main_train.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1 Physical GPUs, 1 Logical GPUs\n"
     ]
    }
   ],
   "source": [
    "# some times Tensorflow has problems with memory growth, if that case happens run this cell.\n",
    "import tensorflow as tf\n",
    "\n",
    "gpus = tf.config.experimental.list_physical_devices('GPU')\n",
    "if gpus:\n",
    "  try:\n",
    "    # Currently, memory growth needs to be the same across GPUs\n",
    "    for gpu in gpus:\n",
    "        tf.config.experimental.set_memory_growth(gpu, True)\n",
    "        #tf.config.experimental.set_virtual_device_configuration(gpu, \n",
    "        #[tf.config.experimental.VirtualDeviceConfiguration(memory_limit=4096)])\n",
    "    logical_gpus = tf.config.experimental.list_logical_devices('GPU')\n",
    "    print(len(gpus), \"Physical GPUs,\", len(logical_gpus), \"Logical GPUs\")\n",
    "  except RuntimeError as e:\n",
    "    # Memory growth must be set before GPUs have been initialized\n",
    "    print(e)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import sys\n",
    "import os\n",
    "import argparse\n",
    "import json\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "import datetime as dt\n",
    "\n",
    "from core.data.data_loader import *\n",
    "from core.models.lstm import ModelLSTM, ModelLSTMParallel, GaussianLSTM\n",
    "from core.models.cnn import ModelCNN, ModelCNNParallel\n",
    "from core.models.mlp import ModelMLP, GaussianMLP\n",
    "from core.models.rnn import ModelRNN\n",
    "\n",
    "from core.utils.metrics import *\n",
    "from core.utils.utils import *\n",
    "\n",
    "import numpy as np\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def manage_models(config):\n",
    "    \n",
    "    type_model = config['model']['name']\n",
    "    model = None\n",
    "\n",
    "    if type_model == 'lstm': #simple LSTM\n",
    "        model = ModelLSTM(config)\n",
    "    elif type_model == 'gaussian-lstm':\n",
    "        model = GaussianLSTM(config)       \n",
    "    elif type_model == 'lstm-parallel':\n",
    "        model = ModelLSTMParallel(config)\n",
    "    elif type_model == 'cnn':\n",
    "        model = ModelCNN(config)\n",
    "    elif type_model == 'cnn-parallel':\n",
    "        model = ModelCNNParallel(config)        \n",
    "    elif type_model == 'mlp':\n",
    "        model = ModelMLP(config)\n",
    "    elif type_model == 'gaussian-mlp':\n",
    "        model = GaussianMLP(config)        \n",
    "    elif type_model == 'simple-rnn':\n",
    "        model = ModelRNN(config)        \n",
    "\n",
    "    return model\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Readed\n"
     ]
    }
   ],
   "source": [
    "# reading config file\n",
    "# load differents configurations \n",
    "#configs = json.load(open('config_lstm_parallel_internal.json', 'r'))\n",
    "configs = json.load(open('config_lstm_parallel_internal_all.json', 'r'))\n",
    "#configs = json.load(open('config_mlp_gaussian.json', 'r'))\n",
    "#configs = json.load(open('config_lstm_gaussian.json', 'r'))\n",
    "\n",
    "\n",
    "# create defaults dirs\n",
    "output_bin = configs['paths']['bin_dir']\n",
    "output_path = configs['paths']['save_dir']\n",
    "output_logs = configs['paths']['log_dir']\n",
    "time_steps =  configs['model']['layers'][0]['input_timesteps']  # the number of points or hits\n",
    "t_features = configs['model']['layers'][0]['input_features']  # the number of features of a tensor\n",
