Skip to content

Advocate99/DragDiffusion

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

DragDiffusion

This is an unofficial code for DragDiffusion.

We show the DragDiffusion in a proof-of-concept way where we present the clean structured code of per-image optimization.

We hope the implementation of the principles helps.

The performances are not comparable with the paper's, and considering the performances, we do not include the GUI version yet.

Demo case of Our Implementation

Environment

conda env create -f environment.yml
conda activate diff

How-to

  1. Put the image file in the ./finetune_data/ and finetune the SD-v1.5 with LoRA.

    python dreambooth_lora.py --pretrained_model_name_or_path 'runwayml/stable-diffusion-v1-5' --instance_data_dir './finetune_data/' --instance_prompt 'xxy5syt00' --num_train_epochs 200 --checkpointing_steps 200 --output_dir 'lora-200'
    
  2. Latent optimization.

    python run_drag.py
    

Acknowledgement

About

Unofficial implementation of DragDiffusion

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages