Ex1_FOOPSI_CNN.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Most basic training example\n",
    "\n",
    "Here we train on data simulated with a simple linear model of fluorescence dynamics.\n",
    "We use a convolutional neural network to get a factorized estimation of the posterior. \n",
    "\n",
    "At the end of the notebook we also show how to perform supervised training.\n",
    "\n",
    "This notebook can be run on a CPU or GPU."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-01-08T13:44:21.701927Z",
     "start_time": "2018-01-08T13:44:16.557189Z"
    },
    "run_control": {
     "frozen": false,
     "read_only": false
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using cuDNN version 6021 on context None\n",
      "Preallocating 10986/12207 Mb (0.900000) on cuda2\n",
      "Mapped name None to device cuda2: GeForce GTX TITAN X (0000:04:00.0)\n"
     ]
    }
   ],
   "source": [
    "import sys\n",
    "import os\n",
    "sys.path.append(\"engine/\")\n",
    "sys.path.append(\"funcs/\")\n",
    "sys.setrecursionlimit(10000)\n",
    "\n",
    "import theano.gpuarray\n",
    "theano.gpuarray.use('cuda2')\n",
    "\n",
    "import theano\n",
    "import theano.tensor as T\n",
    "from theano import config\n",
    "import numpy as np\n",
    "import pickle\n",
    "import copy\n",
    "import time\n",
    "from itertools import *\n",
    "from tqdm import tqdm\n",
    "\n",
    "from RecognitionModel import *\n",
    "from VIMCO_alpha import *\n",
    "from plot_funcs import *\n",
    "\n",
    "from matplotlib import pyplot as plt\n",
    "from matplotlib import gridspec\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Simulate some data from a simple linear model:\n",
    "    \n",
    "    C_(t) = s_t + sigmoid(g) * C_(t-1)\n",
    "    F_(t) = softplus(α) * C_(t) + softplus(σ)\n",
    "    \n",
    "Softplus and sigmoid functions are used to keep the parameters in the allowed range and avoid the training to crash.\n",
    "We simulate 100 cells with parameters drawn randomly from a certain parameter range. For every cell we sample three traces with different spiking rates."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-01-08T13:44:43.138249Z",
     "start_time": "2018-01-08T13:44:21.704483Z"
    },
    "run_control": {
     "frozen": false,
     "read_only": false
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 100/100 [00:17<00:00,  5.80it/s]\n"
     ]
    }
   ],
   "source": [
    "rng = np.random.RandomState(1)\n",
    "lasagne.random.set_rng(rng)\n",
    "\n",
    "n_cells = 100\n",
    "timebins = 10000\n",
    "fps = 60\n",
    "fRate = [0.7,1.,1.2]\n",
    "fProb = [f/fps for f in fRate]\n",
    "\n",
    "traces_foopsi = []\n",
    "spikes_foopsi = []\n",
    "\n",
    "TrueParams_foopsi = []\n",
    "InitParameters = {}\n",
    "\n",
    "low = {\"gamma\": 2.75, \"alpha\": -1.25, \"beta\": -0.3, \"sigma\": 0.07}\n",
    "high = {\"gamma\": 4.6, \"alpha\": -0.43, \"beta\": -0.3, \"sigma\": 0.15}\n",
    "\n",
    "gen_model = FOOPSI(low)\n",
    "\n",
    "for i in tqdm(range(n_cells)):\n",
    "    \n",
    "    InitParameters = {}\n",
    "    for p in low:\n",
    "        if low[p] < high[p]:\n",
    "            InitParameters.update({p:rng.randint(1000*low[p],1000*high[p])/1000}) \n",
    "        else:\n",
    "            InitParameters.update({p:low[p]})\n",
    "    \n",
    "    gen_model.set_params(InitParameters)\n",
    "    sim_spikes = []\n",
    "    for f in fProb:\n",
    "        sim_spikes.append(rng.binomial(1, f, (timebins)))\n",
    "        \n",
    "    sim_spikes = np.array(sim_spikes).astype(config.floatX)\n",
    "    sim_fluor = gen_model.genfunc(sim_spikes) + rng.normal(size=(sim_spikes.shape)) * InitParameters['sigma']\n",
    "    traces_foopsi.append(copy.deepcopy(sim_fluor))\n",
    "    spikes_foopsi.append(copy.deepcopy(sim_spikes))\n",
    "    TrueParams_foopsi.append(copy.deepcopy(InitParameters))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-01-08T13:44:43.304160Z",
     "start_time": "2018-01-08T13:44:43.140549Z"
    },
    "run_control": {
     "frozen": false,
     "read_only": false
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f083ecd4c50>]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f08426fbbe0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize = (10,5))\n",
    "plt.plot(sim_fluor[0][:1000])\n",
    "plt.plot(sim_spikes[0][:1000]/5-0.7)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Save dataset for later use."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-01-08T13:44:43.589332Z",
     "start_time": "2018-01-08T13:44:43.305780Z"
    },
    "run_control": {
     "frozen": false,
     "read_only": false
    }
   },
   "outputs": [],
   "source": [
    "data = {'traces': np.array(traces_foopsi).astype(config.floatX), 'spikes': np.array(spikes_foopsi).astype(config.floatX), 'params': TrueParams_foopsi, 'fps':[fps for _ in range(n_cells)], 'spike_fps':60}\n",
    "\n",
    "with open('foopsi_100.pkl', 'wb') as f:\n",
    "    pickle.dump(data, f, pickle.HIGHEST_PROTOCOL)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Prepare data for training: we use 50 cells, and chop the traces into snippets with a length of 300 timesteps.\n",
    "train_spikes is empty because we perform unsupervised learning."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-01-08T13:44:43.903090Z",
     "start_time": "2018-01-08T13:44:43.592380Z"
    },
    "run_control": {
     "frozen": false,
     "read_only": false
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(198, 300)\n",
      "[[], [], [], [], [], [], [], [], [], [], [], [], [], [], [], [], [], [], [], [], [], [], [], [], [], [], [], [], [], []]\n"
     ]
    }
   ],
   "source": [
    "timebins = 300\n",
    "\n",
    "with open('foopsi_100.pkl', 'rb') as f:\n",
    "    data = pickle.load(f)\n",
    "\n",
    "n_cells = 30\n",
    "\n",
    "Traces_train = data['traces']\n",
    "Spikes_train = data['spikes']\n",
    "truths = data['params']\n",
    "\n",
    "train_traces, train_spikes = data_chop(timebins, Traces_train[:n_cells], Spikes_train[:n_cells], sv_inds = [])\n",
    "\n",
    "print(train_traces[0].shape)\n",
    "print(train_spikes)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Initialize the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-01-08T13:44:49.889386Z",
     "start_time": "2018-01-08T13:44:43.905763Z"
    },
    "run_control": {
     "frozen": false,
     "read_only": false
    }
   },
   "outputs": [],
   "source": [
    "init_params = {'gamma': 3.5, 'alpha': 0, 'beta': -0.3, 'sigma': -2.5}\n",
    "inf_params = None\n",
    "\n",
    "# Initialize recognition model network (1D CNN)\n",
    "recDict = set_rec_net(num_filters=[30,25,25,20,15], filtsize=[31,21,21,13,11])\n",
    "# Initialize model including recognition and generative model\n",
    "model = VIMCO(init_params, inf_params, FOOPSI, recDict, BernoulliRecognition, batch_size=20, n_samples = 64, n_cells = n_cells, rng = rng, filename = 'foopsi_model')\n",
    "\n",
    "model.true_params = [t for t in truths] # List of true parametervalues used to simulate data. Used to verify quality of generative model fit\n",
    "model.gen_lr_fac = 30 # We mulitply the learning rate by this parameter for adjusting the generative model parameters. \n",
    "model.lr_decay = np.array(0.995, dtype=theano.config.floatX) # Learning rate decay (after every evaluation)\n",
    "model.buffers = [100,50] # Buffer regions at the beginning and end of every trace. These are excluded when obtaining gradients.\n",
    "model.facs = [1] # Downsample predictions and ground truth by that factor before evaluation. Useful to compare performance with other methods.\n",
    "model.mrec.pz = 0.015 # Prior on the firing rate. This prior is rather 'weak' and doesn't influence training to much. initialization of gen. parameters has a much bigger impact. \n",
    "model.gradnorm_tot = True # Whether gradients are clipped when their norm exceeds a certain threshold. This proved to be important for good performance.\n",
    "model.gradnorm_clip = 0.02 # Threshold value for clipping\n",
    "model.validation_score = 'synchro_0' # During training the current as well as the best model are stored. The best model is determined by this metric.\n",
    "# model.conc_cache = True\n",
    "\n",
    "# Provide dataset for evaluation during training. Here we include ground truth spikes. We evaluate the first 'eval_T' tiemsteps of every trace provided.\n",
    "# 'eval_rep' determines the number of samples taken for evaluation, multiple samples are only needed for correlated posteriors. \n",
    "# 'test_gen' determines wheter we calculate the quality of the reconstruction\n",
    "# Multiple dicts can be provided to test on training and validation data.\n",
    "model.eval_sets = [{'traces':Traces_train[:n_cells,:,:1000], 'spikes':Spikes_train[:n_cells,:,:1000], 'eval_T':1000, 'eval_rep': 1, 'test_gen':True}]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Training:\n",
    "'epochs' are not passes through the whole dataset but number of evaluations that come after every 'print_freq' iterations. So here we train for 500*100 = 50000 iterations. \n",
    "\n",
    "The following metrics are displayed during training, all of them are calculated on the training set (because we can only evaluate the gen. model fit on these cells). The method has proven to be very robust to overfitting though.\n",
    "    * Corr. Train: Correlation of the predictions with the true spike train.\n",
    "    * Synchro: Spike synchronization: http://www.scholarpedia.org/article/Measures_of_spike_train_synchrony#SPIKE_synchronization\n",
    "            Unlike the correlation this metric operates on predicted spike trains and not firing rates. Therefore it is specifically valuable when comparing the performance of correlated and factorized posterior estimates.\n",
    "            It takes values between 0 and 1 (higher is better)\n",
    "    * MSE Pred./Truth. : Mean squared error of the reconstruction, divided by the best possible reconstruction (i.e. for true spikes and gen. model parameters, limited by the noise), so that the optimal value is 1.\n",
