无论如何要在 Mac 中与 Apple Silicon（M1、M1 Pro、M1 Max）GPU 一起使用 Tensorflow？

Question

我有一台配备 M1 Max 处理器的 MacBook Pro，我想运行此 GPU 上的 Tensorflow。我已按照 google 官方页面中 https://developer.apple.com/metal/tensorflow-plugin but I don't know why it runs slower on my GPU. I tested with MNIST tutorial 的步骤进行操作。

我试过的代码

import tensorflow as tf
import tensorflow_datasets as tfds

DISABLE_GPU = False

if DISABLE_GPU:
    try:
        # Disable all GPUS
        tf.config.set_visible_devices([], 'GPU')
        visible_devices = tf.config.get_visible_devices()
        for device in visible_devices:
            assert device.device_type != 'GPU'
    except:
        # Invalid device or cannot modify virtual devices once initialized.
        pass

print(tf.__version__)

(ds_train, ds_test), ds_info = tfds.load('mnist', split=['train', 'test'], shuffle_files=True, as_supervised=True,
                                         with_info=True)


def normalize_img(image, label):
    return tf.cast(image, tf.float32) / 255., label


ds_train = ds_train.map(normalize_img, num_parallel_calls=tf.data.AUTOTUNE)
ds_train = ds_train.cache()
ds_train = ds_train.shuffle(ds_info.splits['train'].num_examples)
ds_train = ds_train.batch(128)
ds_train = ds_train.prefetch(tf.data.AUTOTUNE)
ds_test = ds_test.map(
    normalize_img, num_parallel_calls=tf.data.AUTOTUNE)
ds_test = ds_test.batch(128)
ds_test = ds_test.cache()
ds_test = ds_test.prefetch(tf.data.AUTOTUNE)

model = tf.keras.models.Sequential([
    tf.keras.layers.Flatten(input_shape=(28, 28)),
    tf.keras.layers.Dense(128, activation='relu'),
    tf.keras.layers.Dense(10)
])
model.compile(
    optimizer=tf.keras.optimizers.Adam(0.001),
    loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),
    metrics=[tf.keras.metrics.SparseCategoricalAccuracy()],
)

model.fit(ds_train, epochs=6, validation_data=ds_test, )

输出（GPU）：

462/469 [============================>.] - ETA: 0s - loss: 0.3619 - sparse_categorical_accuracy: 0.9003
469/469 [==============================] - 4s 5ms/step - loss: 0.3595 - sparse_categorical_accuracy: 0.9008 - val_loss: 0.1963 - val_sparse_categorical_accuracy: 0.9432
Epoch 2/6
469/469 [==============================] - 2s 5ms/step - loss: 0.1708 - sparse_categorical_accuracy: 0.9514 - val_loss: 0.1392 - val_sparse_categorical_accuracy: 0.9606
Epoch 3/6
469/469 [==============================] - 2s 5ms/step - loss: 0.1224 - sparse_categorical_accuracy: 0.9651 - val_loss: 0.1233 - val_sparse_categorical_accuracy: 0.9650
Epoch 4/6
469/469 [==============================] - 2s 5ms/step - loss: 0.0956 - sparse_categorical_accuracy: 0.9725 - val_loss: 0.0988 - val_sparse_categorical_accuracy: 0.9696
Epoch 5/6
469/469 [==============================] - 2s 5ms/step - loss: 0.0766 - sparse_categorical_accuracy: 0.9780 - val_loss: 0.0875 - val_sparse_categorical_accuracy: 0.9727
Epoch 6/6
469/469 [==============================] - 2s 5ms/step - loss: 0.0633 - sparse_categorical_accuracy: 0.9813 - val_loss: 0.0842 - val_sparse_categorical_accuracy: 0.9745

