Update app.py

app.py

@@ -3,12 +3,12 @@ import pandas as pd
 import plotly.express as px
 
 CITATION_BUTTON_LABEL = "Copy the following snippet to cite these results"
-CITATION_BUTTON_TEXT = r"""@misc{energystarai-leaderboard,
+CITATION_BUTTON_TEXT = r"""@misc{aienergyscore-leaderboard,
     author = {Sasha Luccioni and Boris Gamazaychikov and Emma Strubell and Sara Hooker and Yacine Jernite and Carole-Jean Wu and Margaret Mitchell},
-    title = {AI Energy Score Leaderboard v.0},
-    year = {2024},
+    title = {AI Energy Score Leaderboard - February 2025},
+    year = {2025},
     publisher = {Hugging Face},
-    howpublished = "\url{https://huggingface.co/spaces/EnergyStarAI/2024_Leaderboard}",
+    howpublished = "\url{https://huggingface.co/spaces/AIEnergyScore/Leaderboard}",
 }"""
 
 # List of tasks (CSV filenames)
@@ -28,17 +28,18 @@ tasks = [
 def format_stars(score):
     """
     Convert the energy_score (assumed to be an integer from 1 to 5)
-    into that many star characters wrapped in a span with the given color.
+    into that many star characters wrapped in a span styled with color #3fa45bff
+    and with a font size increased to 2em.
     """
     try:
         score_int = int(score)
     except Exception:
         score_int = 0
-    return f'<span style="color: #3fa45bff; font-size:1.2em;">{"★" * score_int}</span>'
+    return f'<span style="color: #3fa45bff; font-size:2em;">{"★" * score_int}</span>'
 
 def make_link(mname):
     """
-    Create a markdown link from the model identifier.
+    Create a markdown link for the model.
     For example, if mname is "org/model", display "model" and link to its HF page.
     """
     parts = str(mname).split('/')
@@ -48,13 +49,13 @@ def make_link(mname):
 def get_plots(task):
     """
     Read the energy CSV for a given task and return a Plotly scatter plot.
-    The y-axis shows the total GPU energy (Wh) and the color is determined by energy_score.
+    Now the x-axis is the numeric energy (GPU Energy (Wh)) and
+    the y-axis displays the model name.
     """
     df = pd.read_csv('data/energy/' + task)
-    # Ensure energy_score is an integer (for discrete color mapping)
     df['energy_score'] = df['energy_score'].astype(int)
-    # Convert kWh to Wh and round to 4 decimal places.
-    df['Total GPU Energy (Wh)'] = (df['total_gpu_energy'] * 1000).round(4)
+    # Do not multiply by 1000; simply round to 4 decimals
+    df['GPU Energy (Wh)'] = df['total_gpu_energy'].round(4)
     
     # Define a 5-level color mapping: 1 = red, 5 = green.
     color_map = {
@@ -65,10 +66,11 @@ def get_plots(task):
         5: "green"
     }
     
+    # Create a horizontal scatter plot: x is the energy, y is the model.
     fig = px.scatter(
         df,
-        x="model",
-        y="Total GPU Energy (Wh)",
+        x="GPU Energy (Wh)",
+        y="model",
         custom_data=['energy_score'],
         height=500,
         width=800,
@@ -77,24 +79,24 @@ def get_plots(task):
     )
     fig.update_traces(
         hovertemplate="<br>".join([
-            "Model: %{x}",
-            "Total Energy (Wh): %{y}",
+            "Model: %{y}",
+            "GPU Energy (Wh): %{x}",
             "Energy Score: %{customdata[0]}"
         ])
     )
-    fig.update_layout(xaxis_title="Model", yaxis_title="Total GPU Energy (Wh)")
+    fig.update_layout(xaxis_title="GPU Energy (Wh)", yaxis_title="Model")
     return fig
 
 def get_all_plots():
     """
     Combine data from all tasks and return a scatter plot.
-    Duplicate models (if any) are dropped.
+    Duplicate models are dropped.
     """
     all_df = pd.DataFrame()
     for task in tasks:
         df = pd.read_csv('data/energy/' + task)
         df['energy_score'] = df['energy_score'].astype(int)
-        df['Total GPU Energy (Wh)'] = (df['total_gpu_energy'] * 1000).round(4)
+        df['GPU Energy (Wh)'] = df['total_gpu_energy'].round(4)
         all_df = pd.concat([all_df, df], ignore_index=True)
     all_df = all_df.drop_duplicates(subset=['model'])
     
