app.py

from flask import Flask, request, jsonify
import torch
from PIL import Image
from transformers import AutoModel, AutoTokenizer, BitsAndBytesConfig

# Get API token from environment variable
api_token = os.getenv("HF_TOKEN").strip()

# Model configuration
bnb_config = BitsAndBytesConfig(
    load_in_4bit=True, 
    bnb_4bit_quant_type="nf4", 
    bnb_4bit_use_double_quant=True, 
    bnb_4bit_compute_dtype=torch.float16,
)

# Model and tokenizer loading
model = AutoModel.from_pretrained(
    "ContactDoctor/Bio-Medical-MultiModal-Llama-3-8B-V1", 
    quantization_config=bnb_config, 
    device_map="auto", 
    torch_dtype=torch.float16, 
    trust_remote_code=True, 
    attn_implementation="flash_attention_2",
)
tokenizer = AutoTokenizer.from_pretrained(
    "ContactDoctor/Bio-Medical-MultiModal-Llama-3-8B-V1", 
    trust_remote_code=True
)



app = Flask(__name__)

# Model configuration and loading (same as before)

@app.route('/analyze', methods=['POST'])
def analyze():
    image = request.files['image']
    question = request.form['question']

    # Preprocess image
    image = Image.open(image).convert('RGB')

    # Prepare input
    msgs = [{'role': 'user', 'content': [image, question]}]

    # Generate response
    res = model.chat(
        image=image, 
        msgs=msgs, 
        tokenizer=tokenizer, 
        sampling=True, 
        temperature=0.95, 
        stream=True
    )

    # Process response
    generated_text = ""
    for new_text in res:
        generated_text += new_text

    return jsonify({'response': generated_text})

if __name__ == '__main__':
    app.run(debug=True)