{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import torch\n",
    "torch.set_printoptions(edgeitems=2, threshold=50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(720, 1280, 3)"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import imageio\n",
    "\n",
    "img_arr = imageio.imread('../data/p1ch4/image-dog/bobby.jpg')\n",
    "img_arr.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "img = torch.from_numpy(img_arr)\n",
    "out = img.permute(2, 0, 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "batch_size = 3\n",
    "batch = torch.zeros(batch_size, 3, 256, 256, dtype=torch.uint8)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "import os\n",
    "\n",
    "data_dir = '../data/p1ch4/image-cats/'\n",
    "filenames = [name for name in os.listdir(data_dir)\n",
    "             if os.path.splitext(name)[-1] == '.png']\n",
    "for i, filename in enumerate(filenames):\n",
    "    img_arr = imageio.imread(os.path.join(data_dir, filename))\n",
    "    img_t = torch.from_numpy(img_arr)\n",
    "    img_t = img_t.permute(2, 0, 1)\n",
    "    img_t = img_t[:3] # <1>\n",
    "    batch[i] = img_t"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "batch = batch.float()\n",
    "batch /= 255.0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "n_channels = batch.shape[1]\n",
    "for c in range(n_channels):\n",
    "    mean = torch.mean(batch[:, c])\n",
    "    std = torch.std(batch[:, c])\n",
    "    batch[:, c] = (batch[:, c] - mean) / std"
   ]
  }
 ],
 "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": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}