This model can be used to generate a SMILES string from an input caption.
Example Usage
from transformers import T5Tokenizer, T5ForConditionalGeneration
tokenizer = T5Tokenizer.from_pretrained("laituan245/molt5-small-caption2smiles", model_max_length=512)
model = T5ForConditionalGeneration.from_pretrained('laituan245/molt5-small-caption2smiles')
input_text = 'The molecule is a monomethoxybenzene that is 2-methoxyphenol substituted by a hydroxymethyl group at position 4. It has a role as a plant metabolite. It is a member of guaiacols and a member of benzyl alcohols.'
input_ids = tokenizer(input_text, return_tensors="pt").input_ids
outputs = model.generate(input_ids, num_beams=5, max_length=512)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))
# The model will generate "COC1=C(C=CC(=C1)CCCO)O". The ground-truth is "COC1=C(C=CC(=C1)CO)O".
Paper
For more information, please take a look at our paper.
Paper: Translation between Molecules and Natural Language
Authors: Carl Edwards*, Tuan Lai*, Kevin Ros, Garrett Honke, Heng Ji