added app

app.py

@@ -4,6 +4,7 @@ import requests
 import pandas as pd
 from io import StringIO
 import plotly.graph_objs as go
+import plotly.express as px
 
 
 def convert_markdown_table_to_dataframe(md_content):
@@ -23,7 +24,7 @@ def convert_markdown_table_to_dataframe(md_content):
     return df
 
 
-def create_bar_chart(df, metric):
+def create_bar_chart(df, metric, color_map, key_suffix):
     """
     Creates and displays a bar chart for a given metric.
     """
@@ -37,19 +38,67 @@ def create_bar_chart(df, metric):
         x=sorted_df[metric],
         y=sorted_df['Model'],
         orientation='h',
-        marker=dict(color=sorted_df[metric], colorscale='Inferno')
+        marker=dict(color=sorted_df[metric], colorscale=color_map)
     ))
     fig.update_layout(margin=dict(l=20, r=20, t=20, b=20))
-    st.plotly_chart(fig, use_container_width=True)
+    st.plotly_chart(fig, use_container_width=True, key=f"bar_chart_{metric}_{key_suffix}")
+
+
+def create_radar_chart(df, metric_columns):
+    """
+    Create a radar chart for the top 10 models by "Average" score.
+    """
+    st.write("### Radar Chart (Top 10 Models by Average Score)")
+
+    if 'Average' not in df.columns:
+        st.write("Average column not found.")
+        return
+
+    top_10_df = df.nlargest(10, 'Average')
+    if top_10_df.empty:
+        st.write("No models available for the radar chart.")
+        return
+
+    radar_data = top_10_df[['Model'] + metric_columns].set_index('Model')
+
+    fig = go.Figure()
+
+    for model_name, row in radar_data.iterrows():
+        fig.add_trace(go.Scatterpolar(
+            r=row.values,
+            theta=metric_columns,
+            fill='toself',
+            name=model_name
+        ))
+
+    fig.update_layout(
+        polar=dict(
+            radialaxis=dict(visible=True, range=[0, 1.2 * radar_data.max().max()])  # Dynamic range
+        ),
+        showlegend=True
+    )
+
+    st.plotly_chart(fig, use_container_width=True, key="radar_chart")
 
 
 def main():
     st.set_page_config(page_title="LLM Leaderboard", layout="wide")
-    st.title("🏆 LLM Leaderboard")
+    st.title("🏆 SLM Leaderboard")
+    st.markdown("We record Nous benchmark results for various SLMs. Please submit a PR to this [repo](https://github.com/amazon-science/aws-research-science/tree/main/SLMleaderboard) to inlcude your model! Heavily Inspired by [YALB](https://huggingface.co/spaces/mlabonne/Yet_Another_LLM_Leaderboard) ")
 
     # URL to your markdown file
-    md_url = st.text_input("Enter the URL to the markdown file", "https://raw.githubusercontent.com/yourrepo/README.md")
+    md_url = st.text_input("This the default location of the bechmarks and can be changed", 
+                           "https://raw.githubusercontent.com/amazon-science/aws-research-science/refs/heads/main/SLMleaderboard/nous.md")
     
+
+    st.markdown("""
+                Copy the following links into the textbox above and refresh dashboard:
+
+                - [Nous benchmark results](https://raw.githubusercontent.com/amazon-science/aws-research-science/refs/heads/main/SLMleaderboard/nous.md)
+                - [Standard LLM benchmarks](https://raw.githubusercontent.com/amazon-science/aws-research-science/refs/heads/main/SLMleaderboard/standard.md) """)
+
+
+
     if not md_url:
         st.error("Please provide a valid URL to a markdown file containing the leaderboard table.")
         return
@@ -68,6 +117,14 @@ def main():
         for col in metric_columns:
             df[col] = pd.to_numeric(df[col], errors='coerce')
 
+        # Calculate "Average" score as a new column (exclude 'Model' and 'URL')
+        df['Average'] = df[metric_columns].mean(axis=1, skipna=True)
+        if 'Average' not in metric_columns:
+            metric_columns.append('Average')
+
+        # Dropdown to select color map
+        color_map = st.selectbox("Select Color Map for Bar Charts", options=['Inferno', 'Viridis', 'Cividis'])
+
         # Sortable leaderboard table
         st.dataframe(
             df[['Model'] + metric_columns + ['URL']],
@@ -76,8 +133,14 @@ def main():
         )
 
         # Bar charts for each metric
-        for metric in metric_columns:
-            create_bar_chart(df, metric)
+        for i, metric in enumerate(metric_columns):
+            create_bar_chart(df, metric, color_map, key_suffix=i)
+
+        # Extra bar chart for the "Average" score
+        create_bar_chart(df, 'Average', color_map, key_suffix="average")
+
+        # Radar chart for the top 10 models by "Average" score
+        create_radar_chart(df, metric_columns)
 
     except Exception as e:
         st.error(f"An error occurred while processing the markdown table: {e}")
@@ -85,4 +148,3 @@ def main():
 
 if __name__ == "__main__":
     main()
-

requirements.txt

@@ -0,0 +1,5 @@
+streamlit
+pandas
+requests
+plotly
+