ONM-snnTorch.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "8w6lhn7H8fW5"
   },
   "source": [
    "# Open Neuromorphic Seminar - Hands-On with snnTorch\n",
    "## By Jason K. Eshraghian (www.ncg.ucsc.edu)\n",
    "\n",
    "\n",
    "[<img src='https://github.com/jeshraghian/snntorch/blob/master/docs/_static/img/snntorch_alpha_w.png?raw=true' width=\"200\">](https://github.com/jeshraghian/snntorch/)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 11928,
     "status": "ok",
     "timestamp": 1677777128539,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "1BlTqunB73-t",
    "outputId": "7332f708-5bc6-410f-b973-df669f8f002b"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "ERROR: Invalid requirement: '#'\n"
     ]
    }
   ],
   "source": [
    "!pip install snntorch # shift + enter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "cellView": "form",
    "executionInfo": {
     "elapsed": 166,
     "status": "ok",
     "timestamp": 1677777175586,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "PhO1wRCqFPlr"
   },
   "outputs": [],
   "source": [
    "#@title Plotting Settings\n",
    "def plot_cur_mem_spk(cur, mem, spk, thr_line=False, vline=False, title=False, ylim_max1=1.25, ylim_max2=1.25):\n",
    "  # Generate Plots\n",
    "  fig, ax = plt.subplots(3, figsize=(8,6), sharex=True, \n",
    "                        gridspec_kw = {'height_ratios': [1, 1, 0.4]})\n",
    "\n",
    "  # Plot input current\n",
    "  ax[0].plot(cur, c=\"tab:orange\")\n",
    "  ax[0].set_ylim([0, ylim_max1])\n",
    "  ax[0].set_xlim([0, 200])\n",
    "  ax[0].set_ylabel(\"Input Current ($I_{in}$)\")\n",
    "  if title:\n",
    "    ax[0].set_title(title)\n",
    "\n",
    "  # Plot membrane potential\n",
    "  ax[1].plot(mem)\n",
    "  ax[1].set_ylim([0, ylim_max2]) \n",
    "  ax[1].set_ylabel(\"Membrane Potential ($U_{mem}$)\")\n",
    "  if thr_line:\n",
    "    ax[1].axhline(y=thr_line, alpha=0.25, linestyle=\"dashed\", c=\"black\", linewidth=2)\n",
    "  plt.xlabel(\"Time step\")\n",
    "\n",
    "  # Plot output spike using spikeplot\n",
    "  splt.raster(spk, ax[2], s=400, c=\"black\", marker=\"|\")\n",
    "  if vline:\n",
    "    ax[2].axvline(x=vline, ymin=0, ymax=6.75, alpha = 0.15, linestyle=\"dashed\", c=\"black\", linewidth=2, zorder=0, clip_on=False)\n",
    "  plt.ylabel(\"Output spikes\")\n",
    "  plt.yticks([]) \n",
    "\n",
    "  plt.show()\n",
    "\n",
    "def plot_snn_spikes(spk_in, spk1_rec, spk2_rec, title):\n",
    "  # Generate Plots\n",
    "  fig, ax = plt.subplots(3, figsize=(8,7), sharex=True, \n",
    "                        gridspec_kw = {'height_ratios': [1, 1, 0.4]})\n",
    "\n",
    "  # Plot input spikes\n",
    "  splt.raster(spk_in[:,0], ax[0], s=0.03, c=\"black\")\n",
    "  ax[0].set_ylabel(\"Input Spikes\")\n",
    "  ax[0].set_title(title)\n",
    "\n",
    "  # Plot hidden layer spikes\n",
    "  splt.raster(spk1_rec.reshape(num_steps, -1), ax[1], s = 0.05, c=\"black\")\n",
    "  ax[1].set_ylabel(\"Hidden Layer\")\n",
    "\n",
    "  # Plot output spikes\n",
    "  splt.raster(spk2_rec.reshape(num_steps, -1), ax[2], c=\"black\", marker=\"|\")\n",
    "  ax[2].set_ylabel(\"Output Spikes\")\n",
    "  ax[2].set_ylim([0, 10])\n",
    "\n",
    "  plt.show()\n",
    "\n",
    "def dvs_animator(spike_data):\n",
    "  fig, ax = plt.subplots()\n",
    "  anim = splt.animator((spike_data[:,0] + spike_data[:,1]), fig, ax)\n",
    "  return anim\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "xLlnINAI9mgJ"
   },
   "source": [
    "*What will I learn?*\n",
    "\n",
    "**The Basics:**\n",
    "1. The leaky integrate-and-fire (`snn.Leaky`) neuron in 5 mins\n",
    "2. Train an SNN classifier using surrogate gradient descent\n",
    "\n",
    "**The Fancier Stuff**\n",
    "3. Population Codes\n",
    "4. Recurrent SNNs and Spiking LSTMs\n",
    "5. Learning precise spike-timing\n",
    "6. Hardware Friendly Training\n",
    "  - Weight Quantization with Brevitas\n",
    "  - State Quantization\n",
    "\n",
    "*If there's time...*\n",
    "7. SNNs can sometimes be a pain in the ass to train. Hyperparameter Optimization with Optuna."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "-CS50cwuCW6n"
   },
   "source": [
    "# 1. The Leaky Integrate-and-Fire Neuron\n",
    "### (...in 20 mins)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "executionInfo": {
     "elapsed": 4045,
     "status": "ok",
     "timestamp": 1677777438351,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "H_TzogsCCcSe"
   },
   "outputs": [],
   "source": [
    "# imports\n",
    "import snntorch as snn\n",
    "from snntorch import spikeplot as splt\n",
    "import torch\n",
    "\n",
    "# plotting\n",
    "import matplotlib.pyplot as plt\n",
    "from IPython.display import HTML"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "executionInfo": {
     "elapsed": 340,
     "status": "ok",
     "timestamp": 1677777546784,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "Ao9LfqkeEw-5"
   },
   "outputs": [],
   "source": [
    "lif = snn.Leaky(beta=0.8) # LIF neuron with a decay rate of 0.8"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "executionInfo": {
     "elapsed": 220,
     "status": "ok",
     "timestamp": 1677778103163,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "bixHpBamE1AS"
   },
   "outputs": [],
   "source": [
    "# setup inputs\n",
    "num_steps = 200 # number of time-steps to simulate\n",
    "\n",
    "x = torch.cat((torch.zeros(10), torch.ones(190)*0.21))\n",
    "\n",
    "mem = torch.zeros(1)\n",
    "spk = torch.zeros(1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 404
    },
    "executionInfo": {
     "elapsed": 668,
     "status": "ok",
     "timestamp": 1677778106699,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "5wTggA5XFJuI",
    "outputId": "095d961b-4b37-46eb-89db-2f6fa3e7b018"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x600 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "mem_rec = []\n",
    "spk_rec = []\n",
    "\n",
    "# neuron simulation\n",
    "for step in range(num_steps):\n",
    "  # lif neuron loop\n",
    "   spk, mem = lif(x[step], mem)\n",
    "   mem_rec.append(mem)\n",
    "   spk_rec.append(spk)\n",
    "\n",
    "# convert lists to tensors\n",
    "mem_rec = torch.stack(mem_rec)\n",
    "spk_rec = torch.stack(spk_rec)\n",
    "\n",
    "plot_cur_mem_spk(x, mem_rec, spk_rec, thr_line=1, ylim_max1=1.0,\n",
    "                 title=\"snn.Leaky Neuron Model\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "LEKAbrAACdDM"
   },
   "source": [
    "# 2. Train an SNN Classifier \n",
    "### (...in 10-ish mins)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "executionInfo": {
     "elapsed": 527,
     "status": "ok",
     "timestamp": 1677778188841,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "Z0DGl4FtChWO"
   },
   "outputs": [],
   "source": [
    "# import\n",
    "import torch.nn as nn\n",
    "from torch.utils.data import DataLoader\n",
    "from torchvision import datasets, transforms\n",
    "import numpy as np\n",
    "import itertools"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "nftOdpyAGv7D"
   },
   "source": [
    "### Boilerplate: DataLoading the MNIST Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "executionInfo": {
     "elapsed": 148,
     "status": "ok",
     "timestamp": 1677778264343,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "SsM2Z5NXGu5z"
   },
   "outputs": [],
   "source": [
    "# dataloader arguments\n",
    "batch_size = 128\n",
    "data_path='/data/mnist'\n",
    "\n",
    "dtype = torch.float\n",
    "device = torch.device(\"cuda\") if torch.cuda.is_available() else torch.device(\"cpu\")\n",
    "## if you're on M1 or M2 GPU:\n",
    "# device = torch.device(\"mps\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 443,
     "referenced_widgets": [
      "fc1cea022fd64d2580b636ad29ce74cb",
      "dab3e4d911d847128ae0a18d9041847f",
      "5163362add9c49258343be1043f69f6f",
      "bcce9f99717b4dae884fcdbe2d5ecce0",
      "175878491cd44bf6a30a707de283c7b0",
      "606551ca726b4998a55c7885b1c13e56",
      "c91df8aec8204e30837776b10c4ebe75",
      "ac978e74c8174866a72ae760b64f22cd",
      "1ba6a17991704c5ca752233b815fa3fc",
      "37653a294ae645b1b64363d27aaf496e",
      "37898c9ed3374007ac38913b228b5bb8",
      "ae7dcbfa1dbc426697a0d98608f5c6c3",
      "9c47c04ca5cf424da91533412e4e16a4",
      "4953e7632467486a9e83a71c02f2c81b",
      "5133365e8fc849f5ae14ed583c156efb",
      "ee77acb5fac84e2ab27cc310dd88182d",
      "2e78af0a8dfe4d1084bfd6974dc38268",
      "60e62bba3b9a499ea407ece948946349",
      "58e57832bb60410bb21b6e56013ac7e3",
      "8476b70863e94334ae796f5ecade108e",
      "1346b7bb93a943fe86d30de2f4dd5c7f",
      "f9cc6c36fec84e6b962d3f23f32156b4",
      "af7998e749d24f1687bb1d90c7774165",
      "828d6cac2cf54bcc88b33a1bda0d0bb7",
      "09c0667250db4a10af4b6540d847589e",
      "3288ef76407441c7a06b35e45fd5bcbd",
      "4a74bb5f467f4ef8bfcca8fd77c749cf",
      "5b542a1db8904adebe23631200aa0b06",
      "7b5359a82a114f1f9c7bb7eb2306dedc",
      "4dc765eb8fdd40379331c34b1163c7fa",
      "45b5ab7c9350401fa9f3d902c770b318",
      "e3b431043c144785a95d0e387b3f65ac",
      "02ae05cdb81f40cfb3a61915c40a113d",
      "1d0f4a482fab425b9fe3642ad8cdca88",
      "4b5e1725eaf44fa9af29c23065998cbd",
      "3a95403d6c374148b27635dde7a479e4",
      "cc751284ebfc47668e22b184d2b88432",
      "b2ff14ee0abf4ccc8463a92615805c21",
      "de12dfeeaa0d43ffb70dc035b5703fca",
      "51442d3c20444c8b82d4131a64a7f0a0",
      "ddedd2970a32439fb11dc57a9f675b42",
      "cd564c42dc3c4ed5a41df9eaa6ca9b2e",
      "69446de0b5c146c6b0d3a1fa26d09c99",
      "eacf93019f9e44bb826bff75b052ed72"
     ]
    },
    "executionInfo": {
     "elapsed": 1720,
     "status": "ok",
     "timestamp": 1677778307520,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "XqbYptgDHUPg",
    "outputId": "4300e7ae-c1b3-45d2-a6ad-499fa310615b"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading http://yann.lecun.com/exdb/mnist/train-images-idx3-ubyte.gz\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "28.4%"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading http://yann.lecun.com/exdb/mnist/train-images-idx3-ubyte.gz to /data/mnist\\MNIST\\raw\\train-images-idx3-ubyte.gz\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100.0%\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Extracting /data/mnist\\MNIST\\raw\\train-images-idx3-ubyte.gz to /data/mnist\\MNIST\\raw\n",
      "\n",
      "Downloading http://yann.lecun.com/exdb/mnist/train-labels-idx1-ubyte.gz\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100.0%"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading http://yann.lecun.com/exdb/mnist/train-labels-idx1-ubyte.gz to /data/mnist\\MNIST\\raw\\train-labels-idx1-ubyte.gz\n",
      "Extracting /data/mnist\\MNIST\\raw\\train-labels-idx1-ubyte.gz to /data/mnist\\MNIST\\raw\n",
      "\n",
      "Downloading http://yann.lecun.com/exdb/mnist/t10k-images-idx3-ubyte.gz\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "100.0%"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading http://yann.lecun.com/exdb/mnist/t10k-images-idx3-ubyte.gz to /data/mnist\\MNIST\\raw\\t10k-images-idx3-ubyte.gz\n",
      "Extracting /data/mnist\\MNIST\\raw\\t10k-images-idx3-ubyte.gz to /data/mnist\\MNIST\\raw\n",
      "\n",
      "Downloading http://yann.lecun.com/exdb/mnist/t10k-labels-idx1-ubyte.gz\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "100.0%"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading http://yann.lecun.com/exdb/mnist/t10k-labels-idx1-ubyte.gz to /data/mnist\\MNIST\\raw\\t10k-labels-idx1-ubyte.gz\n",
      "Extracting /data/mnist\\MNIST\\raw\\t10k-labels-idx1-ubyte.gz to /data/mnist\\MNIST\\raw\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "# Define a transform\n",
    "transform = transforms.Compose([\n",
    "            transforms.Resize((28, 28)),\n",
    "            transforms.Grayscale(),\n",
    "            transforms.ToTensor(),\n",
    "            transforms.Normalize((0,), (1,))])\n",
    "\n",
    "mnist_train = datasets.MNIST(data_path, train=True, download=True, transform=transform)\n",
    "mnist_test = datasets.MNIST(data_path, train=False, download=True, transform=transform)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "executionInfo": {
     "elapsed": 2,
     "status": "ok",
     "timestamp": 1677778316817,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "jSlS3gWZHXI0"
   },
   "outputs": [],
   "source": [
    "# Create DataLoaders\n",
    "train_loader = DataLoader(mnist_train, batch_size=batch_size, shuffle=True, drop_last=True)\n",
    "test_loader = DataLoader(mnist_test, batch_size=batch_size, shuffle=True, drop_last=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "SJVhfNukHbsp"
   },
   "source": [
    "## Construct a Fully Connected SNN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "executionInfo": {
     "elapsed": 185,
     "status": "ok",
     "timestamp": 1677779136871,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "uu324fr_HhxV"
   },
   "outputs": [],
   "source": [
    "# Network Architecture\n",
    "num_inputs = 28*28\n",
    "num_hidden = 1000\n",
    "num_outputs = 10\n",
    "\n",
    "# Temporal Dynamics\n",
    "num_steps = 25\n",
    "beta = 0.95"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "executionInfo": {
     "elapsed": 5302,
     "status": "ok",
     "timestamp": 1677779648808,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "CkM1Z1EjHeW8"
   },
   "outputs": [],
   "source": [
    "from snntorch import surrogate\n",
    "\n",
    "# Define Network\n",
    "class Net(nn.Module):\n",
    "    def __init__(self):\n",
    "        super().__init__()\n",
    "\n",
    "        # Initialize layers\n",
    "        self.fc1 = nn.Linear(num_inputs, num_hidden)\n",
    "        self.lif1 = snn.Leaky(beta=beta)\n",
    "        self.fc2 = nn.Linear(num_hidden, num_outputs)\n",
    "        self.lif2 = snn.Leaky(beta=beta)\n",
    "\n",
    "    def forward(self, x):\n",
    "\n",
    "        # Initialize hidden states at t=0\n",
    "        mem1 = self.lif1.init_leaky()\n",
    "        mem2 = self.lif2.init_leaky()\n",
    "        \n",
    "        # Record the final layer\n",
    "        spk2_rec = []\n",
    "        mem2_rec = []\n",
    "\n",
    "        # time-loop\n",
    "        for step in range(num_steps):\n",
    "          cur1 = self.fc1(x.flatten(1)) # batch128 x 784\n",
    "          spk1, mem1 = self.lif1(cur1, mem1)\n",
    "          cur2 = self.fc2(spk1)\n",
    "          spk2, mem2 = self.lif2(cur2, mem2)\n",
    "          \n",
    "          # store in list\n",
    "          spk2_rec.append(spk2)\n",
    "          mem2_rec.append(mem2)\n",
    "\n",
    "        return torch.stack(spk2_rec, dim=0), torch.stack(mem2_rec, dim=0) # time-steps x batch x num_out\n",
    "        \n",
    "# Load the network onto CUDA if available\n",
    "net = Net().to(device)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "p8qBw03rHpn3"
   },
   "source": [
    "## Training the SNN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 11870,
     "status": "ok",
     "timestamp": 1677780142773,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "1telBMU-HrIg",
    "outputId": "df7256f5-0460-48c0-83a0-224c0f1f14f1"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Iteration: 0 \t Train Loss: 2.7357726097106934\n",
      "Iteration: 10 \t Train Loss: 0.7314499616622925\n",
      "Iteration: 20 \t Train Loss: 0.5139610767364502\n",
      "Iteration: 30 \t Train Loss: 0.4090139865875244\n",
      "Iteration: 40 \t Train Loss: 0.40244409441947937\n",
      "Iteration: 50 \t Train Loss: 0.2850271463394165\n",
      "Iteration: 60 \t Train Loss: 0.406260222196579\n",
      "Iteration: 70 \t Train Loss: 0.2408822625875473\n",
      "Iteration: 80 \t Train Loss: 0.2902601361274719\n",
      "Iteration: 90 \t Train Loss: 0.17323006689548492\n"
     ]
    }
   ],
   "source": [
    "loss = nn.CrossEntropyLoss()\n",
    "optimizer = torch.optim.Adam(net.parameters(), lr=5e-4, betas=(0.9, 0.999))\n",
    "\n",
    "num_epochs = 1 # 60000 / 128 = 468\n",
    "counter = 0\n",
    "\n",
    "# Outer training loop\n",
    "for epoch in range(num_epochs):\n",
    "    train_batch = iter(train_loader)\n",
    "\n",
    "    # Minibatch training loop\n",
    "    for data, targets in train_batch:\n",
    "        data = data.to(device)\n",
    "        targets = targets.to(device)\n",
    "\n",
    "        # forward pass\n",
    "        net.train()\n",
    "        spk_rec, _ = net(data)\n",
    "\n",
    "        # initialize the loss & sum over time\n",
    "        loss_val = torch.zeros((1), dtype=dtype, device=device)\n",
    "        loss_val = loss(spk_rec.sum(0), targets) # batch x num_out\n",
    "\n",
