* Compiles the inference graph and returns the minimal set of nodes that are * required for execution, in the correct execution order. * @return

* Returns the singleton instance of the map.

* Provided `f(x)`, returns another function `g(x, dy?)`, which gives the * gradient of `f(x)` with respect to `x`. * * If `dy` is provided, the gra

* Load a model from local storage. * * See the documentation to `browserLocalStorage` for details on the saved * artifacts. * * @return

* Reset the state tensors of the RNN. * * If the `states` argument is `undefined` or `null`, will set the * state tensor(s) of the RNN to all

FullName after removing the local path to the repository. If we have a real absolute path name here we can try to do something smart: detecti

* Creates a `tf.Tensor` filled with a scalar value. * * ```js * tf.fill([2, 2], 4).print(); * ``` * * @param shape An array of integers defining

@nocollapse

* Wrap the model's predict function (`model.predict` for tf.LayersModel * and `model.executeAsync` for tf.GraphModel) with the input. * * @param mo

* Returns gradients of `f` with respect to each of the `xs`. The gradients * returned are of the same length as `xs`, but some might be null if `f`

* Write a histogram summary, for later analysis in TensorBoard's 'Histograms' * and 'Distributions' dashboards (data written using this API will ap

* Synchronously construct the in memory weight map and * compile the inference graph. * * @doc {heading: 'Models', subheading: 'Classes', ign