Testo
Testo
I tryed it on an old commit, but still get the error.
> Are you using --medvram as a startup option ? Yes, and I see that without this option creating embeddings works.
Look here: https://github.com/AUTOMATIC1111/stable-diffusion-webui/issues/3066#issuecomment-1284931959
Wenn I run the code a second time in a new notebook, I get this output: ``` Wed Dec 22 12:28:31 2021 +-----------------------------------------------------------------------------+ | NVIDIA-SMI 495.44 Driver Version: 495.44 CUDA...
I have to add that after the training is started, more gpu memory is used. When I increase the batch size until I get an oom-error gpu memory usage is...
Hello and sorry for my late answer. I tested the links above, but I was not able to get any new information. Most links where just to old to work...
Hello sushreebarsa, its a similar output: ``` TensorFlow: 2.9.0-dev20211227 Python: 3.8.10 (default, Nov 26 2021, 20:14:08) [GCC 9.3.0] ***CUDA*** nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2021 NVIDIA Corporation...