Download app.py from livecodebench/code_generation_samples: direct link, hf CLI and curl.
- Browser
- Download file 4.03 kB
-
https://huggingface.co/spaces/livecodebench/code_generation_samples/resolve/main/app.py
- Command line
-
hf download hf://spaces/livecodebench/code_generation_samples/app.py
-
curl -L -o app.py https://huggingface.co/spaces/livecodebench/code_generation_samples/resolve/main/app.py
4.03 kB
| import os | |
| import random | |
| import glob | |
| import json | |
| import numpy as np | |
| from flask import Flask, render_template, request | |
| app = Flask(__name__) | |
| with open("problems.json") as f: | |
| problems = json.load(f) | |
| problem_choices = [q["question_title"] for q in problems] | |
| random_idxs = list(range(len(problems))) | |
| # random.seed(42) | |
| # random.shuffle(random_idxs) | |
| problems = [problems[idx] for idx in random_idxs] | |
| with open("all_outputs.json") as f: | |
| all_outputs = json.load(f) | |
| all_models = list(all_outputs.keys()) | |
| num_questions_filtered = len(problems) | |
| all_correctness_by_problem = { | |
| idx: {model: np.mean(all_outputs[model][idx]["pass1_list"]) for model in all_models} | |
| for idx in random_idxs | |
| } | |
| def calculate_color(performance): | |
| # Convert performance to a value between 0 and 1 | |
| # Calculate the red and green components of the color | |
| if performance > 0.75: | |
| return f"rgba(0, 150, 0, 0.5)" | |
| elif performance > 0.5: | |
| return f"rgba(50, 150, 0, {performance})" | |
| elif performance > 0.25: | |
| return f"rgba(150, 50, 0, {1-performance})" | |
| else: | |
| return f"rgba(150, 0, 0, 0.5)" | |
| all_evaluations_by_problem_colored = [ | |
| ( | |
| trueidx, | |
| { | |
| model: { | |
| "correctness": f"{all_correctness_by_problem[idx][model]*100:.1f}", | |
| "correctness_color": calculate_color( | |
| all_correctness_by_problem[idx][model] | |
| ), | |
| } | |
| for model in all_models | |
| }, | |
| problems[idx]["difficulty"], | |
| problems[idx]["question_id"], | |
| ) | |
| for trueidx, idx in enumerate(random_idxs) | |
| ] | |
| all_data_for_view_formatted = { | |
| model: [ | |
| [ | |
| {"code": a, "pass1": b, "metadata": c} | |
| for a, b, c in zip( | |
| row["code_list"], row["pass1_list"], row["metadata_list"] | |
| ) | |
| ] | |
| # print(row) | |
| for idx in random_idxs | |
| for row in [resp[idx]] | |
| ] | |
| for model, resp in all_outputs.items() | |
| } | |
| def home(): | |
| # Fetch your data here | |
| print(all_models) | |
| return render_template( | |
| "index.html", models=all_models, problems=all_evaluations_by_problem_colored | |
| ) | |
| def problem(problem_idx): | |
| # Fetch your data here | |
| data = { | |
| model: all_data_for_view_formatted[model][problem_idx] for model in all_models | |
| } | |
| evaluation = all_evaluations_by_problem_colored[problem_idx][1] | |
| question = problems[problem_idx] | |
| # print(data) | |
| return render_template( | |
| "problem.html", | |
| problem_idx=problem_idx, | |
| question_id=all_evaluations_by_problem_colored[problem_idx][3], | |
| evaluation=evaluation, | |
| models=all_models, | |
| question=question, | |
| data=data, | |
| ) | |
| mini_models = [ | |
| # "DeepSeek-V2", | |
| "DeepSeek-V3", | |
| "DeepSeek-R1-Preview", | |
| # "DSCoder-33b-Ins", | |
| # "GPT-4-Turbo-2024-04-09", | |
| "GPT-4O-2024-05-13", | |
| "Claude-3.5-Sonnet-20240620", | |
| "Gemini-Flash-2.0-Thinking", | |
| # "Gemini-Exp-1206", | |
| # "Claude-3-Sonnet", | |
| "O1-2024-12-17 (N=1) (High)", | |
| "QwQ-32B-Preview (N=1)", | |
| ] | |
| def mini(): | |
| # Fetch your data here | |
| return render_template( | |
| "index_mini.html", | |
| models=mini_models, | |
| problems=all_evaluations_by_problem_colored, | |
| ) | |
| def problem_mini(problem_idx): | |
| # Fetch your data here | |
| data = { | |
| model: all_data_for_view_formatted[model][problem_idx] for model in mini_models | |
| } | |
| evaluation = all_evaluations_by_problem_colored[problem_idx][1] | |
| question = problems[problem_idx] | |
| # print(data) | |
| return render_template( | |
| "problem_mini.html", | |
| problem_idx=problem_idx, | |
| question_id=all_evaluations_by_problem_colored[problem_idx][3], | |
| evaluation=evaluation, | |
| models=mini_models, | |
| question=question, | |
| data=data, | |
| ) | |
| if __name__ == "__main__": | |
| app.run(host="0.0.0.0", port=7860) | |