1H-NMR model trainig checkpoints based on https://github.com/rxn4chemistry/multimodal-spectroscopic-dataset

These are .pt checkpoints generated by OpenNMT during training of the 1H-NMR - to - SMILES model roughly as defined in this repository [no affiliation]. Once the Python environment is activated you should be able to use the checkpoints as follows. Tested with OpenNMT-py==3.5.1. onmt_translate \ --model model_step_110000.pt \ --src src-test_samples.txt \ --output predictions.txt \ --gpu -1 \ --verbose Further documentation is forthcoming. Note that licenses from the cited repository might apply. Training took ~40h on an NVIDIA A10. Many thanks to the generous support of Lambda "On-Demand Cloud".