Feature scaling for LSTM time series classification

Question

I am trying to prepare input data for LSTM time series classification model. I wanted to scale the features using sklearn MaxAbsScaler(), but I was having a hard time to incorporate scaling in my following code. I am using tensorflow implementation of LSTM network.

import keras
from keras.layers import LSTM, Dropout, Dense
import tensorflow as tf
import numpy as np
import pandas as pd
df = pd.read_excel('/content/hdds.xlsx')

def generate_data(X, y, sequence_length=2, step = 1):
    X_local = []
    y_local = []
    for start in range(0, len(df) - sequence_length, step):
        end = start + sequence_length
        X_local.append(X[start:end])
        y_local.append(y[end-1])
    return np.array(X_local), np.array(y_local)

X_sequence, y = generate_data(df.loc[:, "V1":"V3"].values, df.Class)

X_sequence.shape, y.shape
((16, 2, 3), (16,))

training_size = int(len(X_sequence) * 0.7)
X_train, y_train = X_sequence[:training_size], y[:training_size]
X_test, y_test = X_sequence[training_size:], y[training_size:]

X_train.shape, X_test.shape
((11, 2, 3), (5, 2, 3))

I intended to use the following codes to scale the input data, but it is not working. I appreciate your suggestions. Thanks!

from sklearn.preprocessing import MaxAbsScaler
scaler =  MaxAbsScaler()
X_train = scaler.fit_transform(X_train)
X_test = scaler.transform(X_test)
X_train = np.clip(X_train, -5, 5)
X_test = np.clip(X_test, -5, 5)

Sample data: