vllm/model_executor/layers/quantization/fp8.py · 0130223bd9900710a0d93e46a4255ec5d1a077a8 · OpenDAS / vllm_cscc · GitLab

Find file Blame History Permalink

fix memory for online fp8 quantization with streaming weight load (#31914) · 0130223b
Vasiliy Kuznetsov authored Feb 02, 2026
```
Signed-off-by: vasiliy <vasiliy@fb.com>
```
0130223b

fp8.py 49.8 KB