metadata
dataset_info:
features:
- name: query
dtype: string
- name: source
dtype: string
- name: image_filename
dtype: string
- name: negative_passages
sequence: int64
- name: negatives
sequence: string
- name: positive_passages
sequence: int64
- name: gold_in_top_100
dtype: bool
- name: language
dtype: string
splits:
- name: VisRAG_Ret_Train_Synthetic_data_ICML_Papers
num_bytes: 100436695
num_examples: 29824
- name: VisRAG_Ret_Train_Synthetic_data_Manuallib
num_bytes: 382494378
num_examples: 119934
- name: VisRAG_Ret_Train_Synthetic_data_NeurIPS_Papers
num_bytes: 110097215
num_examples: 29952
- name: VisRAG_Ret_Train_Synthetic_data_Textbooks
num_bytes: 189278144
num_examples: 59648
download_size: 244671188
dataset_size: 782306432
configs:
- config_name: default
data_files:
- split: VisRAG_Ret_Train_Synthetic_data_ICML_Papers
path: data/VisRAG_Ret_Train_Synthetic_data_ICML_Papers-*
- split: VisRAG_Ret_Train_Synthetic_data_Manuallib
path: data/VisRAG_Ret_Train_Synthetic_data_Manuallib-*
- split: VisRAG_Ret_Train_Synthetic_data_NeurIPS_Papers
path: data/VisRAG_Ret_Train_Synthetic_data_NeurIPS_Papers-*
- split: VisRAG_Ret_Train_Synthetic_data_Textbooks
path: data/VisRAG_Ret_Train_Synthetic_data_Textbooks-*