Cells Distribution analysis.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/fmena/anaconda3/envs/tesis/lib/python3.5/site-packages/h5py/__init__.py:34: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.\n",
      "  from ._conv import register_converters as _register_converters\n",
      "Using TensorFlow backend.\n",
      "/home/fmena/.local/lib/python3.5/site-packages/tensorflow/python/framework/dtypes.py:469: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint8 = np.dtype([(\"qint8\", np.int8, 1)])\n",
      "/home/fmena/.local/lib/python3.5/site-packages/tensorflow/python/framework/dtypes.py:470: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint8 = np.dtype([(\"quint8\", np.uint8, 1)])\n",
      "/home/fmena/.local/lib/python3.5/site-packages/tensorflow/python/framework/dtypes.py:471: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint16 = np.dtype([(\"qint16\", np.int16, 1)])\n",
      "/home/fmena/.local/lib/python3.5/site-packages/tensorflow/python/framework/dtypes.py:472: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint16 = np.dtype([(\"quint16\", np.uint16, 1)])\n",
      "/home/fmena/.local/lib/python3.5/site-packages/tensorflow/python/framework/dtypes.py:473: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint32 = np.dtype([(\"qint32\", np.int32, 1)])\n",
      "/home/fmena/.local/lib/python3.5/site-packages/tensorflow/python/framework/dtypes.py:476: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  np_resource = np.dtype([(\"resource\", np.ubyte, 1)])\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package reuters to /home/fmena/nltk_data...\n",
      "[nltk_data]   Package reuters is already up-to-date!\n",
      "[nltk_data] Downloading package wordnet to /home/fmena/nltk_data...\n",
      "[nltk_data]   Package wordnet is already up-to-date!\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import keras,gc, os, time, sys, nltk\n",
    "from keras.layers import *\n",
    "from keras.models import Sequential,Model\n",
    "from keras import backend as K\n",
    "from astropy.table import Table\n",
    "import matplotlib.pyplot as plt\n",
    "from scipy.special import expit\n",
    "import seaborn as sns\n",
    "import pandas as pd\n",
    "from scipy.io import loadmat\n",
    "nltk.download('reuters')\n",
    "nltk.download('wordnet')\n",
    "%matplotlib inline\n",
    "\n",
    "from base_networks import *\n",
    "from utils import check_availability, load_imgs_mask\n",
    "from utils import get_topK_labels,set_newlabel_list, enmask_data\n",
    "\n",
    "np.random.seed(20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def plot_cells_distribution(Nb,train_hash,test_hash=[]):\n",
    "    print(\"Cantidad de memorias en tabla hash: \",2**Nb)\n",
    "\n",
    "    valores_unicos, count_hash =  hash_analysis(train_hash)\n",
    "    print(\"Entrenamiento----\")\n",
    "    print(\"Cantidad de datos a llenar la tabla hash: \",train_hash.shape[0])\n",
    "    plt.figure(figsize=(14,4))\n",
    "    plt.plot(sorted(list(count_hash.values()))[::-1],'bo-')\n",
    "    plt.title(\"Cantidad de datos por celda/memoria en tabla hash\")\n",
    "    plt.show()\n",
    "    print(\"Cantidad de memorias ocupadas: \",len(valores_unicos))\n",
    "    if len(test_hash) != 0:\n",
    "        valores_unicos, count_hash =  hash_analysis(test_hash)\n",
    "        print(\"Pruebas-----\")\n",
    "        print(\"Cantidad de datos a llenar la tabla hash: \",test_hash.shape[0])\n",
    "        plt.figure(figsize=(14,4))\n",
    "        plt.plot(sorted(list(count_hash.values()))[::-1],'bo-')\n",
    "        plt.title(\"Cantidad de datos por celda/memoria en tabla hash\")\n",
    "        plt.show()\n",
    "        print(\"Cantidad de memorias ocupadas: \",len(valores_unicos))\n",
    "        \n",
    "def hash_analysis(hash_data):\n",
    "    hash_string = []\n",
    "    for valor in hash_data:\n",
    "        hash_string.append(str(valor)[1:-1].replace(' ',''))\n",
    "    valores_unicos = set(hash_string)\n",
    "    count_hash = {valor: hash_string.count(valor) for valor in valores_unicos}\n",
    "    return valores_unicos, count_hash\n",
    "\n",
    "\n",
    "def measure_metrics(hash_data):\n",
    "    valores_unicos, count_hash =  hash_analysis(hash_data)\n",
    "    dis_cell = np.asarray(list(count_hash.values()))\n",
    "    \n",
    "    N_data, B_data = hash_data.shape\n",
    "    \n",
    "    norm_dis_cell = dis_cell/N_data #np.sum(dis_cell)\n",
    "    \n",
    "    N_hash = 2.**B_data\n",
    "    N_hash_used = len(dis_cell)\n",
    "    N_hash_empty = N_hash-N_hash_used\n",
    "\n",
    "    top10_dis = sorted(dis_cell)[::-1][:10]\n",
    "    mean_N_xcell = np.mean(dis_cell)\n",
    "    std_N_xcell = np.std(dis_cell)\n",
    "    \n",
    "    #calcular varianza y cosas..\n",
    "    \n",
    "    ref_dis = np.tile(1./len(norm_dis_cell), len(norm_dis_cell))\n",
    "    \n",
    "    H_cells = -np.sum(norm_dis_cell* np.log(norm_dis_cell+ 1e-7))\n",
    "    H_balanced = -np.sum(ref_dis* np.log(ref_dis+ 1e-7))\n",
    "    \n",
    "    G_indx = 1 - np.sum(norm_dis_cell**2)\n",
    "    G_indx_balanced = 1 - np.sum(ref_dis**2)    \n",
    "    return N_data, N_hash, N_hash_used, N_hash_empty, top10_dis,mean_N_xcell,   std_N_xcell, H_cells/H_balanced,G_indx/G_indx_balanced\n",
    "\n",
    "def compare_cells_dist(train_hash_, train_hash_B, test_hash_, test_hash_B):\n",
    "    N_data = []\n",
    "    Np_hash_empty = []\n",
    "    top10_dis = []\n",
    "    mean_N_xcell = []\n",
    "    std_N_xcell = []\n",
    "    H_N = []\n",
    "    Gini_N = []\n",
    "    \n",
    "    n, a, b, c, d,e , f, g, h = measure_metrics(train_hash_)\n",
    "    N_data.append(n)\n",
    "    Np_hash_empty.append(c/a)\n",
    "    top10_dis.append(d)\n",
    "    mean_N_xcell.append(e)\n",
    "    std_N_xcell.append(f)\n",
    "    H_N.append(g)   \n",
    "    Gini_N.append(h)\n",
    "    \n",
    "    n, a, b, c, d,e , f, g, h = measure_metrics(train_hash_B)\n",
    "    N_data.append(n)\n",
    "    Np_hash_empty.append(c/a)\n",
    "    top10_dis.append(d)\n",
    "    mean_N_xcell.append(e)\n",
    "    std_N_xcell.append(f)\n",
    "    H_N.append(g)  \n",
    "    Gini_N.append(h)\n",
    "    \n",
    "    n, a, b, c, d,e , f, g, h = measure_metrics(test_hash_)\n",
    "    N_data.append(n)\n",
    "    Np_hash_empty.append(c/a)\n",
    "    top10_dis.append(d)\n",
    "    mean_N_xcell.append(e)\n",
    "    std_N_xcell.append(f)\n",
    "    H_N.append(g)  \n",
    "    Gini_N.append(h)\n",
    "    \n",
    "    n, a, b, c, d,e , f, g, h = measure_metrics(test_hash_B)\n",
    "    N_data.append(n)\n",
    "    Np_hash_empty.append(c/a)\n",
    "    top10_dis.append(d)\n",
    "    mean_N_xcell.append(e)\n",
    "    std_N_xcell.append(f)\n",
    "    H_N.append(g)  \n",
    "    Gini_N.append(h)\n",
    "    \n",
