data_augmentation/quick_start.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Quick Start Guide\n",
    "\n",
    "Here we show how to use some of the tools created in this repo. Let's start with the resource estimation."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Resource Estimation\n",
    "Let's construct a very simple circuit. In this case, we'll consider a VQC that contains 10 parameterised RX, RY or RZ rotations and 5 parameterised CRZ, CRY or CRZ rotations."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "4100000"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from ml_training import MlTrainingResourceEstimator, sim_ansatz, DifferentiationMethod\n",
    "\n",
    "simple_circuit_resources = MlTrainingResourceEstimator(num_rot_gates=10, num_c_rot_gates= 5,diff_method = DifferentiationMethod.PARAMETER_SHIFT) \n",
    "simple_circuit_resources.estimate_num_training_circuits(training_set_size=1000, n_epochs=100)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Feel free to have a play with some of the parameters if you've got a circuit in mind.\n",
    "\n",
    "We can also use some pre-made circuits, the [Sim ansatze](https://arxiv.org/abs/1905.10876). These allow us to build large standard quantum circuits quickly."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2201000000"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sim_circuit = sim_ansatz(ansatz_idx=5, n_qubits=10, layers=5)\n",
    "sim_circuit_resources = MlTrainingResourceEstimator(**sim_circuit,diff_method= DifferentiationMethod.PARAMETER_SHIFT)\n",
    "sim_circuit_resources.estimate_num_training_circuits(training_set_size=50000, n_epochs=20)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This framework also allows us to estimate the time it'll take to run these calculations. For instance, if we take the time to run one of these circuits as 0.1s:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2,547 days\n"
     ]
    }
   ],
   "source": [
    "time_in_seconds = sim_circuit_resources.estimate_training_time_in_seconds(time_per_circuit_in_seconds=0.1,training_set_size=50000,n_epochs=20)\n",
    "print(f\"{time_in_seconds/(60*60*24):,.0f} days\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Augmentation\n",
    "Now let's have a play with some data augmentation techniques. The first image in this dataset is a frog. Though any will do."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7f16e83da640>"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "from myrtle_torch_backend import *\n",
    "from myrtle_core import *\n",
    "from quantum_blur import *\n",
    "from blur_transform import GaussianBlur, QuantumBlur\n",
    "\n",
    "DATA_DIR = \"../data\"\n",
    "dataset = cifar10(root=DATA_DIR)\n",
    "\n",
    "# Pick the image you want to follow through data augmentation techniques\n",
    "img_idx = 0\n",
    "\n",
    "img1 = dataset[\"train\"][\"data\"][img_idx]\n",
    "plt.axis('off')\n",
