Upload 7 files

app.py

@@ -84,7 +84,7 @@ with hf_app:
                 elem_id="text-leaderboard-table",
                 interactive=False,
                 visible=True,
-                height=dataframe_height
+                # height=dataframe_height
             )
 
             # Show information about the clemscore and last updated date below the table
@@ -124,7 +124,7 @@ with hf_app:
                 elem_id="mm-leaderboard-table",
                 interactive=False,
                 visible=True,
-                height=dataframe_height
+                # height=dataframe_height
             )
 
             # Show information about the clemscore and last updated date below the table
@@ -407,7 +407,7 @@ with hf_app:
                 elem_id="version-leaderboard-table",
                 interactive=False,
                 visible=True,
-                height=dataframe_height
+                # height=dataframe_height
             )
 
             dummy_prev_table = gr.Dataframe(

requirements.txt

@@ -1,4 +1,4 @@
-gradio==4.40.0
+gradio==5.8.0
 pandas==2.2.2
 plotly==5.18.0
 apscheduler==3.10.4

src/assets/text_content.py

@@ -1,7 +1,7 @@
 TITLE = """<h1 align="center" id="space-title"> 🏆 CLEM Leaderboard</h1>"""
 
-REPO = "https://raw.githubusercontent.com/clembench/clembench-runs/main/"
-REGISTRY_URL = "https://raw.githubusercontent.com/kushal-10/clembench/feat/registry/backends/model_registry_updated.json"
+REPO = "https://raw.githubusercontent.com/kushal-10/clembench-runs/main/"
+REGISTRY_URL = "https://raw.githubusercontent.com/clp-research/clembench/refs/heads/main/backends/model_registry.json"
 
 HF_REPO = "colab-potsdam/clem-leaderboard"
 

src/leaderboard_utils.py

@@ -10,6 +10,7 @@ from src.assets.text_content import REPO
 def get_github_data():
     """
     Read and process data from CSV files hosted on GitHub. - https://github.com/clembench/clembench-runs
+    Set the path in src/assets/text_content/REPO
 
     Returns:
         github_data (dict): Dictionary containing:
@@ -60,7 +61,7 @@ def get_github_data():
                 text_dfs.append(df)
                 if text_flag:
                     text_flag = False
-                    text_date = next(ver['date'] for ver in versions if ver['version'] == version)
+                    text_date = next(ver['last_updated'] for ver in versions if ver['version'] == version)
                     text_date = datetime.strptime(text_date, "%Y-%m-%d").strftime("%d %b %Y")  
 
             else:
@@ -77,7 +78,7 @@ def get_github_data():
             mm_dfs.append(df)
             if mm_flag:
                 mm_flag = False
-                mm_date = next(ver['date'] for ver in versions if ver['version'] == version)
+                mm_date = next(ver['last_updated'] for ver in versions if ver['version'] == version)
                 mm_date = datetime.strptime(mm_date, "%Y-%m-%d").strftime("%d %b %Y")
 
       

src/trend_utils.py

@@ -346,7 +346,7 @@ def get_final_trend_plot(benchmark: str = "Text", mobile_view: bool = False) ->
         benchmark_ticks = {}
         for ver in versions:
             if 'multimodal' not in ver['version']: # Skip MM specific benchmark dates
-                benchmark_ticks[pd.to_datetime(ver['date'])] = ver['version']
+                benchmark_ticks[pd.to_datetime(ver['release_date'])] = ver['version']
         fig =  get_plot(text_result_df, start_date=START_DATE, end_date=datetime.now().strftime('%Y-%m-%d'), benchmark_ticks=benchmark_ticks, **plot_kwargs)
     else:
         mm_dfs = get_github_data()['multimodal']
@@ -358,8 +358,8 @@ def get_final_trend_plot(benchmark: str = "Text", mobile_view: bool = False) ->
         for ver in versions:
             if 'multimodal' in ver['version']:
                 ver['version'] = ver['version'].replace('_multimodal', '')
-            if date_difference(ver['date'], '2024-07-15') >= 0:
-                benchmark_ticks[pd.to_datetime(ver['date'])] = ver['version'] ## MM benchmark dates considered after v1.6 (incl.)
+            if date_difference(ver['release_date'], '2024-07-15') >= 0:
+                benchmark_ticks[pd.to_datetime(ver['release_date'])] = ver['version'] ## MM benchmark dates considered after v1.6 (incl.)
         fig = get_plot(df, start_date=START_DATE, end_date=datetime.now().strftime('%Y-%m-%d'), benchmark_ticks=benchmark_ticks, **plot_kwargs)
 
     return fig

src/version_utils.py

@@ -40,7 +40,7 @@ def get_versions_data():
     # Get Last updated date of the latest version
     latest_version = version_names[0]
     latest_date = next(
-        ver['date'] for ver in versions if ver['version'] == latest_version
+        ver['last_updated'] for ver in versions if ver['version'] == latest_version
     )
     formatted_date = datetime.strptime(latest_date, "%Y-%m-%d").strftime("%d %b %Y")