CINXE.COM

<!DOCTYPE html> <html lang="en"> <head> <meta charset="utf-8"> <meta name="viewport" content="width=device-width, initial-scale=1"> <meta name="description" content="Keras documentation"> <meta name="author" content="Keras Team"> <link rel="shortcut icon" href="https://keras.io/img/favicon.ico"> <link rel="canonical" href="https://keras.io/getting_started/intro_to_keras_for_engineers/" />  <meta property="og:title" content="Keras documentation: Introduction to Keras for engineers"> <meta property="og:image" content="https://keras.io/img/logo-k-keras-wb.png"> <meta name="twitter:title" content="Keras documentation: Introduction to Keras for engineers"> <meta name="twitter:image" content="https://keras.io/img/k-keras-social.png"> <meta name="twitter:card" content="summary"> <title>Introduction to Keras for engineers</title>  <link href="/css/bootstrap.min.css" rel="stylesheet">  <link href="https://fonts.googleapis.com/css2?family=Open+Sans:wght@400;600;700;800&display=swap" rel="stylesheet">  <link href="/css/docs.css" rel="stylesheet"> <link href="/css/monokai.css" rel="stylesheet">  <script>(function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start': new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0], j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src= 'https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f); })(window,document,'script','dataLayer','GTM-5DNGF4N'); </script> <script> (function(i,s,o,g,r,a,m){i['GoogleAnalyticsObject']=r;i[r]=i[r]||function(){ (i[r].q=i[r].q||[]).push(arguments)},i[r].l=1*new Date();a=s.createElement(o), m=s.getElementsByTagName(o)[0];a.async=1;a.src=g;m.parentNode.insertBefore(a,m) })(window,document,'script','https://www.google-analytics.com/analytics.js','ga'); ga('create', 'UA-175165319-128', 'auto'); ga('send', 'pageview'); </script>  <script async defer src="https://buttons.github.io/buttons.js"></script> </head> <body>  <noscript><iframe src="https://www.googletagmanager.com/ns.html?id=GTM-5DNGF4N" height="0" width="0" style="display:none;visibility:hidden"></iframe></noscript>  <div class='k-page'> <div class="k-nav" id="nav-menu"> <a href='/'><img src='/img/logo-small.png' class='logo-small' /></a> <div class="nav flex-column nav-pills" role="tablist" aria-orientation="vertical"> <a class="nav-link" href="/about/" role="tab" aria-selected="">About Keras</a> <a class="nav-link active" href="/getting_started/" role="tab" aria-selected="">Getting started</a> <a class="nav-sublink active" href="/getting_started/intro_to_keras_for_engineers/">Introduction to Keras for engineers</a> <a class="nav-sublink" href="/getting_started/benchmarks/">Keras 3 benchmarks</a> <a class="nav-sublink" href="/getting_started/ecosystem/">The Keras ecosystem</a> <a class="nav-sublink" href="/getting_started/faq/">Frequently Asked Questions</a> <a class="nav-link" href="/guides/" role="tab" aria-selected="">Developer guides</a> <a class="nav-link" href="/api/" role="tab" aria-selected="">Keras 3 API documentation</a> <a class="nav-link" href="/2.18/api/" role="tab" aria-selected="">Keras 2 API documentation</a> <a class="nav-link" href="/examples/" role="tab" aria-selected="">Code examples</a> <a class="nav-link" href="/keras_tuner/" role="tab" aria-selected="">KerasTuner: Hyperparameter Tuning</a> <a class="nav-link" href="/keras_hub/" role="tab" aria-selected="">KerasHub: Pretrained Models</a> <a class="nav-link" href="/keras_cv/" role="tab" aria-selected="">KerasCV: Computer Vision Workflows</a> <a