Classification with Gated Residual and Variable Selection Networks

<!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/examples/structured_data/classification_with_grn_and_vsn/" />  <meta property="og:title" content="Keras documentation: Classification with Gated Residual and Variable Selection Networks"> <meta property="og:image" content="https://keras.io/img/logo-k-keras-wb.png"> <meta name="twitter:title" content="Keras documentation: Classification with Gated Residual and Variable Selection Networks"> <meta name="twitter:image" content="https://keras.io/img/k-keras-social.png"> <meta name="twitter:card" content="summary"> <title>Classification with Gated Residual and Variable Selection Networks</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" href="/getting_started/" role="tab" aria-selected="">Getting started</a> <a class="nav-link" href="/guides/" role="tab" aria-selected="">Developer guides</a> <a class="nav-link active" href="/examples/" role="tab" aria-selected="">Code examples</a> <a class="nav-sublink" href="/examples/vision/">Computer Vision</a> <a class="nav-sublink" href="/examples/nlp/">Natural Language Processing</a> <a class="nav-sublink active" href="/examples/structured_data/">Structured Data</a> <a class="nav-sublink2" href="/examples/structured_data/structured_data_classification_with_feature_space/">Structured data classification with FeatureSpace</a> <a class="nav-sublink2" href="/examples/structured_data/feature_space_advanced/">FeatureSpace advanced use cases</a> <a class="nav-sublink2" href="/examples/structured_data/imbalanced_classification/">Imbalanced classification: credit card fraud detection</a> <a class="nav-sublink2" href="/examples/structured_data/structured_data_classification_from_scratch/">Structured data classification from scratch</a> <a class="nav-sublink2" href="/examples/structured_data/wide_deep_cross_networks/">Structured data learning with Wide, Deep, and Cross networks</a> <a class="nav-sublink2" href="/examples/structured_data/customer_lifetime_value/">Deep Learning for Customer Lifetime Value</a> <a class="nav-sublink2 active" href="/examples/structured_data/classification_with_grn_and_vsn/">Classification with Gated Residual and Variable Selection Networks</a> <a class="nav-sublink2" href="/examples/structured_data/classification_with_tfdf/">Classification with TensorFlow Decision Forests</a> <a class="nav-sublink2" href="/examples/structured_data/deep_neural_decision_forests/">Classification with Neural Decision Forests</a> <a class="nav-sublink2" href="/examples/structured_data/tabtransformer/">Structured data learning with TabTransformer</a> <a class="nav-sublink2" href="/examples/structured_data/collaborative_filtering_movielens/">Collaborative Filtering for Movie Recommendations</a> <a class="nav-sublink2" href="/examples/structured_data/movielens_recommendations_transformers/">A Transformer-based recommendation system</a> <a class="nav-sublink" href="/examples/timeseries/">Timeseries</a> <a class="nav-sublink" href="/examples/generative/">Generative Deep Learning</a> <a class="nav-sublink" href="/examples/audio/">Audio Data</a> <a class="nav-sublink" href="/examples/rl/">Reinforcement Learning</a> <a class="nav-sublink" href="/examples/graph/">Graph Data</a> <a class="nav-sublink" href="/examples/keras_recipes/">Quick Keras Recipes</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="/keras_tuner/" role="tab" aria-selected="">KerasTuner: Hyperparam Tuning</a> <a class="nav-link" href="/keras_hub/" role="tab" aria-selected="">KerasHub: Pretrained Models</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='/examples/'>Code examples</a> / <a href='/examples/structured_data/'>Structured Data</a> / Classification with Gated Residual and Variable Selection Networks </div> <div class='k-content'> <h1 id="classification-with-gated-residual-and-variable-selection-networks">Classification with Gated Residual and Variable Selection Networks</h1> Author: <a href="https://www.linkedin.com/in/khalid-salama-24403144/">Khalid Salama</a> Date created: 2021/02/10 Last modified: 2025/01/08 Description: Using Gated Residual and Variable Selection Networks for income level prediction. <div class='example_version_banner keras_3'>ⓘ This example uses Keras 3</div> <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/examples/structured_data/ipynb/classification_with_grn_and_vsn.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/examples/structured_data/classification_with_grn_and_vsn.py">GitHub source</a> <hr /> <h2 id="introduction">Introduction</h2> This example demonstrates the use of Gated Residual Networks (GRN) and Variable Selection Networks (VSN), proposed by Bryan Lim et al. in <a href="https://arxiv.org/abs/1912.09363">Temporal Fusion Transformers (TFT) for Interpretable Multi-horizon Time Series Forecasting</a>, for structured data classification. GRNs give the flexibility to the model to apply non-linear processing only where needed. VSNs allow the model to softly remove any unnecessary noisy inputs which could negatively impact performance. Together, those techniques help improving the learning capacity of deep neural network models. Note that this example implements only the GRN and VSN components described in in the paper, rather than the whole TFT model, as GRN and VSN can be useful on their own for structured data learning tasks. To run the code you need to use TensorFlow 2.3 or higher. <hr /> <h2 id="the-dataset">The dataset</h2> This example uses the <a href="https://archive.ics.uci.edu/ml/datasets/Census-Income+%28KDD%29">United States Census Income Dataset</a> provided by the <a href="https://archive.ics.uci.edu/ml/index.php">UC Irvine Machine Learning Repository</a>. The task is binary classification to determine whether a person makes over 50K a year. The dataset includes ~300K instances with 41 input features: 7 numerical features and 34 categorical features. <hr /> <h2 id="setup">Setup</h2> <div class="codehilite"><pre><code>import os import subprocess import tarfile os.environ["KERAS_BACKEND"] = "torch" # or jax, or tensorflow import numpy as np import pandas as pd import keras from keras import layers </code></pre></div> <hr /> <h2 id="prepare-the-data">Prepare the data</h2> First we load the data from the UCI Machine Learning Repository into a Pandas DataFrame. <div class="codehilite"><pre><code># Column names. CSV_HEADER = [ "age", "class_of_worker", "detailed_industry_recode", "detailed_occupation_recode", "education", "wage_per_hour", "enroll_in_edu_inst_last_wk", "marital_stat", "major_industry_code", "major_occupation_code", "race", "hispanic_origin", "sex", "member_of_a_labor_union", "reason_for_unemployment", "full_or_part_time_employment_stat", "capital_gains", "capital_losses", "dividends_from_stocks", "tax_filer_stat", "region_of_previous_residence", "state_of_previous_residence", "detailed_household_and_family_stat", "detailed_household_summary_in_household", "instance_weight", "migration_code-change_in_msa", "migration_code-change_in_reg", "migration_code-move_within_reg", "live_in_this_house_1_year_ago", "migration_prev_res_in_sunbelt", "num_persons_worked_for_employer", "family_members_under_18", "country_of_birth_father", "country_of_birth_mother", "country_of_birth_self", "citizenship", "own_business_or_self_employed", "fill_inc_questionnaire_for_veterans_admin", "veterans_benefits", "weeks_worked_in_year", "year", "income_level", ] data_url = "https://archive.ics.uci.edu/static/public/117/census+income+kdd.zip" keras.utils.get_file(origin=data_url, extract=True) </code></pre></div> <div class="k-default-codeblock"> <div class="codehilite"><pre><code>'/home/humbulani/.keras/datasets/census+income+kdd.zip' </code></pre></div> </div> Determine the downloaded .tar.gz file path and extract the files from the downloaded .tar.gz file <div class="codehilite"><pre><code>extracted_path = os.path.join( os.path.expanduser("~"), ".keras", "datasets", "census+income+kdd.zip" ) for root, dirs, files in os.walk(extracted_path): for file in files: if file.endswith(".tar.gz"): tar_gz_path = os.path.join(root, file) with tarfile.open(tar_gz_path, "r:gz") as tar: tar.extractall(path=root) train_data_path = os.path.join( os.path.expanduser("~"), ".keras", "datasets", "census+income+kdd.zip", "census-income.data", ) test_data_path = os.path.join( os.path.expanduser("~"), ".keras", "datasets", "census+income+kdd.zip", "census-income.test", ) data = pd.read_csv(train_data_path, header=None, names=CSV_HEADER) test_data = pd.read_csv(test_data_path, header=None, names=CSV_HEADER) print(f"Data shape: {data.shape}") print(f"Test data shape: {test_data.shape}") </code></pre></div> <div class="k-default-codeblock"> <div class="codehilite"><pre><code>Data shape: (199523, 42) Test data shape: (99762, 42) </code></pre></div> </div> We convert the target column from string to integer. <div class="codehilite"><pre><code>data["income_level"] = data["income_level"].apply( lambda x: 0 if x == " - 50000." else 1 ) test_data["income_level"] = test_data["income_level"].apply( lambda x: 0 if x == " - 50000." else 1 ) </code></pre></div> Then, We split the dataset into train and validation sets. <div class="codehilite"><pre><code>random_selection = np.random.rand(len(data.index)) <= 0.85 train_data = data[random_selection] valid_data = data[~random_selection] </code></pre></div> Finally we store the train and test data splits locally to CSV files. <div class="codehilite"><pre><code>train_data_file = "train_data.csv" valid_data_file = "valid_data.csv" test_data_file = "test_data.csv" train_data.to_csv(train_data_file, index=False, header=False) valid_data.to_csv(valid_data_file, index=False, header=False) test_data.to_csv(test_data_file, index=False, header=False) </code></pre></div> <hr /> <h2 id="define-dataset-metadata">Define dataset metadata</h2> Here, we define the metadata of the dataset that will be useful for reading and parsing the data into input features, and encoding the input features with respect to their types. <div class="codehilite"><pre><code># Target feature name. TARGET_FEATURE_NAME = "income_level" # Weight column name. WEIGHT_COLUMN_NAME = "instance_weight" # Numeric feature names. NUMERIC_FEATURE_NAMES = [ "age", "wage_per_hour", "capital_gains", "capital_losses", "dividends_from_stocks", "num_persons_worked_for_employer", "weeks_worked_in_year", ] # Categorical features and their vocabulary lists. # Note that we add 'v=' as a prefix to all categorical feature values to make # sure that they are treated as strings. CATEGORICAL_FEATURES_WITH_VOCABULARY = { feature_name: sorted([str(value) for value in list(data[feature_name].unique())]) for feature_name in CSV_HEADER if feature_name not in list(NUMERIC_FEATURE_NAMES + [WEIGHT_COLUMN_NAME, TARGET_FEATURE_NAME]) } # All features names. FEATURE_NAMES = NUMERIC_FEATURE_NAMES + list( CATEGORICAL_FEATURES_WITH_VOCABULARY.keys() ) # Feature default values. COLUMN_DEFAULTS = [ ( [0.0] if feature_name