{ "cells": [ { "cell_type": "code", "execution_count": null, "metadata": { "id": "b518b04cbfe0" }, "outputs": [], "source": [ "##### Copyright 2020 The TensorFlow Authors." ] }, { "cell_type": "code", "execution_count": null, "metadata": { "cellView": "form", "id": "906e07f6e562", "vscode": { "languageId": "python" } }, "outputs": [], "source": [ "#@title Licensed under the Apache License, Version 2.0 (the \"License\");\n", "# you may not use this file except in compliance with the License.\n", "# You may obtain a copy of the License at\n", "#\n", "# https://www.apache.org/licenses/LICENSE-2.0\n", "#\n", "# Unless required by applicable law or agreed to in writing, software\n", "# distributed under the License is distributed on an \"AS IS\" BASIS,\n", "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n", "# See the License for the specific language governing permissions and\n", "# limitations under the License." ] }, { "cell_type": "markdown", "metadata": { "id": "a5620ee4049e" }, "source": [ "# 自定义 Model.fit 的内容" ] }, { "cell_type": "markdown", "metadata": { "id": "0a56ffedf331" }, "source": [ "\n", " \n", " \n", " \n", " \n", "
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
" ] }, { "cell_type": "markdown", "metadata": { "id": "7ebb4e65ef9b" }, "source": [ "## 简介\n", "\n", "您在进行监督学习时可以使用 `fit()`，一切都可以顺利完成。\n", "\n", "需要从头开始编写自己的训练循环时，您可以使用 `GradientTape` 并控制每个微小的细节。\n", "\n", "但如果您需要自定义训练算法，又想从 `fit()` 的便捷功能（例如回调、内置分布支持或步骤融合）中受益，那么该怎么做？\n", "\n", "Keras 的核心原则是**渐进式呈现复杂性**。您应当始终能够以渐进的方式习惯较低级别的工作流。如果高级功能并不完全符合您的用例，那么您就不应深陷其中。您应当能够从容地控制微小的细节，同时保留与之相称的高级便利性。\n", "\n", "需要自定义 `fit()` 的功能时，您应**重写 `Model` 类的训练步骤函数**。此函数是 `fit()` 会针对每批次数据调用的函数。然后，您将能够像往常一样调用 `fit()`，它将运行您自己的学习算法。\n", "\n", "请注意，此模式不会妨碍您使用函数式 API 构建模型。无论是构建 `Sequential` 模型、函数式 API 模型还是子类模型，均可采用这种模式。\n", "\n", "让我们了解一下它的工作方式。" ] }, { "cell_type": "markdown", "metadata": { "id": "2849e371b9b6" }, "source": [ "## 设置\n", "\n", "需要 TensorFlow 2.2 或更高版本。" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "4dadb6688663", "vscode": { "languageId": "python" } }, "outputs": [], "source": [ "import tensorflow as tf\n", "from tensorflow import keras" ] }, { "cell_type": "markdown", "metadata": { "id": "9022333acaa7" }, "source": [ "## 第一个简单的示例\n", "\n", "让我们从一个简单的示例开始：\n", "\n", "- 创建一个将 `keras.Model` 子类化的新类。\n", "- 仅重写 `train_step(self, data)` 方法。\n", "- 返回一个将指标名称（包括损失）映射到其当前值的字典。\n", "\n", "输入参数 `data` 是传递以拟合训练数据的数据：\n", "\n", "- 如果通过调用 `fit(x, y, ...)` 传递 Numpy 数组，则 `data` 将为元祖 `(x, y)`。\n", "- 如果通过调用 `fit(dataset, ...)` 传递 `tf.data.Dataset`，则 `data` 将为每批次 `dataset` 产生的数据。\n", "\n", "我们在 `train_step` 方法的主体中实现了定期的训练更新，类似于您已经熟悉的内容。重要的是，**我们通过 `self.compiled_loss` 计算损失**，它会封装传递给 `compile()` 的损失函数。\n", "\n", "同样，我们调用 `self.compiled_metrics.update_state(y, y_pred)` 来更新在 `compile()` 中传递的指标的状态，并在最后从 `self.metrics` 中查询结果以检索其当前值。" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "060c8bf4150d", "vscode": { "languageId": "python" } }, "outputs": [], "source": [ "class CustomModel(keras.Model):\n", " def train_step(self, data):\n", " # Unpack the data. Its structure depends on your model and\n", " # on what you pass to `fit()`.