    "plt.imshow(img1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First let's try the random cropping. Run this kernel multiple times to see how it changes the data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7f16e83c5f70>"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "transforms = [partial(transpose, source=\"NHWC\", target=\"NCHW\")]\n",
    "crop_train_set = list(zip(*preprocess(dataset[\"train\"], [partial(pad, border=4)] + transforms).values()))\n",
    "cropped_set = Transform(train_set, [Crop(32,32)])\n",
    "cropped_set.set_random_choices()\n",
    "plt.axis('off')\n",
    "plt.imshow(colours_channel_last(cropped_set.__getitem__(img_idx)[0]))\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can also try the flipping"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7f16e83daee0>"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "train_set = list(zip(*preprocess(dataset[\"train\"], transforms).values()))\n",
    "flipped_set = Transform(train_set, [FlipLR()])\n",
    "flipped_set.set_random_choices()\n",
    "plt.axis('off')\n",
    "plt.imshow(colours_channel_last(flipped_set.__getitem__(img_idx)[0]))\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Cutout"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7f16e84e94c0>"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "cutout_set = Transform(train_set, [Cutout(h=8,w=8)])\n",
    "cutout_set.set_random_choices()\n",
    "plt.axis('off')\n",
    "plt.imshow(colours_channel_last(cutout_set.__getitem__(img_idx)[0]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "And Gaussian blur (though this is far more subtle and could do with variance tuning)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7f16eb744430>"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "gaussian_blur_set = Transform(train_set, [GaussianBlur(h=8,w=8)])\n",
    "gaussian_blur_set.set_random_choices()\n",
    "plt.axis('off')\n",
    "plt.imshow(colours_channel_last(gaussian_blur_set.__getitem__(img_idx)[0]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Finally, we get to Quantum Blur"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7f16eb6c18b0>"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAGFCAYAAAASI+9IAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy88F64QAAAACXBIWXMAAA9hAAAPYQGoP6dpAAAZeUlEQVR4nO3cy44kB3rd8S8i75WVWbeuS7O7SQ7ZJIcihhxJI0GgZUgDbeyNYK/8EH4Mv4RX1gsYhiAYBgzYsDCARxDkwQiWyRnRHF6mr9VVXZWV91tkeGHg81LnAATsMf6/9ddfR0ZE1slYxCnquq4DAICIKP9vHwAA4P8dhAIAIBEKAIBEKAAAEqEAAEiEAgAgEQoAgEQoAABSUx38wz/6Y2vxaHQjz3bKnbX7uK2/b/fmyZ61+/S4L8/eO9y3drcbLXm22elZu6MhX8qIiLi5Hcmz6633fuPR4YE8W1Yba/dqtZJnl8ultbvb61rzVVTy7HwxtXYfHA714Vo/joiI9WotzzZCv2cjIhqNhjw72Pe+P/2+/t2MiGi19Ou5MM5JRERdGL+nS++76VyfbV1Yu//lv/rX/+AMTwoAgEQoAAASoQAASIQCACARCgCARCgAABKhAABIhAIAIBEKAIBEKAAAEqEAAEhyKcdnn39mLR5dX8uzx17lTBQn+j+4Vw283b0zeXa20/udIiKmld4hVBdta/d86XW3zBd6h9Cm8rqprht6H0u36fUqbbf6sTTMzplOp2PNz5czeXa7865PsTyRZ0u9bigiIjZGf1Sv6X05p0Zvz021tXbv7XndR0Wp9zYVRi9ZRESU+u/