metadata
dataset_info:
features:
- name: smiles
dtype: large_string
- name: zinc_id
dtype: int64
- name: SELFIES
dtype: string
splits:
- name: train
num_bytes: 393170565049
num_examples: 1538340669
- name: val
num_bytes: 47753116448
num_examples: 192292584
- name: test
num_bytes: 46114402425
num_examples: 192292584
download_size: 174349539018
dataset_size: 487038083922
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: val
path: data/val-*
- split: test
path: data/test-*
license: mit
tags:
- chemistry
- biology
- medical
size_categories:
- 1B<n<10B
ZINC20 Dataset with SELFIES added. Any smile that could not be successfully converted was dropped from the dataset.
Every tranch was downloaded, this is not the ~1B example ML subset from https://files.docking.org/zinc20-ML/.
The dataset was entirely shuffled then split into 80%/10%/10% splits for train/val/test.
A file vocab.csv is in the root of the reposity that contains all of the SELFIES tokens found in the data, with [START], [STOP], and [PAD] added.