three-layer-e2e.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os,sys\n",
    "import cPickle as pickle\n",
    "import numpy as np                                       # fast vectors and matrices\n",
    "import matplotlib.pyplot as plt                          # plotting\n",
    "from scipy.fftpack import fft\n",
    "\n",
    "from time import time\n",
    "\n",
    "sys.path.insert(0,'lib/')\n",
    "import config\n",
    "import diagnostics\n",
    "import base_model\n",
    "\n",
    "from sklearn.metrics import average_precision_score\n",
    "\n",
    "import tensorflow as tf\n",
    "\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(config.labels_path) as f:\n",
    "    labels = pickle.load(f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "class Spectrograms(base_model.Model):\n",
    "    def __init__(self, *args, **kwargs):\n",
    "        super(Spectrograms, self).__init__(*args, **kwargs)\n",
    "\n",
    "    def define_graph(self):\n",
    "        super(Spectrograms, self).define_graph()\n",
    "        \n",
    "        # lvl1 convolutions are shared between regions\n",
    "        self.k = 512              # lvl1 nodes\n",
    "        self.d = 4096              # lvl1 receptive field\n",
    "        \n",
    "        self.k2 = 512              # lvl2 nodes\n",
    "\n",
    "        # number of lvl1 features\n",
    "        regions = 1 + (self.window - self.d)/self.stride\n",
    "        print 'Number of V1 feature regions: {}'.format(regions)\n",
    "\n",
    "        wscale = 10e-3\n",
    "        with tf.variable_scope('parameters'):\n",
    "            w = tf.Variable(wscale*tf.random_normal([1,self.d,1,self.k],seed=999))\n",
    "            wavg = self.register_weights(w,'w',average=.9998)\n",
    "            w1 = tf.Variable(wscale*tf.random_normal([regions*self.k,self.k2],seed=999))\n",
    "            w1avg = self.register_weights(w1,'w1',average=.9998)\n",
    "            beta = tf.Variable(wscale*tf.random_normal([self.k2,self.m],seed=999))\n",
    "            betaavg = self.register_weights(beta,'beta',average=.9998)\n",
    "\n",
    "        with tf.variable_scope('queued_model'):\n",
    "            zx = tf.log(tf.constant(1.) + tf.nn.relu(tf.nn.conv2d(self.xq,w,strides=[1,1,self.stride,1],padding='VALID')))\n",
    "            z2 = tf.nn.relu(tf.matmul(tf.reshape(zx,[self.batch_size,regions*self.k]),w1))\n",
    "            y = tf.matmul(z2,beta)\n",
    "            self.loss = tf.reduce_mean(tf.nn.l2_loss(y-tf.reshape(self.yq,[self.batch_size,self.m])))\n",
    "\n",
    "        with tf.variable_scope('direct_model'):\n",
    "            zx = tf.log(tf.constant(1.) + tf.nn.relu(tf.nn.conv2d(self.xd,wavg,strides=[1,1,self.stride,1],padding='VALID')))\n",
    "            z2 = tf.nn.relu(tf.matmul(tf.reshape(zx,[tf.shape(self.xd)[0],regions*self.k]),w1avg))\n",
    "            self.y_direct = tf.matmul(z2,betaavg)\n",
    "            self.loss_direct = tf.reduce_mean(tf.nn.l2_loss(self.y_direct-self.yd))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of V1 feature regions: 25\n"
     ]
    }
   ],
   "source": [
    "model = Spectrograms(labels,checkpoint_path='convnet_mlp_raw/', outputs=1, window=16384, mmap=True, batch_size=100,\n",
    "                    normalize=True, extended_test_set=True, use_mirex=False, init=False, pitch_transforms=5, jitter=.1,\n",
    "                    restrict=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of V1 feature regions: 25\n",
      "avp_test\tavp_train\tmse_test\tmse_train\ttime\tutime\n",
      "0.715536        0.721713        0.527816        0.669378        18      18      \n",
      "0.715707        0.721725        0.527762        0.669348        45      17      \n",
      "0.715556        0.721724        0.527837        0.669337        47      19      \n",
      "0.715377        0.721738        0.527920        0.669287        46      19      \n",
      "0.715295        0.721811        0.527912        0.669211        45      21      \n",
