Is it possible to terminate an AI training process in django?

I am developing a webpage using Django. In the admin site, there is a section called: Train Model, in here, there is a button called "Train Model" this button runs a python script that has a simulation of a training process, it is just a sleep(30) for now, while the sleeping time is happening the user should be able no cancel this process, in case the user forgot something, is it possible to do so? I have tried some things, but the logs show that even if I press the cancel button, the "training" process runs until it is completed.

import time
import multiprocessing
import os
import signal
import logging

# Global variables to manage the training state
is_training = False
training_process = None
training_pid = None  # Store the PID of the training process
stop_event = None
logger = logging.getLogger()


def train_task(stop_event):
    """simulates the training task with stop event support"""
    for i in range(30):
        if stop_event.is_set():
            print("Training cancelled.")
            return
        time.sleep(1)
    print("Training done.")

def train_model():
    global is_training, training_process, stop_event

    # If training is already in progress, return immediately
    if is_training:
        print("Training already in progress...")
        return "Training already in progress"

    #mark the start of the training
    is_training = True
    stop_event = multiprocessing.Event() #creates a stop event
    print("Training started...")

    #start the training process with stop event support
    training_process = multiprocessing.Process(target = train_task, args=(stop_event,))
    training_process.start()
    #stores the pid of the process
    training_pid = training_process.pid

    return "Training started."

def cancel_training_process():
    global training_pid, is_training, stop_event

    if is_training and training_pid:
        try:
            os.kill(training_pid, signal.SIGTERM) #sends the SIGTERM signal to the process
            is_training = False #resets the training state
            training_pid = None
            stop_event.set() # stops the training event
            return "Training cancelled."
        except OSError as e:
            return f"Error while cancelling the process: {e}"

        return "No training in progress"






def check_training_status():
    global is_training
    if is_training:
        return "Training in progress"
    else:
        return "No training in progress"

Those are the functions that simulates the training and try to cancel the process.

Any advice in this regard? In future, the training will be a whole different script that will grab data from a data base and it will train an AI, but such a script is not yet written, this is just me trying to implement what I was requested to do hehehe

I tried different solutions, none of the seem to be working, when I want to cancel the process it still runs until the end, I am sure there might be a solution, but this is definetely out of my knowledge for now

Back to Top