CUDA: faster MMVQ for low-bit types by am17an · Pull Request #30517 · ggml-org/llama.cpp

🌐 链接: https://www.reddit.com/r/LocalLLaMA/comments/1x8bqqs/cuda_faster_mmvq_for_lowbit_types_by_am17an_pull/
🏷️ 分组: Reddit｜AI·自托管·VPS
🕒 时间: 2026-10-17 22:33:53
Reddit
From the LocalLLaMA community on Reddit: CUDA: faster MMVQ for low-bit types by am17an · Pull Request #30517 · ggml-org/llama.cpp
From the LocalLLaMA community on Reddit: CUDA: faster MMVQ for low-bit types by am17an · Pull Request #30517 · ggml-org/llama.cpp
Explore this post and more from the LocalLLaMA community