enhanced_quantum_convolutional_filter(1).ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pennylane as qml\n",
    "from pennylane import numpy as np\n",
    "from pennylane.templates import RandomLayers\n",
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "import matplotlib.pyplot as plt\n",
    "import time"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Quantum hyperparameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "filter_size_h = 2\n",
    "filter_size_v = 2\n",
    "wires = filter_size_h * filter_size_v\n",
    "filters = 1\n",
    "layers = 1\n",
    "params = tf.Variable(tf.random.uniform([filters,layers, wires], dtype = tf.float64))\n",
    "\n",
    "#Here dev_amazon_braket\n",
    "'''\n",
    "my_bucket = f\"amazon-braket-Your-Bucket-Name\" # the name of the bucket\n",
    "my_prefix = \"Your-Folder-Name\" # the name of the folder in the bucket\n",
    "s3_folder = (my_bucket, my_prefix)\n",
    "device_arn = \"arn:aws:braket:::device/quantum-simulator/amazon/sv1\"\n",
    "\n",
    "dev_amazon_braket_remote = qml.device(\n",
    "  \"braket.aws.qubit\",\n",
    "device_arn=device_arn,\n",
    "wires=wires,\n",
    "s3_destination_folder=s3_folder,\n",
    "parallel=True,\n",
    ")\n",
    "'''\n",
    "\n",
    "dev_braket_local = qml.device(\"braket.local.qubit\", wires=wires)\n",
    "dev_pennylane_local = qml.device(\"default.qubit\", wires=wires)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Classical hyperparameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "n_epochs = 30   # Number of optimization epochs\n",
    "n_layers = 1    # Number of random layers\n",
    "n_train = 50    # Size of the train dataset\n",
    "n_test = 30      # Size of the test dataset\n",
    "np.random.seed(0)           # Seed for NumPy random number generator\n",
    "tf.random.set_seed(0)       # Seed for TensorFlow random number generator\n",
    "\n",
    "\n",
    "# Hidden layers weights and bias\n",
    "out_h1 = 10\n",
    "W1 = tf.Variable(tf.random.uniform([(28 // filter_size_v)* (28 // filter_size_h) * filters, out_h1], dtype = tf.float64))\n",
    "b1 = tf.Variable(tf.random.uniform([out_h1,], dtype = tf.float64))\n",
    "# W2 = tf.Variable(tf.cast(tf.random.normal((out_h1, out_h2)), dtype = tf.float64))\n",
    "# b2 = tf.Variable(tf.cast(tf.random.normal((out_h2,)), dtype = tf.float64))\n",
    "# out_h2 = 10"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load MNIST dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "mnist_dataset = keras.datasets.mnist\n",
    "(train_images, train_labels), (test_images, test_labels) = mnist_dataset.load_data()\n",
    "\n",
    "# Reduce dataset size\n",
    "train_images = train_images[:n_train]\n",
    "train_labels = tf.convert_to_tensor(train_labels[:n_train])\n",
    "test_images = test_images[:n_test]\n",
    "test_labels = tf.convert_to_tensor(test_labels[:n_test])\n",
    "\n",
    "# Normalize pixel values within 0 and 1\n",
    "train_images = train_images / 255\n",
    "test_images = test_images / 255\n",
    "\n",
    "# Add extra dimension for convolution channels\n",
    "train_images = tf.convert_to_tensor(train_images[..., tf.newaxis])\n",
    "test_images = tf.convert_to_tensor(test_images[..., tf.newaxis])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Hybrid quantum-classical CCN model architecture"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "def quanv(image, params, filters = 1):\n",
    "\n",
    "    if filter_size_h == 2:\n",
    "        out = tf.zeros((14, 14, filters))\n",
    "        # Loop over the coordinates of the top-left pixel of 2X2 squares\n",
    "        I = []\n",
    "        for i in range(filters):\n",
    "            J = []\n",
    "            for j in range(0, 28, 2):\n",
    "                K = []\n",
    "                for k in range(0, 28, 2):\n",
    "                    # Process a squared 2x2 region of the image with a quantum circuit\n",
    "\n",
    "                    q_results = quantum_scalar_prod(\n",
    "                        tf.convert_to_tensor([\n",
    "                            image[j, k, 0],\n",
    "                            image[j, k + 1, 0],\n",
    "                            image[j + 1, k, 0],\n",
    "                            image[j + 1, k + 1, 0]\n",
    "                        ]), params[i]\n",
    "                    )\n",
    "                    # Assign expectation values to different channels of the output pixel (j/2, k/2)\n",
    "                    #relu\n",
    "                    if q_results > 0:\n",
    "                        K.append(q_results)\n",
