The Validation results are odd, but the confusion matrix for the new dataset is okay, it would be ideal if I can improve on these results…
This is the training
This is the test
Validation
New Data