| commit |
1.00 |
Skip caching_allocator_warmup on Neuron (no reuse pool to wa |
|
Commit message contains explicit AI assi |
2026-07-08 |
| commit |
1.00 |
Fix MoE decode optimization for experts living in a submodel |
|
Commit message contains explicit AI assi |
2026-07-07 |
| commit |
1.00 |
processor tests: use tiny Hub repos to reduce CI memory (#47 |
|
Commit message contains explicit AI assi |
2026-07-07 |
| commit |
1.00 |
docs-fix: IMDb dataset link in sequence classification guide |
|
Commit message contains explicit AI assi |
2026-07-06 |
| commit |
1.00 |
Add new model: Kimi2-6 (#45630) |
|
Commit message contains explicit AI assi |
2026-07-03 |
| commit |
1.00 |
Add tiny_model_id support to ProcessorTesterMixin for memory |
|
Commit message contains explicit AI assi |
2026-07-03 |
| pr |
1.00 |
fix: resolve device mismatch in Bark semantic model by offlo |
|
PR body explicitly mentions AI collabora |
2026-07-08 |
| pr |
1.00 |
Fix FP8 dequantization for frozen compressed-tensors models |
|
PR body explicitly mentions AI collabora |
2026-07-08 |
| pr |
1.00 |
Optimize flash attention max seqlen computation in vision at |
|
PR body explicitly mentions AI collabora |
2026-07-08 |
| pr |
1.00 |
Add Unlimited OCR |
|
PR body explicitly mentions AI collabora |
2026-06-23 |
| pr |
1.00 |
Add exporters min version requirements and test skip |
|
PR body explicitly mentions AI collabora |
2026-07-08 |
| pr |
1.00 |
[Mistral] Move MistralConverter into integrations/mistral/ p |
|
PR body explicitly mentions AI collabora |
2026-06-12 |
| pr |
1.00 |
Better deepgemm errors |
|
PR body explicitly mentions AI collabora |
2026-07-08 |
| pr |
1.00 |
Fix AqlmConfig error messages to say "int" instead of "float |
|
PR body explicitly mentions AI collabora |
2026-07-05 |
| pr |
1.00 |
Clarify input device placement in the Quicktour inference ex |
|
PR body explicitly mentions AI collabora |
2026-07-07 |
| pr |
1.00 |
nCompass triton-moe experts integration |
|
PR body explicitly mentions AI collabora |
2026-07-08 |
| pr |
1.00 |
Clarify that max_steps is required for datasets without __le |
|
PR body explicitly mentions AI collabora |
2026-07-08 |
| pr |
1.00 |
Improve Trainer DataLoader Controls for Streaming and Multip |
|
PR body explicitly mentions AI collabora |
2026-07-08 |
| pr |
1.00 |
[generation] Fix misleading synced_gpus warning in continuou |
|
PR body explicitly mentions AI collabora |
2026-07-08 |
| pr |
1.00 |
Fix MoE decode optimization for experts living in a submodel |
|
PR body explicitly mentions AI collabora |
2026-07-06 |
| pr |
1.00 |
Adapt voxtral_realtime to mistral-common 1.11.5 audio API |
|
PR body explicitly mentions AI collabora |
2026-06-26 |
| commit |
0.20 |
Bump min peft 0.19.1 remove weight conversion duplicate code |
|
Casual tone, internal context, and typic |
2026-07-01 |
| pr |
0.20 |
add_axk1 |
|
PR body uses formal explanation but incl |
2026-06-24 |
| commit |
0.10 |
Add Xiaomi MiMo-V2 (#45144) |
|
Informal commit messages with domain-spe |
2026-07-01 |
| commit |
0.10 |
[new model] Add Zyphra/ZAYA1-8B (#45862) |
|
Commit-style lines, domain terms, and mi |
2026-07-01 |
| commit |
0.10 |
Fix accelerate hooks for all models using linear attention ( |
|
Informal, terse commit list with domain |
2026-07-01 |
| commit |
0.10 |
Add support for RADIO models (#46425) |
|
Terse commits, technical details, no AI |
2026-07-01 |
| pr |
0.10 |
tests: reduce processor test memory usage and use tiny test |
|
Contains brief summary, but informal and |
2026-07-08 |
| pr |
0.10 |
