examples/Convallaria-2-Training.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# PN2V Training\n",
    "Here we will use the generated noise model and train a PN2V network on single noisy images.\n",
    "Please run the 'Convallaria-1-CreateNoiseModel' notebook first."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CUDA available? False\n",
      "WARNING: CUDA not available. Using CPU.\n"
     ]
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "from pn2v.unet import UNet\n",
    "\n",
    "from pn2v import utils\n",
    "from pn2v import histNoiseModel\n",
    "from pn2v import training\n",
    "from tifffile import imread\n",
    "\n",
    "# See if we can use a GPU\n",
    "device=utils.getDevice()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Load Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "path='data/Convallaria_diaphragm/'\n",
    "\n",
    "# Load the training data\n",
    "data=imread(path+'20190520_tl_25um_50msec_05pc_488_130EM_Conv.tif')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# We are loading the histogram from the 'Convallaria-1-CreateNoiseModel' notebook\n",
    "histogram=np.load(path+'noiseModel.npy')\n",
    "\n",
    "# Create a NoiseModel object from the histogram.\n",
    "noiseModel=histNoiseModel.NoiseModel(histogram, device=device)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Create the Network and Train it\n",
    "This can take a while."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/krull/projects/pn2vCheck/pn2v/unet/model.py:215: UserWarning: nn.init.xavier_normal is now deprecated in favor of nn.init.xavier_normal_.\n",
      "  init.xavier_normal(m.weight)\n",
      "/home/krull/projects/pn2vCheck/pn2v/unet/model.py:216: UserWarning: nn.init.constant is now deprecated in favor of nn.init.constant_.\n",
      "  init.constant(m.bias, 0)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 0 finished      \n",
      "avg. loss: 4.089759075641632+-(2SEM)0.12026341607949252      \n",
      "Epoch 1 finished      \n",
      "avg. loss: 3.639772868156433+-(2SEM)0.11503529183822184      \n",
      "Epoch 2 finished      \n",
      "avg. loss: 3.3973047852516176+-(2SEM)0.07890816372323652      \n",
      "Epoch 3 finished      \n",
      "avg. loss: 3.2314443588256836+-(2SEM)0.08818115890623167      \n",
      "Epoch 4 finished      \n",
      "avg. loss: 3.334096837043762+-(2SEM)0.14057031427579395      \n",
      "Epoch 5 finished      \n",
      "avg. loss: 3.1251014113426208+-(2SEM)0.12321941811626058      \n",
      "Epoch 6 finished      \n",
      "avg. loss: 3.1607281804084777+-(2SEM)0.1026232046439824      \n",
      "Epoch 7 finished      \n",
      "avg. loss: 3.0416570782661436+-(2SEM)0.08162758531355314      \n",
      "Epoch 8 finished      \n",
      "avg. loss: 2.9617292165756224+-(2SEM)0.0941533955891487      \n",
      "Epoch 9 finished      \n",
      "avg. loss: 3.0457193732261656+-(2SEM)0.10268263079265588      \n",
      "Epoch 10 finished      \n",
      "avg. loss: 3.025261652469635+-(2SEM)0.08855933389405661      \n",
      "Epoch 11 finished      \n",
      "avg. loss: 2.9870091915130614+-(2SEM)0.08896224698616402      \n",
      "Epoch 12 finished      \n",
      "avg. loss: 3.051213276386261+-(2SEM)0.0881302175373901      \n",
      "Epoch 13 finished      \n",
      "avg. loss: 2.9096617102622986+-(2SEM)0.0977623208284172      \n",
      "Epoch 14 finished      \n",
      "avg. loss: 2.997691571712494+-(2SEM)0.11812253481325903      \n",
      "Epoch 15 finished      \n",
      "avg. loss: 3.0157735466957094+-(2SEM)0.13413440979063318      \n",
      "Epoch 16 finished      \n",
      "avg. loss: 2.9635536313056945+-(2SEM)0.06400925765932906      \n",