p+dLr99pu9PlG07tnFTwpAAASoQAASIQCACARCgCARCgAABKhAABIhAIAIBEKAIBEKAAAEqEAAEhyD0CvqVcXRESE8fb1W0ZtRUTE2+cH8uzZ6bG1u2e8Sl8U3jlZrJby7HKjVxFERNTmsbR7PX1461VR1Dv92A+O96zd241+LO2W8Rkjoqqs8Wi09Zt8tdavfUTEZqtfzz3jOCIimn39vHTN3dtCr/4oa68+ZRvePW60rcR+37sPp7O5PLvZejUXpXHck/GdtVv6/7/zjQCA31iEAgAgEQoAgEQoAAASoQAASIQCACARCgCARCgAABKhAABIhAIAIBEKAIAkdx91i621eDCQV8f7D46s3Se9hjzb2nmdM9ObtTxb7bxMXcz1c1i2rdUxPNy35ptGp83obuLt1i99HA+8zpnJWO/WWS/12YiIxdLrqKmNLp79vt6pFRGxWS/k2bIyTnhEtDr6ta8q75w0jcKh1crb3W55X4pyp3/fVtNba3dUegdXR/9zFRER253eCXU38zrSFDwpAAASoQAASIQCACARCgCARCgAABKhAABIhAIAIBEKAIBEKAAAEqEAAEjy+/FHHe9V+p7xKv1Bv2ftPh225NlqV1m7nelG03x/vdQzeLUz6wWcbomIaNb6q/TVSq9ciIioG/rnfPVqZO2uNvoVmszn1u55pVecRETs94b68Mq7DxuhX5+y0CsXIiIana48u5h5NTF7Lf2cNGvvuJdL7/osNnrNxS68YxlN9fMymnvf5alRh7PcfPe/63lSAAAkQgEAkAgFAEAiFAAAiVAAACRCAQCQCAUAQCIUAACJUAAAJEIBAJAIBQBAkgtzTg/1vpSIiEFL7wXqdr0OobKh95T0el6v0mard9TsorB217Xe3bLeel0s1drrV9nV+nxtdgLVzbY8O1nPrN1Vpd8r80rvD4qI2Jrzk5l+Dp/deJ+zVerHMpx69+Hm5bU8u7jz+qPevPdYnj07e2jtLgZ31vzq9rU8O5161+duoncfXd953WHfPNE/Z9XwOs8UPCkAABKhAABIhAIAIBEKAIBEKAAAEqEAAEiEAgAgEQoAgEQoAAASoQAASPI70m+c9q3Fw/ZWnt3f02sRIiIKo6IhwquLKGq9XmC18CoASqMW42RwYO3u970akvGdXnVwMBxauydL/fp8+0w/joiI6UqvuWh7rRXxYM+rDGi29PqCb16PrN2rWv+crcK7xw+GA3n209/6kbV7/EKviann5nHfa1nzq7l+PadT7/dxp6Ufy6ML/XxHRJydncuzl2O9bkPFkwIAIBEKAIBEKAAAEqEAAEiEAgAgEQoAgEQoAAASoQAASIQCACARCgCARCgAAJJcDnI86HmL1yN5ttPyOmf2Onvy7Grh9CRFbHZ6Z9Ph4ZG1u671rpd15eX1ZuN1oOzt78uzz69W1u5ffXsnz15N9PMdETE3xt/q6f1BERH/7B//0Jp/eF8/h//2Z19Zu//qy5fy7Ha3tnY3S/0+nIyurN3zqX6vDAZel1FUendYRES3q+9vd717Za/Qd28r7x5/89Eb8uzgZmLtVvCkAABIhAIAIBEKAIBEKAAAEqEAAEiEAgAgEQoAgEQoAAASoQAASIQCACDJ/RJnxyfW4sWNXrtQFl7NxXSuV1cs1t4r5s1Cf919vqms3U4CLzZedcHh0dCaX1d61cFXT59bu2/G+nmpm21rd6Ohn8Vh17s+Z02vMqB7o1c6vDe8sHa/ONY/5+XolbV7NdfvrZ9/8YW1u9zu5NlN37tn4+Dcmy/1vysHB3p1TkTEYKd/f5Zrr2qnXo/l2bdP+9ZuBU8KAIBEKAAAEqEAAEiEAgAgEQoAgEQoAAASoQAASIQCACARCgCARCgAABKhAABIcjnI0b1Ta/HRfk+eLcuWtXs0vpVnN7Optbus9L6cXeg9LxERdUvvYtnf71q7N+HN/+IrvdNmtppZu7vdjj7b9nqven29o+ao4fVe/ezLS2t+u9aPfXXgdR+dHunXswivQ2iz1XvJ5uuFtXs21zuB1lvv+hRmH1gU+mirNIYjoi71jrRW07vHtyu9U6s2OsxUPCkAABKhAABIhAIAIBEKAIBEKAAAEqEAAEiEAgAgEQoAgEQoAAASoQAASIQCACDppRxmP1HR8uYdna6+ey/61u6mkZNl6WXqxuhK6vQOrN3XLyfW/Pxa749659jrVVrp1TrRNbqMIiI+ePeBPFs6BxIR24Z3z46NDq5m487aPWjr9+3J0bvW7nffe1Oe/frXf2Pt/uUXz+TZdlPv+ImIqGuvx2y7Nf68NdvW7lZbv1d2O68jbWeUNhXFd/+7nicFAEAiFAAAiVAAACRCAQCQCAUAQCIUAACJUAAAJEIBAJAIBQBAIhQAAEl+D3yx3FiLi83CmN5au2ezsTy73ni5ty31Sofp3KuWGBvzDx7pr+hHRNRb71jeuqe/Sv/uG179w3yp737w/ifW7natV1fc3nn3bO/wxJqP1w159NHFfWv1aDaTZ9/5/nvW7uGRXi0yPPrQ2n17pd+Ht3de9UfLqP6IiCjrjjy72VXWbqe5otp4f99K/esTdV1bu6X//zvfCAD4jUUoAAASoQAASIQCACARCgCARCgAABKhAABIhAIAIBEKAIBEKAAAEqEAAEhywU5VeN0gdaX3fbj9Hb1uT57dH+g9LxERz6/0zqavn15Zu5st/XO2L59bu5eX3rG8d6b3Gf3JH3vdOr96diPPDh6cWrvvnVzIs6+uLq3dh4dmt85OP4ftUu9Jioh4dfVMnm12R9buq9ELefbZi6m1u9XSv2+HQ6NAKCIWC+/vRN3Uf/MWTuFQROyMrqSy8HYXpX7c1XdffcSTAgDg/yAUAACJUAAAJEIBAJAIBQBAIhQAAIlQAAAkQgEAkAgFAEAiFAAASa65ODzctxZvm3rNxXS6tHbXG/0V87vJnbX721/r1QjTqVcB0OvqGfzi67G1+7zbtuYfPHhLnj1843vW7tbEqC/o6lUREREPP/l9ffVLvSoiIqK39apCqtDv29nMu8fv7+n1H+vKq4so+vp3+WH/DWv34FCvIZm8fmntfnX52prfFPq9tVyvrN1R6v0S/U7XWr1e6H9XWm3v+6PgSQEAkAgFAEAiFAAAiVAAACRCAQCQCAUAQCIUAACJUAAAJEIBAJAIBQBAIhQAAEnuPpqMvN6R5noiz7YKM5saxnE0jOGImE/1rqSjQd/afdjXO1AWt1730dkbJ9b8g4//SJ79H0/X1u4vvtTnP71/bO0ejfTd5+9+Yu0uY27Nr1d6V9Jh7fUTjV/p37feemPtvn+sn/NR1bF2tz4+kmcXoxfW7v/6H/7Cmn/6RL8+DbtDqJAnF3pNUkREbIzf6uXGu/bSzu98IwDgNxahAABIhAIAIBEKAIBEKAAAEqEAAEiEAgAgEQoAgEQoAAASoQAASHLNRUN/qzsiIqrFVJ6tjVfGIyLK2OrHUXg1F7fGW+Pjsff+er3SKxruH3gVGr/34x9b8w8/+AN59t/92b+xdl/09+XZxnph7X721a/043jnt6zd3ZPH1ny/1qtc5jevvGNpvCfP1vXS2r2u9e/bg/P3rd2d7oU8u1rq1zIi4u+//GtrvjzVv5/ffO7V+Gw2+ne52FbW7qLW57db+U+4jCcFAEAiFAAAiVAAACRCAQCQCAUAQCIUAACJUAAAJEIBAJAIBQBAIhQAAIlQAAAkuTij8Gp+otroJUJF6WVT0xivF0aZUUQUO332+GTP2n2xp3c2/c6PvM6ZDz/Vu4wiIm5f6d1Une2dtfudhw/l2Z1zwiPi4uxUnt0u9fMdETEf6X02ERHrrb5/s/A6anatrjz74tlLa/fLl1/Lsx/svOvz1nv6cY9nXh9U2WtZ8916IM8eve3d49Xa6CcyOs8iIu6uRvLsauL9DVLwpAAASIQCACARCgCARCgAABKhAABIhAIAIBEKAIBEKAAAEqEAAEiEAgAgEQoAgCQXsuy2etdHRMRipXemtPv71u5mU+9AaZRe78jjiyN5ttvzMvXttx7Js5/84Y+t3fc/+Nia/9u/+jN59s1H+jmJiLj46AfybPv0XWt3c+9Anp0v9X6niIjFeGLNXz5/Is/eXj61dreMSptOf2jt7vb0zqbR+Btr98ndsTzb3HndVOVqbs0vjXNejb3rUxtlcL2O19nUvtDnx53C2q3gSQEAkAgFAEAiFAAAiVAAACRCAQCQCAUAQCIUAACJUAAAJEIBAJAIBQBAkmsuWg15NCIibif6K+nV0ntVu7fXk2cbpf46ekTE2YneL/Dkxcja/e7v/BN59uEP9Nn/zaui2Exm8uzBQK+WiIg4ff+H8uysqdciRER89vO/kWdXC/0zRkSMxyNr/vrZr+XZRuXVrbz5+KE82++8ae3+3ve/L882ut71GXTP5Nmi7Fi7a/P6LC/H8uz66ZW1e2v8nJ42GtbuvZO+PHv+xom1W8GTAgAgEQoAgEQoAAASoQAASIQCACARCgCARCgAABKhAABIhAIAIBEKAIBEKAAAklxotFosrcV7Hb0rqeh63SCtcivP1pU+GxHR29eP5U//xZ9auz/9p38izw7vnVu7L7/6hTXfMM7haHJn7b765u/l2eeTytr9l3/+5/Lsfq9l7V6uptb8xbneCTUc6H02ERFPjHN4N51Yuw8v3pFnz9+6b+2e3Op/J2Y3r6zd4+nGO5aNfizjtfc3aLnYybPT2utfq6f6cX94aK2W8KQAAEiEAgAgEQoAgEQoAAASoQAASIQCACARCgCARCgAABKhAABIhAIAIMldFLt67W3e6fUFxVZ/ZTwiYlvrr7sXhfeKebczlGd/+Lu/a+3utPTahc//9ufW7tvnv7LmVyv9VfrJ7Y21+8mXn8uz07pn7W5V+nHvN73qgmHXq6I4PdJrLl5cvrR2bzv67uXKq3948vXX8ux87v1u3O30c94qve/9qjyx5sfrb+XZWfvY2r030O/bXrNj7Z7Mx/LsdufV+Ch4UgAAJEIBAJAIBQBAIhQAAIlQAAAkQgEAkAgFAEAiFAAAiVAAACRCAQCQCAUAQJK7jyK8npLdVu9Karb2rN3VVu9VWofXDXJ+cCTP/se/+PfW7uPzz+TZs/uPrN3r+Z0132rpfSz7fb0PKiKiWer9N32jDyoi4uJM779ZTG6t3b2G11Hz+upant2s9Xs2IqI/0M/haqx35URE/OKvfyLPHg3/ztq9Neqm6vB6ye4W3vWMQSGPlgPv+pQdvYOra/YTHYXeq/ThR9+zdit4UgAAJEIBAJAIBQBAIhQAAIlQAAAkQgEAkAgFAEAiFAAAiVAAACRCAQCQ5JqL3U5/ZTwiot3U33fvNr0KjSj1Y6kbfWv1br2RZ6+vX1q7p1f6fG/jVRfswugXiIjjI70u4vCNU2v3tlrJs8+ee+fQqUYoS6PFJSLWW6+OoFHoFR39rlfl0m7rs03zu1kY7TadlndflQ39+zOfz63d6/Ura77Zmcmz5ZFXoTHZ6TU+y5n32/tk+I48e8+ofVHxpAAASIQCACARCgCARCgAABKhAABIhAIAIBEKAIBEKAAAEqEAAEiEAgAgEQoAgCSXoJRFx1rc7fTk2Tq8zpl+T++R6Q/uWbvnm6U8ezIwCmoioml8zvXdpbV7V3rHMm/pfVPn59/zjmWt98J88PFDa/dP/8t/lmfXtdet0yq8DqHFVN8/HAyt3duR3h9V7RbW7ulSv8d/9uJza/dopN/jq0LvJoqIOH3f+w374CP9b1Cr9r4/t9f6tW8v9Y6siIj+A73PaDGvrN0KnhQAAIlQAAAkQgEAkAgFAEAiFAAAiVAAACRCAQCQCAUAQCIUAACJUAAAJLnmot308mO+0l/Tb3T71u5dQ6/cmG+8CoBGq5ZnO239NfqIiFZL/5ztvQNr98HQO4cvr/QajfkDr4ri7NFjefbZq2tr90e/94/k2enVc