Skip to content

Latest commit

 

History

History
73 lines (48 loc) · 2.45 KB

README.md

File metadata and controls

73 lines (48 loc) · 2.45 KB
Raw

A Dense Reward View on Aligning Text-to-Image Diffusion with Preference (ICML'24)

Source code for the single and multiple experiments in A Dense Reward View on Aligning Text-to-Image Diffusion with Preference. [Paper].

Bibtex:

@inproceedings{
yang2024adensereward,
title={A Dense Reward View on Aligning Text-to-Image Diffusion with Preference},
author={Shentao Yang and Tianqi Chen and Mingyuan Zhou},
booktitle={Forty-first International Conference on Machine Learning},
year={2024},
url={https://openreview.net/forum?id=xVXnXk9I3I}
}

Dependency

To install the required packages, please run the following command:

bash install_packages.sh

Experiments

Single Prompt Experiments

As a minimal example, our single prompt experiments can be run by the following command

accelerate launch train_t2i.py --expid="single"

We provide our checkpoints in ./ckpts. To evaluate our checkpoints, please use the following command

cd eval_ckpts
torchrun --nproc_per_node $NUM_GPUS_TO_USE --standalone main.py --type="both_seen_unseen" --eval_generated_imgs=1 --metrics="image_reward,aesthetic" --outdir="./outputs"

Note:

  • $NUM_GPUS_TO_USE is the number of gpus you want to use.
  • Set --eval_generated_imgs=0 if evaluating the generated images is not needed, i.e. only want to generate images.
  • Add the flag --return_traj if you want to store the generation trajectory corresponding to each image as well.

Multiple Prompt Experiments

As a minimal example, our multiple prompt experiments can be run by the following command

accelerate launch train_t2i.py --single_flag=0 --expid="multiple"

The above command is a minimal example, please check parse_args.py for available flags.

Declaimer: The HPSv2 train set has not been officially released at this moment. We are currently in the process of consulting with the HPSv2's authors on including those prompts in our repository. For now, we temporarily use the drawbench prompts as a substitution

Acknowledgement

This codebase builds on the following codebases: