metadata
dataset_info:
features:
- name: image
dtype: image
- name: file_path
dtype: string
- name: findings
dtype: string
- name: impression
dtype: string
- name: combined_text
dtype: string
- name: predicted_labels
dtype: string
- name: clip_vector
sequence: float32
- name: clip_vector_img
sequence: float32
- name: clip_vector_txt
sequence: float32
- name: clip_vector_joint
sequence: float32
splits:
- name: train
num_bytes: 14957576333.25
num_examples: 43598
- name: test
num_bytes: 9750523368.125
num_examples: 17911
download_size: 34477777012
dataset_size: 24708099701.375
configs:
- config_name: default
data_files:
- split: test
path: data/test-*
- split: train
path: data/train-*