comfyui-Image-reward

[a/ImageReward](https://github.com/THUDM/ImageReward): Human preference learning in text-to-image generation. This is a [a/paper](https://arxiv.org/abs/2304.05977) from…

By shadowcz007·Created 2 years ago·Updated 2 years ago· 32

shadowcz007/comfyui-Image-reward

Nodes2

On cloudLocal install

Category♾️Mixlab/Image

Stars32

Updated2 years ago

Nodes (2)

Image Batch To List

Split a batch so ComfyUI runs one image at a time

♾️Mixlab/Image

Image Reward

♾️Mixlab/Image

Readme

comfyui-Image-reward

1711282071529 1711282130123

workflow example

Collaborate with mixlab-nodes to convert the workflow into an app.

配合mixlab-nodes，把workflow转为app使用。

ImageReward

Human preference learning in text-to-image generation. This is a paper for NeurIPS 2023, trained using the professional large-scale dataset ImageRewardDB: approximately 137,000 comparison pairs.

文本到图像生成中的人类偏好学习。这是NeurIPS 2023的一篇论文，训练使用专业的大规模数据集ImageRewardDB：约13.7万个⽐较pairs。

install

requirement.txt

models

ImageReward 放到 ComfyUI/models/image_reward/ImageReward.pt ComfyUI/models/image_reward/med_config.json

bert-base-uncased 放到 ComfyUI/models/image_reward/bert-base-uncased/