    "        # Gradient calculation + weight update\n",
    "        optimizer.zero_grad()\n",
    "        loss_val.backward()\n",
    "        optimizer.step()\n",
    "\n",
    "        # Print train/test loss/accuracy\n",
    "        if counter % 10 == 0:\n",
    "            print(f\"Iteration: {counter} \\t Train Loss: {loss_val.item()}\")\n",
    "        counter += 1\n",
    "\n",
    "        if counter == 100:\n",
    "          break"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "executionInfo": {
     "elapsed": 164,
     "status": "ok",
     "timestamp": 1677780404506,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "nHsdkuSlIS1E"
   },
   "outputs": [],
   "source": [
    "def measure_accuracy(model, dataloader):\n",
    "  with torch.no_grad():\n",
    "    model.eval()\n",
    "    running_length = 0\n",
    "    running_accuracy = 0\n",
    "\n",
    "    for data, targets in iter(dataloader):\n",
    "      data = data.to(device)\n",
    "      targets = targets.to(device)\n",
    "\n",
    "      # forward-pass\n",
    "      spk_rec, _ = model(data)\n",
    "      spike_count = spk_rec.sum(0) # batch x num_outputs\n",
    "      _, max_spike = spike_count.max(1)\n",
    "\n",
    "      # correct classes for one batch\n",
    "      num_correct = (max_spike == targets).sum()\n",
    "\n",
    "      # total accuracy\n",
    "      running_length += len(targets)\n",
    "      running_accuracy += num_correct\n",
    "    \n",
    "    accuracy = (running_accuracy / running_length)\n",
    "\n",
    "    return accuracy.item()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 4737,
     "status": "ok",
     "timestamp": 1677780415197,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "oJHAltRCKGyx",
    "outputId": "998dcc8a-0cbb-4110-8417-38dd9735eb09"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test set accuracy: 0.9321914911270142\n"
     ]
    }
   ],
   "source": [
    "print(f\"Test set accuracy: {measure_accuracy(net, test_loader)}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "eFvZh1neMLtq"
   },
   "source": [
    "## Convolutional SNN Classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "executionInfo": {
     "elapsed": 177,
     "status": "ok",
     "timestamp": 1677780860094,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "cjnd_xroMt0o"
   },
   "outputs": [],
   "source": [
    "# Define Network\n",
    "class ConvNet(nn.Module):\n",
    "    def __init__(self):\n",
    "        super().__init__()\n",
    "\n",
    "        # Initialize layers\n",
    "        self.conv1 = nn.Conv2d(1, 8, 5, padding=\"same\")\n",
    "        self.lif1 = snn.Leaky(beta=beta)\n",
    "        self.mp1 = nn.MaxPool2d(2)\n",
    "\n",
    "        self.conv2 = nn.Conv2d(8, 24, 5, padding=\"same\")\n",
    "        self.lif2 = snn.Leaky(beta=beta)\n",
    "        self.mp2 = nn.MaxPool2d(2)\n",
    "\n",
    "        self.fc = nn.Linear(24 * 7 * 7, 10)\n",
    "        self.lif3 = snn.Leaky(beta=beta)\n",
    "\n",
    "\n",
    "    def forward(self, x):\n",
    "\n",
    "        # Initialize hidden states at t=0\n",
    "        mem1 = self.lif1.init_leaky()\n",
    "        mem2 = self.lif2.init_leaky()\n",
    "        mem3 = self.lif3.init_leaky()\n",
    "        \n",
    "        # Record the final layer\n",
    "        spk3_rec = []\n",
    "        mem3_rec = []\n",
    "\n",
    "        # time-loop\n",
    "        for step in range(num_steps):\n",
    "            cur1 = self.conv1(x)\n",
    "            spk1, mem1 = self.lif1(self.mp1(cur1), mem1)\n",
    "            cur2 = self.conv2(spk1)\n",
    "            spk2, mem2 = self.lif2(self.mp2(cur2), mem2)\n",
    "            cur3 = self.fc(spk2.flatten(1))\n",
    "            spk3, mem3 = self.lif3(cur3, mem3)\n",
    "\n",
    "            spk3_rec.append(spk3)\n",
    "            mem3_rec.append(mem3)\n",
    "\n",
    "        return torch.stack(spk3_rec, dim=0), torch.stack(mem3_rec, dim=0)\n",
    "        \n",
    "# Load the network onto CUDA if available\n",
    "convnet = ConvNet().to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 16612,
     "status": "ok",
     "timestamp": 1677780910089,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "5fGM99IBOnuO",
    "outputId": "6a16caee-1bc6-4f26-ccb4-0b80ca5bd914"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Iteration: 0 \t Train Loss: 2.4256603717803955\n",
      "Iteration: 10 \t Train Loss: 2.3025853633880615\n",
      "Iteration: 20 \t Train Loss: 1.9744020700454712\n",
      "Iteration: 30 \t Train Loss: 0.9479921460151672\n",
      "Iteration: 40 \t Train Loss: 0.553931474685669\n",
      "Iteration: 50 \t Train Loss: 0.24587005376815796\n",
      "Iteration: 60 \t Train Loss: 0.3799990713596344\n",
      "Iteration: 70 \t Train Loss: 0.18676407635211945\n",
      "Iteration: 80 \t Train Loss: 0.26160889863967896\n",
      "Iteration: 90 \t Train Loss: 0.18616953492164612\n"
     ]
    }
   ],
   "source": [
    "loss = nn.CrossEntropyLoss()\n",
    "optimizer = torch.optim.Adam(convnet.parameters(), lr=1e-2, betas=(0.9, 0.999))\n",
    "\n",
    "num_epochs = 1\n",
    "loss_hist = []\n",
    "test_loss_hist = []\n",
    "counter = 0\n",
    "\n",
    "# Outer training loop\n",
    "for epoch in range(num_epochs):\n",
    "    train_batch = iter(train_loader)\n",
    "\n",
    "    # Minibatch training loop\n",
    "    for data, targets in train_batch:\n",
    "        data = data.to(device)\n",
    "        targets = targets.to(device)\n",
    "\n",
    "        # forward pass\n",
    "        convnet.train()\n",
    "        spk_rec, _ = convnet(data)\n",
    "\n",
    "        # initialize the loss & sum over time\n",
    "        loss_val = torch.zeros((1), dtype=dtype, device=device)\n",
    "        loss_val = loss(spk_rec.sum(0), targets)\n",
    "\n",
    "        # Gradient calculation + weight update\n",
    "        optimizer.zero_grad()\n",
    "        loss_val.backward()\n",
    "        optimizer.step()\n",
    "\n",
    "        # Store loss history for future plotting\n",
    "        loss_hist.append(loss_val.item())\n",
    "\n",
    "        # Print train/test loss/accuracy\n",
    "        if counter % 10 == 0:\n",
    "            print(f\"Iteration: {counter} \\t Train Loss: {loss_val.item()}\")\n",
    "        counter += 1\n",
    "\n",
    "        if counter == 100:\n",
    "          break"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 6084,
     "status": "ok",
     "timestamp": 1677780932861,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "j9i9VIMeQGNO",
    "outputId": "776a969e-1773-417f-9b99-c64f03053f7e"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ConvNet Accuracy: 0.9357972741127014\n"
     ]
    }
   ],
   "source": [
    "def measure_accuracy(model, dataloader):\n",
    "  with torch.no_grad():\n",
    "    model.eval()\n",
    "    running_length = 0\n",
    "    running_accuracy = 0\n",
    "\n",
    "    for data, targets in iter(dataloader):\n",
    "      data = data.to(device)\n",
    "      targets = targets.to(device)\n",
    "\n",
    "      # forward-pass\n",
    "      spk_rec, _ = model(data)\n",
    "      spike_count = spk_rec.sum(0)\n",
    "      _, max_spike = spike_count.max(1)\n",
    "\n",
    "      # correct classes for one batch\n",
    "      num_correct = (max_spike == targets).sum()\n",
    "\n",
    "      # total accuracy\n",
    "      running_length += len(targets)\n",
    "      running_accuracy += num_correct\n",
    "    \n",
    "    accuracy = (running_accuracy / running_length)\n",
    "\n",
    "    return accuracy.item()\n",
    "\n",
    "print(f\"ConvNet Accuracy: {measure_accuracy(convnet, test_loader)}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "CwlMRIFpCiXq"
   },
   "source": [
    "<img src='https://i.ytimg.com/vi/Rt8LTxysC2s/maxresdefault.jpg' width=\"400\">\n",
    "\n",
    "# 3. Population Codes\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "executionInfo": {
     "elapsed": 167,
     "status": "ok",
     "timestamp": 1677781129545,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "ZAXtcMpRDwk4"
   },
   "outputs": [],
   "source": [
    "# Define Network\n",
    "class Net(nn.Module):\n",
    "    def __init__(self):\n",
    "        super().__init__()\n",
    "\n",
    "        num_steps = 1\n",
    "        num_outputs = 10\n",
    "        population = 100\n",
    "\n",
    "        # Initialize layers\n",
    "        self.fc1 = nn.Linear(num_inputs, num_hidden)\n",
    "        self.lif1 = snn.Leaky(beta=beta)\n",
    "        self.fc2 = nn.Linear(num_hidden, num_outputs*population)\n",
    "        self.lif2 = snn.Leaky(beta=beta)\n",
    "\n",
    "    def forward(self, x):\n",
    "\n",
    "        # Initialize hidden states at t=0\n",
    "        mem1 = self.lif1.init_leaky()\n",
    "        mem2 = self.lif2.init_leaky()\n",
    "        \n",
    "        # Record the final layer\n",
    "        spk2_rec = []\n",
    "        mem2_rec = []\n",
    "\n",
    "        for step in range(num_steps):\n",
    "            cur1 = self.fc1(x.flatten(1))\n",
    "            spk1, mem1 = self.lif1(cur1, mem1)\n",
    "            cur2 = self.fc2(spk1)\n",
    "            spk2, mem2 = self.lif2(cur2, mem2)\n",
    "            spk2_rec.append(spk2)\n",
    "            mem2_rec.append(mem2)\n",
    "\n",
    "        return torch.stack(spk2_rec, dim=0), torch.stack(mem2_rec, dim=0)\n",
    "\n",
    "        \n",
    "# Load the network onto CUDA if available\n",
    "popnet = Net().to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "executionInfo": {
     "elapsed": 2,
     "status": "ok",
     "timestamp": 1677781147559,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "5BO9chvmXhmv"
   },
   "outputs": [],
   "source": [
    "from snntorch import functional as SF"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 9067,
     "status": "ok",
     "timestamp": 1677781257214,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "qeQ9o28dXNqT",
    "outputId": "976ce458-aee4-413d-a30b-734761b931d6"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([25, 128, 1000]) torch.Size([128])\n",
      "tensor([[[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        ...,\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]]], grad_fn=<StackBackward0>)\n",
      "Iteration: 0 \t Train Loss: 0.2676335275173187\n",
      "torch.Size([25, 128, 1000]) torch.Size([128])\n",
      "tensor([[[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        ...,\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]]], grad_fn=<StackBackward0>)\n",
      "torch.Size([25, 128, 1000]) torch.Size([128])\n",
      "tensor([[[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        ...,\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]]], grad_fn=<StackBackward0>)\n",
      "torch.Size([25, 128, 1000]) torch.Size([128])\n",
      "tensor([[[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        ...,\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]]], grad_fn=<StackBackward0>)\n",
      "torch.Size([25, 128, 1000]) torch.Size([128])\n",
      "tensor([[[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        ...,\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]]], grad_fn=<StackBackward0>)\n",
      "torch.Size([25, 128, 1000]) torch.Size([128])\n",
      "tensor([[[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        ...,\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]]], grad_fn=<StackBackward0>)\n",
      "torch.Size([25, 128, 1000]) torch.Size([128])\n",
      "tensor([[[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        ...,\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]]], grad_fn=<StackBackward0>)\n",
      "torch.Size([25, 128, 1000]) torch.Size([128])\n",
      "tensor([[[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        ...,\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]]], grad_fn=<StackBackward0>)\n",
      "torch.Size([25, 128, 1000]) torch.Size([128])\n",
      "tensor([[[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        ...,\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]]], grad_fn=<StackBackward0>)\n",
      "torch.Size([25, 128, 1000]) torch.Size([128])\n",
      "tensor([[[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        ...,\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]]], grad_fn=<StackBackward0>)\n",
      "torch.Size([25, 128, 1000]) torch.Size([128])\n",
      "tensor([[[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        ...,\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]]], grad_fn=<StackBackward0>)\n",
      "Iteration: 10 \t Train Loss: 0.4306793808937073\n",
      "torch.Size([25, 128, 1000]) torch.Size([128])\n",
      "tensor([[[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        ...,\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]]], grad_fn=<StackBackward0>)\n",
      "torch.Size([25, 128, 1000]) torch.Size([128])\n",
      "tensor([[[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        ...,\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]],\n",
      "\n",
      "        [[0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         ...,\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.],\n",
      "         [0., 0., 0.,  ..., 0., 0., 0.]]], grad_fn=<StackBackward0>)\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "Cell \u001b[1;32mIn[25], line 29\u001b[0m\n\u001b[0;32m     27\u001b[0m \u001b[38;5;66;03m# Gradient calculation + weight update\u001b[39;00m\n\u001b[0;32m     28\u001b[0m optimizer\u001b[38;5;241m.\u001b[39mzero_grad()\n\u001b[1;32m---> 29\u001b[0m \u001b[43mloss_val\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mbackward\u001b[49m\u001b[43m(\u001b[49m\u001b[43m)\u001b[49m\n\u001b[0;32m     30\u001b[0m optimizer\u001b[38;5;241m.\u001b[39mstep()\n\u001b[0;32m     32\u001b[0m \u001b[38;5;66;03m# Store loss history for future plotting\u001b[39;00m\n",
      "File \u001b[1;32m~\\anaconda3\\envs\\classifier\\lib\\site-packages\\torch\\_tensor.py:492\u001b[0m, in \u001b[0;36mTensor.backward\u001b[1;34m(self, gradient, retain_graph, create_graph, inputs)\u001b[0m\n\u001b[0;32m    482\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m has_torch_function_unary(\u001b[38;5;28mself\u001b[39m):\n\u001b[0;32m    483\u001b[0m     \u001b[38;5;28;01mreturn\u001b[39;00m handle_torch_function(\n\u001b[0;32m    484\u001b[0m         Tensor\u001b[38;5;241m.\u001b[39mbackward,\n\u001b[0;32m    485\u001b[0m         (\u001b[38;5;28mself\u001b[39m,),\n\u001b[1;32m   (...)\u001b[0m\n\u001b[0;32m    490\u001b[0m         inputs\u001b[38;5;241m=\u001b[39minputs,\n\u001b[0;32m    491\u001b[0m     )\n\u001b[1;32m--> 492\u001b[0m \u001b[43mtorch\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mautograd\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mbackward\u001b[49m\u001b[43m(\u001b[49m\n\u001b[0;32m    493\u001b[0m \u001b[43m    \u001b[49m\u001b[38;5;28;43mself\u001b[39;49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mgradient\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mretain_graph\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43mcreate_graph\u001b[49m\u001b[43m,\u001b[49m\u001b[43m \u001b[49m\u001b[43minputs\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[43minputs\u001b[49m\n\u001b[0;32m    494\u001b[0m \u001b[43m\u001b[49m\u001b[43m)\u001b[49m\n",
      "File \u001b[1;32m~\\anaconda3\\envs\\classifier\\lib\\site-packages\\torch\\autograd\\__init__.py:251\u001b[0m, in \u001b[0;36mbackward\u001b[1;34m(tensors, grad_tensors, retain_graph, create_graph, grad_variables, inputs)\u001b[0m\n\u001b[0;32m    246\u001b[0m     retain_graph \u001b[38;5;241m=\u001b[39m create_graph\n\u001b[0;32m    248\u001b[0m \u001b[38;5;66;03m# The reason we repeat the same comment below is that\u001b[39;00m\n\u001b[0;32m    249\u001b[0m \u001b[38;5;66;03m# some Python versions print out the first line of a multi-line function\u001b[39;00m\n\u001b[0;32m    250\u001b[0m \u001b[38;5;66;03m# calls in the traceback and some print out the last line\u001b[39;00m\n\u001b[1;32m--> 251\u001b[0m \u001b[43mVariable\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43m_execution_engine\u001b[49m\u001b[38;5;241;43m.\u001b[39;49m\u001b[43mrun_backward\u001b[49m\u001b[43m(\u001b[49m\u001b[43m  \u001b[49m\u001b[38;5;66;43;03m# Calls into the C++ engine to run the backward pass\u001b[39;49;00m\n\u001b[0;32m    252\u001b[0m \u001b[43m    \u001b[49m\u001b[43mtensors\u001b[49m\u001b[43m,\u001b[49m\n\u001b[0;32m    253\u001b[0m \u001b[43m    \u001b[49m\u001b[43mgrad_tensors_\u001b[49m\u001b[43m,\u001b[49m\n\u001b[0;32m    254\u001b[0m \u001b[43m    \u001b[49m\u001b[43mretain_graph\u001b[49m\u001b[43m,\u001b[49m\n\u001b[0;32m    255\u001b[0m \u001b[43m    \u001b[49m\u001b[43mcreate_graph\u001b[49m\u001b[43m,\u001b[49m\n\u001b[0;32m    256\u001b[0m \u001b[43m    \u001b[49m\u001b[43minputs\u001b[49m\u001b[43m,\u001b[49m\n\u001b[0;32m    257\u001b[0m \u001b[43m    \u001b[49m\u001b[43mallow_unreachable\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;28;43;01mTrue\u001b[39;49;00m\u001b[43m,\u001b[49m\n\u001b[0;32m    258\u001b[0m \u001b[43m    \u001b[49m\u001b[43maccumulate_grad\u001b[49m\u001b[38;5;241;43m=\u001b[39;49m\u001b[38;5;28;43;01mTrue\u001b[39;49;00m\u001b[43m,\u001b[49m\n\u001b[0;32m    259\u001b[0m \u001b[43m\u001b[49m\u001b[43m)\u001b[49m\n",