    "                       Reconstruction is either based on the prediction (testing the quality of both the recogntion model and generative model fit) or on the true spike train (just testing the generative model fit)\n",
    "    * Cost: Objective that gets optimized. Not a good measurement of perfance.\n",
    "    * Factor: Number of inferred spikes / number of true spikes. This is usefull to see whether the generative model has roughly the right amplitude. For bad initialization training might diverge to a solution with small amplitudes and many spikes that fit the noise.\n",
    "    * Time upd./Eval: Time required per training iteration / per evaluation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-01-08T14:48:36.416672Z",
     "start_time": "2018-01-08T13:44:49.891830Z"
    },
    "run_control": {
     "frozen": false,
     "read_only": false
    },
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/theano/tensor/basic.py:5262: UserWarning: flatten outdim parameter is deprecated, use ndim instead.\n",
      "  \"flatten outdim parameter is deprecated, use ndim instead.\")\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Corr. Train: 0.748 Synchro: 0.670 || MSE Pred./Truth: 2.195 2.179 || Cost: -5.284 || Factor: 0.805 || Time upd./Eval.: 69.8 ms 2.1 s || BatchNr.: 200\n",
      "Corr. Train: 0.753 Synchro: 0.672 || MSE Pred./Truth: 1.981 1.732 || Cost: 0.545 || Factor: 0.842 || Time upd./Eval.: 73.4 ms 1.5 s || BatchNr.: 400\n",
      "Corr. Train: 0.767 Synchro: 0.681 || MSE Pred./Truth: 1.870 1.442 || Cost: 1.266 || Factor: 0.866 || Time upd./Eval.: 72.3 ms 1.6 s || BatchNr.: 600\n",
      "Corr. Train: 0.813 Synchro: 0.779 || MSE Pred./Truth: 1.591 1.286 || Cost: 1.420 || Factor: 0.826 || Time upd./Eval.: 66.6 ms 1.7 s || BatchNr.: 800\n",
      "Corr. Train: 0.819 Synchro: 0.821 || MSE Pred./Truth: 1.487 1.184 || Cost: 1.630 || Factor: 1.001 || Time upd./Eval.: 68.9 ms 1.4 s || BatchNr.: 1000\n",
      "Corr. Train: 0.820 Synchro: 0.804 || MSE Pred./Truth: 1.512 1.088 || Cost: 1.730 || Factor: 1.011 || Time upd./Eval.: 71.2 ms 1.2 s || BatchNr.: 1200\n",
      "Corr. Train: 0.835 Synchro: 0.818 || MSE Pred./Truth: 1.524 1.078 || Cost: 1.616 || Factor: 0.991 || Time upd./Eval.: 69.5 ms 1.0 s || BatchNr.: 1400\n",
      "Corr. Train: 0.828 Synchro: 0.791 || MSE Pred./Truth: 1.729 1.056 || Cost: 1.757 || Factor: 1.267 || Time upd./Eval.: 71.5 ms 1.1 s || BatchNr.: 1600\n",
      "Corr. Train: 0.820 Synchro: 0.839 || MSE Pred./Truth: 1.409 1.040 || Cost: 1.759 || Factor: 1.021 || Time upd./Eval.: 67.4 ms 1.5 s || BatchNr.: 1800\n",
      "Corr. Train: 0.817 Synchro: 0.783 || MSE Pred./Truth: 1.941 1.045 || Cost: 1.764 || Factor: 1.226 || Time upd./Eval.: 69.6 ms 1.1 s || BatchNr.: 2000\n",
      "Corr. Train: 0.841 Synchro: 0.850 || MSE Pred./Truth: 1.366 1.023 || Cost: 1.745 || Factor: 0.983 || Time upd./Eval.: 65.2 ms 1.5 s || BatchNr.: 2200\n",
      "Corr. Train: 0.846 Synchro: 0.852 || MSE Pred./Truth: 1.383 1.041 || Cost: 1.785 || Factor: 0.990 || Time upd./Eval.: 68.6 ms 1.4 s || BatchNr.: 2400\n",
      "Corr. Train: 0.831 Synchro: 0.828 || MSE Pred./Truth: 1.578 1.036 || Cost: 1.820 || Factor: 1.092 || Time upd./Eval.: 72.0 ms 1.1 s || BatchNr.: 2600\n",
      "Corr. Train: 0.834 Synchro: 0.839 || MSE Pred./Truth: 1.489 1.038 || Cost: 1.913 || Factor: 1.171 || Time upd./Eval.: 63.6 ms 1.0 s || BatchNr.: 2800\n",
      "Corr. Train: 0.842 Synchro: 0.831 || MSE Pred./Truth: 1.498 1.036 || Cost: 1.865 || Factor: 1.074 || Time upd./Eval.: 66.5 ms 1.0 s || BatchNr.: 3000\n",
      "Corr. Train: 0.843 Synchro: 0.825 || MSE Pred./Truth: 1.440 1.047 || Cost: 1.855 || Factor: 0.972 || Time upd./Eval.: 67.5 ms 1.1 s || BatchNr.: 3200\n",
      "Corr. Train: 0.836 Synchro: 0.815 || MSE Pred./Truth: 1.581 1.027 || Cost: 1.861 || Factor: 1.116 || Time upd./Eval.: 63.3 ms 1.1 s || BatchNr.: 3400\n",
      "Corr. Train: 0.840 Synchro: 0.840 || MSE Pred./Truth: 1.463 1.043 || Cost: 1.858 || Factor: 1.096 || Time upd./Eval.: 74.0 ms 1.1 s || BatchNr.: 3600\n",
      "Corr. Train: 0.848 Synchro: 0.865 || MSE Pred./Truth: 1.451 1.039 || Cost: 1.921 || Factor: 1.055 || Time upd./Eval.: 70.6 ms 1.5 s || BatchNr.: 3800\n",
      "Corr. Train: 0.839 Synchro: 0.828 || MSE Pred./Truth: 1.519 1.017 || Cost: 1.849 || Factor: 1.082 || Time upd./Eval.: 64.1 ms 0.9 s || BatchNr.: 4000\n",
      "Corr. Train: 0.842 Synchro: 0.848 || MSE Pred./Truth: 1.356 1.026 || Cost: 1.900 || Factor: 0.916 || Time upd./Eval.: 65.4 ms 1.0 s || BatchNr.: 4200\n",
      "Corr. Train: 0.841 Synchro: 0.843 || MSE Pred./Truth: 1.435 1.047 || Cost: 1.802 || Factor: 1.004 || Time upd./Eval.: 65.1 ms 1.0 s || BatchNr.: 4400\n",
      "Corr. Train: 0.849 Synchro: 0.839 || MSE Pred./Truth: 1.431 1.033 || Cost: 1.809 || Factor: 1.002 || Time upd./Eval.: 71.8 ms 0.9 s || BatchNr.: 4600\n",
      "Corr. Train: 0.847 Synchro: 0.858 || MSE Pred./Truth: 1.345 1.021 || Cost: 1.862 || Factor: 1.001 || Time upd./Eval.: 73.2 ms 1.3 s || BatchNr.: 4800\n",
      "Corr. Train: 0.848 Synchro: 0.831 || MSE Pred./Truth: 1.465 1.024 || Cost: 1.862 || Factor: 1.009 || Time upd./Eval.: 66.5 ms 1.0 s || BatchNr.: 5000\n",
      "Corr. Train: 0.840 Synchro: 0.826 || MSE Pred./Truth: 1.565 1.024 || Cost: 1.856 || Factor: 1.177 || Time upd./Eval.: 66.7 ms 1.0 s || BatchNr.: 5200\n",
      "Corr. Train: 0.838 Synchro: 0.803 || MSE Pred./Truth: 1.608 1.018 || Cost: 1.883 || Factor: 1.144 || Time upd./Eval.: 71.3 ms 1.3 s || BatchNr.: 5400\n",
      "Corr. Train: 0.840 Synchro: 0.824 || MSE Pred./Truth: 1.527 1.034 || Cost: 1.869 || Factor: 1.086 || Time upd./Eval.: 68.0 ms 1.1 s || BatchNr.: 5600\n",
      "Corr. Train: 0.837 Synchro: 0.853 || MSE Pred./Truth: 1.404 1.025 || Cost: 1.838 || Factor: 1.070 || Time upd./Eval.: 67.8 ms 1.0 s || BatchNr.: 5800\n",
      "Corr. Train: 0.840 Synchro: 0.850 || MSE Pred./Truth: 1.340 1.024 || Cost: 1.839 || Factor: 1.051 || Time upd./Eval.: 77.9 ms 1.2 s || BatchNr.: 6000\n",
      "Corr. Train: 0.845 Synchro: 0.852 || MSE Pred./Truth: 1.407 1.026 || Cost: 1.826 || Factor: 1.008 || Time upd./Eval.: 70.9 ms 1.1 s || BatchNr.: 6200\n",
      "Corr. Train: 0.833 Synchro: 0.819 || MSE Pred./Truth: 1.482 1.030 || Cost: 1.860 || Factor: 1.047 || Time upd./Eval.: 72.6 ms 1.3 s || BatchNr.: 6400\n",
      "Corr. Train: 0.841 Synchro: 0.853 || MSE Pred./Truth: 1.376 1.018 || Cost: 1.818 || Factor: 0.991 || Time upd./Eval.: 66.4 ms 1.1 s || BatchNr.: 6600\n",
      "Corr. Train: 0.849 Synchro: 0.853 || MSE Pred./Truth: 1.381 1.035 || Cost: 1.863 || Factor: 0.973 || Time upd./Eval.: 68.0 ms 1.1 s || BatchNr.: 6800\n",
      "Corr. Train: 0.849 Synchro: 0.836 || MSE Pred./Truth: 1.408 1.023 || Cost: 1.818 || Factor: 1.059 || Time upd./Eval.: 65.6 ms 1.1 s || BatchNr.: 7000\n",
      "Corr. Train: 0.846 Synchro: 0.850 || MSE Pred./Truth: 1.428 1.018 || Cost: 1.868 || Factor: 1.081 || Time upd./Eval.: 65.4 ms 1.1 s || BatchNr.: 7200\n",
      "Corr. Train: 0.848 Synchro: 0.851 || MSE Pred./Truth: 1.344 1.026 || Cost: 1.890 || Factor: 0.982 || Time upd./Eval.: 62.8 ms 1.1 s || BatchNr.: 7400\n",
      "Corr. Train: 0.847 Synchro: 0.843 || MSE Pred./Truth: 1.410 1.031 || Cost: 1.862 || Factor: 1.032 || Time upd./Eval.: 67.9 ms 1.1 s || BatchNr.: 7600\n",
      "Corr. Train: 0.847 Synchro: 0.830 || MSE Pred./Truth: 1.472 1.015 || Cost: 1.908 || Factor: 1.047 || Time upd./Eval.: 63.9 ms 1.1 s || BatchNr.: 7800\n",
      "Corr. Train: 0.846 Synchro: 0.855 || MSE Pred./Truth: 1.374 1.035 || Cost: 1.865 || Factor: 0.967 || Time upd./Eval.: 68.3 ms 1.2 s || BatchNr.: 8000\n",
      "Corr. Train: 0.842 Synchro: 0.807 || MSE Pred./Truth: 1.642 1.027 || Cost: 1.847 || Factor: 1.111 || Time upd./Eval.: 64.3 ms 1.1 s || BatchNr.: 8200\n",
      "Corr. Train: 0.851 Synchro: 0.839 || MSE Pred./Truth: 1.408 1.019 || Cost: 1.900 || Factor: 1.040 || Time upd./Eval.: 63.7 ms 1.2 s || BatchNr.: 8400\n",
      "Corr. Train: 0.847 Synchro: 0.807 || MSE Pred./Truth: 1.586 1.023 || Cost: 1.793 || Factor: 1.176 || Time upd./Eval.: 64.0 ms 0.8 s || BatchNr.: 8600\n",
      "Corr. Train: 0.850 Synchro: 0.865 || MSE Pred./Truth: 1.331 1.017 || Cost: 1.885 || Factor: 1.044 || Time upd./Eval.: 69.4 ms 1.0 s || BatchNr.: 8800\n",
      "Corr. Train: 0.836 Synchro: 0.781 || MSE Pred./Truth: 1.812 1.013 || Cost: 1.864 || Factor: 1.344 || Time upd./Eval.: 61.7 ms 1.3 s || BatchNr.: 9000\n",
      "Corr. Train: 0.846 Synchro: 0.848 || MSE Pred./Truth: 1.380 1.021 || Cost: 1.843 || Factor: 0.997 || Time upd./Eval.: 71.7 ms 1.1 s || BatchNr.: 9200\n",
      "Corr. Train: 0.849 Synchro: 0.808 || MSE Pred./Truth: 1.547 1.019 || Cost: 1.851 || Factor: 1.072 || Time upd./Eval.: 64.8 ms 1.1 s || BatchNr.: 9400\n",
      "Corr. Train: 0.848 Synchro: 0.872 || MSE Pred./Truth: 1.290 1.016 || Cost: 1.866 || Factor: 0.998 || Time upd./Eval.: 68.3 ms 1.5 s || BatchNr.: 9600\n",
      "Corr. Train: 0.845 Synchro: 0.854 || MSE Pred./Truth: 1.351 1.023 || Cost: 1.880 || Factor: 0.983 || Time upd./Eval.: 65.5 ms 0.9 s || BatchNr.: 9800\n",
      "Corr. Train: 0.844 Synchro: 0.814 || MSE Pred./Truth: 1.610 1.016 || Cost: 1.893 || Factor: 1.233 || Time upd./Eval.: 69.0 ms 1.1 s || BatchNr.: 10000\n",
      "Corr. Train: 0.848 Synchro: 0.834 || MSE Pred./Truth: 1.553 1.020 || Cost: 1.893 || Factor: 1.119 || Time upd./Eval.: 63.9 ms 1.1 s || BatchNr.: 10200\n",
      "Corr. Train: 0.852 Synchro: 0.795 || MSE Pred./Truth: 1.540 1.020 || Cost: 1.842 || Factor: 1.087 || Time upd./Eval.: 65.7 ms 1.0 s || BatchNr.: 10400\n",
      "Corr. Train: 0.847 Synchro: 0.859 || MSE Pred./Truth: 1.387 1.033 || Cost: 1.852 || Factor: 1.070 || Time upd./Eval.: 72.6 ms 1.0 s || BatchNr.: 10600\n",
      "Corr. Train: 0.849 Synchro: 0.838 || MSE Pred./Truth: 1.422 1.021 || Cost: 1.841 || Factor: 1.089 || Time upd./Eval.: 70.0 ms 1.1 s || BatchNr.: 10800\n",
      "Corr. Train: 0.847 Synchro: 0.823 || MSE Pred./Truth: 1.442 1.017 || Cost: 1.936 || Factor: 1.124 || Time upd./Eval.: 64.1 ms 1.1 s || BatchNr.: 11000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Corr. Train: 0.850 Synchro: 0.829 || MSE Pred./Truth: 1.449 1.016 || Cost: 1.856 || Factor: 1.091 || Time upd./Eval.: 63.4 ms 1.1 s || BatchNr.: 11200\n",
      "Corr. Train: 0.846 Synchro: 0.832 || MSE Pred./Truth: 1.519 1.025 || Cost: 1.906 || Factor: 1.156 || Time upd./Eval.: 68.8 ms 1.1 s || BatchNr.: 11400\n",
