Entrenamiento distribuido con Keras

Ver en TensorFlow.org Ejecutar en Google Colab Ver fuente en GitHub Descargar cuaderno

Visión general

El tf.distribute.Strategy API proporciona una abstracción para la distribución de su formación a través de múltiples unidades de procesamiento. Le permite realizar entrenamiento distribuido utilizando modelos existentes y código de entrenamiento con cambios mínimos.

En este tutorial se muestra cómo utilizar el tf.distribute.MirroredStrategy para llevar a cabo la replicación en el gráfico con la formación sincrónica en muchas GPU en una máquina. La estrategia esencialmente copia todas las variables del modelo a cada procesador. A continuación, utiliza todo reducir combinar los gradientes de todos los procesadores, y se aplica el valor combinado de todas las copias del modelo.

Que va a utilizar las tf.keras API para construir el modelo y Model.fit para la formación de la misma. (Para obtener información acerca de la formación distribuida con un bucle de entrenamiento personalizado y el MirroredStrategy , echa un vistazo a este tutorial .)

MirroredStrategy entrena su modelo en varias GPU en una sola máquina. Para la formación sincrónica en muchas GPU en múltiples trabajadores, utilice el tf.distribute.MultiWorkerMirroredStrategy con el Model.fit Keras o un bucle de formación personalizada . Para otras opciones, consulte la guía de capacitación distribuida .

Para obtener información sobre otras estrategias diversas, existe la formación distribuida con TensorFlow guía.

Configuración

import tensorflow_datasets as tfds
import tensorflow as tf

import os

# Load the TensorBoard notebook extension.
%load_ext tensorboard
2021-08-04 01:24:55.165631: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Successfully opened dynamic library libcudart.so.11.0
print(tf.__version__)
2.5.0

Descarga el conjunto de datos

Cargar el conjunto de datos de MNIST TensorFlow conjuntos de datos . Esto devuelve un conjunto de datos en el tf.data formato.

Ajuste de la with_info argumento para True incluye los metadatos para todo el conjunto de datos, que se está guardando aquí para info . Entre otras cosas, este objeto de metadatos incluye el número de ejemplos de prueba y entrenamiento.

datasets, info = tfds.load(name='mnist', with_info=True, as_supervised=True)

mnist_train, mnist_test = datasets['train'], datasets['test']
2021-08-04 01:25:00.048530: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Successfully opened dynamic library libcuda.so.1
2021-08-04 01:25:00.691099: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2021-08-04 01:25:00.691993: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1733] Found device 0 with properties: 
pciBusID: 0000:00:05.0 name: Tesla V100-SXM2-16GB computeCapability: 7.0
coreClock: 1.53GHz coreCount: 80 deviceMemorySize: 15.78GiB deviceMemoryBandwidth: 836.37GiB/s
2021-08-04 01:25:00.692033: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Successfully opened dynamic library libcudart.so.11.0
2021-08-04 01:25:00.695439: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Successfully opened dynamic library libcublas.so.11
2021-08-04 01:25:00.695536: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Successfully opened dynamic library libcublasLt.so.11
2021-08-04 01:25:00.696685: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Successfully opened dynamic library libcufft.so.10
2021-08-04 01:25:00.697009: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Successfully opened dynamic library libcurand.so.10
2021-08-04 01:25:00.698067: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Successfully opened dynamic library libcusolver.so.11
2021-08-04 01:25:00.698998: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Successfully opened dynamic library libcusparse.so.11
2021-08-04 01:25:00.699164: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Successfully opened dynamic library libcudnn.so.8
2021-08-04 01:25:00.699264: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2021-08-04 01:25:00.700264: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2021-08-04 01:25:00.701157: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1871] Adding visible gpu devices: 0
2021-08-04 01:25:00.701928: I tensorflow/core/platform/cpu_feature_guard.cc:142] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 AVX512F FMA
To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.
2021-08-04 01:25:00.702642: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2021-08-04 01:25:00.703535: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1733] Found device 0 with properties: 
pciBusID: 0000:00:05.0 name: Tesla V100-SXM2-16GB computeCapability: 7.0
coreClock: 1.53GHz coreCount: 80 deviceMemorySize: 15.78GiB deviceMemoryBandwidth: 836.37GiB/s
2021-08-04 01:25:00.703621: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2021-08-04 01:25:00.704507: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2021-08-04 01:25:00.705349: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1871] Adding visible gpu devices: 0
2021-08-04 01:25:00.705388: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Successfully opened dynamic library libcudart.so.11.0
2021-08-04 01:25:01.356483: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1258] Device interconnect StreamExecutor with strength 1 edge matrix:
2021-08-04 01:25:01.356521: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1264]      0 
2021-08-04 01:25:01.356530: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1277] 0:   N 
2021-08-04 01:25:01.356777: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2021-08-04 01:25:01.357792: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2021-08-04 01:25:01.358756: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2021-08-04 01:25:01.359641: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1418] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 14646 MB memory) -> physical GPU (device: 0, name: Tesla V100-SXM2-16GB, pci bus id: 0000:00:05.0, compute capability: 7.0)

