1_stable_rank_though_layers.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This notebook shows that stable rank is increasing layer by layer when neural network is learning."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from pylab import *\n",
    "from skimage.color import rgb2gray\n",
    "from skimage.transform import rescale,resize\n",
    "from torch.autograd import Variable\n",
    "import torch, time,os\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "import torch.optim as optim\n",
    "\n",
    "\n",
    "# Set the CUDA flag\n",
    "cuda_flag = True\n",
    "os.environ['CUDA_VISIBLE_DEVICES'] = '1'\n",
    "criterion = torch.nn.MSELoss()\n",
    "epochs = 2000\n",
    "lr1 = 1e-3\n",
    "lr2= 1e-4\n",
    "\n",
    "N_test = 10\n",
    "N_repeat = 10\n",
    "\n",
    "kk = torch.randint(512,(N_test,))\n",
    "kk = [158, 329,  47, 344, 257, 410, 426, 199, 426, 295]\n",
    "#kk = [ 88, 269, 409,  59, 245, 382, 277, 362,  23, 308, 504, 448, 411, 304, 365, 200,  92, 350, 405,  77]\n",
    "d = 20\n",
    "\n",
    "losses_ = np.zeros((2,N_test,N_repeat,epochs))\n",
    "sr_ = np.zeros((2,N_test,N_repeat,5,int(epochs/d)))\n",
    "trn_psnr_ = np.zeros((2,N_test,N_repeat,epochs))\n",
    "tst_psnr_ = np.zeros((2,N_test,N_repeat,epochs))\n",
    "\n",
    "file_path = 'Final_1n/'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "rgb = imread('4.2.07.tiff');im = rgb2gray(rgb);\n",
    "signal_true = torch.from_numpy(im).type(torch.FloatTensor)\n",
    "signal_random = torch.randn((512,512))\n",
    "signals = [signal_true,signal_random]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "width = 256\n",
    "\n",
    "# Here define the MLP\n",
    "class MLP(nn.Module):\n",
    "    def __init__(self,input_dim):\n",
    "        super(MLP, self).__init__()\n",
    "        self.fc1 = nn.Linear(input_dim, width)\n",
    "        self.fc2 = nn.Linear(width, width)\n",
    "        self.fc3 = nn.Linear(width, width)\n",
    "        self.fc4 = nn.Linear(width, width)\n",
    "        self.fco = nn.Linear(width, 1)\n",
    "        self.relu = nn.ReLU()\n",
    "    def forward(self, x):\n",
    "        x1 = self.fc1(x)\n",
    "        e1 = self.relu(x1)\n",
    "        x2 = self.fc2(e1)\n",
    "        e2 = self.relu(x2)\n",
    "        x3 = self.fc3(e2)\n",
    "        e3 = self.relu(x3)\n",
    "        x4 = self.fc4(e3)\n",
    "        e4 = self.relu(x4)\n",
    "        out = self.fco(e4)\n",
    "        return out,e4,e3,e2,e1\n",
    "    def name(self):\n",
    "        return \"MLP\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def srank_func(X):\n",
    "    # Get the SVD\n",
    "    (u,s,v) = torch.svd(X)\n",
    "    sr = (s*s).sum()/s[0]/s[0]\n",
    "    return sr\n",
    "\n",
    "def psnr_func(x,y):\n",
    "    diff = x - y\n",
    "    err = (diff*diff).flatten().mean().log10()\n",
    "    return -10*err\n",
    "\n",
    "def rff_embedding(x, b):\n",
    "    return torch.cat((torch.sin((2.*np.pi*x) @ b.T),torch.cos((2.*np.pi*x) @ b.T)),1)\n",
    "\n",
    "def basic_encoding(x):\n",
    "    return torch.cat((torch.sin((2.*np.pi*x)),torch.cos((2.*np.pi*x))),1)\n",
    "\n",
    "def rbf_embedding(x,dic,sig):\n",
    "    return (-0.5*(x-dic)**2/sig**2).exp()\n",
    "\n",
    "def norm_func(x):\n",
    "    # Flatten the data \n",
    "    #x = x.flatten(1,3)\n",
    "    \n",
    "    # Normalize for gain and bias\n",
    "    y = x-x.mean(1).unsqueeze(-1)\n",
    "    y = x/x.norm(dim=1).unsqueeze(-1)    \n",
    "    return y"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# no encoding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "==>>> E: 0, F: 0, train psnr: 18.1771--- , test psnr: 18.1705--- , time: 18.4692 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 18.2618--- , test psnr: 18.321--- , time: 32.3927 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 18.5272--- , test psnr: 18.5752--- , time: 46.7099 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 17.781--- , test psnr: 17.8098--- , time: 61.08 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 18.7177--- , test psnr: 18.7645--- , time: 75.4474 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 19.7685--- , test psnr: 19.8365--- , time: 89.8637 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 19.3584--- , test psnr: 19.3481--- , time: 104.182 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 19.4963--- , test psnr: 19.5712--- , time: 118.524 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 18.406--- , test psnr: 18.5391--- , time: 130.26 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 18.1003--- , test psnr: 18.1749--- , time: 144.736 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 22.8638--- , test psnr: 22.5428--- , time: 14.4911 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 23.6025--- , test psnr: 23.0736--- , time: 28.9952 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 24.4291--- , test psnr: 24.051--- , time: 43.526 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 22.6816--- , test psnr: 22.3469--- , time: 58.0623 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 23.2591--- , test psnr: 22.7428--- , time: 72.4857 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 22.0239--- , test psnr: 21.7008--- , time: 86.5923 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 22.5694--- , test psnr: 22.2891--- , time: 100.766 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 22.5677--- , test psnr: 22.2613--- , time: 117.821 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 23.3402--- , test psnr: 22.7879--- , time: 134.075 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 21.5965--- , test psnr: 21.3324--- , time: 150.216 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 21.1135--- , test psnr: 21.5852--- , time: 17.0188 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 21.0657--- , test psnr: 21.2096--- , time: 33.3669 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 20.08--- , test psnr: 20.2811--- , time: 50.6021 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 22.9923--- , test psnr: 22.812--- , time: 66.3351 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 22.7654--- , test psnr: 22.3812--- , time: 80.6165 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 21.7824--- , test psnr: 22.201--- , time: 94.8654 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 22.1911--- , test psnr: 23.0032--- , time: 110.127 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 20.219--- , test psnr: 20.5643--- , time: 127.773 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 21.9072--- , test psnr: 22.4569--- , time: 144.704 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 22.6381--- , test psnr: 22.6401--- , time: 161.452 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 23.458--- , test psnr: 23.1365--- , time: 16.3801 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 23.9859--- , test psnr: 23.6067--- , time: 32.4707 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 24.5926--- , test psnr: 23.9285--- , time: 48.1222 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 25.2039--- , test psnr: 24.643--- , time: 62.1599 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 22.9148--- , test psnr: 22.6351--- , time: 75.1613 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 25.0784--- , test psnr: 24.4371--- , time: 89.6496 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 23.5035--- , test psnr: 23.115--- , time: 105.409 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 24.8095--- , test psnr: 24.3573--- , time: 122.046 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 23.4928--- , test psnr: 23.1696--- , time: 138.579 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 24.585--- , test psnr: 24.055--- , time: 155.273 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 18.7819--- , test psnr: 18.6347--- , time: 16.2131 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 19.2291--- , test psnr: 19.0436--- , time: 32.5195 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 18.9958--- , test psnr: 18.8318--- , time: 49.2664 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 19.8772--- , test psnr: 19.6745--- , time: 63.1291 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 19.4542--- , test psnr: 19.2326--- , time: 78.3222 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 19.4636--- , test psnr: 19.262--- , time: 92.7005 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 19.3366--- , test psnr: 19.1045--- , time: 106.38 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 17.8165--- , test psnr: 17.7164--- , time: 120.763 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 19.2287--- , test psnr: 19.0608--- , time: 135.164 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 18.673--- , test psnr: 18.5015--- , time: 149.604 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 24.9657--- , test psnr: 25.381--- , time: 15.894 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 24.5048--- , test psnr: 25.0714--- , time: 30.3029 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 26.0327--- , test psnr: 26.0064--- , time: 44.8035 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 24.9469--- , test psnr: 25.4461--- , time: 59.2411 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 24.3988--- , test psnr: 25.2262--- , time: 73.6328 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 24.3505--- , test psnr: 25.0572--- , time: 87.9519 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 24.8439--- , test psnr: 24.8256--- , time: 102.321 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 24.4875--- , test psnr: 25.1209--- , time: 116.428 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 23.7382--- , test psnr: 24.4237--- , time: 130.728 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 24.5557--- , test psnr: 25.3776--- , time: 145.202 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 25.2174--- , test psnr: 25.3674--- , time: 14.4025 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 23.4714--- , test psnr: 24.9316--- , time: 28.7461 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 23.9805--- , test psnr: 25.0533--- , time: 43.1719 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 25.01--- , test psnr: 25.7716--- , time: 57.5031 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 24.7352--- , test psnr: 25.865--- , time: 71.8472 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 23.4204--- , test psnr: 24.5797--- , time: 87.6401 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 24.07--- , test psnr: 25.1112--- , time: 102.012 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 24.1316--- , test psnr: 25.1277--- , time: 116.173 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 25.1589--- , test psnr: 25.6301--- , time: 130.396 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 24.8931--- , test psnr: 25.5088--- , time: 144.574 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 20.6459--- , test psnr: 20.4265--- , time: 14.2993 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 20.7972--- , test psnr: 20.5787--- , time: 28.5484 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 19.7212--- , test psnr: 19.5887--- , time: 42.6098 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 20.0161--- , test psnr: 19.8738--- , time: 56.8348 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 20.1131--- , test psnr: 19.9366--- , time: 71.0139 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 19.7482--- , test psnr: 19.6455--- , time: 85.2761 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 20.3417--- , test psnr: 20.2026--- , time: 99.5613 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 22.0205--- , test psnr: 21.6464--- , time: 113.904 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 18.6137--- , test psnr: 18.5477--- , time: 128.182 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 19.9728--- , test psnr: 19.8068--- , time: 142.416 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 25.5673--- , test psnr: 26.0636--- , time: 14.5524 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 24.0933--- , test psnr: 25.4561--- , time: 28.9572 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 25.1719--- , test psnr: 26.1669--- , time: 43.4244 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 24.8729--- , test psnr: 25.3262--- , time: 57.9219 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 23.6506--- , test psnr: 24.3226--- , time: 72.5009 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 22.7462--- , test psnr: 24.2976--- , time: 88.1153 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 24.5857--- , test psnr: 26.1548--- , time: 104.473 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 24.2677--- , test psnr: 25.6863--- , time: 119.309 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 23.6119--- , test psnr: 24.4945--- , time: 133.491 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 25.6639--- , test psnr: 26.2104--- , time: 147.86 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 22.2067--- , test psnr: 22.2629--- , time: 14.2087 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 22.4303--- , test psnr: 22.4008--- , time: 28.5307 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 23.4455--- , test psnr: 23.5307--- , time: 45.1147 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 23.7243--- , test psnr: 23.8497--- , time: 60.985 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 23.4304--- , test psnr: 23.498--- , time: 75.7552 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 22.8133--- , test psnr: 22.9277--- , time: 90.2468 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 21.9958--- , test psnr: 22.0033--- , time: 104.635 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 22.5927--- , test psnr: 22.663--- , time: 119.013 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 22.1463--- , test psnr: 22.125--- , time: 133.482 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 23.1802--- , test psnr: 23.1958--- , time: 147.896 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 0.578099--- , test psnr: -0.552303--- , time: 14.4452 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 0.631344--- , test psnr: -0.569958--- , time: 29.0315 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 0.613434--- , test psnr: -0.541953--- , time: 43.5695 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 0.581237--- , test psnr: -0.537382--- , time: 58.0343 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 0.577234--- , test psnr: -0.520767--- , time: 72.5794 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 0.626937--- , test psnr: -0.556824--- , time: 87.0758 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 0.607013--- , test psnr: -0.548179--- , time: 101.591 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 0.671833--- , test psnr: -0.623322--- , time: 116.245 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 0.591273--- , test psnr: -0.557781--- , time: 130.806 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 0.661862--- , test psnr: -0.587026--- , time: 145.397 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 0.945615--- , test psnr: -0.387613--- , time: 14.5885 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 1.03276--- , test psnr: -0.460221--- , time: 29.7828 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 0.915411--- , test psnr: -0.343384--- , time: 44.2858 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 0.924201--- , test psnr: -0.371912--- , time: 58.8686 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 1.00206--- , test psnr: -0.4284--- , time: 73.386 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 1.08498--- , test psnr: -0.42035--- , time: 87.8183 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 0.921346--- , test psnr: -0.394732--- , time: 102.323 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 1.05181--- , test psnr: -0.403969--- , time: 