\n", " x, y = data\n", "\n", " with tf.GradientTape() as tape:\n", " y_pred = self(x, training=True) # Forward pass\n", " # Compute the loss value\n", " # (the loss function is configured in `compile()`)\n", " loss = self.compiled_loss(y, y_pred, regularization_losses=self.losses)\n", "\n", " # Compute gradients\n", " trainable_vars = self.trainable_variables\n", " gradients = tape.gradient(loss, trainable_vars)\n", " # Update weights\n", " self.optimizer.apply_gradients(zip(gradients, trainable_vars))\n", " # Update metrics (includes the metric that tracks the loss)\n", " self.compiled_metrics.update_state(y, y_pred)\n", " # Return a dict mapping metric names to current value\n", " return {m.name: m.result() for m in self.metrics}\n" ] }, { "cell_type": "markdown", "metadata": { "id": "c9d2cc7a7014" }, "source": [ "我们来试一下：" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "5e6bd7b554f6", "vscode": { "languageId": "python" } }, "outputs": [], "source": [ "import numpy as np\n", "\n", "# Construct and compile an instance of CustomModel\n", "inputs = keras.Input(shape=(32,))\n", "outputs = keras.layers.Dense(1)(inputs)\n", "model = CustomModel(inputs, outputs)\n", "model.compile(optimizer=\"adam\", loss=\"mse\", metrics=[\"mae\"])\n", "\n", "# Just use `fit` as usual\n", "x = np.random.random((1000, 32))\n", "y = np.random.random((1000, 1))\n", "model.fit(x, y, epochs=3)" ] }, { "cell_type": "markdown", "metadata": { "id": "a882cb6467d6" }, "source": [ "## 在更低级别上操作\n", "\n", "当然，您可以直接跳过在 `compile()` 中传递损失函数，而在 `train_step` 中*手动*完成所有内容。指标也是如此。\n", "\n", "以下是一个较低级别的示例，仅使用 `compile()` 配置优化器：\n", "\n", "- 我们从创建 `Metric` 实例以跟踪我们的损失和 MAE 得分开始。\n", "- 我们实现可更新这些指标状态（通过对指标调用 `update_state()`）的自定义 `train_step()` ，然后对其进行查询（通过 `result()`）以返回其当前平均值，由进度条显示并传递给任何回调。\n", "- 请注意，需要在每个周期之间对指标调用 `reset_states()`！否则，调用 `result()` 会返回自训练开始以来的平均值，但我们通常要使用的是每个周期的平均值。幸运的是，该框架可以帮助我们实现：只需在模型的 `metrics` 属性中列出要重置的任何指标。模型将在每个 `fit()` 周期开始时或在开始调用 `evaluate()` 时对其中列出的任何对象调用 `reset_states()`。" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "2308abf5fe7d", "vscode": { "languageId": "python" } }, "outputs": [], "source": [ "loss_tracker = keras.metrics.Mean(name=\"loss\")\n", "mae_metric = keras.metrics.MeanAbsoluteError(name=\"mae\")\n", "\n", "\n", "class CustomModel(keras.Model):\n", " def train_step(self, data):\n", " x, y = data\n", "\n", " with tf.GradientTape() as tape:\n", " y_pred = self(x, training=True) # Forward pass\n", " # Compute our own loss\n", " loss = keras.losses.mean_squared_error(y, y_pred)\n", "\n", " # Compute gradients\n", " trainable_vars = self.trainable_variables\n", " gradients = tape.gradient(loss, trainable_vars)\n", "\n", " # Update weights\n", " self.optimizer.apply_gradients(zip(gradients, trainable_vars))\n", "\n", " # Compute our own metrics\n", " loss_tracker.update_state(loss)\n", " mae_metric.update_state(y, y_pred)\n", " return {\"loss\": loss_tracker.result(), \"mae\": mae_metric.result()}\n", "\n", " @property\n", " def metrics(self):\n", " # We list our `Metric` objects here so that `reset_states()` can be\n", " # called automatically at the start of each epoch\n", " # or at the start of `evaluate()`.\n", " # If you don't implement this property, you have to call\n", " # `reset_states()` yourself at the time of your choosing.