train_model.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# ImageNet"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torchvision\n",
    "from torch.utils.data import DataLoader\n",
    "import torchvision.transforms as transforms\n",
    "from torchsummary import summary\n",
    "import numpy as np\n",
    "from models import train_and_test\n",
    "from models import AlexNet_imagenet, VGG11_imagenet, ResNet18_imagenet\n",
    "data_transform = transforms.Compose([\n",
    "        transforms.Resize(256),\n",
    "        transforms.CenterCrop(224),\n",
    "        transforms.ToTensor(),\n",
    "        transforms.Normalize(mean=[0.485, 0.456, 0.406],\n",
    "                             std=[0.229, 0.224, 0.225])\n",
    "    ])\n",
    "device = 'cuda' if torch.cuda.is_available() else 'cpu'\n",
    "train_dataset =torchvision.datasets.ImageFolder(root='./datasets/ImageNet/train',transform=data_transform)\n",
    "test_dataset = torchvision.datasets.ImageFolder(root='./datasets/ImageNet/val',transform=data_transform)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------------------------------------------------\n",
      "        Layer (type)               Output Shape         Param #\n",
      "================================================================\n",
      "            Conv2d-1           [-1, 64, 55, 55]          23,296\n",
      "              ReLU-2           [-1, 64, 55, 55]               0\n",
      "         MaxPool2d-3           [-1, 64, 27, 27]               0\n",
      "            Conv2d-4          [-1, 192, 27, 27]         307,392\n",
      "              ReLU-5          [-1, 192, 27, 27]               0\n",
      "         MaxPool2d-6          [-1, 192, 13, 13]               0\n",
      "            Conv2d-7          [-1, 384, 13, 13]         663,936\n",
      "              ReLU-8          [-1, 384, 13, 13]               0\n",
      "            Conv2d-9          [-1, 256, 13, 13]         884,992\n",
      "             ReLU-10          [-1, 256, 13, 13]               0\n",
      "           Conv2d-11          [-1, 256, 13, 13]         590,080\n",
      "             ReLU-12          [-1, 256, 13, 13]               0\n",
      "        MaxPool2d-13            [-1, 256, 6, 6]               0\n",
      "AdaptiveAvgPool2d-14            [-1, 256, 6, 6]               0\n",
      "          Dropout-15                 [-1, 9216]               0\n",
      "           Linear-16                 [-1, 4096]      37,752,832\n",
      "             ReLU-17                 [-1, 4096]               0\n",
      "           Linear-18                   [-1, 10]          40,970\n",
      "================================================================\n",
      "Total params: 40,263,498\n",
      "Trainable params: 40,263,498\n",
      "Non-trainable params: 0\n",
      "----------------------------------------------------------------\n",
      "Input size (MB): 0.57\n",
      "Forward/backward pass size (MB): 8.28\n",
      "Params size (MB): 153.59\n",
      "Estimated Total Size (MB): 162.44\n",
      "----------------------------------------------------------------\n",
      "Epoch 1\n",
      "-------------------------------\n",
      "loss:2.302833 [    0/13000]\n",
      "loss:1.519196 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 44.2%, Avg loss: 1.524913 \n",
      "\n",
      "Epoch 2\n",
      "-------------------------------\n",
      "loss:1.519548 [    0/13000]\n",
      "loss:1.183344 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 57.2%, Avg loss: 1.148726 \n",
      "\n",
      "Epoch 3\n",
      "-------------------------------\n",
      "loss:1.100953 [    0/13000]\n",
      "loss:1.173143 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 64.2%, Avg loss: 1.041110 \n",
      "\n",
      "Epoch 4\n",
      "-------------------------------\n",
      "loss:1.026734 [    0/13000]\n",
      "loss:0.955860 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 75.8%, Avg loss: 0.794213 \n",
      "\n",
      "Epoch 5\n",
      "-------------------------------\n",
      "loss:0.848215 [    0/13000]\n",
      "loss:0.755196 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 75.8%, Avg loss: 0.730222 \n",
      "\n",
      "Epoch 6\n",
      "-------------------------------\n",
      "loss:0.742704 [    0/13000]\n",
      "loss:0.659527 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 73.4%, Avg loss: 0.780190 \n",
      "\n",
      "Epoch 7\n",
      "-------------------------------\n",
      "loss:0.607903 [    0/13000]\n",
      "loss:0.534192 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 79.0%, Avg loss: 0.662876 \n",
      "\n",
      "Epoch 8\n",
      "-------------------------------\n",
      "loss:0.516372 [    0/13000]\n",
      "loss:0.562488 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 76.0%, Avg loss: 0.694512 \n",
      "\n",
      "Epoch 9\n",
      "-------------------------------\n",
      "loss:0.287672 [    0/13000]\n",
      "loss:0.355157 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 80.0%, Avg loss: 0.672684 \n",
      "\n",
      "Epoch 10\n",
      "-------------------------------\n",
      "loss:0.391431 [    0/13000]\n",
      "loss:0.442767 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 81.0%, Avg loss: 0.585398 \n",
      "\n",
      "Epoch 11\n",
      "-------------------------------\n",
      "loss:0.347941 [    0/13000]\n",
      "loss:0.355012 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 79.0%, Avg loss: 0.677627 \n",
      "\n",
      "Epoch 12\n",
      "-------------------------------\n",
      "loss:0.222284 [    0/13000]\n",
      "loss:0.312700 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 76.8%, Avg loss: 0.742325 \n",
      "\n",
      "Epoch 13\n",
      "-------------------------------\n",
      "loss:0.343419 [    0/13000]\n",
      "loss:0.325134 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 77.8%, Avg loss: 0.769010 \n",
      "\n",
      "Epoch 14\n",
      "-------------------------------\n",
      "loss:0.253720 [    0/13000]\n",
      "loss:0.347559 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 76.8%, Avg loss: 0.902158 \n",
      "\n",
      "Epoch 15\n",
      "-------------------------------\n",
      "loss:0.165903 [    0/13000]\n",
      "loss:0.216307 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 77.8%, Avg loss: 0.786525 \n",
      "\n",
      "Epoch 16\n",
      "-------------------------------\n",
      "loss:0.228767 [    0/13000]\n",
      "loss:0.231927 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 78.4%, Avg loss: 0.719160 \n",
      "\n",
      "Epoch 17\n",
      "-------------------------------\n",
      "loss:0.156461 [    0/13000]\n",
      "loss:0.077600 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 77.2%, Avg loss: 0.898726 \n",
      "\n",
      "Epoch 18\n",
      "-------------------------------\n",
      "loss:0.113504 [    0/13000]\n",
      "loss:0.173572 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 76.8%, Avg loss: 0.898588 \n",
      "\n",
      "Epoch 19\n",
      "-------------------------------\n",
      "loss:0.101595 [    0/13000]\n",
      "loss:0.150830 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 79.6%, Avg loss: 0.895515 \n",
      "\n",
      "Epoch 20\n",
      "-------------------------------\n",
      "loss:0.097031 [    0/13000]\n",
      "loss:0.107100 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 78.0%, Avg loss: 1.066057 \n",
      "\n",
      "Epoch 21\n",
      "-------------------------------\n",
      "loss:0.065478 [    0/13000]\n",
      "loss:0.363405 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 76.6%, Avg loss: 1.220493 \n",
      "\n",
      "Epoch 22\n",
      "-------------------------------\n",
      "loss:0.174902 [    0/13000]\n",
      "loss:0.082419 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 78.6%, Avg loss: 1.169202 \n",
      "\n",
      "Epoch 23\n",
      "-------------------------------\n",
      "loss:0.103744 [    0/13000]\n",
      "loss:0.153305 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 79.4%, Avg loss: 1.074556 \n",
      "\n",
      "Epoch 24\n",
      "-------------------------------\n",
      "loss:0.071406 [    0/13000]\n",
      "loss:0.134619 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 80.4%, Avg loss: 1.003078 \n",
      "\n",
      "Epoch 25\n",
      "-------------------------------\n",
      "loss:0.050144 [    0/13000]\n",
      "loss:0.064067 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 77.2%, Avg loss: 1.241357 \n",
      "\n",
      "Epoch 26\n",
      "-------------------------------\n",
      "loss:0.102180 [    0/13000]\n",
      "loss:0.125737 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 79.0%, Avg loss: 1.291067 \n",
      "\n",
      "Epoch 27\n",
      "-------------------------------\n",
      "loss:0.068958 [    0/13000]\n",
      "loss:0.036423 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 76.6%, Avg loss: 1.269210 \n",
      "\n",
      "Epoch 28\n",
      "-------------------------------\n",
      "loss:0.047772 [    0/13000]\n",
      "loss:0.213399 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 76.8%, Avg loss: 1.329898 \n",
      "\n",
      "Epoch 29\n",
      "-------------------------------\n",
      "loss:0.052765 [    0/13000]\n",
      "loss:0.130105 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 78.8%, Avg loss: 1.178636 \n",
      "\n",
      "Epoch 30\n",
      "-------------------------------\n",
      "loss:0.107302 [    0/13000]\n",
      "loss:0.124502 [12800/13000]\n",
      "Test Error: \n",
      " Accuracy: 79.2%, Avg loss: 1.157395 \n",
      "\n",
      "Save model: alexnet_imagenet29.pth\n"
     ]
    }
   ],
   "source": [
    "#执行每个opoch\n",
    "train_dataloader =DataLoader(train_dataset,batch_size=128, shuffle=True)\n",
    "test_dataloader = DataLoader(test_dataset,batch_size=128, shuffle=True) \n",
    "\n",
    "alexnet = AlexNet_imagenet().to(device)\n",
    "summary(alexnet, (3,224,224))\n",
    "train_and_test(train_dataloader, test_dataloader, alexnet, 'alexnet_imagenet', epochs=30, learning_rate=1e-3, batch_size=128,imagenet_optimize=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------------------------------------------------\n",
      "        Layer (type)               Output Shape         Param #\n",
      "================================================================\n",
      "            Conv2d-1         [-1, 64, 224, 224]           1,792\n",
      "              ReLU-2         [-1, 64, 224, 224]               0\n",
      "         MaxPool2d-3         [-1, 64, 112, 112]               0\n",
      "            Conv2d-4        [-1, 128, 112, 112]          73,856\n",
      "              ReLU-5        [-1, 128, 112, 112]               0\n",
      "         MaxPool2d-6          [-1, 128, 56, 56]               0\n",
      "            Conv2d-7          [-1, 256, 56, 56]         295,168\n",
      "              ReLU-8          [-1, 256, 56, 56]               0\n",
      "            Conv2d-9          [-1, 256, 56, 56]         590,080\n",
      "             ReLU-10          [-1, 256, 56, 56]               0\n",
      "        MaxPool2d-11          [-1, 256, 28, 28]               0\n",
      "           Conv2d-12          [-1, 512, 28, 28]       1,180,160\n",
      "             ReLU-13          [-1, 512, 28, 28]               0\n",
      "           Conv2d-14          [-1, 512, 28, 28]       2,359,808\n",
      "             ReLU-15          [-1, 512, 28, 28]               0\n",
      "        MaxPool2d-16          [-1, 512, 14, 14]               0\n",
      "           Conv2d-17          [-1, 512, 14, 14]       2,359,808\n",
      "             ReLU-18          [-1, 512, 14, 14]               0\n",
      "           Conv2d-19          [-1, 512, 14, 14]       2,359,808\n",
      "             ReLU-20          [-1, 512, 14, 14]               0\n",
      "        MaxPool2d-21            [-1, 512, 7, 7]               0\n",
      "AdaptiveAvgPool2d-22            [-1, 512, 7, 7]               0\n",
      "           Linear-23                 [-1, 4096]     102,764,544\n",
      "             ReLU-24                 [-1, 4096]               0\n",
      "          Dropout-25                 [-1, 4096]               0\n",
      "           Linear-26                   [-1, 10]          40,970\n",
      "================================================================\n",
      "Total params: 112,025,994\n",
      "Trainable params: 112,025,994\n",
      "Non-trainable params: 0\n",
      "----------------------------------------------------------------\n",
      "Input size (MB): 0.57\n",
      "Forward/backward pass size (MB): 125.27\n",
      "Params size (MB): 427.35\n",
      "Estimated Total Size (MB): 553.19\n",
      "----------------------------------------------------------------\n",
      "Epoch 1\n",
      "-------------------------------\n",
      "loss:2.316895 [    0/13000]\n",
      "loss:2.289169 [ 3200/13000]\n",
      "loss:2.291216 [ 6400/13000]\n",
      "loss:2.227903 [ 9600/13000]\n",
      "loss:2.240573 [12800/13000]\n",
      "loss:2.254447 [12960/13000]\n",
      "Test Error: \n",
      " Accuracy: 27.6%, Avg loss: 2.202906 \n",
      "\n",
      "Epoch 2\n",
      "-------------------------------\n",
      "loss:2.194835 [    0/13000]\n",
      "loss:2.067638 [ 3200/13000]\n",
      "loss:1.766509 [ 6400/13000]\n",
      "loss:1.950108 [ 9600/13000]\n",
      "loss:1.694599 [12800/13000]\n",
      "loss:1.790712 [12960/13000]\n",
      "Test Error: \n",
      " Accuracy: 36.2%, Avg loss: 1.792799 \n",
      "\n",
      "Epoch 3\n",
      "-------------------------------\n",
      "loss:2.006464 [    0/13000]\n",
      "loss:1.639353 [ 3200/13000]\n",
      "loss:1.406296 [ 6400/13000]\n",
      "loss:1.456913 [ 9600/13000]\n",
      "loss:1.736100 [12800/13000]\n",
      "loss:1.635872 [12960/13000]\n",
      "Test Error: \n",
      " Accuracy: 40.4%, Avg loss: 1.601537 \n",
      "\n",
      "Epoch 4\n",
      "-------------------------------\n",
      "loss:1.560057 [    0/13000]\n",
      "loss:1.482667 [ 3200/13000]\n",
      "loss:1.661521 [ 6400/13000]\n",
      "loss:1.410365 [ 9600/13000]\n",
      "loss:1.444065 [12800/13000]\n",
      "loss:1.495144 [12960/13000]\n",
      "Test Error: \n",
      " Accuracy: 48.0%, Avg loss: 1.513824 \n",
      "\n",
      "Epoch 5\n",
      "-------------------------------\n",
      "loss:1.533008 [    0/13000]\n",
      "loss:1.462623 [ 3200/13000]\n",
      "loss:1.384118 [ 6400/13000]\n",
      "loss:1.539192 [ 9600/13000]\n",
      "loss:1.263631 [12800/13000]\n",
      "loss:1.173091 [12960/13000]\n",
      "Test Error: \n",
      " Accuracy: 45.2%, Avg loss: 1.435469 \n",
      "\n",
      "Epoch 6\n",
      "-------------------------------\n",
      "loss:1.266654 [    0/13000]\n",
      "loss:1.181696 [ 3200/13000]\n",
      "loss:1.647480 [ 6400/13000]\n",
      "loss:1.125730 [ 9600/13000]\n",
      "loss:1.294291 [12800/13000]\n",
      "loss:1.324910 [12960/13000]\n",
      "Test Error: \n",
      " Accuracy: 49.2%, Avg loss: 1.520991 \n",
      "\n",
      "Epoch 7\n",
      "-------------------------------\n",
      "loss:1.245773 [    0/13000]\n",
      "loss:1.240380 [ 3200/13000]\n",
      "loss:0.985645 [ 6400/13000]\n",
      "loss:1.522057 [ 9600/13000]\n",
      "loss:1.166717 [12800/13000]\n",
      "loss:0.948333 [12960/13000]\n",
      "Test Error: \n",
      " Accuracy: 53.4%, Avg loss: 1.239860 \n",
      "\n",
      "Epoch 8\n",
      "-------------------------------\n",
      "loss:1.236250 [    0/13000]\n",
      "loss:1.446579 [ 3200/13000]\n",
      "loss:1.150012 [ 6400/13000]\n",
      "loss:1.329874 [ 9600/13000]\n",
      "loss:1.159279 [12800/13000]\n",
      "loss:1.107114 [12960/13000]\n",
      "Test Error: \n",
      " Accuracy: 57.2%, Avg loss: 1.178109 \n",
      "\n",
      "Epoch 9\n",
      "-------------------------------\n",
      "loss:1.177768 [    0/13000]\n",
      "loss:1.374946 [ 3200/13000]\n",
      "loss:1.202853 [ 6400/13000]\n",
      "loss:1.117039 [ 9600/13000]\n",
      "loss:1.302199 [12800/13000]\n",
      "loss:1.173187 [12960/13000]\n",
      "Test Error: \n",
      " Accuracy: 45.8%, Avg loss: 1.580022 \n",
      "\n",
      "Epoch 10\n",
      "-------------------------------\n",
      "loss:1.309307 [    0/13000]\n",
      "loss:0.743344 [ 3200/13000]\n",
      "loss:0.988000 [ 6400/13000]\n",
      "loss:0.985731 [ 9600/13000]\n",
      "loss:1.148625 [12800/13000]\n",
      "loss:0.718772 [12960/13000]\n",
      "Test Error: \n",
      " Accuracy: 60.2%, Avg loss: 1.103012 \n",
