feat: add Jupyter notebook for simple neural network implementation with CUDA support

16.ipynb

@@ -0,0 +1,117 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "id": "initial_id",
+   "metadata": {
+    "collapsed": true
+   },
+   "source": [
+    "import torch\n",
+    "\n",
+    "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")"
+   ],
+   "outputs": [],
+   "execution_count": null
+  },
+  {
+   "metadata": {},
+   "cell_type": "code",
+   "source": [
+    "data = torch.ones(3, 3)\n",
+    "print(data.device)"
+   ],
+   "id": "7a630763614905d",
+   "outputs": [],
+   "execution_count": null
+  },
+  {
+   "metadata": {},
+   "cell_type": "code",
+   "source": [
+    "device = torch.device(\"cuda:0\")\n",
+    "data_gpu = data.to(device)\n",
+    "print(data_gpu.device)"
+   ],
+   "id": "e2a2d8a6d60231c",
+   "outputs": [],
+   "execution_count": null
+  },
+  {
+   "metadata": {},
+   "cell_type": "code",
+   "source": [
+    "import torch.nn as nn\n",
+    "\n",
+    "net = nn.Sequential(nn.Linear(3, 3))\n",
+    "net.to(device)"
+   ],
+   "id": "458ea27224fd0061",
+   "outputs": [],
+   "execution_count": null
+  },
+  {
+   "metadata": {},
+   "cell_type": "code",
+   "source": [
+    "from torch import nn\n",
+    "\n",
+    "\n",
+    "class ASimpleNet(nn.Module):\n",
+    "    def __init__(self, layers=3):\n",
+    "        super(ASimpleNet, self).__init__()\n",
+    "        self.linears = nn.ModuleList([nn.Linear(3, 3, bias=False) for i in range(layers)])\n",
+    "\n",
+    "    def forward(self, x):\n",
+    "        print(\"forward batchsize is: {}\".format(x.size()[0]))\n",
+    "        x = self.linears(x)\n",
+    "        x = torch.relu(x)\n",
+    "        return x"
+   ],
+   "id": "4859aa95dd22d01d",
+   "outputs": [],
+   "execution_count": null
+  },
+  {
+   "metadata": {},
+   "cell_type": "code",
+   "source": [
+    "batch_size = 16\n",
+    "inputs = torch.randn(batch_size, 3)\n",
+    "labels = torch.randn(batch_size, 3)\n",
+    "inputs, labels = inputs.to(device), labels.to(device)\n",
+    "net = ASimpleNet()\n",
+    "net = nn.DataParallel(net)\n",
+    "net.to(device)\n",
+    "# print(\"CUDA_VISIBLE_DEVICES :{}\".format(os.environ[\"CUDA_VISIBLE_DEVICES\"]))\n",
+    "\n",
+    "for epoch in range(1):\n",
+    "    outputs = net(inputs)"
+   ],
+   "id": "d3eeb897f7f0ee68",
+   "outputs": [],
+   "execution_count": null
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 2
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython2",
+   "version": "2.7.6"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

main.py

@@ -1,21 +0,0 @@
-import torch
-import numpy as np
-
-print(torch.__version__)
-
-
-A = torch.tensor([[1, 2, 3], [4, 5, 6]])  # shape = (2, 3)
-
-B0 = torch.stack([A, A], dim=0)
-print("dim=0:\n", B0)
-print("shape:", B0.shape)
-
-B1 = torch.stack([A, A], dim=1)
-print("dim=1:\n", B1)
-print("shape:", B1.shape)
-
-B2 = torch.stack([A, A], dim=2)
-print("dim=2:\n", B2)
-print("shape:", B2.shape)
-
-np.max()