class="nav-link" href="/keras_nlp/" role="tab" aria-selected="">KerasNLP: Natural Language Workflows</a> </div> </div> <div class='k-main'> <div class='k-main-top'> <script> function displayDropdownMenu() { e = document.getElementById("nav-menu"); if (e.style.display == "block") { e.style.display = "none"; } else { e.style.display = "block"; document.getElementById("dropdown-nav").style.display = "block"; } } function resetMobileUI() { if (window.innerWidth <= 840) { document.getElementById("nav-menu").style.display = "none"; document.getElementById("dropdown-nav").style.display = "block"; } else { document.getElementById("nav-menu").style.display = "block"; document.getElementById("dropdown-nav").style.display = "none"; } var navmenu = document.getElementById("nav-menu"); var menuheight = navmenu.clientHeight; var kmain = document.getElementById("k-main-id"); kmain.style.minHeight = (menuheight + 100) + 'px'; } window.onresize = resetMobileUI; window.addEventListener("load", (event) => { resetMobileUI() }); </script> <div id='dropdown-nav' onclick="displayDropdownMenu();"> <svg viewBox="-20 -20 120 120" width="60" height="60"> <rect width="100" height="20"></rect> <rect y="30" width="100" height="20"></rect> <rect y="60" width="100" height="20"></rect> </svg> </div> <form class="bd-search d-flex align-items-center k-search-form" id="search-form"> <input type="search" class="k-search-input" id="search-input" placeholder="Search Keras documentation..." aria-label="Search Keras documentation..." autocomplete="off"> <button class="k-search-btn"> <svg width="13" height="13" viewBox="0 0 13 13"><title>search</title><path d="m4.8495 7.8226c0.82666 0 1.5262-0.29146 2.0985-0.87438 0.57232-0.58292 0.86378-1.2877 0.87438-2.1144 0.010599-0.82666-0.28086-1.5262-0.87438-2.0985-0.59352-0.57232-1.293-0.86378-2.0985-0.87438-0.8055-0.010599-1.5103 0.28086-2.1144 0.87438-0.60414 0.59352-0.8956 1.293-0.87438 2.0985 0.021197 0.8055 0.31266 1.5103 0.87438 2.1144 0.56172 0.60414 1.2665 0.8956 2.1144 0.87438zm4.4695 0.2115 3.681 3.6819-1.259 1.284-3.6817-3.7 0.0019784-0.69479-0.090043-0.098846c-0.87973 0.76087-1.92 1.1413-3.1207 1.1413-1.3553 0-2.5025-0.46363-3.4417-1.3909s-1.4088-2.0686-1.4088-3.4239c0-1.3553 0.4696-2.4966 1.4088-3.4239 0.9392-0.92727 2.0864-1.3969 3.4417-1.4088 1.3553-0.011889 2.4906 0.45771 3.406 1.4088 0.9154 0.95107 1.379 2.0924 1.3909 3.4239 0 1.2126-0.38043 2.2588-1.1413 3.1385l0.098834 0.090049z"></path></svg> </button> </form> <script> var form = document.getElementById('search-form'); form.onsubmit = function(e) { e.preventDefault(); var query = document.getElementById('search-input').value; window.location.href = '/search.html?query=' + query; return False } </script> </div> <div class='k-main-inner' id='k-main-id'> <div class='k-location-slug'> ► <a href='/getting_started/'>Getting started</a> / Introduction to Keras for engineers </div> <div class='k-content'> <h1 id="introduction-to-keras-for-engineers">Introduction to Keras for engineers</h1> Author: <a href="https://twitter.com/fchollet">fchollet</a> Date created: 2023/07/10 Last modified: 2023/07/10 Description: First contact with Keras 3. <img class="k-inline-icon" src="https://colab.research.google.com/img/colab_favicon.ico"/> <a href="https://colab.research.google.com/github/keras-team/keras-io/blob/master/guides/ipynb/intro_to_keras_for_engineers.ipynb">View in Colab</a> •<img class="k-inline-icon" src="https://github.com/favicon.ico"/> <a href="https://github.com/keras-team/keras-io/blob/master/guides/intro_to_keras_for_engineers.py">GitHub