Contribute to EsmailLeath/Alemdar development by creating an account on GitHub.
Production-grade KV-cache and weight quantization for llama.cpp, with cross-backend kernel support for Apple Silicon, NVIDIA CUDA, AMD ROCm, and Vulkan.
一些您可能无法访问的结果已被隐去。
显示无法访问的结果