_chars": 76936, "estimated_input_tokens": 19234, "estimated_eval_tokens": 38724 }, @@ -16518,300 +16470,330 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a wave of per-model refactors to standardized output tracing/capture_outputs, all orbiting external tracking issue #43979. Most pairs are only thematically similar; the real duplicate candidates are the GPT-J/CodeGen PRs and the GPT-Neo PRs.", - "confidence": 0.92, - "canonical_issue_reason": "No issue artifact is present in the input; the PRs only reference external tracker #43979.", - "canonical_pr_reason": "PR 44066 is the best canonical anchor: it directly targets #43979, covers the shared GPT-J/CodeGen output-tracing refactor, and is the clearest representative of the duplicate GPT-J family.", - "best_issue_reason": "No issue artifact is present in the input; the cluster is organized around external tracker #43979 only.", - "best_pr_reason": "PR 44066 is the strongest overall representative of the cluster because it is the most central concrete code change with explicit issue linkage; the other items are mostly separate per-model refactors rather than true duplicates.", + "summary": "This cluster is mostly a set of model-specific output-tracing refactors toward `capture_outputs`/`can_return_tuple`, not one shared bug. The only clear duplicate PR pairs are the GPT-J/CodeGen refactor set and the GPT-Neo refactor set; the rest are separate ports to different model codepaths.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44066 is the best anchor: it is explicitly issue-linked, touches the core GPT-J/CodeGen output-tracing refactor, and is the clearest representative of the shared change that later GPT-J PRs repeat.", + "best_issue_reason": null, + "best_pr_reason": "PR 44066 is the most representative and broadly scoped standardized-output-tracing PR in the cluster; it captures the main concrete codepath change better than the more fragmented model-by-model follow-ups.", "soft_edge_verdicts": [ { "left": "pull_request:43996", "right": "pull_request:44085", "accept": false, - "reason": "Different models and code paths; only the generic output-tracing theme matches." + "reason": "Different model families and codepaths (CVT/FNet vs GPT-J); same umbrella refactor theme is not enough." }, { "left": "pull_request:43996", "right": "pull_request:44044", "accept": false, - "reason": "Different models and code paths; same refactor theme only." + "reason": "Different models and implementations (CVT/FNet vs DeBERTa); not the same underlying change." }, { "left": "pull_request:44066", "right": "pull_request:44085", "accept": true, - "reason": "Both are GPT-J output-tracing refactors on the same file family and could be merged into one PR." + "reason": "Both target GPT-J output tracing on the same concrete block codepath; 44066 is the broader GPT-J/CodeGen refactor and 44085 is the overlapping GPT-J-only variant." }, { "left": "pull_request:44007", "right": "pull_request:44072", "accept": false, - "reason": "Different vision backbones and code paths; same standardized-output pattern only." + "reason": "Both are output-tracing refactors, but they touch unrelated model stacks (RegNet/ResNet/RT-DETR vs EfficientNet)." }, { "left": "pull_request:44072", "right": "pull_request:44722", "accept": false, - "reason": "Different models and code paths; both are output-tracing refactors but not the same change." + "reason": "EfficientNet and GPT-J/CodeGen are unrelated codepaths; only the refactor pattern is similar." }, { "left": "pull_request:44066", "right": "pull_request:44072", "accept": false, - "reason": "GPT-J/CodeGen versus EfficientNet; same initiative, not the same concrete bug or change." + "reason": "Different model families and concrete implementations; no shared bug/change." }, { "left": "pull_request:44013", "right": "pull_request:44044", "accept": false, - "reason": "Different model families and implementation paths; only the refactor pattern is shared." + "reason": "MobileNetV2 and DeBERTa are unrelated codepaths; same tracking issue does not make them duplicates." }, { "left": "pull_request:44066", "right": "pull_request:44086", "accept": false, - "reason": "Different models and code paths; same output-tracing migration theme only." + "reason": "Different models (GPT-J/CodeGen vs MGP-STR); only the standardized-output-tracing theme is shared." }, { "left": "pull_request:44066", "right": "pull_request:44071", "accept": false, - "reason": "Different models and code paths; not a mergeable duplicate." + "reason": "Different codepaths and model families (GPT-J/CodeGen vs MPT)." }, { "left": "pull_request:44018", "right": "pull_request:44068", "accept": true, - "reason": "Both are GPT-Neo output-tracing refactors on the same model file and appear to be the same underlying change." + "reason": "Same GPT-Neo refactor, same files, and same output-tracing change; these are duplicate implementations of the same concrete fix." }, { "left": "pull_request:44066", "right": "pull_request:44068", "accept": false, - "reason": "GPT-J/CodeGen versus GPT-Neo; shared theme only." + "reason": "GPT-J/CodeGen and GPT-Neo are separate model implementations; not the same PR change." }, { "left": "pull_request:44018", "right": "pull_request:44066", "accept": false, - "reason": "Different models and code paths; same broad standardized-output effort only." + "reason": "GPT-Neo vs GPT-J/CodeGen are distinct refactors on different codepaths." }, { "left": "pull_request:44019", "right": "pull_request:44722", "accept": false, - "reason": "ResNet and GPT-J are unrelated code paths despite both being output-tracing refactors." + "reason": "ResNet and GPT-J/CodeGen are unrelated model codepaths." }, { "left": "pull_request:44086", "right": "pull_request:44722", "accept": false, - "reason": "Different models and implementations; only the decorator migration is shared." + "reason": "MGP-STR and GPT-J/CodeGen are different concrete implementations." }, { "left": "pull_request:44019", "right": "pull_request:44085", "accept": false, - "reason": "ResNet versus GPT-J; not the same bug or change." + "reason": "ResNet and GPT-J are unrelated; same output-tracing pattern is not enough." }, { "left": "pull_request:44018", "right": "pull_request:44019", "accept": false, - "reason": "GPT-Neo and ResNet are distinct model implementations." + "reason": "GPT-Neo and ResNet are different model stacks with different fixes." }, { "left": "pull_request:44019", "right": "pull_request:44071", "accept": false, - "reason": "Different model families and code paths." + "reason": "ResNet and MPT are unrelated codepaths." }, { "left": "pull_request:44068", "right": "pull_request:44116", "accept": false, - "reason": "GPT-Neo and Flaubert are separate refactors in different code paths." + "reason": "GPT-Neo and Flaubert are different models; these are separate standardized-output-tracing ports." }, { "left": "pull_request:44068", "right": "pull_request:44722", "accept": false, - "reason": "Different model families; same standardized-output theme only." + "reason": "GPT-Neo and GPT-J/CodeGen are separate model refactors." }, { "left": "pull_request:44071", "right": "pull_request:44722", "accept": false, - "reason": "MPT and GPT-J/CodeGen are unrelated concrete changes." + "reason": "MPT and GPT-J/CodeGen are unrelated concrete codepaths." }, { "left": "pull_request:44059", "right": "pull_request:44086", "accept": false, - "reason": "GPT-2 and MGP-STR are unrelated model paths." + "reason": "GPT-2 and MGP-STR are unrelated model implementations." }, { "left": "pull_request:44085", "right": "pull_request:44086", "accept": false, - "reason": "GPT-J and MGP-STR are different changes." + "reason": "GPT-J and MGP-STR are different codepaths; only the refactor style matches." }, { "left": "pull_request:44059", "right": "pull_request:44722", "accept": false, - "reason": "GPT-2 versus GPT-J/CodeGen; same pattern only." + "reason": "GPT-2 and GPT-J/CodeGen are separate output-tracing refactors." }, { "left": "pull_request:44098", "right": "pull_request:44101", "accept": false, - "reason": "ViLT and XLM/Flaubert are different model code paths." - }, - { - "left": "pull_request:44073", - "right": "pull_request:44101", - "accept": false, - "reason": "VisualBert and XLM/Flaubert are unrelated concrete changes." + "reason": "ViLT and Flaubert are unrelated model codepaths." }, { - "left": "pull_request:44056", - "right": "pull_request:44161", + "left": "pull_request:44018", + "right": "pull_request:44071", "accept": false, - "reason": "MPNet and LongT5 are different model implementations." + "reason": "GPT-Neo and MPT are different models with separate changes." }, { "left": "pull_request:44017", "right": "pull_request:44722", "accept": false, - "reason": "SegFormer and GPT-J/CodeGen are unrelated code paths." + "reason": "Segformer and GPT-J/CodeGen are unrelated concrete implementations." }, { "left": "pull_request:44017", "right": "pull_request:44161", "accept": false, - "reason": "SegFormer and LongT5 are different model refactors." + "reason": "Segformer and LongT5 are different model families; not a duplicate change." }, { "left": "pull_request:44017", "right": "pull_request:44073", "accept": false, - "reason": "SegFormer and VisualBert are distinct changes." + "reason": "Segformer and VisualBert are separate model codepaths." }, { "left": "pull_request:44068", - "right": "pull_request:44086", + "right": "pull_request:44071", "accept": false, - "reason": "GPT-Neo and MGP-STR are unrelated model paths." + "reason": "GPT-Neo and MPT are unrelated implementations." }, { - "left": "pull_request:44073", - "right": "pull_request:44161", + "left": "pull_request:44071", + "right": "pull_request:44086", "accept": false, - "reason": "VisualBert and LongT5 are not the same change." + "reason": "MPT and MGP-STR are different codepaths." }, { - "left": "pull_request:44071", - "right": "pull_request:44086", + "left": "pull_request:43997", + "right": "pull_request:44073", "accept": false, - "reason": "MPT and MGP-STR are different concrete implementations." + "reason": "RegNet and VisualBert are unrelated model families." }, { "left": "pull_request:43997", + "right": "pull_request:44017", + "accept": false, + "reason": "RegNet and Segformer are separate model implementations." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44071", + "accept": false, + "reason": "SqueezeBert and MPT are unrelated codepaths." + }, + { + "left": "pull_request:44010", "right": "pull_request:44073", "accept": false, - "reason": "RegNet and VisualBert are separate model code paths." + "reason": "SqueezeBert and VisualBert are different model families." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44129", + "accept": false, + "reason": "GPT-Neo and SpeechT5 are separate concrete refactors." }, { "left": "pull_request:44074", "right": "pull_request:44101", "accept": false, - "reason": "TextNet and XLM/Flaubert are unrelated concrete changes." + "reason": "TextNet and Flaubert are unrelated implementations." }, { "left": "pull_request:44056", "right": "pull_request:44161", "accept": false, - "reason": "MPNet and LongT5 are different models and refactors." + "reason": "MPNet and LongT5 are different codepaths." }, { "left": "pull_request:43997", "right": "pull_request:44018", "accept": false, - "reason": "RegNet and GPT-Neo are unrelated code paths." + "reason": "RegNet and GPT-Neo are unrelated model refactors." }, { "left": "pull_request:43997", "right": "pull_request:44161", "accept": false, - "reason": "RegNet and LongT5 are different model implementations." + "reason": "RegNet and LongT5 are distinct model implementations." }, { "left": "pull_request:44071", "right": "pull_request:44129", "accept": false, - "reason": "MPT and SpeechT5 are unrelated concrete changes." + "reason": "MPT and SpeechT5 are unrelated concrete codepaths." }, { - "left": "pull_request:44010", - "right": "pull_request:44071", + "left": "pull_request:43996", + "right": "pull_request:44161", "accept": false, - "reason": "SqueezeBert and MPT are different model paths." + "reason": "CVT/FNet and LongT5 are different models; not the same change." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44017", + "accept": false, + "reason": "CVT/FNet and Segformer are unrelated implementations." }, { "left": "pull_request:44010", - "right": "pull_request:44073", + "right": "pull_request:44044", "accept": false, - "reason": "SqueezeBert and VisualBert are distinct refactors." + "reason": "SqueezeBert and DeBERTa are separate model codepaths." }, { - "left": "pull_request:44018", - "right": "pull_request:44129", + "left": "pull_request:44044", + "right": "pull_request:44074", "accept": false, - "reason": "GPT-Neo and SpeechT5 are unrelated code paths." + "reason": "DeBERTa and TextNet are unrelated models." }, { - "left": "pull_request:44074", - "right": "pull_request:44161", + "left": "pull_request:44066", + "right": "pull_request:44722", + "accept": true, + "reason": "Same GPT-J/CodeGen output-tracing refactor; 44722 is a near-duplicate follow-up on the same concrete change." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44722", + "accept": true, + "reason": "Both are GPT-J output-tracing refactors on the same underlying codepath and could plausibly be merged into one PR." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44129", "accept": false, - "reason": "TextNet and LongT5 are different model changes." + "reason": "GPT-J and SpeechT5 are unrelated model implementations." }, { - "left": "pull_request:44056", - "right": "pull_request:44161", + "left": "pull_request:43996", + "right": "pull_request:44007", "accept": false, - "reason": "MPNet and LongT5 are separate concrete refactors." + "reason": "CVT/FNet and RegNet/ResNet/RT-DETR are different model families." }, { - "left": "pull_request:44018", - "right": "pull_request:44059", + "left": "pull_request:43996", + "right": "pull_request:44129", "accept": false, - "reason": "GPT-Neo and GPT-2 are different model code paths." + "reason": "CVT/FNet and SpeechT5 are unrelated codepaths." }, { - "left": "pull_request:44010", - "right": "pull_request:44044", + "left": "pull_request:44044", + "right": "pull_request:44129", "accept": false, - "reason": "SqueezeBert and DeBERTa are unrelated implementations." + "reason": "DeBERTa and SpeechT5 are separate model implementations." }, { "left": "pull_request:44044", - "right": "pull_request:44074", + "right": "pull_request:44154", "accept": false, - "reason": "DeBERTa and TextNet are different model paths." + "reason": "DeBERTa and ViT-S are unrelated model codepaths." }, { "left": "pull_request:44024", "right": "pull_request:44076", "accept": false, - "reason": "FocalNet and ImageGPT are different model refactors." + "reason": "FocalNet and ImageGPT are different model families; same output-tracing theme does not imply duplication." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet, the canonical choices are conservative, and the soft-edge verdicts avoid overstating cross-model similarity. The two accepted pairs are both supported as same-family, same-file refactor candidates that could plausibly be merged into a single bugfix PR." + "feedback": "The summary is well grounded in the packet and stays conservative about the soft edges. The accepted pairs are all within the same concrete model/codepath families (GPT-Neo and GPT-J/CodeGen), and the rejections appropriately avoid treating shared output-tracing style or shared issue-targets as duplicates across unrelated models. No clear overstatement of evidence." }, "evaluator_used": true, "retried": false, @@ -16846,12 +16828,12 @@ "pull_request:43997|pull_request:44010", "pull_request:44010|pull_request:44129" ], - "prepared_review_unit_hash": "70ac57ded28e5532df8f52d574ed54738fab8691c9a0f022e76d265eff73037c", + "prepared_review_unit_hash": "126c7d405f80fdc7403b5cdc7097d4c584642317aed1ef568699a1d5cc227e4e", "component_budget": { "node_count": 28, "item_count": 28, "soft_pair_count": 77, - "serialized_chars": 84501, + "serialized_chars": 84502, "estimated_input_tokens": 21126, "estimated_eval_tokens": 42508 }, @@ -16859,7 +16841,7 @@ "node_count": 6, "item_count": 6, "soft_pair_count": 5, - "serialized_chars": 13067, + "serialized_chars": 13066, "estimated_input_tokens": 3267, "estimated_eval_tokens": 6790 }, @@ -16868,42 +16850,42 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These PRs share a common refactoring theme around standardized output tracing / capture_outputs, but they target different model implementations and code paths, so they are not duplicates of the same concrete change.", - "confidence": 0.9, + "summary": "These PRs are all model-specific refactors for standardized output tracing/capture, but they touch different architectures and code paths, so they should not be deduplicated as the same change.", + "confidence": 0.92, "canonical_issue_reason": null, - "canonical_pr_reason": "No PR is a true canonical duplicate here; if one representative is needed, PR 44129 is the most concrete because it is tied to an explicit issue and shows a model-specific output-tracing refactor.", + "canonical_pr_reason": "PR #44129 is the strongest representative: it has an explicit issue target and a concrete, self-contained SpeechT5 tracing refactor. That said, it is not a duplicate of the others.", "best_issue_reason": null, - "best_pr_reason": "PR 44129 is the best single representative of the cluster because it has the clearest problem statement (explicit issue link) and a substantive, model-specific output-tracing implementation change.", + "best_pr_reason": "PR #44129 is the best single PR to keep as the representative example of this refactor pattern because it is well-scoped and linked to an issue. The others are similarly scoped but model-specific.", "soft_edge_verdicts": [ { "left": "pull_request:44073", "right": "pull_request:44074", "accept": false, - "reason": "Both are output-tracing refactors, but they affect different models (VisualBert vs TextNet) and different forward/output code paths; too broad to treat as the same underlying change." + "reason": "Both are output-tracing refactors, but one is for VisualBert and the other for TextNet. Different models, different files, and different implementations; not the same underlying change." }, { "left": "pull_request:44056", "right": "pull_request:44073", "accept": false, - "reason": "MPNet vs VisualBert; same migration pattern, but not the same bug or mergeable code change." + "reason": "MPNet vs VisualBert. Same broad refactor theme, but they modify separate model code paths and are independent changes." }, { "left": "pull_request:44056", "right": "pull_request:44074", "accept": false, - "reason": "MPNet vs TextNet; these are separate model-specific refactors, not duplicate fixes." + "reason": "MPNet and TextNet are distinct model implementations with unrelated refactors despite similar decorator/capture naming." }, { "left": "pull_request:43997", "right": "pull_request:44010", "accept": false, - "reason": "RegNet and SqueezeBert are unrelated implementations; the similar wording only indicates a shared refactor style, not the same change." + "reason": "RegNet and SqueezeBert are unrelated model-specific output tracing updates; they do not appear mergeable into one PR for the same bug or change." }, { "left": "pull_request:44010", "right": "pull_request:44129", "accept": false, - "reason": "SqueezeBert and SpeechT5 modify different code paths and even different output/capture mechanics; they should not be merged as duplicates." + "reason": "SqueezeBert vs SpeechT5, different model internals and different output-tracing logic. Similar theme only at a high level, not the same concrete fix." } ] }, @@ -16983,268 +16965,268 @@ "pull_request:44028|pull_request:44030", "pull_request:44029|pull_request:44030" ], - "prepared_review_unit_hash": "43467ca809c6aa50563a041c3223408a71a0a19b35877b263d1a60954bdbf607", + "prepared_review_unit_hash": "636272954d30916491a475487422f1a8c03c8b5eaf656b8dca64a8cd3dbbe155", "component_budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 40, - "serialized_chars": 40177, - "estimated_input_tokens": 10045, - "estimated_eval_tokens": 20346 + "serialized_chars": 40173, + "estimated_input_tokens": 10044, + "estimated_eval_tokens": 20344 }, "budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 40, - "serialized_chars": 40177, - "estimated_input_tokens": 10045, - "estimated_eval_tokens": 20346 + "serialized_chars": 40173, + "estimated_input_tokens": 10044, + "estimated_eval_tokens": 20344 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is a set of model-specific refactors around output tracing/can_return_tuple, but they target different architectures and code paths, so they do not look like duplicate PRs.", - "confidence": 0.9, + "summary": "This cluster is a set of separate, model-specific output-tracing refactors. They are thematically similar, but they do not describe the same concrete bug or change, so they should not be merged as duplicates.", + "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44025 is the best representative of the pattern: it applies the output-tracing refactor cleanly within the closely related Depth Anything family and is self-contained.", + "canonical_pr_reason": "PR 44025 is the best representative of the theme because it is a clean, self-contained output-tracing refactor and spans a closely related pair of files. That said, the cluster is not duplicate-equivalent overall; the PRs target different model implementations.", "best_issue_reason": null, - "best_pr_reason": "44025 is the tightest, clearest example of the change style in this cluster; the other PRs span unrelated model implementations and are less suitable as a canonical representative.", + "best_pr_reason": "PR 44025 is the most representative hub for the cluster, but only as a thematic example rather than a duplicate canonical. It is not evidence that the other PRs are the same change.", "soft_edge_verdicts": [ { "left": "pull_request:43998", "right": "pull_request:44002", "accept": false, - "reason": "Different model families (timm_backbone vs upernet); same refactor theme, but not the same concrete code path." + "reason": "Same refactor pattern, but timm_backbone and upernet are different model code paths." }, { "left": "pull_request:43998", "right": "pull_request:44000", "accept": false, - "reason": "Different architectures (timm_backbone vs vision_text_dual_encoder); too broad to treat as the same change." + "reason": "Both are output-tracing refactors, but they touch unrelated model implementations." }, { "left": "pull_request:43998", "right": "pull_request:44001", "accept": false, - "reason": "Unrelated models (timm_backbone vs univnet); shared output-tracing pattern is not enough for duplicate triage." + "reason": "Different architectures and forward paths; not the same underlying change." }, { "left": "pull_request:43998", "right": "pull_request:44003", "accept": false, - "reason": "Timm backbone and mamba/falcon_mamba are different code paths; similar refactor style only." + "reason": "timm_backbone is unrelated to mamba/falcon_mamba; same theme, not same code path." }, { "left": "pull_request:43998", "right": "pull_request:44029", "accept": false, - "reason": "Different subsystems (timm_backbone vs rwkv); same theme, but not one mergeable PR." + "reason": "Unrelated model families; these are not mergeable as one concrete fix." }, { "left": "pull_request:43999", "right": "pull_request:44001", "accept": false, - "reason": "MobileNetV1 and UnivNet are unrelated models; both refactor output handling, but not the same bug/change." + "reason": "Different model implementations; only the refactor pattern matches." }, { "left": "pull_request:43999", "right": "pull_request:44002", "accept": false, - "reason": "MobileNetV1 vs UperNet touches different model code paths; reject as duplicate." + "reason": "MobileNetV1 and UperNet are separate model-specific changes." }, { "left": "pull_request:43999", "right": "pull_request:44025", "accept": false, - "reason": "Different model families (MobileNetV1 vs Depth Anything); same refactor motif only." + "reason": "Different model families and forward logic; not a duplicate." }, { "left": "pull_request:43999", "right": "pull_request:44026", "accept": false, - "reason": "MobileNetV1 and vision-encoder-decoder are unrelated; not a shared concrete fix." + "reason": "MobileNetV1 vs vision-encoder-decoder wrappers are distinct code paths." }, { "left": "pull_request:43999", "right": "pull_request:44027", "accept": false, - "reason": "MobileNetV1 vs speech-encoder-decoder is too far apart to merge as one PR." + "reason": "Separate model wrappers; same refactor style is not enough to merge." }, { "left": "pull_request:44000", "right": "pull_request:44001", "accept": false, - "reason": "Vision-text dual encoder and UnivNet are different models; same output-tracing refactor pattern only." + "reason": "Vision-text dual encoder and UnivNet are unrelated implementations." }, { "left": "pull_request:44000", "right": "pull_request:44002", "accept": false, - "reason": "Different architectures (vision-text dual encoder vs UperNet); not the same underlying change." + "reason": "Different model subsystems; not the same bug or change." }, { "left": "pull_request:44000", "right": "pull_request:44003", "accept": false, - "reason": "Unrelated code paths across vision-text dual encoder and mamba/falcon_mamba." + "reason": "Vision-text dual encoder and mamba are unrelated model paths." }, { "left": "pull_request:44000", "right": "pull_request:44025", "accept": false, - "reason": "Different model families; this is only a broad refactor theme, not a duplicate PR." + "reason": "These are separate model-specific refactors, not one concrete fix." }, { "left": "pull_request:44001", "right": "pull_request:44002", "accept": false, - "reason": "UnivNet and UperNet are unrelated; no shared concrete bug or fix." + "reason": "UnivNet and UperNet are different models with distinct forward paths." }, { "left": "pull_request:44001", "right": "pull_request:44003", "accept": false, - "reason": "Different subsystems (audio generation vs mamba/falcon_mamba); reject." + "reason": "Different architectures; the similarity is only at the refactor level." }, { "left": "pull_request:44001", "right": "pull_request:44025", "accept": false, - "reason": "UnivNet vs Depth Anything are unrelated implementations, even if both adjust return-tracing." + "reason": "Separate model families; not a duplicate change." }, { "left": "pull_request:44002", - "right": "pull_request:44029", + "right": "pull_request:44003", "accept": false, - "reason": "UperNet and RWKV do not share a concrete code path; same refactor style only." + "reason": "UperNet and mamba/falcon_mamba are unrelated code paths." }, { "left": "pull_request:44002", - "right": "pull_request:44003", + "right": "pull_request:44029", "accept": false, - "reason": "UperNet vs mamba/falcon_mamba are different model stacks; not mergeable as one PR." + "reason": "Different models; same output-tracing theme does not imply duplication." }, { "left": "pull_request:44003", "right": "pull_request:44028", "accept": false, - "reason": "Mamba/FalconMamba and SuperPoint are unrelated models; too broad for duplicate triage." + "reason": "Mamba/falcon_mamba and SuperPoint are unrelated implementations." }, { "left": "pull_request:44003", "right": "pull_request:44004", "accept": false, - "reason": "Mamba/FalconMamba vs CodeGen are distinct code paths; reject despite similar refactor intent." + "reason": "Mamba/falcon_mamba changes are not the same change as CodeGen output tracing." }, { "left": "pull_request:43998", "right": "pull_request:44004", "accept": false, - "reason": "Timm backbone and CodeGen are unrelated; output-tracing refactor is not enough to unify them." + "reason": "timm_backbone and CodeGen are different model code paths." }, { "left": "pull_request:44001", "right": "pull_request:44004", "accept": false, - "reason": "UnivNet and CodeGen touch different model implementations and cannot plausibly be one PR." + "reason": "UnivNet and CodeGen are unrelated; no shared concrete bug." }, { "left": "pull_request:44002", "right": "pull_request:44004", "accept": false, - "reason": "UperNet and CodeGen are unrelated; same broad cleanup theme only." + "reason": "UperNet and CodeGen are distinct implementations with different fixes." }, { "left": "pull_request:44025", "right": "pull_request:44026", "accept": false, - "reason": "Depth Anything and vision-encoder-decoder are different model families; not the same concrete fix." + "reason": "DepthAnything and vision-encoder-decoder are different model paths; same refactor pattern only." }, { "left": "pull_request:44025", "right": "pull_request:44027", "accept": false, - "reason": "Depth Anything vs speech-encoder-decoder are separate code paths; reject." + "reason": "DepthAnything and SpeechEncoderDecoder are unrelated concrete changes." }, { "left": "pull_request:44025", "right": "pull_request:44028", "accept": false, - "reason": "Depth Anything and SuperPoint are unrelated models, even though both change output tracing." + "reason": "DepthAnything and SuperPoint do not fix the same underlying issue." }, { "left": "pull_request:44025", "right": "pull_request:44029", "accept": false, - "reason": "Depth Anything and RWKV are different architectures; not a duplicate change." + "reason": "DepthAnything and RWKV are unrelated code paths." }, { "left": "pull_request:44025", "right": "pull_request:44030", "accept": false, - "reason": "Depth Anything and DPR are unrelated model code paths." + "reason": "DepthAnything and DPR are different model implementations." }, { "left": "pull_request:44026", "right": "pull_request:44029", "accept": false, - "reason": "vision-encoder-decoder and RWKV are unrelated; same refactor style only." + "reason": "Vision-encoder-decoder and RWKV are unrelated; not the same fix." }, { "left": "pull_request:44026", "right": "pull_request:44027", "accept": false, - "reason": "Vision encoder-decoder vs speech encoder-decoder are separate implementations; not the same bug." + "reason": "Both are encoder-decoder wrappers, but they are separate model-specific forward methods and not one concrete bug." }, { "left": "pull_request:44026", "right": "pull_request:44028", "accept": false, - "reason": "Different model families (vision-encoder-decoder vs SuperPoint); reject." + "reason": "Vision-encoder-decoder and SuperPoint are different code paths." }, { "left": "pull_request:44026", "right": "pull_request:44030", "accept": false, - "reason": "Vision encoder-decoder and DPR touch different code paths; not mergeable into one PR." + "reason": "Encoder-decoder wrapper refactor vs DPR wrapper refactor are not the same change." }, { "left": "pull_request:44027", "right": "pull_request:44028", "accept": false, - "reason": "Speech encoder-decoder and SuperPoint are unrelated models; too broad for duplicate triage." + "reason": "SpeechEncoderDecoder and SuperPoint are unrelated implementations." }, { "left": "pull_request:44027", "right": "pull_request:44029", "accept": false, - "reason": "Speech encoder-decoder vs RWKV are different architectures and fixes." + "reason": "SpeechEncoderDecoder and RWKV are unrelated model code paths." }, { "left": "pull_request:44027", "right": "pull_request:44030", "accept": false, - "reason": "Speech encoder-decoder and DPR do not share a concrete bug or code path." + "reason": "SpeechEncoderDecoder and DPR do not share the same concrete fix." }, { "left": "pull_request:44028", "right": "pull_request:44029", "accept": false, - "reason": "SuperPoint and RWKV are unrelated; same cleanup pattern only." + "reason": "SuperPoint and RWKV are distinct model-specific changes." }, { "left": "pull_request:44028", "right": "pull_request:44030", "accept": false, - "reason": "SuperPoint vs DPR are different implementations; not the same underlying change." + "reason": "SuperPoint and DPR are unrelated implementations." }, { "left": "pull_request:44029", "right": "pull_request:44030", "accept": false, - "reason": "RWKV and DPR are different model code paths; output-tracing refactor alone is insufficient." + "reason": "RWKV and DPR are different models with different forward logic." } ] }, @@ -17277,52 +17259,52 @@ "pull_request:44662|pull_request:45075", "pull_request:45075|pull_request:45401" ], - "prepared_review_unit_hash": "1d945e31114ec032c7ebea6f8e56041daa5f2e10dcdf792d3aade67caf9f6ba9", + "prepared_review_unit_hash": "29283d69ec866496e4c2137c385e87156608f0b8d667abd94b444c4d1d5f92c3", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 11070, - "estimated_input_tokens": 2768, - "estimated_eval_tokens": 5792 + "serialized_chars": 11073, + "estimated_input_tokens": 2769, + "estimated_eval_tokens": 5794 }, "budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 11070, - "estimated_input_tokens": 2768, - "estimated_eval_tokens": 5792 + "serialized_chars": 11073, + "estimated_input_tokens": 2769, + "estimated_eval_tokens": 5794 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are three distinct model-addition pull requests for different models (PenguinVL, DeepSeek-OCR-2, Voxtral TTS). They share only common Transformers boilerplate and docs/indexing changes, not the same underlying bug or change set.", - "confidence": 0.97, + "summary": "These are three separate model-addition PRs for different models (PenguinVL, DeepSeek-OCR-2, and Voxtral TTS). The overlap is mostly shared Transformers boilerplate for new model integration, not the same underlying change.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "No canonical PR: although the PRs share new-model scaffolding, they implement different models and are not duplicates.", "best_issue_reason": null, - "best_pr_reason": "PR 45075 is the strongest standalone candidate: it is a complete, non-draft model addition with conversion tooling and tests. The other PRs are separate model integrations and do not form a duplicate cluster.", + "best_pr_reason": "No single PR is a best duplicate representative here; each PR targets a different model and could not be merged into one change.", "soft_edge_verdicts": [ { "left": "pull_request:44662", "right": "pull_request:45401", "accept": false, - "reason": "Different model integrations for unrelated models; shared auto-registry/docs scaffolding is expected and does not indicate the same change or code-path bug." + "reason": "Different models and codepaths: PenguinVL vs Voxtral TTS. Shared auto-registration/docs files are generic new-model plumbing, not the same fix." }, { "left": "pull_request:44662", "right": "pull_request:45075", "accept": false, - "reason": "Both are independent new-model implementations with different architectures and files; overlap is limited to generic Transformers integration boilerplate." + "reason": "Different model implementations: PenguinVL and DeepSeek-OCR-2. The overlap is only shared framework registration and docs, so they are not duplicates." }, { "left": "pull_request:45075", "right": "pull_request:45401", "accept": false, - "reason": "Separate model add-ons for different products; they do not fix or implement the same concrete functionality and should not be merged as duplicates." + "reason": "Different model additions for DeepSeek-OCR-2 and Voxtral TTS. Shared Transformers scaffolding does not indicate the same underlying bug or change." } ] }, @@ -17352,12 +17334,12 @@ "soft_pairs": [ "pull_request:44979|pull_request:45363" ], - "prepared_review_unit_hash": "4dfb6fa51145423a12ffc2b965787f8f0956aec4ffb3af5aa2a4744a1c3d3091", + "prepared_review_unit_hash": "5c60281b491d0b5f0ac2a5454cc92f029ee68617c43e1f78eb08002e3a18f362", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4408, + "serialized_chars": 4407, "estimated_input_tokens": 1102, "estimated_eval_tokens": 2460 }, @@ -17365,7 +17347,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4408, + "serialized_chars": 4407, "estimated_input_tokens": 1102, "estimated_eval_tokens": 2460 }, @@ -17374,18 +17356,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Related fusion features, but not the same underlying change: one adds a standalone module-fusion API, the other adds kernel-config-based n-to-1 fusion plumbing and an example. They overlap conceptually but target different code paths and would not naturally merge into one PR.", - "confidence": 0.93, + "summary": "These are related fusion-feature PRs, but they target different layers and are not duplicates: one adds a standalone Module Fusion API, while the other wires n-to-1 kernel fusion through KernelConfig and model loading internals.", + "confidence": 0.87, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44979 is the cleaner core API proposal: it introduces the standalone module-fusion machinery and tests, while 45363 is a separate integration layer built around KernelConfig and hub-kernel plumbing.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR 44979 is the better representative of the cluster because it is the more self-contained foundational fusion feature; 45363 is more of an application/integration PR than the core abstraction.", + "best_pr_reason": "PR 45363 is the broader, more end-to-end change for kernel fusion, touching configuration, integrations, and model loading; it is the better representative of the cluster's main theme.", "soft_edge_verdicts": [ { "left": "pull_request:44979", "right": "pull_request:45363", "accept": false, - "reason": "Reject: these are both fusion-related, but 44979 adds a generic module-fusion API and 45363 implements n-to-1 kernel fusion via KernelConfig. They touch different abstractions and code paths, so they are not the same concrete change." + "reason": "Different concrete changes and code paths: 44979 introduces a new Module Fusion API module and tests, while 45363 adds KernelConfig-based n-to-1 kernel fusion plumbing, integration hooks, and an example. Related topic, but not the same mergeable change." } ] }, @@ -17415,7 +17397,7 @@ "soft_pairs": [ "pull_request:45086|pull_request:45317" ], - "prepared_review_unit_hash": "e05a247d607f84a870d35f36ac7379a1339167b002816fe1daaa0db0cb05913c", + "prepared_review_unit_hash": "220462e86df87d807814f14bebe628ba16f9a3c29fb55dc25ad5cb568d840aa5", "component_budget": { "node_count": 2, "item_count": 2, @@ -17437,18 +17419,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs target the same Mistral regex patch bug: an AttributeError caused by using `tokenizer.backend_tokenizer.pre_tokenizer` instead of `tokenizer.pre_tokenizer` in `_patch_mistral_regex`. PR 45317 is the more complete version because it includes the fix plus tests; PR 45086 is effectively a subset of the same change.", + "summary": "Both PRs fix the same Mistral regex patching AttributeError in the same file and target the same issue; 45317 looks like the fuller follow-up with tests.", "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45317 is the best canonical PR because it fixes the same concrete code path as 45086 and also adds a test update, making it the more complete implementation.", + "canonical_pr_reason": "PR 45317 is the better canonical PR because it covers the same fix as 45086 but is more complete and includes test coverage.", "best_issue_reason": null, - "best_pr_reason": "PR 45317 is the strongest PR candidate: same bug, same file, same issue target, and it adds regression coverage beyond the minimal code-only patch.", + "best_pr_reason": "PR 45317 is the best representative of the cluster: same concrete bug fix, clearer title, and additional tests make it the strongest standalone artifact.", "soft_edge_verdicts": [ { "left": "pull_request:45086", "right": "pull_request:45317", "accept": true, - "reason": "Accept: both PRs make the same fix in `_patch_mistral_regex` for the same AttributeError, differing only in completeness (45317 adds tests)." + "reason": "Same underlying AttributeError fix in _patch_mistral_regex, same target issue, and the diff shows the same code-path correction; 45317 is essentially the more complete version of 45086." } ] }, @@ -17478,7 +17460,7 @@ "soft_pairs": [ "pull_request:45115|pull_request:45569" ], - "prepared_review_unit_hash": "19cc4dec6aef7b1349823262c61c0bc431904c3287011fa33f6b1cdf112a09d6", + "prepared_review_unit_hash": "150f3d80335a91aecbd7cabd6fe0d8c8114de376ef811205a0309fa3248fd650", "component_budget": { "node_count": 2, "item_count": 2, @@ -17500,18 +17482,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open NemotronH-related PRs overlap on shared model files, but they target different changes: one is a refactor to inherit GraniteMoeHybrid, the other splits NemotronH into dense/sparse model variants and updates docs/tests. They are related but not the same underlying change.", - "confidence": 0.91, + "summary": "Two open PRs both touch Nemotron H internals, but they target different changes: one refactors NemotronH to inherit GraniteMoeHybrid and add conversion mapping, while the other splits Nemotron H into dense/sparse model variants and updates docs/tests. They are related but not the same fix/change.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45569 is the better representative because it implements the broader end-state change: dedicated NemotronHDense/NemotronHSparse packages, auto-mapping updates, docs, and tests. PR #45115 is a narrower refactor step focused on GraniteMoeHybrid inheritance and conversion mapping.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR #45569 is the most complete and user-facing artifact in the cluster, with the clearest scope and strongest alignment to the final architecture change. PR #45115 appears to be an earlier structural refactor rather than the main implementation.", + "best_pr_reason": "PR #45569 is the better cluster representative because it covers the larger, end-to-end Nemotron H architecture change, including model split, auto-mapping, docs, and tests. PR #45115 is a narrower refactor focused on inheritance and weight conversion.", "soft_edge_verdicts": [ { "left": "pull_request:45115", "right": "pull_request:45569", "accept": false, - "reason": "Shared NemotronH files alone are not enough: #45115 refactors NemotronH to inherit GraniteMoeHybrid and adds conversion mapping, while #45569 restructures NemotronH into dense/sparse model families with docs and tests. Different concrete code-path changes, so not duplicates." + "reason": "Shared Nemotron H files alone are not enough: #45115 is a structural refactor/conversion-mapping PR, while #45569 introduces separate dense/sparse model classes plus docs/tests. They do not look like the same concrete change and would not plausibly be merged as one PR." } ] }, @@ -17541,7 +17523,7 @@ "soft_pairs": [ "pull_request:45332|pull_request:45333" ], - "prepared_review_unit_hash": "e504be2940d40e71fcd5b050618f5adcc12adaf93f7fde276a9d582fad9ac571", + "prepared_review_unit_hash": "a924295be58fe005afd8cdcea24072d977f42e21d22780879e9c84a8c45df93f", "component_budget": { "node_count": 2, "item_count": 2, @@ -17563,18 +17545,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related feature PRs for heterogeneous model support, but they are not duplicates: one adds per-layer config plumbing, the other adds model/cache/masking behavior built on top of it.", - "confidence": 0.87, + "summary": "These are related but distinct PRs: one adds per-layer heterogeneous config plumbing, while the other adds broader heterogeneous model/cache support built on top of that config work. They look complementary rather than duplicate.", + "confidence": 0.82, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45332 is the broader, more concrete implementation of heterogeneous model behavior across cache/model code paths; #45333 is narrower config infrastructure.", + "canonical_pr_reason": "PR #45332 is the broader, more central change: it implements heterogeneous model support across cache/modeling code and includes the config hooks it needs.", "best_issue_reason": null, - "best_pr_reason": "PR #45332 is the best representative because it covers the end-user-facing model-side support, not just config API additions.", + "best_pr_reason": "PR #45332 is the best representative of the cluster because it spans the full code-path impact and is the more complete end-to-end implementation.", "soft_edge_verdicts": [ { "left": "pull_request:45332", "right": "pull_request:45333", "accept": false, - "reason": "Same feature area, but not the same underlying change: #45333 introduces per-layer heterogeneous config support, while #45332 updates cache/modeling logic to consume it. They are related and potentially sequential, not duplicate PRs that should be merged into one." + "reason": "They share heterogeneity plumbing, but #45333 is config-only while #45332 extends model/cache behavior; they are related groundwork and follow-on work, not the same concrete change that should be merged as one PR." } ] }, @@ -17613,7 +17595,7 @@ "pull_request:45592|pull_request:45615", "pull_request:45596|pull_request:45615" ], - "prepared_review_unit_hash": "4e11b82ca52fa7fe27c929da54b11c88f068dfc2509fe6072efa45e73378a360", + "prepared_review_unit_hash": "40f8da80ebff371273d5c3387a02dc195c9aa0368513048dff465def359654eb", "component_budget": { "node_count": 5, "item_count": 5, @@ -17635,54 +17617,54 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is mixed: one substantive model-parallel fix PR plus several unrelated test-only XPU expectation/padding updates across different models. No pair looks like the same concrete bug/change closely enough to merge as duplicates.", - "confidence": 0.82, + "summary": "These PRs are not true duplicates overall: one substantive model-parallel/code-path fix (45487), one XCLIP test/model-parallel fix (45394), one FastVLM test-padding fix (45592), and two separate XPU expectation updates for unrelated models (45596, 45615).", + "confidence": 0.92, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45487 is the broadest and most substantive code fix in the set, with actual model code changes and tests, so it is the best representative if one PR must anchor the cluster.", + "canonical_pr_reason": "PR 45487 is the strongest representative fix: it changes actual model-parallel behavior in AltCLIP/ChineseCLIP (device placement for buffered token type IDs and split-module lists), whereas the others are narrower model-specific test updates or unrelated fixes.", "best_issue_reason": null, - "best_pr_reason": "PR 45487 is the strongest global representative because it fixes a real code-path issue rather than only updating test expectations, and it covers multiple affected model files.", + "best_pr_reason": "PR 45487 is the best standalone PR in the set because it addresses a concrete runtime bug in shared code paths and is more substantive than the test-only expectation changes.", "soft_edge_verdicts": [ { "left": "pull_request:45394", "right": "pull_request:45596", "accept": false, - "reason": "XCLIP model-parallel no_split_modules fix vs BLT XPU test expectations; different model families and different failure modes." + "reason": "Different models and different problems: XCLIP split-module/test failures vs BLT XPU expectation updates." }, { "left": "pull_request:45487", "right": "pull_request:45592", "accept": false, - "reason": "AltCLIP/ChineseCLIP model-parallel/token_type_ids fix vs FastVLM test padding-side update; same general area of tests, but not the same concrete bug." + "reason": "Both mention device/test failures, but 45487 is an AltCLIP/ChineseCLIP model-parallel code fix while 45592 is only a FastVLM test padding-side adjustment." }, { "left": "pull_request:45487", "right": "pull_request:45596", "accept": false, - "reason": "Code-path fix for AltCLIP/ChineseCLIP vs BLT test expectation changes on XPU; unrelated models and not mergeable as one PR." + "reason": "Different model families and change types; 45487 fixes AltCLIP/ChineseCLIP runtime behavior, 45596 only updates BLT XPU test expectations." }, { "left": "pull_request:45487", "right": "pull_request:45615", "accept": false, - "reason": "AltCLIP/ChineseCLIP model code fix vs Qianfan OCR XPU expectation updates; different underlying issues." + "reason": "Not the same bug: 45487 is a model code fix, while 45615 only adds XPU expectations for Qianfan OCR tests." }, { "left": "pull_request:45592", "right": "pull_request:45596", "accept": false, - "reason": "Both are test changes, but they target different models and distinct behaviors (padding side vs XPU output expectations)." + "reason": "Both are test-related, but they target different models and distinct failures; no shared concrete code-path." }, { "left": "pull_request:45592", "right": "pull_request:45615", "accept": false, - "reason": "FastVLM test padding-side issue and Qianfan OCR XPU expectations are separate model-specific test fixes." + "reason": "Different test suites and outputs; these are separate device-specific expectations, not one underlying change." }, { "left": "pull_request:45596", "right": "pull_request:45615", "accept": false, - "reason": "Both add XPU expectations, but for different models and outputs; same backend theme is too broad to treat as duplicate." + "reason": "Both add XPU expectations, but for unrelated models and outputs, so they are not the same underlying bug or change." } ] }, @@ -17712,7 +17694,7 @@ "soft_pairs": [ "pull_request:45415|pull_request:45425" ], - "prepared_review_unit_hash": "00c9979ed975c493175698f5caab1041e4a29199e20ae30d8ea8bce4e793bf29", + "prepared_review_unit_hash": "8b4294fa11631aa382e969d55cd3cce9d4c25c2c9eecc76d5582adbf0c0f7cd4", "component_budget": { "node_count": 2, "item_count": 2, @@ -17734,18 +17716,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two separate typing-related pull requests, not duplicates: one is a broad type-checking sweep across many `src/transformers/*` files, while the other adds typing helper protocols and targeted fixes in a smaller set of modules.", - "confidence": 0.91, + "summary": "Both PRs are typing-related, but they address different scopes: one is a broad type-checking sweep across many Transformers modules, while the other adds typing helper protocols and a narrower modeling_utils-related cleanup. They share some files, but not the same concrete change.", + "confidence": 0.88, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: the PRs solve different typing tasks and are not the same concrete change.", + "canonical_pr_reason": "PR #45425 is the better canonical representative because it is a focused typing infrastructure change with a coherent purpose, whereas PR #45415 is a much broader multi-file type-checking sweep with mixed, unrelated edits.", "best_issue_reason": null, - "best_pr_reason": "PR #45425 is the more focused and concrete typing fix; PR #45415 is a broad repository-wide type-checking cleanup rather than the same underlying change.", + "best_pr_reason": "PR #45425 is the best overall fit for the cluster because it has the clearest single change theme and the least heterogeneous scope.", "soft_edge_verdicts": [ { "left": "pull_request:45415", "right": "pull_request:45425", "accept": false, - "reason": "Shared files are incidental; #45415 is a wide mypy/type-checking sweep across many modules, while #45425 adds specific typing aliases/protocols and targeted fixes in a different code path. They are not the same bug or change and would not be merged as one PR." + "reason": "Reject: the overlap is only in typing-related files and shared cleanup areas, but the PRs do not appear to fix the same concrete code-path problem or be safely mergeable into one PR." } ] }, diff --git a/analysis/current/manifest.json b/analysis/current/manifest.json index b9e03596f78b5ba731f87f36ec32c4d47ddcf934..24be8725030cb7ae653728a41e35cba71c71931f 100644 --- a/analysis/current/manifest.json +++ b/analysis/current/manifest.json @@ -1,8 +1,8 @@ { - "analysis_id": "hybrid-model-20260425t000023z", + "analysis_id": "hybrid-model-20260425t060019z", "archived_artifacts": { - "hybrid": "snapshots/20260425T000023Z/analysis-runs/hybrid-model-20260425t000023z/analysis-report-hybrid.json", - "hybrid_reviews": "snapshots/20260425T000023Z/analysis-runs/hybrid-model-20260425t000023z/analysis-report-hybrid.llm-reviews.json" + "hybrid": "snapshots/20260425T060019Z/analysis-runs/hybrid-model-20260425t060019z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260425T060019Z/analysis-runs/hybrid-model-20260425t060019z/analysis-report-hybrid.llm-reviews.json" }, "artifacts": { "hybrid": "analysis/current/analysis-report-hybrid.json", @@ -10,9 +10,9 @@ }, "channel": "canonical", "model": null, - "published_at": "2026-04-25T00:03:43Z", + "published_at": "2026-04-25T06:10:12Z", "repo": "huggingface/transformers", "schema_version": 1, - "snapshot_id": "20260425T000023Z", + "snapshot_id": "20260425T060019Z", "variant": "hybrid" } diff --git a/snapshots/20260425T060019Z/analysis-runs/hybrid-model-20260425t060019z/analysis-report-hybrid.json b/snapshots/20260425T060019Z/analysis-runs/hybrid-model-20260425t060019z/analysis-report-hybrid.json new file mode 100644 index 0000000000000000000000000000000000000000..c049595e0ac61e0bf1dc9512b54f8ce83e85490d --- /dev/null +++ b/snapshots/20260425T060019Z/analysis-runs/hybrid-model-20260425t060019z/analysis-report-hybrid.json @@ -0,0 +1,1750 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260425T060019Z", + "generated_at": "2026-04-25T06:10:11Z", + "evidence_quality": "full", + "llm_enrichment": true, + "meta_bugs": [ + { + "cluster_id": "cluster-43979-11", + "summary": "Cluster of 1 issues and 10 PRs centered on issue #43979.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 43979, + "canonical_pr_number": 44007, + "issue_numbers": [ + 43979 + ], + "pr_numbers": [ + 43996, + 44007, + 44013, + 44044, + 44066, + 44072, + 44085, + 44129, + 44154, + 44722 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 43996, + "right_pr_number": 44007, + "code_similarity": 0.179, + "size_similarity": 0.576, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.429, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44013, + "code_similarity": 0.122, + "size_similarity": 0.318, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.392, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44044, + "code_similarity": 0.245, + "size_similarity": 0.864, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.479, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44066, + "code_similarity": 0.225, + "size_similarity": 0.818, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.408, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44072, + "code_similarity": 0.14, + "size_similarity": 0.303, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.528, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44085, + "code_similarity": 0.216, + "size_similarity": 0.783, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.398, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44129, + "code_similarity": 0.163, + "size_similarity": 0.643, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.229, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44154, + "code_similarity": 0.153, + "size_similarity": 0.535, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.31, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44722, + "code_similarity": 0.225, + "size_similarity": 0.848, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.368, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44013, + "code_similarity": 0.19, + "size_similarity": 0.553, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.531, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44044, + "code_similarity": 0.186, + "size_similarity": 0.667, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.354, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44066, + "code_similarity": 0.188, + "size_similarity": 0.704, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.315, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44072, + "code_similarity": 0.212, + "size_similarity": 0.526, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.708, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44085, + "code_similarity": 0.195, + "size_similarity": 0.735, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.318, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44129, + "code_similarity": 0.103, + "size_similarity": 0.37, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.191, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44154, + "code_similarity": 0.238, + "size_similarity": 0.93, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.344, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44722, + "code_similarity": 0.178, + "size_similarity": 0.679, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.28, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44044, + "code_similarity": 0.126, + "size_similarity": 0.368, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.351, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44066, + "code_similarity": 0.127, + "size_similarity": 0.389, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.325, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44072, + "code_similarity": 0.29, + "size_similarity": 0.952, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.667, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44085, + "code_similarity": 0.131, + "size_similarity": 0.406, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.329, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44129, + "code_similarity": 0.07, + "size_similarity": 0.205, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.192, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44154, + "code_similarity": 0.177, + "size_similarity": 0.594, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.389, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44722, + "code_similarity": 0.118, + "size_similarity": 0.375, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.287, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44066, + "code_similarity": 0.25, + "size_similarity": 0.947, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.404, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44072, + "code_similarity": 0.136, + "size_similarity": 0.351, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.442, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44085, + "code_similarity": 0.24, + "size_similarity": 0.906, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.394, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44129, + "code_similarity": 0.147, + "size_similarity": 0.555, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.243, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44154, + "code_similarity": 0.17, + "size_similarity": 0.62, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.306, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44722, + "code_similarity": 0.257, + "size_similarity": 0.982, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.402, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44072, + "code_similarity": 0.133, + "size_similarity": 0.37, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.393, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44085, + "code_similarity": 0.763, + "size_similarity": 0.957, + "file_overlap": 0.5, + "area_overlap": 0.825, + "patch_similarity": 0.887, + "shared_filenames": [ + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 41 + ], + [ + 174, + 181 + ], + [ + 250, + 257 + ], + [ + 398, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 466 + ], + [ + 471, + 488 + ], + [ + 518, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 606 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ], + "right_ranges": [ + [ + 33, + 40 + ], + [ + 173, + 180 + ], + [ + 249, + 256 + ], + [ + 397, + 405 + ], + [ + 408, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 465 + ], + [ + 470, + 489 + ], + [ + 517, + 539 + ], + [ + 553, + 559 + ], + [ + 565, + 574 + ], + [ + 579, + 597 + ], + [ + 600, + 611 + ], + [ + 633, + 639 + ], + [ + 645, + 652 + ], + [ + 657, + 673 + ], + [ + 716, + 728 + ], + [ + 737, + 743 + ], + [ + 748, + 771 + ], + [ + 789, + 794 + ] + ] + } + ] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44129, + "code_similarity": 0.145, + "size_similarity": 0.526, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.263, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44154, + "code_similarity": 0.174, + "size_similarity": 0.654, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.286, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44722, + "code_similarity": 0.907, + "size_similarity": 0.964, + "file_overlap": 1.0, + "area_overlap": 0.808, + "patch_similarity": 0.874, + "shared_filenames": [ + "src/transformers/models/codegen/modeling_codegen.py", + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/codegen/modeling_codegen.py", + "left_ranges": [ + [ + 245, + 266 + ] + ], + "right_ranges": [ + [ + 228, + 234 + ] + ] + }, + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 41 + ], + [ + 174, + 181 + ], + [ + 250, + 257 + ], + [ + 398, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 466 + ], + [ + 471, + 488 + ], + [ + 518, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 606 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ], + "right_ranges": [ + [ + 33, + 42 + ], + [ + 175, + 182 + ], + [ + 251, + 258 + ], + [ + 399, + 420 + ], + [ + 426, + 435 + ], + [ + 460, + 467 + ], + [ + 472, + 489 + ], + [ + 519, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 612 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 663 + ], + [ + 666, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ] + } + ] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44085, + "code_similarity": 0.137, + "size_similarity": 0.387, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.398, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44129, + "code_similarity": 0.074, + "size_similarity": 0.195, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.231, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44154, + "code_similarity": 0.175, + "size_similarity": 0.566, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.414, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44722, + "code_similarity": 0.124, + "size_similarity": 0.357, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.347, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44129, + "code_similarity": 0.141, + "size_similarity": 0.503, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.272, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44154, + "code_similarity": 0.18, + "size_similarity": 0.684, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.289, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44722, + "code_similarity": 0.728, + "size_similarity": 0.923, + "file_overlap": 0.5, + "area_overlap": 0.791, + "patch_similarity": 0.78, + "shared_filenames": [ + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 40 + ], + [ + 173, + 180 + ], + [ + 249, + 256 + ], + [ + 397, + 405 + ], + [ + 408, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 465 + ], + [ + 470, + 489 + ], + [ + 517, + 539 + ], + [ + 553, + 559 + ], + [ + 565, + 574 + ], + [ + 579, + 597 + ], + [ + 600, + 611 + ], + [ + 633, + 639 + ], + [ + 645, + 652 + ], + [ + 657, + 673 + ], + [ + 716, + 728 + ], + [ + 737, + 743 + ], + [ + 748, + 771 + ], + [ + 789, + 794 + ] + ], + "right_ranges": [ + [ + 33, + 42 + ], + [ + 175, + 182 + ], + [ + 251, + 258 + ], + [ + 399, + 420 + ], + [ + 426, + 435 + ], + [ + 460, + 467 + ], + [ + 472, + 489 + ], + [ + 519, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 612 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 663 + ], + [ + 666, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ] + } + ] + }, + { + "left_pr_number": 44129, + "right_pr_number": 44154, + "code_similarity": 0.099, + "size_similarity": 0.344, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.199, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44129, + "right_pr_number": 44722, + "code_similarity": 0.146, + "size_similarity": 0.545, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.247, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44154, + "right_pr_number": 44722, + "code_similarity": 0.164, + "size_similarity": 0.631, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.255, + "shared_filenames": [], + "shared_file_areas": [] + } + ] + }, + { + "cluster_id": "cluster-41211-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #41211.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 41211, + "canonical_pr_number": 44339, + "issue_numbers": [ + 41211 + ], + "pr_numbers": [ + 41356, + 44339 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 41356, + "right_pr_number": 44339, + "code_similarity": 0.155, + "size_similarity": 0.094, + "file_overlap": 0.273, + "area_overlap": 0.078, + "patch_similarity": 0.183, + "shared_filenames": [ + "docs/source/en/model_doc/deimv2.md", + "src/transformers/models/deimv2/__init__.py", + "src/transformers/models/deimv2/configuration_deimv2.py", + "src/transformers/models/deimv2/modeling_deimv2.py", + "tests/models/deimv2/__init__.py", + "tests/models/deimv2/test_modeling_deimv2.py" + ], + "shared_file_areas": [ + { + "filename": "docs/source/en/model_doc/deimv2.md", + "left_ranges": [ + [ + 1, + 132 + ] + ], + "right_ranges": [ + [ + 1, + 65 + ] + ] + }, + { + "filename": "src/transformers/models/deimv2/__init__.py", + "left_ranges": [ + [ + 1, + 15 + ] + ], + "right_ranges": [ + [ + 1, + 29 + ] + ] + }, + { + "filename": "src/transformers/models/deimv2/configuration_deimv2.py", + "left_ranges": [ + [ + 1, + 74 + ] + ], + "right_ranges": [ + [ + 1, + 266 + ] + ] + }, + { + "filename": "tests/models/deimv2/test_modeling_deimv2.py", + "left_ranges": [ + [ + 1, + 15 + ] + ], + "right_ranges": [ + [ + 1, + 1734 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-43240-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #43240.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 43240, + "canonical_pr_number": 43251, + "issue_numbers": [ + 43240 + ], + "pr_numbers": [ + 43251, + 43254 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 43251, + "right_pr_number": 43254, + "code_similarity": 0.794, + "size_similarity": 0.64, + "file_overlap": 1.0, + "area_overlap": 0.667, + "patch_similarity": 0.882, + "shared_filenames": [ + "src/transformers/loss/loss_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/loss/loss_utils.py", + "left_ranges": [ + [ + 30, + 50 + ] + ], + "right_ranges": [ + [ + 30, + 43 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-43824-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #43824.", + "status": "open", + "confidence": 0.55, + "canonical_issue_number": 43824, + "canonical_pr_number": 43836, + "issue_numbers": [ + 43824 + ], + "pr_numbers": [ + 43836, + 43842 + ], + "evidence_types": [ + "closing_reference" + ], + "pr_comparisons": [ + { + "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-45081-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45081.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 45081, + "canonical_pr_number": 45317, + "issue_numbers": [ + 45081 + ], + "pr_numbers": [ + 45086, + 45317 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 45086, + "right_pr_number": 45317, + "code_similarity": 0.589, + "size_similarity": 0.34, + "file_overlap": 0.5, + "area_overlap": 1.0, + "patch_similarity": 0.136, + "shared_filenames": [ + "src/transformers/tokenization_utils_tokenizers.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/tokenization_utils_tokenizers.py", + "left_ranges": [ + [ + 1360, + 1370 + ], + [ + 1374, + 1380 + ] + ], + "right_ranges": [ + [ + 1360, + 1370 + ], + [ + 1374, + 1380 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-43698-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #43698.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 43698, + "canonical_pr_number": 43779, + "issue_numbers": [ + 43698 + ], + "pr_numbers": [ + 43779, + 43816 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 43779, + "right_pr_number": 43816, + "code_similarity": 0.418, + "size_similarity": 0.538, + "file_overlap": 1.0, + "area_overlap": 0.02, + "patch_similarity": 0.02, + "shared_filenames": [ + "src/transformers/integrations/integration_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/integrations/integration_utils.py", + "left_ranges": [ + [ + 2227, + 2241 + ], + [ + 2303, + 2309 + ] + ], + "right_ranges": [ + [ + 2278, + 2291 + ], + [ + 2309, + 2322 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-44018-2", + "summary": "Cluster of 2 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 44068, + "issue_numbers": [], + "pr_numbers": [ + 44018, + 44068 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 44018, + "right_pr_number": 44068, + "code_similarity": 0.766, + "size_similarity": 0.939, + "file_overlap": 1.0, + "area_overlap": 0.425, + "patch_similarity": 0.866, + "shared_filenames": [ + "src/transformers/models/gpt_neo/modeling_gpt_neo.py", + "tests/models/gpt_neo/test_modeling_gpt_neo.