metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: loss_weight
sequence: float64
- name: aux_features
sequence:
sequence: uint8
- name: chromatin_features
sequence: float64
- name: flip_p
sequence: float64
splits:
- name: train
num_bytes: 1550800000
num_examples: 100000
download_size: 242922569
dataset_size: 1550800000
configs:
- config_name: default
data_files:
- split: train
path: data/train-*