      "Epoch 17 finished      \n",
      "avg. loss: 2.9656522035598756+-(2SEM)0.08645031995813347      \n",
      "Epoch 18 finished      \n",
      "avg. loss: 2.927971827983856+-(2SEM)0.13055766545727737      \n",
      "Epoch 19 finished      \n",
      "avg. loss: 2.995678687095642+-(2SEM)0.11014297336291633      \n",
      "Epoch 20 finished      \n",
      "avg. loss: 2.988750946521759+-(2SEM)0.08052880931019418      \n",
      "Epoch 21 finished      \n",
      "avg. loss: 3.0664520263671875+-(2SEM)0.08604592753320843      \n",
      "Epoch 22 finished      \n",
      "avg. loss: 3.0337172746658325+-(2SEM)0.08370304116065565      \n",
      "Epoch 23 finished      \n",
      "avg. loss: 2.997710943222046+-(2SEM)0.1014597047306437      \n",
      "Epoch 24 finished      \n",
      "avg. loss: 2.9983714699745176+-(2SEM)0.0904080531823383      \n",
      "Epoch 25 finished      \n",
      "avg. loss: 2.9178090333938598+-(2SEM)0.10584815819328003      \n",
      "Epoch 26 finished      \n",
      "avg. loss: 2.9022466063499452+-(2SEM)0.09111648666722665      \n",
      "Epoch 27 finished      \n",
      "avg. loss: 2.9287564158439636+-(2SEM)0.12909194328131446      \n",
      "Epoch 28 finished      \n",
      "avg. loss: 3.027549624443054+-(2SEM)0.11459828150850462      \n",
      "Epoch 29 finished      \n",
      "avg. loss: 3.0246501088142397+-(2SEM)0.10887209209847269      \n",
      "Epoch 30 finished      \n",
      "avg. loss: 2.96717187166214+-(2SEM)0.10739364882706078      \n",
      "Epoch 31 finished      \n",
      "avg. loss: 2.9544616460800173+-(2SEM)0.11405254699325551      \n",
      "Epoch 32 finished      \n",
      "avg. loss: 2.931500828266144+-(2SEM)0.08307307013663191      \n",
      "Epoch 33 finished      \n",
      "avg. loss: 2.9029569745063784+-(2SEM)0.08851462558699896      \n",
      "Epoch 34 finished      \n",
      "avg. loss: 3.007542669773102+-(2SEM)0.14001407958491321      \n",
      "Epoch 35 finished      \n",
      "avg. loss: 2.9935431838035584+-(2SEM)0.09598404987266555      \n",
      "Epoch 36 finished      \n",
      "avg. loss: 2.951803183555603+-(2SEM)0.08176128409536591      \n",
      "Epoch    36: reducing learning rate of group 0 to 5.0000e-04.\n",
      "Epoch 37 finished      \n",
      "avg. loss: 3.0197945713996885+-(2SEM)0.0990595647321066      \n",
      "Epoch 38 finished      \n",
      "avg. loss: 3.020003283023834+-(2SEM)0.11451035584360768      \n",
      "Epoch 39 finished      \n",
      "avg. loss: 2.892862045764923+-(2SEM)0.10120118578102297      \n",
      "Epoch 40 finished      \n",
      "avg. loss: 3.022839605808258+-(2SEM)0.08264252363867167      \n",
      "Epoch 41 finished      \n",
      "avg. loss: 2.918239426612854+-(2SEM)0.08953168669491553      \n",
      "Epoch 42 finished      \n",
      "avg. loss: 3.0198025584220884+-(2SEM)0.0961474203479445      \n",
      "Epoch 43 finished      \n",
      "avg. loss: 2.928452789783478+-(2SEM)0.12455311152672134      \n",
      "Epoch 44 finished      \n",
      "avg. loss: 2.9206276893615724+-(2SEM)0.12289149631474616      \n",
      "Epoch 45 finished      \n",
      "avg. loss: 2.998249554634094+-(2SEM)0.11175756779348466      \n",
      "Epoch 46 finished      \n",
      "avg. loss: 2.930358648300171+-(2SEM)0.09022654862911957      \n",
      "Epoch 47 finished      \n",
      "avg. loss: 3.0269295930862428+-(2SEM)0.07647065707562624      \n",
      "Epoch 48 finished      \n",
      "avg. loss: 2.929220449924469+-(2SEM)0.06290119713166868      \n",
      "Epoch 49 finished      \n",
      "avg. loss: 2.969872844219208+-(2SEM)0.13444405207930554      \n",
      "Epoch 50 finished      \n",
      "avg. loss: 2.9737980127334596+-(2SEM)0.09209894826667894      \n",
      "Epoch 51 finished      \n",
      "avg. loss: 2.9641520023345946+-(2SEM)0.09798419941889283      \n",
      "Epoch 52 finished      \n",
      "avg. loss: 2.9324525237083434+-(2SEM)0.09206343913889588      \n",
      "Epoch 53 finished      \n",
      "avg. loss: 2.9484452843666076+-(2SEM)0.11917628437746233      \n",
      "Epoch 54 finished      \n",
      "avg. loss: 2.967315149307251+-(2SEM)0.10648853559026822      \n",
      "Epoch 55 finished      \n",
      "avg. loss: 2.8632136821746825+-(2SEM)0.07750972082694904      \n",
      "Epoch 56 finished      \n",
      "avg. loss: 2.888731133937836+-(2SEM)0.10394449174803998      \n",
      "Epoch 57 finished      \n",
      "avg. loss: 2.9218475103378294+-(2SEM)0.050729481637047595      \n",
      "Epoch 58 finished      \n",
      "avg. loss: 2.9589900255203245+-(2SEM)0.12558451763561543      \n",
      "Epoch 59 finished      \n",
      "avg. loss: 3.0167863368988037+-(2SEM)0.12967192471884412      \n",
      "Epoch 60 finished      \n",
      "avg. loss: 2.983299136161804+-(2SEM)0.08858588098855863      \n",
      "Epoch 61 finished      \n",
      "avg. loss: 2.9855345129966735+-(2SEM)0.09519505687622325      \n",
      "Epoch 62 finished      \n",
      "avg. loss: 2.9410678863525392+-(2SEM)0.12253680992814611      \n",
      "Epoch 63 finished      \n",
      "avg. loss: 2.9415801763534546+-(2SEM)0.08682237715956978      \n",
      "Epoch 64 finished      \n",
      "avg. loss: 2.9200940847396852+-(2SEM)0.10551902214755247      \n",
      "Epoch    64: reducing learning rate of group 0 to 2.5000e-04.\n",
      "Epoch 65 finished      \n",
      "avg. loss: 2.9952526330947875+-(2SEM)0.10985422549065603      \n",
      "Epoch 66 finished      \n",
      "avg. loss: 3.001185750961304+-(2SEM)0.09271204845090937      \n",
      "Epoch 67 finished      \n",
      "avg. loss: 2.8931732773780823+-(2SEM)0.09343820817004249      \n",
      "Epoch 68 finished      \n",
      "avg. loss: 2.996389961242676+-(2SEM)0.08180237316048478      \n",
      "Epoch 69 finished      \n",
      "avg. loss: 2.9134845972061156+-(2SEM)0.08536764646820404      \n",
      "Epoch 70 finished      \n",
      "avg. loss: 2.907607614994049+-(2SEM)0.09127392144537626      \n",
      "Epoch 71 finished      \n",
      "avg. loss: 2.9305078744888307+-(2SEM)0.10023540293779712      \n",
      "Epoch 72 finished      \n",
      "avg. loss: 3.0122191429138185+-(2SEM)0.0918383243756239      \n",
      "Epoch 73 finished      \n",
      "avg. loss: 2.970137321949005+-(2SEM)0.098488873481677      \n",
      "Epoch 74 finished      \n",
      "avg. loss: 3.032272529602051+-(2SEM)0.11286121416410165      \n",
      "Epoch 75 finished      \n",
      "avg. loss: 2.9161924004554747+-(2SEM)0.07211180498506733      \n",
      "Epoch 76 finished      \n",
      "avg. loss: 2.935866963863373+-(2SEM)0.10488215375729902      \n",
      "Epoch 77 finished      \n",
      "avg. loss: 2.9782342314720154+-(2SEM)0.1379591878352872      \n",
      "Epoch 78 finished      \n",
      "avg. loss: 2.889685130119324+-(2SEM)0.09883110626729966      \n",
      "Epoch    78: reducing learning rate of group 0 to 1.2500e-04.\n",
      "Epoch 79 finished      \n",
      "avg. loss: 2.9279857873916626+-(2SEM)0.10862439654222374      \n",
      "Epoch 80 finished      \n",
      "avg. loss: 3.0403643250465393+-(2SEM)0.08393977371316047      \n",
      "Epoch 81 finished      \n",
      "avg. loss: 2.953695058822632+-(2SEM)0.11909248436538357      \n",
      "Epoch 82 finished      \n",
      "avg. loss: 2.970161497592926+-(2SEM)0.09477417339837603      \n",
      "Epoch 83 finished      \n",