\n", " return [loss_tracker, mae_metric]\n", "\n", "\n", "# Construct an instance of CustomModel\n", "inputs = keras.Input(shape=(32,))\n", "outputs = keras.layers.Dense(1)(inputs)\n", "model = CustomModel(inputs, outputs)\n", "\n", "# We don't passs a loss or metrics here.\n", "model.compile(optimizer=\"adam\")\n", "\n", "# Just use `fit` as usual -- you can use callbacks, etc.\n", "x = np.random.random((1000, 32))\n", "y = np.random.random((1000, 1))\n", "model.fit(x, y, epochs=5)\n" ] }, { "cell_type": "markdown", "metadata": { "id": "f451e382c6a8" }, "source": [ "## 支持 `sample_weight` 和 `class_weight`\n", "\n", "您可能已经注意到，我们的第一个基本示例并没有提及样本加权。如果要支持 `fit()` 参数 `sample_weight` 和 `class_weight`，只需执行以下操作：\n", "\n", "- 从 `data` 参数中解包 `sample_weight`\n", "- 将其传递给 `compiled_loss` 和 `compiled_metrics`（当然，如果您不依赖 `compile()` 来获取损失和指标，也可以手动应用）\n", "- 就是这么简单。" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "522d7281f948", "vscode": { "languageId": "python" } }, "outputs": [], "source": [ "class CustomModel(keras.Model):\n", " def train_step(self, data):\n", " # Unpack the data. Its structure depends on your model and\n", " # on what you pass to `fit()`.\n", " if len(data) == 3:\n", " x, y, sample_weight = data\n", " else:\n", " sample_weight = None\n", " x, y = data\n", "\n", " with tf.GradientTape() as tape:\n", " y_pred = self(x, training=True) # Forward pass\n", " # Compute the loss value.\n", " # The loss function is configured in `compile()`.\n", " loss = self.compiled_loss(\n", " y,\n", " y_pred,\n", " sample_weight=sample_weight,\n", " regularization_losses=self.losses,\n", " )\n", "\n", " # Compute gradients\n", " trainable_vars = self.trainable_variables\n", " gradients = tape.gradient(loss, trainable_vars)\n", "\n", " # Update weights\n", " self.optimizer.apply_gradients(zip(gradients, trainable_vars))\n", "\n", " # Update the metrics.\n", " # Metrics are configured in `compile()`.\n", " self.compiled_metrics.update_state(y, y_pred, sample_weight=sample_weight)\n", "\n", " # Return a dict mapping metric names to current value.\n", " # Note that it will include the loss (tracked in self.metrics).\n", " return {m.name: m.result() for m in self.metrics}\n", "\n", "\n", "# Construct and compile an instance of CustomModel\n", "inputs = keras.Input(shape=(32,))\n", "outputs = keras.layers.Dense(1)(inputs)\n", "model = CustomModel(inputs, outputs)\n", "model.compile(optimizer=\"adam\", loss=\"mse\", metrics=[\"mae\"])\n", "\n", "# You can now use sample_weight argument\n", "x = np.random.random((1000, 32))\n", "y = np.random.random((1000, 1))\n", "sw = np.random.random((1000, 1))\n", "model.fit(x, y, sample_weight=sw, epochs=3)" ] }, { "cell_type": "markdown", "metadata": { "id": "03000c5590db" }, "source": [ "## 提供您自己的评估步骤\n", "\n", "如何对调用 `model.evaluate()` 进行相同的处理？您需要以完全相同的方式重写 `test_step`。如下所示：" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "999edb22c50e", "vscode": { "languageId": "python" } }, "outputs": [], "source": [ "class CustomModel(keras.Model):\n", " def test_step(self, data):\n", " # Unpack the data\n", " x, y = data\n", " # Compute predictions\n", " y_pred = self(x, training=False)\n", " # Updates the metrics tracking the loss\n", " self.compiled_loss(y, y_pred, regularization_losses=self.losses)\n", " # Update the metrics.