by: Christoph Schuhmann, 8 Aug, 2022
We present LAION-Aesthetics, several collections of subsets from LAION 5B with high visual quality.
To create LAION-Aesthetics we trained several lightweight models that predict the rating people gave when they were asked “How much do you like this image on a scale from 1 to 10?”.
We started with training a linear model on 5000 image-rating pairs from the SAC dataset (which only contained 5000 samples at that time).
Simulacra Aesthetic Captions is a dataset of over 238000 synthetic images generated with AI models such as CompVis latent GLIDE and Stable Diffusion from over forty thousand user submitted prompts.
As inputs this model uses not the images themselves, but their CLIP Image embeddings produced with the Open AI CLIP VIT L 14 model. We call this model LAION-Aesthetics_Predictor V1.
Its results were so encouraging, that we decided to produce 8M and 120M sample subsets of the LAION 5B images with the highest predicted scores, of those that have english texts.
We call the dataset consisting of these 2 subsets LAION-Aesthetics V1.
The model used for creating this subset can be found here.
The LAION-Aesthetics V1 dataset & further details about it can be found here.
After these very encouraging results, we continued to experiment and gathered the following data to train more improved MLP (multi-layer perceptron) models:
- More samples from the SAC dataset, which had grown in the meanwhile to 176000 image - rating pairs
- LAION-Logos, a dataset of 15.000 logo image-text pairs with aesthetic ratings from 1 to 10. We collected this dataset to improve the models abilities to evaluate images with more or less aesthetic texts in them.
- The Aesthetic Visual Analysis (AVA) dataset, which is a large-Scale database for aesthetic visual analysis that contains 250000 photos from dpchallenge.com with several aesthetic ratings from 1 to 10 for most images.
- After training several MLPs with different numbers of layers and parameters and different activation functions, we found that a simple linear model on the top of CLIP ViT/14 produced in our subjective view the visually most appealing results when used to rank images of LAION-5B. (Even though other MLPs with e.g. Relu functions produced slightly lower MSE and MAE loss values.) We call the resulting model trained on SAC, LAION-Logos and AVA LAION-Aesthetics_Predictor V2.
- Visualizations of sorting all 2.37B images from LAION 5B that have English captions into 40 buckets with the LAION-Aesthetics_Predictor V2 can be found here.
Using LAION-Aesthetics_Predictor V2, we created the following subsets of the LAION 5B samples with English captions:
- 1,2B image-text pairs with predicted aesthetics scores of 4.5 or higher: browse huggingface
- 939M image-text pairs with predicted aesthetics scores of 4.75 or higher: browse huggingface
- 600M image-text pairs with predicted aesthetics scores of 5 or higher: browse huggingface
- 12M image-text pairs with predicted aesthetics scores of 6 or higher: browse huggingface
- 3M image-text pairs with predicted aesthetics scores of 6.25 or higher: browse huggingface
- 625K image-text pairs with predicted aesthetics scores of 6.5 or higher: browse huggingface
These subsets overlap. 5 fully includes 6 which includes 6.25 and so on. We call the collection of these subsetsLAION-Aesthetics V2.
At the moment we are translating all 2,15B samples from LAION 5B of the multilingual subset to English using the 1,2B parameter M2M-100 model .
This will allow us to roughly double the size of V2.
Additionally, we are already working on new multimodal large-scale dataset, this time at webpage-level, similar to the interleaved image-text dataset Deepmind used for Flamingo, but also with audio & video files ... and much, much bigger. :)
Stay tuned & keep checking our blog for more datasets in the near future.
If you have any questions or comments or the wish to support our efforts, don’t hesitate to join our Discord community and contact us.
Christoph Schuhmann ( spirit-from-germany#1488 ) and Romain Beaumont ( rom1504#5008 )