    "    df = pd.DataFrame({\"#data\":N_data, \"#empty cells(%)\":Np_hash_empty, \n",
    "                       \"mean #data x cell\":mean_N_xcell, \"std #data x cell\":std_N_xcell,\n",
    "                       \"H cells dis\":H_N, \"Gini cells dis\": Gini_N,\n",
    "                      \"top10\":top10_dis})\n",
    "    df[\"Type\"] = [\"VDSH train\",\"B-VAE train\", \"VDSH test\",\"B-VAE test\"]\n",
    "\n",
    "    return df\n",
    "\n",
    "#cells \t#data \t#empty cells \t#empty cells(%) \t#used cells/#data \tmean #data x cell \ttop10 \tType\n",
    "\n",
    "from utils import calculate_hash, compare_cells_plot"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Choose load data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cantidad de datos Entrenamiento:  11016\n",
      "Cantidad de datos Validación:  3667\n",
      "Cantidad de datos Pruebas:  3668\n"
     ]
    }
   ],
   "source": [
    "name_dat = \"20News\"\n",
    "\n",
    "from sklearn.datasets import fetch_20newsgroups\n",
    "newsgroups_t = fetch_20newsgroups(subset='train')\n",
    "labels = newsgroups_t.target_names\n",
    "\n",
    "from utils import Load_Dataset\n",
    "\n",
    "filename = 'Data/ng20.tfidf.mat'\n",
    "data = Load_Dataset(filename)\n",
    "\n",
    "X_train_input = data[\"train\"]\n",
    "X_train = X_train_input \n",
    "X_val_input = data[\"cv\"]\n",
    "X_val = X_val_input \n",
    "X_test_input = data[\"test\"]\n",
    "X_test = X_test_input\n",
    "labels_train = np.asarray([labels[value.argmax(axis=-1)] for value in data[\"gnd_train\"]])\n",
    "labels_val = np.asarray([labels[value.argmax(axis=-1)] for value in data[\"gnd_cv\"]])\n",
    "labels_test = np.asarray([labels[value.argmax(axis=-1)] for value in data[\"gnd_test\"]])\n",
    "\n",
    "\n",
    "print(\"Cantidad de datos Entrenamiento: \",len(labels_train))\n",
    "print(\"Cantidad de datos Validación: \",len(labels_val))\n",
    "print(\"Cantidad de datos Pruebas: \",len(labels_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cantidad de datos Entrenamiento:  7752\n",
      "Cantidad de datos Validación:  964\n",
      "Cantidad de datos Pruebas:  967\n"
     ]
    }
   ],
   "source": [
    "name_dat = \"Reuters\"\n",
    "labels = ['a','b','c','d','e','f','e','g','h','i','j','k','l','m','n','o','p','q','r','s']\n",
    "\n",
    "filename = 'Data/reuters.tfidf.mat'\n",
    "data = Load_Dataset(filename)\n",
    "\n",
    "X_train_input = data[\"train\"]\n",
    "X_train = X_train_input \n",
    "X_val_input = data[\"cv\"]\n",
    "X_val = X_val_input \n",
    "X_test_input = data[\"test\"]\n",
    "X_test = X_test_input\n",
    "\n",
    "labels = np.asarray(labels)\n",
    "labels_train = np.asarray([labels[value.astype(bool)] for value in data[\"gnd_train\"]])\n",
    "labels_val = np.asarray([labels[value.astype(bool)] for value in data[\"gnd_cv\"]])\n",
    "labels_test = np.asarray([labels[value.astype(bool)] for value in data[\"gnd_test\"]])\n",
    "\n",
    "print(\"Cantidad de datos Entrenamiento: \",len(labels_train))\n",
    "print(\"Cantidad de datos Validación: \",len(labels_val))\n",
    "print(\"Cantidad de datos Pruebas: \",len(labels_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cantidad de datos Entrenamiento:  21286\n",
      "Cantidad de datos Validación:  3498\n",
      "Cantidad de datos Pruebas:  3498\n"
     ]
    }
   ],
   "source": [
    "name_dat = \"TMC\"\n",
    "\n",
    "labels = ['a','b','c','d','e','f','e','g','h','i','j','k','l','m','n','o','p','q','r','s','t','u']\n",
    "\n",
    "filename = 'Data/tmc.tfidf.mat'\n",
    "data = Load_Dataset(filename)\n",
    "\n",
    "X_train_input = data[\"train\"]\n",
    "X_train = X_train_input \n",
    "X_val_input = data[\"cv\"]\n",
    "X_val = X_val_input \n",
    "X_test_input = data[\"test\"]\n",
    "X_test = X_test_input\n",
    "\n",
    "labels = np.asarray(labels)\n",
    "labels_train = np.asarray([labels[value.astype(bool)] for value in data[\"gnd_train\"]])\n",
    "labels_val = np.asarray([labels[value.astype(bool)] for value in data[\"gnd_cv\"]])\n",
    "labels_test = np.asarray([labels[value.astype(bool)] for value in data[\"gnd_test\"]])\n",
    "\n",
    "print(\"Cantidad de datos Entrenamiento: \",len(labels_train))\n",
    "print(\"Cantidad de datos Validación: \",len(labels_val))\n",
    "print(\"Cantidad de datos Pruebas: \",len(labels_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Datos de entrenamiento:  10060\n",
      "Datos de pruebas:  2280\n",
      "Cantidad de datos Entrenamiento:  9054\n",
      "Cantidad de datos Validación:  1006\n",
      "Cantidad de datos Pruebas:  2280\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/fmena/.local/lib/python3.5/site-packages/sklearn/feature_extraction/text.py:507: UserWarning: The parameter 'token_pattern' will not be used since 'tokenizer' is not None'\n",
      "  warnings.warn(\"The parameter 'token_pattern' will not be used\"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 1.63 s, sys: 5.74 ms, total: 1.63 s\n",
      "Wall time: 1.62 s\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAD8CAYAAACMwORRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd0be02ff60>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "name_dat = \"Snippets\"\n",
    "\n",
    "def read_file(archivo,symb=' '):\n",
    "    with open(archivo,'r') as f:\n",
    "        lineas = f.readlines()\n",
    "        tokens_f = [linea.strip().split(symb) for linea in lineas]\n",
    "        labels = [tokens[-1] for tokens in tokens_f]\n",
    "        tokens = [' '.join(tokens[:-1]) for tokens in tokens_f]\n",
    "    return labels,tokens\n",
    "labels_t,texts_t = read_file(\"Data/data-web-snippets/train.txt\")\n",
    "labels_test,texts_test = read_file(\"Data/data-web-snippets/test.txt\")\n",
    "print(\"Datos de entrenamiento: \",len(texts_t))\n",
    "print(\"Datos de pruebas: \",len(texts_test))\n",
    "\n",
    "labels = list(set(labels_t))\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "labels_t = np.asarray(labels_t)\n",
    "labels_test = np.asarray(labels_test)\n",
    "texts_train,texts_val,labels_train,labels_val  = train_test_split(texts_t,labels_t,random_state=20,test_size=0.1)\n",
    "\n",
    "print(\"Cantidad de datos Entrenamiento: \",len(texts_train))\n",
    "print(\"Cantidad de datos Validación: \",len(texts_val))\n",
    "print(\"Cantidad de datos Pruebas: \",len(texts_test))\n",
    "\n",
    "\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "from nltk.stem.snowball import SnowballStemmer\n",
    "from nltk.stem import WordNetLemmatizer \n",
    "tokenizer = TfidfVectorizer().build_tokenizer()\n",
    "stemmer = SnowballStemmer(\"english\") \n",
    "lemmatizer = WordNetLemmatizer()\n",
    "\n",
    "\"\"\"Extract features from raw input\"\"\"\n",
    "def preProcess(s): #String processor\n",
    "    return s.lower().strip().strip('-').strip('_')\n",