Definir la estrategia de distribución

Crear un MirroredStrategy objeto. Este se encargará de la distribución y proporcionar un gestor de contexto ( MirroredStrategy.scope ) para construir su interior modelo.

strategy = tf.distribute.MirroredStrategy()
WARNING:tensorflow:Collective ops is not configured at program startup. Some performance features may not be enabled.
WARNING:tensorflow:Collective ops is not configured at program startup. Some performance features may not be enabled.
INFO:tensorflow:Using MirroredStrategy with devices ('/job:localhost/replica:0/task:0/device:GPU:0',)
INFO:tensorflow:Using MirroredStrategy with devices ('/job:localhost/replica:0/task:0/device:GPU:0',)
print('Number of devices: {}'.format(strategy.num_replicas_in_sync))
Number of devices: 1

Configurar la canalización de entrada

Al entrenar un modelo con varias GPU, puede utilizar la potencia de cálculo adicional de forma eficaz aumentando el tamaño del lote. En general, utilice el tamaño de lote más grande que se ajuste a la memoria de la GPU y ajuste la velocidad de aprendizaje en consecuencia.

# You can also do info.splits.total_num_examples to get the total
# number of examples in the dataset.

num_train_examples = info.splits['train'].num_examples
num_test_examples = info.splits['test'].num_examples

BUFFER_SIZE = 10000

BATCH_SIZE_PER_REPLICA = 64
BATCH_SIZE = BATCH_SIZE_PER_REPLICA * strategy.num_replicas_in_sync

Definir una función que normaliza los valores de pixel de imagen a partir de la [0, 255] gama a la [0, 1] rango ( escala característica ):

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

  return image, label

Aplicar esta scale la función de los datos de entrenamiento y de prueba, y luego usar los tf.data.Dataset API para mezclar los datos de entrenamiento ( Dataset.shuffle ), y el lote que ( Dataset.batch ). Tenga en cuenta que también se están manteniendo un caché en memoria de los datos de entrenamiento para mejorar el rendimiento ( Dataset.cache ).

train_dataset = mnist_train.map(scale).cache().shuffle(BUFFER_SIZE).batch(BATCH_SIZE)
eval_dataset = mnist_test.map(scale).batch(BATCH_SIZE)

Crea el modelo

Crear y compilar el modelo Keras en el contexto de Strategy.scope :

with strategy.scope():
  model = tf.keras.Sequential([
      tf.keras.layers.Conv2D(32, 3, activation='relu', input_shape=(28, 28, 1)),
      tf.keras.layers.MaxPooling2D(),
      tf.keras.layers.Flatten(),
      tf.keras.layers.Dense(64, activation='relu'),
      tf.keras.layers.Dense(10)
  ])

  model.compile(loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),
                optimizer=tf.keras.optimizers.Adam(),
                metrics=['accuracy'])
INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).
INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).
INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).
INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).
INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).
INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).
INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).
INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).

