consistencydecoder

consistencydecoder copied to clipboard

What losses are involved when training consistency decoder vae?

2

I'm recently trying to use this work to my current research, but have problems reproducing the code.❤️

diaoquesang

Open-source model definitions

7

Hey @gabgoh, Super cool that you're open-sourcing the consistency decoder of Dalle-3 with a MIT license :heart: Any chance you can also add the model definitions of the torch.jit binary?...

patrickvonplaten

Can we call consistency_decoder in context of torch.enable_grad() ?

zwx8981

Hi @gabgoh , I attempted to reproduce the results of the code provided in the Usage section of the README, but I encountered some discrepancies compared to the expected output...

LaLaLailalai

No difference when using consistencydecoder in diffusers

1

Here is my code, however, there seems no difference between gan_vae and cosistency_vae. What's wrong? `import torch` `from diffusers import StableDiffusionPipeline, ConsistencyDecoderVAE` `vae = ConsistencyDecoderVAE.from_pretrained("openai/consistency-decoder", torch_dtype=torch.float16)` `pipe_gan = StableDiffusionPipeline.from_pretrained( "runwayml/stable-diffusion-v1-5",...

howardgriffin

It's not for SD1, or it should use specified libraries versions, or something else? Code: ``` latent = txt2img( prompt=prompt, width=width, height=height, num_images_per_prompt = num_image, guidance_scale = guidance, num_inference_steps =...

Eliont

Significant performance problems, see profiler screenshot

3

 Running the consistency decoder takes several seconds and most of this time is spent in a stalled state and reducing the number of diffusion steps leads to no meaningful...

felix-red-panda

512x512 output

1

Hi, Does it work for 512x512 images? Thanks!

nicolas-dufour

How to use this VAE on LCM

5

Thanks to open your work. Btw, I ran the code: ```python from diffusers import DiffusionPipeline,StableDiffusionPipeline import torch from consistencydecoder import ConsistencyDecoder, save_image, load_image pipe = DiffusionPipeline.from_pretrained("SimianLuo/LCM_Dreamshaper_v7", custom_pipeline="latent_consistency_txt2img", custom_revision="main", revision="fb9c5d") decoder_consistency...

alfredplpl

Running in Automatic1111?

3

Can we use this in automatic1111? or does an extension need to be created first?

zono50