2v3V198Zs3PpiN5ttnw7sODA70Wowi9siQi4sUz/bz8+ts7a3fZ0e/D4bleVxMRcXrsVYUURp1HceN9f45u5T+d8eDs2Nr98FD/vn35+Utr94//+T88w5MCACARCgCARCgAABKhAABIhAIAIBEKAIBEKAAAEqEAAEiEAgAgEQoAgEQoAACSXOBxfurlx+b1a3l2UXndLbOZPluXlbW72dQ7TYbDE2t3u9WSZxezsbW719KPOyIi1vr8f/vpT63V73yg9yo9fep1t5RlIc/udfTzHRHRMDq1IiJ6Pb0vZzb1uo8WC31+u11bu/d7+uf89Lfft3Z3B3o/0baxtXZXm7k1v3iidx+Vk661+2xvIM/+9vsfebsPz+XZn7342tqt4EkBAJAIBQBAIhQAAIlQAAAkQgEAkAgFAEAiFAAAiVAAACRCAQCQCAUAQCIUAABJLsB581HbWnxQ6F0iXz7xOk0ur2p5dl15fTb7+3on0Gx+Z+2udlN5tmHm9c2V3jUVETGZ6r0zy433ORu1Pj/YP7J2X768kWefzvTum4iIXa33KkVEnJ/q3VfFbmPtvh3dyrOdvnePHx7ovT3thncfrtZG11jT66aarbxjWU/1/f2dt/vxowt59o0LryPtyVO9O+z1lfe3U8GTAgAgEQoAgEQoAAASoQAASIQCACARCgCARCgAABKhAABIhAIAIBEKAIAkdzoMj7xX0hfG69dHZw1rd/T35NHry5W1erley7PN9tDabayO3caoC4iITeV9zruFXqPQ73k1Csu5Xi+xWF5bu9fGeanMc1jX3n04Hev3+HDYs3YPhwfy7GLhVR1cv9av/f5+39pdlPrvzGKr19VERLSb3jns6E070W571/7tx2/Ls4u59zl/8pPP5dn//sUra7eCJwUAQCIUAACJUAAAJEIBAJAIBQBAIhQAAIlQAAAkQgEAkAgFAEAiFAAAiVAAACS5+6jZlUcjIqI7bMuzx/teNjUXes9Pq7ezdo9vjc9Zecfd657pq1vecVerkTXf3tM/Z6upX8uIiEZD76Za1d7nXG/0Aqm6LqzdhVdRE/Va73iq9NGIiGg1ja6xttdNNbrVu48W6421++BQ7wNrGj1JERGleR/OYyvPXl5PrN23U333ZHZn7f5Pf/lLefbSq72S8KQAAEiEAgAgEQoAgEQoAAASoQAASIQCACARCgCARCgAABKhAABIhAIAIMldB9Op8dp9RERjXx7d73sdAK2e3kfQ73St3QcHeu3CdLywdk/Hl/rsvLJ2b5be/KB9Is92W9613670GpJm0/td0jbGW52GtbsovGPZ29erQkqvJSa2lV6j0O55y4eHeg3JzY1X/zAxakuGx/o9GBEx3+oVJxER//Ob1/LsL//uibX7/Fiv8zh/qJ/viIgo9XN472Dg7Vb+++98IwDgNxahAABIhAIAIBEKAIBEKAAAEqEAAEiEAgAgEQoAgEQoAAASoQAASIQCACDJpSlPv/UWr0Z659DgVO95iYjo9jby7IFewRQREcfHeo/MdDa3do9G+vzt67a1+1aveYmIiMZO7wXa1XrXVEREVRk9TDuvs8n5FVOUhbW70fQ6hBaVfjS1d4tHa6ff49v5jbW7Wuj3YdX0eq9GU3332rv0cWN2jX3zpf6lGL2eWbvXM/3gLw4urN0fvvVAnjVPiYQnBQBAIhQAAIlQAAAkQgEAkAgFAEAiFAAAiVAAACRCAQCQCAUAQCIUAABJfq+/at2zFm/aP5JnV7uVtbvcXsuz3QOv6uDwVK/nOCq97oLj+U6eHd30rN2ja722IiJiMdMrHaqtV7kRtf5bY7fVz0lExHKxlGfbbe+4G03vHE6W+rEvpvpxR0S06rU8OygH1u5dOZZnNxuv+qPT1ytRuq2OtfuwrZ+TiIh34lCe/cEnfWv3Bx9/Is++/fixtfv3/0CvCnn6fGrtVvCkAABIhAIAIBEKAIBEKAAAEqEAAEiEAgAgEQoAgEQoAAASoQAASIQCACARCgCAVNR1rZeVAAD+v8aTAgAgEQoAgEQoAAASoQAASIQCACARCgCARCgAABKhAABIhAIAIP0vUIvnZau3BOsAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "blur_factor = 0.1\n",
    "quantum_blur_set = Transform(train_set, [QuantumBlur(h=8,w=8,alpha=blur_factor, normalise=False)])\n",
    "quantum_blur_set.set_random_choices()\n",
    "plt.axis('off')\n",
    "plt.imshow(colours_channel_last(quantum_blur_set.__getitem__(img_idx)[0]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "All of the heights and widths of the methods above can also be adjusted, and we can combine these methods. Run the above methods a few times with different values for `img_idx` to see how these transformations affect a lot of images in this set."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7f16e9b60af0>"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "multiple_transforms = [Crop(32, 32), FlipLR(), Cutout(8, 8)]\n",
    "multi_transform_dataset = Transform(train_set, multiple_transforms)\n",
