app.py

import gradio as gr
import pandas as pd
import plotly.express as px
import plotly.graph_objects as go
from huggingface_hub import HfApi
from datetime import datetime
import numpy as np

def format_number(num):
    """Format large numbers with K, M suffix"""
    if num >= 1e6:
        return f"{num/1e6:.1f}M"
    elif num >= 1e3:
        return f"{num/1e3:.1f}K"
    return str(num)

def fetch_stats():
    """Fetch all Meta Llama 3 model statistics"""
    api = HfApi()
    
    # Fetch original models
    original_models = [
        "meta-llama/Llama-3.3-70B-Instruct",
        "meta-llama/Meta-Llama-3-70B-Instruct",
        "meta-llama/Llama-3.1-70B-Instruct",
        "meta-llama/Llama-3.1-405B-FP8",
        "meta-llama/Llama-3.2-90B-Vision-Instruct",
        "meta-llama/Llama-3.2-11B-Vision-Instruct",
        "meta-llama/Llama-3.2-3B-Instruct-QLORA_INT4_EO8",
        "meta-llama/Llama-3.2-3B-Instruct-SpinQuant_INT4_EO8",
        "meta-llama/Llama-3.2-1B-Instruct-SpinQuant_INT4_EO8",
        "meta-llama/Llama-3.2-1B-Instruct-QLORA_INT4_EO8",
        "meta-llama/Llama-Guard-3-11B-Vision",
        "meta-llama/Llama-3.2-1B",
        "meta-llama/Llama-3.2-1B-Instruct",
        "meta-llama/Llama-3.2-3B",
        "meta-llama/Llama-3.2-3B-Instruct",
        "meta-llama/Llama-3.1-8B",
        "meta-llama/Llama-Guard-3-8B",
        "meta-llama/Meta-Llama-3-70B",
        "meta-llama/Meta-Llama-3-8B-Instruct",
        "meta-llama/Meta-Llama-3-8B",
        "meta-llama/Llama-3.2-90B-Vision",
        "meta-llama/Llama-3.2-11B-Vision",
        "meta-llama/Llama-Guard-3-1B",
        "meta-llama/Llama-Guard-3-1B-INT4",
        "meta-llama/Llama-3.1-405B-Instruct-FP8",
        "meta-llama/Llama-3.1-405B-Instruct",
        "meta-llama/Llama-3.1-405B",
        "meta-llama/Llama-3.1-70B",
        "meta-llama/Llama-3.1-8B-Instruct",
        "meta-llama/Llama-Guard-3-8B-INT8"
    ]
    
    original_stats = []
    for model_id in original_models:
        try:
            info = api.model_info(model_id)
            original_stats.append({
                'model_id': model_id,
                'downloads_30d': info.downloads if hasattr(info, 'downloads') else 0,
                'likes': info.likes if hasattr(info, 'likes') else 0
            })
        except Exception as e:
            print(f"Error fetching {model_id}: {str(e)}")
    
    model_types = ["adapter", "finetune", "merge", "quantized"]
    base_models = [
        "Llama-3.3-70B-Instruct",
        "Meta-Llama-3-70B-Instruct",
        "Llama-3.1-70B-Instruct",
        "Llama-3.1-405B-FP8",
        "Llama-3.2-90B-Vision-Instruct",
        "Llama-3.2-11B-Vision-Instruct",
        "Llama-3.2-3B-Instruct-QLORA_INT4_EO8",
        "Llama-3.2-3B-Instruct-SpinQuant_INT4_EO8",
        "Llama-3.2-1B-Instruct-SpinQuant_INT4_EO8",
        "Llama-3.2-1B-Instruct-QLORA_INT4_EO8",
        "Llama-Guard-3-11B-Vision",
        "Llama-3.2-1B",
        "Llama-3.2-1B-Instruct",
        "Llama-3.2-3B",
        "Llama-3.2-3B-Instruct",
        "Llama-3.1-8B",
        "Llama-Guard-3-8B",
        "Meta-Llama-3-70B",
        "Meta-Llama-3-8B-Instruct",
        "Meta-Llama-3-8B",
        "Llama-3.2-90B-Vision",
        "Llama-3.2-11B-Vision",
        "Llama-Guard-3-1B",
        "Llama-Guard-3-1B-INT4",
        "Llama-3.1-405B-Instruct-FP8",
        "Llama-3.1-405B-Instruct",
        "Llama-3.1-405B",
        "Llama-3.1-70B",
        "Llama-3.1-8B-Instruct",
        "Llama-Guard-3-8B-INT8"
    ]
    
    derivative_stats = []
    
    for base_model in base_models:
        for model_type in model_types:
            try:
                models = list(api.list_models(
                    filter=f"base_model:{model_type}:meta-llama/{base_model}",
                    full=True
                ))
                
