MultiMolecule How to use multimolecule/dnabert2 with MultiMolecule:
pip install multimolecule
from multimolecule import AutoModel, AutoTokenizer
tokenizer = AutoTokenizer.from_pretrained("multimolecule/dnabert2")
model = AutoModel.from_pretrained("multimolecule/dnabert2")
inputs = tokenizer("ACTCCCCTGCCCTCAACAAGATGTTTTGCCAACTGGCCAAGACCTGCCCTGTGCAGCTGTGGGTTGATTCCACACCCCCGCCCGGCACCCGCGTCCGCGCCATGGCCATCTACAAGCAGTCACAGCACATGACGGAGGTTGTGAGGCGCTGCCCCCACCATGAGCGCTGCTCAGATAGCGATGG", return_tensors="pt")
outputs = model(**inputs)
embeddings = outputs.last_hidden_state import multimolecule
from transformers import pipeline
predictor = pipeline("fill-mask", model="multimolecule/dnabert2")
output = predictor("ACTCCCCTGCCCTC<mask>ACAAGATGTTTTGCCAACTGGCCAAGACCTGCCCTGTGCAGCTGTGGGTTGATTCCACACCCCCGCCCGGCACCCGCGTCCGCGCCATGGCCATCTACAAGCAGTCACAGCACATGACGGAGGTTGTGAGGCGCTGCCCCCACCATGAGCGCTGCTCAGATAGCGATGG") 
