import streamlit as st
from streamlit_extras.switch_page_button import switch_page
translations = {
'en': {'title': 'MiniGemini',
'original_tweet':
"""
[Original tweet](https://x.com/mervenoyann/status/1783864388249694520) (April 26, 2024)
""",
'tweet_1':
"""
MiniGemini is the coolest VLM, let's explain 🧶
""",
'tweet_2':
"""
MiniGemini is a vision language model that understands both image and text and also generates text and an image that goes best with the context! 🤯
""",
'tweet_3':
"""
This model has two image encoders (one CNN and one ViT) in parallel to capture the details in the images.
I saw the same design in DocOwl 1.5 then it has a decoder to output text and also a prompt to be sent to SDXL for image generation (which works very well!)
""",
'tweet_4':
"""
They adopt CLIP's ViT for low resolution visual embedding encoder and a CNN-based one for high resolution image encoding (precisely a pre-trained ConvNeXt).
""",
'tweet_5':
"""
Thanks to the second encoder it can grasp details in images, which also comes in handy for e.g. document tasks (but see below the examples are mindblowing IMO).
""",
'tweet_6':
"""
According to their reporting the model performs very well across many benchmarks compared to LLaVA 1.5 and Gemini Pro.
""",
'ressources':
"""
Resources:
[Mini-Gemini: Mining the Potential of Multi-modality Vision Language Models](https://huggingface.co/papers/2403.18814)
by Yanwei Li, Yuechen Zhang, Chengyao Wang, Zhisheng Zhong, Yixin Chen, Ruihang Chu, Shaoteng Liu, Jiaya Jia (2024)
[GitHub](https://github.com/dvlab-research/MGM)
[Model Repository](https://huggingface.co/YanweiLi/MGM-13B-HD)
"""
},
'fr': {
'title': 'MiniGemini',
'original_tweet':
"""
[Tweet de base](https://x.com/mervenoyann/status/1783864388249694520) (26 avril 2024)
""",
'tweet_1':
"""
MiniGemini est le VLM le plus cool, voici pourquoi 🧶
""",
'tweet_2':
"""
MiniGemini est un modèle de langage/vision qui comprend à la fois l'image et le texte et qui génère également le texte et l'image qui s'accordent le mieux avec le contexte ! 🤯 """,
'tweet_3':
"""
Ce modèle possède deux encodeurs d'images (un ConvNet et un ViT) en parallèle pour capturer les détails dans les images.
J'ai vu la même conception dans DocOwl 1.5 où il y a un décodeur pour produire du texte et aussi un prompt à envoyer au SDXL pour la génération d'images (qui fonctionne très bien !). """,
'tweet_4':
"""
Les auteurs adoptent le ViT de CLIP pour les enchâssements visuels de basse résolution et un ConvNet pour les images en haute résolution (précisément un ConvNeXt pré-entraîné).
""",
'tweet_5':
"""
Grâce au second encodeur, il peut saisir des détails dans les images, ce qui s'avère également utile pour les tâches documentaires (voir ci-dessous les exemples époustouflants). """,
'tweet_6':
"""
D'après leur rapport, le modèle est très performant dans de nombreux benchmarks par rapport à LLaVA 1.5 et Gemini Pro.
""",
'ressources':
"""
Resources :
[Mini-Gemini: Mining the Potential of Multi-modality Vision Language Models](https://huggingface.co/papers/2403.18814)
de Yanwei Li, Yuechen Zhang, Chengyao Wang, Zhisheng Zhong, Yixin Chen, Ruihang Chu, Shaoteng Liu, Jiaya Jia (2024)
[GitHub](https://github.com/dvlab-research/MGM)
[Modèle](https://huggingface.co/YanweiLi/MGM-13B-HD)
"""
}
}
def language_selector():
languages = {'EN': '🇬🇧', 'FR': '🇫🇷'}
selected_lang = st.selectbox('', options=list(languages.keys()), format_func=lambda x: languages[x], key='lang_selector')
return 'en' if selected_lang == 'EN' else 'fr'
left_column, right_column = st.columns([5, 1])
# Add a selector to the right column
with right_column:
lang = language_selector()
# Add a title to the left column
with left_column:
st.title(translations[lang]["title"])
st.success(translations[lang]["original_tweet"], icon="ℹ️")
st.markdown(""" """)
st.markdown(translations[lang]["tweet_1"], unsafe_allow_html=True)
st.markdown(""" """)
st.image("pages/MiniGemini/image_1.jpg", use_column_width=True)
st.markdown(""" """)
st.markdown(translations[lang]["tweet_2"], unsafe_allow_html=True)
st.markdown(""" """)
st.image("pages/MiniGemini/image_2.jpg", use_column_width=True)
st.markdown(""" """)
st.markdown(translations[lang]["tweet_3"], unsafe_allow_html=True)
st.markdown(""" """)
st.image("pages/MiniGemini/image_3.jpg", use_column_width=True)
st.markdown(""" """)
st.markdown(translations[lang]["tweet_4"], unsafe_allow_html=True)
st.markdown(""" """)
st.image("pages/MiniGemini/image_4.jpg", use_column_width=True)
st.markdown(""" """)
st.markdown(translations[lang]["tweet_5"], unsafe_allow_html=True)
st.markdown(""" """)
st.image("pages/MiniGemini/image_5.jpg", use_column_width=True)
st.markdown(""" """)
st.markdown(translations[lang]["tweet_6"], unsafe_allow_html=True)
st.markdown(""" """)
st.image("pages/MiniGemini/image_6.jpg", use_column_width=True)
st.markdown(""" """)
st.info(translations[lang]["ressources"], icon="📚")
st.markdown(""" """)
st.markdown(""" """)
st.markdown(""" """)
col1, col2, col3= st.columns(3)
with col1:
if lang == "en":
if st.button('Previous paper', use_container_width=True):
switch_page("DocOwl 1.5")
else:
if st.button('Papier précédent', use_container_width=True):
switch_page("DocOwl 1.5")
with col2:
if lang == "en":
if st.button("Home", use_container_width=True):
switch_page("Home")
else:
if st.button("Accueil", use_container_width=True):
switch_page("Home")
with col3:
if lang == "en":
if st.button("Next paper", use_container_width=True):
switch_page("CuMo")
else:
if st.button("Papier suivant", use_container_width=True):
switch_page("PLLaVA")