    "n_features = configs['data']['features']  # the number of features of data input\n",
    "data_file = configs['data']['filename']\n",
    "split = configs['data']['train_split']  # the number of features of each hits\n",
    "cylindrical = configs['data']['cylindrical']  # set to polar or cartesian coordenates\n",
    "normalise = configs['data']['normalise'] \n",
    "num_hits = configs['data']['num_hits']\n",
    "type_norm = configs['data']['type_norm']\n",
    "points_3d = configs['data']['points_3d'] # what kind of points: (rho, eta, phi) or (eta, phi)    \n",
    "\n",
    "type_model = configs['model']['name']\n",
    "optim = configs['model']['optimizer']\n",
    "arch = configs['model']['layers']\n",
    "is_parallel = configs['model']['isparallel']\n",
    "over_write = configs['model']['overwrite']\n",
    "\n",
    "loadModel = configs['training']['load_model']\n",
    "validation_split = configs['training']['validation']\n",
    "epochs = configs['training']['epochs']\n",
    "batch = configs['training']['batch_size']\n",
    "shuffle_train = configs['training']['shuffle']\n",
    "\n",
    "#create a encryp name for dataset\n",
    "path_to, filename = os.path.split(data_file)\n",
    "\n",
    "orig_ds_name = filename\n",
    "\n",
    "encryp_ds_name = get_unique_name(orig_ds_name)\n",
    "decryp_ds_name = get_decryp_name(encryp_ds_name)\n",
    "\n",
    "output_encry = os.path.join(output_path, encryp_ds_name)  \n",
    "if os.path.isdir(output_bin) == False:\n",
    "    os.mkdir(output_bin)\n",
    "\n",
    "if os.path.isdir(output_path) == False: \n",
    "    os.mkdir(output_path)\n",
    "\n",
    "if os.path.isdir(output_encry) == False: \n",
    "    os.mkdir(output_encry)\n",
    "\n",
    "if os.path.isdir(output_logs) == False:\n",
    "    os.mkdir(output_logs)        \n",
    "print('Readed')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[Data] Data loaded from  /data/track-ml/bracis/datasets/eta_n0.5-0.5_phi_n0.5-0.5/eta_n0.5-0.5_phi_n0.5-0.5_20200518171238_tracks.csv\n",
      "[Data] Data set shape  (30000, 30)\n",
      "[Data] Data train shape  (24000, 30)\n",
      "[Data] Data test shape  (6000, 30)\n",
      "[Data] Data coordinates  cylin\n",
      "[Data] Data normalization type  KindNormalization.Scaling\n",
      "[Data] Shape supervised: X(144000, 12) y(144000, 3) :\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>10</th>\n",
       "      <th>11</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-0.994612</td>\n",
       "      <td>0.191916</td>\n",
       "      <td>0.056131</td>\n",
       "      <td>-0.834097</td>\n",
       "      <td>0.273532</td>\n",
       "      <td>0.054053</td>\n",
       "      <td>-0.850836</td>\n",
       "      <td>0.208417</td>\n",
       "      <td>0.051335</td>\n",
       "      <td>-0.844055</td>\n",
       "      <td>0.084845</td>\n",
       "      <td>0.045496</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-0.759050</td>\n",
       "      <td>0.217656</td>\n",
       "      <td>0.052578</td>\n",
       "      <td>-0.643922</td>\n",
       "      <td>0.282194</td>\n",
       "      <td>0.050409</td>\n",
       "      <td>-0.665107</td>\n",
       "      <td>0.213588</td>\n",
       "      <td>0.046944</td>\n",
       "      <td>-0.585823</td>\n",
       "      <td>0.089207</td>\n",
       "      <td>0.038217</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-0.489969</td>\n",
       "      <td>0.225938</td>\n",