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/gpt_neo/modeling_gpt_neo.py", + "left_ranges": [ + [ + 26, + 46 + ], + [ + 138, + 143 + ], + [ + 182, + 187 + ], + [ + 283, + 295 + ], + [ + 327, + 341 + ], + [ + 348, + 354 + ], + [ + 360, + 369 + ], + [ + 399, + 406 + ], + [ + 411, + 419 + ], + [ + 428, + 433 + ], + [ + 472, + 492 + ], + [ + 507, + 513 + ], + [ + 519, + 528 + ], + [ + 541, + 559 + ], + [ + 562, + 567 + ], + [ + 595, + 601 + ], + [ + 607, + 614 + ], + [ + 627, + 643 + ], + [ + 685, + 690 + ], + [ + 708, + 714 + ], + [ + 720, + 727 + ], + [ + 740, + 757 + ], + [ + 761, + 766 + ], + [ + 780, + 786 + ], + [ + 791, + 798 + ], + [ + 807, + 822 + ], + [ + 840, + 845 + ] + ], + "right_ranges": [ + [ + 26, + 31 + ], + [ + 34, + 44 + ], + [ + 136, + 141 + ], + [ + 180, + 185 + ], + [ + 281, + 293 + ], + [ + 325, + 339 + ], + [ + 346, + 352 + ], + [ + 358, + 367 + ], + [ + 397, + 404 + ], + [ + 409, + 417 + ], + [ + 426, + 434 + ], + [ + 463, + 483 + ], + [ + 498, + 504 + ], + [ + 510, + 519 + ], + [ + 532, + 550 + ], + [ + 553, + 564 + ], + [ + 586, + 592 + ], + [ + 598, + 605 + ], + [ + 618, + 634 + ], + [ + 676, + 688 + ], + [ + 699, + 705 + ], + [ + 711, + 718 + ], + [ + 731, + 748 + ], + [ + 752, + 762 + ], + [ + 771, + 777 + ], + [ + 782, + 789 + ], + [ + 798, + 813 + ], + [ + 831, + 836 + ] + ] + }, + { + "filename": "tests/models/gpt_neo/test_modeling_gpt_neo.py", + "left_ranges": [ + [ + 458, + 464 + ] + ], + "right_ranges": [ + [ + 458, + 464 + ] + ] + } + ] + } + ] + } + ], + "duplicate_issues": [], + "duplicate_prs": [ + { + "cluster_id": "cluster-41211-3", + "canonical_pr_number": 44339, + "duplicate_pr_numbers": [ + 41356 + ], + "target_issue_number": 41211, + "reason": "PRs in cluster-41211-3 are treated as duplicates because they converge on issue #41211 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-43240-3", + "canonical_pr_number": 43251, + "duplicate_pr_numbers": [ + 43254 + ], + "target_issue_number": 43240, + "reason": "PRs in cluster-43240-3 are treated as duplicates because they converge on issue #43240 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-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-43824-3", + "canonical_pr_number": 43836, + "duplicate_pr_numbers": [ + 43842 + ], + "target_issue_number": 43824, + "reason": "PRs in cluster-43824-3 are treated as duplicates because they converge on issue #43824 with closing_reference evidence." + }, + { + "cluster_id": "cluster-43979-11", + "canonical_pr_number": 44007, + "duplicate_pr_numbers": [ + 43996, + 44013, + 44044, + 44066, + 44072, + 44085, + 44129, + 44154, + 44722 + ], + "target_issue_number": 43979, + "reason": "PRs in cluster-43979-11 are treated as duplicates because they converge on issue #43979 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-44018-2", + "canonical_pr_number": 44068, + "duplicate_pr_numbers": [ + 44018 + ], + "target_issue_number": null, + "reason": "PRs in cluster-44018-2 are treated as duplicates because they share soft_similarity evidence." + }, + { + "cluster_id": "cluster-45081-3", + "canonical_pr_number": 45317, + "duplicate_pr_numbers": [ + 45086 + ], + "target_issue_number": 45081, + "reason": "PRs in cluster-45081-3 are treated as duplicates because they converge on issue #45081 with closing_reference, shared_issue_target, soft_similarity evidence." + } + ], + "best_issue": { + "cluster_id": "cluster-41211-3", + "issue_number": 41211, + "reason": "Issue #41211 is the strongest global issue candidate because it is open, belongs to a cluster with 3 artifacts, and carries 6 discussion comments plus 6 inbound references.", + "score": 148.0 + }, + "best_pr": { + "cluster_id": "cluster-43979-11", + "pr_number": 44007, + "reason": "PR #44007 is the strongest global PR candidate because it is open, belongs to a cluster with 11 artifacts, links to 1 issues, and carries 2 review/discussion events.", + "score": 238.886 + } +} diff --git a/snapshots/20260425T060019Z/analysis-runs/hybrid-model-20260425t060019z/analysis-report-hybrid.llm-reviews.json b/snapshots/20260425T060019Z/analysis-runs/hybrid-model-20260425t060019z/analysis-report-hybrid.llm-reviews.json new file mode 100644 index 0000000000000000000000000000000000000000..715dc1a474709dfd6884f6651904c566f7b1f83c --- /dev/null +++ b/snapshots/20260425T060019Z/analysis-runs/hybrid-model-20260425t060019z/analysis-report-hybrid.llm-reviews.json @@ -0,0 +1,17742 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260425T060019Z", + "generated_at": "2026-04-25T06:10:11Z", + "model": "gpt-5.4-mini?service_tier=flex", + "reviews": [ + { + "label": "issue", + "component_index": 1, + "component_count": 13, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-13244-3", + "nodes": [ + "issue:13244", + "issue:15980", + "issue:16618" + ], + "soft_pairs": [ + "issue:13244|issue:16618", + "issue:15980|issue:16618" + ], + "prepared_review_unit_hash": "9c0637be12a32d10044e3b0909db2a845cb6aa29108062979602cbe64c140d1f", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 2477, + "estimated_input_tokens": 620, + "estimated_eval_tokens": 1496 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 2477, + "estimated_input_tokens": 620, + "estimated_eval_tokens": 1496 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These three issues are not duplicates of one another: one is about Tapas tokenization mismatch, one about a bad auth/error message for private models, and one about failing to load bert-base-uncased. The cluster looks spurious rather than a single underlying bug.", + "confidence": 0.98, + "canonical_issue_reason": "No single canonical issue fits here; the reports describe unrelated problems.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 16618 is the broadest and most active report in the set, but it still does not represent the other two bugs.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:13244", + "right": "issue:16618", + "accept": false, + "reason": "Tapas tokenization mismatch vs. model loading failure for bert-base-uncased are different code-path problems." + }, + { + "left": "issue:15980", + "right": "issue:16618", + "accept": false, + "reason": "Private-model login/error-message handling is unrelated to the general 'can't load bert-base-uncased' report." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 1, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:24643", + "issue:30277", + "issue:34634", + "issue:34928", + "issue:35545", + "issue:36010", + "issue:36331", + "issue:39290", + "issue:41093", + "issue:41720", + "issue:41762", + "issue:42915", + "issue:43039", + "issue:43296", + "issue:43366", + "issue:43404", + "issue:43425", + "issue:43531", + "issue:43541", + "issue:43596", + "issue:43638", + "issue:43716", + "issue:43792", + "issue:43827", + "issue:43828", + "issue:43854", + "issue:43866", + "issue:43901", + "issue:43975", + "issue:44292", + "issue:44322", + "issue:44457", + "issue:44509", + "issue:44512", + "issue:44560", + "issue:44661", + "issue:44805", + "issue:44841", + "issue:44863", + "issue:44918", + "issue:45070", + "issue:45081", + "issue:45084", + "issue:45161", + "issue:45237", + "issue:45362", + "issue:45464", + "issue:45507" + ], + "soft_pairs": [ + "issue:44322|issue:45464", + "issue:43716|issue:45237", + "issue:43828|issue:45237", + "issue:45237|issue:45507", + "issue:44292|issue:45237", + "issue:30277|issue:35545", + "issue:35545|issue:42915", + "issue:35545|issue:41720", + "issue:44509|issue:44512", + "issue:35545|issue:36010", + "issue:43638|issue:44805", + "issue:43596|issue:44805", + "issue:44805|issue:45161", + "issue:44805|issue:44918", + "issue:44661|issue:44805", + "issue:30277|issue:43638", + "issue:34634|issue:35545", + "issue:43596|issue:43638", + "issue:43975|issue:44457", + "issue:43039|issue:45161", + "issue:44292|issue:44457", + "issue:34928|issue:44805", + "issue:24643|issue:30277", + "issue:43975|issue:45237", + "issue:43638|issue:45161", + "issue:34928|issue:36331", + "issue:43596|issue:45161", + "issue:43716|issue:45507", + "issue:39290|issue:45081", + "issue:43828|issue:45161", + "issue:43425|issue:44292", + "issue:44918|issue:45161", + "issue:34928|issue:41720", + "issue:43541|issue:43716", + "issue:43828|issue:44292", + "issue:43716|issue:44292", + "issue:44292|issue:44918", + "issue:43039|issue:44918", + "issue:41762|issue:43638", + "issue:43827|issue:44512", + "issue:43716|issue:44560", + "issue:43716|issue:44918", + "issue:43039|issue:43828", + "issue:30277|issue:36331", + "issue:43039|issue:44805", + "issue:34928|issue:36010", + "issue:41762|issue:43596", + "issue:43901|issue:44512", + "issue:39290|issue:41720", + "issue:43039|issue:44560", + "issue:43866|issue:44863", + "issue:43854|issue:43866", + "issue:43425|issue:45070", + "issue:41093|issue:43638", + "issue:39290|issue:43531", + "issue:43404|issue:43866", + "issue:45362|issue:45464", + "issue:43792|issue:45070", + "issue:43541|issue:43828", + "issue:36010|issue:39290", + "issue:44841|issue:45084", + "issue:41093|issue:41762", + "issue:43296|issue:43366" + ], + "prepared_review_unit_hash": "0803d07dfe8d664e5b8b7f87a0c8722b9cd99c34a2bcdf649913c3fa521b2477", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 63, + "serialized_chars": 46481, + "estimated_input_tokens": 11621, + "estimated_eval_tokens": 23498 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous, but there is a small clear duplicate family around DeepSpeed ZeRO-3 loading failures that produce \"index 0 is out of bounds for dimension 0 with size 0\". Most other soft pairs are only loosely similar and should stay separate.", + "confidence": 0.76, + "canonical_issue_reason": "Issue 43596 is the clearest representative of the ZeRO-3 index-out-of-bounds loading bug and matches the strongest near-duplicate reports in this set.", + "canonical_pr_reason": null, + "best_issue_reason": "43596 is the most central issue to anchor the duplicate family because it names the concrete DeepSpeed ZeRO-3 failure mode without being overly model-specific.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44322", + "right": "issue:45464", + "accept": false, + "reason": "Both are Qwen3.5-related, but they describe different failures in different entrypoints and are not the same concrete bug." + }, + { + "left": "issue:43716", + "right": "issue:45237", + "accept": false, + "reason": "One is a dtype mismatch in image preprocessing, the other is a GPU/AMD compatibility problem; too different." + }, + { + "left": "issue:43828", + "right": "issue:45237", + "accept": false, + "reason": "Autocast dtype mismatch is not the same underlying issue as model failing on AMD GPUs." + }, + { + "left": "issue:45237", + "right": "issue:45507", + "accept": false, + "reason": "Different model families and different failure modes; no clear shared code-path bug." + }, + { + "left": "issue:44292", + "right": "issue:45237", + "accept": false, + "reason": "Both mention Qwen-3 variants, but the reported errors are unrelated." + }, + { + "left": "issue:30277", + "right": "issue:35545", + "accept": false, + "reason": "DeepSpeed collective mismatch and ONNX export failure are unrelated." + }, + { + "left": "issue:35545", + "right": "issue:42915", + "accept": false, + "reason": "Different models and different export/runtime problems; not the same bug." + }, + { + "left": "issue:35545", + "right": "issue:41720", + "accept": false, + "reason": "ONNX export error and CUDA assert under auto device mapping are different issues." + }, + { + "left": "issue:44509", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs regressions, but they cover different removed commands/tasks and do not look like one change." + }, + { + "left": "issue:35545", + "right": "issue:36010", + "accept": false, + "reason": "Export to ONNX and missing GenerationMixin import are unrelated." + }, + { + "left": "issue:43638", + "right": "issue:44805", + "accept": false, + "reason": "Both are index/shape errors, but the titles point to different concrete failure patterns." + }, + { + "left": "issue:43596", + "right": "issue:44805", + "accept": false, + "reason": "Zero-3 init index-out-of-bounds and mask-shape mismatch are not the same bug." + }, + { + "left": "issue:44805", + "right": "issue:45161", + "accept": false, + "reason": "Different error signatures and different subsystems." + }, + { + "left": "issue:44805", + "right": "issue:44918", + "accept": false, + "reason": "These are separate training/runtime issues with different error types." + }, + { + "left": "issue:44661", + "right": "issue:44805", + "accept": false, + "reason": "Tokenizer mapping logic and mask/tensor shape mismatch are unrelated." + }, + { + "left": "issue:30277", + "right": "issue:43638", + "accept": false, + "reason": "Both involve DeepSpeed, but one is a collective mismatch and the other is a size-0 index error." + }, + { + "left": "issue:34634", + "right": "issue:35545", + "accept": false, + "reason": "Processor voice_preset and ONNX export problems are unrelated." + }, + { + "left": "issue:43596", + "right": "issue:43638", + "accept": true, + "reason": "These describe the same ZeRO-3 index-0-out-of-bounds loading failure with only model-specific wording differences." + }, + { + "left": "issue:43975", + "right": "issue:44457", + "accept": false, + "reason": "Different language/model-specific inference behavior; no evidence they are the same bug." + }, + { + "left": "issue:43039", + "right": "issue:45161", + "accept": false, + "reason": "Liger Kernel cross-entropy invocation and GPT-OSS tensor-parallel failure are different problems." + }, + { + "left": "issue:44292", + "right": "issue:44457", + "accept": false, + "reason": "Different Qwen3.5/LoRA-related reports with different symptoms." + }, + { + "left": "issue:34928", + "right": "issue:44805", + "accept": false, + "reason": "Activation-checkpointing/FSDP recomputation mismatch is distinct from the mask index error." + }, + { + "left": "issue:24643", + "right": "issue:30277", + "accept": false, + "reason": "Both mention DeepSpeed, but the training weight-dimension error and collective mismatch are different bugs." + }, + { + "left": "issue:43975", + "right": "issue:45237", + "accept": false, + "reason": "Unrelated model detokenization vs AMD GPU runtime issue." + }, + { + "left": "issue:43638", + "right": "issue:45161", + "accept": false, + "reason": "Zero-3/BERT load failure is unrelated to GPT-OSS tensor-parallel behavior." + }, + { + "left": "issue:34928", + "right": "issue:36331", + "accept": false, + "reason": "FSDP checkpointing mismatch and CustomTrainer signature change are unrelated." + }, + { + "left": "issue:43596", + "right": "issue:45161", + "accept": false, + "reason": "Different model-loading bug versus tensor-parallel runtime issue." + }, + { + "left": "issue:43716", + "right": "issue:44292", + "accept": false, + "reason": "Different models and different error contexts; not the same underlying defect." + }, + { + "left": "issue:39290", + "right": "issue:45081", + "accept": false, + "reason": "Gemma3/vLLM attribute error and Mistral tokenizer patch crash are unrelated." + }, + { + "left": "issue:43828", + "right": "issue:45161", + "accept": false, + "reason": "Autocast dtype mismatch and GPT-OSS TP failure do not match." + }, + { + "left": "issue:30277", + "right": "issue:36331", + "accept": false, + "reason": "Collective mismatch in DeepSpeed and trainer kwarg incompatibility are separate issues." + }, + { + "left": "issue:43039", + "right": "issue:44805", + "accept": false, + "reason": "Liger Kernel cross-entropy invocation is unrelated to mask/tensor indexing errors." + }, + { + "left": "issue:34928", + "right": "issue:36010", + "accept": false, + "reason": "Activation checkpointing/FSDP issue does not match the import error." + }, + { + "left": "issue:41762", + "right": "issue:43596", + "accept": true, + "reason": "Both are the same DeepSpeed ZeRO-3 index-0-out-of-bounds failure appearing during model loading." + }, + { + "left": "issue:43827", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs regressions, but they target different removed guidance and do not look like one duplicate." + }, + { + "left": "issue:43716", + "right": "issue:44560", + "accept": false, + "reason": "Different model families and different symptom classes." + }, + { + "left": "issue:43716", + "right": "issue:44918", + "accept": false, + "reason": "Image-preprocessor dtype mismatch and Qwen3.5 SFT embedding unpacking are unrelated." + }, + { + "left": "issue:43039", + "right": "issue:43828", + "accept": false, + "reason": "Both touch loss/dtype behavior, but they are not the same concrete bug." + }, + { + "left": "issue:30277", + "right": "issue:36331", + "accept": false, + "reason": "DeepSpeed rank collective mismatch and CustomTrainer API mismatch are unrelated." + }, + { + "left": "issue:43039", + "right": "issue:44560", + "accept": false, + "reason": "Liger Kernel cross-entropy usage and Qwen3-vl embedding video StopIteration are different failures." + }, + { + "left": "issue:43866", + "right": "issue:44863", + "accept": false, + "reason": "Corrupted checkpoint and checkpoint loading failure are too vague and not clearly the same artifact." + }, + { + "left": "issue:43854", + "right": "issue:43866", + "accept": false, + "reason": "Unit test load failure and corrupted checkpoint are different problems." + }, + { + "left": "issue:43425", + "right": "issue:45070", + "accept": false, + "reason": "Torch compatibility and pydantic model field regression are unrelated." + }, + { + "left": "issue:41093", + "right": "issue:43638", + "accept": false, + "reason": "Mask shape mismatch and ZeRO-3 zero-size index error are different error signatures." + }, + { + "left": "issue:39290", + "right": "issue:43531", + "accept": false, + "reason": "Both involve recent Qwen/Gemma-style models, but the reported missing attribute differs and there is no clear same bug." + }, + { + "left": "issue:43404", + "right": "issue:43866", + "accept": false, + "reason": "Untied lm_head weights and corrupted checkpoint are unrelated." + }, + { + "left": "issue:45362", + "right": "issue:45464", + "accept": false, + "reason": "Both are Qwen3.5, but different sizes and different runtime paths make this too weak to merge." + }, + { + "left": "issue:43792", + "right": "issue:45070", + "accept": false, + "reason": "Whisper inference failure and pydantic field regression are unrelated." + }, + { + "left": "issue:43541", + "right": "issue:43828", + "accept": false, + "reason": "Grouped_mm tracing error and autocast dtype mismatch are different runtime issues." + }, + { + "left": "issue:36010", + "right": "issue:39290", + "accept": false, + "reason": "ImportError for GenerationMixin and Gemma3 config attribute error are unrelated." + }, + { + "left": "issue:44841", + "right": "issue:45084", + "accept": false, + "reason": "Processor failure and template compilation error are not the same bug." + }, + { + "left": "issue:41093", + "right": "issue:41762", + "accept": false, + "reason": "Both are index-related, but one is a mask-size mismatch and the other is a zero-size load failure." + }, + { + "left": "issue:43296", + "right": "issue:43366", + "accept": false, + "reason": "Different model-support requests/errors; not the same concrete issue." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The two accepted ZeRO-3 index-out-of-bounds pairs are well supported by the titles, and the rest of the soft-edge rejections do not overclaim duplication from the available evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 2, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:36331", + "issue:39186", + "issue:39290", + "issue:39692", + "issue:40990", + "issue:41093", + "issue:41553", + "issue:41762", + "issue:42915", + "issue:43054", + "issue:43366", + "issue:43404", + "issue:43531", + "issue:43541", + "issue:43572", + "issue:43596", + "issue:43644", + "issue:43645", + "issue:43742", + "issue:43828", + "issue:43866", + "issue:43873", + "issue:43883", + "issue:43950", + "issue:43975", + "issue:43994", + "issue:44291", + "issue:44292", + "issue:44360", + "issue:44387", + "issue:44403", + "issue:44492", + "issue:44512", + "issue:44530", + "issue:44534", + "issue:44560", + "issue:44589", + "issue:44661", + "issue:44863", + "issue:44918", + "issue:44928", + "issue:44960", + "issue:45005", + "issue:45070", + "issue:45161", + "issue:45399", + "issue:45464", + "issue:45507" + ], + "soft_pairs": [ + "issue:43572|issue:45070", + "issue:44387|issue:45005", + "issue:43541|issue:45161", + "issue:43828|issue:45464", + "issue:44291|issue:45070", + "issue:43366|issue:45464", + "issue:43531|issue:45070", + "issue:44661|issue:45464", + "issue:39290|issue:43742", + "issue:43054|issue:43994", + "issue:43541|issue:44589", + "issue:41093|issue:43596", + "issue:42915|issue:43541", + "issue:40990|issue:44960", + "issue:39186|issue:44918", + "issue:43404|issue:45005", + "issue:43645|issue:45399", + "issue:44530|issue:45005", + "issue:44863|issue:45005", + "issue:43873|issue:45005", + "issue:44360|issue:44512", + "issue:42915|issue:45399", + "issue:44918|issue:45507", + "issue:44492|issue:44512", + "issue:43975|issue:44292", + "issue:43950|issue:44534", + "issue:43644|issue:43950", + "issue:44292|issue:45507", + "issue:39692|issue:43054", + "issue:43828|issue:44928", + "issue:41553|issue:43883", + "issue:43866|issue:44403", + "issue:36331|issue:41762", + "issue:44560|issue:45507" + ], + "prepared_review_unit_hash": "8e4a76b427a830fbff24d9f89ece1457b2939ffc99ee7f27ae39ec5dec96a71e", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37569, + "estimated_input_tokens": 9393, + "estimated_eval_tokens": 19042 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most pairs are not true duplicates; the only clear duplicate cluster is the non-persistent buffer corruption regression in Transformers v5 (`43644`, `43950`, `44534`).", + "confidence": 0.84, + "canonical_issue_reason": "`issue:43950` is the clearest and most specific statement of the underlying bug (`from_pretrained` corrupting `register_buffer(persistent=False)` buffers), and it matches the other buffer-corruption reports.", + "canonical_pr_reason": null, + "best_issue_reason": "`issue:43950` is the most precise, actionable representative of the duplicate cluster and best captures the concrete regression.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43572", + "right": "issue:45070", + "accept": false, + "reason": "Both are v5 config-related, but one is a missing `pad_token_idx` field in StableLM and the other is a pydantic `PretrainedConfig` regression; not the same bug." + }, + { + "left": "issue:44387", + "right": "issue:45005", + "accept": false, + "reason": "Both mention v5/quantization or weights, but the symptoms and code paths differ: int4 memory/OOM vs tied weights in translation models." + }, + { + "left": "issue:43541", + "right": "issue:45161", + "accept": false, + "reason": "Both involve MoE models, but one is a torch dynamo grouped_mm failure and the other is a TP-only support issue; not the same underlying problem." + }, + { + "left": "issue:43828", + "right": "issue:45464", + "accept": false, + "reason": "Different failures and subsystems: autocast dtype mismatch vs chat/completions streaming inference." + }, + { + "left": "issue:44291", + "right": "issue:45070", + "accept": false, + "reason": "`init_empty_weights` unexpected-arg error is a constructor/init path bug, while 45070 is a pydantic config-field regression." + }, + { + "left": "issue:43366", + "right": "issue:45464", + "accept": false, + "reason": "GGUF gpt-oss support and Qwen3.5 streaming API failures are unrelated changes with different targets." + }, + { + "left": "issue:43531", + "right": "issue:45070", + "accept": false, + "reason": "Qwen3-MoE sliding-window handling and pydantic `PretrainedConfig` field issues are not the same bug." + }, + { + "left": "issue:44661", + "right": "issue:45464", + "accept": false, + "reason": "Tokenizer mapping/add-new-model-like logic is unrelated to the Qwen3.5 streaming inference failure." + }, + { + "left": "issue:39290", + "right": "issue:43742", + "accept": false, + "reason": "Gemma3/vLLM missing attribute and MobileLLM loading key error are different model-specific loading bugs." + }, + { + "left": "issue:43054", + "right": "issue:43994", + "accept": false, + "reason": "Both are about SigLIP2, but one reports embedding quality degradation and the other a pipeline/AutoModel output issue; not clearly the same concrete bug." + }, + { + "left": "issue:43541", + "right": "issue:44589", + "accept": false, + "reason": "Dynamo tracing on Mixtral float32 and missing Float8 storage are different runtime failures." + }, + { + "left": "issue:41093", + "right": "issue:43596", + "accept": false, + "reason": "Both are index errors, but they arise in different paths: masking mismatch vs DeepSpeed ZeRO-3/BertModel init." + }, + { + "left": "issue:42915", + "right": "issue:43541", + "accept": false, + "reason": "Qwen3Moe FP8 failure and Mixtral torch dynamo grouped_mm failure are unrelated model/runtime bugs." + }, + { + "left": "issue:40990", + "right": "issue:44960", + "accept": false, + "reason": "Perplexity regression on gpt-oss-20b and the generic `GLM5` issue do not appear to describe the same bug." + }, + { + "left": "issue:39186", + "right": "issue:44918", + "accept": false, + "reason": "FSDP `'weight' must be 2-D` and Qwen3.5 embedding unpacking with TRL SFT trainer are different training/loading issues." + }, + { + "left": "issue:43404", + "right": "issue:45005", + "accept": false, + "reason": "Both mention tied weights, but one is a Mistral3 lm_head tying bug and the other is a broader translation-model v5 report; not the same concrete failure." + }, + { + "left": "issue:43645", + "right": "issue:45399", + "accept": false, + "reason": "Jupyter custom-model initialization and flash-attn2 fallback checks are unrelated regressions." + }, + { + "left": "issue:44530", + "right": "issue:45005", + "accept": false, + "reason": "PagedAttentionCache/linear_attention on Qwen3.5 is unrelated to tied-weights handling in translation models." + }, + { + "left": "issue:44863", + "right": "issue:45005", + "accept": false, + "reason": "NemotronH checkpoint loading and tied-weights translation-model issues are different model-specific bugs." + }, + { + "left": "issue:43873", + "right": "issue:45005", + "accept": false, + "reason": "Quantization/offloading behavior and tied-weights handling are different failure modes." + }, + { + "left": "issue:44360", + "right": "issue:44512", + "accept": false, + "reason": "DSA indexer behavior and removed-run-command docs are unrelated." + }, + { + "left": "issue:42915", + "right": "issue:45399", + "accept": false, + "reason": "Qwen3Moe FP8 support and flash-attn2 fallback gating are different code paths." + }, + { + "left": "issue:44918", + "right": "issue:45507", + "accept": false, + "reason": "TRL SFT unpacking and GraniteMoEHybrid invalid method calls are unrelated trainer/model bugs." + }, + { + "left": "issue:44492", + "right": "issue:44512", + "accept": false, + "reason": "Cache-strategy typo and docs about a removed command are both documentation-level but not the same issue." + }, + { + "left": "issue:43975", + "right": "issue:44292", + "accept": false, + "reason": "Detokenization on DeepSeek Coder and Qwen-3-8B-NVFP4 runtime errors are different model problems." + }, + { + "left": "issue:43950", + "right": "issue:44534", + "accept": true, + "reason": "Same regression: both report v5 corrupting non-persistent buffers / `register_buffer(persistent=False)` state during loading." + }, + { + "left": "issue:43644", + "right": "issue:43950", + "accept": true, + "reason": "Same underlying bug: Transformers v5/from_pretrained is filling or corrupting non-persistent buffers with junk." + }, + { + "left": "issue:44292", + "right": "issue:45507", + "accept": false, + "reason": "Qwen-3-8B-NVFP4 runtime error and GraniteMoEHybrid invalid method are unrelated model/runtime failures." + }, + { + "left": "issue:39692", + "right": "issue:43054", + "accept": false, + "reason": "Both involve SigLIP2, but one is docs-example/quantization issues and the other is a text-embedding quality complaint; not the same bug." + }, + { + "left": "issue:43828", + "right": "issue:44928", + "accept": false, + "reason": "Autocast dtype mismatch and RLHF gradient explosion from 3D position ids are distinct failures." + }, + { + "left": "issue:41553", + "right": "issue:43883", + "accept": false, + "reason": "Bad AutoTokenizer error messaging for Voxtral and missing `all_tied_weights_keys` on MolmoForCausalLM are unrelated." + }, + { + "left": "issue:43866", + "right": "issue:44403", + "accept": false, + "reason": "Corrupted Ovis2 checkpoint and generic loading noise are not the same concrete bug." + }, + { + "left": "issue:36331", + "right": "issue:41762", + "accept": false, + "reason": "CustomTrainer `compute_loss` signature mismatch and Gemma3 ZeRO-3 loading index error are unrelated training/loading bugs." + }, + { + "left": "issue:44560", + "right": "issue:45507", + "accept": false, + "reason": "Qwen3-vl-embedding video StopIteration and GraniteMoEHybrid invalid method calls are different errors." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The cluster summary is grounded in the packet, and the accepted soft-edge links are conservative: both accepted pairs clearly point to the same Transformers v5 non-persistent buffer corruption regression. The rejections are also generally cautious and the reasoning does not overstate equivalence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 3, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:38175", + "issue:41553", + "issue:41762", + "issue:42491", + "issue:42915", + "issue:42947", + "issue:43054", + "issue:43257", + "issue:43278", + "issue:43404", + "issue:43493", + "issue:43643", + "issue:43742", + "issue:43782", + "issue:43792", + "issue:43824", + "issue:43827", + "issue:43854", + "issue:43856", + "issue:43866", + "issue:43872", + "issue:43881", + "issue:43883", + "issue:43940", + "issue:43950", + "issue:43975", + "issue:44220", + "issue:44291", + "issue:44368", + "issue:44387", + "issue:44403", + "issue:44451", + "issue:44488", + "issue:44509", + "issue:44589", + "issue:44661", + "issue:44863", + "issue:44928", + "issue:44960", + "issue:44991", + "issue:45020", + "issue:45081", + "issue:45305", + "issue:45313", + "issue:45356", + "issue:45399", + "issue:45520", + "issue:45588" + ], + "soft_pairs": [ + "issue:41553|issue:43881", + "issue:43872|issue:44589", + "issue:38175|issue:43054", + "issue:42947|issue:44387", + "issue:43054|issue:43493", + "issue:43881|issue:43883", + "issue:44403|issue:44451", + "issue:43975|issue:45520", + "issue:43643|issue:45020", + "issue:42947|issue:43856", + "issue:44403|issue:44488", + "issue:43782|issue:44960", + "issue:42491|issue:44368", + "issue:41762|issue:42915", + "issue:43975|issue:44451", + "issue:45020|issue:45356", + "issue:44661|issue:45020", + "issue:43054|issue:43278", + "issue:43824|issue:43872", + "issue:44991|issue:45399", + "issue:43827|issue:44509", + "issue:43792|issue:44220", + "issue:45081|issue:45399", + "issue:43404|issue:43950", + "issue:43257|issue:43866", + "issue:42915|issue:45020", + "issue:43742|issue:45399", + "issue:41553|issue:44991", + "issue:43742|issue:45020", + "issue:43940|issue:45313", + "issue:43872|issue:44291", + "issue:43854|issue:44863", + "issue:44928|issue:45588", + "issue:41553|issue:44488", + "issue:42915|issue:43872", + "issue:43257|issue:45305" + ], + "prepared_review_unit_hash": "cd8879a6f1e6dd337c20e23631136f7b305c5486979dda1591d027ee5aeb276f", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 36, + "serialized_chars": 38601, + "estimated_input_tokens": 9651, + "estimated_eval_tokens": 19558 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous, but a few soft pairs are true duplicates: SigLIP2 implementation mismatch, stale v5 docs, Whisper fbank extraction, Qwen3 ZeRO-3 weight loading, and the _is_hf_initialized/init_empty_weights TypeError. Most other pairs are only broadly related by subsystem, not the same bug.", + "confidence": 0.89, + "canonical_issue_reason": "issue:45020 is the broadest representative of the cluster\u2019s main theme: recent Transformers v5 loading/tokenizer regressions, especially around remote_code and model initialization.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45020 is the best single representative because it is broad, current, and central to the model-loading regression theme spanning several items in this cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41553", + "right": "issue:43881", + "accept": false, + "reason": "Different models and failures: Voxtral AutoTokenizer error-message issue vs GLM-4v-9b loading failure." + }, + { + "left": "issue:43872", + "right": "issue:44589", + "accept": false, + "reason": "Both are quantization/storage-related, but the root causes differ: bitsandbytes _is_hf_initialized vs missing Float8 storage type." + }, + { + "left": "issue:38175", + "right": "issue:43054", + "accept": false, + "reason": "Same model family, but zero probabilities vs degraded text embeddings are not clearly the same underlying bug." + }, + { + "left": "issue:42947", + "right": "issue:44387", + "accept": false, + "reason": "Both concern training/memory behavior, but gradient checkpointing/LoRA and int4 CUDA reserved memory are distinct issues." + }, + { + "left": "issue:43054", + "right": "issue:43493", + "accept": true, + "reason": "Both point to a SigLIP2 implementation mismatch versus reference behavior, with degraded embeddings as a likely symptom of the same bug." + }, + { + "left": "issue:43881", + "right": "issue:43883", + "accept": false, + "reason": "Unrelated failures: GLM-4v loading vs Molmo missing all_tied_weights_keys." + }, + { + "left": "issue:44403", + "right": "issue:44451", + "accept": false, + "reason": "Generic loading noise is not the same as a specific ScandiBERT loading regression." + }, + { + "left": "issue:43975", + "right": "issue:45520", + "accept": false, + "reason": "Different areas: tokenizer detokenization regression vs flash_attn import KeyError on Python 3.13." + }, + { + "left": "issue:43643", + "right": "issue:45020", + "accept": false, + "reason": "Both involve remote_code/config handling, but one is missing fields in AutoConfig and the other is broader model breakage; not the same concrete bug." + }, + { + "left": "issue:42947", + "right": "issue:43856", + "accept": false, + "reason": "LoRA gradient checkpointing and Qwen3 MoE memory usage are different training bugs." + }, + { + "left": "issue:44403", + "right": "issue:44488", + "accept": false, + "reason": "Both are loading-related, but the underlying models and symptoms differ." + }, + { + "left": "issue:43782", + "right": "issue:44960", + "accept": false, + "reason": "Qwen3VL weight_only loading error and GLM5 loading are unrelated model-specific problems." + }, + { + "left": "issue:42491", + "right": "issue:44368", + "accept": false, + "reason": "LoRA model compatibility on hf4.x/hf5.x is not the same as a tie_word_embeddings warning." + }, + { + "left": "issue:41762", + "right": "issue:42915", + "accept": false, + "reason": "Both mention DeepSpeed, but Gemma3 ZeRO-3 index error and Qwen3 MoE FP8 failure are different code paths." + }, + { + "left": "issue:43975", + "right": "issue:44451", + "accept": false, + "reason": "Different tokenizer/load failures for different models." + }, + { + "left": "issue:45020", + "right": "issue:45356", + "accept": false, + "reason": "Broad remote_code regressions and Kimi-K2.5 tokenizer codec/fix_mistral_regex regression are distinct bugs." + }, + { + "left": "issue:44661", + "right": "issue:45020", + "accept": false, + "reason": "add-new-model-like failure in tokenizer mapping is a separate codegen/tooling bug, not the same as remote_code model breakage." + }, + { + "left": "issue:43054", + "right": "issue:43278", + "accept": false, + "reason": "SigLIP2 quality discrepancy and BF16-to-FP32 dtype change are different symptoms with no clear shared root cause." + }, + { + "left": "issue:43824", + "right": "issue:43872", + "accept": false, + "reason": "Qwen2.5-VL import error and bitsandbytes init_empty_weights incompatibility are unrelated." + }, + { + "left": "issue:44991", + "right": "issue:45399", + "accept": false, + "reason": "Tokenizer loading failure for est-roberta and flash-attn2 fallback blocking are unrelated." + }, + { + "left": "issue:43827", + "right": "issue:44509", + "accept": true, + "reason": "Same documentation regression: stale docs still reference pipeline tasks removed in v5." + }, + { + "left": "issue:43792", + "right": "issue:44220", + "accept": true, + "reason": "Both describe the same Whisper audio-feature path; the fbank extraction helper is the likely concrete cause of Whisper failing to run." + }, + { + "left": "issue:45081", + "right": "issue:45399", + "accept": false, + "reason": "Mistral regex patch crash and flash-attn2 fallback blocking are separate issues." + }, + { + "left": "issue:43404", + "right": "issue:43950", + "accept": false, + "reason": "lm_head tying in Mistral3 and silent corruption of non-persistent buffers are unrelated model-loading bugs." + }, + { + "left": "issue:43257", + "right": "issue:43866", + "accept": false, + "reason": "Qwen3 MoE weight conversion with accelerate+deepspeed and Ovis2 checkpoint corruption are different problems." + }, + { + "left": "issue:42915", + "right": "issue:45020", + "accept": false, + "reason": "Qwen3 MoE FP8 training failure is narrower and different from the broad remote_code loading regressions." + }, + { + "left": "issue:43742", + "right": "issue:45399", + "accept": false, + "reason": "MobileLLM loading failure and flash-attn2 fallback checks are unrelated." + }, + { + "left": "issue:41553", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer-loading issues, but they affect different models and likely different root causes." + }, + { + "left": "issue:43742", + "right": "issue:45020", + "accept": false, + "reason": "MobileLLM KeyError is a specific model load bug, not the same as the broad remote_code regression bucket." + }, + { + "left": "issue:43940", + "right": "issue:45313", + "accept": true, + "reason": "Same concrete failure class: Qwen3-family DeepSpeed ZeRO-3 weight loading leaves parameters missing." + }, + { + "left": "issue:43872", + "right": "issue:44291", + "accept": true, + "reason": "These are the same TypeError class: loading with init_empty_weights hits unexpected _is_hf_initialized handling." + }, + { + "left": "issue:43854", + "right": "issue:44863", + "accept": false, + "reason": "GLM-4.7-Flash unit-test loading problem and NemotronH checkpoint loading failure are distinct model-specific bugs." + }, + { + "left": "issue:44928", + "right": "issue:45588", + "accept": false, + "reason": "RLHF NaN from 3D position_ids/SDPA fallback and flash_attention s_aux=None crash are different failures." + }, + { + "left": "issue:41553", + "right": "issue:44488", + "accept": false, + "reason": "Different tokenizer/model-loading failures for Voxtral and sleng-bert." + }, + { + "left": "issue:42915", + "right": "issue:43872", + "accept": false, + "reason": "Qwen3 MoE FP8 training failure and bitsandbytes init_empty_weights incompatibility are unrelated." + }, + { + "left": "issue:43257", + "right": "issue:45305", + "accept": false, + "reason": "Qwen3 MoE weight conversion/loading and gradient averaging under DeepSpeed are different bugs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative on the soft edges. The accepted pairs are plausibly the same bug class, and the rejected pairs are mostly only subsystem-level matches rather than clear duplicates. No overstatement stands out." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 4, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:24643", + "issue:30064", + "issue:36010", + "issue:36331", + "issue:38175", + "issue:39692", + "issue:41093", + "issue:42491", + "issue:42617", + "issue:42915", + "issue:43317", + "issue:43366", + "issue:43531", + "issue:43550", + "issue:43596", + "issue:43638", + "issue:43644", + "issue:43645", + "issue:43646", + "issue:43720", + "issue:43824", + "issue:43828", + "issue:43872", + "issue:43950", + "issue:43957", + "issue:43994", + "issue:44155", + "issue:44479", + "issue:44530", + "issue:44534", + "issue:44560", + "issue:44589", + "issue:44661", + "issue:44756", + "issue:44805", + "issue:44863", + "issue:44877", + "issue:44910", + "issue:44918", + "issue:44960", + "issue:44977", + "issue:45030", + "issue:45313", + "issue:45325", + "issue:45357", + "issue:45538", + "issue:45584" + ], + "soft_pairs": [ + "issue:36331|issue:41093", + "issue:41093|issue:44805", + "issue:43872|issue:45538", + "issue:44479|issue:44560", + "issue:44530|issue:44863", + "issue:43366|issue:43828", + "issue:24643|issue:30064", + "issue:44155|issue:45584", + "issue:43366|issue:43531", + "issue:43550|issue:44910", + "issue:43645|issue:43646", + "issue:43950|issue:45325", + "issue:44560|issue:44918", + "issue:43366|issue:44589", + "issue:43824|issue:43957", + "issue:42915|issue:43957", + "issue:42617|issue:43366", + "issue:38175|issue:43994", + "issue:43646|issue:43950", + "issue:44560|issue:44805", + "issue:36010|issue:42915", + "issue:43638|issue:44661", + "issue:44877|issue:45030", + "issue:43596|issue:44661", + "issue:43317|issue:44756", + "issue:43644|issue:44534", + "issue:45313|issue:45357", + "issue:38175|issue:39692", + "issue:42491|issue:43720", + "issue:44960|issue:44977" + ], + "prepared_review_unit_hash": "aa7006a996d29ff284f956a94d66ee4d595c3f1c16fbf2747192deb628227d94", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36151, + "estimated_input_tokens": 9038, + "estimated_eval_tokens": 18332 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly a bag of unrelated issues, with only two clear duplicate pairs: the Transformers 5.0 custom model initialization regression, and the non-persistent buffer corruption regression. The rest are only superficially similar by title and should stay separate.", + "confidence": 0.86, + "canonical_issue_reason": "No single canonical issue exists for the full set because the items span many unrelated bugs; if a duplicate anchor is required, issue 43644 is the best representative of the non-persistent-buffer regression and is essentially duplicated by 44534.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43644 is the strongest issue candidate because it is the clearest, earliest report of the non-persistent buffer junk regression and has a near-identical duplicate in 44534.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36331", + "right": "issue:41093", + "accept": false, + "reason": "Different failures: trainer API mismatch vs mask/tensor shape IndexError." + }, + { + "left": "issue:41093", + "right": "issue:44805", + "accept": false, + "reason": "Only the error wording overlaps; the reported shapes/models and likely root causes differ." + }, + { + "left": "issue:43872", + "right": "issue:45538", + "accept": false, + "reason": "Unrelated bugs: bitsandbytes initialization error vs tokenizer model_max_length behavior." + }, + { + "left": "issue:44479", + "right": "issue:44560", + "accept": false, + "reason": "Same broad Qwen3/VL area, but different concrete bugs and code paths (video regression vs embedding/tracks issue)." + }, + { + "left": "issue:44530", + "right": "issue:44863", + "accept": false, + "reason": "Different subsystems and failures: cache/group-type crash vs checkpoint loading for a different model family." + }, + { + "left": "issue:43366", + "right": "issue:43828", + "accept": false, + "reason": "Unrelated: GGUF gpt-oss architecture support vs autocast dtype mismatch." + }, + { + "left": "issue:24643", + "right": "issue:30064", + "accept": false, + "reason": "DeepSpeed training weight-shape error is not the same bug as void segmentation map processing." + }, + { + "left": "issue:44155", + "right": "issue:45584", + "accept": false, + "reason": "Different models and mechanics: batched AudioFlamingo track leakage vs Whisper empty-transcription handling." + }, + { + "left": "issue:43366", + "right": "issue:43531", + "accept": false, + "reason": "Both mention Qwen-related topics, but one is GGUF support and the other is a sliding_window issue; not the same bug." + }, + { + "left": "issue:43550", + "right": "issue:44910", + "accept": false, + "reason": "Different attention/backend crashes with different models and error modes." + }, + { + "left": "issue:43645", + "right": "issue:43646", + "accept": true, + "reason": "Same underlying Transformers 5.0 custom model initialization regression; the notebook report is a context-specific manifestation of the general bug." + }, + { + "left": "issue:43950", + "right": "issue:45325", + "accept": false, + "reason": "Non-persistent buffer corruption regression vs Qwen2.5-VL rope scaling bug; different code paths." + }, + { + "left": "issue:44560", + "right": "issue:44918", + "accept": false, + "reason": "Both involve Qwen models, but one is a video StopIteration regression and the other is a TRL embedding unpacking failure." + }, + { + "left": "issue:43366", + "right": "issue:44589", + "accept": false, + "reason": "No shared bug: GGUF support request vs missing Float8 storage type." + }, + { + "left": "issue:43824", + "right": "issue:43957", + "accept": false, + "reason": "ImportError for Qwen2.5-VL is unrelated to meta-device loading regressions." + }, + { + "left": "issue:42915", + "right": "issue:43957", + "accept": false, + "reason": "Qwen3Moe FP8 failure and meta-device loading problems are different issues." + }, + { + "left": "issue:42617", + "right": "issue:43366", + "accept": false, + "reason": "Different problems: 3d_parallel.py execution vs GGUF architecture support." + }, + { + "left": "issue:38175", + "right": "issue:43994", + "accept": false, + "reason": "Same model family, but one is zero-probabilities behavior and the other is an AutoModel/pipeline nonsensical-results report with different root cause." + }, + { + "left": "issue:43646", + "right": "issue:43950", + "accept": false, + "reason": "Custom model init regression is distinct from the non-persistent buffer corruption regression." + }, + { + "left": "issue:44560", + "right": "issue:44805", + "accept": false, + "reason": "Different Qwen video issues; the reported failures and symptoms do not line up." + }, + { + "left": "issue:36010", + "right": "issue:42915", + "accept": false, + "reason": "ImportError for GenerationMixin is unrelated to Qwen3Moe FineGrainedFP8 loading." + }, + { + "left": "issue:43638", + "right": "issue:44661", + "accept": false, + "reason": "DeepSpeed/Bert index error vs tokenizer mapping-name constraint are unrelated." + }, + { + "left": "issue:44877", + "right": "issue:45030", + "accept": false, + "reason": "Both are config-loading regressions, but the models and failure mechanisms differ." + }, + { + "left": "issue:43596", + "right": "issue:44661", + "accept": false, + "reason": "Different bugs: ZeRO-3 init index error vs add-new-model-like failing in TOKENIZER_MAPPING_NAMES." + }, + { + "left": "issue:43317", + "right": "issue:44756", + "accept": false, + "reason": "device_map/offload loading failure is not the same as mmap/OOM on Strix Halo." + }, + { + "left": "issue:43644", + "right": "issue:44534", + "accept": true, + "reason": "Same underlying Transformers v5 regression: non-persistent buffers get corrupted/filled with junk." + }, + { + "left": "issue:45313", + "right": "issue:45357", + "accept": false, + "reason": "Different Qwen3.5 regressions: ZeRO-3 language_model weight loading vs save_pretrained visual encoder keys." + }, + { + "left": "issue:38175", + "right": "issue:39692", + "accept": false, + "reason": "SigLIP2 output anomaly vs docs example/model-processor mismatch and quantization failure are not the same bug." + }, + { + "left": "issue:42491", + "right": "issue:43720", + "accept": false, + "reason": "LoRA checkpoint compatibility on hf5.x is unrelated to packed-weight handling in BitNet AutoBitLinear." + }, + { + "left": "issue:44960", + "right": "issue:44977", + "accept": false, + "reason": "GLM5 and Qwen3.5 flash-attention failures are different model-specific bugs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The analyst output is grounded in the packet and stays conservative overall. The two accepted duplicate pairs are well supported by the titles, and the rejected soft-edge pairs mostly hinge on clearly different model/subsystem behaviors. No mergeability issues apply here because there are no PR pairs." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 5, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:36010", + "issue:39404", + "issue:39692", + "issue:41720", + "issue:42491", + "issue:42915", + "issue:42947", + "issue:43257", + "issue:43381", + "issue:43404", + "issue:43425", + "issue:43454", + "issue:43644", + "issue:43645", + "issue:43646", + "issue:43708", + "issue:43854", + "issue:43856", + "issue:43873", + "issue:43931", + "issue:43957", + "issue:43994", + "issue:44075", + "issue:44155", + "issue:44164", + "issue:44291", + "issue:44292", + "issue:44387", + "issue:44402", + "issue:44451", + "issue:44488", + "issue:44661", + "issue:44683", + "issue:44743", + "issue:44756", + "issue:44910", + "issue:44912", + "issue:44928", + "issue:44960", + "issue:44991", + "issue:45216", + "issue:45276", + "issue:45292", + "issue:45335", + "issue:45362", + "issue:45406", + "issue:45588" + ], + "soft_pairs": [ + "issue:44451|issue:44488", + "issue:43708|issue:44075", + "issue:43854|issue:43931", + "issue:43856|issue:43873", + "issue:42491|issue:45276", + "issue:41720|issue:42915", + "issue:43257|issue:43854", + "issue:39692|issue:43994", + "issue:44488|issue:44991", + "issue:45362|issue:45406", + "issue:43856|issue:44387", + "issue:43644|issue:43645", + "issue:42915|issue:44756", + "issue:44661|issue:44683", + "issue:42491|issue:43931", + "issue:44910|issue:45588", + "issue:43404|issue:43454", + "issue:42491|issue:43856", + "issue:45276|issue:45335", + "issue:36010|issue:39404", + "issue:44164|issue:44291", + "issue:44402|issue:44488", + "issue:41720|issue:44155", + "issue:43873|issue:44387", + "issue:39692|issue:44960", + "issue:43425|issue:43957", + "issue:44387|issue:44756", + "issue:43854|issue:45216", + "issue:42947|issue:43381", + "issue:44292|issue:44912", + "issue:44291|issue:44756", + "issue:44743|issue:44960", + "issue:45276|issue:45292", + "issue:43381|issue:44928", + "issue:43957|issue:44756", + "issue:44164|issue:44756", + "issue:43646|issue:43708", + "issue:45292|issue:45335" + ], + "prepared_review_unit_hash": "9c84d8d354c202417f84c58d5517d2e55f9e2a68aaff7647ce8319de72f6fe11", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 38, + "serialized_chars": 38392, + "estimated_input_tokens": 9598, + "estimated_eval_tokens": 19452 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly unrelated. The only clear duplicate set is the `resize_token_embeddings` regression around model-specific embeddings/output heads; the rest are separate model-loading, quantization, checkpointing, or runtime bugs that only overlap at a broad subsystem level.", + "confidence": 0.92, + "canonical_issue_reason": "Issue 45276 is the best canonical issue: it is the broadest and most representative report for the `resize_token_embeddings` propagation bug, and it predates the closest follow-up reports.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45276 is the strongest global issue representative because it captures the core regression most generally and matches the other accepted duplicate reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44451", + "right": "issue:44488", + "accept": false, + "reason": "Both are model-loading complaints, but they involve different models and no evidence of the same root cause." + }, + { + "left": "issue:43708", + "right": "issue:44075", + "accept": false, + "reason": "Different bugs: checkpoint resume step \u0917\u0923culation vs optimizer arguments being ignored." + }, + { + "left": "issue:43854", + "right": "issue:43931", + "accept": false, + "reason": "Different models and different failure modes; not the same underlying defect." + }, + { + "left": "issue:43856", + "right": "issue:43873", + "accept": false, + "reason": "One is training memory usage, the other is quantization/offloading behavior; too different to be duplicates." + }, + { + "left": "issue:42491", + "right": "issue:45276", + "accept": false, + "reason": "LoRA hf4.x/hf5.x compatibility is unrelated to the `resize_token_embeddings` regression." + }, + { + "left": "issue:41720", + "right": "issue:42915", + "accept": false, + "reason": "Both mention Qwen3/MoE, but the concrete failures are different code paths and symptoms." + }, + { + "left": "issue:43257", + "right": "issue:43854", + "accept": false, + "reason": "Accelerate/deepspeed conversion issue vs model-load test failure; not the same bug." + }, + { + "left": "issue:39692", + "right": "issue:43994", + "accept": false, + "reason": "Same model family, but one is a docs/example problem and the other is inference quality; insufficient to merge." + }, + { + "left": "issue:44488", + "right": "issue:44991", + "accept": false, + "reason": "Different affected models and different tokenizer/load regressions." + }, + { + "left": "issue:45362", + "right": "issue:45406", + "accept": false, + "reason": "Different products and failures: chat crash vs processor attribute error." + }, + { + "left": "issue:43856", + "right": "issue:44387", + "accept": false, + "reason": "Both are memory-related, but the reported causes are different (training footprint vs int4 OOM regression)." + }, + { + "left": "issue:43644", + "right": "issue:43645", + "accept": false, + "reason": "Both affect Transformers 5 custom-model scenarios, but the specific defects are different." + }, + { + "left": "issue:42915", + "right": "issue:44756", + "accept": false, + "reason": "Unrelated: Qwen3 MoE FP8 failure vs mmap OOM mitigation on Strix Halo." + }, + { + "left": "issue:44661", + "right": "issue:44683", + "accept": false, + "reason": "Different areas entirely: tokenizer mapping validation vs compiled flex_attention on newer torch." + }, + { + "left": "issue:42491", + "right": "issue:43931", + "accept": false, + "reason": "Different models and unrelated failures." + }, + { + "left": "issue:44910", + "right": "issue:45588", + "accept": false, + "reason": "Both touch flash attention, but one is a packed-sequence/position_ids bug and the other is an s_aux None AttributeError." + }, + { + "left": "issue:43404", + "right": "issue:43454", + "accept": false, + "reason": "Both mention untied lm_head weights, but the model families and reported behavior differ enough that this is not clearly the same bug." + }, + { + "left": "issue:45276", + "right": "issue:45335", + "accept": true, + "reason": "Same underlying bug: `resize_token_embeddings` does not propagate to model-specific embedding/output modules." + }, + { + "left": "issue:36010", + "right": "issue:39404", + "accept": false, + "reason": "Unrelated import error vs Whisper pipeline regression." + }, + { + "left": "issue:44164", + "right": "issue:44291", + "accept": false, + "reason": "Different save/load initialization failures with different triggers." + }, + { + "left": "issue:44402", + "right": "issue:44488", + "accept": false, + "reason": "Different models and different tokenizer/vocab symptoms; not a duplicate pair." + }, + { + "left": "issue:41720", + "right": "issue:44155", + "accept": false, + "reason": "Different domains: Qwen3 device-mapping crash vs AudioFlamingo3 batched inference leak." + }, + { + "left": "issue:43873", + "right": "issue:44387", + "accept": false, + "reason": "Both are quantization/memory issues, but they describe different regressions and code paths." + }, + { + "left": "issue:39692", + "right": "issue:44960", + "accept": false, + "reason": "SigLIP2 docs/inference issue is unrelated to GLM5." + }, + { + "left": "issue:43425", + "right": "issue:43957", + "accept": false, + "reason": "Torch version incompatibility is unrelated to meta-device loading failures." + }, + { + "left": "issue:44387", + "right": "issue:44756", + "accept": false, + "reason": "Different OOM scenarios with different proposed fixes; not the same defect." + }, + { + "left": "issue:43854", + "right": "issue:45216", + "accept": false, + "reason": "Different models and different regressions: load failure in tests vs incorrect save_pretrained checkpoint." + }, + { + "left": "issue:42947", + "right": "issue:43381", + "accept": false, + "reason": "Both mention gradient checkpointing, but one is inefficacy under LoRA and the other is eval-mode prohibition." + }, + { + "left": "issue:44292", + "right": "issue:44912", + "accept": false, + "reason": "Different quantization formats and different loading failures; not mergeable as one bug." + }, + { + "left": "issue:44291", + "right": "issue:44756", + "accept": false, + "reason": "Unexpected `_is_hf_initialized` argument is unrelated to mmap/OOM on Strix Halo." + }, + { + "left": "issue:44743", + "right": "issue:44960", + "accept": false, + "reason": "Recurrent-state cache bug vs GLM5 issue; unrelated." + }, + { + "left": "issue:45276", + "right": "issue:45292", + "accept": true, + "reason": "Same `resize_token_embeddings` propagation bug, with `output_embeddings` called out in both reports." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative on the soft edges. The two accepted duplicates around `resize_token_embeddings` are well-supported by the issue titles and similarity, and the other pairwise rejections do not overclaim shared root cause. No mergeability problems for PR pairs here." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 6, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:32090", + "issue:36032", + "issue:39404", + "issue:40990", + "issue:42947", + "issue:43257", + "issue:43299", + "issue:43421", + "issue:43425", + "issue:43452", + "issue:43582", + "issue:43596", + "issue:43611", + "issue:43632", + "issue:43644", + "issue:43645", + "issue:43716", + "issue:43828", + "issue:43856", + "issue:43883", + "issue:43901", + "issue:43906", + "issue:43931", + "issue:43957", + "issue:43994", + "issue:44079", + "issue:44164", + "issue:44265", + "issue:44402", + "issue:44451", + "issue:44479", + "issue:44488", + "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:44991", + "issue:45081", + "issue:45084", + "issue:45161" + ], + "soft_pairs": [ + "issue:44479|issue:44877", + "issue:44402|issue:44451", + "issue:43645|issue:44164", + "issue:43257|issue:44863", + "issue:43425|issue:43828", + "issue:43828|issue:43957", + "issue:44589|issue:44756", + "issue:43931|issue:44863", + "issue:43906|issue:44530", + "issue:43611|issue:43645", + "issue:39404|issue:44912", + "issue:43425|issue:43716", + "issue:44530|issue:45084", + "issue:43901|issue:44509", + "issue:43299|issue:43931", + "issue:44265|issue:44898", + "issue:44479|issue:44805", + "issue:43994|issue:44079", + "issue:44671|issue:44991", + "issue:43856|issue:44756", + "issue:43425|issue:44918", + "issue:39404|issue:44671", + "issue:40990|issue:42947", + "issue:43596|issue:44683", + "issue:44589|issue:45084", + "issue:36032|issue:43452", + "issue:44991|issue:45081", + "issue:43582|issue:43632", + "issue:32090|issue:39404", + "issue:43957|issue:44164", + "issue:40990|issue:43421", + "issue:43883|issue:44488", + "issue:44164|issue:44479", + "issue:44488|issue:45081", + "issue:43425|issue:45161", + "issue:43611|issue:43644" + ], + "prepared_review_unit_hash": "5b5edd9395ff00742810924d51557e560100c9789e579dbe13df3b617f3446d0", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 36, + "serialized_chars": 38484, + "estimated_input_tokens": 9621, + "estimated_eval_tokens": 19498 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose grab-bag of Transformers v5 regressions, mostly around model loading, tokenizers, and a few unrelated runtime/docs bugs. The most representative issue is the broad v5 model-loading regression around `base_model_prefix`.", + "confidence": 0.61, + "canonical_issue_reason": "Issue 43611 is the broadest, most representative v5 regression in the set: it directly summarizes a general model-loading breakage that overlaps the cluster\u2019s dominant theme.", + "canonical_pr_reason": null, + "best_issue_reason": "43611 is the best cluster anchor because its scope is general enough to represent the recurring Transformers v5 loading failures without being tied to one specific model family.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44479", + "right": "issue:44877", + "accept": false, + "reason": "Both are v5 regressions, but one is a video-input bug for Qwen VL models and the other is a config-loading issue for granite_speech; different code paths." + }, + { + "left": "issue:44402", + "right": "issue:44451", + "accept": false, + "reason": "Both involve tokenizer/model loading failures, but the symptoms and likely causes differ: vocab-size mismatch vs inability to load a specific tokenizer." + }, + { + "left": "issue:43645", + "right": "issue:44164", + "accept": false, + "reason": "Custom-model notebook initialization and save/from_pretrained extra_state handling are different loading/saving problems." + }, + { + "left": "issue:43257", + "right": "issue:44863", + "accept": false, + "reason": "Different model families and failure modes: Qwen3 MoE conversion with accelerate/deepspeed vs NemotronH checkpoint loading." + }, + { + "left": "issue:43425", + "right": "issue:43828", + "accept": false, + "reason": "Torch-version incompatibility is not the same as an autocast dtype mismatch in a specific model." + }, + { + "left": "issue:43828", + "right": "issue:43957", + "accept": false, + "reason": "One is an autocast dtype mismatch; the other is meta-device loading breakage. Related only at a high level." + }, + { + "left": "issue:44589", + "right": "issue:44756", + "accept": false, + "reason": "Storage lookup failure and mmap/OOM behavior are unrelated bugs." + }, + { + "left": "issue:43931", + "right": "issue:44863", + "accept": false, + "reason": "Both are model-loading errors, but they concern different architectures and likely different conversion/shape issues." + }, + { + "left": "issue:43906", + "right": "issue:44530", + "accept": false, + "reason": "A reproduction of an older issue and a PagedAttentionCache crash are not the same underlying bug." + }, + { + "left": "issue:43611", + "right": "issue:43645", + "accept": false, + "reason": "Both are v5 regressions, but one is about `base_model_prefix` loading and the other about notebook-defined custom models." + }, + { + "left": "issue:39404", + "right": "issue:44912", + "accept": false, + "reason": "Whisper pipeline language return and GPT-OSS MXFP4 quantization fallback are unrelated." + }, + { + "left": "issue:43425", + "right": "issue:43716", + "accept": false, + "reason": "Torch compatibility and image-preprocessor/model dtype mismatch are distinct issues." + }, + { + "left": "issue:44530", + "right": "issue:45084", + "accept": false, + "reason": "PagedAttentionCache group-type handling and template-node compilation errors are different failures." + }, + { + "left": "issue:43901", + "right": "issue:44509", + "accept": false, + "reason": "Both are docs-related, but they target different APIs and removals." + }, + { + "left": "issue:43299", + "right": "issue:43931", + "accept": false, + "reason": "Both mention Qwen3-VL loading, but one is about MoE weight conversion and the other about weight-shape mismatch; not enough to treat as the same bug." + }, + { + "left": "issue:44265", + "right": "issue:44898", + "accept": false, + "reason": "Torch export with `torch_compilable_check` and Perceiver non-default resolution loading are different code paths." + }, + { + "left": "issue:44479", + "right": "issue:44805", + "accept": false, + "reason": "A video-input regression and a mask shape IndexError are unrelated." + }, + { + "left": "issue:43994", + "right": "issue:44079", + "accept": false, + "reason": "Incorrect SigLIP2 outputs and `ModelOutput` key assignment are different problems." + }, + { + "left": "issue:44671", + "right": "issue:44991", + "accept": false, + "reason": "CamemBERT masked-LM prediction errors and tokenizer loading failures for a different model are not the same bug." + }, + { + "left": "issue:43856", + "right": "issue:44756", + "accept": false, + "reason": "Memory inefficiency in Qwen3 MoE training is unrelated to Strix Halo mmap OOM avoidance." + }, + { + "left": "issue:43425", + "right": "issue:44918", + "accept": false, + "reason": "Torch version incompatibility and TRL SFT embedding unpacking are different failures." + }, + { + "left": "issue:39404", + "right": "issue:44671", + "accept": false, + "reason": "Whisper return_language and CamemBERT masked-LM regressions are unrelated model-specific bugs." + }, + { + "left": "issue:40990", + "right": "issue:42947", + "accept": false, + "reason": "High perplexity on GPT-OSS and ineffective gradient checkpointing with LoRA are not the same issue." + }, + { + "left": "issue:43596", + "right": "issue:44683", + "accept": false, + "reason": "DeepSpeed ZeRO-3 initialization IndexError and compiled flex_attention on torch>=2.9 are unrelated." + }, + { + "left": "issue:43883", + "right": "issue:44488", + "accept": false, + "reason": "Missing `all_tied_weights_keys` on Molmo and loading `cjvt/sleng-bert` are different model-loading problems." + }, + { + "left": "issue:44488", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer load failure for `sleng-bert` and the Mistral regex patch crash are distinct tokenizer bugs." + }, + { + "left": "issue:36032", + "right": "issue:43452", + "accept": false, + "reason": "T5 tokenizer special-token naming conflict and `gguf_file` loading breakage are unrelated." + }, + { + "left": "issue:44991", + "right": "issue:45081", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different tokenizers and likely different fix paths." + }, + { + "left": "issue:43582", + "right": "issue:43632", + "accept": false, + "reason": "Apple Silicon allocator warmup TypeError and `_is_hf_initialized` breakage are different runtime issues." + }, + { + "left": "issue:32090", + "right": "issue:39404", + "accept": false, + "reason": "Trainer GPU broadcast TypeError and Whisper pipeline language return are unrelated." + }, + { + "left": "issue:43957", + "right": "issue:44164", + "accept": false, + "reason": "Meta-device loading issues and extra_state save/load handling are distinct code paths." + }, + { + "left": "issue:40990", + "right": "issue:43421", + "accept": false, + "reason": "Perplexity anomaly and runtime special-token post-processor updates are unrelated." + }, + { + "left": "issue:43883", + "right": "issue:44488", + "accept": false, + "reason": "These are separate tokenizer/model-loading failures with no clear shared root cause." + }, + { + "left": "issue:44589", + "right": "issue:45084", + "accept": false, + "reason": "Float8 storage lookup and template-node compilation errors are unrelated." + }, + { + "left": "issue:43611", + "right": "issue:43644", + "accept": false, + "reason": "Both are broad v5 regressions, but one is about `base_model_prefix` and the other about non-persistent buffer initialization; not the same bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 7, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:34567", + "issue:36032", + "issue:39404", + "issue:40444", + "issue:41553", + "issue:42915", + "issue:42947", + "issue:43262", + "issue:43317", + "issue:43322", + "issue:43377", + "issue:43404", + "issue:43452", + "issue:43482", + "issue:43540", + "issue:43541", + "issue:43582", + "issue:43643", + "issue:43644", + "issue:43645", + "issue:43646", + "issue:43716", + "issue:43819", + "issue:43825", + "issue:43883", + "issue:43994", + "issue:44206", + "issue:44220", + "issue:44403", + "issue:44451", + "issue:44466", + "issue:44488", + "issue:44530", + "issue:44534", + "issue:44610", + "issue:44683", + "issue:44756", + "issue:44811", + "issue:44849", + "issue:44987", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45216", + "issue:45325", + "issue:45357", + "issue:45507", + "issue:45584" + ], + "soft_pairs": [ + "issue:44811|issue:45584", + "issue:44206|issue:44220", + "issue:44488|issue:44987", + "issue:43716|issue:45084", + "issue:34567|issue:40444", + "issue:43883|issue:44991", + "issue:42947|issue:44756", + "issue:43644|issue:43646", + "issue:43643|issue:44530", + "issue:44451|issue:44987", + "issue:43452|issue:43482", + "issue:43317|issue:43482", + "issue:36032|issue:42915", + "issue:43541|issue:45084", + "issue:39404|issue:43825", + "issue:44849|issue:45325", + "issue:42915|issue:44530", + "issue:43404|issue:44466", + "issue:43262|issue:43377", + "issue:36032|issue:43645", + "issue:44987|issue:44991", + "issue:45084|issue:45507", + "issue:43646|issue:44534", + "issue:43994|issue:44610", + "issue:43645|issue:44534", + "issue:45216|issue:45357", + "issue:43322|issue:43540", + "issue:43377|issue:43819", + "issue:44403|issue:45081", + "issue:44451|issue:45081", + "issue:41553|issue:43582", + "issue:44220|issue:44683" + ], + "prepared_review_unit_hash": "a5b18d7e3de6320ee75afcbb76ef1a71fb1ecb2dd09f23756acea42d923f91e2", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37198, + "estimated_input_tokens": 9300, + "estimated_eval_tokens": 18856 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly heterogeneous, with many unrelated Transformer bugs. The only clear duplicate-like pair is the Qwen3.5 save_pretrained regression (#45216 / #45357); the rest are too model- or symptom-specific to merge together.", + "confidence": 0.82, + "canonical_issue_reason": "No single canonical issue across the full set: most items describe unrelated bugs. If a duplicate anchor is needed, #45216 is the best one because #45357 is a later, narrower follow-up of the same Qwen3.5 save_pretrained regression.", + "canonical_pr_reason": null, + "best_issue_reason": "#45216 is the best representative issue for the only strong duplicate pair, since it is earlier and describes the broader regression; #45357 is a narrower persistence of the same problem.