metadata
dataset_info:
features:
- name: material_id
dtype: string
- name: numbers
sequence: int64
- name: positions
sequence:
sequence: float64
- name: cell
sequence:
sequence: float64
- name: pbc
sequence: bool
- name: num_atoms
dtype: int64
- name: composition
sequence: int64
- name: energy
dtype: float64
- name: forces
sequence:
sequence: float64
- name: stress
sequence:
sequence: float64
splits:
- name: train
num_bytes: 220552014352
num_examples: 111771592
download_size: 70745823829
dataset_size: 220552014352
configs:
- config_name: default
data_files:
- split: train
path: data/train-*