in NUMERIC_FEATURE_NAMES + [TARGET_FEATURE_NAME, WEIGHT_COLUMN_NAME] else ["NA"] ) for feature_name in CSV_HEADER ] </code></pre></div> <hr /> <h2 id="tfdatadataset">Create a <a href="https://www.tensorflow.org/api_docs/python/tf/data/Dataset"><code>tf.data.Dataset</code></a> for training and evaluation</h2> We create an input function to read and parse the file, and convert features and labels into a <a href="https://www.tensorflow.org/guide/datasets">[<code>tf.data.Dataset</code>](https://www.tensorflow.org/api_docs/python/tf/data/Dataset)</a> for training and evaluation. <div class="codehilite"><pre><code># Tensorflow required for tf.data.Datasets import tensorflow as tf # We process our datasets elements here (categorical) and convert them to indices to avoid this step # during model training since only tensorflow support strings. def process(features, target): for feature_name in features: if feature_name in CATEGORICAL_FEATURES_WITH_VOCABULARY: # Cast categorical feature values to string. features[feature_name] = tf.cast(features[feature_name], "string") vocabulary = CATEGORICAL_FEATURES_WITH_VOCABULARY[feature_name] # Create a lookup to convert a string values to an integer indices. # Since we are not using a mask token nor expecting any out of vocabulary # (oov) token, we set mask_token to None and num_oov_indices to 0. index = layers.StringLookup( vocabulary=vocabulary, mask_token=None, num_oov_indices=0, output_mode="int", ) # Convert the string input values into integer indices. value_index = index(features[feature_name]) features[feature_name] = value_index else: # Do nothing for numerical features pass # Get the instance weight. weight = features.pop(WEIGHT_COLUMN_NAME) # Change features from OrderedDict to Dict to match Inputs as they are Dict. return dict(features), target, weight def get_dataset_from_csv(csv_file_path, shuffle=False, batch_size=128): dataset = tf.data.experimental.make_csv_dataset( csv_file_path, batch_size=batch_size, column_names=CSV_HEADER, column_defaults=COLUMN_DEFAULTS, label_name=TARGET_FEATURE_NAME, num_epochs=1, header=False, shuffle=shuffle, ).map(process) return dataset </code></pre></div> <hr /> <h2 id="create-model-inputs">Create model inputs</h2> <div class="codehilite"><pre><code>def create_model_inputs(): inputs = {} for feature_name in FEATURE_NAMES: if feature_name in CATEGORICAL_FEATURES_WITH_VOCABULARY: # Make them int64, they are Categorical (whole units) inputs[feature_name] = layers.Input( name=feature_name, shape=(), dtype="int64" ) else: # Make them float32, they are Real numbers inputs[feature_name] = layers.Input( name=feature_name, shape=(), dtype="float32" ) return inputs </code></pre></div> <hr /> <h2 id="implement-the-gated-linear-unit">Implement the Gated Linear Unit</h2> <a href="https://arxiv.org/abs/1612.08083">Gated Linear Units (GLUs)</a> provide the flexibility to suppress input that are not relevant for a given task. <div class="codehilite"><pre><code>class GatedLinearUnit(layers.Layer): def __init__(self, units): super().__init__() self.linear = layers.Dense(units) self.sigmoid = layers.Dense(units, activation="sigmoid") def call(self, inputs): return self.linear(inputs) * self.sigmoid(inputs) # Remove build warnings def build(self): self.built = True </code></pre></div> <hr /> <h2 id="implement-the-gated-residual-network">Implement the Gated Residual Network</h2> The Gated Residual Network (GRN) works as follows: <ol> <li>Applies the nonlinear ELU transformation to the inputs.</li> <li>Applies linear transformation followed by dropout.</li> <li>Applies GLU and adds the original inputs to the output of the GLU to perform skip (residual) connection.</li> <li>Applies layer normalization and produces the output.</li> </ol> <div class="codehilite"><pre><code>class GatedResidualNetwork(layers.Layer): def __init__(self, units, dropout_rate): super().__init__() self.units = units self.elu_dense = layers.Dense(units, activation="elu") self.linear_dense = layers.Dense(units) self.dropout = layers.Dropout(dropout_rate) self.gated_linear_unit = GatedLinearUnit(units) self.layer_norm = layers.LayerNormalization() self.project = layers.Dense(units) def call(self, inputs): x = self.elu_dense(inputs) x = self.linear_dense(x) x = self.dropout(x) if inputs.shape[-1] != self.units: inputs = self.project(inputs) x = inputs + self.gated_linear_unit(x) x = self.layer_norm(x) return x # Remove build warnings def build(self): self.built = True </code></pre></div> <hr /> <h2 id="implement-the-variable-selection-network">Implement the Variable Selection Network</h2> The Variable Selection Network (VSN) works as follows: <ol> <li>Applies a GRN to each feature individually.</li> <li>Applies a GRN on the concatenation of all the features, followed by a softmax to produce feature weights.</li> <li>Produces a weighted sum of the output of the individual GRN.</li> </ol> Note that the output of the VSN is [batch_size, encoding_size], regardless of the number of the input features. For categorical features, we encode them using <code>layers.Embedding</code> using the <code>encoding_size</code> as the embedding dimensions. For the numerical features, we apply linear transformation using <code>layers.Dense</code> to project each feature into <code>encoding_size</code>-dimensional vector. Thus, all the encoded features will have the same dimensionality. <div class="codehilite"><pre><code>class VariableSelection(layers.Layer): def __init__(self, num_features, units, dropout_rate): super().__init__() self.units = units # Create an embedding layers with the specified dimensions self.embeddings = dict() for input_ in CATEGORICAL_FEATURES_WITH_VOCABULARY: vocabulary = CATEGORICAL_FEATURES_WITH_VOCABULARY[input_] embedding_encoder = layers.Embedding( input_dim=len(vocabulary), output_dim=self.units, name=input_ ) self.embeddings[input_] = embedding_encoder # Projection layers for numeric features self.proj_layer = dict() for input_ in NUMERIC_FEATURE_NAMES: proj_layer = layers.Dense(units=self.units) self.proj_layer[input_] = proj_layer self.grns = list() # Create a GRN for each feature independently for idx in range(num_features): grn = GatedResidualNetwork(units, dropout_rate) self.grns.append(grn) # Create a GRN for the concatenation of all the features self.grn_concat = GatedResidualNetwork(units, dropout_rate) self.softmax = layers.Dense(units=num_features, activation="softmax") def call(self, inputs): concat_inputs = [] for input_ in inputs: if input_ in CATEGORICAL_FEATURES_WITH_VOCABULARY: max_index = self.embeddings[input_].input_dim - 1 # Clamp the indices # torch had some index errors during embedding hence the clip function embedded_feature = self.embeddings[input_]( keras.ops.clip(inputs[input_], 0, max_index) ) concat_inputs.append(embedded_feature) else: # Project the numeric feature to encoding_size using linear transformation. proj_feature = keras.ops.expand_dims(inputs[input_], -1) proj_feature = self.proj_layer[input_](proj_feature) concat_inputs.append(proj_feature) v = layers.concatenate(concat_inputs) v = self.grn_concat(v) v = keras.ops.expand_dims(self.softmax(v), axis=-1) x = [] for idx, input in enumerate(concat_inputs): x.append(self.grns[idx](input)) x = keras.ops.stack(x, axis=1) return keras.ops.squeeze( keras.ops.matmul(keras.ops.transpose(v, axes=[0, 2, 1]), x), axis=1 ) # to remove the build warnings def build(self): self.built = True </code></pre></div> <hr /> <h2 id="create-gated-residual-and-variable-selection-networks-model">Create Gated Residual and Variable Selection Networks model</h2> <div class="codehilite"><pre><code>def create_model(encoding_size): inputs = create_model_inputs() num_features = len(inputs) features = VariableSelection(num_features, encoding_size, dropout_rate)(inputs) outputs = layers.Dense(units=1, activation="sigmoid")(features) # Functional model model = keras.Model(inputs=inputs, outputs=outputs) return model </code></pre></div> <hr /> <h2 id="compile-train-and-evaluate-the-model">Compile, train, and evaluate the model</h2> <div class="codehilite"><pre><code>learning_rate = 0.001 dropout_rate = 0.15 batch_size = 265 num_epochs = 20 # may be adjusted to a desired value encoding_size = 16 model = create_model(encoding_size) model.compile( optimizer=keras.optimizers.Adam(learning_rate=learning_rate), loss=keras.losses.BinaryCrossentropy(), metrics=[keras.metrics.BinaryAccuracy(name="accuracy")], ) </code></pre></div> Let's visualize our connectivity graph: <div class="codehilite"><pre><code># `rankdir='LR'` is to make the graph horizontal. keras.utils.plot_model(model, show_shapes=True, show_layer_names=True, rankdir="LR") # Create an early stopping callback. early_stopping = keras.callbacks.EarlyStopping( monitor="val_loss", patience=5, restore_best_weights=True ) print("Start training the model...") train_dataset = get_dataset_from_csv( train_data_file, shuffle=True, batch_size=batch_size ) valid_dataset = get_dataset_from_csv(valid_data_file, batch_size=batch_size) model.fit( train_dataset, epochs=num_epochs, validation_data=valid_dataset, callbacks=[early_stopping], ) print("Model training finished.") print("Evaluating model performance...") test_dataset = get_dataset_from_csv(test_data_file, batch_size=batch_size) _, accuracy = model.evaluate(test_dataset) print(f"Test accuracy: {round(accuracy * 100, 2)}%") </code></pre></div> <div class="k-default-codeblock"> <div class="codehilite"><pre><code>Start training the model... </code></pre></div> </div> <div class="k-default-codeblock"> <div class="codehilite"><pre><code> 1/Unknown 1s 698ms/step - accuracy: 0.4717 - loss: 1212.3043 2/Unknown 1s 200ms/step - accuracy: 0.5745 - loss: 1141.6052 3/Unknown 1s 195ms/step - accuracy: 0.6388 - loss: 1084.4358 4/Unknown 1s 199ms/step - accuracy: 0.6822 - loss: 1031.0354 5/Unknown 2s 201ms/step - accuracy: 0.7131 - loss: 986.4984 6/Unknown 2s 197ms/step - accuracy: 0.7363 - loss: 947.2644 7/Unknown 2s 190ms/step - accuracy: 0.7546 - loss: 912.4213 8/Unknown 2s 188ms/step - accuracy: 0.7698 - loss: 881.4526 9/Unknown 2s 186ms/step - accuracy: 0.7824 - loss: 853.8523 10/Unknown 2s 184ms/step - accuracy: 0.7932 - loss: 829.0496 11/Unknown 3s 183ms/step - accuracy: 0.8022 - loss: 807.4752 12/Unknown 3s 184ms/step - accuracy: 0.8100 - loss: 788.1222 13/Unknown 3s 187ms/step - accuracy: 0.8170 - loss: 770.3723 14/Unknown 3s 187ms/step - accuracy: 0.8233 - loss: 753.6734 15/Unknown 3s 186ms/step - accuracy: 0.8289 - loss: 737.9523 16/Unknown 3s 186ms/step - accuracy: 0.8342 - loss: 723.0760 17/Unknown 4s 186ms/step - accuracy: 0.8389 - loss: 709.2202 18/Unknown 4s 202ms/step - accuracy: 0.8432 - loss: 696.8585 19/Unknown 4s 200ms/step - accuracy: 0.8470 - loss: 685.7762 20/Unknown 4s 198ms/step - accuracy: 0.8505 - loss: 675.3044 21/Unknown 5s 197ms/step - accuracy: 0.8537 - loss: 665.8409 22/Unknown 5s 196ms/step - accuracy: 0.8566 - loss: 657.3629 23/Unknown 5s 195ms/step - accuracy: 0.8593 - loss: 649.5444 24/Unknown 5s 195ms/step - accuracy: 0.8618 - loss: 642.1780 25/Unknown 5s 194ms/step - accuracy: 0.8641 - loss: 635.1900 26/Unknown 6s 195ms/step - accuracy: 0.8662 - loss: 628.5919 27/Unknown 6s 195ms/step - accuracy: 0.8683 - loss: 622.2363 28/Unknown 6s 195ms/step - accuracy: 0.8702 - loss: 616.1565 29/Unknown 6s 194ms/step - accuracy: 0.8720 - loss: 610.3881 30/Unknown 6s 194ms/step - accuracy: 0.8737 - loss: 604.7990 31/Unknown 6s 193ms/step - accuracy: 0.8753 - loss: 599.5613 32/Unknown 7s 194ms/step - accuracy: 0.8769 - loss: 594.4847 33/Unknown 7s 194ms/step - accuracy: 0.8783 - loss: 589.5745 34/Unknown 7s 194ms/step - accuracy: 0.8797 - loss: 584.9431 35/Unknown 7s 194ms/step - accuracy: 0.8810 - loss: 580.5197 36/Unknown 7s 193ms/step - accuracy: 0.8822 - loss: 576.2609 37/Unknown 8s 193ms/step - accuracy: 0.8834 - loss: 572.0708 38/Unknown 8s 194ms/step - accuracy: 0.8845 - loss: 567.9126 39/Unknown 8s 194ms/step - accuracy: 0.8856 - loss: 563.8269 40/Unknown 8s 194ms/step - accuracy: 0.8867 - loss: 559.9911 41/Unknown 8s 194ms/step - accuracy: 0.8877 - loss: 556.2637 42/Unknown 9s 193ms/step - accuracy: 0.8886 - loss: 552.6080 43/Unknown 9s 193ms/step - accuracy: 0.8896 - loss: 549.0726 44/Unknown 9s 193ms/step - accuracy: 0.8905 - loss: 545.6210 45/Unknown 9s 193ms/step - accuracy: 0.8913 - loss: 542.2662 46/Unknown 9s 193ms/step - accuracy: 0.8921 - loss: 539.0649 47/Unknown 10s 193ms/step - accuracy: 0.8929 - loss: 535.9783 48/Unknown 10s 193ms/step - accuracy: 0.8936 - loss: 532.9994 49/Unknown 10s 193ms/step - accuracy: 0.8944 - loss: 530.0856 50/Unknown 10s 193ms/step - accuracy: 0.8951 - loss: 527.2556 51/Unknown 10s 194ms/step - accuracy: 0.8957 - loss: 524.4853 52/Unknown 11s 194ms/step - accuracy: 0.8964 - loss: 521.8221 53/Unknown 11s 194ms/step - accuracy: 0.8970 - loss: 519.2384 54/Unknown 11s 194ms/step - accuracy: 0.8976 - loss: 516.6887 55/Unknown 11s 195ms/step - accuracy: 0.8982 - loss: 514.2283 56/Unknown 11s 195ms/step - accuracy: 0.8987 - loss: 511.8073 57/Unknown 12s 195ms/step - accuracy: 0.8993 - loss: 509.4113 58/Unknown 12s 194ms/step - accuracy: 0.8998 - loss: 507.0705 59/Unknown 12s 194ms/step - accuracy: 0.9004 - loss: 504.7713 60/Unknown 12s 194ms/step - accuracy: 0.9009 - loss: 502.5121 61/Unknown 12s 195ms/step - accuracy: 0.9014 - loss: 500.2973 62/Unknown 13s 195ms/step - accuracy: 0.9019 - loss: 498.1272 63/Unknown 13s 196ms/step - accuracy: 0.9023 - loss: 496.0018 64/Unknown 13s 196ms/step - accuracy: 0.9028 - loss: 493.9293 65/Unknown 13s 197ms/step - accuracy: 0.9032 - loss: 491.9118 66/Unknown 14s 197ms/step - accuracy: 0.9037 - loss: 489.9484 67/Unknown 14s 197ms/step - accuracy: 0.9041 - loss: 488.0164 68/Unknown 14s 197ms/step - accuracy: 0.9045 - loss: 486.1193 69/Unknown 14s 197ms/step - accuracy: 0.9049 - loss: 484.2630 70/Unknown 14s 197ms/step - accuracy: 0.9053 - loss: 482.4265 71/Unknown 14s 197ms/step - accuracy: 0.9057 - loss: 480.6362 72/Unknown 15s 197ms/step - accuracy: 0.9061 - loss: 478.8780 73/Unknown 15s 197ms/step - accuracy: 0.9064 - loss: 477.1625 74/Unknown 15s 197ms/step - accuracy: 0.9068 - loss: 475.4860 75/Unknown 15s 198ms/step - accuracy: 0.9071 - loss: 473.8222 76/Unknown 15s 197ms/step - accuracy: 0.9075 - loss: 472.2155 77/Unknown 16s 197ms/step - accuracy: 0.9078 - loss: 470.6271 78/Unknown 16s 196ms/step - accuracy: 0.9081 - loss: 469.0505 79/Unknown 16s 196ms/step - accuracy: 0.9084 - loss: 467.4939 80/Unknown 16s 196ms/step - accuracy: 0.9087 - loss: 465.9711 81/Unknown 16s 196ms/step - accuracy: 0.9090 - loss: 464.4900 82/Unknown 17s 196ms/step - accuracy: 0.9093 - loss: 463.0288 83/Unknown 17s 196ms/step - accuracy: 0.9096 - loss: 461.5836 84/Unknown 17s 196ms/step - accuracy: 0.9099 - loss: 460.1690 85/Unknown 17s 196ms/step - accuracy: 0.9101 - loss: 458.7745 86/Unknown 17s 196ms/step - accuracy: 0.9104 - loss: 457.3958 87/Unknown 18s 196ms/step - accuracy: 0.9107 - loss: 456.0403 88/Unknown 18s 196ms/step - accuracy: 0.9109 - loss: 454.6969 89/Unknown 18s 195ms/step - accuracy: 0.9112 - loss: 453.3802 90/Unknown 18s 195ms/step - accuracy: 0.9115 - loss: 452.0960 91/Unknown 18s 195ms/step - accuracy: 0.9117 - loss: 450.8279 92/Unknown 18s 195ms/step - accuracy: 0.9119 - loss: 449.5715 93/Unknown 19s 195ms/step - accuracy: 0.9122 - loss: 448.3546 94/Unknown 19s 195ms/step - accuracy: 0.9124 - loss: 447.1467 95/Unknown 19s 195ms/step - accuracy: 0.9126 - loss: 445.9574 96/Unknown 19s 195ms/step - accuracy: 0.9129 - loss: 444.7892 97/Unknown 19s 195ms/step - accuracy: 0.9131 - loss: 443.6310 98/Unknown 20s 195ms/step - accuracy: 0.9133 - loss: 442.5005 99/Unknown 20s 195ms/step - accuracy: 0.9135 - loss: 441.3832 100/Unknown 20s 195ms/step - accuracy: 0.9137 - loss: 440.2770 101/Unknown 20s 194ms/step - accuracy: 0.9139 - loss: 439.1806 102/Unknown 20s 194ms/step - accuracy: 0.9141 - loss: 438.0936 103/Unknown 21s 194ms/step - accuracy: 0.9144 - loss: 437.0181 104/Unknown 21s 195ms/step - accuracy: 0.9146 - loss: 435.9646 105/Unknown 21s 195ms/step - accuracy: 0.9148 - loss: 434.9277 106/Unknown 21s 195ms/step - accuracy: 0.9150 - loss: 433.8949 107/Unknown 21s 195ms/step - accuracy: 0.9151 - loss: 432.8877 108/Unknown 22s 195ms/step - accuracy: 0.9153 - loss: 431.8965 109/Unknown 22s 195ms/step - accuracy: 0.9155 - loss: 430.9133 110/Unknown 22s 196ms/step - accuracy: 0.9157 - loss: 429.9397 111/Unknown 22s 196ms/step - accuracy: 0.9159 - loss: 428.9818 112/Unknown 22s 196ms/step - accuracy: 0.9161 - loss: 428.0353 113/Unknown 23s 197ms/step - accuracy: 0.9162 - loss: 427.0999 114/Unknown 23s 197ms/step - accuracy: 0.9164 - loss: 426.1697 115/Unknown 23s 197ms/step - accuracy: 0.9166 - loss: 425.2458 116/Unknown 23s 197ms/step - accuracy: 0.9168 - loss: 424.3345 117/Unknown 24s 197ms/step - accuracy: 0.9169 - loss: 423.4386 118/Unknown 24s 197ms/step - accuracy: 0.9171 - loss: 422.5567 119/Unknown 24s 197ms/step - accuracy: 0.9173 - loss: 421.6823 120/Unknown 24s 197ms/step - accuracy: 0.9174 - loss: 420.8182 121/Unknown 24s 197ms/step - accuracy: 0.9176 - loss: 419.9664 122/Unknown 25s 197ms/step - accuracy: 0.9177 - loss: 419.1238 123/Unknown 25s 197ms/step - accuracy: 0.9179 - loss: 418.2940 124/Unknown 25s 197ms/step - accuracy: 0.9181 - loss: 417.4785 125/Unknown 25s 197ms/step - accuracy: 0.9182 - loss: 416.6722 126/Unknown 25s 197ms/step - accuracy: 0.9183 - loss: 415.8714 127/Unknown 26s 197ms/step - accuracy: 0.9185 - loss: 415.0771 128/Unknown 26s 197ms/step - accuracy: 0.9186 - loss: 414.2919 129/Unknown 26s 197ms/step - accuracy: 0.9188 - loss: 413.5163 130/Unknown 26s 197ms/step - accuracy: 0.9189 - loss: 412.7452 131/Unknown 26s 198ms/step - accuracy: 0.9191 - loss: 411.9837 132/Unknown 27s 197ms/step - accuracy: 0.9192 - loss: 411.2362 133/Unknown 27s 198ms/step - accuracy: 0.9193 - loss: 410.4987 134/Unknown 27s 198ms/step - accuracy: 0.9195 - loss: 409.7713 135/Unknown 27s 198ms/step - accuracy: 0.9196 - loss: 409.0497 136/Unknown 27s 198ms/step - accuracy: 0.9197 - loss: 408.3427 137/Unknown 28s 198ms/step - accuracy: 0.9199 - loss: 407.6457 138/Unknown 28s 198ms/step - accuracy: 0.9200 - loss: 406.9554 139/Unknown 28s 198ms/step - accuracy: 0.9201 - loss: 406.2739 140/Unknown 28s 198ms/step - accuracy: 0.9202 - loss: 405.6025 141/Unknown 28s 198ms/step - accuracy: 0.9204 - loss: 404.9378 142/Unknown 29s 198ms/step - accuracy: 0.9205 - loss: 404.2800 143/Unknown 29s 198ms/step - accuracy: 0.9206 - loss: 403.6313 144/Unknown 29s 198ms/step - accuracy: 0.9207 - loss: 402.9886 145/Unknown 29s 198ms/step - accuracy: 0.9208 - loss: 402.3509 146/Unknown 29s 198ms/step - accuracy: 0.9209 - loss: 401.7167 147/Unknown 30s 199ms/step - accuracy: 0.9211 - loss: 401.0935 148/Unknown 30s 199ms/step - accuracy: 0.9212 - loss: 400.4745 149/Unknown 30s 200ms/step - accuracy: 0.9213 - loss: 399.8633 150/Unknown 30s 200ms/step - accuracy: 0.9214 - loss: 399.2570 151/Unknown 31s 200ms/step - accuracy: 0.9215 - loss: 398.6591 152/Unknown 31s 200ms/step - accuracy: 0.9216 - loss: 398.0708 153/Unknown 31s 201ms/step - accuracy: 0.9217 - loss: 397.4871 154/Unknown 31s 201ms/step - accuracy: 0.9218 - loss: 396.9082 155/Unknown 32s 201ms/step - accuracy: 0.9219 - loss: 396.3336 156/Unknown 32s 201ms/step - accuracy: 0.9220 - loss: 395.7641 157/Unknown 32s 202ms/step - accuracy: 0.9222 - loss: 395.1993 158/Unknown 32s 202ms/step - accuracy: 0.9223 - loss: 394.6414 159/Unknown 33s 202ms/step - accuracy: 0.9224 - loss: 394.0871 160/Unknown 33s 202ms/step - accuracy: 0.9225 - loss: 393.5349 161/Unknown 33s 202ms/step - accuracy: 0.9226 - loss: 392.9877 162/Unknown 33s 202ms/step - accuracy: 0.9227 - loss: 392.4430 163/Unknown 33s 202ms/step - accuracy: 0.9228 - loss: 391.9062 164/Unknown 34s 202ms/step - accuracy: 0.9229 - loss: 391.3741 165/Unknown 34s 202ms/step - accuracy: 0.9230 - loss: 390.8450 166/Unknown 34s 202ms/step - accuracy: 0.9231 - loss: 390.3184 167/Unknown 34s 202ms/step - accuracy: 0.9232 - loss: 389.7947 168/Unknown 34s 202ms/step - accuracy: 0.9233 - loss: 389.2769 169/Unknown 35s 201ms/step - accuracy: 0.9234 - loss: 388.7642 170/Unknown 35s 201ms/step - accuracy: 0.9234 - loss: 388.2601 171/Unknown 35s 202ms/step - accuracy: 0.9235 - loss: 387.7563 172/Unknown 35s 202ms/step - accuracy: 0.9236 - loss: 387.2565 173/Unknown 35s 202ms/step - accuracy: 0.9237 - loss: 386.7589 174/Unknown 36s 202ms/step - accuracy: 0.9238 - loss: 386.2676 175/Unknown 36s 202ms/step - accuracy: 0.9239 - loss: 385.7783 176/Unknown 36s 202ms/step - accuracy: 0.9240 - loss: 385.2943 177/Unknown 36s 202ms/step - accuracy: 0.9241 - loss: 384.8183 178/Unknown 36s 202ms/step - accuracy: 0.9242 - loss: 384.3459 179/Unknown 37s 202ms/step - accuracy: 0.9243 - loss: 383.8763 180/Unknown 37s 202ms/step - accuracy: 0.9243 - loss: 383.4090 181/Unknown 37s 202ms/step - accuracy: 0.9244 - loss: 382.9447 182/Unknown 37s 202ms/step - accuracy: 0.9245 - loss: 382.4837 183/Unknown 37s 202ms/step - accuracy: 0.9246 - loss: 382.0276 184/Unknown 38s 202ms/step - accuracy: 0.9247 - loss: 381.5741 185/Unknown 38s 202ms/step - accuracy: 0.9248 - loss: 381.1227 186/Unknown 38s 203ms/step - accuracy: 0.9248 - loss: 380.6739 187/Unknown 38s 203ms/step - accuracy: 0.9249 - loss: 380.2263 188/Unknown 39s 203ms/step - accuracy: 0.9250 - loss: 379.7803 189/Unknown 39s 203ms/step - accuracy: 0.9251 - loss: 379.3362 190/Unknown 39s 203ms/step - accuracy: 0.9252 - loss: 378.8958 191/Unknown 39s 203ms/step - accuracy: 0.9252 - loss: 378.4580 192/Unknown 40s 203ms/step - accuracy: 0.9253 - loss: 378.0284 193/Unknown 40s 203ms/step - accuracy: 0.9254 - loss: 377.6024 194/Unknown 40s 203ms/step - accuracy: 0.9255 - loss: 377.1786 195/Unknown 40s 203ms/step - accuracy: 0.9256 - loss: 376.7578 196/Unknown 40s 203ms/step - accuracy: 0.9256 - loss: 376.3396 197/Unknown 41s 203ms/step - accuracy: 0.9257 - loss: 375.9260 198/Unknown 41s 203ms/step - accuracy: 0.9258 - loss: 375.5150 199/Unknown 41s 203ms/step - accuracy: 0.9259 - loss: 375.1063 200/Unknown 41s 203ms/step - accuracy: 0.9259 - loss: 374.6986 