|
import gradio as gr |
|
|
|
from src.leaderboard import BGB_COLUMN_MAPPING, get_bgb_leaderboard_df, get_leaderboard_df |
|
from src.llm_perf import get_eval_df, get_llm_perf_df |
|
|
|
|
|
def select_columns_fn(machine, columns, search, llm_perf_df=None): |
|
if llm_perf_df is None: |
|
llm_perf_df = get_llm_perf_df(machine=machine) |
|
|
|
selected_leaderboard_df = get_leaderboard_df(llm_perf_df) |
|
selected_leaderboard_df = selected_leaderboard_df[ |
|
selected_leaderboard_df["Model π€"].str.contains(search, case=False) |
|
] |
|
selected_leaderboard_df = selected_leaderboard_df[columns] |
|
|
|
return selected_leaderboard_df |
|
|
|
|
|
def select_columns_bgb_fn(machine, columns, search, type_checkboxes, param_slider, eval_df=None): |
|
if eval_df is None: |
|
eval_df = get_eval_df(machine) |
|
|
|
selected_leaderboard_df = get_bgb_leaderboard_df(eval_df) |
|
selected_leaderboard_df = selected_leaderboard_df[ |
|
selected_leaderboard_df["Model π€"].str.contains(search, case=False) |
|
] |
|
|
|
print(param_slider) |
|
|
|
import pdb |
|
|
|
pdb.set_trace() |
|
|
|
columns = ["Model π€"] + columns + type_checkboxes |
|
|
|
return selected_leaderboard_df[columns] |
|
|
|
|
|
def create_select_callback( |
|
|
|
machine_textbox, |
|
|
|
columns_checkboxes, |
|
search_bar, |
|
type_checkboxes, |
|
param_slider, |
|
|
|
leaderboard_table, |
|
): |
|
columns_checkboxes.change( |
|
fn=select_columns_bgb_fn, |
|
inputs=[machine_textbox, columns_checkboxes, search_bar, type_checkboxes, param_slider], |
|
outputs=[leaderboard_table], |
|
) |
|
search_bar.change( |
|
fn=select_columns_bgb_fn, |
|
inputs=[machine_textbox, columns_checkboxes, search_bar, type_checkboxes, param_slider], |
|
outputs=[leaderboard_table], |
|
) |
|
|