Edit model card

The pretrained model for the paper "Learning Molecular Representation in a Cell".

The model is trained with the following configurations:

  • emb_dim: 300
  • model: gin-virtual
  • norm_layer: batch_norm
  • num_layer: 5
  • prior: 1.0e-09
  • readout: sum
  • threshold: 0.8
  • walk_length: 4

Please place the config.YAML file in the ckpt folder and rename it to pretrain.yaml.

Downloads last month
0
Unable to determine this model's library. Check the docs .

Dataset used to train liuganghuggingface/InfoAlign-Pretrained