Skip to content

Instantly share code, notes, and snippets.

@FoobarProtocol
Created October 21, 2023 23:45
Show Gist options
  • Save FoobarProtocol/fd72079d619737fa8a190fa2a5836a80 to your computer and use it in GitHub Desktop.
Save FoobarProtocol/fd72079d619737fa8a190fa2a5836a80 to your computer and use it in GitHub Desktop.
This is the `main.py` from the Evol_Instruct repo that got removed by the researchers (this makes calls to the different transformative prompts; this is meant to be separated from the Auroboros technique)
import json
import random
from openai_access import call_chatgpt
from depth import createConstraintsPrompt, createDeepenPrompt, createConcretizingPrompt, createReasoningPrompt
from breadth import createBreadthPrompt
fr = open('alpaca_data.json','r')
all_objs = json.load(fr)
evol_objs = []
for cur_obj in all_objs:
instruction = cur_obj['instruction'].strip() + '\r\n'+ cur_obj['input'].strip()
evol_prompts = []
evol_prompts.append(createConstraintsPrompt(instruction))
evol_prompts.append(createDeepenPrompt(instruction))
evol_prompts.append(createConcretizingPrompt(instruction))
evol_prompts.append(createReasoningPrompt(instruction))
evol_prompts.append(createBreadthPrompt(instruction))
selected_evol_prompt = random.choice(evol_prompts)
evol_instruction = call_chatgpt(selected_evol_prompt)
answer = call_chatgpt(evol_instruction)
evol_objs.append({"instruction":evol_instruction,"output":answer})
with open('alpaca_data_evol.json', 'w') as f:
json.dump(evol_objs, f, indent=4)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment