Feat. req: Save model on ctrl-c key press

Sample code to do this:

```
import signal
import sys

def signal_handler(sig, frame):
print('You pressed Ctrl+C!')
# Insert additional function execution here
run_before_exit()
sys.exit(0)

def run_before_exit():
print('Saving model before exiting.')
save_model_before_exiting()

signal.signal(signal.SIGINT, signal_handler)
```

This would be a great new feature and would allow to interrupt training without losing the trained model at that time.

The issue provides a Python signal-handler example using SIGINT and describes saving the model before exiting training. Start by locating the training entry point and determining where model checkpoints are saved. Done means pressing Ctrl+C preserves the currently trained model before the process exits.

Written by the indexing model from the issue text.