116.861 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 0.997816--- , test psnr: -0.445603--- , time: 133.132 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 0.900249--- , test psnr: -0.304271--- , time: 150.217 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: -0.0363507--- , test psnr: 0.739575--- , time: 16.699 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 0.121366--- , test psnr: 0.710658--- , time: 33.6796 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 0.0452936--- , test psnr: 0.634701--- , time: 50.4111 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 0.134492--- , test psnr: 0.58713--- , time: 67.8874 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 0.00218854--- , test psnr: 0.706276--- , time: 84.7481 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 0.216535--- , test psnr: 0.64645--- , time: 102.764 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 0.0103587--- , test psnr: 0.710782--- , time: 119.177 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 0.118328--- , test psnr: 0.713113--- , time: 135.906 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 0.00255352--- , test psnr: 0.760981--- , time: 152.494 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 0.1253--- , test psnr: 0.67719--- , time: 169.777 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 0.052494--- , test psnr: -0.689501--- , time: 17.7019 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 0.0185349--- , test psnr: -0.667194--- , time: 34.8932 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 0.0265265--- , test psnr: -0.674782--- , time: 51.8208 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 0.0761306--- , test psnr: -0.710416--- , time: 68.7161 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 0.0629277--- , test psnr: -0.670985--- , time: 85.7395 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 0.0475671--- , test psnr: -0.688763--- , time: 101.964 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 0.0505915--- , test psnr: -0.688367--- , time: 118.991 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 0.0771532--- , test psnr: -0.69146--- , time: 135.384 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 0.0391124--- , test psnr: -0.66354--- , time: 152.052 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 0.0579269--- , test psnr: -0.704666--- , time: 168.29 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 1.15552--- , test psnr: -0.736117--- , time: 16.7868 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 1.05555--- , test psnr: -0.685192--- , time: 33.6659 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 1.14444--- , test psnr: -0.793571--- , time: 51.249 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 1.03515--- , test psnr: -0.692095--- , time: 68.0808 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 1.19712--- , test psnr: -0.754728--- , time: 85.3319 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 1.03245--- , test psnr: -0.684046--- , time: 102.184 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 1.13508--- , test psnr: -0.698554--- , time: 119.592 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 1.06405--- , test psnr: -0.731447--- , time: 136.798 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 1.05392--- , test psnr: -0.695197--- , time: 154.372 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 1.14922--- , test psnr: -0.732546--- , time: 170.56 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 0.609755--- , test psnr: -0.922422--- , time: 16.7045 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 0.576304--- , test psnr: -0.942524--- , time: 33.7906 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 0.584625--- , test psnr: -0.91164--- , time: 51.08 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 0.646793--- , test psnr: -1.00996--- , time: 69.3944 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 0.575998--- , test psnr: -0.93119--- , time: 86.6531 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 0.600735--- , test psnr: -0.922357--- , time: 106.214 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 0.690855--- , test psnr: -1.01884--- , time: 123.826 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 0.609983--- , test psnr: -0.947623--- , time: 140.362 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 0.605671--- , test psnr: -1.05181--- , time: 158.261 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 0.701573--- , test psnr: -1.06439--- , time: 175.249 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 0.380497--- , test psnr: -0.00561744--- , time: 16.9888 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 0.338133--- , test psnr: -0.00832164--- , time: 34.8845 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 0.411373--- , test psnr: -0.0783165--- , time: 51.8775 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 0.374967--- , test psnr: -0.0221094--- , time: 68.4081 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 0.334196--- , test psnr: -0.0120336--- , time: 85.5693 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 0.365786--- , test psnr: -0.0424158--- , time: 102.169 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 0.338033--- , test psnr: -0.0249631--- , time: 120.005 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 0.38285--- , test psnr: -0.0575934--- , time: 136.568 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 0.410357--- , test psnr: -0.0882197--- , time: 152.649 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 0.389267--- , test psnr: -0.0760614--- , time: 170.16 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 1.82386--- , test psnr: 0.367152--- , time: 17.3894 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 1.75857--- , test psnr: 0.417148--- , time: 33.7434 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 1.89534--- , test psnr: 0.308753--- , time: 50.8464 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 1.76019--- , test psnr: 0.416366--- , time: 68.085 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 1.83471--- , test psnr: 0.393989--- , time: 85.3744 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 1.82934--- , test psnr: 0.356096--- , time: 102.768 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 1.84932--- , test psnr: 0.346796--- , time: 119.407 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 1.8072--- , test psnr: 0.357958--- , time: 135.217 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 1.87231--- , test psnr: 0.363201--- , time: 151.233 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 1.8322--- , test psnr: 0.37849--- , time: 169.002 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 0.385852--- , test psnr: -0.0716425--- , time: 18.4317 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 0.351948--- , test psnr: -0.012516--- , time: 35.4325 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 0.342703--- , test psnr: -0.0292297--- , time: 52.1486 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 0.399317--- , test psnr: -0.0848569--- , time: 69.1303 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 0.394664--- , test psnr: -0.0583294--- , time: 85.9576 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 0.323576--- , test psnr: -0.010569--- , time: 102.942 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 0.373918--- , test psnr: -0.0412585--- , time: 120.51 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 0.344554--- , test psnr: 0.0263373--- , time: 137.498 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 0.364955--- , test psnr: -0.0431729--- , time: 153.34 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 0.355257--- , test psnr: -0.0284404--- , time: 171.241 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 0.212781--- , test psnr: -0.61501--- , time: 17.2036 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 0.226133--- , test psnr: -0.610162--- , time: 34.8079 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 0.256829--- , test psnr: -0.61607--- , time: 51.5933 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 0.23687--- , test psnr: -0.623758--- , time: 68.2497 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 0.243604--- , test psnr: -0.541314--- , time: 85.5348 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 0.149429--- , test psnr: -0.545113--- , time: 103.302 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 0.249313--- , test psnr: -0.619615--- , time: 120.615 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 0.195668--- , test psnr: -0.579768--- , time: 137.247 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 0.184438--- , test psnr: -0.564406--- , time: 154.75 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 0.322126--- , test psnr: -0.673409--- , time: 172.194 seconds ---\n"
     ]
    }
   ],
   "source": [
    "sample_N = 512\n",
    "sample_d = 1\n",
    "\n",
    "\n",
    "all_data = torch.from_numpy(np.linspace(0, 1, sample_N+1)[:-1].reshape(-1,1)).type(torch.FloatTensor)\n",
    "\n",
    "train_data = all_data[::2*sample_d]\n",
    "test_data = all_data[sample_d::2*sample_d]\n",
    "\n",
    "# Calculate the input stable rank\n",
    "sr_[:,:,:,0] = srank_func(train_data)\n",
    "\n",
    "for i in range(2):\n",
    "    for j in range(N_test):\n",
    "        start_time = time.time()\n",
    "        \n",
    "        \n",
    "        # Prepare the data\n",
    "        all_target = signals[i][kk[j]].reshape(-1,1)\n",
    "        train_label = all_target[::2*sample_d]\n",
    "        test_label = all_target[sample_d::2*sample_d]\n",
    "        \n",
    "        xt, yt = train_data.cuda(),train_label.cuda()\n",
    "        xt, yt = Variable(xt), Variable(yt)\n",
    "\n",
    "        xts, yts = test_data.cuda(),test_label.cuda()\n",
    "        \n",
    "        \n",
    "        for k in range(N_repeat):\n",
    "            # Initialize classification model to learn\n",
    "            model = MLP(1).cuda()\n",
    "            # Set the optimization\n",
    "            optimizer = optim.Adam(model.parameters(), lr1, betas=(0.9, 0.999),weight_decay=1e-8)\n",
    "            \n",
    "\n",
    "            for epoch in range(epochs):\n",
    "                model.train()\n",
    "                optimizer.zero_grad()\n",
    "\n",
    "                out,e4,e3,e2,e1 = model(xt)\n",
    "                loss = criterion(out, yt)\n",
    "\n",
    "                loss.backward()\n",
    "                losses_[i,j,k,epoch] = loss.item()\n",
    "                optimizer.step()\n",
    "\n",
    "                model.eval()\n",
    "                trn_psnr_[i,j,k,epoch] = psnr_func(model(xt)[0],yt)\n",
    "                tst_psnr_[i,j,k,epoch] = psnr_func(model(xts)[0],yts)\n",
    "\n",
    "                if epoch%d == 0:\n",
    "                    es = [e1,e2,e3,e4]\n",
    "                    for l in range(4):\n",
    "                        sr_[i,j,k,l+1,epoch//d] = srank_func(es[l].detach().cpu())\n",
    "\n",
    "            print(\"==>>> E: %g, F: %g, train psnr: %g--- , test psnr: %g--- , time: %g seconds ---\" \n",
    "                    % (i,j, np.mean(trn_psnr_[i,j,k,:-100]),np.mean(tst_psnr_[i,j,k,:-100]),time.time() - start_time))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAD4CAYAAAD8Zh1EAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjAsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+17YcXAAAgAElEQVR4nO3deXQcZ53v//fT+6rW0tp32/JuJ3a8JQ44OyETsjEQciGBDEOGYQiEGeY3cO49BH4z3Lmc2WAOhNwQQgiEwLAEMgkQsi+OtySON3mXbVn72up9f+4fLcubZMl2Sy21vq9z+nR3VXXVt9zSx6WnnnpKaa0RQggx8xlyXYAQQojskEAXQog8IYEuhBB5QgJdCCHyhAS6EELkCVOuNuz1enVDQ0OuNi+EEDPSO++806e1Lh1tXs4CvaGhgbfffjtXmxdCiBlJKXVsrHnS5CKEEHlCAl0IIfKEBLoQQuQJCXQhhMgTEuhCCJEnJNCFECJPSKALIUSeyFk/dCGEyGdaa3QkQioYJB0IkOztJdHZRbKrE9vy5bjWr8/6NiXQhRBiWGpoiERXN+lQkHQoRDocQcfj6HiMdDBIoqOTREcHyf5+lNmMslpQJjOpoSFSg4OkfD50LIZOJNCJBIxxv4mSz3xGAl0IIS6ETiRIR6OZkA4ESAWDJLt7iB87lnkcPUr86FFSAwPnXI+y2TBXVWEqLUWnkqQHQuhkEmNBAdamJoyFhRhsVpTZDGYzRqcTg8uFwenCVFqKubICU3k5BpttUvZTAl0IMSOlgiGSXZ2ZZozuLtKxGGggnSbR0UHs4EFihw6RHBiARGLM9RhLvVjq63Ffew2WhkbM1VUYnC4MTicGhx1lsWKwWlAOB8bCQpRSU7eT50kCXQgxrejhQI4fyRw1J9qOkxrykwoGSPsDJPv6SHZ3kw6FxlyHslqxzp2Lc91aTGXlmWC22zE4HBjdbgwuF6aSEsx19Rhdzincu8k1bqArpR4DbgZ6tNZLx1jmKuDbgBno01pvyGaRQoj8oVOpTJtzfz/Jvj4SHR0k2jtItLcRO3iIWEsLOhodWV7Z7RgLCzG6XBjcbqxNTTjXr8dUVoq5ohJzVSWm8goMdhsoBUphLChAGY053MvcmMgR+uPAd4EnRpuplCoEHgJu1Fq3KqXKsleeEGKm0uk0idZWos3NRPftJ3boELGDB0m0t0M6ffrCSmEqL8c6dy5Fd96JZd5crI2NWOrrMXq907qZYzoZN9C11q8rpRrOscj/AH6jtW4dXr4nO6UJIaYTnUyS8vtJ+XyZo+vhR9rvz3TNC4Uyr30+koM+Eq2tJ5tFTCYsDfXYli6h4OY/w1TixVRchLHEi7m6CnNZGcpiye0O5oFstKHPB8xKqVcBN/AdrfVYR/P3AfcB1NXVZWHTQoiLoZPJTJt0Tw+J7m5SfX1gMGKw21AmE7HDLUR27yLa3Eyqt2/M9SirFYPTidHtxlhUhLmiAseKFdiWLMa2aBHWefMksKdANgLdBFwGXAvYgU1Kqc1a6wNnLqi1fgR4BGDVqlWjd9AUQlw0rTXJ7m6i+/YR3bOHaPNe4kePogyGTJc6yAR5X9/ZzR+nUgrL3Dm4rliPuaYm05ZdWIippBhjSQmmkhKMHs/IOkVuZSPQ28icCA0BIaXU68AlwFmBLoS4eDqdJtnZmek/3dZG4ngbyb6+4Qta4iQHB4kdPER6aCjzAaWwNDRgnTcXUJkLZXQa68KFmCvKMZWVYyovw1xejrHEC2h0NEo6FsNcVZ1XvUDyXTYC/XfAd5VSJsACrAX+IwvrFUIA6VCI0KZNBF59leiu3cSPHTutFwgmEyavF2W1YLBYMbjdFNx4I9b5TdgWLMC6cJGE8iwxkW6LTwFXAV6lVBvwIJnuiWitH9Za71VK/RHYCaSBR7XWuyevZCHyi04kSPb1EWtpyVwMc/Agqb5+0uEwqVCQ+MFD6EQCg8uFfeUKnOvWYWlsxNJQj6W2FlNFxazsoifONpFeLndNYJl/Af4lKxUJkWfS4TDR/fuJNjcT27ef5OAAaX+AlN9Psq+PVH//aWN+GEtKMJeXY3A4MJeW4Vy7DteGDTguWylt1eKc5EpRIS5SOholuns3kd27SbS2Em89TqK9nZTfTzoQQMfjI8saCwsxlZVhLCjAXFmJfdkyTGVlmMrKsNTXY53fhKm4OId7I2YyCXQhzoNOp0m0tRHZsZPIjh1EduwguncvJJMAGNxuLHV1WOfPzwzU5HJiLPBgnTcX25IlmMrL5SIZMWkk0IUYprVGx+OZ9mijkVR/P5Gdu4js3EFs/wHira0k2trQsRgAyuHAvmQJJffei33FpdiXL8dYUiKBLXJGAl3MSjqZJNHVRfzYMaLNzUTe3U7k3XdJnejqdyqjEeucOVjnNOLasAFLQz325cszF8uY5FdITB/y0yjymk6liB0+THTXbmKHDhE/ciTzaG8faSYBsDQ24rruWix19ZBOoVNpDC4n9mXLsC1ejMFuz+FeCDExEugib2itSbS1Ed21i8iu3Znn5mZ0OAxkLk+31NdjXbgQ9wc+gKWuFnNtHdameXIiUuQFCXQxY+l4nMjOnYS2bs2cnNy5i9TgIADKYsG6aCGFd9yBfdlSbMuWYWloQBnkvugif0mgi2lPa03i2DGie/dmLnXv6CB+9CiR93agI5GT441cfTX25cuwLVuGralJBoMSs44Euph2kv39RHbuJLJzJ9Gdu4js3n1yXBLA6PFgrqmh8MMfxrluLY5VqzAWFuawYiGmBwl0MS2kgiECz/8R369+TWT79sxEgwFrUxMFN9yAbfky7EuXYq6tk3FJhBiDBLqYMrGWI0R37SQVCqHDYVI+H4mOThJdXUT37UOHw1gaGyl94AEcqy7L9C5xOHJdthAzhgS6mFRaa0JvbmTgiScIvfHGafOU2YypshJzRQWeWz6E55ZbsK9YIRfmCHGBJNBFVmmtSRw/TnjrVkJbtxLespVkdzdGrxfvF+6n4MYbMRYUYHA4UHa7hLcQWSSBLi6KTqWIHTxI5L33CG97m/Dbb5Ps7gYyowY61qzGffXVuG+8EYP0OhFiUkmgi/Om02lCb7zB4M+eIrxtG+nhC3dMpaU4Vq/GsWY1jlWrsMydK0fgQkwhCXQxITqdJnbwEKG33sL3858TP3YMU2kpnttuxb5iBfZLL8VcUyMBLkQOSaCLMaWjUYKvvIL/938gvHXryMBVtkuWU3X/v1Jww/Vy8Y4Q04gEujhNOh4nvHkz/uefJ/D8n0gHg5jKynBddy2OVatxrF6NpaY612UKIUYhgS7QySTBN9/E/8x/E3ztNdKhEAanE/f11+O59RYca9bIPSuFmAEk0GexeFs7gz/7GUPPPEOqrw9jUREFN30Q9/XX41i3TnqlCJFFWmv8fVF6WwMUeG2U1RdkfRsS6LNQdN8++h/9If4//AGUwnXVBgpvvx3X+94nbeJCjCOVShMLJVEGMBoNaMDfF2GoJ4K/L0LYHycaTBANJUgl06RTmlQyzWBXmHgkMwb/8qtrJNDFxYns2UPfd79H8JVXMDgcFN9zD8Wf+iTm8vJclybElNJaEwsnhx8JYuEkkWAmiGPhJMl4mmQ8RSKeIh5OEoskiYYShHwxIsEE6LHXbbIasTvN2FxmTGYDyqAwW400rS6ntNZFaZ2b4qrJGY9IAj2Paa1JtLcT3bWLoWefI/jSSxgKCvB+4X6KP/5xjB5PrksUImuS8RT9HSH624LEIsnMPWLTGq0ZDmBN0Benvy3IQEeQeDQ15roMRoXJYsRkMWB1mLHaTTgLrZTVF+D0WLC5LGitSac0WmsKSux4yux4Su1YbLmL1XG3rJR6DLgZ6NFaLx1l/lXA74Ajw5N+o7X+/7NZpDg/8aNH6f/R4wReeIHUwABAJsjv/zzF99yD0e3OcYVCTIzWmpAvTjSUIB5NEg8nCfpiBAejhAYzR8vRUIJIMEGgL5IJ73OwOkwUVzlZsLaCglJ7JqwdJmxOEzaXBbvLjMVhwmicmTdCmch/JY8D3wWeOMcyb2itb85KReKC6HSa8NZtDD75UwIvvoQym3F/4AM4LlspN3wQ04LWmlQiTSqlMRgUygBoiIUzzRnxSJJ0SpNOa6KhBG37BznePECgP3rWupRB4fRYsLst2JwmCkpszF9djrfWhbfGhd1lAUVmO0qB4vT3eWrcQNdav66Uapj8UsSFiLe14/uv/2Lov/+bZGcnBo+Hkr+6j+JPfAKT15vr8sQspbVmoCPEkZ19HN3ZR397kGQ8fV7rsNiM1Cws5tLr6nAWWrDYTFhsJpyFFhweKwZD/gbzhcpWY8/lSqkdQAfwZa31ntEWUkrdB9wHUFdXl6VNz05aa4Z+8zRd3/wmOhbDeeV6yr78d7ivuUbuUC8mVTKRIuSLEwnECQ/FGegM0XPMT29rgPBQnLTWp500LKt3s+R91ZhtRswWIwajQqchnc4EvM1pxuowY7EbMRoNKKPCbDFSUu3EMEObPnIlG4H+LlCvtQ4qpW4Cfgs0jbag1voR4BGAVatWjdPaJcaSGhqi88GvE/jjH3GsXk3V//lnzNVy9aa4MOm0pu94gLZ9g3Qf8WNzmSmqcOApc5CMpwgPxQn6Yvi6wwx0hvD3Rc7q5VFY7qCqqRBXsW24WQNcRTbql5Xg9Fhzs2Oz0EUHutbaf8rr3yulHlJKebXWfRe7bnG6eGsrg0/+DN9vfkM6EqH0S1+i5C8/LVdxijGFhmIcbx6gtXmA7qN+EtFMl7xUIo0yqOGjZU0ykTlaLii1Ex9u0z6V0WTAU2antNbNgjXluEvsOAosOAosmZOLdukwNx1c9LeglKoAurXWWim1BjAA/RddmQAgHYsRfOUVfE8/Tej1N8BopOCGGyj5y09jW7w41+WJHPH3RxjsDJ+coMBoVBhMBqLBBO37B2nbP8hARwgAu9tMVVMhNqcZk8WI0WwArUmlMofaZfVuqucXjRxNR4MJfL1hzBYjzkIrVocpr08m5ouJdFt8CrgK8Cql2oAHATOA1vph4M+Bv1ZKJYEI8DGtx+s8JMaT7Ouj75FHGPrdM6SHhjCVl+P9689SeOfHMJeX5bo8MUXSqTSxcJJELEUilqLzkI8DW7vpPDx0zs+ZzAYq53lYsLaC2kXFeGtcqPM4iWhzmalwyXUKM81EerncNc7875Lp1iiyIB0OM/DjH9P/g0dJx2IU3Hgjnttvx3n5OmlayWOxSBJfd5ihnjBDvREGOkMMdITwdYdJp04/PiqqdLL21jlUzy/KdP2DzEnGVJp0UmO0GCivL8gchYtZRRq+ppHovn0c/9znSHZ04rruWsr+9u+wzmnMdVniIiXiKaLBTD/rWCQ5Ms5HNJigry1IzzE/Qz2Rkx9QUFBio7jSScOyEpyFNsxWI2arkcJyOyXVLmn+EKOSQJ8mAi+/QvuXv4yxoID6nzyBY/XqXJckLlAinqLzoI+2fZl27N7jgTHH/nAVWSmtc7NwXSXFVc7M5eNeOyaL/DUmzp8Eeo5prRl84gm6/8+3sC1eTM33H8JcJm3k01E0lGCgM0TIF8NoMmA0GzAoRSKead8O+WK07Rug4+AQqWQag0lR0ehh1U0NuAqtI32tbU5z5uEy53TcD5F/5Kcph3QiQdc/fRPfL36B+/rrqPrWtzA4HLkuSwzTWtN5yMfetzppbR4gPBQf9zNFlU6WbqimbnExlU2FmOVIW0whCfQcSQ0N0fbAA4Q3babkvvsofeCLKIOcxJpqiXiKQF+UsD9G2B8feUT8cToO+fD3RTHbjDQs8+KtcVFc5cRdbCOdzoxLkk5rzJZM+7bVacqMISJEjkig50Cis5PWv/g08bY2Kv/5nym8/bZclzRrhHwxuo/46Tzso+PQEH2tAdLp0xu4DUaFo8BCUYWDNTc3MmdFGWarHGmL6U8CfYol2ts59slPkfL5qH/sh3LycxLEo0nCQ3FCQzGGeiP4usP4usP0tgYIDsaAzJWPZQ1uVtxQR0m1C0eBBfvwlY9yEY2YqSTQp1C8rZ3WT36SVCBA3Y8ew75sWa5LyhshX4xdr7Wxd2MnYf/pbd0Go8JTaqdyrofyRg/ljQWU1rqln7bIOxLoUyR+7BjH7r2XdChM3WOPYV+6JNclzVg6rUcuvvF1h+k56ufIjj7SWtOwzEvlXA9OT2aI1QKvHXeJTYZaFbOCBPoUiDY30/qZ+yCdpv5Hj8kYLBfA3xfhwLZuug4P0dUyRCycHJnn9FhYdnUNy66qwVMqQweL2UsCfZKFtm6l7XN/g8Htpu6Hj2KdMyfXJU1rJ46+TwgMRtn9ajtHdvSidaZb4JwVpVQ0eiiudlJU4ZSR/oQYJr8Jkyj45kbaPvc5zLW11P3wUcwVFbkuadpKpdIc2NLN9j8dY7ArfNo8m9PMig/Us/T91biLbTmqUIjpTwJ9koTf3U7b/fdjmTOHuh89hqmoKNclTQuJeIqB9hB9bQH8fVFikSTxcILOliGCAzFKql1suGs+5uErKM0WI3VLiuVSeCEmQAJ9EkT37+f4Zz+LuayMukd/MOvDPJVI0/JeL7tfb6fzkG/kzuwGg8LiMGF1mCiqcLLhrgXULy2RLoNCXCAJ9CyLt7XT+um/xOBwUPfYD2ftjZq11nQf9XP4nR72b+kiEkhQ4LWx8sZ6SuvceGtcFJTYz2uMbiHEuUmgZ1n3P/0TOhym/le/nHX3+QwMROk87KPr0BBHdvURHIhhMCrql5aw5P3V1C0qlgAXYhJJoGdR8LXXCL76KmV///ezojdLKpWm46CPozv6OLqrD39fFACT1UjN/ELW3jKHhmVebE5zjisVYnaQQM+SdDxO1//+31gaGym++xO5LmdS+brD7H6jnf2buoiGEhjNBmoWFrH86loq53nw1rgwGOUqTCHGorWelHNFEuhZMvD4j0kca6X2Bz9AWfJvxL1YJEnL9h72b+6i/YAPg0HReGkp89eUU7uoWAavEnktrdME4gH8cT/+uJ9APEAoESKcCBNMBPHFfPiiPobiQ4QSISLJCNFkNPNIZZ5jqRixVIx4Ks69S+/liyu/mPU6JdCzINHdTd/DD+O69lpc77sy1+VkjdaajgM+dr/ezpEdfaSSaQpK7ay9dQ6LrqgcuUO8EDOJ1ppwMsxgdJC+SB/9kX76o5nHYHQQX8xHKBEiGA/ij/tHpqV06pzrdZvdeKwenGYndpMdu8lOka0Iu9GO1WTFajz5WFWxalL2TQI9C3q//R1IJCj/yj/kupSsiATiHN7ey+7X2uhvD2F1mlh8ZRXz15ZT3lAg3QrFtBBLxQjGg0SSESLJCL6Yj8HoIAPRAfqj/fRF+uiL9OGP+QklQpmQTgQJxANjhrPb4qbQWojL7MJtcVPnruOS0ksothVTaC3EY/XgtrhxW9w4zc6Rh8fiwWzM/bkiCfSLFDt4kKHf/Y7ie+7BUlub63IuiNaa/vYQrc39HN3ZR9fhIbQGb62Lq+9eyPzV5XJhj5hU8VSc/kj/SJOGP+7HF/UxGBtkKDZEMBEcOWruCffQHe7GF/ONuT6FoshWhNfuxWP1UOmsxGF24La4KbAUjAS31+7Fa/dSYi+hyFo0LUL5YkigX6Se//g2BoeDkr+6L9elnLfuo352vdLG8b0DI0POemtdrLqpgcZLSvHWyt3lxcQk00kiyQihRIhAPEAgHqAv0kdHsIOOUAdDsaGRZdM6PXJUHYgH6A53MxAdGHPdNqPttKPhcmc5l5ReQrmzHLfFPdK84bF6KLIWZY6mbYWYDTM7nC/EuIGulHoMuBno0VovPcdyq4HNwJ1a619lr8TpK/zuuwRffpnSBx6YUVeD9rcH2fJMC0d29GF1mKhbUkLtoiJqFxXjKpKxUmY7rTWBRICByACDsUwTxkB0gJ5wD12hLrpCXfhiPgLxwMiRczKdHHN9LrOLIlsRiszBgUEZRkLYa/eyxLuEckc5ZY4yPJaTTRpFtiI8Vg92k4ygOVETOUJ/HPgu8MRYCyiljMC3gOezU9b0p7Wm59/+HWOpl+J77s51OeeUTqXpPR7k2O5+Wvf0033Ej8VuYu0tjSy/plbuPJ/n4qn4yIm9E4+h2BD90X4GIpmwPjF9MDrIYGxw1IBWKErtpVQ4Kyh3lDOvcB4us2vkJKDNlDmSLrAU4LK4KLYVU+WqosBSkIO9np3G/U3WWr+ulGoYZ7H7gV8Ds+Z+asFXXyXyzjtUfP1BDA5Hrss5jU5r2g8McnR3Pz1H/fS2BkjG06CgvKGAtbfOYen7q+WCnxkuEA9wyHeIg4MH6Qp1kdIpUukU4WSYjlAHncFOusPdhBKhMddRYCmg2FZMka2IGlcNS71LKbIWUWTLNF2cmFdsK6bEXjIrmzFmkos+NFNKVQO3A9cwTqArpe4D7gOoq6u72E3njNaavocfxlxbS+GHP5zrckYM9YbZt6mLfZs7CQ7EMJoNlNa6WLy+ivI5BdQuKpa70k9z0WSU3nBvpkkjESAYDzIUH8ocOUcH6Qn30BnqpCPYQX+0f+RzRmXMPAxGbEYbFc4K5njmcEXVFRTZiii0FlJoLRxpxvBYPBTbimf8SUBxumz8rf1t4B+01qnxTqBprR8BHgFYtWqVPufC01h42zaiO3ZS8eDXUObc/kKkkmmO7OhjzxvttO0bRCmoXVzMFXfMo3G5V3qnTBNaa4ZiQ3SGOukKddEd7qY73E1PuIfecC+9kV56wj344/4x1+E0O/HavVQ6K9lQu4Fady3zi+bTVNhEhbNCTmCLrAT6KuDnwz9MXuAmpVRSa/3bLKx7Wup/9FGMJSV4br89ZzVorTn0Tg+bnj5MoD+Kq9jKmg81suiKSjmxOYUS6QQDkUy/5/5I/0gf6IHIAH3RTD/ovnAfHaEOIsnIaZ81KRNeh5cyexl17jpWla+izFFGqaOUImsRLosLl9mV6b1hK8JqlAu5xLlddKBrrRtPvFZKPQ48m89hHt2/n9Drb1D6wBcx2HITnF0tQ7z5y4N0H/FTUu3ips8tp35pidwIeRKkdZquUBftwXb6I5mLVTpDnbQMtXBk6AjtwfZRP2c1Wkf6ODd4Gri86nIqnZVUuiqpdFZS7iinxF6CQcmYNyJ7JtJt8SngKsCrlGoDHgTMAFrrhye1ummo/9EfYnA4KLrrrinf9lBvmE1Pt3D43R4cHgvX3LOQBesqJcgvgtYaX8w30gzSEeygLdhGW6CN44HjtPpbiafjp33GarTSUNDA8tLl3DL3lpELU0psmUexvRiHySFNIGLKTaSXy4STS2v9qYuqZpqLt7Xj//3vKb7nHowez5RsM53WdB32cWBrN3vf6sRgVKy+uZEV19fJgFgTkNZpfDEfveFeWoZa2D+wn/2D+0f6UvuiPpL69C56dpOdalc1te5arqy+krqCOmpcNZTaS/HavRRYC+TIWkxL0gH5PAz+5AkwGCj+1CcnbRutzf30tQWJBhKE/fGRqziNZgML11Ww5pY5MijWKMKJMEeGjnBg8AAtQy0c9h2mZaiF7lD3aYFtMpiY65k7MkZHka2IElsJFc6KkUeJTW6DJ2YmCfQJSofD+H7zNAUf+ADm8vKsr1+nNW89fZj3XmgFwGgyYHebqZjrYd5lZdQvLZnVFwCFEiFafC10hjJ9q7tCXfSGe0d6i3QEO9BkOk5ZDBYaPY0s9y6nqqGKUkcppfZS6gvqmeOZI131RN6avQlxnoaefZZ0IEDR/8h+23kykeLFH+3l8Ls9LN1Qzbrb5mKxGWflUWIilaBlqIWDvoMc9h3m4OBBDvkOnXXy0Wa0Ue4sp9ReyiWll3DbvNtoKmxibuFcatw1mAzyoy1mH/mpnwCtNYNP/RzrggXYV6zI6rr9/RFe+GEzXS1DXHHHPC69vjbvg1xrjT/upyvURWuglVZ/60j79uGhwyOXnZuUiQZPA8u9y7mj6Q7mFs6l2lVNhaMCj9WT9/9OQpwvCfQJiLz3HrG9e6n4+tezFiJaa/Zu7OTNXx0EDR/4zFLmXVaWlXVPJ2md5uDgQbZ2bWVr11YODR6iN9JLLBU7bTmv3cuC4gVcWX0lC4oX0FTYRH1BvTSPCHEeJNAnYPCppzA4nXg+dHNW1tdzzM+W37XQ2jxA9YJCrrl7EQXemT+iXDAeZO/AXvb07RlpMmkZahm5oKbWXcsy77KRi2fKneXUueuoc9fhsrhyXL0QM58E+jiSAwME/vBHCj/6UQxO5wWvR2vN8eYBtr/QStu+QSw2I++7cz7LNlSjZmA/8t5wLzv7dnJg4AAHfQc5MHiAY/5jI/PL7GXMLZzLh5s+zMLihaypWEOlqzKHFQuR/yTQx+H79a/RiQRFd33sgj4fDSXYt6mTPW904OsO4/RYuOKOeSx5XxUW+8z45x+IDrCvfx97B/bS3N/M7r7ddIQ6gMyQqnUFdTQVNnHznJtZ6l3K4pLFFNuKc1y1ELPPzEiUHPL/97PYV67EOm/eeX1Oa83Ol9vY9NvDpBJpKuZ4uO5Ti5h3WTlG8/S9KCWZTrKzdyebOjext38vewf20hPuGZlf7apmWekyPr7o4ywvXc78ovk4zNNr+GAhZisJ9HOIHz1K7MAByr/6lfP6XCQQ56Uf7+XY7n4alntZe8scvDXTs424L9LH/oH9HBg8wO6+3Wzq3EQgHsCgDMzxzGF1xWoWFS9iUfEiFhQvwGOdmitkhRDnTwL9HAIvvgiA+7rrJvyZ/vYgz/zne8RCSd7/sfks3VA97brXdYe6+ePRP/KHI39gT/+ekekVzgqurbuWK6uv5PKqy+VOM0LMMBLo5+B/4QVsS5Zgrq6e0PKpRJoXHtuD1vDnX7kMb417kiscn9aatkAbO/p2sL17O+/2vMsh3yEAFpcs5oGVD4w0ncjRtxAzmwT6GBJdXUR37KT0S1+a8Ge2PneE/vYQf/Y3y3Ma5h3BDl449gJvdbzFnv49I3dcd5qdXFp6KTc13sT19dfT4GnIWY1CiOyTQB9D4MWXAHBff/2Elu9qGWL788dYtL6ShmXeySxtVMf8x3i59WVePPYiO/t2AjCvcB7X1V3HEu8SlnmX0VTYhNEgIzQKka8k0McQeOEFLKvQdbgAABikSURBVPPmYp3TOO6yiViKFx9vxlVk48o/b5qC6jJ6wj38+uCvef7I8xweOgzAouJFPLDyAW5ouIFad+2U1SKEyD0J9FEkBwcJb9tGyV/dN+6yWmte/dk+hnoi3PqlFZPetzyajLK1aytPH3yaV46/QkqnWFOxho8s+AhX115NlatqUrcvhJi+JNBHEXz5ZUinKZhAc8s7fzjGgS3drL2lkZoFRZNSz0B0gJdbX+bV46+ypXML0VSUQmsh9yy+h4/M/wi1BXIkLoSQQB9V4E8vYK6uxrpo0TmXO/ROD1ueaWH+2nIu+2BDVmsIJ8IjXQu3dW0jpVNUu6q5o+kONtRsYFXFKixGS1a3KYSY2STQz5CORAht3kzhRz96zv7jPcf8vPh4MxVzPFz9iYVZ62veMtTCL/b9gmcOP0MwEaS+oJ6/WPoX3NBwAwuKFky7Pu1CiOlDAv0Moc2b0bEYrqs2jLlMPJrk+Uf3YHeZ+eBnl2EyX3zPkZahFr7/3vd5/ujzmAwmrq+/no8t/BiXll4qIS6EmBAJ9DMEX3sN5XDgWL16zGXe/K+D+Psi3P63K3AUXFyzR0ewg+9u/y7PHXkOq9HKp5d9mk8s+gQl9pKLWq8QYvaRQD+F1prgq6/hWn8FBsvoQX14ew973+pk5Y31VDVd+EnQQDzAD3b9gCebn0QpxT2L7+HepffKKIVCiAsmgX6K2IEDJLu6cH3+b0adH/LFeOWn+yitc7Pm5vH7p49Ga82zLc/yL9v+BV/Mx4fmfoj7V9xPhbPiYkoXQojxA10p9RhwM9CjtV46yvxbgX8E0kASeEBr/Wa2C50KwVdfA8D5/vefNU9rzatP7iMVT3P9XyzGaDr/IXDbg+3846Z/ZGPHRpaXLufhtQ+zuGTxRdcthBAwsSP0x4HvAk+MMf8l4BmttVZKLQf+C1iYnfKmVvDVVzODcZWdfW/Plu29HN3Vz/o/n0dRxfnduSit0/xi/y/4j3f+A4Xiq2u+yp0L7pTL8IUQWTVuoGutX1dKNZxjfvCUt05AX3xZUy85OEhkxw68f/3XZ82LR5K88YsDeGtdLL+65rzW2xZo42tvfY1tXdtYX7WeBy9/UG7FJoSYFFlpQ1dK3Q78M1AG/Nk5lrsPuA+grq4uG5vOmtAbb0A6PWp3xc3PtBDyx/ngXy/HYJxYU0tap/nl/l/yb+/8GwZl4OuXf507mu6QLohCiEmTlUDXWj8NPK2Uej+Z9vRR7wihtX4EeARg1apV0+pIPvjqaxi9XmxLlpw2vfuon12vtrHsqhrKGyZ2w4fjgeM8+NaDbOvaxuWVl/ONK74hR+VCiEmX1V4uw80zc5VSXq11XzbXPZl0KkVw40bc11yDMpw8Atda8+Z/HcRRYGHdLXMmtK7X217ny699WY7KhRBT7qIDXSk1Dzg8fFJ0JWAB+i+6sikU3bOH9NAQzivXnza9be8gXS1DbLhr/oRGUdzWtY2/ffVvmVs4l+9c/R3piiiEmFIT6bb4FHAV4FVKtQEPAmYArfXDwIeBe5RSCSAC3Km1nlbNKeMJbdwISuG84oqRaVprtj13BFeRlUVXjD8kbXN/M/e/fD9Vrioevu5himyTM/KiEEKMZSK9XO4aZ/63gG9lraIcCL65EdvixZiKToZw+/5BOg8P8f6PzcdoPveJ0CNDR/jsC5+lwFLAI9c/ImEuhMiJ8786Js+kgkEi772H88orT5u+7bmjOD0WFq0/98nMnnAPn33hsyileOT6R6SZRQiRM7M+0MObN0MqhXP9yeaW9v2DdBz0sfLG+nOOpOiP+/nsi5/FF/Px0HUPyU2XhRA5NevHcglu3IjB4cBx6aUj07b9/igOj4XFV47ddh5LxfjCy1/gyNARvnft91hSsmTMZYUQYirM+iP00Ma3cKxdixoeXbG3NUD7/kEuubZ2zKNzrTUPvvUg73S/wzfXf5Mrqq4YdTkhhJhKszrQ462tJFpbca4/2V1x+wutmG1GlryveszP/aT5JzzX8hyfv/Tz3DTnpqkoVQghxjWrAz20cSPASPu5vz/CoXd6WHJlFdYx+p1v6dzCv7/z71xbdy2fWf6ZKatVCCHGM6sDPbhxI+bqaiwNDQDsfKkNBSy/pnbU5TuCHfz9a39PfUE937zymxjUrP7nE0JMM7M2kXQySXjzFpxXXIFSimgowZ6NHcxbXYa72HbW8ql0iq+88RUS6QTfvvrbOM3nN4SuEEJMtlkb6NE9e0gHgzivuByAPW+0k4ylWHH96KNAPrn3Sbb3bOera79Ko+fC7lYkhBCTadYGemjzFgAca9eSTqXZ9Wo7NQuL8Na4z1r26NBR/nP7f7KhZgMfmvOhqS5VCCEmZBYH+iasCxZgKi6mtXmAkC/G0g1n92xJpVP8r43/C4vRwtcu/5qMnCiEmLZmZaCnYzEi727HuW4tAM1vdmB3m2lY5j1r2Sf3PsmO3h18dc1XKXOcfWs6IYSYLmZloEe2v4eOxXCsW0doKMbRXf0svLzyrBs/D8WGeHjnw6yvXs/Nc27OUbVCCDExszLQQ1s2g9GIY/Vq9m3qRKc1i9effZn/j3b/iGA8yJdWfkmaWoQQ096sDPTw5i3Yli7B4HTSvLGTqqZCCssdpy3TG+7lyb1P8sHGD7KgeEGOKhVCiImbdYGeCoaI7NqFc93ltB/w4e+NjDoI1//d+X9JppN8/tLP56BKIYQ4f7Mu0CPvvA3JJM51a2l+swOrw8TcFaWnLXM8cJxfH/g1dzTdQW3B6FeNCiHEdDPrAj20aTPKYsG0eDlH3uulaVU5Jsvpoyo+9N5DmAwm/uqSv8pRlUIIcf5mX6Bv2YJ9xQqOHwqRTKSZt+r0roiHfYd5ruU57lp4l3RTFELMKLMq0JMDA8T27sV5+ToOv9uD3W2mcl7hacs89N5D2E127l16b46qFEKICzOrAj28eTMA1tXrOLq7nzkryjAYTnZH3D+wnz8d+xMfX/RxudGzEGLGmVWBHtq0CYPbTY+qJBlLnXUy9HvvfQ+32c0nl3wyRxUKIcSFGzfQlVKPKaV6lFK7x5j/caXUzuHHW0qpS7Jf5sXTWhPa+BbOdWtpea8Pm9NM1fyTzS17+vbwyvFXuHvJ3XisnhxWKoQQF2YiR+iPAzeeY/4RYIPWejnwj8AjWagr6xKtrSQ6OrCtu4IjO/tovNSL0Xhy9x/a8RAeq4e7F92dwyqFEOLCjRvoWuvXgYFzzH9Laz04/HYzUJOl2rIqtGkTAIPly0lEU8xdebIHy4HBA7ze9jp3L7obl8WVqxKFEOKiZLsN/dPAH8aaqZS6Tyn1tlLq7d7e3ixv+txCG9/CVFXJsTaF1WGiZsHJk54/3vNj7CY7dy64c0prEkKIbMpaoCulriYT6P8w1jJa60e01qu01qtKS0vHWizrdCqV6X9++XqO7uyjYbl3ZGTFrlAXv2/5PbfPu51CW+E4axJCiOkrK4GulFoOPArcqrXuz8Y6sym6Zw9pv5+huVcQCyeZd9nJ5pYn9z5JmjR3L5a2cyHEzHbRga6UqgN+A9yttT5w8SVlX+itTPt5e7wCq8NE7aJiAALxAL888EtuqL+BGve0bPoXQogJM423gFLqKeAqwKuUagMeBMwAWuuHga8BJcBDw2OGJ7XWqyar4AsR2rQJ06KlHN3nZ95lZSPNLb868CtCiRCfWvqp3BYohBBZMG6ga63vGmf+XwJ/mbWKsiwdiRB5913Ct3+eRE+KpsvKAUikE/x0709ZU7GGJSVLclylEEJcvLy/UjT8zrvoRIJOWxN2t5nqBZkTny8ee5GecA/3LL4nxxUKIUR25H+gb95EyuqkrUsxd0UZhuGLiX7a/FPq3HW8r+Z9Oa5QCCGyI+8DPbRpM/4VN5GMnxwqd0fvDnb27eTjiz6OQeX9P4EQYpbI6zRL+XxEm5vpKbsMh8cyMlTuT5t/itvs5rZ5t+W4QiGEyJ68DvTQlq0kDVa6Qm7mrcwMldsV6uKFYy9wR9MdOMyO8VcihBAzRH4H+uZN9FetIpWCeasyvVt+vu/naDR3LTpn5x0hhJhx8jrQw5s20zfn/biKrFQ0FhBOhPnlgV9yde3VVLuqc12eEEJkVd4GeqKri/DxLnoNVcxdWYYyKH576Lf44365gYUQIi/lbaCHNm2m17uctFbMu6yMVDrFT5p/wvLS5VxaemmuyxNCiKzL20APb95Eb9UaXEVWyhsKePn4y7QF2/jUkk8xPESBEELklbwMdK01vs3bGfDMz9zIQsHjux+nxlXDNbXX5Lo8IYSYFHkZ6PEjR+iikjRG5l1Wxnu977Gzbyf3LLkHo8GY6/KEEGJS5GWgh7dupad0Ja4CE+WNBTy++3E8Vg+3zr0116UJIcSkyctA921+h4GSxcxdXcnxwHFeOf4KH53/UbmQSAiR1/Iu0LXWtO0fQisjcy4t5ad7f4rRYOSuhXIhkRAiv+VdoMcPH6bXUo/FlMZRk+l7flPjTZQ6pu4epkIIkQt5F+jBLVvpL15MzfwCnj78GyLJiNwvVAgxK+RdoHdu3k/c6qFuVRU/2/czVlesZmHxwlyXJYQQky6vAl1rTVtrAoAjBbvpCnXxiUWfyHFVQggxNfIq0OOHDtHnaKS4IMlTx35CrbuWDTUbcl2WEEJMibwK9IE3tzFU0EjBfNPIHYnkQiIhxGyRV4F+7O02UAZ2V+zGYrBw85ybc12SEEJMmbwJdJ1O09FrxKLiPBv4JdfUXYPH6sl1WUIIMWXGDXSl1GNKqR6l1O4x5i9USm1SSsWUUl/OfokTE2neS59rHs4iH764j1vnyWX+QojZZSJH6I8DN55j/gDwBeBfs1HQhTr27EYSFjf7ag5RZi/j8srLc1mOEEJMuXEDXWv9OpnQHmt+j9Z6G5DIZmHn69iOXtBp/mh9jg/N/ZCcDBVCzDpT2oaulLpPKfW2Uurt3t7erK03dugQPYZqrPZBIqagNLcIIWalKQ10rfUjWutVWutVpaXZG1ul+9mXCLjrOFjWzCWll9DoaczauoUQYqbIi14uRzYdBWBL4VvcNu+2nNYihBC5MuMDPdZyhO6kF2UIEvP4uanxplyXJIQQOWEabwGl1FPAVYBXKdUGPAiYAbTWDyulKoC3gQIgrZR6AFistfZPWtWn8P3xTwwULaSl8D1umXeL3MRCCDFrjRvoWutz3hlCa90F1GStovPU+tpuUkVN7C/dzXcWfjNXZQghRM7N6CaX+PHjdAYL0CSpWlDIHM+cXJckhBA5M6MDPbxlC/0lS+h1HOTOpR/JdTlCCJFTMzrQe9/ZR9hRQXdNmwyTK4SY9WZ0oLcdzFzAunLlQrkyVAgx683YQE/5/fijDrSOcdtq6aoohBAzNtDDO3YQcNWRdAxQ5szeVadCCDFTjdttcbpqe+tlgq7VuOfkdEwwIcRMoPXwI515cMprnT593onXZy1z5vJnzB91nWOsq6AairM/RMmMDfSOd/eSKnofly2fl+tSRD5Jp0GnMr906eFnnRp+rc+Ylzr5C5pOn1z2tM+mT//cWfPOeJxr3lnzT9Q01mfPmD/aemCcbZ5R+6jzx5t24v1Y6xkrOPXJ53OG6wTqQOfyp+ps6x+A67+R9dXOyEDX6TSpPgsUQePcqlyXk1snfmjTyeFHKvM8Mi11ct6Z03TqZICdmH4ivE5b5owQG/fzp4Zg6oyaUmfMG2XZE0F01rKjhOlp6zljXSNBO9q09OnrPDEvrykwGEEZzvFQoIyjvz7xWc58bTj5eqzPG4ygzJnnsz5vzCw72vTT1nnitWGMOk5Zzzn3h3Nsw3DKNsaoaWRfDWPXdGbdp/47KQN4JudazBkZ6IfefYWorRalUhRXZvlSf60hGYNkNPNIRE55H4NUbPj98LRUfHh64uS8VCIzPRU/+TqdGH6dGH6dPDktnTz5nE6cHsLp5PCyZwT2SEinsrv/2aaMmR9wg2n4teHktNOeT51uOPeyJssoy54SVCeWG3k92jTjKSF1xrIj2zaM/pmxtnHa9k/57IlfbMOpoWDk9BAxnP4Lr87cH3WOecOnwsZc74lQEflu5gV6IsLuF39C2LWOoqIkhkPPZ0I3FR8O3ygkwpnXI++jJwN65H3k5HMydsqyEbLy55nBDEYLGE2ZZ4P57NcGMxjNmWeLIxN6I/NMJ98bToSieXjaKSE5EpTGU+admG44OU0ZM9s68Qt+2ueMjBq8py5zamiO1DZaGJ8R0kKIKTPzAn3fcwS2bSFc9BEWRl6Gpx4ZY0EFZjuYbCefTTYwWTPvbYXgsoHZBib78LPt7M+Y7ZnPnPis8ZTXJutwaFtOvj7xLEdEQogpNuMC/VBBGSUDpQRK7ZRtuAlW33VKYA+HtWk4hCVUhRCzyIwL9NaAD2eyFoDSlaug2p3jioQQYnqYcY2cawYKCbhrMRiguMqZ63KEEGLamHFH6MaSEiJzV1NS7sRomnH/HwkhxKSZcYloW7wYv7mU0kZPrksRQohpZcYFeqA/SiycpLRW2s6FEOJUMy7Qe1sDAJTVS6ALIcSpZlyglzUUcPUnFlJS5cp1KUIIMa3MuJOi7mIbi6+c5eO3CCHEKGbcEboQQojRSaALIUSeGDfQlVKPKaV6lFK7x5ivlFL/qZQ6pJTaqZRamf0yhRBCjGciR+iPAzeeY/4Hgabhx33A9y++LCGEEOdr3JOiWuvXlVIN51jkVuAJrbUGNiulCpVSlVrrzizVeJpv/Pcemjv8k7FqIYSYEourCnjwQ0uyvt5stKFXA8dPed82PO0sSqn7lFJvK6Xe7u3tzcKmhRBCnJCNboujjVE76h0itNaPAI8ArFq16oLuIjEZ/6sJIUQ+yMYRehtQe8r7GqAjC+sVQghxHrIR6M8A9wz3dlkHDE1W+7kQQoixjdvkopR6CrgK8Cql2oAHATOA1vph4PfATcAhIAzcO1nFCiGEGNtEerncNc58DfxN1ioSQghxQeRKUSGEyBMS6EIIkSck0IUQIk9IoAshRJ5QmXOaOdiwUr3AsQv8uBfoy2I5M8Vs3O/ZuM8wO/d7Nu4znP9+12utS0ebkbNAvxhKqbe11qtyXcdUm437PRv3GWbnfs/GfYbs7rc0uQghRJ6QQBdCiDwxUwP9kVwXkCOzcb9n4z7D7Nzv2bjPkMX9npFt6EIIIc42U4/QhRBCnEECXQgh8sSMC3Sl1I1Kqf3DN6X+Sq7rmQxKqVql1CtKqb1KqT1KqS8OTy9WSr2glDo4/FyU61ong1LKqJTarpR6dvh9o1Jqy/B+/0IpZcl1jdk0fNvGXyml9g1/55fPhu9aKfWl4Z/v3Uqpp5RStnz8rpVSjymlepRSu0+ZNur3OzwM+X8O59tOpdTK89nWjAp0pZQR+B6ZG1MvBu5SSi3ObVWTIgn8ndZ6EbAO+Jvh/fwK8JLWugl4afh9PvoisPeU998C/mN4vweBT+ekqsnzHeCPWuuFwCVk9j2vv2ulVDXwBWCV1nopYAQ+Rn5+148DN54xbazv94NA0/DjPuD757OhGRXowBrgkNa6RWsdB35O5ibVeUVr3am1fnf4dYDML3g1mX398fBiPwZuy02Fk0cpVQP8GfDo8HsFXAP8aniRvNpvpVQB8H7ghwBa67jW2scs+K7JDN9tV0qZAAfQSR5+11rr14GBMyaP9f3eCjyhMzYDhUqpyolua6YF+oRvSJ0vlFINwApgC1B+4m5Qw89luats0nwb+P+A9PD7EsCntU4Ov8+373wO0Av8aLiZ6VGllJM8/6611u3AvwKtZIJ8CHiH/P6uTzXW93tRGTfTAn3CN6TOB0opF/Br4AGttT/X9Uw2pdTNQI/W+p1TJ4+yaD595yZgJfB9rfUKIESeNa+MZrjN+FagEagCnGSaG86UT9/1RFzUz/tMC/RZc0NqpZSZTJg/qbX+zfDk7hN/fg0/9+SqvkmyHrhFKXWUTHPaNWSO2AuH/yyH/PvO24A2rfWW4fe/IhPw+f5dXwcc0Vr3aq0TwG+AK8jv7/pUY32/F5VxMy3QtwFNw2fCLWROojyT45qybrjd+IfAXq31v58y6xngk8OvPwn8bqprm0xa669qrWu01g1kvtuXtdYfB14B/nx4sbzab611F3BcKbVgeNK1QDN5/l2TaWpZp5RyDP+8n9jvvP2uzzDW9/sMcM9wb5d1wNCJppkJ0VrPqAeZG1IfAA4D/zPX9UzSPl5J5s+sncB7w4+byLQnvwQcHH4uznWtk/hvcBXw7PDrOcBWMjci/yVgzXV9Wd7XS4G3h7/v3wJFs+G7Br4B7AN2Az8BrPn4XQNPkTlPkCBzBP7psb5fMk0u3xvOt11kegFNeFty6b8QQuSJmdbkIoQQYgwS6EIIkSck0IUQIk9IoAshRJ6QQBdCiDwhgS6EEHlCAl0IIfLE/wNJvI2DfHHVRgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for i in range(2):\n",
    "    figure()\n",
    "    plot(np.mean(trn_psnr_[i,:,:],(0,1)))\n",
    "    plot(np.mean(tst_psnr_[i,:,:],(0,1)))\n",
    "\n",
    "for i in range(2):\n",
    "    figure()\n",
    "    for j in range(5):\n",
    "        plot(np.mean(sr_[i,:,:,j,:],(0,1)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "file_name = 'no'\n",
    "np.save(file_path+file_name+'_losses',losses_)\n",
    "np.save(file_path+file_name+'_sr',sr_)\n",
    "np.save(file_path+file_name+'_trn',trn_psnr_)\n",
