ImageReward

ImageReward is the first general-purpose text-to-image human preference RM which is trained on in total 137k pairs of expert comparisons, based on text prompts and corresponding model outputs from DiffusionDB. We demonstrate that ImageReward outperforms existing text-image scoring methods, such as CLIP, Aesthetic, and BLIP, in terms of understanding human preference in text-to-image synthesis through extensive analysis and experiments.

Approach

Setup

Environment: install dependencies via pip install -r requirements.txt.

Usage

import os
import torch
import ImageReward as reward

if __name__ == "__main__":
    prompt = "a painting of an ocean with clouds and birds, day time, low depth field effect"
    img_prefix = "assets/images"
    generations = [f"{pic_id}.webp" for pic_id in range(1, 5)]
    img_list = [os.path.join(img_prefix, img) for img in generations]
    model = reward.load()
    with torch.no_grad():
        ranking, rewards = model.inference_rank(prompt, img_list)
        # Print the result
        print("\nPreference predictions:\n")
        print(f"ranking = {ranking}")
        print(f"rewards = {rewards}")
        for index in range(len(img_list)):
            score = model.score(prompt, img_list[index])
            print(f"{generations[index]:>16s}: {score:.2f}")

The output will look like the following (the exact numbers may be slightly different depending on the compute device):

Preference predictions:

ranking = [1, 2, 3, 4]
rewards = [[0.5811622738838196], [0.2745276093482971], [-1.4131819009780884], [-2.029569625854492]]
          1.webp: 0.58
          2.webp: 0.27
          3.webp: -1.41
          4.webp: -2.03

Test

Setup for baselines

Environment

$ pip install git+https://github.com/openai/CLIP.git

Checkpoint

Models	File Paths	Download Links
ImageReward	checkpoint/	Download
CLIP Score	checkpoint/clip/	Download
BLIP Score	checkpoint/blip/	Download
Aesthetic	checkpoint/aesthetic/	Download

Data

Data	File Paths	Download Links
test_images	data/	Download

Download test_images.zip and unzip it to data/test_images/

One step for test

$ python test.py

The test result is:

Models	Preference Acc.
CLIP Score	54.82
Aesthetic Score	57.35
BLIP Score	57.76
ImageReward (Ours)	65.14