      "File \u001b[1;32m~\\anaconda3\\envs\\classifier\\lib\\site-packages\\torch\\autograd\\function.py:276\u001b[0m, in \u001b[0;36mBackwardCFunction.apply\u001b[1;34m(self, *args)\u001b[0m\n\u001b[0;32m    275\u001b[0m \u001b[38;5;28;01mclass\u001b[39;00m \u001b[38;5;21;01mBackwardCFunction\u001b[39;00m(_C\u001b[38;5;241m.\u001b[39m_FunctionBase, FunctionCtx, _HookMixin):\n\u001b[1;32m--> 276\u001b[0m     \u001b[38;5;28;01mdef\u001b[39;00m \u001b[38;5;21mapply\u001b[39m(\u001b[38;5;28mself\u001b[39m, \u001b[38;5;241m*\u001b[39margs):\n\u001b[0;32m    277\u001b[0m         \u001b[38;5;66;03m# _forward_cls is defined by derived class\u001b[39;00m\n\u001b[0;32m    278\u001b[0m         \u001b[38;5;66;03m# The user should define either backward or vjp but never both.\u001b[39;00m\n\u001b[0;32m    279\u001b[0m         backward_fn \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_forward_cls\u001b[38;5;241m.\u001b[39mbackward  \u001b[38;5;66;03m# type: ignore[attr-defined]\u001b[39;00m\n\u001b[0;32m    280\u001b[0m         vjp_fn \u001b[38;5;241m=\u001b[39m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39m_forward_cls\u001b[38;5;241m.\u001b[39mvjp  \u001b[38;5;66;03m# type: ignore[attr-defined]\u001b[39;00m\n",
      "\u001b[1;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "loss_fn = SF.ce_count_loss(population_code=True, num_classes=10)\n",
    "optimizer = torch.optim.Adam(popnet.parameters(), lr=1e-3, betas=(0.9, 0.999))\n",
    "\n",
    "num_epochs = 1\n",
    "loss_hist = []\n",
    "test_loss_hist = []\n",
    "counter = 0\n",
    "\n",
    "# Outer training loop\n",
    "for epoch in range(num_epochs):\n",
    "    train_batch = iter(train_loader)\n",
    "\n",
    "    # Minibatch training loop\n",
    "    for data, targets in train_batch:\n",
    "        data = data.to(device)\n",
    "        targets = targets.to(device)\n",
    "\n",
    "        # forward pass\n",
    "        popnet.train()\n",
    "        spk_rec, _ = popnet(data)\n",
    "\n",
    "        # initialize the loss & sum over time\n",
    "        print(spk_rec.shape, targets.shape)\n",
    "        print(spk_rec)\n",
    "        loss_val = loss_fn(spk_rec, targets)\n",
    "\n",
    "        # Gradient calculation + weight update\n",
    "        optimizer.zero_grad()\n",
    "        loss_val.backward()\n",
    "        optimizer.step()\n",
    "\n",
    "        # Store loss history for future plotting\n",
    "        loss_hist.append(loss_val.item())\n",
    "\n",
    "        # Print train/test loss/accuracy\n",
    "        if counter % 10 == 0:\n",
    "            print(f\"Iteration: {counter} \\t Train Loss: {loss_val.item()}\")\n",
    "        counter += 1\n",
    "\n",
    "        if counter == 100:\n",
    "          break"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "executionInfo": {
     "elapsed": 1,
     "status": "ok",
     "timestamp": 1677781365057,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "9IF5HPHvX3YH"
   },
   "outputs": [],
   "source": [
    "def population_acc(testloader, model):\n",
    "  with torch.no_grad():\n",
    "    model.eval()\n",
    "    running_accuracy = 0\n",
    "    running_length = 0\n",
    "    for data, labels in iter(testloader):\n",
    "      data = data.to(device)\n",
    "      labels = labels.to(device)\n",
    "\n",
    "      # calculate accuracy below\n",
    "      spk_rec, _ = model(data)\n",
    "      running_accuracy += SF.accuracy_rate(spk_rec, labels, population_code=True, num_classes=10)\n",
    "      running_length += len(labels)\n",
    "    \n",
    "    accuracy = running_accuracy * 100 / running_length\n",
    "    \n",
    "    return accuracy "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 5106,
     "status": "ok",
     "timestamp": 1677781373396,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "0zwNSfORY-Rl",
    "outputId": "e51ab260-213b-4f10-c196-869b52d67972"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Population Code Accuracy: 0.720527844551282%\n"
     ]
    }
   ],
   "source": [
    "accuracy = population_acc(test_loader, popnet)\n",
    "print(f\"Population Code Accuracy: {accuracy}%\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "_3-y-Z6pDw-W"
   },
   "source": [
    "# 4. Recurrent SNNs and Spiking LSTMs\n",
    "## 4.1 Neuromorphic DataLoading with Tonic\n",
    "\n",
    "[<img src='https://github.com/neuromorphs/tonic/blob/develop/docs/_static/tonic-logo-white.png?raw=true' width=\"150\">](https://github.com/neuromorphs/tonic/)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 6346,
     "status": "ok",
     "timestamp": 1677781406471,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "PgGW1kliBgYK",
    "outputId": "e55513c2-42f1-4136-8046-4105c308b2cf"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m104.1/104.1 KB\u001b[0m \u001b[31m5.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m50.4/50.4 KB\u001b[0m \u001b[31m4.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m112.7/112.7 KB\u001b[0m \u001b[31m11.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[?25h  Preparing metadata (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
      "  Building wheel for importRosbag (setup.py) ... \u001b[?25l\u001b[?25hdone\n"
     ]
    }
   ],
   "source": [
    "!pip install tonic --quiet "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 363,
     "referenced_widgets": [
      "53a57f276a61483aa96fc195ee485187",
      "3bd55950664d4e6e96a3dd553fac73f1",
      "2da74e44fe2b4b50aad73fdf7f02c3b6",
      "e478c3957c964155ab2148e79260ef09",
      "a78f7c36cc7d4f2788f5d3a98d500520",
      "3dc418e477bb4846840ea683f3541c50",
      "c32e21f24e9e4efb80393aabde1b2679",
      "12f42925d27245feacbbb67b71b111c3",
      "649a0f85a7e4403997aaa2b2ada8d887",
      "e08217b98d414bcd8d155181631bc695",
      "f74089a9d8494b0d9e0a365bb8b5ad37",
      "9d2488cd936c487eac69fe639ce6a77d",
      "c4c7fa18359041dfa9960668fb884f69",
      "b622c6be305a4d99a2a3ffda2babbb0e",
      "12c820a3950645d08234b0f62e737a8f",
      "0783e220986744a598f0c7c034b1905d",
      "f1cf4115d2944e40897bf6e29d7dbe8f",
      "85c95cfc8352416fa745f160bce87195",
      "c4631d6cbaa349108dd27edfc95006e3",
      "47ec1d064cbd4fd79628900377476b89",
      "68106930402a441eb79acfd1da06acc8",
      "6943b4e4ae504532a9dc7c67da00b874"
     ]
    },
    "executionInfo": {
     "elapsed": 12540,
     "status": "ok",
     "timestamp": 1677781425470,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "Vz3UPUG-Bfv3",
    "outputId": "54405b16-2192-46e3-99c8-fa23ff5694dd"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading https://www.neuromorphic-vision.com/public/downloads/pips_train.tar.gz to ./data/POKERDVS/pips_train.tar.gz\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "53a57f276a61483aa96fc195ee485187",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/700096 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Extracting ./data/POKERDVS/pips_train.tar.gz to ./data/POKERDVS\n",
      "Downloading https://www.neuromorphic-vision.com/public/downloads/pips_test.tar.gz to ./data/POKERDVS/pips_test.tar.gz\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9d2488cd936c487eac69fe639ce6a77d",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "  0%|          | 0/311022 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Extracting ./data/POKERDVS/pips_test.tar.gz to ./data/POKERDVS\n",
      "[(    0, 20, 21, 0) (  106, 21, 19, 0) (  146, 13,  9, 1) ...\n",
      " (15365, 13, 17, 1) (15384, 16, 16, 1) (15408, 16, 22, 0)]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import tonic\n",
    "\n",
    "dataset = tonic.datasets.POKERDVS(save_to='./data', train=True)\n",
    "testset = tonic.datasets.POKERDVS(save_to='./data', train=False)\n",
    "\n",
    "events, target = dataset[0]\n",
    "print(events)\n",
    "tonic.utils.plot_event_grid(events)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 265,
     "status": "ok",
     "timestamp": 1677781467074,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "xOtYPzuUBlhG",
    "outputId": "1415d99f-382d-4547-818f-a2b1d2da9118"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(35, 35, 2)"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tonic.datasets.POKERDVS.sensor_size"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "executionInfo": {
     "elapsed": 182,
     "status": "ok",
     "timestamp": 1677781645701,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "8JnjGcLSBokg"
   },
   "outputs": [],
   "source": [
    "import torchvision\n",
    "from tonic import DiskCachedDataset\n",
    "from torch.utils.data import DataLoader\n",
    "\n",
    "# time_window\n",
    "frame_transform = tonic.transforms.Compose([tonic.transforms.Denoise(filter_time=10000), \n",
    "                                    tonic.transforms.ToFrame(\n",
    "                                        sensor_size=tonic.datasets.POKERDVS.sensor_size,\n",
    "                                        time_window=1000)\n",
    "                                     ])\n",
    "\n",
    "batch_size = 8\n",
    "cached_trainset = DiskCachedDataset(dataset, transform=frame_transform, cache_path='./cache/pokerdvs/train')\n",
    "cached_testset = DiskCachedDataset(testset, transform=frame_transform, cache_path='./cache/pokerdvs/test')\n",
    "\n",
    "trainloader = DataLoader(cached_trainset, batch_size=batch_size, collate_fn=tonic.collation.PadTensors(batch_first=False), shuffle=True)\n",
    "testloader = DataLoader(cached_testset, batch_size=batch_size, collate_fn=tonic.collation.PadTensors(batch_first=False), shuffle=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 160,
     "status": "ok",
     "timestamp": 1677781646997,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "wfb6ec3lEr6n",
    "outputId": "998d9c62-1931-4c47-d517-5a3ac4f21906"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([20, 8, 2, 35, 35])\n",
      "tensor([1, 0, 1, 0, 2, 3, 3, 1])\n"
     ]
    }
   ],
   "source": [
    "for data, labels in iter(trainloader):\n",
    "  print(data.size())\n",
    "  print(labels)\n",
    "  break"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 540
    },
    "executionInfo": {
     "elapsed": 1903,
     "status": "ok",
     "timestamp": 1677781675733,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "8ZwZLPHfwV5a",
    "outputId": "b1c5a45e-28f4-4f39-b350-b06281034947"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<video width=\"432\" height=\"288\" controls autoplay loop>\n",
       "  <source type=\"video/mp4\" src=\"data:video/mp4;base64,AAAAHGZ0eXBNNFYgAAACAGlzb21pc28yYXZjMQAAAAhmcmVlAAB/pW1kYXQAAAKuBgX//6rcRem9\n",
       "5tlIt5Ys2CDZI+7veDI2NCAtIGNvcmUgMTU1IHIyOTE3IDBhODRkOTggLSBILjI2NC9NUEVHLTQg\n",
       "QVZDIGNvZGVjIC0gQ29weWxlZnQgMjAwMy0yMDE4IC0gaHR0cDovL3d3dy52aWRlb2xhbi5vcmcv\n",
       "eDI2NC5odG1sIC0gb3B0aW9uczogY2FiYWM9MSByZWY9MyBkZWJsb2NrPTE6MDowIGFuYWx5c2U9\n",
       "MHgzOjB4MTEzIG1lPWhleCBzdWJtZT03IHBzeT0xIHBzeV9yZD0xLjAwOjAuMDAgbWl4ZWRfcmVm\n",
       "PTEgbWVfcmFuZ2U9MTYgY2hyb21hX21lPTEgdHJlbGxpcz0xIDh4OGRjdD0xIGNxbT0wIGRlYWR6\n",
       "b25lPTIxLDExIGZhc3RfcHNraXA9MSBjaHJvbWFfcXBfb2Zmc2V0PS0yIHRocmVhZHM9MyBsb29r\n",
       "YWhlYWRfdGhyZWFkcz0xIHNsaWNlZF90aHJlYWRzPTAgbnI9MCBkZWNpbWF0ZT0xIGludGVybGFj\n",
       "ZWQ9MCBibHVyYXlfY29tcGF0PTAgY29uc3RyYWluZWRfaW50cmE9MCBiZnJhbWVzPTMgYl9weXJh\n",
       "bWlkPTIgYl9hZGFwdD0xIGJfYmlhcz0wIGRpcmVjdD0xIHdlaWdodGI9MSBvcGVuX2dvcD0wIHdl\n",
       "aWdodHA9MiBrZXlpbnQ9MjUwIGtleWludF9taW49MjUgc2NlbmVjdXQ9NDAgaW50cmFfcmVmcmVz\n",
       "aD0wIHJjX2xvb2thaGVhZD00MCByYz1jcmYgbWJ0cmVlPTEgY3JmPTIzLjAgcWNvbXA9MC42MCBx\n",
       "cG1pbj0wIHFwbWF4PTY5IHFwc3RlcD00IGlwX3JhdGlvPTEuNDAgYXE9MToxLjAwAIAAAAnXZYiE\n",
       "ADv//vdOvwKbRZdqA5JXCvbKpCZZuVJrAfKmAAADAAARsG58lw15q9IkAACtobhz+EtXgAGbPfnA\n",
       "6vBq2y4KD1VFSYr8QiGU1RZ6il6SV2xj+fxRT9T5XCfupsCCGI9rMzLTf6dgHpGzFSOW3HfPJLi7\n",
       "mZ16P29yPgAbEL8gPtkzMM0eWMliDR1baRxx/vGMIv0YqwzznR6obGUfYwk61NX0dLiTPLUb5tHc\n",
       "U4t36dfjTBzEgnKT5yWzQpkNWiR+AX7n+fKpOoC0itea6Ktx85NxWpqSOHgmQjc2iah/HyxLggnV\n",
       "QHvIjmZzF7JXZiER7dnAAaKCnBulS5QkHjMdERyniFdrN7IGfkc1mDssm0wTZ1eB5AbDPUpjhGFr\n",
       "/o/06ubq2eju4UKuPSoZCP5NkmQfG2h9ziO+5q9Vmn86ZcPy2R5CGQxPLDZyQzAXY6KNUBFxIqQa\n",
       "uWZ8J07N3EGz3kj8o/PIBleuj6m4jSvAMHJKfDqIntLA6dpM60LC3TvfhcbCEsHgannNoTu0vm2g\n",
       "ssQ3+EzwDf6Eb+lHfn/x408E3O/3AT2bLN+xnqgZsL5UoC19zx4ZvXbRzdXgi/bI5MVsxEaFA1uS\n",
       "B2wrpKEwzQ0tnrFvnQGIwwc5LB+ftnphuejZQdGfm+c5q3H4MW8YFBo5d2NGTmZkptCG5dCD5z19\n",
       "YXPKMH5lgsgKfHBru6bmStIp9CYkk94iNYYgUjLVzEySzMulCWJfgsnGAkCjNvaETSgcoEcBpJon\n",
       "oN48jBxTa1YIUkgF0AlGEnLAVVGBzMrjVonXpbVenx4KJ3KDeOAbxhKft4O9FtQYu9CPA8SSWjQS\n",
       "mdJY7hpDjDbrwd+6BrHaOS3diOyr3ONWtFzAlK8/Ooybu12Z8BXW949KSLPYZRWJmGZ8tCfWo2eV\n",
       "tYMLzboQ83lNYbXmcxZ4e0cKx+fUnB+ZT/KM+HqLRB5nr+8ETsBALcoPg5AZvjhCtkVLvH5xxq1Q\n",
       "/JmD0cqgwHUK6SKQovE5vFwgJ7bwQOJ53FsDttGNBnqz2NF2/Bk+ILUTpBEf8BPdvs7I01aztDHM\n",
       "PrYltHluJKfM6ZMaaRDW2Eg05uCLZAB8T2GMCHPVR8gjLzJblSl0xYEF7SlVTM0o6HgiYilg/Ck6\n",
       "FVehbSoph6djf4ClbKgjCNEzescgN97dA4AMhTshf876GZcAcy1GPUNmO8c35wj56xTxlTXt3wJs\n",
       "mEfZkzRz7L5AUierNWgUQCT1SsT6thr+U4E4uNctkH7VLxYJoZxD4Di6PWG/HJ26Q52zrkCxZplq\n",
       "FAOU3r7ncKTvd5yQnaKgCho2IebXOw+V4ogjDako0H3F7A2mW6G3618alKsv5fX4+5Iwr5cKtFSG\n",
       "ypKg7pf04ty8VOw7HRgt+FuKtLx4FSrdor/Xav25R+GTXW4EodlT+rz5P1SQ7/5et9jmjJtQC+n9\n",
       "alXrxrHg+kvtnoBvGVRxwQjrae7T7/4+pUH4HsLisadFwG/l/fVpSe7uAhZqZkajBcgLDEvOowD8\n",
       "WdyZOFDTx275Xb4Ji3Y28kUcRy4MqDeSVjAD5HKB8NZ3eRiVYmVVBF100oj2RHgZnKrHuqT3O9B1\n",
       "J0b3kNjgM7kFvHpphWvGZHU/x3xvk3cfI3m3OWrnBHW3Gpp8oi42k66EbfdGdZJ3x7UscPicRxSw\n",
       "SUd58S7y4TxRNqW6/m0WigXFd1G8dQSsCua8sZSukHBcoC/iWToel0aLfoxiasP5geJXvxTgSWgL\n",
       "JjU/F18JrWlx9t53MFcXehf1wjPCpST2lu+SQVsdGQ56So9POkHsGJ2nC/4c4CCqXwpV2GyBTPp2\n",
       "ODM7JEfZEyc5qoxJCOHKLR96+6unu9b/cLsn4+9xdKL7hCKm2eKTarLeV5b8R545dGy3LAJlkq+W\n",
       "NC/Nwh9AgUs50RHS7qKjc/egPlhAxQQlEatvRjICdYtMe3nVHNUXwJTLQPs6BPQflDNk+T1IWr6s\n",
       "+mA9BEuPuB5a7Uz/gUTUBH3WWgLFatQXsJCWqsnAHBOpauDNiwYripqejuWswLaFbCtKSGAerZ0w\n",
       "QwAYv/2MmQr7ttO9OvdPa4eHEYafElDh7zirql4P3+2q9e9w/Ph4OLVHiu7j2KNYGHKvjXRhLdEr\n",
       "nfum7+5w8nWQxnrkzG6RtD7CCj+eINGn8zpsczc6qYdgSJ9l3KZuRiuq7wJK/ATlr/f+/TqEOrgM\n",
       "rWo1NXClHvB7PkT6qNWZemSa/yovnyEfjYctYqSIPmMk3ELWEYBi05QdI7YH4ozBIfH8Xym9GoYL\n",
       "E4tAe2JsFlbj/MHPasJeNc1crFcaio6JgrXW9BRaCAx/DYgfxiW9cW8rFYhQDwpSjgeiRwEIhR0Z\n",
       "faGMuaUN7hr6/Kcg1kkiRpiwj8rRTC0ByU+TErDEiPuaBe95SZc3nT37dudXr961gLdpZTnXxFTE\n",
       "CiiTWe075GLwTP4Pg4vsOCpZVVJql4NhBbGNzlbyUqm9/ng/aSKVcCYkeQ9hcOL/E87OSfYzjJ7k\n",
       "emuXw0C5zPCXwDueDDz9dsaEOEpUQrWmm+ALnfZ6E2d9lrkUKci9jBA4QJsz+qyszZuTiQSHI0Xk\n",
       "x/HYdiRMSZalRx4zKKJao6Syp80nXg5tA7LFjlN2h5Zbg5uduxWsXDXNdCbnmNyPpxm4o9tKfh1q\n",
       "PQQFqBsDckhPJmu+CpQAkmqEDKFaxyREsVfaWPZ8QytWb4seNsVXTpd/2f6gnzKXOGR4ki9ZfD6F\n",
       "Q59IGPcSM5hDrURlM++sldi9zp/kbbzJdTT37RUzptMnbnbAoL0OdLX4/J97poEGre51BrrFor/s\n",
       "ABNrgg4UlTUALeEZ55PEqjgbMae+voevRhhIfVkHmue0USFJ+VHYzWsAe9njSX9npv1h5up1Hp5O\n",
       "oPVA3KBWntC0F0NTAL1I976KXkd2xSBQfRoI/aDIRaYlfX+lQtVkl4VFISxLcL//NTHZZr5mwFTK\n",
       "/QOTLkY8DOCVejXG+6svfWySH4z4MxJspiYTLxyfsvynh2ueot2f4Y3HyF00OfO2jcfvkZbjgF4H\n",
       "7gwbt1T+hCGJT5nZu6yNx9c8Mz7s9rAayFsoC3p5ZfUfWa2wetKpZJQTzuk3th4tKOu0fJnVmDih\n",
       "fJaOa0aaUbmpo4CdX6ArVCdythWO57V4/3yuinPvMpdO1PSC9C5oNwIwr5ahsGPOHVBEuHCJt+fi\n",
       "UcdyFAuCA98gXLS7VplxCTFjPGfN+EbxVRRS0zB+Pi9SLstBmGHqeJxC4kreS8OtiwW75Mvy9/gj\n",
       "SP1L7LQAux6ucFcl7rdqI9OH30KDfW2Fodf0cmW2LdTGOPt0P+pqs9xdbwo9EOJIa/j7m12xWDMA\n",
       "AAMBnNALKKEAAAbDQZokbEO//qmWANP4+m7rOa1ZN9rEVk0LLBsylTr7pqI+ABo3w0r/0J1qLzcx\n",
       "xNes4oQi3kALFupFXI/eZ5lVLNcgwZJxLDsdifpkKz6wKpysaoA8MvmYCVAzrXgG+uChT3Yja6+1\n",
       "HZiy3bV+E/1/sDQqkKa9EtnqOrxA+P63kbZokfUyMsM4ooIlCr5bS+0nBxKBFUP3GiPHhZT2kI2r\n",
       "SNzGTgNuyXUKO3PoQE79OtZO6NnfiyvXLpMC691PrxVwE4REc2FvEn738RnMVhzX2Tkv8elaFc2t\n",
       "HF/T38pX0lmDcvoaTPNqUDF/jKmYVOf9IE6PV8UlPbVub5n4pLNvk97mVT/hFNv8D8ZTviK5sqyk\n",
       "ooOumGWtVZXDNWkmMVYFDanwrs3Gj8HB6WjxnJrR/Wsm/xy1PFLRlFqroI0P6gQJTnobyf8nKbwc\n",
       "z7Ot+axUN91YwloGZIrX5CtcWJLaJg8TGXGzYEOz2zocNyHMunSPAscPZc5aznZ9Xn9ywlrqOAn3\n",
       "311AtVB84GwJAG4BGY/7iS4lnNOEPjPzL+fRGSHEaGmE5j9//7W/E2szuc3YQ6VFrgk2RQo1QMtT\n",