      "\n",
      "Epoch 11\n",
      "-------------------------------\n",
      "loss:1.151484 [    0/13000]\n",
      "loss:0.962419 [ 3200/13000]\n",
      "loss:0.952165 [ 6400/13000]\n",
      "loss:0.894913 [ 9600/13000]\n",
      "loss:1.129583 [12800/13000]\n",
      "loss:1.177490 [12960/13000]\n",
      "Test Error: \n",
      " Accuracy: 58.6%, Avg loss: 1.181111 \n",
      "\n",
      "Epoch 12\n",
      "-------------------------------\n",
      "loss:1.666844 [    0/13000]\n",
      "loss:0.733405 [ 3200/13000]\n",
      "loss:0.989765 [ 6400/13000]\n",
      "loss:1.023960 [ 9600/13000]\n",
      "loss:1.005590 [12800/13000]\n",
      "loss:1.020647 [12960/13000]\n",
      "Test Error: \n",
      " Accuracy: 64.2%, Avg loss: 1.021391 \n",
      "\n",
      "Epoch 13\n",
      "-------------------------------\n",
      "loss:1.162194 [    0/13000]\n",
      "loss:0.994110 [ 3200/13000]\n",
      "loss:0.951999 [ 6400/13000]\n",
      "loss:0.798367 [ 9600/13000]\n",
      "loss:1.055797 [12800/13000]\n",
      "loss:1.251306 [12960/13000]\n",
      "Test Error: \n",
      " Accuracy: 54.6%, Avg loss: 1.281452 \n",
      "\n",
      "Epoch 14\n",
      "-------------------------------\n",
      "loss:1.298781 [    0/13000]\n",
      "loss:0.727322 [ 3200/13000]\n",
      "loss:0.803818 [ 6400/13000]\n",
      "loss:0.891624 [ 9600/13000]\n",
      "loss:0.777492 [12800/13000]\n",
      "loss:0.915632 [12960/13000]\n",
      "Test Error: \n",
      " Accuracy: 62.8%, Avg loss: 1.038111 \n",
      "\n",
      "Epoch 15\n",
      "-------------------------------\n",
      "loss:0.944364 [    0/13000]\n",
      "loss:0.673074 [ 3200/13000]\n",
      "loss:0.722817 [ 6400/13000]\n",
      "loss:0.784956 [ 9600/13000]\n",
      "loss:1.009081 [12800/13000]\n",
      "loss:0.782008 [12960/13000]\n",
      "Test Error: \n",
      " Accuracy: 65.8%, Avg loss: 1.000627 \n",
      "\n",
      "Epoch 16\n",
      "-------------------------------\n",
      "loss:0.850723 [    0/13000]\n",
      "loss:0.692887 [ 3200/13000]\n",
      "loss:0.663712 [ 6400/13000]\n",
      "loss:0.463341 [ 9600/13000]\n",
      "loss:0.663627 [12800/13000]\n",
      "loss:1.266891 [12960/13000]\n",
      "Test Error: \n",
      " Accuracy: 69.8%, Avg loss: 0.908677 \n",
      "\n",
      "Epoch 17\n",
      "-------------------------------\n",
      "loss:0.626577 [    0/13000]\n",
      "loss:0.783425 [ 3200/13000]\n",
      "loss:0.790091 [ 6400/13000]\n",
      "loss:0.692255 [ 9600/13000]\n",
      "loss:1.153253 [12800/13000]\n",
      "loss:0.782741 [12960/13000]\n",
      "Test Error: \n",
      " Accuracy: 71.0%, Avg loss: 0.823652 \n",
      "\n",
      "Epoch 18\n",
      "-------------------------------\n",
      "loss:0.700954 [    0/13000]\n",
      "loss:0.636150 [ 3200/13000]\n",
      "loss:0.658354 [ 6400/13000]\n",
      "loss:0.829762 [ 9600/13000]\n",
      "loss:0.795267 [12800/13000]\n",
      "loss:1.017397 [12960/13000]\n",
      "Test Error: \n",
      " Accuracy: 70.0%, Avg loss: 0.878617 \n",
      "\n",
      "Epoch 19\n",
      "-------------------------------\n",
      "loss:0.728357 [    0/13000]\n",
      "loss:0.382855 [ 3200/13000]\n",
      "loss:0.671176 [ 6400/13000]\n",
      "loss:0.533884 [ 9600/13000]\n",
      "loss:1.436427 [12800/13000]\n",
      "loss:0.663236 [12960/13000]\n",
      "Test Error: \n",
      " Accuracy: 57.4%, Avg loss: 1.301054 \n",
      "\n",
      "Epoch 20\n",
      "-------------------------------\n",
      "loss:0.998527 [    0/13000]\n",
      "loss:0.634149 [ 3200/13000]\n",
      "loss:0.565137 [ 6400/13000]\n",
      "loss:0.706074 [ 9600/13000]\n",
      "loss:0.746518 [12800/13000]\n",
      "loss:0.518439 [12960/13000]\n",
      "Test Error: \n",
      " Accuracy: 62.8%, Avg loss: 1.013124 \n",
      "\n",
      "Epoch 21\n",
      "-------------------------------\n",
      "loss:0.548825 [    0/13000]\n",
      "loss:0.619890 [ 3200/13000]\n",
      "loss:0.544492 [ 6400/13000]\n",
      "loss:0.528670 [ 9600/13000]\n",
      "loss:0.607947 [12800/13000]\n",
      "loss:0.866407 [12960/13000]\n",
      "Test Error: \n",
      " Accuracy: 71.2%, Avg loss: 0.842492 \n",
      "\n",
      "Epoch 22\n",
      "-------------------------------\n",
      "loss:0.834128 [    0/13000]\n",
      "loss:0.910422 [ 3200/13000]\n",
      "loss:0.497412 [ 6400/13000]\n",
      "loss:0.325372 [ 9600/13000]\n",
      "loss:0.653620 [12800/13000]\n",
      "loss:0.353601 [12960/13000]\n",
      "Test Error: \n",
      " Accuracy: 74.0%, Avg loss: 0.729258 \n",
      "\n",
      "Epoch 23\n",
      "-------------------------------\n",
      "loss:0.314877 [    0/13000]\n",
      "loss:0.654222 [ 3200/13000]\n",
      "loss:0.868784 [ 6400/13000]\n",
      "loss:0.731647 [ 9600/13000]\n",
      "loss:0.583296 [12800/13000]\n",
      "loss:0.580138 [12960/13000]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test Error: \n",
      " Accuracy: 70.4%, Avg loss: 0.875052 \n",
      "\n",
      "Epoch 24\n",
      "-------------------------------\n",
      "loss:0.577963 [    0/13000]\n",
      "loss:0.393215 [ 3200/13000]\n",
      "loss:0.621959 [ 6400/13000]\n",
      "loss:0.822950 [ 9600/13000]\n",
      "loss:0.840658 [12800/13000]\n",
      "loss:0.628458 [12960/13000]\n",
      "Test Error: \n",
      " Accuracy: 73.6%, Avg loss: 0.773735 \n",
      "\n",
      "Epoch 25\n",
      "-------------------------------\n",
      "loss:0.655472 [    0/13000]\n",
      "loss:0.523085 [ 3200/13000]\n",
      "loss:0.559752 [ 6400/13000]\n",
      "loss:0.790259 [ 9600/13000]\n",
      "loss:0.413693 [12800/13000]\n",
      "loss:0.560887 [12960/13000]\n",
      "Test Error: \n",
      " Accuracy: 72.2%, Avg loss: 0.778700 \n",
      "\n",
      "Epoch 26\n",
      "-------------------------------\n",
      "loss:0.667619 [    0/13000]\n",
      "loss:0.605306 [ 3200/13000]\n",
      "loss:0.491239 [ 6400/13000]\n",
      "loss:0.415008 [ 9600/13000]\n",
      "loss:0.572492 [12800/13000]\n",
      "loss:0.486224 [12960/13000]\n",
      "Test Error: \n",
      " Accuracy: 76.6%, Avg loss: 0.671035 \n",
      "\n",
      "Epoch 27\n",
      "-------------------------------\n",
      "loss:0.720583 [    0/13000]\n",
      "loss:0.576119 [ 3200/13000]\n",
      "loss:0.335572 [ 6400/13000]\n",
      "loss:0.507545 [ 9600/13000]\n",
      "loss:0.662614 [12800/13000]\n",
      "loss:0.820273 [12960/13000]\n",
      "Test Error: \n",
      " Accuracy: 73.2%, Avg loss: 0.775032 \n",
      "\n",
      "Epoch 28\n",
      "-------------------------------\n",
      "loss:0.337502 [    0/13000]\n",
      "loss:0.615361 [ 3200/13000]\n",
      "loss:0.170662 [ 6400/13000]\n",
      "loss:0.483945 [ 9600/13000]\n",
      "loss:0.337543 [12800/13000]\n",
      "loss:0.397089 [12960/13000]\n",
      "Test Error: \n",
      " Accuracy: 75.8%, Avg loss: 0.688458 \n",
      "\n",
      "Epoch 29\n",
      "-------------------------------\n",
      "loss:0.226062 [    0/13000]\n",
      "loss:0.570107 [ 3200/13000]\n",
      "loss:0.222710 [ 6400/13000]\n",
      "loss:1.046861 [ 9600/13000]\n",
      "loss:0.252743 [12800/13000]\n",
      "loss:0.384023 [12960/13000]\n",
      "Test Error: \n",
      " Accuracy: 75.8%, Avg loss: 0.729931 \n",
      "\n",
      "Epoch 30\n",
      "-------------------------------\n",
      "loss:0.122498 [    0/13000]\n",
      "loss:0.443148 [ 3200/13000]\n",
      "loss:0.413206 [ 6400/13000]\n",
      "loss:0.202778 [ 9600/13000]\n",
      "loss:0.563037 [12800/13000]\n",
      "loss:0.312326 [12960/13000]\n",
      "Test Error: \n",
      " Accuracy: 72.0%, Avg loss: 1.052369 \n",
      "\n",
      "Save model: vgg11_imagenet29.pth\n"
     ]
    }
   ],
   "source": [
    "#执行每个opoch\n",
    "train_dataloader =DataLoader(train_dataset,batch_size=32, shuffle=True)\n",
    "test_dataloader = DataLoader(test_dataset,batch_size=32, shuffle=True) \n",
    "vgg11 = VGG11_imagenet().to(device)\n",
    "summary(vgg11, (3,224,224))\n",
    "train_and_test(train_dataloader, test_dataloader, vgg11, 'vgg11_imagenet', epochs=30, learning_rate=1e-3, batch_size=32)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x288 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x288 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from utils import draw_chart\n",
    "draw_chart('alexnet_imagenet', True, True )\n",
    "draw_chart('vgg11_imagenet', True, True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Cifar-10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torchvision\n",
    "from torch.utils.data import DataLoader\n",
    "import torchvision.transforms as transforms\n",
    "from torchsummary import summary\n",
    "import numpy as np\n",
    "from models import train_and_test\n",
    "\n",
    "train_data = torchvision.datasets.CIFAR10(\n",
    "    root=\"./datasets\",#路径\n",
    "    train=True,#训练集\n",
    "    download=False,#是否下载\n",
    "    transform=transforms.ToTensor()#数据预处理\n",
    ")\n",
    "\n",
    "test_data = torchvision.datasets.CIFAR10(\n",
    "    root=\"./datasets\",\n",
    "    train=False,\n",
    "    download=False,\n",
    "    transform=transforms.ToTensor()\n",
    ")\n",
    "\n",
    "train_dataloader = DataLoader(train_data, batch_size=128, shuffle=True)\n",
    "test_dataloader = DataLoader(test_data, batch_size=128, shuffle=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Device:  cuda\n",
      "----------------------------------------------------------------\n",
      "        Layer (type)               Output Shape         Param #\n",
      "================================================================\n",
      "            Conv2d-1           [-1, 64, 16, 16]           1,792\n",
      "              ReLU-2           [-1, 64, 16, 16]               0\n",
      "         MaxPool2d-3             [-1, 64, 7, 7]               0\n",
      "            Conv2d-4            [-1, 192, 7, 7]         307,392\n",
      "              ReLU-5            [-1, 192, 7, 7]               0\n",
      "         MaxPool2d-6            [-1, 192, 3, 3]               0\n",
      "            Conv2d-7            [-1, 384, 3, 3]         663,936\n",
      "              ReLU-8            [-1, 384, 3, 3]               0\n",
      "            Conv2d-9            [-1, 256, 3, 3]         884,992\n",
      "             ReLU-10            [-1, 256, 3, 3]               0\n",
      "           Conv2d-11            [-1, 256, 3, 3]         590,080\n",
      "             ReLU-12            [-1, 256, 3, 3]               0\n",
      "AdaptiveAvgPool2d-13            [-1, 256, 1, 1]               0\n",
      "          Dropout-14                  [-1, 256]               0\n",
      "           Linear-15                  [-1, 256]          65,792\n",
      "             ReLU-16                  [-1, 256]               0\n",
      "          Dropout-17                  [-1, 256]               0\n",
      "           Linear-18                  [-1, 256]          65,792\n",
      "             ReLU-19                  [-1, 256]               0\n",
      "           Linear-20                   [-1, 10]           2,570\n",
      "================================================================\n",
      "Total params: 2,582,346\n",
      "Trainable params: 2,582,346\n",
      "Non-trainable params: 0\n",
      "----------------------------------------------------------------\n",
      "Input size (MB): 0.01\n",
      "Forward/backward pass size (MB): 0.57\n",
      "Params size (MB): 9.85\n",
      "Estimated Total Size (MB): 10.43\n",
      "----------------------------------------------------------------\n",
      "Epoch 1\n",
      "-------------------------------\n",
      "loss:2.302582 [    0/50000]\n",
      "loss:1.974325 [12800/50000]\n",
      "loss:1.949205 [25600/50000]\n",
      "loss:1.836081 [38400/50000]\n",
      "loss:1.928786 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 18.6%, Avg loss: 1.937378 \n",
      "\n",
      "Epoch 2\n",
      "-------------------------------\n",
      "loss:1.947721 [    0/50000]\n",
      "loss:1.868811 [12800/50000]\n",
      "loss:1.840943 [25600/50000]\n",
      "loss:1.827373 [38400/50000]\n",
      "loss:1.773886 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 29.3%, Avg loss: 1.704473 \n",
      "\n",
      "Epoch 3\n",
      "-------------------------------\n",
      "loss:1.597623 [    0/50000]\n",
      "loss:1.625944 [12800/50000]\n",
      "loss:1.646720 [25600/50000]\n",
      "loss:1.451116 [38400/50000]\n",
      "loss:1.466204 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 38.9%, Avg loss: 1.639802 \n",
      "\n",
      "Epoch 4\n",
      "-------------------------------\n",
      "loss:1.627277 [    0/50000]\n",
      "loss:1.387623 [12800/50000]\n",
      "loss:1.447935 [25600/50000]\n",
      "loss:1.445212 [38400/50000]\n",
      "loss:1.316455 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 50.4%, Avg loss: 1.376930 \n",
      "\n",
      "Epoch 5\n",
      "-------------------------------\n",
      "loss:1.194720 [    0/50000]\n",
      "loss:1.353607 [12800/50000]\n",
      "loss:1.156036 [25600/50000]\n",
      "loss:1.172619 [38400/50000]\n",
      "loss:1.195368 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 55.2%, Avg loss: 1.270415 \n",
      "\n",
      "Epoch 6\n",
      "-------------------------------\n",
      "loss:1.179176 [    0/50000]\n",
      "loss:1.204458 [12800/50000]\n",
      "loss:1.077455 [25600/50000]\n",
      "loss:1.135038 [38400/50000]\n",
      "loss:1.034570 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 59.4%, Avg loss: 1.159100 \n",
      "\n",
      "Epoch 7\n",
      "-------------------------------\n",
      "loss:0.986060 [    0/50000]\n",
      "loss:0.962885 [12800/50000]\n",
      "loss:1.126168 [25600/50000]\n",
      "loss:1.003774 [38400/50000]\n",
      "loss:1.149805 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 60.3%, Avg loss: 1.155074 \n",
      "\n",
      "Epoch 8\n",
      "-------------------------------\n",
      "loss:0.968781 [    0/50000]\n",
      "loss:1.103768 [12800/50000]\n",
      "loss:1.208881 [25600/50000]\n",
      "loss:0.837588 [38400/50000]\n",
      "loss:0.916941 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 61.4%, Avg loss: 1.131736 \n",
      "\n",
      "Epoch 9\n",
      "-------------------------------\n",
      "loss:0.902589 [    0/50000]\n",
      "loss:0.921128 [12800/50000]\n",
      "loss:0.865930 [25600/50000]\n",
      "loss:0.744417 [38400/50000]\n",
      "loss:0.972235 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 64.6%, Avg loss: 1.045129 \n",
      "\n",
      "Epoch 10\n",
      "-------------------------------\n",
      "loss:0.871337 [    0/50000]\n",
      "loss:0.848193 [12800/50000]\n",
      "loss:0.787217 [25600/50000]\n",
      "loss:0.714832 [38400/50000]\n",
      "loss:0.886017 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 65.0%, Avg loss: 1.081229 \n",
      "\n",
      "Epoch 11\n",
      "-------------------------------\n",
      "loss:0.766654 [    0/50000]\n",
      "loss:0.868651 [12800/50000]\n",
      "loss:0.839080 [25600/50000]\n",
      "loss:0.834967 [38400/50000]\n",
      "loss:0.871224 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 65.2%, Avg loss: 1.070682 \n",
      "\n",
      "Epoch 12\n",
      "-------------------------------\n",
      "loss:0.651879 [    0/50000]\n",
      "loss:0.756108 [12800/50000]\n",
      "loss:1.035889 [25600/50000]\n",
      "loss:0.853780 [38400/50000]\n",
      "loss:0.656086 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 67.0%, Avg loss: 1.053981 \n",
      "\n",
      "Epoch 13\n",
      "-------------------------------\n",
      "loss:0.727398 [    0/50000]\n",
      "loss:0.601581 [12800/50000]\n",
      "loss:0.698332 [25600/50000]\n",
      "loss:0.819706 [38400/50000]\n",
      "loss:0.606188 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 66.9%, Avg loss: 1.052993 \n",
      "\n",
      "Epoch 14\n",
      "-------------------------------\n",
      "loss:0.753448 [    0/50000]\n",
      "loss:0.574459 [12800/50000]\n",
      "loss:0.674319 [25600/50000]\n",
      "loss:0.797983 [38400/50000]\n",
      "loss:0.677513 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 67.0%, Avg loss: 1.085050 \n",
      "\n",
      "Epoch 15\n",
      "-------------------------------\n",
      "loss:0.551230 [    0/50000]\n",
      "loss:0.650286 [12800/50000]\n",
      "loss:0.630392 [25600/50000]\n",
      "loss:0.361235 [38400/50000]\n",
      "loss:0.466802 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 67.0%, Avg loss: 1.073545 \n",
      "\n",
