Tobi Ore
Tobi Ore
> > It might just be easier to start with the pretrained weights and channels, and instead of inputing single-channel greyscale, just duplicate the greyscale values over 3 channels. Then...
> What happen if I have images with 4 or five channels? It’s similar, I have trained a 5 channel diffusion model. Don’t load the data with Pillow. Use numpy...
To train the model for 1 channel, you need to go to the image_datasets.py file and comment out line 97: arr = np.array(pil_image.convert("RGB")). This line forces the input channel to...
clone the repo and install 'improved_diffusion' using: pip install -e .
Refer to the documentation provided in the repo on how to properly run the code. For example the image_train.py is in the scripts folder. You ought to use scripts/image_train.py
Your `pip install -e` call is wrong. There is a full stop “.” after the -e to install the improved_diffusion package from the current directory `pip install -e .`. Make...