@@ -107,8 +109,8 @@ def get_all_plots():
     }
     fig = px.scatter(
         all_df,
-        x="model",
-        y="Total GPU Energy (Wh)",
+        x="GPU Energy (Wh)",
+        y="model",
         custom_data=['energy_score'],
         height=500,
         width=800,
@@ -117,59 +119,68 @@ def get_all_plots():
     )
     fig.update_traces(
         hovertemplate="<br>".join([
-            "Model: %{x}",
-            "Total Energy (Wh): %{y}",
+            "Model: %{y}",
+            "GPU Energy (Wh): %{x}",
             "Energy Score: %{customdata[0]}"
         ])
     )
-    fig.update_layout(xaxis_title="Model", yaxis_title="Total GPU Energy (Wh)")
+    fig.update_layout(xaxis_title="GPU Energy (Wh)", yaxis_title="Model")
     return fig
 
 def get_model_names(task):
     """
-    For a given task, load the energy CSV and return a dataframe with three columns:
-    - Model (a markdown link),
-    - Rating (the star rating based on energy_score),
-    - Total GPU Energy (Wh)
+    For a given task, load the energy CSV and return a dataframe with the following columns:
+      - Model (a markdown link)
+      - GPU Energy (Wh)
+      - Score (a star rating based on energy_score)
+    For text_generation.csv only, also add the "Class" column from the CSV.
+    The final order is: Model, GPU Energy (Wh), Score, [Class].
     """
     df = pd.read_csv('data/energy/' + task)
     df['energy_score'] = df['energy_score'].astype(int)
-    df['Total GPU Energy (Wh)'] = (df['total_gpu_energy'] * 1000).round(4)
+    df['GPU Energy (Wh)'] = df['total_gpu_energy'].round(4)
     df['Model'] = df['model'].apply(make_link)
-    df['Rating'] = df['energy_score'].apply(format_stars)
-    df = df.sort_values(by='Total GPU Energy (Wh)')
-    model_names = df[['Model', 'Rating', 'Total GPU Energy (Wh)']]
-    return model_names
+    df['Score'] = df['energy_score'].apply(format_stars)
+    
+    # If this CSV contains a "class" column (e.g., for Text Generation), add it.
+    if 'class' in df.columns:
+        df['Class'] = df['class']
+        df = df[['Model', 'GPU Energy (Wh)', 'Score', 'Class']]
+    else:
+        df = df[['Model', 'GPU Energy (Wh)', 'Score']]
+        
+    df = df.sort_values(by='GPU Energy (Wh)')
+    return df
 
 def get_all_model_names():
     """
-    Combine data from all tasks and return a table of models.
+    Combine data from all tasks and return a leaderboard table with:
+      - Model, GPU Energy (Wh), Score
     Duplicate models are dropped.
     """
     all_df = pd.DataFrame()
     for task in tasks:
         df = pd.read_csv('data/energy/' + task)
         df['energy_score'] = df['energy_score'].astype(int)
-        df['Total GPU Energy (Wh)'] = (df['total_gpu_energy'] * 1000).round(4)
+        df['GPU Energy (Wh)'] = df['total_gpu_energy'].round(4)
         df['Model'] = df['model'].apply(make_link)
-        df['Rating'] = df['energy_score'].apply(format_stars)
+        df['Score'] = df['energy_score'].apply(format_stars)
         all_df = pd.concat([all_df, df], ignore_index=True)
     all_df = all_df.drop_duplicates(subset=['model'])
-    all_df = all_df.sort_values(by='Total GPU Energy (Wh)')
-    model_names = all_df[['Model', 'Rating', 'Total GPU Energy (Wh)']]
-    return model_names
+    all_df = all_df.sort_values(by='GPU Energy (Wh)')
+    return all_df[['Model', 'GPU Energy (Wh)', 'Score']]
 
 # Build the Gradio interface.
 demo = gr.Blocks()
 
 with demo:
     gr.Markdown(
-        """# AI Energy Score Leaderboard - v.0 (2024) 🌎 💻 🌟
-### Welcome to the leaderboard for the [AI Energy Score Project!](https://huggingface.co/EnergyStarAI)
+        """# AI Energy Score Leaderboard
+### Welcome to the leaderboard for the [AI Energy Score Project!](https://huggingface.co/AIEnergyScore)
 Click through the tasks below to see how different models measure up in terms of energy efficiency."""
     )
     gr.Markdown(
-        """Test your own models via the [submission portal](https://huggingface.co/spaces/AIEnergyScore/submission_portal)!"""
+        """Test your own models via the [submission portal](https://huggingface.co/spaces/AIEnergyScore/submission_portal)"""
     )
     
     with gr.Tabs():
@@ -178,6 +189,7 @@ Click through the tasks below to see how different models measure up in terms of
                 with gr.Column(scale=1.3):
                     plot = gr.Plot(get_plots('text_generation.csv'))
                 with gr.Column(scale=1):
+                    # For text generation, the CSV is assumed to have a "class" column.
                     table = gr.Dataframe(get_model_names('text_generation.csv'), datatype="markdown")
                     
         with gr.TabItem("Image Generation 📷"):
@@ -262,4 +274,4 @@ Click through the tasks below to see how different models measure up in terms of
         """Last updated: February 2025"""
     )
 
-demo.launch()
+demo.launch()