      "Epoch 16\n",
      "-------------------------------\n",
      "loss:0.535980 [    0/50000]\n",
      "loss:0.364948 [12800/50000]\n",
      "loss:0.463312 [25600/50000]\n",
      "loss:0.487847 [38400/50000]\n",
      "loss:0.745528 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 67.7%, Avg loss: 1.076445 \n",
      "\n",
      "Epoch 17\n",
      "-------------------------------\n",
      "loss:0.385002 [    0/50000]\n",
      "loss:0.434118 [12800/50000]\n",
      "loss:0.451777 [25600/50000]\n",
      "loss:0.632590 [38400/50000]\n",
      "loss:0.431774 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 68.0%, Avg loss: 1.101790 \n",
      "\n",
      "Epoch 18\n",
      "-------------------------------\n",
      "loss:0.489239 [    0/50000]\n",
      "loss:0.445400 [12800/50000]\n",
      "loss:0.626904 [25600/50000]\n",
      "loss:0.584220 [38400/50000]\n",
      "loss:0.329615 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 67.4%, Avg loss: 1.192668 \n",
      "\n",
      "Epoch 19\n",
      "-------------------------------\n",
      "loss:0.306510 [    0/50000]\n",
      "loss:0.389530 [12800/50000]\n",
      "loss:0.456732 [25600/50000]\n",
      "loss:0.562042 [38400/50000]\n",
      "loss:0.608001 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 67.6%, Avg loss: 1.248630 \n",
      "\n",
      "Epoch 20\n",
      "-------------------------------\n",
      "loss:0.432908 [    0/50000]\n",
      "loss:0.481244 [12800/50000]\n",
      "loss:0.405441 [25600/50000]\n",
      "loss:0.411299 [38400/50000]\n",
      "loss:0.308178 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 68.0%, Avg loss: 1.199501 \n",
      "\n",
      "Epoch 21\n",
      "-------------------------------\n",
      "loss:0.389836 [    0/50000]\n",
      "loss:0.393070 [12800/50000]\n",
      "loss:0.359149 [25600/50000]\n",
      "loss:0.275027 [38400/50000]\n",
      "loss:0.478610 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 65.9%, Avg loss: 1.307303 \n",
      "\n",
      "Epoch 22\n",
      "-------------------------------\n",
      "loss:0.448092 [    0/50000]\n",
      "loss:0.390054 [12800/50000]\n",
      "loss:0.309160 [25600/50000]\n",
      "loss:0.322509 [38400/50000]\n",
      "loss:0.332058 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 68.0%, Avg loss: 1.265393 \n",
      "\n",
      "Epoch 23\n",
      "-------------------------------\n",
      "loss:0.365997 [    0/50000]\n",
      "loss:0.279254 [12800/50000]\n",
      "loss:0.333599 [25600/50000]\n",
      "loss:0.334212 [38400/50000]\n",
      "loss:0.395789 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 68.3%, Avg loss: 1.329001 \n",
      "\n",
      "Epoch 24\n",
      "-------------------------------\n",
      "loss:0.338625 [    0/50000]\n",
      "loss:0.161400 [12800/50000]\n",
      "loss:0.390491 [25600/50000]\n",
      "loss:0.361475 [38400/50000]\n",
      "loss:0.335008 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 67.8%, Avg loss: 1.441028 \n",
      "\n",
      "Epoch 25\n",
      "-------------------------------\n",
      "loss:0.336137 [    0/50000]\n",
      "loss:0.193650 [12800/50000]\n",
      "loss:0.368218 [25600/50000]\n",
      "loss:0.315837 [38400/50000]\n",
      "loss:0.306565 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 68.5%, Avg loss: 1.377509 \n",
      "\n",
      "Epoch 26\n",
      "-------------------------------\n",
      "loss:0.256909 [    0/50000]\n",
      "loss:0.361565 [12800/50000]\n",
      "loss:0.280506 [25600/50000]\n",
      "loss:0.257417 [38400/50000]\n",
      "loss:0.253980 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 67.5%, Avg loss: 1.458298 \n",
      "\n",
      "Epoch 27\n",
      "-------------------------------\n",
      "loss:0.202601 [    0/50000]\n",
      "loss:0.305824 [12800/50000]\n",
      "loss:0.291316 [25600/50000]\n",
      "loss:0.356639 [38400/50000]\n",
      "loss:0.180965 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 67.8%, Avg loss: 1.555168 \n",
      "\n",
      "Epoch 28\n",
      "-------------------------------\n",
      "loss:0.156586 [    0/50000]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "loss:0.203390 [12800/50000]\n",
      "loss:0.220629 [25600/50000]\n",
      "loss:0.307743 [38400/50000]\n",
      "loss:0.246757 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 68.0%, Avg loss: 1.521504 \n",
      "\n",
      "Epoch 29\n",
      "-------------------------------\n",
      "loss:0.186487 [    0/50000]\n",
      "loss:0.176693 [12800/50000]\n",
      "loss:0.191266 [25600/50000]\n",
      "loss:0.164460 [38400/50000]\n",
      "loss:0.201349 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 67.2%, Avg loss: 1.597249 \n",
      "\n",
      "Epoch 30\n",
      "-------------------------------\n",
      "loss:0.151232 [    0/50000]\n",
      "loss:0.198122 [12800/50000]\n",
      "loss:0.265353 [25600/50000]\n",
      "loss:0.223604 [38400/50000]\n",
      "loss:0.198151 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 67.8%, Avg loss: 1.680739 \n",
      "\n",
      "Save model: alexnet_cifar29.pth\n"
     ]
    }
   ],
   "source": [
    "from models import AlexNet_cifar\n",
    "device = 'cuda' if torch.cuda.is_available() else 'cpu'\n",
    "print('Device: ',device)\n",
    "alexnet_cifar = AlexNet_cifar(10).to(device)\n",
    "summary(alexnet_cifar, (3,32,32))\n",
    "train_and_test(train_dataloader, test_dataloader, alexnet_cifar, 'alexnet_cifar', epochs=30, learning_rate = 1e-3, batch_size = 128, steps=100,imagenet_optimize=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Device:  cuda\n",
      "----------------------------------------------------------------\n",
      "        Layer (type)               Output Shape         Param #\n",
      "================================================================\n",
      "            Conv2d-1           [-1, 64, 32, 32]           1,792\n",
      "              ReLU-2           [-1, 64, 32, 32]               0\n",
      "         MaxPool2d-3           [-1, 64, 16, 16]               0\n",
      "            Conv2d-4          [-1, 128, 16, 16]          73,856\n",
      "              ReLU-5          [-1, 128, 16, 16]               0\n",
      "         MaxPool2d-6            [-1, 128, 8, 8]               0\n",
      "            Conv2d-7            [-1, 256, 8, 8]         295,168\n",
      "              ReLU-8            [-1, 256, 8, 8]               0\n",
      "            Conv2d-9            [-1, 256, 8, 8]         590,080\n",
      "             ReLU-10            [-1, 256, 8, 8]               0\n",
      "        MaxPool2d-11            [-1, 256, 4, 4]               0\n",
      "           Conv2d-12            [-1, 512, 4, 4]       1,180,160\n",
      "             ReLU-13            [-1, 512, 4, 4]               0\n",
      "           Conv2d-14            [-1, 512, 4, 4]       2,359,808\n",
      "             ReLU-15            [-1, 512, 4, 4]               0\n",
      "        MaxPool2d-16            [-1, 512, 2, 2]               0\n",
      "           Conv2d-17            [-1, 512, 2, 2]       2,359,808\n",
      "             ReLU-18            [-1, 512, 2, 2]               0\n",
      "           Conv2d-19            [-1, 512, 2, 2]       2,359,808\n",
      "             ReLU-20            [-1, 512, 2, 2]               0\n",
      "        MaxPool2d-21            [-1, 512, 1, 1]               0\n",
      "          Flatten-22                  [-1, 512]               0\n",
      "           Linear-23                  [-1, 512]         262,656\n",
      "             ReLU-24                  [-1, 512]               0\n",
      "           Linear-25                   [-1, 10]           5,130\n",
      "================================================================\n",
      "Total params: 9,488,266\n",
      "Trainable params: 9,488,266\n",
      "Non-trainable params: 0\n",
      "----------------------------------------------------------------\n",
      "Input size (MB): 0.01\n",
      "Forward/backward pass size (MB): 2.56\n",
      "Params size (MB): 36.19\n",
      "Estimated Total Size (MB): 38.77\n",
      "----------------------------------------------------------------\n",
      "Epoch 1\n",
      "-------------------------------\n",
      "loss:2.302700 [    0/50000]\n",
      "loss:2.302552 [12800/50000]\n",
      "loss:2.019045 [25600/50000]\n",
      "loss:1.780609 [38400/50000]\n",
      "loss:1.695880 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 30.3%, Avg loss: 1.710253 \n",
      "\n",
      "Epoch 2\n",
      "-------------------------------\n",
      "loss:1.823254 [    0/50000]\n",
      "loss:1.597828 [12800/50000]\n",
      "loss:1.427472 [25600/50000]\n",
      "loss:1.488646 [38400/50000]\n",
      "loss:1.367225 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 49.2%, Avg loss: 1.378527 \n",
      "\n",
      "Epoch 3\n",
      "-------------------------------\n",
      "loss:1.233596 [    0/50000]\n",
      "loss:1.094206 [12800/50000]\n",
      "loss:1.162819 [25600/50000]\n",
      "loss:1.225479 [38400/50000]\n",
      "loss:1.285259 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 60.0%, Avg loss: 1.126915 \n",
      "\n",
      "Epoch 4\n",
      "-------------------------------\n",
      "loss:1.031335 [    0/50000]\n",
      "loss:0.956416 [12800/50000]\n",
      "loss:0.892043 [25600/50000]\n",
      "loss:0.998483 [38400/50000]\n",
      "loss:1.007285 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 63.0%, Avg loss: 1.035214 \n",
      "\n",
      "Epoch 5\n",
      "-------------------------------\n",
      "loss:0.860346 [    0/50000]\n",
      "loss:1.025096 [12800/50000]\n",
      "loss:0.934976 [25600/50000]\n",
      "loss:0.817290 [38400/50000]\n",
      "loss:0.802375 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 68.3%, Avg loss: 0.917265 \n",
      "\n",
      "Epoch 6\n",
      "-------------------------------\n",
      "loss:0.895641 [    0/50000]\n",
      "loss:0.777297 [12800/50000]\n",
      "loss:0.608718 [25600/50000]\n",
      "loss:0.903001 [38400/50000]\n",
      "loss:0.645601 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 68.2%, Avg loss: 0.935122 \n",
      "\n",
      "Epoch 7\n",
      "-------------------------------\n",
      "loss:0.719401 [    0/50000]\n",
      "loss:0.802934 [12800/50000]\n",
      "loss:0.646904 [25600/50000]\n",
      "loss:0.709870 [38400/50000]\n",
      "loss:0.688913 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 70.9%, Avg loss: 0.844999 \n",
      "\n",
      "Epoch 8\n",
      "-------------------------------\n",
      "loss:0.569869 [    0/50000]\n",
      "loss:0.533089 [12800/50000]\n",
      "loss:0.579722 [25600/50000]\n",
      "loss:0.651401 [38400/50000]\n",
      "loss:0.501035 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 72.9%, Avg loss: 0.804947 \n",
      "\n",
      "Epoch 9\n",
      "-------------------------------\n",
      "loss:0.472028 [    0/50000]\n",
      "loss:0.521745 [12800/50000]\n",
      "loss:0.573835 [25600/50000]\n",
      "loss:0.551839 [38400/50000]\n",
      "loss:0.434160 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 73.6%, Avg loss: 0.828592 \n",
      "\n",
      "Epoch 10\n",
      "-------------------------------\n",
      "loss:0.459040 [    0/50000]\n",
      "loss:0.495224 [12800/50000]\n",
      "loss:0.608579 [25600/50000]\n",
      "loss:0.461915 [38400/50000]\n",
      "loss:0.475162 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 73.7%, Avg loss: 0.838052 \n",
      "\n",
      "Epoch 11\n",
      "-------------------------------\n",
      "loss:0.437489 [    0/50000]\n",
      "loss:0.390594 [12800/50000]\n",
      "loss:0.316792 [25600/50000]\n",
      "loss:0.528105 [38400/50000]\n",
      "loss:0.422009 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 74.5%, Avg loss: 0.835914 \n",
      "\n",
      "Epoch 12\n",
      "-------------------------------\n",
      "loss:0.244154 [    0/50000]\n",
      "loss:0.314378 [12800/50000]\n",
      "loss:0.220004 [25600/50000]\n",
      "loss:0.504520 [38400/50000]\n",
      "loss:0.340441 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 74.2%, Avg loss: 0.837945 \n",
      "\n",
      "Epoch 13\n",
      "-------------------------------\n",
      "loss:0.255972 [    0/50000]\n",
      "loss:0.207593 [12800/50000]\n",
      "loss:0.424062 [25600/50000]\n",
      "loss:0.360407 [38400/50000]\n",
      "loss:0.369588 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 74.8%, Avg loss: 0.884156 \n",
      "\n",
      "Epoch 14\n",
      "-------------------------------\n",
      "loss:0.234884 [    0/50000]\n",
      "loss:0.302120 [12800/50000]\n",
      "loss:0.202100 [25600/50000]\n",
      "loss:0.298535 [38400/50000]\n",
      "loss:0.327340 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 74.5%, Avg loss: 0.940823 \n",
      "\n",
      "Epoch 15\n",
      "-------------------------------\n",
      "loss:0.190173 [    0/50000]\n",
      "loss:0.275746 [12800/50000]\n",
      "loss:0.314645 [25600/50000]\n",
      "loss:0.272464 [38400/50000]\n",
      "loss:0.173899 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 74.6%, Avg loss: 1.016739 \n",
      "\n",
      "Epoch 16\n",
      "-------------------------------\n",
      "loss:0.151622 [    0/50000]\n",
      "loss:0.214191 [12800/50000]\n",
      "loss:0.159872 [25600/50000]\n",
      "loss:0.209061 [38400/50000]\n",
      "loss:0.291403 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 73.6%, Avg loss: 0.996347 \n",
      "\n",
      "Epoch 17\n",
      "-------------------------------\n",
      "loss:0.131906 [    0/50000]\n",
      "loss:0.180225 [12800/50000]\n",
      "loss:0.188269 [25600/50000]\n",
      "loss:0.184086 [38400/50000]\n",
      "loss:0.231076 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 72.9%, Avg loss: 1.127337 \n",
      "\n",
      "Epoch 18\n",
      "-------------------------------\n",
      "loss:0.238558 [    0/50000]\n",
      "loss:0.158285 [12800/50000]\n",
      "loss:0.191508 [25600/50000]\n",
      "loss:0.271476 [38400/50000]\n",
      "loss:0.320962 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 74.4%, Avg loss: 1.102030 \n",
      "\n",
      "Epoch 19\n",
      "-------------------------------\n",
      "loss:0.178490 [    0/50000]\n",
      "loss:0.182666 [12800/50000]\n",
      "loss:0.154362 [25600/50000]\n",
      "loss:0.146234 [38400/50000]\n",
      "loss:0.178449 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 73.9%, Avg loss: 1.167703 \n",
      "\n",
      "Epoch 20\n",
      "-------------------------------\n",
      "loss:0.120013 [    0/50000]\n",
      "loss:0.097218 [12800/50000]\n",
      "loss:0.138443 [25600/50000]\n",
      "loss:0.330382 [38400/50000]\n",
      "loss:0.187991 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 74.3%, Avg loss: 1.228258 \n",
      "\n",
      "Epoch 21\n",
      "-------------------------------\n",
      "loss:0.028279 [    0/50000]\n",
      "loss:0.165836 [12800/50000]\n",
      "loss:0.202632 [25600/50000]\n",
      "loss:0.129168 [38400/50000]\n",
      "loss:0.137852 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 74.5%, Avg loss: 1.303510 \n",
      "\n",
      "Epoch 22\n",
      "-------------------------------\n",
      "loss:0.045817 [    0/50000]\n",
      "loss:0.069350 [12800/50000]\n",
      "loss:0.162512 [25600/50000]\n",
      "loss:0.067782 [38400/50000]\n",
      "loss:0.144148 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 73.8%, Avg loss: 1.354896 \n",
      "\n",
      "Epoch 23\n",
      "-------------------------------\n",
      "loss:0.097135 [    0/50000]\n",
      "loss:0.112391 [12800/50000]\n",
      "loss:0.116252 [25600/50000]\n",
      "loss:0.143724 [38400/50000]\n",
      "loss:0.065319 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 74.8%, Avg loss: 1.237190 \n",
      "\n",
      "Epoch 24\n",
      "-------------------------------\n",
      "loss:0.022617 [    0/50000]\n",
      "loss:0.076956 [12800/50000]\n",
      "loss:0.129199 [25600/50000]\n",
      "loss:0.139611 [38400/50000]\n",
      "loss:0.082481 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 74.7%, Avg loss: 1.304804 \n",
      "\n",
      "Epoch 25\n",
      "-------------------------------\n",
      "loss:0.062742 [    0/50000]\n",
      "loss:0.124766 [12800/50000]\n",
      "loss:0.077409 [25600/50000]\n",
      "loss:0.047042 [38400/50000]\n",
      "loss:0.158486 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 74.6%, Avg loss: 1.305015 \n",
      "\n",
      "Epoch 26\n",
      "-------------------------------\n",
      "loss:0.075536 [    0/50000]\n",
