Seg_net-flair_t1ne.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "import sklearn\n",
    "import pylab  \n",
    "import datetime\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib.patches as patches\n",
    "import cv2\n",
    "import nibabel as nib\n",
    "import math\n",
    "import random\n",
    "import time\n",
    "from scipy import ndimage\n",
    "from sklearn.metrics import f1_score\n",
    "from scipy.spatial.distance import directed_hausdorff\n",
    "import numpy as np\n",
    "import csv\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "os.environ[\"CUDA_VISIBLE_DEVICES\"] = \"0\"\n",
    "def npmin(dd):\n",
    "    return np.min(dd)\n",
    "def npmax(dd):\n",
    "    return np.max(dd)\n",
    "def npmean(dd):\n",
    "    return np.mean(dd)\n",
    "def norm (data):\n",
    "    return (data - npmin(data)) / (npmax(data) - npmin(data))*1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "isTrain = True\n",
    "checkpoint_dir = \"checkpoint\"\n",
    "epoch = 50\n",
    "batch_size = 1\n",
    "output_class_num = 4"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "def scale_3d(src, md='nearest'):\n",
    "    depth=128\n",
    "    height=128\n",
    "    width=128\n",
    "    out = ndimage.zoom(src,(depth/src.shape[0],height/src.shape[1],width/src.shape[2]),mode = md,order = 0)\n",
    "    return out"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "def value2onehot(label):\n",
    "#     label = label[...,np.newaxis]\n",
    "#     label = np.repeat(label, output_class_num+1, axis=3)\n",
    "    l0=[]\n",
    "    for i in range(label.shape[0]):\n",
    "        l1=[]\n",
    "        for j in range(label.shape[1]):\n",
    "            l2=[]\n",
    "            for k in range(label.shape[2]):\n",
    "                if label[i,j,k]==0:\n",
    "                    l2.append([1,0,0,0])\n",
    "                elif label[i,j,k]==0.25:\n",
    "                    l2.append([0,1,0,0])\n",
    "                elif label[i,j,k]==0.5:\n",
    "                    l2.append([0,0,1,0])\n",
    "                elif label[i,j,k]==1:\n",
    "                    l2.append([0,0,0,1])\n",
    "            l1.append(l2)\n",
    "        l0.append(l1)\n",
    "    l0 = np.asarray(l0, dtype=np.uint8)\n",
    "    #print(l0.shape)\n",
    "    return l0\n",
    "\n",
    "def onehot2value(label_batch):\n",
    "    l0_batch=[]\n",
    "    #for l in range(label_batch.shape[0]):\n",
    "    label = label_batch\n",
    "    l0=[]\n",
    "    for i in range(label.shape[0]):\n",
    "        l1=[]\n",
    "        for j in range(label.shape[1]):\n",
    "            l2=[]\n",
    "            for k in range(label.shape[2]):\n",
    "                if np.argmax(label[i,j,k])==0:\n",
    "                    l2.append(0)\n",
    "                elif np.argmax(label[i,j,k])==1:\n",
    "                    l2.append(0.25)\n",
    "                    #print('!!',label[i,j,k])\n",
    "                elif np.argmax(label[i,j,k])==2:\n",
    "                    l2.append(0.5)\n",
    "                    #print('!!',label[i,j,k])\n",
    "                elif np.argmax(label[i,j,k])==3:\n",
    "                    l2.append(1)\n",
    "                    #print('!!',label[i,j,k])\n",
    "            l1.append(l2)\n",
    "        l0.append(l1)\n",
    "\n",
    "    l0_batch = np.asarray(l0)\n",
    "    #print(l0_batch.shape)\n",
    "    return l0_batch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "def f1score(gt, seg):\n",
    "    if (np.sum(seg) + np.sum(gt)) == 0:\n",
    "        return 1.0\n",
    "    else:\n",
    "        return np.sum(np.multiply(gt,seg))*2.0 / (np.sum(seg) + np.sum(gt))\n",
    "def evaluation_return(y_true, y_pred):\n",
    "    dim = len(y_true.shape)\n",
    "    score=[]\n",
    "    hausdorff=[]\n",
    "    for k in range(y_true.shape[dim-1]):\n",
    "        f1 = f1score(y_true[...,k], y_pred[...,k])\n",
    "        score.append(f1)\n",
    "        sym_hausdorff = max(directed_hausdorff(y_true[...,k],y_pred[...,k])[0], directed_hausdorff(y_pred[...,k],y_true[...,k])[0])\n",
    "        hausdorff.append(sym_hausdorff)\n",
    "    score = np.asarray(score)\n",
    "    hausdorff = np.asarray(hausdorff)\n",
    "\n",
    "#     flat_true = np.asarray(flat_true)\n",
    "#     flat_pred = np.asarray(flat_pred)\n",
    "#     sym_hausdorff = max(directed_hausdorff(flat_true[1:],flat_pred[1:])[0], directed_hausdorff(flat_pred[1:],flat_true[1:])[0])\n",
    "    return np.mean(score[1:]), np.mean(hausdorff[1:])\n",
    "\n",
    "def evaluation(y_true, y_pred):\n",
    "    dim = len(y_true.shape)\n",
    "    score=[]\n",
    "    for k in range(y_true.shape[dim-1]):\n",
    "        f1 = f1score(y_true[...,k], y_pred[...,k])\n",
    "        score.append(f1)\n",
    "        print ('type',k ,'f1 score is',f1)\n",
    "    score = np.asarray(score)\n",
    "    print('whole f1 score is',np.mean(score[1:]))\n",
    "\n",
    "def mse(A,B):\n",
    "    mse = ((A - B)**2).mean(axis=None)\n",
    "    return mse\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/yunyun/anaconda3/envs/py36/lib/python3.6/site-packages/ipykernel_launcher.py:18: DeprecationWarning: get_data() is deprecated in favor of get_fdata(), which has a more predictable return type. To obtain get_data() behavior going forward, use numpy.asanyarray(img.dataobj).\n",
      "\n",
      "* deprecated from version: 3.0\n",
      "* Will raise <class 'nibabel.deprecator.ExpiredDeprecationError'> as of version: 5.0\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(210, 155, 240, 240)\n",
      "(210, 155, 240, 240)\n",
      "(210, 155, 240, 240)\n"
     ]
    }
   ],
   "source": [
    "\n",
    "filedir = \"../MICCAI_BraTS_2018_Data_Training/HGG/\"\n",
    "#filepath += \"ct_train_1001_image.nii\"\n",
    "\n",
    "train_data = os.listdir(filedir)\n",
    "train_array = []\n",
    "t1ce_train_array = []\n",
    "seg_array = []\n",
    "for d in train_data:\n",
    "    #print(d)\n",
    "    nii_dir = os.path.join(filedir, d)\n",
    "    nii_file = os.listdir(nii_dir)\n",
    "    for n in nii_file:\n",
    "        #print(n)\n",
    "        if n.split(\".\")[0].split(\"_\")[-1] == 't1ce' or n.split(\".\")[0].split(\"_\")[-1] == 'seg' or n.split(\".\")[0].split(\"_\")[-1] == 'flair':\n",
    "            filepath = os.path.join(nii_dir, n)\n",
    "            nii = nib.load(filepath)\n",
    "            #print(\"load\", filepath)\n",
    "            img_array = nii.get_data()\n",
    "            affine_array = nii.affine \n",
    "            img_head = nii.header\n",
    "            #print(img_array.shape)\n",
    "            img_array = img_array.transpose(0, 2, 1)\n",
    "            img_array = img_array.transpose(1, 0, 2)\n",
    "            nii_CT = img_array\n",
    "            nii_CT = np.asarray(nii_CT)\n",
    "            nii_CT = np.rot90(nii_CT,axes=(1,2))\n",
    "            #print(nii_CT.shape)\n",
    "            norm_nii_CT = nii_CT\n",
    "            norm_nii_CT = norm(norm_nii_CT)\n",
    "\n",
    "            #print(np.max(nii_CT),np.min(nii_CT))\n",
    "            #print(np.max(norm_nii_CT),np.min(norm_nii_CT))\n",
    "            if n.split(\".\")[0].split(\"_\")[-1] == 't1ce':\n",
    "                t1ce_train_array.append(norm_nii_CT)\n",
    "            elif n.split(\".\")[0].split(\"_\")[-1] == 'flair':\n",
    "                train_array.append(norm_nii_CT)\n",
    "            elif n.split(\".\")[0].split(\"_\")[-1] == 'seg':\n",
    "                seg_array.append(norm_nii_CT)\n",
    "t1ce_train_array = np.array(t1ce_train_array)\n",
    "print(t1ce_train_array.shape)\n",
    "train_array = np.array(train_array)\n",
    "print(train_array.shape)\n",
    "seg_array = np.array(seg_array)\n",
    "print(seg_array.shape)\n",
    "np.save('t1ce_train_array', t1ce_train_array)\n",
    "np.save('flair_train_array', train_array)\n",
    "np.save('seg_array', seg_array)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# back_check_label = onehot2value(onehot_nii_label)\n",
    "# print(np.max(back_check_label),np.min(back_check_label))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "input_height = 128 #onehot_nii_label.shape[1]\n",
    "input_width = 128 #onehot_nii_label.shape[2]\n",
    "input_depth = 128"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(210, 155, 240, 240)\n",
      "(210, 155, 240, 240)\n"
     ]
    }
   ],
   "source": [
    "flair_nii_CT = np.load('./flair_train_array.npy')\n",
    "t1ce_nii_CT = np.load('./t1ce_train_array.npy')\n",
    "print(flair_nii_CT.shape)\n",
    "print(t1ce_nii_CT.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(210, 155, 240, 240)\n"
     ]
    }
   ],
   "source": [
    "nii_label = np.load('./seg_array.npy')\n",
    "print(nii_label.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 0 15  6 36  9 33 12 13 14  5 20 11  2 32 26 34  8  1 27 18  3 22 24 23\n",
      " 16 28 25 10  4 39 29 19 30 35 38 17 37 31 21  7]\n"
     ]
    }
   ],
   "source": [
    "shuff= np.arange(40)\n",
    "np.random.shuffle(shuff)\n",
    "print(shuff)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(40, 155, 240, 240)\n",
      "(40, 155, 240, 240)\n",
      "(35, 128, 128, 128)\n",
      "(35, 128, 128, 128)\n",
      "(5, 128, 128, 128)\n",
      "(5, 128, 128, 128)\n"
     ]
    }
   ],
   "source": [
    "small_nii_CT = np.concatenate((flair_nii_CT[:20], t1ce_nii_CT[:20]),axis=0)\n",
    "small_nii_label = np.concatenate((nii_label[:20], nii_label[:20]), axis=0)\n",
    "print(small_nii_CT.shape)\n",
    "print(small_nii_label.shape)\n",
    "\n",
    "small_nii_CT = small_nii_CT[shuff]\n",
    "small_nii_label = small_nii_label[shuff]\n",
    "\n",
    "n1 = []\n",
    "n2 = []\n",
    "for i in range(small_nii_CT.shape[0]):\n",
    "    image1 = scale_3d(small_nii_CT[i])\n",
    "    image2 = scale_3d(small_nii_label[i])\n",
    "    n1.append(image1)\n",
    "    n2.append(image2)\n",
    "resize_nii_CT = np.asarray(n1[:35])\n",
    "resize_nii_label = np.asarray(n2[:35])\n",
    "valid_nii_CT = np.asarray(n1[35:])\n",
    "valid_nii_label = np.asarray(n2[35:])\n",
    "print(resize_nii_CT.shape)\n",
    "print(resize_nii_label.shape)\n",
    "print(valid_nii_CT.shape)\n",
    "print(valid_nii_label.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "# value to onehot\n",
    "onehot_array = []\n",
    "onehot_array2 = []\n",
    "for i in range(resize_nii_label.shape[0]):\n",
    "    onehot_nii_label = value2onehot(resize_nii_label[i])\n",
    "    onehot_array.append(onehot_nii_label)\n",
    "\n",
    "for i in range(valid_nii_label.shape[0]):\n",
    "    onehot_nii_label = value2onehot(valid_nii_label[i])\n",
    "    onehot_array2.append(onehot_nii_label)    \n",
    "\n",
    "train_onehot_array = np.asarray(onehot_array, dtype=np.uint8)\n",
    "valid_onehot_array = np.asarray(onehot_array2, dtype=np.uint8)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.0\n",
      "0.0\n"
     ]
    }
   ],
   "source": [
    "#check onehot to value\n",
    "value_array = []\n",
    "value_array2 = []\n",
    "for i in range(train_onehot_array.shape[0]):\n",
    "    value_nii_label = onehot2value(train_onehot_array[i])\n",
    "    value_array.append(value_nii_label)\n",
    "for i in range(valid_onehot_array.shape[0]):\n",
    "    value_nii_label = onehot2value(valid_onehot_array[i])\n",
    "    value_array2.append(value_nii_label)\n",
    "value_array = np.asarray(value_array)\n",
    "value_array2 = np.asarray(value_array2)\n",
    "print(np.sum(value_array-resize_nii_label))\n",
    "print(np.sum(value_array2-valid_nii_label))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(35, 128, 128, 128, 4)\n",
      "(5, 128, 128, 128, 4)\n"
     ]
    }
   ],
   "source": [
    "print(train_onehot_array.shape)\n",
    "print(valid_onehot_array.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Flair CT nii image: \n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.6802096985583224 0.0\n",
      "1.0 0.0\n"
     ]
    }
   ],
   "source": [
    "tt = 0\n",
    "sl_num = 50\n",
    "'''\n",
    "ROIrx = 180\n",
    "ROIry = 200\n",
    "ROIbx = 300\n",
    "ROIby = 110\n",
    "ROIs = 15\n",
    "'''\n",
    "print('Flair CT nii image: \\n')\n",
    "fig = plt.figure(figsize=(20, 20))\n",
    "fig.add_subplot(1,2,1)\n",
    "plt.imshow(resize_nii_CT[tt][sl_num],cmap=pylab.cm.gray,vmin = 0,vmax = 1)\n",
    "fig.add_subplot(1,2,2)\n",
    "#rect = patches.Rectangle((ROIrx,ROIry),ROIs,ROIs,linewidth=1,edgecolor='r',facecolor='none')\n",
    "#plt.gca().add_patch(rect)\n",
    "#rect = patches.Rectangle((ROIbx,ROIby),ROIs,ROIs,linewidth=1,edgecolor='b',facecolor='none')\n",
    "#plt.gca().add_patch(rect)\n",
    "plt.imshow(resize_nii_label[tt][sl_num],cmap=pylab.cm.gray,vmin = 0,vmax = 1)\n",
    "# fig.add_subplot(1,3,3)\n",
    "# plt.imshow(norm(back_check_label)[tt],cmap=pylab.cm.gray,vmin = 0,vmax = 1)\n",
    "plt.show()\n",
    "\n",
    "print(np.max(resize_nii_CT[tt][sl_num]),np.min(resize_nii_CT[tt][sl_num]))\n",
    "print(np.max(resize_nii_label[tt][sl_num]),np.min(resize_nii_label[tt][sl_num]))\n",
    "\n",
    "#print(\"red ROI\",np.mean(nii_label[tt][50,ROIry:ROIry+ROIs,ROIrx:ROIrx+ROIs]),\n",
    "#      np.std(nii_label[tt][50,ROIry:ROIry+ROIs,ROIrx:ROIrx+ROIs]))\n",
    "#print(\"blue ROI\",np.mean(nii_label[tt][50,ROIby:ROIby+ROIs,ROIbx:ROIbx+ROIs]),\n",
    "#      np.std(nii_label[tt][50,ROIby:ROIby+ROIs,ROIbx:ROIbx+ROIs]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "t1ne CT nii image: \n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.7955314827352742 0.0\n",
      "0.5 0.0\n"
     ]
    }
   ],
   "source": [
    "tt = 1\n",
    "sl_num = 50\n",
    "\n",
    "print('t1ne CT nii image: \\n')\n",
    "fig = plt.figure(figsize=(20, 20))\n",
    "fig.add_subplot(1,2,1)\n",
    "plt.imshow(resize_nii_CT[tt][sl_num],cmap=pylab.cm.gray,vmin = 0,vmax = 1)\n",
    "fig.add_subplot(1,2,2)\n",
    "#rect = patches.Rectangle((ROIrx,ROIry),ROIs,ROIs,linewidth=1,edgecolor='r',facecolor='none')\n",
    "#plt.gca().add_patch(rect)\n",
    "#rect = patches.Rectangle((ROIbx,ROIby),ROIs,ROIs,linewidth=1,edgecolor='b',facecolor='none')\n",
    "#plt.gca().add_patch(rect)\n",
    "plt.imshow(resize_nii_label[tt][sl_num],cmap=pylab.cm.gray,vmin = 0,vmax = 1)\n",
    "# fig.add_subplot(1,3,3)\n",
    "# plt.imshow(norm(back_check_label)[tt],cmap=pylab.cm.gray,vmin = 0,vmax = 1)\n",
    "plt.show()\n",
    "\n",
    "print(np.max(resize_nii_CT[tt][sl_num]),np.min(resize_nii_CT[tt][sl_num]))\n",
    "print(np.max(resize_nii_label[tt][sl_num]),np.min(resize_nii_label[tt][sl_num]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.layers import UpSampling3D, Conv3D, MaxPool3D, ReLU\n",
    "\n",
    "def Conv3d(input, ch, f, pad= 'same', act = 'leakyrelu'):\n",
    "    with tf.name_scope(\"Unet\"):\n",
    "        x = input\n",
    "        x = Conv3D(ch, 3, padding = pad, use_bias=True)(x)\n",
    "        if act == 'relu':\n",
    "            x = tf.nn.relu(x)\n",
    "        elif act == 'sigmoid':\n",
    "            x = tf.sigmoid(x)\n",
    "        elif act == 'leakyrelu':\n",
    "            x = tf.nn.leaky_relu(x, alpha = 0.01)\n",
    "        return x\n",
    "\n",
    "def MaxPool3d(input, ps):\n",
    "    x = input\n",
    "    x = MaxPool3D(pool_size=ps, strides=ps)(x)\n",
    "    return x\n",
    "\n",
    "def Upsample3d(input, scale):\n",
    "    x = input\n",
    "    x = UpSampling3D(size = (scale,scale,scale), data_format =\"channels_last\")(x)\n",
    "    return x\n",
    "\n",
    "def Deconv3d(input, ch, upscale, batch, xdim, ydim, zdim, pad= 'SAME'):\n",
    "    x = input\n",
    "    x = Conv3DTranspose(x, filter=tf.constant(np.ones([upscale,upscale,upscale,ch,ch], np.float32)),\n",
    "                               output_shape=[batch, xdim, ydim, zdim, ch],\n",
    "                                strides=[1, upscale, upscale, upscale, 1],\n",
    "                                padding= pad)\n",
    "#     x = tf.layers.conv3d_transpose(x, ch, f, strides = [2,2,2], padding = pad, kernel_initializer=tf.contrib.layers.xavier_initializer(), use_bias=True)\n",
    "    return x\n",
    "\n",
    "def Unet(placeholders, batch):\n",
    "    #decode\n",
    "   \n",
    "    x = placeholders['input']\n",
    "    print(\"Input\",x)\n",
    "    x = Conv3d(x, 64, 3)\n",
    "    x = Conv3d(x, 64, 3)\n",
    "    conv1 = x\n",
    "    x = MaxPool3d(x, 2)\n",
    "    print(\"Conv1\", conv1)\n",
    "    \n",
    "    x = Conv3d(x, 128, 3)\n",
    "    x = Conv3d(x, 128, 3)\n",
    "    conv2 = x\n",
    "    x = MaxPool3d(x, 2)\n",
    "    print(\"Conv2\", conv2)\n",
    "    \n",
    "    x = Conv3d(x, 256, 3)\n",
    "    x = Conv3d(x, 256, 3)\n",
    "    conv3 = x\n",
    "    x = MaxPool3d(x, 2)\n",
    "    print(\"Conv3\", conv3)\n",
    "    \n",
    "    x = Conv3d(x, 512, 3)\n",
    "    x = Conv3d(x, 512, 3)\n",
    "    conv4 = x\n",
    "    x = MaxPool3d(x, 2)\n",
    "    print(\"Conv4\", conv4)\n",
    "    \n",
    "    #encode\n",
    "    x = Conv3d(x, 512, 3)\n",
    "    x = Conv3d(x, 512, 3)\n",
    "    conv5 = x\n",
    "    print(\"Conv5\", conv5)\n",
    "#     x = Deconv3d(x, 512, 2, batch, input_depth//8, input_height//8, input_width//8)\n",
    "    x = Upsample3d(x,2)\n",
    "    print(\"Upsample5\", x)\n",
    "    x = tf.concat([x, conv4], axis = 4)\n",
    "    merge5 = x\n",
    "    print(\"Merge5\", merge5)\n",
    "    \n",
    "    x = Conv3d(x, 256, 3)\n",
    "    x = Conv3d(x, 256, 3)\n",
    "    conv6 = x\n",
    "    print(\"Conv6\", conv6)\n",
    "#     x = Deconv3d(x, 256, 2, batch, input_depth//4, input_height//4, input_width//4)\n",
    "    x = Upsample3d(x,2)\n",
    "    x = tf.concat([x, conv3], axis = 4)\n",
    "    merge6 = x\n",
    "    print(\"Merge6\", merge6)\n",
    "    \n",
    "    x = Conv3d(x, 128, 3)\n",
    "    x = Conv3d(x, 128, 3)\n",
    "    conv7 = x\n",
    "    print(\"Conv7\", conv7)\n",
    "#     x = Deconv3d(x, 128, 2, batch, input_depth//2, input_height//2, input_width//2)\n",
    "    x = Upsample3d(x,2)\n",
    "    x = tf.concat([x, conv2], axis = 4)\n",
    "    merge7 = x\n",
    "    print(\"Merge7\", merge7)\n",
    "    \n",
    "    x = Conv3d(x, 64, 3)\n",
    "    x = Conv3d(x, 64, 3)\n",
    "    conv8 = x\n",
    "    print(\"Conv8\", conv8)\n",
    "#     x = Deconv3d(x, 64, 2, batch, input_depth, input_height, input_width)\n",
    "    x = Upsample3d(x,2)\n",
    "    x = tf.concat([x, conv1], axis = 4)\n",
    "    merge8 = x\n",
    "    print(\"Merge8\", merge8)\n",
    "    \n",
    "    x = Conv3d(x, 32, 3)\n",
    "    x = Conv3d(x, 32, 3)\n",
    "    x = Conv3d(x, output_class_num, 1)\n",
    "    conv9 = x\n",
    "    print(\"Conv9\", conv9)\n",
    "    \n",
    "    return x\n",
    "    \n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input Tensor(\"Placeholder_4:0\", shape=(?, 128, 128, 128, 1), dtype=float32)\n",
      "Conv1 Tensor(\"Unet_39/LeakyRelu:0\", shape=(?, 128, 128, 128, 64), dtype=float32)\n",
      "Conv2 Tensor(\"Unet_41/LeakyRelu:0\", shape=(?, 64, 64, 64, 128), dtype=float32)\n",
      "Conv3 Tensor(\"Unet_43/LeakyRelu:0\", shape=(?, 32, 32, 32, 256), dtype=float32)\n",
      "Conv4 Tensor(\"Unet_45/LeakyRelu:0\", shape=(?, 16, 16, 16, 512), dtype=float32)\n",
      "Conv5 Tensor(\"Unet_47/LeakyRelu:0\", shape=(?, 8, 8, 8, 512), dtype=float32)\n",
      "Upsample5 Tensor(\"up_sampling3d_9/concat_2:0\", shape=(?, 16, 16, 16, 512), dtype=float32)\n",
      "Merge5 Tensor(\"concat_8:0\", shape=(?, 16, 16, 16, 1024), dtype=float32)\n",
      "Conv6 Tensor(\"Unet_49/LeakyRelu:0\", shape=(?, 16, 16, 16, 256), dtype=float32)\n",
      "Merge6 Tensor(\"concat_9:0\", shape=(?, 32, 32, 32, 512), dtype=float32)\n",
      "Conv7 Tensor(\"Unet_51/LeakyRelu:0\", shape=(?, 32, 32, 32, 128), dtype=float32)\n",
      "Merge7 Tensor(\"concat_10:0\", shape=(?, 64, 64, 64, 256), dtype=float32)\n",
      "Conv8 Tensor(\"Unet_53/LeakyRelu:0\", shape=(?, 64, 64, 64, 64), dtype=float32)\n",
      "Merge8 Tensor(\"concat_11:0\", shape=(?, 128, 128, 128, 128), dtype=float32)\n",
      "Conv9 Tensor(\"Unet_56/LeakyRelu:0\", shape=(?, 128, 128, 128, 4), dtype=float32)\n"
     ]
    }
   ],
   "source": [
    "placeholders = {\n",
    "    'input': tf.placeholder(tf.float32,shape=(None,input_depth,input_height,input_width,1)),\n",
    "    'label': tf.placeholder(tf.float32,shape=(None,input_depth,input_height,input_width,output_class_num))\n",
    "}\n",
    "\n",
    "\n",
    "#with tf.variable_scope('Unet_model') as scope:\n",
    "Unet_out_tensor = Unet(placeholders, batch_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "def focal_loss(pred, label):\n",
    "    '''\n",
    "    :param pred: B,N,num_class\n",
    "    :param label: B,N\n",
    "    :return: float\n",
    "    '''\n",
    "    label = tf.to_float(label)\n",