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44811", + "right": "issue:45584", + "accept": false, + "reason": "Both are Whisper-related, but one is batch_decode skip-special-token behavior and the other is empty-transcription generation after align_special_tokens; different failure modes." + }, + { + "left": "issue:44206", + "right": "issue:44220", + "accept": false, + "reason": "Both touch audio feature extraction, but one is an unsupported `center` arg in `LasrFeatureExtractor` and the other is a low-level `_torch_extract_fbank_features()` issue; not the same bug." + }, + { + "left": "issue:44488", + "right": "issue:44987", + "accept": false, + "reason": "Both are loading regressions, but they involve different models and likely different code paths (`cjvt/sleng-bert` vs `physical-intelligence/fast`)." + }, + { + "left": "issue:43716", + "right": "issue:45084", + "accept": false, + "reason": "Completely different problems: Mistral-3 image preprocessor dtype mismatch vs a template compilation TypeError." + }, + { + "left": "issue:34567", + "right": "issue:40444", + "accept": false, + "reason": "Unrelated: Trainer token accounting vs Qwen2.5-VL iterable dataset multi-image loading failure." + }, + { + "left": "issue:43883", + "right": "issue:44991", + "accept": false, + "reason": "Both are loading-related, but one is a missing `all_tied_weights_keys` attribute and the other is a tokenizer load failure for a different model; not the same underlying bug." + }, + { + "left": "issue:42947", + "right": "issue:44756", + "accept": false, + "reason": "Different subsystems and symptoms: LoRA gradient checkpointing vs Apple Silicon mmap/OOM handling." + }, + { + "left": "issue:43644", + "right": "issue:43646", + "accept": false, + "reason": "Both are Transformers 5 breakages, but one is non-persistent buffer initialization and the other is custom model initialization; too broad and not the same defect." + }, + { + "left": "issue:43643", + "right": "issue:44530", + "accept": false, + "reason": "Different issues: missing fields from `AutoConfig.from_pretrained(trust_remote_code=True)` vs PagedAttentionCache group-type handling on Qwen3.5." + }, + { + "left": "issue:44451", + "right": "issue:44987", + "accept": false, + "reason": "Both fail on model loading, but they are different models with no evidence of a shared code-path bug." + }, + { + "left": "issue:43452", + "right": "issue:43482", + "accept": false, + "reason": "Both mention GGUF, but one is a broad `gguf_file` API break and the other is a model-specific v5 loading failure; not enough evidence they are the same bug." + }, + { + "left": "issue:43317", + "right": "issue:43482", + "accept": false, + "reason": "Different loading paths: `device_map=auto` offload/dequantization vs GGUF model loading." + }, + { + "left": "issue:36032", + "right": "issue:42915", + "accept": false, + "reason": "Unrelated: T5 tokenizer method conflict vs Qwen3Moe FP8 configuration failure." + }, + { + "left": "issue:43541", + "right": "issue:45084", + "accept": false, + "reason": "Different compilation failures with no shared code path." + }, + { + "left": "issue:39404", + "right": "issue:43825", + "accept": false, + "reason": "Both involve pipeline/Whisper wording, but one is a functional `return_language` regression and the other is an error-message issue about unsupported translation tasks." + }, + { + "left": "issue:44849", + "right": "issue:45325", + "accept": false, + "reason": "Different Qwen variants and different bugs: hidden-states behavior vs image temporal position-id scaling." + }, + { + "left": "issue:42915", + "right": "issue:44530", + "accept": false, + "reason": "Both mention Qwen, but one is FP8 config failure and the other is an invalid cache group type; not the same concrete bug." + }, + { + "left": "issue:43404", + "right": "issue:44466", + "accept": false, + "reason": "Related to tied weights, but one is a missing tie in `Mistral3ForConditionalGeneration` and the other is inconsistent serialization by device; separate issues." + }, + { + "left": "issue:43262", + "right": "issue:43377", + "accept": false, + "reason": "Both are audio-processing issues, but one is chat-template sampling-rate defaults and the other is missing padding-mask support in MIMI batching." + }, + { + "left": "issue:36032", + "right": "issue:43645", + "accept": false, + "reason": "Different Transformers 5 regressions: tokenizer method name collision vs custom model initialization in notebooks." + }, + { + "left": "issue:44987", + "right": "issue:44991", + "accept": false, + "reason": "Both are v5 loading regressions, but they concern different models and different failure modes; no shared underlying bug is evident." + }, + { + "left": "issue:45084", + "right": "issue:45507", + "accept": false, + "reason": "Different template/model-method failures with no concrete overlap." + }, + { + "left": "issue:43646", + "right": "issue:44534", + "accept": false, + "reason": "Both mention custom model/init behavior, but one is generic custom model initialization and the other is non-persistent buffer serialization junk." + }, + { + "left": "issue:43994", + "right": "issue:44610", + "accept": false, + "reason": "Both are vision-related, but one is bad predictions from SigLIP2 and the other is a processor size mismatch for OmDet-Turbo." + }, + { + "left": "issue:43645", + "right": "issue:44534", + "accept": false, + "reason": "Different regressions: notebook custom-model initialization vs junk-filled non-persistent buffers." + }, + { + "left": "issue:45216", + "right": "issue:45357", + "accept": true, + "reason": "Same underlying Qwen3.5 `save_pretrained` regression; #45357 is a later, narrower report that still saves incorrect visual encoder keys." + }, + { + "left": "issue:43322", + "right": "issue:43540", + "accept": false, + "reason": "Both affect multimodal loading/processing, but one is a Llava Next segfault and the other is a Qwen3OmniMoe video-input ValueError." + }, + { + "left": "issue:43377", + "right": "issue:43819", + "accept": false, + "reason": "Different audio-model bugs: missing padding-mask support in MIMI encoder vs DAC latent/forward mismatch without STE." + }, + { + "left": "issue:44403", + "right": "issue:45081", + "accept": false, + "reason": "One is generic noisy model-loading output, the other is a specific Mistral tokenizer regex patch crash." + }, + { + "left": "issue:44451", + "right": "issue:45081", + "accept": false, + "reason": "Both are tokenizer/model-loading failures, but for different models and different root causes." + }, + { + "left": "issue:41553", + "right": "issue:43582", + "accept": false, + "reason": "Unrelated: bad AutoTokenizer error message for Voxtral vs Apple Silicon caching_allocator_warmup TypeError." + }, + { + "left": "issue:44220", + "right": "issue:44683", + "accept": false, + "reason": "Different low-level failures: audio fbank extraction vs compiled flex_attention on newer torch." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays appropriately conservative. The only accepted duplicate-like pair, #45216/#45357, is plausibly the same Qwen3.5 save_pretrained regression and looks mergeable into one bugfix PR. The other soft-edge pairs are rejected with cautious, title-based reasoning rather than overclaiming evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 8, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:33666", + "issue:35707", + "issue:36032", + "issue:40444", + "issue:40990", + "issue:41720", + "issue:42175", + "issue:42915", + "issue:43296", + "issue:43317", + "issue:43404", + "issue:43421", + "issue:43425", + "issue:43482", + "issue:43531", + "issue:43611", + "issue:43632", + "issue:43742", + "issue:43792", + "issue:43818", + "issue:43856", + "issue:44062", + "issue:44079", + "issue:44265", + "issue:44291", + "issue:44403", + "issue:44479", + "issue:44534", + "issue:44568", + "issue:44589", + "issue:44610", + "issue:44683", + "issue:44756", + "issue:44811", + "issue:44829", + "issue:44877", + "issue:44987", + "issue:44991", + "issue:45020", + "issue:45081", + "issue:45084", + "issue:45127", + "issue:45325", + "issue:45357", + "issue:45479", + "issue:45538", + "issue:45542" + ], + "soft_pairs": [ + "issue:43632|issue:44756", + "issue:44265|issue:44610", + "issue:43404|issue:45127", + "issue:42915|issue:45084", + "issue:44987|issue:45081", + "issue:44479|issue:45325", + "issue:43296|issue:43531", + "issue:44062|issue:44987", + "issue:42915|issue:44291", + "issue:40990|issue:44811", + "issue:43611|issue:44534", + "issue:43317|issue:43856", + "issue:42175|issue:45542", + "issue:36032|issue:43742", + "issue:40444|issue:45325", + "issue:43632|issue:44534", + "issue:43421|issue:44568", + "issue:43818|issue:45357", + "issue:43531|issue:44877", + "issue:42915|issue:44589", + "issue:36032|issue:41720", + "issue:43425|issue:44403", + "issue:44683|issue:45020", + "issue:43482|issue:45538", + "issue:44829|issue:45479", + "issue:33666|issue:40444", + "issue:44403|issue:44991", + "issue:43792|issue:44589", + "issue:44079|issue:44534", + "issue:35707|issue:40444" + ], + "prepared_review_unit_hash": "a28d5f7018859e92f8f51d63cf5adfe56345c1daa2ed96da09f5699a5c13ee10", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36162, + "estimated_input_tokens": 9041, + "estimated_eval_tokens": 18338 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: the soft-pair candidates mostly share only broad subsystem or version context, not the same concrete bug or fix path. I reject all soft edges and see no clear canonical duplicate representative.", + "confidence": 0.97, + "canonical_issue_reason": "No single issue clearly represents the cluster; the items span unrelated bugs across tokenizers, Qwen/VL, packaging, loading, and attention behavior.", + "canonical_pr_reason": null, + "best_issue_reason": "No issue is a strong global representative. The cluster does not appear to center on one underlying defect, so a canonical issue choice would be arbitrary.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43632", + "right": "issue:44756", + "accept": false, + "reason": "Unrelated: HF init flag regression vs mmap/OOM on Strix Halo." + }, + { + "left": "issue:44265", + "right": "issue:44610", + "accept": false, + "reason": "Different bugs: torch export/check utility failure vs OmDet-Turbo processor size mismatch." + }, + { + "left": "issue:43404", + "right": "issue:45127", + "accept": false, + "reason": "Distinct causes: lm_head weight tying bug vs LoRA merge with extended vocabulary collapse." + }, + { + "left": "issue:42915", + "right": "issue:45084", + "accept": false, + "reason": "FP8 config failure and template-node compilation error are unrelated code paths." + }, + { + "left": "issue:44987", + "right": "issue:45081", + "accept": false, + "reason": "Both are loading-related, but one is remote-code/model loading and the other is a Mistral regex patch crash." + }, + { + "left": "issue:44479", + "right": "issue:45325", + "accept": false, + "reason": "Same model family, but video-input regression and still-image rope-index scaling are different symptoms and likely different fixes." + }, + { + "left": "issue:43296", + "right": "issue:43531", + "accept": false, + "reason": "Different model families and failure modes: PaddleOCR-VL loading vs Qwen3-MoE sliding_window behavior." + }, + { + "left": "issue:44062", + "right": "issue:44987", + "accept": false, + "reason": "Tokenizer AddedToken argument conflict is unrelated to loading a remote-code model." + }, + { + "left": "issue:42915", + "right": "issue:44291", + "accept": false, + "reason": "Both touch transformers v5-era loading, but FP8 config failure and _is_hf_initialized/init_empty_weights mismatch are different bugs." + }, + { + "left": "issue:40990", + "right": "issue:44811", + "accept": false, + "reason": "Perplexity/eval issue vs Whisper batch_decode special-token handling; no shared bug." + }, + { + "left": "issue:43611", + "right": "issue:44534", + "accept": false, + "reason": "base_model_prefix loading breakage is separate from non-persistent buffer initialization junk." + }, + { + "left": "issue:43317", + "right": "issue:43856", + "accept": false, + "reason": "Loading with gpu+cpu offload is unrelated to training-time memory inefficiency." + }, + { + "left": "issue:42175", + "right": "issue:45542", + "accept": false, + "reason": "Both are TensorFlow packaging/backend issues, but one is missing TF in the torch extra and the other is an undefined backend error with tensorboard-only install." + }, + { + "left": "issue:36032", + "right": "issue:43742", + "accept": false, + "reason": "Tokenizer special-token conflict vs model key error; different underlying failures." + }, + { + "left": "issue:40444", + "right": "issue:45325", + "accept": false, + "reason": "Both involve Qwen VL, but iterable-dataset multi-image finetuning and rope-index scaling for still images are not the same bug." + }, + { + "left": "issue:43632", + "right": "issue:44534", + "accept": false, + "reason": "Unrelated: _is_hf_initialized flag regression vs non-persistent buffer junk in v5." + }, + { + "left": "issue:43421", + "right": "issue:44568", + "accept": false, + "reason": "Shared special-token theme, but runtime post-processor updates and add_special_tokens=True regression are different surfaces and not clearly one fix." + }, + { + "left": "issue:43818", + "right": "issue:45357", + "accept": false, + "reason": "Video-LLaVA tower/temporal-attention issue is unrelated to Qwen3.5 save_pretrained visual-key regression." + }, + { + "left": "issue:43531", + "right": "issue:44877", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and granite_speech strict config loading are unrelated." + }, + { + "left": "issue:42915", + "right": "issue:44589", + "accept": false, + "reason": "FineGrainedFP8Config failure is unrelated to missing Float8 storage type lookup." + }, + { + "left": "issue:36032", + "right": "issue:41720", + "accept": false, + "reason": "Tokenizer load conflict vs CUDA assert during auto device mapping are different problem classes." + }, + { + "left": "issue:43425", + "right": "issue:44403", + "accept": false, + "reason": "Torch compatibility issue is unrelated to noisy transformer loading logs." + }, + { + "left": "issue:44683", + "right": "issue:45020", + "accept": false, + "reason": "Compiled flex_attention on new torch versions is unrelated to remote_code model breakage." + }, + { + "left": "issue:43482", + "right": "issue:45538", + "accept": false, + "reason": "GGUF loading with transformers v5 and CLIPTokenizer model_max_length are different defects." + }, + { + "left": "issue:44829", + "right": "issue:45479", + "accept": false, + "reason": "Both affect training loss, but flash_attention_3 degeneration and num_labels=1 zero-loss behavior have different root causes." + }, + { + "left": "issue:33666", + "right": "issue:40444", + "accept": false, + "reason": "Both mention Qwen VL training, but multi-GPU training and multi-image iterable dataset failures are not the same bug." + }, + { + "left": "issue:44403", + "right": "issue:44991", + "accept": false, + "reason": "Generic loading noise is unrelated to tokenizer loading failure for EMBEDDIA/est-roberta." + }, + { + "left": "issue:43792", + "right": "issue:44589", + "accept": false, + "reason": "Whisper run failure and Float8 storage lookup error are unrelated." + }, + { + "left": "issue:44079", + "right": "issue:44534", + "accept": false, + "reason": "ModelOutput key assignment and non-persistent buffer junk are separate internal behaviors." + }, + { + "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 finetuning." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 9, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:34634", + "issue:35707", + "issue:40444", + "issue:40990", + "issue:41720", + "issue:42175", + "issue:42915", + "issue:43262", + "issue:43296", + "issue:43317", + "issue:43377", + "issue:43454", + "issue:43482", + "issue:43525", + "issue:43526", + "issue:43531", + "issue:43611", + "issue:43632", + "issue:43644", + "issue:43646", + "issue:43701", + "issue:43746", + "issue:43749", + "issue:43792", + "issue:43828", + "issue:43873", + "issue:43881", + "issue:43883", + "issue:43957", + "issue:43994", + "issue:44079", + "issue:44265", + "issue:44291", + "issue:44387", + "issue:44466", + "issue:44488", + "issue:44589", + "issue:44617", + "issue:44683", + "issue:44898", + "issue:44991", + "issue:45072", + "issue:45081", + "issue:45325", + "issue:45381", + "issue:45491", + "issue:45563", + "issue:45584" + ], + "soft_pairs": [ + "issue:43377|issue:45491", + "issue:43482|issue:43525", + "issue:44291|issue:44589", + "issue:43749|issue:44079", + "issue:44079|issue:45072", + "issue:43644|issue:44079", + "issue:40444|issue:45381", + "issue:43632|issue:45072", + "issue:44265|issue:45072", + "issue:43881|issue:44488", + "issue:45325|issue:45381", + "issue:43531|issue:44291", + "issue:43611|issue:43646", + "issue:41720|issue:45081", + "issue:42175|issue:43296", + "issue:43454|issue:44466", + "issue:43994|issue:44898", + "issue:44387|issue:44617", + "issue:43828|issue:44683", + "issue:42915|issue:43482", + "issue:43526|issue:43883", + "issue:43792|issue:43957", + "issue:43701|issue:45325", + "issue:43262|issue:43746", + "issue:43632|issue:45563", + "issue:43881|issue:44991", + "issue:43317|issue:43873", + "issue:40990|issue:45584", + "issue:34634|issue:35707" + ], + "prepared_review_unit_hash": "8036223205b231423116aa0b4e79a355b114956e81d139dea9a4dd8db108e853", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36388, + "estimated_input_tokens": 9097, + "estimated_eval_tokens": 18450 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a heterogeneous set of mostly unrelated issues, with only a few superficial overlaps around Transformers v5 regressions, quantization/offloading, and a couple of model-family-specific bugs. None of the soft edges look like true duplicates.", + "confidence": 0.91, + "canonical_issue_reason": "No single canonical issue is safe here: the items span many distinct bugs and model families, so picking one representative would be misleading.", + "canonical_pr_reason": "No PRs are present in this cluster.", + "best_issue_reason": "There is no strong global best issue; the closest broad items are still too narrow or too different from the rest of the set to serve as a reliable representative.", + "best_pr_reason": "No PRs are present in this cluster.", + "soft_edge_verdicts": [ + { + "left": "issue:43377", + "right": "issue:45491", + "accept": false, + "reason": "Different models and failure modes: batched-vs-single padding-mask mismatch vs mixed-length sliding-window NaNs." + }, + { + "left": "issue:43482", + "right": "issue:43525", + "accept": false, + "reason": "Both are load failures, but one is a Qwen2.5-GGUF/v5 loading issue and the other is a Llama4Config missing `pad_token_id`." + }, + { + "left": "issue:44291", + "right": "issue:44589", + "accept": false, + "reason": "Different code paths: unexpected `_is_hf_initialized` argument vs missing Float8 storage object." + }, + { + "left": "issue:43749", + "right": "issue:44079", + "accept": false, + "reason": "FSDP CPU RAM efficient loading is unrelated to `ModelOutput` key assignment behavior." + }, + { + "left": "issue:44079", + "right": "issue:45072", + "accept": false, + "reason": "`ModelOutput` key handling and bf16 dtype mismatches are unrelated bugs." + }, + { + "left": "issue:43644", + "right": "issue:44079", + "accept": false, + "reason": "Non-persistent buffer serialization/junk is not the same bug as `ModelOutput` key assignment." + }, + { + "left": "issue:40444", + "right": "issue:45381", + "accept": false, + "reason": "Different modalities and symptoms: multi-image finetuning failure vs Qwen2.5-VL video position-id bug." + }, + { + "left": "issue:43632", + "right": "issue:45072", + "accept": false, + "reason": "`_is_hf_initialized` regression and bfloat16 inference dtype mismatch are separate issues." + }, + { + "left": "issue:44265", + "right": "issue:45072", + "accept": false, + "reason": "`torch.export`/`torch_compilable_check` failure is not the same as a bfloat16 dtype mismatch." + }, + { + "left": "issue:43881", + "right": "issue:44488", + "accept": false, + "reason": "Different models and likely different loader failures; no shared concrete bug." + }, + { + "left": "issue:45325", + "right": "issue:45381", + "accept": false, + "reason": "Same model family, but still-image temporal scaling and video vision-position-id errors are different code paths." + }, + { + "left": "issue:43531", + "right": "issue:44291", + "accept": false, + "reason": "Qwen3-MoE sliding-window behavior is unrelated to the `init_empty_weights` argument regression." + }, + { + "left": "issue:43611", + "right": "issue:43646", + "accept": false, + "reason": "Both are Transformers v5 regressions, but one is about `base_model_prefix` loading and the other about custom model initialization; not the same concrete bug." + }, + { + "left": "issue:41720", + "right": "issue:45081", + "accept": false, + "reason": "Qwen3 auto device-map CUDA assert and Mistral tokenizer regex crash are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:43296", + "accept": false, + "reason": "Packaging/backend selection issue is unrelated to PaddleOCR-VL loading failure." + }, + { + "left": "issue:43454", + "right": "issue:44466", + "accept": false, + "reason": "Both mention tied `lm_head` weights, but one is a generation-quality/init issue and the other is device-dependent serialization; not the same fix." + }, + { + "left": "issue:43994", + "right": "issue:44898", + "accept": false, + "reason": "Different vision-model failures with different symptoms and likely different root causes." + }, + { + "left": "issue:44387", + "right": "issue:44617", + "accept": false, + "reason": "Both mention OOM, but one is an int4 quantization memory regression and the other is a model-specific video OOM." + }, + { + "left": "issue:43828", + "right": "issue:44683", + "accept": false, + "reason": "Autocast dtype mismatch and compiled flex_attention failure on torch >= 2.9 are different problems." + }, + { + "left": "issue:42915", + "right": "issue:43482", + "accept": false, + "reason": "Qwen3Moe FP8 config failure and Qwen2.5-GGUF loading failure are unrelated." + }, + { + "left": "issue:43526", + "right": "issue:43883", + "accept": false, + "reason": "Image processor `reduce_labels` bug is unrelated to missing `all_tied_weights_keys` on a model class." + }, + { + "left": "issue:43792", + "right": "issue:43957", + "accept": false, + "reason": "Whisper runtime failure and meta-device model loading regression do not share the same root cause." + }, + { + "left": "issue:43701", + "right": "issue:45325", + "accept": false, + "reason": "Checkpoint resume key mismatch is unrelated to Qwen2.5-VL rope/position-id scaling." + }, + { + "left": "issue:43262", + "right": "issue:43746", + "accept": false, + "reason": "Audio chat-template sampling-rate default and PEFT local-checkpoint loading are different bugs." + }, + { + "left": "issue:43632", + "right": "issue:45563", + "accept": false, + "reason": "An initialization flag regression is unrelated to a stale generate warning." + }, + { + "left": "issue:43881", + "right": "issue:44991", + "accept": false, + "reason": "Different model/tokenizer loading failures with no shared code path." + }, + { + "left": "issue:43317", + "right": "issue:43873", + "accept": false, + "reason": "Both concern offloading/quantization, but one is a dequantized-model load failure with gpu+cpu offload and the other is a broader offloading complaint; not the same concrete bug." + }, + { + "left": "issue:40990", + "right": "issue:45584", + "accept": false, + "reason": "Perplexity/evaluation anomaly and Whisper empty-transcription generation failure are unrelated." + }, + { + "left": "issue:34634", + "right": "issue:35707", + "accept": false, + "reason": "Bark voice_preset support and progressive generation with inputs_embeds/past_key_values are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 10, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:35707", + "issue:36331", + "issue:39692", + "issue:40444", + "issue:41720", + "issue:42617", + "issue:42915", + "issue:43296", + "issue:43316", + "issue:43317", + "issue:43322", + "issue:43377", + "issue:43404", + "issue:43454", + "issue:43526", + "issue:43550", + "issue:43577", + "issue:43606", + "issue:43708", + "issue:43720", + "issue:43749", + "issue:43792", + "issue:43819", + "issue:43873", + "issue:43881", + "issue:44186", + "issue:44265", + "issue:44291", + "issue:44423", + "issue:44466", + "issue:44479", + "issue:44534", + "issue:44610", + "issue:44671", + "issue:44683", + "issue:44743", + "issue:44756", + "issue:44877", + "issue:44898", + "issue:44913", + "issue:44918", + "issue:44977", + "issue:44991", + "issue:45072", + "issue:45325", + "issue:45357", + "issue:45406" + ], + "soft_pairs": [ + "issue:43606|issue:44265", + "issue:44291|issue:44877", + "issue:44186|issue:44265", + "issue:43606|issue:43819", + "issue:44610|issue:45325", + "issue:35707|issue:41720", + "issue:43377|issue:44898", + "issue:43454|issue:43550", + "issue:43377|issue:44610", + "issue:43454|issue:45072", + "issue:39692|issue:40444", + "issue:44466|issue:45357", + "issue:43749|issue:44534", + "issue:43296|issue:44991", + "issue:43792|issue:44683", + "issue:43296|issue:44877", + "issue:41720|issue:43322", + "issue:43454|issue:43720", + "issue:43404|issue:45357", + "issue:41720|issue:44977", + "issue:44423|issue:45406", + "issue:43526|issue:43881", + "issue:44186|issue:45072", + "issue:43873|issue:44756", + "issue:43577|issue:43708", + "issue:42617|issue:42915", + "issue:44877|issue:44991", + "issue:44610|issue:44898", + "issue:43550|issue:43720", + "issue:44479|issue:44918", + "issue:44186|issue:44610", + "issue:36331|issue:42915", + "issue:36331|issue:42617", + "issue:43316|issue:44913", + "issue:44671|issue:44743", + "issue:42915|issue:43317", + "issue:43550|issue:43606", + "issue:41720|issue:42617" + ], + "prepared_review_unit_hash": "7852e5b2837c4dbfd48907421944cee58f7022d1c59336beb3d907bd0f0dae02", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 38, + "serialized_chars": 38640, + "estimated_input_tokens": 9660, + "estimated_eval_tokens": 19576 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: the soft links point to broad thematic overlap around loading/offloading, quantization, serialization, and multimodal preprocessing, but the paired issues do not appear to describe the same underlying bug. I would not merge any of the soft-edge candidates.", + "confidence": 0.76, + "canonical_issue_reason": "issue:43873 is the broadest representative of the recurring loading/offloading theme, but the cluster is not actually a duplicate set and most items have distinct root causes.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43873 is the closest umbrella issue and has the strongest cross-cutting theme among the set, but it still does not subsume the model-specific and API-specific reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43606", + "right": "issue:44265", + "accept": false, + "reason": "CPU-offload device mismatch vs torch.export failure on torch_compilable_check; different bugs." + }, + { + "left": "issue:44291", + "right": "issue:44877", + "accept": false, + "reason": "Unexpected _is_hf_initialized arg during init_empty_weights vs strict config loading; unrelated root causes." + }, + { + "left": "issue:44186", + "right": "issue:44265", + "accept": false, + "reason": "Tokenizer padding/NER crash vs export-time check failure; not the same issue." + }, + { + "left": "issue:43606", + "right": "issue:43819", + "accept": false, + "reason": "CPU offload mismatch vs DAC STE/latents mismatch; different code paths." + }, + { + "left": "issue:44610", + "right": "issue:45325", + "accept": false, + "reason": "Processor/model input-size mismatch vs rope-index temporal scaling regression; different defects." + }, + { + "left": "issue:35707", + "right": "issue:41720", + "accept": false, + "reason": "Progressive generation with inputs_embeds/past_key_values vs auto device-map cuda assert; unrelated." + }, + { + "left": "issue:43377", + "right": "issue:44898", + "accept": false, + "reason": "MIMI padding-mask batching bug vs Perceiver resolution handling bug; different model paths." + }, + { + "left": "issue:43454", + "right": "issue:43550", + "accept": false, + "reason": "lm_head tying/serialization bug vs torch.compile SDPA failure; not the same change." + }, + { + "left": "issue:43377", + "right": "issue:44610", + "accept": false, + "reason": "Padding-mask batching issue vs OmDet image-size mismatch; different preprocessing bugs." + }, + { + "left": "issue:43454", + "right": "issue:45072", + "accept": false, + "reason": "Weight-tying serialization issue vs dtype mismatch in CI; different symptoms and fixes." + }, + { + "left": "issue:39692", + "right": "issue:40444", + "accept": false, + "reason": "Doc example errors vs finetuning failure with multiple images; different reports." + }, + { + "left": "issue:44466", + "right": "issue:45357", + "accept": false, + "reason": "Generic tied-weight serialization inconsistency vs Qwen3.5 visual-key save_pretrained regression; not the same bug." + }, + { + "left": "issue:43749", + "right": "issue:44534", + "accept": false, + "reason": "FSDP RAM-efficient loading failure vs non-persistent buffer junk-initialization; different mechanisms." + }, + { + "left": "issue:43296", + "right": "issue:44991", + "accept": false, + "reason": "PaddleOCR-VL load failure vs tokenizer load regression for est-roberta; unrelated." + }, + { + "left": "issue:43792", + "right": "issue:44683", + "accept": false, + "reason": "Whisper load/run issue vs compiled flex_attention incompatibility; different subsystems." + }, + { + "left": "issue:43296", + "right": "issue:44877", + "accept": false, + "reason": "Model loading failure vs strict config rejection; no shared underlying bug." + }, + { + "left": "issue:41720", + "right": "issue:43322", + "accept": false, + "reason": "Qwen3 device-map cuda assert vs Llava Next segfault; different models and failures." + }, + { + "left": "issue:43454", + "right": "issue:43720", + "accept": false, + "reason": "lm_head tying bug vs packed-weight unpacking during accelerate loading; different code paths." + }, + { + "left": "issue:43404", + "right": "issue:45357", + "accept": false, + "reason": "Mistral3 lm_head tying bug vs Qwen3.5 visual-encoder key serialization bug; not duplicates." + }, + { + "left": "issue:41720", + "right": "issue:44977", + "accept": false, + "reason": "Device-mapping crash vs flash-attention generation regression; different Qwen3.5 issues." + }, + { + "left": "issue:44423", + "right": "issue:45406", + "accept": false, + "reason": "Continuous-batching multimodal crash vs Gemma4Processor missing _tokenizer; similar area, different bug." + }, + { + "left": "issue:43526", + "right": "issue:43881", + "accept": false, + "reason": "BeitImageProcessorFast label reduction bug vs glm-4v loading failure; unrelated." + }, + { + "left": "issue:44186", + "right": "issue:45072", + "accept": false, + "reason": "Tokenizer crash vs dtype mismatch in inference; no common root cause." + }, + { + "left": "issue:43873", + "right": "issue:44756", + "accept": false, + "reason": "Quantization/offloading issue vs mmap OOM workaround; different failure modes." + }, + { + "left": "issue:43577", + "right": "issue:43708", + "accept": false, + "reason": "dtype preservation bug vs resume_from_checkpoint max_steps miscalculation; unrelated." + }, + { + "left": "issue:42617", + "right": "issue:42915", + "accept": false, + "reason": "3d_parallel startup failure vs Qwen3Moe FineGrainedFP8Config failure; different models and errors." + }, + { + "left": "issue:44877", + "right": "issue:44991", + "accept": false, + "reason": "Config-loading strictness vs tokenizer-loading regression; distinct issues." + }, + { + "left": "issue:44610", + "right": "issue:44898", + "accept": false, + "reason": "Processor output size mismatch vs Perceiver non-default resolution failure; not the same bug." + }, + { + "left": "issue:43550", + "right": "issue:43720", + "accept": false, + "reason": "torch.compile SDPA failure vs accelerate packed-weight loading; different code paths." + }, + { + "left": "issue:44479", + "right": "issue:44918", + "accept": false, + "reason": "Video-input regression vs SFT input-embedding unpacking failure; unrelated." + }, + { + "left": "issue:44186", + "right": "issue:44610", + "accept": false, + "reason": "Tokenizer batching crash vs image processor resolution mismatch; different defects." + }, + { + "left": "issue:36331", + "right": "issue:42915", + "accept": false, + "reason": "Trainer compute_loss signature breakage vs Qwen3Moe FP8 failure; no overlap." + }, + { + "left": "issue:36331", + "right": "issue:42617", + "accept": false, + "reason": "CustomTrainer API break vs 3d_parallel startup failure; unrelated." + }, + { + "left": "issue:43316", + "right": "issue:44913", + "accept": false, + "reason": "Gemma3TextConfig API discrepancy vs GPTNeoX rotary_pct not persisting on reload; different config bugs." + }, + { + "left": "issue:44671", + "right": "issue:44743", + "accept": false, + "reason": "CamemBERT masked-LM prediction regression vs qwen3_5 recurrent-state reset; different models and logic." + }, + { + "left": "issue:42915", + "right": "issue:43317", + "accept": false, + "reason": "Qwen3Moe FP8 failure vs device_map=auto dequantized-load offload failure; different bugs." + }, + { + "left": "issue:43550", + "right": "issue:43606", + "accept": false, + "reason": "Bamba torch.compile/SDPA bug vs CPU offload device mismatch; not the same issue." + }, + { + "left": "issue:41720", + "right": "issue:42617", + "accept": false, + "reason": "Qwen3 auto device-map assert vs 3d_parallel startup failure; unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 11, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:30064", + "issue:33357", + "issue:34567", + "issue:34634", + "issue:35707", + "issue:36010", + "issue:36331", + "issue:41720", + "issue:42617", + "issue:42886", + "issue:42915", + "issue:43322", + "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:44062", + "issue:44186", + "issue:44292", + "issue:44361", + "issue:44368", + "issue:44451", + "issue:44479", + "issue:44492", + "issue:44509", + "issue:44514", + "issue:44545", + "issue:44625", + "issue:44843", + "issue:44898", + "issue:44991", + "issue:45059", + "issue:45072", + "issue:45161", + "issue:45357", + "issue:45381", + "issue:45406", + "issue:45464", + "issue:45479" + ], + "soft_pairs": [ + "issue:43531|issue:43572", + "issue:44292|issue:44479", + "issue:41720|issue:45161", + "issue:43540|issue:43782", + "issue:43388|issue:45059", + "issue:44186|issue:44361", + "issue:42915|issue:43322", + "issue:41720|issue:43873", + "issue:36010|issue:42617", + "issue:45406|issue:45464", + "issue:44514|issue:44545", + "issue:36331|issue:41720", + "issue:33357|issue:35707", + "issue:41720|issue:43746", + "issue:30064|issue:34634", + "issue:43606|issue:43720", + "issue:44492|issue:44509", + "issue:43550|issue:43653", + "issue:44368|issue:45357", + "issue:44625|issue:45479", + "issue:43421|issue:43927", + "issue:43606|issue:43653", + "issue:43782|issue:45381", + "issue:34567|issue:35707", + "issue:34634|issue:41720", + "issue:44898|issue:45072", + "issue:42886|issue:44843", + "issue:43421|issue:43653", + "issue:44451|issue:44991", + "issue:44062|issue:44451" + ], + "prepared_review_unit_hash": "23ca7a7f42bc48e7ebde864b67fa218895228b3b0500112b06b5c0abd3128e50", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36737, + "estimated_input_tokens": 9185, + "estimated_eval_tokens": 18626 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Only two soft edges look like true duplicates: the Qwen2_5_VLProcessor batched-padding crash and the HF_HUB_OFFLINE tokenizer-cache failure. The remaining pairs share broad areas but not the same concrete bug/change.", + "confidence": 0.96, + "canonical_issue_reason": "issue:44514 is the cleanest canonical issue for its pair: it is the older report and has an exact title match with 44545, so it best serves as the duplicate anchor.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44843 is the strongest standalone issue overall because it states the offline-mode failure clearly and points to the likely root cause (`_patch_mistral_regex` calling `model_info()` unconditionally), making it the most actionable anchor.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43531", + "right": "issue:43572", + "accept": false, + "reason": "Different models and failure modes: Qwen3-MoE sliding-window behavior vs StableLmConfig pad_token_idx missing after v5 update." + }, + { + "left": "issue:44292", + "right": "issue:44479", + "accept": false, + "reason": "Both involve Qwen models, but one is an NVFP4 loading error and the other is a video-input regression in Qwen VL/MoE paths." + }, + { + "left": "issue:41720", + "right": "issue:45161", + "accept": false, + "reason": "Different concrete problems: auto device mapping CUDA assert on A800 vs GPT-OSS MoE tensor-parallelism failure." + }, + { + "left": "issue:43540", + "right": "issue:43782", + "accept": false, + "reason": "One is a Qwen3OmniMoe video-input processing ValueError; the other is a Qwen3VL from_pretrained weight_only loading error." + }, + { + "left": "issue:43388", + "right": "issue:45059", + "accept": false, + "reason": "Metric gathering label truncation in Trainer/Mask2Former is not the same bug as SAM3 text/bbox weird behavior." + }, + { + "left": "issue:44186", + "right": "issue:44361", + "accept": false, + "reason": "Both are tokenizer crashes, but they affect different models and appear to stem from different code paths and inputs." + }, + { + "left": "issue:42915", + "right": "issue:43322", + "accept": false, + "reason": "Qwen3Moe FineGrainedFP8Config failure and Llava Next segfault are unrelated concrete bugs." + }, + { + "left": "issue:41720", + "right": "issue:43873", + "accept": false, + "reason": "Both relate loosely to loading/offloading, but the reported failures and triggers are different enough to keep separate." + }, + { + "left": "issue:36010", + "right": "issue:42617", + "accept": false, + "reason": "GenerationMixin import error and a 3d_parallel script failure are unrelated." + }, + { + "left": "issue:45406", + "right": "issue:45464", + "accept": false, + "reason": "Gemma4Processor missing _tokenizer in serve is a different issue from Qwen3.5 streaming inference failure." + }, + { + "left": "issue:44514", + "right": "issue:44545", + "accept": true, + "reason": "Exact same title and same crash scenario; 44514 is the older canonical report." + }, + { + "left": "issue:36331", + "right": "issue:41720", + "accept": false, + "reason": "Trainer compute_loss signature mismatch is unrelated to Qwen3 auto device-mapping CUDA asserts." + }, + { + "left": "issue:33357", + "right": "issue:35707", + "accept": false, + "reason": "MacOS bus error with a CLIP model is not the same underlying bug as progressive generation with inputs_embeds/past_key_values." + }, + { + "left": "issue:41720", + "right": "issue:43746", + "accept": false, + "reason": "Different failures: device-mapping CUDA assert vs PEFT-adapter local checkpoint loading for GraniteSpeech." + }, + { + "left": "issue:30064", + "right": "issue:34634", + "accept": false, + "reason": "Image Processor void segmentation maps and BarkProcessor voice_preset are entirely different subsystems and bugs." + }, + { + "left": "issue:43606", + "right": "issue:43720", + "accept": false, + "reason": "Both touch accelerate/offloading/loading, but one is a CPU-offload device mismatch and the other is unpacking packed weights in BitNet." + }, + { + "left": "issue:44492", + "right": "issue:44509", + "accept": false, + "reason": "A cache-strategy typo and docs mentioning removed pipeline tasks are unrelated documentation/config issues." + }, + { + "left": "issue:43550", + "right": "issue:43653", + "accept": false, + "reason": "torch.compile/SDPA failure in Bamba-9B-v2 is unrelated to BigBirdTokenizer special-token registration." + }, + { + "left": "issue:44368", + "right": "issue:45357", + "accept": false, + "reason": "A tie_word_embeddings warning during LoRA fine-tuning is not the same as incorrect visual-encoder keys being saved." + }, + { + "left": "issue:44625", + "right": "issue:45479", + "accept": false, + "reason": "num_labels propagation in Qwen3.5 config is different from the cross-model zero-loss classification bug." + }, + { + "left": "issue:43421", + "right": "issue:43927", + "accept": false, + "reason": "Runtime post-processor updates for special tokens and save/load loss of custom token IDs are related broadly, but not the same concrete bug." + }, + { + "left": "issue:43606", + "right": "issue:43653", + "accept": false, + "reason": "Device mismatch during CPU offload and mask-token registration in a tokenizer are unrelated." + }, + { + "left": "issue:43782", + "right": "issue:45381", + "accept": false, + "reason": "from_pretrained weight_only loading error is distinct from wrong vision_position_ids in Qwen2.5-VL video input." + }, + { + "left": "issue:34567", + "right": "issue:35707", + "accept": false, + "reason": "TrainerState token counting and progressive generation with past_key_values are different issues." + }, + { + "left": "issue:34634", + "right": "issue:41720", + "accept": false, + "reason": "BarkProcessor voice_preset and Qwen3 auto device mapping are unrelated." + }, + { + "left": "issue:44898", + "right": "issue:45072", + "accept": false, + "reason": "Perceiver interpolation at non-default resolution and dtype mismatches in bfloat16 inference do not describe the same bug." + }, + { + "left": "issue:42886", + "right": "issue:44843", + "accept": true, + "reason": "Both describe the same HF_HUB_OFFLINE tokenizer-loading failure; 44843 supplies the likely root cause via an unconditional `model_info()` call." + }, + { + "left": "issue:43421", + "right": "issue:43653", + "accept": false, + "reason": "TokenizersBackend post-processor refresh and BigBirdTokenizer special-token registration are different tokenization bugs." + }, + { + "left": "issue:44451", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer-loading regressions, but for different models and not enough evidence of a shared underlying cause." + }, + { + "left": "issue:44062", + "right": "issue:44451", + "accept": false, + "reason": "AddedToken constructor argument handling is unrelated to ScandiBERT tokenizer loading." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded and conservative overall. The two accepted duplicate pairs are well supported by the packet: 44514/44545 have the same title and scenario, and 42886/44843 describe the same HF_HUB_OFFLINE tokenizer-loading failure with a plausible shared root cause. The remaining soft edges are rejected cautiously without overclaiming similarity." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 12, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:30064", + "issue:36010", + "issue:39692", + "issue:40444", + "issue:43335", + "issue:43381", + "issue:43421", + "issue:43525", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43596", + "issue:43618", + "issue:43653", + "issue:43697", + "issue:43716", + "issue:43720", + "issue:43746", + "issue:43756", + "issue:43819", + "issue:43844", + "issue:43873", + "issue:43994", + "issue:44008", + "issue:44062", + "issue:44077", + "issue:44186", + "issue:44220", + "issue:44488", + "issue:44589", + "issue:44610", + "issue:44625", + "issue:44661", + "issue:44671", + "issue:44811", + "issue:44843", + "issue:44898", + "issue:44918", + "issue:44936", + "issue:44987", + "issue:45042", + "issue:45072", + "issue:45081", + "issue:45183", + "issue:45356", + "issue:45381", + "issue:45479" + ], + "soft_pairs": [ + "issue:43381|issue:43844", + "issue:44918|issue:44936", + "issue:44077|issue:44625", + "issue:44987|issue:45042", + "issue:43531|issue:44589", + "issue:43994|issue:44671", + "issue:39692|issue:44610", + "issue:43421|issue:44186", + "issue:45081|issue:45356", + "issue:43525|issue:43531", + "issue:43596|issue:44220", + "issue:43746|issue:45381", + "issue:45183|issue:45479", + "issue:39692|issue:43873", + "issue:43618|issue:43697", + "issue:43540|issue:44008", + "issue:40444|issue:43746", + "issue:43746|issue:43873", + "issue:44220|issue:44661", + "issue:43716|issue:44008", + "issue:43335|issue:43756", + "issue:43720|issue:43819", + "issue:43653|issue:43720", + "issue:43421|issue:44811", + "issue:30064|issue:36010", + "issue:44843|issue:45081", + "issue:44062|issue:44488", + "issue:44186|issue:44898", + "issue:43540|issue:45381", + "issue:43550|issue:45072" + ], + "prepared_review_unit_hash": "f309753de1960f2da45be8c380ea47234b7a5dcd7928486e16379c9f0cb9bae2", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36405, + "estimated_input_tokens": 9102, + "estimated_eval_tokens": 18460 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous; the soft pairs are mostly broad-topic similarities rather than true duplicates, so I rejected all of them. The closest-looking pair is the Mistral tokenizer regex set, but even there the failures are different enough to keep separate.", + "confidence": 0.95, + "canonical_issue_reason": "issue:45081 is the most concrete, code-path-specific bug in the set (a direct tokenizer loading crash with a clear trigger and traceback), so it is the best representative issue if one must be chosen.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45081 has the strongest actionable reproduction and the most precisely scoped failure mode among these unrelated issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43381", + "right": "issue:43844", + "accept": false, + "reason": "Different subsystems and failure modes: eval-mode gradient checkpointing vs ZeRO-3 gradient explosion during training." + }, + { + "left": "issue:44918", + "right": "issue:44936", + "accept": false, + "reason": "One is an embeddings/SFT trainer failure, the other is evaluate() breaking after train(); not the same bug." + }, + { + "left": "issue:44077", + "right": "issue:44625", + "accept": false, + "reason": "Unrelated model/config problems: patchtsmixer post_init policy vs Qwen3.5 num_labels propagation." + }, + { + "left": "issue:44987", + "right": "issue:45042", + "accept": false, + "reason": "Different bugs and surfaces: model loading regression vs PIL image processor torchvision dependency issue." + }, + { + "left": "issue:43531", + "right": "issue:44589", + "accept": false, + "reason": "Qwen3-MoE sliding_window bug is unrelated to the Float8 storage lookup error." + }, + { + "left": "issue:43994", + "right": "issue:44671", + "accept": false, + "reason": "Both are model-output regressions, but they involve different models and distinct code paths." + }, + { + "left": "issue:39692", + "right": "issue:44610", + "accept": false, + "reason": "Doc/example errors for SigLIP2 do not match an OmDet-Turbo processor/model input-size mismatch." + }, + { + "left": "issue:43421", + "right": "issue:44186", + "accept": false, + "reason": "Tokenizer post-processor update behavior is not the same as LayoutLMv2 NER padding/truncation crashes." + }, + { + "left": "issue:45081", + "right": "issue:45356", + "accept": false, + "reason": "Both mention fix_mistral_regex, but one is a loading crash and the other is a tokenizer regression/warning issue; not the same underlying defect." + }, + { + "left": "issue:43525", + "right": "issue:43531", + "accept": false, + "reason": "Llama4Config pad_token_id absence is unrelated to Qwen3-MoE sliding_window handling." + }, + { + "left": "issue:43596", + "right": "issue:44220", + "accept": false, + "reason": "BertModel/ZeRO-3 initialization failure is unrelated to audio feature extraction." + }, + { + "left": "issue:43746", + "right": "issue:45381", + "accept": false, + "reason": "PEFT checkpoint loading for GraniteSpeech is unrelated to Qwen2.5-VL video vision_position_ids." + }, + { + "left": "issue:45183", + "right": "issue:45479", + "accept": false, + "reason": "Input validation error messages are not the same bug as the degenerate zero-loss classification behavior." + }, + { + "left": "issue:39692", + "right": "issue:43873", + "accept": false, + "reason": "A documentation example bug is not the same as an offloading/quantization runtime bug." + }, + { + "left": "issue:43618", + "right": "issue:43697", + "accept": false, + "reason": "Missing CLIP attentions assignment and RTDetrV2 output differences are different regressions." + }, + { + "left": "issue:43540", + "right": "issue:44008", + "accept": false, + "reason": "Different models and failure types: Qwen3OmniMoe video processing vs Gemma 3n variable collision." + }, + { + "left": "issue:40444", + "right": "issue:43746", + "accept": false, + "reason": "Multi-image Qwen2.5-VL finetuning failure is unrelated to GraniteSpeech PEFT checkpoint loading." + }, + { + "left": "issue:43746", + "right": "issue:43873", + "accept": false, + "reason": "Checkpoint loading with PEFT adapters is unrelated to quantization offloading behavior." + }, + { + "left": "issue:44220", + "right": "issue:44661", + "accept": false, + "reason": "Audio feature extraction and add-new-model-like/tokenizer mapping failure are unrelated." + }, + { + "left": "issue:43716", + "right": "issue:44008", + "accept": false, + "reason": "Mistral-3 image dtype mismatch is unrelated to Gemma 3n audio tensor name collision." + }, + { + "left": "issue:43335", + "right": "issue:43756", + "accept": false, + "reason": "Sparse-layer creation bug and RoPE-layer dropping are different model/config regressions." + }, + { + "left": "issue:43720", + "right": "issue:43819", + "accept": false, + "reason": "Packed-weight loading in BitNet is unrelated to DAC forward/from_latents mismatch." + }, + { + "left": "issue:43653", + "right": "issue:43720", + "accept": false, + "reason": "BigBirdTokenizer special-token decoding bug is unrelated to BitNet packed-weight loading." + }, + { + "left": "issue:43421", + "right": "issue:44811", + "accept": false, + "reason": "Both involve tokenization/decoding, but one is runtime post-processor updates and the other is Whisper batch_decode skip_special_tokens behavior." + }, + { + "left": "issue:30064", + "right": "issue:36010", + "accept": false, + "reason": "Image processor segmentation-map handling is unrelated to a GenerationMixin import error." + }, + { + "left": "issue:44843", + "right": "issue:45081", + "accept": false, + "reason": "Both touch _patch_mistral_regex, but one is unconditional hub metadata access/offline failure and the other is an attribute error on tokenizer loading." + }, + { + "left": "issue:44062", + "right": "issue:44488", + "accept": false, + "reason": "AddedToken special-argument duplication is unrelated to the model-loading problem for cjvt/sleng-bert." + }, + { + "left": "issue:44186", + "right": "issue:44898", + "accept": false, + "reason": "Tokenizer padding/truncation crashes are not the same as Perceiver image resolution interpolation failures." + }, + { + "left": "issue:43540", + "right": "issue:45381", + "accept": false, + "reason": "Both are video-related, but they affect different models and different internal bugs." + }, + { + "left": "issue:43550", + "right": "issue:45072", + "accept": false, + "reason": "Bamba torch.compile SDPA failure is unrelated to dtype mismatches in SwitchTransformers/TimmWrapperModel inference." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 13, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:30064", + "issue:34634", + "issue:36010", + "issue:36331", + "issue:41720", + "issue:42175", + "issue:43262", + "issue:43531", + "issue:43550", + "issue:43577", + "issue:43582", + "issue:43606", + "issue:43643", + "issue:43644", + "issue:43650", + "issue:43653", + "issue:43720", + "issue:43749", + "issue:43782", + "issue:43818", + "issue:43819", + "issue:43825", + "issue:43827", + "issue:43906", + "issue:43927", + "issue:43994", + "issue:44060", + "issue:44186", + "issue:44368", + "issue:44442", + "issue:44448", + "issue:44485", + "issue:44509", + "issue:44561", + "issue:44610", + "issue:44671", + "issue:44805", + "issue:44821", + "issue:44877", + "issue:44936", + "issue:44977", + "issue:44991", + "issue:45020", + "issue:45072", + "issue:45081", + "issue:45084", + "issue:45325", + "issue:45406" + ], + "soft_pairs": [ + "issue:43577|issue:43818", + "issue:43643|issue:44877", + "issue:43550|issue:43927", + "issue:43782|issue:44977", + "issue:43531|issue:43643", + "issue:43577|issue:43582", + "issue:44448|issue:44671", + "issue:43644|issue:43749", + "issue:43653|issue:43819", + "issue:44561|issue:45020", + "issue:43825|issue:43994", + "issue:43550|issue:43819", + "issue:43825|issue:43827", + "issue:44186|issue:44442", + "issue:44821|issue:45084", + "issue:43650|issue:43906", + "issue:41720|issue:44991", + "issue:44805|issue:44936", + "issue:30064|issue:36331", + "issue:44485|issue:44509", + "issue:43606|issue:45072", + "issue:42175|issue:45081", + "issue:43782|issue:45325", + "issue:44610|issue:45072", + "issue:34634|issue:36010", + "issue:44060|issue:44368", + "issue:45081|issue:45406", + "issue:43720|issue:45072", + "issue:43262|issue:43819", + "issue:43653|issue:44671" + ], + "prepared_review_unit_hash": "a40f6d534b3accdb894f805fe61485b9afab9d7fbb8d4c540563e4fd505556c9", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36866, + "estimated_input_tokens": 9217, + "estimated_eval_tokens": 18690 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly a set of unrelated regression reports that share broad Transformers-version or model-family vocabulary rather than the same underlying bug. I found no soft-edge pair that is clearly the same concrete issue/change, so all candidate links are rejected.", + "confidence": 0.71, + "canonical_issue_reason": "issue:45020 is the broadest active report in the set and best fits a loose umbrella role for recent-version remote_code regressions, though the cluster overall is too heterogeneous for a strong canonical choice.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45020 has the widest scope, is still open, and can plausibly serve as the representative tracking issue for the cluster\u2019s loose theme of recent-version breakages.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43577", + "right": "issue:43818", + "accept": false, + "reason": "Different problems: one is a dtype-loading issue in BLIP2, the other is a Video-LLaVA model-structure/weight-sharing bug." + }, + { + "left": "issue:43643", + "right": "issue:44877", + "accept": false, + "reason": "Both involve config loading, but one is about missing fields from trust_remote_code and the other is a strict-config failure for granite_speech; not the same bug." + }, + { + "left": "issue:43550", + "right": "issue:43927", + "accept": false, + "reason": "Unrelated: torch.compile/SDPA failure for Bamba versus DiaConfig losing custom token IDs on save/load." + }, + { + "left": "issue:43782", + "right": "issue:44977", + "accept": false, + "reason": "Different model and failure mode: Qwen3VL weight_only loading error versus Qwen3.5 flash-attention generation issue." + }, + { + "left": "issue:43531", + "right": "issue:43643", + "accept": false, + "reason": "No concrete overlap beyond being Transformer model/config issues; sliding_window behavior and missing config fields are distinct." + }, + { + "left": "issue:43577", + "right": "issue:43582", + "accept": false, + "reason": "BLIP2 dtype mismatch and AppleSilicon caching_allocator_warmup TypeError are unrelated." + }, + { + "left": "issue:44448", + "right": "issue:44671", + "accept": false, + "reason": "Both are v5 output regressions, but for different models and different code paths; too broad to merge." + }, + { + "left": "issue:43644", + "right": "issue:43749", + "accept": false, + "reason": "Non-persistent buffer initialization bug and FSDP CPU RAM efficient loading breakage are separate loading-path problems." + }, + { + "left": "issue:43653", + "right": "issue:43819", + "accept": false, + "reason": "Tokenizer special-token registration and DAC latent reconstruction are unrelated." + }, + { + "left": "issue:44561", + "right": "issue:45020", + "accept": false, + "reason": "45020 is a broad umbrella report, while 44561 is a specific v5 regression from removing is_torch_fx_available; not clearly the same concrete issue." + }, + { + "left": "issue:43825", + "right": "issue:43994", + "accept": false, + "reason": "Pipeline error-message regression and SigLIP2 nonsensical outputs are different issues." + }, + { + "left": "issue:43825", + "right": "issue:43827", + "accept": false, + "reason": "Related only at the documentation/pipeline-removal theme; one is an error message bug, the other is stale docs." + }, + { + "left": "issue:44186", + "right": "issue:44442", + "accept": false, + "reason": "LayoutLMv2Tokenizer NER/padding crash and FastSpeech2ConformerTokenizer loading failure are unrelated tokenizer issues." + }, + { + "left": "issue:44821", + "right": "issue:45084", + "accept": false, + "reason": "AutoImageProcessor URL loading and template-node compilation errors do not share the same failure mode." + }, + { + "left": "issue:43650", + "right": "issue:43906", + "accept": false, + "reason": "A low-content placeholder issue and an isolated reproduction of another issue are not duplicates." + }, + { + "left": "issue:41720", + "right": "issue:44991", + "accept": false, + "reason": "Different symptoms and subsystems: Qwen3 device-mapping CUDA assert versus tokenizer loading for EMBEDDIA/est-roberta." + }, + { + "left": "issue:44805", + "right": "issue:44936", + "accept": false, + "reason": "Mask-shape IndexError and trainer.evaluate after train are not the same bug." + }, + { + "left": "issue:30064", + "right": "issue:36331", + "accept": false, + "reason": "Image processor void segmentation map handling is unrelated to CustomTrainer.compute_loss keyword-arg incompatibility." + }, + { + "left": "issue:44485", + "right": "issue:44509", + "accept": false, + "reason": "GLM-5 RoPE discussion and docs mentioning removed pipeline tasks are different topics." + }, + { + "left": "issue:43606", + "right": "issue:45072", + "accept": false, + "reason": "CPU offload device mismatch and bfloat16 dtype mismatch are separate inference bugs." + }, + { + "left": "issue:42175", + "right": "issue:45081", + "accept": false, + "reason": "TensorFlow extra installation issue and Mistral regex patch crash are unrelated." + }, + { + "left": "issue:43782", + "right": "issue:45325", + "accept": false, + "reason": "Both are Qwen VL issues, but one is a weight_only loading error and the other is a rope-index scaling bug; not the same change." + }, + { + "left": "issue:44610", + "right": "issue:45072", + "accept": false, + "reason": "Processor image-size mismatch and dtype mismatch in inference are different defects." + }, + { + "left": "issue:34634", + "right": "issue:36010", + "accept": false, + "reason": "BarkProcessor voice_preset and GenerationMixin import failure are unrelated." + }, + { + "left": "issue:44060", + "right": "issue:44368", + "accept": false, + "reason": "Both mention tied-weights warnings around Qwen models, but one is about an incorrect internal weight tie and the other is a user-visible warning during LoRA fine-tuning." + }, + { + "left": "issue:45081", + "right": "issue:45406", + "accept": false, + "reason": "Mistral regex patching crash and Gemma4Processor missing _tokenizer in transformers serve are different processor bugs." + }, + { + "left": "issue:43720", + "right": "issue:45072", + "accept": false, + "reason": "BitNet packed-weight unpacking and bfloat16 dtype mismatches are not the same underlying problem." + }, + { + "left": "issue:43262", + "right": "issue:43819", + "accept": false, + "reason": "Audio sampling-rate defaulting and DAC latent reconstruction mismatch are unrelated." + }, + { + "left": "issue:43653", + "right": "issue:44671", + "accept": false, + "reason": "Tokenizer special-token registration and CamemBERT masked-LM prediction regression 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": 2, + "component_count": 13, + "review_unit_index": 14, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:34634", + "issue:36010", + "issue:36331", + "issue:38175", + "issue:39692", + "issue:43257", + "issue:43262", + "issue:43329", + "issue:43452", + "issue:43454", + "issue:43526", + "issue:43531", + "issue:43550", + "issue:43577", + "issue:43582", + "issue:43638", + "issue:43645", + "issue:43653", + "issue:43696", + "issue:43697", + "issue:43720", + "issue:43746", + "issue:43749", + "issue:43756", + "issue:43819", + "issue:43854", + "issue:43927", + "issue:43931", + "issue:44291", + "issue:44479", + "issue:44554", + "issue:44568", + "issue:44589", + "issue:44610", + "issue:44756", + "issue:44811", + "issue:44857", + "issue:44912", + "issue:44936", + "issue:44977", + "issue:44991", + "issue:45042", + "issue:45084", + "issue:45356", + "issue:45488", + "issue:45538", + "issue:45600" + ], + "soft_pairs": [ + "issue:43749|issue:44291", + "issue:43582|issue:44857", + "issue:43526|issue:43697", + "issue:43526|issue:44554", + "issue:45084|issue:45538", + "issue:43329|issue:45600", + "issue:43653|issue:44912", + "issue:43454|issue:43577", + "issue:44568|issue:45488", + "issue:34634|issue:36331", + "issue:43638|issue:44936", + "issue:43452|issue:44991", + "issue:43257|issue:43329", + "issue:43746|issue:44811", + "issue:38175|issue:44610", + "issue:43854|issue:44977", + "issue:44589|issue:45042", + "issue:43257|issue:43696", + "issue:39692|issue:44912", + "issue:44568|issue:45356", + "issue:36010|issue:43645", + "issue:43819|issue:43927", + "issue:43720|issue:44610", + "issue:43262|issue:44811", + "issue:43550|issue:44912", + "issue:43749|issue:44756", + "issue:43531|issue:45042", + "issue:43329|issue:44479", + "issue:43257|issue:43931", + "issue:43257|issue:43756", + "issue:36010|issue:36331" + ], + "prepared_review_unit_hash": "6b59f64fc4b00194160d74c2c96dc296206bbeb3c9a663b8a7af88ab77831f0e", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36684, + "estimated_input_tokens": 9171, + "estimated_eval_tokens": 18598 + }, + "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 regressions across loading, tokenization, processors, and training. I\u2019d only merge one soft edge: the Qwen video-input regression pair (43329, 44479). Everything else looks like broad thematic similarity rather than the same underlying bug.", + "confidence": 0.37, + "canonical_issue_reason": "issue:43329 is the most concrete root-cause report here: it points to the broken video branch in `_get_num_multimodal_tokens` with undefined helpers/variables, which is the clearest diagnostic match for the only plausible duplicate pair.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43329 is the best representative of the cluster\u2019s only real duplicate-like subcase because it is specific, actionable, and pinpoints the failing code path rather than just describing user-visible symptoms.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43749", + "right": "issue:44291", + "accept": false, + "reason": "Both involve model loading, but one is FSDP CPU RAM efficient loading and the other is an `init_empty_weights` argument mismatch; different failure points and fixes." + }, + { + "left": "issue:43582", + "right": "issue:44857", + "accept": false, + "reason": "Apple Silicon allocator warmup TypeError vs CUDA float16 AMP loss crash; unrelated platform and subsystem." + }, + { + "left": "issue:43526", + "right": "issue:43697", + "accept": false, + "reason": "BEiT `reduce_labels` fast-processor bug and RTDetrV2 output drift are both vision-related, but they are different code paths and symptoms." + }, + { + "left": "issue:43526", + "right": "issue:44554", + "accept": false, + "reason": "BEiT image-label handling vs an MPS attention correctness issue; no shared bug." + }, + { + "left": "issue:45084", + "right": "issue:45538", + "accept": false, + "reason": "Template-node compile failure and CLIP tokenizer max-length handling are unrelated." + }, + { + "left": "issue:43329", + "right": "issue:45600", + "accept": false, + "reason": "Video multimodal token counting bug vs stale `auto_mappings.py` config references; different parts of the library." + }, + { + "left": "issue:43653", + "right": "issue:44912", + "accept": false, + "reason": "BigBird special-token registration and GPT-oss MXFP4 loading behavior are unrelated." + }, + { + "left": "issue:43454", + "right": "issue:43577", + "accept": false, + "reason": "AyaVision lm_head tying and BLIP-2 dtype propagation are different model-initialization bugs." + }, + { + "left": "issue:44568", + "right": "issue:45488", + "accept": false, + "reason": "Both are tokenizer regressions, but one is special-token/BOS-EOS handling and the other is pre-tokenizer override behavior." + }, + { + "left": "issue:34634", + "right": "issue:36331", + "accept": false, + "reason": "BarkProcessor voice_preset and CustomTrainer `compute_loss` signature are unrelated." + }, + { + "left": "issue:43638", + "right": "issue:44936", + "accept": false, + "reason": "Zero3 training index error vs evaluate-after-train failure; different trainer issues." + }, + { + "left": "issue:43452", + "right": "issue:44991", + "accept": false, + "reason": "GGUF loading breaks for auto classes, while the other is a tokenizer-loading regression for a specific checkpoint; same broad area only." + }, + { + "left": "issue:43257", + "right": "issue:43329", + "accept": false, + "reason": "Qwen3 MoE weight conversion under accelerate/deepspeed is unrelated to the video multimodal token-count bug." + }, + { + "left": "issue:43746", + "right": "issue:44811", + "accept": false, + "reason": "PEFT checkpoint loading and Whisper batch decode skip-special-tokens are different audio/model-loading bugs." + }, + { + "left": "issue:38175", + "right": "issue:44610", + "accept": false, + "reason": "Unexpected zero probabilities in SigLIP2 vs OmDet-Turbo processor size mismatch; both vision issues, but not the same bug." + }, + { + "left": "issue:43854", + "right": "issue:44977", + "accept": false, + "reason": "GLM-4.7-Flash load test failure vs Qwen3.5 flash-attention generation regression; different models and code paths." + }, + { + "left": "issue:44589", + "right": "issue:45042", + "accept": false, + "reason": "Float8 storage lookup during loading vs PIL backend requiring torchvision; unrelated." + }, + { + "left": "issue:43257", + "right": "issue:43696", + "accept": false, + "reason": "Qwen3 MoE weight conversion and GPT-oss OOM are not the same failure." + }, + { + "left": "issue:39692", + "right": "issue:44912", + "accept": false, + "reason": "SigLIP2 doc-example/model-processor mismatch and GPT-oss MXFP4 fallback are different quantization/model-loading problems." + }, + { + "left": "issue:44568", + "right": "issue:45356", + "accept": false, + "reason": "Special-token addition regression vs Kimi-K2.5 codec-handling regression; distinct tokenizer bugs." + }, + { + "left": "issue:36010", + "right": "issue:43645", + "accept": false, + "reason": "GenerationMixin import breakage and custom-model notebook initialization failure are related to v5 API changes but not the same bug." + }, + { + "left": "issue:43819", + "right": "issue:43927", + "accept": false, + "reason": "DAC `from_latents` mismatch and DiaConfig save/load token-ID loss are unrelated." + }, + { + "left": "issue:43720", + "right": "issue:44610", + "accept": false, + "reason": "BitNet packed-weight loading and OmDet processor image-size mismatch are different problems." + }, + { + "left": "issue:43262", + "right": "issue:44811", + "accept": false, + "reason": "Audio chat-template sample-rate default and Whisper batch decode skip-special-tokens are distinct processor bugs." + }, + { + "left": "issue:43550", + "right": "issue:44912", + "accept": false, + "reason": "Bamba torch.compile/SDPA failure vs GPT-oss quantization fallback are unrelated." + }, + { + "left": "issue:43749", + "right": "issue:44756", + "accept": false, + "reason": "Both mention loading/OOM, but one is FSDP CPU RAM efficient loading and the other is mmap on Strix Halo; different root causes." + }, + { + "left": "issue:43531", + "right": "issue:45042", + "accept": false, + "reason": "Qwen3-MoE sliding window behavior and PIL torchvision dependency handling are unrelated." + }, + { + "left": "issue:43329", + "right": "issue:44479", + "accept": true, + "reason": "These look like the same video-input regression in Qwen VLM processors; 43329 identifies the concrete broken video token-counting path, while 44479 reports the broader symptom across affected models." + }, + { + "left": "issue:43257", + "right": "issue:43931", + "accept": false, + "reason": "Qwen3 MoE weight conversion and Qwen3-VL shape mismatch are different loading bugs." + }, + { + "left": "issue:43257", + "right": "issue:43756", + "accept": false, + "reason": "Qwen3 MoE weight conversion is unrelated to Smollm3 RoPE-layer dropping." + }, + { + "left": "issue:36010", + "right": "issue:36331", + "accept": false, + "reason": "GenerationMixin import failure and `compute_loss` signature mismatch are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The single accepted soft edge (43329, 44479) is plausibly the same video-input regression, and the rejected pairs are reasonably distinguished by different failure modes or code paths. No clear overstatement of evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 15, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:28282", + "issue:33357", + "issue:34567", + "issue:34689", + "issue:36010", + "issue:38175", + "issue:40990", + "issue:41720", + "issue:42175", + "issue:42915", + "issue:43317", + "issue:43329", + "issue:43450", + "issue:43452", + "issue:43526", + "issue:43540", + "issue:43577", + "issue:43582", + "issue:43646", + "issue:43697", + "issue:43742", + "issue:43749", + "issue:43782", + "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" + ], + "soft_pairs": [ + "issue:34567|issue:40990", + "issue:43697|issue:45357", + "issue:43329|issue:43450", + "issue:43577|issue:45072", + "issue:44479|issue:45381", + "issue:41720|issue:43540", + "issue:43452|issue:43582", + "issue:44987|issue:45538", + "issue:42175|issue:44938", + "issue:43317|issue:45538", + "issue:43927|issue:44912", + "issue:44291|issue:45042", + "issue:43906|issue:45381", + "issue:43540|issue:44560", + "issue:33357|issue:34567", + "issue:43540|issue:44479", + "issue:42915|issue:45538", + "issue:42915|issue:43742", + "issue:42175|issue:43824", + "issue:43526|issue:45479", + "issue:42915|issue:44821", + "issue:44062|issue:44991", + "issue:44737|issue:45588", + "issue:42175|issue:44488", + "issue:45137|issue:45405", + "issue:38175|issue:44898", + "issue:43646|issue:43749", + "issue:43582|issue:43782", + "issue:28282|issue:34689", + "issue:36010|issue:42175" + ], + "prepared_review_unit_hash": "f39cab3844e68597ebb04904a026187ed22343b431e2f29d68f78eb151c6ea55", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36212, + "estimated_input_tokens": 9053, + "estimated_eval_tokens": 18362 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly unrelated issue reports with one plausible duplicate pair around the Qwen video-input regression; no PRs are present.", + "confidence": 0.44, + "canonical_issue_reason": "Issue 44479 is the broadest and clearest report: it describes a v5.3.0 video-input regression across multiple Qwen model families, which likely subsumes the more specific qwen2.5-vl manifestation.