vllm/sequence.py · eb6c50cdc2bfb58591bd524ff08c8016e7c0411a · OpenDAS / vllm_cscc · GitLab

Find file Blame History Permalink

[Core] Cross-attention KV caching and memory-management (towards eventual... · 4238bc82
afeldman-nm authored May 29, 2024
```
[Core] Cross-attention KV caching and memory-management (towards eventual encoder/decoder model support) (#4837)
```
4238bc82

sequence.py 32.2 KB