    "    num_class = 4\n",
    "    pred_soft = tf.nn.softmax(pred)\n",
    "    #label_onehot = tf.one_hot(label, depth=num_class, dtype=tf.float32)\n",
    "    probs = pred_soft * label  # (b, n, n_c)\n",
    "    probs = tf.reduce_sum(probs, axis=-1)  # (b, n)\n",
    "    focal_loss = -tf.pow((1-probs), 5.0) * tf.log(probs)\n",
    "    loss = tf.reduce_mean(focal_loss)\n",
    "    return loss\n",
    "\n",
    "cross_entropy_loss = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(labels=placeholders['label'],logits=Unet_out_tensor))\n",
    "\n",
    "#bce = tf.keras.losses.BinaryCrossentropy()\n",
    "#bce_loss = bce(placeholders['label'], Unet_out_tensor)\n",
    "f_loss = focal_loss(Unet_out_tensor, placeholders['label'])\n",
    "\n",
    "out_softmax = tf.nn.softmax(Unet_out_tensor)\n",
    "softmax_label = tf.nn.softmax(placeholders['label'])\n",
    "mse_loss = tf.losses.mean_squared_error(softmax_label,out_softmax)\n",
    "#print(mse_loss.shape)\n",
    "sub_value = softmax_label-out_softmax\n",
    "#print(sub_value.shape)\n",
    "a = sub_value[:,:,:,:,0]*0.05\n",
    "b = sub_value[:,:,:,:,1]*0.2\n",
    "c = sub_value[:,:,:,:,2]*0.15\n",
    "d = sub_value[:,:,:,:,3]*0.6\n",
    "weight_value = tf.stack([a, b, c, d], axis=-1)\n",
    "weighted_mse_loss = tf.reduce_mean(weight_value**2)\n",
    "#print(weighted_mse_loss.shape)\n",
    "inse = tf.reduce_sum(out_softmax * softmax_label)\n",
    "l = tf.reduce_sum(out_softmax)\n",
    "r = tf.reduce_sum(softmax_label)\n",
    "dice = (2. * inse + 1e-5) / (l + r + 1e-5)\n",
    "dice = tf.reduce_mean(dice)\n",
    "dice_loss = (1-dice)\n",
    "\n",
    "loss = weighted_mse_loss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "variables = tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES, scope='Unet_model')\n",
    "trainer = tf.train.AdamOptimizer(1e-4).minimize(loss)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Device mapping:\n",
      "/job:localhost/replica:0/task:0/device:XLA_CPU:0 -> device: XLA_CPU device\n",
      "/job:localhost/replica:0/task:0/device:GPU:0 -> device: 0, name: GeForce RTX 2080 Ti, pci bus id: 0000:01:00.0, compute capability: 7.5\n",
      "/job:localhost/replica:0/task:0/device:XLA_GPU:0 -> device: XLA_GPU device\n",
      "\n"
     ]
    }
   ],
   "source": [
    "\n",
    "\n",
    "saver = tf.train.Saver()\n",
    "sess = tf.Session(config=tf.ConfigProto(log_device_placement=True))\n",
    "sess.run(tf.global_variables_initializer())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(35, 128, 128, 128, 4)\n",
      "(35, 128, 128, 128)\n",
      "(35, 128, 128, 128, 4)\n"
     ]
    }
   ],
   "source": [
    "data_3d = resize_nii_CT\n",
    "label_3d = train_onehot_array \n",
    "print(label_3d.shape)\n",
    "\n",
    "'''\n",
    "data_3d=[]\n",
    "label_3d=[]\n",
    "for i in range(data.shape[0]//input_depth):\n",
    "    for j in range(data.shape[1]//input_height):\n",
    "        for k in range(data.shape[2]//input_width):\n",
    "            data_3d.append(data[i*input_depth:(i+1)*input_depth,j*input_height:(j+1)*input_height,k*input_width:(k+1)*input_width])\n",
    "            label_3d.append(label[i*input_depth:(i+1)*input_depth,j*input_height:(j+1)*input_height,k*input_width:(k+1)*input_width])\n",
    "\n",
    "data_3d = np.asarray(data_3d)\n",
    "label_3d = np.asarray(label_3d)\n",
    "'''\n",
    "print(data_3d.shape)\n",
    "print(label_3d.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch  0  Current time:  Sun Jun 21 22:01:54 2020\n",
      "Epoch 0 Iteration 0\n",
      "Loss: 0.0006728807\n",
      "Epoch 0 Iteration 1\n",
      "Loss: 0.0006428072\n",
      "Epoch 0 Iteration 2\n",
      "Loss: 0.0006364543\n",
      "Epoch 0 Iteration 3\n",
      "Loss: 0.0006378792\n",
      "Epoch 0 Iteration 4\n",
      "Loss: 0.0006264175\n",
      "Epoch 0 Iteration 5\n",
      "Loss: 0.0006065521\n",
      "Epoch 0 Iteration 6\n",
      "Loss: 0.00059579586\n",
      "Epoch 0 Iteration 7\n",
      "Loss: 0.00057430624\n",
      "Epoch 0 Iteration 8\n",
      "Loss: 0.00055925397\n",
      "Epoch 0 Iteration 9\n",
      "Loss: 0.0004915631\n",
      "Epoch 0 Iteration 10\n",
      "Loss: 0.0004256758\n",
      "Epoch 0 Iteration 11\n",
      "Loss: 0.0005673651\n",
      "Epoch 0 Iteration 12\n",
      "Loss: 0.00033551798\n",
      "Epoch 0 Iteration 13\n",
      "Loss: 0.0003080081\n",
      "Epoch 0 Iteration 14\n",
      "Loss: 0.00029859724\n",
      "Epoch 0 Iteration 15\n",
      "Loss: 0.00029331734\n",
      "Epoch 0 Iteration 16\n",
      "Loss: 0.00022054548\n",
      "Epoch 0 Iteration 17\n",
      "Loss: 0.00017539668\n",
      "Epoch 0 Iteration 18\n",
      "Loss: 0.0003349824\n",
      "Epoch 0 Iteration 19\n",
      "Loss: 0.00019130818\n",
      "Epoch 0 Iteration 20\n",
      "Loss: 0.00019338058\n",
      "Epoch 0 Iteration 21\n",
      "Loss: 0.00020841396\n",
      "Epoch 0 Iteration 22\n",
      "Loss: 0.00018717715\n",
      "Epoch 0 Iteration 23\n",
      "Loss: 0.00013430334\n",
      "Epoch 0 Iteration 24\n",
      "Loss: 0.00020161479\n",
      "Epoch 0 Iteration 25\n",
      "Loss: 8.875234e-05\n",
      "Epoch 0 Iteration 26\n",
      "Loss: 0.000108268396\n",
      "Epoch 0 Iteration 27\n",
      "Loss: 0.00011935602\n",
      "Epoch 0 Iteration 28\n",
      "Loss: 0.00011609413\n",
      "Epoch 0 Iteration 29\n",
      "Loss: 6.211955e-05\n",
      "Epoch 0 Iteration 30\n",
      "Loss: 7.556018e-05\n",
      "Epoch 0 Iteration 31\n",
      "Loss: 0.00011052514\n",
      "Epoch 0 Iteration 32\n",
      "Loss: 9.347874e-05\n",
      "Epoch 0 Iteration 33\n",
      "Loss: 5.405371e-05\n",
      "Epoch 0 Iteration 34\n",
      "Loss: 6.789499e-05\n",
      "Valid loss: 8.075881487457082e-05\n",
      "Epoch  1  Current time:  Sun Jun 21 22:02:50 2020\n",
      "Epoch 1 Iteration 0\n",
      "Loss: 7.208235e-05\n",
      "Epoch 1 Iteration 1\n",
      "Loss: 7.1338625e-05\n",
      "Epoch 1 Iteration 2\n",
      "Loss: 7.7975346e-05\n",
      "Epoch 1 Iteration 3\n",
      "Loss: 0.00011985435\n",
      "Epoch 1 Iteration 4\n",
      "Loss: 4.545095e-05\n",
      "Epoch 1 Iteration 5\n",
      "Loss: 7.964787e-05\n",
      "Epoch 1 Iteration 6\n",
      "Loss: 4.7486945e-05\n",
      "Epoch 1 Iteration 7\n",
      "Loss: 5.1578354e-05\n",
      "Epoch 1 Iteration 8\n",
      "Loss: 7.419633e-05\n",
      "Epoch 1 Iteration 9\n",
      "Loss: 6.205898e-05\n",
      "Epoch 1 Iteration 10\n",
      "Loss: 4.6024234e-05\n",
      "Epoch 1 Iteration 11\n",
      "Loss: 4.587141e-05\n",
      "Epoch 1 Iteration 12\n",
      "Loss: 3.8512087e-05\n",
      "Epoch 1 Iteration 13\n",
      "Loss: 3.4594355e-05\n",
      "Epoch 1 Iteration 14\n",
      "Loss: 7.120801e-05\n",
      "Epoch 1 Iteration 15\n",
      "Loss: 5.2230924e-05\n",
      "Epoch 1 Iteration 16\n",
      "Loss: 3.421359e-05\n",
      "Epoch 1 Iteration 17\n",
      "Loss: 3.2513573e-05\n",
      "Epoch 1 Iteration 18\n",
      "Loss: 2.656227e-05\n",
      "Epoch 1 Iteration 19\n",
      "Loss: 6.284317e-05\n",
      "Epoch 1 Iteration 20\n",
      "Loss: 2.5890668e-05\n",
      "Epoch 1 Iteration 21\n",
      "Loss: 3.3987686e-05\n",
      "Epoch 1 Iteration 22\n",
      "Loss: 5.803521e-05\n",
      "Epoch 1 Iteration 23\n",
      "Loss: 2.8366721e-05\n",
      "Epoch 1 Iteration 24\n",
      "Loss: 4.2188625e-05\n",
      "Epoch 1 Iteration 25\n",
      "Loss: 6.194374e-05\n",
      "Epoch 1 Iteration 26\n",
      "Loss: 5.5939203e-05\n",
      "Epoch 1 Iteration 27\n",
      "Loss: 5.4448014e-05\n",
      "Epoch 1 Iteration 28\n",
      "Loss: 4.502789e-05\n",
      "Epoch 1 Iteration 29\n",
      "Loss: 3.0721247e-05\n",
      "Epoch 1 Iteration 30\n",
      "Loss: 6.224992e-05\n",
      "Epoch 1 Iteration 31\n",
      "Loss: 4.8270576e-05\n",
      "Epoch 1 Iteration 32\n",
      "Loss: 5.5671626e-05\n",
      "Epoch 1 Iteration 33\n",
      "Loss: 3.175962e-05\n",
      "Epoch 1 Iteration 34\n",
      "Loss: 3.6833106e-05\n",
      "Valid loss: 5.041688564233482e-05\n",
      "Epoch  2  Current time:  Sun Jun 21 22:03:34 2020\n",
      "Epoch 2 Iteration 0\n",
      "Loss: 3.2261283e-05\n",
      "Epoch 2 Iteration 1\n",
      "Loss: 5.9300746e-05\n",
      "Epoch 2 Iteration 2\n",
      "Loss: 8.7453045e-05\n",
      "Epoch 2 Iteration 3\n",
      "Loss: 5.7365487e-05\n",
      "Epoch 2 Iteration 4\n",
      "Loss: 5.9288996e-05\n",
      "Epoch 2 Iteration 5\n",
      "Loss: 2.5692767e-05\n",
      "Epoch 2 Iteration 6\n",
      "Loss: 2.2706658e-05\n",
      "Epoch 2 Iteration 7\n",
      "Loss: 2.3368237e-05\n",
      "Epoch 2 Iteration 8\n",
      "Loss: 3.3261356e-05\n",
      "Epoch 2 Iteration 9\n",
      "Loss: 4.9721883e-05\n",
      "Epoch 2 Iteration 10\n",
      "Loss: 2.0070267e-05\n",
      "Epoch 2 Iteration 11\n",
      "Loss: 1.3997884e-05\n",
      "Epoch 2 Iteration 12\n",
      "Loss: 2.2175796e-05\n",
      "Epoch 2 Iteration 13\n",
      "Loss: 5.343297e-05\n",
      "Epoch 2 Iteration 14\n",
      "Loss: 1.8233943e-05\n",
      "Epoch 2 Iteration 15\n",
      "Loss: 1.5276224e-05\n",
      "Epoch 2 Iteration 16\n",
      "Loss: 2.1384698e-05\n",
      "Epoch 2 Iteration 17\n",
      "Loss: 2.448094e-05\n",
      "Epoch 2 Iteration 18\n",
      "Loss: 4.7167847e-05\n",
      "Epoch 2 Iteration 19\n",
      "Loss: 4.3574488e-05\n",
      "Epoch 2 Iteration 20\n",
      "Loss: 5.1935716e-05\n",
      "Epoch 2 Iteration 21\n",
      "Loss: 3.7277365e-05\n",
      "Epoch 2 Iteration 22\n",
      "Loss: 3.3871092e-05\n",
      "Epoch 2 Iteration 23\n",
      "Loss: 2.9914343e-05\n",
      "Epoch 2 Iteration 24\n",
      "Loss: 1.3732203e-05\n",
      "Epoch 2 Iteration 25\n",
      "Loss: 4.0308016e-05\n",
      "Epoch 2 Iteration 26\n",
      "Loss: 3.271393e-05\n",
      "Epoch 2 Iteration 27\n",
      "Loss: 3.6472113e-05\n",
      "Epoch 2 Iteration 28\n",
      "Loss: 1.1955804e-05\n",
      "Epoch 2 Iteration 29\n",
      "Loss: 1.9524177e-05\n",
      "Epoch 2 Iteration 30\n",
      "Loss: 4.604434e-05\n",
      "Epoch 2 Iteration 31\n",
      "Loss: 4.9122824e-05\n",
      "Epoch 2 Iteration 32\n",
      "Loss: 3.8845672e-05\n",
      "Epoch 2 Iteration 33\n",
      "Loss: 1.3904788e-05\n",
      "Epoch 2 Iteration 34\n",
      "Loss: 4.0686573e-05\n",
      "Valid loss: 3.970108809880912e-05\n",
      "Epoch  3  Current time:  Sun Jun 21 22:04:18 2020\n",
      "Epoch 3 Iteration 0\n",
      "Loss: 9.655559e-06\n",
      "Epoch 3 Iteration 1\n",
      "Loss: 4.0255447e-05\n",
      "Epoch 3 Iteration 2\n",
      "Loss: 2.7841541e-05\n",
      "Epoch 3 Iteration 3\n",
      "Loss: 1.773245e-05\n",
      "Epoch 3 Iteration 4\n",
      "Loss: 3.429294e-05\n",
      "Epoch 3 Iteration 5\n",
      "Loss: 4.955005e-05\n",
      "Epoch 3 Iteration 6\n",
      "Loss: 4.7332676e-05\n",
      "Epoch 3 Iteration 7\n",
      "Loss: 1.663598e-05\n",
      "Epoch 3 Iteration 8\n",
      "Loss: 4.953806e-05\n",
      "Epoch 3 Iteration 9\n",
      "Loss: 1.2788211e-05\n",
      "Epoch 3 Iteration 10\n",
      "Loss: 1.3794126e-05\n",
      "Epoch 3 Iteration 11\n",
      "Loss: 4.6248824e-05\n",
      "Epoch 3 Iteration 12\n",
      "Loss: 3.6304606e-05\n",
      "Epoch 3 Iteration 13\n",
      "Loss: 4.6857785e-05\n",
      "Epoch 3 Iteration 14\n",
      "Loss: 2.9932124e-05\n",
      "Epoch 3 Iteration 15\n",
      "Loss: 4.054427e-05\n",
      "Epoch 3 Iteration 16\n",
      "Loss: 1.7561913e-05\n",
      "Epoch 3 Iteration 17\n",
      "Loss: 1.2869133e-05\n",
      "Epoch 3 Iteration 18\n",
      "Loss: 7.943254e-06\n",
      "Epoch 3 Iteration 19\n",
      "Loss: 3.9733877e-05\n",
      "Epoch 3 Iteration 20\n",
      "Loss: 1.9327323e-05\n",
      "Epoch 3 Iteration 21\n",
      "Loss: 3.8403272e-05\n",
      "Epoch 3 Iteration 22\n",
      "Loss: 9.307796e-06\n",
      "Epoch 3 Iteration 23\n",
      "Loss: 3.127763e-05\n",
      "Epoch 3 Iteration 24\n",
      "Loss: 4.034102e-05\n",
      "Epoch 3 Iteration 25\n",
      "Loss: 1.9145245e-05\n",
      "Epoch 3 Iteration 26\n",
      "Loss: 7.966779e-06\n",
      "Epoch 3 Iteration 27\n",
      "Loss: 2.8900147e-05\n",
      "Epoch 3 Iteration 28\n",
      "Loss: 8.519669e-06\n",
      "Epoch 3 Iteration 29\n",
      "Loss: 7.339636e-05\n",
      "Epoch 3 Iteration 30\n",
      "Loss: 1.5740772e-05\n",
      "Epoch 3 Iteration 31\n",
      "Loss: 1.0747032e-05\n",
      "Epoch 3 Iteration 32\n",
      "Loss: 4.2432934e-05\n",
      "Epoch 3 Iteration 33\n",
      "Loss: 4.2189415e-05\n",
      "Epoch 3 Iteration 34\n",
      "Loss: 2.4417517e-05\n",
      "Valid loss: 3.550270412233658e-05\n",
      "Epoch  4  Current time:  Sun Jun 21 22:05:02 2020\n",
      "Epoch 4 Iteration 0\n",
      "Loss: 3.0193787e-05\n",
      "Epoch 4 Iteration 1\n",
      "Loss: 2.8741766e-05\n",
      "Epoch 4 Iteration 2\n",
      "Loss: 4.6152316e-05\n",
      "Epoch 4 Iteration 3\n",
      "Loss: 8.414663e-06\n",
      "Epoch 4 Iteration 4\n",
      "Loss: 1.0225348e-05\n",
      "Epoch 4 Iteration 5\n",
      "Loss: 1.5767686e-05\n",
      "Epoch 4 Iteration 6\n",
      "Loss: 3.4801466e-05\n",
      "Epoch 4 Iteration 7\n",
      "Loss: 3.8367256e-05\n",
      "Epoch 4 Iteration 8\n",
      "Loss: 1.6062451e-05\n",
      "Epoch 4 Iteration 9\n",
      "Loss: 1.8390616e-05\n",
      "Epoch 4 Iteration 10\n",
      "Loss: 6.5207173e-06\n",
      "Epoch 4 Iteration 11\n",
      "Loss: 7.4377895e-05\n",
      "Epoch 4 Iteration 12\n",
      "Loss: 4.4033994e-05\n",
      "Epoch 4 Iteration 13\n",
      "Loss: 3.7303642e-05\n",
      "Epoch 4 Iteration 14\n",
      "Loss: 1.242248e-05\n",
      "Epoch 4 Iteration 15\n",
      "Loss: 1.0393878e-05\n",
      "Epoch 4 Iteration 16\n",
      "Loss: 2.653657e-05\n",
      "Epoch 4 Iteration 17\n",
      "Loss: 2.5042054e-05\n",
      "Epoch 4 Iteration 18\n",
      "Loss: 4.244459e-05\n",
      "Epoch 4 Iteration 19\n",
      "Loss: 7.202812e-06\n",
      "Epoch 4 Iteration 20\n",
      "Loss: 4.2050982e-05\n",
      "Epoch 4 Iteration 21\n",
      "Loss: 4.4550758e-05\n",
      "Epoch 4 Iteration 22\n",
      "Loss: 6.816719e-06\n",
      "Epoch 4 Iteration 23\n",
      "Loss: 2.806995e-05\n",
      "Epoch 4 Iteration 24\n",
      "Loss: 1.6607202e-05\n",
      "Epoch 4 Iteration 25\n",
      "Loss: 5.0760664e-06\n",
      "Epoch 4 Iteration 26\n",
      "Loss: 4.4047396e-05\n",
      "Epoch 4 Iteration 27\n",
      "Loss: 4.2292093e-05\n",
      "Epoch 4 Iteration 28\n",
      "Loss: 9.913472e-06\n",
      "Epoch 4 Iteration 29\n",
      "Loss: 2.9295947e-05\n",
      "Epoch 4 Iteration 30\n",
      "Loss: 3.573106e-05\n",
      "Epoch 4 Iteration 31\n",
      "Loss: 1.2887044e-05\n",
      "Epoch 4 Iteration 32\n",
      "Loss: 3.317954e-05\n",
      "Epoch 4 Iteration 33\n",
      "Loss: 1.2550183e-05\n",
      "Epoch 4 Iteration 34\n",
      "Loss: 3.754596e-05\n",
      "Valid loss: 3.358186950208619e-05\n",
      "Epoch  5  Current time:  Sun Jun 21 22:05:47 2020\n",
      "Epoch 5 Iteration 0\n",
      "Loss: 3.6388657e-05\n",
      "Epoch 5 Iteration 1\n",
      "Loss: 3.4403223e-05\n",
      "Epoch 5 Iteration 2\n",
      "Loss: 3.2882555e-05\n",
      "Epoch 5 Iteration 3\n",
      "Loss: 3.4917524e-05\n",
      "Epoch 5 Iteration 4\n",
      "Loss: 2.2919436e-05\n",
      "Epoch 5 Iteration 5\n",
      "Loss: 6.8808745e-06\n",
      "Epoch 5 Iteration 6\n",
      "Loss: 8.650263e-06\n",
      "Epoch 5 Iteration 7\n",
      "Loss: 9.061596e-06\n",
      "Epoch 5 Iteration 8\n",
      "Loss: 4.4192457e-05\n",
      "Epoch 5 Iteration 9\n",
      "Loss: 3.910549e-05\n",
      "Epoch 5 Iteration 10\n",
      "Loss: 4.3564953e-05\n",
      "Epoch 5 Iteration 11\n",
      "Loss: 9.018619e-06\n",
      "Epoch 5 Iteration 12\n",
      "Loss: 5.8017545e-06\n",
      "Epoch 5 Iteration 13\n",
      "Loss: 1.2106753e-05\n",
      "Epoch 5 Iteration 14\n",
      "Loss: 6.933156e-05\n",
      "Epoch 5 Iteration 15\n",
      "Loss: 3.6532994e-05\n",
      "Epoch 5 Iteration 16\n",
      "Loss: 4.1421954e-05\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 5 Iteration 17\n",
      "Loss: 1.1439394e-05\n",
      "Epoch 5 Iteration 18\n",
      "Loss: 4.6269015e-06\n",
      "Epoch 5 Iteration 19\n",
      "Loss: 4.047981e-06\n",
      "Epoch 5 Iteration 20\n",
      "Loss: 4.2187457e-05\n",
      "Epoch 5 Iteration 21\n",
      "Loss: 4.183932e-05\n",
      "Epoch 5 Iteration 22\n",
      "Loss: 1.3620641e-05\n",
      "Epoch 5 Iteration 23\n",
      "Loss: 2.2279377e-05\n",
      "Epoch 5 Iteration 24\n",
      "Loss: 2.8349938e-05\n",
      "Epoch 5 Iteration 25\n",
      "Loss: 1.5406851e-05\n",
      "Epoch 5 Iteration 26\n",
      "Loss: 8.2157e-06\n",
      "Epoch 5 Iteration 27\n",
      "Loss: 4.3990367e-06\n",
      "Epoch 5 Iteration 28\n",
      "Loss: 1.2071764e-05\n",
      "Epoch 5 Iteration 29\n",
      "Loss: 2.5230273e-05\n",
      "Epoch 5 Iteration 30\n",
      "Loss: 3.9898794e-05\n",
      "Epoch 5 Iteration 31\n",
      "Loss: 3.5112364e-05\n",
      "Epoch 5 Iteration 32\n",
      "Loss: 2.7574493e-05\n",
      "Epoch 5 Iteration 33\n",
      "Loss: 1.5165047e-05\n",
      "Epoch 5 Iteration 34\n",
      "Loss: 2.90573e-05\n",
      "Valid loss: 3.172888682456687e-05\n",
      "WARNING:tensorflow:From /home/yunyun/anaconda3/envs/py36/lib/python3.6/site-packages/tensorflow_core/python/training/saver.py:963: remove_checkpoint (from tensorflow.python.training.checkpoint_management) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use standard file APIs to delete files with this prefix.\n",
      "Epoch  6  Current time:  Sun Jun 21 22:06:31 2020\n",
      "Epoch 6 Iteration 0\n",
      "Loss: 4.2038464e-05\n",
      "Epoch 6 Iteration 1\n",
      "Loss: 1.2424056e-05\n",
      "Epoch 6 Iteration 2\n",
      "Loss: 4.4127055e-06\n",
      "Epoch 6 Iteration 3\n",
      "Loss: 2.5908976e-05\n",
      "Epoch 6 Iteration 4\n",
      "Loss: 2.215106e-05\n",
      "Epoch 6 Iteration 5\n",
      "Loss: 1.1051046e-05\n",
      "Epoch 6 Iteration 6\n",
      "Loss: 3.2438016e-05\n",
      "Epoch 6 Iteration 7\n",
      "Loss: 6.844665e-05\n",
      "Epoch 6 Iteration 8\n",
      "Loss: 1.2278277e-05\n",
      "Epoch 6 Iteration 9\n",
      "Loss: 4.404e-06\n",
      "Epoch 6 Iteration 10\n",
      "Loss: 5.2790856e-06\n",
      "Epoch 6 Iteration 11\n",
      "Loss: 3.944703e-05\n",
      "Epoch 6 Iteration 12\n",
      "Loss: 2.2646964e-05\n",
      "Epoch 6 Iteration 13\n",
      "Loss: 4.186763e-05\n",
      "Epoch 6 Iteration 14\n",
      "Loss: 2.6664828e-05\n",
      "Epoch 6 Iteration 15\n",
      "Loss: 4.3231677e-05\n",
      "Epoch 6 Iteration 16\n",
      "Loss: 4.152475e-05\n",
      "Epoch 6 Iteration 17\n",
      "Loss: 6.9604803e-06\n",
      "Epoch 6 Iteration 18\n",
      "Loss: 2.5494934e-05\n",
      "Epoch 6 Iteration 19\n",
      "Loss: 3.831114e-05\n",
      "Epoch 6 Iteration 20\n",
      "Loss: 4.6021814e-06\n",
      "Epoch 6 Iteration 21\n",
      "Loss: 9.02194e-06\n",
      "Epoch 6 Iteration 22\n",
      "Loss: 3.4029385e-05\n",
      "Epoch 6 Iteration 23\n",
      "Loss: 2.8353335e-05\n",
      "Epoch 6 Iteration 24\n",
      "Loss: 3.199658e-05\n",
      "Epoch 6 Iteration 25\n",
      "Loss: 1.4752244e-05\n",
      "Epoch 6 Iteration 26\n",
      "Loss: 3.3037372e-06\n",
      "Epoch 6 Iteration 27\n",
      "Loss: 3.6430265e-05\n",
      "Epoch 6 Iteration 28\n",
      "Loss: 8.2605775e-06\n",
      "Epoch 6 Iteration 29\n",
      "Loss: 8.252329e-06\n",
      "Epoch 6 Iteration 30\n",
      "Loss: 1.1479202e-05\n",
      "Epoch 6 Iteration 31\n",
      "Loss: 4.0219555e-05\n",
      "Epoch 6 Iteration 32\n",
      "Loss: 1.436203e-05\n",
      "Epoch 6 Iteration 33\n",
      "Loss: 3.3626675e-05\n",
      "Epoch 6 Iteration 34\n",
      "Loss: 3.539928e-05\n",
      "Valid loss: 3.1482493795920163e-05\n",
      "Epoch  7  Current time:  Sun Jun 21 22:07:15 2020\n",
      "Epoch 7 Iteration 0\n",
      "Loss: 4.525687e-06\n",
      "Epoch 7 Iteration 1\n",
      "Loss: 3.3762633e-05\n",
      "Epoch 7 Iteration 2\n",
      "Loss: 1.497822e-05\n",
      "Epoch 7 Iteration 3\n",
      "Loss: 3.7780255e-06\n",
      "Epoch 7 Iteration 4\n",
      "Loss: 3.0927302e-05\n",
      "Epoch 7 Iteration 5\n",
      "Loss: 1.2208872e-05\n",
      "Epoch 7 Iteration 6\n",
      "Loss: 2.660165e-05\n",
      "Epoch 7 Iteration 7\n",
      "Loss: 2.096691e-05\n",
      "Epoch 7 Iteration 8\n",
      "Loss: 6.826481e-06\n",
      "Epoch 7 Iteration 9\n",
      "Loss: 2.397387e-05\n",
      "Epoch 7 Iteration 10\n",
      "Loss: 1.981938e-05\n",
      "Epoch 7 Iteration 11\n",
      "Loss: 3.2045325e-06\n",
      "Epoch 7 Iteration 12\n",
      "Loss: 4.277096e-05\n",
      "Epoch 7 Iteration 13\n",
      "Loss: 4.202375e-06\n",
      "Epoch 7 Iteration 14\n",
      "Loss: 1.4061216e-05\n",
      "Epoch 7 Iteration 15\n",
      "Loss: 7.526159e-06\n",
      "Epoch 7 Iteration 16\n",
      "Loss: 3.36515e-06\n",
      "Epoch 7 Iteration 17\n",