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44479 is the best representative issue because it captures the shared regression surface and is broad enough to anchor the related Qwen video-input discussion.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:34567", + "right": "issue:40990", + "accept": false, + "reason": "Unrelated: training-state metric update vs perplexity evaluation behavior." + }, + { + "left": "issue:43697", + "right": "issue:45357", + "accept": false, + "reason": "Different bugs and model paths: RTDetr output drift vs Qwen3.5 save_pretrained key regression." + }, + { + "left": "issue:43329", + "right": "issue:43450", + "accept": false, + "reason": "Both involve video/multimodal code, but one is an undefined-variable token-counting bug and the other is a batched processor shape issue." + }, + { + "left": "issue:43577", + "right": "issue:45072", + "accept": false, + "reason": "Both mention dtype, but they affect different models and different failure modes." + }, + { + "left": "issue:44479", + "right": "issue:45381", + "accept": true, + "reason": "Likely same underlying Qwen video-input regression: 44479 is the umbrella report, while 45381 is a specific qwen2.5-vl symptom around video position IDs." + }, + { + "left": "issue:41720", + "right": "issue:43540", + "accept": false, + "reason": "Different concrete failures: auto device mapping CUDA assert vs video-processing ValueError in a different Qwen model." + }, + { + "left": "issue:43452", + "right": "issue:43582", + "accept": false, + "reason": "No shared code path: gguf loading regression vs Apple Silicon allocator warmup type error." + }, + { + "left": "issue:44987", + "right": "issue:45538", + "accept": false, + "reason": "Unrelated: model-loading failure for a specific repo vs CLIP tokenizer max-length semantics." + }, + { + "left": "issue:42175", + "right": "issue:44938", + "accept": false, + "reason": "Packaging/backend availability issue vs a Python 3.14 loading failure; too generic to treat as duplicates." + }, + { + "left": "issue:43317", + "right": "issue:45538", + "accept": false, + "reason": "Different areas: device_map/offload loading bug vs tokenizer max-length constant." + }, + { + "left": "issue:43927", + "right": "issue:44912", + "accept": false, + "reason": "Different model/config bugs: DiaConfig persistence vs MXFP4 quantization fallback." + }, + { + "left": "issue:44291", + "right": "issue:45042", + "accept": false, + "reason": "Different load-time API mismatch vs image-processor torchvision dependency regression." + }, + { + "left": "issue:43906", + "right": "issue:45381", + "accept": false, + "reason": "A reproduction of another issue, but not the same concrete bug as the qwen2.5-vl video-position-ID regression." + }, + { + "left": "issue:43540", + "right": "issue:44560", + "accept": false, + "reason": "Both are video-related, but they describe different models and distinct observed failures." + }, + { + "left": "issue:33357", + "right": "issue:34567", + "accept": false, + "reason": "Unrelated Mac bus error in CLIP vs TrainerState token-count tracking." + }, + { + "left": "issue:43540", + "right": "issue:44479", + "accept": false, + "reason": "Related theme, but not enough to confirm the same code-path bug; one is a specific model ValueError and the other is a broader regression report." + }, + { + "left": "issue:42915", + "right": "issue:45538", + "accept": false, + "reason": "Different categories: FP8 config loading vs tokenizer max-length behavior." + }, + { + "left": "issue:42915", + "right": "issue:43742", + "accept": false, + "reason": "Different failures in different model-loading paths." + }, + { + "left": "issue:42175", + "right": "issue:43824", + "accept": false, + "reason": "Packaging/backend issue vs missing Qwen2.5-VL class import." + }, + { + "left": "issue:43526", + "right": "issue:45479", + "accept": false, + "reason": "Different model families and different functional bugs." + }, + { + "left": "issue:42915", + "right": "issue:44821", + "accept": false, + "reason": "Unrelated: FP8 Qwen3Moe failure vs AutoImageProcessor URL loading." + }, + { + "left": "issue:44062", + "right": "issue:44991", + "accept": false, + "reason": "Tokenizer-added-token argument bug vs tokenizer-loading regression for a specific model." + }, + { + "left": "issue:44737", + "right": "issue:45588", + "accept": false, + "reason": "XLNet CPU placement bug vs flash-attention sinkless-model AttributeError." + }, + { + "left": "issue:42175", + "right": "issue:44488", + "accept": false, + "reason": "Backend/install packaging issue vs model-loading failure for cjvt/sleng-bert." + }, + { + "left": "issue:45137", + "right": "issue:45405", + "accept": false, + "reason": "DeepSpeed ZeRO3 runtime error vs dependency version bump; unrelated." + }, + { + "left": "issue:38175", + "right": "issue:44898", + "accept": false, + "reason": "Different vision-model bugs with different symptoms and code paths." + }, + { + "left": "issue:43646", + "right": "issue:43749", + "accept": false, + "reason": "Transformers 5 custom init breakage vs FSDP CPU RAM efficient loading regression." + }, + { + "left": "issue:43582", + "right": "issue:43782", + "accept": false, + "reason": "Apple Silicon allocator warmup type error vs Qwen3VL weight-only load error." + }, + { + "left": "issue:28282", + "right": "issue:34689", + "accept": false, + "reason": "Missing PyTorch import vs Llama 3.2 Vision loading breakage; not the same issue." + }, + { + "left": "issue:36010", + "right": "issue:42175", + "accept": false, + "reason": "Import path regression for GenerationMixin vs missing TensorFlow backend in an install extra." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall: the canonical issue choice is supported by the packet, no PRs are present, and the one accepted soft-edge pair (44479/45381) is stated cautiously. The rejected pairs are generally conservative and do not overclaim duplication." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 16, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:28282", + "issue:29127", + "issue:34567", + "issue:36331", + "issue:41762", + "issue:43452", + "issue:43493", + "issue:43531", + "issue:43577", + "issue:43643", + "issue:43653", + "issue:43688", + "issue:43697", + "issue:43720", + "issue:43742", + "issue:43749", + "issue:43756", + "issue:43819", + "issue:43994", + "issue:44062", + "issue:44291", + "issue:44484", + "issue:44488", + "issue:44514", + "issue:44561", + "issue:44568", + "issue:44589", + "issue:44625", + "issue:44704", + "issue:44792", + "issue:44811", + "issue:44821", + "issue:44843", + "issue:44877", + "issue:44898", + "issue:44936", + "issue:44964", + "issue:45042", + "issue:45081", + "issue:45216", + "issue:45290", + "issue:45325", + "issue:45375", + "issue:45381", + "issue:45405", + "issue:45440", + "issue:45600" + ], + "soft_pairs": [ + "issue:43643|issue:45375", + "issue:43756|issue:45325", + "issue:43742|issue:44291", + "issue:44484|issue:44704", + "issue:43493|issue:43994", + "issue:28282|issue:29127", + "issue:34567|issue:36331", + "issue:44062|issue:45081", + "issue:44291|issue:44821", + "issue:43452|issue:44488", + "issue:44792|issue:44964", + "issue:44514|issue:45290", + "issue:45405|issue:45600", + "issue:43643|issue:44561", + "issue:43688|issue:43749", + "issue:44514|issue:45381", + "issue:43697|issue:45216", + "issue:43452|issue:43577", + "issue:45042|issue:45081", + "issue:43742|issue:44589", + "issue:43531|issue:44821", + "issue:43688|issue:45440", + "issue:43452|issue:44843", + "issue:44625|issue:44877", + "issue:43653|issue:44568", + "issue:43720|issue:44898", + "issue:43819|issue:44811", + "issue:41762|issue:44936" + ], + "prepared_review_unit_hash": "497a926599b4d3624f159cbaefa2e132eacd11606be0e0cf46edecd1917a48a5", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35845, + "estimated_input_tokens": 8962, + "estimated_eval_tokens": 18180 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly a loose set of unrelated Transformers bug reports that only share high-level loading/API vocabulary. I rejected all soft edges because the paired items do not describe the same concrete underlying bug or change.", + "confidence": 0.62, + "canonical_issue_reason": "issue:43452 is the most broadly representative loading-related report in the set, spanning core `from_pretrained` behavior for multiple auto classes and fitting the loose thematic center of the cluster better than the more specialized reports.", + "canonical_pr_reason": null, + "best_issue_reason": "It is a clear, user-facing regression in a core loading path and is broad enough to act as the best single representative issue for this heterogeneous cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43643", + "right": "issue:45375", + "accept": false, + "reason": "Both involve missing config fields, but they are different mechanisms on different models: `trust_remote_code` returning incomplete config vs a strict config dropping `deepstack_visual_indexes`." + }, + { + "left": "issue:43756", + "right": "issue:45325", + "accept": false, + "reason": "Both mention RoPE/position IDs, but the bugs are model-specific and distinct: Smollm3 dropping RoPE layers vs Qwen2.5-VL temporal position scaling." + }, + { + "left": "issue:43742", + "right": "issue:44291", + "accept": false, + "reason": "One is a key error when loading MobileLLM, the other is an `init_empty_weights`/`_is_hf_initialized` TypeError. Different failure modes and code paths." + }, + { + "left": "issue:44484", + "right": "issue:44704", + "accept": false, + "reason": "A save-pretrained sharding-size question is unrelated to `AutoProcessor` failing to forward kwargs to `cached_file`." + }, + { + "left": "issue:43493", + "right": "issue:43994", + "accept": false, + "reason": "Both are about SigLIP2, but one is a HF-vs-JAX implementation discrepancy and the other is bad outputs with AutoModel/pipeline; not the same concrete bug." + }, + { + "left": "issue:28282", + "right": "issue:29127", + "accept": false, + "reason": "Torch import failure for AutoModel is unrelated to unclear error messaging for insufficient boxes in LayoutLMv3." + }, + { + "left": "issue:34567", + "right": "issue:36331", + "accept": false, + "reason": "Trainer state tracking and a custom `compute_loss` signature break are separate trainer issues with different symptoms and fixes." + }, + { + "left": "issue:44062", + "right": "issue:45081", + "accept": false, + "reason": "One is an `AddedToken` constructor conflict, the other is a Mistral regex patch crashing on tokenizer internals; they are different tokenizer bugs." + }, + { + "left": "issue:44291", + "right": "issue:44821", + "accept": false, + "reason": "`init_empty_weights` loading breakage and URL loading for `AutoImageProcessor` are unrelated." + }, + { + "left": "issue:43452", + "right": "issue:44488", + "accept": false, + "reason": "Both are loading-related, but one is a GGUF-specific breakage in auto classes and the other is a model-specific inability to load `cjvt/sleng-bert`." + }, + { + "left": "issue:44792", + "right": "issue:44964", + "accept": false, + "reason": "Different multimodal models and different reports; no evidence they share the same underlying defect." + }, + { + "left": "issue:44514", + "right": "issue:45290", + "accept": false, + "reason": "Both crash in `apply_chat_template`, but the triggering conditions differ materially: batched input with `padding=False` vs assistant messages with tool calls and no content." + }, + { + "left": "issue:45405", + "right": "issue:45600", + "accept": false, + "reason": "A PEFT version bump issue is unrelated to stale `auto_mappings` references to removed configs." + }, + { + "left": "issue:43643", + "right": "issue:44561", + "accept": false, + "reason": "Both involve config/remote-code compatibility, but one is missing fields from `AutoConfig`, while the other is removal of `is_torch_fx_available` breaking remote-code models." + }, + { + "left": "issue:43688", + "right": "issue:43749", + "accept": false, + "reason": "Auxiliary-loss normalization and FSDP CPU RAM efficient loading are unrelated subsystems and failure modes." + }, + { + "left": "issue:44514", + "right": "issue:45381", + "accept": false, + "reason": "Both concern Qwen2.5-VL chat/template behavior, but one is a template crash on batched padding and the other is incorrect `vision_position_ids`; not the same bug." + }, + { + "left": "issue:43697", + "right": "issue:45216", + "accept": false, + "reason": "RTDetrV2 output mismatch and Qwen3.5 checkpoint save corruption are different regression classes with different code paths." + }, + { + "left": "issue:43452", + "right": "issue:43577", + "accept": false, + "reason": "`gguf_file`/auto-class loading issues are unrelated to BLIP2 dtype propagation for `qformer`." + }, + { + "left": "issue:45042", + "right": "issue:45081", + "accept": false, + "reason": "Torchvision being incorrectly required for PIL backend image processors is unrelated to the Mistral regex/tokenizer crash." + }, + { + "left": "issue:43742", + "right": "issue:44589", + "accept": false, + "reason": "MobileLLM key error and a missing `Float8_e4m3fnStorage` type are separate loading/runtime failures." + }, + { + "left": "issue:43531", + "right": "issue:44821", + "accept": false, + "reason": "Qwen3-MoE sliding-window behavior and image processor loading from URL are unrelated." + }, + { + "left": "issue:43688", + "right": "issue:45440", + "accept": false, + "reason": "Both involve MoE-style models, but the reported defects are distinct: auxiliary-loss normalization vs implementation divergence from a remote model." + }, + { + "left": "issue:43452", + "right": "issue:44843", + "accept": false, + "reason": "GGUF/auto-class loading and offline `model_info()` calls in `_patch_mistral_regex` are different loading regressions." + }, + { + "left": "issue:44625", + "right": "issue:44877", + "accept": false, + "reason": "Qwen3.5 `num_labels` propagation and strict config preventing `granite_speech` loading are separate config bugs." + }, + { + "left": "issue:43653", + "right": "issue:44568", + "accept": false, + "reason": "Special-token registration causing empty decode output is different from `add_special_tokens=True` failing to add BOS/EOS." + }, + { + "left": "issue:43720", + "right": "issue:44898", + "accept": false, + "reason": "BitNet packed-weight loading and Perceiver non-default-resolution failure are unrelated." + }, + { + "left": "issue:43819", + "right": "issue:44811", + "accept": false, + "reason": "DAC latent decoding mismatch and Whisper `batch_decode` ignoring `skip_special_tokens` are different bugs in different components." + }, + { + "left": "issue:41762", + "right": "issue:44936", + "accept": false, + "reason": "DeepSpeed ZeRO-3 loading for Gemma3 and `trainer.evaluate()` after `train()` are unrelated trainer/model-loading issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 17, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:24643", + "issue:28282", + "issue:29127", + "issue:30064", + "issue:32090", + "issue:33357", + "issue:33666", + "issue:34567", + "issue:35141", + "issue:38175", + "issue:41720", + "issue:43122", + "issue:43262", + "issue:43295", + "issue:43329", + "issue:43452", + "issue:43504", + "issue:43525", + "issue:43531", + "issue:43550", + "issue:43577", + "issue:43643", + "issue:43653", + "issue:43697", + "issue:43720", + "issue:43742", + "issue:43906", + "issue:43927", + "issue:44361", + "issue:44442", + "issue:44451", + "issue:44464", + "issue:44466", + "issue:44479", + "issue:44554", + "issue:44625", + "issue:44743", + "issue:44843", + "issue:44849", + "issue:44936", + "issue:44977", + "issue:44991", + "issue:45005", + "issue:45084", + "issue:45325", + "issue:45356", + "issue:45357", + "issue:45588" + ], + "soft_pairs": [ + "issue:43295|issue:45356", + "issue:44466|issue:45005", + "issue:44464|issue:44743", + "issue:41720|issue:43906", + "issue:24643|issue:29127", + "issue:43525|issue:43643", + "issue:43577|issue:43697", + "issue:38175|issue:43295", + "issue:44361|issue:44442", + "issue:44843|issue:44991", + "issue:43653|issue:44361", + "issue:44479|issue:44936", + "issue:43122|issue:43906", + "issue:44977|issue:45356", + "issue:28282|issue:32090", + "issue:28282|issue:33357", + "issue:28282|issue:35141", + "issue:43577|issue:45588", + "issue:30064|issue:33666", + "issue:43577|issue:44554", + "issue:43653|issue:43927", + "issue:44464|issue:44977", + "issue:43531|issue:43742", + "issue:43329|issue:45325", + "issue:44479|issue:44625", + "issue:43550|issue:44361", + "issue:43452|issue:44451", + "issue:43504|issue:43720", + "issue:41720|issue:44464", + "issue:43504|issue:43550", + "issue:43262|issue:45084", + "issue:44849|issue:45357", + "issue:32090|issue:34567", + "issue:32090|issue:33666" + ], + "prepared_review_unit_hash": "2fd41c04be8e77d6954e741c61678b11eb23dbf55502cc321b32cf5f9dd76f89", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 38189, + "estimated_input_tokens": 9548, + "estimated_eval_tokens": 19352 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is heterogeneous: the soft links are mostly broad subsystem similarity (tokenizers, Qwen, tied weights, generation) rather than the same underlying bug. I would not merge any of these pairs as duplicates.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43295", + "right": "issue:45356", + "accept": false, + "reason": "Both are tokenizer/processor regressions, but they affect different code paths: processor.tokenizer/images handling vs Kimi-K2.5 codec and warning behavior." + }, + { + "left": "issue:44466", + "right": "issue:45005", + "accept": false, + "reason": "Both mention tied weights, but one is a device-dependent lm_head serialization bug and the other is a broader translation-model tied-weights issue." + }, + { + "left": "issue:44464", + "right": "issue:44743", + "accept": false, + "reason": "Different generation bugs: compiled-forward inconsistency vs recurrent state reset with cache and seq_len>1." + }, + { + "left": "issue:41720", + "right": "issue:43906", + "accept": false, + "reason": "No clear match; one is a Qwen3 auto-device-map cuda assert, the other is an isolated reproduction of a different issue." + }, + { + "left": "issue:24643", + "right": "issue:29127", + "accept": false, + "reason": "Completely different failures: DeepSpeed training weight-shape error vs LayoutLMv3 box-validation error message." + }, + { + "left": "issue:43525", + "right": "issue:43643", + "accept": false, + "reason": "Both are config/field problems, but they concern different models and different missing-field behavior." + }, + { + "left": "issue:43577", + "right": "issue:43697", + "accept": false, + "reason": "Unrelated bugs: BLIP2 dtype not propagated vs RTDetrV2 output mismatch in v5." + }, + { + "left": "issue:38175", + "right": "issue:43295", + "accept": false, + "reason": "Different models and symptoms: zero probabilities in SigLIP2 vs processor.tokenizer/images regression." + }, + { + "left": "issue:44361", + "right": "issue:44442", + "accept": false, + "reason": "Both are tokenizer-loading failures, but they are different models and different error modes." + }, + { + "left": "issue:44843", + "right": "issue:44991", + "accept": false, + "reason": "Both involve tokenizer loading, but one is an offline model_info call bug and the other is a model-specific v5 tokenizer load failure." + }, + { + "left": "issue:43653", + "right": "issue:44361", + "accept": false, + "reason": "Different tokenizers and failures: BigBird special-token registration vs MLukeTokenizer attribute error." + }, + { + "left": "issue:44479", + "right": "issue:44936", + "accept": false, + "reason": "Different areas entirely: Qwen3.5 video regression vs Trainer evaluate-after-train behavior." + }, + { + "left": "issue:43122", + "right": "issue:43906", + "accept": false, + "reason": "Only loosely related through tokenization/versioning; the reported failures and targets are not the same bug." + }, + { + "left": "issue:44977", + "right": "issue:45356", + "accept": false, + "reason": "Different regressions: Qwen3.5 flash-attention generation vs Kimi-K2.5 tokenizer codec handling." + }, + { + "left": "issue:28282", + "right": "issue:32090", + "accept": false, + "reason": "ImportError for missing PyTorch is unrelated to Trainer NoneType broadcast during distributed loading." + }, + { + "left": "issue:28282", + "right": "issue:33357", + "accept": false, + "reason": "Different problems: missing PyTorch dependency vs MacOS bus error with a CLIP model." + }, + { + "left": "issue:28282", + "right": "issue:35141", + "accept": false, + "reason": "No shared bug; dependency import failure is unrelated to embedding resize/reinit behavior." + }, + { + "left": "issue:43577", + "right": "issue:45588", + "accept": false, + "reason": "Unrelated: BLIP2 dtype propagation vs flash_attention crash on s_aux=None." + }, + { + "left": "issue:30064", + "right": "issue:33666", + "accept": false, + "reason": "Different subsystems and failures: image processor void segmentation maps vs Qwen2-VL multi-GPU training." + }, + { + "left": "issue:43577", + "right": "issue:44554", + "accept": false, + "reason": "Different code paths: BLIP2 dtype issue vs MPS attention correctness issue when value head dim differs." + }, + { + "left": "issue:43653", + "right": "issue:43927", + "accept": false, + "reason": "Both are tokenizer/config correctness issues, but the underlying bugs are different and not mergeable as one fix." + }, + { + "left": "issue:44464", + "right": "issue:44977", + "accept": false, + "reason": "Different generation regressions: compiled forward chunking vs flash-attention generation on Qwen3.5." + }, + { + "left": "issue:43531", + "right": "issue:43742", + "accept": false, + "reason": "No overlap beyond being model-loading/runtime issues; sliding_window bug and MobileLLM key error are distinct." + }, + { + "left": "issue:43329", + "right": "issue:45325", + "accept": false, + "reason": "Both are Qwen multimodal/video-related, but one is an undefined-symbol video branch bug and the other is rope/position-id scaling." + }, + { + "left": "issue:44479", + "right": "issue:44625", + "accept": false, + "reason": "Different bugs in Qwen3.5: video input regression vs num_labels propagation." + }, + { + "left": "issue:43550", + "right": "issue:44361", + "accept": false, + "reason": "Different components and failures: torch.compile/SDPA on Bamba vs tokenizer attribute error on MLuke." + }, + { + "left": "issue:43452", + "right": "issue:44451", + "accept": false, + "reason": "Both are loading issues, but one is gguf_file handling in Auto* loaders and the other is a model-specific tokenizer load failure." + }, + { + "left": "issue:43504", + "right": "issue:43720", + "accept": false, + "reason": "Unrelated model-loading bugs: legacy field in BEiT vs packed-weight unpacking in BitNet/Aircelerate loading." + }, + { + "left": "issue:41720", + "right": "issue:44464", + "accept": false, + "reason": "Different failures and execution paths: auto device mapping cuda assert vs chunked generation with compiled forward." + }, + { + "left": "issue:43504", + "right": "issue:43550", + "accept": false, + "reason": "Different models and issues; shared only at a very broad \u201cruntime bug\u201d level." + }, + { + "left": "issue:43262", + "right": "issue:45084", + "accept": false, + "reason": "Audio chat-template sampling-rate defaulting is unrelated to compilation failing on non-template nodes." + }, + { + "left": "issue:44849", + "right": "issue:45357", + "accept": false, + "reason": "Different regressions in Qwen3.5: hidden-states behavior vs incorrect visual-encoder keys on save_pretrained." + }, + { + "left": "issue:32090", + "right": "issue:34567", + "accept": false, + "reason": "Trainer distributed broadcast failure is unrelated to TrainerState.num_input_tokens_seen not updating." + }, + { + "left": "issue:32090", + "right": "issue:33666", + "accept": false, + "reason": "Different training-time problems: NoneType broadcast vs multi-GPU training for Qwen2-VL." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 18, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:33666", + "issue:36010", + "issue:36331", + "issue:39692", + "issue:43257", + "issue:43278", + "issue:43295", + "issue:43299", + "issue:43334", + "issue:43493", + "issue:43504", + "issue:43525", + "issue:43540", + "issue:43572", + "issue:43606", + "issue:43646", + "issue:43688", + "issue:43716", + "issue:43720", + "issue:43742", + "issue:43746", + "issue:43756", + "issue:43824", + "issue:43825", + "issue:43844", + "issue:43873", + "issue:43927", + "issue:44060", + "issue:44361", + "issue:44423", + "issue:44466", + "issue:44488", + "issue:44521", + "issue:44589", + "issue:44610", + "issue:44625", + "issue:44734", + "issue:44821", + "issue:44843", + "issue:44913", + "issue:44964", + "issue:44977", + "issue:45200", + "issue:45229", + "issue:45290", + "issue:45405", + "issue:45520", + "issue:45584" + ], + "soft_pairs": [ + "issue:43504|issue:44610", + "issue:43278|issue:43688", + "issue:43334|issue:43572", + "issue:43525|issue:43572", + "issue:43756|issue:44913", + "issue:43540|issue:43716", + "issue:43493|issue:43825", + "issue:43742|issue:44821", + "issue:45200|issue:45229", + "issue:33666|issue:36010", + "issue:43720|issue:43927", + "issue:43606|issue:43927", + "issue:43688|issue:43873", + "issue:43257|issue:43299", + "issue:43278|issue:43844", + "issue:44488|issue:44843", + "issue:43540|issue:44977", + "issue:43720|issue:44361", + "issue:39692|issue:43493", + "issue:44964|issue:45405", + "issue:43493|issue:43844", + "issue:33666|issue:36331", + "issue:43756|issue:44060", + "issue:44589|issue:45520", + "issue:44625|issue:45584", + "issue:44521|issue:45290", + "issue:43646|issue:44466", + "issue:44423|issue:44734", + "issue:43824|issue:44589", + "issue:43688|issue:43746", + "issue:39692|issue:43295" + ], + "prepared_review_unit_hash": "0a8f87fa71a06c139253751805aa1355af77bb1c9044c8e53a7a11b1ec9f1f3b", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37135, + "estimated_input_tokens": 9284, + "estimated_eval_tokens": 18824 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly a grab-bag of distinct model-loading, config, tokenizer, multimodal, and training regressions that only look similar at the title level. I would not merge any of the soft-edge pairs; they describe different concrete bugs. The broadest representative issue is the generic v5/custom-initialization regression.", + "confidence": 0.56, + "canonical_issue_reason": "issue:43646 is the broadest umbrella item in the set: a generic Transformers 5.0 custom model initialization regression that can represent the wider class of load/init failures here.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43646 is the most representative anchor for this cluster because it is broader than the model-specific loading/config bugs and best matches the recurring v5 regression theme.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43504", + "right": "issue:44610", + "accept": false, + "reason": "Different failures: BEiT legacy-field loading vs OmDet processor/model image-size mismatch." + }, + { + "left": "issue:43278", + "right": "issue:43688", + "accept": false, + "reason": "Both mention dtype/normalization issues, but one is an eval dtype regression and the other is auxiliary-loss normalization in MoE models." + }, + { + "left": "issue:43334", + "right": "issue:43572", + "accept": false, + "reason": "Different config fields and models: Qwen3-VL pad_token_id vs StableLm pad_token_idx." + }, + { + "left": "issue:43525", + "right": "issue:43572", + "accept": false, + "reason": "Both are missing config attributes, but they affect different model families and different fields." + }, + { + "left": "issue:43756", + "right": "issue:44913", + "accept": false, + "reason": "Smollm3 RoPE-layer dropping is unrelated to GPTNeoX rotary_pct reload behavior." + }, + { + "left": "issue:43540", + "right": "issue:43716", + "accept": false, + "reason": "Qwen3OmniMoe video-input processing error is not the same as Mistral-3 image-preprocessor dtype mismatch." + }, + { + "left": "issue:43493", + "right": "issue:43825", + "accept": false, + "reason": "SigLIP2 implementation mismatch is unrelated to pipeline translation-task messaging in v5." + }, + { + "left": "issue:43742", + "right": "issue:44821", + "accept": false, + "reason": "MobileLLM key error and AutoImageProcessor URL loading failure are different code paths and model types." + }, + { + "left": "issue:45200", + "right": "issue:45229", + "accept": false, + "reason": "Gemma 4 text-only fine-tuning defaults are unrelated to Gemma 4 multi-GPU inference OOM." + }, + { + "left": "issue:33666", + "right": "issue:36010", + "accept": false, + "reason": "Multi-GPU training for Qwen2-VL is unrelated to an import error for GenerationMixin." + }, + { + "left": "issue:43720", + "right": "issue:43927", + "accept": false, + "reason": "Packed-weight unpacking during accelerate loading is not the same bug as DiaConfig losing custom token IDs on save/load." + }, + { + "left": "issue:43606", + "right": "issue:43927", + "accept": false, + "reason": "CPU offload device mismatch and config token-ID loss are separate failures." + }, + { + "left": "issue:43688", + "right": "issue:43873", + "accept": false, + "reason": "Aux-loss normalization in MoE models is unrelated to quantization/offloading behavior." + }, + { + "left": "issue:43257", + "right": "issue:43299", + "accept": false, + "reason": "Both involve Qwen3 MoE loading, but one is accelerate+deepspeed conversion and the other is a v5 loading regression; not the same concrete bug." + }, + { + "left": "issue:43278", + "right": "issue:43844", + "accept": false, + "reason": "Embedding dtype drift in eval is unrelated to ZeRO-3 gradient inflation." + }, + { + "left": "issue:44488", + "right": "issue:44843", + "accept": false, + "reason": "Model-load failure for cjvt/sleng-bert is unrelated to offline tokenizer hub access in _patch_mistral_regex." + }, + { + "left": "issue:43540", + "right": "issue:44977", + "accept": false, + "reason": "Video-input processing error in Qwen3OmniMoe is not the same as flash-attention generation issues in Qwen3.5." + }, + { + "left": "issue:43720", + "right": "issue:44361", + "accept": false, + "reason": "Both are loader-related, but packed-weight unpacking and MLukeTokenizer task AttributeError are distinct bugs." + }, + { + "left": "issue:39692", + "right": "issue:43493", + "accept": false, + "reason": "SigLIP2 documentation issues are not the same as SigLIP2 implementation-vs-JAX discrepancies." + }, + { + "left": "issue:44964", + "right": "issue:45405", + "accept": false, + "reason": "Phi-4 multimodal load failure and PEFT minimum-version bump are unrelated." + }, + { + "left": "issue:43493", + "right": "issue:43844", + "accept": false, + "reason": "SigLIP2 model mismatch and ZeRO-3 gradient instability are different problem classes." + }, + { + "left": "issue:33666", + "right": "issue:36331", + "accept": false, + "reason": "Multi-GPU training setup for Qwen2-VL is unrelated to the CustomTrainer.compute_loss signature break." + }, + { + "left": "issue:43756", + "right": "issue:44060", + "accept": false, + "reason": "Smollm3 RoPE-layer count bug is unrelated to Qwen3-Next tied-weights warnings." + }, + { + "left": "issue:44589", + "right": "issue:45520", + "accept": false, + "reason": "Float8 storage lookup failure and flash_attn KeyError on Python 3.13 are different import/runtime issues." + }, + { + "left": "issue:44625", + "right": "issue:45584", + "accept": false, + "reason": "Qwen3.5 num_labels propagation and Whisper empty-transcription generation are unrelated." + }, + { + "left": "issue:44521", + "right": "issue:45290", + "accept": false, + "reason": "All-zero assistant masks for multimodal inputs is not the same as tool-call assistant-message crash." + }, + { + "left": "issue:43646", + "right": "issue:44466", + "accept": false, + "reason": "Generic custom model initialization regression is broader, but lm_head.weight serialization/device-dependent tied-weights behavior is a distinct bug." + }, + { + "left": "issue:44423", + "right": "issue:44734", + "accept": false, + "reason": "Continuous-batching multimodal crash and /v1/responses KV-cache indexing bug are different serving-path failures." + }, + { + "left": "issue:43824", + "right": "issue:44589", + "accept": false, + "reason": "Qwen2_5_VL import failure and Float8 storage lookup failure are unrelated." + }, + { + "left": "issue:43688", + "right": "issue:43746", + "accept": false, + "reason": "Aux-loss normalization in MoE models is unrelated to GraniteSpeech PEFT local-checkpoint loading." + }, + { + "left": "issue:39692", + "right": "issue:43295", + "accept": false, + "reason": "SigLIP2 doc-example errors are unrelated to the v4.57.5 processor.tokenizer regression." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 19, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:15354", + "issue:29942", + "issue:30333", + "issue:33357", + "issue:33666", + "issue:37428", + "issue:38175", + "issue:39401", + "issue:43232", + "issue:43278", + "issue:43295", + "issue:43334", + "issue:43504", + "issue:43525", + "issue:43540", + "issue:43572", + "issue:43575", + "issue:43582", + "issue:43606", + "issue:43653", + "issue:43701", + "issue:43844", + "issue:43927", + "issue:43957", + "issue:43994", + "issue:44112", + "issue:44291", + "issue:44315", + "issue:44361", + "issue:44464", + "issue:44493", + "issue:44554", + "issue:44568", + "issue:44610", + "issue:44792", + "issue:44849", + "issue:44964", + "issue:44977", + "issue:45072", + "issue:45084", + "issue:45198", + "issue:45200", + "issue:45325", + "issue:45356", + "issue:45405", + "issue:45446", + "issue:45520", + "issue:45584" + ], + "soft_pairs": [ + "issue:33357|issue:33666", + "issue:43572|issue:44291", + "issue:43994|issue:44493", + "issue:43504|issue:43606", + "issue:44464|issue:45584", + "issue:43504|issue:43653", + "issue:43575|issue:45405", + "issue:43927|issue:44361", + "issue:43278|issue:43701", + "issue:44792|issue:45405", + "issue:43606|issue:44112", + "issue:29942|issue:30333", + "issue:15354|issue:38175", + "issue:44554|issue:45446", + "issue:43540|issue:45200", + "issue:43232|issue:44849", + "issue:45084|issue:45520", + "issue:43844|issue:44964", + "issue:39401|issue:44568", + "issue:43582|issue:44977", + "issue:45072|issue:45198", + "issue:44315|issue:44849", + "issue:37428|issue:43957", + "issue:44554|issue:45356", + "issue:43334|issue:43525", + "issue:43295|issue:44610", + "issue:45200|issue:45325" + ], + "prepared_review_unit_hash": "af3b007793388969c289cd0c34340e224109487c74a3dc4865f23cb3ae9e06cf", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 35862, + "estimated_input_tokens": 8966, + "estimated_eval_tokens": 18188 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "No soft pair looks like a true duplicate: the items mostly share only broad keywords such as model loading, tokenizer, generation, or dtype issues, but they describe different failure modes and code paths.", + "confidence": 0.62, + "canonical_issue_reason": "Issue 43957 is the broadest, most representative report in this set: it describes a general model-loading regression affecting multiple models under transformers 5.x.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43957 is the best standalone representative because it captures a wider cross-model regression than the more model-specific reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:33357", + "right": "issue:33666", + "accept": false, + "reason": "Both mention model usage on specialized hardware/training, but one is a MacOS bus error for CLIP and the other is a Qwen2-VL multi-GPU training request; different bugs and workflows." + }, + { + "left": "issue:43572", + "right": "issue:44291", + "accept": false, + "reason": "One is about missing pad_token_idx after a 5.0 update; the other is an init_empty_weights/_is_hf_initialized argument error. Related release context, but not the same bug." + }, + { + "left": "issue:43994", + "right": "issue:44493", + "accept": false, + "reason": "Both involve wrong outputs, but one is SigLIP2 nonsense results and the other is unexpected position-id keys across many models; different root causes." + }, + { + "left": "issue:43504", + "right": "issue:43606", + "accept": false, + "reason": "Beit pretrained loading with a legacy field is unrelated to a Bark CPU offload device mismatch." + }, + { + "left": "issue:44464", + "right": "issue:45584", + "accept": false, + "reason": "Chunked generation with compiled forward is a generation-path consistency issue; Whisper empty transcription after align_special_tokens is a tokenizer/special-token edge case." + }, + { + "left": "issue:43504", + "right": "issue:43653", + "accept": false, + "reason": "A legacy-field load failure in BEiT is unrelated to BigBirdTokenizer mask-token registration and decode behavior." + }, + { + "left": "issue:43575", + "right": "issue:45405", + "accept": false, + "reason": "Qwen2-57B-A14B-Instruct TP OOM and MIN_PEFT_VERSION being bumped to an unreleased version are unrelated." + }, + { + "left": "issue:43927", + "right": "issue:44361", + "accept": false, + "reason": "DiaConfig losing custom token IDs after save/load is a config persistence bug; MLukeTokenizer failing on tasks is a tokenizer/task-specific AttributeError." + }, + { + "left": "issue:43278", + "right": "issue:43701", + "accept": false, + "reason": "BF16-to-FP32 embedding dtype drift and resume_from_checkpoint key mismatch are different training-state issues." + }, + { + "left": "issue:44792", + "right": "issue:45405", + "accept": false, + "reason": "A janus image-generation test failure is unrelated to a dependency version bump issue." + }, + { + "left": "issue:43606", + "right": "issue:44112", + "accept": false, + "reason": "Both are CI/device-related, but Bark CPU offload mismatch and GraniteSpeech stale device override are distinct bugs." + }, + { + "left": "issue:29942", + "right": "issue:30333", + "accept": false, + "reason": "Failing Flash Attention 2 tests and MLFlow not reporting failed jobs are unrelated." + }, + { + "left": "issue:15354", + "right": "issue:38175", + "accept": false, + "reason": "GeneratorExp/JIT scripting incompatibility is unrelated to SigLIP2 zero-probability outputs." + }, + { + "left": "issue:44554", + "right": "issue:45446", + "accept": false, + "reason": "An MPS attention correctness bug with mismatched head dims is unrelated to a PyTorch version check in flex_attention." + }, + { + "left": "issue:43540", + "right": "issue:45200", + "accept": false, + "reason": "Qwen3OmniMoe video-input validation and Gemma 4 mm_token_type_ids defaults are separate multimodal input bugs." + }, + { + "left": "issue:43232", + "right": "issue:44849", + "accept": false, + "reason": "A generation-state update bug after sync_gpus is not the same as Qwen3.5 output_hidden_states behavior." + }, + { + "left": "issue:45084", + "right": "issue:45520", + "accept": false, + "reason": "Template-node compilation failure and flash_attn import KeyError on Python 3.13 are unrelated failures." + }, + { + "left": "issue:43844", + "right": "issue:44964", + "accept": false, + "reason": "ZeRO-3 gradient growth with DeepSpeed is unrelated to Phi-4 multimodal model loading failures." + }, + { + "left": "issue:39401", + "right": "issue:44568", + "accept": false, + "reason": "Qwen3 tokenizer offset mapping is a tokenizer alignment bug; mdeberta add_special_tokens not adding BOS/EOS is a different special-token handling issue." + }, + { + "left": "issue:43582", + "right": "issue:44977", + "accept": false, + "reason": "Apple Silicon caching_allocator_warmup TypeError and Qwen3.5 flash-attention generation are different runtime bugs." + }, + { + "left": "issue:45072", + "right": "issue:45198", + "accept": false, + "reason": "Both are inference issues, but dtype mismatches in bfloat16 models and Wav2Vec2 save_pretrained/tokenization failures are not the same code path." + }, + { + "left": "issue:44315", + "right": "issue:44849", + "accept": false, + "reason": "Liger Kernel not applied with model_init is about trainer/model initialization hooks, not Qwen3.5 hidden-state outputs." + }, + { + "left": "issue:37428", + "right": "issue:43957", + "accept": false, + "reason": "Flash-support import failure and meta-device model-loading regressions are both loading-related but clearly separate errors." + }, + { + "left": "issue:44554", + "right": "issue:45356", + "accept": false, + "reason": "An MPS attention correctness issue is unrelated to Kimi-K2.5 tokenizer codec/reg-ex warnings." + }, + { + "left": "issue:43334", + "right": "issue:43525", + "accept": false, + "reason": "Qwen3-VL missing pad_token_id and Llama4Config missing pad_token_id are similar symptomatically, but they affect different model families and likely different config paths." + }, + { + "left": "issue:43295", + "right": "issue:44610", + "accept": false, + "reason": "Processor.tokenizer regression with images is unrelated to OmDet-Turbo processor image-size mismatch." + }, + { + "left": "issue:45200", + "right": "issue:45325", + "accept": false, + "reason": "Gemma 4 token-type-id defaults and Qwen2.5-VL rope-index scaling are different multimodal bookkeeping bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 20, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:37428", + "issue:38175", + "issue:42915", + "issue:43122", + "issue:43278", + "issue:43295", + "issue:43388", + "issue:43450", + "issue:43493", + "issue:43526", + "issue:43572", + "issue:43653", + "issue:43701", + "issue:43819", + "issue:43824", + "issue:43827", + "issue:43844", + "issue:43856", + "issue:43873", + "issue:43901", + "issue:43975", + "issue:43994", + "issue:44112", + "issue:44206", + "issue:44360", + "issue:44442", + "issue:44464", + "issue:44466", + "issue:44514", + "issue:44521", + "issue:44561", + "issue:44568", + "issue:44589", + "issue:44792", + "issue:44849", + "issue:44898", + "issue:44936", + "issue:44987", + "issue:44991", + "issue:45083", + "issue:45127", + "issue:45198", + "issue:45245", + "issue:45290", + "issue:45381", + "issue:45412", + "issue:45563", + "issue:45584" + ], + "soft_pairs": [ + "issue:44466|issue:45127", + "issue:44898|issue:45412", + "issue:43493|issue:43701", + "issue:43873|issue:44464", + "issue:44792|issue:45563", + "issue:44514|issue:44521", + "issue:37428|issue:43824", + "issue:43388|issue:43526", + "issue:43572|issue:44987", + "issue:43844|issue:44936", + "issue:43572|issue:44589", + "issue:43824|issue:43975", + "issue:43295|issue:43450", + "issue:43295|issue:43873", + "issue:44442|issue:45198", + "issue:44360|issue:44561", + "issue:38175|issue:45412", + "issue:43295|issue:43994", + "issue:43122|issue:44849", + "issue:44206|issue:45584", + "issue:42915|issue:43824", + "issue:43278|issue:43856", + "issue:38175|issue:43493", + "issue:43827|issue:43901", + "issue:44568|issue:44991", + "issue:43450|issue:43493", + "issue:43653|issue:44112", + "issue:45290|issue:45381", + "issue:45127|issue:45245", + "issue:43819|issue:45083" + ], + "prepared_review_unit_hash": "b47851ceb17d1d6c19f013525e704552c6d07a8d46652b1373a1677b13051bab", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36759, + "estimated_input_tokens": 9190, + "estimated_eval_tokens": 18636 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is a loose collection of unrelated Transformers issues rather than one duplicate cluster. The soft pairs mostly share broad subsystems or model families, but not the same concrete bug/change, so all soft edges should be rejected.", + "confidence": 0.97, + "canonical_issue_reason": "issue:43824 is the best canonical anchor here: it is a concrete, reproducible import regression with a clear failure mode and strong user-facing impact, and it has more discussion/reference signal than most others.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43824 is the strongest representative issue in this mixed set because it is self-contained, clearly actionable, and the most clearly described regression.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44466", + "right": "issue:45127", + "accept": false, + "reason": "Different bugs: tied-weight serialization inconsistency vs LoRA merge collapse after vocab extension." + }, + { + "left": "issue:44898", + "right": "issue:45412", + "accept": false, + "reason": "Different model paths and symptoms: Perceiver resolution failure vs RT-DETR memory leak." + }, + { + "left": "issue:43493", + "right": "issue:43701", + "accept": false, + "reason": "Same repo area but unrelated failures: SigLIP2 output mismatch vs resume checkpoint key mismatch." + }, + { + "left": "issue:43873", + "right": "issue:44464", + "accept": false, + "reason": "Offloading/quantization behavior is unrelated to compiled-forward chunked generation inconsistency." + }, + { + "left": "issue:44792", + "right": "issue:45563", + "accept": false, + "reason": "Janus test failure and stale generate warning are different issues and likely different code paths." + }, + { + "left": "issue:44514", + "right": "issue:44521", + "accept": false, + "reason": "Both touch apply_chat_template, but one crashes on batched padding=False while the other returns all-zero assistant masks." + }, + { + "left": "issue:37428", + "right": "issue:43824", + "accept": false, + "reason": "Both are ImportErrors, but for different missing symbols and different code paths." + }, + { + "left": "issue:43388", + "right": "issue:43526", + "accept": false, + "reason": "Trainer metric gathering bug and BeitImageProcessorFast label reduction bug are distinct." + }, + { + "left": "issue:43572", + "right": "issue:44987", + "accept": false, + "reason": "Missing config field after v5 update is not the same as a model-loading failure for a remote model." + }, + { + "left": "issue:43844", + "right": "issue:44936", + "accept": false, + "reason": "ZeRO-3 gradient growth and trainer.evaluate() failing after train are different training behaviors." + }, + { + "left": "issue:43572", + "right": "issue:44589", + "accept": false, + "reason": "Config API regression vs Float8 storage lookup error; no shared concrete bug." + }, + { + "left": "issue:43824", + "right": "issue:43975", + "accept": false, + "reason": "Both involve v5 regressions, but one is an import failure and the other is incorrect detokenization." + }, + { + "left": "issue:43295", + "right": "issue:43450", + "accept": false, + "reason": "Custom processor/tokenizer crash and batched video shape bug are different multimodal regressions." + }, + { + "left": "issue:43295", + "right": "issue:43873", + "accept": false, + "reason": "Processor API regression is unrelated to quantization/offloading behavior." + }, + { + "left": "issue:44442", + "right": "issue:45198", + "accept": false, + "reason": "Tokenizer load failure and save/tokenization failure are separate tokenizer issues." + }, + { + "left": "issue:44360", + "right": "issue:44561", + "accept": false, + "reason": "DSA indexing discussion is unrelated to removal of is_torch_fx_available breaking trust_remote_code." + }, + { + "left": "issue:38175", + "right": "issue:45412", + "accept": false, + "reason": "SigLIP2 inference anomaly and RT-DETR memory reclamation bug are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:43994", + "accept": false, + "reason": "Different models and failures: custom multimodal processor regression vs SigLIP2 nonsensical outputs." + }, + { + "left": "issue:43122", + "right": "issue:44849", + "accept": false, + "reason": "Tokenizer version drift and Qwen3.5 hidden-state bug are unrelated." + }, + { + "left": "issue:44206", + "right": "issue:45584", + "accept": false, + "reason": "Unsupported feature-extractor arg crash and Whisper empty-transcription failure are different bugs." + }, + { + "left": "issue:42915", + "right": "issue:43824", + "accept": false, + "reason": "Qwen3Moe FP8 config failure is unrelated to a missing exported VL class." + }, + { + "left": "issue:43278", + "right": "issue:43856", + "accept": false, + "reason": "Eval dtype change and MoE memory usage are different training-time issues." + }, + { + "left": "issue:38175", + "right": "issue:43493", + "accept": false, + "reason": "Same model family, but zero probabilities and JAX parity discrepancy are not the same concrete failure." + }, + { + "left": "issue:43827", + "right": "issue:43901", + "accept": false, + "reason": "Both are docs-only regressions, but they refer to different pages and different outdated API mentions." + }, + { + "left": "issue:44568", + "right": "issue:44991", + "accept": false, + "reason": "Special-token handling bug and tokenizer loading failure are distinct." + }, + { + "left": "issue:43450", + "right": "issue:43493", + "accept": false, + "reason": "Batched video shape bug is unrelated to SigLIP2 implementation discrepancy." + }, + { + "left": "issue:43653", + "right": "issue:44112", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and GraniteSpeech CI device override are unrelated." + }, + { + "left": "issue:45290", + "right": "issue:45381", + "accept": false, + "reason": "Chat-template crash with tool-call assistant messages is unrelated to Qwen2.5-VL video vision_position_ids." + }, + { + "left": "issue:45127", + "right": "issue:45245", + "accept": false, + "reason": "LoRA/tied-weights collapse is unrelated to the 2^24 category limit runtime error." + }, + { + "left": "issue:43819", + "right": "issue:45083", + "accept": false, + "reason": "DAC latent/forward mismatch and qwen3_omni_moe feature-length helper behavior are different code paths." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 21, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "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:43723", + "issue:43819", + "issue:43825", + "issue:43931", + "issue:44112", + "issue:44292", + "issue:44351", + "issue:44448", + "issue:44462", + "issue:44493", + "issue:44554", + "issue:44561", + "issue:44568", + "issue:44610", + "issue:44743", + "issue:44779", + "issue:44855", + "issue:44857", + "issue:44977", + "issue:45072", + "issue:45083", + "issue:45200", + "issue:45229", + "issue:45310", + "issue:45381", + "issue:45478", + "issue:45491", + "issue:45520", + "issue:45588", + "issue:45593" + ], + "soft_pairs": [ + "issue:44462|issue:44561", + "issue:44561|issue:44779", + "issue:45200|issue:45588", + "issue:43450|issue:43575", + "issue:44610|issue:44857", + "issue:44857|issue:45593", + "issue:37428|issue:43425", + "issue:43262|issue:43450", + "issue:44448|issue:44743", + "issue:43819|issue:44112", + "issue:45310|issue:45478", + "issue:43606|issue:45083", + "issue:43278|issue:43825", + "issue:29942|issue:32090", + "issue:37428|issue:44351", + "issue:39692|issue:43526", + "issue:43611|issue:43931", + "issue:43335|issue:43550", + "issue:43819|issue:44493", + "issue:45200|issue:45381", + "issue:43450|issue:43526", + "issue:39692|issue:44493", + "issue:44855|issue:45083", + "issue:44448|issue:44779", + "issue:29127|issue:32090", + "issue:43335|issue:43653", + "issue:38175|issue:44493", + "issue:44977|issue:45229", + "issue:44292|issue:45520", + "issue:43335|issue:45083", + "issue:43335|issue:43504", + "issue:43550|issue:44112", + "issue:44568|issue:45491", + "issue:43335|issue:45072", + "issue:43723|issue:44554" + ], + "prepared_review_unit_hash": "135b3d7fbdc3571120a3e7aefe95004d0a76a1e1ff2c72366e962c43fe381b34", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 35, + "serialized_chars": 38239, + "estimated_input_tokens": 9560, + "estimated_eval_tokens": 19376 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly heterogeneous v5/model-specific regressions that should not be merged. I found one clear duplicate pair: the two identical Qwen3.5 MoE `from_pretrained` error reports. No PRs are present in the cluster.", + "confidence": 0.78, + "canonical_issue_reason": "issue:45310 is the clearest canonical issue because it is a concrete, well-scoped bug report and has an exact duplicate in the cluster (`issue:45478`) with the same title/failure mode.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45310 is the best representative issue in this cluster: it is specific, reproducible, and duplicated verbatim by another issue, whereas most other items are unrelated regressions or generic v5 breakages.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44462", + "right": "issue:44561", + "accept": false, + "reason": "Both are v5-era failures, but one is tokenizer file selection and the other is a removed utility breaking trust_remote_code model loading; different root causes." + }, + { + "left": "issue:44561", + "right": "issue:44779", + "accept": false, + "reason": "Remote-code loading breakage and DeepSeek tokenizer output regression are separate tokenizer/model-loading bugs." + }, + { + "left": "issue:45200", + "right": "issue:45588", + "accept": false, + "reason": "Gemma4 text-only token-id defaults and flash-attention sink-less crash are unrelated code paths." + }, + { + "left": "issue:43450", + "right": "issue:43575", + "accept": false, + "reason": "Video processor batching shape bug and Qwen TP OOM are different subsystems and failure modes." + }, + { + "left": "issue:44610", + "right": "issue:44857", + "accept": false, + "reason": "OmDet preprocessing size mismatch and LwDetr AMP/CUDA loss crash are different vision bugs." + }, + { + "left": "issue:44857", + "right": "issue:45593", + "accept": false, + "reason": "Float16 loss crash and missing auxiliary losses when denoising is off are unrelated training issues." + }, + { + "left": "issue:37428", + "right": "issue:43425", + "accept": false, + "reason": "A missing flash-attention import symbol is not the same concrete bug as broad Torch 2.10 incompatibility." + }, + { + "left": "issue:43262", + "right": "issue:43450", + "accept": false, + "reason": "Audio chat-template sampling-rate defaults and video batch-shape handling are different processor bugs." + }, + { + "left": "issue:44448", + "right": "issue:44743", + "accept": false, + "reason": "Pegasus v4/v5 output drift and Qwen3.5 recurrent-state reset are unrelated model-specific regressions." + }, + { + "left": "issue:43819", + "right": "issue:44112", + "accept": false, + "reason": "DAC forward/from_latents mismatch and GraniteSpeech CI device-override test failure are not the same bug." + }, + { + "left": "issue:45310", + "right": "issue:45478", + "accept": true, + "reason": "Same title, same model family, same `from_pretrained` failure; this is a clear duplicate report." + }, + { + "left": "issue:43606", + "right": "issue:45083", + "accept": false, + "reason": "CPU offload device mismatch in Bark and `_get_feat_extract_output_lengths` behavior in qwen3_omni_moe are unrelated." + }, + { + "left": "issue:43278", + "right": "issue:43825", + "accept": false, + "reason": "BF16-to-FP32 eval dtype drift and `pipeline()` translation-task messaging are different issues." + }, + { + "left": "issue:29942", + "right": "issue:32090", + "accept": false, + "reason": "Flash Attention 2 test failures and `_gpu_broadcast_one` NoneType TypeError are different failures." + }, + { + "left": "issue:37428", + "right": "issue:44351", + "accept": false, + "reason": "These are both import errors, but for different missing symbols and different breakages." + }, + { + "left": "issue:39692", + "right": "issue:43526", + "accept": false, + "reason": "Documentation/example errors for SigLIP2 are unrelated to BeitImageProcessorFast `reduce_labels` behavior." + }, + { + "left": "issue:43611", + "right": "issue:43931", + "accept": false, + "reason": "`base_model_prefix` loading regression and Qwen3-VL weight-shape mismatch are distinct model-loading problems." + }, + { + "left": "issue:43335", + "right": "issue:43550", + "accept": false, + "reason": "SwitchTransformers sparse-layer config bug and Bamba `torch.compile`/SDPA failure are different code paths." + }, + { + "left": "issue:43819", + "right": "issue:44493", + "accept": false, + "reason": "DAC latent conversion mismatch and widespread unexpected position-id keys are unrelated." + }, + { + "left": "issue:45200", + "right": "issue:45381", + "accept": false, + "reason": "Both concern multimodal ids, but Gemma4 `mm_token_type_ids` defaults and Qwen2.5-VL video position ids are separate bugs." + }, + { + "left": "issue:43450", + "right": "issue:43526", + "accept": false, + "reason": "Video batching shape handling and Beit fast processor `reduce_labels` are different processor defects." + }, + { + "left": "issue:39692", + "right": "issue:44493", + "accept": false, + "reason": "A SigLIP2 docs/example report is not the same as a generic position-id key regression across many models." + }, + { + "left": "issue:44855", + "right": "issue:45083", + "accept": false, + "reason": "Python 3.13 `@torch.jit.script` parsing/indentation failure and qwen3_omni_moe helper behavior are unrelated." + }, + { + "left": "issue:44448", + "right": "issue:44779", + "accept": false, + "reason": "Both mention v4/v5 output changes, but Pegasus generation drift and DeepSeek tokenizer regression are different underlying bugs." + }, + { + "left": "issue:29127", + "right": "issue:32090", + "accept": false, + "reason": "LayoutLMv3 error-message clarity and Trainer `_gpu_broadcast_one` NoneType crash are unrelated." + }, + { + "left": "issue:43335", + "right": "issue:43653", + "accept": false, + "reason": "SwitchTransformers config and BigBirdTokenizer special-token registration are different model/tokenizer bugs." + }, + { + "left": "issue:38175", + "right": "issue:44493", + "accept": false, + "reason": "SigLIP2 zero-probabilities and position-id key warnings are different symptoms with different likely causes." + }, + { + "left": "issue:44977", + "right": "issue:45229", + "accept": false, + "reason": "Qwen3.5 flash-attention generation bug and Gemma4 multi-GPU OOM are unrelated." + }, + { + "left": "issue:44292", + "right": "issue:45520", + "accept": false, + "reason": "Qwen-3-8B-NVFP4 runtime failure and Python 3.13 flash_attn import key error are different problems." + }, + { + "left": "issue:43335", + "right": "issue:45083", + "accept": false, + "reason": "SwitchTransformers sparse-layer config and qwen3_omni_moe feature-length helper behavior are not the same bug." + }, + { + "left": "issue:43335", + "right": "issue:43504", + "accept": false, + "reason": "Sparse-layer creation in SwitchTransformers and BEiT legacy-field loading failure are unrelated." + }, + { + "left": "issue:43550", + "right": "issue:44112", + "accept": false, + "reason": "Bamba SDPA/compile failure and GraniteSpeech stale CI test are not the same underlying issue." + }, + { + "left": "issue:44568", + "right": "issue:45491", + "accept": false, + "reason": "Tokenizer special-token insertion and Gemma3 NaN embeddings under sliding-window attention are different issues." + }, + { + "left": "issue:43723", + "right": "issue:44554", + "accept": false, + "reason": "AutoTokenizer loading regression and MPS attention correctness issue are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: the cluster is indeed issue-only, and the one duplicate pair (issue:45310 / issue:45478) is a clear match by title and failure mode. The soft-edge rejections are conservative and do not overclaim equivalence, so I don\u2019t see evidence of an overstated mergeability verdict." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 22, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:25251", + "issue:29942", + "issue:32090", + "issue:33357", + "issue:34689", + "issue:35141", + "issue:37428", + "issue:40444", + "issue:41720", + "issue:43065", + "issue:43066", + "issue:43278", + "issue:43329", + "issue:43335", + "issue:43412", + "issue:43526", + "issue:43550", + "issue:43606", + "issue:43653", + "issue:43701", + "issue:43720", + "issue:43873", + "issue:43976", + "issue:43994", + "issue:44112", + "issue:44448", + "issue:44451", + "issue:44462", + "issue:44466", + "issue:44493", + "issue:44743", + "issue:44857", + "issue:44977", + "issue:44987", + "issue:45072", + "issue:45083", + "issue:45229", + "issue:45245", + "issue:45372", + "issue:45406", + "issue:45412", + "issue:45440", + "issue:45459", + "issue:45520", + "issue:45563", + "issue:45588", + "issue:45593" + ], + "soft_pairs": [ + "issue:40444|issue:43526", + "issue:44448|issue:44977", + "issue:43335|issue:45593", + "issue:43720|issue:44112", + "issue:43066|issue:44462", + "issue:45372|issue:45588", + "issue:34689|issue:35141", + "issue:44977|issue:45563", + "issue:44857|issue:45072", + "issue:43606|issue:44857", + "issue:43412|issue:45588", + "issue:43873|issue:45229", + "issue:37428|issue:44987", + "issue:43873|issue:44493", + "issue:45372|issue:45406", + "issue:43701|issue:45083", + "issue:43976|issue:45440", + "issue:29942|issue:35141", + "issue:45245|issue:45563", + "issue:43550|issue:45083", + "issue:41720|issue:45229", + "issue:43278|issue:43329", + "issue:32090|issue:33357", + "issue:45372|issue:45459", + "issue:44451|issue:45520", + "issue:44448|issue:44466", + "issue:43065|issue:43994", + "issue:43065|issue:45412", + "issue:25251|issue:32090", + "issue:43653|issue:44743", + "issue:44857|issue:45588" + ], + "prepared_review_unit_hash": "784d466bd7b168a4f8d2a829a66036c8814c72bd2546f9974e025738aeea3615", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36492, + "estimated_input_tokens": 9123, + "estimated_eval_tokens": 18502 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous: the soft pairs span unrelated bugs across loading, tokenization, multimodal processing, flash-attention, offloading, and CI. I rejected every soft edge as a duplicate false positive. With no PRs present, I selected the broadest active issue as the best issue representative.", + "confidence": 0.69, + "canonical_issue_reason": "Issue 43873 is the broadest and most active open report here, with multiple inbound references and discussion around offloading/quantization behavior, making it the least specific umbrella representative among these unrelated issues.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43873 is the best available issue candidate because it is open, actively discussed, and broadly scoped compared with the other more niche or closed reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:40444", + "right": "issue:43526", + "accept": false, + "reason": "Different subsystems and failures: Qwen2.5-VL iterable dataset finetuning vs BEiT fast image processor label reduction." + }, + { + "left": "issue:44448", + "right": "issue:44977", + "accept": false, + "reason": "Both mention output differences, but they concern different models and code paths: Pegasus v4/v5 mismatch vs Qwen3.5 flash-attention generation." + }, + { + "left": "issue:43335", + "right": "issue:45593", + "accept": false, + "reason": "Unrelated bugs in different models: SwitchTransformers sparse-layer config vs D-FINE auxiliary losses." + }, + { + "left": "issue:43720", + "right": "issue:44112", + "accept": false, + "reason": "Packed-weight loading in BitNet is unrelated to a stale GraniteSpeech CI device-override test." + }, + { + "left": "issue:43066", + "right": "issue:44462", + "accept": false, + "reason": "Both are tokenizer-related, but one is a decoder-type regression and the other is AutoTokenizer ignoring tokenizer.json; different underlying bugs." + }, + { + "left": "issue:45372", + "right": "issue:45588", + "accept": false, + "reason": "Gemma 4 processor import failure and flash-attention s_aux=None crash are different failure modes in different code paths." + }, + { + "left": "issue:34689", + "right": "issue:35141", + "accept": false, + "reason": "Model loading failure for Llama 3.2 Vision is unrelated to post_init reinitializing embeddings after resize." + }, + { + "left": "issue:44977", + "right": "issue:45563", + "accept": false, + "reason": "Qwen3.5 flash-attention generation bug is unrelated to a stale warning about num_return_sequences." + }, + { + "left": "issue:44857", + "right": "issue:45072", + "accept": false, + "reason": "Different numeric/runtime issues in different models: LwDetrImageLoss AMP crash vs dtype mismatches in SwitchTransformers/TimmWrapperModel." + }, + { + "left": "issue:43606", + "right": "issue:44857", + "accept": false, + "reason": "CPU offload device mismatch for bark-small is not the same bug as the LwDetrImageLoss float16 CUDA crash." + }, + { + "left": "issue:43412", + "right": "issue:45588", + "accept": false, + "reason": "Gemma3n ExecuTorch export problems are unrelated to an s_aux=None flash-attention crash." + }, + { + "left": "issue:43873", + "right": "issue:45229", + "accept": false, + "reason": "Quantization/offloading behavior and Gemma4 multi-GPU OOM are different issues, not one concrete code-path bug." + }, + { + "left": "issue:37428", + "right": "issue:44987", + "accept": false, + "reason": "The missing internal flash-attention import and loading physical-intelligence/fast may be adjacent, but the reports do not describe the same concrete bug." + }, + { + "left": "issue:43873", + "right": "issue:44493", + "accept": false, + "reason": "Offloading/quantization behavior is unrelated to unexpected position-id keys during model loading." + }, + { + "left": "issue:45372", + "right": "issue:45406", + "accept": false, + "reason": "Both involve Gemma 4, but one is an external import failure in processor loading and the other is a missing _tokenizer attribute in serve; different root causes." + }, + { + "left": "issue:43701", + "right": "issue:45083", + "accept": false, + "reason": "Checkpoint key mismatch is unrelated to the qwen3_omni_moe feature-extraction length helper bug." + }, + { + "left": "issue:43976", + "right": "issue:45440", + "accept": false, + "reason": "Python version compatibility regression is unrelated to DeepseekV3MoE divergence from the remote implementation." + }, + { + "left": "issue:29942", + "right": "issue:35141", + "accept": false, + "reason": "Flash Attention 2 test failures are unrelated to embedding reinitialization after token resize." + }, + { + "left": "issue:45245", + "right": "issue:45563", + "accept": false, + "reason": "Category-count runtime limit and a stale warning in paged generate() are not the same underlying problem." + }, + { + "left": "issue:43550", + "right": "issue:45083", + "accept": false, + "reason": "Bamba SDPA torch.compile failure is unrelated to the qwen3_omni_moe feature-length helper behavior." + }, + { + "left": "issue:41720", + "right": "issue:45229", + "accept": false, + "reason": "Qwen3 auto device-mapping cudaErrorAssert and Gemma4 multi-GPU OOM are both large-model runtime issues, but not the same bug." + }, + { + "left": "issue:43278", + "right": "issue:43329", + "accept": false, + "reason": "Embedding dtype drift during evaluation is unrelated to undefined variables in multimodal token counting." + }, + { + "left": "issue:32090", + "right": "issue:33357", + "accept": false, + "reason": "Trainer broadcast TypeError and MacOS bus error on CLIP are unrelated failures." + }, + { + "left": "issue:45372", + "right": "issue:45459", + "accept": false, + "reason": "Gemma 4 processor import failure and tokenizer error masking from protobuf absence are different issues." + }, + { + "left": "issue:44451", + "right": "issue:45520", + "accept": false, + "reason": "Loading a specific model and a Python 3.13 flash_attn KeyError are unrelated." + }, + { + "left": "issue:44448", + "right": "issue:44466", + "accept": false, + "reason": "Both involve version-dependent behavior, but Pegasus output differences and lm_head serialization inconsistency are different bugs." + }, + { + "left": "issue:43065", + "right": "issue:43994", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder is unrelated to SigLIP2 nonsensical AutoModel/pipeline outputs." + }, + { + "left": "issue:43065", + "right": "issue:45412", + "accept": false, + "reason": "A placeholder Conv2d in Sam3PixelDecoder is unrelated to RT-DETR memory not being released." + }, + { + "left": "issue:25251", + "right": "issue:32090", + "accept": false, + "reason": "Pipeline top_k nesting behavior and a Trainer broadcast TypeError have no common underlying bug." + }, + { + "left": "issue:43653", + "right": "issue:44743", + "accept": false, + "reason": "BigBirdTokenizer special-token decode bug is unrelated to recurrent state reset in qwen3_5 modular code." + }, + { + "left": "issue:44857", + "right": "issue:45588", + "accept": false, + "reason": "A CUDA AMP loss crash and an s_aux=None flash-attention crash are different model/runtime failures." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 23, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:29127", + "issue:29942", + "issue:30064", + "issue:32090", + "issue:33290", + "issue:33357", + "issue:34567", + "issue:35141", + "issue:36010", + "issue:39692", + "issue:42915", + "issue:42994", + "issue:43257", + "issue:43262", + "issue:43278", + "issue:43334", + "issue:43381", + "issue:43421", + "issue:43526", + "issue:43531", + "issue:43653", + "issue:43701", + "issue:43756", + "issue:43761", + "issue:43824", + "issue:43994", + "issue:44079", + "issue:44186", + "issue:44206", + "issue:44265", + "issue:44279", + "issue:44361", + "issue:44442", + "issue:44464", + "issue:44479", + "issue:44610", + "issue:44871", + "issue:44977", + "issue:45072", + "issue:45103", + "issue:45198", + "issue:45229", + "issue:45245", + "issue:45372", + "issue:45412", + "issue:45440", + "issue:45491", + "issue:45593" + ], + "soft_pairs": [ + "issue:44186|issue:45198", + "issue:43421|issue:45491", + "issue:44265|issue:45198", + "issue:43278|issue:43381", + "issue:43994|issue:45412", + "issue:42915|issue:42994", + "issue:44361|issue:45198", + "issue:44206|issue:44479", + "issue:44871|issue:44977", + "issue:43756|issue:45440", + "issue:43257|issue:43278", + "issue:44610|issue:45412", + "issue:29942|issue:30064", + "issue:45103|issue:45372", + "issue:43824|issue:44279", + "issue:33357|issue:39692", + "issue:45440|issue:45593", + "issue:29127|issue:34567", + "issue:32090|issue:33290", + "issue:45072|issue:45229", + "issue:43653|issue:44442", + "issue:43381|issue:43701", + "issue:43526|issue:44186", + "issue:44464|issue:45491", + "issue:43334|issue:43531", + "issue:44977|issue:45245", + "issue:43262|issue:44206", + "issue:29942|issue:36010", + "issue:29127|issue:33357", + "issue:43761|issue:44079", + "issue:29127|issue:35141" + ], + "prepared_review_unit_hash": "77edb973a1351d1e6c94b2b5aba3a2b1fac90dbb895e76241359e584b61db6d1", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37064, + "estimated_input_tokens": 9266, + "estimated_eval_tokens": 18788 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These issues are largely unrelated Hugging Face Transformers bug reports across different subsystems; I don\u2019t see any convincing duplicate pairs, so no soft edges should be merged.", + "confidence": 0.99, + "canonical_issue_reason": "issue:43824 has the strongest impact signal in this set (6 inbound references) and is a clear, self-contained reproducible import failure.