효과적인 Tensorflow 2

import tensorflow as tf
import tensorflow_datasets as tfds

2022-12-14 21:55:10.524042: W tensorflow/compiler/xla/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libnvinfer.so.7'; dlerror: libnvinfer.so.7: cannot open shared object file: No such file or directory
2022-12-14 21:55:10.524133: W tensorflow/compiler/xla/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libnvinfer_plugin.so.7'; dlerror: libnvinfer_plugin.so.7: cannot open shared object file: No such file or directory
2022-12-14 21:55:10.524142: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Cannot dlopen some TensorRT libraries. If you would like to use Nvidia GPU with TensorRT, please make sure the missing libraries mentioned above are installed properly.

datasets, info = tfds.load(name='mnist', with_info=True, as_supervised=True)
mnist_train, mnist_test = datasets['train'], datasets['test']

BUFFER_SIZE = 10 # Use a much larger value for real code
BATCH_SIZE = 64
NUM_EPOCHS = 5


def scale(image, label):
  image = tf.cast(image, tf.float32)
  image /= 255

  return image, label

train_data = mnist_train.map(scale).shuffle(BUFFER_SIZE).batch(BATCH_SIZE)
test_data = mnist_test.map(scale).batch(BATCH_SIZE)

STEPS_PER_EPOCH = 5

train_data = train_data.take(STEPS_PER_EPOCH)
test_data = test_data.take(STEPS_PER_EPOCH)

image_batch, label_batch = next(iter(train_data))

2022-12-14 21:55:16.307777: W tensorflow/core/kernels/data/cache_dataset_ops.cc:856] The calling iterator did not fully read the dataset being cached. In order to avoid unexpected truncation of the dataset, the partially cached contents of the dataset  will be discarded. This can happen if you have an input pipeline similar to `dataset.cache().take(k).repeat()`. You should use `dataset.take(k).cache().repeat()` instead.

@tf.function
def train(model, dataset, optimizer):
  for x, y in dataset:
    with tf.GradientTape() as tape:
      # training=True is only needed if there are layers with different
      # behavior during training versus inference (e.g. Dropout).
      prediction = model(x, training=True)
      loss = loss_fn(prediction, y)
    gradients = tape.gradient(loss, model.trainable_variables)
    optimizer.apply_gradients(zip(gradients, model.trainable_variables))

model.compile(optimizer=optimizer, loss=loss_fn)
model.fit(dataset)

model = tf.keras.Sequential([
    tf.keras.layers.Conv2D(32, 3, activation='relu',
                           kernel_regularizer=tf.keras.regularizers.l2(0.02),
                           input_shape=(28, 28, 1)),
    tf.keras.layers.MaxPooling2D(),
    tf.keras.layers.Flatten(),
    tf.keras.layers.Dropout(0.1),
    tf.keras.layers.Dense(64, activation='relu'),
    tf.keras.layers.BatchNormalization(),
    tf.keras.layers.Dense(10)
])

# Model is the full model w/o custom layers
model.compile(optimizer='adam',
              loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),
              metrics=['accuracy'])

model.fit(train_data, epochs=NUM_EPOCHS)
loss, acc = model.evaluate(test_data)

print("Loss {}, Accuracy {}".format(loss, acc))