Definir las devoluciones de llamada

Definir los siguientes tf.keras.callbacks :

Para fines ilustrativos, añadir una devolución de llamada personalizado llamado PrintLR para mostrar la tasa de aprendizaje en el cuaderno.

# Define the checkpoint directory to store the checkpoints.
checkpoint_dir = './training_checkpoints'
# Define the name of the checkpoint files.
checkpoint_prefix = os.path.join(checkpoint_dir, "ckpt_{epoch}")
# Define a function for decaying the learning rate.
# You can define any decay function you need.
def decay(epoch):
  if epoch < 3:
    return 1e-3
  elif epoch >= 3 and epoch < 7:
    return 1e-4
  else:
    return 1e-5
# Define a callback for printing the learning rate at the end of each epoch.
class PrintLR(tf.keras.callbacks.Callback):
  def on_epoch_end(self, epoch, logs=None):
    print('\nLearning rate for epoch {} is {}'.format(epoch + 1,
                                                      model.optimizer.lr.numpy()))
# Put all the callbacks together.
callbacks = [
    tf.keras.callbacks.TensorBoard(log_dir='./logs'),
    tf.keras.callbacks.ModelCheckpoint(filepath=checkpoint_prefix,
                                       save_weights_only=True),
    tf.keras.callbacks.LearningRateScheduler(decay),
    PrintLR()
]
2021-08-04 01:25:02.054144: I tensorflow/core/profiler/lib/profiler_session.cc:126] Profiler session initializing.
2021-08-04 01:25:02.054179: I tensorflow/core/profiler/lib/profiler_session.cc:141] Profiler session started.
2021-08-04 01:25:02.054232: I tensorflow/core/profiler/internal/gpu/cupti_tracer.cc:1611] Profiler found 1 GPUs
2021-08-04 01:25:02.098001: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Successfully opened dynamic library libcupti.so.11.2
2021-08-04 01:25:02.288095: I tensorflow/core/profiler/lib/profiler_session.cc:159] Profiler session tear down.
2021-08-04 01:25:02.292220: I tensorflow/core/profiler/internal/gpu/cupti_tracer.cc:1743] CUPTI activity buffer flushed

Capacitar y evaluar

Ahora, entrenar el modelo de la forma habitual, llamando Model.fit en el modelo y que pasa en el conjunto de datos creada al principio del tutorial. Este paso es el mismo tanto si distribuye la formación como si no.