Fix crash in greedy assisted generation with different token |
|
Slightly formal grammar in description, |
2026-06-27 |
| pr |
0.10 |
Raise a clear error for empty token lists in bad_words_ids / |
|
Slightly more textual detail, but tone a |
2026-07-01 |
| pr |
0.10 |
processor tests: use tiny Hub repos to reduce CI memory by ~ |
|
Summary is domain-specific and technical |
2026-07-06 |
| commit |
0.00 |
🌐 [i18n-KO] Translate accelerator_selection.md to Korean (#4 |
|
Standard i18n commit, contains co-author |
2026-07-08 |
| commit |
0.00 |
[i18n-KO] Translate optimum.md to Korean and fix Furiosa typ |
|
Brief translation commit, no AI signals |
2026-07-08 |
| commit |
0.00 |
[docs] fix curly quotes rendering to straight quotes (#47135 |
|
Typical terse docs commit message, fully |
2026-07-08 |
| commit |
0.00 |
Fix custom code which doesn't know about the new linear laye |
|
Standard fix message with Git trailer, n |
2026-07-08 |
| commit |
0.00 |
Reject path traversal in the `transformers_weights` config f |
|
Technical description with domain detail |
2026-07-08 |
| commit |
0.00 |
[docs] Custom code conversion mapping (#47114) |
|
Minimal free-text, standard docs update |
2026-07-08 |
| commit |
0.00 |
Add exporters min version requirements and test skip (#47161 |
|
Human-like terse project maintenance com |
2026-07-08 |
| commit |
0.00 |
tests: reduce processor test memory usage and use tiny test |
|
Terse and repetitive test commit with hu |
2026-07-08 |
| commit |
0.00 |
Clarify input device placement in the Quicktour inference ex |
|
Short human-style commit, no AI phrasing |
2026-07-08 |
| commit |
0.00 |
Extend continuous batching memory prediction test to XPU (#4 |
|
Technical, domain-specific commit, signe |
2026-07-08 |
| commit |
0.00 |
Fix case where `_LazyAutoMapping.register` is passed a `str` |
|
Technical commit message, concise and do |
2026-07-08 |
| commit |
0.00 |
[docs] MoE decode switching (#47149) |
|
Terse commit with domain reference, huma |
2026-07-08 |
| commit |
0.00 |
add XPU output expectations for minicpm3 tests (#47092) |
|
Short, specific, uses abbreviations, typ |
2026-07-08 |
| commit |
0.00 |
Diffusion gemma: fix failed test cases (#47025) |
|
Short, domain-relevant, no AI phrasing p |
2026-07-08 |
| commit |
0.00 |
add XPU Expectation for cosmos3_omni tests (#46880) |
|
Straightforward, terse, domain-specific |
2026-07-08 |
| commit |
0.00 |
Fix IndexError Bug in XLMRoberta/Camembert ForMultipleChoice |
|
Fix summary is direct, uses technical te |
2026-07-08 |
| commit |
0.00 |
Restrict QuantizedCache to only full attention (#47160) |
|
Extremely brief message; typical of huma |
2026-07-08 |
| commit |
0.00 |
[docs] continuous batching (offloading behavior, max batch t |
|
Concise conventional commit messages, hu |
2026-07-07 |
| commit |
0.00 |
[docs] fix autolinks (#46968) |
|
Brief, minimal text; typical human style |
2026-07-07 |
| commit |
0.00 |
Adapt voxtral_realtime to mistral-common 1.11.5 audio API (# |
|
Contains domain-specific changelog, no A |
2026-07-07 |
| commit |
0.00 |
revert #47121 (#47144) |
|
Standard revert message; no signs of AI |
2026-07-07 |
| commit |
0.00 |
Fix output labels for AudioFlamingo3 (and related) models (# |
|
Terse, informal commit chunks; human eng |
2026-07-07 |
| commit |
0.00 |
Fix false __len__ claims in Trainer docstrings (#47131) |
|
Brief, targeted doc corrections; distinc |
2026-07-07 |
| commit |
0.00 |
Fix: accept scalar audio formats in Voxtral transcription re |
|
Direct and concise, no signals of AI-gen |
2026-07-07 |