Epoch 1/5
5/5 [==============================] - 3s 6ms/step - loss: 1.5373 - accuracy: 0.5000
Epoch 2/5
2022-12-14 21:55:19.700111: W tensorflow/core/kernels/data/cache_dataset_ops.cc:856] The calling iterator did not fully read the dataset being cached. In order to avoid unexpected truncation of the dataset, the partially cached contents of the dataset  will be discarded. This can happen if you have an input pipeline similar to `dataset.cache().take(k).repeat()`. You should use `dataset.take(k).cache().repeat()` instead.
5/5 [==============================] - 0s 5ms/step - loss: 0.4361 - accuracy: 0.9312
Epoch 3/5
2022-12-14 21:55:19.979427: W tensorflow/core/kernels/data/cache_dataset_ops.cc:856] The calling iterator did not fully read the dataset being cached. In order to avoid unexpected truncation of the dataset, the partially cached contents of the dataset  will be discarded. This can happen if you have an input pipeline similar to `dataset.cache().take(k).repeat()`. You should use `dataset.take(k).cache().repeat()` instead.
5/5 [==============================] - 0s 5ms/step - loss: 0.2826 - accuracy: 0.9656
Epoch 4/5
2022-12-14 21:55:20.289201: W tensorflow/core/kernels/data/cache_dataset_ops.cc:856] The calling iterator did not fully read the dataset being cached. In order to avoid unexpected truncation of the dataset, the partially cached contents of the dataset  will be discarded. This can happen if you have an input pipeline similar to `dataset.cache().take(k).repeat()`. You should use `dataset.take(k).cache().repeat()` instead.
5/5 [==============================] - 0s 5ms/step - loss: 0.1974 - accuracy: 0.9875
Epoch 5/5
2022-12-14 21:55:20.576118: W tensorflow/core/kernels/data/cache_dataset_ops.cc:856] The calling iterator did not fully read the dataset being cached. In order to avoid unexpected truncation of the dataset, the partially cached contents of the dataset  will be discarded. This can happen if you have an input pipeline similar to `dataset.cache().take(k).repeat()`. You should use `dataset.take(k).cache().repeat()` instead.
5/5 [==============================] - 0s 5ms/step - loss: 0.1499 - accuracy: 1.0000
2022-12-14 21:55:20.848689: W tensorflow/core/kernels/data/cache_dataset_ops.cc:856] The calling iterator did not fully read the dataset being cached. In order to avoid unexpected truncation of the dataset, the partially cached contents of the dataset  will be discarded. This can happen if you have an input pipeline similar to `dataset.cache().take(k).repeat()`. You should use `dataset.take(k).cache().repeat()` instead.
5/5 [==============================] - 0s 3ms/step - loss: 1.6309 - accuracy: 0.7531
Loss 1.6308777332305908, Accuracy 0.753125011920929
2022-12-14 21:55:21.212331: W tensorflow/core/kernels/data/cache_dataset_ops.cc:856] The calling iterator did not fully read the dataset being cached. In order to avoid unexpected truncation of the dataset, the partially cached contents of the dataset  will be discarded. This can happen if you have an input pipeline similar to `dataset.cache().take(k).repeat()`. You should use `dataset.take(k).cache().repeat()` instead.

model = tf.keras.Sequential([
    tf.keras.layers.Conv2D(32, 3, activation='relu',
                           kernel_regularizer=tf.keras.regularizers.l2(0.02),
                           input_shape=(28, 28, 1)),
    tf.keras.layers.MaxPooling2D(),
    tf.keras.layers.Flatten(),
    tf.keras.layers.Dropout(0.1),
    tf.keras.layers.Dense(64, activation='relu'),
    tf.keras.layers.BatchNormalization(),
    tf.keras.layers.Dense(10)
])

optimizer = tf.keras.optimizers.Adam(0.001)
loss_fn = tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True)

@tf.function
def train_step(inputs, labels):
  with tf.GradientTape() as tape:
    predictions = model(inputs, training=True)
    regularization_loss=tf.math.add_n(model.losses)
    pred_loss=loss_fn(labels, predictions)
    total_loss=pred_loss + regularization_loss

  gradients = tape.gradient(total_loss, model.trainable_variables)
  optimizer.apply_gradients(zip(gradients, model.trainable_variables))

for epoch in range(NUM_EPOCHS):
  for inputs, labels in train_data:
    train_step(inputs, labels)
  print("Finished epoch", epoch)

2022-12-14 21:55:22.897407: W tensorflow/core/kernels/data/cache_dataset_ops.cc:856] The calling iterator did not fully read the dataset being cached. In order to avoid unexpected truncation of the dataset, the partially cached contents of the dataset  will be discarded. This can happen if you have an input pipeline similar to `dataset.cache().take(k).repeat()`. You should use `dataset.take(k).cache().repeat()` instead.
Finished epoch 0
2022-12-14 21:55:23.185574: W tensorflow/core/kernels/data/cache_dataset_ops.cc:856] The calling iterator did not fully read the dataset being cached. In order to avoid unexpected truncation of the dataset, the partially cached contents of the dataset  will be discarded. This can happen if you have an input pipeline similar to `dataset.cache().take(k).repeat()`. You should use `dataset.take(k).cache().repeat()` instead.
Finished epoch 1
2022-12-14 21:55:23.467763: W tensorflow/core/kernels/data/cache_dataset_ops.cc:856] The calling iterator did not fully read the dataset being cached. In order to avoid unexpected truncation of the dataset, the partially cached contents of the dataset  will be discarded. This can happen if you have an input pipeline similar to `dataset.cache().take(k).repeat()`. You should use `dataset.take(k).cache().repeat()` instead.
Finished epoch 2
2022-12-14 21:55:23.738925: W tensorflow/core/kernels/data/cache_dataset_ops.cc:856] The calling iterator did not fully read the dataset being cached. In order to avoid unexpected truncation of the dataset, the partially cached contents of the dataset  will be discarded. This can happen if you have an input pipeline similar to `dataset.cache().take(k).repeat()`. You should use `dataset.take(k).cache().repeat()` instead.
Finished epoch 3
Finished epoch 4
2022-12-14 21:55:24.029871: W tensorflow/core/kernels/data/cache_dataset_ops.cc:856] The calling iterator did not fully read the dataset being cached. In order to avoid unexpected truncation of the dataset, the partially cached contents of the dataset  will be discarded. This can happen if you have an input pipeline similar to `dataset.cache().take(k).repeat()`. You should use `dataset.take(k).cache().repeat()` instead.

