vram ~0.5-1Gb vram usage spikes during prompt processing on vulkan llama.cpp 7900xtx

🌐 链接: https://www.reddit.com/r/LocalLLaMA/comments/1xs72gf/vram_051gb_vram_usage_spikes_during_prompt/
🏷️ 分组: Reddit｜AI·自托管·VPS
🕒 时间: 2026-11-10 07:16:47
Reddit
From the LocalLLaMA community on Reddit
From the LocalLLaMA community on Reddit
Explore this post and more from the LocalLLaMA community