      "0.715328        0.721836        0.527821        0.669184        44      17      \n",
      "0.715432        0.721808        0.527687        0.669228        46      18      \n",
      "0.715463        0.721816        0.527634        0.669235        45      19      \n",
      "0.715478        0.721805        0.527580        0.669221        44      19      \n",
      "0.715580        0.721796        0.527487        0.669197        45      17      \n",
      "0.715634        0.721805        0.527526        0.669160        46      18      \n",
      "0.715535        0.721828        0.527615        0.669112        45      18      \n",
      "0.715486        0.721823        0.527639        0.669085        44      18      \n",
      "0.715574        0.721848        0.527580        0.669054        42      13      \n",
      "0.715630        0.721818        0.527522        0.669052        45      17      \n",
      "0.715522        0.721754        0.527534        0.669097        46      18      \n",
      "0.715362        0.721768        0.527591        0.669112        45      18      \n",
      "0.715428        0.721806        0.527540        0.669084        47      18      \n",
      "0.715487        0.721848        0.527439        0.669032        48      19      \n",
      "0.715403        0.721892        0.527455        0.668993        48      20      \n",
      "0.715460        0.721927        0.527380        0.668950        44      17      \n",
      "0.715620        0.721992        0.527338        0.668896        45      16      \n",
      "0.715702        0.722014        0.527323        0.668880        46      16      \n",
      "0.715825        0.722077        0.527234        0.668840        47      17      \n",
      "0.715830        0.722094        0.527242        0.668854        45      17      \n",
      "0.715907        0.722148        0.527169        0.668813        38      11      \n",
      "0.715945        0.722193        0.527091        0.668781        38      10      \n",
      "0.715753        0.722187        0.527183        0.668786        46      19      \n",
      "0.715738        0.722202        0.527199        0.668746        45      19      \n",
      "0.715829        0.722205        0.527138        0.668746        44      18      \n",
      "0.715849        0.722228        0.527115        0.668718        45      17      \n",
      "0.715763        0.722234        0.527195        0.668723        46      19      \n",
      "0.715783        0.722235        0.527224        0.668726        45      20      \n",
      "0.715740        0.722225        0.527225        0.668699        43      19      \n",
      "0.715718        0.722190        0.527254        0.668721        44      17      \n",
      "0.715675        0.722191        0.527245        0.668705        46      19      \n",
      "0.715580        0.722150        0.527353        0.668741        45      19      \n",
      "0.715563        0.722076        0.527403        0.668807        37      12      \n",
      "0.715405        0.722067        0.527507        0.668821        39      11      \n",
      "0.715219        0.722076        0.527664        0.668828        45      18      \n",
      "0.715067        0.722038        0.527732        0.668862        42      16      \n",
      "0.715019        0.722026        0.527747        0.668882        44      17      \n",
      "0.714949        0.721987        0.527716        0.668909        40      12      \n",
      "0.714960        0.721961        0.527780        0.668955        46      18      \n",
      "0.714929        0.721969        0.527829        0.668961        44      17      \n",
      "0.714962        0.721932        0.527808        0.669013        44      17      \n",
      "0.714952        0.721995        0.527879        0.668968        44      15      \n",
      "0.714893        0.721979        0.527947        0.668992        46      17      \n",
      "0.714970        0.721995        0.527907        0.668965        45      17      \n",