| commit |
0.00 |
[serge] Fix 12 integration tests for model `dac` failing wit |
|
— |
2026-07-07 |
| commit |
0.00 |
Fix CLI compatibility with huggingface_hub 1.22 (#47059) (#4 |
|
Detailed technical explanation, human-st |
2026-07-07 |
| commit |
0.00 |
Add TIPSv2 (#46347) |
|
Natural, informal commit fragments; clea |
2026-07-07 |
| commit |
0.00 |
we want to run the CI in the release branches (#47125) |
|
Informal tone; typical team communicatio |
2026-07-07 |
| commit |
0.00 |
Small improvement (#47128) |
|
Terse commit message with minimal conten |
2026-07-07 |
| commit |
0.00 |
[cache] Simplify cache dispatch based on layer_types (#47118 |
|
Succinct, iterative commit messages with |
2026-07-07 |
| commit |
0.00 |
[sdpa] Allow prefill to use FA kernel with StaticCache (#470 |
|
Terse change list and domain-specific co |
2026-07-07 |
| commit |
0.00 |
[Model] Support `use_cache=False` for DeepSeek V4 (#46965) |
|
Technical, to-the-point changelog with h |
2026-07-07 |
| commit |
0.00 |
Fix CANINE fp16 dtype mismatch in attention mask causing ONN |
|
Technical, concise, specific to code; hu |
2026-07-06 |
| commit |
0.00 |
Make doc builds faster (#47099) |
|
Standard changelog entries, informal ton |
2026-07-06 |
| commit |
0.00 |
Fix AltCLIP text embedding resize test (#47079) |
|
Minimal and terse commit messages; clear |
2026-07-06 |
| commit |
0.00 |
fix mask return-type contract regression and add correctness |
|
Domain-specific, informal, emails; stron |
2026-07-06 |
| commit |
0.00 |
Fix save_pretrained with offloading and weight conversions ( |
|
Terse, colloquial commit messages; defin |
2026-07-06 |
| commit |
0.00 |
Update dev (#47044) |
|
Extremely terse; characteristic of human |
2026-07-03 |
| commit |
0.00 |
[`Gemma4`] Update 1 integration test (#47042) |
|
Very short, direct commit message; human |
2026-07-03 |
| commit |
0.00 |
Update workflow callers to use `transformers-ci` (#47040) |
|
Brief, to-the-point; has a 'co-authored- |
2026-07-03 |
| commit |
0.00 |
Fix/qwen3 variants tp plan (#47041) |
|
Technical details and problem explanatio |
2026-07-03 |
| commit |
0.00 |
Add HunYuan VL model (#46417) |
|
Domain jargon, detailed code changes, an |
2026-07-03 |
| commit |
0.00 |
[Gemma4] Fix dtype casting for quantized vision/audio embedd |
|
Short, practical commit; typical human e |
2026-07-03 |
| commit |
0.00 |
chore(linter): add TRF018 modeling rule (#46259) |
|
Informal, stream-of-consciousness tone; |
2026-07-03 |
| commit |
0.00 |
Fix silent SDPA math-kernel fallback for GQA when key/value |
|
Technical content with human-style expla |
2026-07-02 |
| commit |
0.00 |
[PoC] HF exporters (#41992) |
|
Minimal commit messages and highly techn |
2026-07-02 |
| commit |
0.00 |
TST Skip PEFT tests if PEFT version is too low (#47027) |
|
Technical, concise phrasing, with minor |
2026-07-02 |
| commit |
0.00 |
CI Add PEFT integration tests (#47021) |
|
Brief and contextually relevant, follows |
2026-07-02 |
| commit |
0.00 |
[glm-mode-dsa] Indexer uses interleaved rope (#46842) |
|
Technical jargon and informal phrasing t |
2026-07-02 |
| commit |
0.00 |
Use standard arg names in Mllama (#46977) |
|
Single terse word, characteristic of hum |
2026-07-02 |
| commit |
0.00 |
[docs] update cache apis (#46892) |
|
Minimal, terse commit summary, no AI sig |
2026-07-01 |
| commit |
0.00 |
Raise a loud error for missing prefix (#46980) |
|
Concise commit subject/action pairs, no |
2026-07-01 |
| commit |
0.00 |
Fix typo in Qwen3 ASR no_split_module (#47002) |
|
Direct, minimal commit; no AI signals. |
