---
dataset_info:
  features:
  - name: input
    dtype: string
  - name: mc_labels
    sequence: float64
  splits:
  - name: train
    num_bytes: 66709931019
    num_examples: 315151780
  - name: test
    num_bytes: 11625286787
    num_examples: 54872514
  download_size: 17338800261
  dataset_size: 78335217806
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: test
    path: data/test-*
---


Sequential Attachment-based Fragment Embedding (SAFE) is a novel molecular line notation that represents molecules as an unordered sequence of fragment blocks to improve molecule design using generative models.

Find the details and how to use at SAFE in the repo https://github.com/datamol-io/safe or the paper https://arxiv.org/pdf/2310.10773.pdf.

This dataset is a subset of the original SAFE dataset, made for the purpose of training smaller models.