      "0.715010        0.722012        0.527899        0.668948        44      17      \n",
      "0.714899        0.721979        0.527968        0.668983        45      16      \n",
      "0.714895        0.721979        0.527957        0.668973        72      43      \n",
      "0.714987        0.721979        0.527858        0.668944        122     99      \n",
      "0.715126        0.721989        0.527842        0.668928        50      24      \n",
      "0.715148        0.721997        0.527817        0.668939        45      21      \n",
      "0.715217        0.722004        0.527794        0.668987        46      20      \n",
      "0.715238        0.722005        0.527787        0.669006        47      20      \n",
      "0.715177        0.722009        0.527840        0.669007        48      22      \n",
      "0.715138        0.722034        0.527888        0.668993        45      22      \n",
      "0.715296        0.722002        0.527811        0.669012        45      21      \n",
      "0.715390        0.722035        0.527784        0.669019        46      20      \n",
      "0.715564        0.722078        0.527614        0.669017        48      21      \n",
      "0.715582        0.722103        0.527529        0.669018        46      21      \n",
      "0.715604        0.722081        0.527442        0.669012        46      21      \n",
      "0.715702        0.722013        0.527317        0.669042        47      19      \n",
      "0.715824        0.721962        0.527196        0.669103        48      20      \n",
      "0.715799        0.721953        0.527222        0.669131        47      20      \n",
      "0.715777        0.721980        0.527251        0.669098        46      20      \n",
      "0.715822        0.722008        0.527290        0.669080        119     91      \n",
      "0.715738        0.722023        0.527357        0.669067        125     96      \n",
      "0.715672        0.722025        0.527393        0.669079        121     92      \n",
      "0.715652        0.722051        0.527378        0.669048        126     99      \n",
      "Graceful Exit\n"
     ]
    }
   ],
   "source": [
    "lr = .001/3/3/3\n",
    "mom = .95\n",
    "\n",
    "# stop the model if it is already running\n",
    "model.stop()\n",
    "\n",
    "# we have to rebuild the graph every time because input queues can't be reopened\n",
    "model.define_graph()\n",
    "\n",
    "learning_rate = tf.placeholder(tf.float32, shape=[])\n",
    "opt_op = tf.train.MomentumOptimizer(learning_rate,mom).minimize(model.loss)\n",
    "with tf.control_dependencies([opt_op]):\n",
    "    train_step = tf.group(*model.averages)\n",
    "\n",
    "# start up the session, kick off the worker threads, restore checkpoint, etc.\n",
    "model.start()\n",
    "\n",
    "try:\n",
    "    ptime = time()\n",
    "    print model.status_header()\n",
    "    while True:\n",
    "        if model.iter % 1000 == 0:\n",
    "            model.update_status(ptime,time(),lr)\n",
    "            model.checkpoint()\n",
    "            print model.status()\n",
    "            ptime = time()\n",
    "\n",
    "        model.sess.run(train_step, feed_dict={learning_rate: lr})\n",
    "        model.iter += 1\n",
    "\n",
    "except KeyboardInterrupt:\n",
    "    model.checkpoint()\n",
    "    print 'Graceful Exit'\n",
    "finally:\n",
    "    model.stop()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.715944618797\n",
      "0.527091033809\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAuEAAAJOCAYAAAAZCtmpAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nBuEiIiIiImX2f3hxBENhdwLTAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f382b01a810>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "burnin=100\n",
    "fig, ((ax1, ax2),(ax3,ax4),(ax5,ax6)) = plt.subplots(3,2)\n",
    "fig.set_figwidth(12)\n",
    "fig.set_figheight(10)\n",
    "ax1.set_title('average precision')\n",
