I trained a Keras model that gives good validation accuracy.
After saving and loading it, the predictions become completely wrong.
Even training samples are misclassified.
Nothing crashes, but the outputs no longer make sense.
Lost your password? Please enter your email address. You will receive a link and will create a new password via email.
Please briefly explain why you feel this question should be reported.
Please briefly explain why you feel this answer should be reported.
Please briefly explain why you feel this user should be reported.
I trained a Keras model that gives good validation accuracy.
After saving and loading it, the predictions become completely wrong.
Even training samples are misclassified.
Nothing crashes, but the outputs no longer make sense.
This happens because the preprocessing used during inference does not match the preprocessing used during training.
Neural networks learn patterns in the numerical space they were trained on. If you normalize, tokenize, or scale data during training but skip or change it when running inference, the model sees completely unfamiliar values and produces garbage outputs.
You must save and reuse the exact same preprocessing objects — scalers, tokenizers, and transforms — along with the model. For example, in Keras:
joblib.dump(scaler, "scaler.pkl")
...
scaler = joblib.load("scaler.pkl")
X = scaler.transform(X)
The same applies to image transforms and text tokenizers. Even a small difference like missing standardization will break predictions.