huggingface/text-generation-inference
5
0
Fork 0
mirror of https://github.com/huggingface/text-generation-inference.git synced 2025-04-21 06:42:10 +00:00
Code Issues Packages Projects Releases Wiki Activity

Fixing the oom maybe with 2.5.1 change. (#2958)

Browse Source
This commit is contained in:
Nicolas Patry 2025-01-28 10:30:28 +01:00 committed by GitHub
parent c690da5973
commit eb3df0f46f
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
server/text_generation_server/models/globals.py
View File

@ -28,7 +28,7 @@ if PREFIX_CACHING and ATTENTION not in {
raise RuntimeError("Prefix caching is only supported with flashinfer") raise RuntimeError("Prefix caching is only supported with flashinfer")
MEM_POOL = torch.cuda.graph_pool_handle() if torch.cuda.is_available() else None MEM_POOL = torch.cuda.graph_pool_handle() if torch.cuda.is_available() else None
TGI_WIGGLE_ROOM = float(os.getenv("TGI_WIGGLE_ROOM", "0.95")) TGI_WIGGLE_ROOM = float(os.getenv("TGI_WIGGLE_ROOM", "0.93"))
assert TGI_WIGGLE_ROOM > 0 assert TGI_WIGGLE_ROOM > 0
assert TGI_WIGGLE_ROOM < 1 assert TGI_WIGGLE_ROOM < 1