EPOCHS = 12

model.fit(train_dataset, epochs=EPOCHS, callbacks=callbacks)
2021-08-04 01:25:02.342811: W tensorflow/core/grappler/optimizers/data/auto_shard.cc:461] The `assert_cardinality` transformation is currently not handled by the auto-shard rewrite and will be removed.
2021-08-04 01:25:02.389307: I tensorflow/compiler/mlir/mlir_graph_optimization_pass.cc:176] None of the MLIR Optimization Passes are enabled (registered 2)
2021-08-04 01:25:02.389734: I tensorflow/core/platform/profile_utils/cpu_utils.cc:114] CPU Frequency: 2000179999 Hz
Epoch 1/12
INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).
INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).
INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).
INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).
INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).
INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).
INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).
INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).
INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).
INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).
INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).
INFO:tensorflow:Reduce to /job:localhost/replica:0/task:0/device:CPU:0 then broadcast to ('/job:localhost/replica:0/task:0/device:CPU:0',).
2021-08-04 01:25:05.851687: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Successfully opened dynamic library libcudnn.so.8
2021-08-04 01:25:07.965516: I tensorflow/stream_executor/cuda/cuda_dnn.cc:359] Loaded cuDNN version 8100
2021-08-04 01:25:13.166255: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Successfully opened dynamic library libcublas.so.11
2021-08-04 01:25:13.566160: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Successfully opened dynamic library libcublasLt.so.11
1/938 [..............................] - ETA: 3:09:47 - loss: 2.2850 - accuracy: 0.1094
2021-08-04 01:25:14.615346: I tensorflow/core/profiler/lib/profiler_session.cc:126] Profiler session initializing.
2021-08-04 01:25:14.615388: I tensorflow/core/profiler/lib/profiler_session.cc:141] Profiler session started.
3/938 [..............................] - ETA: 4:21 - loss: 2.1694 - accuracy: 0.3333WARNING:tensorflow:Callback method `on_train_batch_begin` is slow compared to the batch time (batch time: 0.0045s vs `on_train_batch_begin` time: 0.0762s). Check your callbacks.
2021-08-04 01:25:15.082713: I tensorflow/core/profiler/lib/profiler_session.cc:66] Profiler session collecting data.
2021-08-04 01:25:15.085886: I tensorflow/core/profiler/internal/gpu/cupti_tracer.cc:1743] CUPTI activity buffer flushed
2021-08-04 01:25:15.122453: I tensorflow/core/profiler/internal/gpu/cupti_collector.cc:673]  GpuTracer has collected 96 callback api events and 93 activity events. 
2021-08-04 01:25:15.126946: I tensorflow/core/profiler/lib/profiler_session.cc:159] Profiler session tear down.
2021-08-04 01:25:15.138108: I tensorflow/core/profiler/rpc/client/save_profile.cc:137] Creating directory: ./logs/train/plugins/profile/2021_08_04_01_25_15
2021-08-04 01:25:15.146767: I tensorflow/core/profiler/rpc/client/save_profile.cc:143] Dumped gzipped tool data for trace.json.gz to ./logs/train/plugins/profile/2021_08_04_01_25_15/kokoro-gcp-ubuntu-prod-1251741625.trace.json.gz
2021-08-04 01:25:15.154434: I tensorflow/core/profiler/rpc/client/save_profile.cc:137] Creating directory: ./logs/train/plugins/profile/2021_08_04_01_25_15
2021-08-04 01:25:15.155169: I tensorflow/core/profiler/rpc/client/save_profile.cc:143] Dumped gzipped tool data for memory_profile.json.gz to ./logs/train/plugins/profile/2021_08_04_01_25_15/kokoro-gcp-ubuntu-prod-1251741625.memory_profile.json.gz
2021-08-04 01:25:15.155597: I tensorflow/core/profiler/rpc/client/capture_profile.cc:251] Creating directory: ./logs/train/plugins/profile/2021_08_04_01_25_15Dumped tool data for xplane.pb to ./logs/train/plugins/profile/2021_08_04_01_25_15/kokoro-gcp-ubuntu-prod-1251741625.xplane.pb
Dumped tool data for overview_page.pb to ./logs/train/plugins/profile/2021_08_04_01_25_15/kokoro-gcp-ubuntu-prod-1251741625.overview_page.pb
Dumped tool data for input_pipeline.pb to ./logs/train/plugins/profile/2021_08_04_01_25_15/kokoro-gcp-ubuntu-prod-1251741625.input_pipeline.pb
Dumped tool data for tensorflow_stats.pb to ./logs/train/plugins/profile/2021_08_04_01_25_15/kokoro-gcp-ubuntu-prod-1251741625.tensorflow_stats.pb
Dumped tool data for kernel_stats.pb to ./logs/train/plugins/profile/2021_08_04_01_25_15/kokoro-gcp-ubuntu-prod-1251741625.kernel_stats.pb

WARNING:tensorflow:Callback method `on_train_batch_begin` is slow compared to the batch time (batch time: 0.0045s vs `on_train_batch_begin` time: 0.0762s). Check your callbacks.
WARNING:tensorflow:Callback method `on_train_batch_end` is slow compared to the batch time (batch time: 0.0045s vs `on_train_batch_end` time: 0.0155s). Check your callbacks.
WARNING:tensorflow:Callback method `on_train_batch_end` is slow compared to the batch time (batch time: 0.0045s vs `on_train_batch_end` time: 0.0155s). Check your callbacks.
938/938 [==============================] - 16s 4ms/step - loss: 0.1997 - accuracy: 0.9421

