From 3684295e5a208da5016d6f0d6d1e9eeb1aed7565 Mon Sep 17 00:00:00 2001
From: NeTT <nett@nett.moe>
Date: Sun, 10 Dec 2023 19:05:55 +0530
Subject: [PATCH] change comments

---
 examples/classification/iris.ts | 12 ++++++------
 examples/classification/spam.ts |  4 ++--
 2 files changed, 8 insertions(+), 8 deletions(-)

diff --git a/examples/classification/iris.ts b/examples/classification/iris.ts
index 8bea622..01f45bc 100644
--- a/examples/classification/iris.ts
+++ b/examples/classification/iris.ts
@@ -58,17 +58,17 @@ const net = new Sequential({
 
   // Define each layer of the network
   layers: [
-    // A dense layer with 4 neurons
+    // A dense layer with 16 neurons
     DenseLayer({ size: [16] }),
-    // A sigmoid activation layer
+    // A ReLu activation layer
     ReluLayer(),
-    // A dense layer with 1 neuron
+    // A dense layer with 3 neurons
     DenseLayer({ size: [3] }),
-    // Another sigmoid layer
+    // A Softmax activation layer
     SoftmaxLayer(),
   ],
   optimizer: AdamOptimizer(),
-  // We are using MSE for finding cost
+  // We are using CrossEntropy for finding cost
   cost: Cost.CrossEntropy,
   scheduler: OneCycle({ max_rate: 0.05, step_size: 50 }),
 });
@@ -83,7 +83,7 @@ net.train(
       outputs: tensor2D(train[1]),
     },
   ],
-  // Train for 10000 epochs
+  // Train for 300 epochs
   300,
   1,
   0.02,
diff --git a/examples/classification/spam.ts b/examples/classification/spam.ts
index 4a14d44..ff5b181 100644
--- a/examples/classification/spam.ts
+++ b/examples/classification/spam.ts
@@ -90,7 +90,7 @@ const net = new Sequential({
     SigmoidLayer(),
   ],
 
-  // We are using MSE for finding cost
+  // We are using Log Loss for finding cost
   cost: Cost.BinCrossEntropy,
   optimizer: AdamOptimizer(),
 });
@@ -106,7 +106,7 @@ net.train(
       outputs: tensor2D(train[1].map((x) => [x])),
     },
   ],
-  // Train for 10000 epochs
+  // Train for 20 epochs
   20,
   1,
   0.01,