source</a> <hr /> <h2 id="introduction">Introduction</h2> Keras 3 is a deep learning framework works with TensorFlow, JAX, and PyTorch interchangeably. This notebook will walk you through key Keras 3 workflows. <hr /> <h2 id="setup">Setup</h2> We're going to be using the JAX backend here – but you can edit the string below to <code>"tensorflow"</code> or <code>"torch"</code> and hit "Restart runtime", and the whole notebook will run just the same! This entire guide is backend-agnostic. <div class="codehilite"><pre><code>import numpy as np import os os.environ["KERAS_BACKEND"] = "jax" # Note that Keras should only be imported after the backend # has been configured. The backend cannot be changed once the # package is imported. import keras </code></pre></div> <hr /> <h2 id="a-first-example-a-mnist-convnet">A first example: A MNIST convnet</h2> Let's start with the Hello World of ML: training a convnet to classify MNIST digits. Here's the data: <div class="codehilite"><pre><code># Load the data and split it between train and test sets (x_train, y_train), (x_test, y_test) = keras.datasets.mnist.load_data() # Scale images to the [0, 1] range x_train = x_train.astype("float32") / 255 x_test = x_test.astype("float32") / 255 # Make sure images have shape (28, 28, 1) x_train = np.expand_dims(x_train, -1) x_test = np.expand_dims(x_test, -1) print("x_train shape:", x_train.shape) print("y_train shape:", y_train.shape) print(x_train.shape[0], "train samples") print(x_test.shape[0], "test samples") </code></pre></div> <div class="k-default-codeblock"> <div class="codehilite"><pre><code>x_train shape: (60000, 28, 28, 1) y_train shape: (60000,) 60000 train samples 10000 test samples </code></pre></div> </div> Here's our model. Different model-building options that Keras offers include: <ul> <li><a href="https://keras.io/guides/sequential_model/">The Sequential API</a> (what we use below)</li> <li><a href="https://keras.io/guides/functional_api/">The Functional API</a> (most typical)</li> <li><a href="https://keras.io/guides/making_new_layers_and_models_via_subclassing/">Writing your own models yourself via subclassing</a> (for advanced use cases)</li> </ul> <div class="codehilite"><pre><code># Model parameters num_classes = 10 input_shape = (28, 28, 1) model = keras.Sequential( [ keras.layers.Input(shape=input_shape), keras.layers.Conv2D(64, kernel_size=(3, 3), activation="relu"), keras.layers.Conv2D(64, kernel_size=(3, 3), activation="relu"), keras.layers.MaxPooling2D(pool_size=(2, 2)), keras.layers.Conv2D(128, kernel_size=(3, 3), activation="relu"), keras.layers.Conv2D(128, kernel_size=(3, 3), activation="relu"), keras.layers.GlobalAveragePooling2D(), keras.layers.Dropout(0.5), keras.layers.Dense(num_classes, activation="softmax"), ] ) </code></pre></div> Here's our model summary: <div class="codehilite"><pre><code>model.summary() </code></pre></div> <pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace">Model: "sequential" </pre> <pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace">┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┓ ┃ Layer (type) ┃ Output Shape ┃ Param # ┃ ┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━┩ │ conv2d (Conv2D) │ (None, 26, 26, 64) │ 640 │ ├─────────────────────────────────┼───────────────────────────┼────────────┤ │ conv2d_1 (Conv2D) │ (None, 24, 24, 64) │ 36,928 │ ├─────────────────────────────────┼───────────────────────────┼────────────┤ │ max_pooling2d (MaxPooling2D) │ (None, 12, 12, 