dalle-mini
dalle-mini copied to clipboard
DALL·E Mini - Generate images from a text prompt
jpeg artifacts can be quite annoying.
Hello, I'm attempting to fine tune dalle-mega but hit this error while trying to stand things up: ``` Traceback (most recent call last): File "/home/netruk44/ml/workspace/repos/dalle-mini/tools/train/train.py", line 1742, in main() File...
__Originally posted by @mirabish in https://github.com/just-the-docs/just-the-docs/pull/917__
Scaling was not applied properly.
Basically what the title says, using images to help fine tune the results
just curios as if i want to host my self hosting how much will it cost
can you please add an option to upload an image and paint an area to edit with text
How difficult would it be to use a different GAN? Instead of VQGAN use for example StyleGAN?
Can someone please remove the wandb login requirement from the notebook?
Errors while trying to generate images with GPU: ``` XlaRuntimeError Traceback (most recent call last) Input In [13], in () 23 encoded_images = encoded_images.sequences[..., 1:] 24 # decode images --->...