    "multi_transform_dataset.set_random_choices()\n",
    "plt.axis('off')\n",
    "plt.imshow(colours_channel_last(multi_transform_dataset.__getitem__(img_idx)[0]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Mixup\n",
    "Let's also explore how we can use mixup with images. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7f16eb6b6790>"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAGFCAYAAAASI+9IAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy88F64QAAAACXBIWXMAAA9hAAAPYQGoP6dpAAAVvUlEQVR4nO3cy69dh3Uf4LXPOfdJ8vIlUbJEG7RcNzGatEDQJEDch+OiQBG4SNECGWXQSfI3dZB2knHaAg3aog0KO3KcxpEdy5JjyZItUSQlvnnvJe/jvDtwsZCZ1ip0E7n5vvHi4j5773N+dw/2b1iv1+sAgIgY/U0fAACfHkIBgCQUAEhCAYAkFABIQgGAJBQASEIBgDQ5q8Wr1eqsVvO3WeNVy2EYWqtPjo5b848ePyzPXrlyubV7OTstz+7s7rZ2jze3yrProfd34yrq53zc2swnYTT6+OvpSQGAJBQASEIBgCQUAEhCAYAkFABIQgGAJBQASEIBgCQUAEhCAYB0Zt1HlY4N+DSZHh+05h/f/kl59tYPe7sPDo/Ks1/+6j9r7d7b2W5M977HQ6P7yC/Ep5PrAkASCgAkoQBAEgoAJKEAQBIKACShAEASCgAkoQBAEgoApDOruViv12e1mr/FOvfVaOjdg3dvvdea//6f/nF5dn5y3Nq9cf5yefbksFehsXflSnl21aitiIhYD/W/M/1C/PUbho+/np4UAEhCAYAkFABIQgGAJBQASEIBgCQUAEhCAYAkFABIQgGAJBQASGfWfVTp2ICudazKs/Npr2/ow1s3W/N7uzvl2d1LF1q77z95Wp599NGd1u4XPvu5+vBo3Nrd6TMaRn4jPo08KQCQhAIASSgAkIQCAEkoAJCEAgBJKACQhAIASSgAkIQCAOnMai6gYr3uFCNEjIb6/IPHj1q733//g9b8tLH/wvZma/fxs8Py7Fuv/0Vr94s3vlCevfTiy63d0biezUuvOueviScFAJJQACAJBQCSUAAgCQUAklAAIAkFAJJQACAJBQCSUAAgCQUAku4j/ob1CnDW62V59s7t263d733Qm7/17k/Ks89dON/aff25c+XZjz642dr9xmt/Xp79h1+51Nq9u3exPqzK6FPJkwIASSgAkIQCAEkoAJCEAgBJKACQhAIASSgAkIQCAEkoAJDOsOZi1Zg9y/fdP0Xv0jcaHdbN+odYd853RAz18zKc6d8OveuzWi3Ks/PFvLX76fFpa/72vcfl2XuN2YiI5fJaefb6td71eevPv12evfbiZ1q7/+4v/0pjuvfzM1r37pWh8xVq3uKdQxm6382zNHz8B/WkAEASCgAkoQBAEgoAJKEAQBIKACShAEASCgAkoQBAEgoAJKEAQDrD7qNmd88ZWZ9l91H3I67r/2DdmP3podQ7gSKafUaNnqSf7u70KnXV/8Xnbtxobd69sNeaPzw6qQ8XOmf+qjdv3S/P7ky2Wrsnp7Py7A++9Y3W7qsvv1CevXz9ldbuYdH7TgyNgqLu78RqVD+WxuiZq3yVPSkAkIQCAEkoAJCEAgBJKACQhAIASSgAkIQCAEkoAJCEAgDpDGsuPh15M5zhK+bdKopY1edX62Vr9XxRry6IiNjc3CzPDu2T2KkX6K4el0cvX36utfof/ZOvtObf+N5b5dn337vZ2r1c1K//u+O7rd3bN16qH8fb77R2v/GNPynP/uq/fL61e2f3fGt+2WiuaDa5tEoxFmdY+dOplImo/eB/On65AfhUEAoAJKEAQBIKACShAEASCgAkoQBAEgoAJKEAQBIKACShAEA6u+6jdad45MyOIqLZT7Ru9JR0D3uxXpRn33m31zlzcnLUmv/5L32pPLu1Ve8biogYdYtkGlbr+rGsmrf3r335H7fmP3jvTnn29/7d77V2L07qXVYfPNhv7d7a3SrPfvFK7+/Gt199rTz7/PVXWrt//su/0po/jvr3bWPV+5ybjXv88fFBa/d0Ni3PdjqyIiI+/8LnP3bGkwIASSgAkIQCAEkoAJCEAgBJKACQhAIASSgAkIQCAEkoAJCEAgDpzLqPVo3OoaFXTxTrxu71st5/EhExdGKy2fFz684H5dn/8l//sLX78LDXr/JrD++XZ3/9n361tXtrq96t07lPIiJWjdnFsjMdcf7Chdb8137za+XZd9/+UWv3H/23/1mePZz37vG37twtz14edlq7t0/rX6D//d//R2v35Or51vzohUvl2aP93vdnY1XvHPro8HZr98HT+rGcnp62dn/+N373Y2c8KQCQhAIASSgAkIQCAEkoAJCEAgBJKACQhAIASSgAkIQCAOnMai4i6q+B97olIp48eVSePXjyuLV7GNerK+4+qFdFRET86WvfLs9+5wevt3YfPt5vzU/ns/Ls3/vFX2jtvvb8c+XZ8bh3Cx4+PS7P7u/vt3bfuH69Nf/S9Wvl2X/7O7/d2n3rzo/Ls3/2+vdbu6dH4/LsO7frlRgREbsv1nc/evPN1u7j/9gajy98+ZfKs0+ePe0dy/FheXY67Ld2z+bT8uxq1ewIKvCkAEASCgAkoQBAEgoAJKEAQBIKACShAEASCgAkoQBAEgoAJKEAQGoUz9T7OCIiVqtO91FrdRwcPizPvvqtb7Z23