|
""" |
|
Main application for Dynamic Highscores system. |
|
This file integrates all components into a unified application. |
|
""" |
|
|
|
import os |
|
import gradio as gr |
|
import threading |
|
import time |
|
from database_schema import DynamicHighscoresDB |
|
from auth import HuggingFaceAuth, create_login_ui, setup_auth_handlers |
|
from benchmark_selection import BenchmarkSelector, create_benchmark_selection_ui |
|
from evaluation_queue import EvaluationQueue, create_model_submission_ui |
|
from leaderboard import Leaderboard, create_leaderboard_ui |
|
from sample_benchmarks import add_sample_benchmarks |
|
|
|
|
|
db = DynamicHighscoresDB() |
|
auth_manager = HuggingFaceAuth(db) |
|
benchmark_selector = BenchmarkSelector(db, auth_manager) |
|
evaluation_queue = EvaluationQueue(db, auth_manager) |
|
leaderboard = Leaderboard(db) |
|
|
|
|
|
print("Checking for existing benchmarks...") |
|
benchmarks = db.get_benchmarks() |
|
if not benchmarks or len(benchmarks) == 0: |
|
print("No benchmarks found. Adding sample benchmarks...") |
|
try: |
|
num_added = add_sample_benchmarks() |
|
print(f"Added {num_added} sample benchmarks.") |
|
except Exception as e: |
|
print(f"Error adding sample benchmarks: {str(e)}") |
|
|
|
|
|
css = """ |
|
.info-text { |
|
background-color: rgba(53, 130, 220, 0.1); |
|
padding: 12px; |
|
border-radius: 8px; |
|
border-left: 4px solid #3498db; |
|
margin: 12px 0; |
|
} |
|
""" |
|
|
|
|
|
with gr.Blocks(css=css, title="Dynamic Highscores") as app: |
|
gr.Markdown("# π Dynamic Highscores", elem_classes=["header"]) |
|
gr.Markdown(""" |
|
Welcome to Dynamic Highscores - a community benchmark platform for evaluating and comparing language models. |
|
|
|
- **Add your own benchmarks** from HuggingFace datasets |
|
- **Submit your models** for CPU-only evaluation |
|
- **Compare performance** across different models and benchmarks |
|
- **Filter results** by model type (Merge, Agent, Reasoning, Coding, etc.) |
|
""", elem_classes=["info-text"]) |
|
|
|
|
|
with gr.Tabs() as tabs: |
|
with gr.TabItem("π Leaderboard", id=0): |
|
leaderboard_ui = create_leaderboard_ui(leaderboard, db) |
|
|
|
with gr.TabItem("π Submit Model", id=1): |
|
submission_ui = create_model_submission_ui(evaluation_queue, auth_manager, db) |
|
|
|
with gr.TabItem("π Benchmarks", id=2): |
|
benchmark_ui = create_benchmark_selection_ui(benchmark_selector, auth_manager) |
|
|
|
|
|
if __name__ == "__main__": |
|
app.launch() |