      "Corr. Train: 0.840 Synchro: 0.804 || MSE Pred./Truth: 1.626 1.018 || Cost: 1.875 || Factor: 1.201 || Time upd./Eval.: 65.8 ms 0.9 s || BatchNr.: 11600\n",
      "Corr. Train: 0.841 Synchro: 0.837 || MSE Pred./Truth: 1.459 1.019 || Cost: 1.865 || Factor: 1.076 || Time upd./Eval.: 70.2 ms 1.0 s || BatchNr.: 11800\n",
      "Corr. Train: 0.850 Synchro: 0.836 || MSE Pred./Truth: 1.499 1.015 || Cost: 1.885 || Factor: 1.061 || Time upd./Eval.: 65.2 ms 0.9 s || BatchNr.: 12000\n",
      "Corr. Train: 0.850 Synchro: 0.863 || MSE Pred./Truth: 1.325 1.018 || Cost: 1.868 || Factor: 0.986 || Time upd./Eval.: 67.8 ms 1.0 s || BatchNr.: 12200\n",
      "Corr. Train: 0.848 Synchro: 0.773 || MSE Pred./Truth: 1.706 1.019 || Cost: 1.857 || Factor: 1.278 || Time upd./Eval.: 67.1 ms 1.1 s || BatchNr.: 12400\n",
      "Corr. Train: 0.847 Synchro: 0.809 || MSE Pred./Truth: 1.512 1.020 || Cost: 1.903 || Factor: 1.135 || Time upd./Eval.: 68.6 ms 1.0 s || BatchNr.: 12600\n",
      "Corr. Train: 0.846 Synchro: 0.836 || MSE Pred./Truth: 1.444 1.020 || Cost: 1.971 || Factor: 1.113 || Time upd./Eval.: 67.2 ms 1.1 s || BatchNr.: 12800\n",
      "Corr. Train: 0.846 Synchro: 0.797 || MSE Pred./Truth: 1.615 1.040 || Cost: 1.831 || Factor: 1.229 || Time upd./Eval.: 65.5 ms 1.1 s || BatchNr.: 13000\n",
      "Corr. Train: 0.833 Synchro: 0.842 || MSE Pred./Truth: 1.444 1.021 || Cost: 1.917 || Factor: 1.096 || Time upd./Eval.: 69.6 ms 1.1 s || BatchNr.: 13200\n",
      "Corr. Train: 0.850 Synchro: 0.816 || MSE Pred./Truth: 1.556 1.025 || Cost: 1.867 || Factor: 1.158 || Time upd./Eval.: 68.3 ms 1.1 s || BatchNr.: 13400\n",
      "Corr. Train: 0.853 Synchro: 0.839 || MSE Pred./Truth: 1.410 1.024 || Cost: 1.878 || Factor: 1.063 || Time upd./Eval.: 67.3 ms 1.0 s || BatchNr.: 13600\n",
      "Corr. Train: 0.854 Synchro: 0.845 || MSE Pred./Truth: 1.348 1.019 || Cost: 1.875 || Factor: 1.012 || Time upd./Eval.: 64.5 ms 1.1 s || BatchNr.: 13800\n",
      "Corr. Train: 0.848 Synchro: 0.822 || MSE Pred./Truth: 1.473 1.013 || Cost: 1.901 || Factor: 1.202 || Time upd./Eval.: 68.0 ms 1.1 s || BatchNr.: 14000\n",
      "Corr. Train: 0.852 Synchro: 0.827 || MSE Pred./Truth: 1.464 1.019 || Cost: 1.836 || Factor: 1.055 || Time upd./Eval.: 64.9 ms 1.3 s || BatchNr.: 14200\n",
      "Corr. Train: 0.852 Synchro: 0.811 || MSE Pred./Truth: 1.440 1.016 || Cost: 1.884 || Factor: 1.048 || Time upd./Eval.: 70.3 ms 1.1 s || BatchNr.: 14400\n",
      "Corr. Train: 0.846 Synchro: 0.822 || MSE Pred./Truth: 1.416 1.022 || Cost: 1.910 || Factor: 0.996 || Time upd./Eval.: 63.4 ms 1.1 s || BatchNr.: 14600\n",
      "Corr. Train: 0.846 Synchro: 0.844 || MSE Pred./Truth: 1.368 1.020 || Cost: 1.884 || Factor: 1.028 || Time upd./Eval.: 69.0 ms 1.1 s || BatchNr.: 14800\n",
      "Corr. Train: 0.850 Synchro: 0.840 || MSE Pred./Truth: 1.379 1.015 || Cost: 1.907 || Factor: 1.006 || Time upd./Eval.: 73.8 ms 1.1 s || BatchNr.: 15000\n",
      "Corr. Train: 0.848 Synchro: 0.834 || MSE Pred./Truth: 1.494 1.023 || Cost: 1.809 || Factor: 1.133 || Time upd./Eval.: 65.8 ms 1.1 s || BatchNr.: 15200\n",
      "Corr. Train: 0.850 Synchro: 0.849 || MSE Pred./Truth: 1.373 1.020 || Cost: 1.888 || Factor: 0.989 || Time upd./Eval.: 72.4 ms 1.0 s || BatchNr.: 15400\n",
      "Corr. Train: 0.848 Synchro: 0.837 || MSE Pred./Truth: 1.417 1.036 || Cost: 1.851 || Factor: 1.036 || Time upd./Eval.: 66.7 ms 1.3 s || BatchNr.: 15600\n",
      "Corr. Train: 0.855 Synchro: 0.807 || MSE Pred./Truth: 1.390 1.020 || Cost: 1.864 || Factor: 0.994 || Time upd./Eval.: 65.4 ms 1.0 s || BatchNr.: 15800\n",
      "Corr. Train: 0.853 Synchro: 0.855 || MSE Pred./Truth: 1.349 1.028 || Cost: 1.864 || Factor: 1.036 || Time upd./Eval.: 63.0 ms 1.0 s || BatchNr.: 16000\n",
      "Corr. Train: 0.850 Synchro: 0.843 || MSE Pred./Truth: 1.422 1.027 || Cost: 1.867 || Factor: 1.101 || Time upd./Eval.: 70.2 ms 1.0 s || BatchNr.: 16200\n",
      "Corr. Train: 0.854 Synchro: 0.860 || MSE Pred./Truth: 1.355 1.014 || Cost: 1.982 || Factor: 1.050 || Time upd./Eval.: 66.9 ms 1.1 s || BatchNr.: 16400\n",
      "Corr. Train: 0.852 Synchro: 0.846 || MSE Pred./Truth: 1.359 1.020 || Cost: 1.870 || Factor: 1.046 || Time upd./Eval.: 68.3 ms 1.0 s || BatchNr.: 16600\n",
      "Corr. Train: 0.846 Synchro: 0.822 || MSE Pred./Truth: 1.568 1.022 || Cost: 1.827 || Factor: 1.157 || Time upd./Eval.: 63.3 ms 1.0 s || BatchNr.: 16800\n",
      "Corr. Train: 0.846 Synchro: 0.837 || MSE Pred./Truth: 1.452 1.029 || Cost: 1.880 || Factor: 1.096 || Time upd./Eval.: 65.6 ms 1.1 s || BatchNr.: 17000\n",
      "Corr. Train: 0.849 Synchro: 0.846 || MSE Pred./Truth: 1.413 1.024 || Cost: 1.868 || Factor: 1.040 || Time upd./Eval.: 75.3 ms 1.2 s || BatchNr.: 17200\n",
      "Corr. Train: 0.853 Synchro: 0.832 || MSE Pred./Truth: 1.486 1.013 || Cost: 1.902 || Factor: 1.134 || Time upd./Eval.: 67.5 ms 1.0 s || BatchNr.: 17400\n",
      "Corr. Train: 0.845 Synchro: 0.827 || MSE Pred./Truth: 1.440 1.018 || Cost: 1.883 || Factor: 1.095 || Time upd./Eval.: 71.0 ms 1.0 s || BatchNr.: 17600\n",
      "Corr. Train: 0.849 Synchro: 0.839 || MSE Pred./Truth: 1.431 1.021 || Cost: 1.867 || Factor: 1.068 || Time upd./Eval.: 64.8 ms 1.1 s || BatchNr.: 17800\n",
      "Corr. Train: 0.853 Synchro: 0.867 || MSE Pred./Truth: 1.347 1.025 || Cost: 1.912 || Factor: 1.052 || Time upd./Eval.: 74.7 ms 1.0 s || BatchNr.: 18000\n",
      "Corr. Train: 0.855 Synchro: 0.841 || MSE Pred./Truth: 1.394 1.021 || Cost: 1.884 || Factor: 1.054 || Time upd./Eval.: 65.5 ms 1.0 s || BatchNr.: 18200\n",
      "Corr. Train: 0.854 Synchro: 0.844 || MSE Pred./Truth: 1.318 1.022 || Cost: 1.826 || Factor: 0.932 || Time upd./Eval.: 67.9 ms 1.1 s || BatchNr.: 18400\n",
      "Corr. Train: 0.848 Synchro: 0.814 || MSE Pred./Truth: 1.448 1.013 || Cost: 1.881 || Factor: 0.961 || Time upd./Eval.: 63.1 ms 1.0 s || BatchNr.: 18600\n",
      "Corr. Train: 0.853 Synchro: 0.860 || MSE Pred./Truth: 1.316 1.018 || Cost: 1.904 || Factor: 1.064 || Time upd./Eval.: 70.0 ms 0.9 s || BatchNr.: 18800\n",
      "Corr. Train: 0.855 Synchro: 0.822 || MSE Pred./Truth: 1.455 1.018 || Cost: 1.859 || Factor: 1.042 || Time upd./Eval.: 63.4 ms 1.0 s || BatchNr.: 19000\n",
      "Corr. Train: 0.857 Synchro: 0.848 || MSE Pred./Truth: 1.385 1.008 || Cost: 1.929 || Factor: 1.061 || Time upd./Eval.: 62.4 ms 1.0 s || BatchNr.: 19200\n",
      "Corr. Train: 0.852 Synchro: 0.839 || MSE Pred./Truth: 1.424 1.018 || Cost: 1.926 || Factor: 1.094 || Time upd./Eval.: 66.7 ms 1.2 s || BatchNr.: 19400\n",
      "Corr. Train: 0.857 Synchro: 0.824 || MSE Pred./Truth: 1.453 1.015 || Cost: 1.896 || Factor: 1.105 || Time upd./Eval.: 65.4 ms 0.9 s || BatchNr.: 19600\n",
      "Corr. Train: 0.851 Synchro: 0.816 || MSE Pred./Truth: 1.527 1.010 || Cost: 1.958 || Factor: 1.154 || Time upd./Eval.: 72.8 ms 1.4 s || BatchNr.: 19800\n",
      "Corr. Train: 0.855 Synchro: 0.817 || MSE Pred./Truth: 1.459 1.011 || Cost: 1.947 || Factor: 1.055 || Time upd./Eval.: 70.6 ms 1.0 s || BatchNr.: 20000\n",
      "Corr. Train: 0.850 Synchro: 0.840 || MSE Pred./Truth: 1.366 1.018 || Cost: 1.860 || Factor: 0.969 || Time upd./Eval.: 63.7 ms 1.1 s || BatchNr.: 20200\n",
      "Corr. Train: 0.853 Synchro: 0.858 || MSE Pred./Truth: 1.388 1.011 || Cost: 1.942 || Factor: 1.063 || Time upd./Eval.: 66.9 ms 1.1 s || BatchNr.: 20400\n",
      "Corr. Train: 0.854 Synchro: 0.849 || MSE Pred./Truth: 1.370 1.017 || Cost: 1.934 || Factor: 1.009 || Time upd./Eval.: 65.9 ms 1.3 s || BatchNr.: 20600\n",
      "Corr. Train: 0.854 Synchro: 0.831 || MSE Pred./Truth: 1.417 1.014 || Cost: 1.925 || Factor: 1.105 || Time upd./Eval.: 61.0 ms 1.0 s || BatchNr.: 20800\n",
      "Corr. Train: 0.855 Synchro: 0.861 || MSE Pred./Truth: 1.370 1.014 || Cost: 1.844 || Factor: 1.069 || Time upd./Eval.: 69.4 ms 1.2 s || BatchNr.: 21000\n",
      "Corr. Train: 0.852 Synchro: 0.849 || MSE Pred./Truth: 1.348 1.017 || Cost: 1.927 || Factor: 0.968 || Time upd./Eval.: 62.8 ms 1.3 s || BatchNr.: 21200\n",
      "Corr. Train: 0.855 Synchro: 0.815 || MSE Pred./Truth: 1.436 1.013 || Cost: 1.960 || Factor: 1.070 || Time upd./Eval.: 68.7 ms 1.0 s || BatchNr.: 21400\n",
      "Corr. Train: 0.854 Synchro: 0.811 || MSE Pred./Truth: 1.448 1.013 || Cost: 1.933 || Factor: 1.113 || Time upd./Eval.: 62.5 ms 1.1 s || BatchNr.: 21600\n",
      "Corr. Train: 0.858 Synchro: 0.844 || MSE Pred./Truth: 1.364 1.011 || Cost: 1.906 || Factor: 1.051 || Time upd./Eval.: 67.1 ms 1.0 s || BatchNr.: 21800\n",
      "Corr. Train: 0.857 Synchro: 0.829 || MSE Pred./Truth: 1.468 1.013 || Cost: 1.927 || Factor: 1.087 || Time upd./Eval.: 67.6 ms 1.0 s || BatchNr.: 22000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Corr. Train: 0.854 Synchro: 0.847 || MSE Pred./Truth: 1.418 1.015 || Cost: 1.912 || Factor: 1.131 || Time upd./Eval.: 66.6 ms 1.3 s || BatchNr.: 22200\n",
      "Corr. Train: 0.857 Synchro: 0.844 || MSE Pred./Truth: 1.352 1.009 || Cost: 1.937 || Factor: 1.089 || Time upd./Eval.: 65.0 ms 1.2 s || BatchNr.: 22400\n",
      "Corr. Train: 0.853 Synchro: 0.839 || MSE Pred./Truth: 1.414 1.017 || Cost: 1.920 || Factor: 1.121 || Time upd./Eval.: 64.9 ms 1.0 s || BatchNr.: 22600\n",
      "Corr. Train: 0.855 Synchro: 0.813 || MSE Pred./Truth: 1.430 1.022 || Cost: 1.912 || Factor: 1.030 || Time upd./Eval.: 65.7 ms 1.0 s || BatchNr.: 22800\n",
      "Corr. Train: 0.854 Synchro: 0.840 || MSE Pred./Truth: 1.437 1.023 || Cost: 1.806 || Factor: 1.058 || Time upd./Eval.: 67.1 ms 1.1 s || BatchNr.: 23000\n",
      "Corr. Train: 0.857 Synchro: 0.836 || MSE Pred./Truth: 1.360 1.018 || Cost: 1.893 || Factor: 1.059 || Time upd./Eval.: 63.6 ms 1.1 s || BatchNr.: 23200\n",
      "Corr. Train: 0.850 Synchro: 0.803 || MSE Pred./Truth: 1.499 1.010 || Cost: 1.937 || Factor: 1.202 || Time upd./Eval.: 64.1 ms 1.1 s || BatchNr.: 23400\n",
      "Corr. Train: 0.853 Synchro: 0.829 || MSE Pred./Truth: 1.375 1.009 || Cost: 1.877 || Factor: 1.058 || Time upd./Eval.: 67.7 ms 0.9 s || BatchNr.: 23600\n",
      "Corr. Train: 0.858 Synchro: 0.843 || MSE Pred./Truth: 1.414 1.011 || Cost: 1.878 || Factor: 1.091 || Time upd./Eval.: 68.2 ms 1.0 s || BatchNr.: 23800\n",
      "Corr. Train: 0.855 Synchro: 0.822 || MSE Pred./Truth: 1.423 1.010 || Cost: 1.928 || Factor: 1.071 || Time upd./Eval.: 71.6 ms 1.0 s || BatchNr.: 24000\n",
      "Corr. Train: 0.857 Synchro: 0.850 || MSE Pred./Truth: 1.350 1.011 || Cost: 1.874 || Factor: 1.077 || Time upd./Eval.: 66.2 ms 1.0 s || BatchNr.: 24200\n",
      "Corr. Train: 0.855 Synchro: 0.848 || MSE Pred./Truth: 1.332 1.010 || Cost: 1.872 || Factor: 0.999 || Time upd./Eval.: 70.2 ms 1.1 s || BatchNr.: 24400\n",
      "Corr. Train: 0.854 Synchro: 0.843 || MSE Pred./Truth: 1.398 1.011 || Cost: 1.937 || Factor: 1.101 || Time upd./Eval.: 65.9 ms 1.1 s || BatchNr.: 24600\n",
      "Corr. Train: 0.852 Synchro: 0.824 || MSE Pred./Truth: 1.431 1.007 || Cost: 1.923 || Factor: 1.113 || Time upd./Eval.: 68.3 ms 1.2 s || BatchNr.: 24800\n",
      "Corr. Train: 0.851 Synchro: 0.815 || MSE Pred./Truth: 1.531 1.013 || Cost: 1.934 || Factor: 1.250 || Time upd./Eval.: 69.6 ms 1.0 s || BatchNr.: 25000\n",
      "Corr. Train: 0.856 Synchro: 0.816 || MSE Pred./Truth: 1.463 1.023 || Cost: 1.927 || Factor: 1.104 || Time upd./Eval.: 69.5 ms 0.9 s || BatchNr.: 25200\n",
      "Corr. Train: 0.851 Synchro: 0.849 || MSE Pred./Truth: 1.334 1.012 || Cost: 1.984 || Factor: 1.089 || Time upd./Eval.: 68.3 ms 1.2 s || BatchNr.: 25400\n",
      "Corr. Train: 0.851 Synchro: 0.855 || MSE Pred./Truth: 1.368 1.011 || Cost: 1.935 || Factor: 1.052 || Time upd./Eval.: 69.1 ms 1.2 s || BatchNr.: 25600\n",
      "Corr. Train: 0.852 Synchro: 0.845 || MSE Pred./Truth: 1.375 1.011 || Cost: 1.902 || Factor: 1.109 || Time upd./Eval.: 65.9 ms 1.1 s || BatchNr.: 25800\n",
      "Corr. Train: 0.851 Synchro: 0.826 || MSE Pred./Truth: 1.525 1.008 || Cost: 1.866 || Factor: 1.173 || Time upd./Eval.: 67.1 ms 1.2 s || BatchNr.: 26000\n",
      "Corr. Train: 0.855 Synchro: 0.849 || MSE Pred./Truth: 1.400 1.009 || Cost: 1.928 || Factor: 1.114 || Time upd./Eval.: 69.3 ms 1.3 s || BatchNr.: 26200\n",
      "Corr. Train: 0.857 Synchro: 0.848 || MSE Pred./Truth: 1.376 1.019 || Cost: 1.973 || Factor: 1.067 || Time upd./Eval.: 64.7 ms 1.0 s || BatchNr.: 26400\n",
      "Corr. Train: 0.853 Synchro: 0.833 || MSE Pred./Truth: 1.413 1.017 || Cost: 1.832 || Factor: 1.063 || Time upd./Eval.: 62.2 ms 1.1 s || BatchNr.: 26600\n",
      "Corr. Train: 0.856 Synchro: 0.838 || MSE Pred./Truth: 1.398 1.008 || Cost: 1.956 || Factor: 1.048 || Time upd./Eval.: 77.4 ms 1.1 s || BatchNr.: 26800\n",
      "Corr. Train: 0.854 Synchro: 0.828 || MSE Pred./Truth: 1.415 1.014 || Cost: 1.895 || Factor: 1.088 || Time upd./Eval.: 67.6 ms 1.1 s || BatchNr.: 27000\n",
      "Corr. Train: 0.854 Synchro: 0.864 || MSE Pred./Truth: 1.311 1.010 || Cost: 1.939 || Factor: 1.058 || Time upd./Eval.: 73.4 ms 1.1 s || BatchNr.: 27200\n",
      "Corr. Train: 0.855 Synchro: 0.843 || MSE Pred./Truth: 1.373 1.009 || Cost: 1.839 || Factor: 1.069 || Time upd./Eval.: 69.9 ms 1.1 s || BatchNr.: 27400\n",
      "Corr. Train: 0.850 Synchro: 0.828 || MSE Pred./Truth: 1.431 1.004 || Cost: 1.834 || Factor: 1.067 || Time upd./Eval.: 66.9 ms 1.1 s || BatchNr.: 27600\n",
      "Corr. Train: 0.851 Synchro: 0.848 || MSE Pred./Truth: 1.340 1.007 || Cost: 1.929 || Factor: 1.039 || Time upd./Eval.: 63.8 ms 1.2 s || BatchNr.: 27800\n",
      "Corr. Train: 0.853 Synchro: 0.860 || MSE Pred./Truth: 1.352 1.014 || Cost: 1.908 || Factor: 1.052 || Time upd./Eval.: 67.0 ms 1.3 s || BatchNr.: 28000\n",
      "Corr. Train: 0.856 Synchro: 0.818 || MSE Pred./Truth: 1.445 1.015 || Cost: 1.943 || Factor: 1.131 || Time upd./Eval.: 70.3 ms 0.9 s || BatchNr.: 28200\n",
      "Corr. Train: 0.855 Synchro: 0.827 || MSE Pred./Truth: 1.484 1.012 || Cost: 1.935 || Factor: 1.098 || Time upd./Eval.: 69.3 ms 1.1 s || BatchNr.: 28400\n",
      "Corr. Train: 0.855 Synchro: 0.839 || MSE Pred./Truth: 1.403 1.011 || Cost: 1.907 || Factor: 1.069 || Time upd./Eval.: 68.3 ms 1.1 s || BatchNr.: 28600\n",
      "Corr. Train: 0.857 Synchro: 0.847 || MSE Pred./Truth: 1.360 1.009 || Cost: 1.906 || Factor: 1.079 || Time upd./Eval.: 68.6 ms 1.0 s || BatchNr.: 28800\n",
      "Corr. Train: 0.856 Synchro: 0.845 || MSE Pred./Truth: 1.365 1.010 || Cost: 1.978 || Factor: 1.051 || Time upd./Eval.: 67.0 ms 0.9 s || BatchNr.: 29000\n",
      "Corr. Train: 0.854 Synchro: 0.845 || MSE Pred./Truth: 1.362 1.013 || Cost: 1.925 || Factor: 1.035 || Time upd./Eval.: 72.3 ms 1.2 s || BatchNr.: 29200\n",
      "Corr. Train: 0.853 Synchro: 0.843 || MSE Pred./Truth: 1.393 1.012 || Cost: 1.866 || Factor: 1.086 || Time upd./Eval.: 70.6 ms 1.0 s || BatchNr.: 29400\n",
      "Corr. Train: 0.853 Synchro: 0.838 || MSE Pred./Truth: 1.359 1.011 || Cost: 1.940 || Factor: 1.020 || Time upd./Eval.: 73.6 ms 1.0 s || BatchNr.: 29600\n",
      "Corr. Train: 0.853 Synchro: 0.830 || MSE Pred./Truth: 1.425 1.023 || Cost: 1.952 || Factor: 0.987 || Time upd./Eval.: 68.6 ms 1.0 s || BatchNr.: 29800\n",
      "Corr. Train: 0.856 Synchro: 0.810 || MSE Pred./Truth: 1.497 1.016 || Cost: 1.955 || Factor: 1.086 || Time upd./Eval.: 69.8 ms 1.1 s || BatchNr.: 30000\n",
      "Corr. Train: 0.856 Synchro: 0.839 || MSE Pred./Truth: 1.380 1.016 || Cost: 1.936 || Factor: 1.081 || Time upd./Eval.: 66.5 ms 1.1 s || BatchNr.: 30200\n",
      "Corr. Train: 0.856 Synchro: 0.840 || MSE Pred./Truth: 1.372 1.015 || Cost: 1.884 || Factor: 1.091 || Time upd./Eval.: 63.4 ms 1.1 s || BatchNr.: 30400\n",
      "Corr. Train: 0.853 Synchro: 0.841 || MSE Pred./Truth: 1.368 1.007 || Cost: 1.896 || Factor: 1.082 || Time upd./Eval.: 66.8 ms 1.1 s || BatchNr.: 30600\n",
      "Corr. Train: 0.856 Synchro: 0.819 || MSE Pred./Truth: 1.448 1.011 || Cost: 1.911 || Factor: 1.156 || Time upd./Eval.: 69.3 ms 1.1 s || BatchNr.: 30800\n",
      "Corr. Train: 0.854 Synchro: 0.819 || MSE Pred./Truth: 1.453 1.018 || Cost: 1.923 || Factor: 1.125 || Time upd./Eval.: 66.6 ms 1.1 s || BatchNr.: 31000\n",
      "Corr. Train: 0.859 Synchro: 0.818 || MSE Pred./Truth: 1.382 1.008 || Cost: 1.975 || Factor: 1.016 || Time upd./Eval.: 71.5 ms 1.1 s || BatchNr.: 31200\n",
      "Corr. Train: 0.853 Synchro: 0.814 || MSE Pred./Truth: 1.557 1.008 || Cost: 1.961 || Factor: 1.203 || Time upd./Eval.: 72.4 ms 1.1 s || BatchNr.: 31400\n",
      "Corr. Train: 0.853 Synchro: 0.861 || MSE Pred./Truth: 1.304 1.006 || Cost: 1.927 || Factor: 0.985 || Time upd./Eval.: 70.8 ms 1.1 s || BatchNr.: 31600\n",
      "Corr. Train: 0.858 Synchro: 0.854 || MSE Pred./Truth: 1.360 1.009 || Cost: 1.902 || Factor: 1.081 || Time upd./Eval.: 65.9 ms 1.0 s || BatchNr.: 31800\n",
      "Corr. Train: 0.858 Synchro: 0.842 || MSE Pred./Truth: 1.384 1.016 || Cost: 1.937 || Factor: 1.087 || Time upd./Eval.: 68.8 ms 0.9 s || BatchNr.: 32000\n",
      "Corr. Train: 0.855 Synchro: 0.850 || MSE Pred./Truth: 1.377 1.009 || Cost: 1.964 || Factor: 1.092 || Time upd./Eval.: 65.0 ms 1.1 s || BatchNr.: 32200\n",
      "Corr. Train: 0.858 Synchro: 0.834 || MSE Pred./Truth: 1.363 1.006 || Cost: 1.910 || Factor: 1.059 || Time upd./Eval.: 64.6 ms 1.2 s || BatchNr.: 32400\n",
      "Corr. Train: 0.856 Synchro: 0.843 || MSE Pred./Truth: 1.357 1.010 || Cost: 1.902 || Factor: 1.043 || Time upd./Eval.: 69.4 ms 1.0 s || BatchNr.: 32600\n",
      "Corr. Train: 0.855 Synchro: 0.826 || MSE Pred./Truth: 1.461 1.013 || Cost: 1.918 || Factor: 1.103 || Time upd./Eval.: 70.2 ms 1.1 s || BatchNr.: 32800\n",
      "Corr. Train: 0.856 Synchro: 0.814 || MSE Pred./Truth: 1.412 1.013 || Cost: 1.985 || Factor: 0.992 || Time upd./Eval.: 75.3 ms 1.3 s || BatchNr.: 33000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Corr. Train: 0.856 Synchro: 0.825 || MSE Pred./Truth: 1.416 1.006 || Cost: 1.988 || Factor: 1.060 || Time upd./Eval.: 66.8 ms 1.1 s || BatchNr.: 33200\n",
      "Corr. Train: 0.856 Synchro: 0.846 || MSE Pred./Truth: 1.354 1.009 || Cost: 1.924 || Factor: 1.005 || Time upd./Eval.: 71.5 ms 1.0 s || BatchNr.: 33400\n",
      "Corr. Train: 0.853 Synchro: 0.857 || MSE Pred./Truth: 1.342 1.010 || Cost: 1.953 || Factor: 1.038 || Time upd./Eval.: 75.1 ms 1.2 s || BatchNr.: 33600\n",
      "Corr. Train: 0.857 Synchro: 0.849 || MSE Pred./Truth: 1.316 1.010 || Cost: 1.990 || Factor: 1.029 || Time upd./Eval.: 70.2 ms 1.1 s || BatchNr.: 33800\n",
      "Corr. Train: 0.855 Synchro: 0.865 || MSE Pred./Truth: 1.317 1.010 || Cost: 1.954 || Factor: 1.035 || Time upd./Eval.: 69.7 ms 1.3 s || BatchNr.: 34000\n",
      "Corr. Train: 0.853 Synchro: 0.837 || MSE Pred./Truth: 1.405 1.007 || Cost: 1.995 || Factor: 1.081 || Time upd./Eval.: 64.3 ms 1.0 s || BatchNr.: 34200\n",
      "Corr. Train: 0.852 Synchro: 0.830 || MSE Pred./Truth: 1.513 1.015 || Cost: 1.869 || Factor: 1.185 || Time upd./Eval.: 67.2 ms 1.1 s || BatchNr.: 34400\n",
      "Corr. Train: 0.857 Synchro: 0.831 || MSE Pred./Truth: 1.425 1.016 || Cost: 1.950 || Factor: 1.095 || Time upd./Eval.: 71.8 ms 1.0 s || BatchNr.: 34600\n",
      "Corr. Train: 0.856 Synchro: 0.830 || MSE Pred./Truth: 1.357 1.009 || Cost: 1.924 || Factor: 1.058 || Time upd./Eval.: 74.5 ms 1.1 s || BatchNr.: 34800\n",
      "Corr. Train: 0.857 Synchro: 0.844 || MSE Pred./Truth: 1.392 1.008 || Cost: 1.957 || Factor: 1.053 || Time upd./Eval.: 66.6 ms 1.1 s || BatchNr.: 35000\n",
      "Corr. Train: 0.854 Synchro: 0.838 || MSE Pred./Truth: 1.322 1.006 || Cost: 1.947 || Factor: 0.997 || Time upd./Eval.: 68.2 ms 1.0 s || BatchNr.: 35200\n",
      "Corr. Train: 0.857 Synchro: 0.831 || MSE Pred./Truth: 1.416 1.008 || Cost: 1.898 || Factor: 1.055 || Time upd./Eval.: 67.1 ms 1.1 s || BatchNr.: 35400\n",
      "Corr. Train: 0.854 Synchro: 0.845 || MSE Pred./Truth: 1.330 1.011 || Cost: 1.952 || Factor: 1.099 || Time upd./Eval.: 69.4 ms 0.9 s || BatchNr.: 35600\n",
      "Corr. Train: 0.855 Synchro: 0.834 || MSE Pred./Truth: 1.405 1.012 || Cost: 1.940 || Factor: 1.045 || Time upd./Eval.: 70.7 ms 1.1 s || BatchNr.: 35800\n",
      "Corr. Train: 0.855 Synchro: 0.846 || MSE Pred./Truth: 1.357 1.009 || Cost: 1.973 || Factor: 1.065 || Time upd./Eval.: 64.7 ms 1.1 s || BatchNr.: 36000\n",
      "Corr. Train: 0.857 Synchro: 0.873 || MSE Pred./Truth: 1.265 1.006 || Cost: 1.885 || Factor: 1.014 || Time upd./Eval.: 73.1 ms 1.7 s || BatchNr.: 36200\n",
      "Corr. Train: 0.853 Synchro: 0.832 || MSE Pred./Truth: 1.451 1.007 || Cost: 2.017 || Factor: 1.136 || Time upd./Eval.: 62.1 ms 1.1 s || BatchNr.: 36400\n",
      "Corr. Train: 0.854 Synchro: 0.820 || MSE Pred./Truth: 1.484 1.008 || Cost: 2.038 || Factor: 1.144 || Time upd./Eval.: 65.0 ms 1.0 s || BatchNr.: 36600\n",
      "Corr. Train: 0.852 Synchro: 0.832 || MSE Pred./Truth: 1.383 1.007 || Cost: 1.930 || Factor: 1.050 || Time upd./Eval.: 71.0 ms 1.1 s || BatchNr.: 36800\n",
      "Corr. Train: 0.857 Synchro: 0.838 || MSE Pred./Truth: 1.350 1.014 || Cost: 1.927 || Factor: 1.054 || Time upd./Eval.: 68.1 ms 0.9 s || BatchNr.: 37000\n",
      "Corr. Train: 0.856 Synchro: 0.839 || MSE Pred./Truth: 1.444 1.012 || Cost: 1.928 || Factor: 1.090 || Time upd./Eval.: 68.1 ms 1.0 s || BatchNr.: 37200\n",
      "Corr. Train: 0.857 Synchro: 0.832 || MSE Pred./Truth: 1.408 1.006 || Cost: 1.938 || Factor: 1.061 || Time upd./Eval.: 67.6 ms 1.1 s || BatchNr.: 37400\n",
      "Corr. Train: 0.856 Synchro: 0.839 || MSE Pred./Truth: 1.344 1.009 || Cost: 1.975 || Factor: 0.996 || Time upd./Eval.: 69.0 ms 1.1 s || BatchNr.: 37600\n",
      "Corr. Train: 0.856 Synchro: 0.840 || MSE Pred./Truth: 1.390 1.007 || Cost: 1.939 || Factor: 1.076 || Time upd./Eval.: 71.7 ms 1.1 s || BatchNr.: 37800\n",
      "Corr. Train: 0.855 Synchro: 0.857 || MSE Pred./Truth: 1.341 1.008 || Cost: 1.957 || Factor: 1.033 || Time upd./Eval.: 65.0 ms 1.1 s || BatchNr.: 38000\n",
      "Corr. Train: 0.855 Synchro: 0.834 || MSE Pred./Truth: 1.337 1.006 || Cost: 1.959 || Factor: 1.048 || Time upd./Eval.: 65.0 ms 1.1 s || BatchNr.: 38200\n",
      "Corr. Train: 0.855 Synchro: 0.843 || MSE Pred./Truth: 1.390 1.005 || Cost: 1.920 || Factor: 1.061 || Time upd./Eval.: 67.9 ms 0.9 s || BatchNr.: 38400\n",
      "Corr. Train: 0.859 Synchro: 0.824 || MSE Pred./Truth: 1.405 1.011 || Cost: 1.940 || Factor: 1.029 || Time upd./Eval.: 65.5 ms 1.1 s || BatchNr.: 38600\n",
      "Corr. Train: 0.849 Synchro: 0.817 || MSE Pred./Truth: 1.457 1.013 || Cost: 1.920 || Factor: 1.148 || Time upd./Eval.: 69.6 ms 1.0 s || BatchNr.: 38800\n",
      "Corr. Train: 0.853 Synchro: 0.819 || MSE Pred./Truth: 1.475 1.011 || Cost: 1.907 || Factor: 1.204 || Time upd./Eval.: 71.0 ms 1.0 s || BatchNr.: 39000\n",
      "Corr. Train: 0.857 Synchro: 0.851 || MSE Pred./Truth: 1.351 1.009 || Cost: 1.906 || Factor: 1.111 || Time upd./Eval.: 68.3 ms 1.0 s || BatchNr.: 39200\n",
      "Corr. Train: 0.856 Synchro: 0.825 || MSE Pred./Truth: 1.364 1.007 || Cost: 1.863 || Factor: 1.045 || Time upd./Eval.: 66.4 ms 1.1 s || BatchNr.: 39400\n",
      "Corr. Train: 0.859 Synchro: 0.850 || MSE Pred./Truth: 1.336 1.007 || Cost: 1.960 || Factor: 1.071 || Time upd./Eval.: 63.6 ms 0.9 s || BatchNr.: 39600\n",
      "Corr. Train: 0.852 Synchro: 0.824 || MSE Pred./Truth: 1.454 1.012 || Cost: 1.981 || Factor: 1.154 || Time upd./Eval.: 69.8 ms 1.2 s || BatchNr.: 39800\n",
      "Corr. Train: 0.856 Synchro: 0.843 || MSE Pred./Truth: 1.361 1.009 || Cost: 1.913 || Factor: 1.061 || Time upd./Eval.: 71.8 ms 1.1 s || BatchNr.: 40000\n",
      "Corr. Train: 0.851 Synchro: 0.848 || MSE Pred./Truth: 1.353 1.008 || Cost: 1.905 || Factor: 1.078 || Time upd./Eval.: 71.0 ms 1.0 s || BatchNr.: 40200\n",
      "Corr. Train: 0.857 Synchro: 0.843 || MSE Pred./Truth: 1.382 1.011 || Cost: 2.042 || Factor: 1.080 || Time upd./Eval.: 67.0 ms 1.1 s || BatchNr.: 40400\n",
      "Corr. Train: 0.859 Synchro: 0.827 || MSE Pred./Truth: 1.372 1.009 || Cost: 1.956 || Factor: 1.043 || Time upd./Eval.: 67.9 ms 1.1 s || BatchNr.: 40600\n",
      "Corr. Train: 0.856 Synchro: 0.835 || MSE Pred./Truth: 1.370 1.006 || Cost: 1.988 || Factor: 1.060 || Time upd./Eval.: 71.3 ms 1.1 s || BatchNr.: 40800\n",
      "Corr. Train: 0.854 Synchro: 0.847 || MSE Pred./Truth: 1.376 1.007 || Cost: 1.941 || Factor: 1.041 || Time upd./Eval.: 65.5 ms 1.1 s || BatchNr.: 41000\n",
      "Corr. Train: 0.860 Synchro: 0.842 || MSE Pred./Truth: 1.374 1.008 || Cost: 2.012 || Factor: 1.060 || Time upd./Eval.: 69.9 ms 1.1 s || BatchNr.: 41200\n",
      "Corr. Train: 0.857 Synchro: 0.835 || MSE Pred./Truth: 1.444 1.014 || Cost: 1.950 || Factor: 1.136 || Time upd./Eval.: 69.5 ms 1.2 s || BatchNr.: 41400\n",
      "Corr. Train: 0.857 Synchro: 0.830 || MSE Pred./Truth: 1.382 1.012 || Cost: 1.914 || Factor: 1.020 || Time upd./Eval.: 67.3 ms 0.9 s || BatchNr.: 41600\n",
      "Corr. Train: 0.855 Synchro: 0.851 || MSE Pred./Truth: 1.376 1.006 || Cost: 1.969 || Factor: 1.101 || Time upd./Eval.: 73.6 ms 1.2 s || BatchNr.: 41800\n",
      "Corr. Train: 0.859 Synchro: 0.839 || MSE Pred./Truth: 1.409 1.010 || Cost: 1.926 || Factor: 1.098 || Time upd./Eval.: 70.9 ms 1.1 s || BatchNr.: 42000\n",
      "Corr. Train: 0.858 Synchro: 0.843 || MSE Pred./Truth: 1.354 1.007 || Cost: 1.962 || Factor: 1.037 || Time upd./Eval.: 64.9 ms 1.1 s || BatchNr.: 42200\n",
      "Corr. Train: 0.857 Synchro: 0.840 || MSE Pred./Truth: 1.335 1.008 || Cost: 1.921 || Factor: 1.056 || Time upd./Eval.: 68.7 ms 0.9 s || BatchNr.: 42400\n",
      "Corr. Train: 0.855 Synchro: 0.840 || MSE Pred./Truth: 1.372 1.012 || Cost: 1.995 || Factor: 1.077 || Time upd./Eval.: 70.3 ms 1.1 s || BatchNr.: 42600\n",
      "Corr. Train: 0.854 Synchro: 0.851 || MSE Pred./Truth: 1.313 1.009 || Cost: 1.886 || Factor: 1.014 || Time upd./Eval.: 64.5 ms 1.1 s || BatchNr.: 42800\n",
      "Corr. Train: 0.856 Synchro: 0.859 || MSE Pred./Truth: 1.335 1.009 || Cost: 1.950 || Factor: 1.049 || Time upd./Eval.: 67.0 ms 1.1 s || BatchNr.: 43000\n",
      "Corr. Train: 0.856 Synchro: 0.844 || MSE Pred./Truth: 1.355 1.006 || Cost: 1.936 || Factor: 1.031 || Time upd./Eval.: 72.7 ms 1.1 s || BatchNr.: 43200\n",
      "Corr. Train: 0.856 Synchro: 0.826 || MSE Pred./Truth: 1.414 1.009 || Cost: 1.956 || Factor: 1.111 || Time upd./Eval.: 67.8 ms 1.1 s || BatchNr.: 43400\n",
      "Corr. Train: 0.858 Synchro: 0.847 || MSE Pred./Truth: 1.344 1.009 || Cost: 1.892 || Factor: 1.102 || Time upd./Eval.: 71.9 ms 1.1 s || BatchNr.: 43600\n",
      "Corr. Train: 0.860 Synchro: 0.836 || MSE Pred./Truth: 1.367 1.009 || Cost: 1.981 || Factor: 1.035 || Time upd./Eval.: 62.8 ms 1.0 s || BatchNr.: 43800\n",
      "Corr. Train: 0.857 Synchro: 0.818 || MSE Pred./Truth: 1.451 1.012 || Cost: 1.967 || Factor: 1.081 || Time upd./Eval.: 69.7 ms 1.0 s || BatchNr.: 44000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Corr. Train: 0.860 Synchro: 0.835 || MSE Pred./Truth: 1.367 1.005 || Cost: 1.890 || Factor: 1.054 || Time upd./Eval.: 71.1 ms 1.1 s || BatchNr.: 44200\n",
      "Corr. Train: 0.857 Synchro: 0.844 || MSE Pred./Truth: 1.352 1.006 || Cost: 1.896 || Factor: 1.082 || Time upd./Eval.: 67.6 ms 1.1 s || BatchNr.: 44400\n",
      "Corr. Train: 0.857 Synchro: 0.851 || MSE Pred./Truth: 1.330 1.008 || Cost: 1.888 || Factor: 1.022 || Time upd./Eval.: 63.9 ms 0.9 s || BatchNr.: 44600\n",
      "Corr. Train: 0.859 Synchro: 0.838 || MSE Pred./Truth: 1.416 1.008 || Cost: 1.800 || Factor: 1.074 || Time upd./Eval.: 69.1 ms 1.0 s || BatchNr.: 44800\n",
      "Corr. Train: 0.859 Synchro: 0.851 || MSE Pred./Truth: 1.324 1.008 || Cost: 1.948 || Factor: 1.024 || Time upd./Eval.: 70.5 ms 1.1 s || BatchNr.: 45000\n",
      "Corr. Train: 0.858 Synchro: 0.836 || MSE Pred./Truth: 1.381 1.009 || Cost: 1.969 || Factor: 1.026 || Time upd./Eval.: 65.5 ms 1.0 s || BatchNr.: 45200\n",
      "Corr. Train: 0.859 Synchro: 0.836 || MSE Pred./Truth: 1.362 1.007 || Cost: 1.973 || Factor: 1.066 || Time upd./Eval.: 70.6 ms 1.1 s || BatchNr.: 45400\n",
      "Corr. Train: 0.860 Synchro: 0.852 || MSE Pred./Truth: 1.317 1.008 || Cost: 1.950 || Factor: 1.049 || Time upd./Eval.: 69.7 ms 1.2 s || BatchNr.: 45600\n",
      "Corr. Train: 0.862 Synchro: 0.849 || MSE Pred./Truth: 1.340 1.006 || Cost: 1.961 || Factor: 1.063 || Time upd./Eval.: 65.5 ms 1.1 s || BatchNr.: 45800\n",
      "Corr. Train: 0.860 Synchro: 0.848 || MSE Pred./Truth: 1.323 1.008 || Cost: 2.009 || Factor: 1.000 || Time upd./Eval.: 74.9 ms 1.1 s || BatchNr.: 46000\n",
      "Corr. Train: 0.859 Synchro: 0.833 || MSE Pred./Truth: 1.407 1.008 || Cost: 1.947 || Factor: 1.077 || Time upd./Eval.: 69.6 ms 1.0 s || BatchNr.: 46200\n",
      "Corr. Train: 0.859 Synchro: 0.844 || MSE Pred./Truth: 1.396 1.008 || Cost: 1.868 || Factor: 1.107 || Time upd./Eval.: 69.7 ms 1.1 s || BatchNr.: 46400\n",
      "Corr. Train: 0.857 Synchro: 0.834 || MSE Pred./Truth: 1.409 1.008 || Cost: 1.958 || Factor: 1.073 || Time upd./Eval.: 69.9 ms 1.0 s || BatchNr.: 46600\n",
      "Corr. Train: 0.862 Synchro: 0.853 || MSE Pred./Truth: 1.312 1.006 || Cost: 2.043 || Factor: 1.004 || Time upd./Eval.: 75.8 ms 1.1 s || BatchNr.: 46800\n",
      "Corr. Train: 0.859 Synchro: 0.830 || MSE Pred./Truth: 1.435 1.010 || Cost: 1.855 || Factor: 1.101 || Time upd./Eval.: 63.3 ms 1.1 s || BatchNr.: 47000\n",
      "Corr. Train: 0.862 Synchro: 0.825 || MSE Pred./Truth: 1.410 1.010 || Cost: 1.878 || Factor: 1.102 || Time upd./Eval.: 70.6 ms 1.1 s || BatchNr.: 47200\n",
      "Corr. Train: 0.862 Synchro: 0.854 || MSE Pred./Truth: 1.347 1.005 || Cost: 1.905 || Factor: 1.077 || Time upd./Eval.: 66.9 ms 1.1 s || BatchNr.: 47400\n",
      "Corr. Train: 0.858 Synchro: 0.840 || MSE Pred./Truth: 1.401 1.009 || Cost: 1.921 || Factor: 1.107 || Time upd./Eval.: 75.8 ms 1.1 s || BatchNr.: 47600\n",
      "Corr. Train: 0.860 Synchro: 0.843 || MSE Pred./Truth: 1.361 1.005 || Cost: 1.952 || Factor: 1.108 || Time upd./Eval.: 68.1 ms 1.3 s || BatchNr.: 47800\n",
      "Corr. Train: 0.860 Synchro: 0.844 || MSE Pred./Truth: 1.363 1.005 || Cost: 1.888 || Factor: 1.051 || Time upd./Eval.: 66.0 ms 1.1 s || BatchNr.: 48000\n",
      "Corr. Train: 0.856 Synchro: 0.836 || MSE Pred./Truth: 1.367 1.009 || Cost: 1.876 || Factor: 1.084 || Time upd./Eval.: 67.0 ms 1.1 s || BatchNr.: 48200\n",
      "Corr. Train: 0.861 Synchro: 0.843 || MSE Pred./Truth: 1.337 1.007 || Cost: 1.942 || Factor: 1.078 || Time upd./Eval.: 63.1 ms 1.1 s || BatchNr.: 48400\n",
      "Corr. Train: 0.861 Synchro: 0.845 || MSE Pred./Truth: 1.354 1.007 || Cost: 1.967 || Factor: 1.074 || Time upd./Eval.: 67.4 ms 1.1 s || BatchNr.: 48600\n",
      "Corr. Train: 0.862 Synchro: 0.868 || MSE Pred./Truth: 1.275 1.004 || Cost: 1.888 || Factor: 1.039 || Time upd./Eval.: 72.1 ms 1.2 s || BatchNr.: 48800\n",
      "Corr. Train: 0.859 Synchro: 0.857 || MSE Pred./Truth: 1.314 1.005 || Cost: 1.966 || Factor: 1.038 || Time upd./Eval.: 64.4 ms 0.9 s || BatchNr.: 49000\n",
      "Corr. Train: 0.859 Synchro: 0.847 || MSE Pred./Truth: 1.422 1.015 || Cost: 1.937 || Factor: 1.119 || Time upd./Eval.: 71.1 ms 1.1 s || BatchNr.: 49200\n",
      "Corr. Train: 0.861 Synchro: 0.829 || MSE Pred./Truth: 1.415 1.004 || Cost: 1.886 || Factor: 1.066 || Time upd./Eval.: 65.6 ms 1.2 s || BatchNr.: 49400\n",
      "Corr. Train: 0.859 Synchro: 0.827 || MSE Pred./Truth: 1.452 1.006 || Cost: 1.907 || Factor: 1.117 || Time upd./Eval.: 72.1 ms 1.1 s || BatchNr.: 49600\n",
      "Corr. Train: 0.854 Synchro: 0.828 || MSE Pred./Truth: 1.459 1.006 || Cost: 1.937 || Factor: 1.112 || Time upd./Eval.: 70.9 ms 1.1 s || BatchNr.: 49800\n",
      "Corr. Train: 0.859 Synchro: 0.836 || MSE Pred./Truth: 1.412 1.010 || Cost: 1.924 || Factor: 1.080 || Time upd./Eval.: 68.3 ms 1.0 s || BatchNr.: 50000\n"
     ]
    }
   ],
   "source": [
    "model.fit(train_traces,train_spikes, max_epochs = 250, learning_rate = 1e-3, print_output = True, print_freq = 200)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "All metrics are stored in ordered dictionaries and can be plotted:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-01-08T14:58:44.023059Z",
     "start_time": "2018-01-08T14:58:43.815885Z"
    },
    "run_control": {
     "frozen": false,
     "read_only": false
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAD1CAYAAACrz7WZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4wLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvpW3flQAAIABJREFUeJzt3Xl8VPW9//HXLJmZ7HtISICwfllFZFFBRVEQl2pte1usrdpV29rbxbbX3lrL1ba/trfX6r1qN9urtbaUa2tLFcVdRFGRTWT5smcn+77MZJbfH2fJJCQQIBA4fJ6Phw8yZ87MnDMm7/nM53zP97hisRhCCCGcxT3cGyCEEGLoSbgLIYQDSbgLIYQDSbgLIYQDSbgLIYQDeYd7A5RSfmAuUAVEhnlzhBDiTOEBCoANWutg3zuHPdwxgv2N4d4IIYQ4Q10MrOu78HQI9yqAJ598kvz8/OHeFiGEOCMcOnSIm266CcwM7et0CPcIQH5+PkVFRcO9LUIIcabpt50tB1SFEMKBJNyFEMKBJNyFEMKBJNyFEMKBJNyFEMKBJNyFEMKBJNyFEKKPT//uHR5+de9hy//n5T38du1+OkJh7vnHByx9YC2bShsHfJ6HXtnDh/5nHcNx3YxBjXNXSi0FHsQ43fVRrfVP+tw/GngcyDDXuUtrvdq87xzg10AaEAXmaq27hmwPhBBiCFU0dfLGnjr0oVa+tHA8brfLvu//NpYTicbwed38YX0JAK/uquG80ZmHPU9JfTv//fJeQpEoFU2dFGUmnbJ9gEFU7kopD/AwcBUwFbhRKTW1z2p3Ayu11rOAZcAj5mO9wB+B27XW04BLge4h23ohhBhib+6pA6CmNch7JT1VeSQao7Kpk4qmTlZvq6IgPcDYnGT2VLcd9hzfe3obH//1esLRKABby5oBCEeip6yKH0xbZh6wV2u9X2sdAlYA1/dZJ4ZRmQOkA5Xmz0uA97XWWwG01vVaa5kcTAhxTDaVNtLSdfx1YVd3hFgsxtrdtfx1Y/kR1123t47sZB9+r5tn3q+0l1c1dxKOGsH8zoEGZo/JZGJeCrtrWnli/UG+9X9bCYWjNHWEePKdUvLTAvzyU7Pxedy8X95EZyjCkl+s5d5ndhz3fhyLwYR7IVAWd7vcXBZvOfAppVQ5sBr4qrl8EhBTSq1RSm1SSn3nBLdXCHGWCYYjLPv12zz0yuE98MGoau5k1r0vsmb7IR58eQ/3PrOjV/XcHgzTGYrYP7+1r45LJuWyZFo+K94t4/XdtQCUNXT2et65xVlMGpFKSX0Hj7y2j6c2lvOdp7ZysL4DgDsWTeTKaflMHZnGlrImHnltL/vr2nnqvXL79U6moTqgeiPwmNa6CLgaeEIp5cbo6V8E3GT+e4NS6vIhek0hzkrVLV1Eo86+sP27Bxp4/oNDANS0BAlFoqzfV39cz/XC9mo6uyO8uKOG7ZXNNHd2U95oBPUvXtzNnB++xO1/3MjOqhYW/PQV6tpCXDU9n/uun8aEvBS+8If3WLenjrJGI7RzUvwAzCnOZOKIFCLRGFXNXYzLTebvWyp5a5/R1hmTbfTYZxals7msiV+9vo/J+am0BsOs2X7ohN6fwRhMuFcAo+JuF5nL4n0OWAmgtV4PBIAcjCp/rda6TmvdgVHVn3eiGy3E2aq6pYuLfvoKzw8QDpFojMhpGPz/9tT7PNVPO+Tvmyt4oc++xGIx7vrb+9z1t/eJmsEJsL2ymVazNdMeDPORR97kNV1z1Nd+cUc1AKu3VdHVbfTAt1U0s6m0kQdf3kNWso/Xd9eyfNV2IpEYf/3SfJZMyycjyceTnz+fcTnJfP4PG3hd1+J2wUfOKyQnxc/k/DQm5qXar/O1yycCsMb8UBqdZYT77OIsQuEoF4zL5k9fuIBRWYn8dVM5lU2dLPzPVyk1K/2hNphw3wBMVEqNVUr5MA6YruqzTilwOYBSagpGuNcCa4AZSqkk8+DqQuDUNJzEWeeJ9Qf58MNv0tU9uK+8HaEwf9tUPizD1I7XjqoWuiMxdle3AtBmVoHbypuJxWJ89rEN/OuKzcO8lb11hMKs3Fh2WBBHojGW/3M7v1m7v9fyzWVN7K9tp6mjm/11bVQ1G1V2NIZ9gHNrWRObSpv41v9tpb7tsOtU2Jo7unl7fz1pAS+dcb8X2yqa+f26A6QGvPz+1rmA0Uf/6OwiZo/pGfmSmezjD5+dRyQa49ltVRSkJ3Lnkkm88I1L8LhdjMtNxu2C8bnJXD5lBC4XbC1vJj8tQCDBA8C1Mwr425fn8/hn5pGV7GPhpFy2ljWx4WADJfUddHSHT+DdHdhRw11rHQbuwAjqnRijYrYrpe5VSl1nrnYn8AWl1Fbgz8CtWuuY1roRuB/jA2ILsElr/ezJ2BFx5tta1sS/P73tuFsOL+2sMXqb5vjkV3fVsK/28JEMlme2VvHNlVt550ADHaEw4Uj0uF73eLQHw4TCx/56+2qM/alsMgLvj2+XcNsTG/nQQ+t4/oNDvLm3juc/OETdEQKvr2g0xv4jvE+WtmCYg3Xt/d7X3NnN9//+ATWth49y3lnVQiwGjR2hXsu3ljfR1NFNTWvvbX1qYzkec/jhhoONVLcYz+lxu3hd1xKLxfigstl+3UX/9TqPvtH7A8Lywo5DhKMxvnLZBAAykhKYUpDGyzuree6DQyybOwqVn8oF47IAuOn80Yc9R15agIWTcgEYlZWI3+shK9kHQCDBw9Lp+dw4bzQpfi9jzGp9dHbPsEe328V5ozPtIZUTclNo6Qrz7oEGXC4ozk7ud9tP1KDGuZtj1lf3WXZP3M87gAUDPPaPGMMhhTiih17dy4s7qvnKZRN48u0Srp5RwPTC9EE/ftehFlwu+NXr+7nx/NF8+clNLJk2ggeXzbLX+ceWCuaNzaIgPZH9ZlC9uquGbz+1lcVT8rnnQ31H+Q4sGI5QWt/BxBGpR185TiwW45OPvkNbVzdP3T6fTDMoBsMadme1KnZXt5KT4qOrO8pPn99lj+Z4Zmslty4Yaz8uGo3hchljuD/9u3d55KbzmFJgDHB7YUc1t/9xIyu+eAEet4uC9ABZyT52HWrtNX77R8/u5LkPqth092I7qDaWNFLb2kVJfQdPvF1CeyjM/R8/t9c2b69sAaCx3Wip1LUF+ebKraT4jcq2uqWLWCyGy2U850s7qlk6PZ939tez4WAD6YkJJPs8zB2bxWNvHWRfbRvpiQkUZiTywLJz+dnzu/jZGs0yM2Djn+vP75YyLjeZmy8s5ucvaGYWZZCfFuAv75WRk+LnsxcZ79F3r5rCptLGAf9fXnduIS/trGFUP2PVH7lptv3zlII0DtZ3UJw98Jj28XkpgNEuKspMtCv8oSZnqIrTQlNHyP7avrGkkUde28c/tvQ9tDOwhvYQ1S1Brj1nJKFIlBXvltHZHelVuZfUt/O1FVv449sl9m2Ax946SFlDJ899UHVYi6a1q5uOUP9fmx9+ZS9X//cb1LcFue2J93j1CP3fru4Ie2uMVsr2yha2ljWxr7adLz+5acDH7KttY+kDa3luWxWPvrGfpQ+sRZvtmAqzcj9Y187EvFQWqlwO1neQ4HExPjeZp7dU0tQRYukDa1n6wFpm3vsCN//+Xd7e38CBunYefeOA/To7q4zwvecfH7DsN29z7z938KvX9vHRX75lv044EuX5D6po6uimqqWnOn/gpd3c8afN/G7dATxuF3/bVMFbe+t67ccHFUaVbVXub+6tY+3uWlZvM3rTwXCUls6w+XOEmtYgk/JSmTMmi/fMyj0/PcAjN53Hly4dzxt76nhxRzXTC9OYW5zFnUsUoXCU13Ut3ZEoC//zNX6+RrPrUAubSpv45LzRJPo8fO/qKdx2yTgum5zHqKxEnvz8+RSkJwIwc1QGn4n7MOzriil55KX6mTkqY8B1APsDc8wRqvHxuUa417QGGZeTcsTnOxES7uK08Oy2KrojRrC+tssISWtEw0Dig3jXISOgbpg1Ep/XzZ/fLQXgQG27vZ51YM2qeg+YlXvQbI9UNXexu88JKbc9sZF/++u2w147Go3x100VdEdirHyvnDXbq/nlq/sG3NbfrTvA4l+sZd2eOp7aWI7P6+b2heNZv7/eDlDLgbp27ly5lU/8ej27DrXyz/creWFHNbsOtbKlrAkw2jKxWIwDde2MzU1mydQRAMwsyuCT549ha1kT//HPHew61Ep+eoCizCTW76tnS5nRs37mfSP8AQ6aH3K7q9uIRGOs21vHs9uqiMXgdW0MA3z3YAONHUblvbemrde2hqMxalqDfP+aKRRlJnLT795h+artBMNGj9uu3M3X21HVQoLHRVrAy9xi45tBSUM7T28up6rJ+H8zMiPA3LFZlDZ0sLm0iYL0RJJ8Xr66aAIpfi/BcJTpI41vdXOLs8hK9rFm+yG2V7ZQ2tDBQ6/u5dO/exef181HzzOu8HbrgrHMn5DD0un5vPGdRaj8wX/jSvJ5eeuuRXzqgjFHXM8Kd+tgan8K0gMk+YxqfVzuyWnJgIS7GAYbDjbwm7W9g/DZ96sYl5Ns9FXNccV9Qy/eE+sPMu/HL9sHT3dVGRXtjMIMZhSm233c9lCE6hbjZyvcDzUbbYCS+g7mmAfPPj7HCID46jsWi7GtvJkdZn833sbSRnv7fv+mUQW/e7CB0voO1u6u5fwfv9TrQN/LO6uJxeArf9rEig2lLJk6ghtmGaeLrNtT2+u5f7N2P6u2VqDyU5k9JpONJY1sK+/ZhlFZiXR1RzlQ105jRzfjcpK5dFIegQQ3l0zK5WOzi0jyeXh6cwWzRmfw2Gfm8fUrJhKOxvjn1iqyk30Ew1H+usn4ZnSgrp3zRmfwsdlFfOOKSXSEIuyrNQL/9d01vLm3jl+/vp8Ej9Hq2GN+ewiGI1Q2dTKzKJ1ZozP4xNzRrP7axdx8wRgee+sg1/73Oh5+dS+7q1vxe910dUfpDEXYUdnCpBGpbLj7Cr595WR7n7/xl632EMHCjEQumZgDGB+6I9ICgBGyH5pZAMD0IiPcPW4XV0zJ49VdNbxh/u5cPDGHsdnJPP6ZecfU9joSr+focXnxxBz+ddEELpucN+A6LpfLrt7H5UrlLk4zh5q7uO6hdfaojcGKxWL84B/b+fHqXaw1/xBburp590ADS6blU5iRSH27UeFVNHby9OZyvtZn9EdzZzc/f2E3ta1Btlc2s+jnr/HAS7vJSfGRm+pnlvnV2ec1fr1//+YBrnrwDTYcbLC3vaY1SGd3hOvOHckvbzqP5ddNY3J+Ks9tq7IP6Na1hWgNhilr7DzsIO+qLZUkJngYl5tMbWuQtIBx+Grle2X8bM0uqluC7DpkvDfNHd1sKWvi6hn5FGUmcuW0fO66ajKTRqSQl+rnjT09bYxoNMZLO6tZPHUET37+Aq6ZUUB1i7Gt1kG8SyYaB/feNMd9j81JJj0pgZfvvJTbFo4jPTHBrlY/Oc84QDizKMN+766YMoKZRek8tbHcrv6nF6bz83+ZyW0LxxFIMN63C8Zl8fLOGm569B1e313LdTMLyUxKYF9tG6/vrmVPdRvRGNy6oJinv7yARJ+HtEAC/3H9dH578xw8bhf/uUYT8Hq45hwjkBs7QuyobGFqQRp+r4d8M7St98D6YC/ISGRCXgqFGUbbpCA9YL9HX7xkPFfPyGdecZa97PpzC2kNhvn12v0UZyfxxOfOZ+XtF3Lh+Owj/DYOvUCCh28uUaT4j3w4c7xZsY+Xyl2cbM0d3fbJF4OxtbyJ98ubuc88lXpjSeOAIxbivVfSaH8tv++ZHYQjUdbtqSMcjbFocl6vr7P17SH+/E4Z/9hSSWcoQiQaY/W2Ku740yaaO40WwVMby9lf105LV5hzzACbZR4EvNysnn77xn7KGjrITwtwqcqlqrnLbskUZydz1YwCknxebp1fzNbyZh5ffxDoaduEwlFq+4w+2VzWyJziTOab4bFQ5XHFlBE89OpePqgw2hBlDcb45Tf31RGNwWcXjOXZf72YB5fNoigzCZfLxUUTc3hzb509Nn1reRO1rUEWm22WWaN7erz3XT+dmUXpXDXdCEprDpSxOUZAFGYYIzkA7lg0gdsuGceHZo4EID89wIg04+SbyQWpfHR2ETurWli3t47WrrA9YiOQ4GHR5DzUiFQ+u2As4WiMK6bkseWexfz8X85hQl4Kz75fxS2/f9c+jX5sP33jxVNH8NzXLmbz9xfz/vIlLJmaD4CubqW+PcTUkUb7Is/cJuv/53sHjbZRQXoAl8vFQpVrb79lbE4yj9w0m+S4AJ0/Ppvxucm0BcPMjQv905V14HaCVO7iZAqGI3zmsXf51KPvDHr+jhrzoNobe+p4Y08t//vmAX747M6jDql7Yn0JaQEv91w7lT01bejqVl7ZVUNawMt5ozPsIWQ+8yvwhhKj2tbVrXz44Tf58pOb2F7ZwjeumESK38vfNxtzf/z1Sxfy04+eA8Dc4kwSEzx85DyjPRGLwecvHstb372ciybk0Nkd4f1yo3cdPwztE3NHcfnkPH7y3C4a20O9hv2VNvScaBKJxthT3YYakcqsUcYHybziTP7nxll8ZFYh5xSl43G77GMGa3fXkhrwcm4/B+OumDKCxo5uvvCH91i+ajvL/7kDj9vFZcr4YJo6Mg2fx01Wso+rZ+TzjzsuYnKBEQxv7qvD43Yxqp/+7oi0AN+9ekqvkRhW9T45P40PnTOSBI+Ln6/RAIyNqyD/82MzWXnbhSyeOoL/vXUuD33yPDKSfLhcLibkGcP4wDiLFGDsAAcPXS4XmcnG4zKTEoxtNj+Qppq96UCCx/7WAxCKRMlJ8dnbbb0PhZmJ/b5G/GvdfGExAHPHnv7hftP5o/nNp2eTlxY4+srHScJdcP8Lu9lU2kQ01lNtHs2hli48bhepfi/Pf3DIHnGx8r3eZyF2R6KUN/Y858aSRhZNzrNHHZQ3dvLm3jounpiL1+O2K/c55oE265jp3zdXsK2ime9dPYX3vncFX7tiIlMKUunsjlCQHmD2mCxyU40qMC8twOZ7FrN46gi7qrX629boiLf21ZPgcTEyo+ePy+VycecSRTAc5Zn3K+2hktD7fSmpbycYjqLyU1k0OY9rZhRw5fR8En0e7v/Euay64yIK0gOUNXbYk1UtGJ/Tb8/2qun53HPtVNbtrWPle2XUtQb5xNxRZCQZLRi/18OF47O5eGKOPbwvO9mHz+umtSvMmKwkEgbRCwbjwKPP42ZqQRqZyT4+fG4hW81efnxAJ/u9pCcl4HK5uGxyXq8PCKtXbLWIMpMSSDeD+0isvvc6cyTNlJFp9n0j+gSc9f8IjG9fv/70bLsVdSSfmDuKu66azLVmC+h0lpHkY8m0/JP6GhLuZ6C9Na1M+f7z9oGtE/Xizmq7p9l3cqSBVLcEyUv1c86odN490GC3MP66qZzuuJOB/rC+hMX3r6UzFKGrO0JFUyfjclPsua331rRR1dxlf023TgK5uM8f83MfVAHw4VmF9hhrq/rrby5tK5AuN8PXGppmfb1ft6eOqSPTDwvcqSPTmJyfyt82V3Cgro3RWUm4XL0rd2320ifnGyH58E3nkZfaO6CKMhMpa+hgX20blc1dXDKp/3ByuVx89qKxbFu+hO3/cSVv3rWIH98wo9c6j94yh//6l5m9HnPzBWP48Lkj+c+45Udzy/xinv/6xXYY333NVPLTAnjdLoqOUhlbLp8ygsVTR9jbWJwzuJ5xpvlhtetQK6OyEkkL9HwgWOFujZyJ/8B1u11cOS3fPqnpSAIJHm5fOJ4k36BO33E8eRfOQO8dbKSzO8KWsqZjPoGmr+aObvbXtvOFi8fy2zcOUNrQzu1PbCQai3Hr/GLmT8jp93HVLV3kpQWYWZTBI68ZI1+WTsvn+e2H2FPdZof1phJjW2tau+x5PYpzkslMSiAxwcObZiVntUfmjs1i4aRcbphVyP0vaqIxSPF7qW4Jkpvqt6tzwH6N+L50X99conrdtsI9HI1x8QD7dsOsQv7fc7vITEpgTnEW3ZEopQ0d7K9t4xcv7aE7HMXlggl5A/dLR2Um8druWl7fbezfJZP6fy2L1SvvT3+V+d3XDv5kK4vP6+41OiM9KYFHb5nDjsqWQY0EAaPf/dub59AdiZIW8DLxCO9BvIy46n5aQe8T06y++9UzCthwsJGRGYP7oBFHJuF+BrJOzCkdZAvlSLaaveeFk/JY8W6Z2UM3Aqmps/uI4T42J9k+iAmwdLoR7hVNnXbwbjeHEda2Bu1T4sdmJ+NyGdWidQDNmkEvJ8XP45+dB8DIjEQSEzyk+L28V9LItLiv8gDzx+cwNieZRUcYdtZXXqofl8to91w0sf99+5c5o/i/jeXsrWljXE4yzZ3dvLW3nhe2V9MWNPrNY3OSSfQNHMijspKobQ3ywvZDjMtNPuVX4Rms6YXpx3QWsCXB42bl7ReSnew/+srm+qkBL61dYft3w3L55BFEojF7TpeR6RLuQ0HaMqeBV3fV8LFfvtWrnXEk1kkkB4dgNrktZU24XHDOqHQKMxN5yxxeN6UgjfIjfHgcMsceWwcJAwlue+SINe9Ja1e3vY01rUEO1Bk/F+cYQVeYmUgo0lPN9/XFS8bxpUvH2y2VvuE+KiuJV7916TGNFU7wuMlN8ZOY4Bmw4s9K9vHMVy/ivuunccv8YkZnJXGopYsJeSl2O2LSiCO/5qgsI6DeOdBgj3xxmsn5ab2+SR2N1Zqx2mmWa84p4MFls5icn8aN80Y79v061aRyPw28c6DBGCJY2cJX/rSJu6+ZytLpAx9s2WtV7vW9J3GKn1MDjFPT739xNxVNnTy47Nx+q8ctZU2Mz00hLZDAqKwkdh1qJZDgZtHkXB55bR+hcJRoLMa//fV9vrpoInlpflq7wrR0hRmRFrCH2OWnBchN9ePzuu2Te3ZW9RwTqG0NcrCunZwUP6lmv9Uaw5yT4u93XPBN5xtnA1oHM6eNPPYKsz8qP5XUgPeIrZBAgodPm6MvPrtgLJPzU7n5wmJ8XjfdkSjTC9MGfCxgz0GSlezjSwvHD8l2n+kyk32UNnQcVrlbfF43/+8jM/q9Txw7CfeTqLWrmx89u5OG9hB3LJrQq4URzzqTccWGMsobO/nZml12uHeEwr0OEHV1R+whdiVxlfXa3bXc8adNvHznpXY19fctFazaWonbBX/ZUMadffrPADsqW+yK2wqkaSPTKc5OJhYzqvBDLV38Y0sl6YkJHKzvsIcRWgfCvn/tVJL9XlwuF4UZiXa4x5/ZWdsaNE6Vz+n5gLE+bI40yRLAeWOMoY39HTg9Hr/59BxcRz8+Z5s6Mq1XIN0yv/ioj5mYl0pmUgL3XDvVHvlytstMSiAjKaHXCUni5JFwH6QtZU20dHYPOOqhP++VNLJig3GFwoL0wMDhbp6R+fRmYxih9bX15Z3VfOnJTTzz1YuYZB443VfbRixmXN1la3kz3/jLFkZlJbG3ppWWrjBby5q4wvxaW98WIiMpgRmF6fx9SwXfXDypV2UPxtmh1jA1a8TEjMJ0e0hiaUMHe8w20KqtlTR3dtvDE62TYq49Z6T9fIUZiXZbZntlC9nJPjxuFzWtXRyob+cy1fP+WWOXjzbiYv74HLYtXzLog35Hc6Re+VBJT0pg0/cXH/Z+n81uubCYK6d1yXtyikjPfZDuf3G3fWbk79YdsCdFOpI6c36T3FQ/Ow8NPGzRqtyt0STWpEKv6VpC4Si/er1nHhar375oshHgT2+u4JFX9/LyTmNOFB03PLKhPUSWOZ65rKGTTaU9V3IHo43T2R2xX886GeaconT757LGDrabs/o1dRjBbq2f388JGCMzAlSY3yw2lTZyTlE6eWl+9tS0Udsa7NUft9oyR6vcYXDzepxuJMR6u2xyHjfOO3y+dHFynHl/MafAjsoW/r6593SzNS1dtHSFufy/XuO+Z3bwmq4d4NE96tqMivziCTnsrGrhibdL+Ogv3zrsggbWepb2oPHBYc2FsmpLpX0i0O7qVrxuF5eaFbDf68btchEMR/G4XfZ8JsbzBslO9nHl9HxS/F4eemVvr5kUu7qjxGI9lez88dncOr+Yy6eMYERaAJ/HTWlDB9srW5gzJpMEj4sLx2Vzw6xCXC4Y0c/X65EZidS0Bqlu6WJfbTtzx2aRlxqwZzOcETcyY9KIFCbnp7JggBE5QojjJ+Hej8feOsD3nu49zasVwNa/rV1HvzRWfVvQ6BWPyaS1K8wDL+5mY0kjH37oTT7x6/XsqGwhFotR3x7sVQW3BcM0d3ajq1v5xJxRhKMx/rHFOM3+g4oWJo5IZeKIFBI8Lj46u4jPLChmbE4yl0zMQZtT34JRuWcnGwcrv37FRF7VtTz3Qc/1Kq15ypPME36S/V6WXzeN9MQEPG4XhZmJ7KtpY29tGxeOz+aXN83mvg9P4ztXTub3t87tdSKKxarG/7nV2N65xVnkpvjtVk78sLvUQALPf/0Sey4YIcTQOSvDfV9t2xEvIlzbGqTdPKMSjLlEGtqDvc6Sq209+mXM6tqC5KT67Dme69tDXH/uSEZlJfFBRTP3v7ibjlCEru4oH5tdxK3zi5lRmE57MMym0kZiMbj+3JFMG5nG67vNy4tVNDN9ZBpJPi9P3T6fu6+Zwl1XTeblby5kSkEa+2vb7cu31beHyEox+um3zC9makEad/xpEz98ZgcdoTAdIWP/BjqjrygzkbV7jEmtpo1M44qpI5iQl0p6UoI950dfVriv2lqJz+NmRmG6fZLK2Jxk0hOPfqq6EOLEnXXhXt7YweL7X7fn9u6PdYCzybw4QUN7iGgM7rhsAo/ePIckn2dQ16isazMq5/iLAnz7SsVfbruQz188jpd2VvP2fmNc+ZjsJJZfN4389ABtwTCbShrxuF2cOzqDhZNy2VTSyJ6aNurbQ3b1O3NUBkk+Y5SK2+1C5acSjsbYX2d8eDV2hMgxD5YmeNysuO0Cbpw3mkfXHeCa/15Hg7mfAx1gnJCXQigcxe91D7q6Ls4xLhj8fnkzM4rSCSR47NE7M47jZBkhxPE568J9b40xB3XfKVzjWQdCrfCzglzlp3LF1BHkpPgHDPemjpBcRLWFAAAbQElEQVR9sLWuLWiP4S7OTmJqQZo9/O+W+cUEEtw8+PIewBjrDcap9u2hMBVNneSnBUjyeVk4KZdwNMavXzem1B3ojMLJ+cY3BH2olaaOELFYzwRPAGmBBH50wwzu+/B0DtS125N9JQ0Q7l+/YhJ/+eIFvPu9Kw6b3GkgIzMSefSWOcwek8nHZhtziuea+3ZOkYS7EKfKWTcUsswcydER7L9nHovFqDND3bosmBXkVgDnpPgGDPfrHnqTD80s4NtXTqauLWSfBflfH59JYkLP252V7GNucZZ9qn+22T5J9ntoD0Zo7ui25+M4b0wmaQEvT28ux+06/Aw/i3UKf0l9B1MKQubzHn4GoTUDoHWFooEq9/TEBM4fd+wXO1g0eYQ9mgeMD0W/1y0HToU4hc66yt0627E91P9QxtZg2O5Z963cc8wAzknxU9caOuyxTR0hShs6qGjstPv01twbs8dkHXZmXnyrwwrhZJ+XtmCYps6ecE/wuPnVp2aTkeRjemH6gGEcSDCubFPa0EG9eeA3u59LjKWY82dbo3ZO9ix643JT2HXfUvvYgxDi5Dvrwr3UnOskvnIPhaM8tbGccCRqhyL0hLt18DTH7B3npPbflrGuO9naFaaxw+jTWx8I/bEmSoKeEE72ewmFo9S1BclI7Hns/Ak5vPqtS/nfW+cecf9GZyVRWt9Bfbuxff1V7tap/tZ1RgdqywwlGfMtxKl1FrZlDq/c73tmB0+8XUJ6YkKvqUl7KvcQfq+bVDMUc1L8NHSECEeivU6usa5C1BoM2x8SOUeYWOncURm4XEa1HogbjghQ1dR1WBtjMCNNRmcnsW5Pnb3tWf1U7qmB3uGemHDyw10IcWqdVeEei8V6KndzjHdXd4Qn3i4BoKWzm0i0Z2ZGu+feahwYtarP3BQfsZgR/vGXybKu3NPWFT6sT9+f9MQEJualEAz3vGaK3wjaUCRKxnEMG7RmMKxo6sTlwr68WTwr3GtbrLaMhLsQTnNWhXtzZzetZjvGOgvUOtkGjN56R7fxlhgXiOhi+artvHOgoVcFbgV2bVuwd7jblXv3YX36gXzxkvE0tve0guIv+psxiMuX9WUdVN1a1kRGYkK/p+0nJnhwu3pGDMmVa4RwnkH9VSullgIPAh7gUa31T/rcPxp4HMgw17lLa726z/07gOVa658P0bYfs/hLyFmV+96aNnxeN26XEe5JZrtmfG4yb+yps0/0ib8MmRX0facN2F8bX7mbbZkjVO6APVzQ0ivcE499NkFrTpitZc0DXlTY5XKR4vfS0hXG5TLmYhdCOMtR/6qVUh7gYeAqYCpwo1Kq7zW+7gZWaq1nAcuAR/rcfz/w3Ilv7omx+u05KX67517bFiQ3xW+OXQ9R3x4kMymB3NSAHezQu3dtBXZd3Fmq4UiUg/U9B1Tr24wzWo/1jMz4ec0Hc+HhvqzZHDu7I1w/c+SA61lzqicmeORgpxAONJjKfR6wV2u9H0AptQK4HqMSt8QAa5xbOmD3OpRSHwYOAL2vLDEM2sz5YEZmBOzRMrWtQXJS/bgwKvfOkJfsFD9ZyUb4jclO4sFls3pdtNc64/JQS88EYKUNHXRHYozLSWZ/XTtVzV1kJvmOOTjj+9/H03PPTvaR6veSm+bniwvHDbie9SEi/XYhnGkw4V4IlMXdLgfO77POcuAFpdRXgWTgCgClVArwb8Bi4FsnurEnyjpzNDPJR32b0R+vawtRmBEAXJQ3dtAViJCT4rPnOJ81KsO+lJwlxe9lfG4y7xxoYHphLZtLG+0LPF8wPpv9de2U1LfbHxDHIr5yz+xnpMvRuFwu/vuTsxidlXTEKw1ZY91PxdzmQohTb6iarTcCj2mti4CrgSeUUm6M0P+F1rptiF7nhFijUjKTEmgP9VTuual+clN91LWFOFjfQWFGElnm1XMGmlPlMpXH2/vrWb5qOw+8tIeXdlb3ulpQaUNHv8MQj6Z3z/34Jtm6TOUx/ijXFbVGzCQlyMFUIZxoMOFeAYyKu11kLov3OWAlgNZ6PRAAcjAq/J8ppQ4CXwf+XSl1x4lt8vGzwz3ZR0cwYp9FmmP33IPUtgaZUpBq99UHuojyZZPzCIWjHDCHP67eVsX0wjQ7kOvaQscV7vGVe9pJnEHReh2p3IVwpsGUbRuAiUqpsRihvgz4ZJ91SoHLgceUUlMwwr1Wa32xtYJSajnQprV+aCg2/HhY0wpkJPoIRaLUtgbNs0j9va6pOTk/jfPGZODzugecyXBOcSbJPg8JXjcel4v69hAzizLsihj6P4HoaPxeNx63C5/HbZ/YdDLYlbuEuxCOdNTKXWsdBu4A1gA7MUbFbFdK3auUus5c7U7gC0qprcCfgVu11gNPmD5MguEoPo/b7jeXmKNbclP9vYYsTi5IJcnn5UMzRw54QNTv9fDtKxXfv2YqC83rqs4clWE/N2C3do6Fy+Ui2ec5rjHux0IOqArhbINquJpj1lf3WXZP3M87gAVHeY7lx7F9QyoUjuLzukk2A62koWdopMVq0QzGrQvGAsbJRs9uq2JucRbdkZ6zTY/ngCgYwXsyWzLGaxjPLycwCeFMZ9VfdjAcwe91k+Q/vHK3ri06pSB1wMcP5PIpI9j6gyUEEjy9zjY9nrYMGAdVT3rlLm0ZIRztrAr3wyr3eqty92H1kNSIYw93wO6Pp5xgzx3gqhkF/c4JM5RSZSikEI7m2HCPxWJsKWsyZ140+uZB85JxVqCVNnTg97pJ8RuXqvvxDTO4eOKJXVAiweMmkOCmqzt63OH+zcWTTmgbBiNVeu5COJpjJxXZUtbEDY+8xXsljfaynsrdCLaDde3kpvbM9vjJ80fbc7OcCOvU/uMN91Ohpy3j2M93Ic5qjg13az51a4pfsHruHpLNaXVbusIUZvQ/udaJsKrizOMYLXOq2OPcZS53IRzJseFunYEaP/9LKGJU7vHV6kBnoJ6I1ICXZJ/npI5TP1Eyzl0IZ3NsuFszOtbEhXuw2+i5J8eF+3kDnIF6IlIC3uMeBnmqFGcnc+v8Yi4xx+gLIZzFsQ3X9mD/lXtKwNtrhMjJqNznFWczKvPEe/cnk9fjZvl104Z7M4QQJ4ljw92q3KtbeuZctyp3n7fnC0vuEa5xery+dsXEIX9OIYQ4Fo5ty1g995rDeu49VbtbrlEhhHAo51bu5jVSa1qDRKMx3G4Xwe4IPvOaon/78nxGpg/9SBkhhDgdODbcrco9HI1R3x4iN9VPKBLFb14v9LyT0GsXQojThWPbMp1x1z+tNlsz1qyQQgjhdI5NuvZQBK/ZVI8Pd6tyF0IIJ3Ns0nUEw4zONoYjVrcEicVihMJR/FK5CyHOAo5NuvZQhDFZVrh3ETLnWfefxmeNCiHEUHFsuHeEwqQnJpCRlEB9e9C+xJ703IUQZwPHJl17MEKS30tOip/6tpB9cWzpuQshzgaOTbqOUJhkn4fsZB91bVK5CyHOLo5Mumg0RkcoQpJPKnchxNnJkUnX2W2McU/2e8hJ6Vu5ywFVIYTzOTLcrbNTk3xeslP8tHSFaQt2A+D3OnKXhRCiF0dOP2DNK5Ps9+B2GbM+VjYZJzL5JNyFEGcBR4Z7fOVuXemusqkTkMpdCHF2cGS4W3O5J/u8JPqMMLfCXSp3IcTZYFDhrpRaCjwIeIBHtdY/6XP/aOBxIMNc5y6t9Wql1GLgJ4APCAHf1lq/MoTb3y/rKkxJfmMoJECF2Zbxe+WAqhDC+Y5axiqlPMDDwFXAVOBGpdTUPqvdDazUWs8ClgGPmMvrgA9prWcAtwBPDNWGH0l85Z6dYvTcq5qlchdCnD0Gk3TzgL1a6/1a6xCwAri+zzoxIM38OR2oBNBab9ZaV5rLtwOJSqmhv65dH3bl7vOQ7PPg97ql5y6EOKsMJukKgbK42+XmsnjLgU8ppcqB1cBX+3mejwKbtNbBfu4bUnbl7vficrnISfHT2CFDIYUQZ4+hSrobgce01kXA1cATSin7uZVS04CfArcN0esd0baKZjxuF8l+o78+MiNg3yc9dyHE2WAw4V4BjIq7XWQui/c5YCWA1no9EAByAJRSRcDTwM1a630nusFH89KOap7aWM5nFxTbQX7uqAz7fum5CyHOBoNJug3ARKXUWKWUD+OA6ao+65QClwMopaZghHutUioDeBZj9MybQ7fZA3t8/UHGZCfxrSuVvWz2mCz7Zwl3IcTZ4KhJp7UOA3cAa4CdGKNitiul7lVKXWeudifwBaXUVuDPwK1a65j5uAnAPUqpLeZ/eSdlT0wH69s5pyijV/tl9piei2F7zEvvCSGEkw1qnLvWejXGgdL4ZffE/bwDWNDP434I/PAEt3HQuiNRKpu6uH5mUq/luaknfYCOEEKcVhzVo6hs6iQSjTE6K+mw+5J8ciBVCHH2cNT0A6UNHQD2hbHjrf3OZVS3dJ3qTRJCiGHhqHAvqTfDvZ/KPSfFT06KtGeEEGcHR7Vlyho68Hnc5KcFjr6yEEI4mKPCvbShg6KsRNwyIkYIcZZzVLiX1Hcwpp+WjBBCnG0cFe7VLV0UZCQO92YIIcSwc1S4d0ei+DyO2iUhhDgujkrCSDSGV/rtQgjhrHAPR2N4PBLuQgjhqHCPRGN4XBLuQgjhmHCPxWKEpS0jhBCAg8I9GjP+9bgds0tCCHHcHJOE4WgUAK/03IUQwjnhbma7zNcuhBA4KNztyl3CXQghnBPuEbPpLpW7EEI4KNzDZrhL5S6EEA4K957K3TG7JIQQx80xSSiVuxBC9HBMuEci0nMXQgiLY8LdGi0j4S6EEA4KdxktI4QQPRwT7tJzF0KIHo4Jd6nchRCih+PCXeaWEUII8A5mJaXUUuBBwAM8qrX+SZ/7RwOPAxnmOndprVeb930X+BwQAf5Va71m6Da/R1jGuQshhO2oSaiU8gAPA1cBU4EblVJT+6x2N7BSaz0LWAY8Yj52qnl7GrAUeMR8viEXkZ67EELYBlPmzgP2aq33a61DwArg+j7rxIA08+d0oNL8+XpghdY6qLU+AOw1n2/IyVBIIYToMZhwLwTK4m6Xm8viLQc+pZQqB1YDXz2Gxw4JqdyFEKLHUDWobwQe01oXAVcDTyilTmnzOyyjZYQQwjaYAK4ARsXdLjKXxfscsBJAa70eCAA5g3zskLCmH/DKAVUhhBjUaJkNwESl1FiMYF4GfLLPOqXA5cBjSqkpGOFeC6wC/qSUuh8YCUwE3h2ibe/Fqtwl24UQYhCVu9Y6DNwBrAF2YoyK2a6UulcpdZ252p3AF5RSW4E/A7dqrWNa6+0YFf0O4HngK1rryMnYkZ6eu6S7EEIMapy7OWZ9dZ9l98T9vANYMMBjfwT86AS2cVBktIwQQvRwTJkro2WEEKKH48JdKnchhHBguMvcMkII4aBwl3HuQgjRwzHhLqNlhBCih2OSUCp3IYTo4Zhwj5hDIWW0jBBCOCjcpXIXQogejgl3a24ZCXchhHBQuNuVu0vCXQghHBPukWgMtwvcUrkLIYRzwj0cjckwSCGEMDkmDaOxmPTbhRDC5JhwD0diMgxSCCFMjgn3SDSKR+aVEUIIwEHhbvTcJdyFEAIcFO6RqPTchRDC4phwl9EyQgjRwzFpKJW7EEL0cEy4hyXchRDC5phwj0SjEu5CCGFyTLjLOHchhOjhmHCXnrsQQvRwTrjHpHIXQgiLc8JdKnchhLA5JtyNnrtjdkcIIU6IdzArKaWWAg8CHuBRrfVP+tz/C+Ay82YSkKe1zjDv+xlwDcYHyYvA17TWsaHZ/B5SuQshRI+jhrtSygM8DCwGyoENSqlVWusd1jpa62/Erf9VYJb583xgAXCOefc6YCHw2hBtvy0cjeJPGNRnlRBCON5g+hjzgL1a6/1a6xCwArj+COvfCPzZ/DkGBAAf4AcSgOrj39yBSeUuhBA9BhPuhUBZ3O1yc9lhlFJjgLHAKwBa6/XAq0CV+d8arfXOE9nggciskEII0WOoj0AuA57SWkcAlFITgClAEcYHwiKl1MVD/JqAdQ1VCXchhIDBhXsFMCrudpG5rD/L6GnJANwAvK21btNatwHPARcez4YeTTgawysX6xBCCGBw4b4BmKiUGquU8mEE+Kq+KymlJgOZwPq4xaXAQqWUVymVgHEw9aS0ZYyeuwyFFEIIGES4a63DwB3AGoxgXqm13q6UulcpdV3cqsuAFX2GOT4F7AO2AVuBrVrrfw7Z1scJR6PScxdCCNOgxg5qrVcDq/ssu6fP7eX9PC4C3HYC2zdo0SgyWkYIIUyO6WNI5S6EED0cE+4yzl0IIXo4JtxlnLsQQvRwTLhHIjJaRgghLI5JQxnnLoQQPRwT7tJzF0KIHo4JdxktI4QQPRwR7tFojGgMmVtGCCFMjgj3SMw4KVYqdyGEMDgj3KNGuHvkgKoQQgAOC3ep3IUQwuCIcA9blbuMcxdCCMAh4S6VuxBC9OaIcA9Ho4DMCimEEBZHhLtU7kII0Zsjwj0csXruEu5CCAEOCffuiNGWkbllhBDC4Ihw7whFAEjyDerCUkII4XiOCPe2YBiAFL+EuxBCgEPCvd0M92QJdyGEABwS7j2Vu2eYt0QIIU4Pjgj39qDRc5fKXQghDA4Jd2nLCCFEPGeEe8gMdxktI4QQgFPCPRgmMcEjJzEJIYTJEeHeFoxIS0YIIeIMKhGVUkuBBwEP8KjW+id97v8FcJl5MwnI01pnmPeNBh4FRgEx4Gqt9cEh2XpTezAsI2WEECLOUcNdKeUBHgYWA+XABqXUKq31DmsdrfU34tb/KjAr7in+APxIa/2iUioFiA7Vxlvag2Gp3IUQIs5gEnEesFdrvR9AKbUCuB7YMcD6NwI/MNedCni11i8CaK3bTniL+9Em4S6EEL0MJhELgbK42+XA+f2tqJQaA4wFXjEXTQKalFJ/M5e/BNyltY4c9xb3oz0UJi81MJRPKYQQZ7ShPqC6DHgqLry9wMXAt4C5wDjg1iF+TdrlgKoQQvQymHCvwDgYaikyl/VnGfDnuNvlwBat9X6tdRj4O3De8WzokbTJAVUhhOhlMOG+AZiolBqrlPJhBPiqvisppSYDmcD6Po/NUErlmrcXMXCv/ri1B8NyApMQQsQ5aribFfcdwBpgJ7BSa71dKXWvUuq6uFWXASu01rG4x0YwWjIvK6W2AS7gt0O5A9FojI5QhCRpywghhG1Qiai1Xg2s7rPsnj63lw/w2BeBc45z+46qo9to70tbRgghepzxZ6jKpGFCCHG4Mz7c5SpMQghxuDM+3O3KXQ6oCiGE7YwP9zZpywghxGHO+HC3rsIkbRkhhOhxxod7a1c3ACkBCXchhLCc8eFe0dgJQEG6zC0jhBCWMz7cD9Z3UJAeIJAg49yFEMJyxod7SX07o7OShnszhBDitHLGh/vB+g6Ks5OHezOEEOK0ckaHe3swTF1bkDE5UrkLIUS8MzrcS+o7ABiTJZW7EELEO8PDvR2AMdlSuQshRLwzOtwPWpW7hLsQQvRyRod7NBZjfG4yqYGE4d4UIYQ4rZzRp3XevnA8n1lQPNybIYQQp50zunL3uF0kyWyQQghxmDM63IUQQvRPwl0IIRxIwl0IIRxIwl0IIRxIwl0IIRxIwl0IIRzodBhH6AE4dOjQcG+HEEKcMeIys9+LWZwO4V4AcNNNNw33dgghxJmoANjXd+HpEO4bgIuBKiAyzNsihBBnCg9GsG/o705XLBY7tZsjhBDipJMDqkII4UCnQ1vmmCmllgIPYnwteVRr/ZNh3qRjppT6PXAtUKO1nm4uywL+AhQDB4GPa60blVIujP29GugAbtVabzIfcwtwt/m0P9RaP24unw08BiQCq4Gvaa2H7WuaUmoU8AdgBBADfqO1ftDh+xwA1gJ+jL+1p7TWP1BKjQVWANnARuDTWuuQUsqP8R7NBuqBT2itD5rP9V3gcxity3/VWq8xl5+WfwtKKQ/wHlChtb72LNnng0ArxvaGtdZzhvP3+4yr3M1fmoeBq4CpwI1KqanDu1XH5TFgaZ9ldwEva60nAi+bt8HY14nmf18Efgn2h8EPgPOBecAPlFKZ5mN+CXwh7nF9X+tUCwN3aq2nAhcAXzH/vzl5n4PAIq31TOBcYKlS6gLgp8AvtNYTgEaMAMP8t9Fc/gtzPcz3aRkwDWOfHlFKeU7zv4WvATvjbp8N+wxwmdb6XK31HPP2sP1+n3HhjrHDe7XW+7XWIYxq4Pph3qZjprVeCzT0WXw98Lj58+PAh+OW/0FrHdNavw1kKKUKgCuBF7XWDVrrRuBFjAApANK01m+bn+x/iHuuYaG1rrIqE611K8YffiHO3ueY1rrNvJlg/hcDFgFPmcv77rP1XjwFXG5WeNcDK7TWQa31AWAvxt/Bafm3oJQqAq4BHjVvu3D4Ph/BsP1+n4nhXgiUxd0uN5c5wQitdZX58yGMFgYMvM9HWl7ez/LTglKqGJgFvIPD99msNrcANRh/qPuAJq112FwlfjvtfTPvb8ZoYxzrezHcHgC+A0TN29k4f5/B+OB+QSm1USn1RXPZsP1+n4nhflYwP50dN5RJKZUC/BX4uta6Jf4+J+6z1jqitT4XKMKoOicP8yadVEop6zjSxuHelmFwkdb6PIyWy1eUUpfE33mqf7/PxHCvAEbF3S4ylzlBtfn1C/PfGnP5QPt8pOVF/SwfVkqpBIxgf1Jr/TdzsaP32aK1bgJeBS7E+ApuDWaI305738z70zEOMh7rezGcFgDXmQcXV2C0Yx7E2fsMgNa6wvy3Bnga48N82H6/z8Rw3wBMVEqNVUr5MA66rBrmbRoqq4BbzJ9vAf4Rt/xmpZTLPCDXbH7VWwMsUUplmgddlgBrzPtalFIXmP3Lm+Oea1iY2/E7YKfW+v64u5y8z7lKqQzz50RgMcaxhleBj5mr9d1n6734GPCKWe2tApYppfzmqJOJwLuchn8LWuvvaq2LtNbF5va8orW+CQfvM4BSKlkplWr9jPF7+QHD+Pt9xg2F1FqHlVJ3YLwJHuD3Wuvtw7xZx0wp9WfgUiBHKVWOcYT8J8BKpdTngBLg4+bqqzGGTO3FGDb1GQCtdYNS6j56zlC7V2ttHaT9Mj3Dpp4z/xtOC4BPA9vMHjTAv+PsfS4AHjdHeLiBlVrrZ5RSO4AVSqkfApsxPvQw/31CKbUX42D7MgCt9Xal1EpgB8aoo69orSMAZ9Dfwr/h7H0eATytlAIjV/+ktX5eKbWBYfr9ljNUhRDCgc7EtowQQoijkHAXQggHknAXQggHknAXQggHknAXQggHknAXQggHknAXQggHknAXQggH+v8tPlY5XvVTFgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f0842746470>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_od(model.col_dict['corr_0'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Fit of generative model parameters for the 50 cells.\n",
