Self_attention2.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%config InlineBackend.figure_format = 'retina'\n",
    "import os\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "from torchvision import transforms\n",
    "from torchvision.datasets import MNIST\n",
    "from torchvision import datasets\n",
    "from torch.utils.data import DataLoader\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "'''\n",
    "# Define data transformer\n",
    "img_transform = transforms.Compose([\n",
    "    transforms.Resize(28),\n",
    "    transforms.ToTensor(),\n",
    "    transforms.Normalize((0.5,), (0.5,))\n",
    "])\n",
    "# Read data and transform\n",
    "dataset = MNIST(root='./data', download=True, train=True, transform=img_transform)\n",
    "dataloader = DataLoader(dataset, batch_size=64, shuffle=False)\n",
    "x,_ = next(iter(dataloader))\n",
    "'''\n",
    "\n",
    "options = []\n",
    "options.append(transforms.CenterCrop(160))\n",
    "options.append(transforms.Resize((64,64)))\n",
    "options.append(transforms.ToTensor())\n",
    "options.append(transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5)))\n",
    "dataset = datasets.ImageFolder(os.getcwd(), transform=transforms.Compose(options))\n",
    "loader = DataLoader(dataset=dataset, batch_size=64)\n",
    "x,_ = next(iter(loader))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "query_conv = nn.Conv2d(in_channels = 3 , out_channels = 1 , kernel_size= 1)\n",
    "key_conv = nn.Conv2d(in_channels = 3 , out_channels = 1 , kernel_size= 1)\n",
    "value_conv = nn.Conv2d(in_channels = 3 , out_channels = 3 , kernel_size= 1)\n",
    "gamma = -0.1\n",
    "softmax = nn.Softmax(dim=-1)\n",
    "\n",
    "m_batchsize,C,width,height = x.size()\n",
    "        \n",
    "proj_query = query_conv(x).view(m_batchsize, -1, width*height).permute(0,2,1) # B * N * C\n",
    "proj_key = key_conv(x).view(m_batchsize, -1, width*height) # B * C * N\n",
    "energy = torch.bmm(proj_query, proj_key) # batch matrix-matrix product\n",
    "attention = softmax(energy) # B * N * N\n",
    "proj_value = value_conv(x).view(m_batchsize, -1, width*height) # B * C * N\n",
    "out = torch.bmm(proj_value, attention.permute(0,2,1)) # batch matrix-matrix product\n",
    "out = out.view(m_batchsize,C,width,height) # B * C * W * H\n",
    "        \n",
    "out1 = gamma*out + x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plotts(index):\n",
    "    fig = plt.figure(figsize=(15, 5))\n",
    "    plt.subplot(131)\n",
    "    plt.imshow(x[index].permute(1,2,0).detach().numpy())\n",
    "    plt.subplot(132)\n",
    "    plt.imshow(out[index].permute(1,2,0).detach().numpy())\n",
    "    plt.subplot(133)\n",
    "    plt.imshow(out1[index].permute(1,2,0).detach().numpy())\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
      "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plotts(6)"
   ]
  },
  {
   "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}