/zwdnn24eF+a/eTo3q/yujcZmv39vRca/7+o845fLW1+8aNz5Znt7a2Wrvv3H5Qnp3P6v1OEREnx/ut+WdP6/MbzZaxL/3yK+XZ7737Rmv37Gm9L+f2fr3jJyJid7N+Pa9f3G7tfu+177bmx1v1v3lHL11p7T5Y1Du46m1Q/9e6/t2fTnu/yxWeFABIQgGAJBQASEIBgCQUAEhCAYAkFABIQgGAJBQASEIBgDSs1+vSO+8/+OG3W4snk43ybLeO4Mn+fnn2u6/3KgDeePut8uzFa1dbuxeTep/H1eeeb+1+8OOPWvM/fLN+Xl5++YXW7ot7O+XZ8aRXAjCd1SsaZtPT1u71sje/0fiT6qXr11q7ty7Wvz+vvfqD1u6/+Obb5dnVsnd9dhvj/+BSr5rl8t6F1vz4uYvl2f3ne5Ubj0er8uzGrLd7MV+UZ4+P63UbERF/9Ad/+LEznhQASEIBgCQUAEhCAYAkFABIQgGAJBQASEIBgCQUAEhCAYAkFABIk+rgt779rdbik8Oj8uy57V4Hyte+9pvl2cV6q7X7O280uo8uXG7tPlnVu3VeutbrG5rfO2nNHxzVO1OO36l35UREXN6q/61x7mLv2p+/XO+E2j5X76eJiLh4qdfzc3Fvrzy7t3e+tXvn/G559itf/dXW7oOHB+XZN9/8SWv3cl7v9/pgv9k1tVHvg4qImNytdwg9fVKfjYhYXKj3e412nmvtvnOr3mN22PidrfKkAEASCgAkoQBAEgoAJKEAQBIKACShAEASCgAkoQBAEgoApHLNxU/e773ufnD/SXn2i5//Ymv3zk69GuHDD++3dt9874Py7Plz9VfdIyKm83q1xHDYq6042e+9ph+jeh3B3/nCK63VX3j+Ynn2wuV6VURExP379YqGy1d6f/N85rO9yo2nh/Xrudlr3IjtVb1yY69xviMi/vm/+PXy7OMnh63d927Xv28Pp72TsnvQO5ZrjRqSybBu7X75wpXy7LkXXmztvvP+++XZ2fHT1u4KTwoAJKEAQBIKACShAEASCgAkoQBAEgoAJKEAQBIKACShAEASCgCkcvfR0UG9cyYi4vi03t2ztbvd2n3wtH4sN2+939p96WK9L2V5dNraPZxOy7Mf3X23tfujDx/2jmVUP5bf+jf/urV79exxefZ/ffPrrd03v3+nPHv14mZr99136n1QEREvv/S58uzB/F5rd2zUO4SuXH2htfoXf+4XyrOzf1X+iYiIiP/w73+/PHvytPf9+XD/WWs+JvXrP531epiePXxUnn2p8ZsSEbG5s1Gefe7apdbuCk8KACShAEASCgAkoQBAEgoAJKEAQBIKACShAEASCgAkoQBAEgoApHKxyWxa7zKKiDieHpVn332v1/Pzn/7zH5Rnv/mNb7R2D+t6/829w14Xy4Obt8qzG70qlpivlq35zRcvlmf/5I9fbe2eHtZ7mP7ynR+1dh/dW5Rn9x/0zsmlq70Orgd368dyeFD/PkREXL60U56dLXvn8Otf/255dmfvamv35eeulWcfzuv9QRERx9P6+Y6IuNPoVlpv9XqvdhvXc/yg3mMVEXHpav27OR73uqkqPCkAkIQCAEkoAJCEAgBJKACQhAIASSgAkIQCAEkoAJCEAgCp/I70xSv1V68jIuaNuDl8dtja/Zff+1559t5777V2j+qnJHYnG63dm6PN8ux6NmvtHkXvNf3rn3m5PHvlwuXW7ifH9UqUV278XGv3zeWT8uz+416NwnLrUmv+3lG9RuH4uFe5sf/4Xnl2GI9bu0+Hxjk8/nFr92izXs+xGte/DxER683e5zyOelfMctHrlTnX+JznL/a+P+Nx/cdzte7dVxWeFABIQgGAJBQASEIBgCQUAEhCAYAkFABIQgGAJBQASEIBgCQUAEjlop/zze6jyYVz5dnZo6PW7oc/ulWe/ez53nEPjX6ipyf17puIiNPRon4cO9ut3VtDrxfmwb3H5dnv/Nnrrd0vXLhQnn30ZL+1++Ck3qv0rFdnEycPex1c0eibmjR7fnY21uXZ02ZP1oP9/fLsctS7r3Yn9U6gYdT7m3S03TuWaHQfxXre2nx0VL8PDw/rsxERl69eqg+vep1nFZ4UAEhCAYAkFABIQgGAJBQASEIBgCQUAEhCAYAkFABIQgGAVK65WG328mO9rL9+vTnu7d6YL8uzn9u70tq9aLzW/7RRuRARMd47X54dbfZqLk7uHbTmp/vH5dmnj562dj9c1a/n/rR+HBERN37p75dn7z541Nq9/6R3Ds+fr1e5nB73qlzmG/Xrfzqt16dERJzM6/UPo1GvRmG7cd+uh161xLJTWxER40n55y1Gi3qtSETEalU/lvsP9lu7F/Wft5hsqrkA4AwJBQCSUAAgCQUAklAAIAkFAJJQACAJBQCSUAAgCQUAklAAIJXLQfb3e/030+NZefbcrN43FBHx/IsvlWcf3bzf2v3u+zfLsw/mp63dV67Ue5hG2zut3UerJ6355bzembI4nrZ2n07r5S2Lodc58+Duw/Ls0bNer9J63juW3a3d8uzspHevDFtb5dnFae/6bJ6rdzatl72+odNp/Xu/GvXO92xR3x0RsbWxWZ7d3K6f74iI87v1HrOdxmxExLxxH45Gn/zf9Z4UAEhCAYAkFABIQgGAJBQASEIBgCQUAEhCAYAkFABIQgGAVK65iJON3ubGm/eLof46ekTEUaMV46OhV6Hx0aL+Wv+zWa8CIB4dlEfHG72KhuNV71jWq3rNxcli0du9rtdcbDaqCCIi7jyo11wsmhUNQ9TPSUTEgyeNapGht3u9rJ/DjZ1eJcreZv2cLxf144iIWK/rFQ3jSe9v0p3o/QaNxvX9G837cGicw3Xzuzk0jns01H/Cyzs/8Y0A/MwSCgAkoQBAEgoAJKEAQBIKACShAEASCgAkoQBAEgoAJKEAQCoXZ0yGXu/IvNGB8uykUZQUEY8PD+uzs97uxUa9