    "def number_normalize(doc):\n",
    "    results = []\n",
    "    for token in tokenizer(doc):\n",
    "        token_pro = preProcess(token)\n",
    "        if len(token_pro) != 0 and not token_pro[0].isdigit():\n",
    "            results.append(token_pro)\n",
    "    return results\n",
    "def stemmed_words(doc):\n",
    "    results = []\n",
    "    for token in tokenizer(doc):\n",
    "        pre_pro = preProcess(token)\n",
    "        #token_pro = stemmer.stem(pre_pro) #aumenta x10 el tiempo de procesamiento\n",
    "        token_pro = lemmatizer.lemmatize(pre_pro) #so can explain/interpretae -- aumenta x5 el tiempo de proce\n",
    "        if len(token_pro) > 2 and not token_pro[0].isdigit(): #elimina palabra largo menor a 2\n",
    "            results.append(token_pro)\n",
    "    return results\n",
    "\n",
    "def get_transform_representation(mode, analizer,min_count,max_feat):\n",
    "    smooth_idf_b = False\n",
    "    use_idf_b = False\n",
    "    binary_b = False\n",
    "\n",
    "    if mode == 'binary':\n",
    "        binary_b = True\n",
    "    elif mode == 'tf':     \n",
    "        pass #default is tf\n",
    "    elif mode == 'tf-idf':\n",
    "        use_idf_b = True\n",
    "        smooth_idf_b = True #inventa 1 conteo imaginario (como priors)--laplace smoothing\n",
    "    return TfidfVectorizer(stop_words='english',tokenizer=analizer,min_df=min_count, max_df=0.8, max_features=max_feat\n",
    "                                ,binary=binary_b, use_idf=use_idf_b, smooth_idf=smooth_idf_b,norm=None\n",
    "                                  ,ngram_range=(1, 3)) \n",
    "\n",
    "min_count = 1 #default = 1\n",
    "max_feat = 10000 #Best: 10000 -- Hinton (2000)\n",
    "vectorizer = get_transform_representation(\"tf\", stemmed_words,min_count,max_feat)\n",
    "\n",
    "%time vectorizer.fit(texts_train)\n",
    "vectors_train = vectorizer.transform(texts_train)\n",
    "vectors_val = vectorizer.transform(texts_val)\n",
    "vectors_test = vectorizer.transform(texts_test)\n",
    "\n",
    "token2idx = vectorizer.vocabulary_\n",
    "idx2token = {idx:token for token,idx in token2idx.items()}\n",
    "\n",
    "\n",
    "\n",
    "#todense --get representation\n",
    "X_train = np.asarray(vectors_train.todense())\n",
    "X_val = np.asarray(vectors_val.todense())\n",
    "X_test = np.asarray(vectors_test.todense())\n",
    "del vectors_train,vectors_val,vectors_test#,vectors_train2,vectors_val2,vectors_test2\n",
    "gc.collect()\n",
    "\n",
    "\n",
    "##representacion soft para TF ---mucho mejor!\n",
    "X_train_input = np.log(X_train+1) \n",
    "X_val_input = np.log(X_val+1) \n",
    "X_test_input = np.log(X_test+1) \n",
    "plt.plot(X_train_input[0])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Para texto"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/fmena/anaconda3/envs/tesis/lib/python3.5/site-packages/ipykernel/__main__.py:2: RuntimeWarning: invalid value encountered in true_divide\n",
      "  from ipykernel import kernelapp as app\n",
      "/home/fmena/anaconda3/envs/tesis/lib/python3.5/site-packages/ipykernel/__main__.py:3: RuntimeWarning: invalid value encountered in true_divide\n",
      "  app.launch_new_instance()\n",
      "/home/fmena/anaconda3/envs/tesis/lib/python3.5/site-packages/ipykernel/__main__.py:4: RuntimeWarning: invalid value encountered in true_divide\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Output target normalizado en dataset  Snippets\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "4935"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#outputs as probabolities -- normalized over datasets..\n",
    "X_train = X_train/X_train.sum(axis=-1,keepdims=True) \n",
    "X_val = X_val/X_val.sum(axis=-1,keepdims=True)\n",
    "X_test = X_test/X_test.sum(axis=-1,keepdims=True)\n",
    "print(\"Output target normalizado en dataset \",name_dat)\n",
    "\n",
    "X_train[np.isnan(X_train)] = 0\n",
    "X_val[np.isnan(X_val)] = 0\n",
    "X_test[np.isnan(X_test)] = 0\n",
    "\n",
    "X_total_input = np.concatenate((X_train_input,X_val_input),axis=0)\n",
    "X_total = np.concatenate((X_train,X_val),axis=0)\n",
    "labels_total = np.concatenate((labels_train,labels_val),axis=0)\n",
    "del X_train_input, X_train, X_val, X_val_input\n",
    "gc.collect()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(70000, 512)"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "name_dat = \"MNIST\"\n",
    "\n",
    "(_, aux_t), (_, aux_test) = keras.datasets.mnist.load_data()\n",
    "labels = [\"0\", \"1\", \"2\", \"3\", \"4\", \"5\", \"6\", \"7\", \"8\", \"9\"]\n",
    "labels_t = np.asarray([labels[value] for value in aux_t])\n",
    "labels_test = np.asarray([labels[value] for value in aux_test])\n",
    "labels_t = np.concatenate((labels_t,labels_test),axis=0)\n",
    "\n",
    "X_t = np.load(\"../AUX/VGG_128/mnist_VGG_avg.npy\")\n",
    "X_t.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 512)"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "name_dat = \"CIFAR-10\"\n",
    "\n",
    "(_, aux_t), (_, aux_test) = keras.datasets.cifar10.load_data()\n",
    "\n",
    "labels = [\"airplane\", \"automobile\",\"bird\", \"cat\",\"deer\",\"dog\",\"frog\",\"horse\",\"ship\",\"truck\"]\n",
    "labels_t = np.asarray([labels[value[0]] for value in aux_t])\n",
    "labels_test = np.asarray([labels[value[0]] for value in aux_test])\n",
    "labels_t = np.concatenate((labels_t,labels_test),axis=0)\n",
    "\n",
    "X_t = np.load(\"../AUX/VGG_224/cifar10_VGG_avg.npy\") #mejora\n",
    "X_t.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cantidad de labels:  81\n",
      "Cantidad de objetos:  169500\n",
      "category with most data (sky) has = 61066, the top-K category (mountain) has = 4232\n",
      "Cantidad de objetos:  158383\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(158383, 512)"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "name_dat = \"Nus-Wide\"\n",
    "\n",
    "folder = \"../Dataset_NUSWIDE/\"\n",
    "mask_av = np.loadtxt(\"./Data/\"+name_dat+\"_mask_avail.txt\").astype(bool)\n",
    "\n",
    "labels = pd.read_csv(folder+'Concepts81.txt',header=None).values.reshape(1,-1)[0]\n",
    "print(\"Cantidad de labels: \",len(labels) )\n",
    "\n",
    "labels_t = [[] for _ in range(269648)]\n",
    "for concept in labels:\n",
    "    aux = pd.read_csv(folder+\"Groundtruth/AllLabels/Labels_\"+concept+\".txt\",header=None)\n",
    "    indexs_true = aux.loc[(aux==1).values[:,0]].index\n",
    "    \n",
    "    for value in indexs_true:\n",
    "        labels_t[value].append(concept)\n",
    "        \n",
    "labels_t = enmask_data(labels_t, mask_av)\n",
    "N_total = len(labels_t)\n",
    "print(\"Cantidad de objetos: \",N_total )\n",
    "\n",
    "new_labels = get_topK_labels(labels_t, labels, K=21)\n",
    "\n",
    "labels_t = set_newlabel_list(new_labels, labels_t)\n",
    "labels = new_labels\n",