    "np.save(file_path+file_name+'_tst',tst_psnr_)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# basic encoding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "==>>> E: 0, F: 0, train psnr: 24.1675--- , test psnr: 24.5159--- , time: 17.5064 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 26.6947--- , test psnr: 25.7769--- , time: 34.5686 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 25.8497--- , test psnr: 25.3021--- , time: 53.3211 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 24.254--- , test psnr: 24.5632--- , time: 71.0978 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 25.3566--- , test psnr: 25.4436--- , time: 88.5656 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 25.5252--- , test psnr: 25.9611--- , time: 104.891 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 27.1904--- , test psnr: 26.4589--- , time: 123.232 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 25.0772--- , test psnr: 25.3485--- , time: 140.482 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 25.6546--- , test psnr: 25.3397--- , time: 158.203 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 25.4822--- , test psnr: 25.2689--- , time: 175.41 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 30.3502--- , test psnr: 28.1547--- , time: 16.9704 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 32.0695--- , test psnr: 29.0465--- , time: 35.9647 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 31.3678--- , test psnr: 28.5646--- , time: 54.1497 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 31.517--- , test psnr: 28.725--- , time: 70.1901 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 31.2757--- , test psnr: 28.7125--- , time: 87.9648 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 30.639--- , test psnr: 28.4006--- , time: 105.313 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 29.2097--- , test psnr: 27.6064--- , time: 122.289 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 30.5536--- , test psnr: 28.4551--- , time: 139.947 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 31.2483--- , test psnr: 28.6983--- , time: 157.237 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 30.897--- , test psnr: 28.4191--- , time: 174.345 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 25.3553--- , test psnr: 27.6421--- , time: 17.7598 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 25.0379--- , test psnr: 27.0304--- , time: 35.957 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 25.225--- , test psnr: 27.4254--- , time: 53.3665 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 25.0166--- , test psnr: 27.338--- , time: 70.8264 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 25.7982--- , test psnr: 27.9739--- , time: 87.5283 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 25.6158--- , test psnr: 27.8942--- , time: 105.375 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 25.5568--- , test psnr: 27.2987--- , time: 122.774 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 25.2173--- , test psnr: 27.7329--- , time: 140.712 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 25.4682--- , test psnr: 27.4969--- , time: 158.407 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 26.1124--- , test psnr: 27.5162--- , time: 176.407 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 30.2762--- , test psnr: 28.5445--- , time: 17.4842 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 28.9372--- , test psnr: 27.6787--- , time: 35.5689 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 30.3121--- , test psnr: 28.2593--- , time: 52.9138 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 31.0853--- , test psnr: 28.5249--- , time: 70.9707 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 30.6501--- , test psnr: 28.5862--- , time: 87.7353 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 31.0761--- , test psnr: 28.9659--- , time: 105.126 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 31.0747--- , test psnr: 28.8938--- , time: 122.884 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 30.8664--- , test psnr: 28.7891--- , time: 140.535 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 29.9816--- , test psnr: 28.2315--- , time: 158.283 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 29.9962--- , test psnr: 28.2604--- , time: 175.817 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 25.7043--- , test psnr: 25.0059--- , time: 17.6364 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 26.2121--- , test psnr: 25.3418--- , time: 35.0246 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 27.2416--- , test psnr: 26.0925--- , time: 52.2438 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 25.8409--- , test psnr: 25.0662--- , time: 68.5604 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 27.3983--- , test psnr: 26.0272--- , time: 86.0637 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 25.4804--- , test psnr: 24.8296--- , time: 102.906 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 25.4587--- , test psnr: 24.64--- , time: 121.125 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 26.9185--- , test psnr: 25.9751--- , time: 138.698 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 25.1206--- , test psnr: 24.472--- , time: 156.653 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 27.1409--- , test psnr: 26.0187--- , time: 172.801 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 31.7721--- , test psnr: 29.2034--- , time: 17.7401 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 32.0849--- , test psnr: 29.0297--- , time: 35.2843 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 31.9213--- , test psnr: 28.8138--- , time: 53.1483 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 31.502--- , test psnr: 29.0107--- , time: 71.1011 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 32.3099--- , test psnr: 29.0178--- , time: 89.2509 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 31.4934--- , test psnr: 28.684--- , time: 107.871 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 30.9479--- , test psnr: 28.6701--- , time: 125.718 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 32.1572--- , test psnr: 28.9481--- , time: 143.553 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 30.953--- , test psnr: 28.8285--- , time: 162.466 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 32.4112--- , test psnr: 28.8008--- , time: 179.943 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 31.6736--- , test psnr: 28.9956--- , time: 17.9278 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 32.0846--- , test psnr: 29.2014--- , time: 34.6826 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 31.9701--- , test psnr: 29.2919--- , time: 52.3251 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 32.4264--- , test psnr: 29.4042--- , time: 69.7675 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 32.3778--- , test psnr: 29.1456--- , time: 88.5425 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 32.3364--- , test psnr: 29.1813--- , time: 106.314 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 31.8441--- , test psnr: 29.0177--- , time: 123.632 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 32.2589--- , test psnr: 29.1577--- , time: 140.879 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 31.7765--- , test psnr: 29.1213--- , time: 158.965 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 32.3452--- , test psnr: 28.8913--- , time: 176.074 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 30.4694--- , test psnr: 26.4142--- , time: 18.7675 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 30.0813--- , test psnr: 26.5795--- , time: 35.4638 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 30.0996--- , test psnr: 26.4816--- , time: 52.3907 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 30.5778--- , test psnr: 26.5267--- , time: 70.6475 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 30.1397--- , test psnr: 26.333--- , time: 88.4782 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 28.9179--- , test psnr: 25.9328--- , time: 105.461 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 30.8341--- , test psnr: 26.7382--- , time: 122.465 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 30.0584--- , test psnr: 26.4954--- , time: 139.266 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 30.3473--- , test psnr: 26.4435--- , time: 156.84 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 30.9551--- , test psnr: 26.8506--- , time: 174.743 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 31.9301--- , test psnr: 29.2404--- , time: 17.4716 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 31.8436--- , test psnr: 28.8794--- , time: 34.6419 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 32.1055--- , test psnr: 29.1803--- , time: 53.0147 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 31.8673--- , test psnr: 29.099--- , time: 71.2414 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 31.8323--- , test psnr: 29.249--- , time: 87.5988 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 32.4253--- , test psnr: 28.8583--- , time: 105.768 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 31.7869--- , test psnr: 28.8172--- , time: 123.097 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 32.1001--- , test psnr: 29.063--- , time: 140.482 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 31.0807--- , test psnr: 28.8835--- , time: 157.671 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 32.0597--- , test psnr: 29.0773--- , time: 175.522 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 31.4975--- , test psnr: 29.0077--- , time: 18.166 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 31.9186--- , test psnr: 28.6808--- , time: 36.3399 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 31.4576--- , test psnr: 29.2938--- , time: 53.12 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 32.1688--- , test psnr: 29.1849--- , time: 70.3628 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 32.3327--- , test psnr: 29.3826--- , time: 88.34 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 31.6636--- , test psnr: 29.0319--- , time: 106.084 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 31.7147--- , test psnr: 28.9178--- , time: 123.474 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 31.1959--- , test psnr: 29.172--- , time: 140.328 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 31.6418--- , test psnr: 29.1146--- , time: 157.915 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 31.2493--- , test psnr: 28.7272--- , time: 174.366 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 1.87696--- , test psnr: -1.01417--- , time: 17.6433 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 1.90285--- , test psnr: -1.0081--- , time: 35.3735 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 1.94018--- , test psnr: -1.02847--- , time: 52.6917 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 2.09542--- , test psnr: -1.04983--- , time: 69.5195 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 2.0499--- , test psnr: -0.979258--- , time: 87.3692 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 2.12195--- , test psnr: -1.05768--- , time: 103.92 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 2.00546--- , test psnr: -0.974003--- , time: 121.124 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 1.92846--- , test psnr: -1.03523--- , time: 138.779 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 1.97837--- , test psnr: -0.971297--- , time: 156.187 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 1.8655--- , test psnr: -1.07122--- , time: 173.662 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 2.20513--- , test psnr: -0.921918--- , time: 17.6754 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 2.47601--- , test psnr: -0.978337--- , time: 36.0487 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 2.49691--- , test psnr: -0.994695--- , time: 54.2028 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 2.59116--- , test psnr: -0.972579--- , time: 71.9597 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 2.2809--- , test psnr: -0.923919--- , time: 89.3646 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 2.26009--- , test psnr: -0.938012--- , time: 107.2 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 2.26312--- , test psnr: -0.848553--- , time: 125.014 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 2.45292--- , test psnr: -0.890831--- , time: 142.439 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 2.55952--- , test psnr: -0.97388--- , time: 159.993 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 2.28171--- , test psnr: -0.898888--- , time: 177.753 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 1.69791--- , test psnr: 0.0385283--- , time: 17.4416 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 1.56016--- , test psnr: 0.111404--- , time: 34.661 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 1.7466--- , test psnr: 0.0733417--- , time: 51.6014 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 1.69841--- , test psnr: 0.046431--- , time: 69.428 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 1.79417--- , test psnr: 0.0869408--- , time: 86.7476 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 1.87491--- , test psnr: -0.0465546--- , time: 105.434 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 1.6852--- , test psnr: 0.0222041--- , time: 123.017 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 1.68172--- , test psnr: 0.0257676--- , time: 140.752 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 1.88351--- , test psnr: 0.0324061--- , time: 157.184 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 1.5178--- , test psnr: 0.0972425--- , time: 175.649 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 1.9602--- , test psnr: -1.27774--- , time: 18.6191 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 1.43455--- , test psnr: -1.08899--- , time: 36.3213 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 1.83414--- , test psnr: -1.2636--- , time: 55.289 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 1.86577--- , test psnr: -1.31187--- , time: 72.4138 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 1.72217--- , test psnr: -1.20452--- , time: 90.2128 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 1.78507--- , test psnr: -1.27633--- , time: 107.738 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 1.68565--- , test psnr: -1.27149--- , time: 125.451 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 2.0253--- , test psnr: -1.31776--- , time: 142.785 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 1.90102--- , test psnr: -1.27687--- , time: 160.871 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 1.87364--- , test psnr: -1.20314--- , time: 180.408 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 2.49771--- , test psnr: -1.33361--- , time: 17.9332 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 2.47844--- , test psnr: -1.3327--- , time: 36.1761 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 2.37187--- , test psnr: -1.26482--- , time: 54.9857 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 2.50822--- , test psnr: -1.32057--- , time: 71.1656 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 2.62936--- , test psnr: -1.34636--- , time: 88.6415 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 2.3228--- , test psnr: -1.31323--- , time: 106.023 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 2.49538--- , test psnr: -1.31114--- , time: 125.008 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 2.43818--- , test psnr: -1.28362--- , time: 142.695 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 2.50562--- , test psnr: -1.31176--- , time: 159.815 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 2.62319--- , test psnr: -1.36019--- , time: 177.783 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 2.2288--- , test psnr: -1.81829--- , time: 17.3817 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 2.09238--- , test psnr: -1.69357--- , time: 35.3114 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 2.06409--- , test psnr: -1.76589--- , time: 53.1675 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 2.15204--- , test psnr: -1.70336--- , time: 71.8907 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 2.20149--- , test psnr: -1.7679--- , time: 89.9147 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 2.04701--- , test psnr: -1.75461--- , time: 108.214 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 1.99744--- , test psnr: -1.6181--- , time: 126.725 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 