S2S96PUqnZ6clmeH6ay1e77u9auMRvVjP3dxr7V7PK7vHk963S3rxp8xnR6eiN5xd+dHo1730ajxOVed4YgYta5P775arupdSevuOWlen1HjvAzNbqoY6rtXjXMSEdGpGls0e8kqPCkAkIQCAEkoAJCEAgBJKACQhAIASSgAkIQCAEkoAJCEAgCp3DHw7Omz1uLDw6Py7NGzk9buo6NGXUTz7fW9S/VKh62drd7yhqFZXbAz2WzNb2zWj71b/7DRqArp1lwsV/XahW7NRURvvrN+3LyeMdSXL5fdGoV6NUL3HM4bu5fN8z2e9O7DSePe6n7O7e3t8uxW4/sQEbFu1GJsbX3yv0GeFABIQgGAJBQASEIBgCQUAEhCAYAkFABIQgGAJBQASEIBgCQUAEjlUo6Hjx61Fs9n9f6O09NZa/dsVp/f2N5o7d7YrncInZz0OptG43oGj0a9npdozq/X9VKoxbLeZxMRMZrUP+fObq+7pdUJ1eyz6fQqdQ3NEq4hmqVdDcfHx+XZbq/SpNHzsx41z0mzP6pzzvs9WY1jb67e3t4pz+o+AuBMCQUAklAAIAkFAJJQACAJBQCSUAAgCQUAklAAIAkFAJJQACCVi0rm814/UazreTOZ9PqJOnUfWzv1HpGIaFWaDPWal4iIGI/r/USrZl/KstFlFNHrtBk3e5XGm/X50Ubv75LNxr3S7bPp9vz0+3LqVo1DGTU7gS5dulSenc/nrd3TRi/Zcuidv25/VOf6LBa9fq/FonFelr1z2ClL6t6zFZ4UAEhCAYAkFABIQgGAJBQASEIBgCQUAEhCAYAkFABIQgGAVC5quHr1amvxKOp1BMtl73X3+WJV3918lf709KQ8O4x7r90PQz2DV6v6Z4yImC178+NVr7qitbtV59F7Tb9z7YdOZ8n/g07rwqrZW7JY1M/Lqvn9GU/q16db/zBvzM9Xvd2jxn0V0avF6FaWdO7xUaO2IqJXXdH9najwpABAEgoAJKEAQBIKACShAEASCgAkoQBAEgoAJKEAQBIKACShAEAqdx/t7e21Fq+WjWKYdS+bprN5efbw+Flr92Sj3mkybsxG9DpNolcJFBuj3jlcNDpTVp3jjmafUaMPKiJiWHcKh3qdM12rRl/OqtlNtW78vbZaN3uyTmbl2fm8/l2LiFh1en5GvW6q7tXs9AKtm9t3t7fLs5uNrqmIiFGjs2kyKf+E1///T3wjAD+zhAIASSgAkIQCAEkoAJCEAgBJKACQhAIASSgAkIQCAKn8jvTQzI9hqL82PptPW7tPpyfl2fm8/kp/RMRoXH8lfdKsllg3qg5mi0Vr93TRq6IYGhUDQ/Nzdl7THzV3rxb1+6pbi9ArXYjolEusG+ckImLZqWgYejUXo0n9WDbGG63dHZ3GkoiIdaNWJCJiuWzUkHRvlka1yKhZ5dLZvZg3+3AKPCkAkIQCAEkoAJCEAgBJKACQhAIASSgAkIQCAEkoAJCEAgBJKACQyt1Hq0YXS0TEdFrvHOr2E81mp/XZxnFERMzm9c6hVaOjJCJiaLTrjBsdTBER21tbrfnRpL5/2exh6nTUdO+rYVQ/7s75juj3MG02r1HH6Wn9Hl80r8+48Tm792Hn2k+nvc6z4+N651lExNDom9re3m7t7pzDxaz3OTtdSdvbve996f//xDcC8DNLKACQhAIASSgAkIQCAEkoAJCEAgBJKACQhAIASSgAkMo1F/P5vLW4U13RfU0/Gq/STyblj/hTrRqFnk5lQLdyYT3qHc28cc6753C5XJZnh6hfy4iI8XijPDtqXMuIXi1CRK/SYd2s89jc3CzPdu+Vs6zQ2NioX59uhUb3c3buw+7n3GzUS+xu7bZ2d+7C7j1b4UkBgCQUAEhCAYAkFABIQgGAJBQASEIBgCQUAEhCAYAkFABIQgGANKw7BS4A/H/NkwIASSgAkIQCAEkoAJCEAgBJKACQhAIASSgAkIQCAOn/AKepKERKaz+KAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "img2 = dataset[\"train\"][\"data\"][2]\n",
    "plt.axis('off')\n",
    "plt.imshow(img2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First, let's apply a classical mixup with alpha = 0.2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7f16eb6b6910>"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "alpha = 0.2\n",
    "plt.axis('off')\n",
    "plt.imshow(np.array(img1*alpha+ img2*(1-alpha),dtype=int))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Sadly the Quantum Mixup is too resource intensive to run in this kernel, so you'll need to try out the code yourself!"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3.9.15 ('.venv': poetry)",
   "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": "3.9.15"
  },
  "orig_nbformat": 4,
  "vscode": {
   "interpreter": {
    "hash": "9a67460380ca0fa1eaf973daa058d2818c5a0bc5a46c644743559835103efd09"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}