    "ax1.plot(model.stats['iter'][2][burnin:],model.stats['avp_test'][2][burnin:],color='g')\n",
    "ax1.plot(model.stats['iter'][2][burnin:],model.stats['avp_train'][2][burnin:],color='b')\n",
    "ax2.set_title('square loss')\n",
    "ax2.plot(model.stats['iter'][2][burnin:],model.stats['mse_test'][2][burnin:],color='g')\n",
    "ax2.plot(model.stats['iter'][2][burnin:],model.stats['mse_train'][2][burnin:],color='b')\n",
    "ax3.set_title('weights top')\n",
    "ax3.plot(model.stats['iter'][2][burnin:],model.stats['nbeta'][2][burnin:],color='g')\n",
    "ax4.set_title('learning rate')\n",
    "ax4.plot(model.stats['iter'][2][burnin:],model.stats['lr'][2][burnin:],color='g')\n",
    "ax5.set_title('weights bottom')\n",
    "ax5.plot(model.stats['iter'][2][burnin:],model.stats['nw'][2][burnin:],color='g')\n",
    "ax6.set_title('weights middle')\n",
    "ax6.plot(model.stats['iter'][2][burnin:],model.stats['nw1'][2][burnin:],color='g')\n",
    "\n",
    "print max(model.stats['avp_test'][2])\n",
    "print min(model.stats['mse_test'][2])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "# Mirex Stats"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.707571258428\n"
     ]
    }
   ],
   "source": [
    "mse_test, Yhat, Y, mse_breakdown, avp_breakdown = model.sample_records(config.test_ids, 7500, fixed_stride=512)\n",
    "avp_test = average_precision_score(Y.flatten(),Yhat.flatten())\n",
    "print avp_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AvgP\tP\tR\tAcc\tETot\tESub\tEmiss\tEfa\n",
      "78.04\t72.10\t73.73\t0.57\t0.44\t0.11\t0.16\t0.18\n",
      "AvgP\tP\tR\tAcc\tETot\tESub\tEmiss\tEfa\n",
      "47.48\t53.70\t47.29\t0.34\t0.76\t0.17\t0.36\t0.24\n",
      "AvgP\tP\tR\tAcc\tETot\tESub\tEmiss\tEfa\n",
      "76.77\t73.12\t69.65\t0.55\t0.47\t0.09\t0.21\t0.16\n",
      "0.488029969265 0.557837001257\n"
     ]
    }
   ],
   "source": [
    "Accavg = Etotavg = 0\n",
    "for i in range(3):\n",
    "    _,_,_,Acc,Etot = diagnostics.mirex_statistics(model,i,threshold=.4)\n",
    "    Accavg += Acc\n",
    "    Etotavg += Etot\n",
    "\n",
    "print Accavg/3.,Etotavg/3."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "# Extended test set stats"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.729065910816\n"
     ]
    }
   ],
   "source": [
    "mse_test, Yhat, Y, mse_breakdown, avp_breakdown = model.sample_records(config.test_ids_ext, 7500, fixed_stride=512)\n",
    "avp_test = average_precision_score(Y.flatten(),Yhat.flatten())\n",
    "print avp_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "AvgP\tP\tR\tAcc\tETot\tESub\tEmiss\tEfa\n",
      "78.04\t72.10\t73.73\t0.57\t0.44\t0.11\t0.16\t0.18\n",
      "AvgP\tP\tR\tAcc\tETot\tESub\tEmiss\tEfa\n",
      "47.48\t53.70\t47.29\t0.34\t0.76\t0.17\t0.36\t0.24\n",
      "AvgP\tP\tR\tAcc\tETot\tESub\tEmiss\tEfa\n",
      "76.77\t73.12\t69.65\t0.55\t0.47\t0.09\t0.21\t0.16\n",
      "AvgP\tP\tR\tAcc\tETot\tESub\tEmiss\tEfa\n",
      "75.06\t70.21\t72.77\t0.56\t0.45\t0.13\t0.14\t0.18\n",
      "AvgP\tP\tR\tAcc\tETot\tESub\tEmiss\tEfa\n",
      "81.43\t71.60\t83.32\t0.63\t0.38\t0.12\t0.05\t0.21\n",
      "AvgP\tP\tR\tAcc\tETot\tESub\tEmiss\tEfa\n",
      "78.38\t73.72\t75.38\t0.59\t0.40\t0.12\t0.13\t0.15\n",
      "AvgP\tP\tR\tAcc\tETot\tESub\tEmiss\tEfa\n",
      "66.69\t59.34\t72.86\t0.49\t0.68\t0.09\t0.18\t0.41\n",
      "AvgP\tP\tR\tAcc\tETot\tESub\tEmiss\tEfa\n",
      "70.80\t68.43\t66.73\t0.51\t0.51\t0.13\t0.20\t0.18\n",
      "AvgP\tP\tR\tAcc\tETot\tESub\tEmiss\tEfa\n",
      "73.12\t69.84\t75.57\t0.57\t0.45\t0.12\t0.13\t0.21\n",
      "AvgP\tP\tR\tAcc\tETot\tESub\tEmiss\tEfa\n",
      "68.97\t69.70\t62.34\t0.49\t0.49\t0.16\t0.22\t0.11\n",
      "0.529693278552 0.503248535322\n"
     ]
    }
   ],
   "source": [
    "Accavg = Etotavg = 0\n",
    "for i in range(10):\n",
    "    _,_,_,Acc,Etot = diagnostics.mirex_statistics(model,i,threshold=.4)\n",
    "    Accavg += Acc\n",
    "    Etotavg += Etot\n",
    "\n",
    "print Accavg/10.,Etotavg/10."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}