this happens. it is really common when you use augmentation on the training data, but also sometimes the test set ends up being a little easier.
#Train Accuracy vs Test Accuracy
willow roost
twin meadow
so is it a bad thing or should i look for any issues in my actual model creation?
willow roost
Really it’s a good thing, if anything. A lot of times my computer vision models start like this because of augmentation then by the end it will be equal out training is a little better. So I wouldn’t worry unless it’s like a really big difference