      "Loss: 2.6173073e-05\n",
      "Epoch 7 Iteration 18\n",
      "Loss: 1.0832708e-05\n",
      "Epoch 7 Iteration 19\n",
      "Loss: 2.4724763e-05\n",
      "Epoch 7 Iteration 20\n",
      "Loss: 3.06418e-05\n",
      "Epoch 7 Iteration 21\n",
      "Loss: 6.441125e-06\n",
      "Epoch 7 Iteration 22\n",
      "Loss: 1.1175059e-05\n",
      "Epoch 7 Iteration 23\n",
      "Loss: 1.0001376e-05\n",
      "Epoch 7 Iteration 24\n",
      "Loss: 3.496139e-05\n",
      "Epoch 7 Iteration 25\n",
      "Loss: 3.2100837e-05\n",
      "Epoch 7 Iteration 26\n",
      "Loss: 3.897575e-05\n",
      "Epoch 7 Iteration 27\n",
      "Loss: 3.3008466e-05\n",
      "Epoch 7 Iteration 28\n",
      "Loss: 6.585045e-05\n",
      "Epoch 7 Iteration 29\n",
      "Loss: 4.0935847e-05\n",
      "Epoch 7 Iteration 30\n",
      "Loss: 3.796499e-05\n",
      "Epoch 7 Iteration 31\n",
      "Loss: 7.39481e-06\n",
      "Epoch 7 Iteration 32\n",
      "Loss: 3.8934184e-05\n",
      "Epoch 7 Iteration 33\n",
      "Loss: 4.0583946e-05\n",
      "Epoch 7 Iteration 34\n",
      "Loss: 4.149188e-05\n",
      "Valid loss: 3.1108658731682226e-05\n",
      "Epoch  8  Current time:  Sun Jun 21 22:08:00 2020\n",
      "Epoch 8 Iteration 0\n",
      "Loss: 1.1204143e-05\n",
      "Epoch 8 Iteration 1\n",
      "Loss: 2.672696e-05\n",
      "Epoch 8 Iteration 2\n",
      "Loss: 3.3791348e-05\n",
      "Epoch 8 Iteration 3\n",
      "Loss: 3.0310166e-05\n",
      "Epoch 8 Iteration 4\n",
      "Loss: 3.309166e-05\n",
      "Epoch 8 Iteration 5\n",
      "Loss: 7.614758e-06\n",
      "Epoch 8 Iteration 6\n",
      "Loss: 3.913458e-06\n",
      "Epoch 8 Iteration 7\n",
      "Loss: 2.0301904e-05\n",
      "Epoch 8 Iteration 8\n",
      "Loss: 4.0921754e-05\n",
      "Epoch 8 Iteration 9\n",
      "Loss: 3.5194975e-05\n",
      "Epoch 8 Iteration 10\n",
      "Loss: 1.5046757e-05\n",
      "Epoch 8 Iteration 11\n",
      "Loss: 4.073679e-05\n",
      "Epoch 8 Iteration 12\n",
      "Loss: 2.4704766e-05\n",
      "Epoch 8 Iteration 13\n",
      "Loss: 6.4803135e-06\n",
      "Epoch 8 Iteration 14\n",
      "Loss: 2.5942847e-05\n",
      "Epoch 8 Iteration 15\n",
      "Loss: 2.0917996e-05\n",
      "Epoch 8 Iteration 16\n",
      "Loss: 6.540373e-05\n",
      "Epoch 8 Iteration 17\n",
      "Loss: 3.9107254e-05\n",
      "Epoch 8 Iteration 18\n",
      "Loss: 3.845628e-05\n",
      "Epoch 8 Iteration 19\n",
      "Loss: 1.0919975e-05\n",
      "Epoch 8 Iteration 20\n",
      "Loss: 4.1412226e-05\n",
      "Epoch 8 Iteration 21\n",
      "Loss: 4.001152e-06\n",
      "Epoch 8 Iteration 22\n",
      "Loss: 3.802166e-05\n",
      "Epoch 8 Iteration 23\n",
      "Loss: 1.1442243e-05\n",
      "Epoch 8 Iteration 24\n",
      "Loss: 3.1249458e-06\n",
      "Epoch 8 Iteration 25\n",
      "Loss: 6.7036553e-06\n",
      "Epoch 8 Iteration 26\n",
      "Loss: 3.285804e-05\n",
      "Epoch 8 Iteration 27\n",
      "Loss: 2.4544247e-06\n",
      "Epoch 8 Iteration 28\n",
      "Loss: 1.33087715e-05\n",
      "Epoch 8 Iteration 29\n",
      "Loss: 3.201608e-05\n",
      "Epoch 8 Iteration 30\n",
      "Loss: 2.4648327e-05\n",
      "Epoch 8 Iteration 31\n",
      "Loss: 7.643871e-06\n",
      "Epoch 8 Iteration 32\n",
      "Loss: 3.1198051e-06\n",
      "Epoch 8 Iteration 33\n",
      "Loss: 1.1769877e-05\n",
      "Epoch 8 Iteration 34\n",
      "Loss: 3.9371917e-05\n",
      "Valid loss: 3.2449377613374963e-05\n",
      "Epoch  9  Current time:  Sun Jun 21 22:08:44 2020\n",
      "Epoch 9 Iteration 0\n",
      "Loss: 3.819237e-05\n",
      "Epoch 9 Iteration 1\n",
      "Loss: 3.4657813e-05\n",
      "Epoch 9 Iteration 2\n",
      "Loss: 1.9479068e-05\n",
      "Epoch 9 Iteration 3\n",
      "Loss: 2.5899244e-05\n",
      "Epoch 9 Iteration 4\n",
      "Loss: 3.1976564e-05\n",
      "Epoch 9 Iteration 5\n",
      "Loss: 6.1970527e-06\n",
      "Epoch 9 Iteration 6\n",
      "Loss: 2.43131e-05\n",
      "Epoch 9 Iteration 7\n",
      "Loss: 7.2623893e-06\n",
      "Epoch 9 Iteration 8\n",
      "Loss: 5.17016e-06\n",
      "Epoch 9 Iteration 9\n",
      "Loss: 2.9326413e-06\n",
      "Epoch 9 Iteration 10\n",
      "Loss: 1.13805745e-05\n",
      "Epoch 9 Iteration 11\n",
      "Loss: 7.654062e-06\n",
      "Epoch 9 Iteration 12\n",
      "Loss: 1.0343036e-05\n",
      "Epoch 9 Iteration 13\n",
      "Loss: 3.510096e-05\n",
      "Epoch 9 Iteration 14\n",
      "Loss: 1.3171772e-05\n",
      "Epoch 9 Iteration 15\n",
      "Loss: 1.3775553e-05\n",
      "Epoch 9 Iteration 16\n",
      "Loss: 1.447706e-05\n",
      "Epoch 9 Iteration 17\n",
      "Loss: 4.4842865e-05\n",
      "Epoch 9 Iteration 18\n",
      "Loss: 4.115876e-05\n",
      "Epoch 9 Iteration 19\n",
      "Loss: 1.7755627e-05\n",
      "Epoch 9 Iteration 20\n",
      "Loss: 7.859732e-06\n",
      "Epoch 9 Iteration 21\n",
      "Loss: 3.675774e-05\n",
      "Epoch 9 Iteration 22\n",
      "Loss: 3.303409e-06\n",
      "Epoch 9 Iteration 23\n",
      "Loss: 3.702411e-05\n",
      "Epoch 9 Iteration 24\n",
      "Loss: 3.069903e-05\n",
      "Epoch 9 Iteration 25\n",
      "Loss: 7.0436254e-06\n",
      "Epoch 9 Iteration 26\n",
      "Loss: 2.0645759e-05\n",
      "Epoch 9 Iteration 27\n",
      "Loss: 4.2342883e-05\n",
      "Epoch 9 Iteration 28\n",
      "Loss: 3.8696024e-05\n",
      "Epoch 9 Iteration 29\n",
      "Loss: 2.6011274e-05\n",
      "Epoch 9 Iteration 30\n",
      "Loss: 6.448028e-05\n",
      "Epoch 9 Iteration 31\n",
      "Loss: 2.7041664e-05\n",
      "Epoch 9 Iteration 32\n",
      "Loss: 3.2349977e-05\n",
      "Epoch 9 Iteration 33\n",
      "Loss: 3.7768375e-05\n",
      "Epoch 9 Iteration 34\n",
      "Loss: 6.269662e-06\n",
      "Valid loss: 3.025455225724727e-05\n",
      "Epoch  10  Current time:  Sun Jun 21 22:09:28 2020\n",
      "Epoch 10 Iteration 0\n",
      "Loss: 2.4851652e-05\n",
      "Epoch 10 Iteration 1\n",
      "Loss: 3.2428914e-05\n",
      "Epoch 10 Iteration 2\n",
      "Loss: 4.1728097e-05\n",
      "Epoch 10 Iteration 3\n",
      "Loss: 4.495643e-06\n",
      "Epoch 10 Iteration 4\n",
      "Loss: 2.2895965e-06\n",
      "Epoch 10 Iteration 5\n",
      "Loss: 4.198949e-05\n",
      "Epoch 10 Iteration 6\n",
      "Loss: 3.420788e-05\n",
      "Epoch 10 Iteration 7\n",
      "Loss: 1.4308984e-05\n",
      "Epoch 10 Iteration 8\n",
      "Loss: 7.54463e-06\n",
      "Epoch 10 Iteration 9\n",
      "Loss: 1.0830108e-05\n",
      "Epoch 10 Iteration 10\n",
      "Loss: 2.086783e-05\n",
      "Epoch 10 Iteration 11\n",
      "Loss: 6.9240996e-06\n",
      "Epoch 10 Iteration 12\n",
      "Loss: 1.1373316e-05\n",
      "Epoch 10 Iteration 13\n",
      "Loss: 4.3564265e-05\n",
      "Epoch 10 Iteration 14\n",
      "Loss: 3.8028167e-05\n",
      "Epoch 10 Iteration 15\n",
      "Loss: 1.4689078e-05\n",
      "Epoch 10 Iteration 16\n",
      "Loss: 1.0739655e-05\n",
      "Epoch 10 Iteration 17\n",
      "Loss: 3.52724e-05\n",
      "Epoch 10 Iteration 18\n",
      "Loss: 2.505399e-06\n",
      "Epoch 10 Iteration 19\n",
      "Loss: 3.3935485e-06\n",
      "Epoch 10 Iteration 20\n",
      "Loss: 7.417281e-06\n",
      "Epoch 10 Iteration 21\n",
      "Loss: 6.5457875e-06\n",
      "Epoch 10 Iteration 22\n",
      "Loss: 4.1123294e-05\n",
      "Epoch 10 Iteration 23\n",
      "Loss: 2.4244071e-05\n",
      "Epoch 10 Iteration 24\n",
      "Loss: 2.093247e-05\n",
      "Epoch 10 Iteration 25\n",
      "Loss: 3.291759e-05\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 10 Iteration 26\n",
      "Loss: 2.3912506e-05\n",
      "Epoch 10 Iteration 27\n",
      "Loss: 4.1943324e-05\n",
      "Epoch 10 Iteration 28\n",
      "Loss: 3.745509e-05\n",
      "Epoch 10 Iteration 29\n",
      "Loss: 6.651653e-05\n",
      "Epoch 10 Iteration 30\n",
      "Loss: 3.06693e-05\n",
      "Epoch 10 Iteration 31\n",
      "Loss: 4.036729e-06\n",
      "Epoch 10 Iteration 32\n",
      "Loss: 1.1604955e-05\n",
      "Epoch 10 Iteration 33\n",
      "Loss: 2.6444346e-05\n",
      "Epoch 10 Iteration 34\n",
      "Loss: 3.272808e-05\n",
      "Valid loss: 3.069291415158659e-05\n",
      "Epoch  11  Current time:  Sun Jun 21 22:10:13 2020\n",
      "Epoch 11 Iteration 0\n",
      "Loss: 3.284521e-05\n",
      "Epoch 11 Iteration 1\n",
      "Loss: 7.151776e-06\n",
      "Epoch 11 Iteration 2\n",
      "Loss: 4.1143227e-05\n",
      "Epoch 11 Iteration 3\n",
      "Loss: 1.0047699e-05\n",
      "Epoch 11 Iteration 4\n",
      "Loss: 6.395062e-05\n",
      "Epoch 11 Iteration 5\n",
      "Loss: 3.170438e-06\n",
      "Epoch 11 Iteration 6\n",
      "Loss: 3.0049987e-06\n",
      "Epoch 11 Iteration 7\n",
      "Loss: 3.777558e-05\n",
      "Epoch 11 Iteration 8\n",
      "Loss: 6.4532364e-06\n",
      "Epoch 11 Iteration 9\n",
      "Loss: 3.3525255e-06\n",
      "Epoch 11 Iteration 10\n",
      "Loss: 2.5915751e-05\n",
      "Epoch 11 Iteration 11\n",
      "Loss: 4.2519874e-05\n",
      "Epoch 11 Iteration 12\n",
      "Loss: 6.42216e-06\n",
      "Epoch 11 Iteration 13\n",
      "Loss: 1.3711441e-05\n",
      "Epoch 11 Iteration 14\n",
      "Loss: 3.8469734e-06\n",
      "Epoch 11 Iteration 15\n",
      "Loss: 3.85908e-05\n",
      "Epoch 11 Iteration 16\n",
      "Loss: 7.3663246e-06\n",
      "Epoch 11 Iteration 17\n",
      "Loss: 4.238462e-05\n",
      "Epoch 11 Iteration 18\n",
      "Loss: 2.2038908e-05\n",
      "Epoch 11 Iteration 19\n",
      "Loss: 1.0941384e-05\n",
      "Epoch 11 Iteration 20\n",
      "Loss: 2.6710182e-05\n",
      "Epoch 11 Iteration 21\n",
      "Loss: 3.568012e-05\n",
      "Epoch 11 Iteration 22\n",
      "Loss: 3.317354e-05\n",
      "Epoch 11 Iteration 23\n",
      "Loss: 2.365748e-05\n",
      "Epoch 11 Iteration 24\n",
      "Loss: 1.0811572e-05\n",
      "Epoch 11 Iteration 25\n",
      "Loss: 1.9187946e-05\n",
      "Epoch 11 Iteration 26\n",
      "Loss: 3.7730602e-05\n",
      "Epoch 11 Iteration 27\n",
      "Loss: 3.1154043e-05\n",
      "Epoch 11 Iteration 28\n",
      "Loss: 2.4102112e-05\n",
      "Epoch 11 Iteration 29\n",
      "Loss: 3.9847214e-06\n",
      "Epoch 11 Iteration 30\n",
      "Loss: 3.9985087e-05\n",
      "Epoch 11 Iteration 31\n",
      "Loss: 1.243704e-05\n",
      "Epoch 11 Iteration 32\n",
      "Loss: 3.034866e-05\n",
      "Epoch 11 Iteration 33\n",
      "Loss: 1.3743536e-05\n",
      "Epoch 11 Iteration 34\n",
      "Loss: 3.176545e-05\n",
      "Valid loss: 3.0320890800794587e-05\n",
      "Epoch  12  Current time:  Sun Jun 21 22:10:57 2020\n",
      "Epoch 12 Iteration 0\n",
      "Loss: 3.2198415e-05\n",
      "Epoch 12 Iteration 1\n",
      "Loss: 3.9861123e-05\n",
      "Epoch 12 Iteration 2\n",
      "Loss: 4.0167502e-05\n",
      "Epoch 12 Iteration 3\n",
      "Loss: 1.19282595e-05\n",
      "Epoch 12 Iteration 4\n",
      "Loss: 1.0999909e-05\n",
      "Epoch 12 Iteration 5\n",
      "Loss: 1.3905816e-05\n",
      "Epoch 12 Iteration 6\n",
      "Loss: 3.873793e-06\n",
      "Epoch 12 Iteration 7\n",
      "Loss: 2.541564e-05\n",
      "Epoch 12 Iteration 8\n",
      "Loss: 4.6005716e-06\n",
      "Epoch 12 Iteration 9\n",
      "Loss: 1.1637187e-05\n",
      "Epoch 12 Iteration 10\n",
      "Loss: 3.8009064e-05\n",
      "Epoch 12 Iteration 11\n",
      "Loss: 3.055361e-05\n",
      "Epoch 12 Iteration 12\n",
      "Loss: 2.4684865e-05\n",
      "Epoch 12 Iteration 13\n",
      "Loss: 4.0884195e-05\n",
      "Epoch 12 Iteration 14\n",
      "Loss: 3.5940423e-06\n",
      "Epoch 12 Iteration 15\n",
      "Loss: 6.770845e-06\n",
      "Epoch 12 Iteration 16\n",
      "Loss: 3.4296303e-05\n",
      "Epoch 12 Iteration 17\n",
      "Loss: 1.4347996e-05\n",
      "Epoch 12 Iteration 18\n",
      "Loss: 7.219636e-06\n",
      "Epoch 12 Iteration 19\n",
      "Loss: 3.041579e-05\n",
      "Epoch 12 Iteration 20\n",
      "Loss: 3.178205e-05\n",
      "Epoch 12 Iteration 21\n",
      "Loss: 4.6897258e-06\n",
      "Epoch 12 Iteration 22\n",
      "Loss: 3.7988295e-05\n",
      "Epoch 12 Iteration 23\n",
      "Loss: 2.0105215e-05\n",
      "Epoch 12 Iteration 24\n",
      "Loss: 1.9069616e-05\n",
      "Epoch 12 Iteration 25\n",
      "Loss: 6.328396e-06\n",
      "Epoch 12 Iteration 26\n",
      "Loss: 1.0132078e-05\n",
      "Epoch 12 Iteration 27\n",
      "Loss: 4.0891257e-05\n",
      "Epoch 12 Iteration 28\n",
      "Loss: 6.868374e-06\n",
      "Epoch 12 Iteration 29\n",
      "Loss: 4.0259813e-05\n",
      "Epoch 12 Iteration 30\n",
      "Loss: 3.251202e-05\n",
      "Epoch 12 Iteration 31\n",
      "Loss: 2.4854922e-05\n",
      "Epoch 12 Iteration 32\n",
      "Loss: 2.4937277e-05\n",
      "Epoch 12 Iteration 33\n",
      "Loss: 2.8610546e-06\n",
      "Epoch 12 Iteration 34\n",
      "Loss: 6.603171e-05\n",
      "Valid loss: 3.070003367611207e-05\n",
      "Epoch  13  Current time:  Sun Jun 21 22:11:41 2020\n",
      "Epoch 13 Iteration 0\n",
      "Loss: 2.6854737e-05\n",
      "Epoch 13 Iteration 1\n",
      "Loss: 4.3494954e-05\n",
      "Epoch 13 Iteration 2\n",
      "Loss: 3.0913372e-05\n",
      "Epoch 13 Iteration 3\n",
      "Loss: 3.903667e-05\n",
      "Epoch 13 Iteration 4\n",
      "Loss: 1.4353192e-05\n",
      "Epoch 13 Iteration 5\n",
      "Loss: 3.5582073e-06\n",
      "Epoch 13 Iteration 6\n",
      "Loss: 1.6202925e-05\n",
      "Epoch 13 Iteration 7\n",
      "Loss: 2.4034462e-05\n",
      "Epoch 13 Iteration 8\n",
      "Loss: 6.103093e-05\n",
      "Epoch 13 Iteration 9\n",
      "Loss: 7.181087e-06\n",
      "Epoch 13 Iteration 10\n",
      "Loss: 4.1720427e-06\n",
      "Epoch 13 Iteration 11\n",
      "Loss: 4.3546906e-06\n",
      "Epoch 13 Iteration 12\n",
      "Loss: 3.2169297e-05\n",
      "Epoch 13 Iteration 13\n",
      "Loss: 2.7299453e-05\n",
      "Epoch 13 Iteration 14\n",
      "Loss: 9.610481e-06\n",
      "Epoch 13 Iteration 15\n",
      "Loss: 2.4462197e-05\n",
      "Epoch 13 Iteration 16\n",
      "Loss: 3.1849268e-06\n",
      "Epoch 13 Iteration 17\n",
      "Loss: 6.484913e-06\n",
      "Epoch 13 Iteration 18\n",
      "Loss: 2.2750639e-05\n",
      "Epoch 13 Iteration 19\n",
      "Loss: 1.0456748e-05\n",
      "Epoch 13 Iteration 20\n",
      "Loss: 4.0927833e-05\n",
      "Epoch 13 Iteration 21\n",
      "Loss: 3.6909664e-06\n",
      "Epoch 13 Iteration 22\n",
      "Loss: 3.475227e-05\n",
      "Epoch 13 Iteration 23\n",
      "Loss: 3.907403e-05\n",
      "Epoch 13 Iteration 24\n",
      "Loss: 3.85509e-05\n",
      "Epoch 13 Iteration 25\n",
      "Loss: 1.2960904e-05\n",
      "Epoch 13 Iteration 26\n",
      "Loss: 1.9067262e-05\n",
      "Epoch 13 Iteration 27\n",
      "Loss: 4.018253e-05\n",
      "Epoch 13 Iteration 28\n",
      "Loss: 3.2265903e-05\n",
      "Epoch 13 Iteration 29\n",
      "Loss: 7.4113186e-06\n",
      "Epoch 13 Iteration 30\n",
      "Loss: 5.487912e-05\n",
      "Epoch 13 Iteration 31\n",
      "Loss: 6.2379813e-06\n",
      "Epoch 13 Iteration 32\n",
      "Loss: 3.1614814e-05\n",
      "Epoch 13 Iteration 33\n",
      "Loss: 1.0562156e-05\n",
      "Epoch 13 Iteration 34\n",
      "Loss: 3.105608e-05\n",
      "Valid loss: 3.12468946503941e-05\n",
      "Epoch  14  Current time:  Sun Jun 21 22:12:26 2020\n",
      "Epoch 14 Iteration 0\n",
      "Loss: 4.204311e-05\n",
      "Epoch 14 Iteration 1\n",
      "Loss: 7.301487e-06\n",
      "Epoch 14 Iteration 2\n",
      "Loss: 3.605263e-06\n",
      "Epoch 14 Iteration 3\n",
      "Loss: 3.822292e-05\n",
      "Epoch 14 Iteration 4\n",
      "Loss: 2.690997e-05\n",
      "Epoch 14 Iteration 5\n",
      "Loss: 3.8750335e-05\n",
      "Epoch 14 Iteration 6\n",
      "Loss: 1.415925e-05\n",
      "Epoch 14 Iteration 7\n",
      "Loss: 4.378096e-06\n",
      "Epoch 14 Iteration 8\n",
      "Loss: 3.6073423e-05\n",
      "Epoch 14 Iteration 9\n",
      "Loss: 3.3111333e-05\n",
      "Epoch 14 Iteration 10\n",
      "Loss: 2.2267908e-05\n",
      "Epoch 14 Iteration 11\n",
      "Loss: 2.4122339e-05\n",
      "Epoch 14 Iteration 12\n",
      "Loss: 3.1727865e-05\n",
      "Epoch 14 Iteration 13\n",
      "Loss: 1.0591384e-05\n",
      "Epoch 14 Iteration 14\n",
      "Loss: 6.909666e-05\n",
      "Epoch 14 Iteration 15\n",
      "Loss: 1.2213508e-05\n",
      "Epoch 14 Iteration 16\n",
      "Loss: 2.582532e-06\n",
      "Epoch 14 Iteration 17\n",
      "Loss: 4.3390806e-05\n",
      "Epoch 14 Iteration 18\n",
      "Loss: 2.0816926e-05\n",
      "Epoch 14 Iteration 19\n",
      "Loss: 4.752994e-06\n",
      "Epoch 14 Iteration 20\n",
      "Loss: 1.514764e-05\n",
      "Epoch 14 Iteration 21\n",
      "Loss: 1.1402704e-05\n",
      "Epoch 14 Iteration 22\n",
      "Loss: 3.9941224e-05\n",
      "Epoch 14 Iteration 23\n",
      "Loss: 4.10488e-05\n",
      "Epoch 14 Iteration 24\n",
      "Loss: 2.7139184e-05\n",
      "Epoch 14 Iteration 25\n",
      "Loss: 3.1161857e-05\n",
      "Epoch 14 Iteration 26\n",
      "Loss: 3.955271e-05\n",
      "Epoch 14 Iteration 27\n",
      "Loss: 6.565409e-06\n",
      "Epoch 14 Iteration 28\n",
      "Loss: 1.0795082e-05\n",
      "Epoch 14 Iteration 29\n",
      "Loss: 2.680882e-06\n",
      "Epoch 14 Iteration 30\n",
      "Loss: 6.289863e-06\n",
      "Epoch 14 Iteration 31\n",
      "Loss: 3.2964832e-05\n",
      "Epoch 14 Iteration 32\n",
      "Loss: 6.6944767e-06\n",
      "Epoch 14 Iteration 33\n",
      "Loss: 2.3410175e-05\n",
      "Epoch 14 Iteration 34\n",
      "Loss: 3.240515e-05\n",
      "Valid loss: 3.0206821975298226e-05\n",
      "Epoch  15  Current time:  Sun Jun 21 22:13:10 2020\n",
      "Epoch 15 Iteration 0\n",
      "Loss: 3.0932493e-05\n",
      "Epoch 15 Iteration 1\n",
      "Loss: 3.4201303e-06\n",
      "Epoch 15 Iteration 2\n",
      "Loss: 1.0392741e-05\n",
      "Epoch 15 Iteration 3\n",
      "Loss: 2.590049e-05\n",
      "Epoch 15 Iteration 4\n",
      "Loss: 4.2565543e-06\n",
      "Epoch 15 Iteration 5\n",
      "Loss: 2.4792134e-05\n",
      "Epoch 15 Iteration 6\n",
      "Loss: 4.0793373e-05\n",
      "Epoch 15 Iteration 7\n",
      "Loss: 3.8246148e-05\n",
      "Epoch 15 Iteration 8\n",
      "Loss: 3.80835e-05\n",
      "Epoch 15 Iteration 9\n",
      "Loss: 1.3782446e-05\n",
      "Epoch 15 Iteration 10\n",
      "Loss: 1.9184146e-05\n",
      "Epoch 15 Iteration 11\n",
      "Loss: 2.465081e-05\n",
      "Epoch 15 Iteration 12\n",
      "Loss: 3.5183064e-05\n",
      "Epoch 15 Iteration 13\n",
      "Loss: 1.1635333e-05\n",
      "Epoch 15 Iteration 14\n",
      "Loss: 3.929287e-05\n",
      "Epoch 15 Iteration 15\n",
      "Loss: 9.713992e-06\n",
      "Epoch 15 Iteration 16\n",
      "Loss: 3.1417017e-05\n",
      "Epoch 15 Iteration 17\n",
      "Loss: 1.0893483e-05\n",
      "Epoch 15 Iteration 18\n",
      "Loss: 6.5169156e-06\n",
      "Epoch 15 Iteration 19\n",
      "Loss: 6.5821528e-06\n",
      "Epoch 15 Iteration 20\n",
      "Loss: 6.314577e-05\n",
      "Epoch 15 Iteration 21\n",
      "Loss: 3.0297324e-05\n",
      "Epoch 15 Iteration 22\n",
      "Loss: 1.5394284e-05\n",
      "Epoch 15 Iteration 23\n",
      "Loss: 2.7545975e-06\n",
      "Epoch 15 Iteration 24\n",
      "Loss: 2.0993932e-05\n",
      "Epoch 15 Iteration 25\n",
      "Loss: 5.936324e-06\n",
      "Epoch 15 Iteration 26\n",
      "Loss: 4.1177893e-05\n",
      "Epoch 15 Iteration 27\n",
      "Loss: 3.8948936e-05\n",
      "Epoch 15 Iteration 28\n",
      "Loss: 7.6277925e-06\n",
      "Epoch 15 Iteration 29\n",
      "Loss: 2.3505661e-05\n",
      "Epoch 15 Iteration 30\n",
      "Loss: 3.9320556e-05\n",
      "Epoch 15 Iteration 31\n",
      "Loss: 3.9837246e-06\n",
      "Epoch 15 Iteration 32\n",
      "Loss: 3.2010856e-05\n",
      "Epoch 15 Iteration 33\n",
      "Loss: 3.9995525e-06\n",
      "Epoch 15 Iteration 34\n",
      "Loss: 3.1329095e-05\n",
      "Valid loss: 2.9365342925302684e-05\n",
      "Epoch  16  Current time:  Sun Jun 21 22:13:55 2020\n",
      "Epoch 16 Iteration 0\n",
      "Loss: 3.7898265e-06\n",
      "Epoch 16 Iteration 1\n",
      "Loss: 1.3616929e-05\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 16 Iteration 2\n",
      "Loss: 3.0630625e-05\n",
      "Epoch 16 Iteration 3\n",
      "Loss: 3.7621012e-05\n",
      "Epoch 16 Iteration 4\n",
      "Loss: 1.0242598e-05\n",
      "Epoch 16 Iteration 5\n",
      "Loss: 2.9444633e-05\n",
      "Epoch 16 Iteration 6\n",
      "Loss: 3.8492868e-05\n",
      "Epoch 16 Iteration 7\n",
      "Loss: 2.24768e-05\n",
      "Epoch 16 Iteration 8\n",
      "Loss: 1.0252785e-05\n",
      "Epoch 16 Iteration 9\n",
      "Loss: 3.6262467e-05\n",
      "Epoch 16 Iteration 10\n",
      "Loss: 4.1906154e-05\n",
      "Epoch 16 Iteration 11\n",
      "Loss: 3.0043971e-05\n",
      "Epoch 16 Iteration 12\n",
      "Loss: 7.2702437e-06\n",
      "Epoch 16 Iteration 13\n",
      "Loss: 3.165414e-06\n",
      "Epoch 16 Iteration 14\n",
      "Loss: 3.786936e-05\n",
      "Epoch 16 Iteration 15\n",
      "Loss: 4.090028e-05\n",
      "Epoch 16 Iteration 16\n",
      "Loss: 2.4541634e-05\n",
      "Epoch 16 Iteration 17\n",
      "Loss: 3.3655902e-05\n",
      "Epoch 16 Iteration 18\n",
      "Loss: 4.0091792e-05\n",
      "Epoch 16 Iteration 19\n",
