utils.py

import pandas as pd
import gradio as gr
import csv
import json
import os
import shutil
from huggingface_hub import Repository

HF_TOKEN = os.environ.get("HF_TOKEN")

SUBJECTS = ["Biology", "Business", "Chemistry", "Computer Science", "Economics", "Engineering",
            "Health", "History", "Law", "Math", "Philosophy", "Physics", "Psychology", "Other"]

MODEL_INFO = [
    "Models",
    "Overall",
    "Biology", "Business", "Chemistry", "Computer Science", "Economics", "Engineering",
    "Health", "History", "Law", "Math", "Philosophy", "Physics", "Psychology", "Other"]

DATA_TITLE_TYPE = ['markdown', 'number', 'number', 'number', 'number', 'number', 'number',
                   'number', 'number', 'number', 'number', 'number', 'number', 'number',
                   'number', 'number']

SUBMISSION_NAME = "mmlu_pro_leaderboard_submission"
SUBMISSION_URL = os.path.join("https://huggingface.co/datasets/TIGER-Lab/", SUBMISSION_NAME)
CSV_DIR = "./mmlu_pro_leaderboard_submission/results.csv"

COLUMN_NAMES = MODEL_INFO

LEADERBOARD_INTRODUCTION = """# MMLU-Pro Leaderboard

Welcome to the MMLU-Pro leaderboard, showcasing the performance of various advanced language models on the MMLU-Pro dataset. The MMLU-Pro dataset is an enhanced version of the original MMLU, specifically engineered to offer a more rigorous and realistic evaluation environment..

The MMLU-Pro dataset consists of approximately 12,000 intricate questions that challenge the comprehension and reasoning abilities of LLMs. Below you can find the accuracies of different models tested on this dataset.

## 1. What's new about MMLU-Pro

Compared to the original MMLU, there are three major differences:

- The original MMLU dataset only contains 4 options, MMLU-Pro increases it to 10 options. The increase in options will make the evaluation more realistic and challenging. The random guessing will lead to a much lower score.
- The original MMLU dataset contains mostly knowledge-driven questions without requiring much reasoning. Therefore, PPL results are normally better than CoT. In our dataset, we increase the problem difficulty and integrate more reasoning-focused problems. In MMLU-Pro, CoT can be 20% higher than PPL. 
- Due to the increase of options, we found that the model performance becomes more robust. For example, Llama-2-7B performance variance on MMLU-Pro is within 1% with several different prompts. In contrast, the performance variance on original MMLU can be as huge as 4-5%.

## 2. Dataset Summary

- **Questions and Options:** Each question within the dataset typically has **ten** multiple-choice options, except for some that were reduced during the manual review process to remove unreasonable choices. This increase from the original **four** options per question is designed to enhance complexity and robustness, necessitating deeper reasoning to discern the correct answer among a larger pool of potential distractors.
  
- **Sources:** The dataset consolidates questions from several sources:
  - **Original MMLU Questions:** Part of the dataset is coming from the original MMLU dataset. We remove the trivial and ambiguous questions.
  - **STEM Website:** Hand picking high-quality STEM problems from the Internet.
  - **TheoremQA:** High-quality human-annotated questions requiring theorems to solve.
  - **Scibench:** Science questions from college exams.

For detailed information about the dataset, visit our page on Hugging Face: MMLU-Pro at Hugging Face. If you are interested in replicating these results or wish to evaluate your models using our dataset, access our evaluation scripts available on GitHub: TIGER-AI-Lab/MMLU-Pro.
"""

TABLE_INTRODUCTION = """
    """

LEADERBOARD_INFO = """
We list the information of the used datasets as follows:<br>

"""

CITATION_BUTTON_LABEL = "Copy the following snippet to cite these results"
CITATION_BUTTON_TEXT = r""""""

SUBMIT_INTRODUCTION = """# Submit on Science Leaderboard Introduction

## ⚠ Please note that you need to submit the json file with following format:

```json
{
    "Model": "[MODEL_NAME]",
    "Overall": 0.5678,
    "Biology": 0.1234, 
    "Business": 0.4567,
    ...,
    "Other: 0.3456"
}
```
After submitting, you can click the "Refresh" button to see the updated leaderboard (it may takes few seconds).

"""


def get_df():
    repo = Repository(local_dir=SUBMISSION_NAME, clone_from=SUBMISSION_URL, use_auth_token=HF_TOKEN)
    repo.git_pull()
    df = pd.read_csv(CSV_DIR)
    df = df.sort_values(by=['Overall'], ascending=False)
    return df[COLUMN_NAMES]


def add_new_eval(
    input_file,
):
    if input_file is None:
        return "Error! Empty file!"

    upload_data = json.loads(input_file)
    print("upload_data:\n", upload_data)
    data_row = [f'{upload_data["Model"]}', upload_data['Overall']]
    for subject in SUBJECTS:
        data_row += [upload_data[subject]]
    print("data_row:\n", data_row)
    submission_repo = Repository(local_dir=SUBMISSION_NAME, clone_from=SUBMISSION_URL,
                                 use_auth_token=HF_TOKEN, repo_type="dataset")
    submission_repo.git_pull()

    already_submitted = []
    with open(CSV_DIR, mode='r') as file:
        reader = csv.reader(file, delimiter=',')
        for row in reader:
            already_submitted.append(row[0])

    if data_row[0] not in already_submitted:
        with open(CSV_DIR, mode='a', newline='') as file:
            writer = csv.writer(file)
            writer.writerow(data_row)
        
        submission_repo.push_to_hub()
        print('Submission Successful')
    else:
        print('The entry already exists')


def refresh_data():
    return get_df()