GitLab

Co-authored-by: Robert Shaw <114415538+robertgshaw2-neuralmagic@users.noreply.github.com>

Co-authored-by: Robert Shaw <rshaw@neuralmagic>

Co-authored-by: Robert Shaw <rshaw@neuralmagic>

Co-authored-by: Robert Shaw <rshaw@neuralmagic>

[Misc] Add channel-wise quantization support for w8a8 dynamic per token activation quantization (#5542)

Co-authored-by: Varun Sundar Rabindranath <varunsundar08@gmail.com>
Co-authored-by: Varun Sundar Rabindranath <varun@neuralmagic.com>

Co-authored-by: Philipp Moritz <pcmoritz@gmail.com>
Co-authored-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
Co-authored-by: mgoin <michael@neuralmagic.com>
Co-authored-by: Tyler Michael Smith <tyler@neuralmagic.com>
Co-authored-by: Cody Yu <hao.yu.cody@gmail.com>

Fixes fp8 iterface which broke in AQLM merge.

Co-authored-by: mgoin <michael@neuralmagic.com>

Provide an initial support to FP8 computation. This PR is inspired by HuggingFace TGI: huggingface/text-generation-inference#1726

This feature can be enabled with --quantization fp8 or -q fp8 when launching an engine.

Algorithm:
We still load a model checkpoint in FP16/BF16. After the weights are loaded, Fp8LinearMethod calculates the per-tensor scaling factor of weights and quantizes the weights accordingly. The scaling factor will then be stored for future use. Meanwhile, the per-tensor scaling factor for activations is calculated in every forward pass.

Initial Results:
Currently tested Mistral-7B on 1xH100. With prompt length ~5 and decoding length 128:

BF16: 1.47s
FP8: 1.66s
I'll try to use larger models and try to find more performance bottleneck. Meanwhile, you're welcome to try this code.

[WIP][Core][Refactor] move vllm/model_executor/parallel_utils into vllm/distributed and vllm/device_communicators (#3950)

Co-authored-by: Robert Shaw <114415538+rib-2@users.noreply.github.com>
Co-authored-by: alexm <alexm@neuralmagic.com>

Co-authored-by: Jiang Li <jiang1.li@intel.com>
Co-authored-by: Kunshang Ji <kunshang.ji@intel.com>