64) │ 0 │ ├─────────────────────────────────┼───────────────────────────┼────────────┤ │ conv2d_2 (Conv2D) │ (None, 10, 10, 128) │ 73,856 │ ├─────────────────────────────────┼───────────────────────────┼────────────┤ │ conv2d_3 (Conv2D) │ (None, 8, 8, 128) │ 147,584 │ ├─────────────────────────────────┼───────────────────────────┼────────────┤ │ global_average_pooling2d │ (None, 128) │ 0 │ │ (GlobalAveragePooling2D) │ │ │ ├─────────────────────────────────┼───────────────────────────┼────────────┤ │ dropout (Dropout) │ (None, 128) │ 0 │ ├─────────────────────────────────┼───────────────────────────┼────────────┤ │ dense (Dense) │ (None, 10) │ 1,290 │ └─────────────────────────────────┴───────────────────────────┴────────────┘ </pre> <pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace"> Total params: 260,298 (1016.79 KB) </pre> <pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace"> Trainable params: 260,298 (1016.79 KB) </pre> <pre style="white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace"> Non-trainable params: 0 (0.00 B) </pre> We use the <code>compile()</code> method to specify the optimizer, loss function, and the metrics to monitor. Note that with the JAX and TensorFlow backends, XLA compilation is turned on by default. <div class="codehilite"><pre><code>model.compile( loss=keras.losses.SparseCategoricalCrossentropy(), optimizer=keras.optimizers.Adam(learning_rate=1e-3), metrics=[ keras.metrics.SparseCategoricalAccuracy(name="acc"), ], ) </code></pre></div> Let's train and evaluate the model. We'll set aside a validation split of 15% of the data during training to monitor generalization on unseen data. <div class="codehilite"><pre><code>batch_size = 128 epochs = 20 callbacks = [ keras.callbacks.ModelCheckpoint(filepath="model_at_epoch_{epoch}.keras"), keras.callbacks.EarlyStopping(monitor="val_loss", patience=2), ] model.fit( x_train, y_train, batch_size=batch_size, epochs=epochs, validation_split=0.15, callbacks=callbacks, ) score = model.evaluate(x_test, y_test, verbose=0) </code></pre></div> <div class="k-default-codeblock"> <div class="codehilite"><pre><code>Epoch 1/20 399/399 ━━━━━━━━━━━━━━━━━━━━ 74s 184ms/step - acc: 0.4980 - loss: 1.3832 - val_acc: 0.9609 - val_loss: 0.1513 Epoch 2/20 399/399 ━━━━━━━━━━━━━━━━━━━━ 74s 186ms/step - acc: 0.9245 - loss: 0.2487 - val_acc: 0.9702 - val_loss: 0.0999 Epoch 3/20 399/399 ━━━━━━━━━━━━━━━━━━━━ 70s 175ms/step - acc: 0.9515 - loss: 0.1647 - val_acc: 0.9816 - val_loss: 0.0608 Epoch 4/20 399/399 ━━━━━━━━━━━━━━━━━━━━ 69s 174ms/step - acc: 0.9622 - loss: 0.1247 - val_acc: 0.9833 - val_loss: 0.0541 Epoch 5/20 399/399 ━━━━━━━━━━━━━━━━━━━━ 68s 171ms/step - acc: 0.9685 - loss: 0.1083 - val_acc: 0.9860 - val_loss: 0.0468 Epoch 6/20 399/399 ━━━━━━━━━━━━━━━━━━━━ 70s 176ms/step - acc: 0.9710 - loss: 0.0955 - val_acc: 0.9897 - val_loss: 0.0400 Epoch 7/20 399/399 ━━━━━━━━━━━━━━━━━━━━ 69s 172ms/step - acc: 0.9742 - loss: 0.0853 - val_acc: 0.9888 - val_loss: 0.0388 Epoch 8/20 399/399 ━━━━━━━━━━━━━━━━━━━━ 68s 169ms/step - acc: 0.9789 - loss: 0.0738 - val_acc: 0.9902 - val_loss: 0.0387 Epoch 9/20 399/399 ━━━━━━━━━━━━━━━━━━━━ 75s 187ms/step - acc: 0.9789 - loss: 0.0691 - val_acc: 0.9907 - val_loss: 0.0341 Epoch 10/20 399/399 ━━━━━━━━━━━━━━━━━━━━ 77s 194ms/step - acc: 0.9806 - loss: 0.0636 - val_acc: 0.9907 - val_loss: 0.0348 Epoch 11/20 399/399 ━━━━━━━━━━━━━━━━━━━━ 74s 186ms/step - acc: 0.9812 - loss: 0.0610 - val_acc: 0.9926 - val_loss: 0.0271 Epoch 12/20 399/399 ━━━━━━━━━━━━━━━━━━━━ 219s 550ms/step - acc: 0.9820 - loss: 0.0590 - val_acc: 0.9912 - val_loss: 0.0294 Epoch 13/20 399/399 ━━━━━━━━━━━━━━━━━━━━ 70s 176ms/step - acc: 0.9843 - loss: 0.0504 - val_acc: 0.9918 - val_loss: 0.0316 </code></pre></div> </div> During training, we were saving a model at the end of each epoch. You can also save the model in its latest state like this: <div class="codehilite"><pre><code>model.save("final_model.keras") </code></pre></div> And reload it like this: <div class="codehilite"><pre><code>model = keras.saving.load_model("final_model.keras") </code></pre></div> Next, you can query predictions of class probabilities with <code>predict()</code>: <div class="codehilite"><pre><code>predictions = model.predict(x_test) </code></pre></div> <div class="k-default-codeblock"> <div class="codehilite"><pre><code> 313/313 ━━━━━━━━━━━━━━━━━━━━ 3s 9ms/step </code></pre></div> </div> That's it for the basics! <hr /> <h2 id="writing-crossframework-custom-components">Writing cross-framework custom components</h2> Keras enables you to write custom Layers, Models, Metrics, Losses, and Optimizers that work across TensorFlow, JAX, and PyTorch with the same codebase. Let's take a look at custom layers first. The <code>keras.ops</code> namespace contains: <ul> <li>An implementation of the NumPy API, e.g. <a href="/api/ops/numpy#stack-function"><code>keras.ops.stack</code></a> or <a href="/api/ops/numpy#matmul-function"><code>keras.ops.matmul</code></a>.</li> <li>A set of neural network specific ops that are absent from NumPy, such as <a href="/api/ops/nn#conv-function"><code>keras.ops.conv</code></a> or <a href="/api/ops/nn#binarycrossentropy-function"><code>keras.ops.binary_crossentropy</code></a>.</li> </ul> Let's make a custom <code>Dense</code> layer that works with all backends: <div class="codehilite"><pre><code>class MyDense(keras.layers.Layer): def __init__(self, units, activation=None, name=None): super().__init__(name=name) self.units = units self.activation = keras.activations.get(activation) def build(self, input_shape): input_dim = input_shape[-1] self.w = self.add_weight( shape=(input_dim, self.units), initializer=keras.initializers.GlorotNormal(), name="kernel", trainable=True, ) self.b = self.add_weight( shape=(self.units,), initializer=keras.initializers.Zeros(), name="bias", trainable=True, ) def call(self, inputs): # Use Keras ops to create backend-agnostic layers/metrics/etc. x = keras.ops.matmul(inputs, self.w) + self.b return self.activation(x) </code></pre></div> Next, let's make a custom <code>Dropout</code> layer that relies on the <code>keras.random</code> namespace: <div class="codehilite"><pre><code>class MyDropout(keras.layers.Layer): def __init__(self, rate, name=None): super().__init__(name=name) self.rate = rate # Use seed_generator for managing RNG state. # It is a state element and its seed variable is # tracked as part of `layer.variables`. self.seed_generator = keras.random.SeedGenerator(1337) def call(self, inputs): # Use `keras.random` for random ops. return keras.random.dropout(inputs, self.rate, seed=self.seed_generator) </code></pre></div> Next, let's write a custom subclassed model that uses our two custom layers: <div class="codehilite"><pre><code>class MyModel(keras.Model): def __init__(self, num_classes): super().