Transfer.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "from Music_Style_Transfer_master.project.midi_handler import score2midi\n",
    "from Music_Style_Transfer_master.project.utils import get_representation, dataset_import\n",
    "from Music_Style_Transfer_master.project.model_transformer_relative import _2way_transformer_wavenet_absolute_pitch, _2way_transformer_wavenet_relative_pitch\n",
    "\n",
    "import tqdm\n",
    "import pypianoroll as pr\n",
    "import tensorflow as tf\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "# 33:97 A1~C7  21:109 A0~C8\n",
    "# 7 octave 21:105  A0~G#7  \n",
    "def midi2score(path, beat_resolution, pitch_range=np.arange(21, 109)):\n",
    "    tmp = pr.Multitrack(filename=path, beat_resolution=beat_resolution)   \n",
    "    tmp.binarize()\n",
    "    tmp.assign_constant(1)\n",
    "    # TODO: use variable to handle pitch range\n",
    "    score = tmp.get_merged_pianoroll()[:, pitch_range]\n",
    "    \n",
    "    return score\n",
    "\n",
    "def load_model(path_model):\n",
    "    model = _2way_transformer_wavenet_absolute_pitch(len_context = 8*4*beat_resolution,\n",
    "                                                 n_octave = 7,\n",
    "                                                 size_embedding = 84,\n",
    "                                                 n_transf_layers = 4,\n",
    "                                                 n_transf_downsample = 0,\n",
    "                                                 n_conv_layers = 7,\n",
    "                                                 context_layers = 1)\n",
    "    model.load_weights(path_model)\n",
    "    \n",
    "    return model\n",
    "\n",
    "def time_index_chosing(_range,\n",
    "                       interval,\n",
    "                       random=True,\n",
    "                       time_list=None):\n",
    "    # For parallel generation\n",
    "    \n",
    "    interval += 1\n",
    "    if (random):\n",
    "        time_index_base = np.random.randint(_range)\n",
    "    else:\n",
    "        time_index_base = time_list[np.random.randint(len(time_list))]\n",
    "\n",
    "    c = time_index_base + np.arange(-interval * 200, interval * 200, interval)\n",
    "\n",
    "    return c[np.where(np.logical_and(c >= 0, c < _range))]\n",
    "\n",
    "\n",
    "def generation(model, score, meta, time_indexes, midi_indexes,\n",
    "                       len_context\n",
    "                       ):\n",
    "    time_indexes = np.array(time_indexes) +  len_context\n",
    "    \n",
    "    left_features = (np.array(score[[np.arange(t -  len_context, t) for t in time_indexes], :]) > 0).astype(int)\n",
    "    left_metas = np.array(meta[[np.arange(t -  len_context, t) for t in time_indexes], :])\n",
    "    \n",
    "    right_features = (np.array(score[[np.arange(t + 1, t + 1 + len_context) for t in time_indexes], :]) > 0).astype(int)\n",