      "Loss: 3.172355e-06\n",
      "Epoch 16 Iteration 20\n",
      "Loss: 6.4046685e-06\n",
      "Epoch 16 Iteration 21\n",
      "Loss: 1.3757712e-05\n",
      "Epoch 16 Iteration 22\n",
      "Loss: 4.0931745e-06\n",
      "Epoch 16 Iteration 23\n",
      "Loss: 6.065295e-06\n",
      "Epoch 16 Iteration 24\n",
      "Loss: 1.0697197e-05\n",
      "Epoch 16 Iteration 25\n",
      "Loss: 6.203333e-05\n",
      "Epoch 16 Iteration 26\n",
      "Loss: 3.0573698e-05\n",
      "Epoch 16 Iteration 27\n",
      "Loss: 1.8371731e-05\n",
      "Epoch 16 Iteration 28\n",
      "Loss: 6.3013167e-06\n",
      "Epoch 16 Iteration 29\n",
      "Loss: 2.593662e-06\n",
      "Epoch 16 Iteration 30\n",
      "Loss: 2.0078613e-05\n",
      "Epoch 16 Iteration 31\n",
      "Loss: 1.1293451e-05\n",
      "Epoch 16 Iteration 32\n",
      "Loss: 2.3324363e-05\n",
      "Epoch 16 Iteration 33\n",
      "Loss: 2.5732828e-05\n",
      "Epoch 16 Iteration 34\n",
      "Loss: 3.129097e-05\n",
      "Valid loss: 2.9213077141321264e-05\n",
      "Epoch  17  Current time:  Sun Jun 21 22:14:39 2020\n",
      "Epoch 17 Iteration 0\n",
      "Loss: 3.4731464e-05\n",
      "Epoch 17 Iteration 1\n",
      "Loss: 2.981667e-05\n",
      "Epoch 17 Iteration 2\n",
      "Loss: 7.06131e-06\n",
      "Epoch 17 Iteration 3\n",
      "Loss: 2.2089032e-05\n",
      "Epoch 17 Iteration 4\n",
      "Loss: 6.5521294e-06\n",
      "Epoch 17 Iteration 5\n",
      "Loss: 8.441475e-06\n",
      "Epoch 17 Iteration 6\n",
      "Loss: 6.327618e-05\n",
      "Epoch 17 Iteration 7\n",
      "Loss: 3.0601463e-05\n",
      "Epoch 17 Iteration 8\n",
      "Loss: 1.0375833e-05\n",
      "Epoch 17 Iteration 9\n",
      "Loss: 2.4041563e-05\n",
      "Epoch 17 Iteration 10\n",
      "Loss: 3.428964e-05\n",
      "Epoch 17 Iteration 11\n",
      "Loss: 7.4486034e-06\n",
      "Epoch 17 Iteration 12\n",
      "Loss: 3.0423034e-05\n",
      "Epoch 17 Iteration 13\n",
      "Loss: 1.4357755e-05\n",
      "Epoch 17 Iteration 14\n",
      "Loss: 2.0181944e-05\n",
      "Epoch 17 Iteration 15\n",
      "Loss: 1.1493598e-05\n",
      "Epoch 17 Iteration 16\n",
      "Loss: 2.5302188e-05\n",
      "Epoch 17 Iteration 17\n",
      "Loss: 2.2320972e-05\n",
      "Epoch 17 Iteration 18\n",
      "Loss: 3.7306515e-06\n",
      "Epoch 17 Iteration 19\n",
      "Loss: 6.6468756e-06\n",
      "Epoch 17 Iteration 20\n",
      "Loss: 3.200324e-05\n",
      "Epoch 17 Iteration 21\n",
      "Loss: 3.9248476e-05\n",
      "Epoch 17 Iteration 22\n",
      "Loss: 3.553313e-05\n",
      "Epoch 17 Iteration 23\n",
      "Loss: 1.8881621e-05\n",
      "Epoch 17 Iteration 24\n",
      "Loss: 4.000293e-05\n",
      "Epoch 17 Iteration 25\n",
      "Loss: 3.8654758e-05\n",
      "Epoch 17 Iteration 26\n",
      "Loss: 4.067383e-05\n",
      "Epoch 17 Iteration 27\n",
      "Loss: 5.901552e-06\n",
      "Epoch 17 Iteration 28\n",
      "Loss: 2.5377722e-06\n",
      "Epoch 17 Iteration 29\n",
      "Loss: 2.8244808e-06\n",
      "Epoch 17 Iteration 30\n",
      "Loss: 1.5237666e-05\n",
      "Epoch 17 Iteration 31\n",
      "Loss: 3.8028124e-05\n",
      "Epoch 17 Iteration 32\n",
      "Loss: 3.372107e-05\n",
      "Epoch 17 Iteration 33\n",
      "Loss: 1.1748149e-05\n",
      "Epoch 17 Iteration 34\n",
      "Loss: 1.1755954e-05\n",
      "Valid loss: 3.1606257834937423e-05\n",
      "Epoch  18  Current time:  Sun Jun 21 22:15:23 2020\n",
      "Epoch 18 Iteration 0\n",
      "Loss: 2.802692e-05\n",
      "Epoch 18 Iteration 1\n",
      "Loss: 3.170479e-05\n",
      "Epoch 18 Iteration 2\n",
      "Loss: 1.0343456e-05\n",
      "Epoch 18 Iteration 3\n",
      "Loss: 3.55611e-05\n",
      "Epoch 18 Iteration 4\n",
      "Loss: 2.0293086e-05\n",
      "Epoch 18 Iteration 5\n",
      "Loss: 1.37209045e-05\n",
      "Epoch 18 Iteration 6\n",
      "Loss: 6.186178e-05\n",
      "Epoch 18 Iteration 7\n",
      "Loss: 2.6370317e-05\n",
      "Epoch 18 Iteration 8\n",
      "Loss: 2.5653022e-05\n",
      "Epoch 18 Iteration 9\n",
      "Loss: 7.670091e-06\n",
      "Epoch 18 Iteration 10\n",
      "Loss: 3.496279e-05\n",
      "Epoch 18 Iteration 11\n",
      "Loss: 3.7126378e-05\n",
      "Epoch 18 Iteration 12\n",
      "Loss: 4.1342457e-05\n",
      "Epoch 18 Iteration 13\n",
      "Loss: 2.2749966e-05\n",
      "Epoch 18 Iteration 14\n",
      "Loss: 6.47662e-06\n",
      "Epoch 18 Iteration 15\n",
      "Loss: 1.8311446e-05\n",
      "Epoch 18 Iteration 16\n",
      "Loss: 3.860778e-05\n",
      "Epoch 18 Iteration 17\n",
      "Loss: 2.6333728e-06\n",
      "Epoch 18 Iteration 18\n",
      "Loss: 4.0998533e-05\n",
      "Epoch 18 Iteration 19\n",
      "Loss: 1.3692126e-05\n",
      "Epoch 18 Iteration 20\n",
      "Loss: 3.4809225e-05\n",
      "Epoch 18 Iteration 21\n",
      "Loss: 3.1051997e-05\n",
      "Epoch 18 Iteration 22\n",
      "Loss: 3.0440118e-05\n",
      "Epoch 18 Iteration 23\n",
      "Loss: 2.629213e-05\n",
      "Epoch 18 Iteration 24\n",
      "Loss: 6.677521e-06\n",
      "Epoch 18 Iteration 25\n",
      "Loss: 3.8878585e-05\n",
      "Epoch 18 Iteration 26\n",
      "Loss: 3.9197885e-06\n",
      "Epoch 18 Iteration 27\n",
      "Loss: 3.621744e-05\n",
      "Epoch 18 Iteration 28\n",
      "Loss: 1.0570935e-05\n",
      "Epoch 18 Iteration 29\n",
      "Loss: 2.6319449e-06\n",
      "Epoch 18 Iteration 30\n",
      "Loss: 3.248775e-05\n",
      "Epoch 18 Iteration 31\n",
      "Loss: 9.941397e-06\n",
      "Epoch 18 Iteration 32\n",
      "Loss: 1.0803428e-05\n",
      "Epoch 18 Iteration 33\n",
      "Loss: 3.415431e-06\n",
      "Epoch 18 Iteration 34\n",
      "Loss: 4.0048258e-06\n",
      "Valid loss: 2.936638156825211e-05\n",
      "Epoch  19  Current time:  Sun Jun 21 22:16:08 2020\n",
      "Epoch 19 Iteration 0\n",
      "Loss: 3.5882436e-05\n",
      "Epoch 19 Iteration 1\n",
      "Loss: 3.2560474e-05\n",
      "Epoch 19 Iteration 2\n",
      "Loss: 4.0522236e-06\n",
      "Epoch 19 Iteration 3\n",
      "Loss: 6.6891034e-06\n",
      "Epoch 19 Iteration 4\n",
      "Loss: 2.3738718e-05\n",
      "Epoch 19 Iteration 5\n",
      "Loss: 5.8976916e-06\n",
      "Epoch 19 Iteration 6\n",
      "Loss: 1.1360358e-05\n",
      "Epoch 19 Iteration 7\n",
      "Loss: 6.247364e-06\n",
      "Epoch 19 Iteration 8\n",
      "Loss: 2.6341619e-05\n",
      "Epoch 19 Iteration 9\n",
      "Loss: 3.778297e-05\n",
      "Epoch 19 Iteration 10\n",
      "Loss: 2.0531313e-06\n",
      "Epoch 19 Iteration 11\n",
      "Loss: 6.146828e-05\n",
      "Epoch 19 Iteration 12\n",
      "Loss: 3.0350368e-05\n",
      "Epoch 19 Iteration 13\n",
      "Loss: 4.2079067e-05\n",
      "Epoch 19 Iteration 14\n",
      "Loss: 3.816537e-05\n",
      "Epoch 19 Iteration 15\n",
      "Loss: 3.6852547e-05\n",
      "Epoch 19 Iteration 16\n",
      "Loss: 1.3732126e-05\n",
      "Epoch 19 Iteration 17\n",
      "Loss: 2.2108232e-05\n",
      "Epoch 19 Iteration 18\n",
      "Loss: 1.0591916e-05\n",
      "Epoch 19 Iteration 19\n",
      "Loss: 2.2772801e-06\n",
      "Epoch 19 Iteration 20\n",
      "Loss: 1.0327567e-05\n",
      "Epoch 19 Iteration 21\n",
      "Loss: 2.4929184e-05\n",
      "Epoch 19 Iteration 22\n",
      "Loss: 1.09626335e-05\n",
      "Epoch 19 Iteration 23\n",
      "Loss: 4.7742074e-06\n",
      "Epoch 19 Iteration 24\n",
      "Loss: 2.9474825e-06\n",
      "Epoch 19 Iteration 25\n",
      "Loss: 1.4007547e-05\n",
      "Epoch 19 Iteration 26\n",
      "Loss: 3.9839e-05\n",
      "Epoch 19 Iteration 27\n",
      "Loss: 2.117234e-05\n",
      "Epoch 19 Iteration 28\n",
      "Loss: 3.8498212e-05\n",
      "Epoch 19 Iteration 29\n",
      "Loss: 1.7903008e-05\n",
      "Epoch 19 Iteration 30\n",
      "Loss: 3.0054754e-05\n",
      "Epoch 19 Iteration 31\n",
      "Loss: 3.805339e-05\n",
      "Epoch 19 Iteration 32\n",
      "Loss: 2.956093e-05\n",
      "Epoch 19 Iteration 33\n",
      "Loss: 3.219233e-05\n",
      "Epoch 19 Iteration 34\n",
      "Loss: 8.085079e-06\n",
      "Valid loss: 2.9543716664193198e-05\n",
      "Epoch  20  Current time:  Sun Jun 21 22:16:52 2020\n",
      "Epoch 20 Iteration 0\n",
      "Loss: 1.123364e-05\n",
      "Epoch 20 Iteration 1\n",
      "Loss: 6.610804e-06\n",
      "Epoch 20 Iteration 2\n",
      "Loss: 7.375396e-06\n",
      "Epoch 20 Iteration 3\n",
      "Loss: 3.204435e-05\n",
      "Epoch 20 Iteration 4\n",
      "Loss: 1.998672e-05\n",
      "Epoch 20 Iteration 5\n",
      "Loss: 2.0421912e-05\n",
      "Epoch 20 Iteration 6\n",
      "Loss: 3.2432217e-05\n",
      "Epoch 20 Iteration 7\n",
      "Loss: 4.060941e-05\n",
      "Epoch 20 Iteration 8\n",
      "Loss: 4.3866917e-06\n",
      "Epoch 20 Iteration 9\n",
      "Loss: 3.961782e-06\n",
      "Epoch 20 Iteration 10\n",
      "Loss: 3.9849016e-05\n",
      "Epoch 20 Iteration 11\n",
      "Loss: 3.766373e-05\n",
      "Epoch 20 Iteration 12\n",
      "Loss: 3.775464e-06\n",
      "Epoch 20 Iteration 13\n",
      "Loss: 3.5523866e-05\n",
      "Epoch 20 Iteration 14\n",
      "Loss: 6.653516e-06\n",
      "Epoch 20 Iteration 15\n",
      "Loss: 1.4827336e-05\n",
      "Epoch 20 Iteration 16\n",
      "Loss: 4.4546246e-06\n",
      "Epoch 20 Iteration 17\n",
      "Loss: 2.206366e-05\n",
      "Epoch 20 Iteration 18\n",
      "Loss: 1.29985665e-05\n",
      "Epoch 20 Iteration 19\n",
      "Loss: 9.8626715e-06\n",
      "Epoch 20 Iteration 20\n",
      "Loss: 3.8465027e-05\n",
      "Epoch 20 Iteration 21\n",
      "Loss: 2.976177e-05\n",
      "Epoch 20 Iteration 22\n",
      "Loss: 2.3165096e-05\n",
      "Epoch 20 Iteration 23\n",
      "Loss: 2.6415064e-06\n",
      "Epoch 20 Iteration 24\n",
      "Loss: 2.4881168e-05\n",
      "Epoch 20 Iteration 25\n",
      "Loss: 1.0208366e-05\n",
      "Epoch 20 Iteration 26\n",
      "Loss: 3.1591233e-05\n",
      "Epoch 20 Iteration 27\n",
      "Loss: 3.077998e-05\n",
      "Epoch 20 Iteration 28\n",
      "Loss: 2.3058308e-05\n",
      "Epoch 20 Iteration 29\n",
      "Loss: 5.938948e-05\n",
      "Epoch 20 Iteration 30\n",
      "Loss: 5.810163e-06\n",
      "Epoch 20 Iteration 31\n",
      "Loss: 3.9330967e-05\n",
      "Epoch 20 Iteration 32\n",
      "Loss: 2.8587447e-05\n",
      "Epoch 20 Iteration 33\n",
      "Loss: 3.585664e-05\n",
      "Epoch 20 Iteration 34\n",
      "Loss: 1.04265855e-05\n",
      "Valid loss: 2.8655606001848355e-05\n",
      "Epoch  21  Current time:  Sun Jun 21 22:17:36 2020\n",
      "Epoch 21 Iteration 0\n",
      "Loss: 2.8036848e-05\n",
      "Epoch 21 Iteration 1\n",
      "Loss: 5.69623e-06\n",
      "Epoch 21 Iteration 2\n",
      "Loss: 3.6826757e-05\n",
      "Epoch 21 Iteration 3\n",
      "Loss: 3.7509188e-05\n",
      "Epoch 21 Iteration 4\n",
      "Loss: 1.8541346e-05\n",
      "Epoch 21 Iteration 5\n",
      "Loss: 3.077909e-05\n",
      "Epoch 21 Iteration 6\n",
      "Loss: 2.2947308e-05\n",
      "Epoch 21 Iteration 7\n",
      "Loss: 2.5944116e-06\n",
      "Epoch 21 Iteration 8\n",
      "Loss: 6.950125e-06\n",
      "Epoch 21 Iteration 9\n",
      "Loss: 1.5572108e-05\n",
      "Epoch 21 Iteration 10\n",
      "Loss: 2.373719e-05\n",
      "Epoch 21 Iteration 11\n",
      "Loss: 3.9703467e-05\n",
      "Epoch 21 Iteration 12\n",
      "Loss: 6.234032e-06\n",
      "Epoch 21 Iteration 13\n",
      "Loss: 1.2487901e-05\n",
      "Epoch 21 Iteration 14\n",
      "Loss: 9.985665e-06\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 21 Iteration 15\n",
      "Loss: 2.992559e-06\n",
      "Epoch 21 Iteration 16\n",
      "Loss: 3.7920538e-05\n",
      "Epoch 21 Iteration 17\n",
      "Loss: 4.011027e-05\n",
      "Epoch 21 Iteration 18\n",
      "Loss: 3.221383e-05\n",
      "Epoch 21 Iteration 19\n",
      "Loss: 2.2417578e-05\n",
      "Epoch 21 Iteration 20\n",
      "Loss: 9.875181e-06\n",
      "Epoch 21 Iteration 21\n",
      "Loss: 1.1597042e-05\n",
      "Epoch 21 Iteration 22\n",
      "Loss: 3.1555814e-05\n",
      "Epoch 21 Iteration 23\n",
      "Loss: 7.0122082e-06\n",
      "Epoch 21 Iteration 24\n",
      "Loss: 1.1099156e-05\n",
      "Epoch 21 Iteration 25\n",
      "Loss: 3.5618035e-05\n",
      "Epoch 21 Iteration 26\n",
      "Loss: 2.9052642e-05\n",
      "Epoch 21 Iteration 27\n",
      "Loss: 3.970531e-06\n",
      "Epoch 21 Iteration 28\n",
      "Loss: 5.8173864e-05\n",
      "Epoch 21 Iteration 29\n",
      "Loss: 3.2388375e-06\n",
      "Epoch 21 Iteration 30\n",
      "Loss: 2.3787596e-05\n",
      "Epoch 21 Iteration 31\n",
      "Loss: 4.226598e-06\n",
      "Epoch 21 Iteration 32\n",
      "Loss: 1.9526295e-05\n",
      "Epoch 21 Iteration 33\n",
      "Loss: 3.865869e-05\n",
      "Epoch 21 Iteration 34\n",
      "Loss: 2.8514645e-05\n",
      "Valid loss: 2.7286116164759733e-05\n",
      "Epoch  22  Current time:  Sun Jun 21 22:18:21 2020\n",
      "Epoch 22 Iteration 0\n",
      "Loss: 2.8978298e-05\n",
      "Epoch 22 Iteration 1\n",
      "Loss: 6.2607096e-06\n",
      "Epoch 22 Iteration 2\n",
      "Loss: 3.4856246e-05\n",
      "Epoch 22 Iteration 3\n",
      "Loss: 4.1056785e-05\n",
      "Epoch 22 Iteration 4\n",
      "Loss: 6.7653277e-06\n",
      "Epoch 22 Iteration 5\n",
      "Loss: 2.1783384e-05\n",
      "Epoch 22 Iteration 6\n",
      "Loss: 9.8074115e-06\n",
      "Epoch 22 Iteration 7\n",
      "Loss: 4.2139714e-06\n",
      "Epoch 22 Iteration 8\n",
      "Loss: 2.9378898e-05\n",
      "Epoch 22 Iteration 9\n",
      "Loss: 6.2798304e-06\n",
      "Epoch 22 Iteration 10\n",
      "Loss: 3.8246544e-05\n",
      "Epoch 22 Iteration 11\n",
      "Loss: 2.6134945e-05\n",
      "Epoch 22 Iteration 12\n",
      "Loss: 6.298877e-05\n",
      "Epoch 22 Iteration 13\n",
      "Loss: 2.3651159e-05\n",
      "Epoch 22 Iteration 14\n",
      "Loss: 7.132893e-06\n",
      "Epoch 22 Iteration 15\n",
      "Loss: 1.858206e-05\n",
      "Epoch 22 Iteration 16\n",
      "Loss: 1.4151436e-05\n",
      "Epoch 22 Iteration 17\n",
      "Loss: 1.979357e-05\n",
      "Epoch 22 Iteration 18\n",
      "Loss: 4.5143206e-06\n",
      "Epoch 22 Iteration 19\n",
      "Loss: 1.1139946e-05\n",
      "Epoch 22 Iteration 20\n",
      "Loss: 1.1560829e-05\n",
      "Epoch 22 Iteration 21\n",
      "Loss: 1.2433074e-05\n",
      "Epoch 22 Iteration 22\n",
      "Loss: 9.948331e-06\n",
      "Epoch 22 Iteration 23\n",
      "Loss: 2.850634e-05\n",
      "Epoch 22 Iteration 24\n",
      "Loss: 3.4969802e-05\n",
      "Epoch 22 Iteration 25\n",
      "Loss: 2.9151565e-06\n",
      "Epoch 22 Iteration 26\n",
      "Loss: 2.3157867e-05\n",
      "Epoch 22 Iteration 27\n",
      "Loss: 2.1035714e-06\n",
      "Epoch 22 Iteration 28\n",
      "Loss: 3.9214126e-05\n",
      "Epoch 22 Iteration 29\n",
      "Loss: 3.5226825e-05\n",
      "Epoch 22 Iteration 30\n",
      "Loss: 2.5714633e-06\n",
      "Epoch 22 Iteration 31\n",
      "Loss: 2.9012248e-05\n",
      "Epoch 22 Iteration 32\n",
      "Loss: 3.8131602e-05\n",
      "Epoch 22 Iteration 33\n",
      "Loss: 2.9929513e-05\n",
      "Epoch 22 Iteration 34\n",
      "Loss: 2.9010436e-05\n",
      "Valid loss: 2.7189773391000926e-05\n",
      "Epoch  23  Current time:  Sun Jun 21 22:19:05 2020\n",
      "Epoch 23 Iteration 0\n",
      "Loss: 1.2049513e-05\n",
      "Epoch 23 Iteration 1\n",
      "Loss: 2.2847871e-05\n",
      "Epoch 23 Iteration 2\n",
      "Loss: 7.2913053e-06\n",
      "Epoch 23 Iteration 3\n",
      "Loss: 4.6183817e-05\n",
      "Epoch 23 Iteration 4\n",
      "Loss: 2.6560965e-06\n",
      "Epoch 23 Iteration 5\n",
      "Loss: 1.0802614e-05\n",
      "Epoch 23 Iteration 6\n",
      "Loss: 3.8393428e-05\n",
      "Epoch 23 Iteration 7\n",
      "Loss: 3.2577864e-05\n",
      "Epoch 23 Iteration 8\n",
      "Loss: 3.997423e-05\n",
      "Epoch 23 Iteration 9\n",
      "Loss: 3.7119604e-05\n",
      "Epoch 23 Iteration 10\n",
      "Loss: 6.5285676e-05\n",
      "Epoch 23 Iteration 11\n",
      "Loss: 3.5598663e-05\n",
      "Epoch 23 Iteration 12\n",
      "Loss: 8.566607e-06\n",
      "Epoch 23 Iteration 13\n",
      "Loss: 6.417891e-06\n",
      "Epoch 23 Iteration 14\n",
      "Loss: 2.5580359e-05\n",
      "Epoch 23 Iteration 15\n",
      "Loss: 1.8073337e-05\n",
      "Epoch 23 Iteration 16\n",
      "Loss: 2.4896495e-05\n",
      "Epoch 23 Iteration 17\n",
      "Loss: 4.5982297e-06\n",
      "Epoch 23 Iteration 18\n",
      "Loss: 3.8251277e-05\n",
      "Epoch 23 Iteration 19\n",
      "Loss: 3.062445e-05\n",
      "Epoch 23 Iteration 20\n",
      "Loss: 1.922964e-05\n",
      "Epoch 23 Iteration 21\n",
      "Loss: 1.3252242e-05\n",
      "Epoch 23 Iteration 22\n",
      "Loss: 4.0443297e-06\n",
      "Epoch 23 Iteration 23\n",
      "Loss: 1.0544715e-05\n",
      "Epoch 23 Iteration 24\n",
      "Loss: 1.1464684e-05\n",
      "Epoch 23 Iteration 25\n",
      "Loss: 3.5173025e-06\n",
      "Epoch 23 Iteration 26\n",
      "Loss: 5.7623524e-06\n",
      "Epoch 23 Iteration 27\n",
      "Loss: 2.9575902e-05\n",
      "Epoch 23 Iteration 28\n",
      "Loss: 3.7973776e-05\n",
      "Epoch 23 Iteration 29\n",
      "Loss: 2.9018229e-05\n",
      "Epoch 23 Iteration 30\n",
      "Loss: 1.4731091e-05\n",
      "Epoch 23 Iteration 31\n",
      "Loss: 2.9006407e-05\n",
      "Epoch 23 Iteration 32\n",
      "Loss: 3.540257e-06\n",
      "Epoch 23 Iteration 33\n",
      "Loss: 2.9819785e-05\n",
      "Epoch 23 Iteration 34\n",
      "Loss: 2.2014481e-05\n",
      "Valid loss: 2.7587422664510086e-05\n",
      "Epoch  24  Current time:  Sun Jun 21 22:19:50 2020\n",
      "Epoch 24 Iteration 0\n",
      "Loss: 7.2146277e-06\n",
      "Epoch 24 Iteration 1\n",
      "Loss: 2.46274e-06\n",
      "Epoch 24 Iteration 2\n",
      "Loss: 2.5168876e-05\n",
      "Epoch 24 Iteration 3\n",
      "Loss: 9.804449e-06\n",
      "Epoch 24 Iteration 4\n",
      "Loss: 1.1589447e-05\n",
      "Epoch 24 Iteration 5\n",
      "Loss: 1.2661685e-05\n",
      "Epoch 24 Iteration 6\n",
      "Loss: 3.606905e-05\n",
      "Epoch 24 Iteration 7\n",
      "Loss: 1.3739783e-05\n",
      "Epoch 24 Iteration 8\n",
      "Loss: 3.4309516e-05\n",
      "Epoch 24 Iteration 9\n",
      "Loss: 5.8170008e-05\n",
      "Epoch 24 Iteration 10\n",
      "Loss: 3.909012e-05\n",
      "Epoch 24 Iteration 11\n",
      "Loss: 2.9672823e-05\n",
      "Epoch 24 Iteration 12\n",
      "Loss: 3.704557e-05\n",
      "Epoch 24 Iteration 13\n",
      "Loss: 7.0493406e-06\n",
      "Epoch 24 Iteration 14\n",
      "Loss: 3.5994704e-05\n",
      "Epoch 24 Iteration 15\n",
      "Loss: 2.1731541e-05\n",
      "Epoch 24 Iteration 16\n",
      "Loss: 4.4959884e-06\n",
      "Epoch 24 Iteration 17\n",
      "Loss: 2.8307912e-05\n",
      "Epoch 24 Iteration 18\n",
      "Loss: 2.724363e-05\n",
      "Epoch 24 Iteration 19\n",
      "Loss: 2.4207286e-05\n",
      "Epoch 24 Iteration 20\n",
      "Loss: 1.23275195e-05\n",
      "Epoch 24 Iteration 21\n",
      "Loss: 1.9499936e-05\n",
      "Epoch 24 Iteration 22\n",
      "Loss: 2.940036e-05\n",
      "Epoch 24 Iteration 23\n",
      "Loss: 1.0825816e-05\n",
      "Epoch 24 Iteration 24\n",
      "Loss: 7.196554e-06\n",
      "Epoch 24 Iteration 25\n",
      "Loss: 3.2555038e-06\n",
      "Epoch 24 Iteration 26\n",
      "Loss: 2.8158395e-06\n",
      "Epoch 24 Iteration 27\n",
      "Loss: 3.2791522e-05\n",
      "Epoch 24 Iteration 28\n",
      "Loss: 1.8497332e-05\n",
      "Epoch 24 Iteration 29\n",
      "Loss: 2.5023472e-05\n",
      "Epoch 24 Iteration 30\n",
      "Loss: 3.9288607e-06\n",
      "Epoch 24 Iteration 31\n",
      "Loss: 3.3562697e-05\n",
      "Epoch 24 Iteration 32\n",
      "Loss: 3.8238773e-05\n",
      "Epoch 24 Iteration 33\n",
      "Loss: 6.9668436e-06\n",
      "Epoch 24 Iteration 34\n",
      "Loss: 3.907955e-05\n",
      "Valid loss: 2.755498280748725e-05\n",
      "Epoch  25  Current time:  Sun Jun 21 22:20:34 2020\n",
      "Epoch 25 Iteration 0\n",
      "Loss: 3.7105237e-05\n",
      "Epoch 25 Iteration 1\n",
      "Loss: 1.4774385e-05\n",
      "Epoch 25 Iteration 2\n",
      "Loss: 3.0439192e-05\n",
      "Epoch 25 Iteration 3\n",
      "Loss: 1.0998097e-05\n",
      "Epoch 25 Iteration 4\n",
      "Loss: 5.488291e-06\n",
      "Epoch 25 Iteration 5\n",
      "Loss: 2.2097438e-05\n",
      "Epoch 25 Iteration 6\n",
      "Loss: 3.73542e-05\n",
      "Epoch 25 Iteration 7\n",
      "Loss: 9.747353e-06\n",
      "Epoch 25 Iteration 8\n",
      "Loss: 3.657627e-05\n",
      "Epoch 25 Iteration 9\n",
      "Loss: 2.7341126e-05\n",
      "Epoch 25 Iteration 10\n",
      "Loss: 1.096248e-05\n",
      "Epoch 25 Iteration 11\n",
      "Loss: 3.580202e-05\n",
      "Epoch 25 Iteration 12\n",
      "Loss: 3.9882638e-05\n",
      "Epoch 25 Iteration 13\n",
      "Loss: 9.395668e-06\n",
      "Epoch 25 Iteration 14\n",
      "Loss: 1.3280305e-05\n",
      "Epoch 25 Iteration 15\n",
      "Loss: 2.9301431e-05\n",
      "Epoch 25 Iteration 16\n",
      "Loss: 3.599043e-05\n",
      "Epoch 25 Iteration 17\n",
      "Loss: 2.2368553e-05\n",
      "Epoch 25 Iteration 18\n",
      "Loss: 6.364139e-06\n",
      "Epoch 25 Iteration 19\n",
      "Loss: 3.8551284e-06\n",
      "Epoch 25 Iteration 20\n",
