带有导出的Keras模型的StandardScaler

Question

I just exported my keras model because I want to use it for my web application. For classification I'm using features from single audio files. The problem ist that with each prediction I only have one row, what makes the output of the StandardScaler 0. Is there any way to export the sklearn objekt that I can re use it in my webapp instead of using all the data from when I trained the model?

Best regards

Answer 1

I haven't used the webapp before, but I can save the trained scaler with joblib and then load it in the future.

from sklearn.preprocessing import StandardScaler
import joblib
from sklearn.datasets import load_breast_cancer

data = load_breast_cancer()

data = data.data

scaler = StandardScaler()

scaler = scaler.fit(data)

data_t = scaler.transform(data)
#saves the trained scaler
joblib.dump(scaler,r"scaler")

#Loads for transforming the new  data
scaler_2 = joblib.load(r"scaler")


data_t_2 = scaler.transform(data)