O Dia da Comunidade de ML é dia 9 de novembro! Junte-nos para atualização de TensorFlow, JAX, e mais Saiba mais

Treinamento distribuído com Keras

Ver no TensorFlow.org Executar no Google Colab Ver fonte no GitHub Baixar caderno

Visão geral

O tf.distribute.Strategy API fornece uma abstração para a distribuição de seu treinamento em várias unidades de processamento. Ele permite que você realize treinamento distribuído usando modelos existentes e código de treinamento com mudanças mínimas.

Este tutorial demonstra como usar o tf.distribute.MirroredStrategy para executar a replicação in-gráfico com formação síncrona em muitas GPUs em uma máquina. A estratégia copia essencialmente todas as variáveis ​​do modelo para cada processador. Em seguida, utiliza todo-reduzir a combinar os gradientes a partir de todos os processadores, e aplica-se o valor combinado para todas as cópias do modelo.

Você usará os tf.keras APIs para construir o modelo e Model.fit para treiná-lo. (Para saber mais sobre o treinamento distribuído com um ciclo de treinamento personalizado eo MirroredStrategy , confira este tutorial .)

MirroredStrategy treina o seu modelo em várias GPUs em uma única máquina. Para a formação síncrona em muitas GPUs em vários trabalhadores, utilize o tf.distribute.MultiWorkerMirroredStrategy com o Keras Model.fit ou um ciclo de treinamento personalizado . Para outras opções, consulte o guia de treinamento distribuído .

Para saber mais sobre várias outras estratégias, há a formação distribuída com TensorFlow guia.

Configurar

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

Baixe o conjunto de dados

Carregar o conjunto de dados MNIST de TensorFlow conjuntos de dados . Isso retorna um conjunto de dados no tf.data formato.

Definir o with_info argumento para True inclui os metadados para todo o conjunto de dados, que está sendo salvo aqui para info . Entre outras coisas, esse objeto de metadados inclui o número de exemplos de treinamento e teste.

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)

Defina a estratégia de distribuição

Criar um MirroredStrategy objeto. Esta cuidará da distribuição e fornecer um gerente de contexto ( MirroredStrategy.scope ) para construir seu modelo dentro.

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

Configure o pipeline de entrada

Ao treinar um modelo com várias GPUs, você pode usar o poder de computação extra de forma eficaz, aumentando o tamanho do lote. Em geral, use o maior tamanho de lote que se ajuste à memória da GPU e ajuste a taxa de aprendizado de acordo.

# 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 uma funo que normaliza os valores de pixel da imagem a partir do [0, 255] intervalo para o [0, 1] intervalo ( escamação característica ):

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

  return image, label

Aplicar esta scale função para os dados de treinamento e teste, e depois usar os tf.data.Dataset APIs para embaralhar os dados de treinamento ( Dataset.shuffle ) e lote-lo ( Dataset.batch ). Observe que você também estão mantendo um cache em memória dos dados de treinamento para melhorar o desempenho ( Dataset.cache ).

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

Crie o modelo

Criar e compilar o modelo Keras no 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',).

Defina os retornos de chamada

Defina as seguintes tf.keras.callbacks :

Para fins ilustrativos, adicionar um callback personalizado chamado PrintLR para exibir a taxa de aprendizagem no notebook.

# 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

Treine e avalie

Agora, treinar o modelo da forma habitual chamando Model.fit sobre o modelo e passando no conjunto de dados criado no início do tutorial. Esta etapa é a mesma, quer você esteja distribuindo o treinamento ou não.

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>

Verifique se há pontos de verificação salvos:

# 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 verificar o quão bem os executa modelo, carregue o último checkpoint e chamar Model.evaluate nos dados de teste:

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 a saída, inicie o TensorBoard e veja os registros:

%tensorboard --logdir=logs

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

Exportar para SavedModel

Exportar o gráfico e as variáveis para o formato SavedModel independente de plataforma usando Model.save . Após o seu modelo é salvo, você pode carregá-lo com ou sem a 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

Agora, carregue o modelo sem 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

Coloque o modelo com 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 adicionais

Mais exemplos que usam diferentes estratégias de distribuição com a Keras Model.fit API:

  1. As tarefas GLUE resolver usando BERT em TPU tutorial usa tf.distribute.MirroredStrategy para treinamento em GPUs e tf.distribute.TPUStrategy -on TPUs.
  2. O Salvar e carregar um modelo usando uma estratégia de distribuição demonstates tutorial como usar as APIs SavedModel com tf.distribute.Strategy .
  3. Os modelos oficiais TensorFlow pode ser configurado para executar múltiplas estratégias de distribuição.

Para saber mais sobre as estratégias de distribuição do TensorFlow:

  1. O treinamento personalizado com tf.distribute.Strategy mostra tutorial como usar o tf.distribute.MirroredStrategy para o treinamento do operador único com um loop de treinamento personalizado.
  2. A formação Multi-trabalhador com Keras mostra tutorial como usar o MultiWorkerMirroredStrategy com Model.fit .
  3. O ciclo de treinamento personalizado com Keras e MultiWorkerMirroredStrategy mostra tutorial como usar o MultiWorkerMirroredStrategy com Keras e um loop de treinamento personalizado.
  4. A formação Distribuído em TensorFlow guia fornece uma visão geral das estratégias de distribuição disponíveis.
  5. O desempenho melhor com tf.function guia fornece informações sobre outras estratégias e ferramentas, como o TensorFlow Profiler você pode usar para otimizar o desempenho de seus modelos TensorFlow.