class DynamicRNN(tf.keras.Model):

  def __init__(self, rnn_cell):
    super(DynamicRNN, self).__init__(self)
    self.cell = rnn_cell

  @tf.function(input_signature=[tf.TensorSpec(dtype=tf.float32, shape=[None, None, 3])])
  def call(self, input_data):

    # [batch, time, features] -> [time, batch, features]
    input_data = tf.transpose(input_data, [1, 0, 2])
    timesteps =  tf.shape(input_data)[0]
    batch_size = tf.shape(input_data)[1]
    outputs = tf.TensorArray(tf.float32, timesteps)
    state = self.cell.get_initial_state(batch_size = batch_size, dtype=tf.float32)
    for i in tf.range(timesteps):
      output, state = self.cell(input_data[i], state)
      outputs = outputs.write(i, output)
    return tf.transpose(outputs.stack(), [1, 0, 2]), state

lstm_cell = tf.keras.layers.LSTMCell(units = 13)

my_rnn = DynamicRNN(lstm_cell)
outputs, state = my_rnn(tf.random.normal(shape=[10,20,3]))
print(outputs.shape)

(10, 20, 13)

cce = tf.keras.losses.CategoricalCrossentropy(from_logits=True)
cce([[1, 0]], [[-1.0,3.0]]).numpy()

4.01815

summary_writer = tf.summary.create_file_writer('/tmp/summaries')
with summary_writer.as_default():
  tf.summary.scalar('loss', 0.1, step=42)

def train(model, optimizer, dataset, log_freq=10):
  avg_loss = tf.keras.metrics.Mean(name='loss', dtype=tf.float32)
  for images, labels in dataset:
    loss = train_step(model, optimizer, images, labels)
    avg_loss.update_state(loss)
    if tf.equal(optimizer.iterations % log_freq, 0):
      tf.summary.scalar('loss', avg_loss.result(), step=optimizer.iterations)
      avg_loss.reset_states()

def test(model, test_x, test_y, step_num):
  # training=False is only needed if there are layers with different
  # behavior during training versus inference (e.g. Dropout).
  loss = loss_fn(model(test_x, training=False), test_y)
  tf.summary.scalar('loss', loss, step=step_num)

train_summary_writer = tf.summary.create_file_writer('/tmp/summaries/train')
test_summary_writer = tf.summary.create_file_writer('/tmp/summaries/test')

with train_summary_writer.as_default():
  train(model, optimizer, dataset)

with test_summary_writer.as_default():
  test(model, test_x, test_y, optimizer.iterations)

tensorboard --logdir /tmp/summaries

# Create the metrics
loss_metric = tf.keras.metrics.Mean(name='train_loss')
accuracy_metric = tf.keras.metrics.SparseCategoricalAccuracy(name='train_accuracy')

@tf.function
def train_step(inputs, labels):
  with tf.GradientTape() as tape:
    predictions = model(inputs, training=True)
    regularization_loss=tf.math.add_n(model.losses)
    pred_loss=loss_fn(labels, predictions)
    total_loss=pred_loss + regularization_loss

  gradients = tape.gradient(total_loss, model.trainable_variables)
  optimizer.apply_gradients(zip(gradients, model.trainable_variables))
  # Update the metrics
  loss_metric.update_state(total_loss)
  accuracy_metric.update_state(labels, predictions)


for epoch in range(NUM_EPOCHS):
  # Reset the metrics
  loss_metric.reset_states()
  accuracy_metric.reset_states()

  for inputs, labels in train_data:
    train_step(inputs, labels)
  # Get the metric results
  mean_loss=loss_metric.result()
  mean_accuracy = accuracy_metric.result()

  print('Epoch: ', epoch)
  print('  loss:     {:.3f}'.format(mean_loss))
  print('  accuracy: {:.3f}'.format(mean_accuracy))