    "                    else:\n",
    "                        K.append(0)\n",
    "                J.append(K)\n",
    "            I.append(J)\n",
    "        return tf.convert_to_tensor(I)\n",
    "\n",
    "\n",
    "    elif filter_size_h == 4:\n",
    "        out = tf.zeros((7, 7, filters))\n",
    "        # Loop over the coordinates of the top-left pixel of 2X2 squares\n",
    "        I = []\n",
    "        for i in range(filters):\n",
    "            J = []\n",
    "            for j in range(0, 28, 4):\n",
    "                K = []\n",
    "                for k in range(0, 28, 4):\n",
    "                    # Process a squared 2x2 region of the image with a quantum circuit\n",
    "\n",
    "                    q_results = quantum_scalar_prod(\n",
    "                        tf.convert_to_tensor([\n",
    "                            image[j, k, 0],\n",
    "                            image[j, k + 1, 0],\n",
    "                            image[j, k + 2, 0],\n",
    "                            image[j, k + 3, 0],\n",
    "                            image[j+1, k, 0],\n",
    "                            image[j+1, k + 1, 0],\n",
    "                            image[j+1, k + 2, 0],\n",
    "                            image[j+1, k + 3, 0],\n",
    "                            image[j+2, k, 0],\n",
    "                            image[j+2, k + 1, 0],\n",
    "                            image[j+2, k + 2, 0],\n",
    "                            image[j+2, k + 3, 0],\n",
    "                            image[j+3, k, 0],\n",
    "                            image[j+3, k + 1, 0],\n",
    "                            image[j+3, k + 2, 0],\n",
    "                            image[j+3, k + 3, 0],\n",
    "\n",
    "\n",
    "                        ]), params[i]\n",
    "                    )\n",
    "                    # Assign expectation values to different channels of the output pixel (j/2, k/2)\n",
    "                    #relu\n",
    "                    if q_results > 0:\n",
    "                        K.append(q_results)\n",
    "                    else:\n",
    "                        K.append(0)\n",
    "                J.append(K)\n",
    "            I.append(J)\n",
    "        return tf.convert_to_tensor(I)\n",
    "\n",
    "    elif filter_size_h == 6:\n",
    "        out = tf.zeros((6, 4, filters))\n",
    "        # Loop over the coordinates of the top-left pixel of 2X2 squares\n",
    "        I = []\n",
    "        for i in range(filters):\n",
    "            J = []\n",
    "            for j in range(0, 28-6, 6):\n",
    "                K = []\n",
    "                for k in range(0, 28-5, 5):\n",
    "                    # Process a squared 2x2 region of the image with a quantum circuit\n",
    "\n",
    "                    #print(j,k)\n",
    "                    q_results = quantum_scalar_prod(\n",
    "                        tf.convert_to_tensor([\n",
    "\n",
    "                            image[j, k, 0],\n",
    "                            image[j, k + 1, 0],\n",
    "                            image[j, k + 2, 0],\n",
    "                            image[j, k + 3, 0],\n",
    "                            image[j, k + 4, 0],\n",
    "\n",
    "                            image[j+1, k, 0],\n",
    "                            image[j+1, k + 1, 0],\n",
    "                            image[j+1, k + 2, 0],\n",
    "                            image[j+1, k + 3, 0],\n",
    "                            image[j+1, k + 4, 0],\n",
    "\n",
    "                            image[j+2, k, 0],\n",
    "                            image[j+2, k + 1, 0],\n",
    "                            image[j+2, k + 2, 0],\n",
    "                            image[j+2, k + 3, 0],\n",
    "                            image[j+2, k + 4, 0],\n",
    "\n",
    "                            image[j+3, k, 0],\n",
    "                            image[j+3, k + 1, 0],\n",
    "                            image[j+3, k + 2, 0],\n",
    "                            image[j+3, k + 3, 0],\n",
    "                            image[j+3, k + 4, 0],\n",
    "\n",
    "                            image[j+4, k, 0],\n",
    "                            image[j+4, k + 1, 0],\n",
    "                            image[j+4, k + 2, 0],\n",
    "                            image[j+4, k + 3, 0],\n",
    "                            image[j+4, k + 4, 0],\n",
    "\n",
    "                            image[j+5, k, 0],\n",
    "                            image[j+5, k + 1, 0],\n",
    "                            image[j+5, k + 2, 0],\n",
    "                            image[j+5, k + 3, 0],\n",
