ML Community Day is November 9! Join us for updates from TensorFlow, JAX, and more Learn more

TensorFlow Addons 回调:TimeStopping

在 TensorFlow.org 上查看 在 Google Colab 中运行 在 GitHub 上查看源代码 下载笔记本

概述

此笔记本将演示如何使用 TensorFlow Addons 中的 TimeStopping 回调。

设置

pip install -U tensorflow-addons
import tensorflow_addons as tfa

from tensorflow.keras.datasets import mnist
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense, Dropout, Flatten
/tmpfs/src/tf_docs_env/lib/python3.7/site-packages/tensorflow_addons/utils/ensure_tf_install.py:67: UserWarning: Tensorflow Addons supports using Python ops for all Tensorflow versions above or equal to 2.3.0 and strictly below 2.6.0 (nightly versions are not supported). 
 The versions of TensorFlow you are currently using is 2.6.0 and is not supported. 
Some things might work, some things might not.
If you were to encounter a bug, do not file an issue.
If you want to make sure you're using a tested and supported configuration, either change the TensorFlow version or the TensorFlow Addons's version. 
You can find the compatibility matrix in TensorFlow Addon's readme:
https://github.com/tensorflow/addons
  UserWarning,

导入并归一化数据

# the data, split between train and test sets
(x_train, y_train), (x_test, y_test) = mnist.load_data()
# normalize data
x_train, x_test = x_train / 255.0, x_test / 255.0
Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/mnist.npz
11493376/11490434 [==============================] - 0s 0us/step
11501568/11490434 [==============================] - 0s 0us/step

构建简单的 MNIST CNN 模型

# build the model using the Sequential API
model = Sequential()
model.add(Flatten(input_shape=(28, 28)))
model.add(Dense(128, activation='relu'))
model.add(Dropout(0.2))
model.add(Dense(10, activation='softmax'))

model.compile(optimizer='adam',
              loss = 'sparse_categorical_crossentropy',
              metrics=['accuracy'])
2021-08-13 19:51:38.498285: 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
2021-08-13 19:51:38.506150: 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
2021-08-13 19:51:38.507036: 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
2021-08-13 19:51:38.508667: 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.
2021-08-13 19:51:38.509232: 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
2021-08-13 19:51:38.510194: 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
2021-08-13 19:51:38.511098: 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
2021-08-13 19:51:39.093394: 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
2021-08-13 19:51:39.094491: 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
2021-08-13 19:51:39.095386: 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
2021-08-13 19:51:39.096269: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1510] Created device /job:localhost/replica:0/task:0/device:GPU:0 with 14648 MB memory:  -> device: 0, name: Tesla V100-SXM2-16GB, pci bus id: 0000:00:05.0, compute capability: 7.0

简单的 TimeStopping 用法

# initialize TimeStopping callback 
time_stopping_callback = tfa.callbacks.TimeStopping(seconds=5, verbose=1)

# train the model with tqdm_callback
# make sure to set verbose = 0 to disable
# the default progress bar.
model.fit(x_train, y_train,
          batch_size=64,
          epochs=100,
          callbacks=[time_stopping_callback],
          validation_data=(x_test, y_test))
2021-08-13 19:51:40.120947: I tensorflow/compiler/mlir/mlir_graph_optimization_pass.cc:185] None of the MLIR Optimization Passes are enabled (registered 2)
Epoch 1/100
938/938 [==============================] - 3s 2ms/step - loss: 0.3375 - accuracy: 0.9032 - val_loss: 0.1676 - val_accuracy: 0.9500
Epoch 2/100
938/938 [==============================] - 2s 2ms/step - loss: 0.1608 - accuracy: 0.9532 - val_loss: 0.1146 - val_accuracy: 0.9639
Epoch 3/100
938/938 [==============================] - 2s 2ms/step - loss: 0.1212 - accuracy: 0.9638 - val_loss: 0.0947 - val_accuracy: 0.9726
Timed stopping at epoch 3 after training for 0:00:05
<keras.callbacks.History at 0x7f7a91776b90>