Spaces:
Running
Running
File size: 2,869 Bytes
c8b83ad 604b05d c8b83ad 604b05d c8b83ad |
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 |
import os
# from dotenv import load_dotenv
import streamlit as st
import PIL.Image
import google.generativeai as genai
from langchain.prompts import ChatPromptTemplate
from langchain_community.llms import Ollama
from transformers import MllamaForConditionalGeneration, AutoProcessor
import torch
from accelerate import init_empty_weights
# Load environment variables
from transformers import AutoProcessor, AutoModelForPreTraining
# Configure Gemini API
# genai.configure(api_key=os.getenv("gkey2"))
# Define the prompt template
# prompt = ChatPromptTemplate.from_messages(
# [
# ("system", "You are a helpful assistant. Please respond to the user's queries."),
# ("user", "Question: {question}")
# ]
# )
prompt="<|image|><|begin_of_text|>You are a helpful assistant. Please respond to the user's queries."
# Initialize the Llama model
# model = Ollama(model="llama3.2")
model_id = "meta-llama/Llama-3.2-11B-Vision"
# model = MllamaForConditionalGeneration.from_pretrained(
# model_id,
# torch_dtype=torch.bfloat16,
# device_map="auto",
# )
# processor = AutoProcessor.from_pretrained(model_id)
processor = AutoProcessor.from_pretrained("meta-llama/Llama-3.2-11B-Vision")
model = AutoModelForPreTraining.from_pretrained("meta-llama/Llama-3.2-11B-Vision")
# Define function to get response from the model
def get_gemin_response(input_text, img):
# complete_prompt = prompt.format(question=input_text)
inputs = processor(images=img, text=prompt, return_tensors="pt").to(model.device)
response=model.generate(**inputs, max_new_tokens=30)
# if input_text != "":
# # Only generate content from input text if present
# response = model.generate([input_text])
# else:
# response = model.generate([img_text])
return response
# Define the main function for the Streamlit app
def main():
st.set_page_config(page_title='Gemini Image & Text')
st.header('Gemini LLM Application')
# Input text
input_text = st.text_input("Input :", key='input')
# Image uploader
imgupload = st.file_uploader('Choose an image file', type=['jpg', 'jpeg', 'png'])
# Display uploaded image and convert to text format (if needed)
img_text = ""
if imgupload is not None:
img = PIL.Image.open(imgupload)
st.image(img, caption='Uploaded Image', use_column_width=True)
img_text = "Image uploaded successfully."
if st.button('Generate Response'):
# Ensure both inputs are provided
if img is not None and input_text:
# Get response from the model
response = get_gemin_response(input_text, img)
st.write(processor.decode(response[0]))
else:
st.error("Please provide both input text and an image before generating a response.")
# Run the app
if __name__ == "__main__":
main() |