vllm/v1/worker/gpu_model_runner.py · c1909e7e8ccd2037e76536a8e726120c85d3754e · OpenDAS / vllm_cscc · GitLab

Find file Blame History Permalink

[Optimization] Cache sampled token ids in model runner (#20291) · 7f280d69
Woosuk Kwon authored Jul 01, 2025
```
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
```
7f280d69

gpu_model_runner.py 119 KB