\n", " self.compiled_metrics.update_state(y, y_pred)\n", " # Return a dict mapping metric names to current value.\n", " # Note that it will include the loss (tracked in self.metrics).\n", " return {m.name: m.result() for m in self.metrics}\n", "\n", "\n", "# Construct an instance of CustomModel\n", "inputs = keras.Input(shape=(32,))\n", "outputs = keras.layers.Dense(1)(inputs)\n", "model = CustomModel(inputs, outputs)\n", "model.compile(loss=\"mse\", metrics=[\"mae\"])\n", "\n", "# Evaluate with our custom test_step\n", "x = np.random.random((1000, 32))\n", "y = np.random.random((1000, 1))\n", "model.evaluate(x, y)" ] }, { "cell_type": "markdown", "metadata": { "id": "9e6a662e6588" }, "source": [ "## 总结：端到端 GAN 示例\n", "\n", "让我们看一个利用您刚刚所学全部内容的端到端示例。\n", "\n", "请考虑：\n", "\n", "- 旨在生成 28x28x1 图像的生成器网络。\n", "- 旨在将 28x28x1 图像分为两类（“fake”和“real”）的鉴别器网络。\n", "- 分别用于两个网络的优化器。\n", "- 训练鉴别器的损失函数。\n" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "6748db01dc7c", "vscode": { "languageId": "python" } }, "outputs": [], "source": [ "from tensorflow.keras import layers\n", "\n", "# Create the discriminator\n", "discriminator = keras.Sequential(\n", " [\n", " keras.Input(shape=(28, 28, 1)),\n", " layers.Conv2D(64, (3, 3), strides=(2, 2), padding=\"same\"),\n", " layers.LeakyReLU(alpha=0.2),\n", " layers.Conv2D(128, (3, 3), strides=(2, 2), padding=\"same\"),\n", " layers.LeakyReLU(alpha=0.2),\n", " layers.GlobalMaxPooling2D(),\n", " layers.Dense(1),\n", " ],\n", " name=\"discriminator\",\n", ")\n", "\n", "# Create the generator\n", "latent_dim = 128\n", "generator = keras.Sequential(\n", " [\n", " keras.Input(shape=(latent_dim,)),\n", " # We want to generate 128 coefficients to reshape into a 7x7x128 map\n", " layers.Dense(7 * 7 * 128),\n", " layers.LeakyReLU(alpha=0.2),\n", " layers.Reshape((7, 7, 128)),\n", " layers.Conv2DTranspose(128, (4, 4), strides=(2, 2), padding=\"same\"),\n", " layers.LeakyReLU(alpha=0.2),\n", " layers.Conv2DTranspose(128, (4, 4), strides=(2, 2), padding=\"same\"),\n", " layers.LeakyReLU(alpha=0.2),\n", " layers.Conv2D(1, (7, 7), padding=\"same\", activation=\"sigmoid\"),\n", " ],\n", " name=\"generator\",\n", ")" ] }, { "cell_type": "markdown", "metadata": { "id": "801e8dd0c92a" }, "source": [ "这是特征齐全的 GAN 类，重写了 `compile()` 以使用其自己的签名，并在 `train_step` 的 17 行中实现了整个 GAN 算法：" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "bc3fb4111393", "vscode": { "languageId": "python" } }, "outputs": [], "source": [ "class GAN(keras.Model):\n", " def __init__(self, discriminator, generator, latent_dim):\n", " super(GAN, self).__init__()\n", " self.discriminator = discriminator\n", " self.generator = generator\n", " self.latent_dim = latent_dim\n", "\n", " def compile(self, d_optimizer, g_optimizer, loss_fn):\n", " super(GAN, self).compile()\n", " self.d_optimizer = d_optimizer\n", " self.g_optimizer = g_optimizer\n", " self.loss_fn = loss_fn\n", "\n", " def train_step(self, real_images):\n", " if isinstance(real_images, tuple):\n", " real_images = real_images[0]\n", " # Sample random points in the latent space\n", " batch_size = tf.shape(real_images)[0]\n", " random_latent_vectors = tf.random.normal(shape=(batch_size, self.latent_dim))\n", "\n", " # Decode them to fake images\n", " generated_images = self.generator(random_latent_vectors)\n", "\n", " # Combine them with real images\n", " combined_images = tf.concat([generated_images, real_images], axis=0)\n", "\n", " # Assemble labels discriminating real from fake images\n", " labels = tf.concat(\n", " [tf.ones((batch_size, 1)), tf.zeros((batch_size, 1))], axis=0\n", " )\n", " # Add random noise to the labels - important trick!\n", " labels += 0.05 * tf.random.uniform(tf.shape(labels))\n", "\n", " # Train the discriminator\n", " with tf.GradientTape() as tape:\n", " predictions = self.discriminator(combined_images)\n", " d_loss = self.loss_fn(labels, predictions)\n", " grads = tape.gradient(d_loss, self.discriminator.trainable_weights)\n", " self.d_optimizer.apply_gradients(\n", " zip(grads, self.discriminator.trainable_weights)\n", " )\n", "\n", " # Sample random points in the latent space\n", " random_latent_vectors = tf.random.normal(shape=(batch_size, self.latent_dim))\n", "\n", " # Assemble labels that say \"all real images\"\n", " misleading_labels = tf.zeros((batch_size, 1))\n", "\n", " # Train the generator (note that we should *not* update the weights\n", " # of the discriminator)!\n", " with tf.GradientTape() as tape:\n", " predictions = self.discriminator(self.generator(random_latent_vectors))\n", " g_loss = self.loss_fn(misleading_labels, predictions)\n", " grads = tape.gradient(g_loss, self.generator.trainable_weights)\n", " self.g_optimizer.apply_gradients(zip(grads, self.generator.trainable_weights))\n", " return {\"d_loss\": d_loss, \"g_loss\": g_loss}\n" ] }, { "cell_type": "markdown", "metadata": { "id": "095c499a6149" }, "source": [ "让我们对其进行测试：" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "id": "46832f2077ac", "vscode": { "languageId": "python" } }, "outputs": [], "source": [ "# Prepare the dataset. We use both the training & test MNIST digits.\n", "batch_size = 64\n", "(x_train, _), (x_test, _) = keras.datasets.mnist.load_data()\n", "all_digits = np.concatenate([x_train, x_test])\n", "all_digits = all_digits.astype(\"float32\") / 255.0\n", "all_digits = np.reshape(all_digits, (-1, 28, 28, 1))\n", "dataset = tf.data.Dataset.from_tensor_slices(all_digits)\n", "dataset = dataset.shuffle(buffer_size=1024).batch(batch_size)\n", "\n", "gan = GAN(discriminator=discriminator, generator=generator, latent_dim=latent_dim)\n", "gan.compile(\n", " d_optimizer=keras.optimizers.Adam(learning_rate=0.0003),\n", " g_optimizer=keras.optimizers.Adam(learning_rate=0.0003),\n", " loss_fn=keras.losses.BinaryCrossentropy(from_logits=True),\n", ")\n", "\n", "# To limit the execution time, we only train on 100 batches. You can train on\n", "# the entire dataset. You will need about 20 epochs to get nice results.\n", "gan.fit(dataset.take(100), epochs=1)" ] }, { "cell_type": "markdown", "metadata": { "id": "2ed211016c96" }, "source": [ "深度学习背后的思想十分简单，那么实现又何必复杂呢？" ] } ], "metadata": { "colab": { "collapsed_sections": [], "name": "customizing_what_happens_in_fit.ipynb", "toc_visible": true }, "kernelspec": { "display_name": "Python 3", "name": "python3" } }, "nbformat": 4, "nbformat_minor": 0 }