       "fymXr0t8FRXwG12MD4T86mQtZwAQ37dzL9MIhTsdEBgcxehQVX+E0me4aCkti+s2ONBmyA76cEYr\n",
       "LHAyQtco+gXMUHskI0RuuXQVVQTH/+1nPuUFUyol/gkPnd/c2u0WtZv2QnhGCYJTOTkruEWoAD64\n",
       "iCm7SRoQ9FoTaaaNyBDYzv7WMmbmnntCmifVIUIt9oYOM44yRIVPdzj6VadnDddwgADZpTY3goX8\n",
       "vTv2J9U4+rRT0qzemEVDCGuZacJTasAoIYA0FOICi2621nT3SEMEdxjDpCEB/5BrVImIStUSoS+u\n",
       "jc/w1fJHnXUdscUbhUGkbEUDyltufKGHIBw3Vib9XthMOem9bH2ziaeZcnSYjYGhk2r+Hn9pTal+\n",
       "nidXAjvbw93133JTmKg35mhlFijRTU7Ca6xiK8BKjlUtOBNsethoWGR/aQLcbzihiAGCAMKSaVXS\n",
       "zYjASrIwG2BbYKpauqnydxGRz5/ahmKX568S5rxjbZCQX0UtZpIucpz6OuBz9VFDqMnjzN8i6LmZ\n",
       "FaMmNjbBNsM3Wa878Aj41byo8+z09Z36xvxvL4iMD+RYcyLMA6IyGRlx8OX53YdvWQ5heZdCvsOq\n",
       "EwgNguT4EPlwPCk2FBnNomL+LMHq+ufpKapk0HsclvARVv0M8W+o9NpoCN0qXmySTADYk6vXeAah\n",
       "gestzIZnEIEw2bytj6Ml9tVht0MalMwEk7vUGPFWV54TXcT0K4pldpIH1oHsi+AoWHicRjyDRVyy\n",
       "+AP4Yj3hOW57vRhtHLwp/pDs30DMmX7k+Q/55Kh7lB4Fs1LjHR+LrL5SLDJBVhXn+8AitbZ5TkWa\n",
       "uGFO9w3wtk3PKQupRyTdW9Q0VzraoucBS1mnRFaYzvvwCGGFHfKMrVErAdHrtAH+jCiNy7YBMq/Z\n",
       "SLxB+sa8XcJYxWkWZ0G0ba8ymJARyGFkJjkkCfAGKQThk4ku52xd67QZOtwC790eB6HoPzCDZXpS\n",
       "4pdv98Ayix1HnNG/FF5vVzCr7DWlzpZEa47n/lhuLshQ4uPkYirQdvLBytdaEXgloymHTDmusZus\n",
       "boKv33zEsRpZi/oWnzGNy09SQ3PH30j8UHuYfGYvZtUkN3aQCpXgD6CeGr6Z4EYlyHMXzZeyUov2\n",
       "1kifwjlgLqoqqD/0zVq1owBdrRvEJV5ICQtt33JImfYBLBtHk6JohQVciz4h8FgEIWaY9s/fvz6x\n",
       "o9/Pa3KuK5zkcjiEtJMcO8OGAVgGk4PTNsnuhTNhcZiW9h/dBdDyr4kRaDz4oriCTcd4K6FoQ6xL\n",
       "rW5XZzzdCAChQOmXUBS9xiPjtpL75CjtY1lz2qdZXchHufyZMb8kA9fwj15ILsTISgGfhQ2jTQ2m\n",
       "44r2v8ASmrIkYn0SwoLCoI/SHfYo1GcZRsfnfXvrZLyOJW3DuZ5cQ0OOGl3f6/wz0u8EPMhvWFOD\n",
       "nykGOF20KBMjB8DlHvfeA9U+sHTFbxWA7AaojJU/c32yI/FNM3d61PASmsST9G0uyt7WMkY8rnft\n",
       "s3Oycycr8FDx0WTZJSsZhFccu34+tz3rUa8OPJoGsdlZqI4JosWiQOd22m6Q8D0ikc7lfc8/CC65\n",
       "dtIkEwfMzaBvTOKnwlA/VVx4hLAIPWeAs8JD9xroe4fifD/aRDQo7vFbqVN6e7Lc5h8AHvHjbT0e\n",
       "6S1s6PNPMjfhCcGqTwa6hxJ65f5QISIKfoQnGaIxhYjgAAAH/UGeQniF/wDnfQdp4Hf8JvnYAGWy\n",
       "VHQhdEusFZsKF9x5K4uRXvs35vKes6cL6IJOPBRC0O02keQjJ0VGRUp6IIpjAas/PdH6lYa/0MzQ\n",
       "tGKisw+xi9Um2WtqwalcIfXacRbb47HOQaAvQNhNXXojy5hRJFBoYsG/N7aoAOd6YLvEt854+zox\n",
       "X6TyNEfgE+xsNdJTEQwAe7xQed4A7xow342gsbSnc+FuJxnVTxosgHnNVdugyPNc5OGlJHDtyNVj\n",
       "a4PzztgPu5p6GmdiikuwO/cYGOpHDmCCJzotoSZaIPIp47lPesSGnpEftWXiYOb9pgHQ1WrQmVOK\n",
       "HGhvtLweU/kGsSc94hV/qCiO+tLhEefmMpTLq3A/KQiL9v+G0lgH/djgDJa/aT0wFttQTcrorZMR\n",
       "krG9mcu1BCSgRQU/3EKTm+ZeYMkOpp7pgXRsCIdCOD54jEQqTvLeFF0BWIRMaohPhmAM3JjWoPJZ\n",
       "lKghpvE36DtJRjmnBW49hNKHuo8gh5+/Kvu6xI7ZorBTEHY5wIBdeI/ASLDUA6tOO0ylOG4QV6x+\n",
       "OW+yVa25AE1xb/bsH100sR3QwMQD+sHLFcH4EWascjFJ5lzakTFgWNr/DLQicxoBi7ddikWqQ2jT\n",
       "7Iz1PFLakovGQcLGbomoQjZQzwr5XTfVMWm0NVT+xX/tsP/PONVvBnD7zGuGw/98behvpxnFuGus\n",
       "pr8PsCI43L5b2kw5HKgTnGfzcsGtCyg+GA9jh+G+7r+pf2mohv/AK8oYlghqtM5vtF8H/RApr8x/\n",
       "Nin7PIAz3/YXpIOVaFFG+gEysS7XqRm+0gVd6OKfH1QSzITWBCmpVkCzI80FqNYXAT9H9z8jyyCE\n",
       "zYWMb1/kUCI42A8Mlx5aZ91EN6+IxtZCwYrv8owBLPwCvtJDILSHTJWSxYk+rxTSDLPc5rl8Ttcf\n",
       "+NICqxazFUdfzDD4gsurzi6cih3c26kO2LEr3CCtx2/nnIgLIdRbBFatNWcFIhFK8zORc6Li+xUb\n",
       "c3gwDMBVbSceqGiZ+fD7f5KSUxAWRWWE3WWsWCInAUlxESD/UqzyTESsW81iTBbhdpu/QlGB4ofa\n",
       "BSbicH/ph3LNQdZRLtlp4SiKpa15uyEx5JXTE6E66ktZMsKG9mEECnDglwbtP9cfIrC2+0e3pgxq\n",
       "3XCbD53rd4U8IUqWn5TBe0+94Zro+x3MoOABJeqYpEwLKza01MBXmHzlUczef/fYwUYCjZR4PR7j\n",
       "Dkf2wEUVSgJwTZeWJy88Z5t9l7t9qOfZx0/dgz+k4gfIoy+IBr8WaqLOgxGMrH/9JsbIqKaCCHq+\n",
       "LPPVREkPxaYAwk/f7ssyAXsFutsdn96VcVWSxcirhnZ9DjMTl1/+ejcNPIrlwTjix8ULH1Jtvwgf\n",
       "Cbp69UPG8TZ4018aiKK47yRnaCK6qNiSMaUl2FqcqxUaIQ9QhKAOseakcvp5qMCccjoUT9ZBDIho\n",
       "k4V/d4jCKBH3PsZgcZ/Of+fsEVP3TLGXCZvv1EONGY52y/XRxDese9TRC0Em/Biq5A5U0lHZa2EX\n",
       "zlqOPnyklDQ4dHl7mL4R+pDNOHzjyEvRjUrOdRz5pp7wCPXjjCFH9h+uQJMrbsbccgcga9t/QaRz\n",
       "C9tcBct8MoUt2Xtp5CpXYVlILazORhDR2Roi3M7GTOXmK4Xv/CFel7xysxrDrt3BVNFEgUtk/pYX\n",
       "Y68bP5HhexC7VRaq/sV02cLtA52lIh/qQK8IvRNp4OJpI8IIcrK2tlimlzI2q8/rZ2jfqrooEey6\n",
       "WEKLZf1z13e5ci3S553QZtjnbL8etxcKD42RPMOHPqaiMt8kqL2c5p0QwHUReTWT2bwX1NPcF4O/\n",
       "r7zpp3d9NYCyP3qvs8yOqSqN1KwixQiPmyuw5d7lPHqox2IiYY1APoU4N71j1OiB0mMjhqkEZNyw\n",
       "JLFFMjbbIkze/ago55JgREwkMel2guy7BglViLMWOK952hyC9mb6ia6p/CTGYWHl1INK+UlXNgUW\n",
       "GHXtaezHMV91OBbPF42RpdCU74HxbvlDg5N5X0DYD8zmeaiiypzAHFxorUHpsPJIDCraXPGGAgfp\n",
       "MoEUJAss0LhDjuovBVqot5VfQ4wqkdPYcu8KQ66EIcyHJplKL4Zck1MamsOYjcCgaBaBdrp0VKDK\n",
       "Deq2EgiJX4vDKYp7z/gD8zDXx1ejxzbNMAALGC5UEi47LSwMnsZN2OEB5/6ZTVACI7mjdnDlhG6O\n",
       "WRycrolPAaQBKX6D3phwBx+8p7fq8+32r8TMkqSeijX8SVKI8UlAUE8ZOm8s3sUKpBB9dxAxsDc8\n",
       "v+ZK7F9P9t7zCWXgqBh6sYgaNX4Ar7mGZtHPp29tI2GZoEVpenWw1Xr569FEzvPbv93GoRJsteBK\n",
       "ocDRdHzUi78eoAOsg3STibUwipLXqsTJEqOkMqq/3ZmENrsBK16e31RUZp96b8MG4IhSd14MTcj6\n",
       "SotS4Y4OouIHaz19HEZ8Fz5KCgIrkU/O0ONLG15FOFNoJKOHS4btThzDVTQRlBd5Ub4HMjB+5+89\n",
       "SIZS6g5zFFaI93L661OblS95/HtnEU+nybffTSDt3R5Qqo4kIvgt+5a58gKaTFzXwBvePAAkp+PI\n",
       "auPioKXUm7shDuQag4fKV3/40icU5sGfEPBXb4v5A4I7mDKEpkVWIfSsE/jjU83+m5mFE1sFBnF/\n",
       "7GF3v2kC9J3Sk6ICyYmD6z5/rx/ZiH64unsYV8ahAAAHAQGeYXRCvwJfDcwkvudemtZf+IVg0ABz\n",
       "haNGV+VyyrySWfya60fQQJTMffPP37CYRNnUUeCxny4Ajn5XX+oFyFKkbEAy/p1s/ipdBCLsz/bq\n",
       "8jXp7mdWiob232dv6N1Cu5IWCuDQ7onhglJCwBybIsTUJLUqB2SAQSyPT5NjhrDqU7ZlRIa0Mp6a\n",
       "M6AwyYWHOD6XSJ2ihWRU8Vim83dMH/F8K2jWPYKKbrQp8HUgxp737u1oAedu/AglkXr+HS3z/hXv\n",
       "N+PIkvLP3pWot3idmnQTGayddB8NVLMTM4gHCgMpizFPVPFHJJHhtP/ksuAlOTUFgBZcIKxAdRIW\n",
       "svHC3C8phgMAD/kj0e94tt4uqnNuoPNhDNRBji6koHTKvVxz2xpP9yHiSldOpRRv0emv4BTpHlur\n",
       "ggkXOB0O+9XmczoEXGjs6eyTslwCHMw0+dNjtiuWToq7KeAOTsTdWgF2OWZritCpxJEE9tQSgNEB\n",
       "Y0+Hxrlh8X6upMFUE44DRQ9cDdwhClBy08mgX5wWGzyMBde0l8dOd7Kfhm2tMKJZ+LTjAbNPMnLM\n",
       "hhz20DZzrpGC4if8GlJPtleCI/WrK/TdYh6+ZOBMu8bG01WNjJokGWPXuK9Sn5MsjEt7M8Cwm/rG\n",
       "k/DeP/HfQgmrZmJ5j//SFZdaBcNh0s32OqN0Ixf1cr+j4g71PcydFMIF54n/7rgOa8e/QccKV7bo\n",
       "Pgh2aGgN+m1/fJ+e16RIHKPQzuyhWQI1LRBkONxIIwpMQ7Ds6fg+ZZTYqvVmY8OLh15VuNnGyMFI\n",
       "vo1ZjwJI66mD/7gKmUTLiIpF7S1tyekxrXLZeiwjJ++sqqRQsLL5SCKBZxmouUaFF7s+ktmCn315\n",
       "0g7z3u8yCAiVW5h3cpFzOsFnH9OZEmSc7vhifLkvsKnZscQNH2/Q9942m9Ihb7W9zbMEBUG96nUe\n",
       "PC211Y62lcVw2L4dbA3M/u1Z782Ml1PpESirdTvKsc1OewCjrcZ2j/0MOfNJskX5foPba/1lBJ5J\n",
       "epav+bfIEjVszKQfieth3v8cLfT1/xJIjkINbk3gBAq+6Klz3x09t+5IA4fdh1NWAHsn5NlWqHre\n",
       "N2D6b+Oj4l6jmQK64lblwkJzGhEyli/ZMW0ARQv/yIdFybSq1640tiupDD/aF0+eR44V+hp5aotd\n",
       "GSPFC5FjbbJIIgAQa2GghERdqO0yc+2dWZ0DzXoOm2gv41rxpc+dIv8KYYwFfC4v4tlr82XcGfLd\n",
       "n78d9j0wQ++8brkxFwcPWV5KXiRBUaw5qJn96WMI7gjbrZNFh1HYrGaD7MyXk3smjB+O9U1A6Xbj\n",
       "Sg6COXIQvV9tIRHBTHucLsS2p5uCXqKB3ENSCW0oLTbaZxIA5n2UgzNcu+RwUGuRdqAE43sLL1JB\n",
       "LkXzsONaPNt+64JcMtRuo5T/XuBEPDvrNPLTUXoi/qP8p5b6bhk1f7VypREyyr1ts885rByB0bMb\n",
       "dJD/8wy3K8b2a6takjkDYc45wJn+bFiEAzYSKh2NOIF1xrlw24cmMcelBA2bF7X38M1yb0/KDeQP\n",
       "tuu46u7DuKcdgvUXdhuQW07yC2TvQe2YFPwEyJKeKqXuJz592GI9ZAidmbbRTd1Kw7jKgA7yAbzb\n",
       "h5vTzMQlt0cZrfReQCPPrxbx5q/DMrACmXD3pAGMN+/qYGg1VQs/8QqzxWVmXBg8/9IDFBUIxben\n",
       "NBO9vVcflKoWpW4J/yPKs/G6zd8sINpEiSH5207n+MTZmfmihD7Om0Hmq/McAlpzMItE1yYHXoAL\n",
       "hwaDuny3cj+QEPzslKmq7ri0c5TRnLrSGPWkBF5GWCEJNYQHtbTspQFpYoOqEaUi6usPwnt8cKdX\n",
       "4zPK+jbj6fWW7Fspp0msmL2uDIXoWAQ2+WtSOeHNks7kJiR+mysGImvgKx10vO1nHLjJt7Jc9z+a\n",
       "IUDlanm3wZRIzCH1wIyOagT/hEHlNFpy7br6tjxIkz1qepnG1jP+CgzIKJUBXRHm09rrFI9SIEqr\n",
       "ioYaEz9ga6O+QfJPXm+dgPjx5reOovrgzFG2lwKoslSDAJ1NZR/io90ztTEYUWQU4gAcv1vAwcbh\n",
       "4JilPTW/5y94qCcgDXi6RGBykzhFb6Hwsm6QnIvzuTDn18ZwYX3D2T6torI3lGd1PsgLE5PPNLN3\n",
       "ldj/EUQLc/Mv/uRO2gxd05En2jJJgGliLURrtWlWLMLr/FmsIQiWfmyf7SEdxcwubsB4WW0iY34m\n",
       "9abRObQHNwVfjNurmZFqdxKM48z3NfFZ6hvyFmsORaJFRZPQ9eMVP2E02HanqIezbAIvvMdkPWFG\n",
       "VIFGJSarEECcnd8R29McbRTpjr3OrzzxbK5fmw0aCFQvx/O7cWedysMncsmVzOxX7dYe2JjTIz3G\n",
       "3ohAAAADCgGeY2pCvwJeA7Qgf5WUug5YNGuVVDfvQHywPIAAU2qBaxhUftWvJJ+VnkfXwMEXaxBb\n",
       "3KiDMWtF0MWSmeiAghEhSVA4xT42igT9L8m8cduyLuBh0WTVzhk4/wvzCIZy33yO5b8Lw+3UtlVh\n",
       "4qcLxiVPYR+8G7Q/QanmlJoFiGYiA88B49MDVpfCTziiys/J+yw3QUq2568VjQrSdsf09+CbsFMM\n",
       "Q+aOmhuq7OIwk2ntXf58SGo3AiIyZjqU3mIlauYOiN1Tz7MwQlT9oXrzL65puoqBP7J5twkKrne7\n",
       "ScVunvcL0iF74zE1fDWgJZGJv+udu6QoOiFhJZbCUKNgcae64muR4GTM73BYcMlt+UI14Lw9yL8L\n",
       "fbVuJbpYiPOeI9oR78SAjDyXfbc4qHAxbjk5i/G24DRI/MiI7sMIWnKSH7ovAtl3AfzMCmsSKwpQ\n",
       "W7KOAAG+w8IU1+VHiVTgW/5c39s7EAdrh1jZSVK6ecE1F27WEng3wX5dZO9RjX41HEwSz8u3ui3Z\n",
       "GNOXgs9eOyGf9FQk7aWFlzunPlNUp3SYu8KCyMRBem8nQ7nPlAlRdmsMnvGGyLdtcvvflvhLKElo\n",
       "E87zmuk+Re6PCNhbfWcsrvovFtNgGfZM0C4/yVJqeAmdNI4ULouNUcOwW7uEsf7n50R6txQRLNs/\n",
       "Fjr6QPbAFDHbLC2W9UQBlEv53olviB16uwWA+kElru2VSyPm0Gum5zi/MrL2aMOmjkYhUQgcSzBP\n",
       "PYpJ/3K4ZURMU+gITVQlhFDSTShFMXzWySlcwtHj5Umidiupce5lOxCJ3V6MGF2d75ScL5DIuwRS\n",
       "fKTe76uFJmfuPEVvBPgHDWch+nCujz4slEH8jcKCBkj/XdAcOwR6YNIsHG61KVncexAxmsSdVMdH\n",
       "MhAyc0QJDSkDAA+0Ny62W5EnjGIiLXFmXC7/j7n8KYbOLsMqf+EPY84CMmQFxexDG3wRTQBayExA\n",
       "a2TljSX4d+6IL/y4DZY/bj/f7wqVih3h7rKOHITssgd8AP2g3b1IJ16T9FMAAAmoQZpoSahBaJlM\n",
       "CHf//qmWAMv6DmvBhgACRxnB+od3ByeCzY/39eJXMJTFH/0T+x70MTZVI/yOcNQwuGvdnOXLn9va\n",
       "UcYYgJiBerQ5JnkO20aZ8wdxMHq+rept/ANX9yfAm6QRwVmi8vTmqTYW7j6up6tsw+HjVT+iy+94\n",
       "uEY++dk3ATUBDKWf8BlRJQIu+4WSN9jq4rWiiT96baPutx+wKYm7CkIXcUh377jvM01v5y0522su\n",
       "zcLTW7q69vGonTIGeIYX1OSIg7xb0m07XzwioaxvnDZ6Ym9ZA7sXFGPqyEyEqenJBv4ynaAwZ/LD\n",
       "xYvL38wqZrTlWsaI7McJjBTTLaJMUAQrjCYZNorVgXuTuN3L6wkOMciSqHPmHJorE+i+jRqKs6Jj\n",
       "Jgl1EUgHzfEbXpK7z21TMt3GPkv5DdAZaC2XnsarbQrpxQkgG5nUdjgYEzA7slbvbtxC55EJUhQV\n",
       "fe/ZLmfnZA9NYzWk2aZcCoJ/EnT99G1bSPSDPUGHiTdAU+7/o1Y2GXdI4uYUiH7gVPvqiZZZc65D\n",
       "1l67+ioCgqEYN1Uu29QTom3rfJjglOaJ9ffR/AI4VBalzodRNdJmh+0qq+xpBMsyglQ9geSojkOB\n",
       "GxBnuMXT6TscOFMeTECaNTjIkUPdDIZx5BJBVxffEXOBVV5ZFUhZW7f5Re4qVIKvLJMAPGK8aehD\n",
       "e2EECJZe7k2kVb/3UxWcsCxY1vxfFLgCCtehFevE4QAGi1wQxOqOdNiaGymVJpokE2BF3ef+o/mR\n",
       "G8kgONMyJA31R++mu5+NLLVqgo+pc94+znCseYjO6l1nmdsEokFVhu6QU7Y6h/xCljbG0UKEOMpk\n",
       "swFspViSQ50q4RG9+pP1VWZHAi4phy5IHBQqlMCTWpIeAIhrsiiCqOQ/sq6IFem5Kdbv/tWdDQTB\n",
       "oJP8XzVLtTOQddLH0b99QM8cbiSAZ2pWTdHQXjOIuQ2V/EUIw/xWzkJ8aj0wZHg3oyTC7K37pXqb\n",
       "I13mq6S0i00yoIXUIX25lbiF71CR5U7KGRCCte9PHbwPMsGHeWcaDQ6zkpx6Sv0hDIiRzaf7/NyD\n",
       "fciM7IS1cxUtba1SPpSYEaTN3auNJq0b2lNojan/2R9IyoLLKHXzF3JGPyLRxnlFVLBttJG/wJN6\n",
       "guIesdqsXe5n1DvgFCgkRngfWhx+Ur2SbBPajAyYmmvfScGW9qBvQ3IaqLWyGCZ3Xp5Id13xL2zG\n",
       "nNzKpXs8Z1+SwSn+0LKR+lbY4s9gVZHM+yI2rhokZpEDCcvHo9X3A+MsLWM6xfN0n7jJbS1N9ui/\n",
       "lr4OXVyt2bnwk0c5pqiqRVodrvIXkAXcDHatt3IH3eEjKbnFCXJP4awfY2ln6oMJf3jSDVdslZki\n",
       "YfHgl/z5ZIhpUIltIVqnraItIF4EcAhRPzSzC1BfWXPvYDeYAfkPoj2VW5BHT259Gc2c5mboXZSG\n",
       "Ie/4SfEUjfO9fiKHaLZ2CdUonsdCGHpx5X1SyGwmrJopZl78Zi6OZspg5w4m4gsoGXz/WJxyTyCB\n",
       "re1i8PliYOJcx9PqiazZmEshYdXo8WtC5XzaCpjyT6YpmqnZuEw6JqEKZguGfKZSBx455vmmePkF\n",
       "SC7zApY/c3ntE3RM5hhfXxSTOm5mRHOSgQPN8yxIoT8ykRbJrrpwYTBCqDPA2Po6XIDY93wbGyex\n",
       "4wmg+WI5Cj45S1Xk2JUjhXut4uob73csu6ab26DccUMd7JYD6nB0UZRNP2Uupu/Fl7JFyM1cmphM\n",
       "npkE574EQX4pETJZ1k/aney7Z1Oo5WFihxXH+DgpbQcJ1YVVLcwi8YzIV+6DfeEIkZ1fKysNKU4S\n",
       "z1nedRsrKmiy6qDFXJsDL8K7NVoJwKMIRK8FHvYtAYOf0M6knU97XK7jTzuVgZIwDhhfZYwy1zIX\n",
       "XSxVz2UdmcKLtQjdX6PGWb2eH0b2fXuN3HEo6Wrih4a8zPVap0KPkaKtVd+rUNsCWhEoMRHdy7fX\n",
       "AQFXI22tb93QVoddYzlvU2D8Q8ys1s7sbcvKe8ArWF8iatHRTclpaYgqtxjYo4p6ywSP0gQr9Neu\n",
       "drLh9rEpChEhUQjELKrUOahtmWIxQ/Zx0PJ1Pc6IQnuI6MHB5dMWIv3QFHU1F7cRHiOZz0KkcR30\n",
       "LyJRDjburdp9Wo2SL7UtJiYrAfb+k2TMQs6nZBqAY3yhXBM3KkXBZH1In3+Jn2q9lBOAOSgsQsub\n",
       "lJ063PEcEw6B1TWN0ZFN1GLNqdXnTfuAjwNesAzGRyXmZ25ZFglSIxGv8JWjVnE+migZhr/BER8f\n",
       "Fh0KSJVda9670C2+IXxzu6dWMEazahOc2Vh27RMsm00yMpLsLFer9gC2l9GA79D+pUltKg8n5BDa\n",
       "xgnhKKRAg4MdN6BZtpj1pDD674ayq/KRqxOeHIHtlZqvThv4LgUeMZiWcy3QDQJpem0a0UQimQie\n",
       "L01NJZzJwuw/ameCLLvqY9EoqzEUyb9z1xdmjsdPQkbwTEWOhXUdFh4zrAFssqdQ3zHgab7cvp3A\n",
       "jctcuEFDH4DtVh5hitzwXOzMNlQ6po4zebGO13xiG1s31HId3MwxNEIXEjzPmER8fuOYtIX31JS6\n",
       "X+E6JtLbTJEjIei9ehvRGaMERtMPw/PKOy/P6+5X8RchqG1/XZJRasxqfFGHXvdLLTh+H69LH3PL\n",
       "F2QLkPZCsD3UDZP3/5BJMNuuXQGen8mMHHxfEI1xKv4PW6WJvq2DTnCibuKo+NOGLcASAPtiJWBN\n",
       "PQoiVhvbMuL76xSmk6AENv/TZCQK4tN8FPSv56DVe/LfscaDlRHsSwYMT3W3oxYzlijFxTJuEz+5\n",
       "cyMBH2FYEfTCMcBO+owpxUGHbmIWETsZMYVi7GBZL45dc22wjvE3jaXWdp0DWLuJgV2fSIaZs7Ux\n",
       "cBL24bGm7sxN5WEUJb7zdEhrhAJSoJV1cKYY/3jhluWWmfkEiHGjDLbqAgJrD6X4fpCGCrgLTNDj\n",
       "vLIo+1iHFcbdoPBSEwi5kLE+cayVpUBvftxJpjboLejY8BdgdKAGJ/0CiRmUJoMzKBImUBTi+atB\n",
       "Wd3IBh7HN77JV2z1SBreX0xGb01jRr8ow9Lem7wNZwqDTpkFVeJ3NFgScQbEuyCjhytWpj+SJ9ki\n",
       "wB7VMznvot6m7o+kmRynxfQUT64ItwHc9mo0XvmQagAbwdvTX17WumDxdWWZpEQ2CC8rTdrGajFn\n",
       "htsYZW+CIN6KG/2p1MGCGBh2WC4a5aG0eK4Jfb2y6FClGXM84bfyzVaK4x/dk7GRzzJrcw1HJEju\n",
       "FcEhH91eUALrZkVBAAAHK0GehkURLC//ANVwhB55ZTbDZ3eHPorDJuG4zXq1oaABs8sXZej7vMDy\n",
       "5Byh8e4n1L0o/vSKhz/x8r5uU25kx8PWqQql0xaZxykt7GhJpH1g/yrYfdFc3NXGIZpAU/GlT4v2\n",
       "9LPwwfLz1lU7BeuvxI/EtQ1tgTHyA6gl37nPoh46BLM2JN+Rj/tm/ax+GoUyzE9n5J/6txD50+fp\n",
       "pm5t14qxxxqS2H4UNaQfSB4enG9bH0Kxd9w/CTS34K2OgjZx6YWg/XBGVpGuXuwnl+oYvU7EfOUx\n",
       "CDyPawM1bidxCezxJPZzadycbG6oLZce2E9Qm1JaR4xqBv61sqXdT72/qUoEcDp4TCFb1kjkF8LE\n",
       "w6jSj7FWn7kEwZHf5tR0KKYAR9pMAztJI/Vd0+e72FP/dOS8FJsfgO4vqWyEwYdIkHQG/X0NgI2E\n",
       "oFJuezkmlH/SrOJcRBiGz0ZAZUpPn6oA0GhGxrRmaHzIa+CEYY9AY6xSwGIImOiG1mG2g+NXk7xs\n",
       "Kt64gJxZgcBlOH7jpSNeJbLuhoGq1GT1+TPcwTY2MxIx6cBNL0hYahvgcWUaY2rVEgOQXOEje3W7\n",
       "mTLXxb687PBD9h8KXGZfPNirXuRMuxkbS9GUgF4WR3xvB6JwCOL3+JMLLQacYUhodlB63y1ZqIYO\n",
       "2tL1H30DAPjH9UPF8FIA4Z/aIA9/tWHLxnPEX9ae5N7pI0ZDKCwRYKLfwEsFS/eaVVMZGUP6XYTM\n",
       "T7fP78M2hqlf/+MFRIqKRhwOEw+UUHg+AgBJN210/LI3fKdEboKx0X8W5YWBNKF6pQdsaa+H5Cr3\n",
       "jhKHpBV2WHIEiEAis7iM8Zhfdt2VMyvb3v/fIaaLc9kxfAw5NwmuZhoukvCrMt58/rVQ8Y5ols6Q\n",
       "ZTKHNES2qhxB/byqqZxg1GBfWFcIGYmUieD20PXj53uLJ34hRK3wfMUEuSs+OsOLoxSVmGqCy4no\n",
       "1beftqaOCoZL0RDUwiFJiUn5Y2f4mgMsW7OU5Z2fpeE9O8zY/6Nw0jyvnzZQGR33BiKM0yPFVRsY\n",
       "fmPJjwUaLNEwmZqSMfsDDUPTN5oOJ94jcWxUSqcKCniUuYySZr9ZgOhy/Y/S8otyxlJtX+xRcVHI\n",
       "ubz97ytnqpZfxceB66XKzP2ziEkXpysC8DweFXCUxldQWlLd6r2baDrN2gemwnkh38kBUOhcXfDb\n",
       "HZ+FInwFhRrTZMjQLkGE3EZLhA1KfifbaiB4VpeK6XfEsB1p6Zz+CDgOciwDFT/n2hz6m71eODF+\n",
       "jEK55pDCSWHlBxCaxE6sEJerY1tUtaZ8JbWaqJivDSh1ejQQJuyFNay2zcdG2mADi0LUKLKrG/4a\n",
       "3osZGUXVVrGEe3/CaYUfw/TtDjzmkxptL9zFFcw+hDhZ8TeNVCZKOlD+1aKtfREHLM8c2mB5nfRZ\n",
       "+XR0zJ3iUIPlXlM53XY4vKeiVc9ivxRlQc0jVZ2t2CkaWedaHpg1SfQQz8wrspsihdtUmjLsJLuf\n",
       "+TGBxMkL1X390AbCqGqH4Xj9eQ0Te6VILMUhI1zQuAiS6A2yCaW41eWa5ULUwmsjLVUzayzjqncf\n",
       "CKQecUmYnWorI4KbfA2JnjghTNgF3OhMiT3VGl7yOzBF+SVU4N9/WIlo7I8VZWD4VmyAk+YJkFne\n",
       "4wdDxpImmxHdW1/cXwcbGZrVQEOHoWXI3FADzHvT8IHiYy5P0DF8CG+40GitXFxxBa2/YfYrMCNS\n",
       "du7RbLV5bXa/b6bWPwxFQujQFOsp7aqsrVCLBf4CikJwYSVLH98L6IX541e4h+cjYxn62SCTspOe\n",
       "6sOPiVEuEA+CsPRM03Ddd3yScHBZYnsV+0PbBWWfPXViYNOtiJqJDwNNl1NuhYb4q8FBQ0veyv1o\n",
       "rAvOcbOoAFeXYNCUFj8r3CscHq/vi+OFTQV+ixIRGK+JnYcmVhETgEKFrNdZ5V6J5DF2i4uzcwyC\n",
       "VpPPURXvBZi7c3NXh7Rl6EPkKGoK/8FwyvYxn3QEhYLeD602bmaAe5vuSqpqWvaK0pUUK4AYBF1/\n",
       "sJ/zH5cFSDPr2nz4EK/ORKTnhyUSEkyt9dIpr7rJYtBqNLVsFku1TxrjKeO/IVMbP57rfZjxPcJf\n",
       "mrnuRcPy6SqZrOVTJGMqngh8SR8h75hTHay1RE5kMoTWEBCFcGGSgyWu/5Tun+7zv70bAnAEA4y3\n",
       "GJtiWoYaGx8Pi/Mz2FUsTq4GLOOQiFzuI6pKseCOc4c93S8Nf1AS77Cw+/SU8SdzXABdgjxQ+J2y\n",
       "BNKcxZV4as6g00H69b90mw4moTW7m71wX5I2ZgnSbw09VePQ+Gv6UCyOyA2eima96u6s8UI0jsR0\n",
       "jIiNZohKB4iSOlLx6ycVpvJuJ2Q5Lm3ADMPW37zxavlUv341MrFdJXkKg4qDy1coxrK97VRrIlEo\n",
       "gDmBlq5kGMcGeF5tzXBkY8QlC7oyBWUdYv+5AAAHGQGepXRCvwCfJbIIw/+h73IJ2bwALqOq+2Lm\n",
       "B4HRZsxst71IoJ0sCB7YJAgLt+HK0BYLCIygpoKPabbRDKgt/5zNTpWqwoLRSzLWSix7+DZzQ/OB\n",
       "+sqkcpH13w6MJ9CYnm5XZLie7ui6AC7wWMKbOOIGVThvUed/uT8l871x/NSCpTfixor0uUDqUb3c\n",
       "bvM2TWO5X/6HN47kl2CV/i/ZMmTbpT3x82Qt5JVUijI1fCqZIatabShov1NQYfSeku950ODlnsgK\n",
       "nLArKYWaFPJlm7o59oQpx2q9KinjUAnoYngbkMTU/BuWB3Q43Fu7MCAl8XtF6EXBnj4T/Yp/zW2/\n",
       "ZKH7ruEviZUHGKys1WO9Rxs1B2kNJ2xfjJVJphXsjh/zimsCOex8Nfgbf9ruNBDpKD2MGAlgdlNu\n",
       "fanROL1zS+wh5/IG7R0Fm5J1ee69Vaq2tlIFEzHzOcWhsM20TPjVDz2kSTIVqVReUHT/0hoLeMAQ\n",
       "NsGQDHMrvFr831BwSKVfjI1B6NGbrS75MrYTazHfhE9pkXhmYAB4x+JJ3AIvOUskAk69ympkD6OM\n",
       "v5lPGk5nqQc787w64Tai8F0ZwCk+J9xEnnEQszu6BDJ/yqefz5y9IA6UG9rhTSixO1MmYMU5/POt\n",
       "CNIzKb+BntOLXDlTPigC0xpTH2BinyR5KNvGR34Fw8jUcCzAVGC0g+oJp7x4mWHwseT6cwt1whbU\n",
       "1P8j2l4OgB/NGWM2BdvlnYLyNLJoS5345WerVCPzrCAwTvWSLgUfu2TyULIKNkvSvle/SGA8o5bJ\n",
