Yolo_v3_weight_initialization

Yolo_v3_weight_initialization.py

import numpy as np
import tensorflow as tf
import matplotlib.pyplot as plt
%matplotlib inline
plt.figure(figsize=(20,10))

x = np.random.randn(2000, 800) * 0.01 # Create input data
stds = [0.1, 0.05, 0.02, 0.01, 0.005, 0.001] # Try to use different standard deviations so that the initial weights are different

for i, std in enumerate(stds):
    # First layer - fully connected layer
    dense_1 = tf.keras.layers.Dense(750, kernel_initializer=tf.random_normal_initializer(std), activation='tanh')
    output_1 = dense_1(x)
    # Second layer - fully connected layer
    dense_2 = tf.keras.layers.Dense(700, kernel_initializer=tf.random_normal_initializer(std), activation='tanh')
    output_2 = dense_2(output_1)
    # Third layer - fully connected layer
    dense_3 = tf.keras.layers.Dense(650, kernel_initializer=tf.random_normal_initializer(std), activation='tanh')
    output_3 = dense_3(output_2).numpy().flatten()

    plt.subplot(1, len(stds), i+1)
    plt.hist(output_3, bins=600, range=[-1, 1])
    plt.xlabel('std = %.3f' %std)
    plt.yticks([])
plt.show()