201/Unknown 41s 203ms/step - accuracy: 0.9260 - loss: 374.2932 202/Unknown 42s 203ms/step - accuracy: 0.9261 - loss: 373.8907 203/Unknown 42s 203ms/step - accuracy: 0.9262 - loss: 373.4908 204/Unknown 42s 203ms/step - accuracy: 0.9262 - loss: 373.0950 205/Unknown 42s 203ms/step - accuracy: 0.9263 - loss: 372.7011 206/Unknown 42s 203ms/step - accuracy: 0.9264 - loss: 372.3101 207/Unknown 43s 203ms/step - accuracy: 0.9264 - loss: 371.9205 208/Unknown 43s 203ms/step - accuracy: 0.9265 - loss: 371.5334 209/Unknown 43s 203ms/step - accuracy: 0.9266 - loss: 371.1480 210/Unknown 43s 203ms/step - accuracy: 0.9266 - loss: 370.7651 211/Unknown 43s 203ms/step - accuracy: 0.9267 - loss: 370.3852 212/Unknown 44s 203ms/step - accuracy: 0.9268 - loss: 370.0093 213/Unknown 44s 203ms/step - accuracy: 0.9268 - loss: 369.6351 214/Unknown 44s 204ms/step - accuracy: 0.9269 - loss: 369.2646 215/Unknown 44s 204ms/step - accuracy: 0.9270 - loss: 368.8969 216/Unknown 45s 204ms/step - accuracy: 0.9270 - loss: 368.5307 217/Unknown 45s 204ms/step - accuracy: 0.9271 - loss: 368.1656 218/Unknown 45s 204ms/step - accuracy: 0.9272 - loss: 367.8022 219/Unknown 45s 204ms/step - accuracy: 0.9272 - loss: 367.4400 220/Unknown 45s 204ms/step - accuracy: 0.9273 - loss: 367.0780 221/Unknown 46s 204ms/step - accuracy: 0.9274 - loss: 366.7194 222/Unknown 46s 204ms/step - accuracy: 0.9274 - loss: 366.3614 223/Unknown 46s 204ms/step - accuracy: 0.9275 - loss: 366.0057 224/Unknown 46s 205ms/step - accuracy: 0.9275 - loss: 365.6517 225/Unknown 47s 205ms/step - accuracy: 0.9276 - loss: 365.2998 226/Unknown 47s 205ms/step - accuracy: 0.9277 - loss: 364.9511 227/Unknown 47s 205ms/step - accuracy: 0.9277 - loss: 364.6082 228/Unknown 47s 205ms/step - accuracy: 0.9278 - loss: 364.2671 229/Unknown 47s 205ms/step - accuracy: 0.9278 - loss: 363.9267 230/Unknown 48s 205ms/step - accuracy: 0.9279 - loss: 363.5866 231/Unknown 48s 205ms/step - accuracy: 0.9280 - loss: 363.2502 232/Unknown 48s 205ms/step - accuracy: 0.9280 - loss: 362.9175 233/Unknown 48s 205ms/step - accuracy: 0.9281 - loss: 362.5866 234/Unknown 49s 205ms/step - accuracy: 0.9281 - loss: 362.2563 235/Unknown 49s 205ms/step - accuracy: 0.9282 - loss: 361.9282 236/Unknown 49s 206ms/step - accuracy: 0.9283 - loss: 361.6011 237/Unknown 49s 206ms/step - accuracy: 0.9283 - loss: 361.2760 238/Unknown 50s 207ms/step - accuracy: 0.9284 - loss: 360.9536 239/Unknown 50s 207ms/step - accuracy: 0.9284 - loss: 360.6325 240/Unknown 50s 207ms/step - accuracy: 0.9285 - loss: 360.3125 241/Unknown 51s 208ms/step - accuracy: 0.9285 - loss: 359.9926 242/Unknown 51s 208ms/step - accuracy: 0.9286 - loss: 359.6747 243/Unknown 51s 208ms/step - accuracy: 0.9287 - loss: 359.3593 244/Unknown 51s 209ms/step - accuracy: 0.9287 - loss: 359.0458 245/Unknown 52s 209ms/step - accuracy: 0.9288 - loss: 358.7349 246/Unknown 52s 209ms/step - accuracy: 0.9288 - loss: 358.4268 247/Unknown 52s 209ms/step - accuracy: 0.9289 - loss: 358.1215 248/Unknown 52s 209ms/step - accuracy: 0.9289 - loss: 357.8191 249/Unknown 53s 209ms/step - accuracy: 0.9290 - loss: 357.5194 250/Unknown 53s 209ms/step - accuracy: 0.9290 - loss: 357.2212 251/Unknown 53s 209ms/step - accuracy: 0.9291 - loss: 356.9238 252/Unknown 53s 209ms/step - accuracy: 0.9291 - loss: 356.6286 253/Unknown 53s 209ms/step - accuracy: 0.9292 - loss: 356.3355 254/Unknown 54s 210ms/step - accuracy: 0.9292 - loss: 356.0449 255/Unknown 54s 210ms/step - accuracy: 0.9293 - loss: 355.7557 256/Unknown 54s 210ms/step - accuracy: 0.9293 - loss: 355.4673 257/Unknown 54s 210ms/step - accuracy: 0.9294 - loss: 355.1799 258/Unknown 55s 210ms/step - accuracy: 0.9294 - loss: 354.8956 259/Unknown 55s 210ms/step - accuracy: 0.9295 - loss: 354.6136 260/Unknown 55s 210ms/step - accuracy: 0.9295 - loss: 354.3326 261/Unknown 55s 210ms/step - accuracy: 0.9296 - loss: 354.0539 262/Unknown 56s 210ms/step - accuracy: 0.9296 - loss: 353.7773 263/Unknown 56s 210ms/step - accuracy: 0.9297 - loss: 353.5032 264/Unknown 56s 211ms/step - accuracy: 0.9297 - loss: 353.2297 265/Unknown 56s 211ms/step - accuracy: 0.9298 - loss: 352.9581 266/Unknown 57s 211ms/step - accuracy: 0.9298 - loss: 352.6891 267/Unknown 57s 211ms/step - accuracy: 0.9299 - loss: 352.4228 268/Unknown 57s 211ms/step - accuracy: 0.9299 - loss: 352.1577 269/Unknown 57s 212ms/step - accuracy: 0.9300 - loss: 351.8944 270/Unknown 58s 212ms/step - accuracy: 0.9300 - loss: 351.6331 271/Unknown 58s 212ms/step - accuracy: 0.9300 - loss: 351.3750 272/Unknown 58s 212ms/step - accuracy: 0.9301 - loss: 351.1187 273/Unknown 59s 213ms/step - accuracy: 0.9301 - loss: 350.8644 274/Unknown 59s 213ms/step - accuracy: 0.9302 - loss: 350.6121 275/Unknown 59s 213ms/step - accuracy: 0.9302 - loss: 350.3619 276/Unknown 59s 213ms/step - accuracy: 0.9303 - loss: 350.1118 277/Unknown 60s 213ms/step - accuracy: 0.9303 - loss: 349.8630 278/Unknown 60s 213ms/step - accuracy: 0.9303 - loss: 349.6153 279/Unknown 60s 213ms/step - accuracy: 0.9304 - loss: 349.3688 280/Unknown 60s 213ms/step - accuracy: 0.9304 - loss: 349.1229 281/Unknown 60s 213ms/step - accuracy: 0.9305 - loss: 348.8784 282/Unknown 61s 213ms/step - accuracy: 0.9305 - loss: 348.6351 283/Unknown 61s 213ms/step - accuracy: 0.9306 - loss: 348.3920 284/Unknown 61s 213ms/step - accuracy: 0.9306 - loss: 348.1502 285/Unknown 61s 214ms/step - accuracy: 0.9306 - loss: 347.9100 286/Unknown 62s 214ms/step - accuracy: 0.9307 - loss: 347.6703 287/Unknown 62s 214ms/step - accuracy: 0.9307 - loss: 347.4321 288/Unknown 62s 214ms/step - accuracy: 0.9308 - loss: 347.1949 289/Unknown 62s 214ms/step - accuracy: 0.9308 - loss: 346.9582 290/Unknown 63s 214ms/step - accuracy: 0.9308 - loss: 346.7231 291/Unknown 63s 214ms/step - accuracy: 0.9309 - loss: 346.4895 292/Unknown 63s 214ms/step - accuracy: 0.9309 - loss: 346.2580 293/Unknown 63s 214ms/step - accuracy: 0.9310 - loss: 346.0265 294/Unknown 63s 214ms/step - accuracy: 0.9310 - loss: 345.7953 295/Unknown 64s 214ms/step - accuracy: 0.9310 - loss: 345.5652 296/Unknown 64s 214ms/step - accuracy: 0.9311 - loss: 345.3354 297/Unknown 64s 214ms/step - accuracy: 0.9311 - loss: 345.1051 298/Unknown 64s 214ms/step - accuracy: 0.9312 - loss: 344.8757 299/Unknown 65s 214ms/step - accuracy: 0.9312 - loss: 344.6488 300/Unknown 65s 215ms/step - accuracy: 0.9312 - loss: 344.4228 301/Unknown 65s 215ms/step - accuracy: 0.9313 - loss: 344.1973 302/Unknown 65s 215ms/step - accuracy: 0.9313 - loss: 343.9730 303/Unknown 66s 215ms/step - accuracy: 0.9314 - loss: 343.7490 304/Unknown 66s 215ms/step - accuracy: 0.9314 - loss: 343.5257 305/Unknown 66s 215ms/step - accuracy: 0.9314 - loss: 343.3047 306/Unknown 66s 215ms/step - accuracy: 0.9315 - loss: 343.0848 307/Unknown 66s 215ms/step - accuracy: 0.9315 - loss: 342.8660 308/Unknown 67s 215ms/step - accuracy: 0.9315 - loss: 342.6490 309/Unknown 67s 215ms/step - accuracy: 0.9316 - loss: 342.4344 310/Unknown 67s 215ms/step - accuracy: 0.9316 - loss: 342.2202 311/Unknown 67s 215ms/step - accuracy: 0.9316 - loss: 342.0071 312/Unknown 67s 215ms/step - accuracy: 0.9317 - loss: 341.7946 313/Unknown 68s 215ms/step - accuracy: 0.9317 - loss: 341.5840 314/Unknown 68s 215ms/step - accuracy: 0.9318 - loss: 341.3745 315/Unknown 68s 215ms/step - accuracy: 0.9318 - loss: 341.1663 316/Unknown 68s 215ms/step - accuracy: 0.9318 - loss: 340.9592 317/Unknown 69s 215ms/step - accuracy: 0.9319 - loss: 340.7528 318/Unknown 69s 215ms/step - accuracy: 0.9319 - loss: 340.5479 319/Unknown 69s 215ms/step - accuracy: 0.9319 - loss: 340.3434 320/Unknown 69s 215ms/step - accuracy: 0.9320 - loss: 340.1393 321/Unknown 69s 215ms/step - accuracy: 0.9320 - loss: 339.9351 322/Unknown 70s 215ms/step - accuracy: 0.9320 - loss: 339.7322 323/Unknown 70s 214ms/step - accuracy: 0.9321 - loss: 339.5305 324/Unknown 70s 214ms/step - accuracy: 0.9321 - loss: 339.3291 325/Unknown 70s 214ms/step - accuracy: 0.9321 - loss: 339.1288 326/Unknown 70s 214ms/step - accuracy: 0.9322 - loss: 338.9296 327/Unknown 71s 214ms/step - accuracy: 0.9322 - loss: 338.7308 328/Unknown 71s 214ms/step - accuracy: 0.9322 - loss: 338.5333 329/Unknown 71s 214ms/step - accuracy: 0.9323 - loss: 338.3367 330/Unknown 71s 214ms/step - accuracy: 0.9323 - loss: 338.1408 331/Unknown 71s 214ms/step - accuracy: 0.9323 - loss: 337.9453 332/Unknown 71s 214ms/step - accuracy: 0.9324 - loss: 337.7515 333/Unknown 72s 214ms/step - accuracy: 0.9324 - loss: 337.5584 334/Unknown 72s 214ms/step - accuracy: 0.9324 - loss: 337.3662 335/Unknown 72s 214ms/step - accuracy: 0.9325 - loss: 337.1747 336/Unknown 72s 214ms/step - accuracy: 0.9325 - loss: 336.9847 337/Unknown 72s 214ms/step - accuracy: 0.9325 - loss: 336.7955 338/Unknown 73s 214ms/step - accuracy: 0.9326 - loss: 336.6066 339/Unknown 73s 214ms/step - accuracy: 0.9326 - loss: 336.4176 340/Unknown 73s 214ms/step - accuracy: 0.9326 - loss: 336.2293 341/Unknown 73s 214ms/step - accuracy: 0.9327 - loss: 336.0417 342/Unknown 74s 214ms/step - accuracy: 0.9327 - loss: 335.8549 343/Unknown 74s 213ms/step - accuracy: 0.9327 - loss: 335.6696 344/Unknown 74s 213ms/step - accuracy: 0.9328 - loss: 335.4847 345/Unknown 74s 213ms/step - accuracy: 0.9328 - loss: 335.3005 346/Unknown 74s 214ms/step - accuracy: 0.9328 - loss: 335.1162 347/Unknown 75s 214ms/step - accuracy: 0.9328 - loss: 334.9317 348/Unknown 75s 214ms/step - accuracy: 0.9329 - loss: 334.7472 349/Unknown 75s 214ms/step - accuracy: 0.9329 - loss: 334.5628 350/Unknown 75s 214ms/step - accuracy: 0.9329 - loss: 334.3791 351/Unknown 75s 214ms/step - accuracy: 0.9330 - loss: 334.1966 352/Unknown 76s 214ms/step - accuracy: 0.9330 - loss: 334.0144 353/Unknown 76s 214ms/step - accuracy: 0.9330 - loss: 333.8329 354/Unknown 76s 214ms/step - accuracy: 0.9331 - loss: 333.6525 355/Unknown 76s 214ms/step - accuracy: 0.9331 - loss: 333.4737 356/Unknown 77s 214ms/step - accuracy: 0.9331 - loss: 333.2956 357/Unknown 77s 214ms/step - accuracy: 0.9332 - loss: 333.1186 358/Unknown 77s 214ms/step - accuracy: 0.9332 - loss: 332.9418 359/Unknown 77s 214ms/step - accuracy: 0.9332 - loss: 332.7658 360/Unknown 77s 214ms/step - accuracy: 0.9332 - loss: 332.5904 361/Unknown 78s 214ms/step - accuracy: 0.9333 - loss: 332.4163 362/Unknown 78s 214ms/step - accuracy: 0.9333 - loss: 332.2424 363/Unknown 78s 214ms/step - accuracy: 0.9333 - loss: 332.0692 364/Unknown 78s 214ms/step - accuracy: 0.9334 - loss: 331.8963 365/Unknown 79s 214ms/step - accuracy: 0.9334 - loss: 331.7242 366/Unknown 79s 214ms/step - accuracy: 0.9334 - loss: 331.5526 367/Unknown 79s 214ms/step - accuracy: 0.9334 - loss: 331.3815 368/Unknown 79s 214ms/step - accuracy: 0.9335 - loss: 331.2112 369/Unknown 79s 214ms/step - accuracy: 0.9335 - loss: 331.0415 370/Unknown 80s 214ms/step - accuracy: 0.9335 - loss: 330.8718 371/Unknown 80s 214ms/step - accuracy: 0.9336 - loss: 330.7039 372/Unknown 80s 214ms/step - accuracy: 0.9336 - loss: 330.5366 373/Unknown 80s 214ms/step - accuracy: 0.9336 - loss: 330.3701 374/Unknown 81s 214ms/step - accuracy: 0.9336 - loss: 330.2039 375/Unknown 81s 214ms/step - accuracy: 0.9337 - loss: 330.0381 