输出（无 GPU）

469/469 [==============================] - 2s 1ms/step - loss: 0.3598 - sparse_categorical_accuracy: 0.9013 - val_loss: 0.1970 - val_sparse_categorical_accuracy: 0.9427
Epoch 2/6
469/469 [==============================] - 0s 933us/step - loss: 0.1705 - sparse_categorical_accuracy: 0.9511 - val_loss: 0.1449 - val_sparse_categorical_accuracy: 0.9589
Epoch 3/6
469/469 [==============================] - 0s 936us/step - loss: 0.1232 - sparse_categorical_accuracy: 0.9642 - val_loss: 0.1146 - val_sparse_categorical_accuracy: 0.9655
Epoch 4/6
469/469 [==============================] - 0s 925us/step - loss: 0.0955 - sparse_categorical_accuracy: 0.9725 - val_loss: 0.1007 - val_sparse_categorical_accuracy: 0.9690
Epoch 5/6
469/469 [==============================] - 0s 946us/step - loss: 0.0774 - sparse_categorical_accuracy: 0.9781 - val_loss: 0.0890 - val_sparse_categorical_accuracy: 0.9732
Epoch 6/6
469/469 [==============================] - 0s 971us/step - loss: 0.0647 - sparse_categorical_accuracy: 0.9811 - val_loss: 0.0844 - val_sparse_categorical_accuracy: 0.9752

Answer 1

它运行变慢的原因可能是因为本教程中使用的批量大小较小。但是，请确保您已正确设置所有内容，如下所示。我们将使用 miniforge 而不是 anaconda，因为它不支持 GPU。

为 TensorFlow 支持设置 miniforge

下载Miniforge3-MacOSX-arm64.sh
运行使用以下命令的文件：-
- ./Miniforge3-MacOSX-arm64.sh
- (不要把上面的运行当成sudo。如果出现权限错误，先运行 chmod +x ./Miniforge3-MacOSX-arm64.sh)
它将在当前目录下下载miniforge。现在你必须激活它。使用以下命令来执行此操作。
- source miniforge3/bin/activate
您应该会在命令行中看到 (conda)。确保它在终端启动期间被激活。使用以下命令。
- conda init
- 或者如果您使用的是 zsh、conda init zsh
确保它已正确激活。要检查它，请使用 which python。它应该显示 .../miniforge3/bin/python。如果没有显示，请先删除 miniforge3 目录，然后尝试从第 2 步开始重新安装。另外，请确保您的 anaconda 环境已禁用。

现在我们将安装 TensorFlow 及其依赖项。

使用以下命令在 conda 环境之上创建一个新环境并激活它。
- conda create -n tensorflow python=<your-python-version
- （用python --version找）
- conda activate tensorflow
现在使用以下命令安装 TensorFlow 依赖项。
- conda install -c apple tensorflow-deps.
使用以下命令为 mac 安装 Tensorflow 和 Tensorflow metal。
- pip install tensorflow-macos
- pip install tensorflow-metal

其他包

使用以下命令安装 jupyter。
- conda install -c conda-forge jupyterlab

疑难解答

'miniforge3/envs/tensorflow/lib/libcblas.3.dylib'（没有这样的文件）或类似的 libcblas 错误。
解决方案：conda install -c conda-forge openblas
/tensorflow/core/framework/tensor.h:880] 检查失败：IsAligned() ptr = 0x101511d60
解决方法: 我在某些程序使用 Tensorflow >2.5.0 时发现这个错误。使用 TensorFlow 版本 2.5.0。要重新安装它，请执行以下操作。
- pip uninstall tensorflow-macos
- pip uninstall tensorflow-metal
- conda install -c apple tensorflow-deps==2.5.0 --force-reinstall（可选，仅在出现错误时尝试）
- pip install tensorflow-mac==2.5.0
- pip install tensorflow-metal
jupyter导入tensorflow时可能会遇到导入错误。这应该通过安装新内核来解决。
- python -m ipykernel install --user --name tensorflow --display-name "Python <your-python-version> (tensorflow)"
- 重要提示： 当你启动 jupyter 时，确保 select 这个内核。此外，tensorflow 环境之外的 jupyter 可以使用此内核导入 tensorflow（即，您不必每次在 jupyter 中使用它时都激活 tensorflow 环境）。

测试中（M1 Max，10 核 CPU，24 核 GPU 版本）

代码：

import tensorflow as tf
import tensorflow_datasets as tfds

DISABLE_GPU = False

if DISABLE_GPU:
    try:
        # Disable all GPUS
        tf.config.set_visible_devices([], 'GPU')
        visible_devices = tf.config.get_visible_devices()
        for device in visible_devices:
            assert device.device_type != 'GPU'
    except:
        # Invalid device or cannot modify virtual devices once initialized.
        pass

print(tf.__version__)