    "# y si quedan datos sin clase?\n",
    "mask_used_t = np.asarray(list(map(len,labels_t))) != 0\n",
    "\n",
    "labels_t = enmask_data(labels_t, mask_used_t)\n",
    "print(\"Cantidad de objetos: \", len(labels_t) )\n",
    "\n",
    "X_t = np.load(\"../AUX/VGG_224/nuswide_VGG_avg.npy\")\n",
    "X_t.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Para imagenes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cantidad de datos Entrenamiento:  69000\n",
      "Cantidad de datos Pruebas:  1000\n"
     ]
    }
   ],
   "source": [
    "from utils import sample_test_mask\n",
    "mask_train = sample_test_mask(labels_t, N=100)\n",
    "\n",
    "## creat test como dicen...\n",
    "X_test = X_t[~mask_train]\n",
    "X_t = X_t[mask_train]\n",
    "labels_test = enmask_data(labels_t, ~mask_train)\n",
    "labels_t = enmask_data(labels_t, mask_train)\n",
    "\n",
    "gc.collect()\n",
    "print(\"Cantidad de datos Entrenamiento: \",len(X_t))\n",
    "print(\"Cantidad de datos Pruebas: \",len(X_test))\n",
    "\n",
    "\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "std = StandardScaler(with_mean=True, with_std=True)\n",
    "std.fit(X_t)\n",
    "\n",
    "X_total = std.transform(X_t)\n",
    "X_test = std.transform(X_test)\n",
    "\n",
    "X_total_input = X_total\n",
    "X_test_input = X_test"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## load model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "TFIDF\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_6 (InputLayer)         (None, 10000)             0         \n",
      "_________________________________________________________________\n",
      "sequential_28 (Sequential)   (None, 500)               5255000   \n",
      "_________________________________________________________________\n",
      "logits-b (Dense)             (None, 32)                16032     \n",
      "=================================================================\n",
      "Total params: 5,271,032\n",
      "Trainable params: 5,269,032\n",
      "Non-trainable params: 2,000\n",
      "_________________________________________________________________\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/fmena/.local/lib/python3.5/site-packages/keras/models.py:251: UserWarning: No training configuration found in save file: the model was *not* compiled. Compile it manually.\n",
      "  warnings.warn('No training configuration found in save file: '\n"
     ]
    }
   ],
   "source": [
    "if \"mnist\" in name_dat.lower() or \"cifar\" in name_dat.lower() or \"nus-wide\" in name_dat.lower():\n",
    "    type_n = \"VGG\"\n",
    "else:\n",
    "    type_n = \"TFIDF\"\n",
    "\n",
    "print(type_n)\n",
    "\n",
    "BITS_S = 32\n",
    "\n",
    "##load model\n",
    "encoder_Tvae = keras.models.load_model(\"saved_models/\"+name_dat+\"_VAE_\"+str(BITS_S)+\"b_E_\"+type_n+\"_B?.h5\")\n",
    "encoder_Bvae = keras.models.load_model(\"saved_models/\"+name_dat+\"_BVAE_\"+str(BITS_S)+\"b_E_\"+type_n+\"_B?.h5\")\n",
    "encoder_Bvae.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "train_hash_ = calculate_hash(encoder_Tvae.predict(X_total_input), from_probas=False)\n",
    "test_hash_ = calculate_hash(encoder_Tvae.predict(X_test_input),  from_probas=False)\n",
    "\n",
    "train_hash_B = calculate_hash(encoder_Bvae.predict(X_total_input), from_probas=True)\n",
    "test_hash_B = calculate_hash(encoder_Bvae.predict(X_test_input),  from_probas=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Numero de Bits=  8\n"
     ]
    }
   ],
   "source": [
    "print(\"Numero de Bits= \",BITS_S)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "20News\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>#data</th>\n",
       "      <th>#empty cells(%)</th>\n",
       "      <th>Gini cells dis</th>\n",
       "      <th>H cells dis</th>\n",
       "      <th>mean #data x cell</th>\n",
       "      <th>std #data x cell</th>\n",
       "      <th>top10</th>\n",
       "      <th>Type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>14683</td>\n",
       "      <td>0.078125</td>\n",
       "      <td>0.987452</td>\n",
       "      <td>0.839131</td>\n",
       "      <td>62.216102</td>\n",
       "      <td>106.836759</td>\n",
       "      <td>[874, 696, 394, 383, 372, 361, 359, 347, 341, ...</td>\n",
       "      <td>VDSH train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>14683</td>\n",
       "      <td>0.148438</td>\n",
       "      <td>0.980620</td>\n",
       "      <td>0.800122</td>\n",
       "      <td>67.353211</td>\n",
       "      <td>138.123097</td>\n",
       "      <td>[1151, 866, 656, 578, 484, 483, 465, 462, 334,...</td>\n",
       "      <td>B-VAE train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3668</td>\n",
       "      <td>0.269531</td>\n",
       "      <td>0.984713</td>\n",
       "      <td>0.836358</td>\n",
       "      <td>19.614973</td>\n",
       "      <td>33.075047</td>\n",
       "      <td>[260, 194, 135, 109, 98, 96, 88, 88, 87, 84]</td>\n",
       "      <td>VDSH test</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3668</td>\n",
       "      <td>0.316406</td>\n",
       "      <td>0.976047</td>\n",
       "      <td>0.797487</td>\n",
       "      <td>20.960000</td>\n",
       "      <td>42.790701</td>\n",
       "      <td>[322, 253, 185, 174, 172, 141, 131, 104, 94, 91]</td>\n",
       "      <td>B-VAE test</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   #data  #empty cells(%)  Gini cells dis  H cells dis  mean #data x cell  \\\n",
       "0  14683         0.078125        0.987452     0.839131          62.216102   \n",
       "1  14683         0.148438        0.980620     0.800122          67.353211   \n",
       "2   3668         0.269531        0.984713     0.836358          19.614973   \n",
       "3   3668         0.316406        0.976047     0.797487          20.960000   \n",
       "\n",
       "   std #data x cell                                              top10  \\\n",
       "0        106.836759  [874, 696, 394, 383, 372, 361, 359, 347, 341, ...   \n",
       "1        138.123097  [1151, 866, 656, 578, 484, 483, 465, 462, 334,...   \n",
       "2         33.075047       [260, 194, 135, 109, 98, 96, 88, 88, 87, 84]   \n",
       "3         42.790701   [322, 253, 185, 174, 172, 141, 131, 104, 94, 91]   \n",
       "\n",
       "          Type  \n",
       "0   VDSH train  \n",
       "1  B-VAE train  \n",
       "2    VDSH test  \n",
       "3   B-VAE test  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#plot together--\n",
    "print(name_dat)\n",
    "compare_cells_dist(train_hash_, train_hash_B, test_hash_, test_hash_B)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reuters\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>#data</th>\n",
       "      <th>#empty cells(%)</th>\n",
       "      <th>Gini cells dis</th>\n",
       "      <th>H cells dis</th>\n",