    "    right_metas = np.array(meta[[np.arange(t + 1, t + 1 + len_context) for t in time_indexes], :])\n",
    "    \n",
    "    central_features = (np.array(score[time_indexes, :]) > 0).astype(int)\n",
    "    central_metas = np.array(meta[time_indexes, :])\n",
    "    \n",
    "    for a, b in enumerate(midi_indexes):\n",
    "        central_features[a, b:] = 2\n",
    "        \n",
    "    central_features = np.reshape(central_features, (len(time_indexes), 1, -1))\n",
    "    central_metas = np.reshape(central_metas, (len(time_indexes), 1, -1))\n",
    "\n",
    "    p = model.predict([left_features,  left_metas, \n",
    "                       central_features, central_metas,\n",
    "                       right_features, right_metas])\n",
    "    \n",
    "    return p\n",
    "\n",
    "def style_transfer(score, meta, score_melody, model, len_context, pitch_range, \n",
    "                   iter_num=25, threshold=0.5\n",
    "                   ):\n",
    "\n",
    "    fixed_rhythm_score = score_melody\n",
    "    original_len = len(score_melody)\n",
    "    new_score = np.array(score)\n",
    "    counter = 0\n",
    "    alpha_initial = 0.6\n",
    "    alpha = alpha_initial\n",
    "    alpha_min = 0\n",
    "    annealing_fraction = 0.6\n",
    "    update_count = 0\n",
    "    \n",
    "    plt.figure(figsize=(20, 20))\n",
    "    plt.imshow(new_score[len_context:-len_context, :pitch_range].T)\n",
    "    plt.pause(1)\n",
    "    \n",
    "    for i in tqdm.tqdm(range(iter_num)):\n",
    "        time_list = np.arange(original_len)\n",
    "        if i < iter_num*0.9:\n",
    "            score[len_context:-len_context] = np.logical_or(score[len_context:-len_context], score_melody).astype(int)\n",
    "        print(\"alpha = \", alpha)\n",
    "        \n",
    "        while (time_list.size > 0):\n",
    "            if(alpha != -1):\n",
    "                alpha = max(0, alpha_initial - update_count * (alpha_initial - alpha_min) / (\n",
    "                    iter_num * original_len * annealing_fraction))\n",
    "            if(alpha == 0):\n",
    "                score = new_score\n",
    "                alpha = -1\n",
    "            elif(counter / original_len > alpha and alpha != -1):\n",
    "                counter = 0\n",
    "                score = np.array(new_score)\n",
    "\n",
    "            time_indexes = time_index_chosing(original_len, len_context, random=False, time_list=time_list)\n",
    "            l = len(time_indexes)\n",
    "            sorter = np.argsort(time_list)\n",
    "            d = sorter[np.searchsorted(time_list, time_indexes, sorter=sorter)]\n",
    "            time_list = np.delete(time_list, d, 0)\n",
    "            counter += l\n",
    "\n",
    "            update_count += l\n",
    "\n",
    "            if(alpha != -1):\n",
    "                midi_indexes = np.arange(pitch_range).tolist() * len(time_indexes)\n",
    "                time_indexes_repeat = np.repeat(time_indexes, pitch_range)\n",
