---
dataset_info:
  features:
  - name: tag
    dtype: string
  - name: model_name
    dtype: string
  - name: model_id
    dtype: int64
  - name: modelVersion_name
    dtype: string
  - name: modelVersion_id
    dtype: int64
  - name: modelVersion_url
    dtype: string
  - name: modelVersion_trainedWords
    dtype: string
  - name: model_download_count
    dtype: int64
  - name: baseModel
    dtype: string
  splits:
  - name: train
    num_bytes: 36188
    num_examples: 200
  download_size: 22662
  dataset_size: 36188
license: openrail
task_categories:
- text-to-image
language:
- en
tags:
- art
- diffusers
size_categories:
- n<1K
---
# GEMRec-18k -- Roster
This is the official model checkpoint metadata dataset for the paper [Towards Personalized Prompt-Model Retrieval for Generative Recommendation](https://github.com/MAPS-research/GEMRec). 

## Dataset Intro
`GEMRec-18K` is a prompt-model interaction dataset with 18K images generated by 200 publicly-available generative models paired with a diverse set of 90 textual prompts. We randomly sampled a subset of 197 models from the full set of models (all finetuned from Stable Diffusion) on [Civitai](https://civitai.com/) according to the popularity distribution (i.e., download counts) and added 3 original Stable Diffusion checkpoints (v1.4, v1.5, v2.1) from HuggingFace. All the model checkpoints have been converted to the [Diffusers](https://huggingface.co/docs/diffusers/index) format. The textual prompts were drawn from three sources: 60 prompts were sampled from [Parti Prompts](https://github.com/google-research/parti); 10 prompts were sampled from [Civitai](https://civitai.com/) by popularity; we also handcrafted 10 prompts following the prompting guide from [DreamStudio](https://beta.dreamstudio.ai/prompt-guide), and then extended them to 20 by creating a shortened and simplified version following the tips from [Midjourney](https://docs.midjourney.com/docs/prompts). The textual prompts were classified into 12 categories: abstract, animal, architecture, art, artifact, food, illustration, people, produce & plant, scenery, vehicle, and world knowledge.

## Links
#### Dataset
- [GEMRec-Promptbook](https://huggingface.co/datasets/MAPS-research/GEMRec-PromptBook): The full version of our GemRec-18k dataset (images & metadata).
- [GEMRec-Metadata](https://huggingface.co/datasets/MAPS-research/GEMRec-Metadata): The pruned version of our GemRec-18k dataset (metadata only).
- [GEMRec-Roster](https://huggingface.co/datasets/MAPS-research/GEMRec-Roster): The metadata for the 200 model checkpoints fetched from [Civitai](https://civitai.com/).

#### Space
- [GEMRec-Gallery](https://huggingface.co/spaces/MAPS-research/GEMRec-Gallery): Our web application for browsing and comparing the generated images.

#### Github Code
- [GEMRec](https://github.com/MAPS-research/GEMRec)


## Acknowledgement
This work was supported through the NYU High Performance Computing resources, services, and staff expertise.

## Citation
If you find our work helpful, please consider cite it as follows:
```bibtex
@article{guo2023towards,
  title={Towards Personalized Prompt-Model Retrieval for Generative Recommendation},
  author={Guo, Yuanhe and Liu, Haoming and Wen, Hongyi},
  journal={arXiv preprint arXiv:2308.02205},
  year={2023}
}
```