this is my flask code is not showing the prediction from the trained model

from flask import Flask, render_template, request, jsonify, session, redirect, url_for
from werkzeug.security import generate_password_hash, check_password_hash
import torch
from transformers import GPT2Tokenizer, GPT2ForSequenceClassification
import requests
import os
import logging

app = Flask(__name__, template_folder=r'D:New folderPyCharm 2023.3.3pythonProject1templates')

# Load environment variables from .env file or system environment
from dotenv import load_dotenv

load_dotenv()  # Take environment variables from .env.

app.secret_key = os.getenv('SECRET_KEY', 'default_secret_key')  # Fallback to 'default_secret_key' if not set
FDA_API_URL = 'https://api.fda.gov/drug/label.json'

# Configure logging
logging.basicConfig(level=logging.DEBUG)

# Load model and tokenizer upon server start
model_path = r'D:New folderPyCharm 2023.3.3pythonProject1templatesLastmodel2.pth'
tokenizer = GPT2Tokenizer.from_pretrained('gpt2', pad_token='<pad>')
model = GPT2ForSequenceClassification.from_pretrained('gpt2', num_labels=2)
model.resize_token_embeddings(len(tokenizer))
model.load_state_dict(torch.load(model_path, map_location=torch.device('cpu')))
model.eval()
try:
    model = GPT2ForSequenceClassification.from_pretrained('gpt2', num_labels=2)
    model.load_state_dict(torch.load(model_path, map_location=torch.device('cpu')))
    model.eval()
    print("Model loaded successfully.")
except Exception as e:
    print(f"Failed to load the model: {str(e)}")

# User authentication setup
users = {'doctor': generate_password_hash('secure123')}


def query_fda_api(drug_names):
    interactions = []
    for drug_name in drug_names:
        params = {'search': f'openfda.brand_name:"{drug_name}"', 'limit': 1}
        response = requests.get(FDA_API_URL, params=params)
        if response.status_code == 200:
            data = response.json()
            if 'results' in data and data['results'] and 'drug_interactions' in data['results'][0]:
                interactions.append(data['results'][0]['drug_interactions'])
            else:
                interactions.append("No known interactions for {}".format(drug_name))
        else:
            interactions.append("Error fetching data for {}".format(drug_name))
    return interactions if interactions else ["No known interactions."]


@app.route('/')
def home():
    return render_template('home.html')


@app.route('/scenario1', methods=['GET', 'POST'])
def scenario1():
    if request.method == 'POST':
        return predict_ddi()
    return render_template('scenario1.html')


@app.route('/predict_ddi', methods=['POST'])
def predict_ddi():
    if request.content_type != 'application/json':
        return jsonify({'error': 'Content-Type must be application/json'}), 415

    try:
        data = request.get_json()
        if not data or 'drugNames' not in data:
            return jsonify({'error': 'Missing drugNames in request'}), 400
        drug_names = data['drugNames']
        if not drug_names or not isinstance(drug_names, list):
            return jsonify({'error': 'Invalid drugNames, must be a list of names'}), 400

        inputs = tokenizer(drug_names, return_tensors="pt", padding=True, truncation=True)
        with torch.no_grad():
            outputs = model(**inputs)
            predictions = torch.argmax(outputs.logits, dim=-1)

        result = "Drug-Drug interaction detected." if predictions.item() == 1 else "No Drug-Drug interaction."
        return jsonify({'prediction': result})
    except Exception as e:
        logging.exception("An error occurred during prediction.")
        return jsonify({'error': 'Internal Server Error', 'message': str(e)}), 500
def test_model_inference():
    test_input = "Example drug interaction text"
    inputs = tokenizer(test_input, return_tensors="pt", padding=True, truncation=True)
    with torch.no_grad():
        outputs = model(**inputs)
        predictions = torch.argmax(outputs.logits, dim=-1)
    print("Test prediction:", predictions.item())

# Call this function after loading the model
test_model_inference()


@app.route('/scenario2', methods=['GET', 'POST'])
def scenario2():
    if 'authenticated' not in session:
        if request.method == 'POST' and 'username' in request.form and 'password' in request.form:
            username = request.form.get('username')
            password = request.form.get('password')
            user_hash = users.get(username)
            if user_hash and check_password_hash(user_hash, password):
                session['authenticated'] = True
                return redirect(url_for('scenario2'))  # Redirect to avoid resubmission
            else:
                return render_template('scenario2.html', error="Invalid username or password")
        return render_template('scenario2.html')

    if request.method == 'POST':
        patient_name = request.form.get('patientName')
        drug_name = request.form.get('drugName')
        if patient_name and drug_name:
            drug_names = [patient_name, drug_name]
            inputs = tokenizer(drug_names, return_tensors="pt", padding=True, truncation=True)
            with torch.no_grad():
                outputs = model(**inputs)
                predictions = torch.argmax(outputs.logits, dim=-1)
            result = "Drug-Drug interaction detected." if predictions.item() == 1 else "No Drug-Drug interaction."
            return jsonify({'prediction': result})
        return render_template('scenario2.html')

    return render_template('scenario2.html')


