fix: use spawn start method for vLLM evaluation to avoid CUDA re-initialization error #2421
fix: use spawn start method for vLLM evaluation to avoid CUDA re-init…
2e1eac14
fix(eval): harden vllm spawn initialization and tests
12facc78
xin3he
approved these changes
on 2026-10-13
xin3he
merged
234ac7af
into main 34 days ago
Assignees
No one assigned
Login to write a write a comment.
Login via GitHub