File size: 4,595 Bytes
bd07176 |
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 97 98 99 100 101 102 103 |
import os
import time
import base64
import requests
import streamlit as st
# Access the secret API key
# if the app is running locally, you can set the API key as an environment variable
api_key = os.getenv("NVIDIA_APP_KEY")
# Function to encode the image
def encode_image(image_path):
with open(image_path, "rb") as image_file:
return base64.b64encode(image_file.read()).decode('utf-8')
# stream the response
stream = True
headers = {
"Authorization": f"Bearer {api_key}",
"Accept": "text/event-stream" if stream else "application/json"
}
def main():
st.title("Multimodal using GPT 4 Turbo Model")
text = """Prof. Louie F. Cervantes, M. Eng. (Information Engineering)
CCS 229 - Intelligent Systems
Department of Computer Science
College of Information and Communications Technology
West Visayas State University
"""
with st.expander("About"):
st.text(text)
st.write("Upload an image and select the image analysis task.")
# File upload for image
uploaded_image = st.file_uploader("Upload an Image", type=["jpg", "jpeg", "png"])
if uploaded_image is not None:
# Encode the uploaded image to base64
base64_image = base64.b64encode(uploaded_image.getvalue()).decode('utf-8')
# Display the uploaded image
st.image(uploaded_image, caption="Uploaded Image", use_container_width=True)
# List of image analysis tasks
analysis_tasks = [
"Scene Analysis: Describe the scene depicted in the image. Identify the objects present, their spatial relationships, and any actions taking place.",
"Object Detection and Classification: Identify and classify all objects present in the image. Provide detailed descriptions of each object, including its size, shape, color, and texture.",
"Image Captioning: Generate a concise and accurate caption that describes the content of the image.",
"Visual Question Answering: Answer specific questions about the image, such as 'What color is the car?' or 'How many people are in the image?'",
"Image Similarity Search: Given a query image, find similar images from a large dataset based on visual features.",
"Image Segmentation: Segment the image into different regions corresponding to objects or areas of interest.",
"Optical Character Recognition (OCR): Extract text from the image, such as printed or handwritten text.",
"Diagram Understanding: Analyze a diagram (e.g., flowchart, circuit diagram) and extract its structure and meaning.",
"Art Analysis: Describe the artistic style, subject matter, and emotional impact of an image.",
"Medical Image Analysis: Analyze medical images (e.g., X-rays, MRIs) to detect abnormalities or diagnose diseases."
]
# Task selection dropdown
selected_task = st.selectbox("Select an image analysis task:", analysis_tasks)
# Button to generate response
if st.button("Generate Response"):
if uploaded_image is None or selected_task == "":
st.error("Please upload an image and sekect a task.")
else:
# Prepare the multimodal prompt
payload = {
"model": 'meta/llama-3.2-90b-vision-instruct',
"messages": [
{
"role": "user",
"content": f'{selected_task} <img src="data:image/png;base64,{base64_image}" />'
}
],
"max_tokens": 512,
"temperature": 1.00,
"top_p": 1.00,
"stream": stream
}
with st.spinner("Processing..."):
try:
# Generate response
response = requests.post("https://ai.api.nvidia.com/v1/gr/meta/llama-3.2-90b-vision-instruct/chat/completions", headers=headers, json=payload)
# Display the response if streaming
if stream:
for line in response.iter_lines():
if line:
st.write(line.decode("utf-8"))
else:
# Show the response content
content = response.json()
contentstring = content['choices'][0]['message']['content']
st.write(f"AI Response: {contentstring}")
st.success("Response generated!")
except Exception as e:
st.error(f"An error occurred: {e}")
if __name__ == "__main__":
main() |