376/Unknown 81s 214ms/step - accuracy: 0.9337 - loss: 329.8729 377/Unknown 81s 214ms/step - accuracy: 0.9337 - loss: 329.7084 378/Unknown 81s 214ms/step - accuracy: 0.9338 - loss: 329.5450 379/Unknown 82s 214ms/step - accuracy: 0.9338 - loss: 329.3824 380/Unknown 82s 214ms/step - accuracy: 0.9338 - loss: 329.2206 381/Unknown 82s 214ms/step - accuracy: 0.9338 - loss: 329.0592 382/Unknown 82s 214ms/step - accuracy: 0.9339 - loss: 328.8988 383/Unknown 83s 214ms/step - accuracy: 0.9339 - loss: 328.7392 384/Unknown 83s 214ms/step - accuracy: 0.9339 - loss: 328.5800 385/Unknown 83s 214ms/step - accuracy: 0.9339 - loss: 328.4217 386/Unknown 83s 214ms/step - accuracy: 0.9340 - loss: 328.2640 387/Unknown 84s 215ms/step - accuracy: 0.9340 - loss: 328.1079 388/Unknown 84s 215ms/step - accuracy: 0.9340 - loss: 327.9520 389/Unknown 84s 215ms/step - accuracy: 0.9340 - loss: 327.7965 390/Unknown 84s 215ms/step - accuracy: 0.9341 - loss: 327.6415 391/Unknown 84s 215ms/step - accuracy: 0.9341 - loss: 327.4871 392/Unknown 85s 215ms/step - accuracy: 0.9341 - loss: 327.3329 393/Unknown 85s 215ms/step - accuracy: 0.9341 - loss: 327.1793 394/Unknown 85s 215ms/step - accuracy: 0.9342 - loss: 327.0259 395/Unknown 85s 215ms/step - accuracy: 0.9342 - loss: 326.8732 396/Unknown 86s 215ms/step - accuracy: 0.9342 - loss: 326.7209 397/Unknown 86s 215ms/step - accuracy: 0.9342 - loss: 326.5693 398/Unknown 86s 215ms/step - accuracy: 0.9343 - loss: 326.4187 399/Unknown 86s 215ms/step - accuracy: 0.9343 - loss: 326.2686 400/Unknown 87s 216ms/step - accuracy: 0.9343 - loss: 326.1197 401/Unknown 87s 216ms/step - accuracy: 0.9343 - loss: 325.9718 402/Unknown 87s 216ms/step - accuracy: 0.9344 - loss: 325.8240 403/Unknown 88s 216ms/step - accuracy: 0.9344 - loss: 325.6769 404/Unknown 88s 216ms/step - accuracy: 0.9344 - loss: 325.5306 405/Unknown 88s 216ms/step - accuracy: 0.9344 - loss: 325.3849 406/Unknown 88s 216ms/step - accuracy: 0.9345 - loss: 325.2393 407/Unknown 88s 216ms/step - accuracy: 0.9345 - loss: 325.0950 408/Unknown 89s 216ms/step - accuracy: 0.9345 - loss: 324.9518 409/Unknown 89s 216ms/step - accuracy: 0.9345 - loss: 324.8096 410/Unknown 89s 216ms/step - accuracy: 0.9346 - loss: 324.6682 411/Unknown 89s 216ms/step - accuracy: 0.9346 - loss: 324.5275 412/Unknown 90s 216ms/step - accuracy: 0.9346 - loss: 324.3878 413/Unknown 90s 216ms/step - accuracy: 0.9346 - loss: 324.2490 414/Unknown 90s 216ms/step - accuracy: 0.9347 - loss: 324.1107 415/Unknown 90s 216ms/step - accuracy: 0.9347 - loss: 323.9733 416/Unknown 90s 216ms/step - accuracy: 0.9347 - loss: 323.8363 417/Unknown 91s 216ms/step - accuracy: 0.9347 - loss: 323.7005 418/Unknown 91s 216ms/step - accuracy: 0.9347 - loss: 323.5650 419/Unknown 91s 216ms/step - accuracy: 0.9348 - loss: 323.4300 420/Unknown 91s 216ms/step - accuracy: 0.9348 - loss: 323.2954 421/Unknown 92s 216ms/step - accuracy: 0.9348 - loss: 323.1616 422/Unknown 92s 216ms/step - accuracy: 0.9348 - loss: 323.0279 423/Unknown 92s 216ms/step - accuracy: 0.9349 - loss: 322.8943 424/Unknown 92s 217ms/step - accuracy: 0.9349 - loss: 322.7610 425/Unknown 93s 217ms/step - accuracy: 0.9349 - loss: 322.6279 426/Unknown 93s 217ms/step - accuracy: 0.9349 - loss: 322.4956 427/Unknown 93s 217ms/step - accuracy: 0.9349 - loss: 322.3640 428/Unknown 93s 217ms/step - accuracy: 0.9350 - loss: 322.2333 429/Unknown 93s 217ms/step - accuracy: 0.9350 - loss: 322.1030 430/Unknown 94s 217ms/step - accuracy: 0.9350 - loss: 321.9735 431/Unknown 94s 217ms/step - accuracy: 0.9350 - loss: 321.8445 432/Unknown 94s 217ms/step - accuracy: 0.9350 - loss: 321.7156 433/Unknown 95s 217ms/step - accuracy: 0.9351 - loss: 321.5872 434/Unknown 95s 217ms/step - accuracy: 0.9351 - loss: 321.4588 435/Unknown 95s 217ms/step - accuracy: 0.9351 - loss: 321.3307 436/Unknown 95s 217ms/step - accuracy: 0.9351 - loss: 321.2033 437/Unknown 95s 217ms/step - accuracy: 0.9352 - loss: 321.0760 438/Unknown 96s 217ms/step - accuracy: 0.9352 - loss: 320.9493 439/Unknown 96s 217ms/step - accuracy: 0.9352 - loss: 320.8226 440/Unknown 96s 217ms/step - accuracy: 0.9352 - loss: 320.6959 441/Unknown 96s 217ms/step - accuracy: 0.9352 - loss: 320.5698 442/Unknown 97s 218ms/step - accuracy: 0.9353 - loss: 320.4443 443/Unknown 97s 218ms/step - accuracy: 0.9353 - loss: 320.3189 444/Unknown 97s 218ms/step - accuracy: 0.9353 - loss: 320.1938 445/Unknown 97s 218ms/step - accuracy: 0.9353 - loss: 320.0695 446/Unknown 98s 218ms/step - accuracy: 0.9353 - loss: 319.9457 447/Unknown 98s 218ms/step - accuracy: 0.9354 - loss: 319.8219 448/Unknown 98s 218ms/step - accuracy: 0.9354 - loss: 319.6988 449/Unknown 98s 218ms/step - accuracy: 0.9354 - loss: 319.5759 450/Unknown 99s 218ms/step - accuracy: 0.9354 - loss: 319.4533 451/Unknown 99s 218ms/step - accuracy: 0.9354 - loss: 319.3317 452/Unknown 99s 218ms/step - accuracy: 0.9355 - loss: 319.2104 453/Unknown 99s 218ms/step - accuracy: 0.9355 - loss: 319.0900 454/Unknown 100s 218ms/step - accuracy: 0.9355 - loss: 318.9702 455/Unknown 100s 218ms/step - accuracy: 0.9355 - loss: 318.8509 456/Unknown 100s 218ms/step - accuracy: 0.9355 - loss: 318.7319 457/Unknown 100s 218ms/step - accuracy: 0.9356 - loss: 318.6129 458/Unknown 101s 219ms/step - accuracy: 0.9356 - loss: 318.4945 459/Unknown 101s 219ms/step - accuracy: 0.9356 - loss: 318.3764 460/Unknown 101s 219ms/step - accuracy: 0.9356 - loss: 318.2592 461/Unknown 101s 219ms/step - accuracy: 0.9356 - loss: 318.1426 462/Unknown 102s 219ms/step - accuracy: 0.9356 - loss: 318.0262 463/Unknown 102s 219ms/step - accuracy: 0.9357 - loss: 317.9101 464/Unknown 102s 219ms/step - accuracy: 0.9357 - loss: 317.7942 465/Unknown 102s 219ms/step - accuracy: 0.9357 - loss: 317.6788 466/Unknown 102s 219ms/step - accuracy: 0.9357 - loss: 317.5635 467/Unknown 103s 219ms/step - accuracy: 0.9357 - loss: 317.4484 468/Unknown 103s 219ms/step - accuracy: 0.9358 - loss: 317.3335 469/Unknown 103s 219ms/step - accuracy: 0.9358 - loss: 317.2188 470/Unknown 103s 219ms/step - accuracy: 0.9358 - loss: 317.1040 471/Unknown 104s 219ms/step - accuracy: 0.9358 - loss: 316.9898 472/Unknown 104s 219ms/step - accuracy: 0.9358 - loss: 316.8763 473/Unknown 104s 219ms/step - accuracy: 0.9359 - loss: 316.7631 474/Unknown 104s 219ms/step - accuracy: 0.9359 - loss: 316.6501 475/Unknown 105s 219ms/step - accuracy: 0.9359 - loss: 316.5378 476/Unknown 105s 219ms/step - accuracy: 0.9359 - loss: 316.4257 477/Unknown 105s 220ms/step - accuracy: 0.9359 - loss: 316.3138 478/Unknown 105s 220ms/step - accuracy: 0.9359 - loss: 316.2022 479/Unknown 106s 220ms/step - accuracy: 0.9360 - loss: 316.0910 480/Unknown 106s 220ms/step - accuracy: 0.9360 - loss: 315.9806 481/Unknown 106s 220ms/step - accuracy: 0.9360 - loss: 315.8706 482/Unknown 106s 220ms/step - accuracy: 0.9360 - loss: 315.7607 483/Unknown 107s 220ms/step - accuracy: 0.9360 - loss: 315.6510 484/Unknown 107s 220ms/step - accuracy: 0.9361 - loss: 315.5415 485/Unknown 107s 220ms/step - accuracy: 0.9361 - loss: 315.4324 486/Unknown 107s 220ms/step - accuracy: 0.9361 - loss: 315.3241 487/Unknown 108s 220ms/step - accuracy: 0.9361 - loss: 315.2159 488/Unknown 108s 220ms/step - accuracy: 0.9361 - loss: 315.1081 489/Unknown 108s 220ms/step - accuracy: 0.9361 - loss: 315.0006 490/Unknown 108s 220ms/step - accuracy: 0.9362 - loss: 314.8934 491/Unknown 109s 220ms/step - accuracy: 0.9362 - loss: 314.7863 492/Unknown 109s 220ms/step - accuracy: 0.9362 - loss: 314.6796 493/Unknown 109s 220ms/step - accuracy: 0.9362 - loss: 314.5731 494/Unknown 109s 220ms/step - accuracy: 0.9362 - loss: 314.4671 495/Unknown 109s 220ms/step - accuracy: 0.9363 - loss: 314.3610 496/Unknown 110s 220ms/step - accuracy: 0.9363 - loss: 314.2551 497/Unknown 110s 220ms/step - accuracy: 0.9363 - loss: 314.1498 498/Unknown 110s 220ms/step - accuracy: 0.9363 - loss: 314.0451 499/Unknown 110s 220ms/step - accuracy: 0.9363 - loss: 313.9405 500/Unknown 111s 220ms/step - accuracy: 0.9363 - loss: 313.8364 501/Unknown 111s 220ms/step - accuracy: 0.9364 - loss: 313.7324 502/Unknown 111s 220ms/step - accuracy: 0.9364 - loss: 313.6290 503/Unknown 111s 220ms/step - accuracy: 0.9364 - loss: 313.5266 504/Unknown 111s 220ms/step - accuracy: 0.9364 - loss: 313.4240 505/Unknown 112s 220ms/step - accuracy: 0.9364 - loss: 313.3215 506/Unknown 112s 220ms/step - accuracy: 0.9364 - loss: 313.2194 507/Unknown 112s 220ms/step - accuracy: 0.9365 - loss: 313.1174 508/Unknown 112s 220ms/step - accuracy: 0.9365 - loss: 313.0163 509/Unknown 113s 220ms/step - accuracy: 0.9365 - loss: 312.9149 510/Unknown 113s 220ms/step - accuracy: 0.9365 - loss: 312.8137 511/Unknown 113s 220ms/step - accuracy: 0.9365 - loss: 312.7127 512/Unknown 113s 220ms/step - accuracy: 0.9365 - loss: 312.6117 513/Unknown 113s 220ms/step - accuracy: 0.9366 - loss: 312.5108 514/Unknown 114s 220ms/step - accuracy: 0.9366 - loss: 312.4105 515/Unknown 114s 220ms/step - accuracy: 0.9366 - loss: 312.3107 516/Unknown 114s 221ms/step - accuracy: 0.9366 - loss: 312.2112 517/Unknown 115s 221ms/step - accuracy: 0.9366 - loss: 312.1120 518/Unknown 115s 221ms/step - accuracy: 0.9366 - loss: 312.0136 519/Unknown 115s 221ms/step - accuracy: 0.9367 - loss: 311.9154 520/Unknown 115s 221ms/step - accuracy: 0.9367 - loss: 311.8175 521/Unknown 116s 221ms/step - accuracy: 0.9367 - loss: 311.7198 522/Unknown 116s 221ms/step - accuracy: 0.9367 - loss: 311.6223 523/Unknown 116s 221ms/step - accuracy: 0.9367 - loss: 311.5254 524/Unknown 116s 221ms/step - accuracy: 0.9367 - loss: 311.4289 525/Unknown 117s 221ms/step - accuracy: 0.9368 - loss: 311.3327 526/Unknown 117s 221ms/step - accuracy: 0.9368 - loss: 311.2370 527/Unknown 117s 221ms/step - accuracy: 0.9368 - loss: 311.1415 528/Unknown 117s 221ms/step - accuracy: 0.9368 - loss: 311.0460 529/Unknown 118s 221ms/step - accuracy: 0.9368 - loss: 310.9506 530/Unknown 118s 221ms/step - accuracy: 0.9368 - loss: 310.8556 531/Unknown 118s 221ms/step - accuracy: 0.9368 - loss: 310.7611 532/Unknown 118s 221ms/step - accuracy: 0.9369 - loss: 310.6666 533/Unknown 119s 222ms/step - accuracy: 0.9369 - loss: 310.5723 534/Unknown 119s 222ms/step - accuracy: 0.9369 - loss: 310.4783 535/Unknown 119s 222ms/step - accuracy: 0.9369 - loss: 310.3845 536/Unknown 119s 222ms/step - accuracy: 0.9369 - loss: 310.2907 537/Unknown 119s 222ms/step - accuracy: 0.9369 - loss: 310.1972 538/Unknown 120s 222ms/step - accuracy: 0.9370 - loss: 310.1041 539/Unknown 120s 222ms/step - accuracy: 0.9370 - loss: 310.0114 540/Unknown 120s 222ms/step - accuracy: 0.9370 - loss: 309.9187 541/Unknown 120s 222ms/step - accuracy: 0.9370 - loss: 309.8263 542/Unknown 121s 221ms/step - accuracy: 0.9370 - loss: 309.7338 543/Unknown 121s 221ms/step - accuracy: 0.9370 - loss: 309.6414 544/Unknown 121s 222ms/step - accuracy: 0.9370 - loss: 309.5492 545/Unknown 121s 222ms/step - accuracy: 0.9371 - loss: 309.4574 546/Unknown 121s 222ms/step - accuracy: 0.9371 - loss: 309.3658 547/Unknown 122s 222ms/step - accuracy: 0.9371 - loss: 309.2743 548/Unknown 122s 222ms/step - accuracy: 0.9371 - loss: 309.1831 549/Unknown 122s 222ms/step - accuracy: 