2.10489--- , test psnr: -1.74958--- , time: 143.915 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 2.04623--- , test psnr: -1.70744--- , time: 161.604 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 2.04457--- , test psnr: -1.67938--- , time: 178.085 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 1.59599--- , test psnr: -0.804084--- , time: 17.9135 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 1.5499--- , test psnr: -0.807084--- , time: 35.6154 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 1.82205--- , test psnr: -0.89692--- , time: 53.767 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 1.51349--- , test psnr: -0.75043--- , time: 71.8463 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 1.46485--- , test psnr: -0.737134--- , time: 89.3666 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 1.77579--- , test psnr: -0.868419--- , time: 106.834 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 1.58436--- , test psnr: -0.806009--- , time: 125.31 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 1.7625--- , test psnr: -0.858517--- , time: 143.051 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 1.706--- , test psnr: -0.824099--- , time: 162.951 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 1.52541--- , test psnr: -0.731073--- , time: 180.443 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 3.3888--- , test psnr: -0.111045--- , time: 18.4585 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 3.52225--- , test psnr: -0.146864--- , time: 36.2804 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 3.50478--- , test psnr: -0.150632--- , time: 53.3886 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 3.41976--- , test psnr: -0.0696927--- , time: 70.7758 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 3.56079--- , test psnr: -0.141674--- , time: 88.0646 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 3.33917--- , test psnr: -0.0916798--- , time: 106.555 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 3.5454--- , test psnr: -0.118008--- , time: 124.903 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 3.42246--- , test psnr: -0.134626--- , time: 142.996 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 3.45463--- , test psnr: -0.168653--- , time: 159.903 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 3.5309--- , test psnr: -0.0755044--- , time: 177.275 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 1.56898--- , test psnr: -0.767839--- , time: 18.1486 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 1.56256--- , test psnr: -0.820973--- , time: 36.3547 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 1.69435--- , test psnr: -0.83885--- , time: 53.9412 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 1.54918--- , test psnr: -0.703893--- , time: 72.8999 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 1.58086--- , test psnr: -0.833832--- , time: 91.5591 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 1.66128--- , test psnr: -0.803285--- , time: 110.051 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 1.77602--- , test psnr: -0.88523--- , time: 128.206 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 1.79452--- , test psnr: -0.838586--- , time: 145.238 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 1.66191--- , test psnr: -0.800808--- , time: 164.654 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 1.76816--- , test psnr: -0.853381--- , time: 183.736 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 1.93432--- , test psnr: -1.27841--- , time: 18.6673 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 1.5713--- , test psnr: -1.2389--- , time: 36.0006 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 1.74573--- , test psnr: -1.33525--- , time: 53.4992 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 1.76244--- , test psnr: -1.28317--- , time: 70.8403 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 1.53438--- , test psnr: -1.21342--- , time: 88.0072 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 1.46976--- , test psnr: -1.18034--- , time: 106.337 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 1.50409--- , test psnr: -1.19384--- , time: 123.456 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 1.39096--- , test psnr: -1.19445--- , time: 141.176 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 1.5297--- , test psnr: -1.13324--- , time: 159.445 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 1.59301--- , test psnr: -1.20771--- , time: 177.216 seconds ---\n"
     ]
    }
   ],
   "source": [
    "sample_N = 512\n",
    "sample_d = 1\n",
    "\n",
    "all_data = torch.from_numpy(np.linspace(0, 1, sample_N+1)[:-1].reshape(-1,1)).type(torch.FloatTensor)\n",
    "\n",
    "train_data = basic_encoding(all_data[::2*sample_d])\n",
    "test_data = basic_encoding(all_data[sample_d::2*sample_d])\n",
    "\n",
    "\n",
    "# Calculate the input stable rank\n",
    "sr_[:,:,:,0] = srank_func(train_data)\n",
    "\n",
    "for i in range(2):\n",
    "    for j in range(N_test):\n",
    "        start_time = time.time()\n",
    "        \n",
    "        \n",
    "        # Prepare the data\n",
    "        all_target = signals[i][kk[j]].reshape(-1,1)\n",
    "        train_label = all_target[::2*sample_d]\n",
    "        test_label = all_target[sample_d::2*sample_d]\n",
    "        \n",
    "        xt, yt = train_data.cuda(),train_label.cuda()\n",
    "        xt, yt = Variable(xt), Variable(yt)\n",
    "\n",
    "        xts, yts = test_data.cuda(),test_label.cuda()\n",
    "        \n",
    "        \n",
    "        for k in range(N_repeat):\n",
    "            # Initialize classification model to learn\n",
    "            model = MLP(2).cuda()\n",
    "            # Set the optimization\n",
    "            optimizer = optim.Adam(model.parameters(), lr1, betas=(0.9, 0.999),weight_decay=1e-8)\n",
    "            \n",
    "\n",
    "            for epoch in range(epochs):\n",
    "                model.train()\n",
    "                optimizer.zero_grad()\n",
    "\n",
    "                out,e4,e3,e2,e1 = model(xt)\n",
    "                loss = criterion(out, yt)\n",
    "\n",
    "                loss.backward()\n",
    "                losses_[i,j,k,epoch] = loss.item()\n",
    "                optimizer.step()\n",
    "\n",
    "                model.eval()\n",
    "                trn_psnr_[i,j,k,epoch] = psnr_func(model(xt)[0],yt)\n",
    "                tst_psnr_[i,j,k,epoch] = psnr_func(model(xts)[0],yts)\n",
    "\n",
    "                if epoch%d == 0:\n",
    "                    es = [e1,e2,e3,e4]\n",
    "                    for l in range(4):\n",
    "                        sr_[i,j,k,l+1,epoch//d] = srank_func(es[l].detach().cpu())\n",
    "\n",
    "            print(\"==>>> E: %g, F: %g, train psnr: %g--- , test psnr: %g--- , time: %g seconds ---\" \n",
    "                    % (i,j, np.mean(trn_psnr_[i,j,k,:-100]),np.mean(tst_psnr_[i,j,k,:-100]),time.time() - start_time))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "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": [
    "for i in range(2):\n",
    "    figure()\n",
    "    plot(np.mean(trn_psnr_[i,:,:],(0,1)))\n",
    "    plot(np.mean(tst_psnr_[i,:,:],(0,1)))\n",
    "\n",
    "for i in range(2):\n",
    "    figure()\n",
    "    for j in range(5):\n",
    "        plot(np.mean(sr_[i,:,:,j,:],(0,1)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "file_name = 'basic'\n",
    "np.save(file_path+file_name+'_losses',losses_)\n",
    "np.save(file_path+file_name+'_sr',sr_)\n",
    "np.save(file_path+file_name+'_trn',trn_psnr_)\n",
    "np.save(file_path+file_name+'_tst',tst_psnr_)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# RFF"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "==>>> E: 0, F: 0, train psnr: 38.0612--- , test psnr: 28.4938--- , time: 17.8489 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 39.8602--- , test psnr: 28.5207--- , time: 35.6407 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 40.1071--- , test psnr: 28.8199--- , time: 53.5956 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 40.4706--- , test psnr: 28.7169--- , time: 71.3107 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 37.2158--- , test psnr: 28.3993--- , time: 89.1371 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 39.3646--- , test psnr: 28.7514--- , time: 107.49 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 40.0753--- , test psnr: 28.7448--- , time: 126.643 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 39.6806--- , test psnr: 28.5833--- , time: 146.643 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 38.422--- , test psnr: 28.3776--- , time: 165.256 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 38.9534--- , test psnr: 28.6588--- , time: 184.681 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 39.4488--- , test psnr: 29.7561--- , time: 19.3655 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 38.2862--- , test psnr: 29.4244--- , time: 37.6007 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 39.6361--- , test psnr: 29.7111--- , time: 55.1549 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 40.4161--- , test psnr: 29.8726--- , time: 74.149 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 39.4011--- , test psnr: 29.5932--- , time: 92.1977 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 39.3165--- , test psnr: 29.6999--- , time: 111.99 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 39.3524--- , test psnr: 29.5131--- , time: 130.191 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 39.999--- , test psnr: 29.6529--- , time: 149.731 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 41.3449--- , test psnr: 29.9838--- , time: 168.995 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 39.2802--- , test psnr: 29.8318--- , time: 185.737 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 37.0378--- , test psnr: 28.4016--- , time: 18.1715 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 37.487--- , test psnr: 28.9189--- , time: 36.4528 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 36.8251--- , test psnr: 28.4306--- , time: 55.5147 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 38.6516--- , test psnr: 29.0771--- , time: 74.1962 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 38.2042--- , test psnr: 28.7728--- , time: 92.0229 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 36.2955--- , test psnr: 27.977--- , time: 109.596 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 37.54--- , test psnr: 28.5462--- , time: 128.282 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 37.9494--- , test psnr: 28.9302--- , time: 147.717 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 37.5334--- , test psnr: 28.3925--- , time: 164.566 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 37.2555--- , test psnr: 28.1454--- , time: 183.038 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 37.2632--- , test psnr: 29.527--- , time: 20.0364 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 36.9484--- , test psnr: 29.5613--- , time: 38.8671 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 38.4499--- , test psnr: 29.6835--- , time: 56.1631 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 36.1306--- , test psnr: 29.3016--- , time: 74.2135 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 37.8249--- , test psnr: 29.5708--- , time: 91.9904 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 38.8979--- , test psnr: 29.7238--- , time: 109.992 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 38.0474--- , test psnr: 29.6609--- , time: 127.615 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 37.2767--- , test psnr: 29.4426--- , time: 144.672 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 39.1805--- , test psnr: 29.7005--- , time: 163.57 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 39.2066--- , test psnr: 29.7439--- , time: 182.479 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 40.2902--- , test psnr: 28.855--- , time: 18.0178 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 41.1058--- , test psnr: 28.9478--- , time: 36.4276 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 37.9764--- , test psnr: 28.3168--- , time: 55.3601 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 39.9439--- , test psnr: 28.6475--- , time: 73.7884 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 38.3748--- , test psnr: 28.5226--- , time: 91.9434 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 38.6972--- , test psnr: 28.6277--- , time: 111.741 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 38.4872--- , test psnr: 28.4492--- , time: 129.995 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 38.5659--- , test psnr: 28.7911--- , time: 148.274 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 40.5263--- , test psnr: 29.0739--- , time: 166.56 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 39.4958--- , test psnr: 28.6741--- , time: 184.469 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 40.1642--- , test psnr: 29.0573--- , time: 19.1013 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 39.6683--- , test psnr: 28.977--- , time: 37.3024 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 40.4294--- , test psnr: 29.027--- , time: 56.3317 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 40.5638--- , test psnr: 29.2526--- , time: 74.2601 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 40.758--- , test psnr: 29.4289--- , time: 93.741 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 42.0353--- , test psnr: 29.3553--- , time: 113.5 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 40.3021--- , test psnr: 29.2826--- , time: 131.706 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 41.1233--- , test psnr: 29.3035--- , time: 150.011 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 40.5168--- , test psnr: 29.0582--- , time: 168.547 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 39.1659--- , test psnr: 29.194--- , time: 187.088 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 39.9611--- , test psnr: 28.9338--- , time: 19.0015 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 42.1569--- , test psnr: 29.4543--- , time: 37.7091 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 39.4268--- , test psnr: 29.0004--- , time: 57.4894 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 41.3127--- , test psnr: 29.422--- , time: 77.1098 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 43.1286--- , test psnr: 29.9294--- , time: 96.9626 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 40.1061--- , test psnr: 29.1956--- , time: 116.603 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 41.2647--- , test psnr: 29.5112--- , time: 135.005 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 40.2927--- , test psnr: 29.3639--- , time: 153.694 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 41.677--- , test psnr: 29.4829--- , time: 172.824 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 40.4561--- , test psnr: 29.5373--- , time: 191.268 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 40.7213--- , test psnr: 29.3626--- , time: 17.8193 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 39.4376--- , test psnr: 29.2511--- , time: 36.1441 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 41.0638--- , test psnr: 29.2169--- , time: 55.3493 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 39.8511--- , test psnr: 29.196--- , time: 73.2313 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 38.0188--- , test psnr: 28.5796--- , time: 92.6964 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 40.6229--- , test psnr: 29.3995--- , time: 111.269 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 38.4216--- , test psnr: 28.8457--- , time: 129.641 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 39.9862--- , test psnr: 29.0591--- , time: 149.01 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 