BioRGM

Issues

• Introduce model evaluation
• Load Model from checkpoints or initialize random model from jupyter notebook for evaluation
• Integrate wandb
• Implement checkpoints
• Track triplets generated
• Combine EmbeddingModel and GINModel

Backlog

• Track augmentations performed for each augmentation
• Parallelize creation of InMemory PubchemDataset
• Add another 50 augmentations
• Configure logging
• Set up pytest
• InMemoryDataset should not create "pre-filter.pt" and "pre_transform.pt" files.