      "loss:0.151837 [12800/50000]\n",
      "loss:0.104326 [25600/50000]\n",
      "loss:0.077141 [38400/50000]\n",
      "loss:0.048363 [49792/50000]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test Error: \n",
      " Accuracy: 74.8%, Avg loss: 1.329198 \n",
      "\n",
      "Epoch 27\n",
      "-------------------------------\n",
      "loss:0.028587 [    0/50000]\n",
      "loss:0.129376 [12800/50000]\n",
      "loss:0.137394 [25600/50000]\n",
      "loss:0.089265 [38400/50000]\n",
      "loss:0.193600 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 74.6%, Avg loss: 1.315265 \n",
      "\n",
      "Epoch 28\n",
      "-------------------------------\n",
      "loss:0.064020 [    0/50000]\n",
      "loss:0.018193 [12800/50000]\n",
      "loss:0.106594 [25600/50000]\n",
      "loss:0.148357 [38400/50000]\n",
      "loss:0.017951 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 74.1%, Avg loss: 1.425452 \n",
      "\n",
      "Epoch 29\n",
      "-------------------------------\n",
      "loss:0.064868 [    0/50000]\n",
      "loss:0.121792 [12800/50000]\n",
      "loss:0.085098 [25600/50000]\n",
      "loss:0.065744 [38400/50000]\n",
      "loss:0.057130 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 74.4%, Avg loss: 1.407961 \n",
      "\n",
      "Epoch 30\n",
      "-------------------------------\n",
      "loss:0.102869 [    0/50000]\n",
      "loss:0.070052 [12800/50000]\n",
      "loss:0.026183 [25600/50000]\n",
      "loss:0.193945 [38400/50000]\n",
      "loss:0.128604 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 73.9%, Avg loss: 1.531922 \n",
      "\n",
      "Save model: vgg11_cifar29.pth\n"
     ]
    }
   ],
   "source": [
    "from models import VGG11_cifar\n",
    "device = 'cuda' if torch.cuda.is_available() else 'cpu'\n",
    "print('Device: ',device)\n",
    "vgg11_cifar = VGG11_cifar(10).to(device)\n",
    "summary(vgg11_cifar, (3,32,32))\n",
    "train_and_test(train_dataloader, test_dataloader, vgg11_cifar, 'vgg11_cifar', epochs=30, learning_rate = 1e-3, batch_size = 128, steps=100,imagenet_optimize=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x288 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x288 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from utils import draw_chart\n",
    "draw_chart('alexnet_cifar', True, True)\n",
    "draw_chart('vgg11_cifar', True, True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Load  VGG11_cifar\n",
      "Epoch 1\n",
      "-------------------------------\n",
      "loss:2.301546 [    0/50000]\n",
      "loss:1.063415 [12800/50000]\n",
      "loss:0.825334 [25600/50000]\n",
      "loss:0.856834 [38400/50000]\n",
      "loss:0.978581 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 69.3%, Avg loss: 0.844478 \n",
      "\n",
      "Epoch 2\n",
      "-------------------------------\n",
      "loss:0.856381 [    0/50000]\n",
      "loss:0.753212 [12800/50000]\n",
      "loss:0.636280 [25600/50000]\n",
      "loss:0.616873 [38400/50000]\n",
      "loss:0.608579 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 72.8%, Avg loss: 0.769198 \n",
      "\n",
      "Epoch 3\n",
      "-------------------------------\n",
      "loss:0.660830 [    0/50000]\n",
      "loss:0.619254 [12800/50000]\n",
      "loss:0.872439 [25600/50000]\n",
      "loss:0.589635 [38400/50000]\n",
      "loss:0.577668 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 76.0%, Avg loss: 0.703064 \n",
      "\n",
      "Epoch 4\n",
      "-------------------------------\n",
      "loss:0.578671 [    0/50000]\n",
      "loss:0.509232 [12800/50000]\n",
      "loss:0.588591 [25600/50000]\n",
      "loss:0.332191 [38400/50000]\n",
      "loss:0.483595 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 77.2%, Avg loss: 0.670738 \n",
      "\n",
      "Epoch 5\n",
      "-------------------------------\n",
      "loss:0.314567 [    0/50000]\n",
      "loss:0.428017 [12800/50000]\n",
      "loss:0.672351 [25600/50000]\n",
      "loss:0.439271 [38400/50000]\n",
      "loss:0.484360 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.3%, Avg loss: 0.652542 \n",
      "\n",
      "Epoch 6\n",
      "-------------------------------\n",
      "loss:0.445669 [    0/50000]\n",
      "loss:0.372606 [12800/50000]\n",
      "loss:0.354407 [25600/50000]\n",
      "loss:0.370807 [38400/50000]\n",
      "loss:0.435889 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.6%, Avg loss: 0.651357 \n",
      "\n",
      "Epoch 7\n",
      "-------------------------------\n",
      "loss:0.308611 [    0/50000]\n",
      "loss:0.422215 [12800/50000]\n",
      "loss:0.343612 [25600/50000]\n",
      "loss:0.238330 [38400/50000]\n",
      "loss:0.300561 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 75.7%, Avg loss: 0.771497 \n",
      "\n",
      "Epoch 8\n",
      "-------------------------------\n",
      "loss:0.400443 [    0/50000]\n",
      "loss:0.355189 [12800/50000]\n",
      "loss:0.408654 [25600/50000]\n",
      "loss:0.328716 [38400/50000]\n",
      "loss:0.435645 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.1%, Avg loss: 0.701810 \n",
      "\n",
      "Epoch 9\n",
      "-------------------------------\n",
      "loss:0.283818 [    0/50000]\n",
      "loss:0.165665 [12800/50000]\n",
      "loss:0.254219 [25600/50000]\n",
      "loss:0.251960 [38400/50000]\n",
      "loss:0.319478 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.1%, Avg loss: 0.749768 \n",
      "\n",
      "Epoch 10\n",
      "-------------------------------\n",
      "loss:0.195440 [    0/50000]\n",
      "loss:0.282220 [12800/50000]\n",
      "loss:0.135702 [25600/50000]\n",
      "loss:0.195656 [38400/50000]\n",
      "loss:0.234406 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 79.1%, Avg loss: 0.755054 \n",
      "\n",
      "Epoch 11\n",
      "-------------------------------\n",
      "loss:0.144043 [    0/50000]\n",
      "loss:0.162316 [12800/50000]\n",
      "loss:0.206329 [25600/50000]\n",
      "loss:0.150420 [38400/50000]\n",
      "loss:0.233811 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.9%, Avg loss: 0.783233 \n",
      "\n",
      "Epoch 12\n",
      "-------------------------------\n",
      "loss:0.120195 [    0/50000]\n",
      "loss:0.173583 [12800/50000]\n",
      "loss:0.170988 [25600/50000]\n",
      "loss:0.152655 [38400/50000]\n",
      "loss:0.263672 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.6%, Avg loss: 0.882418 \n",
      "\n",
      "Epoch 13\n",
      "-------------------------------\n",
      "loss:0.123810 [    0/50000]\n",
      "loss:0.140707 [12800/50000]\n",
      "loss:0.098330 [25600/50000]\n",
      "loss:0.138532 [38400/50000]\n",
      "loss:0.127854 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.5%, Avg loss: 0.888328 \n",
      "\n",
      "Epoch 14\n",
      "-------------------------------\n",
      "loss:0.124163 [    0/50000]\n",
      "loss:0.054600 [12800/50000]\n",
      "loss:0.048472 [25600/50000]\n",
      "loss:0.089036 [38400/50000]\n",
      "loss:0.097971 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 79.5%, Avg loss: 0.975950 \n",
      "\n",
      "Epoch 15\n",
      "-------------------------------\n",
      "loss:0.080706 [    0/50000]\n",
      "loss:0.018167 [12800/50000]\n",
      "loss:0.020892 [25600/50000]\n",
      "loss:0.049372 [38400/50000]\n",
      "loss:0.093523 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.0%, Avg loss: 1.030494 \n",
      "\n",
      "Epoch 16\n",
      "-------------------------------\n",
      "loss:0.076907 [    0/50000]\n",
      "loss:0.060114 [12800/50000]\n",
      "loss:0.080687 [25600/50000]\n",
      "loss:0.084874 [38400/50000]\n",
      "loss:0.057069 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 77.6%, Avg loss: 1.088846 \n",
      "\n",
      "Epoch 17\n",
      "-------------------------------\n",
      "loss:0.069087 [    0/50000]\n",
      "loss:0.018900 [12800/50000]\n",
      "loss:0.053812 [25600/50000]\n",
      "loss:0.055163 [38400/50000]\n",
      "loss:0.045076 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.7%, Avg loss: 1.064434 \n",
      "\n",
      "Epoch 18\n",
      "-------------------------------\n",
      "loss:0.031433 [    0/50000]\n",
      "loss:0.034022 [12800/50000]\n",
      "loss:0.061434 [25600/50000]\n",
      "loss:0.047589 [38400/50000]\n",
      "loss:0.076655 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.9%, Avg loss: 1.081228 \n",
      "\n",
      "Epoch 19\n",
      "-------------------------------\n",
      "loss:0.121466 [    0/50000]\n",
      "loss:0.089260 [12800/50000]\n",
      "loss:0.026106 [25600/50000]\n",
      "loss:0.055010 [38400/50000]\n",
      "loss:0.065865 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.6%, Avg loss: 1.128061 \n",
      "\n",
      "Epoch 20\n",
      "-------------------------------\n",
      "loss:0.105744 [    0/50000]\n",
      "loss:0.024778 [12800/50000]\n",
      "loss:0.063466 [25600/50000]\n",
      "loss:0.017454 [38400/50000]\n",
      "loss:0.068304 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.7%, Avg loss: 1.184739 \n",
      "\n",
      "Epoch 21\n",
      "-------------------------------\n",
      "loss:0.026670 [    0/50000]\n",
      "loss:0.053301 [12800/50000]\n",
      "loss:0.048163 [25600/50000]\n",
      "loss:0.006705 [38400/50000]\n",
      "loss:0.018843 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.8%, Avg loss: 1.268462 \n",
      "\n",
      "Epoch 22\n",
      "-------------------------------\n",
      "loss:0.038764 [    0/50000]\n",
      "loss:0.008044 [12800/50000]\n",
      "loss:0.013704 [25600/50000]\n",
      "loss:0.036876 [38400/50000]\n",
      "loss:0.128289 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.1%, Avg loss: 1.153577 \n",
      "\n",
      "Epoch 23\n",
      "-------------------------------\n",
      "loss:0.037784 [    0/50000]\n",
      "loss:0.038980 [12800/50000]\n",
      "loss:0.006758 [25600/50000]\n",
      "loss:0.036939 [38400/50000]\n",
      "loss:0.060280 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 77.6%, Avg loss: 1.298981 \n",
      "\n",
      "Epoch 24\n",
      "-------------------------------\n",
      "loss:0.145343 [    0/50000]\n",
      "loss:0.011702 [12800/50000]\n",
      "loss:0.075629 [25600/50000]\n",
      "loss:0.059459 [38400/50000]\n",
      "loss:0.154270 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.0%, Avg loss: 1.424964 \n",
      "\n",
      "Epoch 25\n",
      "-------------------------------\n",
      "loss:0.063863 [    0/50000]\n",
      "loss:0.019689 [12800/50000]\n",
      "loss:0.007670 [25600/50000]\n",
      "loss:0.006009 [38400/50000]\n",
      "loss:0.060283 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.9%, Avg loss: 1.326516 \n",
      "\n",
      "Epoch 26\n",
      "-------------------------------\n",
      "loss:0.046865 [    0/50000]\n",
      "loss:0.030290 [12800/50000]\n",
      "loss:0.043325 [25600/50000]\n",
      "loss:0.062218 [38400/50000]\n",
      "loss:0.044174 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 79.2%, Avg loss: 1.254367 \n",
      "\n",
      "Epoch 27\n",
      "-------------------------------\n",
      "loss:0.019867 [    0/50000]\n",
      "loss:0.007659 [12800/50000]\n",
      "loss:0.022037 [25600/50000]\n",
      "loss:0.395815 [38400/50000]\n",
      "loss:0.033173 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.3%, Avg loss: 1.200416 \n",
      "\n",
      "Epoch 28\n",
      "-------------------------------\n",
      "loss:0.075184 [    0/50000]\n",
      "loss:0.021171 [12800/50000]\n",
      "loss:0.009726 [25600/50000]\n",
      "loss:0.005476 [38400/50000]\n",
      "loss:0.015718 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.8%, Avg loss: 1.350360 \n",
      "\n",
      "Epoch 29\n",
      "-------------------------------\n",
      "loss:0.026789 [    0/50000]\n",
      "loss:0.028111 [12800/50000]\n",
      "loss:0.050422 [25600/50000]\n",
      "loss:0.195743 [38400/50000]\n",
      "loss:0.075422 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.4%, Avg loss: 1.275153 \n",
      "\n",
      "Epoch 30\n",
      "-------------------------------\n",
      "loss:0.055753 [    0/50000]\n",
      "loss:0.009121 [12800/50000]\n",
      "loss:0.062145 [25600/50000]\n",
      "loss:0.020958 [38400/50000]\n",
      "loss:0.048172 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.3%, Avg loss: 1.393650 \n",
      "\n",
      "Save model: vgg11_cifar_layer_prune29.pth\n"
     ]
    }
   ],
   "source": [
    "from models import load_pretrained_model\n",
    "import torch\n",
    "import torchvision\n",
    "from torch.utils.data import DataLoader\n",
    "import torchvision.transforms as transforms\n",
    "from torchsummary import summary\n",
    "import numpy as np\n",
    "from models import train_and_test\n",
    "\n",
    "train_data = torchvision.datasets.CIFAR10(\n",
    "    root=\"./datasets\",#路径\n",
    "    train=True,#训练集\n",
    "    download=False,#是否下载\n",
    "    transform=transforms.ToTensor()#数据预处理\n",
    ")\n",
    "\n",
    "test_data = torchvision.datasets.CIFAR10(\n",
    "    root=\"./datasets\",\n",
    "    train=False,\n",
    "    download=False,\n",
    "    transform=transforms.ToTensor()\n",
    ")\n",
    "\n",
    "train_dataloader = DataLoader(train_data, batch_size=128, shuffle=True)\n",
    "test_dataloader = DataLoader(test_data, batch_size=128, shuffle=True)\n",
    "\n",
    "class VGG11_cifar_layer_prune(nn.Module):\n",
    "    def __init__(self, num_classes=10, prune_layers=11, pretrained=False):\n",
    "        super(VGG11_cifar_layer_prune, self).__init__()\n",
    "        pretrained_model = load_pretrained_model('vgg11_cifar')\n",
    "        self.features = pretrained_model.features[:prune_layers]\n",
    "        self.avgpool = nn.AdaptiveAvgPool2d((1, 1))\n",
    "        self.flatten = nn.Flatten()\n",
    "        self.classifier = nn.Sequential(\n",
    "            nn.Linear(256, 512),\n",
    "            nn.ReLU(),\n",
    "            nn.Linear(512,num_classes),\n",
    "        )\n",
    "    \n",
    "    def forward(self, x):\n",
    "        x = self.features(x)\n",
    "        x = self.avgpool(x)\n",
    "        x = self.flatten(x)\n",
    "        x = self.classifier(x)\n",
    "        return x\n",
    "\n",
    "prune_model = VGG11_cifar_layer_prune().cuda()\n",
    "train_and_test(train_dataloader, test_dataloader, prune_model, 'vgg11_cifar_layer_prune', epochs=30, learning_rate = 1e-3, batch_size = 128, steps=100,imagenet_optimize=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x288 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from utils import draw_chart\n",
    "draw_chart('vgg11_cifar_layer_prune', True, True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1\n",
      "-------------------------------\n",
      "loss:2.298866 [    0/50000]\n",
      "loss:1.947591 [12800/50000]\n",
      "loss:1.560428 [25600/50000]\n",
      "loss:1.490746 [38400/50000]\n",
      "loss:1.580029 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 44.4%, Avg loss: 1.497961 \n",
      "\n",
      "Epoch 2\n",
      "-------------------------------\n",
      "loss:1.613284 [    0/50000]\n",
      "loss:1.470234 [12800/50000]\n",
      "loss:1.132734 [25600/50000]\n",
      "loss:1.172507 [38400/50000]\n",
      "loss:1.235841 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 55.9%, Avg loss: 1.205309 \n",
      "\n",
      "Epoch 3\n",
      "-------------------------------\n",
      "loss:1.191912 [    0/50000]\n",
      "loss:0.988573 [12800/50000]\n",
      "loss:1.185650 [25600/50000]\n",
      "loss:1.029407 [38400/50000]\n",
      "loss:0.937489 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 59.6%, Avg loss: 1.090807 \n",
      "\n",
      "Epoch 4\n",
      "-------------------------------\n",
      "loss:1.000427 [    0/50000]\n",
      "loss:0.988725 [12800/50000]\n",
      "loss:0.801850 [25600/50000]\n",
      "loss:1.019836 [38400/50000]\n",
      "loss:0.922550 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 66.2%, Avg loss: 0.954536 \n",
      "\n",
      "Epoch 5\n",
      "-------------------------------\n",
      "loss:1.008922 [    0/50000]\n",
      "loss:0.765641 [12800/50000]\n",
      "loss:0.760709 [25600/50000]\n",
      "loss:0.943044 [38400/50000]\n",
      "loss:0.892104 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 67.6%, Avg loss: 0.919841 \n",
      "\n",
      "Epoch 6\n",
      "-------------------------------\n",
      "loss:0.795105 [    0/50000]\n",
      "loss:0.774643 [12800/50000]\n",
      "loss:0.737244 [25600/50000]\n",