2026-07-01 |
| commit |
0.00 |
only in the original repo (#46982) |
|
Terse, informal commit message; clearly |
2026-06-30 |
| commit |
0.00 |
🚨 Modeling changes for export, compile, and hybrid-attention |
|
Informal, bullet-style commit log; no AI |
2026-06-30 |
| commit |
0.00 |
Fix typos in Gemma 4 Assistant documentation (#46975) |
|
Standard phrasing for typo fixes; human |
2026-06-30 |
| commit |
0.00 |
🚨 [gemma 3/4] Fix bidirectional attention masking crossing s |
|
Commit log with technical detail and hum |
2026-06-30 |
| commit |
0.00 |
the CI status should be a comment (#46976) |
|
Short, informal summary and log; human t |
2026-06-30 |
| commit |
0.00 |
QwenVL model conversion (#46881) |
|
Terse, conversational commit log; clearl |
2026-06-30 |
| commit |
0.00 |
Remove default dtype in FusedRMSNormGated modules (#46953) |
|
Minimal, imperative phrasing; human cont |
2026-06-30 |
| commit |
0.00 |
[docs] EP plan validation (#46900) |
|
Sparse and informal commit/PR messages, |
2026-06-29 |
| commit |
0.00 |
Fix Xcodec2 attention to be non-causal. (#46963) |
|
Brief commit message with a typo, clearl |
2026-06-29 |
| commit |
0.00 |
Fix Mamba2 chunked-prefill / speculative decoding for Zamba2 |
|
Highly technical, informal, and detailed |
2026-06-29 |
| commit |
0.00 |
FIX PEFT test changed error type (#46959) |
|
Clear domain context and personal note, |
2026-06-29 |
| commit |
0.00 |
Fix path traversal via vocab-file arguments in tokenizer_con |
|
In-depth explanation with domain jargon |
2026-06-29 |
| commit |
0.00 |
docs(conditional_detr): fix num_queries default in docstring |
|
Standard docstring update message; conci |
2026-06-29 |
| commit |
0.00 |
Use common floats_list method for feature extractor tests. ( |
|
Generic, terse, and direct: common human |
2026-06-29 |
| commit |
0.00 |
Fix RT-DETR indexing error when num_feature_levels exceeds b |
|
Technical, with edit trail and terse sum |
2026-06-29 |
| commit |
0.00 |
Fix Florence2 training-loss double-shift (same pattern as Mo |
|
Multiple commits, technical fixes, and r |
2026-06-29 |
| commit |
0.00 |
Remove some unnecessary generate warnings (#46955) |
|
Terse commit summary and bullet, matches |
2026-06-29 |
| commit |
0.00 |
[Olmo3] different RoPE per layer type (#46911) |
|
Terse, informal commit messages with nor |
2026-06-29 |
| commit |
0.00 |
Use inspect.getsource instead of open() for source-reading i |
|
Brief, informal commit summaries; no AI- |
2026-06-29 |
| commit |
0.00 |
Reject assisted generation for LFM2 and LFM2-MoE (set _is_st |
|
Contains technical explanation with doma |
2026-06-29 |
| commit |
0.00 |
Don't pin the gated delta net norm to `cuda:0` with a hardco |
|
Technical, domain-specific with minor gr |
2026-06-29 |
| commit |
0.00 |
Fix auto-mappings registration for remote code & fixes a few |
|
Minimal, informal commit messages; typic |
2026-06-29 |
| commit |
0.00 |
Fix broken internal documentation links (#46945) |
|
Concise commit message with domain conte |
2026-06-28 |
| commit |
0.00 |
Insert a Grafana badge in the PR (#46774) |
|
Informal commit message with humanlike t |
2026-06-28 |
| commit |
0.00 |
Add Nemotron 3.5 ASR Streaming (#46565) |
|
Terse commit messages with domain jargon |
2026-06-27 |
| commit |
0.00 |
[NemotronAsrStreaming] fix pipeline (#46870) |
|
Very brief, informal style; typical huma |
2026-06-27 |
| commit |
0.00 |
[NemotronAsrStreaming] processor without modular (#46865) |
|
Short, informal commit log; contains min |
2026-06-27 |
| commit |
0.00 |
[`Dia`] Fix docs (#46923) |