       "      <td>0.048941</td>\n",
       "      <td>-0.411146</td>\n",
       "      <td>0.286473</td>\n",
       "      <td>0.046023</td>\n",
       "      <td>-0.356408</td>\n",
       "      <td>0.217336</td>\n",
       "      <td>0.039704</td>\n",
       "      <td>-0.302234</td>\n",
       "      <td>0.091680</td>\n",
       "      <td>0.030187</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-0.160610</td>\n",
       "      <td>0.230030</td>\n",
       "      <td>0.044562</td>\n",
       "      <td>-0.024250</td>\n",
       "      <td>0.289575</td>\n",
       "      <td>0.038789</td>\n",
       "      <td>-0.017396</td>\n",
       "      <td>0.219461</td>\n",
       "      <td>0.031717</td>\n",
       "      <td>0.090440</td>\n",
       "      <td>0.092994</td>\n",
       "      <td>0.019014</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.386816</td>\n",
       "      <td>0.232996</td>\n",
       "      <td>0.037341</td>\n",
       "      <td>0.400637</td>\n",
       "      <td>0.291333</td>\n",
       "      <td>0.030809</td>\n",
       "      <td>0.452019</td>\n",
       "      <td>0.220590</td>\n",
       "      <td>0.020604</td>\n",
       "      <td>0.534277</td>\n",
       "      <td>0.094430</td>\n",
       "      <td>0.006020</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          0         1         2         3         4         5         6  \\\n",
       "0 -0.994612  0.191916  0.056131 -0.834097  0.273532  0.054053 -0.850836   \n",
       "1 -0.759050  0.217656  0.052578 -0.643922  0.282194  0.050409 -0.665107   \n",
       "2 -0.489969  0.225938  0.048941 -0.411146  0.286473  0.046023 -0.356408   \n",
       "3 -0.160610  0.230030  0.044562 -0.024250  0.289575  0.038789 -0.017396   \n",
       "4  0.386816  0.232996  0.037341  0.400637  0.291333  0.030809  0.452019   \n",
       "\n",
       "          7         8         9        10        11  \n",
       "0  0.208417  0.051335 -0.844055  0.084845  0.045496  \n",
       "1  0.213588  0.046944 -0.585823  0.089207  0.038217  \n",
       "2  0.217336  0.039704 -0.302234  0.091680  0.030187  \n",
       "3  0.219461  0.031717  0.090440  0.092994  0.019014  \n",
       "4  0.220590  0.020604  0.534277  0.094430  0.006020  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# prepare data set\n",
    "if type_norm == \"zscore\":\n",
    "    kind_norm = KindNormalization.Zscore\n",
    "elif type_norm == \"maxmin\":\n",
    "    kind_norm = KindNormalization.Scaling\n",
    "data = Dataset(data_file, split, cylindrical, num_hits, kind_norm, points_3d=points_3d)\n",
    "\n",
    "X_train, y_train = data.get_training_data(n_hit_in=time_steps, n_hit_out=1,\n",
    "                             n_features=n_features, normalise=normalise)\n",
    "\n",
    "print('[Data] Shape supervised: X%s y%s :' % (X_train.shape, y_train.shape))\n",
    "X_train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[Data] Saving distributions at  results/MEnGHa5DCmJGE9C9BpAg4i\n",
      "[Data] Shape Tensor y_train.shape: (144000, 3)\n"
     ]
    }
   ],
   "source": [
    "if normalise:\n",
    "    print('[Data] Saving distributions at ', output_encry)\n",
    "    data.save_scale_param(output_encry)\n",
    "    \n",
    "if type_model == 'lstm' or type_model == 'cnn' or type_model == 'gaussian-lstm':\n",
    "    if not is_parallel:\n",
    "        X_train_ = data.reshape3d(X_train, time_steps, t_features)\n",
    "    print('[Data] Shape Tensor X_train.shape:', X_train.shape)\n",
