CUDA out of memory

[AvadaBlablabla]

Hello,

first thank for your fork, I've just tried it and it work nicely, the generated pictures were very nice.

My problem is that cuda run out of memory after just one prompt.

I tried "PYTORCH_CUDA_ALLOC_CONF=max_split_size_mb:128" but it doesn't seems to solve the problem.
I think it is missing a memory reset or something, because stopping and starting again the UI allows for it to work again (for just one prompt).
I'm just guessing because I know nothing about python or cuda (and pretty much anything about coding).

Until now I've been using https://github.com/basujindal/stable-diffusion , this fork work great for me.

Ps: I'm running it in a conda environment, could that be a limiting factor for my memory ?

[victordibia]

Thanks for the feedback.
There is a chance I should be attempting to flush memory after each generation.
I will look into this.
Conda shold not be a factor.