diff --git a/appendix-A/03_main-chapter-code/exercise-solutions.ipynb b/appendix-A/03_main-chapter-code/exercise-solutions.ipynb
new file mode 100644
index 0000000..4c49997
--- /dev/null
+++ b/appendix-A/03_main-chapter-code/exercise-solutions.ipynb
@@ -0,0 +1,171 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## 练习 A.3"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import torch\n",
+    "\n",
+    "class NeuralNetwork(torch.nn.Module):\n",
+    "    def __init__(self, num_inputs, num_outputs):\n",
+    "        super().__init__()\n",
+    "\n",
+    "        self.layers = torch.nn.Sequential(\n",
+    "                \n",
+    "            # 第一个隐藏层\n",
+    "            torch.nn.Linear(num_inputs, 30),\n",
+    "            torch.nn.ReLU(),\n",
+    "\n",
+    "            # 第二个隐藏层\n",
+    "            torch.nn.Linear(30, 20),\n",
+    "            torch.nn.ReLU(),\n",
+    "\n",
+    "            # 输出层\n",
+    "            torch.nn.Linear(20, num_outputs),\n",
+    "        )\n",
+    "\n",
+    "    def forward(self, x):\n",
+    "        logits = self.layers(x)\n",
+    "        return logits"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Total number of trainable model parameters: 752\n"
+     ]
+    }
+   ],
+   "source": [
+    "model = NeuralNetwork(2, 2)\n",
+    "\n",
+    "num_params = sum(p.numel() for p in model.parameters() if p.requires_grad)\n",
+    "print(\"Total number of trainable model parameters:\", num_params)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## 练习 A.4"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {
+    "id": "qGgnamiyLJxp"
+   },
+   "outputs": [],
+   "source": [
+    "import torch\n",
+    "# 创建随机向量\n",
+    "a = torch.rand(100, 200)\n",
+    "b = torch.rand(200, 300)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {
+    "colab": {
+     "base_uri": "https://localhost:8080/"
+    },
+    "id": "CvGvIeVkLzXE",
+    "outputId": "44d027be-0787-4348-9c06-4e559d94d0e1"
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "63.8 µs ± 8.7 µs per loop (mean ± std. dev. of 7 runs, 10000 loops each)\n"
+     ]
+    }
+   ],
+   "source": [
+    "# 使用 @ 符号进行矩阵相乘，并计算执行时间\n",
+    "# %timeit 是 IPython 提供的魔术命令，用于多次执行代码以获取平均执行时间\n",
+    "# 它会自动选择执行次数以确保结果的准确性\n",
+    "%timeit a @ b"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {
+    "id": "OmRtZLa9L2ZG"
+   },
+   "outputs": [],
+   "source": [
+    "# 将 a 和 b 移动到 CUDA 设备上以利用 GPU 加速计算\n",
+    "a, b = a.to(\"cuda\"), b.to(\"cuda\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "metadata": {
+    "colab": {
+     "base_uri": "https://localhost:8080/"
+    },
+    "id": "duLEhXDPL6k0",
+    "outputId": "3486471d-fd62-446f-9855-2d01f41fd101"
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "13.8 µs ± 425 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each)\n"
+     ]
+    }
+   ],
+   "source": [
+    "%timeit a @ b"
+   ]
+  }
+ ],
+ "metadata": {
+  "accelerator": "GPU",
+  "colab": {
+   "gpuType": "V100",
+   "machine_shape": "hm",
+   "provenance": []
+  },
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.5"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 4
+}