    "elif type_model == 'lstm-parallel' or type_model == 'cnn-parallel':\n",
    "    if not is_parallel:\n",
    "        print('DEBUG')\n",
    "\n",
    "    X_train_ = np.reshape(X_train.values, (X_train.shape[0], time_steps, n_features))\n",
    "    #X_train = data.reshape3d(X_train, time_steps, n_features)\n",
    "    y_train_ = np.reshape(y_train.values, (y_train.shape[0], n_features))\n",
    "\n",
    "    X1 = X_train_[:,:,0].reshape(X_train_.shape[0], X_train_.shape[1], t_features)\n",
    "    X2 = X_train_[:,:,1].reshape(X_train_.shape[0], X_train_.shape[1], t_features)\n",
    "    X3 = X_train_[:,:,2].reshape(X_train_.shape[0], X_train_.shape[1], t_features)\n",
    "\n",
    "    Y1 = y_train_[:,0].reshape(y_train_.shape[0],  t_features)\n",
    "    Y2 = y_train_[:,1].reshape(y_train_.shape[0],  t_features)\n",
    "    Y3 = y_train_[:,2].reshape(y_train_.shape[0],  t_features)\n",
    "\n",
    "    X_train_ = [X1, X2, X3]\n",
    "\n",
    "#print('[Data] Shape Tensor X_train.shape:', X_train.shape)\n",
    "print('[Data] Shape Tensor y_train.shape:', y_train_.shape)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "#!pip install livelossplot --quiet\n",
    "import keras\n",
    "from keras.callbacks import EarlyStopping, ModelCheckpoint, TerminateOnNaN\n",
    "from livelossplot import PlotLossesKeras\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy\n",
      "\ttraining         \t (min:    0.978, max:    0.995, cur:    0.994)\n",
      "\tvalidation       \t (min:    0.993, max:    0.996, cur:    0.996)\n",
      "Loss\n",
      "\ttraining         \t (min:    0.000, max:    0.001, cur:    0.000)\n",
      "\tvalidation       \t (min:    0.000, max:    0.000, cur:    0.000)\n",
      "[Model] Model for inference saved at /data/track-ml/eramia/compiled/model-lstm-parallel-MEnGHa5DCmJGE9C9BpAg4i-coord-cylin-normalise-true-epochs-30-batch-60.h5\n",
      "[Model] Model training stopped at 0 epoch\n",
      "[Model] Training Completed. Model h5 saved as /data/track-ml/eramia/compiled/model-lstm-parallel-MEnGHa5DCmJGE9C9BpAg4i-coord-cylin-normalise-true-epochs-30-batch-60.h5\n",
      "[Model] Model train with structure: [<tf.Tensor 'input_4:0' shape=(None, 4, 1) dtype=float32>, <tf.Tensor 'input_5:0' shape=(None, 4, 1) dtype=float32>, <tf.Tensor 'input_6:0' shape=(None, 4, 1) dtype=float32>]\n",
      "Time taken: 0:19:22.226278\n",
      "Time taken: 0:19:22.228899\n"
     ]
    }
   ],
   "source": [
    "# training model\n",
    "model = manage_models(configs)\n",
    "\n",
    "if model is None:\n",
    "    print('Please instance model')\n",
    "\n",
    "show_metrics = configs['training']['show_metrics']\n",
    "report = \"\"\n",
    "\n",
    "if cylindrical:\n",
    "    coord = 'cylin'\n",
    "else:\n",
    "    coord = 'xyz'\n",
    "\n",
    "ident_name = model.name + \"_\" + coord \n",
    "timer = Timer()\n",
    "\n",
    "# configure callbacks\n",
    "early_stop = EarlyStopping(monitor='val_loss', patience=50, verbose=1)\n",
    "terminate_nan = TerminateOnNaN()\n",
    "plotlosses = PlotLossesKeras()\n",
    "\n",
    "if not loadModel:\n",
    "    if not over_write:\n",
    "        # if exist, please used the compiled model!\n",
    "        if model.exist_model(model.save_fnameh5):\n",
    "            print(\"[Warning] Please there is a previous model compiled (%s) for %s file.\" \n",
    "                % (model.save_fnameh5, data_file))\n",
    "\n",
    "    model.build_model()\n",