      "avg. loss: 2.927616548538208+-(2SEM)0.10342503631604609      \n",
      "Epoch 84 finished      \n",
      "avg. loss: 2.937042200565338+-(2SEM)0.115868540004102      \n",
      "Epoch 85 finished      \n",
      "avg. loss: 3.0033146858215334+-(2SEM)0.10232241674959837      \n",
      "Epoch 86 finished      \n",
      "avg. loss: 2.8677237153053285+-(2SEM)0.10039897146817525      \n",
      "Epoch 87 finished      \n",
      "avg. loss: 3.020893168449402+-(2SEM)0.0927669957751763      \n",
      "Epoch 88 finished      \n",
      "avg. loss: 2.9566489696502685+-(2SEM)0.06505744247399924      \n",
      "Epoch 89 finished      \n",
      "avg. loss: 3.011168348789215+-(2SEM)0.11322789224366803      \n",
      "Epoch    89: reducing learning rate of group 0 to 6.2500e-05.\n",
      "Epoch 90 finished      \n",
      "avg. loss: 2.91367267370224+-(2SEM)0.10101909758718858      \n",
      "Epoch 91 finished      \n",
      "avg. loss: 2.879884123802185+-(2SEM)0.0660901566342759      \n",
      "Epoch 92 finished      \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "avg. loss: 3.0089614391326904+-(2SEM)0.07491113073267992      \n",
      "Epoch 93 finished      \n",
      "avg. loss: 2.9971357703208925+-(2SEM)0.09450119084365638      \n",
      "Epoch 94 finished      \n",
      "avg. loss: 2.920409154891968+-(2SEM)0.07208078122480653      \n",
      "Epoch 95 finished      \n",
      "avg. loss: 3.0453878045082092+-(2SEM)0.10026367465819444      \n",
      "Epoch 96 finished      \n",
      "avg. loss: 2.9590693354606628+-(2SEM)0.07328501094504418      \n",
      "Epoch 97 finished      \n",
      "avg. loss: 2.9206487417221068+-(2SEM)0.11063813904317488      \n",
      "Epoch 98 finished      \n",
      "avg. loss: 2.9798823237419128+-(2SEM)0.059885246063534917      \n",
      "Epoch 99 finished      \n",
      "avg. loss: 2.97254341840744+-(2SEM)0.08273064061293679      \n",
      "Epoch 100 finished      \n",
      "avg. loss: 3.001618003845215+-(2SEM)0.09691676992937631      \n",
      "Epoch   100: reducing learning rate of group 0 to 3.1250e-05.\n",
      "Epoch 101 finished      \n",
      "avg. loss: 2.977292609214783+-(2SEM)0.12349163769554848      \n",
      "Epoch 102 finished      \n",
      "avg. loss: 3.0236170887947083+-(2SEM)0.08818519801089808      \n",
      "Epoch 103 finished      \n",
      "avg. loss: 2.9905953764915467+-(2SEM)0.11713047633086159      \n",
      "Epoch 104 finished      \n",
      "avg. loss: 2.9985088229179384+-(2SEM)0.10631462278400923      \n",
      "Epoch 105 finished      \n",
      "avg. loss: 2.9183295249938963+-(2SEM)0.07029312640571173      \n",
      "Epoch 106 finished      \n",
      "avg. loss: 3.0141526579856874+-(2SEM)0.10975248854595869      \n",
      "Epoch 107 finished      \n",
      "avg. loss: 2.915389156341553+-(2SEM)0.0907509541841944      \n",
      "Epoch 108 finished      \n",
      "avg. loss: 2.9458510279655457+-(2SEM)0.10673232456885816      \n",
      "Epoch 109 finished      \n",
      "avg. loss: 3.0240073204040527+-(2SEM)0.11103904348226154      \n",
      "Epoch 110 finished      \n",
      "avg. loss: 2.875267136096954+-(2SEM)0.08721655324789034      \n",
      "Epoch 111 finished      \n",
      "avg. loss: 3.0131166577339172+-(2SEM)0.09645899199333588      \n",
      "Epoch   111: reducing learning rate of group 0 to 1.5625e-05.\n",
      "Epoch 112 finished      \n",
      "avg. loss: 3.027109694480896+-(2SEM)0.07764170393576166      \n",
      "Epoch 113 finished      \n",
      "avg. loss: 2.980350303649902+-(2SEM)0.10797166373469799      \n",
      "Epoch 114 finished      \n",
      "avg. loss: 2.8325629472732543+-(2SEM)0.08258636295085137      \n",
      "Epoch 115 finished      \n",
      "avg. loss: 2.998828709125519+-(2SEM)0.10630043445957904      \n",
      "Epoch 116 finished      \n",
      "avg. loss: 3.003774440288544+-(2SEM)0.11736443405866422      \n",
      "Epoch 117 finished      \n",
      "avg. loss: 3.0051623940467835+-(2SEM)0.064278699797727      \n",
      "Epoch 118 finished      \n",
      "avg. loss: 3.0444867849349975+-(2SEM)0.12400509249799377      \n",
      "Epoch 119 finished      \n",
      "avg. loss: 3.016837990283966+-(2SEM)0.11288973440518274      \n",
      "Epoch 120 finished      \n",
      "avg. loss: 2.9872026443481445+-(2SEM)0.10675874591242085      \n",
      "Epoch 121 finished      \n",
      "avg. loss: 2.979781150817871+-(2SEM)0.10955531850442106      \n",
      "Epoch 122 finished      \n",
      "avg. loss: 2.888844680786133+-(2SEM)0.09184247651285925      \n",
      "Epoch   122: reducing learning rate of group 0 to 7.8125e-06.\n",
      "Epoch 123 finished      \n",
      "avg. loss: 2.9039318680763246+-(2SEM)0.06895691346849624      \n",
      "Epoch 124 finished      \n",
      "avg. loss: 2.982909059524536+-(2SEM)0.0697591549793121      \n",
      "Epoch 125 finished      \n",
      "avg. loss: 2.9955235958099364+-(2SEM)0.10808265778216673      \n",
      "Epoch 126 finished      \n",
      "avg. loss: 2.9158218264579774+-(2SEM)0.11329833364223338      \n",
      "Epoch 127 finished      \n",
      "avg. loss: 2.938697564601898+-(2SEM)0.09309876629346536      \n",
      "Epoch 128 finished      \n",
      "avg. loss: 2.9811004519462587+-(2SEM)0.06471542770954475      \n",
      "Epoch 129 finished      \n",
      "avg. loss: 2.8753915309906004+-(2SEM)0.10705849370944472      \n",
      "Epoch 130 finished      \n",
      "avg. loss: 2.9987592577934263+-(2SEM)0.10318240767172851      \n",
      "Epoch 131 finished      \n",
      "avg. loss: 2.9432009935379027+-(2SEM)0.11684603458220547      \n",
      "Epoch 132 finished      \n",
      "avg. loss: 2.977895426750183+-(2SEM)0.10190329456411601      \n",
      "Epoch 133 finished      \n",
      "avg. loss: 2.9414501667022703+-(2SEM)0.10698963062079299      \n",
      "Epoch   133: reducing learning rate of group 0 to 3.9063e-06.\n",
      "Epoch 134 finished      \n",
      "avg. loss: 2.9306083798408507+-(2SEM)0.10201099109614492      \n",
      "Epoch 135 finished      \n",
      "avg. loss: 2.9399945616722105+-(2SEM)0.13114161960737208      \n",
      "Epoch 136 finished      \n",
      "avg. loss: 3.00670440196991+-(2SEM)0.08965656216836293      \n",
      "Epoch 137 finished      \n",
      "avg. loss: 2.947170877456665+-(2SEM)0.12397515903462177      \n",
      "Epoch 138 finished      \n",
      "avg. loss: 2.904181218147278+-(2SEM)0.06848556230438888      \n",
      "Epoch 139 finished      \n",
      "avg. loss: 2.9712830901145937+-(2SEM)0.10259056996831885      \n",
      "Epoch 140 finished      \n",
      "avg. loss: 2.9634138226509092+-(2SEM)0.12432872964007328      \n",
      "Epoch 141 finished      \n",
      "avg. loss: 2.840677535533905+-(2SEM)0.08983691457052383      \n",
      "Epoch 142 finished      \n",
      "avg. loss: 2.982657718658447+-(2SEM)0.07461965218044882      \n",
      "Epoch 143 finished      \n",
      "avg. loss: 2.9271892428398134+-(2SEM)0.09882328746971107      \n",
      "Epoch 144 finished      \n",
      "avg. loss: 2.958388066291809+-(2SEM)0.09097974018299909      \n",
      "Epoch   144: reducing learning rate of group 0 to 1.9531e-06.\n",
      "Epoch 145 finished      \n",
      "avg. loss: 2.907707679271698+-(2SEM)0.09238008002340867      \n",
      "Epoch 146 finished      \n",
      "avg. loss: 2.9553367853164674+-(2SEM)0.08946802941871948      \n",
