In [ ]:

    

import numpy as np
import tensorflow as tf
import matplotlib.pyplot as plt

# Generate a sample sin wave
x = np.array([np.linspace(0,100,1000)])
y = np.sin(x)

X = np.empty([1000,1000])
Y = np.empty([1000,1000])

for i in range(0,1000):
  X[i] = x
  Y[i] = y

# Plot the sample sin wave
plt.plot(Y[0])
plt.show()

# mirrored strategy is used to take advantage of local GPUs
mirrored_strategy = tf.distribute.MirroredStrategy()
with mirrored_strategy.scope():
  # Define a model to fit the above data
  model = tf.keras.Sequential([
      tf.keras.layers.Dropout(rate=0.2, input_shape=x.shape[1:]),
      tf.keras.layers.Dense(units=64, activation='sigmoid'),
      tf.keras.layers.Dropout(rate=0.2),
      tf.keras.layers.Dense(units=1000, activation='sigmoid')
  ])
  model.compile(loss='mse',
              optimizer='adam',
              metrics=['accuracy'])

# Fit the model
model.fit(X,Y, epochs=2000, batch_size=1024)

# Get the prediction
y_hat = model.predict(X)

# Plot the results
plt.plot(y_hat[0])
plt.show()


    

In [ ]: