---
license: apache-2.0
dataset_info:
  features:
  - name: smiles
    dtype: string
  - name: activity
    dtype: float64
  splits:
  - name: train
    num_bytes: 13736
    num_examples: 513
  - name: validation
    num_bytes: 1721
    num_examples: 64
  - name: test
    num_bytes: 1780
    num_examples: 65
  download_size: 15823
  dataset_size: 17237
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
  - split: test
    path: data/test-*
---
The dataset is created using data from the deepchem library introduced by moleculenet. 
The dataset is split in a random fashion and normalized using z_score. 
The random seed used for splitting is 123.
The smiles are canonicalized using RDkit.