      "Epoch 147 finished      \n",
      "avg. loss: 2.92761070728302+-(2SEM)0.07303647694112145      \n",
      "Epoch 148 finished      \n",
      "avg. loss: 2.8966461420059204+-(2SEM)0.10702426847745199      \n",
      "Epoch 149 finished      \n",
      "avg. loss: 2.9828192472457884+-(2SEM)0.099961132031664      \n",
      "Epoch 150 finished      \n",
      "avg. loss: 3.0573457002639772+-(2SEM)0.10060097285436709      \n",
      "Epoch 151 finished      \n",
      "avg. loss: 2.9951697111129763+-(2SEM)0.08399612490849817      \n",
      "Epoch 152 finished      \n",
      "avg. loss: 2.9895230650901796+-(2SEM)0.10822317392289887      \n",
      "Epoch 153 finished      \n",
      "avg. loss: 3.0375745058059693+-(2SEM)0.12421810912239446      \n",
      "Epoch 154 finished      \n",
      "avg. loss: 2.9791797399520874+-(2SEM)0.10190545455153598      \n",
      "Epoch 155 finished      \n",
      "avg. loss: 2.9958006620407103+-(2SEM)0.11178481295367318      \n",
      "Epoch   155: reducing learning rate of group 0 to 9.7656e-07.\n",
      "Epoch 156 finished      \n",
      "avg. loss: 2.920079064369202+-(2SEM)0.09179807142786146      \n",
      "Epoch 157 finished      \n",
      "avg. loss: 2.922919309139252+-(2SEM)0.08089467322134677      \n",
      "Epoch 158 finished      \n",
      "avg. loss: 2.9678664445877074+-(2SEM)0.1118334395302015      \n",
      "Epoch 159 finished      \n",
      "avg. loss: 2.955147123336792+-(2SEM)0.10061142639335897      \n",
      "Epoch 160 finished      \n",
      "avg. loss: 2.9343526363372803+-(2SEM)0.08560988359728151      \n",
      "Epoch 161 finished      \n",
      "avg. loss: 2.917695689201355+-(2SEM)0.10865782175115218      \n",
      "Epoch 162 finished      \n",
      "avg. loss: 2.9456443905830385+-(2SEM)0.0680617994091327      \n",
      "Epoch 163 finished      \n",
      "avg. loss: 2.9488882303237913+-(2SEM)0.0597228295187634      \n",
      "Epoch 164 finished      \n",
      "avg. loss: 2.9956018209457396+-(2SEM)0.11237803029298478      \n",
      "Epoch 165 finished      \n",
      "avg. loss: 2.979853641986847+-(2SEM)0.08458026890180008      \n",
      "Epoch 166 finished      \n",
      "avg. loss: 3.0152522683143617+-(2SEM)0.11868563936149154      \n",
      "Epoch   166: reducing learning rate of group 0 to 4.8828e-07.\n",
      "Epoch 167 finished      \n",
      "avg. loss: 3.0616050481796266+-(2SEM)0.1055395981017882      \n",
      "Epoch 168 finished      \n",
      "avg. loss: 2.956546902656555+-(2SEM)0.10475556534640035      \n",
      "Epoch 169 finished      \n",
      "avg. loss: 2.955781090259552+-(2SEM)0.0742156466038424      \n",
      "Epoch 170 finished      \n",
      "avg. loss: 3.00222190618515+-(2SEM)0.09844189494733294      \n",
      "Epoch 171 finished      \n",
      "avg. loss: 2.9644793272018433+-(2SEM)0.10232366091993182      \n",
      "Epoch 172 finished      \n",
      "avg. loss: 2.9830072283744813+-(2SEM)0.1078193417310335      \n",
      "Epoch 173 finished      \n",
      "avg. loss: 2.9357664585113525+-(2SEM)0.09839980588204583      \n",
      "Epoch 174 finished      \n",
      "avg. loss: 2.8397382736206054+-(2SEM)0.0831821828868677      \n",
      "Epoch 175 finished      \n",
      "avg. loss: 3.0157333612442017+-(2SEM)0.08469263941433744      \n",
      "Epoch 176 finished      \n",
      "avg. loss: 2.97558308839798+-(2SEM)0.06542965205700228      \n",
      "Epoch 177 finished      \n",
      "avg. loss: 3.003683340549469+-(2SEM)0.08805350400625367      \n",
      "Epoch   177: reducing learning rate of group 0 to 2.4414e-07.\n",
      "Epoch 178 finished      \n",
      "avg. loss: 2.946373665332794+-(2SEM)0.07779416949284611      \n",
