yassa9/qwen600
A minimalist single-batch CUDA inference engine for Qwen3-0.6B that runs entirely on GPU with no Python dependencies.
Not currently ranked — collecting fresh signals.
star history
This repository implements a static, suckless-style inference engine for the Qwen3-0.6B language model using pure CUDA C/C++. It provides single-batch LLM inference on GPU with compile-time constants for optimization, using cuBLAS and CUB libraries for efficient computation. The engine benchmarks claim 8.5% faster token generation than llama.cpp and 292% faster than HuggingFace with flash-attn. Configuration is done directly in source code to minimize dependencies and abstractions.
Frequently asked
- What is yassa9/qwen600?
- A minimalist single-batch CUDA inference engine for Qwen3-0.6B that runs entirely on GPU with no Python dependencies.
- Is qwen600 open source?
- Yes — yassa9/qwen600 is open source, released under the MIT license.
- What language is qwen600 written in?
- yassa9/qwen600 is primarily written in Cuda.
- How popular is qwen600?
- yassa9/qwen600 has 557 stars on GitHub.
- Where can I find qwen600?
- yassa9/qwen600 is on GitHub at https://github.com/yassa9/qwen600.