biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[V1] Fully Transparent Implementation of CPU Offloading (#15354)

Browse Source 
Signed-off-by: youkaichao <youkaichao@gmail.com>
This commit is contained in:
youkaichao
2025-03-31 20:22:34 +08:00
committed by GitHub
parent e7ae3bf3d6
commit 555aa21905
12 changed files with 148 additions and 25 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
vllm/config.py
View File

@@ -3562,9 +3562,10 @@ class VllmConfig:
if self.cache_config is not None and \
self.cache_config.cpu_offload_gb > 0 and \
self.compilation_config.level != CompilationLevel.NO_COMPILATION:
self.compilation_config.level != CompilationLevel.NO_COMPILATION \
and not envs.VLLM_USE_V1:
logger.warning(
"CPU offload is not supported with `torch.compile` yet."
"CPU offload is not supported with `torch.compile` in v0 yet."
" Disabling `torch.compile`.")
self.compilation_config.level = CompilationLevel.NO_COMPILATION