(ds_train, ds_test), ds_info = tfds.load('mnist', split=['train', 'test'], shuffle_files=True, as_supervised=True,
                                         with_info=True)


def normalize_img(image, label):
    return tf.cast(image, tf.float32) / 255., label


ds_train = ds_train.map(normalize_img, num_parallel_calls=tf.data.AUTOTUNE)
ds_train = ds_train.cache()
ds_train = ds_train.shuffle(ds_info.splits['train'].num_examples)
ds_train = ds_train.batch(128)
ds_train = ds_train.prefetch(tf.data.AUTOTUNE)
ds_test = ds_test.map(
    normalize_img, num_parallel_calls=tf.data.AUTOTUNE)
ds_test = ds_test.batch(128)
ds_test = ds_test.cache()
ds_test = ds_test.prefetch(tf.data.AUTOTUNE)

model = tf.keras.models.Sequential([
    tf.keras.layers.Flatten(input_shape=(28, 28)),
    tf.keras.layers.Dense(128, activation='relu'),
    tf.keras.layers.Dense(10)
])
model.compile(
    optimizer=tf.keras.optimizers.Adam(0.001),
    loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),
    metrics=[tf.keras.metrics.SparseCategoricalAccuracy()],
)

model.fit(ds_train, epochs=6, validation_data=ds_test, )

批量大小 = 128

输出（GPU）

462/469 [============================>.] - ETA: 0s - loss: 0.3619 - sparse_categorical_accuracy: 0.9003
469/469 [==============================] - 4s 5ms/step - loss: 0.3595 - sparse_categorical_accuracy: 0.9008 - val_loss: 0.1963 - val_sparse_categorical_accuracy: 0.9432
Epoch 2/6
469/469 [==============================] - 2s 5ms/step - loss: 0.1708 - sparse_categorical_accuracy: 0.9514 - val_loss: 0.1392 - val_sparse_categorical_accuracy: 0.9606
Epoch 3/6
469/469 [==============================] - 2s 5ms/step - loss: 0.1224 - sparse_categorical_accuracy: 0.9651 - val_loss: 0.1233 - val_sparse_categorical_accuracy: 0.9650
Epoch 4/6
469/469 [==============================] - 2s 5ms/step - loss: 0.0956 - sparse_categorical_accuracy: 0.9725 - val_loss: 0.0988 - val_sparse_categorical_accuracy: 0.9696
Epoch 5/6
469/469 [==============================] - 2s 5ms/step - loss: 0.0766 - sparse_categorical_accuracy: 0.9780 - val_loss: 0.0875 - val_sparse_categorical_accuracy: 0.9727
Epoch 6/6
469/469 [==============================] - 2s 5ms/step - loss: 0.0633 - sparse_categorical_accuracy: 0.9813 - val_loss: 0.0842 - val_sparse_categorical_accuracy: 0.9745

输出（无 GPU）

469/469 [==============================] - 2s 1ms/step - loss: 0.3598 - sparse_categorical_accuracy: 0.9013 - val_loss: 0.1970 - val_sparse_categorical_accuracy: 0.9427
Epoch 2/6
469/469 [==============================] - 0s 933us/step - loss: 0.1705 - sparse_categorical_accuracy: 0.9511 - val_loss: 0.1449 - val_sparse_categorical_accuracy: 0.9589
Epoch 3/6
469/469 [==============================] - 0s 936us/step - loss: 0.1232 - sparse_categorical_accuracy: 0.9642 - val_loss: 0.1146 - val_sparse_categorical_accuracy: 0.9655
Epoch 4/6
469/469 [==============================] - 0s 925us/step - loss: 0.0955 - sparse_categorical_accuracy: 0.9725 - val_loss: 0.1007 - val_sparse_categorical_accuracy: 0.9690
Epoch 5/6
469/469 [==============================] - 0s 946us/step - loss: 0.0774 - sparse_categorical_accuracy: 0.9781 - val_loss: 0.0890 - val_sparse_categorical_accuracy: 0.9732
Epoch 6/6
469/469 [==============================] - 0s 971us/step - loss: 0.0647 - sparse_categorical_accuracy: 0.9811 - val_loss: 0.0844 - val_sparse_categorical_accuracy: 0.9752