39.5388--- , test psnr: 29.1628--- , time: 166.121 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 40.9421--- , test psnr: 29.056--- , time: 184.466 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 40.2159--- , test psnr: 29.4044--- , time: 18.0568 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 40.3045--- , test psnr: 29.3164--- , time: 36.3143 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 40.6729--- , test psnr: 29.3682--- , time: 55.3071 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 39.0687--- , test psnr: 29.0977--- , time: 74.4994 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 40.0402--- , test psnr: 29.47--- , time: 92.0159 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 41.9593--- , test psnr: 29.5482--- , time: 110.992 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 41.2361--- , test psnr: 29.3576--- , time: 129.042 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 41.3485--- , test psnr: 29.4108--- , time: 146.54 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 42.4566--- , test psnr: 29.7918--- , time: 165.107 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 40.8381--- , test psnr: 29.3249--- , time: 184.08 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 40.8855--- , test psnr: 31.5434--- , time: 18.5059 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 38.9997--- , test psnr: 31.6062--- , time: 37.7946 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 38.5343--- , test psnr: 31.4255--- , time: 56.2721 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 40.0762--- , test psnr: 31.3203--- , time: 75.0251 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 39.5936--- , test psnr: 31.4294--- , time: 94.6508 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 38.3528--- , test psnr: 31.1833--- , time: 112.342 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 38.0565--- , test psnr: 31.1864--- , time: 130.617 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 39.1766--- , test psnr: 31.3787--- , time: 148.407 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 38.9356--- , test psnr: 31.3176--- , time: 167.341 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 39.4269--- , test psnr: 31.4714--- , time: 185.005 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 14.2202--- , test psnr: -1.33995--- , time: 19.085 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 13.5422--- , test psnr: -1.29275--- , time: 38.0722 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 14.4199--- , test psnr: -1.32485--- , time: 55.9693 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 14.8792--- , test psnr: -1.33684--- , time: 75.0873 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 14.4994--- , test psnr: -1.26187--- , time: 93.3699 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 13.9611--- , test psnr: -1.25938--- , time: 111.643 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 14.4522--- , test psnr: -1.32485--- , time: 129.073 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 15.1208--- , test psnr: -1.32262--- , time: 147.873 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 14.039--- , test psnr: -1.30207--- , time: 167.194 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 13.9688--- , test psnr: -1.28078--- , time: 186.819 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 14.8266--- , test psnr: -1.82921--- , time: 18.4114 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 15.2832--- , test psnr: -1.83673--- , time: 36.1164 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 14.6574--- , test psnr: -1.82387--- , time: 54.6477 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 14.8774--- , test psnr: -1.80163--- , time: 72.0301 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 15.1771--- , test psnr: -1.80202--- , time: 89.7619 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 14.0615--- , test psnr: -1.80197--- , time: 107.841 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 14.8244--- , test psnr: -1.80841--- , time: 125.73 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 14.5275--- , test psnr: -1.83482--- , time: 143.837 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 14.556--- , test psnr: -1.80392--- , time: 164.129 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 14.9065--- , test psnr: -1.74768--- , time: 182.228 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 14.5633--- , test psnr: -1.16045--- , time: 17.7942 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 15.0434--- , test psnr: -1.1539--- , time: 36.3981 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 16.1688--- , test psnr: -1.11658--- , time: 54.5549 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 14.9592--- , test psnr: -1.16506--- , time: 73.721 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 14.9534--- , test psnr: -1.16938--- , time: 92.8652 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 15.415--- , test psnr: -1.14959--- , time: 111.778 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 15.6423--- , test psnr: -1.19446--- , time: 130.497 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 15.8214--- , test psnr: -1.14376--- , time: 149.29 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 14.0161--- , test psnr: -1.21176--- , time: 168.652 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 14.3538--- , test psnr: -1.10951--- , time: 188.266 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 12.8516--- , test psnr: -2.09863--- , time: 19.9422 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 12.7162--- , test psnr: -2.03206--- , time: 37.997 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 13.1227--- , test psnr: -2.0674--- , time: 55.8869 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 12.9689--- , test psnr: -2.07408--- , time: 74.9752 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 13.2383--- , test psnr: -2.02809--- , time: 94.1443 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 13.0209--- , test psnr: -2.0564--- , time: 112.216 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 13.0968--- , test psnr: -2.0259--- , time: 131.108 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 13.5236--- , test psnr: -2.06932--- , time: 148.446 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 14.0211--- , test psnr: -2.0835--- , time: 167.951 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 12.3773--- , test psnr: -2.08822--- , time: 186.847 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 13.8576--- , test psnr: -2.14962--- , time: 19.4533 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 14.6051--- , test psnr: -2.14369--- , time: 39.2379 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 14.406--- , test psnr: -2.18616--- , time: 57.9961 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 14.2082--- , test psnr: -2.15722--- , time: 75.1455 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 14.3495--- , test psnr: -2.16881--- , time: 93.2055 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 12.7122--- , test psnr: -2.16377--- , time: 112.386 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 13.9935--- , test psnr: -2.20015--- , time: 130.239 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 13.6741--- , test psnr: -2.16888--- , time: 147.955 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 13.9588--- , test psnr: -2.1746--- , time: 166.081 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 14.0991--- , test psnr: -2.16296--- , time: 184.532 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 14.9582--- , test psnr: -2.60406--- , time: 19.4331 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 14.1519--- , test psnr: -2.57462--- , time: 37.1952 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 14.8842--- , test psnr: -2.58478--- , time: 56.0884 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 15.1973--- , test psnr: -2.56236--- , time: 75.617 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 14.597--- , test psnr: -2.57139--- , time: 95.0214 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 16.3523--- , test psnr: -2.55663--- , time: 114.042 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 13.7735--- , test psnr: -2.5762--- , time: 131.547 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 15.6088--- , test psnr: -2.61155--- , time: 149.805 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 14.0301--- , test psnr: -2.58379--- , time: 167.697 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 15.3782--- , test psnr: -2.5907--- , time: 185.917 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 14.4978--- , test psnr: -1.87492--- , time: 18.7205 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 13.413--- , test psnr: -1.83204--- , time: 36.4005 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 14.4868--- , test psnr: -1.79635--- , time: 53.6072 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 14.393--- , test psnr: -1.81967--- , time: 71.9507 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 13.345--- , test psnr: -1.91451--- , time: 90.5171 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 14.342--- , test psnr: -1.84415--- , time: 109.544 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 12.8184--- , test psnr: -1.8823--- , time: 127.691 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 14.3041--- , test psnr: -1.85457--- , time: 146.038 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 13.7062--- , test psnr: -1.86376--- , time: 164.611 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 14.8231--- , test psnr: -1.88342--- , time: 183.091 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 15.1687--- , test psnr: -0.742741--- , time: 19.0741 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 16.338--- , test psnr: -0.728704--- , time: 37.0994 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 15.0197--- , test psnr: -0.731125--- , time: 55.2346 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 16.8745--- , test psnr: -0.699745--- , time: 74.0774 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 15.0443--- , test psnr: -0.702893--- , time: 91.9482 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 15.6266--- , test psnr: -0.707276--- , time: 111.136 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 15.4768--- , test psnr: -0.749612--- , time: 128.751 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 15.6026--- , test psnr: -0.673396--- , time: 147.645 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 15.9048--- , test psnr: -0.691871--- , time: 165.439 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 15.3895--- , test psnr: -0.65323--- , time: 183.218 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 13.6522--- , test psnr: -1.85374--- , time: 19.1805 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 15.3361--- , test psnr: -1.90532--- , time: 37.2105 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 13.8203--- , test psnr: -1.81709--- , time: 55.9075 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 13.7429--- , test psnr: -1.90419--- , time: 75.3016 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 14.2673--- , test psnr: -1.84296--- , time: 94.7984 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 13.7178--- , test psnr: -1.8501--- , time: 113.933 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 13.9213--- , test psnr: -1.87174--- , time: 132.25 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 14.2435--- , test psnr: -1.79344--- , time: 151.802 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 13.7665--- , test psnr: -1.85259--- , time: 171.032 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 14.008--- , test psnr: -1.86504--- , time: 189.801 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 12.1811--- , test psnr: -2.01297--- , time: 18.6419 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 12.9784--- , test psnr: -2.02901--- , time: 36.5541 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 12.64--- , test psnr: -2.07741--- , time: 54.0982 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 12.7777--- , test psnr: -2.04838--- , time: 72.7764 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 13.3569--- , test psnr: -2.03956--- , time: 90.1991 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 12.5546--- , test psnr: -2.04623--- , time: 107.721 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 12.1187--- , test psnr: -2.04221--- , time: 126.552 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 12.7079--- , test psnr: -2.02614--- , time: 145.339 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 13.1735--- , test psnr: -2.03046--- , time: 163.521 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 13.4791--- , test psnr: -2.07681--- , time: 182.66 seconds ---\n"
     ]
    }
   ],
   "source": [
    "sample_N = 512\n",
    "sample_d = 1\n",
    "embedding_dimension = 256\n",
    "\n",
    "all_data = torch.from_numpy(np.linspace(0, 1, sample_N+1)[:-1].reshape(-1,1)).type(torch.FloatTensor)\n",
    "\n",
    "b = 5*torch.randn((int(embedding_dimension/2),1))\n",
    "train_data = rff_embedding(all_data[::2*sample_d],b)\n",
    "test_data = rff_embedding(all_data[sample_d::2*sample_d],b)\n",
    "\n",
    "# Calculate the input stable rank\n",
    "sr_[:,:,:,0] = srank_func(train_data)\n",
    "\n",
    "for i in range(2):\n",
    "    for j in range(N_test):\n",
    "        start_time = time.time()\n",
    "        \n",
    "        \n",
    "        # Prepare the data\n",
    "        all_target = signals[i][kk[j]].reshape(-1,1)\n",
    "        train_label = all_target[::2*sample_d]\n",
    "        test_label = all_target[sample_d::2*sample_d]\n",
    "        \n",
    "        xt, yt = train_data.cuda(),train_label.cuda()\n",
    "        xt, yt = Variable(xt), Variable(yt)\n",
    "\n",
    "        xts, yts = test_data.cuda(),test_label.cuda()\n",
    "        \n",
    "        \n",
    "        for k in range(N_repeat):\n",
    "            # Initialize classification model to learn\n",
    "            model = MLP(embedding_dimension).cuda()\n",
    "            # Set the optimization\n",
    "            optimizer = optim.Adam(model.parameters(), lr2, betas=(0.9, 0.999),weight_decay=1e-8)\n",
    "            \n",
    "\n",
    "            for epoch in range(epochs):\n",
    "                model.train()\n",
    "                optimizer.zero_grad()\n",
    "\n",
    "                out,e4,e3,e2,e1 = model(xt)\n",
    "                loss = criterion(out, yt)\n",
    "\n",
    "                loss.backward()\n",
    "                losses_[i,j,k,epoch] = loss.item()\n",
    "                optimizer.step()\n",
    "\n",
    "                model.eval()\n",
    "                trn_psnr_[i,j,k,epoch] = psnr_func(model(xt)[0],yt)\n",
    "                tst_psnr_[i,j,k,epoch] = psnr_func(model(xts)[0],yts)\n",
    "\n",
    "                if epoch%d == 0:\n",
    "                    es = [e1,e2,e3,e4]\n",
    "                    for l in range(4):\n",
    "                        sr_[i,j,k,l+1,epoch//d] = srank_func(es[l].detach().cpu())\n",
    "\n",
    "            print(\"==>>> E: %g, F: %g, train psnr: %g--- , test psnr: %g--- , time: %g seconds ---\" \n",
    "                    % (i,j, np.mean(trn_psnr_[i,j,k,:-100]),np.mean(tst_psnr_[i,j,k,:-100]),time.time() - start_time))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "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": [
    "for i in range(2):\n",
    "    figure()\n",
    "    plot(np.mean(trn_psnr_[i,:,:],(0,1)))\n",
    "    plot(np.mean(tst_psnr_[i,:,:],(0,1)))\n",
    "\n",
    "for i in range(2):\n",
    "    figure()\n",
    "    for j in range(5):\n",
    "        plot(np.mean(sr_[i,:,:,j,:],(0,1)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "file_name = 'rff'\n",
    "np.save(file_path+file_name+'_losses',losses_)\n",
    "np.save(file_path+file_name+'_sr',sr_)\n",