    "Red are true values, blue are inferred values, blue line is initialization."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-01-08T14:58:50.020733Z",
     "start_time": "2018-01-08T14:58:49.409541Z"
    },
    "run_control": {
     "frozen": false,
     "read_only": false
    },
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f08426fbeb8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f08348eb5c0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f08345aed30>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f0834561160>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "for p in init_params.keys():\n",
    "    if p is not 'sigma':\n",
    "        true = [np.array(truths[c][p]).max() for c in range(n_cells)]\n",
    "    else:\n",
    "        true =  [inv_softp(np.array(truths[c][p]).max()) for c in range(n_cells)]\n",
    "    learned = [np.array(model.mgen.genparam_dicts[c][p]).max() for c in range(n_cells)]\n",
    "    plt.figure()\n",
    "    plt.plot(learned,'bo')\n",
    "    plt.plot(true,'ro')\n",
    "    plt.plot([0,n_cells-1],[init_params[p],init_params[p]])\n",
    "    plt.title(p + '  ' + str(np.corrcoef(true,learned)[0,1]))\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-01-08T15:10:23.529404Z",
     "start_time": "2018-01-08T15:10:22.457279Z"
    },
    "run_control": {
     "frozen": false,
     "read_only": false
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f07f96c2c88>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_preds_bl(model, data = data, cell = 4,trace =1, figsize = (25,8))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For supervised learning we the use the SV class and pass the ground truth spikes. We use two datasets for evaluation to check whether we overfit."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-01-08T15:04:08.962822Z",
     "start_time": "2018-01-08T15:03:29.137368Z"
    },
    "run_control": {
     "frozen": false,
     "read_only": false
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Corr. Train/Val: 0.817/0.837 || Cost: -5.352 || Factor: 1.000 || Time upd./Eval.: 6.5 ms 0.1 s || BatchNr.: 100\n",
      "Corr. Train/Val: 0.836/0.854 || Cost: -2.851 || Factor: 1.016 || Time upd./Eval.: 6.4 ms 0.1 s || BatchNr.: 200\n",
      "Corr. Train/Val: 0.841/0.860 || Cost: -2.630 || Factor: 1.112 || Time upd./Eval.: 6.5 ms 0.1 s || BatchNr.: 300\n",
      "Corr. Train/Val: 0.840/0.857 || Cost: -2.414 || Factor: 1.190 || Time upd./Eval.: 5.5 ms 0.1 s || BatchNr.: 400\n",
      "Corr. Train/Val: 0.851/0.859 || Cost: -2.437 || Factor: 1.123 || Time upd./Eval.: 5.6 ms 0.1 s || BatchNr.: 500\n",
      "Corr. Train/Val: 0.853/0.866 || Cost: -2.491 || Factor: 1.044 || Time upd./Eval.: 6.4 ms 0.1 s || BatchNr.: 600\n",
      "Corr. Train/Val: 0.850/0.870 || Cost: -2.428 || Factor: 0.922 || Time upd./Eval.: 5.7 ms 0.1 s || BatchNr.: 700\n",
      "Corr. Train/Val: 0.849/0.869 || Cost: -2.194 || Factor: 0.970 || Time upd./Eval.: 6.7 ms 0.1 s || BatchNr.: 800\n",
      "Corr. Train/Val: 0.855/0.871 || Cost: -2.573 || Factor: 1.046 || Time upd./Eval.: 6.0 ms 0.1 s || BatchNr.: 900\n",
      "Corr. Train/Val: 0.856/0.869 || Cost: -2.358 || Factor: 1.070 || Time upd./Eval.: 6.1 ms 0.1 s || BatchNr.: 1000\n",
      "Corr. Train/Val: 0.858/0.869 || Cost: -2.303 || Factor: 0.975 || Time upd./Eval.: 6.5 ms 0.1 s || BatchNr.: 1100\n",
      "Corr. Train/Val: 0.855/0.869 || Cost: -2.067 || Factor: 1.091 || Time upd./Eval.: 6.7 ms 0.1 s || BatchNr.: 1200\n",
      "Corr. Train/Val: 0.864/0.870 || Cost: -2.186 || Factor: 1.041 || Time upd./Eval.: 6.8 ms 0.1 s || BatchNr.: 1300\n",
      "Corr. Train/Val: 0.859/0.871 || Cost: -2.375 || Factor: 0.935 || Time upd./Eval.: 6.0 ms 0.1 s || BatchNr.: 1400\n",
      "Corr. Train/Val: 0.855/0.865 || Cost: -2.481 || Factor: 0.913 || Time upd./Eval.: 5.4 ms 0.1 s || BatchNr.: 1500\n",
      "Corr. Train/Val: 0.864/0.871 || Cost: -2.099 || Factor: 0.986 || Time upd./Eval.: 6.4 ms 0.1 s || BatchNr.: 1600\n",
      "Corr. Train/Val: 0.863/0.868 || Cost: -1.970 || Factor: 0.989 || Time upd./Eval.: 6.5 ms 0.1 s || BatchNr.: 1700\n",
      "Corr. Train/Val: 0.869/0.871 || Cost: -1.980 || Factor: 1.032 || Time upd./Eval.: 6.1 ms 0.1 s || BatchNr.: 1800\n",
      "Corr. Train/Val: 0.866/0.870 || Cost: -2.140 || Factor: 0.924 || Time upd./Eval.: 6.5 ms 0.1 s || BatchNr.: 1900\n",
      "Corr. Train/Val: 0.870/0.869 || Cost: -2.258 || Factor: 0.965 || Time upd./Eval.: 5.6 ms 0.1 s || BatchNr.: 2000\n",
      "Corr. Train/Val: 0.862/0.865 || Cost: -2.162 || Factor: 0.908 || Time upd./Eval.: 4.2 ms 0.1 s || BatchNr.: 2100\n",
      "Corr. Train/Val: 0.870/0.868 || Cost: -2.199 || Factor: 0.983 || Time upd./Eval.: 5.2 ms 0.1 s || BatchNr.: 2200\n",
      "Corr. Train/Val: 0.867/0.872 || Cost: -1.946 || Factor: 0.974 || Time upd./Eval.: 5.1 ms 0.1 s || BatchNr.: 2300\n",
      "Corr. Train/Val: 0.870/0.864 || Cost: -2.105 || Factor: 1.043 || Time upd./Eval.: 5.7 ms 0.1 s || BatchNr.: 2400\n",
      "Corr. Train/Val: 0.867/0.860 || Cost: -2.052 || Factor: 1.061 || Time upd./Eval.: 6.0 ms 0.1 s || BatchNr.: 2500\n",
      "Corr. Train/Val: 0.865/0.855 || Cost: -1.909 || Factor: 1.164 || Time upd./Eval.: 5.8 ms 0.1 s || BatchNr.: 2600\n",
      "Corr. Train/Val: 0.874/0.867 || Cost: -2.190 || Factor: 1.014 || Time upd./Eval.: 6.3 ms 0.1 s || BatchNr.: 2700\n",
      "Corr. Train/Val: 0.869/0.871 || Cost: -2.089 || Factor: 0.983 || Time upd./Eval.: 6.4 ms 0.1 s || BatchNr.: 2800\n",
      "Corr. Train/Val: 0.872/0.863 || Cost: -1.696 || Factor: 0.915 || Time upd./Eval.: 6.2 ms 0.1 s || BatchNr.: 2900\n",
      "Corr. Train/Val: 0.873/0.859 || Cost: -1.861 || Factor: 1.170 || Time upd./Eval.: 5.8 ms 0.1 s || BatchNr.: 3000\n",
      "Corr. Train/Val: 0.877/0.865 || Cost: -1.975 || Factor: 0.959 || Time upd./Eval.: 6.3 ms 0.1 s || BatchNr.: 3100\n",
      "Corr. Train/Val: 0.877/0.863 || Cost: -1.903 || Factor: 1.003 || Time upd./Eval.: 6.1 ms 0.1 s || BatchNr.: 3200\n",
      "Corr. Train/Val: 0.874/0.862 || Cost: -2.114 || Factor: 0.897 || Time upd./Eval.: 5.4 ms 0.1 s || BatchNr.: 3300\n",
      "Corr. Train/Val: 0.879/0.862 || Cost: -1.873 || Factor: 0.994 || Time upd./Eval.: 5.7 ms 0.1 s || BatchNr.: 3400\n",
      "Corr. Train/Val: 0.883/0.861 || Cost: -1.872 || Factor: 0.986 || Time upd./Eval.: 5.6 ms 0.1 s || BatchNr.: 3500\n",
      "Corr. Train/Val: 0.881/0.857 || Cost: -1.834 || Factor: 1.066 || Time upd./Eval.: 5.5 ms 0.1 s || BatchNr.: 3600\n",
      "Corr. Train/Val: 0.874/0.848 || Cost: -1.905 || Factor: 1.049 || Time upd./Eval.: 7.2 ms 0.1 s || BatchNr.: 3700\n",
      "Corr. Train/Val: 0.878/0.852 || Cost: -2.025 || Factor: 1.001 || Time upd./Eval.: 6.3 ms 0.1 s || BatchNr.: 3800\n",
      "Corr. Train/Val: 0.885/0.861 || Cost: -1.693 || Factor: 0.984 || Time upd./Eval.: 6.4 ms 0.1 s || BatchNr.: 3900\n",
      "Corr. Train/Val: 0.885/0.863 || Cost: -1.524 || Factor: 1.012 || Time upd./Eval.: 6.8 ms 0.1 s || BatchNr.: 4000\n",
      "Corr. Train/Val: 0.884/0.853 || Cost: -1.638 || Factor: 1.056 || Time upd./Eval.: 5.5 ms 0.1 s || BatchNr.: 4100\n",
      "Corr. Train/Val: 0.895/0.858 || Cost: -1.724 || Factor: 1.015 || Time upd./Eval.: 6.4 ms 0.1 s || BatchNr.: 4200\n",
      "Corr. Train/Val: 0.878/0.841 || Cost: -1.817 || Factor: 1.140 || Time upd./Eval.: 5.6 ms 0.1 s || BatchNr.: 4300\n",
      "Corr. Train/Val: 0.889/0.855 || Cost: -1.569 || Factor: 1.041 || Time upd./Eval.: 5.7 ms 0.1 s || BatchNr.: 4400\n",
      "Corr. Train/Val: 0.886/0.851 || Cost: -1.810 || Factor: 1.082 || Time upd./Eval.: 5.4 ms 0.1 s || BatchNr.: 4500\n",
      "Corr. Train/Val: 0.890/0.852 || Cost: -1.685 || Factor: 0.988 || Time upd./Eval.: 5.8 ms 0.1 s || BatchNr.: 4600\n",
      "Corr. Train/Val: 0.892/0.853 || Cost: -1.768 || Factor: 1.091 || Time upd./Eval.: 5.6 ms 0.1 s || BatchNr.: 4700\n",
      "Corr. Train/Val: 0.895/0.854 || Cost: -1.466 || Factor: 1.032 || Time upd./Eval.: 5.8 ms 0.1 s || BatchNr.: 4800\n",
      "Corr. Train/Val: 0.898/0.847 || Cost: -1.396 || Factor: 1.063 || Time upd./Eval.: 5.8 ms 0.1 s || BatchNr.: 4900\n",
      "Corr. Train/Val: 0.894/0.856 || Cost: -1.398 || Factor: 0.992 || Time upd./Eval.: 5.6 ms 0.1 s || BatchNr.: 5000\n"
     ]
    }
   ],
   "source": [
    "from SV_C import *\n",
    "train_traces, train_spikes = data_chop(timebins, Traces_train[:n_cells], Spikes_train[:n_cells], sv_inds = np.arange(n_cells))\n",
    "\n",
    "recDict = set_rec_net(num_filters=[30,25,25,20,15], filtsize=[31,21,21,13,11])\n",
    "model = SV(recDict, BernoulliRecognition, batch_size=20, rng = rng, filename = None)\n",
    "\n",
    "model.lr_decay = np.array(0.995, dtype=theano.config.floatX) \n",
    "model.buffers = [100,50]\n",
    "model.facs = [1] \n",
    "\n",
    "model.eval_sets = [{'traces':Traces_train[:n_cells,:,:1000], 'spikes':Spikes_train[:n_cells,:,:1000], 'eval_T':1000, 'eval_rep': 1, 'test_gen':False}, {'traces':Traces_train[n_cells:100,:,:1000], 'spikes':Spikes_train[n_cells:100,:,:1000], 'eval_T':1000, 'eval_rep': 1, 'test_gen':False}]\n",
    "model.fit(train_traces,train_spikes,max_epochs= 50, learning_rate = 1e-3, print_output=True, print_freq=100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "run_control": {
     "frozen": false,
     "read_only": false
    }
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "hide_input": false,
  "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": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}