[small][batch invariance] Rename the env and internal flags to simplify usage (#26855)

Bram Wasti committed about 1 year ago

[Bug] Add Assertion for `random-input-len` / `random-output-len` (#26834)

Wentao Ye committed about 1 year ago

[CI] Prune Quantization Tests and skip compilation (#27038)

Michael Goin committed about 1 year ago

[Feature] Migrate DeepGEMM API from `get_m_alignment_for_contiguous_layout` to `get_mk_alignment_for_contiguous_layout` (#26935)

Wentao Ye committed about 1 year ago

[GPTOSS][DP/EP][Marlin] Enable GPTOSS Batched DP/EP using Marlin kernels (#25997)

Varun Sundar Rabindranath committed about 1 year ago

[Bug] Fix batch invariant test `has` to `is` (#27032)

Wentao Ye committed about 1 year ago

Adding Warmup to Benchmark Serving (#26943)

kimbochen committed about 1 year ago

[Compressed Tensors] Always clone output for compile robustness (#26849)

Kyle Sayers committed about 1 year ago

Support block size of 256 used by Intel HPU (#26883)

Mandy Li committed about 1 year ago

[gpt-oss][1/N] EZ: refactor serving_responses for modularity (#26948)

Andrew Xia committed about 1 year ago

Support `set` in the CLI generation (#27031)

Harry Mellor committed about 1 year ago

Fix Qwen2.5 VL image grid docstring (#27033)

ZiTian Zhao committed about 1 year ago

[Model] Fix Qwen3VL mm mapping (#27027)

Jee Jee Li committed about 1 year ago

[Feature] Add process_weights_after_loading to AttentionImpl (#26870)

rongfu.leng committed about 1 year ago

[CI] Replace large models with tiny alternatives in tests (#24057)

Tahsin Tunan committed about 1 year ago

[docs] standardize Hugging Face env var to `HF_TOKEN` (deprecates `HUGGING_FACE_HUB_TOKEN`) (#27020)

Kay Yan committed about 1 year ago

[NIXL] Improve request_finished() debug logs (#25665)

Mark McLoughlin committed about 1 year ago

[Attention] Tune CUTLASS MLA num_splits (#26846)

Matthew Bonanni committed about 1 year ago

[Benchmark] Show E2EL by default for pooling models (#27014)

Cyrus Leung committed about 1 year ago

[Bugfix] Correct LayerNorm epsilon parameter in modernbert.py (#27008)

bogdanm committed about 1 year ago