LocalLLaMA Reddit5d ago
Ryzen 395 (Strix Halo) massive performance degradation at high context with ROCm bug I found, may explain speed differences between ROCm and Vulkan with llama-cpp
AI is generating summary...
Comments
No comments yet
Be the first to comment