|
Commit uses terse messages and domain te |
2026-06-26 |
| commit |
0.00 |
Qwen3 ASR and Forced Aligner (#43838) |
|
Changelog uses natural, human-like phras |
2026-06-26 |
| commit |
0.00 |
[Docs] Fix full disk offloading docs (#46905) |
|
Brief human message with author signatur |
2026-06-26 |
| commit |
0.00 |
Fix dtype casting for quantized multimodal embedders (#46904 |
|
Multiple brief, iterative commit message |
2026-06-26 |
| commit |
0.00 |
Fix flash attention with StaticCache (#46914) |
|
Minimal, non-formal commit messages; hum |
2026-06-26 |
| commit |
0.00 |
[CB] Changes to increase max_batch_tokens (#46712) |
|
Colloquial commit sequence and direct st |
2026-06-26 |
| commit |
0.00 |
Redirect to diffusers pipe in docs for experimental features |
|
Message is short and direct; contains no |
2026-06-26 |
| commit |
0.00 |
Install in docker (#46910) |
|
Minimal, imperative commit message; typi |
2026-06-26 |
| commit |
0.00 |
[CI] Use pre-computed `_OLD_MODELS` in `test_new_models_requ |
|
Terse, informal style typical for humans |
2026-06-26 |
| pr |
0.00 |
Add Cosmos3 Edge model support |
|
Only PR template used, actual content mi |
2026-07-08 |
| pr |
0.00 |
Re-enable VITS model forward test |
|
Only PR template used, actual content mi |
2026-07-06 |
| pr |
0.00 |
🌐 [i18n-KO] Translate accelerator_selection.md to Korean |
|
Only PR template used, actual content mi |
2026-07-08 |
| pr |
0.00 |
[i18n-KO] Translate optimum.md to Korean and fix Furiosa typ |
|
Only a specific PR title; no AI signals |
2026-07-08 |
| pr |
0.00 |
[Generation] Add static ensemble verification for lossy spec |
|
Concise technical title, no AI indicator |
2026-05-14 |
| pr |
0.00 |
Add Granite-swa and Granitemoe-swa model support |
|
Direct and technical; lacks AI hallmarks |
2026-07-08 |
| pr |
0.00 |
Implement VibeVoice |
|
Short, repo-specific title; not AI-gener |
2025-08-29 |
| pr |
0.00 |
[Qwen3ASR] Add hotword parsing, and fix language parsing and |
|
Human style; terse with technical jargon |
2026-07-06 |
| pr |
0.00 |
Fix recurrent-layer padding mask being skipped on continued |
|
Detailed technical phrase, no AI feature |
2026-07-05 |
| pr |
0.00 |
Fix FA performance regression |
|
Template plus brief performance fix; not |
2026-07-07 |
| pr |
0.00 |
T5: add SDPA and Flash Attention 2 support |
|
Technically dense, formatted like human |
2026-06-28 |
| pr |
0.00 |
[docs] fix curly quotes rendering to straight quotes |
|
Just a template and descriptive title; n |
2026-07-07 |
| pr |
0.00 |
Allow metal-flash-sdpa for OpenAIPrivacyFilter on MPS |
|
Contains detailed, technical content typ |
2026-06-18 |
| pr |
0.00 |
prep for 0.1.2 release of mlinter |
|
Brief and informal; clear human authorsh |
2026-07-08 |
| pr |
0.00 |
Cleanup pipelines, stop materializing generators |
|
Direct and succinct; no AI-style phrasin |
2026-07-07 |
| pr |
0.00 |
Fix custom code which doesn't know about the new linear laye |
|
Only a terse, domain-relevant commit tit |
2026-07-08 |
| pr |
0.00 |
Don't crash the ASR pipeline when an installed torchcodec ca |
|
Contains only a concise, technical fix t |
2026-07-04 |
| pr |
0.00 |
Fix hardcoded paths in siglip checkpoint/vocab loading |
|
Just a brief summary line describing the |
2026-07-08 |
| pr |
0.00 |
Reject path traversal in the `transformers_weights` config f |
|
Human-like, short, and specific technica |
2026-06-25 |
| pr |
0.00 |
trainer: clear MPS graph cache via torch_empty_cache_steps |
|
Human style: technical and brief title, |
2026-05-07 |
| pr |
0.00 |
