update

app.py

@@ -7,6 +7,7 @@ import huggingface_hub
 from huggingface_hub.utils._errors import EntryNotFoundError, RepositoryNotFoundError
 from dotenv import load_dotenv
 from matplotlib.colors import LinearSegmentedColormap
+import plotly.express as px
 
 load_dotenv()
 webhook_url = os.environ.get("WEBHOOK_URL")
@@ -269,8 +270,34 @@ for folder in get_folders_matching_format('data'):
             all_data[folder_name][file_name][sheet_name] = rename_columns(
                 pd.read_excel(final_file_name + '.xlsx', sheet_name=sheet_name))
 
+
+def create_scaling_plot(all_data, period):
+    selected_columns = ['Name', 'Parameters Count (B)', 'Average (The lower the better)']
+    target_data = all_data[period]
+    new_df = pd.DataFrame()
+
+    for size in target_data.keys():
+        new_df = pd.concat([new_df, target_data[size]['cr'].loc[:, selected_columns]], axis=0)
+
+    new_df.rename(columns={
+        'Parameters Count (B)': 'Params(B)',
+        'Average (The lower the better)': 'Compression Rate (%)'
+    }, inplace=True)
+
+    fig = px.scatter(new_df,
+                     x='Params(B)',
+                     y='Compression Rate (%)',
+                     title='Compression Rate Scaling Law',
+                     hover_name='Name'
+                     )
+    fig.update_traces(marker=dict(size=12))
+    return fig
+
+
+initial_fig = create_scaling_plot(all_data, time_list[-1])
+
 initial_period = time_list[-1]
-initial_models = model_size_list[:1]
+initial_models = model_size_list
 initial_metric = metric_list[0]
 initial_columns = get_unique_column_names(all_data)
 initial_colors = ['Average']
@@ -300,7 +327,7 @@ with gr.Blocks(css=css) as demo:
             with gr.Row():
                 with gr.Column():
                     period_selector = gr.Dropdown(label="Period", choices=time_list, value=time_list[0])
-                    model_selector = gr.CheckboxGroup(label="Model", choices=model_size_list, value=model_size_list[0])
+                    model_selector = gr.CheckboxGroup(label="Model", choices=model_size_list, value=model_size_list)
                     metric_selector = gr.Dropdown(label="Metric", choices=metric_list, value=metric_list[0])
                 with gr.Column():
                     color_selector = gr.CheckboxGroup(label="Colored Columns",
@@ -312,7 +339,7 @@ with gr.Blocks(css=css) as demo:
 
             table = gr.Dataframe(initial_data, column_widths=[130, 60, 60, 35, 35, 40, 40, 35, 35, 35],
                                  wrap=True,
-                                 height=800
+                                 height=800,
                                  )
 
             period_selector.change(update_table,
@@ -333,6 +360,17 @@ with gr.Blocks(css=css) as demo:
 
         with gr.Tab("🌍 MultiLang"):
             gr.Markdown("## Coming soon...")
+        with gr.Tab("📈 Scaling Law"):
+            period_selector_2 = gr.Dropdown(label="Period", choices=time_list, value=time_list[0])
+
+            def update_plot(period):
+                new_fig = create_scaling_plot(all_data, period)
+                return new_fig
+
+
+            plot = gr.Plot(initial_fig)
+            period_selector_2.change(update_plot, inputs=period_selector_2, outputs=plot)
+
         with gr.Tab("ℹ️ About"):
             gr.Markdown(about_md)
         with gr.Tab("🚀 Submit"):

requirements.txt

@@ -1 +1,2 @@
-openpyxl
+openpyxl
+plotly