Spaces:
Running
Running
File size: 3,361 Bytes
f216956 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 |
import argparse
import ast
import pickle
import os
import threading
import time
import gradio as gr
import numpy as np
import pandas as pd
from serve.model import model_config
def hyperlink(name, link):
return f'<a target="_blank" href="{link}" style="color: var(--link-text-color); \
text-decoration: underline;text-decoration-style: dotted;">{name}</a>'
def amend_model_name(name, rank):
model_name = model_config[name].model_name
if rank==1:
return "🥇 " + model_name
elif rank==2:
return "🥈 " + model_name
elif rank==3:
return '🥉 ' + model_name
else:
return model_name
def get_cfg_info(name):
config = model_config[name]
links = []
if config.page_link:
links.append(hyperlink("Page", config.page_link))
if config.code_link:
links.append(hyperlink("Code", config.code_link))
return ", ".join(links) if links else "N/A", config.organization if config.organization else "N/A"
def get_leaderboard_values(leaderboard_df):
leaderboard_vals = []
for i, row in leaderboard_df.iterrows():
rank = i+1
model_name = row["Method"]
task = row["Task"]
if model_name not in model_config.keys() or model_config[model_name].task != task:
continue
values = [rank, amend_model_name(model_name, rank), task]
values = values + [row.get(dim, np.NaN) for dim in leaderboard_df.columns[2:]]
# values.append(round(np.sum([v for v in values[3:] if pd.notna(v)]), 4))
links, organization = get_cfg_info(model_name)
# values.append(links)
values.append(organization)
leaderboard_vals.append(values)
return leaderboard_vals
def get_topk_ranks(df, k=3):
ranks = {}
for col_idx, col in enumerate(df.columns[2:]): # skip "Model" “Task"
topk = df[col].nlargest(k)
for rank, idx in enumerate(topk.index):
if idx not in ranks:
ranks[idx] = {}
ranks[idx][col_idx] = rank + 1 # 1-based rank
for i in range(k): ranks[i][5] = i + 1
return ranks # dict: row -> {col: rank}
def build_leaderboard_tab(leaderboard_file: str, task: str = ""):
if not isinstance(leaderboard_file, str):
leaderboard_file = leaderboard_file.value
if not isinstance(task, str):
task = task.value
df = pd.read_csv(leaderboard_file)
if task in ["Text-to-3D only", "Image-to-3D only"]:
df = df[df["Task"] == task.split()[0]]
# df = df.drop(df[df["Task"]!=task.split()[0]].index)
leaderboard_df = df.drop(df[df["Method"].isnull()].index)
leaderboard_df = leaderboard_df.reset_index(drop=True)
leaderboard_vals = get_leaderboard_values(leaderboard_df)
leaderboard = gr.Dataframe(
headers = ['Rank', "🤖 Model", "🪧 Task" ]
+ [f"{dim}" for dim in leaderboard_df.keys()[2:-1]]
+ ["⭐ Overall", "🏛️ Orgnization"], # "🔗 Links",
datatype = ["number", "str", "str"]
+ ["number"] * (len(leaderboard_df.columns) - 3)
+ ["number", "str"],
value = leaderboard_vals,
height = 680,
column_widths = [60, 140, 100]
+ [120] * (len(leaderboard_df.columns) - 3)
+ [120, 160],
wrap = True,
)
return leaderboard |