Giszy

When I did the toy regression (see:https://github.com/caffe2/caffe2/blob/master/caffe2/python/tutorials/Toy_Regression.ipynb) ,this bug happened. My caffe2 package is built in windows 10 x64, python 2.7/3.5/3.6. and the bug exists on three versions. But I...

Good. @tomdz Created init net. //--------------------------------------------------------------------------- Exception Traceback (most recent call last) in () 44 45 # Get gradients for all the computations above. ---> 46 gradient_map = train_net.AddGradientOperators([loss]) 47...

OK. @tomdz [G.zip](https://github.com/caffe2/caffe2/files/1058916/G.zip)

@tomdz, my problem has been resolved. I think it was indeed caused by a unicode string.

@fmscole 问题分析： paddle.utils.run_check()运行的时候会搜索libcuda.so，但在WSL2下，该文件是Windows的路径"C:\Windows\System32\lxss\lib\libcuda.so"。所以，出现FatalError: Segmentation fault is detected by the operating system，就是因为在WSL2的Ubuntu系统中缺少libcuda.so所致。 解决方案：将libcuda.so拷贝到/usr/local/cuda/lib64/下即可。测试通过。