学习过程中遇到的bug(持续更新)

• 错误：RuntimeError: CUDA error: unknown error
• 解决方法：sudo rm -rf ~/.nv

• 错误：RuntimeError: Failed to export an ONNX attribute, since it’s not constant, please try to make things (e.g., kernel size) static if possible
• 解决方法：pytorch降级 将pytorch版本从1.1.0降为1.0.1

• 错误：cv2.error: /io/opencv/modules/imgproc/src/filter.cpp:3618: error: (-213) Unsupported combination of source format (=1), and buffer format (=5) in function getLinearRowFilter
• 解决方法：result.astype(‘uint8’)

• 错误：ImportError: cannot import name 'imread'
• 解决办法：查看了挺多都说安装上pillow库就ok了。但是我本身就安装了pillow库，依然报这个错误。这时候需要把scipy降级到1.2.1 pip install scipy==1.2.1

• 错误：UnicodeEncodeError ascii' codec can't encode characters
• 解决办法：export LANG=C.UTF-8 export LC_ALL=C.UTF-8

• 错误：1only batches of spatial targets supported (non-empty 3D tensors) but got targets of size
• 解决办法：我这里报错是因为label本应该是一维的，我在保存的时候处理成了3维。

• 错误：HandShake: client signature does not match
• 解决办法：输入数据格式不对，或者-f flv

• 错误：dpkg-deb: error: paste subprocess was killed by signal (Broken pipe)
• 解决办法：找到Errors were encountered while processing 后面出错的包，然后 sudo dpkg -i --force-overwrite + 相应的包名 最后运行 sudo apt -f install

9 .

• 错误： pytorch loss 更新版本差异错误。UserWarning: invalid index of a 0-dim tensor
• 解决办法：loss.data[0] 改为loss.item()