Popular repositories Loading

1. ds4 ds4 Public

   Run DeepSeek V4 Flash and GLM 5.2 locally with a fast, optimized native inference engine.

   C

2. a70yas6086.github.io a70yas6086.github.io Public

   Accelerate inference for DeepSeek V4 Flash, GLM 5.2, and DeepSeek V4 PRO with a narrow, optimized engine for Metal, CUDA, and ROCm.

   HTML