       "      <th>mean #data x cell</th>\n",
       "      <th>std #data x cell</th>\n",
       "      <th>top10</th>\n",
       "      <th>Type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>8716</td>\n",
       "      <td>0.359375</td>\n",
       "      <td>0.973429</td>\n",
       "      <td>0.791104</td>\n",
       "      <td>53.146341</td>\n",
       "      <td>110.604947</td>\n",
       "      <td>[865, 815, 376, 338, 335, 267, 247, 245, 233, ...</td>\n",
       "      <td>VDSH train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>8716</td>\n",
       "      <td>0.757812</td>\n",
       "      <td>0.952928</td>\n",
       "      <td>0.742352</td>\n",
       "      <td>140.580645</td>\n",
       "      <td>238.216325</td>\n",
       "      <td>[1031, 951, 714, 661, 647, 500, 483, 480, 426,...</td>\n",
       "      <td>B-VAE train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>967</td>\n",
       "      <td>0.652344</td>\n",
       "      <td>0.973432</td>\n",
       "      <td>0.843104</td>\n",
       "      <td>10.865169</td>\n",
       "      <td>16.613377</td>\n",
       "      <td>[109, 82, 57, 40, 37, 37, 33, 32, 28, 22]</td>\n",
       "      <td>VDSH test</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>967</td>\n",
       "      <td>0.843750</td>\n",
       "      <td>0.952819</td>\n",
       "      <td>0.795213</td>\n",
       "      <td>24.175000</td>\n",
       "      <td>32.793206</td>\n",
       "      <td>[118, 114, 102, 82, 79, 64, 59, 48, 40, 31]</td>\n",
       "      <td>B-VAE test</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   #data  #empty cells(%)  Gini cells dis  H cells dis  mean #data x cell  \\\n",
       "0   8716         0.359375        0.973429     0.791104          53.146341   \n",
       "1   8716         0.757812        0.952928     0.742352         140.580645   \n",
       "2    967         0.652344        0.973432     0.843104          10.865169   \n",
       "3    967         0.843750        0.952819     0.795213          24.175000   \n",
       "\n",
       "   std #data x cell                                              top10  \\\n",
       "0        110.604947  [865, 815, 376, 338, 335, 267, 247, 245, 233, ...   \n",
       "1        238.216325  [1031, 951, 714, 661, 647, 500, 483, 480, 426,...   \n",
       "2         16.613377          [109, 82, 57, 40, 37, 37, 33, 32, 28, 22]   \n",
       "3         32.793206        [118, 114, 102, 82, 79, 64, 59, 48, 40, 31]   \n",
       "\n",
       "          Type  \n",
       "0   VDSH train  \n",
       "1  B-VAE train  \n",
       "2    VDSH test  \n",
       "3   B-VAE test  "
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#plot together--\n",
    "print(name_dat)\n",
    "compare_cells_dist(train_hash_, train_hash_B, test_hash_, test_hash_B)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "TMC\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>#data</th>\n",
       "      <th>#empty cells(%)</th>\n",
       "      <th>Gini cells dis</th>\n",
       "      <th>H cells dis</th>\n",
       "      <th>mean #data x cell</th>\n",
       "      <th>std #data x cell</th>\n",
       "      <th>top10</th>\n",
       "      <th>Type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>24784</td>\n",
       "      <td>0.207031</td>\n",
       "      <td>0.987905</td>\n",
       "      <td>0.841057</td>\n",
       "      <td>122.088670</td>\n",
       "      <td>190.836974</td>\n",
       "      <td>[1214, 1137, 1080, 680, 644, 611, 590, 570, 51...</td>\n",
       "      <td>VDSH train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>24784</td>\n",
       "      <td>0.640625</td>\n",
       "      <td>0.973804</td>\n",
       "      <td>0.786794</td>\n",
       "      <td>269.391304</td>\n",
       "      <td>415.931222</td>\n",
       "      <td>[1796, 1602, 1337, 1266, 1184, 1182, 1157, 110...</td>\n",
       "      <td>B-VAE train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3498</td>\n",
       "      <td>0.410156</td>\n",
       "      <td>0.988030</td>\n",
       "      <td>0.867489</td>\n",
       "      <td>23.165563</td>\n",
       "      <td>31.040762</td>\n",
       "      <td>[186, 172, 151, 96, 93, 92, 88, 87, 82, 76]</td>\n",
       "      <td>VDSH test</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3498</td>\n",
       "      <td>0.746094</td>\n",
       "      <td>0.975059</td>\n",
       "      <td>0.826833</td>\n",
       "      <td>53.815385</td>\n",
       "      <td>67.991604</td>\n",
       "      <td>[252, 209, 205, 201, 199, 186, 178, 165, 160, ...</td>\n",
       "      <td>B-VAE test</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   #data  #empty cells(%)  Gini cells dis  H cells dis  mean #data x cell  \\\n",
       "0  24784         0.207031        0.987905     0.841057         122.088670   \n",
       "1  24784         0.640625        0.973804     0.786794         269.391304   \n",
       "2   3498         0.410156        0.988030     0.867489          23.165563   \n",
       "3   3498         0.746094        0.975059     0.826833          53.815385   \n",
       "\n",
       "   std #data x cell                                              top10  \\\n",
       "0        190.836974  [1214, 1137, 1080, 680, 644, 611, 590, 570, 51...   \n",
       "1        415.931222  [1796, 1602, 1337, 1266, 1184, 1182, 1157, 110...   \n",
       "2         31.040762        [186, 172, 151, 96, 93, 92, 88, 87, 82, 76]   \n",
       "3         67.991604  [252, 209, 205, 201, 199, 186, 178, 165, 160, ...   \n",
       "\n",
       "          Type  \n",
       "0   VDSH train  \n",
       "1  B-VAE train  \n",
       "2    VDSH test  \n",
       "3   B-VAE test  "
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#plot together--\n",
    "print(name_dat)\n",
    "compare_cells_dist(train_hash_, train_hash_B, test_hash_, test_hash_B)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Snippets\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>#data</th>\n",
       "      <th>#empty cells(%)</th>\n",
       "      <th>Gini cells dis</th>\n",
       "      <th>H cells dis</th>\n",
       "      <th>mean #data x cell</th>\n",
       "      <th>std #data x cell</th>\n",
       "      <th>top10</th>\n",
       "      <th>Type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>10060</td>\n",
       "      <td>0.050781</td>\n",
       "      <td>0.992044</td>\n",
       "      <td>0.874373</td>\n",
       "      <td>41.399177</td>\n",
       "      <td>57.445780</td>\n",
       "      <td>[475, 299, 273, 205, 204, 201, 179, 176, 161, ...</td>\n",
       "      <td>VDSH train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>10060</td>\n",
       "      <td>0.042969</td>\n",
       "      <td>0.989268</td>\n",
       "      <td>0.859134</td>\n",
       "      <td>41.061224</td>\n",
       "      <td>66.445500</td>\n",
       "      <td>[643, 281, 281, 272, 246, 211, 211, 206, 196, ...</td>\n",