    "                            image[j+5, k + 4, 0],\n",
    "\n",
    "\n",
    "                        ]), params[i]\n",
    "                    )\n",
    "                    # Assign expectation values to different channels of the output pixel (j/2, k/2)\n",
    "                    #relu\n",
    "                    if q_results > 0:\n",
    "                        K.append(q_results)\n",
    "                    else:\n",
    "                        K.append(0)\n",
    "                J.append(K)\n",
    "            I.append(J)\n",
    "        return tf.convert_to_tensor(I)\n",
    "\n",
    "\n",
    "def classical_scalar_prod(phi, params): #Classical\n",
    "\n",
    "    sol = 0\n",
    "\n",
    "    for i,j in zip(phi,params[0]):\n",
    "        sol += i * j\n",
    "\n",
    "    return sol\n",
    "\n",
    "def quantum_scalar_prod(phi, params): ##Quantum \n",
    "\n",
    "    prob = circuit(phi,params)[0]\n",
    "    return prob\n",
    "\n",
    "\n",
    "@qml.qnode(dev_pennylane_local, interface = \"tf\")\n",
    "def circuit(phi, params):\n",
    "\n",
    "    # Encoding of 4 classical input values\n",
    "    for j in range(wires):\n",
    "        qml.RY(-np.pi * phi[j], wires=j)\n",
    "\n",
    "    for layer in range(params.shape[0]):\n",
    "        for i in range(params.shape[1]):\n",
    "            qml.RX(np.pi * params[layer,i], wires = i)\n",
    "            qml.CNOT(wires =[i, (i+1)%wires])\n",
    "            qml.RY(np.pi * params[layer,i], wires = i)\n",
    "\n",
    "    return qml.probs(wires = range(wires))\n",
    "\n",
    "\n",
    "def x_hidden(x,W,b):\n",
    "    # Applies a fully connected layer , i.e out = h(Wx+b) where h(·) is the softmax function\n",
    "    x = tf.cast(x, dtype = tf.float64)\n",
    "    #if last == False:\n",
    "    #return tf.nn.relu(x @ W + b)\n",
    "\n",
    "    sol = tf.nn.softmax(x @ W + b)\n",
    "    #print(\"sol\",sol)\n",
    "    return sol\n",
    "\n",
    "def to_vector(n):\n",
    "    L = [0] * 10\n",
    "    L[n] = 1\n",
    "    return tf.cast(tf.convert_to_tensor(L), dtype = tf.float64)\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Training loop of a quantum convolutional filter layer + fully connected layer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0   tf.Tensor([2.84229564], shape=(1,), dtype=float64)\n",
      "Current Time(s) after loss plus previous gradient:  2.720309257507324\n",
      "1   tf.Tensor([2.39007819], shape=(1,), dtype=float64)\n",
      "Current Time(s) after loss plus previous gradient:  9.154855728149414\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-14-32df440f7ee4>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     34\u001b[0m             \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Current Time(s) after loss plus previous gradient: \"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtime\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtime\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0mstart\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     35\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 36\u001b[0;31m         \u001b[0mgradients\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtape\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgradient\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mloss\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mW1\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mb1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     37\u001b[0m         \u001b[0;31m#print(\"max grad\", max(gradients[0][13]))\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     38\u001b[0m         \u001b[0mopt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mapply_gradients\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mzip\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mgradients\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mW1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mb1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/envs/quantvolution/lib/python3.8/site-packages/tensorflow/python/eager/backprop.py\u001b[0m in \u001b[0;36mgradient\u001b[0;34m(self, target, sources, output_gradients, unconnected_gradients)\u001b[0m\n\u001b[1;32m   1078\u001b[0m                           for x in nest.flatten(output_gradients)]\n\u001b[1;32m   1079\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1080\u001b[0;31m     flat_grad = imperative_grad.imperative_grad(\n\u001b[0m\u001b[1;32m   1081\u001b[0m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_tape\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1082\u001b[0m         \u001b[0mflat_targets\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/envs/quantvolution/lib/python3.8/site-packages/tensorflow/python/eager/imperative_grad.py\u001b[0m in \u001b[0;36mimperative_grad\u001b[0;34m(tape, target, sources, output_gradients, sources_raw, unconnected_gradients)\u001b[0m\n\u001b[1;32m     69\u001b[0m         \"Unknown value for unconnected_gradients: %r\" % unconnected_gradients)\n\u001b[1;32m     70\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 71\u001b[0;31m   return pywrap_tfe.TFE_Py_TapeGradient(\n\u001b[0m\u001b[1;32m     72\u001b[0m       \u001b[0mtape\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_tape\u001b[0m\u001b[0;34m,\u001b[0m  \u001b[0;31m# pylint: disable=protected-access\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     73\u001b[0m       \u001b[0mtarget\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/envs/quantvolution/lib/python3.8/site-packages/tensorflow/python/eager/backprop.py\u001b[0m in \u001b[0;36m_gradient_function\u001b[0;34m(op_name, attr_tuple, num_inputs, inputs, outputs, out_grads, skip_input_indices, forward_pass_name_scope)\u001b[0m\n\u001b[1;32m    160\u001b[0m       \u001b[0mgradient_name_scope\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0mforward_pass_name_scope\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;34m\"/\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    161\u001b[0m     \u001b[0;32mwith\u001b[0m \u001b[0mops\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mname_scope\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mgradient_name_scope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 162\u001b[0;31m       \u001b[0;32mreturn\u001b[0m \u001b[0mgrad_fn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmock_op\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0mout_grads\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    163\u001b[0m   \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    164\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0mgrad_fn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmock_op\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0mout_grads\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/envs/quantvolution/lib/python3.8/site-packages/tensorflow/python/ops/math_grad.py\u001b[0m in \u001b[0;36m_MulGrad\u001b[0;34m(op, grad)\u001b[0m\n\u001b[1;32m   1328\u001b[0m       SmartBroadcastGradientArgs(x, y, grad))\n\u001b[1;32m   1329\u001b[0m   \u001b[0mx\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmath_ops\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconj\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1330\u001b[0;31m   \u001b[0my\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmath_ops\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconj\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1331\u001b[0m   \u001b[0;32mif\u001b[0m \u001b[0mskip_input_indices\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0;36m0\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mskip_input_indices\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1332\u001b[0m     \u001b[0mgx\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/envs/quantvolution/lib/python3.8/site-packages/tensorflow/python/util/dispatch.py\u001b[0m in \u001b[0;36mwrapper\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m    199\u001b[0m     \u001b[0;34m\"\"\"Call target, and fall back on dispatchers if there is a TypeError.\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    200\u001b[0m     \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 201\u001b[0;31m       \u001b[0;32mreturn\u001b[0m \u001b[0mtarget\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    202\u001b[0m     \u001b[0;32mexcept\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mTypeError\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mValueError\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    203\u001b[0m       \u001b[0;31m# Note: convert_to_eager_tensor currently raises a ValueError, not a\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/envs/quantvolution/lib/python3.8/site-packages/tensorflow/python/ops/math_ops.py\u001b[0m in \u001b[0;36mconj\u001b[0;34m(x, name)\u001b[0m\n\u001b[1;32m   3979\u001b[0m   \u001b[0;32mwith\u001b[0m \u001b[0mops\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mname_scope\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mname\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"Conj\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   3980\u001b[0m     \u001b[0mx\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mops\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconvert_to_tensor\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"x\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 3981\u001b[0;31m     \u001b[0;32mif\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdtype\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mis_complex\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdtype\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0mdtypes\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvariant\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   3982\u001b[0m       \u001b[0;32mreturn\u001b[0m \u001b[0mgen_math_ops\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconj\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mname\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   3983\u001b[0m     \u001b[0;32melif\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdtype\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mis_floating\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdtype\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mis_integer\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/envs/quantvolution/lib/python3.8/site-packages/tensorflow/python/framework/ops.py\u001b[0m in \u001b[0;36mdtype\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m   1043\u001b[0m     \u001b[0;31m# Note: using the intern table directly here as this is\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1044\u001b[0m     \u001b[0;31m# performance-sensitive in some models.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1045\u001b[0;31m     \u001b[0;32mreturn\u001b[0m \u001b[0mdtypes\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_INTERN_TABLE\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_datatype_enum\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m  \u001b[0;31m# pylint: disable=protected-access\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1046\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1047\u001b[0m   \u001b[0;32mdef\u001b[0m \u001b[0mnumpy\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "def accuracy():\n",
    "\n",
    "    correct = 0\n",
    "    for j in range(n_test):\n",
    "\n",
    "        image = tf.reshape(quanv(test_images[j], params, filters),[1,-1])\n",
    "        o1 = x_hidden(image,W1,b1)\n",
    "        #o2 = x_hidden(o1,W2,b2, last = True)\n",
    "       # o2=o1\n",
    "\n",
    "        if tf.math.argmax(o1[0]) == tf.cast(test_labels[j], dtype = tf.int64):\n",
    "            correct += 1\n",
    "    return correct / n_test\n",
    "\n",
    "losses = []\n",
    "accuracies = []\n",
    "start = time.time()\n",
    "opt = keras.optimizers.SGD()\n",
    "opt2 = keras.optimizers.Adadelta()\n",
    "for i in range(n_epochs):\n",
    "    epoch_loss = 0\n",
    "    for j in range(n_train):\n",
    "        with tf.GradientTape(persistent=True) as tape:\n",
    "            image = tf.reshape(quanv(train_images[j], params, filters),[1,-1])\n",
    "            o1 = x_hidden(image,W1,b1)\n",
    "            #o2 = x_hidden(o1,W2,b2, last = True)\n",
    "            label = tf.reshape(train_labels[j] , o1.shape[0])\n",
    "            #loss = sum((o2 - label)[0] ** 2)\n",
    "            #loss = tf.reduce_mean(-tf.reduce_sum(label * tf.math.log(o1)))\n",
    "            loss = tf.keras.losses.sparse_categorical_crossentropy(label[0], o1) \n",
    "            # comprobar como esta definido el error, creo que algo falla\n",
    "            epoch_loss += loss\n",
    "            print(j,\" \",loss)\n",
    "            print(\"Current Time(s) after loss plus previous gradient: \", time.time()-start)\n",
    "\n",
    "        gradients = tape.gradient(loss, [W1,b1])\n",
    "        #print(\"max grad\", max(gradients[0][13]))\n",
    "        opt.apply_gradients(zip(gradients, [W1, b1]))\n",
    "        #print(\"random\", W1[0][3])\n",
    "        gradients = tape.gradient(loss, [params])\n",
    "        opt.apply_gradients(zip(gradients, [params]))\n",
    "\n",
    "    losses.append(epoch_loss)\n",
    "    ac = accuracy()\n",
    "    accuracies.append(ac)\n",
    "    print(\"Epoch {}; loss: {} accuracy: {}\".format(i,epoch_loss,ac))\n",
    "    print(\"Current Time(s): \", time.time()-start)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'2.3.1'"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import tensorflow as tf\n",
    "tf.version.VERSION"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "quantvolution",
   "language": "python",
   "name": "quantvolution"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}