Train and save MNIST CNN model. Args: filters: number of filters for convolution layers. kernel_size: kernel size for convolution layers.

Train or re-train a model using data. Args: model: `keras.Model` instance to (re-)train. optimizer: Name of the optimizer to use. train

Train a model for IMDB sentiment classification. Args: model_type: Type of the model to train, as a `str`. vocabulary_size: Vocabulary size

* Warms up the model. Subsequent calls will be faster.

* Calcuate and save the maximally-activating input images for a covn2d layer. * * @param {tf.Model} model The model that the conv2d layer of interes

* Writes an array to an NDJSON file. * * @param {string} path * @param {array} collection

Write the transfer-learning data to .js files. Args: x_train_lt5: x (image) data for training: digits >= 5. y_train_lt5: y (label) data for

* Append an item to the replay buffer. * * @param {any} item The item to append.

* Convert an input monocolor image to color by applying a color map. * * @param {tf.Tensor4d} x Input monocolor image, assumed to be of shape *

* The actual computation performed by an instance of ZLayer. * * @param {Tensor[]} inputs this layer takes two input tensors, z_mean and *

* Centers the input and applies the following function to every element of * the input. * * x => [max(x, 0), max(-x, 0)] * * The th

When leaving a detector result or label hover.

* If a custom layer class is to support serialization, it must implement * the `className` static getter.

* Removes text elements from DOM on a left click.

* What action to take when someone clicks the right-click menu option. * Here it takes the url of the right-clicked image and the current tabId, and

* This layer only works on 4D Tensors [batch, height, width, channels], * and produces output with twice as many channels. * * layer.computeO

* Constructor of SnakeGame. * * @param {object} args Configurations for the game. Fields include: * - height {number} height of the board (

* Constructor of ReplayMemory. * * @param {number} maxLen Maximal buffer length.

* Constructor of SnakeGameAgent. * * @param {SnakeGame} game A game object. * @param {object} config The configuration object with the follow

* Constructor of CartPole.

* Constructor of PolicyNetwork. * * @param {number | number[] | tf.LayersModel} hiddenLayerSizes * Can be any of the following * - Siz

* Constructor of SaveablePolicyNetwork * * @param {number | number[]} hiddenLayerSizesOrModel

* Constructor of NeuralNetworkTextGenerator. * * @param {TextData} textData An instance of `TextData`.

* Constructor of NeuralNetworkTextGenerator. * * @param {TextData} textData An instance of `TextData`.

* Constructor of TextData. * * @param {string} dataIdentifier An identifier for this instance of TextData. * @param {string} textString The t

* Constructor of ObjectDetectionImageSynthesizer. * * @param {} canvas An HTML canvas object or node-canvas object. * @param {*} tensorFlow A

* Constructor of CharacterTable. * @param chars A string that contains the characters that can appear * in the input.

* Constructor of CharacterTable. * @param chars A string that contains the characters that can appear * in the input.

* Copies the embed URL to the clipboard.

* Builds a CSV Dataset object using the URL specified in the UI. Then iterates * over all elements in that dataset to count them. Updates the UI ac

* Custom loss function for object detection. * * The loss function is a sum of two losses * - shape-class loss, computed as binaryCrossentropy and

* Creates a dataset pipeline from GAME_GENERATOR_DATASET by: * 1) Applying the function gameToFeaturesAndlabel * 2) Taking the first N samples of th