dataset_info: | |
features: | |
- name: images | |
sequence: image | |
- name: texts | |
list: | |
- name: user | |
dtype: string | |
- name: assistant | |
dtype: string | |
- name: source | |
dtype: string | |
splits: | |
- name: train | |
num_bytes: 244951255658.16818 | |
num_examples: 565009 | |
download_size: 145422811605 | |
dataset_size: 244951255658.16818 | |
configs: | |
- config_name: default | |
data_files: | |
- split: train | |
path: data/train-* | |
[Docmatix](https://huggingface.co/datasets/HuggingFaceM4/Docmatix), but with multi-image samples filtered out. |