       "      <td>B-VAE train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2280</td>\n",
       "      <td>0.453125</td>\n",
       "      <td>0.980866</td>\n",
       "      <td>0.845335</td>\n",
       "      <td>16.285714</td>\n",
       "      <td>26.559175</td>\n",
       "      <td>[209, 122, 113, 99, 99, 57, 54, 54, 53, 53]</td>\n",
       "      <td>VDSH test</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2280</td>\n",
       "      <td>0.273438</td>\n",
       "      <td>0.987749</td>\n",
       "      <td>0.866294</td>\n",
       "      <td>12.258065</td>\n",
       "      <td>18.454188</td>\n",
       "      <td>[131, 104, 104, 76, 67, 61, 52, 52, 49, 48]</td>\n",
       "      <td>B-VAE test</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   #data  #empty cells(%)  Gini cells dis  H cells dis  mean #data x cell  \\\n",
       "0  10060         0.050781        0.992044     0.874373          41.399177   \n",
       "1  10060         0.042969        0.989268     0.859134          41.061224   \n",
       "2   2280         0.453125        0.980866     0.845335          16.285714   \n",
       "3   2280         0.273438        0.987749     0.866294          12.258065   \n",
       "\n",
       "   std #data x cell                                              top10  \\\n",
       "0         57.445780  [475, 299, 273, 205, 204, 201, 179, 176, 161, ...   \n",
       "1         66.445500  [643, 281, 281, 272, 246, 211, 211, 206, 196, ...   \n",
       "2         26.559175        [209, 122, 113, 99, 99, 57, 54, 54, 53, 53]   \n",
       "3         18.454188        [131, 104, 104, 76, 67, 61, 52, 52, 49, 48]   \n",
       "\n",
       "          Type  \n",
       "0   VDSH train  \n",
       "1  B-VAE train  \n",
       "2    VDSH test  \n",
       "3   B-VAE test  "
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#plot together--\n",
    "print(name_dat)\n",
    "compare_cells_dist(train_hash_, train_hash_B, test_hash_, test_hash_B)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MNIST\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>#data</th>\n",
       "      <th>#empty cells(%)</th>\n",
       "      <th>Gini cells dis</th>\n",
       "      <th>H cells dis</th>\n",
       "      <th>mean #data x cell</th>\n",
       "      <th>std #data x cell</th>\n",
       "      <th>top10</th>\n",
       "      <th>Type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>69000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.992141</td>\n",
       "      <td>0.885454</td>\n",
       "      <td>269.531250</td>\n",
       "      <td>381.551631</td>\n",
       "      <td>[2520, 2477, 2011, 1986, 1780, 1575, 1564, 135...</td>\n",
       "      <td>VDSH train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>69000</td>\n",
       "      <td>0.046875</td>\n",
       "      <td>0.967473</td>\n",
       "      <td>0.743243</td>\n",
       "      <td>282.786885</td>\n",
       "      <td>795.033124</td>\n",
       "      <td>[8063, 6959, 2396, 2303, 2267, 2132, 2125, 205...</td>\n",
       "      <td>B-VAE train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1000</td>\n",
       "      <td>0.222656</td>\n",
       "      <td>0.992204</td>\n",
       "      <td>0.904266</td>\n",
       "      <td>5.025126</td>\n",
       "      <td>6.243338</td>\n",
       "      <td>[41, 35, 31, 31, 28, 24, 22, 20, 18, 18]</td>\n",
       "      <td>VDSH test</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1000</td>\n",
       "      <td>0.515625</td>\n",
       "      <td>0.968277</td>\n",
       "      <td>0.814486</td>\n",
       "      <td>8.064516</td>\n",
       "      <td>15.930161</td>\n",
       "      <td>[125, 101, 39, 37, 37, 34, 31, 29, 28, 26]</td>\n",
       "      <td>B-VAE test</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   #data  #empty cells(%)  Gini cells dis  H cells dis  mean #data x cell  \\\n",
       "0  69000         0.000000        0.992141     0.885454         269.531250   \n",
       "1  69000         0.046875        0.967473     0.743243         282.786885   \n",
       "2   1000         0.222656        0.992204     0.904266           5.025126   \n",
       "3   1000         0.515625        0.968277     0.814486           8.064516   \n",
       "\n",
       "   std #data x cell                                              top10  \\\n",
       "0        381.551631  [2520, 2477, 2011, 1986, 1780, 1575, 1564, 135...   \n",
       "1        795.033124  [8063, 6959, 2396, 2303, 2267, 2132, 2125, 205...   \n",
       "2          6.243338           [41, 35, 31, 31, 28, 24, 22, 20, 18, 18]   \n",
       "3         15.930161         [125, 101, 39, 37, 37, 34, 31, 29, 28, 26]   \n",
       "\n",
       "          Type  \n",
       "0   VDSH train  \n",
       "1  B-VAE train  \n",
       "2    VDSH test  \n",
       "3   B-VAE test  "
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#plot together--\n",
    "print(name_dat)\n",
    "compare_cells_dist(train_hash_, train_hash_B, test_hash_, test_hash_B)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CIFAR-10\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>#data</th>\n",
       "      <th>#empty cells(%)</th>\n",
       "      <th>Gini cells dis</th>\n",
       "      <th>H cells dis</th>\n",
       "      <th>mean #data x cell</th>\n",
       "      <th>std #data x cell</th>\n",
       "      <th>top10</th>\n",
       "      <th>Type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>58100</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.996275</td>\n",
       "      <td>0.937852</td>\n",
       "      <td>226.953125</td>\n",
       "      <td>221.201685</td>\n",
       "      <td>[1444, 1421, 1316, 1210, 1102, 1022, 775, 773,...</td>\n",
       "      <td>VDSH train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>58100</td>\n",
       "      <td>0.003906</td>\n",
       "      <td>0.993046</td>\n",
       "      <td>0.898018</td>\n",
       "      <td>227.843137</td>\n",
       "      <td>302.816062</td>\n",
       "      <td>[2195, 1818, 1681, 1403, 1240, 1228, 1179, 111...</td>\n",
       "      <td>B-VAE train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1900</td>\n",
       "      <td>0.050781</td>\n",
       "      <td>0.992874</td>\n",
       "      <td>0.909589</td>\n",
       "      <td>7.818930</td>\n",
       "      <td>10.268020</td>\n",
       "      <td>[90, 74, 65, 49, 35, 35, 29, 25, 25, 24]</td>\n",
       "      <td>VDSH test</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1900</td>\n",
       "      <td>0.109375</td>\n",
       "      <td>0.990980</td>\n",
       "      <td>0.885086</td>\n",
       "      <td>8.333333</td>\n",
       "      <td>11.924593</td>\n",
       "      <td>[76, 74, 63, 54, 52, 49, 47, 41, 39, 39]</td>\n",
       "      <td>B-VAE test</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   #data  #empty cells(%)  Gini cells dis  H cells dis  mean #data x cell  \\\n",
