[Model] Apply shared experts overlap optimization to all models with shared experts (#26145)

bnellnm committed about 1 year ago

[Attention][DCP] Support DCP with query length > 1 (MTP) with FA3 (#25049)

Ming Yang committed about 1 year ago

[Models][Qwen] Replace `pad` with `cat` for better performance (#26486)

Lukas Geiger committed about 1 year ago

Upgrade Pydantic to v2.12.0 and remove hack for Python 3.13 (#26481)

Harry Mellor committed about 1 year ago

Revert #26113 "[Frontend] CompilationConfig overhaul (#20283): deprecate use_inductor in favor of backend, simplify custom_ops" (#26472)

Jiangyun Zhu committed about 1 year ago

[Feature] Use pydantic validation in parallel.py config (#26417)

Simon Danielsson committed about 1 year ago

[V0 deprecation] Remove `QKVCrossParallelLinear` implementation (#26475)

Isotr0py committed about 1 year ago

[Bugfix] Move current_platform import to avoid python import cache. (#16601)

Wenzheng Bi committed about 1 year ago

[Models][Qwen3VL] Optimise `_validate_and_reshape_mm_tensor` (#26426)

Lukas Geiger committed about 1 year ago

[Bugfix] Fix SHM cache initialization (#26427)

Cyrus Leung committed about 1 year ago

[Misc] Upgrade more code to Python 3.10 (#26463)

Cyrus Leung committed about 1 year ago

[Feature] Use pydantic validation in lora.py and load.py configs (#26413)

Simon Danielsson committed about 1 year ago

[doc] add Volcengine as a compute sponsor (#26477)

youkaichao committed about 1 year ago

[Misc] Misc code simplifications (#26450)

Nick Hill committed about 1 year ago

[torchao] Add support for ModuleFqnToConfig using regex (#26001)

Jerry Zhang committed about 1 year ago

Add: Support for multiple hidden layers in Eagle3 (#26164)

Rahul Tuli committed about 1 year ago

Enable `RMSNorm` substitution for Transformers backend (#26353)

Harry Mellor committed about 1 year ago

[Model] Gemma3: Fix GGUF loading and quantization (#26189)

Luciano Martins committed about 1 year ago

Bump Flashinfer to v0.4.0 (#26326)

elvischenv committed about 1 year ago

Update Dockerfile and install runai-model-streamer[gcs] package (#26464)

pwschuurman committed about 1 year ago