-
Story
-
Resolution: Done
-
Undefined
-
None
-
None
-
None
LoRA layer revealed an issue in training in the Triton kernels. We need to pull up vLLM again and it will be either v0.8.4.1 or v0.8.5.0.
- blocks
-
AIPCC-726 RHEL AI 1.4.3 - vllm fails to start on AMD for SDG - Ready to be tested
-
- Closed
-