      "loss:0.992970 [38400/50000]\n",
      "loss:0.634847 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 71.4%, Avg loss: 0.805670 \n",
      "\n",
      "Epoch 7\n",
      "-------------------------------\n",
      "loss:0.664776 [    0/50000]\n",
      "loss:0.600183 [12800/50000]\n",
      "loss:0.512863 [25600/50000]\n",
      "loss:0.877561 [38400/50000]\n",
      "loss:0.561919 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 73.5%, Avg loss: 0.742352 \n",
      "\n",
      "Epoch 8\n",
      "-------------------------------\n",
      "loss:0.728665 [    0/50000]\n",
      "loss:0.497439 [12800/50000]\n",
      "loss:0.633485 [25600/50000]\n",
      "loss:0.563359 [38400/50000]\n",
      "loss:0.508034 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 76.1%, Avg loss: 0.684549 \n",
      "\n",
      "Epoch 9\n",
      "-------------------------------\n",
      "loss:0.611682 [    0/50000]\n",
      "loss:0.580163 [12800/50000]\n",
      "loss:0.658583 [25600/50000]\n",
      "loss:0.477922 [38400/50000]\n",
      "loss:0.548654 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 75.5%, Avg loss: 0.696030 \n",
      "\n",
      "Epoch 10\n",
      "-------------------------------\n",
      "loss:0.615682 [    0/50000]\n",
      "loss:0.612509 [12800/50000]\n",
      "loss:0.505880 [25600/50000]\n",
      "loss:0.453381 [38400/50000]\n",
      "loss:0.464152 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 77.8%, Avg loss: 0.647021 \n",
      "\n",
      "Epoch 11\n",
      "-------------------------------\n",
      "loss:0.505307 [    0/50000]\n",
      "loss:0.482683 [12800/50000]\n",
      "loss:0.289915 [25600/50000]\n",
      "loss:0.557094 [38400/50000]\n",
      "loss:0.516547 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.0%, Avg loss: 0.638519 \n",
      "\n",
      "Epoch 12\n",
      "-------------------------------\n",
      "loss:0.352501 [    0/50000]\n",
      "loss:0.462225 [12800/50000]\n",
      "loss:0.401677 [25600/50000]\n",
      "loss:0.281888 [38400/50000]\n",
      "loss:0.592284 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.5%, Avg loss: 0.652969 \n",
      "\n",
      "Epoch 13\n",
      "-------------------------------\n",
      "loss:0.310078 [    0/50000]\n",
      "loss:0.271875 [12800/50000]\n",
      "loss:0.312038 [25600/50000]\n",
      "loss:0.413832 [38400/50000]\n",
      "loss:0.319522 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 79.6%, Avg loss: 0.629172 \n",
      "\n",
      "Epoch 14\n",
      "-------------------------------\n",
      "loss:0.306727 [    0/50000]\n",
      "loss:0.286845 [12800/50000]\n",
      "loss:0.232992 [25600/50000]\n",
      "loss:0.376547 [38400/50000]\n",
      "loss:0.263295 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 79.1%, Avg loss: 0.657114 \n",
      "\n",
      "Epoch 15\n",
      "-------------------------------\n",
      "loss:0.279670 [    0/50000]\n",
      "loss:0.252709 [12800/50000]\n",
      "loss:0.324668 [25600/50000]\n",
      "loss:0.165374 [38400/50000]\n",
      "loss:0.196265 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.8%, Avg loss: 0.709021 \n",
      "\n",
      "Epoch 16\n",
      "-------------------------------\n",
      "loss:0.246380 [    0/50000]\n",
      "loss:0.271121 [12800/50000]\n",
      "loss:0.148665 [25600/50000]\n",
      "loss:0.203725 [38400/50000]\n",
      "loss:0.214450 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 79.2%, Avg loss: 0.713299 \n",
      "\n",
      "Epoch 17\n",
      "-------------------------------\n",
      "loss:0.136993 [    0/50000]\n",
      "loss:0.124188 [12800/50000]\n",
      "loss:0.163623 [25600/50000]\n",
      "loss:0.159793 [38400/50000]\n",
      "loss:0.310066 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.6%, Avg loss: 0.780664 \n",
      "\n",
      "Epoch 18\n",
      "-------------------------------\n",
      "loss:0.171066 [    0/50000]\n",
      "loss:0.180695 [12800/50000]\n",
      "loss:0.160889 [25600/50000]\n",
      "loss:0.177848 [38400/50000]\n",
      "loss:0.151691 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 80.0%, Avg loss: 0.762978 \n",
      "\n",
      "Epoch 19\n",
      "-------------------------------\n",
      "loss:0.114805 [    0/50000]\n",
      "loss:0.105686 [12800/50000]\n",
      "loss:0.145907 [25600/50000]\n",
      "loss:0.068033 [38400/50000]\n",
      "loss:0.140511 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.6%, Avg loss: 0.862796 \n",
      "\n",
      "Epoch 20\n",
      "-------------------------------\n",
      "loss:0.085688 [    0/50000]\n",
      "loss:0.070624 [12800/50000]\n",
      "loss:0.043999 [25600/50000]\n",
      "loss:0.162259 [38400/50000]\n",
      "loss:0.084461 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 79.4%, Avg loss: 0.910264 \n",
      "\n",
      "Epoch 21\n",
      "-------------------------------\n",
      "loss:0.061999 [    0/50000]\n",
      "loss:0.057613 [12800/50000]\n",
      "loss:0.127043 [25600/50000]\n",
      "loss:0.216935 [38400/50000]\n",
      "loss:0.059344 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.6%, Avg loss: 0.980875 \n",
      "\n",
      "Epoch 22\n",
      "-------------------------------\n",
      "loss:0.159197 [    0/50000]\n",
      "loss:0.041054 [12800/50000]\n",
      "loss:0.053972 [25600/50000]\n",
      "loss:0.110130 [38400/50000]\n",
      "loss:0.047013 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.7%, Avg loss: 1.015326 \n",
      "\n",
      "Epoch 23\n",
      "-------------------------------\n",
      "loss:0.072947 [    0/50000]\n",
      "loss:0.103213 [12800/50000]\n",
      "loss:0.037163 [25600/50000]\n",
      "loss:0.079522 [38400/50000]\n",
      "loss:0.089415 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.9%, Avg loss: 0.970725 \n",
      "\n",
      "Epoch 24\n",
      "-------------------------------\n",
      "loss:0.056031 [    0/50000]\n",
      "loss:0.113507 [12800/50000]\n",
      "loss:0.123370 [25600/50000]\n",
      "loss:0.061459 [38400/50000]\n",
      "loss:0.078599 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.0%, Avg loss: 1.079220 \n",
      "\n",
      "Epoch 25\n",
      "-------------------------------\n",
      "loss:0.124868 [    0/50000]\n",
      "loss:0.070345 [12800/50000]\n",
      "loss:0.065460 [25600/50000]\n",
      "loss:0.036437 [38400/50000]\n",
      "loss:0.116547 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 79.2%, Avg loss: 1.050896 \n",
      "\n",
      "Epoch 26\n",
      "-------------------------------\n",
      "loss:0.016383 [    0/50000]\n",
      "loss:0.028390 [12800/50000]\n",
      "loss:0.076063 [25600/50000]\n",
      "loss:0.043608 [38400/50000]\n",
      "loss:0.079887 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.5%, Avg loss: 1.096975 \n",
      "\n",
      "Epoch 27\n",
      "-------------------------------\n",
      "loss:0.039452 [    0/50000]\n",
      "loss:0.052508 [12800/50000]\n",
      "loss:0.212144 [25600/50000]\n",
      "loss:0.012892 [38400/50000]\n",
      "loss:0.117705 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 79.2%, Avg loss: 1.045526 \n",
      "\n",
      "Epoch 28\n",
      "-------------------------------\n",
      "loss:0.059098 [    0/50000]\n",
      "loss:0.025913 [12800/50000]\n",
      "loss:0.055378 [25600/50000]\n",
      "loss:0.074848 [38400/50000]\n",
      "loss:0.049936 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.2%, Avg loss: 1.309517 \n",
      "\n",
      "Epoch 29\n",
      "-------------------------------\n",
      "loss:0.073656 [    0/50000]\n",
      "loss:0.071660 [12800/50000]\n",
      "loss:0.045453 [25600/50000]\n",
      "loss:0.072111 [38400/50000]\n",
      "loss:0.075943 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 79.1%, Avg loss: 1.107940 \n",
      "\n",
      "Epoch 30\n",
      "-------------------------------\n",
      "loss:0.025035 [    0/50000]\n",
      "loss:0.037908 [12800/50000]\n",
      "loss:0.046529 [25600/50000]\n",
      "loss:0.025569 [38400/50000]\n",
      "loss:0.045280 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 79.1%, Avg loss: 1.199071 \n",
      "\n",
      "Save model: vgg11_cifar_layer_prune_129.pth\n",
      "Test Accuracy = 0.791 Class Accuracy: [0.786 0.887 0.694 0.671 0.748 0.722 0.832 0.785 0.917 0.872]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(0.7914,\n",
       " array([0.786, 0.887, 0.694, 0.671, 0.748, 0.722, 0.832, 0.785, 0.917,\n",
       "        0.872]))"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from models import load_pretrained_model\n",
    "import torch\n",
    "import torchvision\n",
    "from torch.utils.data import DataLoader\n",
    "import torchvision.transforms as transforms\n",
    "from torchsummary import summary\n",
    "import numpy as np\n",
    "from models import train_and_test\n",
    "\n",
    "train_data = torchvision.datasets.CIFAR10(\n",
    "    root=\"./datasets\",#路径\n",
    "    train=True,#训练集\n",
    "    download=False,#是否下载\n",
    "    transform=transforms.ToTensor()#数据预处理\n",
    ")\n",
    "\n",
    "test_data = torchvision.datasets.CIFAR10(\n",
    "    root=\"./datasets\",\n",
    "    train=False,\n",
    "    download=False,\n",
    "    transform=transforms.ToTensor()\n",
    ")\n",
    "def test(model, test_loader, show=True):\n",
    "    model.eval()\n",
    "    correct = 0\n",
    "    acc = 0\n",
    "    cls_num = 10\n",
    "    cls_acc = np.zeros(cls_num)\n",
    "    size = np.zeros(cls_num)\n",
    "    with torch.no_grad():\n",
    "        for data, target in test_loader:\n",
    "            data, target = data.cuda(), target.cuda()\n",
    "            output = model(data)\n",
    "            correct += (output.argmax(1)==target).type(torch.float).sum().item()\n",
    "\n",
    "            for i in range(data.shape[0]):\n",
    "                    size[target[i]] += 1\n",
    "                    if (output.argmax(1)==target)[i]:\n",
    "                        cls_acc[target[i]]+=1\n",
    "        \n",
    "    acc = correct / size.sum()\n",
    "    cls_acc = cls_acc/size\n",
    "    if show:\n",
    "        print(\"Test Accuracy = {0:.3f} Class Accuracy: {1}\".format(acc, np.round(cls_acc,3)))\n",
    "    return acc, cls_acc\n",
    "train_dataloader = DataLoader(train_data, batch_size=128, shuffle=True)\n",
    "test_dataloader = DataLoader(test_data, batch_size=128, shuffle=True)\n",
    "\n",
    "class VGG11_cifar_layer_prune(nn.Module):\n",
    "    def __init__(self, num_classes=10, prune_layers=11, pretrained=False):\n",
    "        super(VGG11_cifar_layer_prune, self).__init__()\n",
    "        self.features = torchvision.models.vgg11(pretrained).features[:prune_layers]\n",
    "        self.avgpool = nn.AdaptiveAvgPool2d((1, 1))\n",
    "        self.flatten = nn.Flatten()\n",
    "        self.classifier = nn.Sequential(\n",
    "            nn.Linear(256, 512),\n",
    "            nn.ReLU(),\n",
    "            nn.Linear(512,num_classes),\n",
    "        )\n",
    "    \n",
    "    def forward(self, x):\n",
    "        x = self.features(x)\n",
    "        x = self.avgpool(x)\n",
    "        x = self.flatten(x)\n",
    "        x = self.classifier(x)\n",
    "        return x\n",
    "\n",
    "prune_model = VGG11_cifar_layer_prune().cuda()\n",
    "train_and_test(train_dataloader, test_dataloader, prune_model, 'vgg11_cifar_layer_prune_1', epochs=30, learning_rate = 1e-3, batch_size = 128, steps=100,imagenet_optimize=False)\n",
    "test(prune_model, test_dataloader)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x288 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from utils import draw_chart\n",
    "draw_chart('vgg11_cifar_layer_prune_1', True, True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------------------------------------------------\n",
      "        Layer (type)               Output Shape         Param #\n",
      "================================================================\n",
      "            Conv2d-1           [-1, 64, 32, 32]           1,792\n",
      "              ReLU-2           [-1, 64, 32, 32]               0\n",
      "         MaxPool2d-3           [-1, 64, 16, 16]               0\n",
      "            Conv2d-4          [-1, 128, 16, 16]          73,856\n",
      "              ReLU-5          [-1, 128, 16, 16]               0\n",
      "         MaxPool2d-6            [-1, 128, 8, 8]               0\n",
      "            Conv2d-7            [-1, 256, 8, 8]         295,168\n",
      "              ReLU-8            [-1, 256, 8, 8]               0\n",
      "            Conv2d-9            [-1, 256, 8, 8]         590,080\n",
      "             ReLU-10            [-1, 256, 8, 8]               0\n",
      "        MaxPool2d-11            [-1, 256, 4, 4]               0\n",
      "AdaptiveAvgPool2d-12            [-1, 256, 1, 1]               0\n",
      "          Flatten-13                  [-1, 256]               0\n",
      "           Linear-14                  [-1, 512]         131,584\n",
      "             ReLU-15                  [-1, 512]               0\n",
      "           Linear-16                   [-1, 10]           5,130\n",
      "================================================================\n",
      "Total params: 1,097,610\n",
      "Trainable params: 1,097,610\n",
      "Non-trainable params: 0\n",
      "----------------------------------------------------------------\n",
      "Input size (MB): 0.01\n",
      "Forward/backward pass size (MB): 2.23\n",
      "Params size (MB): 4.19\n",
      "Estimated Total Size (MB): 6.43\n",
      "----------------------------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "from torchsummary import summary\n",
    "summary(prune_model, (3,32,32))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------------------------------------------------\n",
      "        Layer (type)               Output Shape         Param #\n",
      "================================================================\n",
      "            Conv2d-1           [-1, 64, 32, 32]           1,792\n",
      "              ReLU-2           [-1, 64, 32, 32]               0\n",
      "         MaxPool2d-3           [-1, 64, 16, 16]               0\n",
      "            Conv2d-4          [-1, 128, 16, 16]          73,856\n",
      "              ReLU-5          [-1, 128, 16, 16]               0\n",
      "         MaxPool2d-6            [-1, 128, 8, 8]               0\n",
      "            Conv2d-7            [-1, 256, 8, 8]         295,168\n",
      "              ReLU-8            [-1, 256, 8, 8]               0\n",
      "            Conv2d-9            [-1, 256, 8, 8]         590,080\n",
      "             ReLU-10            [-1, 256, 8, 8]               0\n",
      "        MaxPool2d-11            [-1, 256, 4, 4]               0\n",
      "           Conv2d-12            [-1, 512, 4, 4]       1,180,160\n",
      "             ReLU-13            [-1, 512, 4, 4]               0\n",
      "AdaptiveAvgPool2d-14            [-1, 512, 1, 1]               0\n",
      "          Flatten-15                  [-1, 512]               0\n",
      "           Linear-16                  [-1, 512]         262,656\n",
      "             ReLU-17                  [-1, 512]               0\n",
      "           Linear-18                   [-1, 10]           5,130\n",
      "================================================================\n",
      "Total params: 2,408,842\n",
      "Trainable params: 2,408,842\n",
      "Non-trainable params: 0\n",
      "----------------------------------------------------------------\n",
      "Input size (MB): 0.01\n",
      "Forward/backward pass size (MB): 2.36\n",
      "Params size (MB): 9.19\n",
      "Estimated Total Size (MB): 11.56\n",
      "----------------------------------------------------------------\n",
      "Epoch 1\n",
      "-------------------------------\n",
      "loss:2.299355 [    0/50000]\n",
      "loss:1.804294 [12800/50000]\n",
      "loss:1.551548 [25600/50000]\n",
      "loss:1.562620 [38400/50000]\n",
      "loss:1.347544 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 45.8%, Avg loss: 1.449232 \n",
      "\n",
      "Epoch 2\n",
      "-------------------------------\n",
      "loss:1.441966 [    0/50000]\n",
      "loss:1.257958 [12800/50000]\n",
      "loss:1.143923 [25600/50000]\n",
      "loss:0.957716 [38400/50000]\n",
      "loss:1.005888 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 61.0%, Avg loss: 1.082830 \n",
      "\n",
      "Epoch 3\n",
      "-------------------------------\n",
      "loss:1.042224 [    0/50000]\n",
      "loss:0.954768 [12800/50000]\n",
