Fix contents
Browse files- app.py +7 -4
- content.py +36 -0
app.py
CHANGED
@@ -3,7 +3,7 @@ import json
|
|
3 |
import glob
|
4 |
from collections import defaultdict
|
5 |
import gradio as gr
|
6 |
-
|
7 |
import glob
|
8 |
|
9 |
ARC = "arc_challenge"
|
@@ -80,7 +80,7 @@ LANG_COL = "Language"
|
|
80 |
AVERAGE_COL = "Average"
|
81 |
ARC_COL = "ARC (25-shot)"
|
82 |
HELLASWAG_COL = "HellaSwag (10-shot)️"
|
83 |
-
MMLU_COL = "MMLU (5-shot)
|
84 |
TRUTHFULQA_COL = "TruthfulQA (0-shot)"
|
85 |
|
86 |
COLS = [MODEL_COL, LANG_COL, AVERAGE_COL, ARC_COL, HELLASWAG_COL, MMLU_COL, TRUTHFULQA_COL]
|
@@ -91,8 +91,9 @@ leaderboard_df = get_leaderboard_df(*args)
|
|
91 |
|
92 |
demo = gr.Blocks()
|
93 |
with demo:
|
94 |
-
gr.HTML(
|
95 |
-
|
|
|
96 |
|
97 |
with gr.Box():
|
98 |
search_bar = gr.Textbox(
|
@@ -107,4 +108,6 @@ with demo:
|
|
107 |
elem_id="leaderboard-table",
|
108 |
)
|
109 |
|
|
|
|
|
110 |
demo.launch()
|
|
|
3 |
import glob
|
4 |
from collections import defaultdict
|
5 |
import gradio as gr
|
6 |
+
from content import *
|
7 |
import glob
|
8 |
|
9 |
ARC = "arc_challenge"
|
|
|
80 |
AVERAGE_COL = "Average"
|
81 |
ARC_COL = "ARC (25-shot)"
|
82 |
HELLASWAG_COL = "HellaSwag (10-shot)️"
|
83 |
+
MMLU_COL = "MMLU (5-shot)"
|
84 |
TRUTHFULQA_COL = "TruthfulQA (0-shot)"
|
85 |
|
86 |
COLS = [MODEL_COL, LANG_COL, AVERAGE_COL, ARC_COL, HELLASWAG_COL, MMLU_COL, TRUTHFULQA_COL]
|
|
|
91 |
|
92 |
demo = gr.Blocks()
|
93 |
with demo:
|
94 |
+
gr.HTML(TITLE)
|
95 |
+
gr.Markdown(INTRO_TEXT, elem_classes="markdown-text")
|
96 |
+
gr.Markdown(HOW_TO, elem_classes="markdown-text")
|
97 |
|
98 |
with gr.Box():
|
99 |
search_bar = gr.Textbox(
|
|
|
108 |
elem_id="leaderboard-table",
|
109 |
)
|
110 |
|
111 |
+
gr.Markdown(CITATION, elem_classes="markdown-text")
|
112 |
+
|
113 |
demo.launch()
|
content.py
ADDED
@@ -0,0 +1,36 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
TITLE = '<h1 align="center" id="space-title">Open Multilingual LLM Evaluation Leaderboard</h1>'
|
2 |
+
|
3 |
+
INTRO_TEXT = f"""
|
4 |
+
## About
|
5 |
+
|
6 |
+
This leaderboard shows the performance of pretrained models in 29 languages on four benchmarks:
|
7 |
+
|
8 |
+
- <a href="https://arxiv.org/abs/1803.05457" target="_blank"> AI2 Reasoning Challenge </a> (25-shot)
|
9 |
+
- <a href="https://arxiv.org/abs/1905.07830" target="_blank"> HellaSwag </a> (10-shot)
|
10 |
+
- <a href="https://arxiv.org/abs/2009.03300" target="_blank"> MMLU </a> (5-shot)
|
11 |
+
- <a href="https://arxiv.org/abs/2109.07958" target="_blank"> TruthfulQA </a> (0-shot)
|
12 |
+
|
13 |
+
The evaluation data was translated into 29 languages using ChatGPT.
|
14 |
+
|
15 |
+
"""
|
16 |
+
|
17 |
+
HOW_TO = f"""
|
18 |
+
## How to list your model performance on this leaderboard:
|
19 |
+
|
20 |
+
Send an email with title [Open mLLM Loaderboard] to vietl@uoregon.edu with the huggingface's model name.
|
21 |
+
|
22 |
+
We will run your model on the four benchmarks and add it to the leaderboard.
|
23 |
+
"""
|
24 |
+
|
25 |
+
CITATION = f"""
|
26 |
+
## Citation
|
27 |
+
|
28 |
+
```
|
29 |
+
|
30 |
+
@misc{{lai2023openllmbenchmark,
|
31 |
+
author = {{Viet Lai and Nghia Trung Ngo and Amir Pouran Ben Veyseh and Franck Dernoncourt and Thien Huu Nguyen}},
|
32 |
+
title={{Open Multilingual LLM Evaluation Leaderboard}},
|
33 |
+
year={{2023}}
|
34 |
+
}}
|
35 |
+
```
|
36 |
+
"""
|