Wixiuenne しかちゃん (@Shikha) 在 在 NVIDIA GeForce 920M 跑 LLM :melting_face: 中发帖
[图片]
硬件:
CPU: i3-5005U
RAM：DDR3 4GB 板载内存
GPU: NVIDIA GeForce 920M
环境:
系统: Arch Linux
CUDA: V10.2.89
Driver Version: 470.256.02
GCC: 8.5.0
模型: qwen2.5-1.5b-instruct-q4_k_m.gguf
显存占用: 1289MiB / 2004MiB
llm_load_tensors: offloading 20 repeating layers to GPU
llm_load_tensors: offloaded 20/29 layers to GPU
llm_load_tensors: CPU_Mapped model buffer size = 1059.89 MiB
llm_load_tensors: ...