app.py

import gradio as gr
from gradio_leaderboard import Leaderboard, SelectColumns, ColumnFilter
from pathlib import Path

from utils import LLM_BENCHMARKS_ABOUT_TEXT, LLM_BENCHMARKS_SUBMIT_TEXT, custom_css, jsonl_to_dataframe, add_average_column_to_df, apply_markdown_format_for_columns, submit, sort_dataframe_by_column



abs_path = Path(__file__).parent

# Any pandas-compatible data
leaderboard_df = jsonl_to_dataframe(str(abs_path / "leaderboard_data.jsonl"))

average_column_name = "Average Accuracy"

all_columns = ["Model", average_column_name, "Precision", "#Params (B)", "MMLU", "GSM8K", "TruthfulQA", "Winogrande", "ARC Easy", "Hellaswag", "Belebele"]
columns_to_average = ["MMLU", "GSM8K", "TruthfulQA", "Winogrande", "ARC Easy", "Hellaswag", "Belebele"]


leaderboard_df = add_average_column_to_df(leaderboard_df, columns_to_average, index=3, average_column_name=average_column_name)
leaderboard_df = apply_markdown_format_for_columns(df=leaderboard_df, model_column_name="Model")
leaderboard_df = sort_dataframe_by_column(leaderboard_df, column_name=average_column_name)

columns_data_type = ["markdown" for i in range(len(leaderboard_df.columns))]
# "str", "number", "bool", "date", "markdown"
# columns_data_type[0] = "markdown"

NUM_MODELS=len(leaderboard_df)

with gr.Blocks(css=custom_css) as demo:
    gr.Markdown("""
    # Open Lithuanian LLM Leaderboard
    """)

    gr.Markdown(f"""
    - **Total Models**: {NUM_MODELS}
    """)

    with gr.Tab("🎖️ Lithuanian Leaderboard"):
        Leaderboard(
        value=leaderboard_df,
        datatype=columns_data_type,
        select_columns=SelectColumns(
            default_selection=all_columns,
            cant_deselect=["Model"],
            label="Select Columns to Show",
        ),
        search_columns=["model_name_for_query"],
        hide_columns=["model_name_for_query",],
        filter_columns=["Precision", "#Params (B)"],
    )
    with gr.TabItem("📝 About"):
        gr.Markdown(LLM_BENCHMARKS_ABOUT_TEXT)

    with gr.Tab("✉️ Submit"):
        gr.Markdown(LLM_BENCHMARKS_SUBMIT_TEXT)
        model_name = gr.Textbox(label="Model name")
        model_id = gr.Textbox(label="username/space e.g neurotechnology/Lt-Llama-2-7b-hf")
        contact_email = gr.Textbox(label="Contact E-Mail")
        submit_btn = gr.Button("Submit")

        submit_btn.click(submit, inputs=[model_name, model_id, contact_email], outputs=[])

        gr.Markdown("""
        Please find more information about Neurotechnology on [www.neurotechnology.com](https://www.neurotechnology.com/natural-language-processing.html)""")

if __name__ == "__main__":
    demo.launch()