@app.route('/about')
def about():
    return render_template('about.html')


@app.route('/predict_pdi', methods=['POST'])
def predict_pdi():
    if 'authenticated' not in session:
        return jsonify({'error': 'User not authenticated'}), 401

    data = request.get_json(silent=True)
    if not data or 'patientName' not in data or 'drugName' not in data:
        return jsonify({'error': 'Missing patientName or drugName in request'}), 400

    patient_name = data['patientName']
    drug_name = data['drugName']
    drug_names = [patient_name, drug_name]
    inputs = tokenizer(drug_names, return_tensors="pt", padding=True, truncation=True)
    with torch.no_grad():
        outputs = model(**inputs)
        predictions = torch.argmax(outputs.logits, dim=-1)

    result = "Drug-Drug interaction detected." if predictions.item() == 1 else "No Drug-Drug interaction."
    return jsonify({'prediction': result})


if __name__ == '__main__':
    app.run(debug=True, host='0.0.0.0', port=63342)

and the html template :

<!DOCTYPE html>
<html lang="en">
<head>
    <meta charset="UTF-8">
    <meta name="viewport" content="width=device-width, initial-scale=1.0">
    <title>Service 1 - DDI Prediction Model</title>
    <style>
      
</head>
<body>

<div class="content content-english">
    <h1>Service 1: Verify Interaction Prediction Between Drugs</h1>

    <form id="drugForm">
    <div id="drugInputsEn">
        <label for="drugNameEn1">Enter drug name 1:</label>
        <input type="text" id="drugNameEn1" name="drugName1">
     </div>
     <input class="button" type="button" value="Add Drug" onclick="addDrugInput('en')">
     <br><br>
     <input class="button" type="button" value="Check DDI" onclick="checkDDI('en')">
     </form>
    <p id="predictionResultEn"></p>
    <p><a href="#" onclick="setLanguage('ar')">العربية</a></p>
</div>
<div class="content content-arabic" style="display: none;">
    <h1>الخدمة 1: التحقق من تنبؤ التفاعل بين الأدوية</h1>
    <form id="drugFormAr">
        <div id="drugInputsAr">
            <label for="drugNameAr1">ادخل اسم الدواء 1:</label>
            <input type="text" id="drugNameAr1" name="drugName1">
        </div>
        <input class="button" value="اضافة دواء" onclick="addDrugInput('ar')">
        <br><br>
        <input class="button" value="تحقق" onclick="checkDDI('ar')">
    </form>
    <p id="predictionResultAr"></p>
    <p><a href="#" onclick="setLanguage('en')">English</a></p>
</div>

<script>
    var drugCounts = {'en': 1, 'ar': 1};

    // Function to add new drug input fields dynamically
    function addDrugInput(lang) {
        drugCounts[lang]++;
        var drugInputsDiv = document.getElementById("drugInputs" + (lang === 'en' ? 'En' : 'Ar'));
        var newLabel = document.createElement("label");
        newLabel.innerHTML = lang === 'en' ? "Enter drug name " + drugCounts[lang] + ":" : "ادخل اسم الدواء " + drugCounts[lang] + ":";
        var newInput = document.createElement("input");
        newInput.type = "text";
        newInput.name = "drugName" + drugCounts[lang];
        newInput.placeholder = lang === 'en' ? "Enter drug name " + drugCounts[lang] : "ادخل اسم الدواء " + drugCounts[lang];
        newInput.id = "drugName" + (lang === 'en' ? 'En' : 'Ar') + drugCounts[lang];

        drugInputsDiv.appendChild(document.createElement("br"));
        drugInputsDiv.appendChild(newLabel);
        drugInputsDiv.appendChild(newInput);
    }

  function checkDDI(lang) {
    const drugInputsDiv = document.getElementById("drugInputs" + (lang === 'en' ? 'En' : 'Ar'));
    const inputFields = drugInputsDiv.getElementsByTagName('input');
    const drugNames = Array.from(inputFields).map(input => input.value);

    console.log("Drug names:", drugNames); // Debug log

    fetch('/predict_ddi', {
        method: 'POST',
        headers: {'Content-Type': 'application/json'},
        body: JSON.stringify({drugNames: drugNames})
    })
    .then(response => {
        if (!response.ok) {
            throw new Error('Network response was not ok ' + response.statusText);
        }
        return response.json();
    })
    .then(data => {
        const predictionResultId = "predictionResult" + (lang === 'en' ? 'En' : 'Ar');
        document.getElementById(predictionResultId).innerText = "Prediction: " + data.prediction;
        console.log("Prediction response:", data);
    })
    .catch(error => console.error('Error from server:', error));
}```

Check DDI button is not responding to send the user input to the model and get the prediction from it . so o think there is problim in sending an getting from the model

the model idea : both scenario1 need to take input to get the prediction by the trained model Lastmodel2.pth the first scenario will take Drug_Name1 + Drug_Name2 and can be more then it will get the prediction from the model we trained if no answer it will take it from FDA api to get there is intraction or not .