|
import openai |
|
import argparse |
|
import os |
|
from cliport import tasks |
|
from cliport.dataset import RavensDataset |
|
from cliport.environments.environment import Environment |
|
|
|
from pygments import highlight |
|
from pygments.lexers import PythonLexer |
|
from pygments.formatters import TerminalFormatter |
|
|
|
import time |
|
import random |
|
import json |
|
|
|
from gensim.utils import set_gpt_model, clear_messages, format_finetune_prompt |
|
|
|
|
|
|
|
if __name__ == "__main__": |
|
parser = argparse.ArgumentParser() |
|
parser.add_argument("--task", type=str, default='build-car') |
|
parser.add_argument("--model", type=str, default='davinci:ft-wang-lab:gensim-2023-08-05-16-54-05') |
|
|
|
args = parser.parse_args() |
|
task = args.task |
|
prompt = format_finetune_prompt(task) |
|
|
|
if True: |
|
response = openai.Completion.create( |
|
model=args.model, |
|
prompt=prompt, |
|
temperature=0, |
|
max_tokens=1024) |
|
res = response["choices"][0]["text"] |
|
else: |
|
params = { |
|
"model": args.model, |
|
"max_tokens": 500, |
|
"temperature": 0.1, |
|
"messages": [prompt] |
|
} |
|
call_res = openai.ChatCompletion.create(**params) |
|
res = call_res["choices"][0]["message"]["content"] |
|
|
|
print("code!:", res) |
|
python_file_path = f"cliport/generated_tasks/finetune_{task.replace('-','_')}.py" |
|
print(f"saving task {args.task} to {python_file_path}") |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
