Update app.py

app.py

@@ -30,12 +30,13 @@ 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 styled with color #3fa45bff
     and with a font size increased to 2em.
+    The use of '!important' forces the styling immediately.
     """
     try:
         score_int = int(score)
     except Exception:
         score_int = 0
-    return f'<span style="color: #3fa45bff; font-size:2em;">{"★" * score_int}</span>'
+    return f'<span style="color: #3fa45bff !important; font-size:2em !important;">{"★" * score_int}</span>'
 
 def make_link(mname):
     """
@@ -49,13 +50,15 @@ def make_link(mname):
 def get_plots(task):
     """
     Read the energy CSV for a given task and return a Plotly scatter plot.
-    Now the x-axis is the numeric energy (GPU Energy (Wh)) and
-    the y-axis displays the model name.
+    X-axis: Numeric GPU Energy (Wh) (rounded to 4 decimals)
+    Y-axis: Display only the model name (extracted from the model field)
     """
     df = pd.read_csv('data/energy/' + task)
     df['energy_score'] = df['energy_score'].astype(int)
-    # Do not multiply by 1000; simply round to 4 decimals
+    # Use the raw energy (no multiplication) rounded to 4 decimals for plotting
     df['GPU Energy (Wh)'] = df['total_gpu_energy'].round(4)
+    # Create a column that displays only the model name (the part after '/')
+    df['Display Model'] = df['model'].apply(lambda m: m.split('/')[-1])
     
     # Define a 5-level color mapping: 1 = red, 5 = green.
     color_map = {
@@ -66,11 +69,10 @@ def get_plots(task):
         5: "green"
     }
     
-    # Create a horizontal scatter plot: x is the energy, y is the model.
     fig = px.scatter(
         df,
         x="GPU Energy (Wh)",
-        y="model",
+        y="Display Model",
         custom_data=['energy_score'],
         height=500,
         width=800,
@@ -97,6 +99,7 @@ def get_all_plots():
         df = pd.read_csv('data/energy/' + task)
         df['energy_score'] = df['energy_score'].astype(int)
         df['GPU Energy (Wh)'] = df['total_gpu_energy'].round(4)
+        df['Display Model'] = df['model'].apply(lambda m: m.split('/')[-1])
         all_df = pd.concat([all_df, df], ignore_index=True)
     all_df = all_df.drop_duplicates(subset=['model'])
     
@@ -110,7 +113,7 @@ def get_all_plots():
     fig = px.scatter(
         all_df,
         x="GPU Energy (Wh)",
-        y="model",
+        y="Display Model",
         custom_data=['energy_score'],
         height=500,
         width=800,
@@ -131,18 +134,18 @@ def get_model_names(task):
     """
     For a given task, load the energy CSV and return a dataframe with the following columns:
       - Model (a markdown link)
-      - GPU Energy (Wh)
+      - GPU Energy (Wh) formatted as a string with 4 decimal places
       - 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].
+    The final column order is: Model, GPU Energy (Wh), Score, [Class].
     """
     df = pd.read_csv('data/energy/' + task)
     df['energy_score'] = df['energy_score'].astype(int)
-    df['GPU Energy (Wh)'] = df['total_gpu_energy'].round(4)
+    # Format the energy as a string with 4 decimals so that very small values display correctly
+    df['GPU Energy (Wh)'] = df['total_gpu_energy'].apply(lambda x: f"{x:.4f}")
     df['Model'] = df['model'].apply(make_link)
     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']]
@@ -162,7 +165,7 @@ def get_all_model_names():
     for task in tasks:
         df = pd.read_csv('data/energy/' + task)
         df['energy_score'] = df['energy_score'].astype(int)
-        df['GPU Energy (Wh)'] = df['total_gpu_energy'].round(4)
+        df['GPU Energy (Wh)'] = df['total_gpu_energy'].apply(lambda x: f"{x:.4f}")
         df['Model'] = df['model'].apply(make_link)
         df['Score'] = df['energy_score'].apply(format_stars)
         all_df = pd.concat([all_df, df], ignore_index=True)
@@ -189,7 +192,6 @@ 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 📷"):
@@ -274,4 +276,4 @@ Click through the tasks below to see how different models measure up in terms of
         """Last updated: February 2025"""
     )
 
-demo.launch()
+demo.launch()