    "np.save(file_path+file_name+'_trn',trn_psnr_)\n",
    "np.save(file_path+file_name+'_tst',tst_psnr_)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Gaussian"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "==>>> E: 0, F: 0, train psnr: 37.501--- , test psnr: 28.1488--- , time: 18.8207 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 37.7388--- , test psnr: 28.2235--- , time: 36.3895 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 35.9945--- , test psnr: 27.9916--- , time: 55.1745 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 37.6136--- , test psnr: 27.8892--- , time: 74.1394 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 35.7194--- , test psnr: 27.9589--- , time: 92.2154 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 37.1377--- , test psnr: 28.2866--- , time: 110.814 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 36.1136--- , test psnr: 28.018--- , time: 129.798 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 37.5545--- , test psnr: 28.1602--- , time: 146.801 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 38.3655--- , test psnr: 28.3934--- , time: 165.845 seconds ---\n",
      "==>>> E: 0, F: 0, train psnr: 37.5162--- , test psnr: 28.2012--- , time: 185.16 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 37.7592--- , test psnr: 29.3452--- , time: 18.022 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 38.2565--- , test psnr: 29.4671--- , time: 36.4152 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 37.899--- , test psnr: 29.2549--- , time: 55.0389 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 38.3946--- , test psnr: 29.4373--- , time: 72.1721 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 37.4753--- , test psnr: 29.1023--- , time: 90.4481 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 37.8893--- , test psnr: 29.2332--- , time: 108.677 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 37.7467--- , test psnr: 29.1451--- , time: 127.84 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 35.8085--- , test psnr: 28.6757--- , time: 147.456 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 37.0762--- , test psnr: 29.1747--- , time: 165.221 seconds ---\n",
      "==>>> E: 0, F: 1, train psnr: 36.9652--- , test psnr: 28.9569--- , time: 183.777 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 37.5854--- , test psnr: 27.9501--- , time: 18.292 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 36.0942--- , test psnr: 27.6272--- , time: 36.4511 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 35.2419--- , test psnr: 27.5055--- , time: 54.0242 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 35.8057--- , test psnr: 27.7618--- , time: 74.005 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 34.6461--- , test psnr: 27.4102--- , time: 92.2556 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 35.4977--- , test psnr: 27.7151--- , time: 110.122 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 36.4068--- , test psnr: 27.6772--- , time: 128.783 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 37.3662--- , test psnr: 27.788--- , time: 146.035 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 36.4953--- , test psnr: 27.7985--- , time: 165.049 seconds ---\n",
      "==>>> E: 0, F: 2, train psnr: 36.2521--- , test psnr: 27.6575--- , time: 184.344 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 36.9619--- , test psnr: 29.128--- , time: 16.7933 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 37.3012--- , test psnr: 29.354--- , time: 34.7864 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 36.4694--- , test psnr: 29.4025--- , time: 52.3301 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 36.9617--- , test psnr: 29.151--- , time: 70.4806 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 36.3187--- , test psnr: 29.1424--- , time: 88.8082 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 35.9931--- , test psnr: 28.9198--- , time: 107.109 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 35.6178--- , test psnr: 29.1285--- , time: 124.061 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 36.0021--- , test psnr: 28.938--- , time: 142.805 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 37.1337--- , test psnr: 29.3496--- , time: 160.154 seconds ---\n",
      "==>>> E: 0, F: 3, train psnr: 35.673--- , test psnr: 28.9145--- , time: 178.778 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 35.846--- , test psnr: 28.0457--- , time: 18.5371 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 36.9103--- , test psnr: 28.1056--- , time: 38.355 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 37.1916--- , test psnr: 28.314--- , time: 57.2863 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 37.5259--- , test psnr: 28.3331--- , time: 75.3349 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 35.2043--- , test psnr: 27.9871--- , time: 93.9602 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 36.68--- , test psnr: 28.2768--- , time: 112.927 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 36.3135--- , test psnr: 28.1617--- , time: 131.444 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 37.1089--- , test psnr: 28.3334--- , time: 149.073 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 36.3078--- , test psnr: 28.2951--- , time: 167.454 seconds ---\n",
      "==>>> E: 0, F: 4, train psnr: 36.8742--- , test psnr: 28.4352--- , time: 185.553 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 37.6346--- , test psnr: 28.4652--- , time: 17.7056 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 38.6938--- , test psnr: 28.9368--- , time: 36.1131 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 38.3534--- , test psnr: 28.5417--- , time: 54.3864 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 39.1809--- , test psnr: 28.7715--- , time: 74.039 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 37.3123--- , test psnr: 28.5011--- , time: 92.9403 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 37.4242--- , test psnr: 28.6147--- , time: 111.41 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 38.4288--- , test psnr: 28.6076--- , time: 128.953 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 37.9749--- , test psnr: 28.6135--- , time: 147.174 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 38.7962--- , test psnr: 28.7567--- , time: 165.218 seconds ---\n",
      "==>>> E: 0, F: 5, train psnr: 37.9679--- , test psnr: 28.6728--- , time: 183.059 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 38.9696--- , test psnr: 28.8471--- , time: 19.1117 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 40.3436--- , test psnr: 28.9389--- , time: 37.6546 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 38.3854--- , test psnr: 28.7005--- , time: 55.393 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 37.7077--- , test psnr: 28.4694--- , time: 74.8737 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 38.4488--- , test psnr: 28.8178--- , time: 93.7676 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 39.1566--- , test psnr: 28.8187--- , time: 111.988 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 38.9862--- , test psnr: 28.6593--- , time: 129.335 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 38.4985--- , test psnr: 28.7564--- , time: 147.632 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 39.0798--- , test psnr: 28.7319--- , time: 165.511 seconds ---\n",
      "==>>> E: 0, F: 6, train psnr: 39.0428--- , test psnr: 28.6286--- , time: 184.086 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 36.7632--- , test psnr: 28.426--- , time: 18.2956 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 37.2339--- , test psnr: 28.3327--- , time: 36.6032 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 37.2962--- , test psnr: 28.7477--- , time: 55.2361 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 37.4828--- , test psnr: 28.6081--- , time: 72.0923 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 36.4737--- , test psnr: 28.4651--- , time: 90.5338 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 37.6242--- , test psnr: 28.6605--- , time: 108.013 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 35.9139--- , test psnr: 28.4465--- , time: 126.542 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 37.9467--- , test psnr: 28.9347--- , time: 144.797 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 36.7156--- , test psnr: 28.3871--- , time: 162.483 seconds ---\n",
      "==>>> E: 0, F: 7, train psnr: 38.4046--- , test psnr: 28.8624--- , time: 181.104 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 38.5208--- , test psnr: 28.6873--- , time: 19.489 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 36.3979--- , test psnr: 28.4325--- , time: 37.2098 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 38.9704--- , test psnr: 28.7817--- , time: 54.8507 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 38.5169--- , test psnr: 28.7347--- , time: 73.3544 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 39.0305--- , test psnr: 28.7513--- , time: 91.7553 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 39.4932--- , test psnr: 28.691--- , time: 109.14 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 38.1918--- , test psnr: 28.6847--- , time: 126.893 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 38.3601--- , test psnr: 28.5671--- , time: 145.369 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 38.5104--- , test psnr: 28.6245--- , time: 163.152 seconds ---\n",
      "==>>> E: 0, F: 8, train psnr: 38.0194--- , test psnr: 28.3861--- , time: 182.006 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 37.4924--- , test psnr: 30.951--- , time: 18.1445 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 35.6125--- , test psnr: 30.468--- , time: 35.576 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 36.5873--- , test psnr: 30.8716--- , time: 53.647 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 35.7583--- , test psnr: 30.7093--- , time: 71.8476 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 36.4481--- , test psnr: 30.8186--- , time: 90.0015 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 37.4859--- , test psnr: 30.7456--- , time: 108.186 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 36.4816--- , test psnr: 30.8166--- , time: 127.364 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 35.5266--- , test psnr: 30.5491--- , time: 145.901 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 37.402--- , test psnr: 30.7813--- , time: 163.523 seconds ---\n",
      "==>>> E: 0, F: 9, train psnr: 36.9875--- , test psnr: 30.7264--- , time: 182.312 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 11.1273--- , test psnr: -1.34312--- , time: 18.2866 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 10.9242--- , test psnr: -1.31682--- , time: 35.5762 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 11.5089--- , test psnr: -1.37017--- , time: 52.8169 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 11.5923--- , test psnr: -1.34141--- , time: 70.4542 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 11.316--- , test psnr: -1.3878--- , time: 89.8088 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 11.4424--- , test psnr: -1.35461--- , time: 108.667 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 11.6301--- , test psnr: -1.31624--- , time: 126.253 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 11.4887--- , test psnr: -1.29517--- , time: 144.138 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 11.2454--- , test psnr: -1.34623--- , time: 162.015 seconds ---\n",
      "==>>> E: 1, F: 0, train psnr: 12.2385--- , test psnr: -1.41415--- , time: 181.053 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 12.1786--- , test psnr: -1.87055--- , time: 17.5272 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 12.0679--- , test psnr: -1.85718--- , time: 35.3389 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 12.3745--- , test psnr: -1.81684--- , time: 53.9553 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 12.8574--- , test psnr: -1.86843--- , time: 72.8687 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 12.3034--- , test psnr: -1.8094--- , time: 91.8244 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 12.3639--- , test psnr: -1.8419--- , time: 109.375 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 12.6608--- , test psnr: -1.83004--- , time: 127.708 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 12.3156--- , test psnr: -1.78816--- , time: 145.496 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 12.6687--- , test psnr: -1.82972--- , time: 164.261 seconds ---\n",
      "==>>> E: 1, F: 1, train psnr: 11.9538--- , test psnr: -1.89292--- , time: 182.765 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 9.95273--- , test psnr: -1.09842--- , time: 18.8965 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 9.82717--- , test psnr: -1.14486--- , time: 38.0024 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 10.5492--- , test psnr: -1.22361--- , time: 56.4861 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 10.1678--- , test psnr: -1.12414--- , time: 74.8647 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 9.97768--- , test psnr: -1.17762--- , time: 94.3503 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 10.7068--- , test psnr: -1.19746--- , time: 111.849 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 10.3702--- , test psnr: -1.1571--- , time: 128.662 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 10.0242--- , test psnr: -1.18056--- , time: 144.342 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 10.4856--- , test psnr: -1.20228--- , time: 159.788 seconds ---\n",
      "==>>> E: 1, F: 2, train psnr: 9.45366--- , test psnr: -1.13511--- , time: 175.205 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 9.64236--- , test psnr: -2.03416--- , time: 15.2508 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 9.95296--- , test psnr: -1.9995--- , time: 31.3525 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 10.8326--- , test psnr: -2.06662--- , time: 46.6447 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 10.3539--- , test psnr: -1.99791--- , time: 62.0787 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 10.264--- , test psnr: -2.01114--- , time: 77.442 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 10.1258--- , test psnr: -2.02146--- , time: 92.8564 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 10.6764--- , test psnr: -2.01358--- , time: 108.171 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 9.60159--- , test psnr: -2.01129--- , time: 123.527 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 9.83143--- , test psnr: -1.9965--- , time: 138.397 seconds ---\n",
      "==>>> E: 1, F: 3, train psnr: 9.78637--- , test psnr: -2.02876--- , time: 153.857 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 11.0527--- , test psnr: -2.16637--- , time: 15.5681 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 11.8819--- , test psnr: -2.16345--- , time: 30.9386 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 11.35--- , test psnr: -2.13254--- , time: 46.3446 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 10.9138--- , test psnr: -2.16261--- , time: 61.6897 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 10.7755--- , test psnr: -2.20328--- , time: 74.7847 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 11.2776--- , test psnr: -2.12423--- , time: 88.1611 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 11.7181--- , test psnr: -2.17404--- , time: 103.508 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 11.4795--- , test psnr: -2.15526--- , time: 118.935 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 11.1047--- , test psnr: -2.11876--- , time: 134.285 seconds ---\n",