                # Add each model to our stats
                for model in models:
                    derivative_stats.append({
                        'base_model': f"meta-llama/{base_model}",
                        'model_type': model_type,
                        'model_id': model.id,
                        'downloads_30d': model.downloads if hasattr(model, 'downloads') else 0,
                        'likes': model.likes if hasattr(model, 'likes') else 0
                    })
            except Exception as e:
                print(f"Error fetching {model_type} models for {base_model}: {str(e)}")
    
    # Create DataFrames
    original_df = pd.DataFrame(original_stats, columns=['model_id', 'downloads_30d', 'likes'])
    derivative_df = pd.DataFrame(derivative_stats, columns=['base_model', 'model_type', 'model_id', 'downloads_30d', 'likes'])
    
    return original_df, derivative_df

def create_stats_html():
    """Create HTML for displaying statistics"""
    original_df, derivative_df = fetch_stats()
    
    # Create summary statistics
    total_originals = len(original_df)
    total_derivatives = len(derivative_df)
    total_downloads_orig = original_df['downloads_30d'].sum()
    total_downloads_deriv = derivative_df['downloads_30d'].sum()
    
    # Create derivative type distribution chart
    if len(derivative_df) > 0:
        # Create distribution by model type
        type_dist = derivative_df.groupby('model_type').agg({
            'model_id': 'count',
            'downloads_30d': 'sum'
        }).reset_index()
        
        type_dist = derivative_df.groupby('model_type').agg({
            'model_id': 'count',
            'downloads_30d': 'sum'
        }).reset_index()

        type_dist['model_type'] = type_dist['model_type'].str.capitalize()

        type_dist = type_dist.sort_values('downloads_30d', ascending=True)

        fig_types = go.Figure(data=[
            go.Bar(
                x=list(type_dist['model_type']),  # Convert to list
                y=list(type_dist['downloads_30d'].values),  # Convert series to list of values
                marker_color='rgb(55, 83, 109)'
            )
        ])

        fig_types.update_layout(
            title='Downloads by Model Type',
            #xaxis_title='Model Type',
            yaxis_title='Downloads',
            plot_bgcolor='white',
            showlegend=False,
            bargap=0.3
        )

        fig_types.update_traces(
            text=type_dist['downloads_30d'].apply(format_number),
            textposition='outside'
        )
    else:
        fig_types = px.bar(title='No data available')
    
    if len(derivative_df) > 0:
        top_models = derivative_df.nlargest(10, 'downloads_30d')[
            ['model_id', 'model_type', 'downloads_30d', 'likes']
        ].copy()  # Create a copy to avoid SettingWithCopyWarning
        
        # Capitalize model types in the table
        top_models['model_type'] = top_models['model_type'].str.capitalize()
        
        # Format download numbers
        top_models['downloads_30d'] = top_models['downloads_30d'].apply(format_number)
        
        # Create clickable links for model_id
        top_models['model_id'] = top_models['model_id'].apply(
            lambda x: f'<a href="https://huggingface.co/{x}" target="_blank" onclick="window.open(\'https://huggingface.co/{x}\', \'_blank\')">{x}</a>'
        )
    else:
        top_models = pd.DataFrame(columns=['model_id', 'model_type', 'downloads_30d', 'likes'])

    summary_html = f"""
    <div style='padding: 20px; background-color: #f5f5f5; border-radius: 10px; margin-bottom: 20px;'>
        <h3 style='color: #333333;'>Summary Statistics</h3>
        <p style='color: #333333;'>Derivative Models Downloads: {format_number(total_downloads_deriv)} ({total_derivatives} models)</p>
        <p style='color: #333333;'>Original Models Downloads: {format_number(total_downloads_orig)} ({total_originals} models)</p>
        <p style='color: #333333;'>Last Updated: {datetime.now().strftime('%Y-%m-%d %H:%M UTC')}</p>
        <p style='color: #333333; font-size: 10px;'>* Last 30 days downloads. Includes models from the 3, 3.1, 3.2, and 3.3 versions</p>
    </div>
    """
    
    return summary_html, fig_types, top_models

def create_interface():
    """Create Gradio interface"""
    with gr.Blocks(theme=gr.themes.Soft()) as interface:
        gr.HTML("<h1 style='text-align: center;'>Meta Llama3 Models Stats</h1>")

        with gr.Row():    
            with gr.Column():
                summary_html = gr.HTML()
            with gr.Column():
                plot = gr.Plot()

        with gr.Row():
            table = gr.DataFrame(
                headers=["Model ID", "Type", "Downloads (30d)", "Likes"],
                label="Top 10 Most Downloaded Models",
                wrap=True,
                datatype=["html", "str", "str", "number"]
            )
        
        def update_stats():
            summary, fig, top_models = create_stats_html()
            return summary, fig, top_models
        
        interface.load(update_stats, 
                      outputs=[summary_html, plot, table])
        
    return interface

# Create and launch the interface
demo = create_interface()
demo.launch()