{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Training\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Building and compiling a model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%capture\n",
    "import h5py\n",
    "from chimeranet.models import ChimeraPPModel\n",
    "\n",
    "# probe shape of dataset and set embedding dimension\n",
    "dataset_path = 'example-dataset-train.hdf5'\n",
    "with h5py.File(dataset_path, 'r') as f:\n",
    "    _, T, F, C = f['y/embedding'].shape\n",
    "D = 20\n",
    "cm = ChimeraPPModel(T, F, C, D)\n",
    "\n",
    "# build_model returns Keras' Model object\n",
    "model = cm.build_model()\n",
    "model.compile(\n",
    "    'rmsprop',\n",
    "    loss={\n",
    "        'embedding': cm.loss_deepclustering(),\n",
    "        'mask': cm.loss_mask()\n",
    "    },\n",
    "    loss_weights={\n",
    "        'embedding': 0.9,\n",
    "        'mask': 0.1\n",
    "    }\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Training a model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 160 samples, validate on 160 samples\n",
      "Epoch 1/10\n",
      "160/160 [==============================] - 261s 2s/step - loss: 18381.2736 - embedding_loss: 18457.8914 - mask_loss: 17691.7186 - val_loss: 12940.3148 - val_embedding_loss: 12466.7975 - val_mask_loss: 17201.9725\n",
      "Epoch 2/10\n",
      "160/160 [==============================] - 258s 2s/step - loss: 12468.0131 - embedding_loss: 12140.8994 - mask_loss: 15412.0373 - val_loss: 11921.1551 - val_embedding_loss: 11286.8416 - val_mask_loss: 17629.9813\n",
      "Epoch 3/10\n",
      "160/160 [==============================] - 266s 2s/step - loss: 12008.5932 - embedding_loss: 11784.1303 - mask_loss: 14028.7604 - val_loss: 11863.6486 - val_embedding_loss: 11170.0996 - val_mask_loss: 18105.5883\n",
      "Epoch 4/10\n",
      "160/160 [==============================] - 266s 2s/step - loss: 12022.4951 - embedding_loss: 11823.9797 - mask_loss: 13809.1371 - val_loss: 11875.8588 - val_embedding_loss: 11263.6596 - val_mask_loss: 17385.6521\n",
      "Epoch 5/10\n",
      "160/160 [==============================] - 266s 2s/step - loss: 11840.2252 - embedding_loss: 11682.9516 - mask_loss: 13255.6879 - val_loss: 11862.4658 - val_embedding_loss: 11327.9975 - val_mask_loss: 16672.6859\n",
      "Epoch 6/10\n",
      "160/160 [==============================] - 273s 2s/step - loss: 11768.4652 - embedding_loss: 11728.4332 - mask_loss: 12128.7551 - val_loss: 11538.6459 - val_embedding_loss: 11066.1010 - val_mask_loss: 15791.5525\n",
      "Epoch 7/10\n",
      "160/160 [==============================] - 270s 2s/step - loss: 11755.5100 - embedding_loss: 11649.7219 - mask_loss: 12707.6035 - val_loss: 11632.5807 - val_embedding_loss: 11140.6088 - val_mask_loss: 16060.3297\n",
      "Epoch 8/10\n",
      "160/160 [==============================] - 276s 2s/step - loss: 11816.1187 - embedding_loss: 11642.1570 - mask_loss: 13381.7742 - val_loss: 11814.3582 - val_embedding_loss: 11381.6398 - val_mask_loss: 15708.8264\n",
      "Epoch 9/10\n",
      "160/160 [==============================] - 270s 2s/step - loss: 11801.6576 - embedding_loss: 11660.4234 - mask_loss: 13072.7674 - val_loss: 11853.9027 - val_embedding_loss: 11215.5012 - val_mask_loss: 17599.5209\n",
      "Epoch 10/10\n",
      "160/160 [==============================] - 283s 2s/step - loss: 11742.7662 - embedding_loss: 11698.4328 - mask_loss: 12141.7742 - val_loss: 11385.7367 - val_embedding_loss: 11052.1635 - val_mask_loss: 14387.8969\n"
     ]
    }
   ],
   "source": [
    "# load train and validation data\n",
    "dataset_validation_path = 'example-dataset-validation.hdf5'\n",
    "with h5py.File(dataset_path, 'r') as f:\n",
    "    x_train = f['x'][()]\n",
    "    y_train = {'mask': f['y/mask'][()], 'embedding': f['y/embedding'][()]}\n",
    "with h5py.File(dataset_validation_path, 'r') as f:\n",
    "    x_validation = f['x'][()]\n",
    "    y_validation = {'mask': f['y/mask'][()], 'embedding': f['y/embedding'][()]}\n",
    "\n",
    "# train model by model.fit function\n",
    "model.fit(\n",
    "    x=x_train,\n",
    "    y=y_train,\n",
    "    validation_data=(x_validation, y_validation),\n",
    "    batch_size=32,\n",
    "    epochs=10\n",
    ")\n",
    "# save the model\n",
    "model_path = 'example-model.hdf5'\n",
    "model.save(model_path)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}