Spaces:
Running
Running
Upload folder using huggingface_hub
Browse files
app/__pycache__/draw_diagram.cpython-312.pyc
CHANGED
|
Binary files a/app/__pycache__/draw_diagram.cpython-312.pyc and b/app/__pycache__/draw_diagram.cpython-312.pyc differ
|
|
|
app/__pycache__/pages.cpython-312.pyc
CHANGED
|
Binary files a/app/__pycache__/pages.cpython-312.pyc and b/app/__pycache__/pages.cpython-312.pyc differ
|
|
|
app/pages.py
CHANGED
|
@@ -2,16 +2,72 @@ import streamlit as st
|
|
| 2 |
from app.draw_diagram import *
|
| 3 |
|
| 4 |
def dashboard():
|
| 5 |
-
|
| 6 |
-
|
| 7 |
-
|
| 8 |
-
|
| 9 |
-
|
| 10 |
-
|
|
|
|
|
|
|
|
|
|
| 11 |
|
| 12 |
seaeval_url = "https://seaeval.github.io/"
|
| 13 |
-
|
| 14 |
-
st.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 15 |
|
| 16 |
def cross_lingual_consistency():
|
| 17 |
st.title("Cross-Lingual Consistency")
|
|
|
|
| 2 |
from app.draw_diagram import *
|
| 3 |
|
| 4 |
def dashboard():
|
| 5 |
+
|
| 6 |
+
with st.container():
|
| 7 |
+
st.title("SeaEval")
|
| 8 |
+
|
| 9 |
+
st.markdown("""
|
| 10 |
+
[gh]: https://github.com/SeaEval/SeaEval
|
| 11 |
+
[][gh]
|
| 12 |
+
[][gh]
|
| 13 |
+
""")
|
| 14 |
|
| 15 |
seaeval_url = "https://seaeval.github.io/"
|
| 16 |
+
|
| 17 |
+
st.divider()
|
| 18 |
+
st.markdown("#### What is [SeaEval](%s)" % seaeval_url)
|
| 19 |
+
|
| 20 |
+
with st.container():
|
| 21 |
+
left_co, cent_co,last_co = st.columns(3)
|
| 22 |
+
with cent_co:
|
| 23 |
+
st.image("./style/seaeval_overall.png",
|
| 24 |
+
# caption="SeaEval data range",
|
| 25 |
+
width=500)
|
| 26 |
+
st.markdown('''
|
| 27 |
+
|
| 28 |
+
''')
|
| 29 |
+
st.markdown("##### A new benchmark for multilingual foundation models consisting of 28 dataset.")
|
| 30 |
+
st.markdown(''':star: How models understand and reason with natural language?
|
| 31 |
+
:balloon: Languages: English, Chinese, Malay, Spainish, Indonedian, Vietnamese, Filipino.
|
| 32 |
+
''')
|
| 33 |
+
|
| 34 |
+
st.markdown(''':star: How models comprehend cultural practices, nuances and values?
|
| 35 |
+
:balloon: 4 new datasets on Cultural Understanding.
|
| 36 |
+
''')
|
| 37 |
+
|
| 38 |
+
st.markdown(''':star: How models perform across languages in terms of consistency?
|
| 39 |
+
:balloon: 2 new datasets with curated metrics for Cross-Linugal Consistency.
|
| 40 |
+
''')
|
| 41 |
+
with st.container():
|
| 42 |
+
left_co, cent_co,last_co = st.columns(3)
|
| 43 |
+
with cent_co:
|
| 44 |
+
st.image("./style/consistency.png",
|
| 45 |
+
# caption="SeaEval data range",
|
| 46 |
+
width=500)
|
| 47 |
+
st.markdown("##### Evaluation with enhanced cross-lingual capabilities.")
|
| 48 |
+
st.markdown(''':star: How models perform according to different (paraphrased) instructions?
|
| 49 |
+
:balloon: Each dataset is equipped with 5 different prompts to avoid randomness introduced by instructions,
|
| 50 |
+
which is non-negligible..
|
| 51 |
+
''')
|
| 52 |
+
|
| 53 |
+
st.markdown(''':star: Multilingual accuracy and performance consistency across languages.
|
| 54 |
+
:balloon: If you can answer the question in your native language, can you answer the same question
|
| 55 |
+
correctly in your second/third language?
|
| 56 |
+
''')
|
| 57 |
+
|
| 58 |
+
st.divider()
|
| 59 |
+
with st.container():
|
| 60 |
+
st.markdown("##### Citations")
|
| 61 |
+
|
| 62 |
+
st.markdown('''
|
| 63 |
+
:round_pushpin: SeaEval Paper \n
|
| 64 |
+
@article{SeaEval2023,
|
| 65 |
+
title={SeaEval for Multilingual Foundation Models: From Cross-Lingual Alignment to Cultural Reasoning},
|
| 66 |
+
author={Wang, Bin and Liu, Zhengyuan and Huang, Xin and Jiao, Fangkai and Ding, Yang and Aw, Ai Ti and Chen, Nancy F},
|
| 67 |
+
journal={arXiv preprint arXiv:2309.04766},
|
| 68 |
+
year={2023}
|
| 69 |
+
}
|
| 70 |
+
''')
|
| 71 |
|
| 72 |
def cross_lingual_consistency():
|
| 73 |
st.title("Cross-Lingual Consistency")
|