      "Loss: 3.176044e-05\n",
      "Epoch 25 Iteration 21\n",
      "Loss: 4.000133e-05\n",
      "Epoch 25 Iteration 22\n",
      "Loss: 5.1695397e-06\n",
      "Epoch 25 Iteration 23\n",
      "Loss: 2.4341112e-05\n",
      "Epoch 25 Iteration 24\n",
      "Loss: 6.6999e-06\n",
      "Epoch 25 Iteration 25\n",
      "Loss: 4.526714e-06\n",
      "Epoch 25 Iteration 26\n",
      "Loss: 3.0364403e-05\n",
      "Epoch 25 Iteration 27\n",
      "Loss: 2.7936821e-05\n",
      "Epoch 25 Iteration 28\n",
      "Loss: 2.4369525e-05\n",
      "Epoch 25 Iteration 29\n",
      "Loss: 5.6674566e-05\n",
      "Epoch 25 Iteration 30\n",
      "Loss: 1.9593374e-05\n",
      "Epoch 25 Iteration 31\n",
      "Loss: 2.1447495e-05\n",
      "Epoch 25 Iteration 32\n",
      "Loss: 1.1244806e-05\n",
      "Epoch 25 Iteration 33\n",
      "Loss: 8.621642e-06\n",
      "Epoch 25 Iteration 34\n",
      "Loss: 7.576564e-06\n",
      "Valid loss: 2.7832171326735988e-05\n",
      "Epoch  26  Current time:  Sun Jun 21 22:21:18 2020\n",
      "Epoch 26 Iteration 0\n",
      "Loss: 3.730199e-05\n",
      "Epoch 26 Iteration 1\n",
      "Loss: 1.4798332e-05\n",
      "Epoch 26 Iteration 2\n",
      "Loss: 2.9972267e-05\n",
      "Epoch 26 Iteration 3\n",
      "Loss: 5.6521865e-05\n",
      "Epoch 26 Iteration 4\n",
      "Loss: 6.941589e-06\n",
      "Epoch 26 Iteration 5\n",
      "Loss: 1.2146748e-05\n",
      "Epoch 26 Iteration 6\n",
      "Loss: 1.948358e-05\n",
      "Epoch 26 Iteration 7\n",
      "Loss: 2.3049315e-05\n",
      "Epoch 26 Iteration 8\n",
      "Loss: 4.258475e-06\n",
      "Epoch 26 Iteration 9\n",
      "Loss: 3.1936743e-05\n",
      "Epoch 26 Iteration 10\n",
      "Loss: 6.731175e-06\n",
      "Epoch 26 Iteration 11\n",
      "Loss: 2.2015623e-05\n",
      "Epoch 26 Iteration 12\n",
      "Loss: 4.750839e-05\n",
      "Epoch 26 Iteration 13\n",
      "Loss: 3.2901382e-06\n",
      "Epoch 26 Iteration 14\n",
      "Loss: 2.8583378e-05\n",
      "Epoch 26 Iteration 15\n",
      "Loss: 2.8968941e-05\n",
      "Epoch 26 Iteration 16\n",
      "Loss: 1.0396827e-05\n",
      "Epoch 26 Iteration 17\n",
      "Loss: 3.0690917e-05\n",
      "Epoch 26 Iteration 18\n",
      "Loss: 9.555522e-06\n",
      "Epoch 26 Iteration 19\n",
      "Loss: 1.8114483e-06\n",
      "Epoch 26 Iteration 20\n",
      "Loss: 3.322621e-05\n",
      "Epoch 26 Iteration 21\n",
      "Loss: 3.8458853e-05\n",
      "Epoch 26 Iteration 22\n",
      "Loss: 2.2573251e-05\n",
      "Epoch 26 Iteration 23\n",
      "Loss: 1.0138964e-05\n",
      "Epoch 26 Iteration 24\n",
      "Loss: 4.0550716e-05\n",
      "Epoch 26 Iteration 25\n",
      "Loss: 3.8544877e-06\n",
      "Epoch 26 Iteration 26\n",
      "Loss: 6.3241027e-06\n",
      "Epoch 26 Iteration 27\n",
      "Loss: 2.9001318e-05\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 26 Iteration 28\n",
      "Loss: 1.1144522e-05\n",
      "Epoch 26 Iteration 29\n",
      "Loss: 2.4041667e-05\n",
      "Epoch 26 Iteration 30\n",
      "Loss: 5.8803316e-06\n",
      "Epoch 26 Iteration 31\n",
      "Loss: 3.9185936e-05\n",
      "Epoch 26 Iteration 32\n",
      "Loss: 1.8201887e-05\n",
      "Epoch 26 Iteration 33\n",
      "Loss: 3.5547335e-05\n",
      "Epoch 26 Iteration 34\n",
      "Loss: 2.8769803e-06\n",
      "Valid loss: 2.7732199669117108e-05\n",
      "Epoch  27  Current time:  Sun Jun 21 22:22:03 2020\n",
      "Epoch 27 Iteration 0\n",
      "Loss: 1.1385543e-05\n",
      "Epoch 27 Iteration 1\n",
      "Loss: 7.5745274e-06\n",
      "Epoch 27 Iteration 2\n",
      "Loss: 3.5440912e-06\n",
      "Epoch 27 Iteration 3\n",
      "Loss: 2.913473e-05\n",
      "Epoch 27 Iteration 4\n",
      "Loss: 3.4525765e-05\n",
      "Epoch 27 Iteration 5\n",
      "Loss: 3.85431e-06\n",
      "Epoch 27 Iteration 6\n",
      "Loss: 2.926867e-05\n",
      "Epoch 27 Iteration 7\n",
      "Loss: 2.8231401e-05\n",
      "Epoch 27 Iteration 8\n",
      "Loss: 2.1229966e-06\n",
      "Epoch 27 Iteration 9\n",
      "Loss: 1.0262441e-05\n",
      "Epoch 27 Iteration 10\n",
      "Loss: 3.7773298e-05\n",
      "Epoch 27 Iteration 11\n",
      "Loss: 5.5457305e-05\n",
      "Epoch 27 Iteration 12\n",
      "Loss: 1.3955083e-05\n",
      "Epoch 27 Iteration 13\n",
      "Loss: 1.0800385e-05\n",
      "Epoch 27 Iteration 14\n",
      "Loss: 9.485699e-06\n",
      "Epoch 27 Iteration 15\n",
      "Loss: 1.7784501e-05\n",
      "Epoch 27 Iteration 16\n",
      "Loss: 2.978419e-05\n",
      "Epoch 27 Iteration 17\n",
      "Loss: 3.6962792e-05\n",
      "Epoch 27 Iteration 18\n",
      "Loss: 2.0806407e-05\n",
      "Epoch 27 Iteration 19\n",
      "Loss: 1.201205e-05\n",
      "Epoch 27 Iteration 20\n",
      "Loss: 6.897693e-06\n",
      "Epoch 27 Iteration 21\n",
      "Loss: 2.7507272e-06\n",
      "Epoch 27 Iteration 22\n",
      "Loss: 2.7480206e-05\n",
      "Epoch 27 Iteration 23\n",
      "Loss: 2.7185173e-05\n",
      "Epoch 27 Iteration 24\n",
      "Loss: 3.5745754e-05\n",
      "Epoch 27 Iteration 25\n",
      "Loss: 3.1495452e-05\n",
      "Epoch 27 Iteration 26\n",
      "Loss: 3.35453e-05\n",
      "Epoch 27 Iteration 27\n",
      "Loss: 1.8322302e-05\n",
      "Epoch 27 Iteration 28\n",
      "Loss: 8.108017e-06\n",
      "Epoch 27 Iteration 29\n",
      "Loss: 2.1805185e-05\n",
      "Epoch 27 Iteration 30\n",
      "Loss: 2.0450323e-05\n",
      "Epoch 27 Iteration 31\n",
      "Loss: 5.674876e-06\n",
      "Epoch 27 Iteration 32\n",
      "Loss: 7.4482946e-06\n",
      "Epoch 27 Iteration 33\n",
      "Loss: 2.214545e-05\n",
      "Epoch 27 Iteration 34\n",
      "Loss: 3.8272898e-05\n",
      "Valid loss: 2.6917789000435732e-05\n",
      "Epoch  28  Current time:  Sun Jun 21 22:22:47 2020\n",
      "Epoch 28 Iteration 0\n",
      "Loss: 2.8895083e-05\n",
      "Epoch 28 Iteration 1\n",
      "Loss: 2.9411405e-05\n",
      "Epoch 28 Iteration 2\n",
      "Loss: 1.7895689e-05\n",
      "Epoch 28 Iteration 3\n",
      "Loss: 3.932136e-05\n",
      "Epoch 28 Iteration 4\n",
      "Loss: 2.364036e-05\n",
      "Epoch 28 Iteration 5\n",
      "Loss: 1.14277545e-05\n",
      "Epoch 28 Iteration 6\n",
      "Loss: 1.861992e-05\n",
      "Epoch 28 Iteration 7\n",
      "Loss: 2.1758937e-05\n",
      "Epoch 28 Iteration 8\n",
      "Loss: 2.682396e-05\n",
      "Epoch 28 Iteration 9\n",
      "Loss: 3.391478e-05\n",
      "Epoch 28 Iteration 10\n",
      "Loss: 8.193215e-06\n",
      "Epoch 28 Iteration 11\n",
      "Loss: 1.2289822e-05\n",
      "Epoch 28 Iteration 12\n",
      "Loss: 2.977093e-05\n",
      "Epoch 28 Iteration 13\n",
      "Loss: 2.1641135e-05\n",
      "Epoch 28 Iteration 14\n",
      "Loss: 3.807551e-05\n",
      "Epoch 28 Iteration 15\n",
      "Loss: 3.7476625e-06\n",
      "Epoch 28 Iteration 16\n",
      "Loss: 2.0737818e-06\n",
      "Epoch 28 Iteration 17\n",
      "Loss: 9.963853e-06\n",
      "Epoch 28 Iteration 18\n",
      "Loss: 3.4350964e-05\n",
      "Epoch 28 Iteration 19\n",
      "Loss: 1.2487462e-05\n",
      "Epoch 28 Iteration 20\n",
      "Loss: 3.5892913e-06\n",
      "Epoch 28 Iteration 21\n",
      "Loss: 2.7901711e-05\n",
      "Epoch 28 Iteration 22\n",
      "Loss: 2.2431514e-06\n",
      "Epoch 28 Iteration 23\n",
      "Loss: 2.8451534e-05\n",
      "Epoch 28 Iteration 24\n",
      "Loss: 3.213233e-05\n",
      "Epoch 28 Iteration 25\n",
      "Loss: 6.9017838e-06\n",
      "Epoch 28 Iteration 26\n",
      "Loss: 6.6466678e-06\n",
      "Epoch 28 Iteration 27\n",
      "Loss: 4.5128327e-06\n",
      "Epoch 28 Iteration 28\n",
      "Loss: 6.7136516e-06\n",
      "Epoch 28 Iteration 29\n",
      "Loss: 1.1657081e-05\n",
      "Epoch 28 Iteration 30\n",
      "Loss: 3.5611e-05\n",
      "Epoch 28 Iteration 31\n",
      "Loss: 1.40448365e-05\n",
      "Epoch 28 Iteration 32\n",
      "Loss: 5.824541e-05\n",
      "Epoch 28 Iteration 33\n",
      "Loss: 3.7106296e-05\n",
      "Epoch 28 Iteration 34\n",
      "Loss: 2.1736196e-05\n",
      "Valid loss: 2.650687747518532e-05\n",
      "Epoch  29  Current time:  Sun Jun 21 22:23:32 2020\n",
      "Epoch 29 Iteration 0\n",
      "Loss: 3.4281693e-05\n",
      "Epoch 29 Iteration 1\n",
      "Loss: 2.1412516e-06\n",
      "Epoch 29 Iteration 2\n",
      "Loss: 2.2180073e-05\n",
      "Epoch 29 Iteration 3\n",
      "Loss: 5.3622396e-05\n",
      "Epoch 29 Iteration 4\n",
      "Loss: 3.9207353e-05\n",
      "Epoch 29 Iteration 5\n",
      "Loss: 2.712842e-06\n",
      "Epoch 29 Iteration 6\n",
      "Loss: 1.0653595e-05\n",
      "Epoch 29 Iteration 7\n",
      "Loss: 3.6805606e-05\n",
      "Epoch 29 Iteration 8\n",
      "Loss: 1.1845643e-05\n",
      "Epoch 29 Iteration 9\n",
      "Loss: 8.050381e-06\n",
      "Epoch 29 Iteration 10\n",
      "Loss: 6.258353e-06\n",
      "Epoch 29 Iteration 11\n",
      "Loss: 2.5735935e-05\n",
      "Epoch 29 Iteration 12\n",
      "Loss: 3.8755247e-06\n",
      "Epoch 29 Iteration 13\n",
      "Loss: 1.0943359e-05\n",
      "Epoch 29 Iteration 14\n",
      "Loss: 3.046549e-05\n",
      "Epoch 29 Iteration 15\n",
      "Loss: 1.4392846e-05\n",
      "Epoch 29 Iteration 16\n",
      "Loss: 1.9026334e-05\n",
      "Epoch 29 Iteration 17\n",
      "Loss: 2.8157541e-05\n",
      "Epoch 29 Iteration 18\n",
      "Loss: 1.109336e-05\n",
      "Epoch 29 Iteration 19\n",
      "Loss: 6.570853e-06\n",
      "Epoch 29 Iteration 20\n",
      "Loss: 1.8498333e-05\n",
      "Epoch 29 Iteration 21\n",
      "Loss: 3.4886478e-05\n",
      "Epoch 29 Iteration 22\n",
      "Loss: 6.72459e-06\n",
      "Epoch 29 Iteration 23\n",
      "Loss: 2.093574e-05\n",
      "Epoch 29 Iteration 24\n",
      "Loss: 1.1859847e-05\n",
      "Epoch 29 Iteration 25\n",
      "Loss: 3.5471345e-05\n",
      "Epoch 29 Iteration 26\n",
      "Loss: 3.1480165e-05\n",
      "Epoch 29 Iteration 27\n",
      "Loss: 2.7905593e-05\n",
      "Epoch 29 Iteration 28\n",
      "Loss: 2.169939e-05\n",
      "Epoch 29 Iteration 29\n",
      "Loss: 3.840339e-05\n",
      "Epoch 29 Iteration 30\n",
      "Loss: 2.7221979e-05\n",
      "Epoch 29 Iteration 31\n",
      "Loss: 3.284294e-06\n",
      "Epoch 29 Iteration 32\n",
      "Loss: 2.6460948e-05\n",
      "Epoch 29 Iteration 33\n",
      "Loss: 3.4587408e-06\n",
      "Epoch 29 Iteration 34\n",
      "Loss: 2.7939797e-05\n",
      "Valid loss: 2.6126299417228438e-05\n",
      "Epoch  30  Current time:  Sun Jun 21 22:24:16 2020\n",
      "Epoch 30 Iteration 0\n",
      "Loss: 2.6361751e-05\n",
      "Epoch 30 Iteration 1\n",
      "Loss: 2.0814154e-05\n",
      "Epoch 30 Iteration 2\n",
      "Loss: 1.2141417e-05\n",
      "Epoch 30 Iteration 3\n",
      "Loss: 2.3515704e-06\n",
      "Epoch 30 Iteration 4\n",
      "Loss: 2.5702893e-05\n",
      "Epoch 30 Iteration 5\n",
      "Loss: 7.118866e-06\n",
      "Epoch 30 Iteration 6\n",
      "Loss: 4.7731955e-06\n",
      "Epoch 30 Iteration 7\n",
      "Loss: 2.7389868e-05\n",
      "Epoch 30 Iteration 8\n",
      "Loss: 1.8277506e-05\n",
      "Epoch 30 Iteration 9\n",
      "Loss: 3.4719003e-05\n",
      "Epoch 30 Iteration 10\n",
      "Loss: 3.4262637e-05\n",
      "Epoch 30 Iteration 11\n",
      "Loss: 1.2029739e-05\n",
      "Epoch 30 Iteration 12\n",
      "Loss: 2.7472117e-05\n",
      "Epoch 30 Iteration 13\n",
      "Loss: 1.3387183e-05\n",
      "Epoch 30 Iteration 14\n",
      "Loss: 6.7565456e-06\n",
      "Epoch 30 Iteration 15\n",
      "Loss: 3.5265102e-06\n",
      "Epoch 30 Iteration 16\n",
      "Loss: 9.699608e-06\n",
      "Epoch 30 Iteration 17\n",
      "Loss: 9.300662e-06\n",
      "Epoch 30 Iteration 18\n",
      "Loss: 2.3951969e-05\n",
      "Epoch 30 Iteration 19\n",
      "Loss: 2.764195e-05\n",
      "Epoch 30 Iteration 20\n",
      "Loss: 1.9197323e-05\n",
      "Epoch 30 Iteration 21\n",
      "Loss: 3.8131082e-05\n",
      "Epoch 30 Iteration 22\n",
      "Loss: 3.0979663e-05\n",
      "Epoch 30 Iteration 23\n",
      "Loss: 2.8335775e-05\n",
      "Epoch 30 Iteration 24\n",
      "Loss: 2.1364478e-05\n",
      "Epoch 30 Iteration 25\n",
      "Loss: 5.3322503e-05\n",
      "Epoch 30 Iteration 26\n",
      "Loss: 3.6338308e-05\n",
      "Epoch 30 Iteration 27\n",
      "Loss: 7.7854065e-06\n",
      "Epoch 30 Iteration 28\n",
      "Loss: 1.0472136e-05\n",
      "Epoch 30 Iteration 29\n",
      "Loss: 2.0360943e-05\n",
      "Epoch 30 Iteration 30\n",
      "Loss: 3.2633714e-05\n",
      "Epoch 30 Iteration 31\n",
      "Loss: 4.0207005e-06\n",
      "Epoch 30 Iteration 32\n",
      "Loss: 1.9709692e-06\n",
      "Epoch 30 Iteration 33\n",
      "Loss: 7.030178e-06\n",
      "Epoch 30 Iteration 34\n",
      "Loss: 3.6765938e-05\n",
      "Valid loss: 2.7315280021866783e-05\n",
      "Epoch  31  Current time:  Sun Jun 21 22:25:00 2020\n",
      "Epoch 31 Iteration 0\n",
      "Loss: 3.3810855e-05\n",
      "Epoch 31 Iteration 1\n",
      "Loss: 3.0418837e-06\n",
      "Epoch 31 Iteration 2\n",
      "Loss: 2.7335504e-05\n",
      "Epoch 31 Iteration 3\n",
      "Loss: 1.1411257e-05\n",
      "Epoch 31 Iteration 4\n",
      "Loss: 3.4097643e-06\n",
      "Epoch 31 Iteration 5\n",
      "Loss: 3.5632132e-05\n",
      "Epoch 31 Iteration 6\n",
      "Loss: 2.270482e-05\n",
      "Epoch 31 Iteration 7\n",
      "Loss: 2.7450242e-06\n",
      "Epoch 31 Iteration 8\n",
      "Loss: 1.0264068e-05\n",
      "Epoch 31 Iteration 9\n",
      "Loss: 3.097951e-05\n",
      "Epoch 31 Iteration 10\n",
      "Loss: 1.5728188e-05\n",
      "Epoch 31 Iteration 11\n",
      "Loss: 3.823361e-05\n",
      "Epoch 31 Iteration 12\n",
      "Loss: 1.1787811e-05\n",
      "Epoch 31 Iteration 13\n",
      "Loss: 1.2036057e-05\n",
      "Epoch 31 Iteration 14\n",
      "Loss: 2.7035705e-05\n",
      "Epoch 31 Iteration 15\n",
      "Loss: 1.9454277e-05\n",
      "Epoch 31 Iteration 16\n",
      "Loss: 2.1514406e-05\n",
      "Epoch 31 Iteration 17\n",
      "Loss: 1.1092133e-05\n",
      "Epoch 31 Iteration 18\n",
      "Loss: 2.1102689e-05\n",
      "Epoch 31 Iteration 19\n",
      "Loss: 3.526859e-05\n",
      "Epoch 31 Iteration 20\n",
      "Loss: 1.759404e-05\n",
      "Epoch 31 Iteration 21\n",
      "Loss: 3.3538818e-05\n",
      "Epoch 31 Iteration 22\n",
      "Loss: 3.5707385e-06\n",
      "Epoch 31 Iteration 23\n",
      "Loss: 6.1597325e-06\n",
      "Epoch 31 Iteration 24\n",
      "Loss: 2.7580667e-05\n",
      "Epoch 31 Iteration 25\n",
      "Loss: 2.519971e-05\n",
      "Epoch 31 Iteration 26\n",
      "Loss: 7.952297e-06\n",
      "Epoch 31 Iteration 27\n",
      "Loss: 2.5991614e-05\n",
      "Epoch 31 Iteration 28\n",
      "Loss: 7.047321e-06\n",
      "Epoch 31 Iteration 29\n",
      "Loss: 5.606089e-05\n",
      "Epoch 31 Iteration 30\n",
      "Loss: 2.7039989e-05\n",
      "Epoch 31 Iteration 31\n",
      "Loss: 8.550827e-06\n",
      "Epoch 31 Iteration 32\n",
      "Loss: 3.5993384e-05\n",
      "Epoch 31 Iteration 33\n",
      "Loss: 2.1405582e-05\n",
      "Epoch 31 Iteration 34\n",
      "Loss: 4.1669473e-06\n",
      "Valid loss: 2.5810188162722625e-05\n",
      "Epoch  32  Current time:  Sun Jun 21 22:25:45 2020\n",
      "Epoch 32 Iteration 0\n",
      "Loss: 1.19074175e-05\n",
      "Epoch 32 Iteration 1\n",
      "Loss: 7.1699274e-06\n",
      "Epoch 32 Iteration 2\n",
      "Loss: 1.4104246e-05\n",
      "Epoch 32 Iteration 3\n",
      "Loss: 2.6929116e-05\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 32 Iteration 4\n",
      "Loss: 5.919614e-06\n",
      "Epoch 32 Iteration 5\n",
      "Loss: 2.2666306e-06\n",
      "Epoch 32 Iteration 6\n",
      "Loss: 3.8948932e-05\n",
      "Epoch 32 Iteration 7\n",
      "Loss: 1.1623193e-05\n",
      "Epoch 32 Iteration 8\n",
      "Loss: 5.3960935e-06\n",
      "Epoch 32 Iteration 9\n",
      "Loss: 1.914707e-06\n",
      "Epoch 32 Iteration 10\n",
      "Loss: 2.7336373e-05\n",
      "Epoch 32 Iteration 11\n",
      "Loss: 1.8780129e-05\n",
      "Epoch 32 Iteration 12\n",
      "Loss: 3.0650954e-05\n",
      "Epoch 32 Iteration 13\n",
      "Loss: 1.0471716e-05\n",
      "Epoch 32 Iteration 14\n",
      "Loss: 3.8115577e-06\n",
      "Epoch 32 Iteration 15\n",
      "Loss: 3.3111686e-05\n",
      "Epoch 32 Iteration 16\n",
      "Loss: 1.9187937e-05\n",
      "Epoch 32 Iteration 17\n",
      "Loss: 2.1186677e-05\n",
      "Epoch 32 Iteration 18\n",
      "Loss: 2.901548e-05\n",
      "Epoch 32 Iteration 19\n",
      "Loss: 3.986138e-06\n",
      "Epoch 32 Iteration 20\n",
      "Loss: 3.769874e-06\n",
      "Epoch 32 Iteration 21\n",
      "Loss: 2.3823502e-05\n",
      "Epoch 32 Iteration 22\n",
      "Loss: 9.449644e-06\n",
      "Epoch 32 Iteration 23\n",
      "Loss: 2.6322166e-05\n",
      "Epoch 32 Iteration 24\n",
      "Loss: 3.381771e-05\n",
      "Epoch 32 Iteration 25\n",
      "Loss: 6.3467146e-06\n",
      "Epoch 32 Iteration 26\n",
      "Loss: 3.2827553e-05\n",
      "Epoch 32 Iteration 27\n",
      "Loss: 3.6261983e-05\n",
      "Epoch 32 Iteration 28\n",
      "Loss: 2.0317588e-05\n",
      "Epoch 32 Iteration 29\n",
      "Loss: 3.8166865e-05\n",
      "Epoch 32 Iteration 30\n",
      "Loss: 2.1160433e-05\n",
      "Epoch 32 Iteration 31\n",
      "Loss: 1.0753218e-05\n",
      "Epoch 32 Iteration 32\n",
      "Loss: 2.0864638e-05\n",
      "Epoch 32 Iteration 33\n",
      "Loss: 5.3105512e-05\n",
      "Epoch 32 Iteration 34\n",
      "Loss: 2.570653e-05\n",
      "Valid loss: 2.568922718637623e-05\n",
      "Epoch  33  Current time:  Sun Jun 21 22:26:29 2020\n",
      "Epoch 33 Iteration 0\n",
      "Loss: 1.8410521e-05\n",
      "Epoch 33 Iteration 1\n",
      "Loss: 3.1241747e-05\n",
      "Epoch 33 Iteration 2\n",
      "Loss: 2.5397801e-05\n",
      "Epoch 33 Iteration 3\n",
      "Loss: 2.1647676e-05\n",
      "Epoch 33 Iteration 4\n",
      "Loss: 1.0544801e-05\n",
      "Epoch 33 Iteration 5\n",
      "Loss: 1.3265604e-05\n",
      "Epoch 33 Iteration 6\n",
      "Loss: 2.5598674e-05\n",
      "Epoch 33 Iteration 7\n",
      "Loss: 3.030119e-06\n",
      "Epoch 33 Iteration 8\n",
      "Loss: 3.3093515e-05\n",
      "Epoch 33 Iteration 9\n",
      "Loss: 6.094966e-06\n",
      "Epoch 33 Iteration 10\n",
      "Loss: 3.3476194e-06\n",
      "Epoch 33 Iteration 11\n",
      "Loss: 2.6261812e-05\n",
      "Epoch 33 Iteration 12\n",
      "Loss: 1.9620518e-05\n",
      "Epoch 33 Iteration 13\n",
      "Loss: 2.6212292e-05\n",
      "Epoch 33 Iteration 14\n",
      "Loss: 1.7433373e-05\n",
      "Epoch 33 Iteration 15\n",
      "Loss: 3.570944e-05\n",
      "Epoch 33 Iteration 16\n",
      "Loss: 3.2812997e-05\n",
      "Epoch 33 Iteration 17\n",
      "Loss: 8.902548e-06\n",
      "Epoch 33 Iteration 18\n",
      "Loss: 1.9637018e-05\n",
      "Epoch 33 Iteration 19\n",
      "Loss: 1.4036473e-05\n",
      "Epoch 33 Iteration 20\n",
      "Loss: 2.0161285e-06\n",
      "Epoch 33 Iteration 21\n",
      "Loss: 1.9826928e-05\n",
      "Epoch 33 Iteration 22\n",
      "Loss: 6.492847e-06\n",
      "Epoch 33 Iteration 23\n",
      "Loss: 3.5075635e-05\n",
      "Epoch 33 Iteration 24\n",
      "Loss: 2.047862e-06\n",
      "Epoch 33 Iteration 25\n",
      "Loss: 2.3778877e-05\n",
      "Epoch 33 Iteration 26\n",
      "Loss: 1.1573281e-05\n",
      "Epoch 33 Iteration 27\n",
      "Loss: 3.6404148e-05\n",
      "Epoch 33 Iteration 28\n",
      "Loss: 2.7450185e-05\n",
      "Epoch 33 Iteration 29\n",
      "Loss: 3.9827737e-06\n",
      "Epoch 33 Iteration 30\n",
      "Loss: 2.8980628e-05\n",
      "Epoch 33 Iteration 31\n",
      "Loss: 6.2984695e-06\n",
      "Epoch 33 Iteration 32\n",
      "Loss: 6.5478157e-06\n",
      "Epoch 33 Iteration 33\n",
      "Loss: 5.3325e-05\n",
      "Epoch 33 Iteration 34\n",
      "Loss: 1.1505256e-05\n",
      "Valid loss: 2.558069900260307e-05\n",
      "Epoch  34  Current time:  Sun Jun 21 22:27:13 2020\n",
      "Epoch 34 Iteration 0\n",
      "Loss: 2.9553974e-05\n",
      "Epoch 34 Iteration 1\n",
      "Loss: 1.9579372e-06\n",
      "Epoch 34 Iteration 2\n",
      "Loss: 2.1150852e-05\n",
      "Epoch 34 Iteration 3\n",
      "Loss: 3.1079988e-05\n",
      "Epoch 34 Iteration 4\n",
      "Loss: 2.6922542e-05\n",
      "Epoch 34 Iteration 5\n",
      "Loss: 3.7831164e-06\n",
      "Epoch 34 Iteration 6\n",
      "Loss: 3.725705e-05\n",
      "Epoch 34 Iteration 7\n",
      "Loss: 2.7782892e-06\n",
      "Epoch 34 Iteration 8\n",
      "Loss: 3.4910317e-06\n",
      "Epoch 34 Iteration 9\n",
      "Loss: 2.0114063e-05\n",
      "Epoch 34 Iteration 10\n",
      "Loss: 2.6458021e-05\n",
      "Epoch 34 Iteration 11\n",
      "Loss: 5.7014677e-06\n",
      "Epoch 34 Iteration 12\n",
      "Loss: 7.3021056e-06\n",
      "Epoch 34 Iteration 13\n",
      "Loss: 3.5361965e-05\n",
      "Epoch 34 Iteration 14\n",
      "Loss: 2.5840749e-05\n",
      "Epoch 34 Iteration 15\n",
      "Loss: 1.1631559e-05\n",
      "Epoch 34 Iteration 16\n",
      "Loss: 1.9320541e-05\n",
      "Epoch 34 Iteration 17\n",
      "Loss: 5.5311684e-06\n",
      "Epoch 34 Iteration 18\n",
      "Loss: 2.6326747e-05\n",
      "Epoch 34 Iteration 19\n",
      "Loss: 1.7656701e-05\n",
      "Epoch 34 Iteration 20\n",