    "                p = generation(model, score, meta, time_indexes_repeat, midi_indexes, len_context)\n",
    "                \n",
    "                # TODO: This for loop seems can be removed\n",
    "                for j, t in enumerate(time_indexes_repeat):\n",
    "                    update = p[0][j]\n",
    "                    if update < threshold :\n",
    "                        new_score[t + len_context, midi_indexes[j]] = 0\n",
    "                    else:\n",
    "                        new_score[t + len_context, midi_indexes[j]] = 1\n",
    "               \n",
    "            else:\n",
    "                for midi_index in range(pitch_range):\n",
    "                    \n",
    "                    midi_indexes = [midi_index] * l\n",
    "                    p = generation(model, score, meta, time_indexes, midi_indexes, len_context)\n",
    "                        \n",
    "                    for j, t in enumerate(time_indexes):\n",
    "                        update = p[0][j]\n",
    "                        if i > (0.8 *iter_num):\n",
    "                            if(fixed_rhythm_score[t, midi_indexes[j]] == 0):\n",
    "                                if update < threshold :\n",
    "                                    new_score[t + len_context, midi_indexes[j]] = 0\n",
    "                                else:\n",
    "                                    new_score[t + len_context, midi_indexes[j]] = 1\n",
    "                        else:   \n",
    "                            if update <threshold :\n",
    "                                new_score[t + len_context, midi_indexes[j]] = 0\n",
    "                            else:\n",
    "                                new_score[t + len_context, midi_indexes[j]] = 1\n",
    "        # For plotting\n",
    "        plt.figure(figsize=(20, 20))\n",
    "        plt.imshow(new_score[len_context:-len_context, :pitch_range].T)\n",
    "        plt.pause(1)\n",
    "    return new_score[len_context:-len_context, :pitch_range]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "  0%|          | 0/10 [00:00<?, ?it/s]\u001b[A"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "alpha =  0.6\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      " 10%|█         | 1/10 [02:36<23:26, 156.25s/it]\u001b[A"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "alpha =  0.5003124999999999\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      " 20%|██        | 2/10 [05:01<20:23, 152.89s/it]\u001b[A"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "alpha =  0.40031249999999996\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      " 30%|███       | 3/10 [07:27<17:35, 150.83s/it]\u001b[A"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "alpha =  0.3003125\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      " 40%|████      | 4/10 [09:53<14:56, 149.35s/it]\u001b[A"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "alpha =  