GitLab

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

Signed-off-by: Prashant Gupta <prashantgupta@us.ibm.com>
Co-authored-by: Travis Johnson <tjohnson31415@gmail.com>

Co-authored-by: DefTruth <31974251+deftruth@users.noreply.github.com>

Co-authored-by: Antoni Baum <antoni.baum@protonmail.com>

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

Co-authored-by: Yun Ding <yunding@nvidia.com>
Co-authored-by: Roger Wang <ywang@roblox.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.

Co-authored-by: Ubuntu <ubuntu@ip-172-31-13-147.ec2.internal>

Co-authored-by: simon-mo <simon.mo@hey.com>

[Core] add an option to log every function call to for debugging hang/crash in distributed inference (#4079)
Co-authored-by: Simon Mo <simon.mo@hey.com>

[Core] Add integrity check during initialization; add test for it (#4155)

[Core] replace narrow-usage RayWorkerVllm to general WorkerWrapper to reduce code duplication (#4024)

Co-authored-by: Simon Mo <simon.mo@hey.com>

Co-authored-by: Dylan Hawk <dylanwawk@gmail.com>