metadata
dataset_info:
features:
- name: image
dtype: image
- name: path
dtype: string
- name: img_id
dtype: int64
- name: land_cover
dtype: string
- name: captions
sequence: string
- name: caption_ids
sequence: int64
splits:
- name: train
num_bytes: 460526263.218
num_examples: 8734
- name: val
num_bytes: 60256331.322
num_examples: 1094
download_size: 473790206
dataset_size: 520782594.53999996
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: val
path: data/val-*