RuntimeError: MAGMA gesdd : the updating process of SBDSDC did not converge #13

Enviroment: build from docker

sudo docker run --runtime=nvidia -v /GITHUB/FastPhotoStyle:/home -i -t fastphotostyle:v1.0 /bin/bash

The pytorch found the GPU
torch.cuda.get_device_name(0) => 'GeForce .....'

python demo.py => I Got the error below, does that mean the code didn't use cuda mostly ?

#######################################################
Intel MKL ERROR: Parameter 4 was incorrect on entry to SLASCL.

Intel MKL ERROR: Parameter 4 was incorrect on entry to SLASCL.
Traceback (most recent call last):
File "demo.py", line 68, in
stylized_img = p_wct.transform(cont_img, styl_img, cont_seg, styl_seg)
File "/home/photo_wct.py", line 47, in transform
csF3 = self.__feature_wct(cF3, sF3, cont_seg, styl_seg)
File "/home/photo_wct.py", line 119, in __feature_wct
tmp_target_feature = self.__wct_core(cont_feat_view, styl_feat_view)
File "/home/photo_wct.py", line 154, in __wct_core
c_u, c_e, c_v = torch.svd(contentConv, some=False)
RuntimeError: MAGMA gesdd : the updating process of SBDSDC did not converge (error: 14) at /opt/conda/conda-bld/pytorch_1518238581238/work/torch/lib/THC/generic/THCTensorMathMagma.cu:325