app.py

from flask import Flask, request, jsonify
from flask_cors import CORS
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
import logging
import os

# Initialize logger
logging.basicConfig(level=logging.DEBUG)

# Load tokenizer and model
logging.info("Loading model...")
model_repo = "hsb06/toghetherAi-model"
tokenizer = AutoTokenizer.from_pretrained(model_repo)
model = AutoModelForCausalLM.from_pretrained(model_repo, torch_dtype=torch.float16).to("cuda" if torch.cuda.is_available() else "cpu")
logging.info("Model loaded successfully.")

# Initialize Flask app
app = Flask(__name__)
CORS(app)

def generate_response(prompt):
    inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
    input_length = inputs.input_ids.shape[1]
    outputs = model.generate(
        **inputs,
        max_new_tokens=128,
        do_sample=True,
        temperature=0.7,
        top_p=0.7,
        top_k=50,
        return_dict_in_generate=True,
    )
    token = outputs.sequences[0, input_length:]
    return tokenizer.decode(token, skip_special_tokens=True)

@app.route("/", methods=["GET"])
def home():
    return jsonify({"message": "Flask app is running!"})

@app.route("/chat", methods=["POST"])
def chat():
    data = request.json
    user_input = data.get("message", "")
    prompt = f"<human>: {user_input}\n<bot>:"  # Ensure prompt format matches your model's expectation
    response = generate_response(prompt)
    return jsonify({"response": response})

if __name__ == "__main__":
    from os import getenv
    app.run(host="0.0.0.0", port=int(getenv("PORT", 8080)))