gpt_dna_eng

brief

gpt2 model trained by DNA sequence and Human language based on gpt2 small, 128 max token input

basic use

import torch
from transformers import GPT2Tokenizer,GPT2Model,GPT2LMHeadModel

model = GPT2LMHeadModel.from_pretrained("dnagpt/gpt_dna_eng")
tokenizer = GPT2Tokenizer.from_pretrained("dnagpt/dna_eng_bpe")

input = "hello TCTTTCTCTTCTGTAT"

input_ids = tokenizer.encode(
      input,
      return_tensors="pt",
      truncation=True,
      max_length=1000
)

# Generate
device = model.device

generated_tokens = model.generate(
input_ids=input_ids.to(device),
max_new_tokens=10,
)

generated_text = tokenizer.decode(generated_tokens[0], skip_special_tokens=True)

print(generated_text)

#hello TCTTTCTCTTCTGTATATCTATCTATCTATCTATCTATCTATCTATCTATCTATCTATCTATCTATCTATCTATCTATCTATCTATCTATCTATCTATC

base_model = GPT2Model.from_pretrained("dnagpt/gpt_dna_eng")
output = base_model(input_ids)
print(output)

github

https://github.com/maris205/DNAHL