0.2003125\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      " 50%|█████     | 5/10 [12:18<12:20, 148.03s/it]\u001b[A"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "alpha =  0.10031249999999997\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      " 60%|██████    | 6/10 [14:43<09:49, 147.32s/it]\u001b[A"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "alpha =  0.00031250000000004885\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      " 70%|███████   | 7/10 [28:09<17:14, 344.84s/it]\u001b[A"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "alpha =  -1\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABIEAAAC1CAYAAADIku9pAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAEhhJREFUeJzt3WusZWd5H/D/kxlfgoMx4yDkCw2uMCCEjEEjDCJC1NMEJ0HYHxAi6sVFlvyFttAkCm6+oESKFKQKQpSWysIkTkQCZCDFQoip7ThK8qEuM+BysblMXS5jG0yKTSiWjN0+/XCW7aNh8Nlzzj7r7L3W7yeNzl5rrz3rXXs96z17/rPed1d3BwAAAIBp+6m9bgAAAAAAu08IBAAAADADQiAAAACAGRACAQAAAMyAEAgAAABgBoRAAAAAADMgBAIAAACYgR2FQFV1VVV9paqOV9UNy2oUAAAAAMtV3b29F1btS/LVJL+Q5ESSzyT51e6+e3nNAwAAAGAZ9u/gta9Mcry7702SqvpwkquT/MQQ6Mw6q8/OOTvYJQAAAACb/SAP/X13P2er7XYSAl2U5Fublk8kueLkjarq+iTXJ8nZeUauqEM72CUAAAAAm93Wh7+xyHa7PjF0d9/Y3Qe7++AZOWu3dwcAAADAKewkBLovyfM2LV88rAMAAABgxewkBPpMkkur6pKqOjPJW5LcspxmAQAAALBM254TqLsfr6p/neRIkn1JPtjdX1paywAAAABYmp1MDJ3u/lSSTy2pLQAAAADskl2fGBoAAACAvScEAgAAAJgBIRAAAADADAiBAAAAAGZACAQAAAAwA0IgAAAAgBkQAgEAAADMgBAIAAAAYAaEQAAAAAAzIAQCAAAAmAEhEAAAAMAMCIEAAAAAZkAIBAAAADADQiAAAACAGRACAQAAAMyAEAgAAABgBoRAAAAAADMgBAIAAACYASEQAAAAwAwIgQAAAABmQAgEAAAAMANCIAAAAIAZEAIBAAAAzMD+MXf2wsseyZEjd425y1G9/sLL97oJu+rI/c4di5lyreyFqdfnlOtl6udubFOulb0w9fqccr1M/dyNbcq1shemXp9Trpepn7uxjV0r+y5YbLvq7t1tySbn1oG+og6Ntj8AAACAqbutDx/r7oNbbbflcLCqel5V3VFVd1fVl6rq7cP6A1V1a1V9bfj57GU0HAAAAIDlW2Q42ONJfr27P1tVz0xyrKpuTfKvktze3b9XVTckuSHJO3evqatvyrcGJtO+PXDsczf2ezn145s6fcv6mvq1N/Xjmzp9y/qa+rU39eObOn3L+pr6tTf141vUlncCdfcD3f3Z4fEPktyT5KIkVye5edjs5iTX7FYjAQAAANiZ05oTqKqen+Rvkrw0yTe7+7xhfSV56Inln8ScQAAAAADLtbQ5gZ5QVT+T5GNJ3tHd/7D5ud5Ikk6ZJlXV9VV1tKqOPpZHF90dAAAAAEu0UAhUVWdkIwD6UHd/fFj9naq6YHj+giQPnuq13X1jdx/s7oNn5KxltBkAAACA07TIt4NVkpuS3NPd79n01C1Jrh0eX5vkE8tvHgAAAADLsMi3g70myb9I8oWqemI67d9K8ntJPlpV1yX5RpI3704TAQAAANipLUOg7v67JPUTnjbLMwAAAMAaWHhiaAAAAADW1yLDwZbmhZc9kiNH7tp6wzX1+gsv3+sm7Koj9zt3LGbKtbIXpl6fU66XqZ+7sU25VvbC1OtzyvUy9XM3tinXyl6Yen1OuV6mfu7GNnat7Ltgse3cCQQAAAAwA9Xdo+3s3DrQV5RphAAAAACW5bY+fKy7D2613ajDwaZuyrcGJtO+PXDq525sU66VvTD1+pxyvUz93I1tyrWyF6Zen1Oul6mfu7FNuVb2wtTrc8r1MvVzN7ZVrRXDwQAAAABmwHAwAAAAgDW26HAwdwIBAAAAzIAQCAAAAGAGhEAAAAAAMyAEAgAAAJgBIRAAAADADAiBAAAAAGZACAQAAAAwA0IgAAAAgBkQAgEAAADMgBAIAAAAYAb2j7mzF172SI4cuWvMXY7q9RdevtdN2FVH7nfulmXs93Lqxzd1+pb1NfVrb+rHN3X6lvU19Wtv6sc3dfqW9TX1a2/qx7fvgsW2GzUE+urnnzHpTmHKHUIy7Q5dB7RcU66VvaBvWV9Tv/amfnxTp29ZX1O/9qZ+fFOnb1lfU7/2pn58yfGFtjIcDAAAAGAGqrtH29m5daCvqEOj7Q8AAABg6m7rw8e6++BW27kTCAAAAGAGhEAAAAAAMyAEAgAAAJgBIRAAAADADAiBAAAAAGZg4RCoqvZV1eeq6pPD8iVVdWdVHa+qj1TVmbvXTAAAAAB24nTuBHp7kns2Lb87yXu7+wVJHkpy3TIbBgAAAMDyLBQCVdXFSX4lyQeG5UpyZZLDwyY3J7lmNxoIAAAAwM7tX3C730/ym0meOSyfn+Th7n58WD6R5KIltw0AAABgko7cf9ePrXv9hZfv6j63vBOoqt6Q5MHuPradHVTV9VV1tKqOPpZHt/NXAAAAALBDi9wJ9Jokb6yqX05ydpJzk7wvyXlVtX+4G+jiJPed6sXdfWOSG5Pk3DrQ22nkqdKxRWw3QRt7fyzPupy77bZzuxzfqelbWNS6nLt1ufa2a12OT9/Cotbl3K3Ltbdd63J8+hYWtS7nbl2uve1a5PhO1abtvi/7Llhsu+pePJepqtcl+Y3ufkNV/UWSj3X3h6vqPyf5fHf/p6d7/bl1oK+oQwvvDwAAAICnd1sfPtbdB7fa7nS+Hexk70zya1V1PBtzBN20g78LAAAAgF206MTQSZLu/uskfz08vjfJK5ffpPW1LrfdrYOney934/1yK+Jyqenl0rcsj75luaZ+fFOnb1kefctyjVGbm/ehppdL37I8+pblmvrxLWondwIBAAAAsCZOa06gnTInEAAAAMByjTEnEAAAAABr4rTmBFoFT4zjW9XxdVM25hjKpzu/Y4+N3Q3Go8JTVr1vWafrR98CT1n1vmWr160SfQs8Rd+yPPqWvWE42ApQ/MBu0LcAu0Hfwqo6ndpUV6tH3wI7YzgYAAAAAE9yJxAAAADAGnMnEAAAAABPGnVi6Bde9kiOHDn9sZ7bHa+53XGlUx8fuozxtqfzHh25/67Jv6fsrbGv9dPd31zqf5lj+Rd9z07e51zea8Yx9rXuc8upjf25BcayaG2P1bfM7TrxbyLmyp1AAAAAADNgTiAAAACANWZOIAAAAACeJAQCAAAAmAEhEAAAAMAMCIEAAAAAZkAIBAAAADADQiAAAACAGRACAQAAAMyAEAgAAABgBoRAAAAAADMgBAIAAACYASEQAAAAwAwIgQAAAABmQAgEAAAAMANCIAAAAIAZEAIBAAAAzMBCIVBVnVdVh6vqy1V1T1W9uqoOVNWtVfW14eezd7uxAAAAAGzPoncCvS/Jp7v7xUleluSeJDckub27L01y+7AMAAAAwAraMgSqqmcleW2Sm5Kku3/U3Q8nuTrJzcNmNye5ZrcaCQAAAMDOLHIn0CVJvpvkj6rqc1X1gao6J8lzu/uBYZtvJ3nubjUSAAAAgJ1ZJATan+QVSd7f3S9P8sOcNPSruztJn+rFVXV9VR2tqqOP5dGdthcAAACAbVgkBDqR5ER33zksH85GKPSdqrogSYafD57qxd19Y3cf7O6DZ+SsZbQZAAAAgNO0ZQjU3d9O8q2qetGw6lCSu5PckuTaYd21ST6xKy0EAAAAYMf2L7jdv0nyoao6M8m9Sd6ajQDpo1V1XZJvJHnz7jQRAAAAgJ1aKATq7ruSHDzFU4eW2xwAAAAAdsMicwIBAAAAsOaEQAAAAAAzIAQCAAAAmAEhEAAAAMAMCIEAAAAAZkAIBAAAADADQiAAAACAGRACAQAAAMyAEAgAAABgBoRAAAAAADMgBAIAAACYASEQAAAAwAwIgQAAAABmQAgEAAAAMANCIAAAAIAZEAIBAAAAzIAQCAAAAGAGhEAAAAAAMyAEAgAAAJgBIRAAAADADAiBAAAAAGZACAQAAAAwA0IgAAAAgBkQAgEAAADMgBAIAAAAYAaEQAAAAAAzsFAIVFX/rqq+VFVfrKo/r6qzq+qSqrqzqo5X1Ueq6szdbiwAAAAA27NlCFRVFyX5t0kOdvdLk+xL8pYk707y3u5+QZKHkly3mw0FAAAAYPsWHQ62P8lPV9X+JM9I8kCSK5McHp6/Ock1y28eAAAAAMuwZQjU3fcl+Q9JvpmN8Of7SY4lebi7Hx82O5Hkot1qJAAAAAA7s8hwsGcnuTrJJUkuTHJOkqsW3UFVXV9VR6vq6GN5dNsNBQAAAGD7FhkO9k+T/K/u/m53P5bk40lek+S8YXhYklyc5L5Tvbi7b+zug9198IyctZRGAwAAAHB6FgmBvpnkVVX1jKqqJIeS3J3kjiRvGra5NskndqeJAAAAAOzUInMC3ZmNCaA/m+QLw2tuTPLOJL9WVceTnJ/kpl1sJwAAAAA7sH/rTZLufleSd520+t4kr1x6iwAAAABYukW/Ih4AAACANVbdPd7Oqr6b5IdJ/n60ncJy/GzULetH3bKu1C7rSN2yrtQu60jd/rif6+7nbLXRqCFQklTV0e4+OOpOYYfULetI3bKu1C7rSN2yrtQu60jdbp/hYAAAAAAzIAQCAAAAmIG9CIFu3IN9wk6pW9aRumVdqV3WkbplXald1pG63abR5wQCAAAAYHyGgwEAAADMwGghUFVdVVVfqarjVXXDWPuFRVTVB6vqwar64qZ1B6rq1qr62vDz2cP6qqo/GGr581X1ir1rOXNWVc+rqjuq6u6q+lJVvX1Yr3ZZWVV1dlX996r6H0Pd/vaw/pKqunOoz49U1ZnD+rOG5ePD88/fy/Yzb1W1r6o+V1WfHJbVLSuvqr5eVV+oqruq6uiwzmcFVlpVnVdVh6vqy1V1T1W9Wt0uxyghUFXtS/Ifk/xSkpck+dWqeskY+4YF/XGSq05ad0OS27v70iS3D8vJRh1fOvy5Psn7R2ojnOzxJL/e3S9J8qokbxv6VrXLKns0yZXd/bIklye5qqpeleTdSd7b3S9I8lCS64btr0vy0LD+vcN2sFfenuSeTcvqlnXxT7r78k1fqe2zAqvufUk+3d0vTvKybPS96nYJxroT6JVJjnf3vd39oyQfTnL1SPuGLXX33yT53kmrr05y8/D45iTXbFr/J73hvyU5r6ouGKel8JTufqC7Pzs8/kE2fjleFLXLChvq7/8Mi2cMfzrJlUkOD+tPrtsn6vlwkkNVVSM1F55UVRcn+ZUkHxiWK+qW9eWzAiurqp6V5LVJbkqS7v5Rdz8cdbsUY4VAFyX51qblE8M6WGXP7e4HhsffTvLc4bF6ZuUMQw1enuTOqF1W3DCk5q4kDya5Ncn/TPJwdz8+bLK5Np+s2+H57yc5f9wWQ5Lk95P8ZpL/NyyfH3XLeugk/7