search

CompVis / latent-diffusion

High-Resolution Image Synthesis with Latent Diffusion Models
MIT License
11.57k stars 1.51k forks source link

Details about training inpainting model #9

Open LigZhong opened 2 years ago

LigZhong commented 2 years ago

Hi, this is an awesome work. I am looking into the code but cannot find the entrance to train the inpainting model. Can you elaborate it a little bit?

roimulia2 commented 2 years ago

The same question as @LigZhong.

Do you happen to know if they have a pre-trained model for the inpainting?

zzw-zwzhang commented 2 years ago

+1 How did you train the inpainting model?

david-svitov commented 2 years ago

+1 How did you train the inpainting model?

It looks like you can use this config (https://github.com/CompVis/latent-diffusion/blob/main/models/ldm/inpainting_big/config.yaml) modifying it by adding 'data' section like in this config (https://github.com/CompVis/latent-diffusion/blob/main/configs/latent-diffusion/ffhq-ldm-vq-4.yaml). But first you need to prepare the data like in this repo: https://github.com/saic-mdal/lama I'm not quite sure it will work, but at least you can try it)

shensongli commented 1 year ago

hello,how did you train the inpainting model?Could you give me some advice

CreamyLong commented 9 months ago

The code and steps of training inpainting model is here