0.9371 - loss: 309.0924 550/Unknown 122s 222ms/step - accuracy: 0.9371 - loss: 309.0020 551/Unknown 123s 222ms/step - accuracy: 0.9371 - loss: 308.9121 552/Unknown 123s 222ms/step - accuracy: 0.9372 - loss: 308.8221 553/Unknown 123s 222ms/step - accuracy: 0.9372 - loss: 308.7321 554/Unknown 123s 222ms/step - accuracy: 0.9372 - loss: 308.6426 555/Unknown 124s 222ms/step - accuracy: 0.9372 - loss: 308.5534 556/Unknown 124s 222ms/step - accuracy: 0.9372 - loss: 308.4645 557/Unknown 124s 222ms/step - accuracy: 0.9372 - loss: 308.3757 558/Unknown 124s 222ms/step - accuracy: 0.9372 - loss: 308.2873 559/Unknown 125s 222ms/step - accuracy: 0.9373 - loss: 308.1992 560/Unknown 125s 222ms/step - accuracy: 0.9373 - loss: 308.1115 561/Unknown 125s 222ms/step - accuracy: 0.9373 - loss: 308.0239 562/Unknown 125s 222ms/step - accuracy: 0.9373 - loss: 307.9366 563/Unknown 126s 222ms/step - accuracy: 0.9373 - loss: 307.8493 564/Unknown 126s 222ms/step - accuracy: 0.9373 - loss: 307.7624 565/Unknown 126s 222ms/step - accuracy: 0.9373 - loss: 307.6758 566/Unknown 126s 222ms/step - accuracy: 0.9374 - loss: 307.5896 567/Unknown 126s 222ms/step - accuracy: 0.9374 - loss: 307.5038 568/Unknown 127s 222ms/step - accuracy: 0.9374 - loss: 307.4184 569/Unknown 127s 222ms/step - accuracy: 0.9374 - loss: 307.3336 570/Unknown 127s 222ms/step - accuracy: 0.9374 - loss: 307.2491 571/Unknown 127s 222ms/step - accuracy: 0.9374 - loss: 307.1649 572/Unknown 128s 222ms/step - accuracy: 0.9374 - loss: 307.0807 573/Unknown 128s 222ms/step - accuracy: 0.9375 - loss: 306.9967 574/Unknown 128s 223ms/step - accuracy: 0.9375 - loss: 306.9132 575/Unknown 129s 223ms/step - accuracy: 0.9375 - loss: 306.8299 576/Unknown 129s 223ms/step - accuracy: 0.9375 - loss: 306.7468 577/Unknown 129s 223ms/step - accuracy: 0.9375 - loss: 306.6639 578/Unknown 129s 223ms/step - accuracy: 0.9375 - loss: 306.5809 579/Unknown 130s 223ms/step - accuracy: 0.9375 - loss: 306.4984 580/Unknown 130s 223ms/step - accuracy: 0.9376 - loss: 306.4160 581/Unknown 130s 223ms/step - accuracy: 0.9376 - loss: 306.3336 582/Unknown 130s 223ms/step - accuracy: 0.9376 - loss: 306.2515 583/Unknown 131s 223ms/step - accuracy: 0.9376 - loss: 306.1693 584/Unknown 131s 223ms/step - accuracy: 0.9376 - loss: 306.0871 585/Unknown 131s 223ms/step - accuracy: 0.9376 - loss: 306.0051 586/Unknown 131s 223ms/step - accuracy: 0.9376 - loss: 305.9232 587/Unknown 132s 223ms/step - accuracy: 0.9376 - loss: 305.8415 588/Unknown 132s 223ms/step - accuracy: 0.9377 - loss: 305.7600 589/Unknown 132s 223ms/step - accuracy: 0.9377 - loss: 305.6787 590/Unknown 132s 223ms/step - accuracy: 0.9377 - loss: 305.5978 591/Unknown 133s 223ms/step - accuracy: 0.9377 - loss: 305.5169 592/Unknown 133s 223ms/step - accuracy: 0.9377 - loss: 305.4363 593/Unknown 133s 223ms/step - accuracy: 0.9377 - loss: 305.3560 594/Unknown 133s 223ms/step - accuracy: 0.9377 - loss: 305.2758 595/Unknown 133s 223ms/step - accuracy: 0.9378 - loss: 305.1960 596/Unknown 134s 223ms/step - accuracy: 0.9378 - loss: 305.1167 597/Unknown 134s 223ms/step - accuracy: 0.9378 - loss: 305.0374 598/Unknown 134s 223ms/step - accuracy: 0.9378 - loss: 304.9583 599/Unknown 134s 223ms/step - accuracy: 0.9378 - loss: 304.8791 600/Unknown 134s 223ms/step - accuracy: 0.9378 - loss: 304.8002 601/Unknown 135s 223ms/step - accuracy: 0.9378 - loss: 304.7217 602/Unknown 135s 223ms/step - accuracy: 0.9378 - loss: 304.6434 603/Unknown 135s 223ms/step - accuracy: 0.9379 - loss: 304.5656 604/Unknown 135s 223ms/step - accuracy: 0.9379 - loss: 304.4879 605/Unknown 136s 223ms/step - accuracy: 0.9379 - loss: 304.4105 606/Unknown 136s 223ms/step - accuracy: 0.9379 - loss: 304.3331 607/Unknown 136s 223ms/step - accuracy: 0.9379 - loss: 304.2561 608/Unknown 136s 223ms/step - accuracy: 0.9379 - loss: 304.1792 609/Unknown 137s 223ms/step - accuracy: 0.9379 - loss: 304.1025 610/Unknown 137s 223ms/step - accuracy: 0.9379 - loss: 304.0259 611/Unknown 137s 223ms/step - accuracy: 0.9380 - loss: 303.9494 612/Unknown 137s 223ms/step - accuracy: 0.9380 - loss: 303.8730 613/Unknown 137s 223ms/step - accuracy: 0.9380 - loss: 303.7966 614/Unknown 138s 223ms/step - accuracy: 0.9380 - loss: 303.7203 615/Unknown 138s 223ms/step - accuracy: 0.9380 - loss: 303.6440 616/Unknown 138s 223ms/step - accuracy: 0.9380 - loss: 303.5678 617/Unknown 138s 224ms/step - accuracy: 0.9380 - loss: 303.4917 618/Unknown 139s 224ms/step - accuracy: 0.9380 - loss: 303.4161 619/Unknown 139s 224ms/step - accuracy: 0.9380 - loss: 303.3405 620/Unknown 139s 224ms/step - accuracy: 0.9381 - loss: 303.2652 621/Unknown 140s 224ms/step - accuracy: 0.9381 - loss: 303.1904 622/Unknown 140s 224ms/step - accuracy: 0.9381 - loss: 303.1159 623/Unknown 140s 224ms/step - accuracy: 0.9381 - loss: 303.0418 624/Unknown 140s 224ms/step - accuracy: 0.9381 - loss: 302.9678 625/Unknown 141s 224ms/step - accuracy: 0.9381 - loss: 302.8940 626/Unknown 141s 224ms/step - accuracy: 0.9381 - loss: 302.8201 627/Unknown 141s 224ms/step - accuracy: 0.9381 - loss: 302.7465 628/Unknown 141s 224ms/step - accuracy: 0.9382 - loss: 302.6730 629/Unknown 141s 224ms/step - accuracy: 0.9382 - loss: 302.5998 630/Unknown 142s 224ms/step - accuracy: 0.9382 - loss: 302.5268 631/Unknown 142s 224ms/step - accuracy: 0.9382 - loss: 302.4539 632/Unknown 142s 224ms/step - accuracy: 0.9382 - loss: 302.3810 633/Unknown 142s 224ms/step - accuracy: 0.9382 - loss: 302.3086 634/Unknown 143s 224ms/step - accuracy: 0.9382 - loss: 302.2364 635/Unknown 143s 224ms/step - accuracy: 0.9382 - loss: 302.1645 636/Unknown 143s 224ms/step - accuracy: 0.9383 - loss: 302.0930 637/Unknown 143s 224ms/step - accuracy: 0.9383 - loss: 302.0216 638/Unknown 144s 224ms/step - accuracy: 0.9383 - loss: 301.9502 639/Unknown 144s 224ms/step - accuracy: 0.9383 - loss: 301.8791 /home/humbulani/tensorflow-env/env/lib/python3.11/site-packages/keras/src/trainers/epoch_iterator.py:151: UserWarning: Your input ran out of data; interrupting training. Make sure that your dataset or generator can generate at least `steps_per_epoch * epochs` batches. You may need to use the `.repeat()` function when building your dataset. self._interrupted_warning() </code></pre></div> </div> 639/639 ━━━━━━━━━━━━━━━━━━━━ 160s 249ms/step - accuracy: 0.9383 - loss: 301.8082 - val_accuracy: 0.9485 - val_loss: 235.7996 <div class="k-default-codeblock"> <div class="codehilite"><pre><code>Model training finished. Evaluating model performance... </code></pre></div> </div> <div class="k-default-codeblock"> <div class="codehilite"><pre><code> 1/Unknown 0s 331ms/step - accuracy: 0.9623 - loss: 160.6135 2/Unknown 0s 119ms/step - accuracy: 0.9557 - loss: 181.4366 3/Unknown 1s 131ms/step - accuracy: 0.9524 - loss: 198.4659 4/Unknown 1s 129ms/step - accuracy: 0.9502 - loss: 209.3009 5/Unknown 1s 133ms/step - accuracy: 0.9499 - loss: 215.6982 6/Unknown 1s 131ms/step - accuracy: 0.9499 - loss: 219.7466 7/Unknown 1s 132ms/step - accuracy: 0.9502 - loss: 220.2296 8/Unknown 1s 132ms/step - accuracy: 0.9504 - loss: 219.6000 9/Unknown 1s 133ms/step - accuracy: 0.9506 - loss: 218.5403 10/Unknown 2s 133ms/step - accuracy: 0.9507 - loss: 217.4007 11/Unknown 2s 134ms/step - accuracy: 0.9507 - loss: 216.4865 12/Unknown 2s 133ms/step - accuracy: 0.9504 - loss: 215.7090 13/Unknown 2s 135ms/step - accuracy: 0.9502 - loss: 215.4628 14/Unknown 2s 135ms/step - accuracy: 0.9500 - loss: 215.0735 15/Unknown 2s 134ms/step - accuracy: 0.9499 - loss: 214.8078 16/Unknown 2s 134ms/step - accuracy: 0.9500 - loss: 214.3558 17/Unknown 2s 134ms/step - accuracy: 0.9500 - loss: 213.9521 18/Unknown 3s 135ms/step - accuracy: 0.9501 - loss: 213.9012 19/Unknown 3s 134ms/step - accuracy: 0.9501 - loss: 214.0063 20/Unknown 3s 135ms/step - accuracy: 0.9501 - loss: 214.2168 21/Unknown 3s 134ms/step - accuracy: 0.9500 - loss: 214.5657 22/Unknown 3s 135ms/step - accuracy: 0.9500 - loss: 214.8618 23/Unknown 3s 134ms/step - accuracy: 0.9500 - loss: 215.1154 24/Unknown 3s 135ms/step - accuracy: 0.9499 - loss: 215.2906 25/Unknown 4s 134ms/step - accuracy: 0.9499 - loss: 215.6145 26/Unknown 4s 135ms/step - accuracy: 0.9499 - loss: 215.8544 27/Unknown 4s 135ms/step - accuracy: 0.9498 - loss: 216.0591 28/Unknown 4s 135ms/step - accuracy: 0.9499 - loss: 216.2666 29/Unknown 4s 135ms/step - accuracy: 0.9499 - loss: 216.4423 30/Unknown 4s 135ms/step - accuracy: 0.9499 - loss: 216.5613 31/Unknown 4s 135ms/step - accuracy: 0.9498 - loss: 216.7220 32/Unknown 5s 135ms/step - accuracy: 0.9498 - loss: 216.8842 33/Unknown 5s 135ms/step - accuracy: 0.9498 - loss: 217.1658 34/Unknown 5s 135ms/step - accuracy: 0.9497 - loss: 217.4608 35/Unknown 5s 135ms/step - accuracy: 0.9496 - loss: 217.7231 36/Unknown 5s 134ms/step - accuracy: 0.9496 - loss: 217.9504 37/Unknown 5s 135ms/step - accuracy: 0.9496 - loss: 218.1658 38/Unknown 5s 134ms/step - accuracy: 0.9495 - loss: 218.3597 39/Unknown 5s 134ms/step - accuracy: 0.9495 - loss: 218.5269 40/Unknown 6s 135ms/step - accuracy: 0.9495 - loss: 218.7106 41/Unknown 6s 135ms/step - accuracy: 0.9494 - loss: 218.8721 42/Unknown 6s 136ms/step - accuracy: 0.9494 - loss: 218.9924 43/Unknown 6s 136ms/step - accuracy: 0.9494 - loss: 219.1745 44/Unknown 6s 137ms/step - accuracy: 0.9494 - loss: 219.3449 45/Unknown 6s 138ms/step - accuracy: 0.9493 - loss: 219.4943 46/Unknown 7s 137ms/step - accuracy: 0.9493 - loss: 219.6201 47/Unknown 7s 138ms/step - accuracy: 0.9493 - loss: 219.7240 48/Unknown 7s 138ms/step - accuracy: 0.9493 - loss: 219.8335 49/Unknown 7s 138ms/step - accuracy: 0.9493 - loss: 219.9367 50/Unknown 7s 138ms/step - accuracy: 0.9493 - loss: 220.0834 51/Unknown 7s 138ms/step - accuracy: 0.9492 - loss: 220.2067 52/Unknown 7s 139ms/step - accuracy: 0.9492 - loss: 220.2963 53/Unknown 8s 138ms/step - accuracy: 0.9492 - loss: 220.3649 54/Unknown 8s 138ms/step - accuracy: 0.9492 - loss: 220.4462 55/Unknown 8s 138ms/step - accuracy: 0.9492 - loss: 220.5459 56/Unknown 8s 138ms/step - accuracy: 0.9492 - loss: 220.6197 57/Unknown 8s 138ms/step - accuracy: 0.9491 - loss: 220.6727 58/Unknown 8s 138ms/step - accuracy: 0.9491 - loss: 220.7652 59/Unknown 8s 138ms/step - accuracy: 0.9491 - loss: 220.8513 60/Unknown 8s 138ms/step - accuracy: 0.9491 - loss: 220.9392 61/Unknown 9s 138ms/step - accuracy: 0.9491 - loss: 221.0330 62/Unknown 9s 137ms/step - accuracy: 0.9491 - loss: 221.1127 63/Unknown 9s 137ms/step - accuracy: 0.9491 - loss: 221.2177 64/Unknown 9s 137ms/step - accuracy: 0.9490 - loss: 221.3382 65/Unknown 9s 137ms/step - accuracy: 0.9490 - loss: 221.4572 66/Unknown 9s 137ms/step - accuracy: 0.9490 - loss: 221.5552 67/Unknown 9s 137ms/step - accuracy: 0.9490 - loss: 221.6626 68/Unknown 9s 136ms/step - accuracy: 0.9490 - loss: 221.7653 69/Unknown 10s 136ms/step - accuracy: 0.9490 - loss: 221.8680 70/Unknown 10s 136ms/step - accuracy: 0.9490 - loss: 221.9582 71/Unknown 10s 136ms/step - accuracy: 0.9489 - loss: 222.0398 72/Unknown 10s 136ms/step - accuracy: 0.9489 - loss: 222.1409 73/Unknown 10s 136ms/step - accuracy: 0.9489 - loss: 222.2496 74/Unknown 10s 136ms/step - accuracy: 0.9489 - loss: 