Generate python

from transformers import AutoTokenizer, AutoModelWithLMHead


tokenizer = AutoTokenizer.from_pretrained("Sentdex/GPyT")
model = AutoModelWithLMHead.from_pretrained("Sentdex/GPyT").to("cuda")


def generate(code, max_length=100):
    '''Takes input code, replaces newline chars with <N>, 
    tokenizes, feeds thru model, decodes, 
    then reformats the newlines back in'''
    newlinechar = "<N>"
    converted = code.replace("\n", newlinechar)
    tokenized = tokenizer.encode(converted, return_tensors='pt').to("cuda")
    resp = model.generate(tokenized, max_length=max_length).to("cuda")

    decoded = tokenizer.decode(resp[0])
    reformatted = decoded.replace("<N>","\n")
    return reformatted


print(generate("import"))

Source: pythonprogramming.net

Similar codes

8 votes139 views4 answers
Reverse dictionary python
Python
3 votes157 views1 answer
Get common elements from two lists
Python
0 vote174 views2 answers
Python object name
Python
0 vote216 views1 answer
How to send file in django response
Python Django
3 votes192 views1 answer
How do you print a integer in python
Python
0 vote181 views1 answer
Pygame kreis definition
Python
0 vote185 views1 answer
Run python script from repl
Python
1 vote166 views1 answer
From time import sleep, time
Python
0 vote160 views1 answer
Cbind arrays python
Python
0 vote185 views1 answer
Numpy linspace of dates
Python