      "Epoch 179 finished      \n",
      "avg. loss: 2.9483773350715636+-(2SEM)0.10244581437152825      \n",
      "Epoch 180 finished      \n",
      "avg. loss: 2.99813392162323+-(2SEM)0.11964944368307134      \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 181 finished      \n",
      "avg. loss: 2.9206062078475954+-(2SEM)0.1185963221519298      \n",
      "Epoch 182 finished      \n",
      "avg. loss: 2.970108759403229+-(2SEM)0.06868438829663272      \n",
      "Epoch 183 finished      \n",
      "avg. loss: 2.9398975372314453+-(2SEM)0.09176110833871043      \n",
      "Epoch 184 finished      \n",
      "avg. loss: 2.9660026907920836+-(2SEM)0.1052472006213557      \n",
      "Epoch 185 finished      \n",
      "avg. loss: 2.901454448699951+-(2SEM)0.10341140633291729      \n",
      "Epoch 186 finished      \n",
      "avg. loss: 2.946291136741638+-(2SEM)0.10480967986960914      \n",
      "Epoch 187 finished      \n",
      "avg. loss: 2.9465633273124694+-(2SEM)0.08113772229513888      \n",
      "Epoch 188 finished      \n",
      "avg. loss: 2.9161665439605713+-(2SEM)0.1113127110756062      \n",
      "Epoch   188: reducing learning rate of group 0 to 1.2207e-07.\n",
      "Epoch 189 finished      \n",
      "avg. loss: 2.9172777414321898+-(2SEM)0.09284095001376395      \n",
      "Epoch 190 finished      \n",
      "avg. loss: 3.002373516559601+-(2SEM)0.08709477385801133      \n",
      "Epoch 191 finished      \n",
      "avg. loss: 2.966343343257904+-(2SEM)0.12040276770614845      \n",
      "Epoch 192 finished      \n",
      "avg. loss: 2.9930580973625185+-(2SEM)0.0839043070264814      \n",
      "Epoch 193 finished      \n",
      "avg. loss: 2.974967324733734+-(2SEM)0.14269362522566412      \n",
      "Epoch 194 finished      \n",
      "avg. loss: 2.9176560401916505+-(2SEM)0.085864218575393      \n",
      "Epoch 195 finished      \n",
      "avg. loss: 3.032939314842224+-(2SEM)0.0986732328334115      \n",
      "Epoch 196 finished      \n",
      "avg. loss: 3.065095841884613+-(2SEM)0.11963269702255767      \n",
      "Epoch 197 finished      \n",
      "avg. loss: 3.0053364753723146+-(2SEM)0.0911657818066349      \n",
      "Epoch 198 finished      \n",
      "avg. loss: 2.958442759513855+-(2SEM)0.10141067428128789      \n",
      "Epoch 199 finished      \n",
      "avg. loss: 2.8893272161483763+-(2SEM)0.10126697371911514      \n",
      "Epoch   199: reducing learning rate of group 0 to 6.1035e-08.\n",
      "Finished Training      \n"
     ]
    }
   ],
   "source": [
    "# Create a network with 800 output channels that are interpreted as samples from the prior.\n",
    "net = UNet(800, depth=3)\n",
    "\n",
    "# Split training and validation data.\n",
    "my_train_data=data[:-5].copy()\n",
    "my_val_data=data[-5:].copy()\n",
    "\n",
    "# Start training.\n",
    "trainHist, valHist = training.trainNetwork(net=net, trainData=my_train_data, valData=my_val_data,\n",
    "                                           postfix='conv', directory=path, noiseModel=noiseModel,\n",
    "                                           device=device, numOfEpochs= 200, stepsPerEpoch=5, virtualBatchSize=20,\n",
    "                                           batchSize=1, learningRate=1e-3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Let's look at the training and validation loss\n",
    "plt.xlabel('epoch')\n",
    "plt.ylabel('loss')\n",
    "plt.plot(valHist, label='validation loss')\n",
    "plt.plot(trainHist, label='training loss')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "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": "3.9.16"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}