222.3526 75/Unknown 10s 136ms/step - accuracy: 0.9489 - loss: 222.4433 76/Unknown 11s 136ms/step - accuracy: 0.9489 - loss: 222.5272 77/Unknown 11s 136ms/step - accuracy: 0.9489 - loss: 222.6031 78/Unknown 11s 136ms/step - accuracy: 0.9488 - loss: 222.6857 79/Unknown 11s 136ms/step - accuracy: 0.9488 - loss: 222.7623 80/Unknown 11s 136ms/step - accuracy: 0.9488 - loss: 222.8322 81/Unknown 11s 136ms/step - accuracy: 0.9488 - loss: 222.8963 82/Unknown 11s 136ms/step - accuracy: 0.9488 - loss: 222.9694 83/Unknown 11s 136ms/step - accuracy: 0.9488 - loss: 223.0455 84/Unknown 12s 136ms/step - accuracy: 0.9488 - loss: 223.1209 85/Unknown 12s 136ms/step - accuracy: 0.9488 - loss: 223.1990 86/Unknown 12s 136ms/step - accuracy: 0.9488 - loss: 223.2825 87/Unknown 12s 136ms/step - accuracy: 0.9487 - loss: 223.3633 88/Unknown 12s 136ms/step - accuracy: 0.9487 - loss: 223.4366 89/Unknown 12s 136ms/step - accuracy: 0.9487 - loss: 223.5151 90/Unknown 12s 136ms/step - accuracy: 0.9487 - loss: 223.5958 91/Unknown 13s 136ms/step - accuracy: 0.9487 - loss: 223.6727 92/Unknown 13s 137ms/step - accuracy: 0.9487 - loss: 223.7505 93/Unknown 13s 137ms/step - accuracy: 0.9487 - loss: 223.8250 94/Unknown 13s 137ms/step - accuracy: 0.9486 - loss: 223.9114 95/Unknown 13s 137ms/step - accuracy: 0.9486 - loss: 223.9948 96/Unknown 13s 138ms/step - accuracy: 0.9486 - loss: 224.0807 97/Unknown 14s 138ms/step - accuracy: 0.9486 - loss: 224.1586 98/Unknown 14s 138ms/step - accuracy: 0.9486 - loss: 224.2289 99/Unknown 14s 138ms/step - accuracy: 0.9486 - loss: 224.2979 100/Unknown 14s 138ms/step - accuracy: 0.9486 - loss: 224.3739 101/Unknown 14s 138ms/step - accuracy: 0.9485 - loss: 224.4488 102/Unknown 14s 138ms/step - accuracy: 0.9485 - loss: 224.5210 103/Unknown 14s 139ms/step - accuracy: 0.9485 - loss: 224.5936 104/Unknown 15s 139ms/step - accuracy: 0.9485 - loss: 224.6630 105/Unknown 15s 139ms/step - accuracy: 0.9485 - loss: 224.7316 106/Unknown 15s 139ms/step - accuracy: 0.9485 - loss: 224.8002 107/Unknown 15s 139ms/step - accuracy: 0.9485 - loss: 224.8736 108/Unknown 15s 139ms/step - accuracy: 0.9484 - loss: 224.9466 109/Unknown 15s 138ms/step - accuracy: 0.9484 - loss: 225.0268 110/Unknown 15s 138ms/step - accuracy: 0.9484 - loss: 225.1065 111/Unknown 16s 138ms/step - accuracy: 0.9484 - loss: 225.1895 112/Unknown 16s 139ms/step - accuracy: 0.9484 - loss: 225.2730 113/Unknown 16s 139ms/step - accuracy: 0.9484 - loss: 225.3562 114/Unknown 16s 139ms/step - accuracy: 0.9484 - loss: 225.4317 115/Unknown 16s 139ms/step - accuracy: 0.9484 - loss: 225.5018 116/Unknown 16s 139ms/step - accuracy: 0.9483 - loss: 225.5749 117/Unknown 16s 139ms/step - accuracy: 0.9483 - loss: 225.6508 118/Unknown 17s 139ms/step - accuracy: 0.9483 - loss: 225.7233 119/Unknown 17s 139ms/step - accuracy: 0.9483 - loss: 225.7933 120/Unknown 17s 139ms/step - accuracy: 0.9483 - loss: 225.8627 121/Unknown 17s 139ms/step - accuracy: 0.9483 - loss: 225.9334 122/Unknown 17s 139ms/step - accuracy: 0.9483 - loss: 226.0017 123/Unknown 17s 139ms/step - accuracy: 0.9483 - loss: 226.0612 124/Unknown 17s 139ms/step - accuracy: 0.9483 - loss: 226.1206 125/Unknown 18s 139ms/step - accuracy: 0.9482 - loss: 226.1817 126/Unknown 18s 139ms/step - accuracy: 0.9482 - loss: 226.2397 127/Unknown 18s 139ms/step - accuracy: 0.9482 - loss: 226.2974 128/Unknown 18s 139ms/step - accuracy: 0.9482 - loss: 226.3512 129/Unknown 18s 139ms/step - accuracy: 0.9482 - loss: 226.4005 130/Unknown 18s 139ms/step - accuracy: 0.9482 - loss: 226.4489 131/Unknown 18s 139ms/step - accuracy: 0.9482 - loss: 226.4945 132/Unknown 19s 139ms/step - accuracy: 0.9482 - loss: 226.5418 133/Unknown 19s 139ms/step - accuracy: 0.9482 - loss: 226.5853 134/Unknown 19s 140ms/step - accuracy: 0.9482 - loss: 226.6308 135/Unknown 19s 140ms/step - accuracy: 0.9482 - loss: 226.6727 136/Unknown 19s 140ms/step - accuracy: 0.9482 - loss: 226.7198 137/Unknown 19s 140ms/step - accuracy: 0.9482 - loss: 226.7688 138/Unknown 19s 140ms/step - accuracy: 0.9482 - loss: 226.8152 139/Unknown 20s 140ms/step - accuracy: 0.9482 - loss: 226.8609 140/Unknown 20s 140ms/step - accuracy: 0.9482 - loss: 226.9056 141/Unknown 20s 140ms/step - accuracy: 0.9481 - loss: 226.9498 142/Unknown 20s 140ms/step - accuracy: 0.9481 - loss: 226.9936 143/Unknown 20s 140ms/step - accuracy: 0.9481 - loss: 227.0366 144/Unknown 20s 139ms/step - accuracy: 0.9481 - loss: 227.0785 145/Unknown 20s 139ms/step - accuracy: 0.9481 - loss: 227.1189 146/Unknown 21s 139ms/step - accuracy: 0.9481 - loss: 227.1589 147/Unknown 21s 139ms/step - accuracy: 0.9481 - loss: 227.1976 148/Unknown 21s 140ms/step - accuracy: 0.9481 - loss: 227.2341 149/Unknown 21s 140ms/step - accuracy: 0.9481 - loss: 227.2737 150/Unknown 21s 140ms/step - accuracy: 0.9481 - loss: 227.3114 151/Unknown 21s 140ms/step - accuracy: 0.9481 - loss: 227.3518 152/Unknown 21s 140ms/step - accuracy: 0.9481 - loss: 227.3940 153/Unknown 22s 140ms/step - accuracy: 0.9481 - loss: 227.4359 154/Unknown 22s 140ms/step - accuracy: 0.9481 - loss: 227.4770 155/Unknown 22s 140ms/step - accuracy: 0.9481 - loss: 227.5203 156/Unknown 22s 140ms/step - accuracy: 0.9481 - loss: 227.5634 157/Unknown 22s 140ms/step - accuracy: 0.9481 - loss: 227.6052 158/Unknown 22s 140ms/step - accuracy: 0.9481 - loss: 227.6457 159/Unknown 22s 140ms/step - accuracy: 0.9481 - loss: 227.6858 160/Unknown 23s 140ms/step - accuracy: 0.9481 - loss: 227.7236 161/Unknown 23s 140ms/step - accuracy: 0.9481 - loss: 227.7598 162/Unknown 23s 140ms/step - accuracy: 0.9481 - loss: 227.7944 163/Unknown 23s 140ms/step - accuracy: 0.9481 - loss: 227.8277 164/Unknown 23s 140ms/step - accuracy: 0.9481 - loss: 227.8596 165/Unknown 23s 140ms/step - accuracy: 0.9480 - loss: 227.8896 166/Unknown 23s 140ms/step - accuracy: 0.9480 - loss: 227.9187 167/Unknown 24s 140ms/step - accuracy: 0.9480 - loss: 227.9463 168/Unknown 24s 140ms/step - accuracy: 0.9480 - loss: 227.9726 169/Unknown 24s 140ms/step - accuracy: 0.9480 - loss: 227.9972 170/Unknown 24s 140ms/step - accuracy: 0.9480 - loss: 228.0219 171/Unknown 24s 140ms/step - accuracy: 0.9480 - loss: 228.0469 172/Unknown 24s 140ms/step - accuracy: 0.9480 - loss: 228.0727 173/Unknown 24s 140ms/step - accuracy: 0.9480 - loss: 228.0992 174/Unknown 25s 140ms/step - accuracy: 0.9480 - loss: 228.1232 175/Unknown 25s 140ms/step - accuracy: 0.9480 - loss: 228.1466 176/Unknown 25s 140ms/step - accuracy: 0.9480 - loss: 228.1700 177/Unknown 25s 140ms/step - accuracy: 0.9480 - loss: 228.1928 178/Unknown 25s 140ms/step - accuracy: 0.9480 - loss: 228.2146 179/Unknown 25s 140ms/step - accuracy: 0.9480 - loss: 228.2334 180/Unknown 25s 140ms/step - accuracy: 0.9480 - loss: 228.2502 181/Unknown 26s 140ms/step - accuracy: 0.9480 - loss: 228.2660 182/Unknown 26s 140ms/step - accuracy: 0.9480 - loss: 228.2829 183/Unknown 26s 140ms/step - accuracy: 0.9480 - loss: 228.3001 184/Unknown 26s 140ms/step - accuracy: 0.9480 - loss: 228.3197 185/Unknown 26s 140ms/step - accuracy: 0.9480 - loss: 228.3382 186/Unknown 26s 140ms/step - accuracy: 0.9480 - loss: 228.3550 187/Unknown 26s 140ms/step - accuracy: 0.9480 - loss: 228.3710 188/Unknown 26s 140ms/step - accuracy: 0.9480 - loss: 228.3866 189/Unknown 27s 140ms/step - accuracy: 0.9480 - loss: 228.4023 190/Unknown 27s 140ms/step - accuracy: 0.9480 - loss: 228.4187 191/Unknown 27s 140ms/step - accuracy: 0.9480 - loss: 228.4339 192/Unknown 27s 139ms/step - accuracy: 0.9480 - loss: 228.4476 193/Unknown 27s 139ms/step - accuracy: 0.9480 - loss: 228.4600 194/Unknown 27s 139ms/step - accuracy: 0.9480 - loss: 228.4714 195/Unknown 27s 139ms/step - accuracy: 0.9480 - loss: 228.4833 196/Unknown 27s 139ms/step - accuracy: 0.9480 - loss: 228.4972 197/Unknown 28s 139ms/step - accuracy: 0.9480 - loss: 228.5084 198/Unknown 28s 139ms/step - accuracy: 0.9480 - loss: 228.5233 199/Unknown 28s 139ms/step - accuracy: 0.9480 - loss: 228.5378 200/Unknown 28s 139ms/step - accuracy: 0.9480 - loss: 228.5527 201/Unknown 28s 139ms/step - accuracy: 0.9480 - loss: 228.5660 202/Unknown 28s 139ms/step - accuracy: 0.9480 - loss: 228.5789 203/Unknown 28s 139ms/step - accuracy: 0.9480 - loss: 228.5915 204/Unknown 29s 139ms/step - accuracy: 0.9480 - loss: 228.6026 205/Unknown 29s 139ms/step - accuracy: 0.9480 - loss: 228.6157 206/Unknown 29s 139ms/step - accuracy: 0.9480 - loss: 228.6292 207/Unknown 29s 139ms/step - accuracy: 0.9481 - loss: 228.6418 208/Unknown 29s 139ms/step - accuracy: 0.9481 - loss: 228.6543 209/Unknown 29s 139ms/step - accuracy: 0.9481 - loss: 228.6688 210/Unknown 29s 139ms/step - accuracy: 0.9481 - loss: 228.6819 211/Unknown 29s 139ms/step - accuracy: 0.9481 - loss: 228.6941 212/Unknown 30s 139ms/step - accuracy: 0.9481 - loss: 228.7058 213/Unknown 30s 139ms/step - accuracy: 0.9481 - loss: 228.7174 214/Unknown 30s 139ms/step - accuracy: 0.9481 - loss: 228.7294 215/Unknown 30s 139ms/step - accuracy: 0.9481 - loss: 228.7410 216/Unknown 30s 139ms/step - accuracy: 0.9481 - loss: 228.7537 217/Unknown 30s 139ms/step - accuracy: 0.9481 - loss: 228.7657 218/Unknown 30s 139ms/step - accuracy: 0.9481 - loss: 228.7765 219/Unknown 31s 139ms/step - accuracy: 0.9481 - loss: 228.7861 220/Unknown 31s 139ms/step - accuracy: 0.9481 - loss: 228.7945 221/Unknown 31s 139ms/step - accuracy: 0.9481 - loss: 228.8020 222/Unknown 31s 139ms/step - accuracy: 0.9481 - loss: 228.8082 223/Unknown 31s 139ms/step - accuracy: 0.9481 - loss: 228.8140 224/Unknown 31s 138ms/step - accuracy: 0.9481 - loss: 228.8199 225/Unknown 31s 138ms/step - accuracy: 0.9481 - loss: 228.8271 226/Unknown 31s 138ms/step - accuracy: 0.9481 - loss: 228.8340 227/Unknown 32s 138ms/step - accuracy: 0.9481 - loss: 228.8403 228/Unknown 32s 138ms/step - accuracy: 0.9481 - loss: 228.8456 229/Unknown 32s 138ms/step - accuracy: 0.9481 - loss: 228.8512 230/Unknown 32s 138ms/step - accuracy: 0.9481 - loss: 228.8562 231/Unknown 32s 138ms/step - accuracy: 0.9481 - loss: 228.8611 232/Unknown 32s 138ms/step - accuracy: 0.9481 - loss: 228.8642 233/Unknown 32s 138ms/step - accuracy: 0.9481 - loss: 228.8672 234/Unknown 33s 138ms/step - accuracy: 0.9481 - loss: 228.8694 235/Unknown 33s 138ms/step - accuracy: 0.9481 - loss: 228.8721 236/Unknown 33s 138ms/step - accuracy: 0.9481 - loss: 228.8762 237/Unknown 33s 138ms/step - accuracy: 0.9481 - loss: 228.8797 238/Unknown 33s 138ms/step - accuracy: 0.9481 - loss: 228.8838 239/Unknown 33s 138ms/step - accuracy: 0.9481 - loss: 228.8885 240/Unknown 33s 138ms/step - accuracy: 0.9481 - loss: 228.8919 241/Unknown 33s 138ms/step - accuracy: 0.9481 - loss: 228.8947 242/Unknown 34s 138ms/step - accuracy: 0.9481 - loss: 228.8974 243/Unknown 34s 138ms/step - accuracy: 0.9481 - loss: 228.9001 244/Unknown 34s 138ms/step - accuracy: 0.9481 - loss: 228.9023 245/Unknown 34s 138ms/step - accuracy: 0.9481 - loss: 228.9047 246/Unknown 34s 138ms/step - accuracy: 0.9481 - loss: 228.9063 247/Unknown 34s 139ms/step - accuracy: 0.9481 - loss: 228.9082 248/Unknown 35s 139ms/step - accuracy: 0.9481 - loss: 228.9095 249/Unknown 35s 139ms/step - accuracy: 0.9482 - loss: 228.9103 250/Unknown 35s 