Search found 4 matches

by lamakaha
Sun Jul 25, 2021 1:59 pm
Forum: Training Support
Topic: crash report while training: Failed to get convolution algorithm. This is probably because cuDNN failed to initialize
Replies: 9
Views: 10459

Re: crash report while training: Failed to get convolution algorithm. This is probably because cuDNN failed to initializ

completely crashed my computer by uninstalling nvidia , runned in circles installing nvidia/cuda, managed to get 10.1 installed using this guide https://malukas.lt/blog/cuda-10-1-anaconda-tensorflow-on-linux-mint-19-3/ still was getting same error.... entered train.ini located in /home/lamakaha/face...
by lamakaha
Sun Jul 25, 2021 11:49 am
Forum: Training Support
Topic: crash report while training: Failed to get convolution algorithm. This is probably because cuDNN failed to initialize
Replies: 9
Views: 10459

Re: crash report while training: Failed to get convolution algorithm. This is probably because cuDNN failed to initializ

Thanks, did following: uninstalled CUDA using sudo apt-get remove nvidia-cuda-toolkit checking if have some leftovers lamakaha@lamakaha-HOME:~$ locate cuda | grep /cuda$ /home/lamakaha/miniconda3/envs/faceswap/lib/python3.8/site-packages/tensorflow/include/external/local_config_cuda/cuda /home/lamak...
by lamakaha
Sun Jul 25, 2021 11:21 am
Forum: Training Support
Topic: crash report while training: Failed to get convolution algorithm. This is probably because cuDNN failed to initialize
Replies: 9
Views: 10459

Re: crash report while training: Failed to get convolution algorithm. This is probably because cuDNN failed to initializ

Hi, thank you for a prompt reply, I did enable Growth, did not help, have same issue, same error. Loading... Setting Faceswap backend to NVIDIA 07/25/2021 13:07:26 INFO Log level set to: INFO 07/25/2021 13:07:27 INFO Model A Directory: '/media/lamakaha/work/projects/deepfake/06_Renders/SanneFace' (1...
by lamakaha
Sun Jul 25, 2021 7:30 am
Forum: Training Support
Topic: crash report while training: Failed to get convolution algorithm. This is probably because cuDNN failed to initialize
Replies: 9
Views: 10459

Caught exception in thread: '_training_0'

Hi folks, doing first steps with a faceswap and hitting the wall here Starting the training process, getting following error Loading... Setting Faceswap backend to NVIDIA 07/25/2021 09:14:05 INFO Log level set to: INFO 07/25/2021 09:14:07 INFO Model A Directory: '/media/lamakaha/work/projects/deepfa...