[Bugfix] Fix broken Qwen2.5-omni tests (#17613)

Isotr0py committed over 1 year ago

[Core] Gate `prompt_embeds` behind a feature flag (#17607)

Cyrus Leung committed over 1 year ago

[Bugfix][ROCm] Using device_type because on ROCm the API is still torch.cuda (#17601)

Gregory Shtrasberg committed over 1 year ago

[Build/CI] Upgrade CUTLASS to 3.9.1 (#17602)

Tyler Michael Smith committed over 1 year ago

[Frontend][TPU] Add TPU default max-num-batched-tokens based on device name (#17508)

Chenyaaang committed over 1 year ago

[Quantizaton] [AMD] Add support for running DeepSeek int8 w8a8 MoE on ROCm (#17558)

rasmith committed over 1 year ago

Update test requirements to CUDA 12.8 (#17576)

22quinn committed over 1 year ago

[Neuron][Build] Require setuptools >= 77.0.3 for PEP 639 (#17603)

Liangfu Chen committed over 1 year ago

[release] Add command to clean up Docker containers/images in TPU release machine (#17606)

Kevin H. Luu committed over 1 year ago

fix typo in logging (#17605)

Eric Hartford committed over 1 year ago

[easy] Print number of needed GPUs in skip message (#17594)

Richard Zou committed over 1 year ago

[Hardware][AMD] Improve OAM device ID + llama4 Maverick MOE tuning (#16263)

Xiaodong Wang committed over 1 year ago

Add NVIDIA TensorRT Model Optimizer in vLLM documentation (#17561)

Zhiyu committed over 1 year ago

permute/unpermute kernel for moe optimization (#14568)

Caleb_Du committed over 1 year ago

[BugFix][Attention] Fix sliding window attention in V1 giving incorrect results (#17574)

Lucas Wilkinson committed over 1 year ago

[Bugfix] fix tmp_out and exp_sums dimensions (#17438)

Hui Liu committed over 1 year ago

[Misc] Clean up input processing (#17582)

Cyrus Leung committed over 1 year ago

[doc] miss result (#17589)

Reid committed over 1 year ago

Support W8A8 INT8 MoE for compressed-tensors (#16745)

Michael Goin committed over 1 year ago

[Security] Fix image hash collision (#17378)

Cyrus Leung committed over 1 year ago