Update README.md

README.md

@@ -11,4 +11,6 @@ size_categories:
 - 1M<n<10M
 ---
 
-This repo contains the vision-language data used in the multimodal pre-training of BitVLA. Please refer to our [paper](https://arxiv.org/abs/2506.07530), [implementation](https://github.com/ustcwhy/BitVLA) and [model weights](https://huggingface.co/collections/hongyuw/bitvla-68468fb1e3aae15dd8a4e36e) for more details.
+This repo contains the vision-language data used in the multimodal pre-training of BitVLA. You can download the images from [MAmmoTH-VL](https://huggingface.co/datasets/MAmmoTH-VL/MAmmoTH-VL-Instruct-12M). 
+
+Please refer to our [paper](https://arxiv.org/abs/2506.07530), [implementation](https://github.com/ustcwhy/BitVLA) and [model weights](https://huggingface.co/collections/hongyuw/bitvla-68468fb1e3aae15dd8a4e36e) for more details.