test_time_scaling/get_response_utils.py

from openai import OpenAI
import os
import logging
import torch
import re
import os
import random
import transformers
from tqdm import tqdm
from transformers import (
    AutoTokenizer,
    AutoModelForCausalLM,
)

def setup_logger(log_path):
    logging.basicConfig(
        filename=log_path,
        filemode="w",
        format="%(asctime)s - %(levelname)s - %(message)s",
        level=logging.INFO
    )

ANS_RE = re.compile(r"#### (\-?[0-9\.\,]+)")
INVALID_ANS = "[invalid]"

ANSWER_TRIGGER = "The answer is"

# prompt="""Please solve the following problem carefully, showing all intermediate steps. At the end, write your final answer in the following format: "#### your_answer_here" where "your_answer_here" is the final numerical result.

# Problem:
# {question}

# Remember: Only the final answer should be placed after "####", and do not include any extra text after it."""

def extract_answer_from_output(completion):
    match = ANS_RE.search(completion)
    if match:
        match_str = match.group(1).strip()
        match_str = match_str.replace(",", "")
        return match_str
    else:
        return INVALID_ANS


def is_correct(model_answer, answer):
    gt_answer = extract_answer_from_output(answer)
    assert gt_answer != INVALID_ANS
    return model_answer == gt_answer
def is_correct_direct(model_answer, answer):
    return model_answer == answer
def is_correct_choice(model_answer, answer):
    return int(model_answer) == answer

# 本地vllm调用
def get_response_template(message,model="meta-llama/Meta-Llama-3-8B-Instruct",client=OpenAI(
        api_key="EMPTY",
        base_url="http://127.0.0.1:8422/v1",
    )):
    
    prompt= message
    chat_response = client.chat.completions.create(
    model=model,
    messages=[
            {"role": "system", "content": "You are a helpful assistant."},
            {"role": "user", "content": prompt},
        ],
    temperature = 0.8,
    )
    print("Chat response:", chat_response.choices[0].message.content)
    return chat_response.choices[0].message.content