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43824 is the best cluster anchor because it is concrete, externally referenced, and easy to identify as a distinct bug report.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44186", + "right": "issue:45198", + "accept": false, + "reason": "Different models and failure modes: LayoutLMv2 tokenizer padding/NER crash vs Wav2Vec2 save/tokenization failure." + }, + { + "left": "issue:43421", + "right": "issue:45491", + "accept": false, + "reason": "Tokenizer post-processor update behavior is unrelated to Gemma3 NaNs from sliding-window/all-padding attention." + }, + { + "left": "issue:44265", + "right": "issue:45198", + "accept": false, + "reason": "torch.export / torch_compilable_check failure is a different code path from Wav2Vec2 tokenization/save_pretrained issues." + }, + { + "left": "issue:43278", + "right": "issue:43381", + "accept": false, + "reason": "Embedding dtype drift in eval is unrelated to the gradient-checkpointing eval-mode limitation." + }, + { + "left": "issue:43994", + "right": "issue:45412", + "accept": false, + "reason": "SigLIP2 bad outputs and RT-DETR memory release are different bugs in different subsystems." + }, + { + "left": "issue:42915", + "right": "issue:42994", + "accept": false, + "reason": "Qwen3Moe FP8 config failure is not the same as generic quantized model saving failure." + }, + { + "left": "issue:44361", + "right": "issue:45198", + "accept": false, + "reason": "MLukeTokenizer AttributeError is unrelated to Wav2Vec2 save/tokenization behavior." + }, + { + "left": "issue:44206", + "right": "issue:44479", + "accept": false, + "reason": "Unsupported feature arg in LasrFeatureExtractor is a different regression from Qwen video-input behavior." + }, + { + "left": "issue:44871", + "right": "issue:44977", + "accept": false, + "reason": "Gemma eos-token config mismatch is unrelated to Qwen3.5 flash-attention generation bugs." + }, + { + "left": "issue:43756", + "right": "issue:45440", + "accept": false, + "reason": "Smollm3 RoPE-layer mismatch and DeepseekV3MoE divergence are separate model-implementation issues." + }, + { + "left": "issue:43257", + "right": "issue:43278", + "accept": false, + "reason": "Qwen3 MoE weight conversion under accelerate+deepspeed is unrelated to training/eval embedding dtype changes." + }, + { + "left": "issue:44610", + "right": "issue:45412", + "accept": false, + "reason": "Processor image-size mismatch is not the same as RT-DETR memory not being released on delete." + }, + { + "left": "issue:29942", + "right": "issue:30064", + "accept": false, + "reason": "Flash Attention 2 test failures and void segmentation map processing are different failures in different areas." + }, + { + "left": "issue:45103", + "right": "issue:45372", + "accept": false, + "reason": "auto_docstring future-annotations crash is unrelated to the mistral_common ReasoningEffort import error." + }, + { + "left": "issue:43824", + "right": "issue:44279", + "accept": false, + "reason": "Missing Qwen2_5_VLForConditionalGeneration import is a specific API/export bug, not a generic dependency issue." + }, + { + "left": "issue:33357", + "right": "issue:39692", + "accept": false, + "reason": "MacOS CLIP bus error and SigLIP2 doc-example errors are different models and failure modes." + }, + { + "left": "issue:45440", + "right": "issue:45593", + "accept": false, + "reason": "DeepseekV3MoE implementation divergence is unrelated to D-FINE auxiliary-loss behavior." + }, + { + "left": "issue:29127", + "right": "issue:34567", + "accept": false, + "reason": "LayoutLMv3 error-message clarity and TrainerState token-count tracking are unrelated bugs." + }, + { + "left": "issue:32090", + "right": "issue:33290", + "accept": false, + "reason": "Trainer NoneType broadcast failure is unrelated to DeepSpeed AdaFactor OOM." + }, + { + "left": "issue:45072", + "right": "issue:45229", + "accept": false, + "reason": "bfloat16 dtype mismatch in inference is not the same as Gemma4 multi-GPU OOM." + }, + { + "left": "issue:43653", + "right": "issue:44442", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and AutoTokenizer loading FastSpeech2ConformerTokenizer are different tokenizer issues." + }, + { + "left": "issue:43381", + "right": "issue:43701", + "accept": false, + "reason": "Gradient checkpointing in eval mode is unrelated to resume-from-checkpoint key mismatch." + }, + { + "left": "issue:43526", + "right": "issue:44186", + "accept": false, + "reason": "BeitImageProcessorFast reduce_labels bug is unrelated to LayoutLMv2Tokenizer crashes on padding/truncation." + }, + { + "left": "issue:44464", + "right": "issue:45491", + "accept": false, + "reason": "Chunked generation with compiled forward is a different path from Gemma3 NaNs in mixed-length batching." + }, + { + "left": "issue:43334", + "right": "issue:43531", + "accept": false, + "reason": "Qwen3-VL pad_token_id loading error and Qwen3-MoE sliding_window behavior are separate issues." + }, + { + "left": "issue:44977", + "right": "issue:45245", + "accept": false, + "reason": "Flash-attention generation regression is unrelated to the category-count limit runtime error." + }, + { + "left": "issue:43262", + "right": "issue:44206", + "accept": false, + "reason": "Audio chat-template sampling-rate default is unrelated to LasrFeatureExtractor's unsupported center arg." + }, + { + "left": "issue:29942", + "right": "issue:36010", + "accept": false, + "reason": "Flash Attention test failures are unrelated to the GenerationMixin import error." + }, + { + "left": "issue:29127", + "right": "issue:33357", + "accept": false, + "reason": "LayoutLMv3 error messaging and CLIP MacOS bus errors are not the same underlying bug." + }, + { + "left": "issue:29127", + "right": "issue:35141", + "accept": false, + "reason": "LayoutLMv3 error clarity and output-embedding reinitialization are different bugs in different code paths." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 24, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:33290", + "issue:34567", + "issue:34689", + "issue:39692", + "issue:41628", + "issue:42915", + "issue:43232", + "issue:43316", + "issue:43388", + "issue:43441", + "issue:43452", + "issue:43493", + "issue:43525", + "issue:43526", + "issue:43575", + "issue:43611", + "issue:43646", + "issue:43673", + "issue:43701", + "issue:43723", + "issue:43761", + "issue:43827", + "issue:43873", + "issue:43931", + "issue:43994", + "issue:44077", + "issue:44188", + "issue:44206", + "issue:44387", + "issue:44442", + "issue:44479", + "issue:44610", + "issue:44743", + "issue:44811", + "issue:44857", + "issue:44898", + "issue:44977", + "issue:45003", + "issue:45072", + "issue:45216", + "issue:45290", + "issue:45372", + "issue:45440", + "issue:45479", + "issue:45491", + "issue:45584", + "issue:45588", + "issue:45600" + ], + "soft_pairs": [ + "issue:43441|issue:45588", + "issue:42915|issue:43931", + "issue:43232|issue:43388", + "issue:43526|issue:44857", + "issue:44442|issue:45072", + "issue:43723|issue:45479", + "issue:44479|issue:45290", + "issue:45003|issue:45440", + "issue:43526|issue:44610", + "issue:44743|issue:44811", + "issue:43452|issue:45588", + "issue:43931|issue:45216", + "issue:44977|issue:45584", + "issue:45588|issue:45600", + "issue:43316|issue:43525", + "issue:43761|issue:44188", + "issue:43646|issue:44077", + "issue:43575|issue:43701", + "issue:43994|issue:44206", + "issue:39692|issue:45491", + "issue:43575|issue:44387", + "issue:43873|issue:45588", + "issue:43673|issue:44743", + "issue:33290|issue:34567", + "issue:44442|issue:44898", + "issue:34689|issue:43611", + "issue:41628|issue:45372", + "issue:43493|issue:43827" + ], + "prepared_review_unit_hash": "d7b00247e999fcfa3808e320b6e13dcfc832c8fa7e45cc8eafbd386f301db5f3", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36198, + "estimated_input_tokens": 9050, + "estimated_eval_tokens": 18356 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The soft pairs mostly span unrelated regressions in different subsystems or models, so I rejected all of them; the only plausible hub is the flash-attention-related issue around `s_aux=None`, but even that is not the same concrete bug as the paired reports.", + "confidence": 0.77, + "canonical_issue_reason": "issue:45588 is the best central anchor in this set: it is open and has multiple flash-attention-related soft links, making it the clearest hub among the provided issues.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45588 is the strongest issue candidate overall because it is the most connected and most general among the soft-linked flash-attention reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43441", + "right": "issue:45588", + "accept": false, + "reason": "Both involve FlashAttention, but they describe different failure modes and likely different fixes: Ministral-3 RC regression vs `s_aux=None` on sink-less models." + }, + { + "left": "issue:42915", + "right": "issue:43931", + "accept": false, + "reason": "Different models and different errors: FineGrainedFP8Config failure vs weight-shape mismatch on Qwen3-VL loading." + }, + { + "left": "issue:43232", + "right": "issue:43388", + "accept": false, + "reason": "Unrelated code paths: generation kwargs after `sync_gpus` vs metric gathering dropping tuple labels." + }, + { + "left": "issue:43526", + "right": "issue:44857", + "accept": false, + "reason": "Different components and symptoms: BeitImageProcessorFast label reduction bug vs LwDetrImageLoss AMP crash." + }, + { + "left": "issue:44442", + "right": "issue:45072", + "accept": false, + "reason": "Tokenizer loading failure and dtype mismatch are unrelated bugs." + }, + { + "left": "issue:43723", + "right": "issue:45479", + "accept": false, + "reason": "Tokenizer loading in v5 and zero-loss classification behavior are not the same issue." + }, + { + "left": "issue:44479", + "right": "issue:45290", + "accept": false, + "reason": "Video-input regression for Qwen VL models is separate from chat-template tool-call handling." + }, + { + "left": "issue:45003", + "right": "issue:45440", + "accept": false, + "reason": "`sys.modules` access in modeling utils is unrelated to DeepSeekV3MoE divergence from remote implementation." + }, + { + "left": "issue:43526", + "right": "issue:44610", + "accept": false, + "reason": "Both are model/processor-related, but one is label reduction and the other is input-size mismatch." + }, + { + "left": "issue:44743", + "right": "issue:44811", + "accept": false, + "reason": "Cache/reset behavior in recurrent states is unrelated to Whisper batch decode special-token handling." + }, + { + "left": "issue:43452", + "right": "issue:45588", + "accept": false, + "reason": "Tokenizer/model loading with `gguf_file` is a different failure mode than FlashAttention `s_aux=None`." + }, + { + "left": "issue:43931", + "right": "issue:45216", + "accept": false, + "reason": "Both concern Qwen models, but one is a loading shape mismatch and the other is a broken `save_pretrained` checkpoint." + }, + { + "left": "issue:44977", + "right": "issue:45584", + "accept": false, + "reason": "Qwen3.5 FlashAttention generation regression and Whisper empty-transcription generation failure are unrelated." + }, + { + "left": "issue:45588", + "right": "issue:45600", + "accept": false, + "reason": "FlashAttention `s_aux=None` crash is unrelated to removed `Sam3LiteText` references in auto mappings." + }, + { + "left": "issue:43316", + "right": "issue:43525", + "accept": false, + "reason": "Different config API mismatches: Gemma3TextConfig discrepancy vs missing `pad_token_id` on Llama4Config." + }, + { + "left": "issue:43761", + "right": "issue:44188", + "accept": false, + "reason": "CLIPVision hidden-state regression and attention-kernel divergence under `torch.compile` are separate bugs." + }, + { + "left": "issue:43646", + "right": "issue:44077", + "accept": false, + "reason": "Generic custom-model init breakage is not the same as the `patchtsmixer` `post_init` constraint." + }, + { + "left": "issue:43575", + "right": "issue:43701", + "accept": false, + "reason": "Tensor-parallel OOM and checkpoint key mismatch are unrelated." + }, + { + "left": "issue:43994", + "right": "issue:44206", + "accept": false, + "reason": "SigLIP2 wrong outputs and LastFeatureExtractor center-arg crash are different problems." + }, + { + "left": "issue:39692", + "right": "issue:45491", + "accept": false, + "reason": "SigLIP2 docs/example errors are unrelated to Gemma3 mixed-length NaN embeddings." + }, + { + "left": "issue:43575", + "right": "issue:44387", + "accept": false, + "reason": "Both mention memory, but one is model loading OOM and the other is int4 reserved-memory growth." + }, + { + "left": "issue:43873", + "right": "issue:45588", + "accept": false, + "reason": "Quantization/offloading behavior is unrelated to the FlashAttention sink-less crash." + }, + { + "left": "issue:43673", + "right": "issue:44743", + "accept": false, + "reason": "Both involve cache/generation, but they affect different mechanisms and code paths." + }, + { + "left": "issue:33290", + "right": "issue:34567", + "accept": false, + "reason": "Deepspeed Adafactor OOM and `num_input_tokens_seen` not updating are unrelated trainer issues." + }, + { + "left": "issue:44442", + "right": "issue:44898", + "accept": false, + "reason": "Tokenizer loading failure is unrelated to Perceiver interpolation/non-default resolution failure." + }, + { + "left": "issue:34689", + "right": "issue:43611", + "accept": false, + "reason": "Different release regressions and different model-loading assumptions; not the same concrete bug." + }, + { + "left": "issue:41628", + "right": "issue:45372", + "accept": false, + "reason": "Both are import errors, but they involve different missing symbols and different loading flows." + }, + { + "left": "issue:43493", + "right": "issue:43827", + "accept": false, + "reason": "SigLIP2 implementation mismatch and docs still referencing removed `pipeline()` are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 25, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:16998", + "issue:22355", + "issue:25251", + "issue:28282", + "issue:33290", + "issue:34689", + "issue:35141", + "issue:36010", + "issue:36331", + "issue:41950", + "issue:42175", + "issue:43065", + "issue:43295", + "issue:43316", + "issue:43381", + "issue:43388", + "issue:43404", + "issue:43441", + "issue:43475", + "issue:43540", + "issue:43643", + "issue:43688", + "issue:43716", + "issue:43720", + "issue:43749", + "issue:43761", + "issue:43825", + "issue:43856", + "issue:43873", + "issue:43931", + "issue:44077", + "issue:44162", + "issue:44190", + "issue:44442", + "issue:44479", + "issue:44496", + "issue:44514", + "issue:44617", + "issue:44792", + "issue:44855", + "issue:44871", + "issue:44877", + "issue:44913", + "issue:45127", + "issue:45137", + "issue:45200", + "issue:45584" + ], + "soft_pairs": [ + "issue:43381|issue:43873", + "issue:33290|issue:34689", + "issue:33290|issue:35141", + "issue:44077|issue:44479", + "issue:43720|issue:44442", + "issue:43643|issue:44913", + "issue:44190|issue:44496", + "issue:43749|issue:43931", + "issue:43873|issue:44871", + "issue:43540|issue:44514", + "issue:44162|issue:44792", + "issue:43316|issue:44877", + "issue:43381|issue:43856", + "issue:33290|issue:36331", + "issue:34689|issue:36010", + "issue:44855|issue:45137", + "issue:42175|issue:43316", + "issue:43404|issue:43441", + "issue:43388|issue:43688", + "issue:43065|issue:44617", + "issue:16998|issue:22355", + "issue:43316|issue:43716", + "issue:41950|issue:43475", + "issue:25251|issue:28282", + "issue:45127|issue:45200", + "issue:43761|issue:44479", + "issue:43295|issue:43825", + "issue:44871|issue:45584" + ], + "prepared_review_unit_hash": "cd53b91dc9e07fd57b93d3e9e41a20c335c06cb91c1d332ce2067861b633f736", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35511, + "estimated_input_tokens": 8878, + "estimated_eval_tokens": 18012 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is not a true duplicate set: the paired items span unrelated bugs across loading, training, tokenization, multimodal processors, and config semantics. I would not merge any soft edge.", + "confidence": 0.14, + "canonical_issue_reason": "No single issue cleanly represents the cluster; the items are too heterogeneous to treat one as canonical.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43749 has the strongest discussion/inbound activity, but it still only represents one of many unrelated failure modes.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43381", + "right": "issue:43873", + "accept": false, + "reason": "Gradient checkpointing in eval mode is unrelated to quantization/offloading behavior." + }, + { + "left": "issue:33290", + "right": "issue:34689", + "accept": false, + "reason": "Different failure modes: Deepspeed/Adafactor OOM vs Llama 3.2 Vision model-loading regression." + }, + { + "left": "issue:33290", + "right": "issue:35141", + "accept": false, + "reason": "Training-memory OOM and token-embedding reinitialization are distinct bugs." + }, + { + "left": "issue:44077", + "right": "issue:44479", + "accept": false, + "reason": "Unrelated model-specific config validation vs video-input regression in Qwen VL models." + }, + { + "left": "issue:43720", + "right": "issue:44442", + "accept": false, + "reason": "BitNet packed-weight loading and tokenizer loading are different code paths." + }, + { + "left": "issue:43643", + "right": "issue:44913", + "accept": false, + "reason": "Missing fields from trust_remote_code config loading is unrelated to GPTNeoX rotary_pct persistence." + }, + { + "left": "issue:44190", + "right": "issue:44496", + "accept": false, + "reason": "Local dataset loading in a training script is unrelated to unrecognized model/config loading." + }, + { + "left": "issue:43749", + "right": "issue:43931", + "accept": false, + "reason": "FSDP CPU RAM-efficient loading and Qwen3-VL shape mismatch are separate load-time issues." + }, + { + "left": "issue:43873", + "right": "issue:44871", + "accept": false, + "reason": "Quantization/offloading behavior is unrelated to Gemma-3 eos_token_id inconsistency." + }, + { + "left": "issue:43540", + "right": "issue:44514", + "accept": false, + "reason": "Both are video-related, but they affect different Qwen processors and different error conditions." + }, + { + "left": "issue:44162", + "right": "issue:44792", + "accept": false, + "reason": "ESM2 breakage and janus image-generation test failure do not share the same underlying bug." + }, + { + "left": "issue:43316", + "right": "issue:44877", + "accept": false, + "reason": "Gemma3TextConfig API mismatch is unrelated to strict config loading for granite_speech." + }, + { + "left": "issue:43381", + "right": "issue:43856", + "accept": false, + "reason": "Eval-mode gradient checkpointing is not the same as MoE training memory inefficiency." + }, + { + "left": "issue:33290", + "right": "issue:36331", + "accept": false, + "reason": "Deepspeed OOM with Adafactor is unrelated to CustomTrainer.compute_loss signature breakage." + }, + { + "left": "issue:34689", + "right": "issue:36010", + "accept": false, + "reason": "Llama 3.2 Vision model loading and GenerationMixin import failure are different regressions." + }, + { + "left": "issue:44855", + "right": "issue:45137", + "accept": false, + "reason": "Python 3.13 import parsing error and DeepSpeed ZeRO3 deque error are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:43316", + "accept": false, + "reason": "Backend packaging omission has no overlap with Gemma3TextConfig API discrepancy." + }, + { + "left": "issue:43404", + "right": "issue:43441", + "accept": false, + "reason": "Same model family, but weight tying and FlashAttention failures are distinct code-path bugs." + }, + { + "left": "issue:43388", + "right": "issue:43688", + "accept": false, + "reason": "Metric gathering label truncation and auxiliary-loss normalization are unrelated training issues." + }, + { + "left": "issue:43065", + "right": "issue:44617", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder is not the same as Sam3Video OOM behavior." + }, + { + "left": "issue:16998", + "right": "issue:22355", + "accept": false, + "reason": "Model_max_length behavior and missing transformers.onnx module are unrelated." + }, + { + "left": "issue:43316", + "right": "issue:43716", + "accept": false, + "reason": "Gemma3TextConfig API differences do not match Mistral-3 image preprocessor dtype mismatch." + }, + { + "left": "issue:41950", + "right": "issue:43475", + "accept": false, + "reason": "Video-classification processor lookup and missing Sam3VisionEncoderOutput attribute are separate issues." + }, + { + "left": "issue:25251", + "right": "issue:28282", + "accept": false, + "reason": "Pipeline output-shape behavior is unrelated to missing PyTorch import errors." + }, + { + "left": "issue:45127", + "right": "issue:45200", + "accept": false, + "reason": "LoRA merge collapse with tied embeddings is not the same as missing mm_token_type_ids defaults." + }, + { + "left": "issue:43761", + "right": "issue:44479", + "accept": false, + "reason": "CLIPVisionModel hidden_states regression and Qwen video-input regression are different models and failures." + }, + { + "left": "issue:43295", + "right": "issue:43825", + "accept": false, + "reason": "Processor.tokenizer regression is unrelated to an inaccurate pipeline error message." + }, + { + "left": "issue:44871", + "right": "issue:45584", + "accept": false, + "reason": "Gemma-3 eos_token_id mismatch and Whisper empty-transcription handling are different generation bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 26, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:32090", + "issue:35141", + "issue:41720", + "issue:42617", + "issue:43065", + "issue:43295", + "issue:43452", + "issue:43550", + "issue:43575", + "issue:43606", + "issue:43646", + "issue:43701", + "issue:43742", + "issue:43761", + "issue:43827", + "issue:43906", + "issue:43931", + "issue:43994", + "issue:44060", + "issue:44077", + "issue:44190", + "issue:44242", + "issue:44464", + "issue:44479", + "issue:44496", + "issue:44521", + "issue:44534", + "issue:44610", + "issue:44617", + "issue:44734", + "issue:44792", + "issue:44871", + "issue:44913", + "issue:44933", + "issue:44938", + "issue:44964", + "issue:44977", + "issue:45127", + "issue:45245", + "issue:45290", + "issue:45357", + "issue:45362", + "issue:45381", + "issue:45406", + "issue:45542", + "issue:45563", + "issue:45584", + "issue:45600" + ], + "soft_pairs": [ + "issue:44734|issue:45406", + "issue:44077|issue:44534", + "issue:43295|issue:44871", + "issue:43550|issue:43701", + "issue:43606|issue:43701", + "issue:43646|issue:44913", + "issue:43906|issue:45290", + "issue:44617|issue:45584", + "issue:43761|issue:44242", + "issue:43931|issue:44060", + "issue:43452|issue:44496", + "issue:45563|issue:45584", + "issue:44521|issue:45381", + "issue:43575|issue:43742", + "issue:42617|issue:43575", + "issue:44479|issue:44734", + "issue:44933|issue:45362", + "issue:41720|issue:43575", + "issue:32090|issue:35141", + "issue:44464|issue:45563", + "issue:43065|issue:44610", + "issue:44464|issue:44871", + "issue:44792|issue:44977", + "issue:43827|issue:43994", + "issue:44534|issue:44913", + "issue:41720|issue:45245", + "issue:44190|issue:44964", + "issue:44938|issue:45542", + "issue:45127|issue:45357", + "issue:44792|issue:45600" + ], + "prepared_review_unit_hash": "f4c2c006e106aa45c9e0e87196e9f7f005ee4e470613b9f418ff392c8509ddb6", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36583, + "estimated_input_tokens": 9146, + "estimated_eval_tokens": 18548 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: the soft-similarity pairs mostly share broad areas (serving, multimodal, generation, config loading) but do not describe the same concrete bug/change. I would reject all proposed merges; no PRs are present. If one issue must anchor the set, issue 44521 is the strongest open representative because it has the most inbound references and broad activity.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 44521 is the best representative by current activity and inbound references, but the cluster has no single true duplicate center.", + "canonical_pr_reason": null, + "best_issue_reason": "Open, actively discussed, and the most referenced issue in the set, so it is the best practical anchor.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44734", + "right": "issue:45406", + "accept": false, + "reason": "Both involve transformers serve, but one is KV-cache tensor indexing and the other is a missing processor attribute; different failures." + }, + { + "left": "issue:44077", + "right": "issue:44534", + "accept": false, + "reason": "Unrelated bugs: patchtsmixer post_init validation vs non-persistent buffer corruption in v5." + }, + { + "left": "issue:43295", + "right": "issue:44871", + "accept": false, + "reason": "Different regressions: processor/tokenizer access breakage vs eos_token_id config mismatch." + }, + { + "left": "issue:43550", + "right": "issue:43701", + "accept": false, + "reason": "torch.compile/SDPA model failure is unrelated to resume_from_checkpoint key mismatches." + }, + { + "left": "issue:43606", + "right": "issue:43701", + "accept": false, + "reason": "CPU offload device mismatch is a different code path from checkpoint key handling." + }, + { + "left": "issue:43646", + "right": "issue:44913", + "accept": false, + "reason": "Custom model init breakage is not the same as rotary_pct not persisting on reload." + }, + { + "left": "issue:43906", + "right": "issue:45290", + "accept": false, + "reason": "The reproduction target and failure mode differ: isolated prior issue reproduction vs chat template tool-call crash." + }, + { + "left": "issue:44617", + "right": "issue:45584", + "accept": false, + "reason": "CUDA OOM in Sam3Video is unrelated to Whisper empty-transcription failure after align_special_tokens." + }, + { + "left": "issue:43761", + "right": "issue:44242", + "accept": false, + "reason": "CLIPVision hidden_states bug is different from MoE load-balancing loss omission." + }, + { + "left": "issue:43931", + "right": "issue:44060", + "accept": false, + "reason": "Model weight shape mismatch is not the same as the incorrect tied-weights warning." + }, + { + "left": "issue:43452", + "right": "issue:44496", + "accept": false, + "reason": "GGUF tokenizer/model loading breakage is a different issue from missing model_type in config.json." + }, + { + "left": "issue:45563", + "right": "issue:45584", + "accept": false, + "reason": "Stale num_return_sequences warning is unrelated to Whisper generation crashing on empty output." + }, + { + "left": "issue:44521", + "right": "issue:45381", + "accept": false, + "reason": "Both are multimodal/VL-related, but assistant_mask generation and vision_position_ids are distinct bugs." + }, + { + "left": "issue:43575", + "right": "issue:43742", + "accept": false, + "reason": "TP OOM on Qwen2-57B and key errors when loading MobileLLM are different load-time failures." + }, + { + "left": "issue:42617", + "right": "issue:43575", + "accept": false, + "reason": "3d_parallel.py runtime failure does not match Qwen2 tensor-parallel OOM." + }, + { + "left": "issue:44479", + "right": "issue:44734", + "accept": false, + "reason": "Video-input regression across Qwen VL models is unrelated to serve-side KV-cache indexing." + }, + { + "left": "issue:44933", + "right": "issue:45362", + "accept": false, + "reason": "Missing image_utils import and Qwen3.5 chat crashes are different defects." + }, + { + "left": "issue:41720", + "right": "issue:43575", + "accept": false, + "reason": "Auto device mapping cudaErrorAssert is not the same as tensor-parallel OOM." + }, + { + "left": "issue:32090", + "right": "issue:35141", + "accept": false, + "reason": "Trainer GPU broadcast NoneType error is unrelated to embedding reinitialization in post_init." + }, + { + "left": "issue:44464", + "right": "issue:45563", + "accept": false, + "reason": "Compiled-forward generation inconsistency is not the same as a stale warning." + }, + { + "left": "issue:43065", + "right": "issue:44610", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder and OmDet-Turbo input-size mismatch are separate model issues." + }, + { + "left": "issue:44464", + "right": "issue:44871", + "accept": false, + "reason": "Chunked generation inconsistency is unrelated to eos_token_id configuration drift." + }, + { + "left": "issue:44792", + "right": "issue:44977", + "accept": false, + "reason": "Janus image-generation test failure is not the same as Qwen3.5 flash-attention generation failure." + }, + { + "left": "issue:43827", + "right": "issue:43994", + "accept": false, + "reason": "Documentation references to removed pipeline() are not the same as SigLIP2 producing bad outputs." + }, + { + "left": "issue:44534", + "right": "issue:44913", + "accept": false, + "reason": "Non-persistent buffer junk in v5 is unrelated to GPTNeoX rotary_pct not persisting." + }, + { + "left": "issue:41720", + "right": "issue:45245", + "accept": false, + "reason": "Qwen3 auto-device-map CUDA assert is unrelated to the 2^24 categories limit error." + }, + { + "left": "issue:44190", + "right": "issue:44964", + "accept": false, + "reason": "Local dataset loading in an example script is a different problem from Phi-4 multimodal model loading." + }, + { + "left": "issue:44938", + "right": "issue:45542", + "accept": false, + "reason": "Python 3.14 import/load failure is unrelated to the TensorBoard-without-TensorFlow backend error." + }, + { + "left": "issue:45127", + "right": "issue:45357", + "accept": false, + "reason": "LoRA merge collapse with extended vocab is a different bug from incorrect visual encoder keys being saved." + }, + { + "left": "issue:44242", + "right": "issue:44521", + "accept": false, + "reason": "MoE load-balancing loss omission is unrelated to assistant mask generation for multimodal chat templates." + }, + { + "left": "issue:44617", + "right": "issue:45584", + "accept": false, + "reason": "Sam3Video OOM and Whisper empty-transcription crash are unrelated generation failures." + }, + { + "left": "issue:44792", + "right": "issue:45600", + "accept": false, + "reason": "Janus image-generation test failure is unrelated to CI breakage from removed Sam3LiteText configs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 27, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:36246", + "issue:38175", + "issue:38617", + "issue:39692", + "issue:41628", + "issue:41950", + "issue:42175", + "issue:43065", + "issue:43116", + "issue:43295", + "issue:43388", + "issue:43404", + "issue:43479", + "issue:43540", + "issue:43550", + "issue:43582", + "issue:43644", + "issue:43650", + "issue:43761", + "issue:43976", + "issue:44016", + "issue:44077", + "issue:44079", + "issue:44112", + "issue:44190", + "issue:44220", + "issue:44246", + "issue:44279", + "issue:44355", + "issue:44466", + "issue:44514", + "issue:44561", + "issue:44683", + "issue:44734", + "issue:44743", + "issue:44855", + "issue:44913", + "issue:45003", + "issue:45216", + "issue:45245", + "issue:45276", + "issue:45375", + "issue:45406", + "issue:45446", + "issue:45584", + "issue:45588", + "issue:45600", + "issue:45632" + ], + "soft_pairs": [ + "issue:43388|issue:43582", + "issue:43761|issue:44743", + "issue:44855|issue:45588", + "issue:43650|issue:44220", + "issue:44355|issue:44855", + "issue:43540|issue:44734", + "issue:43644|issue:44913", + "issue:45276|issue:45584", + "issue:43976|issue:44855", + "issue:43404|issue:45216", + "issue:44077|issue:44079", + "issue:43550|issue:44190", + "issue:43116|issue:44190", + "issue:44112|issue:45600", + "issue:43295|issue:45276", + "issue:45245|issue:45584", + "issue:41950|issue:44016", + "issue:39692|issue:44190", + "issue:36246|issue:41628", + "issue:44734|issue:45632", + "issue:43479|issue:45375", + "issue:45406|issue:45632", + "issue:36246|issue:38617", + "issue:44561|issue:45632", + "issue:44514|issue:44734", + "issue:44016|issue:44220", + "issue:38175|issue:43065", + "issue:42175|issue:44279", + "issue:44683|issue:45446", + "issue:44466|issue:44913", + "issue:44246|issue:45003" + ], + "prepared_review_unit_hash": "35db967a3dfbef3fd5e89fa1b92ba62e26932da9cdb718fe8ab845ad2ee5b754", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37130, + "estimated_input_tokens": 9283, + "estimated_eval_tokens": 18822 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous; the soft edges mostly connect unrelated bugs that only share broad subsystem vocabulary (imports, configs, cache, multimodal, CI, etc.). I would not merge any of the proposed pairs as duplicates.", + "confidence": 0.72, + "canonical_issue_reason": "issue:44561 is the best representative: it describes a clear, high-impact framework regression in `trust_remote_code` handling caused by a removed utility, with broad applicability beyond one model family.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44561 has the strongest combination of clarity, scope, and downstream impact, so it is the best single issue anchor for this cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43388", + "right": "issue:43582", + "accept": false, + "reason": "Different failures: metric gathering last-batch label handling vs Apple Silicon allocator warmup TypeError." + }, + { + "left": "issue:43761", + "right": "issue:44743", + "accept": false, + "reason": "Unrelated model bugs: CLIP vision hidden states regression vs Qwen3 recurrent state reset with cache." + }, + { + "left": "issue:44855", + "right": "issue:45588", + "accept": false, + "reason": "Different code paths and symptoms: Python 3.13 JIT parsing/import indentation issue vs flash-attention `s_aux=None` crash." + }, + { + "left": "issue:43650", + "right": "issue:44220", + "accept": false, + "reason": "No shared underlying bug; one is an unclear data-related issue, the other is a specific `_torch_extract_fbank_features()` problem." + }, + { + "left": "issue:44355", + "right": "issue:44855", + "accept": false, + "reason": "Both involve import/compile-time errors, but the concrete root causes and affected paths are different." + }, + { + "left": "issue:43540", + "right": "issue:44734", + "accept": false, + "reason": "Different multimodal/runtime bugs: Qwen3OmniMoe video processing ValueError vs serve KV-cache tensor indexing crash." + }, + { + "left": "issue:43644", + "right": "issue:44913", + "accept": false, + "reason": "One is about non-persistent buffer serialization junk; the other is GPTNeoX config reload losing `rotary_pct`." + }, + { + "left": "issue:45276", + "right": "issue:45584", + "accept": false, + "reason": "Gemma4 embedding resize mismatch is unrelated to Whisper empty-transcription generation after `align_special_tokens`." + }, + { + "left": "issue:43976", + "right": "issue:44855", + "accept": false, + "reason": "Different regressions: package Python-version compatibility vs a Python 3.13 TorchScript parsing/indentation failure." + }, + { + "left": "issue:43404", + "right": "issue:45216", + "accept": false, + "reason": "Both mention weights/checkpointing, but they affect different models and concrete failure modes; not the same bug." + }, + { + "left": "issue:44077", + "right": "issue:44079", + "accept": false, + "reason": "Unrelated issues: `post_init` validation vs `ModelOutput` key assignment logic." + }, + { + "left": "issue:43550", + "right": "issue:44190", + "accept": false, + "reason": "Different domains: `torch.compile` SDPA failure in Bamba vs local dataset loading in an image-classification example." + }, + { + "left": "issue:43116", + "right": "issue:44190", + "accept": false, + "reason": "Separate example-script bugs with different symptoms and likely different causes." + }, + { + "left": "issue:44112", + "right": "issue:45600", + "accept": false, + "reason": "Both are CI-related, but one is a stale device-override test and the other is a broken auto-mapping reference." + }, + { + "left": "issue:43295", + "right": "issue:45276", + "accept": false, + "reason": "Different processor/tokenizer regression vs Gemma4 embedding resize behavior; no shared concrete code-path." + }, + { + "left": "issue:45245", + "right": "issue:45584", + "accept": false, + "reason": "Unrelated errors: category-cardinality limit vs Whisper empty transcription handling." + }, + { + "left": "issue:41950", + "right": "issue:44016", + "accept": false, + "reason": "A video-classification pipeline bug and a notebook syntax error are not the same underlying issue." + }, + { + "left": "issue:39692", + "right": "issue:44190", + "accept": false, + "reason": "Documentation/example errors versus local dataset loading failure; only broad area overlaps." + }, + { + "left": "issue:36246", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but they concern different symbols and different API surfaces; not the same bug." + }, + { + "left": "issue:44734", + "right": "issue:45632", + "accept": false, + "reason": "Serve KV-cache continuation crash and trust-remote-code cache-path collision are unrelated." + }, + { + "left": "issue:43479", + "right": "issue:45375", + "accept": false, + "reason": "Different configuration bugs in different model families; one is default initialization, the other is a missing strict-serialization field." + }, + { + "left": "issue:45406", + "right": "issue:45632", + "accept": false, + "reason": "Gemma4Processor missing `_tokenizer` in serve is unrelated to trust_remote_code cache collisions." + }, + { + "left": "issue:36246", + "right": "issue:38617", + "accept": false, + "reason": "Different import failures affecting different names/modules; no shared root cause is evident." + }, + { + "left": "issue:44561", + "right": "issue:45632", + "accept": false, + "reason": "One is a removed-API regression breaking trust_remote_code models; the other is a local cache-path collision bug." + }, + { + "left": "issue:44514", + "right": "issue:44734", + "accept": false, + "reason": "Different Qwen2.5-VL chat-template batching issue vs serve KV-cache continuation indexing bug." + }, + { + "left": "issue:44016", + "right": "issue:44220", + "accept": false, + "reason": "Notebook syntax error and audio feature-extraction issue are unrelated." + }, + { + "left": "issue:38175", + "right": "issue:43065", + "accept": false, + "reason": "Different model-specific behavior: SigLIP2 zero probabilities vs Sam3PixelDecoder dummy Conv2d." + }, + { + "left": "issue:42175", + "right": "issue:44279", + "accept": false, + "reason": "Backend/dependency installation issue vs generic dependency issue; no concrete shared bug." + }, + { + "left": "issue:44683", + "right": "issue:45446", + "accept": false, + "reason": "Both touch flex_attention/PyTorch versioning, but the failures are distinct and not mergeable as one fix." + }, + { + "left": "issue:44466", + "right": "issue:44913", + "accept": false, + "reason": "Different state/serialization bugs: tied lm_head weight serialization vs GPTNeoX config reload behavior." + }, + { + "left": "issue:44246", + "right": "issue:45003", + "accept": false, + "reason": "Import performance complaint and unsafe `sys.modules[]` access 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": 2, + "component_count": 13, + "review_unit_index": 28, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:42907", + "issue:43295", + "issue:43299", + "issue:43381", + "issue:43441", + "issue:43479", + "issue:43525", + "issue:43526", + "issue:43531", + "issue:43577", + "issue:43643", + "issue:43673", + "issue:43824", + "issue:43854", + "issue:43901", + "issue:43927", + "issue:43976", + "issue:44016", + "issue:44062", + "issue:44075", + "issue:44112", + "issue:44242", + "issue:44279", + "issue:44336", + "issue:44351", + "issue:44368", + "issue:44462", + "issue:44496", + "issue:44521", + "issue:44683", + "issue:44933", + "issue:44964", + "issue:44987", + "issue:45042", + "issue:45081", + "issue:45200", + "issue:45216", + "issue:45278", + "issue:45325", + "issue:45335", + "issue:45357", + "issue:45405", + "issue:45542", + "issue:45561", + "issue:45584", + "issue:45593", + "issue:45600", + "issue:45632" + ], + "soft_pairs": [ + "issue:43299|issue:45216", + "issue:42907|issue:43441", + "issue:43643|issue:44496", + "issue:44016|issue:44062", + "issue:43295|issue:45335", + "issue:44964|issue:45600", + "issue:43976|issue:44987", + "issue:43673|issue:44336", + "issue:44112|issue:44964", + "issue:45561|issue:45632", + "issue:43531|issue:45600", + "issue:45335|issue:45584", + "issue:43824|issue:44933", + "issue:43976|issue:44279", + "issue:44496|issue:45042", + "issue:43854|issue:44964", + "issue:43577|issue:45216", + "issue:43299|issue:45357", + "issue:43526|issue:43901", + "issue:44075|issue:44368", + "issue:44521|issue:45325", + "issue:44242|issue:45200", + "issue:43381|issue:45593", + "issue:44351|issue:45278", + "issue:43927|issue:45632", + "issue:43901|issue:44496", + "issue:43976|issue:45405", + "issue:44683|issue:45542", + "issue:43479|issue:43525", + "issue:44462|issue:45081" + ], + "prepared_review_unit_hash": "018d0195f305cc6745ec7b4e6ef6e3dbc8633d39a3e09947bdb40dcb132ce65c", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36769, + "estimated_input_tokens": 9193, + "estimated_eval_tokens": 18642 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous; the soft links are driven by superficial similarity (same model family, same subsystem, or generic import/load regressions) rather than the same concrete bug. I would keep the cluster centered on the broad upgrade/import-regression report, but reject the proposed duplicate edges.", + "confidence": 0.8, + "canonical_issue_reason": "issue:45278 is the broadest and most representative report here: a user-facing breakage after upgrading to 5.5.0 that can plausibly absorb several related import-regression reports.", + "canonical_pr_reason": null, + "best_issue_reason": "It is the best global issue representative because it is broad, central to the cluster\u2019s common theme (5.x upgrade/import failures), and more likely than the others to serve as an umbrella canonical issue.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43299", + "right": "issue:45216", + "accept": false, + "reason": "Both are Qwen3-family regressions, but one is a loading failure and the other is a save_pretrained checkpoint regression; different code paths and different symptoms." + }, + { + "left": "issue:42907", + "right": "issue:43441", + "accept": false, + "reason": "Dequantized save failure vs FlashAttention load failure for Ministral-3 are distinct problems." + }, + { + "left": "issue:43643", + "right": "issue:44496", + "accept": false, + "reason": "Both involve config/model loading, but one is missing fields in AutoConfig with trust_remote_code and the other is an unrecognized model due to missing model_type; not the same bug." + }, + { + "left": "issue:44016", + "right": "issue:44062", + "accept": false, + "reason": "A notebook syntax error and an AddedToken argument conflict are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:45335", + "accept": false, + "reason": "processor.tokenizer/image passing regression and t5gemma embedding resize behavior are separate issues." + }, + { + "left": "issue:44964", + "right": "issue:45600", + "accept": false, + "reason": "Phi-4 multimodal loading and stale auto_mappings references to Sam3LiteText configs are different failures." + }, + { + "left": "issue:43976", + "right": "issue:44987", + "accept": false, + "reason": "Python version compatibility and loading physical-intelligence/fast are unrelated." + }, + { + "left": "issue:43673", + "right": "issue:44336", + "accept": false, + "reason": "GenerationMixin cache missing during chunked_prefill and ANSI codes in loading_report are different subsystems and bugs." + }, + { + "left": "issue:44112", + "right": "issue:44964", + "accept": false, + "reason": "CI test flakiness in GraniteSpeech is unrelated to Phi-4 multimodal model loading." + }, + { + "left": "issue:45561", + "right": "issue:45632", + "accept": false, + "reason": "A pytest-xdist file race and a trust_remote_code cache-path collision are different concurrency/path issues." + }, + { + "left": "issue:43531", + "right": "issue:45600", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and removed Sam3LiteText references in auto_mappings are unrelated." + }, + { + "left": "issue:45335", + "right": "issue:45584", + "accept": false, + "reason": "Embedding resize propagation and Whisper empty-transcription handling are distinct bugs." + }, + { + "left": "issue:43824", + "right": "issue:44933", + "accept": false, + "reason": "Both are import-related, but one is a missing Qwen2.5-VL export and the other is a non-existent image_utils import; not the same symbol or code path." + }, + { + "left": "issue:43976", + "right": "issue:44279", + "accept": false, + "reason": "Python support/version mismatch and a generic dependency issue are not the same problem." + }, + { + "left": "issue:44496", + "right": "issue:45042", + "accept": false, + "reason": "Missing model_type config loading and PIL image processors requiring torchvision are unrelated." + }, + { + "left": "issue:43854", + "right": "issue:44964", + "accept": false, + "reason": "Two different model-specific loading failures; same broad area, but not the same concrete bug." + }, + { + "left": "issue:43577", + "right": "issue:45216", + "accept": false, + "reason": "BLIP2 dtype propagation and Qwen3.5 save_pretrained checkpoint correctness are unrelated." + }, + { + "left": "issue:43299", + "right": "issue:45357", + "accept": false, + "reason": "Qwen3VL loading and Qwen3.5 visual-encoder save_pretrained key layout are different regressions." + }, + { + "left": "issue:43526", + "right": "issue:43901", + "accept": false, + "reason": "A BeitImageProcessorFast bug and a docs/behavior mismatch in TextClassificationPipeline are unrelated." + }, + { + "left": "issue:44075", + "right": "issue:44368", + "accept": false, + "reason": "Ignored SGD optimizer args and a Qwen3.5 tie_word_embeddings warning are not the same bug." + }, + { + "left": "issue:44521", + "right": "issue:45325", + "accept": false, + "reason": "All-zero assistant masks in multimodal chat templates and Qwen2.5-VL rope-index scaling are separate multimodal issues." + }, + { + "left": "issue:44242", + "right": "issue:45200", + "accept": false, + "reason": "Missing load-balancing loss and Gemma 4 mm_token_type_ids defaults are different model-training bugs." + }, + { + "left": "issue:43381", + "right": "issue:45593", + "accept": false, + "reason": "Gradient checkpointing in eval mode and D-FINE auxiliary-loss behavior are unrelated." + }, + { + "left": "issue:44351", + "right": "issue:45278", + "accept": false, + "reason": "44351 is one specific missing import (HybridCache); 45278 is a broad complaint about many import errors after upgrade, not the same concrete failure." + }, + { + "left": "issue:43901", + "right": "issue:44496", + "accept": false, + "reason": "A pipeline docs mismatch and an unrecognized model/config loading error are unrelated." + }, + { + "left": "issue:43976", + "right": "issue:45405", + "accept": false, + "reason": "Python compatibility and a PEFT minimum-version bump are different dependency/version issues." + }, + { + "left": "issue:44683", + "right": "issue:45542", + "accept": false, + "reason": "Compiled flex_attention on torch>=2.9 and a TensorFlow backend error from tensorboard installation are unrelated." + }, + { + "left": "issue:43479", + "right": "issue:43525", + "accept": false, + "reason": "Different config attributes on different models; both are config bugs, but not the same one." + }, + { + "left": "issue:44462", + "right": "issue:45081", + "accept": false, + "reason": "AutoTokenizer ignoring tokenizer.json and _patch_mistral_regex crashing on backend_tokenizer access are distinct tokenizer-loading bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 29, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:38175", + "issue:39692", + "issue:41950", + "issue:42994", + "issue:43065", + "issue:43122", + "issue:43262", + "issue:43295", + "issue:43299", + "issue:43317", + "issue:43441", + "issue:43450", + "issue:43650", + "issue:43653", + "issue:43673", + "issue:43723", + "issue:43761", + "issue:43784", + "issue:43854", + "issue:43901", + "issue:44016", + "issue:44162", + "issue:44261", + "issue:44263", + "issue:44279", + "issue:44291", + "issue:44297", + "issue:44462", + "issue:44466", + "issue:44496", + "issue:44521", + "issue:44617", + "issue:44623", + "issue:44671", + "issue:44683", + "issue:44749", + "issue:44779", + "issue:44877", + "issue:44898", + "issue:44987", + "issue:45020", + "issue:45276", + "issue:45290", + "issue:45397", + "issue:45405", + "issue:45542", + "issue:45593" + ], + "soft_pairs": [ + "issue:41950|issue:45020", + "issue:44297|issue:44462", + "issue:44279|issue:45405", + "issue:41950|issue:43650", + "issue:44496|issue:44877", + "issue:43299|issue:43854", + "issue:44462|issue:44779", + "issue:38175|issue:44617", + "issue:43295|issue:44779", + "issue:39692|issue:43901", + "issue:43650|issue:45290", + "issue:43723|issue:44162", + "issue:43450|issue:44279", + "issue:44016|issue:44291", + "issue:43262|issue:44521", + "issue:42994|issue:43317", + "issue:44261|issue:44263", + "issue:43441|issue:44683", + "issue:43065|issue:44898", + "issue:44466|issue:45276", + "issue:44623|issue:45397", + "issue:43122|issue:43650", + "issue:44623|issue:44749", + "issue:43784|issue:45542", + "issue:44671|issue:44779", + "issue:43673|issue:43761", + "issue:44987|issue:45542", + "issue:43653|issue:45593" + ], + "prepared_review_unit_hash": "3445cd7f5940f976dde3c5008537684e22c9204532e3ff04c3c50c2944243ef3", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35392, + "estimated_input_tokens": 8848, + "estimated_eval_tokens": 17952 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is heterogeneous: most items are distinct regressions around tokenizers, processors, model loading, quantization, or docs, not one duplicate cluster.", + "confidence": 0.34, + "canonical_issue_reason": "No single true duplicate stands out from the titles alone. If forced to pick one representative issue, 45020 is the broadest umbrella report about recent transformers versions breaking remote_code models.", + "canonical_pr_reason": null, + "best_issue_reason": "45020 is the most general, user-facing regression report in the set and best represents the cluster\u2019s overall theme better than the narrower subsystem-specific issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41950", + "right": "issue:45020", + "accept": false, + "reason": "Different bugs: video-classification processor selection vs. remote_code model-loading regressions." + }, + { + "left": "issue:44297", + "right": "issue:44462", + "accept": false, + "reason": "Both concern tokenizers, but one is save_pretrained metadata and the other is repo loading behavior; not the same concrete bug." + }, + { + "left": "issue:44279", + "right": "issue:45405", + "accept": false, + "reason": "Unrelated: a generic dependency issue vs. a version pin bump to an unreleased PEFT release." + }, + { + "left": "issue:41950", + "right": "issue:43650", + "accept": false, + "reason": "Completely unrelated: video pipeline processor lookup vs. placeholder issue text." + }, + { + "left": "issue:44496", + "right": "issue:44877", + "accept": false, + "reason": "Both are config-loading failures, but for different models and different strictness checks; not one bug." + }, + { + "left": "issue:43299", + "right": "issue:43854", + "accept": false, + "reason": "Different model-loading failures for different architectures; no evidence of a shared code-path bug." + }, + { + "left": "issue:44462", + "right": "issue:44779", + "accept": false, + "reason": "Tokenizer loading from repo vs. incorrect Deepseek tokenization are different regressions." + }, + { + "left": "issue:38175", + "right": "issue:44617", + "accept": false, + "reason": "Unrelated: SigLIP2 probability output vs. Sam3Video CUDA memory usage." + }, + { + "left": "issue:43295", + "right": "issue:44779", + "accept": false, + "reason": "Custom processor.tokenizer access regression vs. Deepseek tokenizer output regression; distinct failures." + }, + { + "left": "issue:39692", + "right": "issue:43901", + "accept": false, + "reason": "Both are docs-related, but one is a model/processor mismatch and quantization example issue; the other is stale pipeline docs." + }, + { + "left": "issue:43650", + "right": "issue:45290", + "accept": false, + "reason": "Placeholder issue text vs. a specific apply_chat_template crash on tool-call assistant messages." + }, + { + "left": "issue:43723", + "right": "issue:44162", + "accept": false, + "reason": "Tokenizer loading regression vs. an ESM2 model breakage; unrelated." + }, + { + "left": "issue:43450", + "right": "issue:44279", + "accept": false, + "reason": "Video processor shape bug vs. dependency/version issue; not the same underlying change." + }, + { + "left": "issue:44016", + "right": "issue:44291", + "accept": false, + "reason": "Notebook syntax error vs. init_empty_weights loading TypeError; unrelated." + }, + { + "left": "issue:43262", + "right": "issue:44521", + "accept": false, + "reason": "Both mention apply_chat_template, but one is audio sampling-rate defaulting and the other is multimodal assistant mask generation." + }, + { + "left": "issue:42994", + "right": "issue:43317", + "accept": false, + "reason": "Different quantization-related failures: saving quantized models vs. loading dequantized models with device_map=auto." + }, + { + "left": "issue:44261", + "right": "issue:44263", + "accept": false, + "reason": "Precision/config issue in MLA layernorm vs. a torch.split return-value issue in a different model component." + }, + { + "left": "issue:43441", + "right": "issue:44683", + "accept": false, + "reason": "Different attention-backend problems: FlashAttention failure for a model vs. compiled flex_attention failing on newer torch." + }, + { + "left": "issue:43065", + "right": "issue:44898", + "accept": false, + "reason": "Unrelated model/component bugs: dummy Conv2d in Sam3PixelDecoder vs. Perceiver image classification resize handling." + }, + { + "left": "issue:44466", + "right": "issue:45276", + "accept": false, + "reason": "Serialization of lm_head.weight vs. resize_token_embeddings not propagating; related area but different concrete bugs." + }, + { + "left": "issue:44623", + "right": "issue:45397", + "accept": false, + "reason": "processor.save_pretrained missing files vs. gemma-4 zero3 from_pretrained; different failure modes." + }, + { + "left": "issue:43122", + "right": "issue:43650", + "accept": false, + "reason": "Tokenizer behavior regression vs. placeholder issue text; no overlap." + }, + { + "left": "issue:44623", + "right": "issue:44749", + "accept": false, + "reason": "Processor saving bug vs. performance regression during data filtering; unrelated." + }, + { + "left": "issue:43784", + "right": "issue:45542", + "accept": false, + "reason": "Import-time NameError from transformers integration vs. tensorboard/TensorFlow backend detection issue." + }, + { + "left": "issue:44671", + "right": "issue:44779", + "accept": false, + "reason": "Different regressions: CamemBERT masked LM predictions vs. Deepseek tokenizer output." + }, + { + "left": "issue:43673", + "right": "issue:43761", + "accept": false, + "reason": "Cache-missing generation regression vs. CLIP vision hidden_states regression; unrelated." + }, + { + "left": "issue:44987", + "right": "issue:45542", + "accept": false, + "reason": "Model loading under transformers>=5.1.0 vs. tensorboard without TensorFlow backend handling; different areas." + }, + { + "left": "issue:43653", + "right": "issue:45593", + "accept": false, + "reason": "Tokenizer special-token registration bug vs. D-FINE auxiliary-loss behavior; unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 30, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:36246", + "issue:41628", + "issue:41950", + "issue:42175", + "issue:42491", + "issue:42757", + "issue:43295", + "issue:43299", + "issue:43352", + "issue:43475", + "issue:43525", + "issue:43526", + "issue:43531", + "issue:43673", + "issue:43824", + "issue:43881", + "issue:43901", + "issue:43976", + "issue:43994", + "issue:44016", + "issue:44162", + "issue:44188", + "issue:44220", + "issue:44242", + "issue:44246", + "issue:44295", + "issue:44297", + "issue:44336", + "issue:44351", + "issue:44361", + "issue:44373", + "issue:44462", + "issue:44485", + "issue:44496", + "issue:44561", + "issue:44589", + "issue:44617", + "issue:44661", + "issue:45003", + "issue:45200", + "issue:45250", + "issue:45356", + "issue:45362", + "issue:45397", + "issue:45542", + "issue:45561", + "issue:45563", + "issue:45593" + ], + "soft_pairs": [ + "issue:42175|issue:43976", + "issue:43901|issue:43994", + "issue:44297|issue:45397", + "issue:44361|issue:45593", + "issue:43475|issue:44016", + "issue:45561|issue:45563", + "issue:43824|issue:43976", + "issue:43295|issue:44617", + "issue:43525|issue:43881", + "issue:43352|issue:45250", + "issue:44589|issue:45362", + "issue:44188|issue:44336", + "issue:41628|issue:43976", + "issue:44295|issue:45397", + "issue:43901|issue:44246", + "issue:42491|issue:43299", + "issue:41950|issue:44220", + "issue:44336|issue:44373", + "issue:44242|issue:45593", + "issue:44462|issue:45356", + "issue:44661|issue:45542", + "issue:44246|issue:44496", + "issue:45003|issue:45561", + "issue:36246|issue:42757", + "issue:36246|issue:44351", + "issue:43531|issue:45362", + "issue:43526|issue:45003", + "issue:44336|issue:44485", + "issue:43673|issue:44561", + "issue:44162|issue:45200" + ], + "prepared_review_unit_hash": "927a8378370efb0e1783a633764ddf08e8dce0b6a876aef02ed14a6fc2ddf28f", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36334, + "estimated_input_tokens": 9084, + "estimated_eval_tokens": 18424 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is not a true duplicate set; it mixes many unrelated Transformers issues (imports, model loading, generation, tokenizer behavior, docs, tests, and backend/runtime problems). I would not merge any of the soft pairs.", + "confidence": 0.91, + "canonical_issue_reason": "issue:43824 is the strongest representative issue here: it has the highest discussion/inbound signal and a concrete, user-facing import regression. That said, the overall set is heterogeneous rather than one coherent bug family.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43824 is the best single issue to anchor the cluster because it is specific, high-impact, and clearly described. It is still only a representative, not a duplicate target for the rest.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42175", + "right": "issue:43976", + "accept": false, + "reason": "Packaging/backend dependency complaint vs a Python-version compatibility bug; different failure modes and fixes." + }, + { + "left": "issue:43901", + "right": "issue:43994", + "accept": false, + "reason": "Docs mismatch in TextClassificationPipeline vs a model inference correctness issue; not the same underlying bug." + }, + { + "left": "issue:44297", + "right": "issue:45397", + "accept": false, + "reason": "Tokenizer metadata saved wrong vs a zero3 from_pretrained loading failure; unrelated code paths." + }, + { + "left": "issue:44361", + "right": "issue:45593", + "accept": false, + "reason": "Tokenizer/task AttributeError vs missing auxiliary losses in D-FINE; different models and mechanisms." + }, + { + "left": "issue:43475", + "right": "issue:44016", + "accept": false, + "reason": "SAM 3 video model attribute bug vs a notebook syntax error; clearly unrelated." + }, + { + "left": "issue:45561", + "right": "issue:45563", + "accept": false, + "reason": "pytest-xdist file race vs stale generation warning; testing infrastructure vs generation logic." + }, + { + "left": "issue:43824", + "right": "issue:43976", + "accept": false, + "reason": "Import/export regression for a Qwen2.5 VL symbol vs a Python support/version issue; not the same bug." + }, + { + "left": "issue:43295", + "right": "issue:44617", + "accept": false, + "reason": "Processor/tokenizer API regression vs CUDA OOM in SAM3Video; different symptoms and fixes." + }, + { + "left": "issue:43525", + "right": "issue:43881", + "accept": false, + "reason": "Missing config attribute in Llama4 vs a separate model-loading failure for GLM-4V; not one code path." + }, + { + "left": "issue:43352", + "right": "issue:45250", + "accept": false, + "reason": "Model does not support Flash Attention 2.0 vs a broad Flash Attention topic issue; not the same concrete problem." + }, + { + "left": "issue:44589", + "right": "issue:45362", + "accept": false, + "reason": "Storage serialization/type error vs a Qwen3.5 chat crash; unrelated runtime failures." + }, + { + "left": "issue:44188", + "right": "issue:44336", + "accept": false, + "reason": "torch.compile attention-kernel divergence vs ANSI code output formatting; no overlap in underlying bug." + }, + { + "left": "issue:41628", + "right": "issue:43976", + "accept": false, + "reason": "AutoImageProcessor import failure vs Python-version incompatibility; different root causes." + }, + { + "left": "issue:44295", + "right": "issue:45397", + "accept": false, + "reason": "Position_ids buffer access bug vs distributed loading issue; unrelated behavior." + }, + { + "left": "issue:43901", + "right": "issue:44246", + "accept": false, + "reason": "Pipeline documentation mismatch vs slow import performance; not the same issue." + }, + { + "left": "issue:42491", + "right": "issue:43299", + "accept": false, + "reason": "Both involve Qwen MoE/version regressions, but they target different model families and failure points, so they are not safely mergeable." + }, + { + "left": "issue:41950", + "right": "issue:44220", + "accept": false, + "reason": "Video pipeline selecting image processors vs audio fbank extraction; different subsystems." + }, + { + "left": "issue:44336", + "right": "issue:44373", + "accept": false, + "reason": "Terminal ANSI formatting bug vs wrong docstring; docs and runtime output are separate concerns." + }, + { + "left": "issue:44242", + "right": "issue:45593", + "accept": false, + "reason": "Missing load-balancing loss in MoE routing vs missing auxiliary losses in D-FINE; both are loss issues but not the same code path." + }, + { + "left": "issue:44462", + "right": "issue:45356", + "accept": false, + "reason": "AutoTokenizer ignoring tokenizer.json vs tokenizer codec/regex regression; different tokenizer loading bugs." + }, + { + "left": "issue:44661", + "right": "issue:45542", + "accept": false, + "reason": "Model-registration template failure vs missing TensorFlow backend dependency; unrelated setup problems." + }, + { + "left": "issue:44246", + "right": "issue:44496", + "accept": false, + "reason": "Slow import vs unrecognized model/config loading; different failure modes." + }, + { + "left": "issue:36246", + "right": "issue:42757", + "accept": false, + "reason": "Import error in transformers Qwen2.5 VL export vs missing huggingface_hub symbol; same symptom class, different package and symbol." + }, + { + "left": "issue:36246", + "right": "issue:44351", + "accept": false, + "reason": "Different missing exports from transformers; superficial similarity only, not one underlying bug." + }, + { + "left": "issue:43531", + "right": "issue:45362", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior vs a Qwen3.5 chat crash; not the same concrete issue." + }, + { + "left": "issue:43526", + "right": "issue:45003", + "accept": false, + "reason": "BEiT label reduction logic bug vs unsafe sys.modules access in modeling_utils; unrelated code paths." + }, + { + "left": "issue:44336", + "right": "issue:44485", + "accept": false, + "reason": "ANSI terminal output bug vs GLM-5 RoPE implementation discussion; no shared failure." + }, + { + "left": "issue:43673", + "right": "issue:44561", + "accept": false, + "reason": "Chunked-prefill cache regression vs removal of is_torch_fx_available affecting trust_remote_code; both are version regressions, but not the same fix." + }, + { + "left": "issue:44162", + "right": "issue:45200", + "accept": false, + "reason": "ESM2 regression impacting workflows vs Gemma 4 mm_token_type_ids defaulting; different models and failures." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 31, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:36246", + "issue:41628", + "issue:41950", + "issue:42886", + "issue:43065", + "issue:43066", + "issue:43232", + "issue:43257", + "issue:43262", + "issue:43295", + "issue:43408", + "issue:43441", + "issue:43475", + "issue:43540", + "issue:43596", + "issue:43653", + "issue:43901", + "issue:44060", + "issue:44077", + "issue:44162", + "issue:44242", + "issue:44246", + "issue:44295", + "issue:44336", + "issue:44368", + "issue:44466", + "issue:44496", + "issue:44749", + "issue:44779", + "issue:44857", + "issue:44871", + "issue:44877", + "issue:44933", + "issue:44977", + "issue:44987", + "issue:44995", + "issue:45003", + "issue:45137", + "issue:45200", + "issue:45229", + "issue:45292", + "issue:45335", + "issue:45375", + "issue:45459", + "issue:45563", + "issue:45584", + "issue:45588" + ], + "soft_pairs": [ + "issue:44295|issue:44749", + "issue:44242|issue:44857", + "issue:44246|issue:44871", + "issue:44877|issue:45375", + "issue:36246|issue:43596", + "issue:43066|issue:44779", + "issue:43441|issue:43540", + "issue:45292|issue:45584", + "issue:44987|issue:45459", + "issue:44779|issue:45292", + "issue:41628|issue:41950", + "issue:42886|issue:45229", + "issue:43065|issue:43408", + "issue:44995|issue:45200", + "issue:43901|issue:45003", + "issue:43232|issue:44336", + "issue:43257|issue:45137", + "issue:44496|issue:45003", + "issue:43295|issue:45292", + "issue:44995|issue:45588", + "issue:44060|issue:44077", + "issue:43653|issue:44162", + "issue:43262|issue:44246", + "issue:43475|issue:43540", + "issue:44368|issue:45563", + "issue:44466|issue:45335", + "issue:44368|issue:44977", + "issue:44933|issue:45459" + ], + "prepared_review_unit_hash": "753cb1695756858fa3cc9a12133a8900da23d25a7a536affddd195d520694a54", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35635, + "estimated_input_tokens": 8909, + "estimated_eval_tokens": 18074 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: the soft pairs share surface vocabulary (tokenizers, embeddings, generation, import/load paths) but usually describe different concrete bugs. I would not merge any of the suggested soft edges.", + "confidence": 0.89, + "canonical_issue_reason": "If a single issue must anchor the set, issue 45292 is the best fit because it is a concrete, reproducible API regression with broader applicability and external references; however, the overall cluster is not a true duplicate group.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45292 is the strongest general-purpose bug report among the items and the most plausible long-term canonical reference for the resize_token_embeddings-related threads.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44295", + "right": "issue:44749", + "accept": false, + "reason": "Both mention position/processing behavior, but one is a buffer read error and the other is a tokenizer performance regression; different code paths and symptoms." + }, + { + "left": "issue:44242", + "right": "issue:44857", + "accept": false, + "reason": "Load-balancing loss gating in MOE is unrelated to a CUDA float16 crash in a detection loss; no shared underlying bug." + }, + { + "left": "issue:44246", + "right": "issue:44871", + "accept": false, + "reason": "Import latency and eos_token_id mismatch are unrelated issues despite both being configuration/runtime complaints." + }, + { + "left": "issue:44877", + "right": "issue:45375", + "accept": false, + "reason": "Both involve config parsing/strictness, but they target different model families and different missing-field behavior." + }, + { + "left": "issue:36246", + "right": "issue:43596", + "accept": false, + "reason": "One is a missing export/import error, the other is a DeepSpeed ZeRO3 index error in BertModel; no common fix path." + }, + { + "left": "issue:43066", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer-related, but one is a wrong decoder-type warning and the other is incorrect Deepseek tokenization output; not the same defect." + }, + { + "left": "issue:43441", + "right": "issue:43540", + "accept": false, + "reason": "FlashAttention failure in Ministral and video-input validation in Qwen3OmniMoe are different model paths with different failure modes." + }, + { + "left": "issue:45292", + "right": "issue:45584", + "accept": false, + "reason": "resize_token_embeddings behavior and Whisper empty-transcription generation after align_special_tokens are separate regressions." + }, + { + "left": "issue:44987", + "right": "issue:45459", + "accept": false, + "reason": "Model-loading failure for a specific checkpoint is unrelated to protobuf exception handling in tokenizer decoding." + }, + { + "left": "issue:44779", + "right": "issue:45292", + "accept": false, + "reason": "Tokenizer incorrect results and embedding resize not propagating affect different APIs and likely different fixes." + }, + { + "left": "issue:41628", + "right": "issue:41950", + "accept": false, + "reason": "AutoImageProcessor import failure and video-classification pipeline misrouting are both image/video related but not the same bug." + }, + { + "left": "issue:42886", + "right": "issue:45229", + "accept": false, + "reason": "Offline cache loading for tokenizers is unrelated to Gemma4 multi-GPU OOM." + }, + { + "left": "issue:43065", + "right": "issue:43408", + "accept": false, + "reason": "A dummy Conv2d in SAM3 pixel decoder is not the same as a model-type mismatch warning for sam3_video vs sam3_tracker." + }, + { + "left": "issue:44995", + "right": "issue:45200", + "accept": false, + "reason": "A stale cache crash on second forward pass is unrelated to mm_token_type_ids defaulting for text-only fine-tuning." + }, + { + "left": "issue:43901", + "right": "issue:45003", + "accept": false, + "reason": "Documentation drift for return_all_scores and unsafe sys.modules access are unrelated." + }, + { + "left": "issue:43232", + "right": "issue:44336", + "accept": false, + "reason": "Generation kwargs after sync_gpus and stray ANSI output in loading_report do not share a code path." + }, + { + "left": "issue:43257", + "right": "issue:45137", + "accept": false, + "reason": "Qwen3 MOE weight conversion under accelerate/deepspeed and a ZeRO3 deque underflow are separate distributed-training bugs." + }, + { + "left": "issue:44496", + "right": "issue:45003", + "accept": false, + "reason": "Unrecognized model/config loading and unsafe sys.modules access are different model-loading failures." + }, + { + "left": "issue:43295", + "right": "issue:45292", + "accept": false, + "reason": "Processor.tokenizer regression is about multimodal processor API behavior, while 45292 is about token embedding resizing." + }, + { + "left": "issue:44995", + "right": "issue:45588", + "accept": false, + "reason": "A stale indexer cache in GLM MOE is unrelated to flash_attention.py crashing on s_aux=None." + }, + { + "left": "issue:44060", + "right": "issue:44077", + "accept": false, + "reason": "Tied-weights warning behavior and patchtsmixer post_init policy are not the same underlying issue." + }, + { + "left": "issue:43653", + "right": "issue:44162", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and ESM2 breakage are different tokenizer/model regressions." + }, + { + "left": "issue:43262", + "right": "issue:44246", + "accept": false, + "reason": "Audio chat-template sampling-rate defaulting and import latency are unrelated." + }, + { + "left": "issue:43475", + "right": "issue:43540", + "accept": false, + "reason": "Missing fpn_position_embeddings in SAM3 video output is a different bug from Qwen3OmniMoe video input processing." + }, + { + "left": "issue:44368", + "right": "issue:45563", + "accept": false, + "reason": "A tied-embeddings warning in Qwen3.5 and a stale warning for num_return_sequences are different warning-regression reports." + }, + { + "left": "issue:44466", + "right": "issue:45335", + "accept": false, + "reason": "Both concern embedding/tied-weight behavior, but one is serialization inconsistency and the other is resize_token_embeddings not updating decoder.embed_tokens; close but not the same bug." + }, + { + "left": "issue:44368", + "right": "issue:44977", + "accept": false, + "reason": "A tied-weights warning and a FlashAttention generation failure in Qwen3.5 are separate regressions." + }, + { + "left": "issue:44933", + "right": "issue:45459", + "accept": false, + "reason": "Missing import from image_utils and protobuf-related tokenizer exception masking are unrelated import/error-handling bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 32, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:28282", + "issue:30990", + "issue:41628", + "issue:42222", + "issue:42371", + "issue:42673", + "issue:42831", + "issue:42994", + "issue:43232", + "issue:43475", + "issue:43482", + "issue:43525", + "issue:43643", + "issue:43646", + "issue:43673", + "issue:43701", + "issue:43704", + "issue:43873", + "issue:43931", + "issue:44038", + "issue:44162", + "issue:44222", + "issue:44297", + "issue:44351", + "issue:44361", + "issue:44368", + "issue:44623", + "issue:44661", + "issue:44734", + "issue:44749", + "issue:44869", + "issue:44933", + "issue:44945", + "issue:45081", + "issue:45092", + "issue:45103", + "issue:45137", + "issue:45200", + "issue:45230", + "issue:45278", + "issue:45310", + "issue:45356", + "issue:45362", + "issue:45375", + "issue:45381", + "issue:45397", + "issue:45478", + "issue:45600" + ], + "soft_pairs": [ + "issue:43673|issue:44734", + "issue:44869|issue:45356", + "issue:45137|issue:45600", + "issue:43931|issue:44368", + "issue:43232|issue:43673", + "issue:44297|issue:44623", + "issue:43673|issue:45278", + "issue:43646|issue:44162", + "issue:44361|issue:44749", + "issue:42673|issue:43704", + "issue:28282|issue:30990", + "issue:44749|issue:45310", + "issue:45230|issue:45397", + "issue:44038|issue:44623", + "issue:42994|issue:43482", + "issue:43701|issue:44749", + "issue:42371|issue:45092", + "issue:43475|issue:45381", + "issue:45375|issue:45478", + "issue:43475|issue:43643", + "issue:41628|issue:44933", + "issue:45081|issue:45103", + "issue:45310|issue:45375", + "issue:42994|issue:43525", + "issue:42222|issue:45200", + "issue:43873|issue:44945", + "issue:43475|issue:44661", + "issue:42831|issue:44222", + "issue:45310|issue:45397", + "issue:45362|issue:45375", + "issue:41628|issue:44351" + ], + "prepared_review_unit_hash": "a5a66186dc29a33786f6f247e33ebae4fd65af52400a556c82617e046374a327", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36663, + "estimated_input_tokens": 9166, + "estimated_eval_tokens": 18588 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly a grab-bag of unrelated Transformers issues. Only one soft pair is a true duplicate: the Qwen3ForCausalLM VRAM leak reports. The rest are same-subsystem or same-version regressions at best, but not the same underlying bug.", + "confidence": 0.95, + "canonical_issue_reason": "issue:42673 is the better canonical because issue:43704 is an exact duplicate of the same Qwen3ForCausalLM VRAM leak in multiple dataloader threads, and 42673 has the fuller report and more discussion/references.", + "canonical_pr_reason": null, + "best_issue_reason": "Given the cluster is heterogeneous, issue:42673 is the strongest representative only because it is backed by a true duplicate pair; none of the other issues are close enough to serve as a broader cluster canonical.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43673", + "right": "issue:44734", + "accept": false, + "reason": "Both mention generation/cache behavior, but one is chunked_prefill cache handling and the other is /v1/responses KV-cache continuation indexing; different code paths." + }, + { + "left": "issue:44869", + "right": "issue:45356", + "accept": false, + "reason": "Both involve tokenizer/codec edge cases, but they are different models and different failures; no shared underlying bug is evident." + }, + { + "left": "issue:45137", + "right": "issue:45600", + "accept": false, + "reason": "DeepSpeed ZeRO3 deque failure and auto_mappings CI breakage are unrelated bugs." + }, + { + "left": "issue:43931", + "right": "issue:44368", + "accept": false, + "reason": "A weight-shape mismatch and a warning about tie_word_embeddings during LoRA fine-tuning are not the same issue." + }, + { + "left": "issue:43232", + "right": "issue:43673", + "accept": false, + "reason": "Both are generation-related, but one is sync_gpus/_update_model_kwargs_for_generation and the other is chunked_prefill cache missing; not the same defect." + }, + { + "left": "issue:44297", + "right": "issue:44623", + "accept": false, + "reason": "Tokenizer save_pretrained metadata mismatch versus processor.save_pretrained missing files are distinct save-path bugs." + }, + { + "left": "issue:43673", + "right": "issue:45278", + "accept": false, + "reason": "A specific generation cache bug is not the same as a broad set of import errors after upgrading." + }, + { + "left": "issue:43646", + "right": "issue:44162", + "accept": false, + "reason": "Generic custom model initialization breakage and the ESM2-specific regression are not the same underlying problem." + }, + { + "left": "issue:44361", + "right": "issue:44749", + "accept": false, + "reason": "MLukeTokenizer task AttributeError and a performance slowdown after upgrade are unrelated." + }, + { + "left": "issue:42673", + "right": "issue:43704", + "accept": true, + "reason": "Exact same Qwen3ForCausalLM VRAM leak in multiple dataloader threads; 43704 is a duplicate of 42673." + }, + { + "left": "issue:28282", + "right": "issue:30990", + "accept": false, + "reason": "Missing PyTorch import error and a Sentence Transformers loading hang are different failure modes." + }, + { + "left": "issue:44749", + "right": "issue:45310", + "accept": false, + "reason": "Performance regression after upgrading is unrelated to a Qwen3.5 Moe from_pretrained error." + }, + { + "left": "issue:45230", + "right": "issue:45397", + "accept": false, + "reason": "Generic placeholder bug report and gemma-4 zero3 loading failure are not the same bug." + }, + { + "left": "issue:44038", + "right": "issue:44623", + "accept": false, + "reason": "Qwen3-VL-Moe loading failure and processor.save_pretrained missing files touch different code paths." + }, + { + "left": "issue:42994", + "right": "issue:43482", + "accept": false, + "reason": "Quantized model saving failure is distinct from GGUF loading failure." + }, + { + "left": "issue:43701", + "right": "issue:44749", + "accept": false, + "reason": "Checkpoint resume key mismatch and a throughput regression are unrelated." + }, + { + "left": "issue:42371", + "right": "issue:45092", + "accept": false, + "reason": "TF32 API-setting guidance and an old remote-code checkpoint/meta-init incompatibility are different issues." + }, + { + "left": "issue:43475", + "right": "issue:45381", + "accept": false, + "reason": "SAM 3 Video missing attribute and qwen2.5-vl video position-id bug are different model-specific regressions." + }, + { + "left": "issue:45375", + "right": "issue:45478", + "accept": false, + "reason": "A missing Qwen3_5MoeVisionConfig field dropped by strict handling is not the same as a generic Qwen3.5 Moe from_pretrained error." + }, + { + "left": "issue:43475", + "right": "issue:43643", + "accept": false, + "reason": "SAM 3 Video attribute access failure and trust_remote_code missing-fields behavior are unrelated." + }, + { + "left": "issue:41628", + "right": "issue:44933", + "accept": false, + "reason": "Missing AutoImageProcessor import and a different nonexistent import from image_utils are not clearly the same bug." + }, + { + "left": "issue:45081", + "right": "issue:45103", + "accept": false, + "reason": "Tokenizer regex patch crash and a future-annotations docstring helper crash are separate AttributeError sites." + }, + { + "left": "issue:45310", + "right": "issue:45375", + "accept": false, + "reason": "Qwen3.5 Moe from_pretrained error and a missing vision-config field are different failures, even if both involve Qwen configs." + }, + { + "left": "issue:42994", + "right": "issue:43525", + "accept": false, + "reason": "Quantized save failure and missing pad_token_id on Llama4Config are unrelated." + }, + { + "left": "issue:42222", + "right": "issue:45200", + "accept": false, + "reason": "VitPose model breakage and Gemma 4 text-only fine-tuning token-type defaults are different model issues." + }, + { + "left": "issue:43873", + "right": "issue:44945", + "accept": false, + "reason": "Quantization/offloading behavior and pipeline-parallelism output corruption are different problems." + }, + { + "left": "issue:43475", + "right": "issue:44661", + "accept": false, + "reason": "SAM 3 Video attribute failure and add-new-model-like tokenizer-mapping failure are not the same code path." + }, + { + "left": "issue:42831", + "right": "issue:44222", + "accept": false, + "reason": "FineGrainedFP8 accuracy degradation and FP8 save_pretrained bugs are different FP8-related issues." + }, + { + "left": "issue:45310", + "right": "issue:45397", + "accept": false, + "reason": "Qwen3.5 Moe loading error and gemma-4 zero3 loading error are separate model regressions." + }, + { + "left": "issue:45362", + "right": "issue:45375", + "accept": false, + "reason": "Qwen3.5-35B chat crash and missing deepstack_visual_indexes in a vision config are unrelated." + }, + { + "left": "issue:41628", + "right": "issue:44351", + "accept": false, + "reason": "AutoImageProcessor import failure and HybridCache import failure are different missing-symbol regressions." + }, + { + "left": "issue:45081", + "right": "issue:45356", + "accept": false, + "reason": "Mistral regex patch crash and Kimi-K2.5 tokenizer regression are distinct tokenizer bugs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet, and the duplicate call on issue:42673/issue:43704 is supported by the identical title and report pattern. The remaining soft-edge rejections stay conservative and do not overclaim shared root causes." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 33, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:30333", + "issue:30990", + "issue:33290", + "issue:35141", + "issue:36010", + "issue:41628", + "issue:42175", + "issue:42898", + "issue:42994", + "issue:43232", + "issue:43278", + "issue:43377", + "issue:43479", + "issue:43504", + "issue:43526", + "issue:43673", + "issue:43723", + "issue:43784", + "issue:43824", + "issue:43874", + "issue:43976", + "issue:44038", + "issue:44060", + "issue:44079", + "issue:44164", + "issue:44188", + "issue:44222", + "issue:44242", + "issue:44297", + "issue:44373", + "issue:44448", + "issue:44496", + "issue:44568", + "issue:44623", + "issue:44734", + "issue:44779", + "issue:44792", + "issue:44869", + "issue:44933", + "issue:44964", + "issue:45092", + "issue:45103", + "issue:45125", + "issue:45137", + "issue:45310", + "issue:45356", + "issue:45459", + "issue:45561" + ], + "soft_pairs": [ + "issue:43232|issue:44188", + "issue:44222|issue:45310", + "issue:42994|issue:44164", + "issue:44038|issue:45310", + "issue:43232|issue:44792", + "issue:44496|issue:45561", + "issue:43526|issue:45561", + "issue:43377|issue:44242", + "issue:43504|issue:45561", + "issue:36010|issue:41628", + "issue:43673|issue:44188", + "issue:30990|issue:35141", + "issue:42898|issue:44568", + "issue:43479|issue:44079", + "issue:43479|issue:44869", + "issue:43976|issue:44933", + "issue:42175|issue:44933", + "issue:44964|issue:45137", + "issue:43784|issue:43824", + "issue:44623|issue:45310", + "issue:44188|issue:44373", + "issue:44060|issue:45125", + "issue:43723|issue:44568", + "issue:42898|issue:45356", + "issue:43278|issue:45137", + "issue:44734|issue:45103", + "issue:42898|issue:44779", + "issue:43874|issue:45092", + "issue:30333|issue:33290", + "issue:42898|issue:44448", + "issue:41628|issue:45459", + "issue:44222|issue:44297" + ], + "prepared_review_unit_hash": "dc37dc9fd1922b47509c9e4ff77430137fea80f305545fbfe8adb410c1514b9e", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37314, + "estimated_input_tokens": 9329, + "estimated_eval_tokens": 18914 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is a set of mostly unrelated transformer bugs and regressions; none of the soft pairs look like the same underlying issue, so no duplicate merges are justified.", + "confidence": 0.93, + "canonical_issue_reason": "Issue 42898 is the best single representative because it is a clear, user-facing v5 tokenizer regression with a broad and well-scoped description; however, it does not subsume the other reports.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 42898 is the strongest anchor for this cluster if one must be chosen: it is concrete, reproducible, and broadly representative of the v5 behavior-change reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43232", + "right": "issue:44188", + "accept": false, + "reason": "Both mention generation internals, but one is about sync_gpus/update kwargs and the other about torch.compile branching in attention kernels; different code paths and symptoms." + }, + { + "left": "issue:44222", + "right": "issue:45310", + "accept": false, + "reason": "Both involve MoE/FP8 wording, but one is about save_pretrained for FP8 and the other is a from_pretrained loading error for Qwen3.5 MoE; not the same bug." + }, + { + "left": "issue:42994", + "right": "issue:44164", + "accept": false, + "reason": "One is about quantized model saving failure, the other about save/from_pretrained handling extra_state; related persistence area, but distinct failure modes and fixes." + }, + { + "left": "issue:44038", + "right": "issue:45310", + "accept": false, + "reason": "Qwen3-VL-Moe transformers v5 compatibility is a different model-loading issue from Qwen3.5 MoE from_pretrained errors." + }, + { + "left": "issue:43232", + "right": "issue:44792", + "accept": false, + "reason": "Generation cache/update logic is unrelated to a failed janus image-generation test." + }, + { + "left": "issue:44496", + "right": "issue:45561", + "accept": false, + "reason": "Model config auto-discovery and pytest-xdist file races are unrelated." + }, + { + "left": "issue:43526", + "right": "issue:45561", + "accept": false, + "reason": "A BeitImageProcessorFast label-reduction bug is unrelated to a testing-utils race on captured_info.txt." + }, + { + "left": "issue:43377", + "right": "issue:44242", + "accept": false, + "reason": "Missing padding-mask support in MIMI encoder is unrelated to MoE load-balancing loss gating." + }, + { + "left": "issue:43504", + "right": "issue:45561", + "accept": false, + "reason": "A legacy-field model preset loading bug is unrelated to a test-worker file race." + }, + { + "left": "issue:36010", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but they concern different symbols and modules: GenerationMixin vs AutoImageProcessor." + }, + { + "left": "issue:43673", + "right": "issue:44188", + "accept": false, + "reason": "Chunked-prefill cache handling and compile-time attention-kernel branching are different generation bugs." + }, + { + "left": "issue:30990", + "right": "issue:35141", + "accept": false, + "reason": "Sentence Transformers loading hangs and token-embedding resizing reinitialization are unrelated." + }, + { + "left": "issue:42898", + "right": "issue:44568", + "accept": false, + "reason": "Both are tokenizer v5 regressions, but one is about clean_up_tokenization_spaces and the other about add_special_tokens/BOS-EOS behavior; different semantics and code paths." + }, + { + "left": "issue:43479", + "right": "issue:44079", + "accept": false, + "reason": "Default multimodal config initialization and ModelOutput key assignment are unrelated." + }, + { + "left": "issue:43479", + "right": "issue:44869", + "accept": false, + "reason": "Default multimodal config initialization is unrelated to Whisper word-timestamp decoding on trailing replacement characters." + }, + { + "left": "issue:43976", + "right": "issue:44933", + "accept": false, + "reason": "Python version compatibility and a missing image_utils import are different issues with no shared underlying bug." + }, + { + "left": "issue:42175", + "right": "issue:44933", + "accept": false, + "reason": "Backend dependency installation and a nonexistent image_utils import are unrelated." + }, + { + "left": "issue:44964", + "right": "issue:45137", + "accept": false, + "reason": "Phi-4 multimodal loading and DeepSpeed ZeRO3 deque underflow are different failures." + }, + { + "left": "issue:43784", + "right": "issue:43824", + "accept": false, + "reason": "Both are import failures, but one is a NameError during sentence-transformers import and the other is a missing Qwen2.5-VL class export." + }, + { + "left": "issue:44623", + "right": "issue:45310", + "accept": false, + "reason": "Processor save_pretrained missing files is unrelated to Qwen3.5 MoE from_pretrained loading errors." + }, + { + "left": "issue:44188", + "right": "issue:44373", + "accept": false, + "reason": "A generation-kernel divergence bug is unrelated to a wrong docstring for position_ids." + }, + { + "left": "issue:44060", + "right": "issue:45125", + "accept": false, + "reason": "Both mention Qwen3 MoE/tensor parallelism, but one is a tied-weights warning bug and the other is a missing _tp_plan attribute; different concrete problems." + }, + { + "left": "issue:43723", + "right": "issue:44568", + "accept": false, + "reason": "Tokenizer loading in v5 and special-token addition behavior are distinct tokenizer issues." + }, + { + "left": "issue:42898", + "right": "issue:45356", + "accept": false, + "reason": "Both are tokenizer regressions, but one is clean_up_tokenization_spaces and the other is Kimi-K2.5 codec handling with a misleading warning; not the same defect." + }, + { + "left": "issue:43278", + "right": "issue:45137", + "accept": false, + "reason": "Embedding dtype drift in evaluate and a ZeRO3 deque pop error are unrelated." + }, + { + "left": "issue:44734", + "right": "issue:45103", + "accept": false, + "reason": "KV-cache continuation indexing in serving and an auto_docstring AttributeError from future annotations are unrelated." + }, + { + "left": "issue:42898", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions in v5, but they affect different models and different mechanisms: cleanup spaces versus DeepSeek codec output." + }, + { + "left": "issue:43874", + "right": "issue:45092", + "accept": false, + "reason": "A missing image-patch counting method and an old InternVL2 meta-init incompatibility are different multimodal-loading issues." + }, + { + "left": "issue:30333", + "right": "issue:33290", + "accept": false, + "reason": "MLflow job status reporting and Deepspeed/Adafactor OOM are unrelated training issues." + }, + { + "left": "issue:42898", + "right": "issue:44448", + "accept": false, + "reason": "A tokenizer cleanup regression and a Pegasus output mismatch across v4/v5 are different behavior changes." + }, + { + "left": "issue:41628", + "right": "issue:45459", + "accept": false, + "reason": "AutoImageProcessor import failure and protobuf-related tokenizer error masking are different import/runtime issues." + }, + { + "left": "issue:44222", + "right": "issue:44297", + "accept": false, + "reason": "FP8 MoE save_pretrained behavior and tokenizer_class metadata mismatch are unrelated save-path bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 34, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:22355", + "issue:28282", + "issue:30333", + "issue:30990", + "issue:34689", + "issue:35141", + "issue:36331", + "issue:38175", + "issue:41628", + "issue:42757", + "issue:42831", + "issue:42886", + "issue:42898", + "issue:42994", + "issue:43295", + "issue:43479", + "issue:43576", + "issue:43644", + "issue:43723", + "issue:43784", + "issue:43824", + "issue:44038", + "issue:44206", + "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:44623", + "issue:44871", + "issue:44913", + "issue:44977", + "issue:44991", + "issue:45092", + "issue:45103", + "issue:45278", + "issue:45310", + "issue:45357", + "issue:45406", + "issue:45459" + ], + "soft_pairs": [ + "issue:42898|issue:43295", + "issue:43479|issue:43644", + "issue:44442|issue:45459", + "issue:44222|issue:44623", + "issue:42831|issue:44623", + "issue:44206|issue:45103", + "issue:41628|issue:42757", + "issue:44297|issue:45310", + "issue:22355|issue:28282", + "issue:44222|issue:44484", + "issue:44483|issue:45406", + "issue:43723|issue:44977", + "issue:44263|issue:44360", + "issue:30333|issue:36331", + "issue:44263|issue:45092", + "issue:30333|issue:34689", + "issue:30333|issue:35141", + "issue:28282|issue:30333", + "issue:44295|issue:44623", + "issue:44514|issue:45103", + "issue:43723|issue:44871", + "issue:43479|issue:44913", + "issue:44261|issue:45092", + "issue:43723|issue:44991", + "issue:44373|issue:44485", + "issue:43723|issue:44442", + "issue:45092|issue:45357", + "issue:30990|issue:38175", + "issue:42886|issue:42994", + "issue:42831|issue:45310", + "issue:41628|issue:43824", + "issue:42831|issue:44038", + "issue:42757|issue:45278", + "issue:43576|issue:43784" + ], + "prepared_review_unit_hash": "de99eca2efe21f3403b8e850e71dda177ea26d57ce219d326a0714d3373d2da0", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 34, + "serialized_chars": 37052, + "estimated_input_tokens": 9263, + "estimated_eval_tokens": 18782 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: most items are unrelated regressions or import/load failures in different subsystems, so the soft duplicate links are not strong enough to merge. No PRs are present.", + "confidence": 0.95, + "canonical_issue_reason": "No single canonical issue cleanly subsumes the cluster. The closest thematic anchor is issue:43723 (tokenizer loading in v5), but the surrounding items span unrelated bugs and are not duplicate-equivalent.", + "canonical_pr_reason": null, + "best_issue_reason": "If one issue must represent the set, issue:43723 is the closest representative because it captures the broad tokenizer-loading regression theme seen in a few nearby reports, but it still should not be treated as a duplicate hub.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42898", + "right": "issue:43295", + "accept": false, + "reason": "Both are v5 regression-style issues, but one is about `clean_up_tokenization_spaces` behavior and the other about `processor.tokenizer` access; different code paths and fixes." + }, + { + "left": "issue:43479", + "right": "issue:43644", + "accept": false, + "reason": "Different problems: config default initialization for Phi4Multimodal vs non-persistent buffer garbage in Transformers 5.0." + }, + { + "left": "issue:44442", + "right": "issue:45459", + "accept": false, + "reason": "Both involve tokenizer-related failures, but one is AutoTokenizer loading a specific tokenizer class and the other is protobuf error masking; not the same bug." + }, + { + "left": "issue:44222", + "right": "issue:44623", + "accept": false, + "reason": "Both mention saving, but one is FP8/MoE `save_pretrained` and the other is missing processor save files; different save paths." + }, + { + "left": "issue:42831", + "right": "issue:44623", + "accept": false, + "reason": "Accuracy loss in FineGrainedFP8 is unrelated to processor save file omissions." + }, + { + "left": "issue:44206", + "right": "issue:45103", + "accept": false, + "reason": "One is a feature extractor regression with unsupported `center`; the other is an auto-docstring annotation crash." + }, + { + "left": "issue:41628", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but they concern different missing symbols from different packages and likely different dependency/version issues." + }, + { + "left": "issue:44297", + "right": "issue:45310", + "accept": false, + "reason": "Tokenizer config mismatch on save is not the same as Qwen3.5 MoE `from_pretrained` loading failure." + }, + { + "left": "issue:22355", + "right": "issue:28282", + "accept": false, + "reason": "Different import failures: missing `transformers.onnx` vs missing PyTorch for `AutoModel`." + }, + { + "left": "issue:44222", + "right": "issue:44484", + "accept": false, + "reason": "One is a concrete FP8 MoE save failure; the other is a question about `max_shard_size` defaults." + }, + { + "left": "issue:44483", + "right": "issue:45406", + "accept": false, + "reason": "Chat-completions request rejection is unrelated to Gemma4Processor missing `_tokenizer` in serve." + }, + { + "left": "issue:43723", + "right": "issue:44977", + "accept": false, + "reason": "Both are tokenizer/model-loading regressions in v5, but one is generic `AutoTokenizer` loading and the other is Qwen3.5 flash-attention generation; not the same code-path bug." + }, + { + "left": "issue:44263", + "right": "issue:44360", + "accept": false, + "reason": "Both mention GLM DSA internals, but one is a `torch.split` return-value issue and the other is a missing ReLU in the indexer." + }, + { + "left": "issue:30333", + "right": "issue:36331", + "accept": false, + "reason": "MLflow reporting on failed jobs is unrelated to the trainer `compute_loss(num_items_in_batch)` signature break." + }, + { + "left": "issue:44263", + "right": "issue:45092", + "accept": false, + "reason": "GLM indexer behavior and InternVL2 meta-initialization incompatibility are different bugs." + }, + { + "left": "issue:30333", + "right": "issue:34689", + "accept": false, + "reason": "Training-job reporting with MLflow is unrelated to Llama 3.2 Vision model-loading regression." + }, + { + "left": "issue:30333", + "right": "issue:35141", + "accept": false, + "reason": "Failed-job reporting has nothing to do with output embeddings being reinitialized after resizing token embeddings." + }, + { + "left": "issue:28282", + "right": "issue:30333", + "accept": false, + "reason": "PyTorch-missing import error and MLflow failed-job reporting are unrelated." + }, + { + "left": "issue:44295", + "right": "issue:44623", + "accept": false, + "reason": "Registering `position_ids` as a buffer and processor save-file omissions are unrelated." + }, + { + "left": "issue:44514", + "right": "issue:45103", + "accept": false, + "reason": "Qwen2_5_VL batched chat-template crash and auto-docstring annotation crash are unrelated." + }, + { + "left": "issue:43723", + "right": "issue:44871", + "accept": false, + "reason": "Tokenizer loading in v5 and Gemma-3 EOS token config inconsistency are different issues." + }, + { + "left": "issue:43479", + "right": "issue:44913", + "accept": false, + "reason": "Phi4Multimodal config initialization and GPTNeoX rotary_pct reloading are unrelated config bugs." + }, + { + "left": "issue:44261", + "right": "issue:45092", + "accept": false, + "reason": "MLA q_a_layernorm precision error is not the same as old InternVL2 checkpoint meta-init incompatibility." + }, + { + "left": "issue:43723", + "right": "issue:44991", + "accept": false, + "reason": "Both concern tokenizer loading in v5, but the evidence is too weak to say they are the same concrete bug or fixable by one PR." + }, + { + "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 discussion." + }, + { + "left": "issue:43723", + "right": "issue:44442", + "accept": false, + "reason": "Generic tokenizer loading regression and FastSpeech2ConformerTokenizer import/load failure are different failures." + }, + { + "left": "issue:45092", + "right": "issue:45357", + "accept": false, + "reason": "Both involve Qwen3.5-family models, but one is an old remote-code meta-init incompatibility and the other is incorrect visual encoder keys on save." + }, + { + "left": "issue:30990", + "right": "issue:38175", + "accept": false, + "reason": "Sentence Transformers getting stuck loading and SigLIP2 zero-probabilities are unrelated model behaviors." + }, + { + "left": "issue:42886", + "right": "issue:42994", + "accept": false, + "reason": "Offline tokenizer cache loading and quantized model saving are different code paths." + }, + { + "left": "issue:42831", + "right": "issue:45310", + "accept": false, + "reason": "FP8 accuracy issues are unrelated to Qwen3.5 MoE `from_pretrained` loading errors." + }, + { + "left": "issue:41628", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors from transformers, but they involve different missing exported classes and likely different fixes." + }, + { + "left": "issue:42831", + "right": "issue:44038", + "accept": false, + "reason": "FineGrainedFP8 accuracy and Qwen3.5-VL-Moe bug reports are unrelated." + }, + { + "left": "issue:42757", + "right": "issue:45278", + "accept": false, + "reason": "A specific missing `huggingface_hub.is_offline_mode` import is not the same as a broad report of many import errors after upgrading to v5.5.0." + }, + { + "left": "issue:43576", + "right": "issue:43784", + "accept": false, + "reason": "Broken `transformers env` command and `sentence-transformers` import NameError are different regressions." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 35, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "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:43784", + "issue:43937", + "issue:43976", + "issue:43992", + "issue:44038", + "issue:44117", + "issue:44222", + "issue:44261", + "issue:44263", + "issue:44295", + "issue:44297", + "issue:44360", + "issue:44373", + "issue:44485", + "issue:44488", + "issue:44514", + "issue:44521", + "issue:44625", + "issue:44704", + "issue:44743", + "issue:44779", + "issue:44829", + "issue:44945", + "issue:44964", + "issue:45200", + "issue:45245", + "issue:45278", + "issue:45290", + "issue:45593", + "issue:45600" + ], + "soft_pairs": [ + "issue:43122|issue:44521", + "issue:44263|issue:44373", + "issue:43232|issue:45593", + "issue:44117|issue:44704", + "issue:42886|issue:43066", + "issue:43122|issue:44514", + "issue:42371|issue:44263", + "issue:30990|issue:33357", + "issue:44704|issue:45290", + "issue:44295|issue:44297", + "issue:42907|issue:44945", + "issue:43976|issue:44945", + "issue:39401|issue:44488", + "issue:43232|issue:44743", + "issue:43232|issue:45245", + "issue:42907|issue:43937", + "issue:43408|issue:43475", + "issue:41628|issue:45278", + "issue:43784|issue:45600", + "issue:44625|issue:44829", + "issue:44360|issue:44485", + "issue:39401|issue:44521", + "issue:43116|issue:45200", + "issue:43122|issue:44779", + "issue:44038|issue:44297", + "issue:29127|issue:30990", + "issue:44261|issue:44373", + "issue:44038|issue:44295", + "issue:42371|issue:42831", + "issue:42907|issue:44964", + "issue:44038|issue:44222", + "issue:36296|issue:43116", + "issue:42907|issue:43452", + "issue:43992|issue:44704" + ], + "prepared_review_unit_hash": "81d4856a5ac29d80c275ab3660bd5364e9a2220e904c968b21ab9310e1afec43", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 34, + "serialized_chars": 36855, + "estimated_input_tokens": 9214, + "estimated_eval_tokens": 18684 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly false-positive similarity edges across unrelated Transformers bugs; the closest cluster is around tokenizer/regression behavior in v5, but none of the candidate pairs look like true duplicates of the same concrete bug.", + "confidence": 0.23, + "canonical_issue_reason": "issue:43122 is the broadest tokenizer-regression report here and sits near several of the tokenizer/chat-template soft links, so it is the best representative issue for triage.", + "canonical_pr_reason": null, + "best_issue_reason": "Same as canonical: it is the most general issue in the nearest tokenizer-related subset, but the overall cluster is too mixed to treat as one duplicate family.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43122", + "right": "issue:44521", + "accept": false, + "reason": "Both involve tokenizer/template output, but one is a version-wide tokenization change and the other is zero assistant masks for multimodal inputs." + }, + { + "left": "issue:44263", + "right": "issue:44373", + "accept": false, + "reason": "One is a torch.split behavior bug and the other is only a wrong docstring for position_ids; not the same defect." + }, + { + "left": "issue:43232", + "right": "issue:45593", + "accept": false, + "reason": "Generation kwargs after sync_gpus and missing auxiliary losses in D-FINE are unrelated code paths." + }, + { + "left": "issue:44117", + "right": "issue:44704", + "accept": false, + "reason": "Both touch from_pretrained internals, but one is TOKENIZER_MAPPING_NAMES returning None and the other is cached_file kwargs passthrough." + }, + { + "left": "issue:42886", + "right": "issue:43066", + "accept": false, + "reason": "Offline cache loading failure and wrong tokenizer decoder type are different tokenizer bugs." + }, + { + "left": "issue:43122", + "right": "issue:44514", + "accept": false, + "reason": "Both are tokenizer-related, but this is a batched chat-template crash, not the same regression as version-to-version tokenization differences." + }, + { + "left": "issue:42371", + "right": "issue:44263", + "accept": false, + "reason": "TF32 behavior settings and a torch.split return-value issue are unrelated." + }, + { + "left": "issue:30990", + "right": "issue:33357", + "accept": false, + "reason": "Sentence-transformers loading hang and a MacOS bus error with a CLIP model are different failure modes." + }, + { + "left": "issue:44704", + "right": "issue:45290", + "accept": false, + "reason": "AutoProcessor kwargs forwarding and chat-template crashes with tool-call messages are separate bugs." + }, + { + "left": "issue:44295", + "right": "issue:44297", + "accept": false, + "reason": "Reading a buffer-backed position_ids tensor and save_pretrained writing the wrong tokenizer_class are different issues." + }, + { + "left": "issue:42907", + "right": "issue:44945", + "accept": false, + "reason": "Saving dequantized Ministrals/Devstrals and incorrect output under pipeline parallelism do not share the same root cause." + }, + { + "left": "issue:43976", + "right": "issue:44945", + "accept": false, + "reason": "Python version compatibility and pipeline-parallel output corruption are unrelated." + }, + { + "left": "issue:39401", + "right": "issue:44488", + "accept": false, + "reason": "Qwen3 offset_mapping errors and failure to load a specific tokenizer/model are not the same bug." + }, + { + "left": "issue:43232", + "right": "issue:44743", + "accept": false, + "reason": "These are both generation/cache related, but one is sync_gpus kwargs handling and the other is recurrent state reset with cache." + }, + { + "left": "issue:43232", + "right": "issue:45245", + "accept": false, + "reason": "Generation kwargs handling and a categories-limit runtime error are unrelated." + }, + { + "left": "issue:42907", + "right": "issue:43937", + "accept": false, + "reason": "Dequantized save failures and invalid GenerationConfig for GLM-5 are different issues." + }, + { + "left": "issue:43408", + "right": "issue:43475", + "accept": false, + "reason": "A sam3_video/sam3_tracker type warning is not the same as the missing fpn_position_embeddings attribute." + }, + { + "left": "issue:41628", + "right": "issue:45278", + "accept": false, + "reason": "A specific missing AutoImageProcessor import is not enough to equate with broad import failures after upgrade." + }, + { + "left": "issue:43784", + "right": "issue:45600", + "accept": false, + "reason": "Sentence-transformers import NameError and Sam3LiteText auto_mappings breakage are unrelated." + }, + { + "left": "issue:44625", + "right": "issue:44829", + "accept": false, + "reason": "Qwen3.5 num_labels propagation and flash_attention_3 training degeneration are different problems." + }, + { + "left": "issue:44360", + "right": "issue:44485", + "accept": false, + "reason": "DSA indexer missing ReLU and GLM-5 RoPE implementation discussion are unrelated." + }, + { + "left": "issue:39401", + "right": "issue:44521", + "accept": false, + "reason": "Offset_mapping correctness and all-zero assistant_masks are both tokenizer-related, but they are distinct bugs." + }, + { + "left": "issue:43116", + "right": "issue:45200", + "accept": false, + "reason": "Multi-label example-script behavior and Gemma 4 mm_token_type_ids defaults are separate issues." + }, + { + "left": "issue:43122", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenization regressions in v5, but they affect different models and symptoms, so they are not the same bug." + }, + { + "left": "issue:44038", + "right": "issue:44297", + "accept": false, + "reason": "A Qwen3-VL-Moe bug and tokenizer_class mismatch on save_pretrained are unrelated." + }, + { + "left": "issue:29127", + "right": "issue:30990", + "accept": false, + "reason": "LayoutLMv3 error-message clarity and Sentence Transformers loading hang are unrelated." + }, + { + "left": "issue:44261", + "right": "issue:44373", + "accept": false, + "reason": "MLA q_a_layernorm precision/config issue and wrong position_ids docstring are not the same defect." + }, + { + "left": "issue:44038", + "right": "issue:44295", + "accept": false, + "reason": "Qwen3-VL-Moe and position_ids buffer read failure are different bugs." + }, + { + "left": "issue:42371", + "right": "issue:42831", + "accept": false, + "reason": "TF32 API-setting guidance and FineGrainedFP8 accuracy issues are not the same root cause." + }, + { + "left": "issue:42907", + "right": "issue:44964", + "accept": false, + "reason": "Dequantized save failure and inability to load Phi-4 multimodal are unrelated." + }, + { + "left": "issue:44038", + "right": "issue:44222", + "accept": false, + "reason": "Qwen3-VL-Moe regression and FP8 save_pretrained moe are only loosely related by model family, not the same bug." + }, + { + "left": "issue:36296", + "right": "issue:43116", + "accept": false, + "reason": "Tensor parallel training bug and multi-label classification example failure are unrelated." + }, + { + "left": "issue:42907", + "right": "issue:43452", + "accept": false, + "reason": "Dequantized Ministrals/Devstrals save failure and gguf_file load failure are different code paths." + }, + { + "left": "issue:43992", + "right": "issue:44704", + "accept": false, + "reason": "Missing embed_tokens.weight in UMT5Encoder and AutoProcessor cached_file kwargs handling are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 36, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:22355", + "issue:29127", + "issue:33453", + "issue:38617", + "issue:39401", + "issue:42222", + "issue:42371", + "issue:42673", + "issue:42757", + "issue:42907", + "issue:43064", + "issue:43116", + "issue:43122", + "issue:43278", + "issue:43441", + "issue:43493", + "issue:43673", + "issue:43784", + "issue:43824", + "issue:43825", + "issue:43827", + "issue:43874", + "issue:43906", + "issue:43931", + "issue:43937", + "issue:43976", + "issue:44038", + "issue:44075", + "issue:44117", + "issue:44222", + "issue:44291", + "issue:44351", + "issue:44485", + "issue:44779", + "issue:44936", + "issue:44945", + "issue:44977", + "issue:45071", + "issue:45072", + "issue:45083", + "issue:45216", + "issue:45278", + "issue:45341", + "issue:45446", + "issue:45478", + "issue:45561", + "issue:45563", + "issue:45588" + ], + "soft_pairs": [ + "issue:45216|issue:45478", + "issue:43824|issue:44351", + "issue:43122|issue:44977", + "issue:42907|issue:44222", + "issue:42757|issue:44351", + "issue:43278|issue:44945", + "issue:43784|issue:45446", + "issue:42222|issue:43441", + "issue:42673|issue:45072", + "issue:38617|issue:45278", + "issue:43064|issue:43673", + "issue:43937|issue:44485", + "issue:22355|issue:29127", + "issue:42371|issue:43874", + "issue:43116|issue:43493", + "issue:44075|issue:44936", + "issue:44222|issue:45478", + "issue:44038|issue:45478", + "issue:43116|issue:43825", + "issue:43976|issue:45446", + "issue:43937|issue:45563", + "issue:39401|issue:44779", + "issue:33453|issue:39401", + "issue:43931|issue:44945", + "issue:45083|issue:45446", + "issue:44117|issue:45588", + "issue:43827|issue:45446", + "issue:43937|issue:44977", + "issue:45341|issue:45561", + "issue:42673|issue:43906", + "issue:44291|issue:45071" + ], + "prepared_review_unit_hash": "7d59db835bee6fee1d619ce0b55f7178f9d391d16471c5515d84703acded473f", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36612, + "estimated_input_tokens": 9153, + "estimated_eval_tokens": 18562 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is highly heterogeneous: most pairs share only broad symptoms (imports, tokenization, save/load, training) or a common subsystem, but not the same concrete bug. I reject all soft edges and see no true duplicate cluster here.", + "confidence": 0.84, + "canonical_issue_reason": "No single issue cleanly represents a duplicate cluster. If forced, issue:45278 is the broadest umbrella, but the items are too diverse for a real canonical duplicate.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 is the closest thing to a cluster anchor because it describes a broad import-regression symptom, but it is still too generic to serve as a strong canonical for the whole set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45216", + "right": "issue:45478", + "accept": false, + "reason": "Same model family and version range, but one is save_pretrained output corruption and the other is from_pretrained loading failure; not the same bug." + }, + { + "left": "issue:43824", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but for different missing symbols and likely different breakages." + }, + { + "left": "issue:43122", + "right": "issue:44977", + "accept": false, + "reason": "Both concern tokenizer behavior, but one is a generic version-to-version change and the other is a DeepSeek-specific regression." + }, + { + "left": "issue:42907", + "right": "issue:44222", + "accept": false, + "reason": "Both mention saving/quantized models, but they are different model families and different failure modes." + }, + { + "left": "issue:42757", + "right": "issue:44351", + "accept": false, + "reason": "Both are import failures, but they involve different packages/symbols and different root causes." + }, + { + "left": "issue:43278", + "right": "issue:44945", + "accept": false, + "reason": "Different areas: dtype change in evaluation versus incorrect outputs under pipeline parallelism." + }, + { + "left": "issue:43784", + "right": "issue:45446", + "accept": false, + "reason": "One is a missing internal name during import; the other is a bad version gate for a flex_attention import." + }, + { + "left": "issue:42222", + "right": "issue:43441", + "accept": false, + "reason": "Different model families and different subsystems: VitPose breakage versus FlashAttention failure." + }, + { + "left": "issue:42673", + "right": "issue:45072", + "accept": false, + "reason": "VRAM leak in Qwen3ForCausalLM is unrelated to dtype mismatches in SwitchTransformers/TimmWrapperModel." + }, + { + "left": "issue:38617", + "right": "issue:45278", + "accept": false, + "reason": "Both are import-related regressions, but one is a single missing symbol while the other is a broad update report; not the same concrete bug." + }, + { + "left": "issue:43064", + "right": "issue:43673", + "accept": false, + "reason": "Training optimizer-state corruption is unrelated to a missing generation cache in chunked prefill." + }, + { + "left": "issue:43937", + "right": "issue:44485", + "accept": false, + "reason": "GenerationConfig validation failure is unrelated to a RoPE implementation discussion." + }, + { + "left": "issue:22355", + "right": "issue:29127", + "accept": false, + "reason": "Completely different failures: missing transformers.onnx module versus unclear LayoutLMv3 error message." + }, + { + "left": "issue:42371", + "right": "issue:43874", + "accept": false, + "reason": "TF32 settings advice and missing multimodal-token helper are unrelated." + }, + { + "left": "issue:43116", + "right": "issue:43493", + "accept": false, + "reason": "Example-script classification bug versus SigLIP2 implementation discrepancy; different code paths." + }, + { + "left": "issue:44075", + "right": "issue:44936", + "accept": false, + "reason": "Optimizer argument handling and trainer.evaluate after train are different bugs." + }, + { + "left": "issue:44222", + "right": "issue:45478", + "accept": false, + "reason": "Both involve Qwen3.5/MoE save/load flows, but one is FP8 save_pretrained and the other is from_pretrained loading; not the same defect." + }, + { + "left": "issue:44038", + "right": "issue:45478", + "accept": false, + "reason": "Same broad Qwen3.5 area, but different symptoms and likely different failing code paths." + }, + { + "left": "issue:43116", + "right": "issue:43825", + "accept": false, + "reason": "Classification example failure is unrelated to a pipeline() error message regression." + }, + { + "left": "issue:43976", + "right": "issue:45446", + "accept": false, + "reason": "Python-version compatibility issue versus a flex_attention import guard bug." + }, + { + "left": "issue:43937", + "right": "issue:45563", + "accept": false, + "reason": "GenerationConfig invalidity and a stale warning about num_return_sequences are different issues." + }, + { + "left": "issue:39401", + "right": "issue:44779", + "accept": false, + "reason": "Both tokenizer-related, but one is offset_mapping and the other is generic incorrect tokenization in DeepSeek." + }, + { + "left": "issue:33453", + "right": "issue:39401", + "accept": false, + "reason": "Tokenizer-loading regression and Qwen3 offset_mapping bug are not the same concrete failure." + }, + { + "left": "issue:43931", + "right": "issue:44945", + "accept": false, + "reason": "Model weight-shape mismatch and pipeline-parallelism incorrect output are unrelated." + }, + { + "left": "issue:45083", + "right": "issue:45446", + "accept": false, + "reason": "A helper-function behavior change in qwen3_omni_moe is unrelated to a PyTorch version check bug." + }, + { + "left": "issue:44117", + "right": "issue:45588", + "accept": false, + "reason": "Tokenizer mapping fallback bug and flash_attention s_aux=None crash are different subsystems and failures." + }, + { + "left": "issue:43827", + "right": "issue:45446", + "accept": false, + "reason": "Documentation mentioning pipeline() is unrelated to a runtime version-check import bug." + }, + { + "left": "issue:43937", + "right": "issue:44977", + "accept": false, + "reason": "Different generation/tokenization problems with different model contexts." + }, + { + "left": "issue:45341", + "right": "issue:45561", + "accept": false, + "reason": "A vague testing_utils bug is not enough to merge with a specific xdist race on captured_info.txt." + }, + { + "left": "issue:42673", + "right": "issue:43906", + "accept": false, + "reason": "VRAM leak in Qwen3ForCausalLM is unrelated to a reproduction of another issue." + }, + { + "left": "issue:44291", + "right": "issue:45071", + "accept": false, + "reason": "Unexpected _is_hf_initialized argument during init_empty_weights is unrelated to a PretrainedConfig type-checking breakage." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 37, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:38617", + "issue:39401", + "issue:41628", + "issue:42175", + "issue:42222", + "issue:42371", + "issue:42617", + "issue:42673", + "issue:43012", + "issue:43066", + "issue:43097", + "issue:43122", + "issue:43408", + "issue:43421", + "issue:43441", + "issue:43450", + "issue:43577", + "issue:43582", + "issue:43673", + "issue:43742", + "issue:43874", + "issue:43931", + "issue:43937", + "issue:43986", + "issue:44117", + "issue:44220", + "issue:44261", + "issue:44263", + "issue:44410", + "issue:44483", + "issue:44484", + "issue:44485", + "issue:44493", + "issue:44496", + "issue:44829", + "issue:44933", + "issue:45071", + "issue:45081", + "issue:45216", + "issue:45278", + "issue:45310", + "issue:45341", + "issue:45405", + "issue:45446", + "issue:45459", + "issue:45464", + "issue:45478", + "issue:45542" + ], + "soft_pairs": [ + "issue:45405|issue:45446", + "issue:39401|issue:43421", + "issue:43450|issue:43986", + "issue:44484|issue:45310", + "issue:44829|issue:45216", + "issue:38617|issue:41628", + "issue:44410|issue:45310", + "issue:44117|issue:45081", + "issue:43408|issue:43742", + "issue:42673|issue:43582", + "issue:42222|issue:45542", + "issue:42371|issue:44261", + "issue:42673|issue:43577", + "issue:43937|issue:44496", + "issue:38617|issue:44220", + "issue:43012|issue:43577", + "issue:42175|issue:42222", + "issue:44483|issue:45464", + "issue:44933|issue:45278", + "issue:44484|issue:45478", + "issue:42617|issue:45341", + "issue:43066|issue:43122", + "issue:43673|issue:45341", + "issue:44410|issue:45478", + "issue:43097|issue:44493", + "issue:43874|issue:44261", + "issue:42222|issue:45459", + "issue:43441|issue:44829", + "issue:43986|issue:44484", + "issue:43012|issue:43408", + "issue:44263|issue:44485", + "issue:41628|issue:45071", + "issue:45071|issue:45542", + "issue:43931|issue:44410" + ], + "prepared_review_unit_hash": "ac0bbe99f108a1aa6ec2458e234c56532bc2602ef102b24057fc7106ad4908c8", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37633, + "estimated_input_tokens": 9409, + "estimated_eval_tokens": 19074 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab bag of unrelated issues. The only clearly supported duplicate set is the Qwen3.5 MoE `from_pretrained` regression between 45310 and 45478. Issue 44484 is about `save_pretrained()` / `max_shard_size` and is not, by title alone, enough to treat it as the same bug.", + "confidence": 0.82, + "canonical_issue_reason": "issue:45478 is the clearest representative of the Qwen3.5 MoE `from_pretrained` regression; it matches issue:45310 closely and has slightly stronger discussion signal. Issue:44484 does not have enough evidence to join that duplicate set.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45478 is the best issue to keep as the representative of the only clearly duplicated bug in the cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45405", + "right": "issue:45446", + "accept": false, + "reason": "Both are version/dependency-related, but one is about PEFT pinning and the other is about an import/version check for AuxRequest; different bugs." + }, + { + "left": "issue:39401", + "right": "issue:43421", + "accept": false, + "reason": "Both concern tokenizer behavior, but one is wrong `offset_mapping` and the other is runtime special-token post-processor updates; not the same issue." + }, + { + "left": "issue:43450", + "right": "issue:43986", + "accept": false, + "reason": "One is a video-processor batch shape bug; the other is a crash when `torchvision` is missing. Different code paths and symptoms." + }, + { + "left": "issue:44484", + "right": "issue:45310", + "accept": false, + "reason": "`44484` is about `save_pretrained()` / `max_shard_size`, while `45310` is a Qwen3.5 MoE `from_pretrained` error. The titles do not establish the same underlying bug." + }, + { + "left": "issue:44829", + "right": "issue:45216", + "accept": false, + "reason": "FlashAttention-3 training degeneration and Qwen3.5 checkpoint save corruption are distinct problems." + }, + { + "left": "issue:38617", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but for different symbols (`layer_type_validation` vs `AutoImageProcessor`)." + }, + { + "left": "issue:44410", + "right": "issue:45310", + "accept": false, + "reason": "One is missing projections in qwen3next layer 0; the other is a Qwen3.5 MoE loading regression. Different bugs." + }, + { + "left": "issue:44117", + "right": "issue:45081", + "accept": false, + "reason": "`TOKENIZER_MAPPING_NAMES` returning `None` and the Mistral regex patch crash are unrelated tokenizer internals." + }, + { + "left": "issue:43408", + "right": "issue:43742", + "accept": false, + "reason": "A model-type warning for sam3_tracker is not the same as a key error when loading MobileLLM." + }, + { + "left": "issue:42673", + "right": "issue:43582", + "accept": false, + "reason": "A VRAM leak in Qwen3ForCausalLM multi-threaded loading is unrelated to an Apple Silicon allocator `TypeError`." + }, + { + "left": "issue:42222", + "right": "issue:45542", + "accept": false, + "reason": "VitPose breakage and undefined TF backend errors are different dependency/model-loading problems." + }, + { + "left": "issue:42371", + "right": "issue:44261", + "accept": false, + "reason": "TF32 behavior/settings guidance is unrelated to the MLA `rms_norm_eps` precision bug." + }, + { + "left": "issue:42673", + "right": "issue:43577", + "accept": false, + "reason": "The Qwen3 VRAM leak and BLIP2 dtype staying float32 are both about memory/precision, but not the same bug." + }, + { + "left": "issue:43937", + "right": "issue:44496", + "accept": false, + "reason": "GLM-5 invalid `GenerationConfig` and OLMo-Hybrid unrecognized model/config are distinct loading/configuration failures." + }, + { + "left": "issue:38617", + "right": "issue:44220", + "accept": false, + "reason": "An import error from `configuration_utils` is not the same as the `_torch_extract_fbank_features()` issue." + }, + { + "left": "issue:43012", + "right": "issue:43577", + "accept": false, + "reason": "Both involve dtype/precision, but one is a compile-time warning and the other is incorrect loading dtypes in BLIP2." + }, + { + "left": "issue:42175", + "right": "issue:42222", + "accept": false, + "reason": "Missing TensorFlow in the `[torch]` install and broken VitPose models are unrelated." + }, + { + "left": "issue:44483", + "right": "issue:45464", + "accept": false, + "reason": "Both mention chat/completions, but one is a generic endpoint request failure and the other is a model-specific streaming inference problem." + }, + { + "left": "issue:44933", + "right": "issue:45278", + "accept": false, + "reason": "A specific missing import from `image_utils` is not enough to equate it with the broader import-error report after upgrading to 5.5.0." + }, + { + "left": "issue:44484", + "right": "issue:45478", + "accept": false, + "reason": "`44484` is about `save_pretrained()` / `max_shard_size`, while `45478` is a Qwen3.5 MoE `from_pretrained` error. The evidence does not support treating them as the same bug." + }, + { + "left": "issue:42617", + "right": "issue:45341", + "accept": false, + "reason": "`3d_parallel.py` execution failure and a `testing_utils.py` bug are unrelated." + }, + { + "left": "issue:43066", + "right": "issue:43122", + "accept": false, + "reason": "Both are tokenizer regression reports, but one targets decoder-type metadata and the other reports changed tokenization output; not clearly the same concrete bug." + }, + { + "left": "issue:43673", + "right": "issue:45341", + "accept": false, + "reason": "Missing generation cache in chunked prefill and a testing utility bug do not share the same code path." + }, + { + "left": "issue:44410", + "right": "issue:45478", + "accept": false, + "reason": "qwen3next missing attention projections is unrelated to the Qwen3.5 MoE loading regression." + }, + { + "left": "issue:43097", + "right": "issue:44493", + "accept": false, + "reason": "Removal of `tie_embeddings_and_encoder_decoder` and unexpected `position_id` keys are different regressions." + }, + { + "left": "issue:43874", + "right": "issue:44261", + "accept": false, + "reason": "A missing image-patch helper and an `rms_norm_eps` precision mismatch are unrelated bugs." + }, + { + "left": "issue:41628", + "right": "issue:45071", + "accept": false, + "reason": "A missing top-level import and a `PreTrainedConfig` type-checking breakage are different failures." + }, + { + "left": "issue:44263", + "right": "issue:44485", + "accept": false, + "reason": "Both are GLM-related, but one concerns `torch.split` handling and the other RoPE implementation; not the same issue." + }, + { + "left": "issue:45071", + "right": "issue:45542", + "accept": false, + "reason": "`PreTrainedConfig` type checking and undefined TF backend errors are unrelated." + }, + { + "left": "issue:43931", + "right": "issue:44410", + "accept": false, + "reason": "Qwen3-VL weight-shape mismatch and qwen3next missing projections are different model-loading bugs." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The cluster summary is too strong about the Qwen3.5 MoE duplicate set. In the packet, 45310 and 45478 clearly match on `Qwen3.5 Moe from_pretrained error`, but 44484\u2019s title is about `save_pretrained()` / `max_shard_size` and does not, by itself, ground the claim that it is the same bug. The accepted soft-edge verdicts involving 44484 therefore overstate the evidence. The rest of the rejected pairs look conservative, but I can\u2019t accept the cluster with that unsupported equivalence." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 38, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:38617", + "issue:42757", + "issue:42886", + "issue:43012", + "issue:43064", + "issue:43066", + "issue:43097", + "issue:43335", + "issue:43352", + "issue:43381", + "issue:43408", + "issue:43452", + "issue:43502", + "issue:43531", + "issue:43550", + "issue:43824", + "issue:43825", + "issue:43873", + "issue:43874", + "issue:43906", + "issue:43931", + "issue:43976", + "issue:43986", + "issue:43992", + "issue:44038", + "issue:44062", + "issue:44117", + "issue:44261", + "issue:44263", + "issue:44315", + "issue:44351", + "issue:44368", + "issue:44380", + "issue:44410", + "issue:44484", + "issue:44485", + "issue:44655", + "issue:44704", + "issue:44829", + "issue:44987", + "issue:45003", + "issue:45071", + "issue:45103", + "issue:45245", + "issue:45341", + "issue:45372", + "issue:45375", + "issue:45588" + ], + "soft_pairs": [ + "issue:45003|issue:45341", + "issue:44117|issue:44987", + "issue:43986|issue:45341", + "issue:43097|issue:43452", + "issue:43931|issue:44038", + "issue:43335|issue:45071", + "issue:43408|issue:44368", + "issue:43874|issue:44263", + "issue:43976|issue:45071", + "issue:44829|issue:45245", + "issue:43986|issue:45372", + "issue:43064|issue:44315", + "issue:43531|issue:45341", + "issue:38617|issue:42757", + "issue:44261|issue:44485", + "issue:44410|issue:45375", + "issue:38617|issue:44351", + "issue:44038|issue:44410", + "issue:44062|issue:44117", + "issue:43097|issue:45588", + "issue:43986|issue:44117", + "issue:44704|issue:45103", + "issue:43502|issue:43992", + "issue:43992|issue:44484", + "issue:43824|issue:43986", + "issue:43097|issue:43873", + "issue:43066|issue:43906", + "issue:43352|issue:44380", + "issue:43825|issue:44655", + "issue:38617|issue:43097", + "issue:43012|issue:44368", + "issue:42886|issue:43381", + "issue:43012|issue:43550" + ], + "prepared_review_unit_hash": "617e895a6f98e5fe2713f014fb05f78b1ef8463d1823f84152e4298749265309", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37573, + "estimated_input_tokens": 9394, + "estimated_eval_tokens": 19044 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly a noisy mix of unrelated Transformers v5 regressions and model-specific bugs. I do not see any soft pair that is the same underlying bug/change, so all candidate edges are rejected. The closest representative issue is the v5 breaking-change report about `tie_embeddings_and_encoder_decoder` being removed.", + "confidence": 0.66, + "canonical_issue_reason": "Issue 43097 is the most cluster-like representative: it describes a concrete, cross-cutting v5 API/config break with inbound references, while most other issues are narrower model- or dependency-specific failures.", + "canonical_pr_reason": null, + "best_issue_reason": "43097 is the best anchor for this set because it captures a broad Transformers v5 regression that plausibly sits near the center of the cluster\u2019s general theme of breaking changes.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45003", + "right": "issue:45341", + "accept": false, + "reason": "Different bug classes: unsafe `sys.modules` access vs a testing utility issue." + }, + { + "left": "issue:44117", + "right": "issue:44987", + "accept": false, + "reason": "Tokenizer mapping fallback bug vs a model-loading failure for `physical-intelligence/fast`." + }, + { + "left": "issue:43986", + "right": "issue:45341", + "accept": false, + "reason": "AutoProcessor/video-model crash without torchvision vs unrelated testing utility bug." + }, + { + "left": "issue:43097", + "right": "issue:43452", + "accept": false, + "reason": "Removed config flag warning vs `gguf_file` loading breakage; different code paths." + }, + { + "left": "issue:43931", + "right": "issue:44038", + "accept": false, + "reason": "Weight-shape mismatch for one model vs a separate Qwen3-VL-Moe loading problem." + }, + { + "left": "issue:43335", + "right": "issue:45071", + "accept": false, + "reason": "SwitchTransformers layer creation bug vs PretrainedConfig type-checking regression." + }, + { + "left": "issue:43408", + "right": "issue:44368", + "accept": false, + "reason": "Model-type warning mismatch vs tie_word_embeddings warning; not the same failure." + }, + { + "left": "issue:43874", + "right": "issue:44263", + "accept": false, + "reason": "Missing image-patch method vs incorrect `torch.split()` return handling." + }, + { + "left": "issue:43976", + "right": "issue:45071", + "accept": false, + "reason": "Python version compatibility issue vs config type-checking failure." + }, + { + "left": "issue:44829", + "right": "issue:45245", + "accept": false, + "reason": "FA3 training degeneration vs category-cardinality runtime limit; unrelated." + }, + { + "left": "issue:43986", + "right": "issue:45372", + "accept": false, + "reason": "Missing torchvision dependency vs `mistral_common` import failure; different models/dependencies." + }, + { + "left": "issue:43064", + "right": "issue:44315", + "accept": false, + "reason": "FSDP2 optimizer-state corruption vs Liger Kernel not being applied." + }, + { + "left": "issue:43531", + "right": "issue:45341", + "accept": false, + "reason": "Qwen3-MoE `sliding_window` behavior vs a testing utility bug." + }, + { + "left": "issue:38617", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but for different missing names and different modules." + }, + { + "left": "issue:44261", + "right": "issue:44485", + "accept": false, + "reason": "RMS norm epsilon precision issue vs GLM-5 RoPE implementation." + }, + { + "left": "issue:44410", + "right": "issue:45375", + "accept": false, + "reason": "Missing attention projections vs missing config field in a different model family." + }, + { + "left": "issue:38617", + "right": "issue:44351", + "accept": false, + "reason": "Different import failures: `layer_type_validation` vs `HybridCache`." + }, + { + "left": "issue:44038", + "right": "issue:44410", + "accept": false, + "reason": "Qwen3-VL-Moe loading bug vs qwen3next missing projections; not the same underlying defect." + }, + { + "left": "issue:44062", + "right": "issue:44117", + "accept": false, + "reason": "`AddedToken` argument duplication vs tokenizer mapping returning None." + }, + { + "left": "issue:43097", + "right": "issue:45588", + "accept": false, + "reason": "Config-removal warning vs flash_attention crash on `s_aux=None`; different bugs." + }, + { + "left": "issue:43986", + "right": "issue:44117", + "accept": false, + "reason": "Video AutoProcessor dependency crash vs tokenizer mapping fallback bug." + }, + { + "left": "issue:44704", + "right": "issue:45103", + "accept": false, + "reason": "Kwargs not forwarded to `cached_file` vs docstring/annotation-related AttributeError." + }, + { + "left": "issue:43502", + "right": "issue:43992", + "accept": false, + "reason": "Offline/network leakage bug vs missing `embed_tokens.weight` during loading." + }, + { + "left": "issue:43992", + "right": "issue:44484", + "accept": false, + "reason": "Missing state-dict key vs `max_shard_size` default question." + }, + { + "left": "issue:43824", + "right": "issue:43986", + "accept": false, + "reason": "Missing model symbol import vs video processor crash without torchvision." + }, + { + "left": "issue:43097", + "right": "issue:43873", + "accept": false, + "reason": "Removed config attribute warning vs quantization offloading behavior." + }, + { + "left": "issue:43066", + "right": "issue:43906", + "accept": false, + "reason": "Wrong tokenizer decoder type vs a reproduction of a different issue." + }, + { + "left": "issue:43352", + "right": "issue:44380", + "accept": false, + "reason": "Flash Attention support gating vs GPT2 attention scaling being ignored." + }, + { + "left": "issue:43825", + "right": "issue:44655", + "accept": false, + "reason": "Pipeline translation-task messaging vs inability to save Pipeline objects." + }, + { + "left": "issue:38617", + "right": "issue:43097", + "accept": false, + "reason": "ImportError for a missing symbol vs a removed config field warning." + }, + { + "left": "issue:44261", + "right": "issue:44485", + "accept": false, + "reason": "Precision mismatch in MLA q_a_layernorm vs GLM-5 RoPE implementation." + }, + { + "left": "issue:43012", + "right": "issue:44368", + "accept": false, + "reason": "bfloat16 compilation warning vs tie_word_embeddings warning." + }, + { + "left": "issue:42886", + "right": "issue:43381", + "accept": false, + "reason": "Offline cache-loading failure vs gradient checkpointing being disallowed in eval mode." + }, + { + "left": "issue:43012", + "right": "issue:43550", + "accept": false, + "reason": "PyTorch precision warning vs Bamba-9B-v2 compile/SDPA failure." