vllm
https://github.com/vllm-project/vllm
Python
A high-throughput and memory-efficient inference and serving engine for LLMs
Triage Issues!
When you volunteer to triage issues, you'll receive an email each day with a link to an open issue that needs help in this project. You'll also receive instructions on how to triage issues.
Triage Docs!
Receive a documented method or class from your favorite GitHub repos in your inbox every day. If you're really pro, receive undocumented methods or classes and supercharge your commit history.
Python not yet supported24 Subscribers
Add a CodeTriage badge to vllm
Help out
- Issues
- [BugFix] Fix Gemma4 parser delimiter over-consumption bug
- [Metrics] Add num_requests_prefilling/decoding Prometheus gauges
- [Feature] Support custom callable proposer backend for speculative decoding
- Add AEO quality badge to README (score: 50/100)
- [Bug]: vLLM docker container with Qwen3.5 - Connection error
- [Performance]: Qwen/Qwen3.5-35B-A3B-FP8 has a 8.5x perf reggression when using FA3 backend on a Hopper backend
- [Bug]: [CPU] AttributeError: '_OpNamespace' '_C' object has no attribute 'gelu_tanh_and_mul'
- [Bug]: `prefix_caching_hash_algo='xxhash'` without `xxhash` installed cause `generate()` returns outputs with empty content, no exception
- [Bug]: `attention_backend='FLASH_ATTN_DIFFKV'` crashes init on standard models with "too many values to unpack"
- [Bug]: `block_size=8` triggers Triton CompilationError in FlexAttention kernel; other backends correctly reject
- Docs
- Python not yet supported