π Technical report | π€ Bunny-v1.0-3B | π Code | π° Demo
Bunny is a family of lightweight multimodal models.
Bunny-v1.0-data is the training dataset for Bunny-v1.0 series, including Bunny-v1.0-3B.
We use a high-quality coreset with less duplicates and more informative samples of LAION-2B built by this work.
We randomly sample 2 million image-text pairs from the coreset and convert them to training format.
The pretraining data and⦠See the full description on the dataset page:
https://huggingface.co/datasets/BoyaWu10/Bunny-v1_0-data.