Fix DeepSpeed SP loss aggregation and LocalLayerRepository k |
|
Highly condensed technical terminology, |
2026-07-04 |
| pr |
0.00 |
[docs] Custom code conversion mapping |
|
Very brief tag-style docs update, clearl |
2026-07-06 |
| pr |
0.00 |
Olmo3 CI |
|
Project or feature name only, human-like |
2026-07-08 |
| pr |
0.00 |
[docs] MoE decode switching |
|
Template plus a brief docs-related title |
2026-07-07 |
| pr |
0.00 |
model: Add NVIDIA Canary-1B-v2 to Transformers |
|
Concise model addition; typical for huma |
2026-06-22 |
| pr |
0.00 |
[Fix] Make DeepGEMM triton fallback more robust |
|
Standard brief fix notation, no AI phras |
2026-07-07 |
| pr |
0.00 |
wip step 3 7 flash |
|
Very informal, technical with link, huma |
2026-06-15 |
| pr |
0.00 |
Fix CI read-only cache failures by patching cached_files in |
|
Only PR title shown, terse and domain-sp |
2026-07-03 |
| pr |
0.00 |
[DiffusionGemma] Support gradient checkpointing |
|
Brief, technical module title only. |
2026-06-11 |
| pr |
0.00 |
Fix experts implementation in two spots |
|
Concise, technical title typical for PRs |
2026-07-06 |
| pr |
0.00 |
Bmg expectation |
|
Very short, non-formal title only. |
2026-07-08 |
| pr |
0.00 |
Fix get_json_schema crash on non-string docstring choices |
|
Technical title with specific bug refere |
2026-07-04 |
| pr |
0.00 |
Fix RT-DETR v2 converter for current model keys |
|
Brief technical summary with specific co |
2026-07-07 |
| pr |
0.00 |
Add native FSDP2 module + migration |
|
Short PR title with domain abbreviation. |
2026-06-17 |
| pr |
0.00 |
Fix it |
|
Extremely terse; human-like minimal titl |
2026-07-07 |
| pr |
0.00 |
Fix GLM video frame padding for temporal patches |
|
References domain-specific patching; tit |
2026-07-07 |
| pr |
0.00 |
Add support for NeuCodec |
|
Concise technical title, no AI traits in |
2026-07-07 |
| pr |
0.00 |
Add FSDP plans to all models |
|
Technical title, no evidence of AI phras |
2026-07-08 |
| pr |
0.00 |
Extend continuous batching memory prediction test to XPU |
|
Only a concise technical title, no AI-st |
2026-07-08 |
| pr |
0.00 |
[Fix] Remove old automatic cross attn pattern from output re |
|
Technical, patch-focused title; uses rep |
2026-07-07 |
| pr |
0.00 |
Fix case where `_LazyAutoMapping.register` is passed a `str` |
|
Only a brief, technical PR title is pres |
2026-07-07 |
| pr |
0.00 |
🚨 Enable SDPA (and other attention backends) for T5 and prop |
|
Technical title only, no AI hallmarks or |
2026-07-02 |
| pr |
0.00 |
Fix dtype-sensitive vision position ids for Idefics/SmolVLM |
|
Technical, domain-specific title; no AI |
2026-07-07 |
| pr |
0.00 |
extend tests/quantization/gptq/test_gptq.py::GPTQTestCUDA an |
|
Title with domain references, no AI styl |
2026-07-08 |
| pr |
0.00 |
chore(qa): fix Windows compatibility for type checker exclud |
|
Technical, domain-specific 'chore' title |
2026-07-08 |
| pr |
0.00 |
add XPU output expectations for minicpm3 tests |
|
Technical test title, no AI patterns det |
2026-07-06 |
| pr |
0.00 |
Diffusion gemma: fix failed test cases |
|
Concise bugfix title, lacks AI-style phr |
2026-07-02 |
| pr |
0.00 |
add XPU Expectation for cosmos3_omni tests |
|
Informal message and review, clearly hum |
2026-06-25 |
| pr |
0.00 |
Fix IndexError Bug in XLMRoberta/Camembert ForMultipleChoice |
|
Bug-fix phrasing and technical specifici |
2026-07-07 |
| pr |
0.00 |
Normalize multimodal input keys in AnyToAnyPipeline |
|
Only a terse, domain-specific PR title i |
2026-07-04 |
| pr |
0.00 |
add paged attention tests support for XPU |