      "loss:1.028789 [25600/50000]\n",
      "loss:0.903278 [38400/50000]\n",
      "loss:0.724975 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 66.3%, Avg loss: 0.934457 \n",
      "\n",
      "Epoch 4\n",
      "-------------------------------\n",
      "loss:0.970118 [    0/50000]\n",
      "loss:0.927808 [12800/50000]\n",
      "loss:0.911181 [25600/50000]\n",
      "loss:0.887371 [38400/50000]\n",
      "loss:0.773078 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 70.0%, Avg loss: 0.834808 \n",
      "\n",
      "Epoch 5\n",
      "-------------------------------\n",
      "loss:0.711706 [    0/50000]\n",
      "loss:0.589253 [12800/50000]\n",
      "loss:0.698292 [25600/50000]\n",
      "loss:0.569027 [38400/50000]\n",
      "loss:0.539509 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 73.5%, Avg loss: 0.774780 \n",
      "\n",
      "Epoch 6\n",
      "-------------------------------\n",
      "loss:0.696248 [    0/50000]\n",
      "loss:0.604436 [12800/50000]\n",
      "loss:0.659319 [25600/50000]\n",
      "loss:0.848349 [38400/50000]\n",
      "loss:0.546284 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 75.6%, Avg loss: 0.700120 \n",
      "\n",
      "Epoch 7\n",
      "-------------------------------\n",
      "loss:0.540687 [    0/50000]\n",
      "loss:0.385382 [12800/50000]\n",
      "loss:0.435521 [25600/50000]\n",
      "loss:0.503147 [38400/50000]\n",
      "loss:0.670544 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.0%, Avg loss: 0.652793 \n",
      "\n",
      "Epoch 8\n",
      "-------------------------------\n",
      "loss:0.271995 [    0/50000]\n",
      "loss:0.400785 [12800/50000]\n",
      "loss:0.494995 [25600/50000]\n",
      "loss:0.571368 [38400/50000]\n",
      "loss:0.477589 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.4%, Avg loss: 0.650174 \n",
      "\n",
      "Epoch 9\n",
      "-------------------------------\n",
      "loss:0.332041 [    0/50000]\n",
      "loss:0.312033 [12800/50000]\n",
      "loss:0.333024 [25600/50000]\n",
      "loss:0.474947 [38400/50000]\n",
      "loss:0.275948 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 77.8%, Avg loss: 0.739332 \n",
      "\n",
      "Epoch 10\n",
      "-------------------------------\n",
      "loss:0.208810 [    0/50000]\n",
      "loss:0.248602 [12800/50000]\n",
      "loss:0.234674 [25600/50000]\n",
      "loss:0.498082 [38400/50000]\n",
      "loss:0.284283 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 75.5%, Avg loss: 0.794027 \n",
      "\n",
      "Epoch 11\n",
      "-------------------------------\n",
      "loss:0.252930 [    0/50000]\n",
      "loss:0.303882 [12800/50000]\n",
      "loss:0.207278 [25600/50000]\n",
      "loss:0.191519 [38400/50000]\n",
      "loss:0.249311 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.7%, Avg loss: 0.753491 \n",
      "\n",
      "Epoch 12\n",
      "-------------------------------\n",
      "loss:0.112261 [    0/50000]\n",
      "loss:0.094815 [12800/50000]\n",
      "loss:0.188561 [25600/50000]\n",
      "loss:0.243940 [38400/50000]\n",
      "loss:0.180004 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.2%, Avg loss: 0.860574 \n",
      "\n",
      "Epoch 13\n",
      "-------------------------------\n",
      "loss:0.150321 [    0/50000]\n",
      "loss:0.154467 [12800/50000]\n",
      "loss:0.197645 [25600/50000]\n",
      "loss:0.155731 [38400/50000]\n",
      "loss:0.253986 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 77.6%, Avg loss: 0.848482 \n",
      "\n",
      "Epoch 14\n",
      "-------------------------------\n",
      "loss:0.142118 [    0/50000]\n",
      "loss:0.119313 [12800/50000]\n",
      "loss:0.102695 [25600/50000]\n",
      "loss:0.234265 [38400/50000]\n",
      "loss:0.056652 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.6%, Avg loss: 0.942817 \n",
      "\n",
      "Epoch 15\n",
      "-------------------------------\n",
      "loss:0.051268 [    0/50000]\n",
      "loss:0.031285 [12800/50000]\n",
      "loss:0.095803 [25600/50000]\n",
      "loss:0.266561 [38400/50000]\n",
      "loss:0.235618 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 77.3%, Avg loss: 1.005630 \n",
      "\n",
      "Epoch 16\n",
      "-------------------------------\n",
      "loss:0.093359 [    0/50000]\n",
      "loss:0.050450 [12800/50000]\n",
      "loss:0.073978 [25600/50000]\n",
      "loss:0.143811 [38400/50000]\n",
      "loss:0.063589 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 77.6%, Avg loss: 1.121127 \n",
      "\n",
      "Epoch 17\n",
      "-------------------------------\n",
      "loss:0.032735 [    0/50000]\n",
      "loss:0.040049 [12800/50000]\n",
      "loss:0.052137 [25600/50000]\n",
      "loss:0.039280 [38400/50000]\n",
      "loss:0.138437 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 77.8%, Avg loss: 1.123243 \n",
      "\n",
      "Epoch 18\n",
      "-------------------------------\n",
      "loss:0.038392 [    0/50000]\n",
      "loss:0.027496 [12800/50000]\n",
      "loss:0.036157 [25600/50000]\n",
      "loss:0.062916 [38400/50000]\n",
      "loss:0.066509 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 77.2%, Avg loss: 1.103796 \n",
      "\n",
      "Epoch 19\n",
      "-------------------------------\n",
      "loss:0.070744 [    0/50000]\n",
      "loss:0.082589 [12800/50000]\n",
      "loss:0.049113 [25600/50000]\n",
      "loss:0.032306 [38400/50000]\n",
      "loss:0.219785 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 76.4%, Avg loss: 1.146883 \n",
      "\n",
      "Epoch 20\n",
      "-------------------------------\n",
      "loss:0.072997 [    0/50000]\n",
      "loss:0.076858 [12800/50000]\n",
      "loss:0.113278 [25600/50000]\n",
      "loss:0.087328 [38400/50000]\n",
      "loss:0.059271 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 77.5%, Avg loss: 1.233268 \n",
      "\n",
      "Epoch 21\n",
      "-------------------------------\n",
      "loss:0.053671 [    0/50000]\n",
      "loss:0.054767 [12800/50000]\n",
      "loss:0.089795 [25600/50000]\n",
      "loss:0.045324 [38400/50000]\n",
      "loss:0.035856 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 77.8%, Avg loss: 1.233071 \n",
      "\n",
      "Epoch 22\n",
      "-------------------------------\n",
      "loss:0.048746 [    0/50000]\n",
      "loss:0.067733 [12800/50000]\n",
      "loss:0.034098 [25600/50000]\n",
      "loss:0.060858 [38400/50000]\n",
      "loss:0.041147 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 77.7%, Avg loss: 1.194829 \n",
      "\n",
      "Epoch 23\n",
      "-------------------------------\n",
      "loss:0.023651 [    0/50000]\n",
      "loss:0.181658 [12800/50000]\n",
      "loss:0.042446 [25600/50000]\n",
      "loss:0.025138 [38400/50000]\n",
      "loss:0.012544 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.8%, Avg loss: 1.256415 \n",
      "\n",
      "Epoch 24\n",
      "-------------------------------\n",
      "loss:0.035287 [    0/50000]\n",
      "loss:0.052399 [12800/50000]\n",
      "loss:0.018063 [25600/50000]\n",
      "loss:0.048073 [38400/50000]\n",
      "loss:0.075882 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 77.2%, Avg loss: 1.265441 \n",
      "\n",
      "Epoch 25\n",
      "-------------------------------\n",
      "loss:0.027760 [    0/50000]\n",
      "loss:0.025936 [12800/50000]\n",
      "loss:0.014268 [25600/50000]\n",
      "loss:0.035642 [38400/50000]\n",
      "loss:0.020454 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 78.0%, Avg loss: 1.107750 \n",
      "\n",
      "Epoch 26\n",
      "-------------------------------\n",
      "loss:0.043437 [    0/50000]\n",
      "loss:0.034082 [12800/50000]\n",
      "loss:0.009012 [25600/50000]\n",
      "loss:0.023076 [38400/50000]\n",
      "loss:0.033466 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 77.7%, Avg loss: 1.232866 \n",
      "\n",
      "Epoch 27\n",
      "-------------------------------\n",
      "loss:0.020699 [    0/50000]\n",
      "loss:0.077521 [12800/50000]\n",
      "loss:0.020862 [25600/50000]\n",
      "loss:0.025916 [38400/50000]\n",
      "loss:0.063775 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 77.9%, Avg loss: 1.221418 \n",
      "\n",
      "Epoch 28\n",
      "-------------------------------\n",
      "loss:0.011559 [    0/50000]\n",
      "loss:0.077396 [12800/50000]\n",
      "loss:0.077153 [25600/50000]\n",
      "loss:0.018812 [38400/50000]\n",
      "loss:0.042853 [49792/50000]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test Error: \n",
      " Accuracy: 76.8%, Avg loss: 1.212481 \n",
      "\n",
      "Epoch 29\n",
      "-------------------------------\n",
      "loss:0.065969 [    0/50000]\n",
      "loss:0.032525 [12800/50000]\n",
      "loss:0.027884 [25600/50000]\n",
      "loss:0.042386 [38400/50000]\n",
      "loss:0.101375 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 77.6%, Avg loss: 1.281452 \n",
      "\n",
      "Epoch 30\n",
      "-------------------------------\n",
      "loss:0.085128 [    0/50000]\n",
      "loss:0.021367 [12800/50000]\n",
      "loss:0.026189 [25600/50000]\n",
      "loss:0.030419 [38400/50000]\n",
      "loss:0.018185 [49792/50000]\n",
      "Test Error: \n",
      " Accuracy: 77.3%, Avg loss: 1.246867 \n",
      "\n",
      "Save model: vgg11_cifar_layer_prune_229.pth\n",
      "Test Accuracy = 0.773 Class Accuracy: [0.845 0.925 0.759 0.564 0.688 0.726 0.906 0.725 0.844 0.749]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(0.7731,\n",
       " array([0.845, 0.925, 0.759, 0.564, 0.688, 0.726, 0.906, 0.725, 0.844,\n",
       "        0.749]))"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from models import load_pretrained_model\n",
    "import torch\n",
    "import torchvision\n",
    "from torch.utils.data import DataLoader\n",
    "import torchvision.transforms as transforms\n",
    "from torchsummary import summary\n",
    "import numpy as np\n",
    "from models import train_and_test\n",
    "\n",
    "train_data = torchvision.datasets.CIFAR10(\n",
    "    root=\"./datasets\",#路径\n",
    "    train=True,#训练集\n",
    "    download=False,#是否下载\n",
    "    transform=transforms.ToTensor()#数据预处理\n",
    ")\n",
    "\n",
    "test_data = torchvision.datasets.CIFAR10(\n",
    "    root=\"./datasets\",\n",
    "    train=False,\n",
    "    download=False,\n",
    "    transform=transforms.ToTensor()\n",
    ")\n",
    "def test(model, test_loader, show=True):\n",
    "    model.eval()\n",
    "    correct = 0\n",
    "    acc = 0\n",
    "    cls_num = 10\n",
    "    cls_acc = np.zeros(cls_num)\n",
    "    size = np.zeros(cls_num)\n",
    "    with torch.no_grad():\n",
    "        for data, target in test_loader:\n",
    "            data, target = data.cuda(), target.cuda()\n",
    "            output = model(data)\n",
    "            correct += (output.argmax(1)==target).type(torch.float).sum().item()\n",
    "\n",
    "            for i in range(data.shape[0]):\n",
    "                    size[target[i]] += 1\n",
    "                    if (output.argmax(1)==target)[i]:\n",
    "                        cls_acc[target[i]]+=1\n",
    "        \n",
    "    acc = correct / size.sum()\n",
    "    cls_acc = cls_acc/size\n",
    "    if show:\n",
    "        print(\"Test Accuracy = {0:.3f} Class Accuracy: {1}\".format(acc, np.round(cls_acc,3)))\n",
    "    return acc, cls_acc\n",
    "train_dataloader = DataLoader(train_data, batch_size=128, shuffle=True)\n",
    "test_dataloader = DataLoader(test_data, batch_size=128, shuffle=True)\n",
    "\n",
    "class VGG11_cifar_layer_prune(nn.Module):\n",
    "    def __init__(self, num_classes=10, prune_layers=13, pretrained=False):\n",
    "        super(VGG11_cifar_layer_prune, self).__init__()\n",
    "        self.features = torchvision.models.vgg11(pretrained).features[:prune_layers]\n",
    "        self.avgpool = nn.AdaptiveAvgPool2d((1, 1))\n",
    "        self.flatten = nn.Flatten()\n",
    "        self.classifier = nn.Sequential(\n",
    "            nn.Linear(512, 512),\n",
    "            nn.ReLU(),\n",
    "            nn.Linear(512,num_classes),\n",
    "        )\n",
    "    \n",
    "    def forward(self, x):\n",
    "        x = self.features(x)\n",
    "        x = self.avgpool(x)\n",
    "        x = self.flatten(x)\n",
    "        x = self.classifier(x)\n",
    "        return x\n",
    "\n",
    "prune_model = VGG11_cifar_layer_prune().cuda()\n",
    "summary(prune_model, (3,32,32))\n",
    "train_and_test(train_dataloader, test_dataloader, prune_model, 'vgg11_cifar_layer_prune_2', epochs=30, learning_rate = 1e-3, batch_size = 128, steps=100,imagenet_optimize=False)\n",
    "test(prune_model, test_dataloader)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x288 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from utils import draw_chart\n",
    "draw_chart('vgg11_cifar_layer_prune_2', True, True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# MNIST"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torchvision\n",
    "from torch.utils.data import DataLoader\n",
    "import torchvision.transforms as transforms\n",
    "from torchsummary import summary\n",
    "import numpy as np\n",
    "from models import train_and_test\n",
    "\n",
    "train_data = torchvision.datasets.MNIST(\n",
    "    root=\"./datasets\",#路径\n",
    "    train=True,#训练集\n",
    "    download=False,#是否下载\n",
    "    transform=transforms.ToTensor()#数据预处理\n",
    ")\n",
    "\n",
    "test_data = torchvision.datasets.MNIST(\n",
    "    root=\"./datasets\",\n",
    "    train=False,\n",
    "    download=False,\n",
    "    transform=transforms.ToTensor()\n",
    ")\n",
    "\n",
    "train_dataloader = DataLoader(train_data, batch_size=128, shuffle=True)\n",
    "test_dataloader = DataLoader(test_data, batch_size=128, shuffle=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Device:  cuda\n",
      "----------------------------------------------------------------\n",
      "        Layer (type)               Output Shape         Param #\n",
      "================================================================\n",
      "            Conv2d-1           [-1, 64, 16, 16]             640\n",
      "              ReLU-2           [-1, 64, 16, 16]               0\n",
      "         MaxPool2d-3             [-1, 64, 7, 7]               0\n",
      "            Conv2d-4            [-1, 192, 7, 7]         307,392\n",
      "              ReLU-5            [-1, 192, 7, 7]               0\n",
      "         MaxPool2d-6            [-1, 192, 3, 3]               0\n",
      "            Conv2d-7            [-1, 384, 3, 3]         663,936\n",
      "              ReLU-8            [-1, 384, 3, 3]               0\n",
      "            Conv2d-9            [-1, 256, 3, 3]         884,992\n",
      "             ReLU-10            [-1, 256, 3, 3]               0\n",
      "           Conv2d-11            [-1, 256, 3, 3]         590,080\n",
      "             ReLU-12            [-1, 256, 3, 3]               0\n",
      "AdaptiveAvgPool2d-13            [-1, 256, 1, 1]               0\n",
      "          Dropout-14                  [-1, 256]               0\n",
      "           Linear-15                  [-1, 256]          65,792\n",
      "             ReLU-16                  [-1, 256]               0\n",
      "          Dropout-17                  [-1, 256]               0\n",
      "           Linear-18                  [-1, 256]          65,792\n",
      "             ReLU-19                  [-1, 256]               0\n",
      "           Linear-20                   [-1, 10]           2,570\n",
      "================================================================\n",
      "Total params: 2,581,194\n",
      "Trainable params: 2,581,194\n",
      "Non-trainable params: 0\n",
      "----------------------------------------------------------------\n",
      "Input size (MB): 0.00\n",
      "Forward/backward pass size (MB): 0.57\n",
      "Params size (MB): 9.85\n",
      "Estimated Total Size (MB): 10.42\n",
      "----------------------------------------------------------------\n",
      "Epoch 1\n",
      "-------------------------------\n",
      "loss:2.302578 [    0/60000]\n",
      "loss:1.522197 [12800/60000]\n",
      "loss:0.838229 [25600/60000]\n",
      "loss:0.529236 [38400/60000]\n",
      "loss:0.237161 [51200/60000]\n",
      "loss:0.266429 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 94.9%, Avg loss: 0.184954 \n",
      "\n",
      "Epoch 2\n",
      "-------------------------------\n",
      "loss:0.258096 [    0/60000]\n",
      "loss:0.137712 [12800/60000]\n",
      "loss:0.051611 [25600/60000]\n",