    "    save_fname = os.path.join(output_encry, 'architecture_%s.png' % ident_name)\n",
    "    model.save_architecture(save_fname) \n",
    "\n",
    "    timer.start()\n",
    "    \n",
    "    check_point = ModelCheckpoint(filepath=model.save_fnameh5, monitor='val_loss', \n",
    "                    mode='min', save_best_only=True)\n",
    "    # live training\n",
    "    history = model.train(\n",
    "        x=X_train_,\n",
    "        y=y_train_,\n",
    "        validation=0.3,\n",
    "        epochs=50,\n",
    "        verbose=False,\n",
    "        batch_size=60,\n",
    "        shuffle=shuffle_train,\n",
    "        callbacks=[early_stop, plotlosses, terminate_nan]\n",
    "    )\n",
    "    if show_metrics:\n",
    "        report = evaluate_training(history, output_encry, ident_name)\n",
    "    timer.stop()\n",
    "elif loadModel:       \n",
    "    if not model.load_model():\n",
    "        print ('[Error] please change the config file : load_model')\n",
    "#session.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "accuracy: 0.9961110949516296, loss: 8.26694376883097e-05, val_acc: 0.9961110949516296, val_loss: 8.26694376883097e-05\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[Output] Metrics saved at %s results/MEnGHa5DCmJGE9C9BpAg4i/evaluation-loss_lstm-parallel_cylin.png\n",
      "Final cost: 0.0001\n"
     ]
    }
   ],
   "source": [
    "#history.history\n",
    "report = evaluate_training(history, output_encry, ident_name)\n",
    "print('Final cost: {0:.4f}'.format(history.history['loss'][-1]))\n",
    "# 0.002, 0.0011\n",
    "# gaussian 2.2359\n",
    "# lstm\n",
    "#  accuracy: 0.9964583516120911, loss: 0.00020314622088335454, val_acc: 0.9964583516120911, val_loss: 0.00020314622088335454"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[Output] All results saved at results/MEnGHa5DCmJGE9C9BpAg4i directory it results-train.txt file. Please use notebooks/plot_prediction.ipynb\n"
     ]
    }
   ],
   "source": [
    "\n",
    "# save results in a file    \n",
    "orig_stdout = sys.stdout\n",
    "f = open(os.path.join(output_encry, 'results-train.txt'), 'a')\n",
    "sys.stdout = f\n",
    "\n",
    "now = dt.datetime.now()\n",
    "print(\"[Output] Train results \")\n",
    "print(\"---Parameters--- \")\n",
    "print(\"\\t Model Name        : \", model.name)\n",
    "print(\"\\t Dataset           : \", model.orig_ds_name)\n",
    "print(\"\\t Total tracks      : \", len(X_train))\n",
    "print(\"\\t Path saved        : \", model.save_fnameh5) \n",
    "print(\"\\t Coordenate type   : \", coord) \n",
    "print(\"\\t Compiled date     : %s taken %s\" % (timer.start_dt.strftime(\"%d/%m/%Y %H:%M:%S\"), timer.taken()))    \n",
    "print(\"\\t Model scaled      : \", model.normalise)\n",
    "print(\"\\t Model scaled      : \", type_norm)\n",
    "print(\"\\t Model Optimizer   : \", optim)\n",
    "print(\"\\t Model batch_size  : \", batch)\n",
    "print(\"\\t Model epochs      : %s  stopped %s \" % (epochs, model.stopped_epoch))\n",
    "print(\"\\t Accuracy          : \", report)\n",
    "print(\"\\t Architecture      : \", arch)\n",
    "\n",
    "sys.stdout = orig_stdout\n",
    "f.close()    \n",
    "\n",
    "print('[Output] All results saved at %s directory it results-train.txt file. Please use notebooks/plot_prediction.ipynb' % output_encry)    \n"
   ]
  },
  {
   "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": "3.8.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}