float16 example

Author: zsdonghao

example/tutorial_mnist_float16.py

@@ -0,0 +1,104 @@
+#! /usr/bin/python
+# -*- coding: utf8 -*-
+
+import numpy as np
+import tensorflow as tf
+import tensorlayer as tl
+from tensorlayer.layers import *
+import time
+
+X_train, y_train, X_val, y_val, X_test, y_test = \
+                tl.files.load_mnist_dataset(shape=(-1, 28, 28, 1))
+
+sess = tf.InteractiveSession()
+
+batch_size = 128
+D_TYPE = tf.float32 # tf.float32  tf.float16
+
+x = tf.placeholder(D_TYPE, shape=[batch_size, 28, 28, 1])
+y_ = tf.placeholder(tf.int64, shape=[batch_size,])
+
+def model(x, is_train=True, reuse=False):
+    with tf.variable_scope("model", reuse=reuse):
+        tl.layers.set_name_reuse(reuse)
+        n = InputLayer(x, name='input')
+        # cnn
+        n = Conv2d(n, 32, (5, 5), (1, 1), padding='SAME',
+            W_init_args={'dtype': D_TYPE}, b_init_args={'dtype': D_TYPE}, name='cnn1')
+        n = BatchNormLayer(n, act=tf.nn.relu, is_train=is_train, dtype=D_TYPE, name='bn1')
+        n = MaxPool2d(n, (2, 2), (2, 2), padding='SAME', name='pool1')
+        n = Conv2d(n, 64, (5, 5), (1, 1), padding='SAME',
+            W_init_args={'dtype': D_TYPE}, b_init_args={'dtype': D_TYPE}, name='cnn2')
+        n = BatchNormLayer(n, act=tf.nn.relu, is_train=is_train, dtype=D_TYPE, name='bn2')
+        n = MaxPool2d(n, (2, 2), (2, 2), padding='SAME', name='pool2')
+        # mlp
+        n = FlattenLayer(n, name='flatten')
+        n = DropoutLayer(n, 0.5, True, is_train, name='drop1')
+        n = DenseLayer(n, 256, act=tf.nn.relu,
+            W_init_args={'dtype': D_TYPE}, b_init_args={'dtype': D_TYPE}, name='relu1')
+        n = DropoutLayer(n, 0.5, True, is_train, name='drop2')
+        n = DenseLayer(n, 10, act=tf.identity,
+            W_init_args={'dtype': D_TYPE}, b_init_args={'dtype': D_TYPE}, name='output')
+    return n
+
+# define inferences
+net_train = model(x, is_train=True, reuse=False)
+net_test = model(x, is_train=False, reuse=True)
+
+net_train.print_params(False)
+
+# cost for training
+y = net_train.outputs
+cost = tl.cost.cross_entropy(y, y_, name='xentropy')
+
+# cost and accuracy for evalution
+y2 = net_test.outputs
+cost_test = tl.cost.cross_entropy(y2, y_, name='xentropy2')
+correct_prediction = tf.equal(tf.argmax(y2, 1), y_)
+acc = tf.reduce_mean(tf.cast(correct_prediction, D_TYPE))
+
+# define the optimizer
+train_params = tl.layers.get_variables_with_name('model', train_only=True, printable=False)
+train_op = tf.train.AdamOptimizer(learning_rate=0.0001, beta1=0.9, beta2=0.999,
+                # epsilon=1e-08,    # for float32 as default
+                epsilon=1e-4,       # for float16, see https://stackoverflow.com/questions/42064941/tensorflow-float16-support-is-broken
+                use_locking=False).minimize(cost, var_list=train_params)
+
+# initialize all variables in the session
+tl.layers.initialize_global_variables(sess)
+
+# train the network
+n_epoch = 500
+print_freq = 1
+
+for epoch in range(n_epoch):
+    start_time = time.time()
+    for X_train_a, y_train_a in tl.iterate.minibatches(
+                                X_train, y_train, batch_size, shuffle=True):
+        sess.run(train_op, feed_dict={x: X_train_a, y_: y_train_a})
+
+    if epoch + 1 == 1 or (epoch + 1) % print_freq == 0:
+        print("Epoch %d of %d took %fs" % (epoch + 1, n_epoch, time.time() - start_time))
+        train_loss, train_acc, n_batch = 0, 0, 0
+        for X_train_a, y_train_a in tl.iterate.minibatches(
+                                X_train, y_train, batch_size, shuffle=True):
+            err, ac = sess.run([cost_test, acc], feed_dict={x: X_train_a, y_: y_train_a})
+            train_loss += err; train_acc += ac; n_batch += 1
+        print("   train loss: %f" % (train_loss/ n_batch))
+        print("   train acc: %f" % (train_acc/ n_batch))
+        val_loss, val_acc, n_batch = 0, 0, 0
+        for X_val_a, y_val_a in tl.iterate.minibatches(
+                                    X_val, y_val, batch_size, shuffle=True):
+            err, ac = sess.run([cost_test, acc], feed_dict={x: X_val_a, y_: y_val_a})
+            val_loss += err; val_acc += ac; n_batch += 1
+        print("   val loss: %f" % (val_loss/ n_batch))
+        print("   val acc: %f" % (val_acc/ n_batch))
+
+print('Evaluation')
+test_loss, test_acc, n_batch = 0, 0, 0
+for X_test_a, y_test_a in tl.iterate.minibatches(
+                            X_test, y_test, batch_size, shuffle=True):
+    err, ac = sess.run([cost_test, acc], feed_dict={x: X_test_a, y_: y_test_a})
+    test_loss += err; test_acc += ac; n_batch += 1
+print("   test loss: %f" % (test_loss/n_batch))
+print("   test acc: %f" % (test_acc/n_batch))