       "BPZmEBO5Vl2APRvxi4kjz6dawkHaFe7EsPnDj5wM4CmMGVZMFvxIukesWvgKbpltHsn/LuqMcX4q\n",
       "5AgekFUgfEn5W6cs0ngW+oSZfTd9PUIngg4qMNxhL7Ib6g2b9/GLXCt1jUlnjaFaS9kDGFTEq54t\n",
       "2EdIIa7v14s6IhG77Mki6YPbDOX1rLVRER657L5F4UVz6Cvgrk4LxOoaq6Sd+1WtyDfxiY21bOGO\n",
       "wFsrtmqN/ERggBq/bRiFCXj/Sm/auVHA/PPb/sxZnDIGZPHXOllnVP27D5sXztDFXb6EJ3soP/Lb\n",
       "DakFdlBvGdEbjNlBAAD4GrTO0m24RMDsfSow0eFudGyIuz2PjOvTe9JfwcD+33P1n62j23qscbcP\n",
       "CcHt0+xeIwtsakKtNfpGIFGpmI75emBhrhspn4UFghT0zrYcy9YBj06yWAzecYPPbyIELGav+e+I\n",
       "QkDQDcoZ/51Dbk88T0s0/psQDf+UkOYsxFINi3SecdnNeFQ+M7M/m5Qc2vvc1oK5XKiAz42PtAny\n",
       "YPUe0r4C5nqCEEtOl+h4s1viQG3f5gxjjf/qA62vIO4SwI8Db4Nz5tzokDBMe5bVCFAwvZquWRDj\n",
       "ot6Pm4lQOco8S0gBiCoKBR4hXOESVKML7Nj0VU1MbcRU5YcJf2OzWMftKXg4cSLTjNp4uUjbmCi0\n",
       "A9sYOCIol6ves+l9IUvtO+THW9UtPMSVvg+GmQWEyQBLFuSopfGq8hNmZsxUJJ+caEOqi+nmE52P\n",
       "AbbBwne2caZ9uC1o4qpoVqI7Tpgxlc/XCYnHRTWpDxt4G7QgNIgFJc5oSrB+sLcxGVwCkMiaIVKH\n",
       "unjG7Vy39/ehhyWyMU+KsHnnAVnH6gMX+ehNxrZQ8H71LY7DQauX3ssEsy2BDsCVm9NWHiJs7IlP\n",
       "4B+D9ITtZ0xiG1NOv7W+aPTPSQHVlVp4ayTrAHoXAfboEWcpgpgZErdsrvACmHGNcyGGkmT5YKfw\n",
       "Djw4DWK/H8h2hUbkDAHj4uw92ZSEFaZuSEVTdws9jQa798TNjNuHOMBjEo7ZKN21eMAR2QlY//Ij\n",
       "cIAzUyQdqg4Er1P2Pc7iYWLLHvLH/ZCAfS4snt9ys29VuBptIu5nTh7u8EHzlSrOsGciEF9nJfDx\n",
       "g4vwnDLlJjEwWJ4wbf0oVgiW+b7LGaggxy5pLwAeDT9IEeepQP1tui5wPp9qOSXpHT//3jLjaoKb\n",
       "J/E72AKRKTEhnN8pwRoJsWEa9ndgMZHVTGqx4dYSCJgAs0D7itHNAulNsm7suKJQgqn73TZ/6y0M\n",
       "ZUjf8GIyIuElEOZImOKFqoFyzoa4qezcu8Wr8xHmeL+XA6hZX+E1lhxFPEpNSQcsi4+hjkvVBOPR\n",
       "R/dgrNeK10GH62jJlan2+qehV/X2u2JQvoDUe1j/XeDHWRTPxfW2X2aWAyUWWSmRRL3CXCzg2nBL\n",
       "Di5Z349N6zeXI6KUvAyNGsqlMLTlsjyFuuww94eEeHld69YLDEdrHxkTVSXgUzGoqrlAaoUYJzQs\n",
       "k7cs6YowwosX562Q1TsrexcoAxbh/1zSEl3JnEmLbcQ35dWa2dmV+FeZBmfTGXtPzaAGiS2kOJp0\n",
       "+/iGrL/iFKH/g0oMu4ImVAATWzZeDaaxAAAG/wGep2pCvwEm13KjS7bppLidYACvU6KaAIkLG9LR\n",
       "LakgBJvRHXbnxJ+0Yv27qhQN8HYtRJwbc3RoKEmU9x4pFtSXuPie9YvugpNcm/nWaci5iNpDz755\n",
       "urBn3J3C7G+G+TV+flIjnrlqZU1ss/HYAuYeQgDMCnSTzRjxE0Nss5ic6d9+OnkGtcmEfy16f36J\n",
       "Xp8jGEa07t8nzDvvwgREAX9gcKpyjEPTkOBvExIaqBC15jntNPx7a3Gc+3l1o/fEIMd2OOFbDxMH\n",
       "DFw28P5gFhGbZWwifWjOaLzkvoGvK/J+16OD2NkxOqA4U7q9IPBiDuyOFJyqZUesl60iIXcFfnkp\n",
       "Q3QBdR+yBBBWXl1DPfpJw7jAMxtg9PPJbWNTiRO/5n/eO0mtgxXtKYcv23sn31sMuVJYtZZTJf+h\n",
       "jx2dumdom/mf/1r2T20KfE1qUX9s0Ph6cKKg3U7BcMSXJNzxlRd3+VwfN8jXv1IaaCl03E4Pz9KZ\n",
       "I38PrM52qWTnV0XqpbwaTWtwb7O8qY0SlNwxjKn9eLVvIQTJDdVcFpvJ4Z5ny9D5MhooAUeqDFcA\n",
       "rK4Lj3pL6T92iMpGUmdNUUJ2TsNOtvVAwOQ5rPMI2D5x4v6tJ3WmPYhUqgmXsRIAVrDk0Wfrjuf/\n",
       "5dI6DOIgsoqs/oafA2mhrd52xaXVej0pMZv3A0m3XUFubtm76gXxcRzV29HSdVraWhqz2qi5p7Mv\n",
       "PDy+6UUCuc/2JHRL1X4dvp2+Huaut2y0SaEI8/biJNyYLvpDGfTOIESaXhPcvGMMs2m0OOnJ420e\n",
       "SstwtFSwlnRI17DdQuMjrTy+5TlGhQQf6t2JXpzxBEa/UkImDspMEtf/EErhUIBkUCq3vEjMgW44\n",
       "gaMyJ8Pj6tbeyBEiYGR4aufJsvJVxIS5lYyXUNBHKBgxZjrswOr2y18Z4A7Xzamf1abGsUq57Ja5\n",
       "iUsbbtzm3A5zKyoL0IEmig/eQ58OFGSPqCV9TI9tPbV2fL4aLa5jPwXTsWvU2FssY9U4+Xc4XgQJ\n",
       "Emmeo+rzEdttoGxtBfF1OPajcSdPPHos6zByoypml+kEFG0FOg8RH3xPxVLTHJWS5FZ6iR5zDjkE\n",
       "ZW2CIqTD4YUNu4urdXYpVg/HVy7pRlv4DY1+qGMuXOCk1s0LzVz9qCVgAC5AJ0M4OxqqFMwFreVp\n",
       "bcirRYD89Z9J1JCR2YFxwHBF2OW42VX/k60UL+vTgNkWq7+2KcYRX/Yhr6eNrtM0K1EY/8AEovhC\n",
       "SiLrT+DdVu7OFD1KCK8TJhDcy6aTMIQMjhPGHs+T5Qnu6vVn0eUGoYAaXl/gwq5t3uHlFr/cHZcT\n",
       "CZJtOWAl/1MpEu4vneazr7qpQ7nNCByItAf9kp8jiQ5D9pS/QqWlxn4N6jF7VElVhia8DBoCwoPD\n",
       "htPhd0YW6uFewee14yrKtJuV5e0aKsLSlgyhf84QNMaPQIL745WNC2Rdq+XaoF0ufSbLwlDo1IJH\n",
       "jjVbE3K+f5oZnlzlnVzykNlxPKbb8rZFPibQ48kTjSYPRs0UfzmY3leSeEINyALqExfkqF6QZELo\n",
       "ODYKH17s/QZVpaRuMxhBj+I7MCQhAw5/lizQC8vRvwv8LULyNMJe8MyaLc8C2g1bmFZ+C7SHBjNn\n",
       "hXYFg9c3JGhHu3pPAIDUX6N/wCBBlEubzU/XEI9gvZNYT4fb5dyaVgMI2hC+CuuRfNKI/BNvkIg8\n",
       "pG8P+XHqH50fXmqPXT6awgNCltEqUvinpTMmJ8py0rayTbrsd9ADYRVEafEW3pHorOtyN7fBg9Tx\n",
       "xImVJC8Yo9LpXh7ZMPywGfzENcOtn3IPb99zl7BCxTqZ2sLUM2D0Usr5/W1h2JdJxRYZMSkmjF0b\n",
       "BNVtr3B/M+4XJHBOeAEFw7Qfu2BrkXertzIbQP6kVRFnqapV7razJ0jORv+4scvVjJEYLJruDtYm\n",
       "3YWx1WvmbtBW7l+HXHHbMKhm3gIZITkObmrwFrQbj46EdYjNxFoWHQ3AZwvyK4QSsJpvVIxA7wEk\n",
       "5ZsGCCEuWpkTlr22H35E24+dD950w8UQsjZa8MUynwP2beva35Xy+p4DFhW9QoxD3I21Kajo+Z6/\n",
       "ZE6boH7WhgfHrXhgwzwg3Ct6rc/ryez6JHCtAe6ch1ek5gKGM2Krle8pG/nTL+/nUkC0dINafLvv\n",
       "TDd0XI4kkGKJ3WBE8h7qmqL29xbQ8BnoyZX4vi8MDOF+osVsaaM060HJ8pryiLkqqemjbsjSYBEF\n",
       "QSkl4CRab2ieUSSp4gVzF+5AinVuugOUFUDj5F3ccLYORmRhaCtPW7TL/NOL+dPz9CBNEOJDDv5g\n",
       "Fk6AVh9sVNh3/gpzTl0k9sQUucX6egC51/cDXhagQhx6Y844yZchdlxQoDias+oU6ugolAAAB15B\n",
       "mqxJqEFsmUwIb//+p4QBcjLhmmqrPbuMAJan6rINv8T/ltErM0VGcLjKbvfWzIfvaKtnFViePNUQ\n",
       "uBl5BBeRQ6AJqi0hpVND9lp75S2dvfoH5p7jD38R0gfMfVvMPdvF8OvPzc/NAjrRdsFGV0AEjPHX\n",
       "WBlHNrVzDj+usQORej7wVMp1cULUiuOlK3I3h2fscuh3k+dyOeD6T8MKMSnsRRUJXxit9UzvItG/\n",
       "Wa9ZrFYVUgdbtUuXxZTm0KVb3T3wJhMMIHqzRANrSCN8G0xluFS2wa8tPqrwhocIjlSBYeMWREUK\n",
       "yitHBzvYD+46iM50Gxonz8VDD21BDY+s1dPKuVhnUpzmgSfBdl6zAuehocKUbIO4y5rmon4bU8CS\n",
       "n9fbRk+Cym3Y+837qMO/3WaFlvpfrdpb352wuYdxAR2xSy5mZiIZdHJFHBbUtXzdFrJgfvzs01Rb\n",
       "F1KwQS/FCNLP4g+21WMlkxyebVal1YX+F4wYJg267/99bGMr8wukL+4av8DfTKGLvLRySzwoogAy\n",
       "h1AHh2V3k0irb1hr5rc+Oo0AbHRLvemx/4NMUIKT69fJ2D4wpX1ModvUJnxF35zrQ6R97QWcQE6q\n",
       "122G+r76JXZBmkIgwM1C3c64WXSf5fIYeP6nWm/Cm7yAbLP7bnxPS3bsqmrjQte4CEuRh9HssfRJ\n",
       "nAKVNJ2Cd7eQnRJ6rONK61QqVlXHaF+okSuOXBDk3T9I/3OBrpkUieGXy16Y9UEQiUNCsC2qi5Lr\n",
       "uzH9SS0gDFcy0Rg31EbcjxDivX8ExKoozV1ArQ460vJxQe/83pgsEt6OScZ0SE9/k+Ha9GG9IkeH\n",
       "5PYOox+HfnAgMIWoGFgAGg5YaoeBD2qecAfV/gpZhUcgcJnh5z4aamcCRcg3W5tCSfVaWYdQwg0r\n",
       "sYdLqxyb0HBoxicGLBk5pwYX3VZ8KMkpzwQw1vo06mgYaeWLKh/3OjbOp25kTWjnKTop2pqVqAVK\n",
       "8fsFydGSyuMya+atn8/olYcmo6JCHmZF6jyt/7IeD0eV0PlVZiMB4blODKxXFlwSynWAu/cgmJlo\n",
       "5LJMhji7MbxWUP34Kat0pLkLt+5uorrABZUfXjrza1Sr4qxWj6CxMWFSONCvNSqJkg9yv6pg0lmN\n",
       "ClJcTCHVBMu4Se5IY1SdcvgJ41ID8cKZG7mZfH5yzP3Ff5n44yp3F+h5I42P57JVAu3Y3qw1IDOw\n",
       "XR7vchl5+/pb8EI6kd+GdMYap9KDxE+NkSMhT3QqKZm/Iz27PMvCYDtQCrAIdCZHmSqzIrn1ct+B\n",
       "f2dZaL2KD/INKpUa154C9B2zVVRv04r8aKGgEaCekk7vvga3pnpw7alC0o9HVuznbgWfBTnHGklE\n",
       "RvCM3iYlsD95KIQx+ApKrk3K0vsK0KrULKY3s6uuc/qGl4Jfsvjo3E6rO+VrT6sdfgQ8g7j296jJ\n",
       "WMrjEBZCB7dPwhtgcirhzPc3yJQITQMz8p+wZElzdkP2gIP1l35v2ksDCJYMANc1/MqOK9ZdXg0d\n",
       "7uuRChaovKyS4NDD5Fgws7Tv8ohBf9mB3m21eQbMrfcrT40qW2eKXbhl0bX4MdDWB+7U4CDQYMO2\n",
       "/f+JunLSX1TP32E0RVy0tUbVUsW0jUH/dYp4K/fUgYJ3BW2IQRl6qjGX1B3lSm7MHhE3jtG4fiHb\n",
       "PNe0qivNPdj2cw54BAOunZKrcgPlCSTwjLjHb3rpInvECpvOKe/gbGbOj2R2jua0ug+fin833Y7S\n",
       "QzoxB+T3RDQXMFy2edRtH+2i/wFIE8I1OSAeisyU8Apj/r+rmIBaY7cT3m8EO4UCQ5Jn/FL2WO6Z\n",
       "5DIkO9rq8H8sJM8Da2yYAG8xXI0xFOk4EDJzHtBb5J6o4RWr/JRDUVMA4vBsyhLmsxkNqaxY09nA\n",
       "aTzeA7OBW5ib0LB8yECizLFSdxxrPAtefWUNd3qs0jrscX3ZIknziXZQrsZG2VdULnpYiyuod1XI\n",
       "Zd3wKEdk2Vi30ijDW2VBJMEl1MgiBRNIbaJzcuwXDsCIZYJGf1QwCeS9WSe9ZKdqLKL6qYttPDLx\n",
       "uVIiToIUdwFiWub3UB2abTDjCCqDcqvVbQKDKuDZto0ZqwF1yZ6+LINJgPPJ+F4pSlJE8pyezD3G\n",
       "HwVSa+Q8p+b7xe9a3AFzFrIPax7q/CQjzTbcKHv18XlI2OPZw70RXTesS9H+aT8hD2QgIzrV0Ob9\n",
       "rIslwnADSDx1aYSQpl/aR1Sn/WL+S6fQiDB5Kt72Y6Zw6hYZ2DsH+RS4UoMrQ8jJCsxMCb2tPKy3\n",
       "SBK2VPR19JyIBi0f14sTEPAAlZnFhoCvXoOGvKCYoNpe31nVlcKGumMHzg28rG/p659WExjWQUN4\n",
       "GdDwfJB07PaZQIGdIEv158hL1Mb4jErRJMy3BYB2eGkXuFlDqupPVdV5XwYgETuzEk2cPN1TlAaN\n",
       "mzL0b2IFoJT2d0oi3NGb/9IHlHjlucajGCF9KYMef4bS/DUBPUFZmsUmvWwmPB7ee+uiPEpf+394\n",
       "eM89QAAAB7BBnspFFSwv/wDXhjhbmETYdd+NbAJrvgDEkEoAqdqsw7aJrUalwVf9RUHzPAa/L2z1\n",
       "xpF5jT706qshnwj5jDZx/4UqTvwK+HtbeB+zmm2qXHwTC4iZ+QvLZZ3NCrUQKBiGdNZrL+HeIiG2\n",
       "w2d5gytmrHl+K+fvWQKtyl6rdCHiWBMlOXlUgrZQYEZABFGVCDIJgwObKhXAovMLwP/TTI6eGroj\n",
       "g64q32KAEXlHZkMNYYM6RGv3R9JQmBCxHuek0SWfwe5l4F2+q3xgGctF5zl4YzaUP4P0zv/ppdxN\n",
       "3CCtu8Em+wmbhKKEFtqWadxOPLyLIRmQEGjV9Aq7n3Jn5L4Qqtly5/oTVdHyIt6x/KYvqMEHqyYX\n",
       "pwVq2SV0WG+n/YMifFrIq2N9omyscDN1gSvoc/f+pEnV41C4pSZFj+RY6TPl0q61vIQ3tUExWjS3\n",
       "H4PcKjF0v04G1XfWRGCPr9MgL8AFsDAL/fb0L2keIa4vz2T/0DzKCpkNhPo/TonlZiBYFs8ka2rt\n",
       "EY0i0AKu8JM+YEn6NTOCFsP3X+VycU6OJQxUgyw9osqswU6+4Bc03FOy1MV0b0OktJ2Tzz6HJAm7\n",
       "ju/i/TRviTJ8JhveWZr0FBumzR7EOZQCxbl2Gx+/fWaK90112aaNF5G593FYTvllRV7TUv0Q5Cq+\n",
       "A5vZ0mqabm6gpP6uTu/hXo7/LZuu5LZzG0x1fi/f3LycBDrCg/ekIBEouAjaumrnfmRBMjJ7ojPr\n",
       "mAkXopWoJH+xWXACUl6SFvN001mX29zncQeDaJlQF8mTAG+H4zGZ+kYrIwzsl4eQAZt7cm7yDTv4\n",
       "vb6mI0FURrhkOacVnD3HsyClXn4wkXWr7Z5Ets0Y8JbIDI4/jMUHUY68ZZJkw1w509Au9mvhU6bz\n",
       "cDqG895rbGTk7CAzHHktrWCmxNLuaX+dK08y4rBSLVOV56aZCEEBP0Y5wjQUZWB23s9GA4Jsl0zZ\n",
       "pYoG3ANMv17HP9b///+K5Y1h/BjwabElKnSDnQVe6XZ3gBVrO8btGJ84hd0NAf3N2va1GD25XCCe\n",
       "KB3nTh1DeM4t7YfGc2ysMqlHo5khnPmQFRnRCTPm4PBjVJILzvGpfimJiDzxc/iJljOlZlsURa+z\n",
       "MaAKHFThGH7zKMYX6KAGeA4ZjUXVsmIU7umqZpo+n9wlSstI2Es7/lVKCowAQUF21D80sw/lXMLO\n",
       "l5dr3gGAHOEdco84bIW4UH4tvnppasZpRHRqufUNoYuf7RBDbtHn02gY30zX6sIRLJGNzhtHwzUc\n",
       "yGvtLhDjqyAAmmblWokOb9BZRdq4ccv13aF7g4/sjDQurxQ1YizdyB5ByHRsgOm2uzYHrwKYcDx5\n",
       "zVGUAMgzl/5H//RApSrymEIMCH/gm0ih43fAjFF7un9yUVHLqiUyK3rAQwmzc5nm/fNhiVwGpldW\n",
       "a5O9NYLsNwy0b+OCNC3WeucRqHl7hZ/kEQQZKYvsC0NKMrnXGsYjs0x4cIH1UnRjvYAEcEkiE1Tq\n",
       "T7FIbIle4RhCmBbC2aZAbnzluEDpNn20XxDiG1gBOtTJgCwmDd/c7IhLdTTYblT+QkNaqem9rCh0\n",
       "6BFBIf/ZsDrh7mx3KlwRumZc0BzwE2wnMbTge/o5Uy0neUWefhJ+BDhByFnJoZ2xvJTAUW4iFfde\n",
       "wkOEFQVedMTBJaX1fbktGtnSjYn/eTPJ0OwLVeERmBF0SRN/Sb43zNnHd89K6SurDDYgdKi8TyOm\n",
       "AWWIl2ntlhd8VtNq8OJ2erAxfNH/gguZj9g/geBGPlhNA/6kZyDaT46evcLacq0NwS5y1RuFTJ7K\n",
       "DyvWI2ZTXVYiQVtjq5xnNO6F8mo1yNcpNnM4BAuJE3uyYhkHIZIDw6TYAKeS7JJ7isdE+tXFVZeP\n",
       "8imjp2HfBkR83bF47gBZIK0mbpCDGWywhwGFnzKOU0rz0FWeb1Ga0jjZl+Uo90d4TUEgy4AG8QCf\n",
       "kmJ1CsWCjEJvwDOzm67UqXbMmzV3rJXobOv6KcOscBBbhcs9MEm85H6vu20vIH/byP7l+LOOJTPK\n",
       "90h17GaKPE9mLyWVx3si+aREXNLIgwb3yfzOX20itaAILWeFtb+RUWgoZLOu8WHWHpT5HyNkoFQg\n",
       "OLhEMfDSplYRSHxboOUQMr1sCt5rCWIT/nUk33geTshom+yA60a3hkeW3SXoQ+D07jnux50vbUFL\n",
       "4tYSX4uURW/6Y9DfI+t6KWJwQ7dwOMerBnB8mVMgYe8Pxon1uSIZv8Q69u605xtKzjh95VAAg1PL\n",
       "myqxheSA7f3hCR4dBskFpvu9Yfc9kIxs7YhaXGwWLGxR10t84y4/hikRr6gZN7mPnhzaas+FUKVX\n",
       "MmO55jm4juuFL4aOpao50j6yLzTp12hjYGvpJpmYerLmmr6nToN31m/2DF60qr75sGZT4a/Awhic\n",
       "HZEu5TDHtakTDWOXSfXnxgS61SpUmGnMy9VinYm0X9TbSh4rDfcPpiNSpd8lneB7Em0KvhODNbNr\n",
       "yz3F4WIQjn6xuLWGlR3hOPp5k504hoBQ6EDVHi6yJN2ErMIfUxV9EfulWfjZog/9nppX55xHfYIM\n",
       "Zc6+QLkGsotZ9eLCugG1bJpOVMgtJMqbFhXwlG2HMdYCtJwLJVMAAAYpAZ7pdEK/ASZzYoKrDoKj\n",
       "MEUAWRG4ryvvcgMl0i11521f+CHhADhfZPJ3Ll/GgJbhWckBw6WUC3Aqs0QMTImiPlvaGjoa4Fpq\n",
       "xyb6XiSPaI8ycwUL737OkFMQWRrPlGPaNf0gBvl80vTtwYB4cq2hzC/eRGWMzbTy9m3EWI8DB1Pt\n",
       "eRFHw8QMHBfDuvARYO1SzefN1uc/dxPO3LNZOx8Z1aoP9TWseXLlbl/WNHwqLGzzrBHE/8d6rS6J\n",
       "Tj0cbF51hOUaJDNp5d3kb+8qeQFKztDNI91nTbB5Ug3vulkXjW6XJNhrbLrs0weKa+XJiwKnt8B7\n",
       "rRZ2VBlopb9pc6ieWpe097CaFjKhDXb6Om3mmATSdccmr9zaeiZNiZZZBqHwpvQ5gyJML4cYDvxL\n",
       "ScNqqc5J0pJvteyg1ZPwEM705SszD9a+lfY/KPM6dRIKnBvI8EvFs9mcwqxfPGzaBtU5fEHKYOrU\n",
       "PpYCRr2dmNUHcJthHZ952IPamNYprVJdJZDaTGnqP6Z9audCGQJV/nLCBUIz4qkU2UXK/9LQsQpH\n",
       "LvQJOfLtXihcTpuzN5w1RWB+c90FWlgYuxztpqvCz5VYbw25ZL1S/iaCflyvRVVF3jqw9xhVYu2G\n",
       "dKkPa+Nr33cCWopyXMSLTy9BJIgwWAzEFJ4Uu1UmvBnCsvgO6qeiGv7SbEnEA7XuDGQRMKUrhP5X\n",
       "qHT/kZXMMnI+W2dd6HAkZo5Y5wc0WZbUPhv0g0MYs062cD2GXd3g2MeENYwJdAnL02sz/8E2oqqT\n",
       "HpFVIxhOZhIfxXZ7azVNzb+IokBqI24G2YcVqSn7AeypfzPKKtO5kmAUxpggxN7oCahNn6dsmvFR\n",
       "BhDgqsHkIcUdeDJcJMa5G+8OIscTgq40v6yxE36XGrSlHG0GYTjIJNavCAlzB1sS/YDrcnKHMJ3A\n",
       "CbyW3rfaCe6jctMeux9F/D/2ailIXH7oUkPcF0OWCr15YoA1vkCgyv/g8zY0ZMtG64gW+h5pmB3P\n",
       "PkYJJWQTQvKds+vMfa2mw70xAyhhWc5N80sPheE7DRZqpzZj99nTBcEWbQSd1J7o0OHzCGYbvOyc\n",
       "oI6mcliS0TdZP/yXrEUO5HfcuyoqSXNrlAU/O5or4AYE8Jpe69tQ82IeRSDpiCT3S2h+fkpaKjCj\n",
       "TW9aO/Z+OdeajVix1nHlAnWc5HhV3HZsYzKHHyABL4m56Hv37JbOygLkfYB8KwAHj8fBj33P/mN9\n",
       "kHI6uwzBLpmYBjIj6afDWQFBD+pHmVzaBPKWnPBl5xxczHA0pZl1wgbvqGroB2uqWRY1vw+yIRGV\n",
       "VxHc/WchvdxP/gebyS5SDzcScZk3Jp3kAXcHgBQJMcHoelWL7K88F6oU5lHgsyBuSuxMDr2REyTO\n",
       "2dykKZOiqHnexbeHNYnp6gcIaj+lM6gKSQUzTGS5oefyQbwMt+EzizJuHqO05vjoaB/thQePL0gp\n",
       "SHHnvxipUTGgtROyC0BVLmgJlcGsJEVPT7fMr6IMT8xymNCAnJK4fI+duyTcLOfKfF6imAJ2lqig\n",
       "77pDwCtFaOlrib055FKZ7hsfiD8YjBIu9HeQDEig/M391zXZ3/PIDhmFeVIfklA1DlQaDUXDgaGv\n",
       "TVIlCV3Sncxqi+FjkECkGyKjLyGVftON12uPGUAFWVu98y2bacC0JMm5QI/qjcG5urxxakK5I8yG\n",
       "hpj+xXoGvjx7Ph4DU8a+/yvVYvPnpke9y0iBAxsQE10S5zUvDgQA4Mm10kxsyqXVjvtfrc/fuN6H\n",
       "KoxBheR6h7lLA79lHGThJvxKEAXwqJsQS5TYJLSkSgPyc/D6M6BQNFOw9mObXcz0aZ4aMd6OAjAD\n",
       "D97m/VRDRkm3mOWzMTahGU2IZ/MY3A/0o00Ag8RMcgLXPO7G7hU6FHWcfauYVhjBqFaH7Y/ybcMD\n",
       "z4Whhw6qwVo4e1wB5K4ADVWZKElTvz77dtdhmUXwlN7TJYYkX3P4lTKn7/bflL0Ut0Cy1Sgd84X2\n",
       "jyNcPFOxWHmQNA/6suXBBxbgKgMJtTpomac66MKlRBYIJaGcO5gVblkm8gBQgO3u9vdDSsG1UcWF\n",
       "+7sMQXD0YpGJSoZW/iIlDEFxnKzsbNIAAAawAZ7rakK/AR7YFBt8Tlw0z5NkAk0VOPKlfNeyWD9l\n",
       "ztEgFZRfKEJuv6Ww0bKkR65MzYBrHycwc0fL6qTwMMvY9AWQiiaOk0Tfv/hoyiM5H+L3kpaD0C5O\n",
       "Va3dI+SzjJi4T7ll4p2ZvuidOA6YJncwnl/0nZ14PSv3ozMnf8eGAQm7XCkCzn3Pmj6VYhHZHqiF\n",
       "BJinlh8p7saI9D9AyJHuBg58A5sHzj4/2ACKjPwdZ+JzXaQToKuczerHkooh6KI9fqfta8IMVP6T\n",
       "+MSjbK9aiRfkkZ1YPKysV9cY8PFk8cNK60u/HtFVdNrfhHPgRkDV259q2WyOAT7nrQBkZwVwMwsB\n",
       "K9vcswwvpTIy+4TUrhfPURyJbaUjfVpLWDAjWhFuzkpi4t0ytLZlPpZBkwworuNPLi6CIqjqqa8d\n",
       "JmYtgSIzT/sG9sy7mkp+ZZ9dNU3GNqidBpI2nckuncZijusbMO+QsYYvdw41OKyhp9W6sCugQLb9\n",
       "FN56jwF1Iihf9YzrjVVBjA1pl64XVJ7DjnfsGKa2xhfAyNdpP9rtnqSiI4Cxtt8O4AQ1nNEpb3rH\n",
       "BXJaVnV2i7pFeouVzZf0sAKdBn7rvAznkJ80fENkxzWfmrE1CnceFkGvBuXtRrVYtU43dqFxFr0/\n",
       "Yn6Y6FmqqV1I2a0q9lYGQ3D1fE4cZhobzQ5a9mrKgzRmhPazTQ25j2+8rqPxR0Aq2A4Vitp1tr1u\n",
       "xbPwLktcYpzU+YluHdzOKAcu2oHO9hbNKxLvrWSpfB6tJiw7+UxJlG0XR3ptLzjYtMT//89XO2rv\n",
       "nlYxcKt0PANOeFEJGYHBb4R+/I7a4BsTGNqPugqAWikMxfCkz4ok1PnC/qM3VtNLsKKN1xTQ/ToT\n",
       "YlBAyB+jKXMVAsjrBToVx1oFJkFRVlP2ZygC2nlvI39688NfTi3IuCyNnurnYfIF+n7U+WL1j+Nv\n",
       "vjUAvqhcuAO1fBKWPJjT7Jla8IkmaSy/KeT6f8rEWNdiXNqPIz+08LaUbqZRzTruF1WOZStz6g0v\n",
       "yE6LfNiMcwYcyNP4q4oickN5JFMivpfxrgSwc1JQnn3TID2MJt7ZnG9nFUJ4nP+Oho7SGuGyyFdI\n",
       "JP6ZP+N2gRjarNmWTlTpYnWDTiZ0MfwXwSgNhbBcS3F/BsoAPJ3lGrxgamY/hAktI+XhJpe0QtGw\n",
       "Mh2ERVIHv3I5OmOn/WmR5t/aH3woPoI2y4ynqlqDOAaF9lm+5zNWDHYPxGH9dXSRkCN+K5Fvh2G+\n",
       "x/2lLEvjYVyEOXi2foycNG310dV4Bkk5lg062zHzmbSVXfG451tj65FX4w/lR/am0L3dp90p/pHv\n",
       "B8L8JZs0aw0jhs3qzT83qORrf44wgmo/4ImD9ESuhTeESvsCgegDJyB5Xhj4q/Ya1SiObueRNwqC\n",
       "NkhoDWOXnLZh94p2S3XXAuOiwAgKPwrC3sIn7tVQD8iiBWRhZewgqK5ciYR0turfQ2cZ10svP2Ys\n",
       "3RSICjnfzvK+Kesb7UNbNnLgZnhdQBT7LtbsNohxQkXkZdeupjbwynpW1UtodxQRafI5Tg38RMvO\n",
       "V7Ebc0fC/rIfMc16M9coceme0gffe/Wsfu1l//r5Y6LKTEE3s4yxL9+y3+l/y9aDUKdNrCatsfkY\n",
       "LPio1/54DVoA75Ct11TYjYh0xFLTsaIMSgOV6A79qFglNW1ffz92Ii9XZknSCmxl/zrg09mw8pTe\n",
       "ZZZcsO3oIDaax/Vs2vsi1WcerzOwwT/wRiHmPg3pMsX8tL1ximdlfS4LL/TozTRLg64u/lCA+NM6\n",
       "QrzO8i2Pz1e+8oweIK8zAcZNmnsqlx0M20b51vzcKGEtXtkDp1MUG6DxCVYbbVgp5+Nz1Qq8H6PT\n",
       "79E5NNtGGKjEqZJ4bKW/aTZHVlTGJh+dwe1IbIXDlByB//C6fUnYvYCMMp2WxT2XF2xtxVZri2MD\n",
       "Gea098ZRtrbAdQ8mwcUdxJgekDOXq736nn6f5qMAydplnj8e2InP/aADKhcgIhaKIhlRIoVMc0bT\n",
       "E9aGMjScMXcKrbFoJJXL0LL5hqu5/zEaQ5qbLFhg8q267mm2pBMbMls1n822Yj48R+BRq6QxXjwb\n",
       "4EZMmEfJZFRsR9yoz48eqr9Zh0qMMPAsVFVFucxMyYiwQZSlx2RyyFK6RRvcIgJ6pZfA5PKvIDl3\n",
       "ikSvBosIxYzvACSvpz6TSsc45L96ptvuLjZgS95o9x9niVi7vCe5CmDtICthsaactR9rGpevNZ2f\n",
       "X7g2wZjVCj8dQ91J3+OFPcI2PQ1hZAExmnUPq1AAAAYtQZrwSahBbJlMCGf//p4QBRaa+QHEUndt\n",
       "bR1c8Zrl7WubwJhhrruNSw8kDSozFbzY/Yb5MvFtqqL1cQXcCv1VAoz+EB7fnBbGvaTPXa8o7aOk\n",
       "9YgndHitz7+5qJMbtAq7GAPjPoYkUIfNNzzaMFynXzfBnD32o6/LBxfV4Fw3fvc/ZJgrW84ni5Pt\n",
       "VX6YeJzx+MnwZLNx2XDTrv9qAYZpkKXIs3/hcIo85lCsGI2X5daAZHh8Q0TSyRhjEinJ0kPTZYef\n",
       "gsQaGRqOGNkJ2LGlIWwBY/YValV0umJRteHAanNhdMCVrlxq2CCjHbgpErDmhGL/tjPXQNqg9icE\n",
       "6XGwCBymZdiGMb1Z4zp1fPzR7451S/wSD9HFQFDKZr3RxJLqqVi7kJzjWN5Znu7ZMVrDa6wBZF7b\n",
       "rmEb6HcWNd7YMUxpT6g+KFa0ZS3RjuCoFl2BvtVivWAiXrKt+krsFTAAIWAOm5MINH9NS0q5DU6D\n",
       "iIZcG2AXqVPjcu6A6tij/RU5W68blRGqlSLLFB+iTfhDjD+UdmxcUqQu3SUMhsAtjzHGq4pvMMOY\n",
       "rHKz/BiU45paMYZU0yzLSZbQ8XnxaEqoNoaW1EKsKLWAQ9n1vl1ULgq8R69kiu39fbx5yHE4vda9\n",
       "swNTddxZYNjulVVy5EgDkAXygSt245C3vEpMosOo90O8Lrx2Y7Xd6HgGUL8K302kmHVzXeXz77e/\n",
       "CBIyTSnWKCrFzztaTTY1rkWdyESEWz0Q+PD88TUcxwXdPSnsqk6wks7IBCydxMD0v8KP9A9UdImX\n",
       "KKbiMLqvgoaZaR+VJlSPoVIUKNexUjWdKMg4nEUvOZpBmIvc3XIGM+dumvcoOQEiNw1yDtj4SlfR\n",
       "LKOtjM2naxg71Gs8mE9fBKe679IHJzRKRXeZmN1s5FDv9tT81o+YVVN1lmZb6MlnrUXNWWujMTbL\n",
       "CtPYzGDkhKZL5l/y57MOslsCH4PHxXHjcEi8miPYOqoIjNGHzq1F6DJUC5x9i7uJ/u87pJPBIMIl\n",
       "ZsRi0Gf+UyO6ynDGM0p1GBmoY4WCJdl1viuOObWvByLMG1OSdVyueq99ha7Uav3UCvH/B3O93Q34\n",
       "bmhAqTaIn99d9cS5q6/LEXFVMdOXQ+Im1Ntw+EkaWCdAD1kJO5ccRJrwbX7XuyEwDYtQt7RkrlUz\n",
       "SatHIwKD35Byr8eb+86Xjt9bWqJs+qylUMpzCEjG/3GhH5Z6gAsCEHOPV7AIULy+zxnivx1O3+nj\n",
       "EsS4CXlcxTnahL/spVbkfnkXzjDe3uTa27uxSFr2IOOHPJWKjsvt7iFfKIK3wxPXJbYY5f+DhO3A\n",
       "1sGti/i7nRh/GX/q6xwowukBQcku5a5k/mj0BvbFdu0ZNvauPQPLYbjtGCiLsZyLOb/iH3h3LT8c\n",
       "BYWdI+bFD692oionVa4Wx2LIIhCgBDOYRqvkERSJ3Ifme9e3yhOTj2g4vmrYJSGQ6AGovUZkxNmE\n",
       "5tXxXuVvAkO+lHEubG+bwaC7HnBvLzTgZfaf/0YjdIof0ONDM+CWX8jB77ckjU/KAZbBasiG6EEP\n",
       "RSdBsAiZHGTvvZKADNimQOZqwChkITZKB08TS1btMbzzRu/Shs2irFqOSEoZrEdYfTyz8cNxd1FU\n",
       "R8gohtrEEtmbcdhoCEra1kwKcN4lZ3GScAACS0tlhFcDHBZhcNoyy17f1csH/a4gviAWoq9HQRzR\n",
       "ER7cJfUiY2jU8h3NDeSEUktpIW+cJyR3E5Xl1WXwh902IgCA4pxGrEooKc4w3I01js3AjyuQvKwv\n",
       "/wHzV/0SImBWMNkz5b7dShs9QNH3VTLuYAc3TJ9W4j001Vo8RHwnOWX7pMVbZ2ERP/L6mZF9Y/wX\n",
       "O9y6y1YYznIuGttFM3q77YI+BSe/sb5TUsnss57El0IMWghc9IKLdnmNAawpSnqsaxS0fnySAsvx\n",
       "aZmIewn3ru33SLdgTrM6JjkpFT1OJ+WUXl8POSztTVU1lHVmpeUOj9pewrSmd3JY5lLG4mNUnxcs\n",
       "B0mvvIn/Xms6ip8fO4S7YGIe6Fse0YXW2+6w8mO1cbQy8iGo2O2o4fN8MokYQLAPe742+MGgLYGH\n",
       "5caSdxMCOymR59SjxZAOFPeBAAAGJEGfDkUVLC//ANLEdVvI6yB2b0SGl8zdvv6x8B2YpyjLICbA\n",
       "CJVjedKBCXaCrg/jTEyOZnP0IDFRObRtYiGaZkJIwhgob1E+8soSWdIHUvEEAuPAWdrEEia1ECeH\n",
       "eIKSCQPAqMrDbFJzt40KqEMh6sUaivVls/IWiaBVLoGpqk1U2XbOzbxcWCrqQw0IfRcDyvMDhpC1\n",
       "4NUKPMtH2ZhTgU4z+JWGRZW57mu12jAhjMEYHZhHssypuAecY4ZmpwkBQgsZgI9DlhD+0NGzyJv7\n",
       "okjjdohkwm6y+irtd7G3ngVVCuftyFWiYV7rrg9u2W4yUd76+Yn29R9gNDkIvEMIFB239Unz4yEU\n",
       "fBb3sfX1a/+vEVsnIxtGIpf/oHeweynCER6RyAwI4my/vDmlo1o8f+ZpkrsrNeJT0rvtPPkyd/DL\n",
       "RiuvZ2T9zzp1XM63ajlNlNtJ0aadB1oAemRXMmarRtvfW+OjxHiIZc4HErf0o3qzBC/ZNwwQlMYI\n",
       "Fg36XccZCQyOofuIEjIYnNmbQXCdMd0KXtB5QnMVhCvKYVupu8S5U7xqEcVOmOfCuzMgksYWgZdc\n",
       "h7V0invpUMCLRXDNSJYmLqp3eMINrMyqJmy6mq/G8oo3+hEc+TwVuDRpw3GH/gj/Ni0O0UjRh/qu\n",
       "K7GCb3B1OBwz6H7OEEYfyVbT8UcH6Vm8Le4q9wa/2ov+2HDSIpUt80gysiuL7KvJxeUJq2EKVG68\n",
       "Ls745zMkPtr0hdMVyHw6C0RAACTmcSEtvNGQCaXKFUbXRfVYcakOtKpCi8vSJblMG4lugpClqT0s\n",
       "7Ai4THJgUHY+DPruOFG4Da9YJGXUvs6g9v2yb0m5LP5pcXIXBxTH11VGdzbCBGmatM63NVgkMixO\n",
       "/m0qJ4MEtOjt/6mohUpLvt9LWnFtnKq3qbQrCOOiwNKmIqoJRREogs7xYzNby6urIJrvpcs+GtbC\n",
       "AdeSrq8n2pGNFoOy6ov930vvC4uXoEEVbeVqTFfW1aGFhALILFFXQG8HC5JkdpCamlNr9xMpstMP\n",
       "1viu2BKrLtIIaZLZYXvrjWRhSCPJgLFZzD1C48Zj67s2ZaItPsKoLJM/NPAj5Ad2odErs+sn46BG\n",
       "NEq4kZXJ/xXhsmAw88JP7NmCAsPSsd6V7WVOHSTuyxp+V6lTwf/PWKH6lEslmh9qG4LSht14XIpz\n",
       "4aVXWRcLApYip7BXRnVeg6aG0C/mEYfU35SxWAMeX4D+RZ+DTxz+IrPioZYm1WQcKMG8J13SNDT8\n",
       "Z9XZ/hS2METinrYKYtMQV0nCI9H+GB2dOsa8H8c4P7juDpuWuXuXkU5Z9FmC2jVycEKwP/ysj0rf\n",
       "DKjFpamzQvRtn27fALuJIu8h+FZT0TQFKAlpSPNaNNw6SwkD11SIyQASNmL7PI1824NaygA2BcJJ\n",
       "gyhydSndW1aUbB/M0GgMy9WLFX2GUYnpCL6hx9C1kfwVPzqE56QDQn7F9715LI/SpW/LhRsx0Lrs\n",
       "+duFBIy8yFcpp/JgZlPJU48AhpfB+j2T7TIW1TxTSA4sXnH0tSEWT8plBb6fIhE+y54/2RC5d1T7\n",
       "JqiYRJ5ySTdPwBzErpG6twiw6Os4yhnBf9PMOFCVEekH70jSC7ct0OfXYxPKOGUmNAA0lWxHMRnU\n",
       "3OOyhiX5TU3Q+DuPjm1ohLvZ56wCPVd0/z2kqDIaAnZAJtoDFYjiyNEaMo3vGJZsxoT4rXLhn6U5\n",
       "Ran8cxRYnVr+WoWWnAeLpTXgV4lF4jV618ITHrtVQjGwYWWhABlWrPkRWmOzja8NtHK71ZMVJNE+\n",
       "m9zUJLY00aiwc5Ge8A9mjfDnBRDDNsvdppZyfSwmHvFYY5gaFIietUmn4KkwLAnnFIXej0DHL0oo\n",
       "LILVH3HKIkz3KMF8Y/98PqldipV4W8lU2zgIHljHCA7qqpE0wvrvpc+M2Z1c51jDsA3K22p6Mp0T\n",
       "7pKiwsT6VDDw8Go7Me5RORwNdh3+FcmGNctNCgOplezhWm3uJSqxQ+/NQ3k2eWxGgmE/pYTtC0q8\n",
       "HYB3+DKw62UPmH9VkkQFbIo2OYAgGXAmoyIKG8iP3QY8mvIXGMVqbovUinY4FCkMPGUPKgp6gQAA\n",
       "BL4Bny10Qr8BHnRP85fJzFU5qfvjygkAA1EIqWvtqTpxV8L+gyzm8pv3wkNlhCcqYlbM430jnphn\n",
       "FRNxR2XCEzGLcSL2YYKq/csppcG1UjFHA2826NT8HHjJyzr6zriit+iA7Ml/ZNC6jMr2WVssXUm6\n",
       "uBxs/TcyXx2LBdPNgIoDbH6jXDq0cA3Jd8yFCXo7wNCvOfrGcWS8XzigbfuTVn9slMh+UnaQwpuI\n",
       "AlPSag/ndf6i5Qeti4EwOmknh+AzHQ802HOwleJ0shrjFRVsM+Wv5moVj3kvJZ9hJJZ/0au5R8J8\n",
       "eDhwDt8LCh75x5ZBXmtO36hFeWUVnRsEMA23PjmDdn/llMLQPrIj/YQ1fUo+h0nswVLGxk7jiQf9\n",
       "N+4UQ2LOeFUJpPaYmQ/Cek0t7u5II1quEWccbcMZ00kgIwK2mn6e/tPV//wXfFVRguhn0j5RN1lv\n",
       "uEakuKf4vHNoc4zcP9gyi6/bhzEJoomgUenADgFp/6k04w+raa1w7rCIEMpomUUit6H6klPUaDpA\n",
       "Laeq3t1Kh0wslO6FTwZzrDgokMD/5oRXWl6GE0b/5xRxsfEokbozvVqNrEI64SPz+59hV06diBiM\n",
       "itmZham6NSgJm7tf++ptBVKjqPowPh7xDXs7XbYMy1cj40XmFTq24OqpgzDf/QVBoFfvOhk2M2E7\n",
       "QBSF+IqrvdLoCG9Z2dILZbVS9Bw6F3iMM58aHiEn8cLGMUTgNEvPE9PzxUcw1AMiAClMjMbVfJ4R\n",
       "HUniXwqxIFGrOifyCdhr9v+dbywcWtufW++O/FOXjXm8zjKf/Q9LEq15L4WQw8xbAUNmq9ZSKh6m\n",
       "RsY03E974I/FpPAjcqXRCmRRsGsWn2HP1FyXW/KDiGP5c2UcRUwMJVOnNUfOZTwogK0DwTnpueTj\n",
       "EqA4kYCpnKUxNfVsYmMMTMaHx6jAqjJCczxHUovLuHoPCOH7w0hTXEW5eNOL5/Pjom+XQemV94J+\n",
       "reiCFsO8bFZwq2kcSA0ymhtSFukRN1bG1pakdHbuSXhK7BvJlT9PhrYgMLKdKb75Laxw1LHlx3F2\n",
       "DVS+w3h8/9swclc0KVfGNDbkc0DvOj4VHrrItYnJuMC5VRZyV3sxKI+R06PZ7LlVk4+awEJcMAcX\n",
       "PZ+qe/m7j5TRtxwzHqPTVC7Fz6Q3qnbCTfWjDmUXcGK3emlCOPZjt8KwkQG5YDvByBG3bIg83HGR\n",
       "61gxeW/0/hq2I068AFyFIO36UW9JGeHJrR9tV8XUJDxXAmPO7qcFVRXFj6YT/saAe8VVFpCYuIXN\n",
       "bSWhWb99v/xwAQYkSR/eS1X2AlHdh2YB5UYGMcApO02mE0k9aRj94QJ14pcC4kUkfGRKnd8/BQvi\n",
       "Cybs4LFrIblLTd3Vj4beYGVNtwJOn402ytfIXaU9Kueshjpg4PGUeQ3EXco6U3trCTHtSNMGCT01\n",
       "3tuDKecHAdCG7dSI6slpPdMrJXzsxxgbV0lIk1T2748l5hE3hJN6zw9ccp+TS0J55CQq115p0D2c\n",
       "22fiHmPcEoSo3ThvazmM2giB1UFBwwoOdhrge7O4lIHnrA31lTxiC1AJVJZqE3ZFcmmtQWN6XNm3\n",
       "Fr0STwIhR5w5s0EZQSfETQfrswAABUgBny9qQr8BHtgUG3x4TWpWqPwZyB3xN7UkAUGI0/0dW6+H\n",
       "rPy2a/Dz5gCHEO5FIkOrXRSNln4DywyD5vYGCxg7lj5e+fyuk5+SHL8ALDYQfp9plJgSzTpT08c1\n",
       "BR9HM4nS89shO22X7B+uBTRTnnw+9NjeXClqlzPkaD0doo7Twnaac5mg+fmXZ1Jm0oxiGDpJUar5\n",
       "3kZ4awUgCDv3n8qnvVHaN/DjBKoabI+UxjOhrSTGLgusvPDO5uy/kw3GUnXaFz+t/qs5Jpw7Cqsc\n",
       "XYIwYejNC9COjgeW4LU0cPPxdFolAOrx8w7IZHTE2qYeDYX6yq1J/+hz+Xn9pAdW72JtgB+o6yqo\n",
       "b7DrVoTugNyKIo97eUd/Ezh+tpUhXGb37LdwaN6KEnPNyTmqsl7RfN01mZXhH2ko7RPbRTUQxqDj\n",
       "8+6ofxn/WW3R4OFaqOWcXMmg9UV5GnRaCuJ1uYtL3grlaZDLjjFQ7zyJ6zdUa2EcHKOGoMIZG+8P\n",
       "FzliD927sFOHtpqg6R0O0F9qDY5m3n0h9gnvObmaQKAs0Kmt7ELywPioJ9ozOflCotS5Euv+/hKR\n",
       "kwgwyxws8XbgPGWKNojD0fx5tW58bcJNneScJR0XjUl4hPWTWVighay+yaUx0nANHHbNbpDPt1rm\n",
       "r6CY3ZtlZUk1tCcWWdUUPC6K8B9MovCgSifTZ1lM+OfKODPf6SfyW7N4sNfRBEOFz7uan5KYnv+V\n",
       "S/IlLhP94tc/AntP1LdHkvYfJW7jQ4e+C9KUtDmv4yf0828TFolg3YAfLmSkWFEUBQL39ctLFr6I\n",
       "UZPw/8Za8NoOtuRWqYUarIxbXAxyzjQEYh2z+nYougNePdY4PRtm30hNGMGeJsOhft+ak3G7BaIk\n",
       "vmcxX+6RNYrheREgiiCkW5FAg99jyQzMkOOFNsRvx/jZiCuDaTBMOn/wczRSpizxhpSAm/IJANdL\n",
       "e6fnx7UmzMuw0ouFtbo8/4AApxVndA2+4KoAyvn9ZKruFVHvkrMhGPj6JVbht8ZgI+wendLoyPVd\n",
       "Ns6EXlVMC3bD06tVEcBdl40xaKP/p4/1bhIDSNTD3zA/8W+xoC3Jh3vQdhV7C7oZnm/Q9Udlsbac\n",
       "AIq6Qp+ao0fH9++SgdQ2KOeSoq4ylfTeQJPzUzv1YRQGBYkHWklvupE+TS39GnAjXu88/syQDNHE\n",
       "ubj6udnU3vFAfHW2//6rPAqe+P9uYxX9AzlthovxeLEA5eTDrVFHPU0vkGczy0DfPVw/0vZK60ox\n",
       "9LpnxEkGG1WYG+vqYIjUy0vgSjeTxowYpDxsUHbgHoTsfoLDK25sGm7nLk0e6+BM2x2UOPo3an2C\n",
       "Q5MBvuX35zYZO6LUcG4SMwF0cdd+CqhhZtt1blkoIUCtcfWY0MVORS1MVqfse93rjMWH+wMoB29j\n",
       "aKVXIZ0NugSLz2Gmr1mWO7HwgAPgmscdioXJjtYBGc7fqHWfFd2853MswcizPKAnDw1kpDItQZMk\n",
       "fLwHRt4VhbNkv3BQY4wdrzpPiZlH2tE7Yd17mBtb0bo3npWBa8mawZD+Av5nHa9527ixL2ig72YP\n",
       "bLWkMBeqZNL58Q7RjIqkfW3OQdlLtqmIuohUkKYtKpnjpPZC3Uq7eeUluwVNQXbyW/hi3mFLc6j5\n",
       "RKvsIxVwLtynhWkJwE2sIe2U/laaaVOQVVqmqQD8llLtG6/1kDcIie4vsiYV0ezKCn/QGzmpdX5y\n",
       "1NUHpMPGUEzZV48xDZF7+LVXMnPuUb2l+pHq/DvQhsvP0UKl8d1mbhhLSOMhwZopwgnj5G2TvHCC\n",
       "A7NlZMnSvAAAAD5BmzNJqEFsmUwIV//+OEALB22cJxswmiQPIItrO5UKX9X+rCK+9NVMkfBn/zDQ\n",
       "58ACMr7oCVeVgbFCi1FAuAAABS1Bn1FFFSwr/wEfDLbsQATrIa90j87Ey5R5sDvK0Resk0YF1Cm6\n",
       "7oXKKYatID1o3J1q8ifPXbg5/Y7V9IZ4F3S8ZdKqlBQi6W+fG96q3yVKgJoILFq0l13W7OcHE0Mp\n",
       "E+H2b2XjHTjIItciMwWdxxw1Fk0KYKC/WkpmT1Wori7L5zLqVCKIw3uGC8ryFLrGZedwwl9SFP7E\n",
       "UhLAVt0rsU8WfkoMp7BcAB5CuXQZtvFZ+nxlmNUmElQEuHySGOwDou9VLfpCmHJkIGbJU8QcyAPO\n",
       "4neVyKF0dZkWomVqcBtw0owfMHH94SCaIxm7MJQvAk+98HrezDff2vsMhMX9jtQO9ZOy8RCQhEuL\n",
       "hNnofduyB+5DD0DWpZStNGJ9RhQ/F7imjxnfFzrLjmBkPkNLi1OUpTGA1M4bA3M5b5a0tPVkmU+R\n",
       "SFcKQ1XQ/G1cMMC+oxS2jZQWrO+Sp/eEVEuvKYJDqI5kL7VqmDoHXAr/+Fv17bgnZPmehgYyLdst\n",
       "fLm24mjn4qTzKB6Nxggz4q0EWJajfTx6gyh0kCcRiCnMrGlLQ8I0a1CujUaxGa98r/sB9hTce67X\n",
       "nHmn6UWXBT3YmUduSqtM0+62owPisgqSaBB+l0qCupNmQeFrsMbMqR4abBv+M/wCVOFbww+C/knX\n",
       "xljl0ooIXm27s+1fdmI+webQetfzk+Jb/VLbaZkewTlupfg132hbgtuFVjVPGoDLvzg9NHg63POF\n",
       "QgQV+ck1CLJrW3NUQ5T9dxBHXj8CUD/FwHgVRB6SoKiRhYjAdszptV9yDFpSgc6aBW15jF9i4H5e\n",
       "2Oj/ipkIp/sAap2VGRGrQunaCISDerLJ1LutwGLS3itAqXOr7zeMcpAbY3Y45iAx3bwyUqmMsjt7\n",
       "DIU68POeHjcKhpk+6K+epMQzr/1EslUsOa+pnR5Ivq0y2YNm0icYbvu689aWp20mrt/2sh4mkTE9\n",
       "k5561C3R7jJPyC9dt+pDbVGT2a50GbBcWW15ixQyqkDw+TwOqz80/5bi9/nXpco84WDJRfYhl40+\n",
       "P8w9NZqbv1T/fuLVdiYOtlX/MOOKBo/bwBBbZ+0oYDEK5uYVZ7VP4DrBz+XLGnt1g/zd8RjY9Y15\n",
       "xTqB0lalBzHgKkAdIxadIp16g1C0nLv///Hsd2bKE/35uM/Ax+nAhzIyPjdY9Rjumzyj18ZonHkB\n",
       "wAOx01/bJFfHDkPgs9m91JuVd/dpfmbu8FNj8vr2kOAyBFX5hS2vLDoBP0aj2+NTwOcCOQV18617\n",
       "KUNFjLL2beJDuKGjiEkC/wDpZE3rZC838n2b0oql20WugA80HrW0j5/4XsPUtTQb/qbZ0Ozi8hGj\n",
       "Gs3+xJ2F/hGGvNw3uqJJGTyHXvBWAJvT57V2K706RB4JdETIPA9235I6XBlPPEE2uwM1ciE/0ivJ\n",
       "fVs9zB9q/MInWqy/TtfykFEoGHwEblz3+5AFdy9c8i9gv1wGVN8zOvSxic0RZqWFi4+rxbdlzE1K\n",
       "U9pFoeFjfvz4rCnP4sF0KvJAwE332uWe4aLcfYodUDt90NQ5wQyE0BgNccj+QChjBbv4212Vs/Fm\n",
       "MqvRVeIdBSwbyYKVVpcyUsUZqgT3gtHOEip7kCk1dvPakSJPFestBbh/NS5Gg3+nMY4KCFGZP9zC\n",
       "dfYmOF8YXhYKWnFXptSaZNO/fKyJTF8ki+UZDqxTdcpnvvM68QzqF/LhNZKy4SZgHmzqQhVTy6HU\n",
       "Sp0Bg1stUEw5QX1kVqO2+qDWmfObWIwaOHc4UfN7AumTiQAAA2UBn3JqQr8BHtdyphNABzhp4AWf\n",
       "HK33tT1JPWe3+t6jqXzP7uRdakf4W2ALC3d1QU4zzV4/ZdN6c3CewBI8v/aUT41ydNXnN14dKipD\n",
       "CQBOfu4t5bQMFX++L8pihJynPmKpdjZ274mqnq+NMTMXHFsfy8VdUqavswXoTRBMgR5dvJxsNWnU\n",
       "EtdvL+2Z5VVeMsCZR0XkLAlpXhJQyQgzwaDT/50H0Lw7kxB5gpIq0lawwdH5PSQcM0x3KJ771mQQ\n",
       "Ywi1gMnWowP+++wEobWzaB1kXNJzsN8r7X+Idtvg5U4XbNqre1QrILcUdsddM0Rs7Mefd5Nj4HfE\n",
       "mBvN1fHHprMKUPqyVXL8M6nASJgZoCInusc6eRTkSx7s5g5EjHbhZewM3Q/NSXZxQkanO2Iwzgzr\n",
       "8P3wVh9mWkblqcpRjGWXRgZUKBj5XczNmYM0HeBUjKRclp6g0aX2u17jttUOf097APFpH6UQ/dzg\n",
       "GE3yloeFSMcmuPx488pX0bUB2JroX7Jqi0gqoiTxOKGcEM1HQEvGiHnhOao8zv7Wwf7ONw91DUBY\n",
       "GqW7cYdQfqpbhmq2vOdZQcsBXdyJfiBvs30y/opR7pMFSE/jol0EwI6WQOYjhfXexjwNtHDwa168\n",
       "ZbLxokikXRUGSxl/fwf+hvDbqHlBNj4j7HanZ0Vxq0cudyGyN8MJY6p9pt8Sl3KLAV0eC+ryRDqW\n",
       "4c/uBBUIYXv3re3Ve9EK6CAp3cJwYaDG23ha+gYHrdyCD/rqmW9uE40ZyF28Zah6JwyQh6FWjH1Y\n",
       "I9V15TaUZQEIu4W2FOt45J+cqF2KpjLMzyiB+0JMT5LDwUfK693s2oIOSsiyWqlDBooq8q6K4Qhl\n",
       "zzn4Ci1Jt+TiYyvc67x7NGoO1mejsch/E7JnGyGv8y7OpH+ff/n7ELIocnjTkPtBwuSKsJPh8sbw\n",
       "JC9fUoir5IqWzufzxwTKY42kp4L87Mt0G2MCIuqbIyuAvo6NRwTIVUseiuWHiMzmq+Zaq5WYGfjW\n",
       "dfBsuauHYNvuLfJg0/D2Md08DcACBrqDLuvT8ojhOGCMJN6tdMcRLFuUWMTOInjXPz3UEVc61Wgx\n",
       "SPcJOoGhtmtlGA99UPuK6mp6UztBCENAWIJZQRSqTpemdwYjoGeWaBIUzrRhif5QflG+RwAABBZt\n",
       "b292AAAAbG12aGQAAAAAAAAAAAAAAAAAAAPoAAADIAABAAABAAAAAAAAAAAAAAAAAQAAAAAAAAAA\n",
       "AAAAAAAAAAEAAAAAAAAAAAAAAAAAAEAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAACAAAD\n",
       "QHRyYWsAAABcdGtoZAAAAAMAAAAAAAAAAAAAAAEAAAAAAAADIAAAAAAAAAAAAAAAAAAAAAAAAQAA\n",
       "AAAAAAAAAAAAAAAAAAEAAAAAAAAAAAAAAAAAAEAAAAABsAAAASAAAAAAACRlZHRzAAAAHGVsc3QA\n",
       "AAAAAAAAAQAAAyAAAAQAAAEAAAAAArhtZGlhAAAAIG1kaGQAAAAAAAAAAAAAAAAAADIAAAAoAFXE\n",
       "AAAAAAAtaGRscgAAAAAAAAAAdmlkZQAAAAAAAAAAAAAAAFZpZGVvSGFuZGxlcgAAAAJjbWluZgAA\n",
       "ABR2bWhkAAAAAQAAAAAAAAAAAAAAJGRpbmYAAAAcZHJlZgAAAAAAAAABAAAADHVybCAAAAABAAAC\n",
       "I3N0YmwAAACzc3RzZAAAAAAAAAABAAAAo2F2YzEAAAAAAAAAAQAAAAAAAAAAAAAAAAAAAAABsAEg\n",
       "AEgAAABIAAAAAAAAAAEAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAY//8AAAAxYXZj\n",
       "QwFkABX/4QAYZ2QAFazZQbCWhAAAAwAEAAADAMg8WLZYAQAGaOvjyyLAAAAAHHV1aWRraEDyXyRP\n",
       "xbo5pRvPAyPzAAAAAAAAABhzdHRzAAAAAAAAAAEAAAAUAAACAAAAABRzdHNzAAAAAAAAAAEAAAAB\n",
       "AAAAqGN0dHMAAAAAAAAAEwAAAAEAAAQAAAAAAQAACgAAAAABAAAEAAAAAAEAAAAAAAAAAQAAAgAA\n",
       "AAABAAAKAAAAAAEAAAQAAAAAAQAAAAAAAAABAAACAAAAAAEAAAoAAAAAAQAABAAAAAABAAAAAAAA\n",
       "AAEAAAIAAAAAAQAACgAAAAABAAAEAAAAAAEAAAAAAAAAAQAAAgAAAAABAAAIAAAAAAIAAAIAAAAA\n",
       "HHN0c2MAAAAAAAAAAQAAAAEAAAAUAAAAAQAAAGRzdHN6AAAAAAAAAAAAAAAUAAAMjQAABscAAAgB\n",
       "AAAHBQAAAw4AAAmsAAAHLwAABx0AAAcDAAAHYgAAB7QAAAYtAAAGtAAABjEAAAYoAAAEwgAABUwA\n",
       "AABCAAAFMQAAA2kAAAAUc3RjbwAAAAAAAAABAAAALAAAAGJ1ZHRhAAAAWm1ldGEAAAAAAAAAIWhk\n",
       "bHIAAAAAAAAAAG1kaXJhcHBsAAAAAAAAAAAAAAAALWlsc3QAAAAlqXRvbwAAAB1kYXRhAAAAAQAA\n",
       "AABMYXZmNTguMjkuMTAw\n",
       "\">\n",
       "  Your browser does not support the video tag.\n",
       "</video>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "animation = dvs_animator(data[:, 1])\n",
    "HTML(animation.to_html5_video())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "hR3Xg09JF0Vf"
   },
   "source": [
    "## 4.2 Recurrent SNN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "executionInfo": {
     "elapsed": 146,
     "status": "ok",
     "timestamp": 1677782250690,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "YRvbELuvD1Gw"
   },
   "outputs": [],
   "source": [
    "# Define Network\n",
    "class RSNN(nn.Module):\n",
    "    def __init__(self):\n",
    "        super().__init__()\n",
    "\n",
    "        # hyperparam stuff\n",
    "        num_inputs = 35 * 35 * 2\n",
    "        num_hidden = 512\n",
    "        num_outputs = 4\n",
    "        beta = 0.9\n",
    "\n",
    "        # Initialize layers\n",
    "        self.fc1 = nn.Linear(num_inputs, num_hidden)\n",
    "        self.lif1 = snn.RLeaky(beta=beta, linear_features=num_hidden)\n",
    "\n",
    "        self.fc2 = nn.Linear(num_hidden, num_outputs)\n",
    "        self.lif2 = snn.RLeaky(beta=beta, linear_features=num_outputs)\n",
    "\n",
    "    def forward(self, x):\n",
    "\n",
    "        # Initialize hidden states at t=0\n",
    "        spk1, mem1 = self.lif1.init_rleaky()\n",
    "        spk2, mem2 = self.lif2.init_rleaky()\n",
    "        \n",
    "        # Record the final layer\n",
    "        spk2_rec = []\n",
    "        mem2_rec = []\n",
    "\n",
    "        for step in range(x.size(0)):  # time x batch x num_inputs\n",
    "            cur1 = self.fc1(x[step].flatten(1))\n",
    "            spk1, mem1 = self.lif1(cur1, spk1, mem1)\n",
    "            cur2 = self.fc2(spk1)\n",
    "            spk2, mem2 = self.lif2(cur2, spk2, mem2)\n",
    "            spk2_rec.append(spk2)\n",
    "            mem2_rec.append(mem2)\n",
    "\n",
    "        return torch.stack(spk2_rec, dim=0), torch.stack(mem2_rec, dim=0)\n",
    "\n",
    "        \n",
    "# Load the network onto CUDA if available\n",
    "rnet = RSNN().to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 8888,
     "status": "ok",
     "timestamp": 1677782269499,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "2wrm2w7lbG3m",
    "outputId": "dbdbf8b7-34db-4903-a1c2-454fb14f2a06"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Iteration: 0 \t Train Loss: 1.4043248891830444\n",
      "Iteration: 10 \t Train Loss: 0.8664901256561279\n",
      "Iteration: 20 \t Train Loss: 0.7414827346801758\n",
      "Iteration: 30 \t Train Loss: 0.7298647165298462\n",
      "Iteration: 40 \t Train Loss: 0.5865459442138672\n",
      "Iteration: 50 \t Train Loss: 0.36410242319107056\n"
     ]
    }
   ],
   "source": [
    "loss = nn.CrossEntropyLoss()\n",
    "optimizer = torch.optim.Adam(rnet.parameters(), lr=1e-3, betas=(0.9, 0.999))\n",
    "\n",
    "num_epochs = 10\n",
    "loss_hist = []\n",
    "test_loss_hist = []\n",
    "counter = 0\n",
    "\n",
    "# Outer training loop\n",
    "for epoch in range(num_epochs):\n",
    "\n",
    "    # Minibatch training loop\n",
    "    for data, targets in iter(trainloader):\n",
    "        data = data.to(device)\n",
    "        targets = targets.to(device)\n",
    "\n",
    "        # forward pass\n",
    "        rnet.train()\n",
    "        spk_rec, _ = rnet(data)\n",
    "\n",
    "        # initialize the loss & sum over time\n",
    "        loss_val = torch.zeros((1), dtype=dtype, device=device)\n",
    "        loss_val += loss(spk_rec.sum(0), targets)\n",
    "\n",
    "        # Gradient calculation + weight update\n",
    "        optimizer.zero_grad()\n",
    "        loss_val.backward()\n",
    "        optimizer.step()\n",
    "\n",
    "        # Store loss history for future plotting\n",
    "        loss_hist.append(loss_val.item())\n",
    "\n",
    "        # Print train/test loss/accuracy\n",
    "        if counter % 10 == 0:\n",
    "            print(f\"Iteration: {counter} \\t Train Loss: {loss_val.item()}\")\n",
    "        counter += 1\n",
    "\n",
    "        if counter == 100:\n",
    "          break"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 459,
     "status": "ok",
     "timestamp": 1677782281360,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "9hVQ6dezduj9",
    "outputId": "bb0bcf12-e559-400a-fc3a-4bfc7592beda"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RNet Accuracy: 0.949999988079071\n"
     ]
    }
   ],
   "source": [
    "print(f\"RNet Accuracy: {measure_accuracy(rnet, testloader)}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "3Ue7gXe7Rayn"
   },
   "source": [
    "## 4.3 Spiking Convolutional LSTM Unit (SCLU)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "uqZZNjcPRadp"
   },
   "outputs": [],
   "source": [
    "# Define Network\n",
    "class SCLU(nn.Module):\n",
    "    def __init__(self):\n",
    "        super().__init__()\n",
    "\n",
    "        num_outputs = 4\n",
    "        beta = 0.9\n",
    "\n",
    "        # Initialize layers\n",
    "        self.sclu1 = snn.SConv2dLSTM(2, 8, 5, max_pool=2, threshold=0.1)\n",
    "        self.sclu2 = snn.SConv2dLSTM(8, 16, 5, max_pool=2, threshold=0.1)\n",
    "        self.fc = nn.Linear(8*8*16, 4)\n",
    "        self.lif = snn.Leaky(beta=beta)\n",
    "\n",
    "    def forward(self, x):\n",
    "\n",
    "        # Initialize hidden states at t=0\n",
    "        syn1, mem1 = self.sclu1.init_sconv2dlstm()\n",
    "        syn2, mem2 = self.sclu2.init_sconv2dlstm()\n",
    "        mem3 = self.lif.init_leaky()\n",
    "        \n",
    "        # Record the final layer\n",
    "        spk3_rec = []\n",
    "        mem3_rec = []\n",
    "\n",
    "        for step in range(x.size(0)):\n",
    "            spk1, syn1, mem1 = self.sclu1(x[step], syn1, mem1)\n",
    "            spk2, syn2, mem2 = self.sclu2(spk1, syn2, mem2)\n",
    "            cur2 = self.fc(spk2.flatten(1))\n",
    "            spk3, mem3 = self.lif(cur2, mem3)\n",
    "\n",
    "            spk3_rec.append(spk3)\n",
    "            mem3_rec.append(mem3)\n",
    "\n",
    "        return torch.stack(spk3_rec, dim=0), torch.stack(mem3_rec, dim=0)\n",
    "\n",
    "        \n",
    "# Load the network onto CUDA if available\n",
    "sclu = SCLU().to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "FSyvR8ZIRWfj"
   },
   "outputs": [],
   "source": [
    "loss = nn.CrossEntropyLoss()\n",
    "optimizer = torch.optim.Adam(sclu.parameters(), lr=1e-3, betas=(0.9, 0.999))\n",
    "\n",
    "num_epochs = 10\n",
    "loss_hist = []\n",
    "test_loss_hist = []\n",
    "counter = 0\n",
    "\n",
    "# Outer training loop\n",
    "for epoch in range(num_epochs):\n",
    "\n",
    "    # Minibatch training loop\n",
    "    for data, targets in iter(trainloader):\n",
    "        data = data.to(device)\n",
    "        targets = targets.to(device)\n",
    "\n",
    "        # forward pass\n",
    "        sclu.train()\n",
    "        spk_rec, _ = sclu(data)\n",
    "\n",
    "        # initialize the loss & sum over time\n",
    "        loss_val = torch.zeros((1), dtype=dtype, device=device)\n",
    "        loss_val += loss(spk_rec.sum(0), targets)\n",
    "\n",
    "        # Gradient calculation + weight update\n",
    "        optimizer.zero_grad()\n",
    "        loss_val.backward()\n",
    "        optimizer.step()\n",
    "\n",
    "        # Store loss history for future plotting\n",
    "        loss_hist.append(loss_val.item())\n",
    "\n",
    "        # Print train/test loss/accuracy\n",
    "        if counter % 10 == 0:\n",
    "            print(f\"Iteration: {counter} \\t Train Loss: {loss_val.item()}\")\n",
    "        counter += 1\n",
    "\n",
    "        if counter == 100:\n",
    "          break"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "Nx11Ubl6U13N"
   },
   "outputs": [],
   "source": [
    "print(f\"SCLU Accuracy: {measure_accuracy(sclu, testloader)}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "8IHkC8-9aOWD"
   },
   "outputs": [],
   "source": [
    "spk_rec.size()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "56wHzZEqage_"
   },
   "outputs": [],
   "source": [
    "print(targets[0])\n",
    "spk_rec[:, 0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "3EVG5EGdD3zD"
   },
   "source": [
    "# 5. Precise Spike-Timing\n",
    "\n",
    "We're still figuring this bit out!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "6hPzs0J9Y8x6"
   },
   "outputs": [],
   "source": [
    "net = Net().to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "MP-BLuS6D5vR"
   },
   "outputs": [],
   "source": [
    "# spike-timing loss??? \n",
    "loss = SF.ce_temporal_loss(on_target=5, off_target=20)\n",
    "\n",
    "\n",
    "optimizer = torch.optim.Adam(net.parameters(), lr=1e-3, betas=(0.9, 0.999))\n",
    "\n",
    "num_epochs = 100\n",
    "loss_hist = []\n",
    "test_loss_hist = []\n",
    "counter = 0\n",
    "\n",
    "# Outer training loop\n",
    "for epoch in range(num_epochs):\n",
    "\n",
    "    # Minibatch training loop\n",
    "    for data, targets in iter(train_loader):\n",
    "        data = data.to(device)\n",
    "        targets = targets.to(device)\n",
    "\n",
    "        # forward pass\n",
    "        sclu.train()\n",
    "        spk_rec, _ = sclu(data)\n",
    "\n",
    "        # initialize the loss & sum over time\n",
    "        # loss_val = torch.zeros((1), dtype=dtype, device=device)\n",
    "        loss_val = loss(spk_rec, targets)\n",
    "\n",
    "        # Gradient calculation + weight update\n",
    "        optimizer.zero_grad()\n",
    "        loss_val.backward()\n",
    "        optimizer.step()\n",
    "\n",
    "        # Store loss history for future plotting\n",
    "        loss_hist.append(loss_val.item())\n",
    "\n",
    "        # Print train/test loss/accuracy\n",
    "        if counter % 10 == 0:\n",
    "            print(f\"Iteration: {counter} \\t Train Loss: {loss_val.item()}\")\n",
    "        counter += 1\n",
    "\n",
    "        # if counter == 100:\n",
    "        #   break"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "f8Z9k4mYYDFG"
   },
   "outputs": [],
   "source": [
    "def temporal_accuracy(model, dataloader):\n",
    "  with torch.no_grad():\n",
    "    model.eval()\n",
    "    running_length = 0\n",
    "    running_accuracy = 0\n",
    "\n",
    "    for data, targets in iter(dataloader):\n",
    "      data = data.to(device)\n",
    "      targets = targets.to(device)\n",
    "\n",
    "      # forward-pass\n",
    "      spk_rec, _ = model(data)\n",
    "\n",
    "      # total accuracy\n",
    "      running_length += len(targets)\n",
    "      running_accuracy += SF.accuracy_temporal(spk_rec, targets)\n",
    "\n",
    "    \n",
    "    accuracy = (running_accuracy / running_length)\n",
    "\n",
    "    return accuracy.item()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "D-FwmkUeXqKP"
   },
   "outputs": [],
   "source": [
    "print(f\"SCLU Accuracy: {temporal_accuracy(sclu, testloader)}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Z9vrb2zUD6S-"
   },
   "source": [
    "# 6. Hardware Friendly Training\n",
    "## 6.1 Weight Quantization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 4229,
     "status": "ok",
     "timestamp": 1677782612415,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "icK4WzuL-2QA",
    "outputId": "04ddf5de-2875-4cbe-df09-8c58a1e45d9a"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[?25l     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m0.0/357.3 KB\u001b[0m \u001b[31m?\u001b[0m eta \u001b[36m-:--:--\u001b[0m\r",
      "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m357.3/357.3 KB\u001b[0m \u001b[31m10.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[?25h"
     ]
    }
   ],
   "source": [
    "!pip install brevitas --quiet"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "PiZYTMA6D-YL"
   },
   "outputs": [],
   "source": [
    "import brevitas.nn as qnn\n",
    "from snntorch.functional import quant\n",
    "\n",
    "# Define Network\n",
    "class Net(nn.Module):\n",
    "    def __init__(self):\n",
    "        super().__init__()\n",
    "\n",
    "        qlif = quant.state_quant(num_bits=8, uniform=False, thr_centered=True)\n",
    "\n",
    "        # Initialize layers\n",
    "        self.fc1 = qnn.QuantLinear(num_inputs, num_hidden, weight_bit_width=8, bias=False)\n",
    "        self.lif1 = snn.Leaky(beta=beta, state_quant=qlif)\n",
    "        self.fc2 = qnn.QuantLinear(num_hidden, num_outputs, weight_bit_width=8, bias=False)\n",
    "        self.lif2 = snn.Leaky(beta=beta, state_quant=qlif)\n",
    "\n",
    "    def forward(self, x):\n",
    "\n",
    "        # Initialize hidden states at t=0\n",
    "        mem1 = self.lif1.init_leaky()\n",
    "        mem2 = self.lif2.init_leaky()\n",
    "        \n",
    "        # Record the final layer\n",
    "        spk2_rec = []\n",
    "        mem2_rec = []\n",
    "\n",
    "        for step in range(num_steps):\n",
    "            cur1 = self.fc1(x)\n",
    "            spk1, mem1 = self.lif1(cur1, mem1)\n",
    "            cur2 = self.fc2(spk1)\n",
    "            spk2, mem2 = self.lif2(cur2, mem2)\n",
    "            spk2_rec.append(spk2)\n",
    "            mem2_rec.append(mem2)\n",
    "\n",
    "        return torch.stack(spk2_rec, dim=0), torch.stack(mem2_rec, dim=0)\n",
    "        \n",
    "# Load the network onto CUDA if available\n",
    "net = Net().to(device)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "JvRzLHdjD-6S"
   },
   "source": [
    "## 6.2 State Quantization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "tAaafUczEAu_"
   },
   "outputs": [],
   "source": [
    "from snntorch.functional import quant\n",
    "\n",
    "# Define Network\n",
    "class Net(nn.Module):\n",
    "    def __init__(self):\n",
    "        super().__init__()\n",
    "\n",
    "\n",
    "        return torch.stack(spk2_rec, dim=0), torch.stack(mem2_rec, dim=0)\n",
    "        \n",
    "# Load the network onto CUDA if available\n",
    "net = Net().to(device)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "rNV1aGoRECig"
   },
   "source": [
    "# 7. Hyperparameter Optimization with Optuna"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 18062,
     "status": "ok",
     "timestamp": 1677774600533,
     "user": {
      "displayName": "Jason Eshraghian",
      "userId": "17871655902437888594"
     },
     "user_tz": 480
    },
    "id": "j7BmK8d7EHTS",
    "outputId": "17e8201a-99be-4354-f5bb-636e410a7924"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m365.3/365.3 KB\u001b[0m \u001b[31m12.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m210.5/210.5 KB\u001b[0m \u001b[31m2.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m78.7/78.7 KB\u001b[0m \u001b[31m4.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[?25hLooking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
      "Collecting optuna-dashboard\n",
      "  Downloading optuna_dashboard-0.8.1-py3-none-any.whl (2.5 MB)\n",
      "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m2.5/2.5 MB\u001b[0m \u001b[31m42.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[?25hRequirement already satisfied: packaging in /usr/local/lib/python3.8/dist-packages (from optuna-dashboard) (23.0)\n",
      "Requirement already satisfied: optuna>=2.4.0 in /usr/local/lib/python3.8/dist-packages (from optuna-dashboard) (3.1.0)\n",
      "Collecting bottle\n",
      "  Downloading bottle-0.12.24-py3-none-any.whl (90 kB)\n",
      "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m90.2/90.2 KB\u001b[0m \u001b[31m9.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
      "\u001b[?25hRequirement already satisfied: scikit-learn in /usr/local/lib/python3.8/dist-packages (from optuna-dashboard) (1.0.2)\n",
      "Requirement already satisfied: tqdm in /usr/local/lib/python3.8/dist-packages (from optuna>=2.4.0->optuna-dashboard) (4.64.1)\n",
      "Requirement already satisfied: colorlog in /usr/local/lib/python3.8/dist-packages (from optuna>=2.4.0->optuna-dashboard) (6.7.0)\n",
      "Requirement already satisfied: cmaes>=0.9.1 in /usr/local/lib/python3.8/dist-packages (from optuna>=2.4.0->optuna-dashboard) (0.9.1)\n",
      "Requirement already satisfied: sqlalchemy>=1.3.0 in /usr/local/lib/python3.8/dist-packages (from optuna>=2.4.0->optuna-dashboard) (1.4.46)\n",
      "Requirement already satisfied: alembic>=1.5.0 in /usr/local/lib/python3.8/dist-packages (from optuna>=2.4.0->optuna-dashboard) (1.9.4)\n",
      "Requirement already satisfied: PyYAML in /usr/local/lib/python3.8/dist-packages (from optuna>=2.4.0->optuna-dashboard) (6.0)\n",
      "Requirement already satisfied: numpy in /usr/local/lib/python3.8/dist-packages (from optuna>=2.4.0->optuna-dashboard) (1.22.4)\n",
      "Requirement already satisfied: scipy>=1.1.0 in /usr/local/lib/python3.8/dist-packages (from scikit-learn->optuna-dashboard) (1.7.3)\n",
      "Requirement already satisfied: joblib>=0.11 in /usr/local/lib/python3.8/dist-packages (from scikit-learn->optuna-dashboard) (1.2.0)\n",
      "Requirement already satisfied: threadpoolctl>=2.0.0 in /usr/local/lib/python3.8/dist-packages (from scikit-learn->optuna-dashboard) (3.1.0)\n",
      "Requirement already satisfied: importlib-resources in /usr/local/lib/python3.8/dist-packages (from alembic>=1.5.0->optuna>=2.4.0->optuna-dashboard) (5.12.0)\n",
      "Requirement already satisfied: importlib-metadata in /usr/local/lib/python3.8/dist-packages (from alembic>=1.5.0->optuna>=2.4.0->optuna-dashboard) (6.0.0)\n",
      "Requirement already satisfied: Mako in /usr/local/lib/python3.8/dist-packages (from alembic>=1.5.0->optuna>=2.4.0->optuna-dashboard) (1.2.4)\n",
      "Requirement already satisfied: greenlet!=0.4.17 in /usr/local/lib/python3.8/dist-packages (from sqlalchemy>=1.3.0->optuna>=2.4.0->optuna-dashboard) (2.0.2)\n",
      "Requirement already satisfied: zipp>=0.5 in /usr/local/lib/python3.8/dist-packages (from importlib-metadata->alembic>=1.5.0->optuna>=2.4.0->optuna-dashboard) (3.15.0)\n",
      "Requirement already satisfied: MarkupSafe>=0.9.2 in /usr/local/lib/python3.8/dist-packages (from Mako->alembic>=1.5.0->optuna>=2.4.0->optuna-dashboard) (2.1.2)\n",
      "Installing collected packages: bottle, optuna-dashboard\n",
      "Successfully installed bottle-0.12.24 optuna-dashboard-0.8.1\n"
     ]
    }
   ],
   "source": [
    "!pip install optuna --quiet\n",
    "!pip install optuna-dashboard"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "Tg_y39e87-TH"
   },
   "outputs": [],
   "source": [
    "# imports\n",
    "import optuna\n",
    "import joblib"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "el71FWNiKlFC"
   },
   "outputs": [],
   "source": [
    "def sample_parameters(trial):\n",
    "    \"\"\"Initialize randomly sampled parameters.\"\"\"\n",
    "\n",
    "    batch_size = trial.suggest_categorical('batch_size', [32, 64, 128, 256])\n",
    "    lr = trial.suggest_float('lr', 5e-5, 1e-2, log=True)\n",
    "    dropout_1 = trial.suggest_float('dropout_1', 0, 1)\n",
    "    dropout_2 = trial.suggest_float('dropout_2', 0, 1)\n",
    "    beta_1 = trial.suggest_float('beta_1', 0, 1)\n",
    "    beta_2 = trial.suggest_float('beta_2', 0, 1)\n",
    "    threshold_1 = trial.suggest_float('threshold_1', 1e-3, 1, log=True)\n",
    "    threshold_2 = trial.suggest_float('threshold_2', 1e-3, 1, log=True)\n",
    "\n",
    "    return (batch_size, lr, dropout_1, dropout_2, beta_1, beta_2, threshold_1, threshold_2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "ODBSkTteKrbg"
   },
   "outputs": [],
   "source": [
    "# parameters to optimize\n",
    "\n",
    "def config(trial):\n",
    "    params = sample_parameters(trial)\n",
    "    batch_size, lr, dropout_1, dropout_2, beta_1, beta_2, threshold_1, threshold_2 = params\n",
    "\n",
    "\n",
    "    config = {\n",
    "        'batch_size' : batch_size,\n",
    "        'lr' : lr,\n",
    "        'dropout_1' : dropout_1,\n",
    "        'dropout_2' : dropout_2,\n",
    "        'beta_1' : beta_1,\n",
    "        'beta_2' : beta_2,\n",
    "        'threshold_1' : threshold_1,\n",
    "        'threshold_2' : threshold_2\n",
    "    }\n",
    "\n",
    "    measure_accuracy(convnet, test_loader)}\"\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "soZj4vLnKfF-"
   },
   "outputs": [],
   "source": [
    "def objective(trial):\n",
    "\n",
    "    # Training loop - optimize for test-acc\n",
    "\n",