      "Loss: 1.1310039e-05\n",
      "Epoch 34 Iteration 21\n",
      "Loss: 1.1018385e-05\n",
      "Epoch 34 Iteration 22\n",
      "Loss: 2.0652544e-05\n",
      "Epoch 34 Iteration 23\n",
      "Loss: 1.8155513e-06\n",
      "Epoch 34 Iteration 24\n",
      "Loss: 3.067468e-05\n",
      "Epoch 34 Iteration 25\n",
      "Loss: 4.9967883e-05\n",
      "Epoch 34 Iteration 26\n",
      "Loss: 1.0536093e-05\n",
      "Epoch 34 Iteration 27\n",
      "Loss: 1.5328462e-05\n",
      "Epoch 34 Iteration 28\n",
      "Loss: 2.4529838e-05\n",
      "Epoch 34 Iteration 29\n",
      "Loss: 9.1264e-06\n",
      "Epoch 34 Iteration 30\n",
      "Loss: 1.9555064e-05\n",
      "Epoch 34 Iteration 31\n",
      "Loss: 3.2079843e-05\n",
      "Epoch 34 Iteration 32\n",
      "Loss: 3.4933088e-05\n",
      "Epoch 34 Iteration 33\n",
      "Loss: 2.3616105e-05\n",
      "Epoch 34 Iteration 34\n",
      "Loss: 6.2471154e-06\n",
      "Valid loss: 2.4903498342609964e-05\n",
      "Epoch  35  Current time:  Sun Jun 21 22:27:58 2020\n",
      "Epoch 35 Iteration 0\n",
      "Loss: 2.4928042e-05\n",
      "Epoch 35 Iteration 1\n",
      "Loss: 1.35956725e-05\n",
      "Epoch 35 Iteration 2\n",
      "Loss: 2.5936473e-05\n",
      "Epoch 35 Iteration 3\n",
      "Loss: 1.5781374e-05\n",
      "Epoch 35 Iteration 4\n",
      "Loss: 2.450258e-05\n",
      "Epoch 35 Iteration 5\n",
      "Loss: 3.311871e-05\n",
      "Epoch 35 Iteration 6\n",
      "Loss: 6.0807315e-06\n",
      "Epoch 35 Iteration 7\n",
      "Loss: 2.8360973e-05\n",
      "Epoch 35 Iteration 8\n",
      "Loss: 6.314291e-06\n",
      "Epoch 35 Iteration 9\n",
      "Loss: 8.758652e-06\n",
      "Epoch 35 Iteration 10\n",
      "Loss: 3.4449185e-05\n",
      "Epoch 35 Iteration 11\n",
      "Loss: 1.36701765e-05\n",
      "Epoch 35 Iteration 12\n",
      "Loss: 2.4854327e-05\n",
      "Epoch 35 Iteration 13\n",
      "Loss: 1.0416992e-05\n",
      "Epoch 35 Iteration 14\n",
      "Loss: 2.024217e-05\n",
      "Epoch 35 Iteration 15\n",
      "Loss: 1.7881646e-06\n",
      "Epoch 35 Iteration 16\n",
      "Loss: 1.9846566e-05\n",
      "Epoch 35 Iteration 17\n",
      "Loss: 1.0700973e-05\n",
      "Epoch 35 Iteration 18\n",
      "Loss: 3.7693022e-05\n",
      "Epoch 35 Iteration 19\n",
      "Loss: 4.1372273e-06\n",
      "Epoch 35 Iteration 20\n",
      "Loss: 1.9213774e-05\n",
      "Epoch 35 Iteration 21\n",
      "Loss: 3.6670963e-06\n",
      "Epoch 35 Iteration 22\n",
      "Loss: 2.3081153e-05\n",
      "Epoch 35 Iteration 23\n",
      "Loss: 1.8111463e-05\n",
      "Epoch 35 Iteration 24\n",
      "Loss: 2.7458543e-06\n",
      "Epoch 35 Iteration 25\n",
      "Loss: 3.085297e-05\n",
      "Epoch 35 Iteration 26\n",
      "Loss: 3.4607322e-06\n",
      "Epoch 35 Iteration 27\n",
      "Loss: 1.8631426e-05\n",
      "Epoch 35 Iteration 28\n",
      "Loss: 5.4828643e-06\n",
      "Epoch 35 Iteration 29\n",
      "Loss: 2.8101225e-05\n",
      "Epoch 35 Iteration 30\n",
      "Loss: 1.0433056e-05\n",
      "Epoch 35 Iteration 31\n",
      "Loss: 2.4563196e-05\n",
      "Epoch 35 Iteration 32\n",
      "Loss: 3.4578952e-05\n",
      "Epoch 35 Iteration 33\n",
      "Loss: 6.257916e-06\n",
      "Epoch 35 Iteration 34\n",
      "Loss: 5.162075e-05\n",
      "Valid loss: 2.5302641006419435e-05\n",
      "Epoch  36  Current time:  Sun Jun 21 22:28:42 2020\n",
      "Epoch 36 Iteration 0\n",
      "Loss: 3.5745885e-05\n",
      "Epoch 36 Iteration 1\n",
      "Loss: 1.6380794e-05\n",
      "Epoch 36 Iteration 2\n",
      "Loss: 3.6003963e-05\n",
      "Epoch 36 Iteration 3\n",
      "Loss: 2.5618128e-05\n",
      "Epoch 36 Iteration 4\n",
      "Loss: 1.1726405e-05\n",
      "Epoch 36 Iteration 5\n",
      "Loss: 9.442778e-06\n",
      "Epoch 36 Iteration 6\n",
      "Loss: 1.40835355e-05\n",
      "Epoch 36 Iteration 7\n",
      "Loss: 4.5396402e-05\n",
      "Epoch 36 Iteration 8\n",
      "Loss: 1.515586e-05\n",
      "Epoch 36 Iteration 9\n",
      "Loss: 5.2894943e-06\n",
      "Epoch 36 Iteration 10\n",
      "Loss: 2.222376e-06\n",
      "Epoch 36 Iteration 11\n",
      "Loss: 2.1707483e-05\n",
      "Epoch 36 Iteration 12\n",
      "Loss: 2.3959034e-05\n",
      "Epoch 36 Iteration 13\n",
      "Loss: 1.8024697e-05\n",
      "Epoch 36 Iteration 14\n",
      "Loss: 2.9752473e-05\n",
      "Epoch 36 Iteration 15\n",
      "Loss: 3.149355e-05\n",
      "Epoch 36 Iteration 16\n",
      "Loss: 3.092537e-05\n",
      "Epoch 36 Iteration 17\n",
      "Loss: 2.5463189e-05\n",
      "Epoch 36 Iteration 18\n",
      "Loss: 6.196146e-06\n",
      "Epoch 36 Iteration 19\n",
      "Loss: 4.028846e-05\n",
      "Epoch 36 Iteration 20\n",
      "Loss: 2.158579e-05\n",
      "Epoch 36 Iteration 21\n",
      "Loss: 3.0775496e-05\n",
      "Epoch 36 Iteration 22\n",
      "Loss: 3.561667e-05\n",
      "Epoch 36 Iteration 23\n",
      "Loss: 4.3383556e-05\n",
      "Epoch 36 Iteration 24\n",
      "Loss: 3.0803822e-06\n",
      "Epoch 36 Iteration 25\n",
      "Loss: 1.4240166e-05\n",
      "Epoch 36 Iteration 26\n",
      "Loss: 6.5094273e-06\n",
      "Epoch 36 Iteration 27\n",
      "Loss: 5.9426293e-06\n",
      "Epoch 36 Iteration 28\n",
      "Loss: 7.335776e-06\n",
      "Epoch 36 Iteration 29\n",
      "Loss: 1.1425822e-05\n",
      "Epoch 36 Iteration 30\n",
      "Loss: 3.178703e-05\n",
      "Epoch 36 Iteration 31\n",
      "Loss: 3.2386888e-05\n",
      "Epoch 36 Iteration 32\n",
      "Loss: 6.4491724e-06\n",
      "Epoch 36 Iteration 33\n",
      "Loss: 2.4225417e-05\n",
      "Epoch 36 Iteration 34\n",
      "Loss: 3.1997656e-06\n",
      "Valid loss: 2.9222856028354727e-05\n",
      "Epoch  37  Current time:  Sun Jun 21 22:29:27 2020\n",
      "Epoch 37 Iteration 0\n",
      "Loss: 3.8676277e-05\n",
      "Epoch 37 Iteration 1\n",
      "Loss: 2.2781638e-05\n",
      "Epoch 37 Iteration 2\n",
      "Loss: 3.1228523e-05\n",
      "Epoch 37 Iteration 3\n",
      "Loss: 1.8727806e-05\n",
      "Epoch 37 Iteration 4\n",
      "Loss: 9.567892e-06\n",
      "Epoch 37 Iteration 5\n",
      "Loss: 3.3864844e-06\n",
      "Epoch 37 Iteration 6\n",
      "Loss: 3.3332126e-05\n",
      "Epoch 37 Iteration 7\n",
      "Loss: 4.158902e-05\n",
      "Epoch 37 Iteration 8\n",
      "Loss: 2.0589516e-05\n",
      "Epoch 37 Iteration 9\n",
      "Loss: 3.3782476e-06\n",
      "Epoch 37 Iteration 10\n",
      "Loss: 2.599884e-05\n",
      "Epoch 37 Iteration 11\n",
      "Loss: 4.3467658e-06\n",
      "Epoch 37 Iteration 12\n",
      "Loss: 9.48206e-06\n",
      "Epoch 37 Iteration 13\n",
      "Loss: 1.1653084e-05\n",
      "Epoch 37 Iteration 14\n",
      "Loss: 2.8279152e-05\n",
      "Epoch 37 Iteration 15\n",
      "Loss: 6.8646536e-06\n",
      "Epoch 37 Iteration 16\n",
      "Loss: 2.2021974e-05\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 37 Iteration 17\n",
      "Loss: 4.131334e-06\n",
      "Epoch 37 Iteration 18\n",
      "Loss: 1.8318082e-05\n",
      "Epoch 37 Iteration 19\n",
      "Loss: 3.7853333e-05\n",
      "Epoch 37 Iteration 20\n",
      "Loss: 7.89193e-06\n",
      "Epoch 37 Iteration 21\n",
      "Loss: 2.7259633e-05\n",
      "Epoch 37 Iteration 22\n",
      "Loss: 4.0645926e-05\n",
      "Epoch 37 Iteration 23\n",
      "Loss: 2.3752982e-05\n",
      "Epoch 37 Iteration 24\n",
      "Loss: 2.9326522e-05\n",
      "Epoch 37 Iteration 25\n",
      "Loss: 1.3277648e-05\n",
      "Epoch 37 Iteration 26\n",
      "Loss: 2.9019096e-05\n",
      "Epoch 37 Iteration 27\n",
      "Loss: 1.9257499e-05\n",
      "Epoch 37 Iteration 28\n",
      "Loss: 1.1160694e-05\n",
      "Epoch 37 Iteration 29\n",
      "Loss: 1.07593305e-05\n",
      "Epoch 37 Iteration 30\n",
      "Loss: 6.532221e-05\n",
      "Epoch 37 Iteration 31\n",
      "Loss: 1.3034008e-05\n",
      "Epoch 37 Iteration 32\n",
      "Loss: 1.3158806e-05\n",
      "Epoch 37 Iteration 33\n",
      "Loss: 3.70806e-06\n",
      "Epoch 37 Iteration 34\n",
      "Loss: 3.2378244e-05\n",
      "Valid loss: 2.791843689919915e-05\n",
      "Epoch  38  Current time:  Sun Jun 21 22:30:11 2020\n",
      "Epoch 38 Iteration 0\n",
      "Loss: 4.1166633e-05\n",
      "Epoch 38 Iteration 1\n",
      "Loss: 1.5067511e-05\n",
      "Epoch 38 Iteration 2\n",
      "Loss: 2.3478222e-05\n",
      "Epoch 38 Iteration 3\n",
      "Loss: 1.2118373e-05\n",
      "Epoch 38 Iteration 4\n",
      "Loss: 3.6087406e-06\n",
      "Epoch 38 Iteration 5\n",
      "Loss: 9.823967e-06\n",
      "Epoch 38 Iteration 6\n",
      "Loss: 5.0886785e-05\n",
      "Epoch 38 Iteration 7\n",
      "Loss: 8.413326e-06\n",
      "Epoch 38 Iteration 8\n",
      "Loss: 1.6530603e-05\n",
      "Epoch 38 Iteration 9\n",
      "Loss: 3.5295197e-05\n",
      "Epoch 38 Iteration 10\n",
      "Loss: 8.286189e-06\n",
      "Epoch 38 Iteration 11\n",
      "Loss: 1.790207e-05\n",
      "Epoch 38 Iteration 12\n",
      "Loss: 3.339008e-05\n",
      "Epoch 38 Iteration 13\n",
      "Loss: 2.9640356e-05\n",
      "Epoch 38 Iteration 14\n",
      "Loss: 1.937553e-05\n",
      "Epoch 38 Iteration 15\n",
      "Loss: 3.687816e-05\n",
      "Epoch 38 Iteration 16\n",
      "Loss: 1.9875451e-05\n",
      "Epoch 38 Iteration 17\n",
      "Loss: 2.5827107e-05\n",
      "Epoch 38 Iteration 18\n",
      "Loss: 2.4534238e-05\n",
      "Epoch 38 Iteration 19\n",
      "Loss: 2.6529236e-05\n",
      "Epoch 38 Iteration 20\n",
      "Loss: 1.32997375e-05\n",
      "Epoch 38 Iteration 21\n",
      "Loss: 3.7063376e-06\n",
      "Epoch 38 Iteration 22\n",
      "Loss: 2.6578118e-05\n",
      "Epoch 38 Iteration 23\n",
      "Loss: 3.4667446e-06\n",
      "Epoch 38 Iteration 24\n",
      "Loss: 7.431445e-06\n",
      "Epoch 38 Iteration 25\n",
      "Loss: 2.9631206e-06\n",
      "Epoch 38 Iteration 26\n",
      "Loss: 1.1347465e-05\n",
      "Epoch 38 Iteration 27\n",
      "Loss: 3.2410942e-05\n",
      "Epoch 38 Iteration 28\n",
      "Loss: 1.8949771e-05\n",
      "Epoch 38 Iteration 29\n",
      "Loss: 2.9407194e-05\n",
      "Epoch 38 Iteration 30\n",
      "Loss: 2.5786587e-05\n",
      "Epoch 38 Iteration 31\n",
      "Loss: 1.1858766e-05\n",
      "Epoch 38 Iteration 32\n",
      "Loss: 4.5368515e-06\n",
      "Epoch 38 Iteration 33\n",
      "Loss: 6.508364e-06\n",
      "Epoch 38 Iteration 34\n",
      "Loss: 2.4459621e-05\n",
      "Valid loss: 2.50858338404214e-05\n",
      "Epoch  39  Current time:  Sun Jun 21 22:30:56 2020\n",
      "Epoch 39 Iteration 0\n",
      "Loss: 1.6306702e-05\n",
      "Epoch 39 Iteration 1\n",
      "Loss: 2.457231e-05\n",
      "Epoch 39 Iteration 2\n",
      "Loss: 1.8689243e-05\n",
      "Epoch 39 Iteration 3\n",
      "Loss: 2.2064682e-05\n",
      "Epoch 39 Iteration 4\n",
      "Loss: 2.3222608e-05\n",
      "Epoch 39 Iteration 5\n",
      "Loss: 2.5016036e-05\n",
      "Epoch 39 Iteration 6\n",
      "Loss: 3.3497781e-06\n",
      "Epoch 39 Iteration 7\n",
      "Loss: 2.5496682e-05\n",
      "Epoch 39 Iteration 8\n",
      "Loss: 3.3417105e-05\n",
      "Epoch 39 Iteration 9\n",
      "Loss: 6.23886e-06\n",
      "Epoch 39 Iteration 10\n",
      "Loss: 8.824023e-06\n",
      "Epoch 39 Iteration 11\n",
      "Loss: 2.495771e-05\n",
      "Epoch 39 Iteration 12\n",
      "Loss: 1.4932325e-05\n",
      "Epoch 39 Iteration 13\n",
      "Loss: 1.8814872e-05\n",
      "Epoch 39 Iteration 14\n",
      "Loss: 2.5342213e-05\n",
      "Epoch 39 Iteration 15\n",
      "Loss: 3.5708148e-05\n",
      "Epoch 39 Iteration 16\n",
      "Loss: 6.5786353e-06\n",
      "Epoch 39 Iteration 17\n",
      "Loss: 1.8351035e-05\n",
      "Epoch 39 Iteration 18\n",
      "Loss: 2.7951876e-06\n",
      "Epoch 39 Iteration 19\n",
      "Loss: 1.8237874e-06\n",
      "Epoch 39 Iteration 20\n",
      "Loss: 1.5641363e-05\n",
      "Epoch 39 Iteration 21\n",
      "Loss: 5.0277104e-06\n",
      "Epoch 39 Iteration 22\n",
      "Loss: 3.7330237e-06\n",
      "Epoch 39 Iteration 23\n",
      "Loss: 2.921658e-05\n",
      "Epoch 39 Iteration 24\n",
      "Loss: 1.9532745e-06\n",
      "Epoch 39 Iteration 25\n",
      "Loss: 1.2736274e-05\n",
      "Epoch 39 Iteration 26\n",
      "Loss: 5.9681415e-06\n",
      "Epoch 39 Iteration 27\n",
      "Loss: 2.5216832e-05\n",
      "Epoch 39 Iteration 28\n",
      "Loss: 3.436146e-05\n",
      "Epoch 39 Iteration 29\n",
      "Loss: 4.0404047e-05\n",
      "Epoch 39 Iteration 30\n",
      "Loss: 1.2322685e-05\n",
      "Epoch 39 Iteration 31\n",
      "Loss: 1.0780179e-05\n",
      "Epoch 39 Iteration 32\n",
      "Loss: 1.7228911e-05\n",
      "Epoch 39 Iteration 33\n",
      "Loss: 5.5769426e-05\n",
      "Epoch 39 Iteration 34\n",
      "Loss: 3.4699995e-05\n",
      "Valid loss: 2.566499460954219e-05\n",
      "Epoch  40  Current time:  Sun Jun 21 22:31:40 2020\n",
      "Epoch 40 Iteration 0\n",
      "Loss: 7.4941267e-06\n",
      "Epoch 40 Iteration 1\n",
      "Loss: 3.6250256e-06\n",
      "Epoch 40 Iteration 2\n",
      "Loss: 1.930618e-05\n",
      "Epoch 40 Iteration 3\n",
      "Loss: 4.926345e-06\n",
      "Epoch 40 Iteration 4\n",
      "Loss: 2.480058e-05\n",
      "Epoch 40 Iteration 5\n",
      "Loss: 2.160039e-05\n",
      "Epoch 40 Iteration 6\n",
      "Loss: 1.7661741e-05\n",
      "Epoch 40 Iteration 7\n",
      "Loss: 1.0769523e-05\n",
      "Epoch 40 Iteration 8\n",
      "Loss: 1.1382925e-05\n",
      "Epoch 40 Iteration 9\n",
      "Loss: 2.6033225e-05\n",
      "Epoch 40 Iteration 10\n",
      "Loss: 2.519198e-05\n",
      "Epoch 40 Iteration 11\n",
      "Loss: 2.9937971e-06\n",
      "Epoch 40 Iteration 12\n",
      "Loss: 3.946832e-05\n",
      "Epoch 40 Iteration 13\n",
      "Loss: 3.0054575e-06\n",
      "Epoch 40 Iteration 14\n",
      "Loss: 6.2528457e-06\n",
      "Epoch 40 Iteration 15\n",
      "Loss: 2.8549555e-05\n",
      "Epoch 40 Iteration 16\n",
      "Loss: 1.3456752e-05\n",
      "Epoch 40 Iteration 17\n",
      "Loss: 2.5687728e-05\n",
      "Epoch 40 Iteration 18\n",
      "Loss: 1.9223246e-05\n",
      "Epoch 40 Iteration 19\n",
      "Loss: 7.096186e-06\n",
      "Epoch 40 Iteration 20\n",
      "Loss: 8.66265e-06\n",
      "Epoch 40 Iteration 21\n",
      "Loss: 3.2118383e-05\n",
      "Epoch 40 Iteration 22\n",
      "Loss: 1.1480177e-05\n",
      "Epoch 40 Iteration 23\n",
      "Loss: 3.622828e-05\n",
      "Epoch 40 Iteration 24\n",
      "Loss: 5.7209677e-06\n",
      "Epoch 40 Iteration 25\n",
      "Loss: 1.7835715e-05\n",
      "Epoch 40 Iteration 26\n",
      "Loss: 2.7665066e-05\n",
      "Epoch 40 Iteration 27\n",
      "Loss: 3.436557e-05\n",
      "Epoch 40 Iteration 28\n",
      "Loss: 3.5252106e-05\n",
      "Epoch 40 Iteration 29\n",
      "Loss: 3.0808143e-05\n",
      "Epoch 40 Iteration 30\n",
      "Loss: 2.1253654e-06\n",
      "Epoch 40 Iteration 31\n",
      "Loss: 2.0424217e-05\n",
      "Epoch 40 Iteration 32\n",
      "Loss: 5.3291275e-05\n",
      "Epoch 40 Iteration 33\n",
      "Loss: 2.3699113e-05\n",
      "Epoch 40 Iteration 34\n",
      "Loss: 1.9494311e-05\n",
      "Valid loss: 2.4831049813656136e-05\n",
      "Epoch  41  Current time:  Sun Jun 21 22:32:24 2020\n",
      "Epoch 41 Iteration 0\n",
      "Loss: 2.7278815e-05\n",
      "Epoch 41 Iteration 1\n",
      "Loss: 4.8553135e-05\n",
      "Epoch 41 Iteration 2\n",
      "Loss: 3.313788e-05\n",
      "Epoch 41 Iteration 3\n",
      "Loss: 1.7914535e-05\n",
      "Epoch 41 Iteration 4\n",
      "Loss: 1.6060887e-05\n",
      "Epoch 41 Iteration 5\n",
      "Loss: 3.4146982e-05\n",
      "Epoch 41 Iteration 6\n",
      "Loss: 1.8532965e-05\n",
      "Epoch 41 Iteration 7\n",
      "Loss: 2.7684018e-05\n",
      "Epoch 41 Iteration 8\n",
      "Loss: 2.4253745e-06\n",
      "Epoch 41 Iteration 9\n",
      "Loss: 6.884032e-06\n",
      "Epoch 41 Iteration 10\n",
      "Loss: 1.4286616e-05\n",
      "Epoch 41 Iteration 11\n",
      "Loss: 1.8794122e-05\n",
      "Epoch 41 Iteration 12\n",
      "Loss: 8.614879e-06\n",
      "Epoch 41 Iteration 13\n",
      "Loss: 1.7544387e-05\n",
      "Epoch 41 Iteration 14\n",
      "Loss: 1.0538621e-05\n",
      "Epoch 41 Iteration 15\n",
      "Loss: 1.2009042e-05\n",
      "Epoch 41 Iteration 16\n",
      "Loss: 2.6302619e-06\n",
      "Epoch 41 Iteration 17\n",
      "Loss: 2.7661974e-05\n",
      "Epoch 41 Iteration 18\n",
      "Loss: 2.4903042e-05\n",
      "Epoch 41 Iteration 19\n",
      "Loss: 3.367397e-05\n",
      "Epoch 41 Iteration 20\n",
      "Loss: 3.902349e-06\n",
      "Epoch 41 Iteration 21\n",
      "Loss: 1.8190227e-05\n",
      "Epoch 41 Iteration 22\n",
      "Loss: 2.7178678e-05\n",
      "Epoch 41 Iteration 23\n",
      "Loss: 4.1505e-06\n",
      "Epoch 41 Iteration 24\n",
      "Loss: 1.5104845e-05\n",
      "Epoch 41 Iteration 25\n",
      "Loss: 6.1044425e-06\n",
      "Epoch 41 Iteration 26\n",
      "Loss: 2.148039e-06\n",
      "Epoch 41 Iteration 27\n",
      "Loss: 6.218909e-06\n",
      "Epoch 41 Iteration 28\n",
      "Loss: 2.8213834e-05\n",
      "Epoch 41 Iteration 29\n",
      "Loss: 2.6464117e-05\n",
      "Epoch 41 Iteration 30\n",
      "Loss: 2.3139608e-05\n",
      "Epoch 41 Iteration 31\n",
      "Loss: 2.5433097e-05\n",
      "Epoch 41 Iteration 32\n",
      "Loss: 3.0503666e-05\n",
      "Epoch 41 Iteration 33\n",
      "Loss: 1.083049e-05\n",
      "Epoch 41 Iteration 34\n",
      "Loss: 7.5996777e-06\n",
      "Valid loss: 2.3919634259073064e-05\n",
      "Epoch  42  Current time:  Sun Jun 21 22:33:09 2020\n",
      "Epoch 42 Iteration 0\n",
      "Loss: 4.2973857e-05\n",
      "Epoch 42 Iteration 1\n",
      "Loss: 6.0012544e-06\n",
      "Epoch 42 Iteration 2\n",
      "Loss: 3.4969125e-05\n",
      "Epoch 42 Iteration 3\n",
      "Loss: 1.7388533e-05\n",
      "Epoch 42 Iteration 4\n",
      "Loss: 1.7512066e-05\n",
      "Epoch 42 Iteration 5\n",
      "Loss: 4.977712e-06\n",
      "Epoch 42 Iteration 6\n",
      "Loss: 2.5298612e-05\n",
      "Epoch 42 Iteration 7\n",
      "Loss: 2.8954473e-06\n",
      "Epoch 42 Iteration 8\n",
      "Loss: 3.5606595e-06\n",
      "Epoch 42 Iteration 9\n",
      "Loss: 3.181526e-05\n",
      "Epoch 42 Iteration 10\n",
      "Loss: 3.725674e-05\n",
      "Epoch 42 Iteration 11\n",
      "Loss: 1.8569994e-05\n",
      "Epoch 42 Iteration 12\n",
      "Loss: 1.7373422e-05\n",
      "Epoch 42 Iteration 13\n",
      "Loss: 2.0043466e-05\n",
      "Epoch 42 Iteration 14\n",
      "Loss: 5.8607143e-06\n",
      "Epoch 42 Iteration 15\n",
      "Loss: 1.0172472e-05\n",
      "Epoch 42 Iteration 16\n",
      "Loss: 2.5771158e-05\n",
      "Epoch 42 Iteration 17\n",
      "Loss: 1.2764314e-05\n",
      "Epoch 42 Iteration 18\n",
      "Loss: 1.06980315e-05\n",
      "Epoch 42 Iteration 19\n",
      "Loss: 2.8013179e-05\n",
      "Epoch 42 Iteration 20\n",
      "Loss: 6.2534946e-06\n",
      "Epoch 42 Iteration 21\n",
      "Loss: 1.1021702e-05\n",
      "Epoch 42 Iteration 22\n",
      "Loss: 2.060505e-06\n",
      "Epoch 42 Iteration 23\n",
      "Loss: 2.8550336e-05\n",
      "Epoch 42 Iteration 24\n",
      "Loss: 2.8590923e-05\n",
      "Epoch 42 Iteration 25\n",
      "Loss: 2.125199e-05\n",
      "Epoch 42 Iteration 26\n",
      "Loss: 3.7344747e-05\n",
      "Epoch 42 Iteration 27\n",
      "Loss: 1.0865717e-05\n",
      "Epoch 42 Iteration 28\n",
      "Loss: 6.3887783e-06\n",
      "Epoch 42 Iteration 29\n",
      "Loss: 2.307628e-05\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 42 Iteration 30\n",
      "Loss: 2.507111e-05\n",
      "Epoch 42 Iteration 31\n",
      "Loss: 2.857378e-06\n",
      "Epoch 42 Iteration 32\n",
      "Loss: 3.2021548e-05\n",
      "Epoch 42 Iteration 33\n",
      "Loss: 1.8009532e-05\n",
      "Epoch 42 Iteration 34\n",
      "Loss: 1.5019946e-05\n",
      "Valid loss: 2.3198981580208056e-05\n",
      "Epoch  43  Current time:  Sun Jun 21 22:33:53 2020\n",
      "Epoch 43 Iteration 0\n",
      "Loss: 3.2871543e-05\n",
      "Epoch 43 Iteration 1\n",
      "Loss: 1.8334913e-05\n",
      "Epoch 43 Iteration 2\n",
      "Loss: 1.8832086e-06\n",
      "Epoch 43 Iteration 3\n",
      "Loss: 7.3488704e-06\n",
      "Epoch 43 Iteration 4\n",
      "Loss: 3.669994e-05\n",
      "Epoch 43 Iteration 5\n",
      "Loss: 3.7906827e-06\n",
      "Epoch 43 Iteration 6\n",
      "Loss: 1.2322919e-05\n",
      "Epoch 43 Iteration 7\n",
      "Loss: 1.8808794e-05\n",
      "Epoch 43 Iteration 8\n",
      "Loss: 2.9621115e-06\n",
      "Epoch 43 Iteration 9\n",
      "Loss: 2.6714963e-05\n",
      "Epoch 43 Iteration 10\n",
      "Loss: 9.386067e-06\n",
      "Epoch 43 Iteration 11\n",
      "Loss: 2.4950645e-05\n",
      "Epoch 43 Iteration 12\n",
      "Loss: 4.4446842e-05\n",
      "Epoch 43 Iteration 13\n",
      "Loss: 3.104759e-05\n",
      "Epoch 43 Iteration 14\n",
      "Loss: 1.0359316e-05\n",
      "Epoch 43 Iteration 15\n",
      "Loss: 2.354636e-05\n",
      "Epoch 43 Iteration 16\n",
      "Loss: 2.4698578e-05\n",
      "Epoch 43 Iteration 17\n",
      "Loss: 2.7200238e-05\n",
      "Epoch 43 Iteration 18\n",
      "Loss: 6.3104703e-06\n",
      "Epoch 43 Iteration 19\n",
      "Loss: 1.1063725e-05\n",
      "Epoch 43 Iteration 20\n",
