commit

benchmark_results.csv

@@ -0,0 +1,10 @@
+model,armenian_language_score,armenian_history_score,mathematics_score,average_score
+claude-3-7-sonnet-20250219,10.5,7.75,15.0,11.08
+claude-3-5-sonnet-20241022,10.0,9.25,12.75,10.67
+gemini-2.0-flash,5.5,6.75,17.25,9.83
+gpt-4o,6.75,6.75,13.25,8.92
+qwen-max-2025-01-25,7.25,4.5,14.25,8.67
+gemini-1.5-flash,4.75,3.75,15.0,7.83
+deepseek-ai/DeepSeek-V3,5.25,5.0,12.25,7.5
+Meta-Llama-3.3-70B-Instruct,4.5,5.25,11.5,7.08
+claude-3-5-haiku-20241022,5.0,3.75,10.75,6.5

main.py

@@ -0,0 +1,45 @@
+import gradio as gr
+import pandas as pd
+import plotly.express as px
+
+def display_table():
+    df = pd.read_csv('benchmark_results.csv')
+    return df
+
+def create_bar_chart():
+    df = pd.read_csv('benchmark_results.csv')
+    fig = px.bar(df,
+             x='average_score',
+             y='model',
+             color='average_score',
+             color_continuous_scale='tealrose',
+             hover_data=['armenian_language_score', 'armenian_history_score', 'mathematics_score'],
+             labels={'average_score': 'Average Score', 'model': 'Model'},
+             title='Average Score per Model',
+             orientation='h',
+             range_color=[0, 20])
+
+    fig.update_layout(
+        xaxis=dict(range=[0, 20]),
+        title=dict(text='Average Score per Model', font=dict(size=16)),
+        xaxis_title=dict(font=dict(size=12)),
+        yaxis_title=dict(font=dict(size=12)),
+        yaxis=dict(autorange="reversed"),
+        hoverlabel=dict(
+            bgcolor="white",  
+            font_size=12,
+            font_family="Arial",
+            font_color="black"  
+        )
+    )
+
+    return fig
+
+with gr.Blocks() as app:
+    gr.Markdown("# ArmBench Leaderboard")
+    table_output = gr.DataFrame(value=display_table())
+    plot_output = gr.Plot(create_bar_chart)
+
+app.launch(share=True)
+
+

requirements.txt

@@ -0,0 +1,3 @@
+gradio==5.19.0
+pandas==2.2.3
+plotly==6.0.0