Guidance on Training Binary Latent Diffusion Models

[rutuja1409]

Hey @ZeWang95 I have a dataset comprising text prompts paired with corresponding images, and I would like to use your implementation to train a model. However, I am a bit uncertain about how to begin the training process. Could you kindly guide me on the recommended workflow?
Specifically: Should I first run the binary autoencoder script, followed by the Bernoulli diffusion code? Are there any additional steps or best practices that I should follow to ensure effective training?

Your insights and guidance would be helpful in my work.