Skip to content

Conversation

@pandyamarut
Copy link
Contributor

No description provided.

Signed-off-by: pandyamarut <pandyamarut@gmail.com>
Signed-off-by: pandyamarut <pandyamarut@gmail.com>
Signed-off-by: pandyamarut <pandyamarut@gmail.com>
Signed-off-by: pandyamarut <pandyamarut@gmail.com>
Signed-off-by: pandyamarut <pandyamarut@gmail.com>
Signed-off-by: pandyamarut <pandyamarut@gmail.com>
endolith added a commit to endolith/worker-vllm that referenced this pull request Nov 14, 2025
Implement changes from runpod-workers/worker-vllm PRs runpod-workers#234, runpod-workers#236, and runpod-workers#138:

- Remove space from gpuIds in hub.json (PR runpod-workers#234)
- Remove unsupported CUDA versions 12.1-12.4 from hub.json and tests.json (PR runpod-workers#236)
- Add error handling for engine initialization and handler exceptions (PR runpod-workers#138 style)
- Ensure all errors return proper ErrorResponse format

These fixes address GPU allocation, CUDA compatibility, and error handling
issues that cause the testing phase to hang indefinitely.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants