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: int64
- name: formatted_sentence
dtype: string
- name: modified_sentence
dtype: string
splits:
- name: train
num_bytes: 2064254
num_examples: 3770
download_size: 747061
dataset_size: 2064254
configs:
- config_name: default
data_files:
- split: train
path: data/train-*