Learning rate for epoch 1 is 0.0010000000474974513
Epoch 2/12
938/938 [==============================] - 3s 3ms/step - loss: 0.0656 - accuracy: 0.9805

Learning rate for epoch 2 is 0.0010000000474974513
Epoch 3/12
938/938 [==============================] - 3s 3ms/step - loss: 0.0461 - accuracy: 0.9857

Learning rate for epoch 3 is 0.0010000000474974513
Epoch 4/12
938/938 [==============================] - 3s 3ms/step - loss: 0.0244 - accuracy: 0.9935

Learning rate for epoch 4 is 9.999999747378752e-05
Epoch 5/12
938/938 [==============================] - 3s 3ms/step - loss: 0.0217 - accuracy: 0.9943

Learning rate for epoch 5 is 9.999999747378752e-05
Epoch 6/12
938/938 [==============================] - 3s 3ms/step - loss: 0.0199 - accuracy: 0.9948

Learning rate for epoch 6 is 9.999999747378752e-05
Epoch 7/12
938/938 [==============================] - 3s 3ms/step - loss: 0.0182 - accuracy: 0.9955

Learning rate for epoch 7 is 9.999999747378752e-05
Epoch 8/12
938/938 [==============================] - 3s 3ms/step - loss: 0.0156 - accuracy: 0.9963

Learning rate for epoch 8 is 9.999999747378752e-06
Epoch 9/12
938/938 [==============================] - 3s 3ms/step - loss: 0.0154 - accuracy: 0.9964

Learning rate for epoch 9 is 9.999999747378752e-06
Epoch 10/12
938/938 [==============================] - 3s 3ms/step - loss: 0.0152 - accuracy: 0.9965

Learning rate for epoch 10 is 9.999999747378752e-06
Epoch 11/12
938/938 [==============================] - 3s 3ms/step - loss: 0.0150 - accuracy: 0.9966

Learning rate for epoch 11 is 9.999999747378752e-06
Epoch 12/12
938/938 [==============================] - 3s 3ms/step - loss: 0.0149 - accuracy: 0.9967

Learning rate for epoch 12 is 9.999999747378752e-06
<tensorflow.python.keras.callbacks.History at 0x7f4e5c176dd0>

Compruebe los puntos de control guardados:

# Check the checkpoint directory.
ls {checkpoint_dir}
checkpoint           ckpt_4.data-00000-of-00001
ckpt_1.data-00000-of-00001   ckpt_4.index
ckpt_1.index             ckpt_5.data-00000-of-00001
ckpt_10.data-00000-of-00001  ckpt_5.index
ckpt_10.index            ckpt_6.data-00000-of-00001
ckpt_11.data-00000-of-00001  ckpt_6.index
ckpt_11.index            ckpt_7.data-00000-of-00001
ckpt_12.data-00000-of-00001  ckpt_7.index
ckpt_12.index            ckpt_8.data-00000-of-00001
ckpt_2.data-00000-of-00001   ckpt_8.index
ckpt_2.index             ckpt_9.data-00000-of-00001
ckpt_3.data-00000-of-00001   ckpt_9.index
ckpt_3.index

Para comprobar si funciona bien el modelo, cargar el último punto de control y llamar Model.evaluate en los datos de prueba:

model.load_weights(tf.train.latest_checkpoint(checkpoint_dir))

eval_loss, eval_acc = model.evaluate(eval_dataset)

print('Eval loss: {}, Eval accuracy: {}'.format(eval_loss, eval_acc))
2021-08-04 01:25:49.277864: W tensorflow/core/grappler/optimizers/data/auto_shard.cc:461] The `assert_cardinality` transformation is currently not handled by the auto-shard rewrite and will be removed.
157/157 [==============================] - 2s 4ms/step - loss: 0.0371 - accuracy: 0.9875
Eval loss: 0.03712465986609459, Eval accuracy: 0.987500011920929

