ord-training-simple / src /__init__.py
Vaishnav14220
Orchestrate full ORD training pipeline in Space
29a351f
raw
history blame contribute delete
650 Bytes
"""
ORD Reaction Prediction Pipeline
A complete end-to-end pipeline for training AI models on the Open Reaction Database (ORD)
for forward synthesis and retrosynthesis prediction.
Modules:
- config: Configuration constants
- utils: Utility functions for SMILES processing
- dataset_prepare: Data streaming and preprocessing
- tokenizer_train: BPE tokenizer training
- data_utils: Tokenization and loading utilities
- train_forward: Forward synthesis model training
- train_retro: Retrosynthesis model training
- evaluate: Model evaluation
"""
__version__ = "1.0.0"
__author__ = "ORD Reaction Pipeline Contributors"