139ms/step - accuracy: 0.9482 - loss: 228.9118 251/Unknown 35s 139ms/step - accuracy: 0.9482 - loss: 228.9147 252/Unknown 35s 139ms/step - accuracy: 0.9482 - loss: 228.9173 253/Unknown 35s 139ms/step - accuracy: 0.9482 - loss: 228.9200 254/Unknown 36s 139ms/step - accuracy: 0.9482 - loss: 228.9222 255/Unknown 36s 139ms/step - accuracy: 0.9482 - loss: 228.9243 256/Unknown 36s 139ms/step - accuracy: 0.9482 - loss: 228.9268 257/Unknown 36s 139ms/step - accuracy: 0.9482 - loss: 228.9288 258/Unknown 36s 140ms/step - accuracy: 0.9482 - loss: 228.9315 259/Unknown 36s 140ms/step - accuracy: 0.9482 - loss: 228.9344 260/Unknown 36s 140ms/step - accuracy: 0.9482 - loss: 228.9363 261/Unknown 37s 140ms/step - accuracy: 0.9482 - loss: 228.9379 262/Unknown 37s 140ms/step - accuracy: 0.9482 - loss: 228.9381 263/Unknown 37s 140ms/step - accuracy: 0.9482 - loss: 228.9378 264/Unknown 37s 140ms/step - accuracy: 0.9482 - loss: 228.9372 265/Unknown 37s 140ms/step - accuracy: 0.9482 - loss: 228.9366 266/Unknown 37s 140ms/step - accuracy: 0.9482 - loss: 228.9373 267/Unknown 37s 140ms/step - accuracy: 0.9482 - loss: 228.9385 268/Unknown 38s 140ms/step - accuracy: 0.9482 - loss: 228.9391 269/Unknown 38s 140ms/step - accuracy: 0.9482 - loss: 228.9401 270/Unknown 38s 140ms/step - accuracy: 0.9482 - loss: 228.9409 271/Unknown 38s 139ms/step - accuracy: 0.9482 - loss: 228.9415 272/Unknown 38s 139ms/step - accuracy: 0.9482 - loss: 228.9413 273/Unknown 38s 139ms/step - accuracy: 0.9482 - loss: 228.9413 274/Unknown 38s 139ms/step - accuracy: 0.9482 - loss: 228.9406 275/Unknown 38s 139ms/step - accuracy: 0.9482 - loss: 228.9396 276/Unknown 39s 139ms/step - accuracy: 0.9482 - loss: 228.9395 277/Unknown 39s 139ms/step - accuracy: 0.9482 - loss: 228.9393 278/Unknown 39s 139ms/step - accuracy: 0.9482 - loss: 228.9390 279/Unknown 39s 139ms/step - accuracy: 0.9482 - loss: 228.9383 280/Unknown 39s 139ms/step - accuracy: 0.9483 - loss: 228.9368 281/Unknown 39s 139ms/step - accuracy: 0.9483 - loss: 228.9358 282/Unknown 39s 139ms/step - accuracy: 0.9483 - loss: 228.9339 283/Unknown 40s 139ms/step - accuracy: 0.9483 - loss: 228.9312 284/Unknown 40s 139ms/step - accuracy: 0.9483 - loss: 228.9295 285/Unknown 40s 139ms/step - accuracy: 0.9483 - loss: 228.9271 286/Unknown 40s 139ms/step - accuracy: 0.9483 - loss: 228.9249 287/Unknown 40s 139ms/step - accuracy: 0.9483 - loss: 228.9228 288/Unknown 40s 139ms/step - accuracy: 0.9483 - loss: 228.9200 289/Unknown 40s 139ms/step - accuracy: 0.9483 - loss: 228.9166 290/Unknown 40s 139ms/step - accuracy: 0.9483 - loss: 228.9136 291/Unknown 41s 139ms/step - accuracy: 0.9483 - loss: 228.9105 292/Unknown 41s 139ms/step - accuracy: 0.9483 - loss: 228.9076 293/Unknown 41s 139ms/step - accuracy: 0.9483 - loss: 228.9045 294/Unknown 41s 139ms/step - accuracy: 0.9483 - loss: 228.9017 295/Unknown 41s 139ms/step - accuracy: 0.9483 - loss: 228.8983 296/Unknown 41s 139ms/step - accuracy: 0.9483 - loss: 228.8967 297/Unknown 41s 139ms/step - accuracy: 0.9483 - loss: 228.8956 298/Unknown 42s 139ms/step - accuracy: 0.9483 - loss: 228.8946 299/Unknown 42s 139ms/step - accuracy: 0.9483 - loss: 228.8931 300/Unknown 42s 139ms/step - accuracy: 0.9483 - loss: 228.8911 301/Unknown 42s 139ms/step - accuracy: 0.9483 - loss: 228.8904 302/Unknown 42s 139ms/step - accuracy: 0.9483 - loss: 228.8897 303/Unknown 42s 139ms/step - accuracy: 0.9483 - loss: 228.8902 304/Unknown 42s 139ms/step - accuracy: 0.9484 - loss: 228.8911 305/Unknown 43s 139ms/step - accuracy: 0.9484 - loss: 228.8920 306/Unknown 43s 139ms/step - accuracy: 0.9484 - loss: 228.8923 307/Unknown 43s 139ms/step - accuracy: 0.9484 - loss: 228.8938 308/Unknown 43s 139ms/step - accuracy: 0.9484 - loss: 228.8946 309/Unknown 43s 139ms/step - accuracy: 0.9484 - loss: 228.8958 310/Unknown 43s 139ms/step - accuracy: 0.9484 - loss: 228.8979 311/Unknown 44s 139ms/step - accuracy: 0.9484 - loss: 228.9006 312/Unknown 44s 139ms/step - accuracy: 0.9484 - loss: 228.9027 313/Unknown 44s 139ms/step - accuracy: 0.9484 - loss: 228.9051 314/Unknown 44s 139ms/step - accuracy: 0.9484 - loss: 228.9074 315/Unknown 44s 140ms/step - accuracy: 0.9484 - loss: 228.9098 316/Unknown 44s 140ms/step - accuracy: 0.9484 - loss: 228.9117 317/Unknown 44s 140ms/step - accuracy: 0.9484 - loss: 228.9142 318/Unknown 45s 140ms/step - accuracy: 0.9484 - loss: 228.9171 319/Unknown 45s 140ms/step - accuracy: 0.9484 - loss: 228.9203 320/Unknown 45s 140ms/step - accuracy: 0.9484 - loss: 228.9232 321/Unknown 45s 140ms/step - accuracy: 0.9484 - loss: 228.9262 322/Unknown 45s 140ms/step - accuracy: 0.9484 - loss: 228.9293 323/Unknown 45s 140ms/step - accuracy: 0.9484 - loss: 228.9324 324/Unknown 46s 140ms/step - accuracy: 0.9484 - loss: 228.9353 325/Unknown 46s 140ms/step - accuracy: 0.9484 - loss: 228.9387 326/Unknown 46s 140ms/step - accuracy: 0.9484 - loss: 228.9429 327/Unknown 46s 140ms/step - accuracy: 0.9484 - loss: 228.9469 328/Unknown 46s 140ms/step - accuracy: 0.9484 - loss: 228.9508 329/Unknown 46s 140ms/step - accuracy: 0.9484 - loss: 228.9543 330/Unknown 46s 140ms/step - accuracy: 0.9484 - loss: 228.9582 331/Unknown 46s 140ms/step - accuracy: 0.9484 - loss: 228.9617 332/Unknown 47s 140ms/step - accuracy: 0.9484 - loss: 228.9662 333/Unknown 47s 140ms/step - accuracy: 0.9484 - loss: 228.9707 334/Unknown 47s 140ms/step - accuracy: 0.9484 - loss: 228.9747 335/Unknown 47s 140ms/step - accuracy: 0.9484 - loss: 228.9783 336/Unknown 47s 140ms/step - accuracy: 0.9485 - loss: 228.9814 337/Unknown 47s 140ms/step - accuracy: 0.9485 - loss: 228.9847 338/Unknown 47s 140ms/step - accuracy: 0.9485 - loss: 228.9882 339/Unknown 47s 140ms/step - accuracy: 0.9485 - loss: 228.9913 340/Unknown 48s 139ms/step - accuracy: 0.9485 - loss: 228.9950 341/Unknown 48s 139ms/step - accuracy: 0.9485 - loss: 228.9996 342/Unknown 48s 139ms/step - accuracy: 0.9485 - loss: 229.0038 343/Unknown 48s 139ms/step - accuracy: 0.9485 - loss: 229.0080 344/Unknown 48s 139ms/step - accuracy: 0.9485 - loss: 229.0122 345/Unknown 48s 139ms/step - accuracy: 0.9485 - loss: 229.0163 346/Unknown 48s 139ms/step - accuracy: 0.9485 - loss: 229.0210 347/Unknown 49s 139ms/step - accuracy: 0.9485 - loss: 229.0256 348/Unknown 49s 139ms/step - accuracy: 0.9485 - loss: 229.0313 349/Unknown 49s 139ms/step - accuracy: 0.9485 - loss: 229.0372 350/Unknown 49s 139ms/step - accuracy: 0.9485 - loss: 229.0429 351/Unknown 49s 139ms/step - accuracy: 0.9485 - loss: 229.0492 352/Unknown 49s 139ms/step - accuracy: 0.9485 - loss: 229.0551 353/Unknown 49s 139ms/step - accuracy: 0.9485 - loss: 229.0615 354/Unknown 49s 139ms/step - accuracy: 0.9485 - loss: 229.0690 355/Unknown 50s 139ms/step - accuracy: 0.9485 - loss: 229.0767 356/Unknown 50s 139ms/step - accuracy: 0.9485 - loss: 229.0841 357/Unknown 50s 139ms/step - accuracy: 0.9485 - loss: 229.0911 358/Unknown 50s 139ms/step - accuracy: 0.9485 - loss: 229.0985 359/Unknown 50s 139ms/step - accuracy: 0.9485 - loss: 229.1061 360/Unknown 50s 139ms/step - accuracy: 0.9485 - loss: 229.1136 361/Unknown 50s 139ms/step - accuracy: 0.9485 - loss: 229.1216 362/Unknown 51s 139ms/step - accuracy: 0.9485 - loss: 229.1297 363/Unknown 51s 139ms/step - accuracy: 0.9485 - loss: 229.1375 364/Unknown 51s 139ms/step - accuracy: 0.9485 - loss: 229.1453 365/Unknown 51s 139ms/step - accuracy: 0.9485 - loss: 229.1526 366/Unknown 51s 139ms/step - accuracy: 0.9485 - loss: 229.1591 367/Unknown 51s 139ms/step - accuracy: 0.9485 - loss: 229.1653 368/Unknown 51s 139ms/step - accuracy: 0.9485 - loss: 229.1717 369/Unknown 52s 139ms/step - accuracy: 0.9485 - loss: 229.1778 370/Unknown 52s 139ms/step - accuracy: 0.9485 - loss: 229.1837 371/Unknown 52s 139ms/step - accuracy: 0.9485 - loss: 229.1893 372/Unknown 52s 139ms/step - accuracy: 0.9485 - loss: 229.1947 373/Unknown 52s 139ms/step - accuracy: 0.9485 - loss: 229.1999 374/Unknown 52s 139ms/step - accuracy: 0.9485 - loss: 229.2054 375/Unknown 52s 139ms/step - accuracy: 0.9485 - loss: 229.2110 376/Unknown 53s 139ms/step - accuracy: 0.9486 - loss: 229.2163 377/Unknown 53s 139ms/step - accuracy: 0.9486 - loss: 229.2217 </code></pre></div> </div> 377/377 ━━━━━━━━━━━━━━━━━━━━ 53s 139ms/step - accuracy: 0.9486 - loss: 229.2270 <div class="k-default-codeblock"> <div class="codehilite"><pre><code>Test accuracy: 94.94% </code></pre></div> </div> You should achieve more than 95% accuracy on the test set. To increase the learning capacity of the model, you can try increasing the <code>encoding_size</code> value, or stacking multiple GRN layers on top of the VSN layer. This may require to also increase the <code>dropout_rate</code> value to avoid overfitting. Example available on HuggingFace <table> <thead> <tr> <th style="text-align: center;">Trained Model</th> <th style="text-align: center;">Demo</th> </tr> </thead> <tbody> <tr> <td style="text-align: center;"><a href="https://huggingface.co/keras-io/structured-data-classification-grn-vsn"><img alt="Generic badge" src="https://img.shields.io/badge/%F0%9F%A4%97%20Model-Classification%20With%20GRN%20%26%20VSN-red" /></a></td> <td style="text-align: center;"><a href="https://huggingface.co/spaces/keras-io/structured-data-classification-grn-vsn"><img alt="Generic badge" src="https://img.shields.io/badge/%F0%9F%A4%97%20Space-Classification%20With%20GRN%20%26%20VSN-red" /></a></td> </tr> </tbody> </table> </div> <div class='k-outline'> <div class='k-outline-depth-1'> <a href='#classification-with-gated-residual-and-variable-selection-networks'>Classification with Gated Residual and Variable Selection Networks</a> </div> <div class='k-outline-depth-2'> ◆ <a href='#introduction'>Introduction</a> </div> <div class='k-outline-depth-2'> ◆ <a href='#the-dataset'>The dataset</a> </div> <div class='k-outline-depth-2'> ◆ <a href='#setup'>Setup</a> </div> <div class='k-outline-depth-2'> ◆ <a href='#prepare-the-data'>Prepare the data</a> </div> <div class='k-outline-depth-2'> ◆ <a href='#define-dataset-metadata'>Define dataset metadata</a> </div> <div class='k-outline-depth-2'> ◆ <a href='#create-a-tfdatadataset-for-training-and-evaluation'>Create a <code>tf.data.Dataset</code> for training and evaluation</a> </div> <div class='k-outline-depth-2'> ◆ <a href='#create-model-inputs'>Create model inputs</a> </div> <div class='k-outline-depth-2'> ◆ <a href='#implement-the-gated-linear-unit'>Implement the Gated Linear Unit</a> </div> <div class='k-outline-depth-2'> ◆ <a href='#implement-the-gated-residual-network'>Implement the Gated Residual Network</a> </div> <div class='k-outline-depth-2'> ◆ <a href='#implement-the-variable-selection-network'>Implement the Variable Selection Network</a> </div> <div class='k-outline-depth-2'> ◆ <a href='#create-gated-residual-and-variable-selection-networks-model'>Create Gated Residual and Variable Selection Networks model</a> </div> <div class='k-outline-depth-2'> ◆ <a href='#compile-train-and-evaluate-the-model'>Compile, train, and evaluate the model</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>

CINXE.COM

Classification with Gated Residual and Variable Selection Networks