Para visualizar el resultado, inicie TensorBoard y vea los registros:

%tensorboard --logdir=logs

ls -sh ./logs
total 4.0K
4.0K train

Exportar a modelo guardado

Exportar el gráfico y las variables en el formato SavedModel independiente de la plataforma que utilizan Model.save . Después de que su modelo se guarda, puede cargarla con o sin la Strategy.scope .

path = 'saved_model/'
model.save(path, save_format='tf')
2021-08-04 01:25:51.983973: W tensorflow/python/util/util.cc:348] Sets are not currently considered sequences, but this may change in the future, so consider avoiding using them.
INFO:tensorflow:Assets written to: saved_model/assets
INFO:tensorflow:Assets written to: saved_model/assets

Ahora, cargue el modelo sin Strategy.scope :

unreplicated_model = tf.keras.models.load_model(path)

unreplicated_model.compile(
    loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),
    optimizer=tf.keras.optimizers.Adam(),
    metrics=['accuracy'])

eval_loss, eval_acc = unreplicated_model.evaluate(eval_dataset)

print('Eval loss: {}, Eval Accuracy: {}'.format(eval_loss, eval_acc))
157/157 [==============================] - 0s 2ms/step - loss: 0.0371 - accuracy: 0.9875
Eval loss: 0.03712465986609459, Eval Accuracy: 0.987500011920929

Cargar el modelo con Strategy.scope :

with strategy.scope():
  replicated_model = tf.keras.models.load_model(path)
  replicated_model.compile(loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),
                           optimizer=tf.keras.optimizers.Adam(),
                           metrics=['accuracy'])

  eval_loss, eval_acc = replicated_model.evaluate(eval_dataset)
  print ('Eval loss: {}, Eval Accuracy: {}'.format(eval_loss, eval_acc))
2021-08-04 01:25:53.544239: W tensorflow/core/grappler/optimizers/data/auto_shard.cc:461] The `assert_cardinality` transformation is currently not handled by the auto-shard rewrite and will be removed.
157/157 [==============================] - 2s 2ms/step - loss: 0.0371 - accuracy: 0.9875
Eval loss: 0.03712465986609459, Eval Accuracy: 0.987500011920929

Recursos adicionales

Más ejemplos que utilizan diferentes estrategias de distribución con la Keras Model.fit API:

  1. Las tareas COLA Resolver utilizando BERT TPU guía de aprendizaje utiliza tf.distribute.MirroredStrategy para la formación en las GPU y tf.distribute.TPUStrategy -on TPU.
  2. La guardar y cargar un modelo utilizando una estrategia de distribución demonstates tutorial de cómo utilizar las API SavedModel con tf.distribute.Strategy .
  3. Los modelos oficiales TensorFlow se pueden configurar para ejecutar varias estrategias de distribución.

Para obtener más información sobre las estrategias de distribución de TensorFlow:

  1. La formación de encargo con tf.distribute.Strategy espectáculos tutorial de cómo utilizar el tf.distribute.MirroredStrategy para la formación de un solo trabajador con un bucle de entrenamiento personalizado.
  2. El entrenamiento del Multi-trabajador con Keras espectáculos tutorial de cómo utilizar el MultiWorkerMirroredStrategy con Model.fit .
  3. El circuito de entrenamiento personalizado con Keras y MultiWorkerMirroredStrategy espectáculos tutorial de cómo utilizar el MultiWorkerMirroredStrategy con Keras y un circuito de entrenamiento personalizado.
  4. La formación distribuida en TensorFlow guía proporciona una visión general de las estrategias de distribución disponibles.
  5. El rendimiento mejor con tf.function guía proporciona información sobre otras estrategias y herramientas, tales como la TensorFlow Profiler se puede utilizar para optimizar el rendimiento de sus modelos TensorFlow.