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 39, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:36010", + "issue:36683", + "issue:38617", + "issue:42757", + "issue:42907", + "issue:42994", + "issue:43012", + "issue:43064", + "issue:43066", + "issue:43208", + "issue:43299", + "issue:43352", + "issue:43475", + "issue:43502", + "issue:43525", + "issue:43526", + "issue:43550", + "issue:43582", + "issue:43618", + "issue:43756", + "issue:43761", + "issue:43824", + "issue:43867", + "issue:43901", + "issue:43937", + "issue:43976", + "issue:43992", + "issue:44038", + "issue:44077", + "issue:44188", + "issue:44220", + "issue:44230", + "issue:44263", + "issue:44380", + "issue:44483", + "issue:44485", + "issue:44492", + "issue:44514", + "issue:44655", + "issue:44683", + "issue:44938", + "issue:45216", + "issue:45290", + "issue:45310", + "issue:45381", + "issue:45440", + "issue:45479", + "issue:45593" + ], + "soft_pairs": [ + "issue:42907|issue:44655", + "issue:43208|issue:43550", + "issue:44380|issue:45479", + "issue:43352|issue:43976", + "issue:43064|issue:44380", + "issue:44483|issue:44514", + "issue:43208|issue:43526", + "issue:43618|issue:44380", + "issue:36683|issue:43582", + "issue:43618|issue:43761", + "issue:44483|issue:45290", + "issue:43012|issue:44683", + "issue:42757|issue:43824", + "issue:44038|issue:44230", + "issue:44483|issue:45381", + "issue:43299|issue:43867", + "issue:38617|issue:43502", + "issue:43756|issue:44485", + "issue:43901|issue:44655", + "issue:42994|issue:44655", + "issue:43867|issue:45216", + "issue:36683|issue:43475", + "issue:44220|issue:44655", + "issue:43066|issue:45381", + "issue:43352|issue:43937", + "issue:44077|issue:44380", + "issue:43064|issue:44492", + "issue:44263|issue:45440", + "issue:43976|issue:44938", + "issue:36683|issue:44077", + "issue:36683|issue:43525", + "issue:36010|issue:42757", + "issue:44188|issue:44380", + "issue:43618|issue:45216", + "issue:43352|issue:45593", + "issue:43992|issue:45310" + ], + "prepared_review_unit_hash": "189834f88c865a0bd70f83c6075a362629939539de30f3be8e7505f5a68fc38c", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 36, + "serialized_chars": 38094, + "estimated_input_tokens": 9524, + "estimated_eval_tokens": 19304 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly heterogeneous: many title-level similarities (imports, saving, Qwen/CLIP/model regressions) but the underlying bugs and code paths differ. I would not merge any soft pair as a duplicate.", + "confidence": 0.88, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "issue:43502 is the strongest standalone representative here: concrete, reproducible API regression, broad user impact, and more discussion/inbound references than most others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42907", + "right": "issue:44655", + "accept": false, + "reason": "Both involve saving, but one is dequantized model saving and the other is Pipeline.save_pretrained; different failure paths." + }, + { + "left": "issue:43208", + "right": "issue:43550", + "accept": false, + "reason": "xLSTM training bugs vs Bamba torch.compile/SDPA regression; different models and code paths." + }, + { + "left": "issue:44380", + "right": "issue:45479", + "accept": false, + "reason": "Attention scaling under SDPA/FlashAttention is unrelated to single-label classification zero-loss behavior." + }, + { + "left": "issue:43352", + "right": "issue:43976", + "accept": false, + "reason": "Nemotron Flash Attention support and Python/version compatibility are unrelated issues." + }, + { + "left": "issue:43064", + "right": "issue:44380", + "accept": false, + "reason": "FSDP/PEFT optimizer-state bug is unrelated to GPT2 attention scaling under backend kernels." + }, + { + "left": "issue:44483", + "right": "issue:44514", + "accept": false, + "reason": "Chat-completions request rejection and batched chat template padding crash are different layers of the stack." + }, + { + "left": "issue:43208", + "right": "issue:43526", + "accept": false, + "reason": "xLSTM training bugs vs BeitImageProcessorFast label reduction bug; no shared underlying defect." + }, + { + "left": "issue:43618", + "right": "issue:44380", + "accept": false, + "reason": "CLIP attentions assignment regression is separate from GPT2 attention scaling config being ignored." + }, + { + "left": "issue:36683", + "right": "issue:43582", + "accept": false, + "reason": "Gemma3 vocab_size attribute error and AppleSilicon TypeError in caching_allocator_warmup are unrelated." + }, + { + "left": "issue:43618", + "right": "issue:43761", + "accept": false, + "reason": "Both are CLIP regressions, but one drops attentions and the other drops hidden_states; not the same bug." + }, + { + "left": "issue:44483", + "right": "issue:45290", + "accept": false, + "reason": "Server-side request validation issue vs apply_chat_template crash on tool-call assistant messages." + }, + { + "left": "issue:43012", + "right": "issue:44683", + "accept": false, + "reason": "PyTorch warning on bfloat16 compile and flex_attention failure on torch>=2.9 are distinct compiler/backend problems." + }, + { + "left": "issue:42757", + "right": "issue:43824", + "accept": false, + "reason": "Different missing imports from different modules; same symptom, different symbols and root causes." + }, + { + "left": "issue:44038", + "right": "issue:44230", + "accept": false, + "reason": "Qwen3-VL-Moe loading regression vs fp8 inference support request; not the same change or bug." + }, + { + "left": "issue:44483", + "right": "issue:45381", + "accept": false, + "reason": "Chat/completions request handling and Qwen2.5-VL video vision_position_ids are unrelated." + }, + { + "left": "issue:43299", + "right": "issue:43867", + "accept": false, + "reason": "Qwen3VL-Moe loading regression and state_dict-sorted load error differ in model path and failure mode." + }, + { + "left": "issue:38617", + "right": "issue:43502", + "accept": false, + "reason": "Missing configuration_utils import and local_files_only still making network calls are unrelated." + }, + { + "left": "issue:43756", + "right": "issue:44485", + "accept": false, + "reason": "Smollm3 RoPE-layer mismatch is unrelated to GLM-5 RoPE implementation discussion." + }, + { + "left": "issue:43901", + "right": "issue:44655", + "accept": false, + "reason": "Documentation mismatch for return_all_scores and Pipeline save_pretrained failure are different concerns." + }, + { + "left": "issue:42994", + "right": "issue:44655", + "accept": false, + "reason": "Quantized model save failure is not the same as saving Pipeline objects." + }, + { + "left": "issue:43867", + "right": "issue:45216", + "accept": false, + "reason": "State_dict ordering load error and Qwen3.5 checkpoint corruption after save_pretrained are different bugs." + }, + { + "left": "issue:36683", + "right": "issue:43475", + "accept": false, + "reason": "Gemma3 vocab_size attribute error and SAM3 video missing fpn_position_embeddings are unrelated model config issues." + }, + { + "left": "issue:44220", + "right": "issue:44655", + "accept": false, + "reason": "_torch_extract_fbank_features bug and Pipeline serialization bug are different subsystems." + }, + { + "left": "issue:43066", + "right": "issue:45381", + "accept": false, + "reason": "Tokenizer decoder-type regression and Qwen2.5-VL video position-id bug are unrelated." + }, + { + "left": "issue:43352", + "right": "issue:43937", + "accept": false, + "reason": "FlashAttention support gating for Nemotron and GLM-5 GenerationConfig validation are different issues." + }, + { + "left": "issue:44077", + "right": "issue:44380", + "accept": false, + "reason": "patchtsmixer post_init allowance and GPT2 attention scaling under SDPA/FlashAttention are unrelated." + }, + { + "left": "issue:43064", + "right": "issue:44492", + "accept": false, + "reason": "Distributed optimizer-state bug and a cache-strategy typo are not the same underlying problem." + }, + { + "left": "issue:44263", + "right": "issue:45440", + "accept": false, + "reason": "torch.split return handling in GLM-Moe indexer is unrelated to DeepseekV3MoE implementation divergence." + }, + { + "left": "issue:43976", + "right": "issue:44938", + "accept": false, + "reason": "Both are environment-compatibility reports, but Python 3.9/3.10 and Python 3.14 failures have different causes." + }, + { + "left": "issue:36683", + "right": "issue:44077", + "accept": false, + "reason": "Gemma3 config attribute error is unrelated to patchtsmixer post_init policy." + }, + { + "left": "issue:36683", + "right": "issue:43525", + "accept": false, + "reason": "Gemma3 vocab_size missing and Llama4 pad_token_id missing are different config attribute regressions." + }, + { + "left": "issue:36010", + "right": "issue:42757", + "accept": false, + "reason": "Different import failures from different modules and packages; not the same missing symbol." + }, + { + "left": "issue:44188", + "right": "issue:44380", + "accept": false, + "reason": "Torch.compile branching divergence in attention kernels is not the same as GPT2 attention scaling config being ignored." + }, + { + "left": "issue:43618", + "right": "issue:45216", + "accept": false, + "reason": "CLIP attentions regression and Qwen3.5 checkpoint save corruption are unrelated." + }, + { + "left": "issue:43352", + "right": "issue:45593", + "accept": false, + "reason": "FlashAttention support for Nemotron and D-FINE auxiliary-loss behavior are different model bugs." + }, + { + "left": "issue:43992", + "right": "issue:45310", + "accept": false, + "reason": "Missing embed_tokens.weight in UMT5Encoder and Qwen3.5 Moe from_pretrained error are different load paths 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": 13, + "review_unit_index": 40, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:28282", + "issue:33453", + "issue:36296", + "issue:36683", + "issue:41669", + "issue:42548", + "issue:42890", + "issue:43010", + "issue:43064", + "issue:43065", + "issue:43232", + "issue:43352", + "issue:43502", + "issue:43575", + "issue:43618", + "issue:43723", + "issue:43867", + "issue:43976", + "issue:43992", + "issue:44062", + "issue:44075", + "issue:44077", + "issue:44112", + "issue:44222", + "issue:44242", + "issue:44246", + "issue:44351", + "issue:44355", + "issue:44360", + "issue:44387", + "issue:44484", + "issue:44556", + "issue:44749", + "issue:44857", + "issue:44938", + "issue:44964", + "issue:45071", + "issue:45125", + "issue:45278", + "issue:45341", + "issue:45357", + "issue:45375", + "issue:45381", + "issue:45440", + "issue:45478", + "issue:45561", + "issue:45588", + "issue:45600" + ], + "soft_pairs": [ + "issue:36683|issue:44242", + "issue:43010|issue:45278", + "issue:36683|issue:43352", + "issue:41669|issue:44246", + "issue:43867|issue:45375", + "issue:36683|issue:43064", + "issue:43992|issue:45478", + "issue:42890|issue:45561", + "issue:44938|issue:45071", + "issue:43867|issue:45357", + "issue:43010|issue:43976", + "issue:44938|issue:45381", + "issue:44360|issue:45440", + "issue:43618|issue:44077", + "issue:36296|issue:43575", + "issue:43867|issue:44484", + "issue:43618|issue:45588", + "issue:43618|issue:44857", + "issue:28282|issue:33453", + "issue:42890|issue:44964", + "issue:42890|issue:45600", + "issue:42890|issue:44112", + "issue:43010|issue:44075", + "issue:43502|issue:44062", + "issue:43723|issue:44556", + "issue:44222|issue:45125", + "issue:42890|issue:45341", + "issue:44387|issue:44556", + "issue:43010|issue:43065", + "issue:43010|issue:43232", + "issue:42548|issue:44351", + "issue:45125|issue:45478", + "issue:44355|issue:44749", + "issue:45125|issue:45375" + ], + "prepared_review_unit_hash": "680a4d8cfe5f60e8baff9bffaea25f07d7d0f8dc3eb86667a4eb383ca4f1736f", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37356, + "estimated_input_tokens": 9339, + "estimated_eval_tokens": 18934 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous and centered on unrelated loading/import/model-specific regressions. Only the import-time slowdown pair looks like a true duplicate; the rest are better treated as distinct bugs.", + "confidence": 0.71, + "canonical_issue_reason": "issue:45278 is the best umbrella issue: it captures the broad post-upgrade import-breakage theme that most closely represents this mixed cluster, whereas the other items are narrower model- or test-specific regressions.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 is the strongest global representative because it is broad, user-facing, and aligned with the dominant import/load regression theme across the set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36683", + "right": "issue:44242", + "accept": false, + "reason": "Gemma3Config vocab_size missing is a model-config bug; load-balancing loss omission is a different MoE training issue." + }, + { + "left": "issue:43010", + "right": "issue:45278", + "accept": false, + "reason": "Cache update/no_grad behavior is unrelated to broad import errors after upgrading." + }, + { + "left": "issue:36683", + "right": "issue:43352", + "accept": false, + "reason": "Different subsystems and symptoms: Gemma3Config attribute error vs Flash Attention 2.0 support gate." + }, + { + "left": "issue:41669", + "right": "issue:44246", + "accept": true, + "reason": "Both describe the same import-time slowdown regression; 41669 identifies the `import *` cause and 44246 reports the user-visible import slowness." + }, + { + "left": "issue:43867", + "right": "issue:45375", + "accept": false, + "reason": "State-dict loading failure and a missing config field are different bugs in different code paths." + }, + { + "left": "issue:36683", + "right": "issue:43064", + "accept": false, + "reason": "Gemma3Config missing vocab_size is unrelated to the FSDP2/PEFT optimizer-state issue." + }, + { + "left": "issue:43992", + "right": "issue:45478", + "accept": false, + "reason": "Both concern model loading, but they affect different models and likely different root causes." + }, + { + "left": "issue:42890", + "right": "issue:45561", + "accept": false, + "reason": "One is flaky test seeding, the other is an xdist file-race in testing utils; not the same bug." + }, + { + "left": "issue:44938", + "right": "issue:45071", + "accept": false, + "reason": "Python 3.14 import failure and PretrainedConfig type-checking regression are distinct compatibility issues." + }, + { + "left": "issue:43867", + "right": "issue:45357", + "accept": false, + "reason": "A state_dict sort/load issue is not the same as incorrect visual encoder keys in save_pretrained." + }, + { + "left": "issue:43010", + "right": "issue:43976", + "accept": false, + "reason": "Cache update decoration behavior is unrelated to Python-version support in Transformers 5.1.0." + }, + { + "left": "issue:44938", + "right": "issue:45381", + "accept": false, + "reason": "Python 3.14 load failure and Qwen2.5-VL vision_position_ids are unrelated." + }, + { + "left": "issue:44360", + "right": "issue:45440", + "accept": false, + "reason": "DSA indexer activation bug and DeepSeekV3MoE divergence are different implementation issues." + }, + { + "left": "issue:43618", + "right": "issue:44077", + "accept": false, + "reason": "CLIPOutput attentions regression and patchtsmixer post_init policy are unrelated." + }, + { + "left": "issue:36296", + "right": "issue:43575", + "accept": false, + "reason": "Tensor-parallel training bug and TP OOM on a specific large model are not the same concrete failure." + }, + { + "left": "issue:43867", + "right": "issue:44484", + "accept": false, + "reason": "State_dict loading error is not the same as the default max_shard_size question." + }, + { + "left": "issue:43618", + "right": "issue:45588", + "accept": false, + "reason": "Different failures: CLIPOutput attentions missing vs flash_attention.py crashing on s_aux=None." + }, + { + "left": "issue:43618", + "right": "issue:44857", + "accept": false, + "reason": "CLIPOutput attentions regression and AMP/CUDA crash in LwDetrImageLoss are unrelated." + }, + { + "left": "issue:28282", + "right": "issue:33453", + "accept": false, + "reason": "AutoModel missing PyTorch dependency and tokenizer loading regression are different import/load problems." + }, + { + "left": "issue:42890", + "right": "issue:44964", + "accept": false, + "reason": "Seed-related test flakiness is unrelated to Phi-4 multimodal model loading." + }, + { + "left": "issue:42890", + "right": "issue:45600", + "accept": false, + "reason": "Missing set_seed in a test is unrelated to removed Sam3LiteText config references in auto_mappings." + }, + { + "left": "issue:42890", + "right": "issue:44112", + "accept": false, + "reason": "Flaky integration test seeding and GraniteSpeech device-override CI failure are different test bugs." + }, + { + "left": "issue:43010", + "right": "issue:44075", + "accept": false, + "reason": "Cache update decoration is unrelated to SGD args not being applied." + }, + { + "left": "issue:43502", + "right": "issue:44062", + "accept": false, + "reason": "Ignoring local_files_only and AddedToken special-arg duplication are different tokenizer/file-fetch bugs." + }, + { + "left": "issue:43723", + "right": "issue:44556", + "accept": false, + "reason": "Tokenizer loading regression and checkpoint reload incompatibility are related only loosely; the concrete failures differ." + }, + { + "left": "issue:44222", + "right": "issue:45125", + "accept": false, + "reason": "FP8 save_pretrained behavior and missing `_tp_plan` for tensor parallelism are different MoE issues." + }, + { + "left": "issue:42890", + "right": "issue:45341", + "accept": false, + "reason": "Missing test seeding is not the same as the generic testing_utils bug referenced here." + }, + { + "left": "issue:44387", + "right": "issue:44556", + "accept": false, + "reason": "Int4 quantization OOM and checkpoint reload failure are distinct regressions." + }, + { + "left": "issue:43010", + "right": "issue:43065", + "accept": false, + "reason": "Cache update no_grad behavior and the Sam3PixelDecoder dummy Conv2d issue are unrelated." + }, + { + "left": "issue:43010", + "right": "issue:43232", + "accept": false, + "reason": "Cache update no_grad behavior and generation kwarg synchronization are different code paths." + }, + { + "left": "issue:42548", + "right": "issue:44351", + "accept": false, + "reason": "Missing PreTrainedModel export and missing HybridCache export are both import errors, but not the same missing symbol or root cause." + }, + { + "left": "issue:44355", + "right": "issue:44749", + "accept": false, + "reason": "Compiled Python-file errors and slow filtering after upgrade are unrelated." + }, + { + "left": "issue:45125", + "right": "issue:45478", + "accept": false, + "reason": "Missing `_tp_plan` and Qwen3.5 Moe from_pretrained error are related only at a high level; not the same concrete bug." + }, + { + "left": "issue:44222", + "right": "issue:45375", + "accept": false, + "reason": "FP8 MoE save_pretrained and missing deepstack_visual_indexes are different Qwen3.5 vision/config issues." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: issue:45278 is a reasonable umbrella for the import/load-related regressions, and the rest of the cluster is mostly heterogeneous. The only accepted soft edge, issue:41669 \u2194 issue:44246, looks like a plausible duplicate import-slowdown pair. The other pairwise rejections are conservative and consistent with the issue titles." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 41, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:30333", + "issue:30990", + "issue:31795", + "issue:33453", + "issue:34689", + "issue:36296", + "issue:38617", + "issue:41628", + "issue:41669", + "issue:42548", + "issue:42757", + "issue:42831", + "issue:43010", + "issue:43299", + "issue:43502", + "issue:43526", + "issue:43576", + "issue:43673", + "issue:43824", + "issue:43827", + "issue:43874", + "issue:43901", + "issue:44038", + "issue:44075", + "issue:44222", + "issue:44230", + "issue:44261", + "issue:44273", + "issue:44315", + "issue:44351", + "issue:44355", + "issue:44360", + "issue:44368", + "issue:44393", + "issue:44485", + "issue:44492", + "issue:44556", + "issue:44623", + "issue:44655", + "issue:44871", + "issue:44908", + "issue:44961", + "issue:45003", + "issue:45071", + "issue:45125", + "issue:45230", + "issue:45468", + "issue:45542" + ], + "soft_pairs": [ + "issue:43502|issue:44485", + "issue:30990|issue:31795", + "issue:44075|issue:44871", + "issue:44315|issue:44360", + "issue:30333|issue:33453", + "issue:43010|issue:44368", + "issue:43502|issue:44222", + "issue:43299|issue:44556", + "issue:43502|issue:44961", + "issue:44222|issue:44230", + "issue:44355|issue:44655", + "issue:30333|issue:36296", + "issue:44355|issue:45071", + "issue:42831|issue:43874", + "issue:33453|issue:34689", + "issue:42757|issue:43576", + "issue:33453|issue:36296", + "issue:43827|issue:44961", + "issue:43576|issue:45071", + "issue:43576|issue:43824", + "issue:38617|issue:43576", + "issue:43673|issue:44492", + "issue:44961|issue:45468", + "issue:31795|issue:38617", + "issue:43576|issue:44351", + "issue:41628|issue:42548", + "issue:44908|issue:44961", + "issue:44485|issue:44961", + "issue:36296|issue:43901", + "issue:36296|issue:43526", + "issue:41628|issue:44355", + "issue:44273|issue:45003", + "issue:44261|issue:44360", + "issue:44623|issue:45230", + "issue:44273|issue:45542", + "issue:43576|issue:44556", + "issue:44038|issue:45125", + "issue:44393|issue:45071", + "issue:41669|issue:43673" + ], + "prepared_review_unit_hash": "99194bfbff11ba0d515efd1661cab690c54cadf7855fb5ed9774db2c75d3382a", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 39, + "serialized_chars": 38649, + "estimated_input_tokens": 9663, + "estimated_eval_tokens": 19582 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a loose set of Transformers v5 regression reports, especially import/loading failures, but the soft pairs do not look like true duplicates of the same concrete bug. I would not merge any of the proposed soft edges.", + "confidence": 0.64, + "canonical_issue_reason": "issue:43576 is the most central and broadly connected report in the set, with multiple soft similarities to other v5 breakage/import-loading issues. It best represents the cluster theme, even though it is not a duplicate of the others.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43576 is the best single issue to anchor this cluster because it is the most connected to the other v5-related reports and captures the shared breakage theme more broadly than the narrower model-specific issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43502", + "right": "issue:44485", + "accept": false, + "reason": "Different problems: offline/local-files-only networking behavior vs GLM-5 RoPE implementation." + }, + { + "left": "issue:30990", + "right": "issue:31795", + "accept": false, + "reason": "Unrelated topics: Sentence Transformers load hang vs documentation confusion about forward arguments." + }, + { + "left": "issue:44075", + "right": "issue:44871", + "accept": false, + "reason": "Different bugs: SGD optimizer args ignored vs Gemma-3 EOS token config mismatch." + }, + { + "left": "issue:44315", + "right": "issue:44360", + "accept": false, + "reason": "No shared code-path or symptom; model_init/Liger Kernel issue is unrelated to the DSA indexer ReLU discussion." + }, + { + "left": "issue:30333", + "right": "issue:33453", + "accept": false, + "reason": "Different failures: MLFlow reporting on training failure vs tokenizer loading regression." + }, + { + "left": "issue:43010", + "right": "issue:44368", + "accept": false, + "reason": "Different areas: cache/layer no_grad decoration vs tie_word_embeddings warning during Qwen fine-tuning." + }, + { + "left": "issue:43502", + "right": "issue:44222", + "accept": false, + "reason": "Different issues: local_files_only network requests vs FP8 moe save_pretrained behavior." + }, + { + "left": "issue:43299", + "right": "issue:44556", + "accept": false, + "reason": "Both are v5-related loading regressions, but they affect different models and scenarios; not the same concrete bug." + }, + { + "left": "issue:43502", + "right": "issue:44961", + "accept": false, + "reason": "No meaningful overlap; local_files_only networking bug vs unrelated 'racoon' issue." + }, + { + "left": "issue:44222", + "right": "issue:44230", + "accept": false, + "reason": "Related FP8/moe area, but one is save_pretrained and the other is inference support; not the same bug." + }, + { + "left": "issue:44355", + "right": "issue:44655", + "accept": false, + "reason": "Different functionality: compiled Python file errors vs saving Pipeline objects." + }, + { + "left": "issue:30333", + "right": "issue:36296", + "accept": false, + "reason": "Different subsystems and symptoms: MLFlow reporting vs tensor parallel training." + }, + { + "left": "issue:44355", + "right": "issue:45071", + "accept": false, + "reason": "Different issues: compiled file execution error vs PretrainedConfig type-checking regression." + }, + { + "left": "issue:42831", + "right": "issue:43874", + "accept": false, + "reason": "Different model-specific bugs: FineGrainedFP8 accuracy vs GLM46V image processor token counting." + }, + { + "left": "issue:33453", + "right": "issue:34689", + "accept": false, + "reason": "Both are loading regressions, but for different models and likely different root causes." + }, + { + "left": "issue:42757", + "right": "issue:43576", + "accept": false, + "reason": "Both are v5 compatibility complaints, but one is a missing import and the other is the env command being broken; not the same bug." + }, + { + "left": "issue:33453", + "right": "issue:36296", + "accept": false, + "reason": "Unrelated: tokenizer loading regression vs tensor parallel training bug." + }, + { + "left": "issue:43827", + "right": "issue:44961", + "accept": false, + "reason": "Docs reference to pipeline() is unrelated to the 'racoon' issue." + }, + { + "left": "issue:43576", + "right": "issue:45071", + "accept": false, + "reason": "Both involve v5 breakage, but one is the env command and the other is PretrainedConfig typing; different concrete failures." + }, + { + "left": "issue:43576", + "right": "issue:43824", + "accept": false, + "reason": "Both are import/loading regressions, but the missing symbol and the env command bug are not the same change or code-path." + }, + { + "left": "issue:38617", + "right": "issue:43576", + "accept": false, + "reason": "Different specific failures: missing layer_type_validation import vs broken env command." + }, + { + "left": "issue:43673", + "right": "issue:44492", + "accept": false, + "reason": "Cache-missing during chunked prefill is not the same as a typo in cache strategies." + }, + { + "left": "issue:44961", + "right": "issue:45468", + "accept": false, + "reason": "No overlap: 'racoon' vs Gemma-4 audio positional encoding." + }, + { + "left": "issue:31795", + "right": "issue:38617", + "accept": false, + "reason": "Documentation confusion vs import error; not duplicate bugs." + }, + { + "left": "issue:43576", + "right": "issue:44351", + "accept": false, + "reason": "Both are v5 import-related, but they concern different missing symbols and likely different fixes." + }, + { + "left": "issue:41628", + "right": "issue:42548", + "accept": false, + "reason": "Both are import errors, but for different classes and missing symbols; not enough evidence of the same underlying bug." + }, + { + "left": "issue:44908", + "right": "issue:44961", + "accept": false, + "reason": "Scheduler kwargs bug is unrelated to the 'racoon' issue." + }, + { + "left": "issue:44485", + "right": "issue:44961", + "accept": false, + "reason": "RoPE implementation discussion is unrelated to 'racoon'." + }, + { + "left": "issue:36296", + "right": "issue:43901", + "accept": false, + "reason": "Tensor parallel training bug is unrelated to a docs mismatch in TextClassificationPipeline." + }, + { + "left": "issue:36296", + "right": "issue:43526", + "accept": false, + "reason": "Different model/training bug vs BeitImageProcessorFast reduce_labels issue." + }, + { + "left": "issue:41628", + "right": "issue:44355", + "accept": false, + "reason": "Import error for AutoImageProcessor is unrelated to compiled Python file execution errors." + }, + { + "left": "issue:44273", + "right": "issue:45003", + "accept": false, + "reason": "Both mention loading behavior, but lazy loading problems and unsafe sys.modules access are not clearly the same bug." + }, + { + "left": "issue:44261", + "right": "issue:44360", + "accept": false, + "reason": "MLA q_a_layernorm precision/config issue is unrelated to the DSA indexer ReLU topic." + }, + { + "left": "issue:44623", + "right": "issue:45230", + "accept": false, + "reason": "Generic 'Bug report' is too vague to treat as the same issue as missing processor.save_pretrained files." + }, + { + "left": "issue:44273", + "right": "issue:45542", + "accept": false, + "reason": "Lazy loading behavior is unrelated to the TensorFlow backend error when only tensorboard is installed." + }, + { + "left": "issue:43576", + "right": "issue:44556", + "accept": false, + "reason": "Both are v5 breakages, but one is the env command and the other is checkpoint reload compatibility; not the same concrete bug." + }, + { + "left": "issue:44038", + "right": "issue:45125", + "accept": false, + "reason": "Different Qwen/moe and tensor-parallel problems; not the same underlying change." + }, + { + "left": "issue:44393", + "right": "issue:45071", + "accept": false, + "reason": "Qwen3-VL bounding-box output issue is unrelated to PretrainedConfig type checking." + }, + { + "left": "issue:41669", + "right": "issue:43673", + "accept": false, + "reason": "Import-star cleanup/perf issue is unrelated to a missing cache during chunked prefill." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 42, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:31515", + "issue:31795", + "issue:33453", + "issue:34689", + "issue:41669", + "issue:42371", + "issue:42890", + "issue:42913", + "issue:43066", + "issue:43352", + "issue:43408", + "issue:43502", + "issue:43519", + "issue:43576", + "issue:43606", + "issue:43704", + "issue:43723", + "issue:43827", + "issue:43976", + "issue:44077", + "issue:44246", + "issue:44261", + "issue:44263", + "issue:44273", + "issue:44295", + "issue:44297", + "issue:44315", + "issue:44351", + "issue:44360", + "issue:44393", + "issue:44485", + "issue:44492", + "issue:44556", + "issue:44623", + "issue:44655", + "issue:44704", + "issue:44829", + "issue:44861", + "issue:44908", + "issue:44945", + "issue:44998", + "issue:45071", + "issue:45092", + "issue:45230", + "issue:45310", + "issue:45341", + "issue:45468", + "issue:45478" + ], + "soft_pairs": [ + "issue:44556|issue:45092", + "issue:42371|issue:44492", + "issue:42890|issue:44908", + "issue:44704|issue:44908", + "issue:44246|issue:44273", + "issue:43408|issue:44315", + "issue:44297|issue:45230", + "issue:43352|issue:44393", + "issue:43576|issue:44908", + "issue:43976|issue:44393", + "issue:44908|issue:45071", + "issue:44295|issue:45230", + "issue:41669|issue:44351", + "issue:41669|issue:45341", + "issue:44492|issue:44945", + "issue:41669|issue:44492", + "issue:44315|issue:44829", + "issue:44393|issue:45478", + "issue:31795|issue:33453", + "issue:43827|issue:44908", + "issue:44077|issue:44861", + "issue:41669|issue:44908", + "issue:43723|issue:44393", + "issue:44655|issue:44861", + "issue:43576|issue:44273", + "issue:44360|issue:45468", + "issue:45230|issue:45310", + "issue:44273|issue:44908", + "issue:41669|issue:43704", + "issue:43519|issue:44485", + "issue:43519|issue:44908", + "issue:31795|issue:34689", + "issue:31515|issue:31795", + "issue:44315|issue:45092", + "issue:43704|issue:44485", + "issue:44485|issue:45468", + "issue:42371|issue:43704", + "issue:43502|issue:43519", + "issue:43519|issue:44623", + "issue:43519|issue:43704", + "issue:44263|issue:45468", + "issue:44261|issue:45468", + "issue:43606|issue:45468", + "issue:42890|issue:43704", + "issue:42913|issue:43066", + "issue:44908|issue:44998" + ], + "prepared_review_unit_hash": "49fde7466f69de5fd53f79114bc3d20afb46bc84605228a528fecfbe42ffe323", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 46, + "serialized_chars": 40783, + "estimated_input_tokens": 10196, + "estimated_eval_tokens": 20648 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a mixed set of mostly unrelated Transformers issues, with several loose clusters around tokenizer/model loading in v5, but no soft pair is strong enough to treat as the same underlying bug under strict duplicate triage.", + "confidence": 0.76, + "canonical_issue_reason": "issue:43723 is the clearest broad anchor in the set because it directly describes an AutoTokenizer.from_pretrained loading regression in v5, which matches the repeated tokenizer-loading theme better than the narrower or more model-specific reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43723 is also the best overall issue to keep as the representative reference for this cluster because it is the most central and reusable description of the shared v5 tokenizer-loading problem area.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44556", + "right": "issue:45092", + "accept": false, + "reason": "Both are v5 checkpoint-loading complaints, but one is a generic reload failure and the other is a specific InternVL2 remote-code/meta-init incompatibility; not the same concrete bug." + }, + { + "left": "issue:42371", + "right": "issue:44492", + "accept": false, + "reason": "TF32 behavior settings and a cache-strategy typo are unrelated problems." + }, + { + "left": "issue:42890", + "right": "issue:44908", + "accept": false, + "reason": "A flaky integration test missing set_seed is unrelated to the inverse_sqrt scheduler ignoring kwargs." + }, + { + "left": "issue:44704", + "right": "issue:44908", + "accept": false, + "reason": "AutoProcessor kwargs forwarding and scheduler kwargs handling are different code paths and different bugs." + }, + { + "left": "issue:44246", + "right": "issue:44273", + "accept": false, + "reason": "Import-time slowness and lazy-loading correctness are related only at a high level, not the same bug." + }, + { + "left": "issue:43408", + "right": "issue:44315", + "accept": false, + "reason": "A model-type warning for sam3_video/sam3_tracker is unrelated to Liger Kernel application with model_init." + }, + { + "left": "issue:44297", + "right": "issue:45230", + "accept": false, + "reason": "A tokenizer_class mismatch in save_pretrained is not the same as an unspecified generic bug report." + }, + { + "left": "issue:43352", + "right": "issue:44393", + "accept": false, + "reason": "Flash Attention 2 support for Nemotron and Qwen3-VL bounding-box output issues are unrelated." + }, + { + "left": "issue:43576", + "right": "issue:44908", + "accept": false, + "reason": "A broken transformers env command is unrelated to scheduler kwargs handling." + }, + { + "left": "issue:43976", + "right": "issue:44393", + "accept": false, + "reason": "Python version compatibility and Qwen3-VL output errors are unrelated." + }, + { + "left": "issue:44908", + "right": "issue:45071", + "accept": false, + "reason": "Scheduler kwargs handling and PretrainedConfig type checking are different regressions." + }, + { + "left": "issue:44295", + "right": "issue:45230", + "accept": false, + "reason": "A position_ids buffer access error is not the same as a generic bug report." + }, + { + "left": "issue:41669", + "right": "issue:44351", + "accept": false, + "reason": "Removing import * for performance is not the same as failing to import HybridCache." + }, + { + "left": "issue:41669", + "right": "issue:45341", + "accept": false, + "reason": "Import cleanup/performance and a testing_utils bug are unrelated." + }, + { + "left": "issue:44492", + "right": "issue:44945", + "accept": false, + "reason": "A cache-strategy typo does not match an incorrect pipeline-parallelism output bug." + }, + { + "left": "issue:41669", + "right": "issue:44492", + "accept": false, + "reason": "Import cleanup and cache-strategy wording are unrelated." + }, + { + "left": "issue:44315", + "right": "issue:44829", + "accept": false, + "reason": "Liger Kernel not being applied with model_init and flash_attention_3 causing degenerate training are different issues." + }, + { + "left": "issue:44393", + "right": "issue:45478", + "accept": false, + "reason": "Qwen3-VL 2D bounding-box hallucinations and Qwen3.5 MoE from_pretrained loading failures are unrelated." + }, + { + "left": "issue:31795", + "right": "issue:33453", + "accept": false, + "reason": "Documentation confusion about forward arguments is not the same as a tokenizer-loading regression." + }, + { + "left": "issue:43827", + "right": "issue:44908", + "accept": false, + "reason": "Docs still referencing pipeline() after removals is unrelated to inverse_sqrt scheduler kwargs." + }, + { + "left": "issue:44077", + "right": "issue:44861", + "accept": false, + "reason": "An optional post_init restriction and a tied-weights AttributeError are different bugs." + }, + { + "left": "issue:41669", + "right": "issue:44908", + "accept": false, + "reason": "Import-time cleanup and scheduler behavior are unrelated." + }, + { + "left": "issue:43723", + "right": "issue:44393", + "accept": false, + "reason": "Tokenizer loading in v5 and Qwen3-VL bounding-box output errors are unrelated." + }, + { + "left": "issue:44655", + "right": "issue:44861", + "accept": false, + "reason": "Saving pipeline objects and _get_tied_weight_keys crashing on a list are different failure modes." + }, + { + "left": "issue:43576", + "right": "issue:44273", + "accept": false, + "reason": "The env command issue is unrelated to lazy loading not working properly." + }, + { + "left": "issue:44360", + "right": "issue:45468", + "accept": false, + "reason": "A missing ReLU in the DSA indexer and Gemma-4 audio positional encoding are unrelated." + }, + { + "left": "issue:45230", + "right": "issue:45310", + "accept": false, + "reason": "A generic bug report is not the same as a specific Qwen3.5 MoE from_pretrained error." + }, + { + "left": "issue:44273", + "right": "issue:44908", + "accept": false, + "reason": "Lazy loading failures and scheduler kwargs handling are unrelated." + }, + { + "left": "issue:41669", + "right": "issue:43704", + "accept": false, + "reason": "Import * cleanup/performance and VRAM leakage in dataloader threads are different issues." + }, + { + "left": "issue:43519", + "right": "issue:44485", + "accept": false, + "reason": "Qwen3VL timestamp calculation and GLM-5 RoPE implementation are separate model bugs." + }, + { + "left": "issue:43519", + "right": "issue:44908", + "accept": false, + "reason": "Processor timestamp calculation and scheduler kwargs handling are unrelated." + }, + { + "left": "issue:31795", + "right": "issue:34689", + "accept": false, + "reason": "Documentation confusion and a Llama 3.2 model-loading regression are not the same bug." + }, + { + "left": "issue:31515", + "right": "issue:31795", + "accept": false, + "reason": "Slow checkpoint loading and forward-argument documentation confusion are unrelated." + }, + { + "left": "issue:44315", + "right": "issue:45092", + "accept": false, + "reason": "Liger Kernel application with model_init and old checkpoint/meta-init incompatibility are different problems." + }, + { + "left": "issue:43704", + "right": "issue:44485", + "accept": false, + "reason": "VRAM leakage across dataloader threads is unrelated to GLM-5 RoPE implementation." + }, + { + "left": "issue:44485", + "right": "issue:45468", + "accept": false, + "reason": "GLM-5 RoPE and Gemma-4 audio positional encoding are unrelated model-specific bugs." + }, + { + "left": "issue:42371", + "right": "issue:43704", + "accept": false, + "reason": "TF32 settings and VRAM leakage in dataloader threads are unrelated." + }, + { + "left": "issue:43502", + "right": "issue:43519", + "accept": false, + "reason": "Unexpected API requests with local_files_only and Qwen3VL timestamp calculation are unrelated." + }, + { + "left": "issue:43519", + "right": "issue:44623", + "accept": false, + "reason": "Processor timestamp math and missing files in processor.save_pretrained are different bugs." + }, + { + "left": "issue:43519", + "right": "issue:43704", + "accept": false, + "reason": "Timestamp calculation and VRAM leakage across threads are unrelated." + }, + { + "left": "issue:44263", + "right": "issue:45468", + "accept": false, + "reason": "A torch.split return-value issue and Gemma-4 audio positional encoding are different bugs." + }, + { + "left": "issue:44261", + "right": "issue:45468", + "accept": false, + "reason": "An MLA q_a_layernorm precision issue is unrelated to Gemma-4 audio positional encoding." + }, + { + "left": "issue:43606", + "right": "issue:45468", + "accept": false, + "reason": "CPU offload device mismatch and Gemma-4 audio positional encoding are unrelated." + }, + { + "left": "issue:42890", + "right": "issue:43704", + "accept": false, + "reason": "A flaky integration test and VRAM leakage in dataloader threads are unrelated." + }, + { + "left": "issue:42913", + "right": "issue:43066", + "accept": false, + "reason": "Both mention tokenizer changes in v5, but one is a broad behavior-difference report and the other is a specific decoder-type mismatch; not enough to call them the same bug." + }, + { + "left": "issue:44908", + "right": "issue:44998", + "accept": false, + "reason": "A scheduler kwargs bug and an unrelated placeholder issue are not the same thing." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 43, + "review_unit_count": 43, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-374", + "nodes": [ + "issue:30333", + "issue:31515", + "issue:33453", + "issue:36296", + "issue:38617", + "issue:43298", + "issue:43525", + "issue:43572", + "issue:43618", + "issue:44230", + "issue:44492", + "issue:44556", + "issue:44623", + "issue:44861", + "issue:44998", + "issue:45290", + "issue:45310", + "issue:45341" + ], + "soft_pairs": [ + "issue:44861|issue:45290", + "issue:44998|issue:45341", + "issue:44492|issue:44998", + "issue:43618|issue:44998", + "issue:44556|issue:44998", + "issue:43525|issue:44861", + "issue:43572|issue:44861", + "issue:31515|issue:38617", + "issue:44230|issue:44623", + "issue:31515|issue:36296", + "issue:30333|issue:31515", + "issue:44230|issue:45310", + "issue:31515|issue:33453", + "issue:43298|issue:44230" + ], + "prepared_review_unit_hash": "04281dd5a6716af263c30c1d3922adecaa5126f2841630e3c8d4107beac9bffc", + "component_budget": { + "node_count": 374, + "item_count": 374, + "soft_pair_count": 1395, + "serialized_chars": 720157, + "estimated_input_tokens": 180040, + "estimated_eval_tokens": 360336 + }, + "budget": { + "node_count": 18, + "item_count": 18, + "soft_pair_count": 14, + "serialized_chars": 14255, + "estimated_input_tokens": 3564, + "estimated_eval_tokens": 7384 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly mixed: the issues share superficial wording around crashes, loading, or config names, but they do not describe the same underlying bug. None of the soft pairs look safe to merge as duplicates.", + "confidence": 0.95, + "canonical_issue_reason": "No single canonical issue stands out as a true duplicate hub; the items span unrelated failures across loading, configs, templates, training, and misc. bugs.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:31515 is the closest thing to a representative because it is a concrete from_pretrained/loading report and has the most superficial overlap with other load-related items, but it still is not a true duplicate center.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44861", + "right": "issue:45290", + "accept": false, + "reason": "Both are crashes, but one is about tied weights internals and the other about chat template tool-call handling; different code paths and failure modes." + }, + { + "left": "issue:44998", + "right": "issue:45341", + "accept": false, + "reason": "'Unemployment' is unrelated to a small testing_utils bug; no shared underlying behavior." + }, + { + "left": "issue:44492", + "right": "issue:44998", + "accept": false, + "reason": "A cache-strategy typo and an unrelated issue title do not describe the same bug or change." + }, + { + "left": "issue:43618", + "right": "issue:44998", + "accept": false, + "reason": "CLIPOutput attentions assignment is a specific model-output regression, unrelated to the other item." + }, + { + "left": "issue:44556", + "right": "issue:44998", + "accept": false, + "reason": "Checkpoint reload breakage after version upgrade is unrelated to the other issue; no concrete overlap." + }, + { + "left": "issue:43525", + "right": "issue:44861", + "accept": false, + "reason": "Missing pad_token_id on Llama4Config and tied-weight-key AttributeError are different config/model bugs." + }, + { + "left": "issue:43572", + "right": "issue:44861", + "accept": false, + "reason": "StableLmConfig pad_token_idx regression is unrelated to the tied-weight-key list crash." + }, + { + "left": "issue:31515", + "right": "issue:38617", + "accept": false, + "reason": "Slow checkpoint loading and an ImportError for layer_type_validation are distinct problems in different code paths." + }, + { + "left": "issue:44230", + "right": "issue:44623", + "accept": false, + "reason": "FP8 support for a model family and missing processor.save_pretrained files are not the same bug." + }, + { + "left": "issue:31515", + "right": "issue:36296", + "accept": false, + "reason": "from_pretrained loading slowness is unrelated to a tensor-parallel training bug." + }, + { + "left": "issue:30333", + "right": "issue:31515", + "accept": false, + "reason": "MLflow reporting on failed training jobs and checkpoint loading speed are unrelated concerns." + }, + { + "left": "issue:44230", + "right": "issue:45310", + "accept": false, + "reason": "FP8 inference support for Qwen3-VL/MoE and a Qwen3.5 MoE from_pretrained error are different concrete issues." + }, + { + "left": "issue:31515", + "right": "issue:33453", + "accept": false, + "reason": "Tokenizer loading regression and checkpoint loading slowness are different failure modes, not the same duplicate." + }, + { + "left": "issue:43298", + "right": "issue:44230", + "accept": false, + "reason": "Continuous batching support for audio models and FP8 support for Qwen models are separate feature/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": 3, + "component_count": 13, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-26350-2", + "nodes": [ + "issue:26350", + "issue:33260" + ], + "soft_pairs": [ + "issue:26350|issue:33260" + ], + "prepared_review_unit_hash": "e0ac32febd7dbe0a2803a61f6aec3c590807364ff4bccf125285591426951abd", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1590, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1590, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two distinct community-contribution enhancement issues: one for Flash Attention 2 support on more architectures, and one for GGUF support on more architectures. They share only a broad theme and are not duplicates.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:26350", + "right": "issue:33260", + "accept": false, + "reason": "Different feature requests for different backends/code paths (Flash Attention 2 vs GGUF); same broad \u2018more architectures\u2019 framing is too generic to treat as the same issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 4, + "component_count": 13, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-27088-4", + "nodes": [ + "issue:27088", + "issue:43834", + "issue:44365", + "issue:44370" + ], + "soft_pairs": [ + "issue:27088|issue:44370", + "issue:27088|issue:44365", + "issue:27088|issue:43834", + "issue:43834|issue:44370", + "issue:43834|issue:44365", + "issue:44365|issue:44370" + ], + "prepared_review_unit_hash": "ebff596b7c9a49a465b1291325e51513294e09c9226da383fa7f0bc2a055c984", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4455, + "estimated_input_tokens": 1114, + "estimated_eval_tokens": 2484 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4455, + "estimated_input_tokens": 1114, + "estimated_eval_tokens": 2484 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All items are about translating docs, but the placeholder-titled issues likely refer to different languages rather than the same request. The tightest near-duplicate group is the generic template issues (43834/44365/44370), while 27088 is a concrete Turkish-specific outlier.", + "confidence": 0.78, + "canonical_issue_reason": "43834 is the earliest of the near-identical template-style issues and the best representative of the generic i18n translation-request pattern.", + "canonical_pr_reason": null, + "best_issue_reason": "43834 is the strongest cluster representative because it sits in the tightest similarity group and is the oldest item among the template-matching issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:27088", + "right": "issue:44370", + "accept": false, + "reason": "Turkish-specific request vs. a generic placeholder language request; same topic area, but not enough evidence they are the same underlying issue." + }, + { + "left": "issue:27088", + "right": "issue:44365", + "accept": false, + "reason": "Different target language context; these look like separate translation requests, not duplicates." + }, + { + "left": "issue:27088", + "right": "issue:43834", + "accept": false, + "reason": "The Turkish issue is concrete, while 43834 is a generic template-style request; they do not appear to be the same issue." + }, + { + "left": "issue:43834", + "right": "issue:44370", + "accept": false, + "reason": "Very similar template text, but likely different language targets, so they are not clearly the same underlying request." + }, + { + "left": "issue:43834", + "right": "issue:44365", + "accept": false, + "reason": "High textual similarity alone is not enough here; the placeholder language fields suggest separate translation requests." + }, + { + "left": "issue:44365", + "right": "issue:44370", + "accept": false, + "reason": "These may share the same template, but there is no evidence they target the same language or same request." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 5, + "component_count": 13, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-36817-14", + "nodes": [ + "issue:36817", + "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:37015|issue:43837", + "issue:37015|issue:44121", + "issue:37015|issue:43909", + "issue:37015|issue:42503", + "issue:37015|issue:45027", + "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:43837|issue:45177" + ], + "prepared_review_unit_hash": "6bfbcd9cc0fffd721ccbbf98669bbd8dddcc1f3c86c5c206ed0d3c05d2935e36", + "component_budget": { + "node_count": 14, + "item_count": 14, + "soft_pair_count": 28, + "serialized_chars": 17646, + "estimated_input_tokens": 4412, + "estimated_eval_tokens": 9080 + }, + "budget": { + "node_count": 14, + "item_count": 14, + "soft_pair_count": 28, + "serialized_chars": 17646, + "estimated_input_tokens": 4412, + "estimated_eval_tokens": 9080 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is mostly a set of separate model-addition requests, not duplicates. The similarity signal is driven by generic wording like \u201cadd model\u201d rather than the same underlying bug/change.", + "confidence": 0.95, + "canonical_issue_reason": "No single canonical issue: each item requests a different model or modality (BERT variants, audio/TTS/ASR, OCR, vision/text), so they are not duplicates.", + "canonical_pr_reason": null, + "best_issue_reason": "No strong duplicate candidate exists. If a representative issue were needed, issue 43671 is one of the more specific and active requests, but it is still not a duplicate of the others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36817", + "right": "issue:41211", + "accept": false, + "reason": "Different model families (EuroBert vs DEIMv2); generic \u201cadd model\u201d similarity only." + }, + { + "left": "issue:36817", + "right": "issue:43671", + "accept": false, + "reason": "EuroBert request is unrelated to Qwen3-TTS support; different model and task." + }, + { + "left": "issue:37015", + "right": "issue:43837", + "accept": false, + "reason": "NeoBERT and Qwen3-ASR are different models with different modalities/tasks." + }, + { + "left": "issue:37015", + "right": "issue:44121", + "accept": false, + "reason": "NeoBERT request is unrelated to OpenAI weight-sparse transformer support." + }, + { + "left": "issue:37015", + "right": "issue:43909", + "accept": false, + "reason": "NeoBERT and LFM2.5 Audio are different model requests, not the same change." + }, + { + "left": "issue:37015", + "right": "issue:42503", + "accept": false, + "reason": "NeoBERT vs ModernVBERT: different model families; no shared concrete code path." + }, + { + "left": "issue:37015", + "right": "issue:45027", + "accept": false, + "reason": "NeoBERT and Voxtral-4B-TTS are separate model additions, not duplicates." + }, + { + "left": "issue:41115", + "right": "issue:43837", + "accept": false, + "reason": "MiniCPM3 architecture support and Qwen3-ASR support are different models/tasks." + }, + { + "left": "issue:41115", + "right": "issue:44121", + "accept": false, + "reason": "MiniCPM3 and circuit-sparsity/circuitgpt are unrelated model requests." + }, + { + "left": "issue:41115", + "right": "issue:42503", + "accept": false, + "reason": "MiniCPM3 vs ModernVBERT: different concrete model additions." + }, + { + "left": "issue:41115", + "right": "issue:45027", + "accept": false, + "reason": "MiniCPM3 and Voxtral-TTS are distinct model families and modalities." + }, + { + "left": "issue:41211", + "right": "issue:43671", + "accept": false, + "reason": "DEIMv2 and Qwen3-TTS are different models with different target behaviors." + }, + { + "left": "issue:42503", + "right": "issue:43837", + "accept": false, + "reason": "ModernVBERT models and Qwen3-ASR support are unrelated model additions." + }, + { + "left": "issue:42503", + "right": "issue:43671", + "accept": false, + "reason": "ModernVBERT and Qwen3-TTS are different model families and tasks." + }, + { + "left": "issue:42503", + "right": "issue:42738", + "accept": false, + "reason": "ModernVBERT and generic BERT-like RoPE support are not the same underlying change." + }, + { + "left": "issue:42738", + "right": "issue:45027", + "accept": false, + "reason": "BERT-like RoPE support and Voxtral TTS are different problem domains." + }, + { + "left": "issue:43671", + "right": "issue:43837", + "accept": false, + "reason": "Qwen3-TTS and Qwen3-ASR are related by name but are distinct models and separate support requests." + }, + { + "left": "issue:43671", + "right": "issue:45027", + "accept": false, + "reason": "Both are audio-related, but they target different models and integrations." + }, + { + "left": "issue:43671", + "right": "issue:45177", + "accept": false, + "reason": "Qwen3-TTS support is unrelated to DeepSeek-OCR2." + }, + { + "left": "issue:43671", + "right": "issue:44121", + "accept": false, + "reason": "Qwen3-TTS and weight-sparse transformer support are different model changes." + }, + { + "left": "issue:43671", + "right": "issue:43909", + "accept": false, + "reason": "Qwen3-TTS and LFM2.5 Audio 1.5B are separate audio-model requests." + }, + { + "left": "issue:43837", + "right": "issue:43909", + "accept": false, + "reason": "Qwen3-ASR and LFM2.5 Audio are distinct audio-model support requests." + }, + { + "left": "issue:43837", + "right": "issue:45322", + "accept": false, + "reason": "Qwen3-ASR support and EUPE are different models with no shared concrete bug/change." + }, + { + "left": "issue:43837", + "right": "issue:44121", + "accept": false, + "reason": "ASR model support is unrelated to the weight-sparse transformer request." + }, + { + "left": "issue:43909", + "right": "issue:44205", + "accept": false, + "reason": "LFM2.5 Audio 1.5B and SAM3-LiteText are different model additions." + }, + { + "left": "issue:44121", + "right": "issue:45322", + "accept": false, + "reason": "Weight-sparse transformer/circuitgpt and EUPE are unrelated model requests." + }, + { + "left": "issue:43837", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-ASR and Voxtral-4B-TTS are different speech-model support requests." + }, + { + "left": "issue:43837", + "right": "issue:45177", + "accept": false, + "reason": "Qwen3-ASR support is unrelated to DeepSeek-OCR2." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 6, + "component_count": 13, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-37934-6", + "nodes": [ + "issue:37934", + "issue:42936", + "issue:43240", + "issue:43668", + "issue:43813", + "issue:44031" + ], + "soft_pairs": [ + "issue:37934|issue:43668", + "issue:42936|issue:43240", + "issue:42936|issue:44031", + "issue:42936|issue:43813", + "issue:42936|issue:43668", + "issue:43240|issue:43668", + "issue:43668|issue:44031", + "issue:43668|issue:43813" + ], + "prepared_review_unit_hash": "5b1f6386955b5ba32a0377b1a492ffb28f6b886be3f8b15dd0673e606ef09c45", + "component_budget": { + "node_count": 6, + "item_count": 6, + "soft_pair_count": 8, + "serialized_chars": 6288, + "estimated_input_tokens": 1572, + "estimated_eval_tokens": 3400 + }, + "budget": { + "node_count": 6, + "item_count": 6, + "soft_pair_count": 8, + "serialized_chars": 6288, + "estimated_input_tokens": 1572, + "estimated_eval_tokens": 3400 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These items are a heterogeneous set of unrelated issues (model norm behavior, Mask2Former loss args, config typing, a typo, and tokenizer regex warnings). The soft-similarity signals are too weak to treat any pair as duplicates.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:37934", + "right": "issue:43668", + "accept": false, + "reason": "Different concerns: Llama4TextL2Norm implementation semantics vs ModernBERTConfig type-hint correctness." + }, + { + "left": "issue:42936", + "right": "issue:43240", + "accept": false, + "reason": "Mask2Former ignore_value handling is a model-specific bug; kwargs loss forwarding is a broader API wiring issue." + }, + { + "left": "issue:42936", + "right": "issue:44031", + "accept": false, + "reason": "Unrelated subsystems: Mask2Former loss parameter usage vs tokenizer regex warning regression." + }, + { + "left": "issue:42936", + "right": "issue:43813", + "accept": false, + "reason": "Different bug class and area: model loss config usage vs a typo in a quantization field name." + }, + { + "left": "issue:42936", + "right": "issue:43668", + "accept": false, + "reason": "No shared underlying defect; one is runtime behavior in a model, the other is a config annotation issue." + }, + { + "left": "issue:43240", + "right": "issue:43668", + "accept": false, + "reason": "Loss kwargs forwarding and incorrect type hints are separate issues with no common code-path." + }, + { + "left": "issue:43668", + "right": "issue:44031", + "accept": false, + "reason": "ModernBERT config typing is unrelated to tokenizer regex warnings." + }, + { + "left": "issue:43668", + "right": "issue:43813", + "accept": false, + "reason": "A type-hint mismatch and a typo in a field path are distinct, non-duplicate 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": 13, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41084-2", + "nodes": [ + "issue:41084", + "issue:43037" + ], + "soft_pairs": [ + "issue:41084|issue:43037" + ], + "prepared_review_unit_hash": "ee99f254c5eef4357c55fdf3b8a5b18278fe1d7880966fd926d5d5b65a7f5df4", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1480, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1480, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues are unrelated: one requests block decoding behavior, the other requests DeepSeek v3.2 model support. They should not be clustered as duplicates.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41084", + "right": "issue:43037", + "accept": false, + "reason": "Different topics and likely different workstreams: block decoding vs DeepSeek v3.2 support; no evidence they describe the same underlying bug or feature." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 8, + "component_count": 13, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43598-2", + "nodes": [ + "issue:43598", + "issue:43600" + ], + "soft_pairs": [ + "issue:43598|issue:43600" + ], + "prepared_review_unit_hash": "4fc6b68ff6bdc51c8e2bf39fd4be5dacd45bd1756e920b0db6f2790f88bf47b1", + "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": "The two closed issues are about different Trainer-related customization problems: one concerns `remove_unused_column`, the other concerns FLOPS estimation for custom trainers. They share a broad area but not the same underlying bug or fix.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "Neither issue is a duplicate of the other. If one must be picked as the representative, issue #43598 is the more concrete Trainer customization problem, but the cluster does not justify a single canonical issue.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43598", + "right": "issue:43600", + "accept": false, + "reason": "Both mention Trainer customization, but #43598 is about `remove_unused_column` behavior while #43600 is about estimating FLOPS for custom trainers. They describe different code paths and different fixes." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 9, + "component_count": 13, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43684-2", + "nodes": [ + "issue:43684", + "issue:44405" + ], + "soft_pairs": [ + "issue:43684|issue:44405" + ], + "prepared_review_unit_hash": "581542f42486834982c1808003961f2e5b9891a7119ddd0536d4a8a1afa2cc0b", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1611, + "estimated_input_tokens": 403, + "estimated_eval_tokens": 1062 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1611, + "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 two separate model-support requests: one for Qwen3-Omni auto-class registration, and one for Qwen3.5 sequence-classification support. They share the general theme of adding AutoModel support, but not the same concrete bug/change.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43684 is the broader auto-registration request and is the clearer representative of the Qwen3-Omni gap; issue 44405 is a distinct Qwen3.5 sequence-classification support request, not a duplicate.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43684", + "right": "issue:44405", + "accept": false, + "reason": "Different model families and different auto classes: Qwen3-Omni registration for AutoModel/AutoModelForConditionalGeneration vs Qwen3.5 sequence-classification support. They are related only at a high level, not the same underlying change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 10, + "component_count": 13, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43835-4", + "nodes": [ + "issue:43835", + "issue:43845", + "issue:43846", + "issue:43859" + ], + "soft_pairs": [ + "issue:43835|issue:43845", + "issue:43835|issue:43846", + "issue:43835|issue:43859", + "issue:43845|issue:43846", + "issue:43845|issue:43859", + "issue:43846|issue:43859" + ], + "prepared_review_unit_hash": "4aceadcf55ba70db28212dd61d0de279a6ca17591306cd06ab01a0a2f3431209", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4265, + "estimated_input_tokens": 1067, + "estimated_eval_tokens": 2390 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4265, + "estimated_input_tokens": 1067, + "estimated_eval_tokens": 2390 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All four issues are extremely low-signal and mostly consist of vague or near-empty titles. Three share the same title ('huggingface '), while 'Fett' looks unrelated. There is not enough concrete content to confidently establish true duplicates, but 43845 is the best placeholder representative among the repeated-title issues.", + "confidence": 0.41, + "canonical_issue_reason": "Issue 43845 is the earliest of the repeated 'huggingface ' reports and is the cleanest placeholder representative for that cluster, though the bodies are too sparse to prove a real underlying bug.", + "canonical_pr_reason": null, + "best_issue_reason": "43845 is the most suitable global issue representative because it is earliest among the near-identical 'huggingface ' issues and has no evidence of being different from 43846 or 43859.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43835", + "right": "issue:43845", + "accept": false, + "reason": "Titles are generic and unrelated ('Fett' vs 'huggingface '); no shared body details or targets to support a duplicate determination." + }, + { + "left": "issue:43835", + "right": "issue:43846", + "accept": false, + "reason": "Same as above: no concrete evidence that 'Fett' and 'huggingface ' describe the same underlying issue." + }, + { + "left": "issue:43835", + "right": "issue:43859", + "accept": false, + "reason": "No meaningful overlap beyond both being vague; insufficient evidence of the same bug." + }, + { + "left": "issue:43845", + "right": "issue:43846", + "accept": false, + "reason": "The titles match, but the issue bodies are too thin to verify they are the same underlying problem; identical generic titles alone are not enough." + }, + { + "left": "issue:43845", + "right": "issue:43859", + "accept": false, + "reason": "All we have is the same generic title; there are no concrete symptoms, targets, or code references to confirm duplication." + }, + { + "left": "issue:43846", + "right": "issue:43859", + "accept": false, + "reason": "Identical vague titles are suggestive, but the lack of substantive description makes a strict duplicate match unjustified." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 11, + "component_count": 13, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44371-5", + "nodes": [ + "issue:44371", + "issue:44418", + "issue:44572", + "issue:44573", + "issue:44937" + ], + "soft_pairs": [ + "issue:44371|issue:44572", + "issue:44371|issue:44573", + "issue:44371|issue:44418", + "issue:44371|issue:44937", + "issue:44418|issue:44572", + "issue:44418|issue:44573", + "issue:44572|issue:44573", + "issue:44572|issue:44937", + "issue:44573|issue:44937" + ], + "prepared_review_unit_hash": "4d32738a40266427712efb1b54036146fd7e4aa3c6838bca7f06614f037eb622", + "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": "All items look like unrelated low-signal/spam issues rather than duplicates of a real bug report. No soft edge should be merged.", + "confidence": 0.88, + "canonical_issue_reason": "issue:44371 is the simplest generic spam placeholder and the earliest item, so it is the least misleading representative for the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44371 is the best available issue representative because it is generic, minimal, and there is no substantive technical content to anchor a more specific choice.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44371", + "right": "issue:44572", + "accept": false, + "reason": "Both are spam-like placeholders, but there is no evidence they describe the same underlying report or change." + }, + { + "left": "issue:44371", + "right": "issue:44573", + "accept": false, + "reason": "Same generic spam style, but no shared bug, feature, or reproduction details to justify duplication." + }, + { + "left": "issue:44371", + "right": "issue:44418", + "accept": false, + "reason": "One is a generic spam title and the other is a documentation suggestion; too little evidence they are the same issue." + }, + { + "left": "issue:44371", + "right": "issue:44937", + "accept": false, + "reason": "Completely different titles and no technical details tying them to the same underlying problem." + }, + { + "left": "issue:44418", + "right": "issue:44572", + "accept": false, + "reason": "Documentation suggestion versus generic spam; not the same underlying issue." + }, + { + "left": "issue:44418", + "right": "issue:44573", + "accept": false, + "reason": "Different apparent intent and no shared concrete content beyond weak similarity signals." + }, + { + "left": "issue:44572", + "right": "issue:44573", + "accept": false, + "reason": "Both are spam placeholders, but that is not enough to conclude they are duplicate reports of the same thing." + }, + { + "left": "issue:44572", + "right": "issue:44937", + "accept": false, + "reason": "Generic spam title versus a specific quoted name; no evidence of the same underlying issue." + }, + { + "left": "issue:44573", + "right": "issue:44937", + "accept": false, + "reason": "No concrete overlap in problem statement, targets, or artifacts; reject as duplicate." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 12, + "component_count": 13, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45145-2", + "nodes": [ + "issue:45145", + "issue:45151" + ], + "soft_pairs": [ + "issue:45145|issue:45151" + ], + "prepared_review_unit_hash": "350f17c51c7a1e5c256a646be320d3d7a061b73b25c701bc88548180a0457d1f", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1646, + "estimated_input_tokens": 412, + "estimated_eval_tokens": 1080 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1646, + "estimated_input_tokens": 412, + "estimated_eval_tokens": 1080 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two closed issues appear to be likely duplicates or near-duplicate reports: they have the exact same title, similar body sizes, and a soft-similarity signal, with no distinguishing links, files, or targets shown. #45145 is the better representative because it was created earlier and has an inbound reference.", + "confidence": 0.88, + "canonical_issue_reason": "#45145 is the better canonical pick because it was created earlier and has an inbound reference, while #45151 appears to be a near-duplicate report with the same title.", + "canonical_pr_reason": null, + "best_issue_reason": "#45145 is the best issue to keep as the representative item because it is earlier and already has an inbound reference, making it the more established anchor for this duplicate pair.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45145", + "right": "issue:45151", + "accept": true, + "reason": "Exact same title plus similar body length and a soft-similarity signal make these look like duplicate reports of the same underlying issue." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The cluster call is a bit too strong for the evidence shown. The packet supports that these issues have the same title, similar sizes, and a soft-similarity signal, but it does not expose the bodies, so statements like \"highly similar content\" / \"substantially similar body content\" are not directly grounded. Also, the best-issue rationale overstates linkage/activity a bit: #45145 has the inbound reference, but it does not have more discussion activity than #45151. A more conservative summary would say they appear likely duplicate or near-duplicate reports, with #45145 as the better representative because it is earlier and has an inbound reference." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 13, + "component_count": 13, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45295-2", + "nodes": [ + "issue:45295", + "issue:45373" + ], + "soft_pairs": [], + "prepared_review_unit_hash": "f5fd95fada2f530ab0999e343c17c16266cde0c7182bb22509972ca44827cd49", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 0, + "serialized_chars": 1286, + "estimated_input_tokens": 322, + "estimated_eval_tokens": 900 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 0, + "serialized_chars": 1286, + "estimated_input_tokens": 322, + "estimated_eval_tokens": 900 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues describe the same missing Gemma 4 sequence-classification support; the later, more specific open issue is the better cluster representative.", + "confidence": 0.94, + "canonical_issue_reason": "issue #45373 is the clearest canonical tracker: it is open, explicitly names the missing class `Gemma4ForSequenceClassification`, and states the module-level gap relative to Gemma 2/3.", + "canonical_pr_reason": null, + "best_issue_reason": "issue #45373 best captures the concrete request and likely remaining work, while #45295 is a shorter earlier duplicate with less detail and is already closed.", + "best_pr_reason": null, + "soft_edge_verdicts": [] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 1, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-39895-13", + "nodes": [ + "pull_request:39895", + "pull_request: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:45587", + "pull_request:45622" + ], + "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:43995", + "pull_request:43823|pull_request:45622", + "pull_request:43995|pull_request:45213", + "pull_request:43995|pull_request:44827", + "pull_request:44827|pull_request:45622", + "pull_request:44827|pull_request:45587", + "pull_request:44827|pull_request:45170", + "pull_request:44827|pull_request:45213", + "pull_request:43823|pull_request:45213", + "pull_request:43251|pull_request:43254" + ], + "prepared_review_unit_hash": "4a04e67c3f56402c39cfd7388722cbeb74f085903a1eaee1b16a634240919561", + "component_budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 18, + "serialized_chars": 35176, + "estimated_input_tokens": 8794, + "estimated_eval_tokens": 17844 + }, + "budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 18, + "serialized_chars": 35176, + "estimated_input_tokens": 8794, + "estimated_eval_tokens": 17844 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Only one pair looks like a true duplicate: the two near-identical loss-utils PRs (43251/43254). The rest are unrelated same-subsystem changes, broad model additions, or docs/test-only updates and should not be merged as duplicates.