Text-to-Image
Diffusers
Safetensors
StableDiffusionXLPipeline
playground

Using DDIMScheduler in Playground v2.5

#35
by XinyangATK - opened

Hi everyone,

Recently I'd like to visualize the intermediate image during each step, I need to decode "pred_x0" based on current latent. However, when I use "DDIMScheduler" to get the "pred_x0", the results always failed that the model always generated bad images. Is there any solution to this issue with DDIMScheduler or is there any other methods to achieve this function with other scheduler?

Sign up or log in to comment