Skip to content

Instantly share code, notes, and snippets.

@Jeksonic
Forked from 3h4/1-10-vanilla-rnn.py
Last active February 3, 2019 22:08
Show Gist options
  • Save Jeksonic/7d08d107aa02b2cb207a5571506d14e8 to your computer and use it in GitHub Desktop.
Save Jeksonic/7d08d107aa02b2cb207a5571506d14e8 to your computer and use it in GitHub Desktop.
from __future__ import print_function, division
import numpy as np
import tensorflow as tf
import matplotlib.pyplot as plt
# hyperparameters
state_size = 4 # size of hidden layer of neurons
seq_length = 15 # number of steps to unroll the RNN for
batch_size = 5 # number of samples to learn simultaneously
learn_rate = 3e-1 # learning rate
num_epochs = 100
total_series_length = 50000
num_classes = 2
num_features = 1
echo_step = 3
def generateData():
x = np.array(np.random.choice(2, total_series_length, p=[0.5, 0.5]))
y = np.roll(x, echo_step)
y[0:echo_step] = 0
x = x.reshape((batch_size, num_features, -1))
y = y.reshape((batch_size, -1))
return (x, y)
def plot(loss_list, predictions_series, batchX, batchY):
plt.subplot(2, 3, 1)
plt.cla()
plt.plot(loss_list)
for batch_series_idx in range(batch_size):
one_hot_output_series = np.array(predictions_series)[:, batch_series_idx, :]
single_output_series = np.array([(1 if out[0] < 0.5 else 0) for out in one_hot_output_series])
plt.subplot(2, 3, batch_series_idx + 2)
plt.cla()
plt.axis([0, seq_length, 0, 2])
left_offset = range(seq_length)
plt.bar(left_offset, batchX[batch_series_idx, :], width=1, color="blue")
plt.bar(left_offset, batchY[batch_series_idx, :] * 0.5, width=1, color="red")
plt.bar(left_offset, single_output_series * 0.3, width=1, color="green")
plt.draw()
plt.pause(0.0001)
def run_vanilla_basicRNN(inputs_series, init_state):
W = tf.Variable(np.random.rand(state_size + num_features, state_size), dtype=tf.float32)
b = tf.Variable(np.zeros((num_features, state_size)), dtype=tf.float32)
current_state = init_state
states_series = []
for current_input in inputs_series:
input_and_state_concatenated = tf.concat([current_input, current_state], 1) # Increasing number of columns
next_state = tf.tanh(tf.matmul(input_and_state_concatenated, W) + b) # Broadcasted addition
states_series.append(next_state)
current_state = next_state
return states_series, current_state
tf.reset_default_graph()
batchX_placeholder = tf.placeholder(tf.float32, [batch_size, num_features, seq_length])
batchY_placeholder = tf.placeholder(tf.int32, [batch_size, seq_length])
init_state = tf.placeholder(tf.float32, [batch_size, state_size])
W = tf.Variable(np.random.rand(state_size, num_classes),dtype=tf.float32)
b = tf.Variable(np.zeros((1,num_classes)), dtype=tf.float32)
# Unpack columns
inputs_series = tf.unstack(batchX_placeholder, axis=2)
labels_series = tf.unstack(batchY_placeholder, axis=1)
# Forward pass
states_series, current_state = run_vanilla_basicRNN(inputs_series, init_state)
logits_series = [tf.matmul(state, W) + b for state in states_series] #Broadcasted addition
predictions_series = [tf.nn.softmax(logits) for logits in logits_series]
losses = [tf.nn.sparse_softmax_cross_entropy_with_logits(logits=logits, labels=labels) for logits, labels in zip(logits_series,labels_series)]
total_loss = tf.reduce_mean(losses)
train_step = tf.train.AdagradOptimizer(learn_rate).minimize(total_loss)
with tf.Session() as sess:
sess.run(tf.initialize_all_variables())
plt.ion()
plt.figure()
plt.show()
loss_list = []
num_batches = total_series_length//batch_size//seq_length
for epoch_idx in range(num_epochs):
x,y = generateData()
_current_state = np.zeros((batch_size, state_size))
print("New data, epoch", epoch_idx)
for batch_idx in range(num_batches):
start_idx = batch_idx * seq_length
end_idx = start_idx + seq_length
batchX = x[:,:,start_idx:end_idx]
batchY = y[:,start_idx:end_idx]
_total_loss, _train_step, _current_state, _predictions_series = sess.run(
[total_loss, train_step, current_state, predictions_series],
feed_dict={
batchX_placeholder:batchX,
batchY_placeholder:batchY,
init_state:_current_state
})
loss_list.append(_total_loss)
if batch_idx%100 == 0:
print("Step",batch_idx, "Loss", _total_loss)
plot(loss_list, _predictions_series, batchX[:,0,:], batchY) # plot only first feature
plt.ioff()
plt.show()
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment