metadata
dataset_info:
features:
- name: file_name
dtype: string
- name: im_path
dtype: string
- name: caption
dtype: string
- name: bbox
dtype: string
- name: prompt
dtype: string
- name: texts
dtype: string
- name: sentence
dtype: string
- name: outp_org
sequence: float64
- name: formatted_sentence
dtype: string
- name: modified_sentence
dtype: string
- name: target
sequence: float64
splits:
- name: train
num_bytes: 6038086.5
num_examples: 2250
- name: test
num_bytes: 670898.5
num_examples: 250
download_size: 1103214
dataset_size: 6708985
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*