__init__() self.conv_base = keras.Sequential( [ keras.layers.Conv2D(64, kernel_size=(3, 3), activation="relu"), keras.layers.Conv2D(64, kernel_size=(3, 3), activation="relu"), keras.layers.MaxPooling2D(pool_size=(2, 2)), keras.layers.Conv2D(128, kernel_size=(3, 3), activation="relu"), keras.layers.Conv2D(128, kernel_size=(3, 3), activation="relu"), keras.layers.GlobalAveragePooling2D(), ] ) self.dp = MyDropout(0.5) self.dense = MyDense(num_classes, activation="softmax") def call(self, x): x = self.conv_base(x) x = self.dp(x) return self.dense(x) </code></pre></div> Let's compile it and fit it: <div class="codehilite"><pre><code>model = MyModel(num_classes=10) model.compile( loss=keras.losses.SparseCategoricalCrossentropy(), optimizer=keras.optimizers.Adam(learning_rate=1e-3), metrics=[ keras.metrics.SparseCategoricalAccuracy(name="acc"), ], ) model.fit( x_train, y_train, batch_size=batch_size, epochs=1, # For speed validation_split=0.15, ) </code></pre></div> <div class="k-default-codeblock"> <div class="codehilite"><pre><code> 399/399 ━━━━━━━━━━━━━━━━━━━━ 70s 174ms/step - acc: 0.5104 - loss: 1.3473 - val_acc: 0.9256 - val_loss: 0.2484 <keras.src.callbacks.history.History at 0x105608670> </code></pre></div> </div> <hr /> <h2 id="training-models-on-arbitrary-data-sources">Training models on arbitrary data sources</h2> All Keras models can be trained and evaluated on a wide variety of data sources, independently of the backend you're using. This includes: <ul> <li>NumPy arrays</li> <li>Pandas dataframes</li> <li>TensorFlow <a href="https://www.tensorflow.org/api_docs/python/tf/data/Dataset"><code>tf.data.Dataset</code></a> objects</li> <li>PyTorch <code>DataLoader</code> objects</li> <li>Keras <code>PyDataset</code> objects</li> </ul> They all work whether you're using TensorFlow, JAX, or PyTorch as your Keras backend. Let's try it out with PyTorch <code>DataLoaders</code>: <div class="codehilite"><pre><code>import torch # Create a TensorDataset train_torch_dataset = torch.utils.data.TensorDataset( torch.from_numpy(x_train), torch.from_numpy(y_train) ) val_torch_dataset = torch.utils.data.TensorDataset( torch.from_numpy(x_test), torch.from_numpy(y_test) ) # Create a DataLoader train_dataloader = torch.utils.data.DataLoader( train_torch_dataset, batch_size=batch_size, shuffle=True ) val_dataloader = torch.utils.data.DataLoader( val_torch_dataset, batch_size=batch_size, shuffle=False ) model = MyModel(num_classes=10) model.compile( loss=keras.losses.SparseCategoricalCrossentropy(), optimizer=keras.optimizers.Adam(learning_rate=1e-3), metrics=[ keras.metrics.SparseCategoricalAccuracy(name="acc"), ], ) model.fit(train_dataloader, epochs=1, validation_data=val_dataloader) </code></pre></div> <div class="k-default-codeblock"> <div class="codehilite"><pre><code> 469/469 ━━━━━━━━━━━━━━━━━━━━ 81s 172ms/step - acc: 0.5502 - loss: 1.2550 - val_acc: 0.9419 - val_loss: 0.1972 <keras.src.callbacks.history.History at 0x2b3385480> </code></pre></div> </div> Now let's try this out with <a href="https://www.tensorflow.org/api_docs/python/tf/data"><code>tf.data</code></a>: <div class="codehilite"><pre><code>import tensorflow as tf train_dataset = ( tf.data.Dataset.from_tensor_slices((x_train, y_train)) .batch(batch_size) .prefetch(tf.data.AUTOTUNE) ) test_dataset = ( tf.data.Dataset.from_tensor_slices((x_test, y_test)) .batch(batch_size) .prefetch(tf.data.AUTOTUNE) ) model = MyModel(num_classes=10) model.compile( loss=keras.losses.SparseCategoricalCrossentropy(), optimizer=keras.optimizers.Adam(learning_rate=1e-3), metrics=[ keras.metrics.SparseCategoricalAccuracy(name="acc"), ], ) model.fit(train_dataset, epochs=1, validation_data=test_dataset) </code></pre></div> <div class="k-default-codeblock"> <div class="codehilite"><pre><code> 469/469 ━━━━━━━━━━━━━━━━━━━━ 81s 172ms/step - acc: 0.5771 - loss: 1.1948 - val_acc: 0.9229 - val_loss: 0.2502 <keras.src.callbacks.history.History at 0x2b33e7df0> </code></pre></div> </div> <hr /> <h2 id="further-reading">Further reading</h2> This concludes our short overview of the new multi-backend capabilities of Keras 3. Next, you can learn about: <h3 id="how-to-customize-what-happens-in-fit">How to customize what happens in <code>fit()</code></h3> Want to implement a non-standard training algorithm yourself but still want to benefit from the power and usability of <code>fit()</code>? It's easy to customize <code>fit()</code> to support arbitrary use cases: <ul> <li><a href="http://keras.io/guides/custom_train_step_in_tensorflow/">Customizing what happens in <code>fit()</code> with TensorFlow</a></li> <li><a href="http://keras.io/guides/custom_train_step_in_jax/">Customizing what happens in <code>fit()</code> with JAX</a></li> <li><a href="http://keras.io/guides/custom_train_step_in_torch/">Customizing what happens in <code>fit()</code> with PyTorch</a></li> </ul> <hr /> <h2 id="how-to-write-custom-training-loops">How to write custom training loops</h2> <ul> <li><a href="http://keras.io/guides/writing_a_custom_training_loop_in_tensorflow/">Writing a training loop from scratch in TensorFlow</a></li> <li><a href="http://keras.io/guides/writing_a_custom_training_loop_in_jax/">Writing a training loop from scratch in JAX</a></li> <li><a href="http://keras.io/guides/writing_a_custom_training_loop_in_torch/">Writing a training loop from scratch in PyTorch</a></li> </ul> <hr /> <h2 id="how-to-distribute-training">How to distribute training</h2> <ul> <li><a href="http://keras.io/guides/distributed_training_with_tensorflow/">Guide to distributed training with TensorFlow</a></li> <li><a href="https://github.com/keras-team/keras/blob/master/examples/demo_jax_distributed.py">JAX distributed training example</a></li> <li><a href="https://github.com/keras-team/keras/blob/master/examples/demo_torch_multi_gpu.py">PyTorch distributed training example</a></li> </ul> Enjoy the library! 🚀 </div> <div class='k-outline'> <div class='k-outline-depth-1'> <a href='#introduction-to-keras-for-engineers'>Introduction to Keras for engineers</a> </div> <div class='k-outline-depth-2'> ◆ <a href='#introduction'>Introduction</a> </div> <div class='k-outline-depth-2'> ◆ <a href='#setup'>Setup</a> </div> <div class='k-outline-depth-2'> ◆ <a href='#a-first-example-a-mnist-convnet'>A first example: A MNIST convnet</a> </div> <div class='k-outline-depth-2'> ◆ <a href='#writing-crossframework-custom-components'>Writing cross-framework custom components</a> </div> <div class='k-outline-depth-2'> ◆ <a href='#training-models-on-arbitrary-data-sources'>Training models on arbitrary data sources</a> </div> <div class='k-outline-depth-2'> ◆ <a href='#further-reading'>Further reading</a> </div> <div class='k-outline-depth-3'> <a href='#how-to-customize-what-happens-in-fit'>How to customize what happens in <code>fit()</code></a> </div> <div class='k-outline-depth-2'> ◆ <a href='#how-to-write-custom-training-loops'>How to write custom training loops</a> </div> <div class='k-outline-depth-2'> ◆ <a href='#how-to-distribute-training'>How to distribute training</a> </div> </div> </div> </div> </div> </body> <footer style="float: left; width: 100%; padding: 1em; border-top: solid 1px #bbb;"> <a href="https://policies.google.com/terms">Terms</a> | <a href="https://policies.google.com/privacy">Privacy</a> </footer> </html>