KB-VQA-E / app.py
m7mdal7aj's picture
Update app.py
4c2fb5b verified
raw
history blame
7.06 kB
import streamlit as st
import torch
import bitsandbytes
import accelerate
import scipy
import copy
from PIL import Image
import torch.nn as nn
from my_model.object_detection import detect_and_draw_objects
from my_model.captioner.image_captioning import get_caption
from my_model.utilities import free_gpu_resources
from my_model.KBVQA import KBVQA, prepare_kbvqa_model
def answer_question(image, question, caption, detected_objects_str, model):
answer = model.generate_answer(question, caption, detected_objects_str)
st.image(image)
st.write(caption)
st.write("----------------")
st.write(detected_objects_str)
return answer
def get_caption(image):
return "Generated caption for the image"
def free_gpu_resources():
pass
# Sample images (assuming these are paths to your sample images)
sample_images = ["Files/sample1.jpg", "Files/sample2.jpg", "Files/sample3.jpg",
"Files/sample4.jpg", "Files/sample5.jpg", "Files/sample6.jpg",
"Files/sample7.jpg"]
def analyze_image(image, model, show_processed_image=False):
img = copy.deepcopy(image)
caption = model.get_caption(img)
image_with_boxes, detected_objects_str = model.detect_objects(img)
if show_processed_image:
st.image(image_with_boxes)
return caption, detected_objects_str
def image_qa_app(kbvqa):
# Initialize session state for storing the current image and its Q&A history
if 'current_image' not in st.session_state:
st.session_state['current_image'] = None
if 'qa_history' not in st.session_state:
st.session_state['qa_history'] = []
if 'analysis_done' not in st.session_state:
st.session_state['analysis_done'] = False
if 'answer_in_progress' not in st.session_state:
st.session_state['answer_in_progress'] = False
# Display sample images as clickable thumbnails
st.write("Choose from sample images:")
cols = st.columns(len(sample_images))
for idx, sample_image_path in enumerate(sample_images):
with cols[idx]:
image = Image.open(sample_image_path)
st.image(image, use_column_width=True)
if st.button(f'Select Sample Image {idx + 1}', key=f'sample_{idx}'):
st.session_state['current_image'] = image
st.session_state['qa_history'] = []
st.session_state['analysis_done'] = False
st.session_state['answer_in_progress'] = False
# Image uploader
uploaded_image = st.file_uploader("Or upload an Image", type=["png", "jpg", "jpeg"])
if uploaded_image is not None:
image = Image.open(uploaded_image)
st.session_state['current_image'] = image
st.session_state['qa_history'] = []
st.session_state['analysis_done'] = False
st.session_state['answer_in_progress'] = False
if st.session_state.get('current_image') and not st.session_state.get('analysis_done', False):
if st.button('Analyze Image'):
caption, detected_objects_str = analyze_image(st.session_state['current_image'], kbvqa)
st.session_state['caption'] = caption
st.session_state['detected_objects_str'] = detected_objects_str
st.session_state['analysis_done'] = True
# Get Answer button
if st.session_state.get('analysis_done', False):
question = st.text_input("Ask a question about this image:")
if st.button('Get Answer'):
answer = answer_question(st.session_state['current_image'], question, st.session_state.get('caption', ''), st.session_state.get('detected_objects_str', ''), kbvqa)
st.session_state['qa_history'].append((question, answer))
# Display all Q&A
for q, a in st.session_state.get('qa_history', []):
st.text(f"Q: {q}\nA: {a}\n")
# Reset the answer_in_progress flag after displaying the answer
if st.session_state['answer_in_progress']:
st.session_state['answer_in_progress'] = False
def run_inference():
st.title("Run Inference")
method = st.selectbox(
"Choose a method:",
["Fine-Tuned Model", "In-Context Learning (n-shots)"],
index=0 # Default to the first option
)
detection_model = st.selectbox(
"Choose a model for object detection:",
["yolov5", "detic"],
index=0 # Default to the first option
)
# Initialize session state for the model
if method == "Fine-Tuned Model":
if 'kbvqa' not in st.session_state:
st.session_state['kbvqa'] = None
# Button to load KBVQA models
if st.button('Load KBVQA Model'):
if st.session_state['kbvqa'] is not None:
st.write("Model already loaded.")
else:
# Call the function to load models and show progress
st.session_state['kbvqa'] = prepare_kbvqa_model(detection_model)
if st.session_state['kbvqa']:
st.write("Model is ready for inference.")
# Set default confidence based on the selected model
default_confidence = 0.2 if detection_model == "yolov5" else 0.4
# Slider for confidence level
confidence_level = st.slider(
"Select Detection Confidence Level",
min_value=0.1,
max_value=0.9,
value=default_confidence,
step=0.1
)
st.session_state['kbvqa'].detection_confidence = confidence_level
if st.session_state['kbvqa']:
image_qa_app(st.session_state['kbvqa'])
else:
st.write(f'{method} model is not ready for inference yet')
# Main function
def main():
st.sidebar.title("Navigation")
selection = st.sidebar.radio("Go to", ["Home", "Dataset Analysis", "Evaluation Results", "Run Inference", "Dissertation Report"])
if selection == "Home":
st.title("MultiModal Learning for Knowledg-Based Visual Question Answering")
st.write("Home page content goes here...")
elif selection == "Dissertation Report":
st.title("Dissertation Report")
st.write("Click the link below to view the PDF.")
# Example to display a link to a PDF
st.download_button(
label="Download PDF",
data=open("Files/Dissertation Report.pdf", "rb"),
file_name="example.pdf",
mime="application/octet-stream"
)
elif selection == "Evaluation Results":
st.title("Evaluation Results")
st.write("This is a Place Holder until the contents are uploaded.")
elif selection == "Dataset Analysis":
st.title("OK-VQA Dataset Analysis")
st.write("This is a Place Holder until the contents are uploaded.")
elif selection == "Run Inference":
run_inference()
elif selection == "Object Detection":
run_object_detection()
if __name__ == "__main__":
main()