WqjlXV9cM6nxVYZZck+W6SPxqG4H6gqs6Jul0KE0PDAnrja/R8lR4rqap+JsnHkryju/9h83Nql1XU3f+3uy9PcnE27hZ+8R43CZ5WVb0hyYPdfWyv2wLb8PPd/YpsDJl5W1W9dvOTPiuwgvYneUWS93f3y5P8ME8N/UqibndirBDoviTP27R88bAOVtl3nriNcPj54LBePbMyquqMbARAH+rujw+r1S5rYbi1+44kr87Grdv7h6c21+aTdTs8/6wk/3vkpsJrkryxqr6ejWkNrszGfBXqlpXX3fcNPx9M8pfZCN99VmCVnUhyorvvHJYPZyMUUrdLMFYI9Jkklw7foHBmkrckuWWkfcN23ZLk2uHxtUk+sWn9vxxmoX9Vku9vui0RRjPML3FTknu6+z2bnlK7rKyqek5VnTc8/ukkv5CN+azuSPKmYbOT6/aJen5Tkr8a/vcPRtPd/767L+7u52fjc+xfdfc/i7plxVXVOVX1zCceJ/nFJF+MzwqssO7+dpJvVdWLhlWHktwddbsUNdbvo6r65WyMpd6X5IPd/buj7BgWUFV/nuR1SX42yXeSvCvJf0ny0ST/KMk3kry5u783/MP7D7PxbWKPJHlrdx/di3Yzb1X180n+NskX8tQcFb+VjXmB1C4rqaouy8Zkjvuy8Z9RH+3u36mqf5yNOywOJPlckn/e3Y9W1dlJ/jQbc159L8lbuvvevWk9JFX1uiS/0d1vULesuqFG/3JY3J/kz7r7d6vq/PiswAqrqsuzMRH/mUnuTfLWDJ8bom53ZLQQCAAAAIC9Y2JoAAAAgBkQAgEAAADMgBAIAAAAYAaEQAAAAAAzIAQCAAAAmAEhEAAAAMAMCIEAAAAAZkAIBAAAADAD/x8S/PtM4oVSBgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 1440x1440 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      " 80%|████████  | 8/10 [41:29<16:02, 481.48s/it]\u001b[A"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "alpha =  -1\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      " 90%|█████████ | 9/10 [54:53<09:38, 578.06s/it]\u001b[A"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "alpha =  -1\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "100%|██████████| 10/10 [1:08:12<00:00, 644.34s/it]\u001b[A"
     ]
    }
   ],
   "source": [
    "beat_resolution = 8\n",
    "pitch_range = 84\n",
    "len_context = beat_resolution*4*8\n",
    "max_len = beat_resolution*4*4\n",
    "\n",
    "path_model = \"../../../scratch/wtl272/projects/Music_style_transfer/runs/run_61/model/model.hdf5\"\n",
    "\n",
    "# load model\n",
    "model = load_model(path_model)\n",
    "\n",
    "path = \"input_midi/0716/train_harmonized.mid\"\n",
    "path_melody = \"input_midi/0716/train.mid\"\n",
    "\n",
    "score = midi2score(path, beat_resolution, pitch_range=np.arange(21, 105))\n",
    "\n",
    "if path_melody is None:\n",
    "    score_melody = np.zeros(score.shape)\n",
    "else:\n",
    "    score_melody = midi2score(path_melody, beat_resolution, pitch_range=np.arange(21, 105))\n",
    " \n",
    "score = np.array(score[:640])\n",
    "score_melody = np.array(score_melody[:640])\n",
    "score, score_meta = get_representation(score, pitch_range, beat_resolution, len_context)\n",
    "\n",
    "#generation\n",
    "result = style_transfer((np.array(score[:, :, 2]) > 0).astype(int), score_meta, \n",
    "                         score_melody, model, len_context, pitch_range, iter_num=15)\n",
    "\n",
    "#save result\n",
    "score2midi(\"output_midi/train_bah_model61.mid\", result[:len(score_melody)], beat_resolution, \n",
    "           120, pitch_range=np.arange(21, 105),melody_constraint=True, melody=score_melody)"
   ]
  }
 ],
 "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": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}