|
BRIEF domain-specific title only, not AI |
2026-07-08 |
| pr |
0.00 |
Restore dynamic_full full-layer cache guard for assisted/con |
|
Technical title, no AI-specific free tex |
2026-07-08 |
| pr |
0.00 |
Fix Aria checkpoint key conversion mapping |
|
Only a terse, highly technical PR title |
2026-07-08 |
| pr |
0.00 |
Fix off-by-one in MoonshineStreaming sliding_window_mask_fun |
|
Brief, domain-specific PR title; no evid |
2026-07-08 |
| pr |
0.00 |
Restrict QuantizedCache to only full attention |
|
Concise, technical-only title; no AI sig |
2026-07-08 |
| pr |
0.00 |
Raise early for unsupported QuantizedCache layer types |
|
Technical title only, no AI-generated st |
2026-06-20 |
| pr |
0.00 |
Fix device_map="auto" under-packing that spills quantized mo |
|
Technical device_map title with specific |
2026-07-04 |
| pr |
0.00 |
Add linear_attn entries to Qwen3.5 base_model_tp_plan |
|
Short, technical PR title; human style i |
2026-07-01 |
| pr |
0.00 |
Fix Qwen Omni batched text post-processing |
|
Only contains a technical PR title with |
2026-07-04 |
| pr |
0.00 |
Forward direct generation kwargs in multimodal generation pi |
|
PR content is a concise technical title |
2026-07-04 |
| pr |
0.00 |
Fix prompt lookup decoding generating past max_length |
|
Short, domain-specific PR title; typical |
2026-07-01 |
| pr |
0.00 |
Add Gemma4Unified sequence classification |
|
Technical title only, no extra phrasing |
2026-06-29 |
| pr |
0.00 |
Port ESMC and ESMFold2 to Transformers |
|
Technical PR title only, no AI-style nar |
2026-06-04 |
| pr |
0.00 |
FSDP orchestration: apply + loading/saving |
|
Only a terse, domain-specific title; no |
2026-07-01 |
| pr |
0.00 |
Shard on read Dtensor aware |
|
PR content is a minimal technical title; |
2026-06-17 |
| pr |
0.00 |
🚨[wav2vec2] Support attn_implementation=sdpa dispatch |
|
Short, emoji-prefixed title with jargon, |
2026-05-25 |
| pr |
0.00 |
Skip caching_allocator_warmup on Neuron (no reuse pool to wa |
|
Technical problem as title, typical huma |
2026-07-02 |
| pr |
0.00 |
Fix xLSTM train mode crash when return_last_states=False |
|
Concise, technical bugfix title; structu |
2026-07-02 |
| pr |
0.00 |
[serge] Fix 20 integration tests for model `whisper` failing |
|
— |
2026-07-08 |
| pr |
0.00 |
[serge] Fix 8 integration tests for model `falcon_mamba` fai |
|
— |
2026-07-07 |
| pr |
0.00 |
Add Fun-ASR-Nano model |
|
Only PR template and a model addition; n |
2026-05-24 |
| pr |
0.00 |
[fix] update integration test values |
|
Simple update title, only template, no s |
2026-07-07 |
| pr |
0.00 |
Add eval_epochs to TrainingArguments to support evaluating e |
|
Template only; no author free-text to an |
2026-07-05 |
| pr |
0.00 |
[docs] continuous batching (offloading behavior, max batch t |
|
Brief domain-jargon title and clear bull |
2026-06-26 |
| pr |
0.00 |
processor tests: use tiny Hub repos to reduce CI memory |
|
Template only; no free-text content from |
2026-07-06 |
| pr |
0.00 |
[docs] fix autolinks |
|
Minimalist, conventional PR title; no si |
2026-06-29 |
| pr |
0.00 |
Fix CohereAsr training-loss double-shift (train against labe |
|
Technical title with code-like context, |
2026-07-05 |
| pr |
0.00 |
Update serve chat parsing |
|
Conventional, succinct PR title; no AI-t |
2026-05-28 |
| pr |
0.00 |
fix: allow num_labels property to return None when id2label |
|
Only a brief technical title with no AI |
2026-07-04 |
| pr |
0.00 |
[xLSTM] Fix train mode crashes with return_last_states=False |
|
Technical title with domain specifics, n |
2026-07-03 |