此时我们可以看到 CPU 中的运行ning 比 GPU 上的运行ning 快 (x5)，但不要因此而失望。我们将运行用于大批量。

批量大小 = 1024

输出（GPU）

58/59 [============================>.] - ETA: 0s - loss: 0.4862 - sparse_categorical_accuracy: 0.8680
59/59 [==============================] - 2s 11ms/step - loss: 0.4839 - sparse_categorical_accuracy: 0.8686 - val_loss: 0.2269 - val_sparse_categorical_accuracy: 0.9362
Epoch 2/6
59/59 [==============================] - 0s 8ms/step - loss: 0.1964 - sparse_categorical_accuracy: 0.9442 - val_loss: 0.1610 - val_sparse_categorical_accuracy: 0.9543
Epoch 3/6
59/59 [==============================] - 1s 9ms/step - loss: 0.1408 - sparse_categorical_accuracy: 0.9605 - val_loss: 0.1292 - val_sparse_categorical_accuracy: 0.9624
Epoch 4/6
59/59 [==============================] - 1s 9ms/step - loss: 0.1067 - sparse_categorical_accuracy: 0.9707 - val_loss: 0.1055 - val_sparse_categorical_accuracy: 0.9687
Epoch 5/6
59/59 [==============================] - 1s 9ms/step - loss: 0.0845 - sparse_categorical_accuracy: 0.9767 - val_loss: 0.0912 - val_sparse_categorical_accuracy: 0.9723
Epoch 6/6
59/59 [==============================] - 1s 9ms/step - loss: 0.0683 - sparse_categorical_accuracy: 0.9814 - val_loss: 0.0827 - val_sparse_categorical_accuracy: 0.9747

输出（无 GPU）

59/59 [==============================] - 2s 15ms/step - loss: 0.4640 - sparse_categorical_accuracy: 0.8739 - val_loss: 0.2280 - val_sparse_categorical_accuracy: 0.9338
Epoch 2/6
59/59 [==============================] - 1s 12ms/step - loss: 0.1962 - sparse_categorical_accuracy: 0.9450 - val_loss: 0.1626 - val_sparse_categorical_accuracy: 0.9537
Epoch 3/6
59/59 [==============================] - 1s 12ms/step - loss: 0.1411 - sparse_categorical_accuracy: 0.9602 - val_loss: 0.1304 - val_sparse_categorical_accuracy: 0.9613
Epoch 4/6
59/59 [==============================] - 1s 12ms/step - loss: 0.1091 - sparse_categorical_accuracy: 0.9700 - val_loss: 0.1020 - val_sparse_categorical_accuracy: 0.9698
Epoch 5/6
59/59 [==============================] - 1s 12ms/step - loss: 0.0864 - sparse_categorical_accuracy: 0.9764 - val_loss: 0.0912 - val_sparse_categorical_accuracy: 0.9716
Epoch 6/6
59/59 [==============================] - 1s 12ms/step - loss: 0.0697 - sparse_categorical_accuracy: 0.9812 - val_loss: 0.0834 - val_sparse_categorical_accuracy: 0.9749

如您现在所见，GPU 上的运行ning 比 CPU 上的运行ning 更快 (x1.3)。增加batch size可以显着提高GPU的性能。

无论如何要在 Mac 中与 Apple Silicon（M1、M1 Pro、M1 Max）GPU 一起使用 Tensorflow？

Anyway to work with Tensorflow in Mac with Apple Silicon (M1, M1 Pro, M1 Max) GPU?

tensorflow

apple-silicon

我试过的代码

输出（GPU）：

输出（无 GPU）

为 TensorFlow 支持设置 miniforge

现在我们将安装 TensorFlow 及其依赖项。

其他包

疑难解答

测试中（M1 Max，10 核 CPU，24 核 GPU 版本）

代码：

批量大小 = 128

输出（GPU）

输出（无 GPU）

批量大小 = 1024

输出（GPU）

输出（无 GPU）

图：运行各种批量大小的时间图。

来源：