      "loss:0.202075 [38400/60000]\n",
      "loss:0.015648 [51200/60000]\n",
      "loss:0.357799 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 96.8%, Avg loss: 0.126823 \n",
      "\n",
      "Epoch 3\n",
      "-------------------------------\n",
      "loss:0.112786 [    0/60000]\n",
      "loss:0.048934 [12800/60000]\n",
      "loss:0.045217 [25600/60000]\n",
      "loss:0.095994 [38400/60000]\n",
      "loss:0.091738 [51200/60000]\n",
      "loss:0.067425 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 98.4%, Avg loss: 0.066276 \n",
      "\n",
      "Epoch 4\n",
      "-------------------------------\n",
      "loss:0.155697 [    0/60000]\n",
      "loss:0.134742 [12800/60000]\n",
      "loss:0.010438 [25600/60000]\n",
      "loss:0.072011 [38400/60000]\n",
      "loss:0.111358 [51200/60000]\n",
      "loss:0.108274 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 98.3%, Avg loss: 0.067415 \n",
      "\n",
      "Epoch 5\n",
      "-------------------------------\n",
      "loss:0.046446 [    0/60000]\n",
      "loss:0.103882 [12800/60000]\n",
      "loss:0.028469 [25600/60000]\n",
      "loss:0.021966 [38400/60000]\n",
      "loss:0.045129 [51200/60000]\n",
      "loss:0.005041 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 98.9%, Avg loss: 0.051455 \n",
      "\n",
      "Epoch 6\n",
      "-------------------------------\n",
      "loss:0.062159 [    0/60000]\n",
      "loss:0.040075 [12800/60000]\n",
      "loss:0.050420 [25600/60000]\n",
      "loss:0.021292 [38400/60000]\n",
      "loss:0.007540 [51200/60000]\n",
      "loss:0.015519 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 98.9%, Avg loss: 0.056168 \n",
      "\n",
      "Epoch 7\n",
      "-------------------------------\n",
      "loss:0.002384 [    0/60000]\n",
      "loss:0.031476 [12800/60000]\n",
      "loss:0.031004 [25600/60000]\n",
      "loss:0.021019 [38400/60000]\n",
      "loss:0.039257 [51200/60000]\n",
      "loss:0.019988 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 98.8%, Avg loss: 0.059503 \n",
      "\n",
      "Epoch 8\n",
      "-------------------------------\n",
      "loss:0.097776 [    0/60000]\n",
      "loss:0.054168 [12800/60000]\n",
      "loss:0.012010 [25600/60000]\n",
      "loss:0.028992 [38400/60000]\n",
      "loss:0.007287 [51200/60000]\n",
      "loss:0.124471 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 98.9%, Avg loss: 0.049245 \n",
      "\n",
      "Epoch 9\n",
      "-------------------------------\n",
      "loss:0.043949 [    0/60000]\n",
      "loss:0.028549 [12800/60000]\n",
      "loss:0.015731 [25600/60000]\n",
      "loss:0.018138 [38400/60000]\n",
      "loss:0.055446 [51200/60000]\n",
      "loss:0.001801 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 98.8%, Avg loss: 0.058817 \n",
      "\n",
      "Epoch 10\n",
      "-------------------------------\n",
      "loss:0.026255 [    0/60000]\n",
      "loss:0.047296 [12800/60000]\n",
      "loss:0.067395 [25600/60000]\n",
      "loss:0.005622 [38400/60000]\n",
      "loss:0.019111 [51200/60000]\n",
      "loss:0.070148 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 98.6%, Avg loss: 0.075099 \n",
      "\n",
      "Epoch 11\n",
      "-------------------------------\n",
      "loss:0.035959 [    0/60000]\n",
      "loss:0.110684 [12800/60000]\n",
      "loss:0.005775 [25600/60000]\n",
      "loss:0.008278 [38400/60000]\n",
      "loss:0.017837 [51200/60000]\n",
      "loss:0.034068 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 98.6%, Avg loss: 0.070940 \n",
      "\n",
      "Epoch 12\n",
      "-------------------------------\n",
      "loss:0.001034 [    0/60000]\n",
      "loss:0.052307 [12800/60000]\n",
      "loss:0.000893 [25600/60000]\n",
      "loss:0.008439 [38400/60000]\n",
      "loss:0.085831 [51200/60000]\n",
      "loss:0.004967 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 98.5%, Avg loss: 0.081541 \n",
      "\n",
      "Epoch 13\n",
      "-------------------------------\n",
      "loss:0.001433 [    0/60000]\n",
      "loss:0.000712 [12800/60000]\n",
      "loss:0.036345 [25600/60000]\n",
      "loss:0.022092 [38400/60000]\n",
      "loss:0.017018 [51200/60000]\n",
      "loss:0.070417 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 98.5%, Avg loss: 0.085425 \n",
      "\n",
      "Epoch 14\n",
      "-------------------------------\n",
      "loss:0.029295 [    0/60000]\n",
      "loss:0.005218 [12800/60000]\n",
      "loss:0.006035 [25600/60000]\n",
      "loss:0.002785 [38400/60000]\n",
      "loss:0.000648 [51200/60000]\n",
      "loss:0.004682 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 98.4%, Avg loss: 0.081657 \n",
      "\n",
      "Epoch 15\n",
      "-------------------------------\n",
      "loss:0.020875 [    0/60000]\n",
      "loss:0.002914 [12800/60000]\n",
      "loss:0.015822 [25600/60000]\n",
      "loss:0.004445 [38400/60000]\n",
      "loss:0.001598 [51200/60000]\n",
      "loss:0.008679 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 98.5%, Avg loss: 0.073702 \n",
      "\n",
      "Epoch 16\n",
      "-------------------------------\n",
      "loss:0.039377 [    0/60000]\n",
      "loss:0.001175 [12800/60000]\n",
      "loss:0.001824 [25600/60000]\n",
      "loss:0.008340 [38400/60000]\n",
      "loss:0.002658 [51200/60000]\n",
      "loss:0.003006 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 98.7%, Avg loss: 0.077902 \n",
      "\n",
      "Epoch 17\n",
      "-------------------------------\n",
      "loss:0.003030 [    0/60000]\n",
      "loss:0.000089 [12800/60000]\n",
      "loss:0.026658 [25600/60000]\n",
      "loss:0.032230 [38400/60000]\n",
      "loss:0.012752 [51200/60000]\n",
      "loss:0.003068 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 98.9%, Avg loss: 0.050325 \n",
      "\n",
      "Epoch 18\n",
      "-------------------------------\n",
      "loss:0.000145 [    0/60000]\n",
      "loss:0.000200 [12800/60000]\n",
      "loss:0.002139 [25600/60000]\n",
      "loss:0.001356 [38400/60000]\n",
      "loss:0.016002 [51200/60000]\n",
      "loss:0.011882 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 98.9%, Avg loss: 0.065362 \n",
      "\n",
      "Epoch 19\n",
      "-------------------------------\n",
      "loss:0.003666 [    0/60000]\n",
      "loss:0.104315 [12800/60000]\n",
      "loss:0.004596 [25600/60000]\n",
      "loss:0.008757 [38400/60000]\n",
      "loss:0.043926 [51200/60000]\n",
      "loss:0.008810 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 98.9%, Avg loss: 0.063651 \n",
      "\n",
      "Epoch 20\n",
      "-------------------------------\n",
      "loss:0.078657 [    0/60000]\n",
      "loss:0.003820 [12800/60000]\n",
      "loss:0.014194 [25600/60000]\n",
      "loss:0.008184 [38400/60000]\n",
      "loss:0.000147 [51200/60000]\n",
      "loss:0.002365 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 99.1%, Avg loss: 0.048099 \n",
      "\n",
      "Epoch 21\n",
      "-------------------------------\n",
      "loss:0.001497 [    0/60000]\n",
      "loss:0.013383 [12800/60000]\n",
      "loss:0.006674 [25600/60000]\n",
      "loss:0.000274 [38400/60000]\n",
      "loss:0.015153 [51200/60000]\n",
      "loss:0.000682 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 98.9%, Avg loss: 0.069430 \n",
      "\n",
      "Epoch 22\n",
      "-------------------------------\n",
      "loss:0.031266 [    0/60000]\n",
      "loss:0.001724 [12800/60000]\n",
      "loss:0.075334 [25600/60000]\n",
      "loss:0.000805 [38400/60000]\n",
      "loss:0.001044 [51200/60000]\n",
      "loss:0.000603 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 99.0%, Avg loss: 0.054393 \n",
      "\n",
      "Epoch 23\n",
      "-------------------------------\n",
      "loss:0.005563 [    0/60000]\n",
      "loss:0.000015 [12800/60000]\n",
      "loss:0.000403 [25600/60000]\n",
      "loss:0.000781 [38400/60000]\n",
      "loss:0.000221 [51200/60000]\n",
      "loss:0.013179 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 98.9%, Avg loss: 0.062143 \n",
      "\n",
      "Epoch 24\n",
      "-------------------------------\n",
      "loss:0.022084 [    0/60000]\n",
      "loss:0.000283 [12800/60000]\n",
      "loss:0.017641 [25600/60000]\n",
      "loss:0.003954 [38400/60000]\n",
      "loss:0.000230 [51200/60000]\n",
      "loss:0.001889 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 98.9%, Avg loss: 0.066916 \n",
      "\n",
      "Epoch 25\n",
      "-------------------------------\n",
      "loss:0.004904 [    0/60000]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "loss:0.083278 [12800/60000]\n",
      "loss:0.000678 [25600/60000]\n",
      "loss:0.002561 [38400/60000]\n",
      "loss:0.000023 [51200/60000]\n",
      "loss:0.003709 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 99.0%, Avg loss: 0.059313 \n",
      "\n",
      "Epoch 26\n",
      "-------------------------------\n",
      "loss:0.002318 [    0/60000]\n",
      "loss:0.001030 [12800/60000]\n",
      "loss:0.001638 [25600/60000]\n",
      "loss:0.092127 [38400/60000]\n",
      "loss:0.000297 [51200/60000]\n",
      "loss:0.000549 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 99.0%, Avg loss: 0.071207 \n",
      "\n",
      "Epoch 27\n",
      "-------------------------------\n",
      "loss:0.000231 [    0/60000]\n",
      "loss:0.000664 [12800/60000]\n",
      "loss:0.005357 [25600/60000]\n",
      "loss:0.026246 [38400/60000]\n",
      "loss:0.016418 [51200/60000]\n",
      "loss:0.005010 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 98.8%, Avg loss: 0.071526 \n",
      "\n",
      "Epoch 28\n",
      "-------------------------------\n",
      "loss:0.002571 [    0/60000]\n",
      "loss:0.017347 [12800/60000]\n",
      "loss:0.007408 [25600/60000]\n",
      "loss:0.000007 [38400/60000]\n",
      "loss:0.000721 [51200/60000]\n",
      "loss:0.005163 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 99.0%, Avg loss: 0.068224 \n",
      "\n",
      "Epoch 29\n",
      "-------------------------------\n",
      "loss:0.000262 [    0/60000]\n",
      "loss:0.000039 [12800/60000]\n",
      "loss:0.000944 [25600/60000]\n",
      "loss:0.000100 [38400/60000]\n",
      "loss:0.085863 [51200/60000]\n",
      "loss:0.056115 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 99.0%, Avg loss: 0.059766 \n",
      "\n",
      "Epoch 30\n",
      "-------------------------------\n",
      "loss:0.003258 [    0/60000]\n",
      "loss:0.010503 [12800/60000]\n",
      "loss:0.000602 [25600/60000]\n",
      "loss:0.004685 [38400/60000]\n",
      "loss:0.000178 [51200/60000]\n",
      "loss:0.001055 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 99.1%, Avg loss: 0.060600 \n",
      "\n",
      "Save model: alexnet_mnist29.pth\n"
     ]
    }
   ],
   "source": [
    "from models import AlexNet_mnist\n",
    "device = 'cuda' if torch.cuda.is_available() else 'cpu'\n",
    "print('Device: ',device)\n",
    "alexnet_mnist = AlexNet_mnist(10).to(device)\n",
    "summary(alexnet_mnist, (1,32,32))\n",
    "train_and_test(train_dataloader, test_dataloader, alexnet_mnist, 'alexnet_mnist', epochs=30, learning_rate = 1e-3, batch_size = 128, steps=100,imagenet_optimize=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Device:  cuda\n",
      "----------------------------------------------------------------\n",
      "        Layer (type)               Output Shape         Param #\n",
      "================================================================\n",
      "            Conv2d-1           [-1, 64, 32, 32]             640\n",
      "              ReLU-2           [-1, 64, 32, 32]               0\n",
      "         MaxPool2d-3           [-1, 64, 16, 16]               0\n",
      "            Conv2d-4          [-1, 128, 16, 16]          73,856\n",
      "              ReLU-5          [-1, 128, 16, 16]               0\n",
      "         MaxPool2d-6            [-1, 128, 8, 8]               0\n",
      "            Conv2d-7            [-1, 256, 8, 8]         295,168\n",
      "              ReLU-8            [-1, 256, 8, 8]               0\n",
      "            Conv2d-9            [-1, 256, 8, 8]         590,080\n",
      "             ReLU-10            [-1, 256, 8, 8]               0\n",
      "        MaxPool2d-11            [-1, 256, 4, 4]               0\n",
      "           Conv2d-12            [-1, 512, 4, 4]       1,180,160\n",
      "             ReLU-13            [-1, 512, 4, 4]               0\n",
      "           Conv2d-14            [-1, 512, 4, 4]       2,359,808\n",
      "             ReLU-15            [-1, 512, 4, 4]               0\n",
      "        MaxPool2d-16            [-1, 512, 2, 2]               0\n",
      "           Conv2d-17            [-1, 512, 2, 2]       2,359,808\n",
      "             ReLU-18            [-1, 512, 2, 2]               0\n",
      "           Conv2d-19            [-1, 512, 2, 2]       2,359,808\n",
      "             ReLU-20            [-1, 512, 2, 2]               0\n",
      "AdaptiveAvgPool2d-21            [-1, 512, 1, 1]               0\n",
      "           Linear-22                  [-1, 512]         262,656\n",
      "             ReLU-23                  [-1, 512]               0\n",
      "           Linear-24                   [-1, 10]           5,130\n",
      "================================================================\n",
      "Total params: 9,487,114\n",
      "Trainable params: 9,487,114\n",
      "Non-trainable params: 0\n",
      "----------------------------------------------------------------\n",
      "Input size (MB): 0.00\n",
      "Forward/backward pass size (MB): 2.56\n",
      "Params size (MB): 36.19\n",
      "Estimated Total Size (MB): 38.75\n",
      "----------------------------------------------------------------\n",
      "Epoch 1\n",
      "-------------------------------\n",
      "loss:2.302592 [    0/60000]\n",
      "loss:1.367428 [12800/60000]\n",
      "loss:0.147596 [25600/60000]\n",
      "loss:0.055777 [38400/60000]\n",
      "loss:0.119713 [51200/60000]\n",
      "loss:0.170974 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 97.8%, Avg loss: 0.085728 \n",
      "\n",
      "Epoch 2\n",
      "-------------------------------\n",
      "loss:0.144759 [    0/60000]\n",
      "loss:0.121946 [12800/60000]\n",
      "loss:0.020023 [25600/60000]\n",
      "loss:0.010326 [38400/60000]\n",
      "loss:0.044821 [51200/60000]\n",
      "loss:0.051941 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 98.8%, Avg loss: 0.046666 \n",
      "\n",
      "Epoch 3\n",
      "-------------------------------\n",
      "loss:0.078156 [    0/60000]\n",
      "loss:0.011413 [12800/60000]\n",
      "loss:0.038163 [25600/60000]\n",
      "loss:0.056072 [38400/60000]\n",
      "loss:0.009303 [51200/60000]\n",
      "loss:0.008713 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 98.8%, Avg loss: 0.041916 \n",
      "\n",
      "Epoch 4\n",
      "-------------------------------\n",
      "loss:0.048864 [    0/60000]\n",
      "loss:0.050097 [12800/60000]\n",
      "loss:0.064370 [25600/60000]\n",
      "loss:0.002120 [38400/60000]\n",
      "loss:0.006034 [51200/60000]\n",
      "loss:0.003170 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 98.8%, Avg loss: 0.053445 \n",
      "\n",
      "Epoch 5\n",
      "-------------------------------\n",
      "loss:0.015443 [    0/60000]\n",
      "loss:0.015755 [12800/60000]\n",
      "loss:0.002563 [25600/60000]\n",
      "loss:0.009347 [38400/60000]\n",
      "loss:0.066133 [51200/60000]\n",
      "loss:0.147360 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 97.8%, Avg loss: 0.086286 \n",
      "\n",
      "Epoch 6\n",
      "-------------------------------\n",
      "loss:0.010224 [    0/60000]\n",
      "loss:0.005622 [12800/60000]\n",
      "loss:0.019902 [25600/60000]\n",
      "loss:0.006012 [38400/60000]\n",
      "loss:0.005151 [51200/60000]\n",
      "loss:0.006002 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 98.9%, Avg loss: 0.046209 \n",
      "\n",
      "Epoch 7\n",
      "-------------------------------\n",
      "loss:0.043234 [    0/60000]\n",
      "loss:0.024746 [12800/60000]\n",
      "loss:0.056360 [25600/60000]\n",
      "loss:0.071593 [38400/60000]\n",
      "loss:0.046844 [51200/60000]\n",
      "loss:0.053100 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 98.7%, Avg loss: 0.084272 \n",
      "\n",
      "Epoch 8\n",
      "-------------------------------\n",
      "loss:0.016617 [    0/60000]\n",
      "loss:0.089430 [12800/60000]\n",
      "loss:0.000628 [25600/60000]\n",
      "loss:0.053147 [38400/60000]\n",
      "loss:0.002514 [51200/60000]\n",
