openinfer-project/openinfer - 5 Help Wanted Issues

Open Issues Need Help

qwen35: migrate GDN prefill to FlashInfer and retire Triton AOT where supported about 2 hours ago

enhancement help wanted qwen35 hw:1-gpu

542

Pure Rust + CUDA LLM inference engine — no PyTorch, OpenAI-compatible, serves Qwen3 to Kimi-K2

Rust

#cuda#cuda-kernels#deepseek#gpu#inference#inference-engine#kimi#kimi-k2#kv-cache#llm#llm-inference#llm-serving#model-serving#moe#openai-api#paged-attention#qwen#qwen3#rust#vllm

bench: add HTTP serving benchmark coverage for Qwen mixed sampling 28 days ago

good first issue

openinfer-project/openinfer

542

Pure Rust + CUDA LLM inference engine — no PyTorch, OpenAI-compatible, serves Qwen3 to Kimi-K2

Rust

#cuda#cuda-kernels#deepseek#gpu#inference#inference-engine#kimi#kimi-k2#kv-cache#llm#llm-inference#llm-serving#model-serving#moe#openai-api#paged-attention#qwen#qwen3#rust#vllm

qwen35: RoPE cache covers 4096 positions but config admits 262144 about 1 month ago

good first issue qwen35

openinfer-project/openinfer

542

Pure Rust + CUDA LLM inference engine — no PyTorch, OpenAI-compatible, serves Qwen3 to Kimi-K2

Rust

#cuda#cuda-kernels#deepseek#gpu#inference#inference-engine#kimi#kimi-k2#kv-cache#llm#llm-inference#llm-serving#model-serving#moe#openai-api#paged-attention#qwen#qwen3#rust#vllm

[Feature] deepseek-v2-lite: bringup and weight loading produce zero log output about 1 month ago

enhancement good first issue

openinfer-project/openinfer

542

Pure Rust + CUDA LLM inference engine — no PyTorch, OpenAI-compatible, serves Qwen3 to Kimi-K2

Rust

#cuda#cuda-kernels#deepseek#gpu#inference#inference-engine#kimi#kimi-k2#kv-cache#llm#llm-inference#llm-serving#model-serving#moe#openai-api#paged-attention#qwen#qwen3#rust#vllm

Preserve streaming usage in the vLLM frontend about 1 month ago

bug good first issue

openinfer-project/openinfer

542

Pure Rust + CUDA LLM inference engine — no PyTorch, OpenAI-compatible, serves Qwen3 to Kimi-K2

Rust

#cuda#cuda-kernels#deepseek#gpu#inference#inference-engine#kimi#kimi-k2#kv-cache#llm#llm-inference#llm-serving#model-serving#moe#openai-api#paged-attention#qwen#qwen3#rust#vllm