Create config.py

config.py

@@ -0,0 +1,74 @@
+import gradio as gr
+from gradio_leaderboard import Leaderboard, SelectColumns, ColumnFilter
+import config
+from pathlib import Path
+import pandas as pd
+from datetime import datetime
+
+abs_path = Path(__file__).parent
+
+
+df = pd.read_json(str(abs_path / "leader_board.json"))
+# Randomly set True/ False for the "MOE" column
+#
+# print(df.info())
+# print(df.columns)
+# print(df.head(1))
+head_content = """
+    # 🏅 BlinkCode Leaderboard
+    ### Welcome to the BlinkCode Leaderboard! On this leaderboard we share the evaluation results of MLLMs obtained by the [OpenSource Framework](github.link).
+
+    ### Currently, BlinkCode Leaderboard covers <model num> different VLMs (including GPT-4v, Gemini, QwenVLMAX, LLaVA, etc.) and 9 different task.
+    ## Main Evaluation Results
+    - Metrics:
+      - Avg Score: The average score on all task (normalized to 0 - 100, the higher the better). 
+      - The scores in the 5 tasks (HumanEval-V, MBPP-V, GSM8K-V, MATH-V, VP) represent the percentage of accuracy.
+      - The scores in the image reconstruction tasks (Matplotlib, SVG, TikZ, Webpage) represent the similarity between the reconstructed images and the original images (normalized to 0 - 100, the higher the better).
+    - By default, we present the unrefined evaluation results,, sorted by the descending order of Avg Score⬆️.   
+      -  The ⭐ symbol indicates results that have undergone two rounds of refinement.  
+        
+        
+    This leaderboard was last updated: <nowtime>.
+    """
+CITATION_BUTTON_TEXT = r"""@misc{2023opencompass,
+    title={OpenCompass: A Universal Evaluation Platform for Foundation Models},
+    author={OpenCompass Contributors},
+    howpublished = {\url{https://github.com/open-compass/opencompass}},
+    year={2023}
+}"""
+CITATION_BUTTON_LABEL = "Copy the following snippet to cite these results"
+unique_models_count = df["Model"].nunique()
+# print(unique_models_count)
+nowtime = datetime.now()
+formatted_time = nowtime.strftime("%y.%m.%d %H:%M:%S")
+head_content = head_content.replace("<nowtime>", formatted_time).replace('<model num>', str(unique_models_count))
+
+with gr.Blocks() as demo:
+    gr.Markdown(head_content)
+    with gr.Tabs():
+        Leaderboard(
+            value=df,
+            select_columns=SelectColumns(
+                default_selection=config.ON_LOAD_COLUMNS,
+                cant_deselect=["Rank", "Model"],
+                label="Select Columns to Display:",
+            ),
+            search_columns=["Model", "Model Type"],
+            hide_columns=["Model Size", "Model Type", "Supports multiple images"],
+            filter_columns=[
+                "Model Size",
+                "Model Type",
+                "Supports multiple images"
+                # ColumnFilter("Params (B)", default=[0, 20]),
+            ],
+            datatype=config.TYPES,
+            column_widths=["5%", "15%"],
+        )
+    with gr.Row():
+        with gr.Accordion('Citation', open=False):
+            citation_button = gr.Textbox(
+                value=CITATION_BUTTON_TEXT,
+                label=CITATION_BUTTON_LABEL,
+                elem_id='citation-button')
+if __name__ == "__main__":
+    demo.launch()