    "    return test_acc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "CPGprbkp9Y-y"
   },
   "outputs": [],
   "source": [
    "n_trials = 5\n",
    "study_name = \"optuna_test\"\n",
    "\n",
    "\n",
    "pruner = optuna.pruners.SuccessiveHalvingPruner()\n",
    "sampler = optuna.samplers.TPESampler()\n",
    "storage_name = \"sqlite:///{}.db\".format(study_name)\n",
    "study = optuna.create_study(study_name=study_name, sampler=sampler, direction='maximize', pruner=pruner, storage=storage_name, load_if_exists=True)\n",
    "study.optimize(objective, n_trials=n_trials)\n",
    "\n",
    "# Create a dataframe from the study.\n",
    "df = study.trials_dataframe()\n",
    "\n",
    "print(f\"Best config: {study.best_params}\")\n",
    "joblib.dump(study, 'optuna.pkl')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "DXEMfIdj-SKd"
   },
   "source": [
    "## Interested in Becoming a Contributor?\n",
    "\n",
    "Random Ideas:\n",
    "\n",
    "* Neurons with adaptive thresholds\n",
    "\n",
    "* Different types of neurons, e.g., Legendre Memory Units\n",
    "\n",
    "* Write up tutorials! From integrating new learning rules, combining snnTorch with liquid state machines, or a spiking version of the forward-forward algorithm?\n",
    "\n",
    "* Dendritic Computation. Integration with Dendrify?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "pNrFs3ro-xUm"
   },
   "source": [
    "## Current Opportunities for Neuromorphs:\n",
    "\n",
    "* Telluride 2023 Applications are open! Call for applications [here](https://sites.google.com/view/telluride-2023/home?authuser=0).\n",
    "* Submit your work to [APL Machine Learning](https://aip.scitation.org/journal/aml)\n",
    "* Submit your work to the Special Issue on Linking Neuro-AI with Neuromorphic Computing to the [IEEE Journal on Emerging Topics in Circuits and Systems](https://ieee-cas.org/files/ieeecass/2023-02/JETCAS_CFP_23-Q4-r2.pdf)\n",
    "\n",
    "* ISCAS 2023 Tutorial: [How to Build Open-Source Neuromorphic Hardware and Algorithms](https://iscas2023.org/tutorials) by Jason Eshraghian & Charlotte Frenkel\n"
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "provenance": [
    {
     "file_id": "10VOBz8G0dFupvaGmRhdL5kj3bfAiCz0Z",
     "timestamp": 1677775596774
    },
    {
     "file_id": "1JqZzSJZk7CXXJA5vBOEKnpGVQUpHCQ9o",
     "timestamp": 1677721881214
    }
   ]
  },
  "gpuClass": "standard",
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.8.18"
  },
  "widgets": {
   "application/vnd.jupyter.widget-state+json": {
    "02ae05cdb81f40cfb3a61915c40a113d": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "0783e220986744a598f0c7c034b1905d": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "09c0667250db4a10af4b6540d847589e": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "FloatProgressModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "FloatProgressModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "ProgressView",
      "bar_style": "success",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_4dc765eb8fdd40379331c34b1163c7fa",
      "max": 1648877,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_45b5ab7c9350401fa9f3d902c770b318",
      "value": 1648877
     }
    },
    "12c820a3950645d08234b0f62e737a8f": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_68106930402a441eb79acfd1da06acc8",
      "placeholder": "​",
      "style": "IPY_MODEL_6943b4e4ae504532a9dc7c67da00b874",
      "value": " 311296/? [00:00&lt;00:00, 552882.73it/s]"
     }
    },
    "12f42925d27245feacbbb67b71b111c3": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "1346b7bb93a943fe86d30de2f4dd5c7f": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "175878491cd44bf6a30a707de283c7b0": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "1ba6a17991704c5ca752233b815fa3fc": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "ProgressStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "ProgressStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "bar_color": null,
      "description_width": ""
     }
    },
    "1d0f4a482fab425b9fe3642ad8cdca88": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HBoxModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HBoxModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HBoxView",
      "box_style": "",
      "children": [
       "IPY_MODEL_4b5e1725eaf44fa9af29c23065998cbd",
       "IPY_MODEL_3a95403d6c374148b27635dde7a479e4",
       "IPY_MODEL_cc751284ebfc47668e22b184d2b88432"
      ],
      "layout": "IPY_MODEL_b2ff14ee0abf4ccc8463a92615805c21"
     }
    },
    "2da74e44fe2b4b50aad73fdf7f02c3b6": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "FloatProgressModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "FloatProgressModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "ProgressView",
      "bar_style": "success",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_12f42925d27245feacbbb67b71b111c3",
      "max": 700096,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_649a0f85a7e4403997aaa2b2ada8d887",
      "value": 700096
     }
    },
    "2e78af0a8dfe4d1084bfd6974dc38268": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "3288ef76407441c7a06b35e45fd5bcbd": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_e3b431043c144785a95d0e387b3f65ac",
      "placeholder": "​",
      "style": "IPY_MODEL_02ae05cdb81f40cfb3a61915c40a113d",
      "value": " 1648877/1648877 [00:00&lt;00:00, 24922580.21it/s]"
     }
    },
    "37653a294ae645b1b64363d27aaf496e": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "37898c9ed3374007ac38913b228b5bb8": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "3a95403d6c374148b27635dde7a479e4": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "FloatProgressModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "FloatProgressModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "ProgressView",
      "bar_style": "success",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_ddedd2970a32439fb11dc57a9f675b42",
      "max": 4542,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_cd564c42dc3c4ed5a41df9eaa6ca9b2e",
      "value": 4542
     }
    },
    "3bd55950664d4e6e96a3dd553fac73f1": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_3dc418e477bb4846840ea683f3541c50",
      "placeholder": "​",
      "style": "IPY_MODEL_c32e21f24e9e4efb80393aabde1b2679",
      "value": ""
     }
    },
    "3dc418e477bb4846840ea683f3541c50": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "45b5ab7c9350401fa9f3d902c770b318": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "ProgressStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "ProgressStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "bar_color": null,
      "description_width": ""
     }
    },
    "47ec1d064cbd4fd79628900377476b89": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "ProgressStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "ProgressStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "bar_color": null,
      "description_width": ""
     }
    },
    "4953e7632467486a9e83a71c02f2c81b": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "FloatProgressModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "FloatProgressModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "ProgressView",
      "bar_style": "success",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_58e57832bb60410bb21b6e56013ac7e3",
      "max": 28881,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_8476b70863e94334ae796f5ecade108e",
      "value": 28881
     }
    },
    "4a74bb5f467f4ef8bfcca8fd77c749cf": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "4b5e1725eaf44fa9af29c23065998cbd": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_de12dfeeaa0d43ffb70dc035b5703fca",
      "placeholder": "​",
      "style": "IPY_MODEL_51442d3c20444c8b82d4131a64a7f0a0",
      "value": "100%"
     }
    },
    "4dc765eb8fdd40379331c34b1163c7fa": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "5133365e8fc849f5ae14ed583c156efb": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_1346b7bb93a943fe86d30de2f4dd5c7f",
      "placeholder": "​",
      "style": "IPY_MODEL_f9cc6c36fec84e6b962d3f23f32156b4",
      "value": " 28881/28881 [00:00&lt;00:00, 1348964.84it/s]"
     }
    },
    "51442d3c20444c8b82d4131a64a7f0a0": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "5163362add9c49258343be1043f69f6f": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "FloatProgressModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "FloatProgressModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "ProgressView",
      "bar_style": "success",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_ac978e74c8174866a72ae760b64f22cd",
      "max": 9912422,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_1ba6a17991704c5ca752233b815fa3fc",
      "value": 9912422
     }
    },
    "53a57f276a61483aa96fc195ee485187": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HBoxModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HBoxModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HBoxView",
      "box_style": "",
      "children": [
       "IPY_MODEL_3bd55950664d4e6e96a3dd553fac73f1",
       "IPY_MODEL_2da74e44fe2b4b50aad73fdf7f02c3b6",
       "IPY_MODEL_e478c3957c964155ab2148e79260ef09"
      ],
      "layout": "IPY_MODEL_a78f7c36cc7d4f2788f5d3a98d500520"
     }
    },
    "58e57832bb60410bb21b6e56013ac7e3": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "5b542a1db8904adebe23631200aa0b06": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "606551ca726b4998a55c7885b1c13e56": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "60e62bba3b9a499ea407ece948946349": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "649a0f85a7e4403997aaa2b2ada8d887": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "ProgressStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "ProgressStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "bar_color": null,
      "description_width": ""
     }
    },
    "68106930402a441eb79acfd1da06acc8": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "6943b4e4ae504532a9dc7c67da00b874": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "69446de0b5c146c6b0d3a1fa26d09c99": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "7b5359a82a114f1f9c7bb7eb2306dedc": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "828d6cac2cf54bcc88b33a1bda0d0bb7": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_5b542a1db8904adebe23631200aa0b06",
      "placeholder": "​",
      "style": "IPY_MODEL_7b5359a82a114f1f9c7bb7eb2306dedc",
      "value": "100%"
     }
    },
    "8476b70863e94334ae796f5ecade108e": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "ProgressStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "ProgressStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "bar_color": null,
      "description_width": ""
     }
    },
    "85c95cfc8352416fa745f160bce87195": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "9c47c04ca5cf424da91533412e4e16a4": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_2e78af0a8dfe4d1084bfd6974dc38268",
      "placeholder": "​",
      "style": "IPY_MODEL_60e62bba3b9a499ea407ece948946349",
      "value": "100%"
     }
    },
    "9d2488cd936c487eac69fe639ce6a77d": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HBoxModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HBoxModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HBoxView",
      "box_style": "",
      "children": [
       "IPY_MODEL_c4c7fa18359041dfa9960668fb884f69",
       "IPY_MODEL_b622c6be305a4d99a2a3ffda2babbb0e",
       "IPY_MODEL_12c820a3950645d08234b0f62e737a8f"
      ],
      "layout": "IPY_MODEL_0783e220986744a598f0c7c034b1905d"
     }
    },
    "a78f7c36cc7d4f2788f5d3a98d500520": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "ac978e74c8174866a72ae760b64f22cd": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "ae7dcbfa1dbc426697a0d98608f5c6c3": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HBoxModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HBoxModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HBoxView",
      "box_style": "",
      "children": [
       "IPY_MODEL_9c47c04ca5cf424da91533412e4e16a4",
       "IPY_MODEL_4953e7632467486a9e83a71c02f2c81b",
       "IPY_MODEL_5133365e8fc849f5ae14ed583c156efb"
      ],
      "layout": "IPY_MODEL_ee77acb5fac84e2ab27cc310dd88182d"
     }
    },
    "af7998e749d24f1687bb1d90c7774165": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HBoxModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HBoxModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HBoxView",
      "box_style": "",
      "children": [
       "IPY_MODEL_828d6cac2cf54bcc88b33a1bda0d0bb7",
       "IPY_MODEL_09c0667250db4a10af4b6540d847589e",
       "IPY_MODEL_3288ef76407441c7a06b35e45fd5bcbd"
      ],
      "layout": "IPY_MODEL_4a74bb5f467f4ef8bfcca8fd77c749cf"
     }
    },
    "b2ff14ee0abf4ccc8463a92615805c21": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "b622c6be305a4d99a2a3ffda2babbb0e": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "FloatProgressModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "FloatProgressModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "ProgressView",
      "bar_style": "success",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_c4631d6cbaa349108dd27edfc95006e3",
      "max": 311022,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_47ec1d064cbd4fd79628900377476b89",
      "value": 311022
     }
    },
    "bcce9f99717b4dae884fcdbe2d5ecce0": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_37653a294ae645b1b64363d27aaf496e",
      "placeholder": "​",
      "style": "IPY_MODEL_37898c9ed3374007ac38913b228b5bb8",
      "value": " 9912422/9912422 [00:00&lt;00:00, 119626843.27it/s]"
     }
    },
    "c32e21f24e9e4efb80393aabde1b2679": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "c4631d6cbaa349108dd27edfc95006e3": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "c4c7fa18359041dfa9960668fb884f69": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_f1cf4115d2944e40897bf6e29d7dbe8f",
      "placeholder": "​",
      "style": "IPY_MODEL_85c95cfc8352416fa745f160bce87195",
      "value": ""
     }
    },
    "c91df8aec8204e30837776b10c4ebe75": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "cc751284ebfc47668e22b184d2b88432": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_69446de0b5c146c6b0d3a1fa26d09c99",
      "placeholder": "​",
      "style": "IPY_MODEL_eacf93019f9e44bb826bff75b052ed72",
      "value": " 4542/4542 [00:00&lt;00:00, 232213.08it/s]"
     }
    },
    "cd564c42dc3c4ed5a41df9eaa6ca9b2e": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "ProgressStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "ProgressStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "bar_color": null,
      "description_width": ""
     }
    },
    "dab3e4d911d847128ae0a18d9041847f": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_606551ca726b4998a55c7885b1c13e56",
      "placeholder": "​",
      "style": "IPY_MODEL_c91df8aec8204e30837776b10c4ebe75",
      "value": "100%"
     }
    },
    "ddedd2970a32439fb11dc57a9f675b42": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "de12dfeeaa0d43ffb70dc035b5703fca": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "e08217b98d414bcd8d155181631bc695": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "e3b431043c144785a95d0e387b3f65ac": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "e478c3957c964155ab2148e79260ef09": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_e08217b98d414bcd8d155181631bc695",
      "placeholder": "​",
      "style": "IPY_MODEL_f74089a9d8494b0d9e0a365bb8b5ad37",
      "value": " 700416/? [00:00&lt;00:00, 774707.95it/s]"
     }
    },
    "eacf93019f9e44bb826bff75b052ed72": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "ee77acb5fac84e2ab27cc310dd88182d": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "f1cf4115d2944e40897bf6e29d7dbe8f": {
     "model_module": "@jupyter-widgets/base",
     "model_module_version": "1.2.0",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "f74089a9d8494b0d9e0a365bb8b5ad37": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "f9cc6c36fec84e6b962d3f23f32156b4": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "fc1cea022fd64d2580b636ad29ce74cb": {
     "model_module": "@jupyter-widgets/controls",
     "model_module_version": "1.5.0",
     "model_name": "HBoxModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HBoxModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HBoxView",
      "box_style": "",
      "children": [
       "IPY_MODEL_dab3e4d911d847128ae0a18d9041847f",
       "IPY_MODEL_5163362add9c49258343be1043f69f6f",
       "IPY_MODEL_bcce9f99717b4dae884fcdbe2d5ecce0"
      ],
      "layout": "IPY_MODEL_175878491cd44bf6a30a707de283c7b0"
     }
    }
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}