      "loss:0.084950 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 99.0%, Avg loss: 0.050929 \n",
      "\n",
      "Epoch 9\n",
      "-------------------------------\n",
      "loss:0.022508 [    0/60000]\n",
      "loss:0.000756 [12800/60000]\n",
      "loss:0.038727 [25600/60000]\n",
      "loss:0.000936 [38400/60000]\n",
      "loss:0.002024 [51200/60000]\n",
      "loss:0.103538 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 99.2%, Avg loss: 0.032844 \n",
      "\n",
      "Epoch 10\n",
      "-------------------------------\n",
      "loss:0.005049 [    0/60000]\n",
      "loss:0.031500 [12800/60000]\n",
      "loss:0.035762 [25600/60000]\n",
      "loss:0.003275 [38400/60000]\n",
      "loss:0.013049 [51200/60000]\n",
      "loss:0.036894 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 99.1%, Avg loss: 0.044606 \n",
      "\n",
      "Epoch 11\n",
      "-------------------------------\n",
      "loss:0.103187 [    0/60000]\n",
      "loss:0.000491 [12800/60000]\n",
      "loss:0.027434 [25600/60000]\n",
      "loss:0.000390 [38400/60000]\n",
      "loss:0.004216 [51200/60000]\n",
      "loss:0.006923 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 99.0%, Avg loss: 0.046187 \n",
      "\n",
      "Epoch 12\n",
      "-------------------------------\n",
      "loss:0.017874 [    0/60000]\n",
      "loss:0.016211 [12800/60000]\n",
      "loss:0.041401 [25600/60000]\n",
      "loss:0.040164 [38400/60000]\n",
      "loss:0.001870 [51200/60000]\n",
      "loss:0.160102 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 99.1%, Avg loss: 0.057166 \n",
      "\n",
      "Epoch 13\n",
      "-------------------------------\n",
      "loss:0.004525 [    0/60000]\n",
      "loss:0.002718 [12800/60000]\n",
      "loss:0.000275 [25600/60000]\n",
      "loss:0.028334 [38400/60000]\n",
      "loss:0.001299 [51200/60000]\n",
      "loss:0.001463 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 99.4%, Avg loss: 0.033223 \n",
      "\n",
      "Epoch 14\n",
      "-------------------------------\n",
      "loss:0.000398 [    0/60000]\n",
      "loss:0.002855 [12800/60000]\n",
      "loss:0.003011 [25600/60000]\n",
      "loss:0.033994 [38400/60000]\n",
      "loss:0.003162 [51200/60000]\n",
      "loss:0.000163 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 99.2%, Avg loss: 0.045592 \n",
      "\n",
      "Epoch 15\n",
      "-------------------------------\n",
      "loss:0.032146 [    0/60000]\n",
      "loss:0.002402 [12800/60000]\n",
      "loss:0.007066 [25600/60000]\n",
      "loss:0.000244 [38400/60000]\n",
      "loss:0.078545 [51200/60000]\n",
      "loss:0.003081 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 99.4%, Avg loss: 0.041530 \n",
      "\n",
      "Epoch 16\n",
      "-------------------------------\n",
      "loss:0.000177 [    0/60000]\n",
      "loss:0.003797 [12800/60000]\n",
      "loss:0.000019 [25600/60000]\n",
      "loss:0.001262 [38400/60000]\n",
      "loss:0.000140 [51200/60000]\n",
      "loss:0.006627 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 99.4%, Avg loss: 0.046533 \n",
      "\n",
      "Epoch 17\n",
      "-------------------------------\n",
      "loss:0.001745 [    0/60000]\n",
      "loss:0.000841 [12800/60000]\n",
      "loss:0.005357 [25600/60000]\n",
      "loss:0.003265 [38400/60000]\n",
      "loss:0.002209 [51200/60000]\n",
      "loss:0.030862 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 99.3%, Avg loss: 0.038933 \n",
      "\n",
      "Epoch 18\n",
      "-------------------------------\n",
      "loss:0.038244 [    0/60000]\n",
      "loss:0.035189 [12800/60000]\n",
      "loss:0.001957 [25600/60000]\n",
      "loss:0.028599 [38400/60000]\n",
      "loss:0.000493 [51200/60000]\n",
      "loss:0.060324 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 99.4%, Avg loss: 0.041532 \n",
      "\n",
      "Epoch 19\n",
      "-------------------------------\n",
      "loss:0.000013 [    0/60000]\n",
      "loss:0.038678 [12800/60000]\n",
      "loss:0.048416 [25600/60000]\n",
      "loss:0.015040 [38400/60000]\n",
      "loss:0.089642 [51200/60000]\n",
      "loss:0.002309 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 99.0%, Avg loss: 0.074607 \n",
      "\n",
      "Epoch 20\n",
      "-------------------------------\n",
      "loss:0.000159 [    0/60000]\n",
      "loss:0.015766 [12800/60000]\n",
      "loss:0.002172 [25600/60000]\n",
      "loss:0.018681 [38400/60000]\n",
      "loss:0.000418 [51200/60000]\n",
      "loss:0.000017 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 99.2%, Avg loss: 0.055799 \n",
      "\n",
      "Epoch 21\n",
      "-------------------------------\n",
      "loss:0.000022 [    0/60000]\n",
      "loss:0.000812 [12800/60000]\n",
      "loss:0.002879 [25600/60000]\n",
      "loss:0.000018 [38400/60000]\n",
      "loss:0.000533 [51200/60000]\n",
      "loss:0.043077 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 99.3%, Avg loss: 0.058452 \n",
      "\n",
      "Epoch 22\n",
      "-------------------------------\n",
      "loss:0.000047 [    0/60000]\n",
      "loss:0.000120 [12800/60000]\n",
      "loss:0.003977 [25600/60000]\n",
      "loss:0.001166 [38400/60000]\n",
      "loss:0.005700 [51200/60000]\n",
      "loss:0.000321 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 99.4%, Avg loss: 0.037674 \n",
      "\n",
      "Epoch 23\n",
      "-------------------------------\n",
      "loss:0.000143 [    0/60000]\n",
      "loss:0.000276 [12800/60000]\n",
      "loss:0.000235 [25600/60000]\n",
      "loss:0.000452 [38400/60000]\n",
      "loss:0.000328 [51200/60000]\n",
      "loss:0.000144 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 98.8%, Avg loss: 0.078633 \n",
      "\n",
      "Epoch 24\n",
      "-------------------------------\n",
      "loss:0.010737 [    0/60000]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "loss:0.000037 [12800/60000]\n",
      "loss:0.030117 [25600/60000]\n",
      "loss:0.017918 [38400/60000]\n",
      "loss:0.013890 [51200/60000]\n",
      "loss:0.023275 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 99.4%, Avg loss: 0.041902 \n",
      "\n",
      "Epoch 25\n",
      "-------------------------------\n",
      "loss:0.016372 [    0/60000]\n",
      "loss:0.000078 [12800/60000]\n",
      "loss:0.001243 [25600/60000]\n",
      "loss:0.011287 [38400/60000]\n",
      "loss:0.000019 [51200/60000]\n",
      "loss:0.031200 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 99.2%, Avg loss: 0.046944 \n",
      "\n",
      "Epoch 26\n",
      "-------------------------------\n",
      "loss:0.041537 [    0/60000]\n",
      "loss:0.000341 [12800/60000]\n",
      "loss:0.007319 [25600/60000]\n",
      "loss:0.000725 [38400/60000]\n",
      "loss:0.000893 [51200/60000]\n",
      "loss:0.000346 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 99.4%, Avg loss: 0.049126 \n",
      "\n",
      "Epoch 27\n",
      "-------------------------------\n",
      "loss:0.025062 [    0/60000]\n",
      "loss:0.001643 [12800/60000]\n",
      "loss:0.000359 [25600/60000]\n",
      "loss:0.000475 [38400/60000]\n",
      "loss:0.002937 [51200/60000]\n",
      "loss:0.008342 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 99.3%, Avg loss: 0.037567 \n",
      "\n",
      "Epoch 28\n",
      "-------------------------------\n",
      "loss:0.000071 [    0/60000]\n",
      "loss:0.000633 [12800/60000]\n",
      "loss:0.001719 [25600/60000]\n",
      "loss:0.000212 [38400/60000]\n",
      "loss:0.003228 [51200/60000]\n",
      "loss:0.052737 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 99.4%, Avg loss: 0.045362 \n",
      "\n",
      "Epoch 29\n",
      "-------------------------------\n",
      "loss:0.000001 [    0/60000]\n",
      "loss:0.000007 [12800/60000]\n",
      "loss:0.000292 [25600/60000]\n",
      "loss:0.000270 [38400/60000]\n",
      "loss:0.000130 [51200/60000]\n",
      "loss:0.002004 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 99.2%, Avg loss: 0.058590 \n",
      "\n",
      "Epoch 30\n",
      "-------------------------------\n",
      "loss:0.022632 [    0/60000]\n",
      "loss:0.000002 [12800/60000]\n",
      "loss:0.001543 [25600/60000]\n",
      "loss:0.004814 [38400/60000]\n",
      "loss:0.014580 [51200/60000]\n",
      "loss:0.004242 [59776/60000]\n",
      "Test Error: \n",
      " Accuracy: 99.1%, Avg loss: 0.049832 \n",
      "\n",
      "Save model: vgg11_mnist29.pth\n"
     ]
    }
   ],
   "source": [
    "from models import VGG11_mnist\n",
    "device = 'cuda' if torch.cuda.is_available() else 'cpu'\n",
    "print('Device: ',device)\n",
    "vgg11_mnist = VGG11_mnist(10).to(device)\n",
    "summary(vgg11_mnist, (1,32,32))\n",
    "train_and_test(train_dataloader, test_dataloader, vgg11_mnist, 'vgg11_mnist', epochs=30, learning_rate = 1e-3, batch_size = 128, steps=100,imagenet_optimize=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x288 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x288 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from utils import draw_chart\n",
    "draw_chart('alexnet_mnist', True, save=True)\n",
    "draw_chart('vgg11_mnist', True, save=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Standard imports \n",
    "import numpy as np\n",
    "from matplotlib import pyplot as plt\n",
    "from PIL import Image\n",
    "import colorcet as cc\n",
    "import torchvision\n",
    "import torch\n",
    "import torchvision.transforms as transforms\n",
    "from torch.utils.data import DataLoader\n",
    "import torch.nn as nn\n",
    "# From the repository\n",
    "from utils import load_imagenet_img, make_grayscale, make_black_white, show_mask_on_image, cut_image_with_mask, show_img\n",
    "from vis import VanillaGradient, Deconvolution, GuidedBackprop, IntegratedGradients, visualize_imagenet, show_mask\n",
    "from models import load_pretrained_model, test_loop\n",
    "import time\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "\n",
    "def test(model, test_loader, show=True):\n",
    "    model.eval()\n",
    "    correct = 0\n",
    "    acc = 0\n",
    "    cls_num = 10\n",
    "    cls_acc = np.zeros(cls_num)\n",
    "    size = np.zeros(cls_num)\n",
    "    with torch.no_grad():\n",
    "        for data, target in test_loader:\n",
    "            data, target = data.cuda(), target.cuda()\n",
    "            output = model(data)\n",
    "            correct += (output.argmax(1)==target).type(torch.float).sum().item()\n",
    "\n",
    "            for i in range(data.shape[0]):\n",
    "                    size[target[i]] += 1\n",
    "                    if (output.argmax(1)==target)[i]:\n",
    "                        cls_acc[target[i]]+=1\n",
    "        \n",
    "    acc = correct / size.sum()\n",
    "    cls_acc = cls_acc/size\n",
    "    if show:\n",
    "        print(\"Test Accuracy = {0:.3f} Class Accuracy: {1}\".format(acc, np.round(cls_acc,3)))\n",
    "    return acc, cls_acc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Load  AlexNet_imagenet\n",
      "Test Error: \n",
      " Accuracy: 81.4%, Avg loss: 0.994053 \n",
      "\n",
      "Test Accuracy = 0.814 Class Accuracy: [0.82 0.58 0.82 0.88 0.84 0.84 0.68 0.92 0.92 0.84]\n",
      "Load  VGG11_imagenet\n",
      "Test Error: \n",
      " Accuracy: 73.0%, Avg loss: 0.991643 \n",
      "\n",
      "Test Accuracy = 0.730 Class Accuracy: [0.94 0.56 0.66 0.86 0.72 0.82 0.68 0.86 0.8  0.4 ]\n"
     ]
    }
   ],
   "source": [
    "data_transform = transforms.Compose([\n",
    "        transforms.Resize(256),\n",
    "        transforms.CenterCrop(224),\n",
    "        transforms.ToTensor(),\n",
    "        transforms.Normalize(mean=[0.485, 0.456, 0.406],\n",
    "                             std=[0.229, 0.224, 0.225])\n",
    "    ])\n",
    "test_dataset = torchvision.datasets.ImageFolder(root='./datasets/ImageNet/val',transform=data_transform)\n",
    "test_dataloader = DataLoader(test_dataset,batch_size=128, shuffle=True)   \n",
    "\n",
    "model_name_list = ['AlexNet_imagenet','VGG11_imagenet']\n",
    "for i in range(2):\n",
    "    model = load_pretrained_model(model_name_list[i].lower())\n",
    "    model.eval()\n",
    "    loss_fn = nn.CrossEntropyLoss()\n",
    "\n",
    "    test_loss, acc = test_loop(test_dataloader, model, loss_fn)\n",
    "    test(model, test_dataloader)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Load  AlexNet_cifar\n",
      "Test Error: \n",
      " Accuracy: 67.9%, Avg loss: 1.670880 \n",
      "\n",
      "Test Accuracy = 0.679 Class Accuracy: [0.729 0.784 0.525 0.464 0.573 0.618 0.774 0.729 0.846 0.751]\n",
      "Load  VGG11_cifar\n",
      "Test Error: \n",
      " Accuracy: 73.9%, Avg loss: 1.566129 \n",
      "\n",
      "Test Accuracy = 0.739 Class Accuracy: [0.79  0.883 0.518 0.559 0.639 0.627 0.852 0.829 0.831 0.866]\n"
     ]
    }
   ],
   "source": [
    "test_data = torchvision.datasets.CIFAR10(\n",
    "    root=\"./datasets\",#路径\n",
    "    train=False,#训练集\n",
    "    download=False,#是否下载\n",
    "    transform=transforms.ToTensor()#数据预处理\n",
    ")\n",
    "test_dataloader = DataLoader(test_data, batch_size=256, shuffle=True)\n",
    "\n",
    "model_name_list = ['AlexNet_cifar','VGG11_cifar']\n",
    "for i in range(2):\n",
    "    model = load_pretrained_model(model_name_list[i].lower())\n",
    "    model.eval()\n",
    "    loss_fn = nn.CrossEntropyLoss()\n",
    "\n",
    "    test_loss, acc = test_loop(test_dataloader, model, loss_fn)\n",
    "    test(model, test_dataloader)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Load  VGG11_cifar\n",
      "VGG11_cifar(\n",
      "  (features): Sequential(\n",
      "    (0): Conv2d(3, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
      "    (1): ReLU(inplace=True)\n",
      "    (2): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
      "    (3): Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
      "    (4): ReLU(inplace=True)\n",
      "    (5): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
      "    (6): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
      "    (7): ReLU(inplace=True)\n",
      "    (8): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
      "    (9): ReLU(inplace=True)\n",
      "    (10): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
      "    (11): Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
      "    (12): ReLU(inplace=True)\n",
      "    (13): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
      "    (14): ReLU(inplace=True)\n",
      "    (15): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
      "    (16): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
      "    (17): ReLU(inplace=True)\n",
      "    (18): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
      "    (19): ReLU(inplace=True)\n",
      "    (20): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
      "  )\n",
      "  (flatten): Flatten(start_dim=1, end_dim=-1)\n",
      "  (classifier): Sequential(\n",
      "    (0): Linear(in_features=512, out_features=512, bias=True)\n",
      "    (1): ReLU()\n",
      "    (2): Linear(in_features=512, out_features=10, bias=True)\n",
      "  )\n",
      ")\n"
     ]
    }
   ],
   "source": [
    "model_name_list = ['AlexNet_cifar','VGG11_cifar']\n",
    "model = load_pretrained_model(model_name_list[1].lower())\n",
    "print(model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Load  AlexNet_mnist\n",
      "Test Error: \n",
      " Accuracy: 99.1%, Avg loss: 0.056052 \n",
      "\n",
      "Test Accuracy = 0.991 Class Accuracy: [0.998 0.997 0.993 0.997 0.989 0.991 0.982 0.988 0.991 0.979]\n",
      "Load  VGG11_mnist\n",
      "Test Error: \n",
      " Accuracy: 99.1%, Avg loss: 0.056914 \n",
      "\n",
      "Test Accuracy = 0.991 Class Accuracy: [0.99  0.996 0.996 0.998 0.984 0.99  0.992 0.985 0.996 0.985]\n"
     ]
    }
   ],
   "source": [
    "test_data = torchvision.datasets.MNIST(\n",
    "    root=\"./datasets\",#路径\n",
    "    train=False,#训练集\n",
    "    download=False,#是否下载\n",
    "    transform=transforms.ToTensor()#数据预处理\n",
    ")\n",
    "test_dataloader = DataLoader(test_data, batch_size=256, shuffle=True)\n",
    "\n",
    "model_name_list = ['AlexNet_mnist','VGG11_mnist']\n",
    "for i in range(2):\n",
    "    model = load_pretrained_model(model_name_list[i].lower())\n",
    "    model.eval()\n",
    "    loss_fn = nn.CrossEntropyLoss()\n",
    "\n",
    "    test_loss, acc = test_loop(test_dataloader, model, loss_fn)\n",
    "    test(model, test_dataloader)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [conda env:pytorch] *",
   "language": "python",
   "name": "conda-env-pytorch-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}