From 80a15966ba561e4d77a83bd325ebabc6bfb9d1f6 Mon Sep 17 00:00:00 2001
From: dhirensr <dhirensr@gmail.com>
Date: Wed, 18 Mar 2020 21:44:22 +0100
Subject: [PATCH] added docstring for compute and apply gradients

---
 tfjs-core/src/optimizers/optimizer.ts | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/tfjs-core/src/optimizers/optimizer.ts b/tfjs-core/src/optimizers/optimizer.ts
index d7fe027986a..fc41ddac062 100644
--- a/tfjs-core/src/optimizers/optimizer.ts
+++ b/tfjs-core/src/optimizers/optimizer.ts
@@ -103,6 +103,7 @@ export abstract class Optimizer extends Serializable {
    * respect to. If specified, only the trainable variables in varList will have
    * gradients computed with respect to. Defaults to all trainable variables.
    */
+  /** @doc {heading: 'Training', subheading: 'Optimizers'} */
   computeGradients(f: () => Scalar, varList?: Variable[]):
       {value: Scalar, grads: NamedTensorMap} {
     return variableGrads(f, varList);
@@ -113,6 +114,7 @@ export abstract class Optimizer extends Serializable {
    *
    * @param variableGradients A mapping of variable name to its gradient value.
    */
+  /** @doc {heading: 'Training', subheading: 'Optimizers'} */
   abstract applyGradients(variableGradients: NamedTensorMap|
                           NamedTensor[]): void;