      "Loss: 2.0474965e-05\n",
      "Epoch 43 Iteration 21\n",
      "Loss: 4.2639517e-06\n",
      "Epoch 43 Iteration 22\n",
      "Loss: 2.8164048e-05\n",
      "Epoch 43 Iteration 23\n",
      "Loss: 3.413896e-05\n",
      "Epoch 43 Iteration 24\n",
      "Loss: 2.9605137e-05\n",
      "Epoch 43 Iteration 25\n",
      "Loss: 6.1231563e-06\n",
      "Epoch 43 Iteration 26\n",
      "Loss: 2.2856924e-05\n",
      "Epoch 43 Iteration 27\n",
      "Loss: 2.8067934e-05\n",
      "Epoch 43 Iteration 28\n",
      "Loss: 1.7026985e-05\n",
      "Epoch 43 Iteration 29\n",
      "Loss: 3.9858536e-05\n",
      "Epoch 43 Iteration 30\n",
      "Loss: 2.9446328e-06\n",
      "Epoch 43 Iteration 31\n",
      "Loss: 1.7003285e-05\n",
      "Epoch 43 Iteration 32\n",
      "Loss: 6.7101964e-06\n",
      "Epoch 43 Iteration 33\n",
      "Loss: 1.1145134e-05\n",
      "Epoch 43 Iteration 34\n",
      "Loss: 2.5676392e-05\n",
      "Valid loss: 2.5070830815820955e-05\n",
      "Epoch  44  Current time:  Sun Jun 21 22:34:38 2020\n",
      "Epoch 44 Iteration 0\n",
      "Loss: 6.4564874e-06\n",
      "Epoch 44 Iteration 1\n",
      "Loss: 5.9907334e-06\n",
      "Epoch 44 Iteration 2\n",
      "Loss: 3.1807896e-05\n",
      "Epoch 44 Iteration 3\n",
      "Loss: 3.262041e-05\n",
      "Epoch 44 Iteration 4\n",
      "Loss: 2.9610305e-06\n",
      "Epoch 44 Iteration 5\n",
      "Loss: 1.0501848e-05\n",
      "Epoch 44 Iteration 6\n",
      "Loss: 4.8249578e-05\n",
      "Epoch 44 Iteration 7\n",
      "Loss: 2.6835e-05\n",
      "Epoch 44 Iteration 8\n",
      "Loss: 2.5816587e-05\n",
      "Epoch 44 Iteration 9\n",
      "Loss: 1.1757567e-05\n",
      "Epoch 44 Iteration 10\n",
      "Loss: 1.8149229e-05\n",
      "Epoch 44 Iteration 11\n",
      "Loss: 6.8724016e-06\n",
      "Epoch 44 Iteration 12\n",
      "Loss: 4.0213063e-06\n",
      "Epoch 44 Iteration 13\n",
      "Loss: 2.7277129e-05\n",
      "Epoch 44 Iteration 14\n",
      "Loss: 1.5518775e-05\n",
      "Epoch 44 Iteration 15\n",
      "Loss: 2.3654563e-06\n",
      "Epoch 44 Iteration 16\n",
      "Loss: 1.8277287e-05\n",
      "Epoch 44 Iteration 17\n",
      "Loss: 2.1196338e-05\n",
      "Epoch 44 Iteration 18\n",
      "Loss: 3.3481698e-05\n",
      "Epoch 44 Iteration 19\n",
      "Loss: 1.4298796e-05\n",
      "Epoch 44 Iteration 20\n",
      "Loss: 2.384679e-05\n",
      "Epoch 44 Iteration 21\n",
      "Loss: 8.27791e-06\n",
      "Epoch 44 Iteration 22\n",
      "Loss: 1.8939934e-05\n",
      "Epoch 44 Iteration 23\n",
      "Loss: 1.0158265e-05\n",
      "Epoch 44 Iteration 24\n",
      "Loss: 3.4048496e-06\n",
      "Epoch 44 Iteration 25\n",
      "Loss: 1.078832e-05\n",
      "Epoch 44 Iteration 26\n",
      "Loss: 2.2642575e-05\n",
      "Epoch 44 Iteration 27\n",
      "Loss: 7.0440783e-06\n",
      "Epoch 44 Iteration 28\n",
      "Loss: 1.8804599e-06\n",
      "Epoch 44 Iteration 29\n",
      "Loss: 3.2500746e-05\n",
      "Epoch 44 Iteration 30\n",
      "Loss: 1.8692139e-05\n",
      "Epoch 44 Iteration 31\n",
      "Loss: 2.1926851e-05\n",
      "Epoch 44 Iteration 32\n",
      "Loss: 1.6820422e-05\n",
      "Epoch 44 Iteration 33\n",
      "Loss: 2.3747762e-05\n",
      "Epoch 44 Iteration 34\n",
      "Loss: 3.381424e-05\n",
      "Valid loss: 2.3884291294962168e-05\n",
      "Epoch  45  Current time:  Sun Jun 21 22:35:22 2020\n",
      "Epoch 45 Iteration 0\n",
      "Loss: 3.9215356e-06\n",
      "Epoch 45 Iteration 1\n",
      "Loss: 9.458099e-06\n",
      "Epoch 45 Iteration 2\n",
      "Loss: 1.7055094e-05\n",
      "Epoch 45 Iteration 3\n",
      "Loss: 3.929573e-06\n",
      "Epoch 45 Iteration 4\n",
      "Loss: 1.2702782e-05\n",
      "Epoch 45 Iteration 5\n",
      "Loss: 2.5567682e-05\n",
      "Epoch 45 Iteration 6\n",
      "Loss: 1.8936398e-05\n",
      "Epoch 45 Iteration 7\n",
      "Loss: 2.7699307e-05\n",
      "Epoch 45 Iteration 8\n",
      "Loss: 2.5525524e-05\n",
      "Epoch 45 Iteration 9\n",
      "Loss: 2.1484471e-05\n",
      "Epoch 45 Iteration 10\n",
      "Loss: 8.788082e-06\n",
      "Epoch 45 Iteration 11\n",
      "Loss: 1.0251507e-05\n",
      "Epoch 45 Iteration 12\n",
      "Loss: 2.5104493e-05\n",
      "Epoch 45 Iteration 13\n",
      "Loss: 6.751106e-06\n",
      "Epoch 45 Iteration 14\n",
      "Loss: 2.3045424e-05\n",
      "Epoch 45 Iteration 15\n",
      "Loss: 1.6878332e-05\n",
      "Epoch 45 Iteration 16\n",
      "Loss: 6.416488e-06\n",
      "Epoch 45 Iteration 17\n",
      "Loss: 3.188757e-05\n",
      "Epoch 45 Iteration 18\n",
      "Loss: 2.9331886e-05\n",
      "Epoch 45 Iteration 19\n",
      "Loss: 3.443959e-05\n",
      "Epoch 45 Iteration 20\n",
      "Loss: 2.5011579e-05\n",
      "Epoch 45 Iteration 21\n",
      "Loss: 2.8170812e-06\n",
      "Epoch 45 Iteration 22\n",
      "Loss: 2.1442524e-05\n",
      "Epoch 45 Iteration 23\n",
      "Loss: 1.740501e-05\n",
      "Epoch 45 Iteration 24\n",
      "Loss: 5.3386633e-05\n",
      "Epoch 45 Iteration 25\n",
      "Loss: 9.668798e-06\n",
      "Epoch 45 Iteration 26\n",
      "Loss: 2.225853e-06\n",
      "Epoch 45 Iteration 27\n",
      "Loss: 8.757226e-06\n",
      "Epoch 45 Iteration 28\n",
      "Loss: 1.0871868e-05\n",
      "Epoch 45 Iteration 29\n",
      "Loss: 5.595626e-06\n",
      "Epoch 45 Iteration 30\n",
      "Loss: 2.1255608e-05\n",
      "Epoch 45 Iteration 31\n",
      "Loss: 1.000354e-05\n",
      "Epoch 45 Iteration 32\n",
      "Loss: 1.837272e-05\n",
      "Epoch 45 Iteration 33\n",
      "Loss: 1.636285e-05\n",
      "Epoch 45 Iteration 34\n",
      "Loss: 2.0779187e-05\n",
      "Valid loss: 2.325428431504406e-05\n",
      "Epoch  46  Current time:  Sun Jun 21 22:36:06 2020\n",
      "Epoch 46 Iteration 0\n",
      "Loss: 3.5502802e-05\n",
      "Epoch 46 Iteration 1\n",
      "Loss: 3.1025316e-05\n",
      "Epoch 46 Iteration 2\n",
      "Loss: 1.9126848e-05\n",
      "Epoch 46 Iteration 3\n",
      "Loss: 5.979328e-06\n",
      "Epoch 46 Iteration 4\n",
      "Loss: 3.2375178e-06\n",
      "Epoch 46 Iteration 5\n",
      "Loss: 1.0534776e-05\n",
      "Epoch 46 Iteration 6\n",
      "Loss: 1.6898033e-05\n",
      "Epoch 46 Iteration 7\n",
      "Loss: 5.174228e-06\n",
      "Epoch 46 Iteration 8\n",
      "Loss: 2.5759391e-05\n",
      "Epoch 46 Iteration 9\n",
      "Loss: 1.9243469e-05\n",
      "Epoch 46 Iteration 10\n",
      "Loss: 2.6827565e-06\n",
      "Epoch 46 Iteration 11\n",
      "Loss: 1.8117156e-05\n",
      "Epoch 46 Iteration 12\n",
      "Loss: 2.2689674e-05\n",
      "Epoch 46 Iteration 13\n",
      "Loss: 1.9801359e-05\n",
      "Epoch 46 Iteration 14\n",
      "Loss: 2.3616283e-05\n",
      "Epoch 46 Iteration 15\n",
      "Loss: 4.4306053e-06\n",
      "Epoch 46 Iteration 16\n",
      "Loss: 2.4103809e-05\n",
      "Epoch 46 Iteration 17\n",
      "Loss: 4.2613538e-06\n",
      "Epoch 46 Iteration 18\n",
      "Loss: 1.5563368e-05\n",
      "Epoch 46 Iteration 19\n",
      "Loss: 6.221707e-06\n",
      "Epoch 46 Iteration 20\n",
      "Loss: 1.6126793e-05\n",
      "Epoch 46 Iteration 21\n",
      "Loss: 8.365706e-06\n",
      "Epoch 46 Iteration 22\n",
      "Loss: 7.2283237e-06\n",
      "Epoch 46 Iteration 23\n",
      "Loss: 1.8056357e-06\n",
      "Epoch 46 Iteration 24\n",
      "Loss: 4.2174688e-05\n",
      "Epoch 46 Iteration 25\n",
      "Loss: 2.0719235e-05\n",
      "Epoch 46 Iteration 26\n",
      "Loss: 6.625552e-05\n",
      "Epoch 46 Iteration 27\n",
      "Loss: 3.8542847e-05\n",
      "Epoch 46 Iteration 28\n",
      "Loss: 2.9526687e-05\n",
      "Epoch 46 Iteration 29\n",
      "Loss: 2.08255e-05\n",
      "Epoch 46 Iteration 30\n",
      "Loss: 2.0163763e-05\n",
      "Epoch 46 Iteration 31\n",
      "Loss: 8.771796e-06\n",
      "Epoch 46 Iteration 32\n",
      "Loss: 3.0310997e-05\n",
      "Epoch 46 Iteration 33\n",
      "Loss: 1.6913635e-05\n",
      "Epoch 46 Iteration 34\n",
      "Loss: 1.0516855e-05\n",
      "Valid loss: 2.4976712666102685e-05\n",
      "Epoch  47  Current time:  Sun Jun 21 22:36:51 2020\n",
      "Epoch 47 Iteration 0\n",
      "Loss: 2.0977943e-05\n",
      "Epoch 47 Iteration 1\n",
      "Loss: 2.4072971e-05\n",
      "Epoch 47 Iteration 2\n",
      "Loss: 2.5326806e-05\n",
      "Epoch 47 Iteration 3\n",
      "Loss: 1.747775e-05\n",
      "Epoch 47 Iteration 4\n",
      "Loss: 4.178317e-05\n",
      "Epoch 47 Iteration 5\n",
      "Loss: 9.416482e-06\n",
      "Epoch 47 Iteration 6\n",
      "Loss: 5.116703e-05\n",
      "Epoch 47 Iteration 7\n",
      "Loss: 2.531857e-05\n",
      "Epoch 47 Iteration 8\n",
      "Loss: 3.5026034e-05\n",
      "Epoch 47 Iteration 9\n",
      "Loss: 1.8880546e-05\n",
      "Epoch 47 Iteration 10\n",
      "Loss: 6.0016428e-06\n",
      "Epoch 47 Iteration 11\n",
      "Loss: 2.307437e-05\n",
      "Epoch 47 Iteration 12\n",
      "Loss: 1.2709842e-05\n",
      "Epoch 47 Iteration 13\n",
      "Loss: 3.3117783e-05\n",
      "Epoch 47 Iteration 14\n",
      "Loss: 2.2559127e-05\n",
      "Epoch 47 Iteration 15\n",
      "Loss: 3.852417e-06\n",
      "Epoch 47 Iteration 16\n",
      "Loss: 9.220366e-06\n",
      "Epoch 47 Iteration 17\n",
      "Loss: 1.5439931e-05\n",
      "Epoch 47 Iteration 18\n",
      "Loss: 8.226958e-06\n",
      "Epoch 47 Iteration 19\n",
      "Loss: 3.0836265e-05\n",
      "Epoch 47 Iteration 20\n",
      "Loss: 6.3184184e-06\n",
      "Epoch 47 Iteration 21\n",
      "Loss: 1.7041802e-05\n",
      "Epoch 47 Iteration 22\n",
      "Loss: 9.8055325e-06\n",
      "Epoch 47 Iteration 23\n",
      "Loss: 7.156665e-06\n",
      "Epoch 47 Iteration 24\n",
      "Loss: 1.4124666e-05\n",
      "Epoch 47 Iteration 25\n",
      "Loss: 2.4860583e-05\n",
      "Epoch 47 Iteration 26\n",
      "Loss: 3.8044877e-06\n",
      "Epoch 47 Iteration 27\n",
      "Loss: 1.7492275e-05\n",
      "Epoch 47 Iteration 28\n",
      "Loss: 1.0776439e-05\n",
      "Epoch 47 Iteration 29\n",
      "Loss: 2.6763084e-06\n",
      "Epoch 47 Iteration 30\n",
      "Loss: 2.7931052e-05\n",
      "Epoch 47 Iteration 31\n",
      "Loss: 2.2501355e-05\n",
      "Epoch 47 Iteration 32\n",
      "Loss: 3.054027e-05\n",
      "Epoch 47 Iteration 33\n",
      "Loss: 4.1657536e-06\n",
      "Epoch 47 Iteration 34\n",
      "Loss: 1.3061202e-05\n",
      "Valid loss: 2.4680764909135178e-05\n",
      "Epoch  48  Current time:  Sun Jun 21 22:37:35 2020\n",
      "Epoch 48 Iteration 0\n",
      "Loss: 2.3221044e-05\n",
      "Epoch 48 Iteration 1\n",
      "Loss: 4.7706977e-05\n",
      "Epoch 48 Iteration 2\n",
      "Loss: 2.165123e-05\n",
      "Epoch 48 Iteration 3\n",
      "Loss: 3.863414e-06\n",
      "Epoch 48 Iteration 4\n",
      "Loss: 1.6823877e-05\n",
      "Epoch 48 Iteration 5\n",
      "Loss: 1.677858e-05\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 48 Iteration 6\n",
      "Loss: 3.229705e-05\n",
      "Epoch 48 Iteration 7\n",
      "Loss: 4.582542e-06\n",
      "Epoch 48 Iteration 8\n",
      "Loss: 1.5522255e-05\n",
      "Epoch 48 Iteration 9\n",
      "Loss: 1.8136701e-05\n",
      "Epoch 48 Iteration 10\n",
      "Loss: 8.670194e-06\n",
      "Epoch 48 Iteration 11\n",
      "Loss: 9.750468e-06\n",
      "Epoch 48 Iteration 12\n",
      "Loss: 2.485323e-05\n",
      "Epoch 48 Iteration 13\n",
      "Loss: 3.6981677e-05\n",
      "Epoch 48 Iteration 14\n",
      "Loss: 1.2912361e-05\n",
      "Epoch 48 Iteration 15\n",
      "Loss: 1.840819e-05\n",
      "Epoch 48 Iteration 16\n",
      "Loss: 6.572167e-06\n",
      "Epoch 48 Iteration 17\n",
      "Loss: 3.641912e-05\n",
      "Epoch 48 Iteration 18\n",
      "Loss: 2.7074118e-05\n",
      "Epoch 48 Iteration 19\n",
      "Loss: 2.0230656e-05\n",
      "Epoch 48 Iteration 20\n",
      "Loss: 1.4117498e-05\n",
      "Epoch 48 Iteration 21\n",
      "Loss: 2.7664146e-05\n",
      "Epoch 48 Iteration 22\n",
      "Loss: 1.0821823e-05\n",
      "Epoch 48 Iteration 23\n",
      "Loss: 5.442555e-06\n",
      "Epoch 48 Iteration 24\n",
      "Loss: 1.8187699e-05\n",
      "Epoch 48 Iteration 25\n",
      "Loss: 6.8859413e-06\n",
      "Epoch 48 Iteration 26\n",
      "Loss: 1.9600388e-05\n",
      "Epoch 48 Iteration 27\n",
      "Loss: 2.109707e-05\n",
      "Epoch 48 Iteration 28\n",
      "Loss: 1.2731704e-05\n",
      "Epoch 48 Iteration 29\n",
      "Loss: 3.847421e-06\n",
      "Epoch 48 Iteration 30\n",
      "Loss: 3.1298445e-05\n",
      "Epoch 48 Iteration 31\n",
      "Loss: 5.002982e-06\n",
      "Epoch 48 Iteration 32\n",
      "Loss: 2.6773714e-06\n",
      "Epoch 48 Iteration 33\n",
      "Loss: 2.4963343e-05\n",
      "Epoch 48 Iteration 34\n",
      "Loss: 1.799445e-06\n",
      "Valid loss: 2.5509729312034324e-05\n",
      "Epoch  49  Current time:  Sun Jun 21 22:38:20 2020\n",
      "Epoch 49 Iteration 0\n",
      "Loss: 1.9716788e-05\n",
      "Epoch 49 Iteration 1\n",
      "Loss: 6.538017e-06\n",
      "Epoch 49 Iteration 2\n",
      "Loss: 1.5585738e-05\n",
      "Epoch 49 Iteration 3\n",
      "Loss: 7.929859e-06\n",
      "Epoch 49 Iteration 4\n",
      "Loss: 5.776811e-06\n",
      "Epoch 49 Iteration 5\n",
      "Loss: 4.9794377e-05\n",
      "Epoch 49 Iteration 6\n",
      "Loss: 2.1977385e-06\n",
      "Epoch 49 Iteration 7\n",
      "Loss: 3.2236032e-05\n",
      "Epoch 49 Iteration 8\n",
      "Loss: 2.4384142e-05\n",
      "Epoch 49 Iteration 9\n",
      "Loss: 1.4232632e-05\n",
      "Epoch 49 Iteration 10\n",
      "Loss: 1.6602156e-05\n",
      "Epoch 49 Iteration 11\n",
      "Loss: 3.149618e-05\n",
      "Epoch 49 Iteration 12\n",
      "Loss: 1.9376726e-05\n",
      "Epoch 49 Iteration 13\n",
      "Loss: 9.537825e-06\n",
      "Epoch 49 Iteration 14\n",
      "Loss: 1.39827835e-05\n",
      "Epoch 49 Iteration 15\n",
      "Loss: 1.9736393e-05\n",
      "Epoch 49 Iteration 16\n",
      "Loss: 9.252241e-06\n",
      "Epoch 49 Iteration 17\n",
      "Loss: 3.979946e-06\n",
      "Epoch 49 Iteration 18\n",
      "Loss: 2.2911938e-05\n",
      "Epoch 49 Iteration 19\n",
      "Loss: 1.8112949e-05\n",
      "Epoch 49 Iteration 20\n",
      "Loss: 1.6170787e-05\n",
      "Epoch 49 Iteration 21\n",
      "Loss: 2.6253547e-06\n",
      "Epoch 49 Iteration 22\n",
      "Loss: 1.6167705e-05\n",
      "Epoch 49 Iteration 23\n",
      "Loss: 2.2840099e-05\n",
      "Epoch 49 Iteration 24\n",
      "Loss: 3.136439e-06\n",
      "Epoch 49 Iteration 25\n",
      "Loss: 1.0159119e-05\n",
      "Epoch 49 Iteration 26\n",
      "Loss: 2.0110576e-05\n",
      "Epoch 49 Iteration 27\n",
      "Loss: 3.0222196e-05\n",
      "Epoch 49 Iteration 28\n",
      "Loss: 2.4263568e-06\n",
      "Epoch 49 Iteration 29\n",
      "Loss: 4.763412e-06\n",
      "Epoch 49 Iteration 30\n",
      "Loss: 1.6271502e-05\n",
      "Epoch 49 Iteration 31\n",
      "Loss: 1.6251925e-05\n",
      "Epoch 49 Iteration 32\n",
      "Loss: 5.131673e-06\n",
      "Epoch 49 Iteration 33\n",
      "Loss: 2.8886618e-05\n",
      "Epoch 49 Iteration 34\n",
      "Loss: 1.4865132e-05\n",
      "Valid loss: 2.550942917878274e-05\n"
     ]
    }
   ],
   "source": [
    "loss_list=[]\n",
    "valid_loss_in_epoch = []\n",
    "if isTrain:\n",
    "    nowtime = time.time()\n",
    "    for e in range(epoch):\n",
    "        print(\"Epoch \",e,\" Current time: \", time.ctime())\n",
    "        sel = np.arange((int)(data_3d.shape[0]/batch_size))\n",
    "        np.random.shuffle(sel)\n",
    "        iteration = (int)(data_3d.shape[0]/batch_size)\n",
    "        loss_in_epoch = []            \n",
    "        loss_v = []\n",
    "        for it in range(iteration):\n",
    "\n",
    "            input_batch = []\n",
    "            label_batch = []\n",
    "            for b in range(batch_size):\n",
    "                input_batch.append(data_3d[sel[it*batch_size+b]])\n",
    "                label_batch.append(label_3d[sel[it*batch_size+b]])\n",
    "            input_batch = np.asarray(input_batch).reshape([batch_size,input_depth,input_height,input_width,1])\n",
    "            label_batch = np.asarray(label_batch).reshape([batch_size,input_depth,input_height,input_width,output_class_num])\n",
    "\n",
    "            feed_dict=dict()\n",
    "            feed_dict.update({placeholders['input']:input_batch})\n",
    "            feed_dict.update({placeholders['label']:label_batch})\n",
    "\n",
    "            outs = sess.run([trainer, loss, Unet_out_tensor],feed_dict=feed_dict)\n",
    "            \n",
    "            cur_loss = outs[1]\n",
    "\n",
    "            print(\"Epoch\", e, \"Iteration\", it)\n",
    "            print(\"Loss:\", cur_loss)\n",
    "            loss_in_epoch.append(cur_loss)\n",
    "        loss_list.append(np.mean(loss_in_epoch))\n",
    "        if(e%1==0):\n",
    "\n",
    "            for i in range(valid_nii_CT.shape[0]):\n",
    "            #         print(\"Testing batch \", i)\n",
    "\n",
    "                input_batch = valid_nii_CT[i]  \n",
    "                label_batch = valid_onehot_array[i]\n",
    "\n",
    "                feed_dict=dict()\n",
    "                feed_dict.update({placeholders['input']:input_batch.reshape(1,input_depth,input_height,input_width,1)})\n",
    "                feed_dict.update({placeholders['label']:label_batch.reshape(1,input_depth,input_height,input_width,output_class_num)})\n",
    "                # print valid loss\n",
    "                outs = sess.run([loss, Unet_out_tensor],feed_dict=feed_dict)\n",
    "                loss_v.append(outs[0])\n",
    "                \n",
    "            valid_loss_in_epoch.append(np.mean(loss_v))\n",
    "            print(\"Valid loss: {}\".format(np.mean(loss_v)))\n",
    "            #logger.info(\"Valid loss: {loss_v}\".format(loss_v = np.mean(valid_loss)))\n",
    "        if not os.path.exists(checkpoint_dir):\n",
    "            os.makedirs(checkpoint_dir)\n",
    "        saver.save(sess, \"./\"+checkpoint_dir+\"/\"+\"Unet_epoch_\"+str(e)+\".ckpt\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training time: 39 min, 6 sec\n"
     ]
    }
   ],
   "source": [
    "if isTrain:\n",
    "    ti = time.time()-nowtime\n",
    "    print(\"Training time: %d min, %d sec\"% (ti/60, ti%60))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(50,)\n"
     ]
    },
    {
     "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": [
    "print(np.asarray(valid_loss_in_epoch).shape)\n",
    "if isTrain:\n",
    "    x = np.linspace(0,50,50)\n",
    "    x1 = np.linspace(0,50,50)\n",
    "    plt.plot(x, loss_list, label='loss')\n",
    "    plt.plot(x1, valid_loss_in_epoch, label='valid loss')\n",
    "    plt.legend(loc='upper right')\n",
    "    plt.ticklabel_format(useOffset=False)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(35, 128, 128, 128)\n",
      "(35, 128, 128, 128, 4)\n"
     ]
    }
   ],
   "source": [
    "test_dataset = data_3d\n",
    "ground_truth = label_3d\n",
    "#test_dataset = valid_nii_CT\n",
    "#ground_truth = valid_nii_label\n",
    "print(test_dataset.shape)\n",
    "print(ground_truth.shape)\n",
    "np.save('test_nii_data', data_3d)\n",
    "np.save('test_nii_label', label_3d)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Restoring parameters from ./checkpoint/Unet_epoch_49.ckpt\n",
      "Model restored.\n",
      "Initialized\n",
      "Testing batch  0\n",
      "test_output size (1, 128, 128, 128)\n",
      "Testing batch  1\n",
      "test_output size (2, 128, 128, 128)\n",
      "Testing batch  2\n",
      "test_output size (3, 128, 128, 128)\n",
      "Testing batch  3\n",
      "test_output size (4, 128, 128, 128)\n",
      "Testing batch  4\n",
      "test_output size (5, 128, 128, 128)\n",
      "Testing batch  5\n",
      "test_output size (6, 128, 128, 128)\n",
      "Testing batch  6\n",
      "test_output size (7, 128, 128, 128)\n",
      "Testing batch  7\n",
      "test_output size (8, 128, 128, 128)\n",
      "Testing batch  8\n",
      "test_output size (9, 128, 128, 128)\n",
      "Testing batch  9\n",
      "test_output size (10, 128, 128, 128)\n",
      "Testing batch  10\n",
      "test_output size (11, 128, 128, 128)\n",
      "Testing batch  11\n",
      "test_output size (12, 128, 128, 128)\n",
      "Testing batch  12\n",
      "test_output size (13, 128, 128, 128)\n",
      "Testing batch  13\n",
      "test_output size (14, 128, 128, 128)\n",
      "Testing batch  14\n",
      "test_output size (15, 128, 128, 128)\n",
      "Testing batch  15\n",
      "test_output size (16, 128, 128, 128)\n",
      "Testing batch  16\n",
      "test_output size (17, 128, 128, 128)\n",
      "Testing batch  17\n",
      "test_output size (18, 128, 128, 128)\n",
      "Testing batch  18\n",
      "test_output size (19, 128, 128, 128)\n",
      "Testing batch  19\n",
      "test_output size (20, 128, 128, 128)\n",
      "Testing batch  20\n",
      "test_output size (21, 128, 128, 128)\n",
      "Testing batch  21\n",
      "test_output size (22, 128, 128, 128)\n",
      "Testing batch  22\n",
      "test_output size (23, 128, 128, 128)\n",
      "Testing batch  23\n",
      "test_output size (24, 128, 128, 128)\n",
      "Testing batch  24\n",
      "test_output size (25, 128, 128, 128)\n",
      "Testing batch  25\n",
      "test_output size (26, 128, 128, 128)\n",
      "Testing batch  26\n",
      "test_output size (27, 128, 128, 128)\n",
      "Testing batch  27\n",
      "test_output size (28, 128, 128, 128)\n",
      "Testing batch  28\n",
      "test_output size (29, 128, 128, 128)\n",
      "Testing batch  29\n",
      "test_output size (30, 128, 128, 128)\n",
      "Testing batch  30\n",
      "test_output size (31, 128, 128, 128)\n",
      "Testing batch  31\n",
      "test_output size (32, 128, 128, 128)\n",
      "Testing batch  32\n",
      "test_output size (33, 128, 128, 128)\n",
      "Testing batch  33\n",
      "test_output size (34, 128, 128, 128)\n",
      "Testing batch  34\n",
      "test_output size (35, 128, 128, 128)\n",
      "Testing time: 1 min, 39 sec\n"
     ]
    }
   ],
   "source": [
    "nowtime = time.time()\n",
    "\n",
    "with tf.Session() as session:\n",
    "    if isTrain:\n",
    "        saver.restore(session, \"./\"+checkpoint_dir+\"/\"+\"Unet_epoch_\"+str(50-1)+\".ckpt\")\n",
    "    else:\n",
    "        saver.restore(session, \"./\"+checkpoint_dir+\"/\"+\"Unet_epoch_\"+str(26)+\".ckpt\")\n",
    "    print(\"Model restored.\") \n",
    "    print('Initialized')\n",
    "\n",
    "    for i in range(test_dataset.shape[0]):\n",
    "        print(\"Testing batch \", i)\n",
    "        test_image = test_dataset[i]\n",
    "        feed_dict=dict()\n",
    "        feed_dict.update({placeholders['input']:test_image.reshape(batch_size,input_depth,input_height,input_width,1)})\n",
    "\n",
    "        gen = session.run(tf.nn.softmax(Unet_out_tensor),feed_dict=feed_dict)\n",
    "    \n",
    "        #print(np.unique(gen), np.unique(gen))\n",
    "        gen = gen.reshape([input_depth,input_height,input_width,output_class_num])\n",
    "        #print(gen[0][0])\n",
    "        value_gen = onehot2value(gen)\n",
    "        value_gen = value_gen.reshape([batch_size, input_depth,input_height,input_width])\n",
    "        if i == 0 :\n",
    "            test_output = value_gen\n",
    "        else:\n",
    "            test_output = np.concatenate((test_output,value_gen), axis = 0)\n",
    "        \n",
    "    \n",
    "#         test_output = np.asarray(test_output)\n",
    "        print(\"test_output size\", test_output.shape)\n",
    "\n",
    "ti = time.time()-nowtime\n",
    "print(\"Testing time: %d min, %d sec\"% (ti/60, ti%60))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "np.save('testing_result_flair_t1ne', test_output)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "'''\n",
    "print('Whole volume f1 score')\n",
    "seg = output_onehot\n",
    "gt = ground_truth_onehot\n",
    "evaluation(gt, seg)\n",
    "'''\n",
    "\n",
    "output_onehot = []\n",
    "for i in range(test_output.shape[0]):\n",
    "    onehot_array = value2onehot(test_output[i])\n",
    "    output_onehot.append(onehot_array)\n",
    "output_onehot = np.asarray(output_onehot, dtype=np.uint8)\n",
    "ground_truth_onehot = ground_truth\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "patient idx:  0\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.6807681304955998 STD: 0.3626141346451423\n",
      "Hausdorff mean: 1.1700957085416546 STD: 1.4120922527458342\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  1\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.7503674045562074 STD: 0.30961438617376574\n",
      "Hausdorff mean: 0.8360020478613133 STD: 1.0572606068864172\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  2\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.6426528530784501 STD: 0.39830315890859475\n",
      "Hausdorff mean: 1.0157447730378293 STD: 1.195790954624197\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  3\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.7617619287543653 STD: 0.3571505490020753\n",
      "Hausdorff mean: 0.7539176039175626 STD: 1.1884598103100552\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  4\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.8854166666666667 STD: 0.2714336295999038\n",
      "Hausdorff mean: 0.2879575780421117 STD: 0.6850697809485278\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  5\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.6436106554408503 STD: 0.40919141458704444\n",
      "Hausdorff mean: 1.000554805849249 STD: 1.2018160975327303\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  6\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.8254168030510897 STD: 0.30651224530958254\n",
      "Hausdorff mean: 0.49403509773237747 STD: 0.897569712226706\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  7\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.6237157073366304 STD: 0.4287474736697943\n",
      "Hausdorff mean: 1.0173837010985636 STD: 1.2226488176070383\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  8\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.7503502208768343 STD: 0.30342152321935834\n",
      "Hausdorff mean: 0.9764435667940474 STD: 1.2521418375757218\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  9\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.6122960823928063 STD: 0.4473456007107438\n",
      "Hausdorff mean: 1.2574176903689769 STD: 1.5224590628490282\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  10\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.6971541630618822 STD: 0.3534948182575851\n",
      "Hausdorff mean: 1.1061978035944493 STD: 1.355665686122265\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  11\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.6120624596998318 STD: 0.4116499185284273\n",
      "Hausdorff mean: 1.4698380612278115 STD: 1.6381316996484432\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  12\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.7359363146270174 STD: 0.3370199459630864\n",
      "Hausdorff mean: 0.8781202387887068 STD: 1.057988283319868\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  13\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.8177083333333334 STD: 0.3250517453464568\n",
      "Hausdorff mean: 0.5075835389021164 STD: 0.9328699997140553\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  14\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.6399219897205684 STD: 0.40199798354733673\n",
      "Hausdorff mean: 1.0529094687687182 STD: 1.258580558497623\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  15\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.7390117980772808 STD: 0.3169707167061711\n",
      "Hausdorff mean: 1.0381246906136705 STD: 1.3393714431599457\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  16\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.7561171458941867 STD: 0.40760703668428055\n",
      "Hausdorff mean: 0.6103802069319976 STD: 1.0601212839834506\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  17\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.6021761711788708 STD: 0.40332086606892664\n",
      "Hausdorff mean: 1.527366830308233 STD: 1.6434661546856766\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  18\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.8020833333333334 STD: 0.3546259277197751\n",
      "Hausdorff mean: 0.4118023695456654 STD: 0.7370470292814684\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  19\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.7552083333333334 STD: 0.387487398988664\n",
      "Hausdorff mean: 0.6542881158416127 STD: 1.0529257742099545\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  20\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.7067525456355591 STD: 0.3041955061444166\n",
      "Hausdorff mean: 1.1787057467693105 STD: 1.3674960267455165\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  21\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.7620760477743523 STD: 0.28304126427430504\n",
      "Hausdorff mean: 0.8720258548931609 STD: 1.0439112497322869\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  22\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.750799320616983 STD: 0.3393383888181792\n",
      "Hausdorff mean: 0.9002053927947227 STD: 1.2659614620831101\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  23\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.7049622577096485 STD: 0.3090953117829719\n",
      "Hausdorff mean: 1.1915425333438248 STD: 1.3873771875011325\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  24\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.773981204027754 STD: 0.33665166612937014\n",
      "Hausdorff mean: 0.7452690390642589 STD: 1.1775095721515918\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  25\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.7552083333333333 STD: 0.4092769178516072\n",
      "Hausdorff mean: 0.647759755014214 STD: 1.1330929279819326\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  26\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.6539488521904029 STD: 0.39611956002295223\n",
      "Hausdorff mean: 1.2289669098275768 STD: 1.497019765137785\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  27\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.5719851836026941 STD: 0.46120241314905114\n",
      "Hausdorff mean: 1.6153614146966242 STD: 1.7971693678384186\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  28\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.7246062411787366 STD: 0.3842261575191459\n",
      "Hausdorff mean: 0.9010564716659294 STD: 1.266507245699617\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  29\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.7701486865028893 STD: 0.28615745742081145\n",
      "Hausdorff mean: 1.03866349949521 STD: 1.3511102433605027\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  30\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.8854166666666667 STD: 0.2714336295999038\n",
      "Hausdorff mean: 0.2879575780421117 STD: 0.6850697809485278\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  31\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.7686280934837879 STD: 0.2929933799989545\n",
      "Hausdorff mean: 1.0249766271255878 STD: 1.3446335236852924\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  32\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.5860404031441414 STD: 0.44574746405265303\n",
      "Hausdorff mean: 1.601618253709757 STD: 1.7827835044506397\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  33\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.7591184188746396 STD: 0.29851489692463146\n",
      "Hausdorff mean: 0.8414747919804728 STD: 1.0720289242860863\n",
      "--------------------------------\n",
      "\n",
      "patient idx:  34\n",
      "dice shape:  (128,)\n",
      "hausdorff shape:  (128,)\n",
      "Dice mean: 0.7799460717377991 STD: 0.34207789030503477\n",
      "Hausdorff mean: 0.6239820678523362 STD: 0.9930505997141355\n",
      "--------------------------------\n",
      "\n"
     ]
    }
   ],
   "source": [
    "for idx in range(output_onehot.shape[0]):\n",
    "    dice = []\n",
    "    hausdorff = []\n",
    "    for sl in range(output_onehot.shape[1]):\n",
    "        seg = output_onehot[idx][sl]\n",
    "        gt = ground_truth_onehot[idx][sl]\n",
    "        d, h = evaluation_return(gt,seg)\n",
    "        dice.append(d)\n",
    "        hausdorff.append(h)\n",
    "\n",
    "    dice = np.asarray(dice)\n",
    "    print('patient idx: ', idx)\n",
    "    print('dice shape: ',dice.shape)\n",
    "    hausdorff = np.asarray(hausdorff)\n",
    "    print('hausdorff shape: ', hausdorff.shape)\n",
    "    print(\"Dice mean:\", np.mean(dice), \"STD:\",np.std(dice))\n",
    "    print(\"Hausdorff mean:\", np.mean(hausdorff), \"STD:\",np.std(hausdorff))\n",
    "    print(\"--------------------------------\\n\")\n",
    "    with open('testing_output_flair_t1ne.csv', 'a', newline='') as csvfile:\n",
    "        writer = csv.writer(csvfile)\n",
    "        writer.writerow([idx, np.mean(dice), np.std(dice), np.mean(hausdorff), np.std(hausdorff)])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n",
      "[0.   0.25 0.5  1.  ]\n",
      "[0. 1.]\n"
     ]
    }
   ],
   "source": [
    "for i in range(test_output.shape[0]):\n",
    "    print(np.unique(test_output[i]))\n",
    "    for j in range(128):\n",
    "        for k in range(128):\n",
    "            for l in range(128):\n",
    "                if test_output[i,j,k,l]!=0:\n",
    "                    test_output[i,j,k,l] = 1\n",
    "    print(np.unique(test_output[i]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Patient: 2\n",
      "Segmentation resullts: \n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1728x1728 with 25 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "rows = 5\n",
    "cols = 5\n",
    "fig,ax = plt.subplots(rows,cols,figsize=[24,24])\n",
    "num = 2\n",
    "print('Patient: {}'.format(num))\n",
    "print('Segmentation resullts: \\n')\n",
    "for i in range(rows*cols):\n",
    "    #ind = start_with + i*show_every\n",
    "    #ax[int(i/rows),int(i % rows)].set_title('slice %d' % ind)\n",
    "    ax[int(i/rows),int(i % rows)].imshow(test_output[num][i+60],cmap=pylab.cm.gray,vmin = 0,vmax = 1)\n",
    "    #ax[int(i/rows),int(i % rows)].axis('off')\n",
    "    ax[int(i/rows),int(i % rows)].set_xlabel('CT idx: {}'.format(i+60))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "### flair CT model result\n",
    "num=0\n",
    "slice_num=50\n",
    "fig = plt.figure(figsize=(20, 20))\n",
    "fig.add_subplot(1,3,1)\n",
    "plt.imshow(data_3d[num][slice_num],cmap=pylab.cm.gray,vmin = 0,vmax = 1)\n",
    "plt.xlabel('original CT')\n",
    "fig.add_subplot(1,3,2)\n",
    "plt.imshow(resize_nii_label[num][slice_num],cmap=pylab.cm.gray,vmin = 0,vmax = 1)\n",
    "plt.xlabel('Ground truth CT')\n",
    "fig.add_subplot(1,3,3)\n",
    "plt.imshow(test_output[num][slice_num],cmap=pylab.cm.gray,vmin = 0,vmax = 1)\n",
    "plt.xlabel('Ouput segmentation CT')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1\n",
      "1\n",
      "(16384, 128)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# flair + t1ce CT model result\n",
    "\n",
    "image_seg_h = data_3d.shape[2]//input_height\n",
    "image_seg_w = data_3d.shape[3]//input_width\n",
    "\n",
    "print(image_seg_h)\n",
    "print(image_seg_w)\n",
    "for i in range(test_output.shape[1]//(image_seg_h*image_seg_w)):\n",
    "    for j in range(image_seg_h):\n",
    "        for k in range(image_seg_w):\n",
    "            if k == 0:\n",
    "                ww = test_output[0][i*(image_seg_h*image_seg_w)+j*image_seg_w+k]\n",
    "            else:\n",
    "                ww = np.concatenate((ww,test_output[0][i*(image_seg_h*image_seg_w)+j*image_seg_w+k]),axis = 2)\n",
    "        if j == 0:\n",
    "            hh = ww\n",
    "        else:\n",
    "            hh = np.concatenate((hh,ww), axis = 1)\n",
    "    if i == 0:\n",
    "        test_resized = hh\n",
    "    else:\n",
    "        test_resized = np.concatenate((test_resized,hh), axis = 0)\n",
    "\n",
    "print(test_resized.shape)\n",
    "\n",
    "num=1\n",
    "slice_num=56\n",
    "fig = plt.figure(figsize=(20, 20))\n",
    "fig.add_subplot(1,3,1)\n",
    "plt.imshow(data_3d[num][slice_num],cmap=pylab.cm.gray,vmin = 0,vmax = 1)\n",
    "plt.xlabel('original CT')\n",
    "fig.add_subplot(1,3,2)\n",
    "plt.imshow(resize_nii_label[num][slice_num],cmap=pylab.cm.gray,vmin = 0,vmax = 1)\n",
    "plt.xlabel('Ground truth CT')\n",
    "fig.add_subplot(1,3,3)\n",
    "plt.imshow(test_output[num][slice_num],cmap=pylab.cm.gray,vmin = 0,vmax = 1)\n",
    "plt.xlabel('Ouput segmentation CT')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([], dtype=int64), array([], dtype=int64))"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.where(test_output[4][60]!=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x720 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1770.0 -1024.0\n",
      "600 0\n",
      "red ROI 550.0 0.0\n",
      "red ROI 0.0 0.0\n",
      "blue ROI 600.0 0.0\n",
      "blue ROI 0.0 0.0\n",
      "green ROI 205.0 0.0\n",
      "green ROI 0.0 0.0\n"
     ]
    }
   ],
   "source": [
    "t1 = 100\n",
    "ROIrx = 180\n",
    "ROIry = 200\n",
    "ROIbx = 300\n",
    "ROIby = 110\n",
    "ROIgx = 330\n",
    "ROIgy = 230\n",
    "ROIs = 15\n",
    "\n",
    "fig=plt.figure(figsize=(20,10))\n",
    "fig.add_subplot(1,3,1)\n",
    "plt.imshow(norm_nii_CT[t1],cmap = plt.cm.gray,vmin= 0,vmax = 1)\n",
    "fig.add_subplot(1,3,2)\n",
    "rect = patches.Rectangle((ROIrx,ROIry),ROIs,ROIs,linewidth=1,edgecolor='r',facecolor='none')\n",
    "plt.gca().add_patch(rect)\n",
    "rect = patches.Rectangle((ROIbx,ROIby),ROIs,ROIs,linewidth=1,edgecolor='b',facecolor='none')\n",
    "plt.gca().add_patch(rect)\n",
    "rect = patches.Rectangle((ROIgx,ROIgy),ROIs,ROIs,linewidth=1,edgecolor='g',facecolor='none')\n",
    "plt.gca().add_patch(rect)\n",
    "plt.imshow(nii_label[t1]/850,cmap = plt.cm.gray,vmin= 0,vmax = 1)\n",
    "fig.add_subplot(1,3,3)\n",
    "plt.imshow(test_resized[t1]/850,cmap = plt.cm.gray,vmin= 0,vmax = 1)\n",
    "plt.show()\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "print(np.max(nii_CT[t1]),np.min(nii_CT[t1]))\n",
    "print(np.max(nii_label[t1]),np.min(nii_label[t1]))\n",
    "\n",
    "print(\"red ROI\",np.mean(nii_label[t1,ROIry:ROIry+ROIs,ROIrx:ROIrx+ROIs]),\n",
    "      np.std(nii_label[t1,ROIry:ROIry+ROIs,ROIrx:ROIrx+ROIs]))\n",
    "print(\"red ROI\",np.mean(test_resized[t1,ROIry:ROIry+ROIs,ROIrx:ROIrx+ROIs]),\n",
    "      np.std(test_resized[t1,ROIry:ROIry+ROIs,ROIrx:ROIrx+ROIs]))\n",
    "print(\"blue ROI\",np.mean(nii_label[t1,ROIby:ROIby+ROIs,ROIbx:ROIbx+ROIs]),\n",
    "      np.std(nii_label[t1,ROIby:ROIby+ROIs,ROIbx:ROIbx+ROIs]))\n",
    "print(\"blue ROI\",np.mean(test_resized[t1,ROIby:ROIby+ROIs,ROIbx:ROIbx+ROIs]),\n",
    "      np.std(test_resized[t1,ROIby:ROIby+ROIs,ROIbx:ROIbx+ROIs]))\n",
    "print(\"green ROI\",np.mean(nii_label[t1,ROIgy:ROIgy+ROIs,ROIgx:ROIgx+ROIs]),\n",
    "      np.std(nii_label[t1,ROIgy:ROIgy+ROIs,ROIgx:ROIgx+ROIs]))\n",
    "print(\"green ROI\",np.mean(test_resized[t1,ROIgy:ROIgy+ROIs,ROIgx:ROIgx+ROIs]),\n",
    "      np.std(test_resized[t1,ROIgy:ROIgy+ROIs,ROIgx:ROIgx+ROIs]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x720 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1508.0 -1024.0\n",
      "820 0\n",
      "red ROI 820.0 0.0\n",
      "red ROI 0.0 0.0\n",
      "blue ROI 600.0 0.0\n",
      "blue ROI 0.0 0.0\n",
      "green ROI 420.0 0.0\n",
      "green ROI 0.0 0.0\n"
     ]
    }
   ],
   "source": [
    "t2 = 200\n",
    "ROIrx = 180\n",
    "ROIry = 200\n",
    "ROIbx = 270\n",
    "ROIby = 120\n",
    "ROIgx = 190\n",
    "ROIgy = 300\n",
    "ROIts = 15\n",
    "\n",
    "fig=plt.figure(figsize=(20,10))\n",
    "fig.add_subplot(1,3,1)\n",
    "plt.imshow(norm_nii_CT[t2],cmap = plt.cm.gray,vmin= 0,vmax = 1)\n",
    "fig.add_subplot(1,3,2)\n",
    "rect = patches.Rectangle((ROIrx,ROIry),ROIs,ROIs,linewidth=1,edgecolor='r',facecolor='none')\n",
    "plt.gca().add_patch(rect)\n",
    "rect = patches.Rectangle((ROIbx,ROIby),ROIs,ROIs,linewidth=1,edgecolor='b',facecolor='none')\n",
    "plt.gca().add_patch(rect)\n",
    "rect = patches.Rectangle((ROIgx,ROIgy),ROIs,ROIs,linewidth=1,edgecolor='g',facecolor='none')\n",
    "plt.gca().add_patch(rect)\n",
    "plt.imshow(nii_label[t2]/850,cmap = plt.cm.gray,vmin= 0,vmax = 1)\n",
    "fig.add_subplot(1,3,3)\n",
    "plt.imshow(test_resized[t2]/850,cmap = plt.cm.gray,vmin= 0,vmax = 1)\n",
    "plt.show()\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "print(np.max(nii_CT[t2]),np.min(nii_CT[t2]))\n",
    "print(np.max(nii_label[t2]),np.min(nii_label[t2]))\n",
    "\n",
    "print(\"red ROI\",np.mean(nii_label[t2,ROIry:ROIry+ROIs,ROIrx:ROIrx+ROIs]),\n",
    "      np.std(nii_label[t2,ROIry:ROIry+ROIs,ROIrx:ROIrx+ROIs]))\n",
    "print(\"red ROI\",np.mean(test_resized[t2,ROIry:ROIry+ROIs,ROIrx:ROIrx+ROIs]),\n",
    "      np.std(test_resized[t2,ROIry:ROIry+ROIs,ROIrx:ROIrx+ROIs]))\n",
    "print(\"blue ROI\",np.mean(nii_label[t2,ROIby:ROIby+ROIs,ROIbx:ROIbx+ROIs]),\n",
    "      np.std(nii_label[t2,ROIby:ROIby+ROIs,ROIbx:ROIbx+ROIs]))\n",
    "print(\"blue ROI\",np.mean(test_resized[t2,ROIby:ROIby+ROIs,ROIbx:ROIbx+ROIs]),\n",
    "      np.std(test_resized[t2,ROIby:ROIby+ROIs,ROIbx:ROIbx+ROIs]))\n",
    "print(\"green ROI\",np.mean(nii_label[t2,ROIgy:ROIgy+ROIts,ROIgx:ROIgx+ROIts]),\n",
    "      np.std(nii_label[t2,ROIgy:ROIgy+ROIts,ROIgx:ROIgx+ROIts]))\n",
    "print(\"green ROI\",np.mean(test_resized[t2,ROIgy:ROIgy+ROIts,ROIgx:ROIgx+ROIts]),\n",
    "      np.std(test_resized[t2,ROIgy:ROIgy+ROIts,ROIgx:ROIgx+ROIts]))"
   ]
  },
  {
   "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": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}