simple gpu benchmark code with pytorch for common CNN netoworks

cnn_dummy_bench.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SqueezeNet(\n",
       "  (features): Sequential(\n",
       "    (0): Conv2d(3, 96, kernel_size=(7, 7), stride=(2, 2))\n",
       "    (1): ReLU(inplace=True)\n",
       "    (2): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=True)\n",
       "    (3): Fire(\n",
       "      (squeeze): Conv2d(96, 16, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (squeeze_activation): ReLU(inplace=True)\n",
       "      (expand1x1): Conv2d(16, 64, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (expand1x1_activation): ReLU(inplace=True)\n",
       "      (expand3x3): Conv2d(16, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "      (expand3x3_activation): ReLU(inplace=True)\n",
       "    )\n",
       "    (4): Fire(\n",
       "      (squeeze): Conv2d(128, 16, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (squeeze_activation): ReLU(inplace=True)\n",
       "      (expand1x1): Conv2d(16, 64, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (expand1x1_activation): ReLU(inplace=True)\n",
       "      (expand3x3): Conv2d(16, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "      (expand3x3_activation): ReLU(inplace=True)\n",
       "    )\n",
       "    (5): Fire(\n",
       "      (squeeze): Conv2d(128, 32, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (squeeze_activation): ReLU(inplace=True)\n",
       "      (expand1x1): Conv2d(32, 128, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (expand1x1_activation): ReLU(inplace=True)\n",
       "      (expand3x3): Conv2d(32, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "      (expand3x3_activation): ReLU(inplace=True)\n",
       "    )\n",
       "    (6): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=True)\n",
       "    (7): Fire(\n",
       "      (squeeze): Conv2d(256, 32, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (squeeze_activation): ReLU(inplace=True)\n",
       "      (expand1x1): Conv2d(32, 128, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (expand1x1_activation): ReLU(inplace=True)\n",
       "      (expand3x3): Conv2d(32, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "      (expand3x3_activation): ReLU(inplace=True)\n",
       "    )\n",
       "    (8): Fire(\n",
       "      (squeeze): Conv2d(256, 48, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (squeeze_activation): ReLU(inplace=True)\n",
       "      (expand1x1): Conv2d(48, 192, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (expand1x1_activation): ReLU(inplace=True)\n",
       "      (expand3x3): Conv2d(48, 192, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "      (expand3x3_activation): ReLU(inplace=True)\n",
       "    )\n",
       "    (9): Fire(\n",
       "      (squeeze): Conv2d(384, 48, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (squeeze_activation): ReLU(inplace=True)\n",
       "      (expand1x1): Conv2d(48, 192, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (expand1x1_activation): ReLU(inplace=True)\n",
       "      (expand3x3): Conv2d(48, 192, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "      (expand3x3_activation): ReLU(inplace=True)\n",
       "    )\n",
       "    (10): Fire(\n",
       "      (squeeze): Conv2d(384, 64, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (squeeze_activation): ReLU(inplace=True)\n",
       "      (expand1x1): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (expand1x1_activation): ReLU(inplace=True)\n",
       "      (expand3x3): Conv2d(64, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "      (expand3x3_activation): ReLU(inplace=True)\n",
       "    )\n",
       "    (11): MaxPool2d(kernel_size=3, stride=2, padding=0, dilation=1, ceil_mode=True)\n",
       "    (12): Fire(\n",
       "      (squeeze): Conv2d(512, 64, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (squeeze_activation): ReLU(inplace=True)\n",
       "      (expand1x1): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1))\n",
       "      (expand1x1_activation): ReLU(inplace=True)\n",
       "      (expand3x3): Conv2d(64, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "      (expand3x3_activation): ReLU(inplace=True)\n",
       "    )\n",
       "  )\n",
       "  (classifier): Sequential(\n",
       "    (0): Dropout(p=0.5, inplace=False)\n",
       "    (1): Conv2d(512, 1000, kernel_size=(1, 1), stride=(1, 1))\n",
       "    (2): ReLU(inplace=True)\n",
       "    (3): AdaptiveAvgPool2d(output_size=(1, 1))\n",
       "  )\n",
       ")"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import torch\n",
    "import torch.nn as nn\n",
    "from torch.utils.data import DataLoader\n",
    "\n",
    "import torchvision.models as models\n",
    "import torchvision.datasets as datasets\n",
    "import torchvision.transforms as transforms\n",
    "\n",
    "\n",
    "model = models.squeezenet1_0()\n",
    "criterion = nn.CrossEntropyLoss()\n",
    "optimizer = torch.optim.SGD(model.parameters(), lr=1e-3)\n",
    "\n",
    "dataset = datasets.FakeData(\n",
    "    size=1000,\n",
    "    transform=transforms.ToTensor())\n",
    "loader = DataLoader(\n",
    "    dataset,\n",
    "    num_workers=1,\n",
    "    pin_memory=True\n",
    ")\n",
    "\n",
    "device = 'cuda:1'\n",
    "\n",
    "model.to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Done\n",
      "CPU times: user 9.62 s, sys: 947 ms, total: 10.6 s\n",
      "Wall time: 9.84 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "for data, target in loader:\n",
    "    data = data.to(device, non_blocking=True)\n",
    "    target = target.to(device, non_blocking=True)\n",
    "    optimizer.zero_grad()\n",
    "    output = model(data)\n",
    "    loss = criterion(output, target)\n",
    "    loss.backward()\n",
    "    optimizer.step()\n",
    "    \n",
    "print('Done')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}