Swin Transformer进行语义分割，Ubuntu系统，解决RuntimeError: one of the variables needed for gradient computation

1、问题描述

        最近在用Swin Transformer做语义分割的任务，基本的配置环境参考的是GitHub的基础安装步骤，主要的是环境的配置问题，mmcv的配置要求，不允许高于1.4.0，所以，在train.py文件进行训练运行的时候，报错：RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation: [torch.cuda.FloatTensor [2, 512, 32, 32]], which is output 0 of ReluBackward0, is at version 1; expected version 0 instead. Hint: enable anomaly detection to find the operation that failed to compute its gradient, with torch.autograd.set_detect_anomaly(True).

2、解决方案与配置单

        1、python=3.8
        2、conda install pytorch==1.8.0 torchvision==0.9.0 torchaudio==0.8.0 cudatoolkit=11.1 -c         pytorch -c conda-forge
        3、pip install mmcv-full==1.4.0 -f         https://download.openmmlab.com/mmcv/dist/cu111/torch1.8.0/index.html

3、总结

        就是环境问题，这是我浪费了一天的时间，总结出来的配置单，大家有问题也可以私信我