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43251 is the best canonical representative of the duplicate loss-utils fix: it has the explicit issue link, directly patches fixed_cross_entropy, and is the clearer of the two near-identical implementations.", + "best_issue_reason": null, + "best_pr_reason": "PR 43251 is the strongest overall PR in the cluster for duplicate-triage purposes because it most cleanly captures the shared bug in fixed_cross_entropy and has the most concrete issue alignment.", + "soft_edge_verdicts": [ + { + "left": "pull_request:39895", + "right": "pull_request:43424", + "accept": false, + "reason": "Videoprism model addition vs executorch dynamic-shape test; different features and code paths." + }, + { + "left": "pull_request:41855", + "right": "pull_request:43823", + "accept": false, + "reason": "Mistral tokenizer methods vs adding MobileLLM; unrelated changes with only superficial tokenizer/model overlap." + }, + { + "left": "pull_request:41855", + "right": "pull_request:43424", + "accept": false, + "reason": "Tokenizer API additions vs executorch export test; no shared bug or implementation change." + }, + { + "left": "pull_request:42668", + "right": "pull_request:43823", + "accept": false, + "reason": "Processor auto-loading robustness vs adding a new model; same broad area of model support, but not the same fix." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43823", + "accept": false, + "reason": "Loss-function kwargs forwarding vs new MobileLLM model files; completely different code paths." + }, + { + "left": "pull_request:43251", + "right": "pull_request:44827", + "accept": false, + "reason": "Cross-entropy kwargs fix vs Mistral4 test/model refactor; unrelated bugs and changes." + }, + { + "left": "pull_request:43424", + "right": "pull_request:44827", + "accept": false, + "reason": "Executorch dynamic-shape test vs Mistral4 test/model refactor; different functionality and no common fix." + }, + { + "left": "pull_request:43424", + "right": "pull_request:43823", + "accept": false, + "reason": "Executorch export testing vs MobileLLM model addition; unrelated changes." + }, + { + "left": "pull_request:43823", + "right": "pull_request:43995", + "accept": false, + "reason": "New MobileLLM model implementation vs Falcon output-collection refactor; different model families and problems." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45622", + "accept": false, + "reason": "MobileLLM model addition vs PEFT constructor fix; unrelated code paths." + }, + { + "left": "pull_request:43995", + "right": "pull_request:45213", + "accept": false, + "reason": "Falcon refactor vs model-creation skill workflow; not the same bug or change." + }, + { + "left": "pull_request:43995", + "right": "pull_request:44827", + "accept": false, + "reason": "Falcon output interface refactor vs Mistral4 tests/compatibility fix; different models and goals." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45622", + "accept": false, + "reason": "Mistral4 test/model fixes vs PEFT constructor fix; unrelated changes despite both touching framework internals." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45587", + "accept": false, + "reason": "Mistral4-related fix vs continuous batching docs; one is code/tests, the other is documentation." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45170", + "accept": false, + "reason": "Mistral4 test/model work vs CLIP-like weight-renaming fix; different subsystems and concrete issues." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45213", + "accept": false, + "reason": "Mistral4 fix vs a draft model-creation skill PR; not the same underlying change." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45213", + "accept": false, + "reason": "MobileLLM model addition vs model-creation skill scaffolding; one creates a model, the other adds workflow tooling." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43254", + "accept": true, + "reason": "Both PRs patch the same fixed_cross_entropy code path to forward supported cross_entropy kwargs; the diffs are near-identical and could be merged into one fix." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded and conservative. The only accepted duplicate pair, 43251/43254, is supported by the shared loss_utils file, the shared issue target, and near-identical fixed_cross_entropy patches. The other pairs are reasonably rejected as different bugfixes, features, or docs/tests rather than mergeable duplicates." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 2, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-40546-2", + "nodes": [ + "pull_request:40546", + "pull_request:42781" + ], + "soft_pairs": [ + "pull_request:40546|pull_request:42781" + ], + "prepared_review_unit_hash": "1584b7ac07cc91c523316348d74a8bea0ef16322c7d9b77f3ae17705ba132ace", + "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 core VibeVoice implementation, the other is a draft for VibeVoice Realtime that adds extra realtime-specific files and changes. The shared filenames mostly reflect the common model family, not the same exact change.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #40546 is the better canonical representative because it is the fuller, reviewed core VibeVoice implementation with broad model and docs integration.", + "best_issue_reason": null, + "best_pr_reason": "PR #40546 is the best overall representative of the cluster: it is the main implementation PR, while #42781 is a draft follow-up for a more specific realtime variant.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40546", + "right": "pull_request:42781", + "accept": false, + "reason": "Reject: they share the VibeVoice codebase, but #42781 is a separate draft for VibeVoice Realtime with additional realtime-specific files and changes, not the same concrete code-path fix/change as #40546." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 3, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-40755-2", + "nodes": [ + "pull_request:40755", + "pull_request:41224" + ], + "soft_pairs": [ + "pull_request:40755|pull_request:41224" + ], + "prepared_review_unit_hash": "c78a4e744c187bc6677f105e311280c04222ff551f2e406baf665f4f8b2f3f0e", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4675, + "estimated_input_tokens": 1169, + "estimated_eval_tokens": 2594 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4675, + "estimated_input_tokens": 1169, + "estimated_eval_tokens": 2594 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The pair is not a duplicate cluster: one PR adds TimesFM covariate forecasting, the other adds DINOv3 image classification support. They target different models and different code paths.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 41224 is the better canonical representative because it is a focused, self-contained model feature with substantial review activity and a clear code/test surface.", + "best_issue_reason": null, + "best_pr_reason": "PR 41224 is the strongest standalone PR in the cluster: it is narrowly scoped, implementation-backed, and has more review evidence than the broader TimesFM feature PR.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40755", + "right": "pull_request:41224", + "accept": false, + "reason": "Different features for different models (TimesFM covariates vs DINOv3 image classification); no shared bug or change to 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": 4, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41356-2", + "nodes": [ + "pull_request:41356", + "pull_request:44339" + ], + "soft_pairs": [ + "pull_request:41356|pull_request:44339" + ], + "prepared_review_unit_hash": "5ca5622c9633bbe7eef9d70a2eb0e3cf7982d42dc10a27ffb1990f021d804327", + "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": "Both items are PRs for adding DEIMv2 support to Transformers. PR #44339 is the more complete later submission and appears to subsume the earlier #41356 work, so it is the better canonical PR candidate. There is no issue in this cluster.", + "confidence": 0.86, + "canonical_issue_reason": null, + "canonical_pr_reason": "#44339 is the broader, later DEIMv2 implementation and includes the core model integration plus additional framework wiring (auto-mappings, conversion, loss/modular updates), making it the best representative PR.", + "best_issue_reason": null, + "best_pr_reason": "#44339 is the most complete and maintainable PR for the DEIMv2 addition; it covers the same feature area as #41356 but with substantially more finished integration work.", + "soft_edge_verdicts": [ + { + "left": "pull_request:41356", + "right": "pull_request:44339", + "accept": true, + "reason": "Both PRs target the same DEIMv2 model addition, and #44339 looks like a superset that could plausibly replace the earlier implementation in one merged change." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: both items are PRs tied to the same issue target (41211) and the later PR #44339 clearly adds broader DEIMv2 integration while overlapping core files with #41356. The soft-edge verdict is conservative, phrased as a plausible supersession rather than a certainty. No issue is present in this cluster." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 5, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41797-2", + "nodes": [ + "pull_request:41797", + "pull_request:45186" + ], + "soft_pairs": [ + "pull_request:41797|pull_request:45186" + ], + "prepared_review_unit_hash": "61c25d0acdfc8da32f5afbba697ca7ba09864588c591c5a3f93f2131a9c1dbff", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6722, + "estimated_input_tokens": 1681, + "estimated_eval_tokens": 3618 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6722, + "estimated_input_tokens": 1681, + "estimated_eval_tokens": 3618 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two distinct model-addition pull requests: one for DeepSeek OCR and one for Isaac. They share only common Transformers integration boilerplate, not the same underlying code change.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "PR 41797 is the stronger standalone representative of the cluster because it has broader implementation scope and more review activity, but it is not a duplicate of PR 45186.", + "soft_edge_verdicts": [ + { + "left": "pull_request:41797", + "right": "pull_request:45186", + "accept": false, + "reason": "Different models and different codepaths: DeepSeek OCR vs Isaac. Shared files are generic registry/docs plumbing, not evidence of the same fix or mergeable change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 6, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-42166-2", + "nodes": [ + "pull_request:42166", + "pull_request:45426" + ], + "soft_pairs": [ + "pull_request:42166|pull_request:45426" + ], + "prepared_review_unit_hash": "42e6a3ab10d2a44c0bd004cfe31c9983aa8b6a61bfd6c988d5a1f98e4a44d7f5", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5407, + "estimated_input_tokens": 1352, + "estimated_eval_tokens": 2960 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5407, + "estimated_input_tokens": 1352, + "estimated_eval_tokens": 2960 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two PRs are separate model-addition workstreams (InternVLFlash vs AXK1). They overlap only in shared registry/boilerplate files, not in the underlying change or code path.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: these are different model integrations with different implementation files and different goals, so they are not duplicates.", + "best_issue_reason": null, + "best_pr_reason": "No single best PR for deduplication. If forced to pick a representative, 42166 is more mature because it includes docs and tests, but it is still unrelated to 45426.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42166", + "right": "pull_request:45426", + "accept": false, + "reason": "Shared auto-registry and package plumbing only; one adds InternVLFlash and the other adds AXK1, so they do not fix or implement the same underlying code-path change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 7, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-42227-4", + "nodes": [ + "pull_request:42227", + "pull_request:43291", + "pull_request:43665", + "pull_request:44815" + ], + "soft_pairs": [ + "pull_request:42227|pull_request:43291", + "pull_request:42227|pull_request:43665", + "pull_request:42227|pull_request:44815", + "pull_request:43291|pull_request:43665", + "pull_request:43291|pull_request:44815", + "pull_request:43665|pull_request:44815" + ], + "prepared_review_unit_hash": "5f5636f86a017ef44cd3ec8295f538556a3761fb687246323de9490bc94a8e1a", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 10169, + "estimated_input_tokens": 2543, + "estimated_eval_tokens": 5342 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 10169, + "estimated_input_tokens": 2543, + "estimated_eval_tokens": 5342 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs do not look like duplicates of the same underlying change. Two are Whisper-related, one is generic test noise, and one is an unrelated FP8/dequantization fix. Only the Whisper return-language PR appears to be a real bug fix; the others are too different in code path and intent.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 42227 is the clearest substantive fix: it targets a specific Whisper ASR bug, has an explicit issue link, and changes the actual pipeline code path rather than only tests or unrelated subsystems.", + "best_issue_reason": null, + "best_pr_reason": "PR 42227 is the best representative artifact in this cluster because it is a concrete production fix with issue linkage and clear behavior change; the others are either test-only/junk-like or unrelated to the same bug.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42227", + "right": "pull_request:43291", + "accept": false, + "reason": "Both are Whisper-related, but they fix different code paths: 42227 propagates return_language through the ASR pipeline, while 43291 changes tokenizer decode/timestamp handling and tests. They do not look mergeable as one duplicate PR." + }, + { + "left": "pull_request:42227", + "right": "pull_request:43665", + "accept": false, + "reason": "42227 is a real Whisper pipeline bug fix; 43665 is just placeholder test edits across CLIP/ViT and does not address the same underlying change." + }, + { + "left": "pull_request:42227", + "right": "pull_request:44815", + "accept": false, + "reason": "These are unrelated subsystems and problems: Whisper ASR parameter propagation versus FP8/dequant/model-loading changes. No shared concrete bug." + }, + { + "left": "pull_request:43291", + "right": "pull_request:43665", + "accept": false, + "reason": "43291 modifies Whisper tokenizer behavior; 43665 is test noise in vision model tests. Different intent and no shared underlying bug." + }, + { + "left": "pull_request:43291", + "right": "pull_request:44815", + "accept": false, + "reason": "Whisper tokenizer decode/timestamp logic is unrelated to FP8 dequantization/model loading. They do not fix the same issue." + }, + { + "left": "pull_request:43665", + "right": "pull_request:44815", + "accept": false, + "reason": "One is placeholder test edits, the other is a substantive quantization/dequant fix. Not the same bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 8, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-42975-2", + "nodes": [ + "pull_request:42975", + "pull_request:42976" + ], + "soft_pairs": [ + "pull_request:42975|pull_request:42976" + ], + "prepared_review_unit_hash": "92ef2a134638e3629d7417b08d4d00d2953b55d7b40a09cfb4cf2a4f1ad25b20", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5935, + "estimated_input_tokens": 1484, + "estimated_eval_tokens": 3224 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5935, + "estimated_input_tokens": 1484, + "estimated_eval_tokens": 3224 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related workflow-maintenance PRs, but they change different GitHub Actions dependencies and solve different upgrade concerns. They are similar in theme, not the same underlying change.", + "confidence": 0.89, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #42975 is the broader, more specific workflow-runtime upgrade: it updates several core actions for Node 24 compatibility and touches more workflow entry points. It is a better umbrella representative of this maintenance cluster than the narrower Docker-action bump in #42976.", + "best_issue_reason": null, + "best_pr_reason": "PR #42975 is the best single representative because its title and diff clearly describe the main compatibility-driven Actions upgrade, while #42976 is a smaller, separate dependency refresh focused on docker/build-push-action.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42975", + "right": "pull_request:42976", + "accept": false, + "reason": "Both are GitHub Actions version bumps, but they update different actions in different workflows (core checkout/cache/upload-artifact vs docker/build-push-action). They are related maintenance, not the same concrete bug or change, so I would not merge them as duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 9, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-42978-9", + "nodes": [ + "pull_request:42978", + "pull_request:43448", + "pull_request:43451", + "pull_request:43973", + "pull_request:44601", + "pull_request:45283", + "pull_request:45490", + "pull_request:45626", + "pull_request:45630" + ], + "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:45283", + "pull_request:44601|pull_request:45626", + "pull_request:45283|pull_request:45626", + "pull_request:45490|pull_request:45626", + "pull_request:45490|pull_request:45630", + "pull_request:45626|pull_request:45630" + ], + "prepared_review_unit_hash": "f1e612125d6dfcc47d82456fe0eeb0b1ba89e20ee1de622adcdaabf7fb26103a", + "component_budget": { + "node_count": 9, + "item_count": 9, + "soft_pair_count": 12, + "serialized_chars": 29350, + "estimated_input_tokens": 7338, + "estimated_eval_tokens": 14932 + }, + "budget": { + "node_count": 9, + "item_count": 9, + "soft_pair_count": 12, + "serialized_chars": 29350, + "estimated_input_tokens": 7338, + "estimated_eval_tokens": 14932 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are not duplicates: they cover different model additions and one distributed-loading feature, with only shared boilerplate in docs/auto registration. All soft edges should be rejected.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR fits this cluster; the items are distinct feature/model additions rather than the same underlying change.", + "best_issue_reason": null, + "best_pr_reason": "No PR is a good global representative because the similarities are superficial (docs, auto-mapping, tests) and the concrete code paths differ.", + "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 boilerplate." + }, + { + "left": "pull_request:42978", + "right": "pull_request:43451", + "accept": false, + "reason": "Different models (ViT NEPA vs Molmo2) with distinct files and code paths." + }, + { + "left": "pull_request:42978", + "right": "pull_request:45490", + "accept": false, + "reason": "ViT NEPA model addition vs CTSM model addition; not the same bug/change." + }, + { + "left": "pull_request:43448", + "right": "pull_request:43451", + "accept": false, + "reason": "Molmo and Molmo2 are separate model support PRs, not one fix split across two changes." + }, + { + "left": "pull_request:43448", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo vs CTSM are unrelated model additions; shared auto/model-doc edits are boilerplate." + }, + { + "left": "pull_request:43451", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo2 and CTSM are different models with different implementation files and targets." + }, + { + "left": "pull_request:43973", + "right": "pull_request:45283", + "accept": false, + "reason": "Lfm2Audio model support and Qwen3.5 GGUF loading are different feature areas and code paths." + }, + { + "left": "pull_request:44601", + "right": "pull_request:45626", + "accept": false, + "reason": "Pipeline-parallel infrastructure work is unrelated to PP-FormulaNet model support." + }, + { + "left": "pull_request:45283", + "right": "pull_request:45626", + "accept": false, + "reason": "GGUF loading support for Qwen3.5 and PP-FormulaNet model integration are distinct changes." + }, + { + "left": "pull_request:45490", + "right": "pull_request:45626", + "accept": false, + "reason": "CTSM model support and PP-FormulaNet model support are separate model additions." + }, + { + "left": "pull_request:45490", + "right": "pull_request:45630", + "accept": false, + "reason": "CTSM and Kimi2-6 are different models; similarity is limited to standard auto/docs scaffolding." + }, + { + "left": "pull_request:45626", + "right": "pull_request:45630", + "accept": false, + "reason": "PP-FormulaNet and Kimi2-6 are unrelated model support PRs with different implementation files." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 10, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43488-3", + "nodes": [ + "pull_request:43488", + "pull_request:43532", + "pull_request:45476" + ], + "soft_pairs": [ + "pull_request:43488|pull_request:43532", + "pull_request:43488|pull_request:45476", + "pull_request:43532|pull_request:45476" + ], + "prepared_review_unit_hash": "2d6920781c3cc68e5504d58e25f0d34fe74baf8a6f33939e3ddfa6159af59186", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 7468, + "estimated_input_tokens": 1867, + "estimated_eval_tokens": 3990 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 7468, + "estimated_input_tokens": 1867, + "estimated_eval_tokens": 3990 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All three are unrelated synthetic \"don't merge\" PRs that only overlap on broad test/workflow areas and a shared vit test file; none look like the same underlying change or bug fix.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: these are separate, non-duplicative bot/demo changes with different purposes (setup/versioning, workflow display tweaks, and CI/workflow invocation plus unrelated code noise).", + "best_issue_reason": null, + "best_pr_reason": "No single PR is a good cluster representative for deduplication because the overlap is superficial and the concrete code paths differ.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43488", + "right": "pull_request:43532", + "accept": false, + "reason": "Shared vit test filename is not enough: 43488 is a setup/dependency/version-formatting PR, while 43532 changes workflow automation and testing utilities. Different underlying changes." + }, + { + "left": "pull_request:43488", + "right": "pull_request:45476", + "accept": false, + "reason": "43488 only tweaks setup metadata and dependency listing; 45476 adds a workflow caller and unrelated tokenization/test noise. They do not fix the same concrete problem." + }, + { + "left": "pull_request:43532", + "right": "pull_request:45476", + "accept": false, + "reason": "Both are synthetic PRs with some overlapping test/workflow files, but they touch different code paths and goals; no evidence they could be merged into one 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": 11, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43612-2", + "nodes": [ + "pull_request:43612", + "pull_request:43613" + ], + "soft_pairs": [ + "pull_request:43612|pull_request:43613" + ], + "prepared_review_unit_hash": "04e50e9bc2e5702c735c83fcbae1eb6f68ec7ee18f74340263f330d25072ba0d", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6083, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6083, + "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 closely named but distinct pipeline additions: one for Promptable Concept Segmentation and one for Promptable Visual Segmentation. They share documentation and registry plumbing, but the actual pipeline implementations, model integrations, and tests target different tasks, so they should not be deduplicated.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "No true duplicate PR pair here; if selecting a single representative, PR 43613 is the stronger/broader one because it adds the visual-segmentation pipeline with multiple model integrations and dedicated tests/docs.", + "best_issue_reason": null, + "best_pr_reason": "PR 43613 is the best representative of the cluster because it is the more complete feature branch and has clearer task scope with broader supporting changes.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43612", + "right": "pull_request:43613", + "accept": false, + "reason": "Different underlying features and code paths: concept segmentation vs visual segmentation. Shared docs/registry updates and similar naming are not enough to treat them as the same change, and they would not plausibly be merged as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 12, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43651-2", + "nodes": [ + "pull_request:43651", + "pull_request:43663" + ], + "soft_pairs": [ + "pull_request:43651|pull_request:43663" + ], + "prepared_review_unit_hash": "e3384a568fbb81b6b0f7327a98b63eb89495861b768e696e2a60b7b6abd621a2", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4293, + "estimated_input_tokens": 1074, + "estimated_eval_tokens": 2404 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4293, + "estimated_input_tokens": 1074, + "estimated_eval_tokens": 2404 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two open PRs in the same trainer file, but they address unrelated Trainer extensibility points: gradient-accumulation loss scaling vs. signature-column filtering. The soft similarity is file-level only, not a shared underlying bug/change.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "PR #43651 is the more concrete change: it fixes a specific training-step loss-scaling behavior and introduces a narrowly scoped hook. PR #43663 is also valid, but it targets a different Trainer concern.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43651", + "right": "pull_request:43663", + "accept": false, + "reason": "Different code paths and behaviors: #43651 changes gradient accumulation loss normalization, while #43663 changes how Trainer derives signature columns for input filtering. They are not the same bug or mergeable into one 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": 13, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43656-3", + "nodes": [ + "pull_request:43656", + "pull_request:43836", + "pull_request:43842" + ], + "soft_pairs": [ + "pull_request:43656|pull_request:43842", + "pull_request:43836|pull_request:43842" + ], + "prepared_review_unit_hash": "466e6bdca13e43bf071fdca69e92afa07d3be700cbbd2cdb5a0544abdf7fc0d3", + "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": "PR #43842 is the strongest canonical choice: it makes the most targeted fix for the `TypeAdapter` NameError in `serve.py`. PRs #43656 and #43836 are related in the same area, but the packet does not conservatively establish that they are the same underlying fix, so they should be treated as possibly overlapping rather than definitively duplicate.", + "confidence": 0.87, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43842 is the cleanest minimal fix: it directly addresses the `TypeAdapter` runtime/import issue with a small, explicit change in `src/transformers/cli/serve.py` and avoids the broader refactoring present in the other PRs.", + "best_issue_reason": null, + "best_pr_reason": "PR #43842 is the best representative of the cluster because it most narrowly and directly fixes the concrete `TypeAdapter` NameError case.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43656", + "right": "pull_request:43842", + "accept": false, + "reason": "The packet shows only shared filename and broad thematic similarity; the different explicit issue target and limited diff evidence do not establish the same concrete failure mode." + }, + { + "left": "pull_request:43836", + "right": "pull_request:43842", + "accept": false, + "reason": "Although both target the same issue, #43836 appears to be a much broader `serve.py` rewrite, so the evidence is insufficient to say it is the same specific bugfix or that the PRs could cleanly merge as one change." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The main conclusion that PR #43842 is the cleanest/minimal fix is grounded. However, the soft-edge verdicts are too strong. For 43656 vs 43842, the packet only shows soft filename similarity plus a different explicit issue target; it does not establish that both PRs fix the same TypeAdapter failure mode. For 43836 vs 43842, they do share the same issue target, but the preview shows 43836 is a much broader serve.py rewrite, so calling it simply another variant of the same bugfix is not sufficiently conservative. The summary should frame these as related or possibly overlapping, not definitively the same underlying fix." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 14, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "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": 15, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 2, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43996-28", + "nodes": [ + "pull_request:43996", + "pull_request:43997", + "pull_request:44007", + "pull_request:44010", + "pull_request:44013", + "pull_request:44017", + "pull_request:44018", + "pull_request:44019", + "pull_request:44024", + "pull_request:44044", + "pull_request:44056", + "pull_request:44059", + "pull_request:44066", + "pull_request:44068", + "pull_request:44071", + "pull_request:44072", + "pull_request:44073", + "pull_request:44074", + "pull_request:44076", + "pull_request:44085", + "pull_request:44086", + "pull_request:44098", + "pull_request:44101", + "pull_request:44116", + "pull_request:44129", + "pull_request:44154", + "pull_request:44161", + "pull_request:44722" + ], + "soft_pairs": [ + "pull_request:43996|pull_request:44085", + "pull_request:43996|pull_request:44044", + "pull_request:44066|pull_request:44085", + "pull_request:44007|pull_request:44072", + "pull_request:44072|pull_request:44722", + "pull_request:44066|pull_request:44072", + "pull_request:44013|pull_request:44044", + "pull_request:44066|pull_request:44086", + "pull_request:44066|pull_request:44071", + "pull_request:44018|pull_request:44068", + "pull_request:44066|pull_request:44068", + "pull_request:44018|pull_request:44066", + "pull_request:44019|pull_request:44722", + "pull_request:44086|pull_request:44722", + "pull_request:44019|pull_request:44085", + "pull_request:44018|pull_request:44019", + "pull_request:44019|pull_request:44071", + "pull_request:44068|pull_request:44116", + "pull_request:44068|pull_request:44722", + "pull_request:44071|pull_request:44722", + "pull_request:44018|pull_request:44116", + "pull_request:44018|pull_request:44086", + "pull_request:44018|pull_request:44085", + "pull_request:44068|pull_request:44086", + "pull_request:44071|pull_request:44085", + "pull_request:44059|pull_request:44086", + "pull_request:44085|pull_request:44086", + "pull_request:44059|pull_request:44722", + "pull_request:44098|pull_request:44101", + "pull_request:44018|pull_request:44071", + "pull_request:44017|pull_request:44722", + "pull_request:44017|pull_request:44161", + "pull_request:44017|pull_request:44073", + "pull_request:44068|pull_request:44071", + "pull_request:44071|pull_request:44086", + "pull_request:43997|pull_request:44073", + "pull_request:44017|pull_request:44071", + "pull_request:44059|pull_request:44071", + "pull_request:44073|pull_request:44101", + "pull_request:44059|pull_request:44068", + "pull_request:44018|pull_request:44059", + "pull_request:44056|pull_request:44101", + "pull_request:44073|pull_request:44161", + "pull_request:44101|pull_request:44161", + "pull_request: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:43997|pull_request:44071", + "pull_request:44010|pull_request:44071", + "pull_request:44010|pull_request:44073", + "pull_request:44018|pull_request:44129", + "pull_request:43997|pull_request:44074", + "pull_request:43997|pull_request:44129", + "pull_request:44010|pull_request:44018", + "pull_request:44071|pull_request:44129", + "pull_request:43996|pull_request:44161", + "pull_request:43996|pull_request:44017", + "pull_request:44010|pull_request:44044", + "pull_request:44044|pull_request:44074", + "pull_request:44066|pull_request:44722", + "pull_request:44085|pull_request:44722", + "pull_request:44085|pull_request:44129", + "pull_request:43996|pull_request:44007", + "pull_request:43996|pull_request:44129", + "pull_request:44044|pull_request:44129", + "pull_request:44044|pull_request:44154", + "pull_request:44024|pull_request:44076" + ], + "prepared_review_unit_hash": "709cb79e419cbe2e4b361692350a70867de96495b094ce811b93d9ea4b8ea596", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 77, + "serialized_chars": 84502, + "estimated_input_tokens": 21126, + "estimated_eval_tokens": 42508 + }, + "budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 72, + "serialized_chars": 76936, + "estimated_input_tokens": 19234, + "estimated_eval_tokens": 38724 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a set of model-specific output-tracing refactors toward `capture_outputs`/`can_return_tuple`, not one shared bug. The only clear duplicate PR pairs are the GPT-J/CodeGen refactor set and the GPT-Neo refactor set; the rest are separate ports to different model codepaths.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44066 is the best anchor: it is explicitly issue-linked, touches the core GPT-J/CodeGen output-tracing refactor, and is the clearest representative of the shared change that later GPT-J PRs repeat.", + "best_issue_reason": null, + "best_pr_reason": "PR 44066 is the most representative and broadly scoped standardized-output-tracing PR in the cluster; it captures the main concrete codepath change better than the more fragmented model-by-model follow-ups.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43996", + "right": "pull_request:44085", + "accept": false, + "reason": "Different model families and codepaths (CVT/FNet vs GPT-J); same umbrella refactor theme is not enough." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44044", + "accept": false, + "reason": "Different models and implementations (CVT/FNet vs DeBERTa); not the same underlying change." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44085", + "accept": true, + "reason": "Both target GPT-J output tracing on the same concrete block codepath; 44066 is the broader GPT-J/CodeGen refactor and 44085 is the overlapping GPT-J-only variant." + }, + { + "left": "pull_request:44007", + "right": "pull_request:44072", + "accept": false, + "reason": "Both are output-tracing refactors, but they touch unrelated model stacks (RegNet/ResNet/RT-DETR vs EfficientNet)." + }, + { + "left": "pull_request:44072", + "right": "pull_request:44722", + "accept": false, + "reason": "EfficientNet and GPT-J/CodeGen are unrelated codepaths; only the refactor pattern is similar." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44072", + "accept": false, + "reason": "Different model families and concrete implementations; no shared bug/change." + }, + { + "left": "pull_request:44013", + "right": "pull_request:44044", + "accept": false, + "reason": "MobileNetV2 and DeBERTa are unrelated codepaths; same tracking issue does not make them duplicates." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44086", + "accept": false, + "reason": "Different models (GPT-J/CodeGen vs MGP-STR); only the standardized-output-tracing theme is shared." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44071", + "accept": false, + "reason": "Different codepaths and model families (GPT-J/CodeGen vs MPT)." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44068", + "accept": true, + "reason": "Same GPT-Neo refactor, same files, and same output-tracing change; these are duplicate implementations of the same concrete fix." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44068", + "accept": false, + "reason": "GPT-J/CodeGen and GPT-Neo are separate model implementations; not the same PR change." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44066", + "accept": false, + "reason": "GPT-Neo vs GPT-J/CodeGen are distinct refactors on different codepaths." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44722", + "accept": false, + "reason": "ResNet and GPT-J/CodeGen are unrelated model codepaths." + }, + { + "left": "pull_request:44086", + "right": "pull_request:44722", + "accept": false, + "reason": "MGP-STR and GPT-J/CodeGen are different concrete implementations." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44085", + "accept": false, + "reason": "ResNet and GPT-J are unrelated; same output-tracing pattern is not enough." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44019", + "accept": false, + "reason": "GPT-Neo and ResNet are different model stacks with different fixes." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44071", + "accept": false, + "reason": "ResNet and MPT are unrelated codepaths." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo and Flaubert are different models; these are separate standardized-output-tracing ports." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-Neo and GPT-J/CodeGen are separate model refactors." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44722", + "accept": false, + "reason": "MPT and GPT-J/CodeGen are unrelated concrete codepaths." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-2 and MGP-STR are unrelated model implementations." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J and MGP-STR are different codepaths; only the refactor style matches." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-2 and GPT-J/CodeGen are separate output-tracing refactors." + }, + { + "left": "pull_request:44098", + "right": "pull_request:44101", + "accept": false, + "reason": "ViLT and Flaubert are unrelated model codepaths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-Neo and MPT are different models with separate changes." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44722", + "accept": false, + "reason": "Segformer and GPT-J/CodeGen are unrelated concrete implementations." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44161", + "accept": false, + "reason": "Segformer and LongT5 are different model families; not a duplicate change." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44073", + "accept": false, + "reason": "Segformer and VisualBert are separate model codepaths." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-Neo and MPT are unrelated implementations." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44086", + "accept": false, + "reason": "MPT and MGP-STR are different codepaths." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44073", + "accept": false, + "reason": "RegNet and VisualBert are unrelated model families." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44017", + "accept": false, + "reason": "RegNet and Segformer are separate model implementations." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44071", + "accept": false, + "reason": "SqueezeBert and MPT are unrelated codepaths." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44073", + "accept": false, + "reason": "SqueezeBert and VisualBert are different model families." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44129", + "accept": false, + "reason": "GPT-Neo and SpeechT5 are separate concrete refactors." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44101", + "accept": false, + "reason": "TextNet and Flaubert are unrelated implementations." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44161", + "accept": false, + "reason": "MPNet and LongT5 are different codepaths." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44018", + "accept": false, + "reason": "RegNet and GPT-Neo are unrelated model refactors." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44161", + "accept": false, + "reason": "RegNet and LongT5 are distinct model implementations." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44129", + "accept": false, + "reason": "MPT and SpeechT5 are unrelated concrete codepaths." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44161", + "accept": false, + "reason": "CVT/FNet and LongT5 are different models; not the same change." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44017", + "accept": false, + "reason": "CVT/FNet and Segformer are unrelated implementations." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44044", + "accept": false, + "reason": "SqueezeBert and DeBERTa are separate model codepaths." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44074", + "accept": false, + "reason": "DeBERTa and TextNet are unrelated models." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44722", + "accept": true, + "reason": "Same GPT-J/CodeGen output-tracing refactor; 44722 is a near-duplicate follow-up on the same concrete change." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44722", + "accept": true, + "reason": "Both are GPT-J output-tracing refactors on the same underlying codepath and could plausibly be merged into one PR." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44129", + "accept": false, + "reason": "GPT-J and SpeechT5 are unrelated model implementations." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44007", + "accept": false, + "reason": "CVT/FNet and RegNet/ResNet/RT-DETR are different model families." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44129", + "accept": false, + "reason": "CVT/FNet and SpeechT5 are unrelated codepaths." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44129", + "accept": false, + "reason": "DeBERTa and SpeechT5 are separate model implementations." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44154", + "accept": false, + "reason": "DeBERTa and ViT-S are unrelated model codepaths." + }, + { + "left": "pull_request:44024", + "right": "pull_request:44076", + "accept": false, + "reason": "FocalNet and ImageGPT are different model families; same output-tracing theme does not imply duplication." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet and stays conservative about the soft edges. The accepted pairs are all within the same concrete model/codepath families (GPT-Neo and GPT-J/CodeGen), and the rejections appropriately avoid treating shared output-tracing style or shared issue-targets as duplicates across unrelated models. No clear overstatement of evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 15, + "component_count": 23, + "review_unit_index": 2, + "review_unit_count": 2, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43996-28", + "nodes": [ + "pull_request:43997", + "pull_request:44010", + "pull_request:44056", + "pull_request:44073", + "pull_request:44074", + "pull_request:44129" + ], + "soft_pairs": [ + "pull_request:44073|pull_request:44074", + "pull_request:44056|pull_request:44073", + "pull_request:44056|pull_request:44074", + "pull_request:43997|pull_request:44010", + "pull_request:44010|pull_request:44129" + ], + "prepared_review_unit_hash": "126c7d405f80fdc7403b5cdc7097d4c584642317aed1ef568699a1d5cc227e4e", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 77, + "serialized_chars": 84502, + "estimated_input_tokens": 21126, + "estimated_eval_tokens": 42508 + }, + "budget": { + "node_count": 6, + "item_count": 6, + "soft_pair_count": 5, + "serialized_chars": 13066, + "estimated_input_tokens": 3267, + "estimated_eval_tokens": 6790 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These PRs are all model-specific refactors for standardized output tracing/capture, but they touch different architectures and code paths, so they should not be deduplicated as the same change.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #44129 is the strongest representative: it has an explicit issue target and a concrete, self-contained SpeechT5 tracing refactor. That said, it is not a duplicate of the others.", + "best_issue_reason": null, + "best_pr_reason": "PR #44129 is the best single PR to keep as the representative example of this refactor pattern because it is well-scoped and linked to an issue. The others are similarly scoped but model-specific.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44073", + "right": "pull_request:44074", + "accept": false, + "reason": "Both are output-tracing refactors, but one is for VisualBert and the other for TextNet. Different models, different files, and different implementations; not the same underlying change." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44073", + "accept": false, + "reason": "MPNet vs VisualBert. Same broad refactor theme, but they modify separate model code paths and are independent changes." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44074", + "accept": false, + "reason": "MPNet and TextNet are distinct model implementations with unrelated refactors despite similar decorator/capture naming." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44010", + "accept": false, + "reason": "RegNet and SqueezeBert are unrelated model-specific output tracing updates; they do not appear mergeable into one PR for the same bug or change." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44129", + "accept": false, + "reason": "SqueezeBert vs SpeechT5, different model internals and different output-tracing logic. Similar theme only at a high level, not the same 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": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43998-13", + "nodes": [ + "pull_request:43998", + "pull_request:43999", + "pull_request:44000", + "pull_request:44001", + "pull_request:44002", + "pull_request:44003", + "pull_request:44004", + "pull_request:44025", + "pull_request:44026", + "pull_request:44027", + "pull_request:44028", + "pull_request:44029", + "pull_request:44030" + ], + "soft_pairs": [ + "pull_request:43998|pull_request:44002", + "pull_request:43998|pull_request:44000", + "pull_request:43998|pull_request:44001", + "pull_request:43998|pull_request:44003", + "pull_request:43998|pull_request:44029", + "pull_request:43999|pull_request:44001", + "pull_request:43999|pull_request:44002", + "pull_request:43999|pull_request:44025", + "pull_request:43999|pull_request:44026", + "pull_request:43999|pull_request:44027", + "pull_request:44000|pull_request:44001", + "pull_request:44000|pull_request:44002", + "pull_request:44000|pull_request:44003", + "pull_request:44000|pull_request:44025", + "pull_request:44001|pull_request:44002", + "pull_request:44001|pull_request:44003", + "pull_request:44001|pull_request:44025", + "pull_request:44002|pull_request:44029", + "pull_request:44002|pull_request:44003", + "pull_request:44003|pull_request: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": "636272954d30916491a475487422f1a8c03c8b5eaf656b8dca64a8cd3dbbe155", + "component_budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 40, + "serialized_chars": 40173, + "estimated_input_tokens": 10044, + "estimated_eval_tokens": 20344 + }, + "budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 40, + "serialized_chars": 40173, + "estimated_input_tokens": 10044, + "estimated_eval_tokens": 20344 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is a set of separate, model-specific output-tracing refactors. They are thematically similar, but they do not describe the same concrete bug or change, so they should not be merged as duplicates.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44025 is the best representative of the theme because it is a clean, self-contained output-tracing refactor and spans a closely related pair of files. That said, the cluster is not duplicate-equivalent overall; the PRs target different model implementations.", + "best_issue_reason": null, + "best_pr_reason": "PR 44025 is the most representative hub for the cluster, but only as a thematic example rather than a duplicate canonical. It is not evidence that the other PRs are the same change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43998", + "right": "pull_request:44002", + "accept": false, + "reason": "Same refactor pattern, but timm_backbone and upernet are different model code paths." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44000", + "accept": false, + "reason": "Both are output-tracing refactors, but they touch unrelated model implementations." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44001", + "accept": false, + "reason": "Different architectures and forward paths; not the same underlying change." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44003", + "accept": false, + "reason": "timm_backbone is unrelated to mamba/falcon_mamba; same theme, not same code path." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44029", + "accept": false, + "reason": "Unrelated model families; these are not mergeable as one concrete fix." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44001", + "accept": false, + "reason": "Different model implementations; only the refactor pattern matches." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44002", + "accept": false, + "reason": "MobileNetV1 and UperNet are separate model-specific changes." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44025", + "accept": false, + "reason": "Different model families and forward logic; not a duplicate." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44026", + "accept": false, + "reason": "MobileNetV1 vs vision-encoder-decoder wrappers are distinct code paths." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44027", + "accept": false, + "reason": "Separate model wrappers; same refactor style is not enough to merge." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44001", + "accept": false, + "reason": "Vision-text dual encoder and UnivNet are unrelated implementations." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44002", + "accept": false, + "reason": "Different model subsystems; not the same bug or change." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44003", + "accept": false, + "reason": "Vision-text dual encoder and mamba are unrelated model paths." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44025", + "accept": false, + "reason": "These are separate model-specific refactors, not one concrete fix." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44002", + "accept": false, + "reason": "UnivNet and UperNet are different models with distinct forward paths." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44003", + "accept": false, + "reason": "Different architectures; the similarity is only at the refactor level." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44025", + "accept": false, + "reason": "Separate model families; not a duplicate change." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44003", + "accept": false, + "reason": "UperNet and mamba/falcon_mamba are unrelated code paths." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44029", + "accept": false, + "reason": "Different models; same output-tracing theme does not imply duplication." + }, + { + "left": "pull_request:44003", + "right": "pull_request:44028", + "accept": false, + "reason": "Mamba/falcon_mamba and SuperPoint are unrelated implementations." + }, + { + "left": "pull_request:44003", + "right": "pull_request:44004", + "accept": false, + "reason": "Mamba/falcon_mamba changes are not the same change as CodeGen output tracing." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44004", + "accept": false, + "reason": "timm_backbone and CodeGen are different model code paths." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44004", + "accept": false, + "reason": "UnivNet and CodeGen are unrelated; no shared concrete bug." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44004", + "accept": false, + "reason": "UperNet and CodeGen are distinct implementations with different fixes." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44026", + "accept": false, + "reason": "DepthAnything and vision-encoder-decoder are different model paths; same refactor pattern only." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44027", + "accept": false, + "reason": "DepthAnything and SpeechEncoderDecoder are unrelated concrete changes." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44028", + "accept": false, + "reason": "DepthAnything and SuperPoint do not fix the same underlying issue." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44029", + "accept": false, + "reason": "DepthAnything and RWKV are unrelated code paths." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44030", + "accept": false, + "reason": "DepthAnything and DPR are different model implementations." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44029", + "accept": false, + "reason": "Vision-encoder-decoder and RWKV are unrelated; not the same fix." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44027", + "accept": false, + "reason": "Both are encoder-decoder wrappers, but they are separate model-specific forward methods and not one concrete bug." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44028", + "accept": false, + "reason": "Vision-encoder-decoder and SuperPoint are different code paths." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44030", + "accept": false, + "reason": "Encoder-decoder wrapper refactor vs DPR wrapper refactor are not the same change." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44028", + "accept": false, + "reason": "SpeechEncoderDecoder and SuperPoint are unrelated implementations." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44029", + "accept": false, + "reason": "SpeechEncoderDecoder and RWKV are unrelated model code paths." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44030", + "accept": false, + "reason": "SpeechEncoderDecoder and DPR do not share the same concrete fix." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44029", + "accept": false, + "reason": "SuperPoint and RWKV are distinct model-specific changes." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44030", + "accept": false, + "reason": "SuperPoint and DPR are unrelated implementations." + }, + { + "left": "pull_request:44029", + "right": "pull_request:44030", + "accept": false, + "reason": "RWKV and DPR are different models with different forward logic." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 17, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-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": "29283d69ec866496e4c2137c385e87156608f0b8d667abd94b444c4d1d5f92c3", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 11073, + "estimated_input_tokens": 2769, + "estimated_eval_tokens": 5794 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 11073, + "estimated_input_tokens": 2769, + "estimated_eval_tokens": 5794 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are three separate model-addition PRs for different models (PenguinVL, DeepSeek-OCR-2, and Voxtral TTS). The overlap is mostly shared Transformers boilerplate for new model integration, not the same underlying change.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: although the PRs share new-model scaffolding, they implement different models and are not duplicates.", + "best_issue_reason": null, + "best_pr_reason": "No single PR is a best duplicate representative here; each PR targets a different model and could not be merged into one change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44662", + "right": "pull_request:45401", + "accept": false, + "reason": "Different models and codepaths: PenguinVL vs Voxtral TTS. Shared auto-registration/docs files are generic new-model plumbing, not the same fix." + }, + { + "left": "pull_request:44662", + "right": "pull_request:45075", + "accept": false, + "reason": "Different model implementations: PenguinVL and DeepSeek-OCR-2. The overlap is only shared framework registration and docs, so they are not duplicates." + }, + { + "left": "pull_request:45075", + "right": "pull_request:45401", + "accept": false, + "reason": "Different model additions for DeepSeek-OCR-2 and Voxtral TTS. Shared Transformers scaffolding does not indicate the same underlying bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 18, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44979-2", + "nodes": [ + "pull_request:44979", + "pull_request:45363" + ], + "soft_pairs": [ + "pull_request:44979|pull_request:45363" + ], + "prepared_review_unit_hash": "5c60281b491d0b5f0ac2a5454cc92f029ee68617c43e1f78eb08002e3a18f362", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4407, + "estimated_input_tokens": 1102, + "estimated_eval_tokens": 2460 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4407, + "estimated_input_tokens": 1102, + "estimated_eval_tokens": 2460 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related fusion-feature PRs, but they target different layers and are not duplicates: one adds a standalone Module Fusion API, while the other wires n-to-1 kernel fusion through KernelConfig and model loading internals.", + "confidence": 0.87, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "PR 45363 is the broader, more end-to-end change for kernel fusion, touching configuration, integrations, and model loading; it is the better representative of the cluster's main theme.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44979", + "right": "pull_request:45363", + "accept": false, + "reason": "Different concrete changes and code paths: 44979 introduces a new Module Fusion API module and tests, while 45363 adds KernelConfig-based n-to-1 kernel fusion plumbing, integration hooks, and an example. Related topic, but not the same mergeable change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 19, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45086-2", + "nodes": [ + "pull_request:45086", + "pull_request:45317" + ], + "soft_pairs": [ + "pull_request:45086|pull_request:45317" + ], + "prepared_review_unit_hash": "220462e86df87d807814f14bebe628ba16f9a3c29fb55dc25ad5cb568d840aa5", + "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 fix the same Mistral regex patching AttributeError in the same file and target the same issue; 45317 looks like the fuller follow-up with tests.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45317 is the better canonical PR because it covers the same fix as 45086 but is more complete and includes test coverage.", + "best_issue_reason": null, + "best_pr_reason": "PR 45317 is the best representative of the cluster: same concrete bug fix, clearer title, and additional tests make it the strongest standalone artifact.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45086", + "right": "pull_request:45317", + "accept": true, + "reason": "Same underlying AttributeError fix in _patch_mistral_regex, same target issue, and the diff shows the same code-path correction; 45317 is essentially the more complete version of 45086." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 20, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45115-2", + "nodes": [ + "pull_request:45115", + "pull_request:45569" + ], + "soft_pairs": [ + "pull_request:45115|pull_request:45569" + ], + "prepared_review_unit_hash": "150f3d80335a91aecbd7cabd6fe0d8c8114de376ef811205a0309fa3248fd650", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6066, + "estimated_input_tokens": 1517, + "estimated_eval_tokens": 3290 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6066, + "estimated_input_tokens": 1517, + "estimated_eval_tokens": 3290 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two open PRs both touch Nemotron H internals, but they target different changes: one refactors NemotronH to inherit GraniteMoeHybrid and add conversion mapping, while the other splits Nemotron H into dense/sparse model variants and updates docs/tests. They are related but not the same fix/change.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "PR #45569 is the better cluster representative because it covers the larger, end-to-end Nemotron H architecture change, including model split, auto-mapping, docs, and tests. PR #45115 is a narrower refactor focused on inheritance and weight conversion.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45115", + "right": "pull_request:45569", + "accept": false, + "reason": "Shared Nemotron H files alone are not enough: #45115 is a structural refactor/conversion-mapping PR, while #45569 introduces separate dense/sparse model classes plus docs/tests. They do not look like the same concrete change and would not plausibly be merged as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 21, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45332-2", + "nodes": [ + "pull_request:45332", + "pull_request:45333" + ], + "soft_pairs": [ + "pull_request:45332|pull_request:45333" + ], + "prepared_review_unit_hash": "a924295be58fe005afd8cdcea24072d977f42e21d22780879e9c84a8c45df93f", + "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 but distinct PRs: one adds per-layer heterogeneous config plumbing, while the other adds broader heterogeneous model/cache support built on top of that config work. They look complementary rather than duplicate.", + "confidence": 0.82, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45332 is the broader, more central change: it implements heterogeneous model support across cache/modeling code and includes the config hooks it needs.", + "best_issue_reason": null, + "best_pr_reason": "PR #45332 is the best representative of the cluster because it spans the full code-path impact and is the more complete end-to-end implementation.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45332", + "right": "pull_request:45333", + "accept": false, + "reason": "They share heterogeneity plumbing, but #45333 is config-only while #45332 extends model/cache behavior; they are related groundwork and follow-on work, not the same concrete change that should be merged as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 22, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45394-5", + "nodes": [ + "pull_request:45394", + "pull_request:45487", + "pull_request:45592", + "pull_request:45596", + "pull_request:45615" + ], + "soft_pairs": [ + "pull_request:45394|pull_request:45596", + "pull_request:45487|pull_request:45592", + "pull_request:45487|pull_request:45596", + "pull_request:45487|pull_request:45615", + "pull_request:45592|pull_request:45596", + "pull_request:45592|pull_request:45615", + "pull_request:45596|pull_request:45615" + ], + "prepared_review_unit_hash": "40f8da80ebff371273d5c3387a02dc195c9aa0368513048dff465def359654eb", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 7, + "serialized_chars": 13274, + "estimated_input_tokens": 3319, + "estimated_eval_tokens": 6894 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 7, + "serialized_chars": 13274, + "estimated_input_tokens": 3319, + "estimated_eval_tokens": 6894 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are not true duplicates overall: one substantive model-parallel/code-path fix (45487), one XCLIP test/model-parallel fix (45394), one FastVLM test-padding fix (45592), and two separate XPU expectation updates for unrelated models (45596, 45615).", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45487 is the strongest representative fix: it changes actual model-parallel behavior in AltCLIP/ChineseCLIP (device placement for buffered token type IDs and split-module lists), whereas the others are narrower model-specific test updates or unrelated fixes.", + "best_issue_reason": null, + "best_pr_reason": "PR 45487 is the best standalone PR in the set because it addresses a concrete runtime bug in shared code paths and is more substantive than the test-only expectation changes.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45394", + "right": "pull_request:45596", + "accept": false, + "reason": "Different models and different problems: XCLIP split-module/test failures vs BLT XPU expectation updates." + }, + { + "left": "pull_request:45487", + "right": "pull_request:45592", + "accept": false, + "reason": "Both mention device/test failures, but 45487 is an AltCLIP/ChineseCLIP model-parallel code fix while 45592 is only a FastVLM test padding-side adjustment." + }, + { + "left": "pull_request:45487", + "right": "pull_request:45596", + "accept": false, + "reason": "Different model families and change types; 45487 fixes AltCLIP/ChineseCLIP runtime behavior, 45596 only updates BLT XPU test expectations." + }, + { + "left": "pull_request:45487", + "right": "pull_request:45615", + "accept": false, + "reason": "Not the same bug: 45487 is a model code fix, while 45615 only adds XPU expectations for Qianfan OCR tests." + }, + { + "left": "pull_request:45592", + "right": "pull_request:45596", + "accept": false, + "reason": "Both are test-related, but they target different models and distinct failures; no shared concrete code-path." + }, + { + "left": "pull_request:45592", + "right": "pull_request:45615", + "accept": false, + "reason": "Different test suites and outputs; these are separate device-specific expectations, not one underlying change." + }, + { + "left": "pull_request:45596", + "right": "pull_request:45615", + "accept": false, + "reason": "Both add XPU expectations, but for unrelated models and outputs, so they are 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": "pull_request", + "component_index": 23, + "component_count": 23, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45415-2", + "nodes": [ + "pull_request:45415", + "pull_request:45425" + ], + "soft_pairs": [ + "pull_request:45415|pull_request:45425" + ], + "prepared_review_unit_hash": "8b4294fa11631aa382e969d55cd3cce9d4c25c2c9eecc76d5582adbf0c0f7cd4", + "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 address different scopes: one is a broad type-checking sweep across many Transformers modules, while the other adds typing helper protocols and a narrower modeling_utils-related cleanup. They share some files, but not the same concrete change.", + "confidence": 0.88, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45425 is the better canonical representative because it is a focused typing infrastructure change with a coherent purpose, whereas PR #45415 is a much broader multi-file type-checking sweep with mixed, unrelated edits.", + "best_issue_reason": null, + "best_pr_reason": "PR #45425 is the best overall fit for the cluster because it has the clearest single change theme and the least heterogeneous scope.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45415", + "right": "pull_request:45425", + "accept": false, + "reason": "Reject: the overlap is only in typing-related files and shared cleanup areas, but the PRs do not appear to fix the same concrete code-path problem or be safely mergeable into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + } + ] +} diff --git a/snapshots/20260425T060019Z/analysis-runs/hybrid-model-20260425t060019z/manifest.json b/snapshots/20260425T060019Z/analysis-runs/hybrid-model-20260425t060019z/manifest.json new file mode 100644 index 0000000000000000000000000000000000000000..7e6ab0d3a8695ad5e0540921e1b1200ac31c2d76 --- /dev/null +++ b/snapshots/20260425T060019Z/analysis-runs/hybrid-model-20260425t060019z/manifest.json @@ -0,0 +1,14 @@ +{ + "analysis_id": "hybrid-model-20260425t060019z", + "artifacts": { + "hybrid": "snapshots/20260425T060019Z/analysis-runs/hybrid-model-20260425t060019z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260425T060019Z/analysis-runs/hybrid-model-20260425t060019z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "model": null, + "published_at": "2026-04-25T06:10:12Z", + "repo": "huggingface/transformers", + "schema_version": 1, + "snapshot_id": "20260425T060019Z", + "variant": "hybrid" +} diff --git a/snapshots/20260425T060019Z/manifest.json b/snapshots/20260425T060019Z/manifest.json index c4974c13715132956a08987f2336417749498a21..b4ebb21f8be1867c8dcc7b34117eec654e8e67cc 100644 --- a/snapshots/20260425T060019Z/manifest.json +++ b/snapshots/20260425T060019Z/manifest.json @@ -32,6 +32,24 @@ "timeline_events": 233 }, "extracted_at": "2026-04-25T06:00:19Z", + "published_analysis": { + "canonical_analysis_id": "hybrid-model-20260425t060019z", + "runs": { + "hybrid-model-20260425t060019z": { + "analysis_id": "hybrid-model-20260425t060019z", + "artifacts": { + "hybrid": "snapshots/20260425T060019Z/analysis-runs/hybrid-model-20260425t060019z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260425T060019Z/analysis-runs/hybrid-model-20260425t060019z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "manifest_path": "snapshots/20260425T060019Z/analysis-runs/hybrid-model-20260425t060019z/manifest.json", + "model": null, + "published_at": "2026-04-25T06:10:12Z", + "variant": "hybrid" + } + }, + "schema_version": 1 + }, "repo": "huggingface/transformers", "snapshot_id": "20260425T060019Z", "watermark": {