README.md

metadata

dataset_info:
  features:
    - name: smiles
      dtype: string
    - name: instructions
      dtype: string
    - name: descriptions
      dtype: string
    - name: image
      dtype: image
  splits:
    - name: train
      num_bytes: 187676711
      num_examples: 18000
  download_size: 179877348
  dataset_size: 187676711
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*

This dataset repository contains data for fine-tuning a visual language model (VLM) based on molecular structure images.