2022-12-14 21:55:24.963973: W tensorflow/core/kernels/data/cache_dataset_ops.cc:856] The calling iterator did not fully read the dataset being cached. In order to avoid unexpected truncation of the dataset, the partially cached contents of the dataset  will be discarded. This can happen if you have an input pipeline similar to `dataset.cache().take(k).repeat()`. You should use `dataset.take(k).cache().repeat()` instead.
Epoch:  0
  loss:     0.133
  accuracy: 0.991
2022-12-14 21:55:25.238740: W tensorflow/core/kernels/data/cache_dataset_ops.cc:856] The calling iterator did not fully read the dataset being cached. In order to avoid unexpected truncation of the dataset, the partially cached contents of the dataset  will be discarded. This can happen if you have an input pipeline similar to `dataset.cache().take(k).repeat()`. You should use `dataset.take(k).cache().repeat()` instead.
Epoch:  1
  loss:     0.113
  accuracy: 1.000
2022-12-14 21:55:25.538122: W tensorflow/core/kernels/data/cache_dataset_ops.cc:856] The calling iterator did not fully read the dataset being cached. In order to avoid unexpected truncation of the dataset, the partially cached contents of the dataset  will be discarded. This can happen if you have an input pipeline similar to `dataset.cache().take(k).repeat()`. You should use `dataset.take(k).cache().repeat()` instead.
Epoch:  2
  loss:     0.101
  accuracy: 1.000
2022-12-14 21:55:25.844832: W tensorflow/core/kernels/data/cache_dataset_ops.cc:856] The calling iterator did not fully read the dataset being cached. In order to avoid unexpected truncation of the dataset, the partially cached contents of the dataset  will be discarded. This can happen if you have an input pipeline similar to `dataset.cache().take(k).repeat()`. You should use `dataset.take(k).cache().repeat()` instead.
Epoch:  3
  loss:     0.088
  accuracy: 1.000
Epoch:  4
  loss:     0.079
  accuracy: 1.000
2022-12-14 21:55:26.131845: W tensorflow/core/kernels/data/cache_dataset_ops.cc:856] The calling iterator did not fully read the dataset being cached. In order to avoid unexpected truncation of the dataset, the partially cached contents of the dataset  will be discarded. This can happen if you have an input pipeline similar to `dataset.cache().take(k).repeat()`. You should use `dataset.take(k).cache().repeat()` instead.

model.compile(
    optimizer = tf.keras.optimizers.Adam(0.001),
    loss = tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),
    metrics = ['acc', 'accuracy', tf.keras.metrics.SparseCategoricalAccuracy(name="my_accuracy")])
history = model.fit(train_data)

5/5 [==============================] - 2s 5ms/step - loss: 0.0902 - acc: 0.9969 - accuracy: 0.9969 - my_accuracy: 0.9969
2022-12-14 21:55:27.804136: W tensorflow/core/kernels/data/cache_dataset_ops.cc:856] The calling iterator did not fully read the dataset being cached. In order to avoid unexpected truncation of the dataset, the partially cached contents of the dataset  will be discarded. This can happen if you have an input pipeline similar to `dataset.cache().take(k).repeat()`. You should use `dataset.take(k).cache().repeat()` instead.

history.history.keys()

dict_keys(['loss', 'acc', 'accuracy', 'my_accuracy'])

@tf.function
def f(x):
  if x > 0:
    import pdb
    pdb.set_trace()
    x = x + 1
  return x

tf.config.run_functions_eagerly(True)
f(tf.constant(1))

>>> f()
-> x = x + 1
(Pdb) l
  6     @tf.function
  7     def f(x):
  8       if x > 0:
  9         import pdb
 10         pdb.set_trace()
 11  ->     x = x + 1
 12       return x
 13
 14     tf.config.run_functions_eagerly(True)
 15     f(tf.constant(1))
[EOF]

class CustomModel(tf.keras.models.Model):

  @tf.function
  def call(self, input_data):
    if tf.reduce_mean(input_data) > 0:
      return input_data
    else:
      import pdb
      pdb.set_trace()
      return input_data // 2


tf.config.run_functions_eagerly(True)
model = CustomModel()
model(tf.constant([-2, -4]))

>>> call()
-> return input_data // 2
(Pdb) l
 10         if tf.reduce_mean(input_data) > 0:
 11           return input_data
 12         else:
 13           import pdb
 14           pdb.set_trace()
 15  ->       return input_data // 2
 16
 17
 18     tf.config.run_functions_eagerly(True)
 19     model = CustomModel()
 20     model(tf.constant([-2, -4]))