      "==>>> E: 1, F: 4, train psnr: 11.1594--- , test psnr: -2.11193--- , time: 149.436 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 11.3486--- , test psnr: -2.64422--- , time: 15.1599 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 10.2127--- , test psnr: -2.62463--- , time: 30.5749 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 11.1116--- , test psnr: -2.65804--- , time: 45.9919 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 10.8731--- , test psnr: -2.65558--- , time: 61.2505 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 10.3953--- , test psnr: -2.55219--- , time: 76.6967 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 10.2635--- , test psnr: -2.6095--- , time: 92.3045 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 9.71394--- , test psnr: -2.57148--- , time: 105.265 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 10.401--- , test psnr: -2.5757--- , time: 118.73 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 10.29--- , test psnr: -2.56639--- , time: 134.16 seconds ---\n",
      "==>>> E: 1, F: 5, train psnr: 10.9064--- , test psnr: -2.62684--- , time: 149.638 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 9.7559--- , test psnr: -1.87063--- , time: 14.6483 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 9.83789--- , test psnr: -1.83727--- , time: 29.4771 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 9.91016--- , test psnr: -1.86348--- , time: 44.7926 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 10.0682--- , test psnr: -1.94269--- , time: 60.528 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 9.73966--- , test psnr: -1.88592--- , time: 76.1008 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 10.2635--- , test psnr: -1.921--- , time: 91.5309 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 9.80062--- , test psnr: -1.8671--- , time: 107.561 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 10.0773--- , test psnr: -1.93292--- , time: 123.676 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 9.73369--- , test psnr: -1.88109--- , time: 139.039 seconds ---\n",
      "==>>> E: 1, F: 6, train psnr: 10.6042--- , test psnr: -1.91774--- , time: 154.64 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 11.5205--- , test psnr: -0.748032--- , time: 15.5975 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 11.4848--- , test psnr: -0.811027--- , time: 31.1706 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 11.7356--- , test psnr: -0.785959--- , time: 46.8247 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 12.6193--- , test psnr: -0.749868--- , time: 62.205 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 11.8872--- , test psnr: -0.787795--- , time: 77.7418 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 12.9331--- , test psnr: -0.713068--- , time: 93.5821 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 12.345--- , test psnr: -0.698329--- , time: 109.172 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 12.1082--- , test psnr: -0.762831--- , time: 124.564 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 11.4283--- , test psnr: -0.727676--- , time: 139.403 seconds ---\n",
      "==>>> E: 1, F: 7, train psnr: 12.2418--- , test psnr: -0.679502--- , time: 155.124 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 9.53068--- , test psnr: -1.82457--- , time: 15.7889 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 9.84635--- , test psnr: -1.81567--- , time: 31.771 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 9.57663--- , test psnr: -1.82782--- , time: 47.1536 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 10.7301--- , test psnr: -1.89818--- , time: 62.0598 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 9.92827--- , test psnr: -1.86696--- , time: 77.1469 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 10.1832--- , test psnr: -1.88361--- , time: 91.9942 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 10.2908--- , test psnr: -1.90726--- , time: 107.416 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 10.1396--- , test psnr: -1.90724--- , time: 122.773 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 9.57182--- , test psnr: -1.85271--- , time: 138.391 seconds ---\n",
      "==>>> E: 1, F: 8, train psnr: 10.8673--- , test psnr: -1.85696--- , time: 153.713 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 10.441--- , test psnr: -2.05353--- , time: 15.4372 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 10.9294--- , test psnr: -2.05037--- , time: 31.2382 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 11.1474--- , test psnr: -2.09058--- , time: 47.0754 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 10.3354--- , test psnr: -2.0417--- , time: 62.3397 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 10.2457--- , test psnr: -1.98366--- , time: 77.5972 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 11.0363--- , test psnr: -2.03369--- , time: 93.0089 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 10.1537--- , test psnr: -2.04273--- , time: 108.904 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 9.88183--- , test psnr: -1.96547--- , time: 124.439 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 10.3873--- , test psnr: -1.95524--- , time: 140.05 seconds ---\n",
      "==>>> E: 1, F: 9, train psnr: 10.8602--- , test psnr: -1.97413--- , time: 154.624 seconds ---\n"
     ]
    }
   ],
   "source": [
    "sample_N = 512\n",
    "sample_d = 1\n",
    "embedding_dimension = 256\n",
    "\n",
    "all_data = torch.from_numpy(np.linspace(0, 1, sample_N+1)[:-1].reshape(-1,1)).type(torch.FloatTensor)\n",
    "\n",
    "dic = torch.zeros((1,embedding_dimension))\n",
    "sig = 0.02*torch.ones((1,embedding_dimension))\n",
    "for i in range(embedding_dimension):\n",
    "    dic[0,i] = i/embedding_dimension\n",
    "\n",
    "train_data = rbf_embedding(all_data[::2*sample_d],dic,sig)\n",
    "test_data = rbf_embedding(all_data[sample_d::2*sample_d],dic,sig)\n",
    "\n",
    "# Calculate the input stable rank\n",
    "sr_[:,:,:,0] = srank_func(train_data)\n",
    "\n",
    "for i in range(2):\n",
    "    for j in range(N_test):\n",
    "        start_time = time.time()\n",
    "        \n",
    "        \n",
    "        # Prepare the data\n",
    "        all_target = signals[i][kk[j]].reshape(-1,1)\n",
    "        train_label = all_target[::2*sample_d]\n",
    "        test_label = all_target[sample_d::2*sample_d]\n",
    "        \n",
    "        xt, yt = train_data.cuda(),train_label.cuda()\n",
    "        xt, yt = Variable(xt), Variable(yt)\n",
    "\n",
    "        xts, yts = test_data.cuda(),test_label.cuda()\n",
    "        \n",
    "        \n",
    "        for k in range(N_repeat):\n",
    "            # Initialize classification model to learn\n",
    "            model = MLP(embedding_dimension).cuda()\n",
    "            # Set the optimization\n",
    "            optimizer = optim.Adam(model.parameters(), lr2, betas=(0.9, 0.999),weight_decay=1e-8)\n",
    "            \n",
    "\n",
    "            for epoch in range(epochs):\n",
    "                model.train()\n",
    "                optimizer.zero_grad()\n",
    "\n",
    "                out,e4,e3,e2,e1 = model(xt)\n",
    "                loss = criterion(out, yt)\n",
    "\n",
    "                loss.backward()\n",
    "                losses_[i,j,k,epoch] = loss.item()\n",
    "                optimizer.step()\n",
    "\n",
    "                model.eval()\n",
    "                trn_psnr_[i,j,k,epoch] = psnr_func(model(xt)[0],yt)\n",
    "                tst_psnr_[i,j,k,epoch] = psnr_func(model(xts)[0],yts)\n",
    "\n",
    "                if epoch%d == 0:\n",
    "                    es = [e1,e2,e3,e4]\n",
    "                    for l in range(4):\n",
    "                        sr_[i,j,k,l+1,epoch//d] = srank_func(es[l].detach().cpu())\n",
    "\n",
    "            print(\"==>>> E: %g, F: %g, train psnr: %g--- , test psnr: %g--- , time: %g seconds ---\" \n",
    "                    % (i,j, np.mean(trn_psnr_[i,j,k,:-100]),np.mean(tst_psnr_[i,j,k,:-100]),time.time() - start_time))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXAAAAD4CAYAAAD1jb0+AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjAsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+17YcXAAAVMklEQVR4nO3df4xd5X3n8ff33vnhH2Bsx4ONbcAQXAghbaCTBkKUrYBUNMlCV4p2E21b2iIhrVabNOqKTZQ/omq30q626o/VVqyskEJ3KdkNyW7SSG2DaKpoRUIydtwEMIEEUjA2eLzGxrHHnpl7v/vHvWPfuZ4fd+be8fCM3y/p6p7znOc853vmzHzm3DP3zonMRJJUnspyFyBJWhwDXJIKZYBLUqEMcEkqlAEuSYXqO58b27RpU+7YseN8blKSird79+7DmTnU3n5eA3zHjh2MjIycz01KUvEi4h9navcSiiQVygCXpEIZ4JJUKANckgplgEtSoeYN8Ij4QkQcioinZ1j2byMiI2LT0pQnSZpNJ2fgDwF3tjdGxOXAB4GXe1yTJKkD874PPDO/FRE7Zlj0x8D9wFd7XNM5fv+vnuHZA28u9WYkaclcv3Udn/un7+zpmIu6Bh4RdwGvZuY/dND3vogYiYiR0dHRxWxOkjSDBX8SMyLWAJ8FfqWT/pm5C9gFMDw8vKi7R/T6t5YkrQSLOQN/O3AV8A8R8VNgO7AnIrb0sjBJ0twWfAaemT8ELp2ab4b4cGYe7mFdkqR5dPI2wkeBbwPXRsT+iLh36cuSJM2nk3ehfHye5Tt6Vo0kqWN+ElOSCmWAS1KhDHBJKpQBLkmFMsAlqVAGuCQVygCXpEIZ4JJUKANckgplgEtSoQxwSSqUAS5JhTLAJalQBrgkFcoAl6RCGeCSVCgDXJIKZYBLUqEMcEkqlAEuSYUywCWpUAa4JBVq3gCPiC9ExKGIeLql7T9HxHMR8YOI+N8RsX5py5QktevkDPwh4M62tseBGzLz54Hngc/0uC5J0jzmDfDM/BZwpK3tG5k52Zz9DrB9CWqTJM2hF9fAfwf469kWRsR9ETESESOjo6M92JwkCboM8Ij4LDAJPDJbn8zclZnDmTk8NDTUzeYkSS36FrtiRNwDfAS4PTOzdyVJkjqxqACPiDuBfwf8k8w82duSJEmd6ORthI8C3waujYj9EXEv8F+Bi4HHI2JvRPy3Ja5TktRm3jPwzPz4DM0PLkEtkqQF8JOYklQoA1ySCmWAS1KhDHBJKpQBLkmFMsAlqVAGuCQVygCXpEIZ4JJUKANckgplgEtSoQxwSSqUAS5JhTLAJalQBrgkFcoAl6RCGeCSVCgDXJIKZYBLUqEMcEkqlAEuSYWaN8Aj4gsRcSginm5p2xgRj0fEC83nDUtbpiSpXSdn4A8Bd7a1fRp4IjN3Ak805yVJ59G8AZ6Z3wKOtDXfDTzcnH4Y+LUe1yVJmsdir4FvzsyDAM3nS2frGBH3RcRIRIyMjo4ucnOSpHZL/kfMzNyVmcOZOTw0NLTUm5OkC8ZiA/z1iLgMoPl8qHclSZI6sdgA/xpwT3P6HuCrvSlHktSpTt5G+CjwbeDaiNgfEfcC/xH4YES8AHywOS9JOo/65uuQmR+fZdHtPa5FkrQAfhJTkgplgEtSoQxwSSqUAS5JhTLAJalQBrgkFcoAl6RCGeCSVCgDXJIKZYBLUqEMcEkqlAEuSYUywCWpUAa4JBXKAJekQhngklQoA1ySCmWAS1KhDHBJKpQBLkmFMsAlqVBdBXhEfCoinomIpyPi0YhY1avCJElzW3SAR8Q24BPAcGbeAFSBj/WqMEnS3Lq9hNIHrI6IPmANcKD7kiRJnVh0gGfmq8AfAi8DB4FjmfmNXhUmSZpbN5dQNgB3A1cBW4G1EfHrM/S7LyJGImJkdHR08ZVKkqbp5hLKHcBLmTmamRPAV4D3tXfKzF2ZOZyZw0NDQ11sTpLUqpsAfxm4OSLWREQAtwP7elOWJGk+3VwDfwp4DNgD/LA51q4e1SVJmkdfNytn5ueAz/WoFknSAvhJTEkqlAEuSYUywCWpUAa4JBXKAJekQhngklQoA1ySCmWAS1KhDHBJKpQBLkmFMsAlqVAGuCQVygCXpEIZ4JJUKANckgplgEtSoQxwSSqUAS5JhTLAJalQBrgkFcoAl6RCGeCSVKiuAjwi1kfEYxHxXETsi4hbelWYJGlufV2u/6fA32TmRyNiAFjTg5okSR1YdIBHxDrgA8BvAWTmODDem7IkSfPp5hLK1cAo8OcR8f2I+HxErG3vFBH3RcRIRIyMjo52sTlJUqtuArwPuAl4IDNvBE4An27vlJm7MnM4M4eHhoa62JwkqVU3Ab4f2J+ZTzXnH6MR6JKk82DRAZ6ZrwGvRMS1zabbgWd7UpUkaV7dvgvl3wCPNN+B8iLw292XJEnqRFcBnpl7geEe1SJJWgA/iSlJhTLAJalQBrgkFcoAl6RCGeCSVCgDXJIKZYBLUqEMcEkqlAEuSYUywCWpUAa4JBXKAJekQhngklQoA1ySCmWAS1KhDHBJKpQBLkmFMsAlqVAGuCQVygCXpEIZ4JJUqK4DPCKqEfH9iPh6LwqSJHWmF2fgnwT29WAcSdICdBXgEbEd+DDw+d6UI0nqVLdn4H8C3A/UZ+sQEfdFxEhEjIyOjna5OUnSlEUHeER8BDiUmbvn6peZuzJzODOHh4aGFrs5SVKbbs7AbwXuioifAl8EbouI/9GTqiRJ81p0gGfmZzJze2buAD4G/F1m/nrPKpMkzcn3gUtSofp6MUhm/j3w970YS5LUGc/AJalQBrgkFcoAl6RCGeCSVCgDXJIKZYBLUqEMcEkqlAEuSYUywCWpUAa4JBXKAJekQhngklQoA1ySCmWAS1KhDHBJKpQBLkmFMsAlqVA9uSPPkqtNAglRAQIiILOxLJrzknSBKSPA//p+GHlw5mVRhYu3wLqtsG4bbNoJm34ONr4d1myE1eth8BKovIVebGQ2fwElZL35aJkmz86fmc6z02fWm2uaOdZd6DPTx2t9ntqfuaanPXPu9JmnDsc7M91eVwdjtK/bPs6s22lfZ5Y6Ou0zV00zjTHnODNtexE1LbQ9O+hzTnsX21uy9vbyZlvWxfYAhu+FS6+bfbuLUEaAX/shWHfZ2VDLbJ51B0yeguOvwfED8NoPYN9fQdbaBggYvLj5WAd9g1AdaDxaz95bQ/TMowb1WmN66vlMW/N5avpMnzw7P1tAS2pq+Rmc9mp6udrblk1rnm39WdZt7X/dhy/QAN95R+PRicnTcOQlOPIijL0Bp47C2FE4fRxOv9l4TI5Drfmo18+uGxWoVCH6Gmf2UWlpqzbO4qfaz7T1nW2vtKwz1Z9oaYuWcePsc3sfZlo2NR1na52zL3Os2zq/gOdp6zJ9zHmnZxnnzPRM6zHHeMA5PyjzjTfDD1/retPW5dzttK8/ax2d9pmlpjn7MHOfmba9qJoW2t7JPrSPo14pI8AXom+w8Vuux7/pJOmtZtEXhiPi8oj4ZkTsi4hnIuKTvSxMkjS3bs7AJ4Hfy8w9EXExsDsiHs/MZ3tUmyRpDos+A8/Mg5m5pzl9HNgHbOtVYZKkufXkvXURsQO4EXhqhmX3RcRIRIyMjo72YnOSJHoQ4BFxEfBl4Hcz88325Zm5KzOHM3N4aGio281Jkpq6CvCI6KcR3o9k5ld6U5IkqRPdvAslgAeBfZn5R70rSZLUiW7OwG8FfgO4LSL2Nh8f6lFdkqR5LPpthJn5f5n2sStJ0vn0FvoPT5KkhSjio/S1eo2xyTFOTp5kbHKM1068xp5De9jz+h5ePPoitayRJKuqq7hp80380oZb2HJwJ/ysn58dPc2p4xMMrO5j9cUDrFnXz8bLLmLT5Rdx8dtWEf6PBkmFKiLA/+CpP+BLz39pWlsQ7Nywk1u23kJ/tZ8geHPsZxz+TvLiT4MDtYPUYpLamtMMrK0yWFtFjlUZP3n2PxUOru1j8451bLn6ErZcdQmbr17HwKoiviSSVEaA33HFHVy57kpW961mdd9qNq7ayLuG3sW6gXVn+hzc+zJ/+8XnOHG6j40Do6yv/C2HLj7AMxcdZc/qQ7y+AYhgy8BWbqj8Ilec3sma45cx+nqdl589QuN+EcHQ5Rex9ec2sHXnei57+yWsWtu/fDsuSXOInOsfm/fY8PBwjoyM9HTM+okTvPTAX/LEvi1Ua6e47kd/yabTLxP9/dSOHTvTLzes4413bOWFHYN8d9tJvjvwKqfr4wAMTK5i28lruPrkDWx+82oufmOIqDf+PFDZOMHAljqrhyqsHepjzVA/azf0MzDQT1+lj2pUqVaqVKNKJSozPoKgEo3xgiAizjxPmdbe8rfh2S7xtK87m7mWLUYu8H+Zd9p/qb8PFzp+x3V30K+TbS/067pU2zgftS7F9nq1b52ONVuXudbdtHoTa/rXdFRDu4jYnZnD7e1FnIHPpH7yJEcfe4xXHvpffO/K36YyWOWuf/UuNt34GFGtAlA7dozxl17i1PPPM7Z7N/3f/R4bnzzIe4G+LVvgPT/P0eu28eqOtfxk3Rivjx3i+ZPf4PDxI1QPr+WSI1vYcvxq3vbCVtY+u543gMb/8JpkrO8QJwaOcbL/OGP9xznVf4LTfSc41XeS030nGa+eYrxvjPHqaSYrp5monmayMs5kZcL37kgXoAfueID3b3t/T8cs5gy8Pj5O7fBhJg8f5sSTT3Lk4b/g+Kk+fvieTzGxaj3/7P73sGn7RXOOkZlMvPIKJ578NieefJITTz1FvXmWXl2/nsF3XMfgNTsZvOYaBq68gsqWzZx+20WMxSTHf3aCIwdPcPzwacaOTjD2Ro3x4zUmTiSTJ2DyZJITHSZzXxLNx9Q0VZrPjUdUs/EeoWoSVaCSzcfZ6ahARhKt7dF4jmgsowJEnr1vQbT0i8ayqelonaelPy39YfpYU/PMsC6t457bfka0PM10f4DW8VsXVWJ6v9axevBLstNXL538IbyTsWbr08s/tC91rQvdVqdjdTROj/at12NNee9l7+XSNZcuaJ2Weso9A3/t3/8H3njkkTPzxy/azoFfvI8DlSuoDlS56xPvnje8oXFQBq64goErrmDDx/4FWa8z/uKLjO3dy8m9ezn9o+c5+uUvkydPtq5E9ZJLqK5fz5ZLLmHrunVUVq+msno1sWY1lYFBYvMgMTBAvTrABANMRD8T9DNR72OSKpP1KrV6MFGrUM8Kk/VgshbUskKtFtRqUKtDfSKYPAX1bNyZrVZv3DCo8ciW6blv6adznXNTmvbpZoeZfm7jnImpfrOsE+d0n/FVV+sYnfeffX72jc+8D/Nsbp7+M/zynEOnv4M66tdFTT198dvJfrdMT/7zhHf2soBCAvwnG97Hj39lmEn6mKxXmJyE/lVV3v2BbfzCbZezdv3gosaNSoXBa65h8JprWP/RjwKQ9ToTBw4ysX8/EwcOMHHgALUj/4/a0aONx7FjTBw8QJ4coz42Ro6PUx8fh4mJc8avAAPNR68lkFEho9p8bp+ukDRu05YRZDRu7za1bPo0Lf0hCZhqb54ST3s+p625ty3rNtpal0+tx9kx4MxP7Nk+kC2n+Nly666cFnhnfzRa+5zbr3XdljpmWfdse2v/6dOz1TF9vFbzjTFdzhhEbfXNtOKs6Tf1Ne4k1eZPpZnrm2vdtq/RbAN3kd7Txpzn69ALs34t2zbXWtfEC3V45+ae1QCFBPiGG6/jslVH6F/VR/+qKhdvXMV1N29hcE3v3yESlQoD27cxsH1h/9o863WYnCQnJhqPWo2crMHkRGNZvU7WGjc1zlrtzGl0Y1nzbuH1emN+6m7fU30yz/Y5c5f55h9lztyUfIY7jrecpmfLetO/22e5o3b7ujPu9Kx/yVlY/44HWMxYi+w/51DL9PLHV11FW/Oe3t/msYgAv/7WrVx/69blLmNOUanAwAAxsBTn25J0Lj9KL0mFMsAlqVAGuCQVygCXpEIZ4JJUKANckgplgEtSoQxwSSrUef1nVhExCvzjIlffBBzuYTmluBD3+0LcZ7gw9/tC3GdY+H5fmZlD7Y3nNcC7EREjM/03rpXuQtzvC3Gf4cLc7wtxn6F3++0lFEkqlAEuSYUqKcB3LXcBy+RC3O8LcZ/hwtzvC3GfoUf7Xcw1cEnSdCWdgUuSWhjgklSoIgI8Iu6MiB9FxI8j4tPLXc9SiIjLI+KbEbEvIp6JiE822zdGxOMR8ULzecNy19prEVGNiO9HxNeb81dFxFPNff6fEbHi7pIREesj4rGIeK55zG9Z6cc6Ij7V/N5+OiIejYhVK/FYR8QXIuJQRDzd0jbjsY2G/9LMth9ExE0L2dZbPsAjogr8GfCrwPXAxyPi+uWtaklMAr+Xme8Abgb+dXM/Pw08kZk7gSea8yvNJ4F9LfP/Cfjj5j6/Ady7LFUtrT8F/iYzrwN+gcb+r9hjHRHbgE8Aw5l5A1AFPsbKPNYPAXe2tc12bH8V2Nl83Ac8sJANveUDHPgl4MeZ+WJmjgNfBO5e5pp6LjMPZuae5vRxGj/Q22js68PNbg8Dv7Y8FS6NiNgOfBj4fHM+gNuAx5pdVuI+rwM+ADwIkJnjmXmUFX6sadzCcXVE9AFrgIOswGOdmd8CjrQ1z3Zs7wb+Ihu+A6yPiMs63VYJAb4NeKVlfn+zbcWKiB3AjcBTwObMPAiNkAcuXb7KlsSfAPcD9eb824CjmTnZnF+Jx/tqYBT48+alo89HxFpW8LHOzFeBPwRephHcx4DdrPxjPWW2Y9tVvpUQ4DFD24p972NEXAR8GfjdzHxzuetZShHxEeBQZu5ubZ6h60o73n3ATcADmXkjcIIVdLlkJs1rvncDVwFbgbU0Lh+0W2nHej5dfb+XEOD7gctb5rcDB5apliUVEf00wvuRzPxKs/n1qZdUzedDy1XfErgVuCsifkrj0thtNM7I1zdfZsPKPN77gf2Z+VRz/jEagb6Sj/UdwEuZOZqZE8BXgPex8o/1lNmObVf5VkKAfw/Y2fxr9QCNP3x8bZlr6rnmtd8HgX2Z+Ucti74G3NOcvgf46vmubalk5mcyc3tm7qBxXP8uM/8l8E3go81uK2qfATLzNeCViLi22XQ78Cwr+FjTuHRyc0SsaX6vT+3zij7WLWY7tl8DfrP5bpSbgWNTl1o6kplv+QfwIeB54CfAZ5e7niXax/fTeOn0A2Bv8/EhGteEnwBeaD5vXO5al2j/fxn4enP6auC7wI+BLwGDy13fEuzvu4GR5vH+P8CGlX6sgd8HngOeBv47MLgSjzXwKI3r/BM0zrDvne3Y0riE8mfNbPshjXfpdLwtP0ovSYUq4RKKJGkGBrgkFcoAl6RCGeCSVCgDXJIKZYBLUqEMcEkq1P8HePyB6IQc+AcAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for i in range(2):\n",
    "    figure()\n",
    "    plot(np.mean(trn_psnr_[i,:,:],(0,1)))\n",
    "    plot(np.mean(tst_psnr_[i,:,:],(0,1)))\n",
    "\n",
    "for i in range(2):\n",
    "    figure()\n",
    "    for j in range(5):\n",
    "        plot(np.mean(sr_[i,:,:,j,:],(0,1)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "file_name = 'rbf'\n",
    "np.save(file_path+file_name+'_losses',losses_)\n",
    "np.save(file_path+file_name+'_sr',sr_)\n",
    "np.save(file_path+file_name+'_trn',trn_psnr_)\n",
    "np.save(file_path+file_name+'_tst',tst_psnr_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}