       "0  58100         0.000000        0.996275     0.937852         226.953125   \n",
       "1  58100         0.003906        0.993046     0.898018         227.843137   \n",
       "2   1900         0.050781        0.992874     0.909589           7.818930   \n",
       "3   1900         0.109375        0.990980     0.885086           8.333333   \n",
       "\n",
       "   std #data x cell                                              top10  \\\n",
       "0        221.201685  [1444, 1421, 1316, 1210, 1102, 1022, 775, 773,...   \n",
       "1        302.816062  [2195, 1818, 1681, 1403, 1240, 1228, 1179, 111...   \n",
       "2         10.268020           [90, 74, 65, 49, 35, 35, 29, 25, 25, 24]   \n",
       "3         11.924593           [76, 74, 63, 54, 52, 49, 47, 41, 39, 39]   \n",
       "\n",
       "          Type  \n",
       "0   VDSH train  \n",
       "1  B-VAE train  \n",
       "2    VDSH test  \n",
       "3   B-VAE test  "
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#plot together--\n",
    "print(name_dat)\n",
    "compare_cells_dist(train_hash_, train_hash_B, test_hash_, test_hash_B)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Nus-Wide\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>#data</th>\n",
       "      <th>#empty cells(%)</th>\n",
       "      <th>Gini cells dis</th>\n",
       "      <th>H cells dis</th>\n",
       "      <th>mean #data x cell</th>\n",
       "      <th>std #data x cell</th>\n",
       "      <th>top10</th>\n",
       "      <th>Type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>156283</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.993023</td>\n",
       "      <td>0.907571</td>\n",
       "      <td>610.480469</td>\n",
       "      <td>814.272139</td>\n",
       "      <td>[7455, 5927, 4787, 4363, 2891, 2667, 2356, 221...</td>\n",
       "      <td>VDSH train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>156283</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.955354</td>\n",
       "      <td>0.748948</td>\n",
       "      <td>610.480469</td>\n",
       "      <td>2059.828910</td>\n",
       "      <td>[28525, 9328, 7761, 6489, 6484, 4344, 3823, 37...</td>\n",
       "      <td>B-VAE train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2100</td>\n",
       "      <td>0.140625</td>\n",
       "      <td>0.979651</td>\n",
       "      <td>0.844905</td>\n",
       "      <td>9.545455</td>\n",
       "      <td>20.150631</td>\n",
       "      <td>[220, 125, 107, 88, 56, 49, 41, 36, 33, 33]</td>\n",
       "      <td>VDSH test</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2100</td>\n",
       "      <td>0.300781</td>\n",
       "      <td>0.895384</td>\n",
       "      <td>0.663068</td>\n",
       "      <td>11.731844</td>\n",
       "      <td>50.626308</td>\n",
       "      <td>[622, 231, 109, 94, 68, 43, 43, 41, 38, 36]</td>\n",
       "      <td>B-VAE test</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    #data  #empty cells(%)  Gini cells dis  H cells dis  mean #data x cell  \\\n",
       "0  156283         0.000000        0.993023     0.907571         610.480469   \n",
       "1  156283         0.000000        0.955354     0.748948         610.480469   \n",
       "2    2100         0.140625        0.979651     0.844905           9.545455   \n",
       "3    2100         0.300781        0.895384     0.663068          11.731844   \n",
       "\n",
       "   std #data x cell                                              top10  \\\n",
       "0        814.272139  [7455, 5927, 4787, 4363, 2891, 2667, 2356, 221...   \n",
       "1       2059.828910  [28525, 9328, 7761, 6489, 6484, 4344, 3823, 37...   \n",
       "2         20.150631        [220, 125, 107, 88, 56, 49, 41, 36, 33, 33]   \n",
       "3         50.626308        [622, 231, 109, 94, 68, 43, 43, 41, 38, 36]   \n",
       "\n",
       "          Type  \n",
       "0   VDSH train  \n",
       "1  B-VAE train  \n",
       "2    VDSH test  \n",
       "3   B-VAE test  "
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#plot together--\n",
    "print(name_dat)\n",
    "compare_cells_dist(train_hash_, train_hash_B, test_hash_, test_hash_B)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Numero de Bits=  32\n"
     ]
    }
   ],
   "source": [
    "print(\"Numero de Bits= \",BITS_S)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "20News\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>#data</th>\n",
       "      <th>#empty cells(%)</th>\n",
       "      <th>Gini cells dis</th>\n",
       "      <th>H cells dis</th>\n",
       "      <th>mean #data x cell</th>\n",
       "      <th>std #data x cell</th>\n",
       "      <th>top10</th>\n",
       "      <th>Type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>14683</td>\n",
       "      <td>0.999997</td>\n",
       "      <td>0.999949</td>\n",
       "      <td>0.989856</td>\n",
       "      <td>1.129635</td>\n",
       "      <td>0.919626</td>\n",
       "      <td>[72, 30, 19, 17, 16, 15, 15, 15, 15, 14]</td>\n",
       "      <td>VDSH train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>14683</td>\n",
       "      <td>0.999997</td>\n",
       "      <td>0.999991</td>\n",
       "      <td>0.996095</td>\n",
       "      <td>1.063986</td>\n",
       "      <td>0.383207</td>\n",
       "      <td>[12, 11, 11, 10, 8, 7, 7, 7, 7, 6]</td>\n",
       "      <td>B-VAE train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3668</td>\n",
       "      <td>0.999999</td>\n",
       "      <td>0.999926</td>\n",
       "      <td>0.993338</td>\n",
       "      <td>1.076923</td>\n",
       "      <td>0.539635</td>\n",
       "      <td>[20, 12, 8, 8, 7, 7, 7, 5, 5, 5]</td>\n",
       "      <td>VDSH test</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3668</td>\n",
       "      <td>0.999999</td>\n",
       "      <td>0.999991</td>\n",
       "      <td>0.998691</td>\n",
       "      <td>1.021442</td>\n",
       "      <td>0.183827</td>\n",
       "      <td>[5, 5, 4, 4, 3, 3, 3, 3, 3, 2]</td>\n",
       "      <td>B-VAE test</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   #data  #empty cells(%)  Gini cells dis  H cells dis  mean #data x cell  \\\n",
       "0  14683         0.999997        0.999949     0.989856           1.129635   \n",
       "1  14683         0.999997        0.999991     0.996095           1.063986   \n",
       "2   3668         0.999999        0.999926     0.993338           1.076923   \n",
       "3   3668         0.999999        0.999991     0.998691           1.021442   \n",
       "\n",
       "   std #data x cell                                     top10         Type  \n",
       "0          0.919626  [72, 30, 19, 17, 16, 15, 15, 15, 15, 14]   VDSH train  \n",
       "1          0.383207        [12, 11, 11, 10, 8, 7, 7, 7, 7, 6]  B-VAE train  \n",
