前言
记录自己常用的指令
Linux
vncserver :99
vncserver kill:99
查看当前运行的matlab进程
ps -ef | grep MATLAB
kill/
查看GPU的利用
nvidia-smi
conda
安装package
conda activate base
pip install XXXX
PyTorch
2021.08.18
使用GPU训练：
/home/qzhang/anaconda3/lib/python3.8/site-packages/torch/cuda/init.py:106: UserWarning:
RTX A6000 with CUDA capability sm_86 is not compatible with the current PyTorch installation.
The current PyTorch install supports CUDA capabilities sm_37 sm_50 sm_60 sm_61 sm_70 sm_75 compute_37.
If you want to use the RTX A6000 GPU with PyTorch, please check the instructions at https://pytorch.org/get-started/locally/
warnings.warn(incompatible_device_warn.format(device_name, capability, " ".join(arch_list), device_name))
之前出现过，使用
vit_model.load_state_dict(torch.load(model_load_path))
时，Python崩溃的情况。应该是因为cuda的版本问题。
找打了如下文章：
nvidia显卡和cuda版本关系
尝试：更新cuda
然后发现
|