Trident: tri-modal molecular representation learning with taxonomic annotations and local correspondence

Molecular representation learning with SMILES, text, HTA, and functional groups.

File Structure

├── main.py          # Main entry
├── config.py        # Configuration
├── dataset.py       # Data processing
├── models.py        # Model definitions
├── train.py         # Training logic
├── utils.py         # Utilities
├── data/            # Data directory
└── save_model/      # Model outputs

Environment Setup

# Create a new conda environment
conda create -n TRIDENT python=3.10

# Activate the environment
conda activate TRIDENT

# Install dependencies
pip install -r requirements.txt

Usage

# Basic training
CUDA_VISIBLE_DEVICES=0,1 python main.py

# Custom settings
CUDA_VISIBLE_DEVICES=0,1 python main.py --batch_size 32 --epochs 50

Inference - Getting Embeddings

Quick Start

cd inference/
python get_embeddings.py