Popular repositories Loading

1. qwen-mtp-optimizations qwen-mtp-optimizations Public

   Multi-Token Prediction for Qwen3.5-27B in llama.cpp

   2

2. qwen-mtp-research qwen-mtp-research Public

   Multi-Token Prediction for Qwen3.5-27B in llama.cpp

   Python 2

3. fast-forth fast-forth Public

   A modern, high-performance ANS Forth compiler with LLVM backend, type safety, and world-class developer tools

   Rust 1

4. getattrlistbulk-rs getattrlistbulk-rs Public

   Safe Rust bindings for macOS getattrlistbulk() - high-performance bulk metadata retrieval

   Rust 1

5. qwen-inference-lab qwen-inference-lab Public

   Research log: optimizing Qwen3.5-27B from 29.5 to 42.7 tok/s on Apple Silicon

   Python 1

6. vllm-qwen-patches vllm-qwen-patches Public

   8 bug fixes for vLLM 0.19.0 speculative decoding on Qwen 3.5-27B (MTP + tree attention + modal_mtp)

   Python 1