https://github.com/laion-ai/dataset-usage
This repository is a summary of all systems and scientific papers that use LAION datasets.
https://github.com/laion-ai/dataset-usage
Last synced: 5 months ago
JSON representation
This repository is a summary of all systems and scientific papers that use LAION datasets.
- Host: GitHub
- URL: https://github.com/laion-ai/dataset-usage
- Owner: LAION-AI
- License: mit
- Created: 2022-04-27T14:40:08.000Z (about 4 years ago)
- Default Branch: main
- Last Pushed: 2022-04-27T15:13:27.000Z (about 4 years ago)
- Last Synced: 2025-05-07T18:13:51.105Z (about 1 year ago)
- Size: 10.7 KB
- Stars: 6
- Watchers: 3
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# dataset-usage
This repository is a summary of all systems and scientific papers that use LAION datasets.
## Papers citing LAION datasets
### LAION-5B
T.B.C.
### LAION-400M
After the release of LAION-400M, several papers used LAION-400M for image generation, text to image generation, image to text generation and text image matching:
- [Vector Quantized Diffusion Model for Text-to-Image Synthesis](https://arxiv.org/abs/2111.14822.pdf) used LAION-400M to train VQ diffusion text to image generation models
- [High-Resolution Image Synthesis with Latent Diffusion Models](https://arxiv.org/abs/2112.10752.pdf) used a subset of LAION-400M to train latent diffusion models
- [General Facial Representation Learning in a Visual-Linguistic Manner](https://arxiv.org/abs/2112.03109.pdf) LAION-400M face subset to train a face clip
- [BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation](https://arxiv.org/abs/2201.12086) image captioning using LAION-400M subset
- [MAGMA – Multimodal Augmentation of Generative Models through Adapter-based Finetuning](https://arxiv.org/pdf/2112.05253.pdf) was trained on image question answering using a LAION-400M subset
## Models trained on LAION datasets
### GLIDE finetuning
Clay Mullis (alias afiaka87) used subsets of LAON-2B to fine-tune the Open-AI Glide model and managed to reintroduce human generations.
- https://replicate.com/afiaka87/laionide-v3
- https://wandb.ai/afiaka87/glide_compare/reports/Finetuning-GLIDE-on-Laion5B–VmlldzoxNTg3MTkz
- https://wandb.ai/afiaka87/laionide-v3-glide/reports/Laionide-Version-3-Benchmark–VmlldzoxNjE0MTE3
### CLOOB
https://github.com/crowsonkb/cloob-training
# future projects