       "2          0.539635          [20, 12, 8, 8, 7, 7, 7, 5, 5, 5]    VDSH test  \n",
       "3          0.183827            [5, 5, 4, 4, 3, 3, 3, 3, 3, 2]   B-VAE test  "
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#plot together--\n",
    "print(name_dat)\n",
    "compare_cells_dist(train_hash_, train_hash_B, test_hash_, test_hash_B)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reuters\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>#data</th>\n",
       "      <th>#empty cells(%)</th>\n",
       "      <th>Gini cells dis</th>\n",
       "      <th>H cells dis</th>\n",
       "      <th>mean #data x cell</th>\n",
       "      <th>std #data x cell</th>\n",
       "      <th>top10</th>\n",
       "      <th>Type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>8716</td>\n",
       "      <td>0.999998</td>\n",
       "      <td>0.999725</td>\n",
       "      <td>0.971207</td>\n",
       "      <td>1.312651</td>\n",
       "      <td>1.775203</td>\n",
       "      <td>[63, 60, 56, 39, 27, 23, 22, 18, 16, 16]</td>\n",
       "      <td>VDSH train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>8716</td>\n",
       "      <td>0.999998</td>\n",
       "      <td>0.999878</td>\n",
       "      <td>0.981857</td>\n",
       "      <td>1.217488</td>\n",
       "      <td>1.138561</td>\n",
       "      <td>[39, 36, 26, 22, 20, 19, 18, 14, 13, 13]</td>\n",
       "      <td>B-VAE train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>967</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.999445</td>\n",
       "      <td>0.984872</td>\n",
       "      <td>1.136310</td>\n",
       "      <td>0.780331</td>\n",
       "      <td>[15, 12, 8, 6, 5, 5, 4, 4, 4, 3]</td>\n",
       "      <td>VDSH test</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>967</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.999847</td>\n",
       "      <td>0.994713</td>\n",
       "      <td>1.061471</td>\n",
       "      <td>0.395582</td>\n",
       "      <td>[10, 4, 4, 3, 3, 3, 2, 2, 2, 2]</td>\n",
       "      <td>B-VAE test</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   #data  #empty cells(%)  Gini cells dis  H cells dis  mean #data x cell  \\\n",
       "0   8716         0.999998        0.999725     0.971207           1.312651   \n",
       "1   8716         0.999998        0.999878     0.981857           1.217488   \n",
       "2    967         1.000000        0.999445     0.984872           1.136310   \n",
       "3    967         1.000000        0.999847     0.994713           1.061471   \n",
       "\n",
       "   std #data x cell                                     top10         Type  \n",
       "0          1.775203  [63, 60, 56, 39, 27, 23, 22, 18, 16, 16]   VDSH train  \n",
       "1          1.138561  [39, 36, 26, 22, 20, 19, 18, 14, 13, 13]  B-VAE train  \n",
       "2          0.780331          [15, 12, 8, 6, 5, 5, 4, 4, 4, 3]    VDSH test  \n",
       "3          0.395582           [10, 4, 4, 3, 3, 3, 2, 2, 2, 2]   B-VAE test  "
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#plot together--\n",
    "print(name_dat)\n",
    "compare_cells_dist(train_hash_, train_hash_B, test_hash_, test_hash_B)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "TMC\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>#data</th>\n",
       "      <th>#empty cells(%)</th>\n",
       "      <th>Gini cells dis</th>\n",
       "      <th>H cells dis</th>\n",
       "      <th>mean #data x cell</th>\n",
       "      <th>std #data x cell</th>\n",
       "      <th>top10</th>\n",
       "      <th>Type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>24784</td>\n",
       "      <td>0.999995</td>\n",
       "      <td>0.999976</td>\n",
       "      <td>0.987607</td>\n",
       "      <td>1.215021</td>\n",
       "      <td>0.852803</td>\n",
       "      <td>[25, 21, 20, 20, 20, 19, 17, 16, 15, 14]</td>\n",
       "      <td>VDSH train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>24784</td>\n",
       "      <td>0.999995</td>\n",
       "      <td>0.999996</td>\n",
       "      <td>0.997142</td>\n",
       "      <td>1.068690</td>\n",
       "      <td>0.319516</td>\n",
       "      <td>[13, 13, 12, 10, 9, 9, 7, 6, 6, 6]</td>\n",
       "      <td>B-VAE train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3498</td>\n",
       "      <td>0.999999</td>\n",
       "      <td>0.999962</td>\n",
       "      <td>0.995061</td>\n",
       "      <td>1.080964</td>\n",
       "      <td>0.377351</td>\n",
       "      <td>[6, 6, 6, 5, 5, 5, 5, 4, 4, 4]</td>\n",
       "      <td>VDSH test</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3498</td>\n",
       "      <td>0.999999</td>\n",
       "      <td>0.999992</td>\n",
       "      <td>0.998823</td>\n",
       "      <td>1.021612</td>\n",
       "      <td>0.171239</td>\n",
       "      <td>[6, 4, 3, 2, 2, 2, 2, 2, 2, 2]</td>\n",
       "      <td>B-VAE test</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   #data  #empty cells(%)  Gini cells dis  H cells dis  mean #data x cell  \\\n",
       "0  24784         0.999995        0.999976     0.987607           1.215021   \n",
       "1  24784         0.999995        0.999996     0.997142           1.068690   \n",
       "2   3498         0.999999        0.999962     0.995061           1.080964   \n",
       "3   3498         0.999999        0.999992     0.998823           1.021612   \n",
       "\n",
       "   std #data x cell                                     top10         Type  \n",
       "0          0.852803  [25, 21, 20, 20, 20, 19, 17, 16, 15, 14]   VDSH train  \n",
       "1          0.319516        [13, 13, 12, 10, 9, 9, 7, 6, 6, 6]  B-VAE train  \n",
       "2          0.377351            [6, 6, 6, 5, 5, 5, 5, 4, 4, 4]    VDSH test  \n",
       "3          0.171239            [6, 4, 3, 2, 2, 2, 2, 2, 2, 2]   B-VAE test  "
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#plot together--\n",
    "print(name_dat)\n",
    "compare_cells_dist(train_hash_, train_hash_B, test_hash_, test_hash_B)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Snippets\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>#data</th>\n",
       "      <th>#empty cells(%)</th>\n",
       "      <th>Gini cells dis</th>\n",
       "      <th>H cells dis</th>\n",
       "      <th>mean #data x cell</th>\n",
       "      <th>std #data x cell</th>\n",
       "      <th>top10</th>\n",
       "      <th>Type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>10060</td>\n",
       "      <td>0.999998</td>\n",
       "      <td>0.999980</td>\n",
       "      <td>0.994205</td>\n",
       "      <td>1.098373</td>\n",
       "      <td>0.469900</td>\n",
       "      <td>[15, 13, 8, 7, 7, 7, 7, 7, 7, 7]</td>\n",
       "      <td>VDSH train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>10060</td>\n",
       "      <td>0.999998</td>\n",
       "      <td>0.999975</td>\n",
       "      <td>0.993307</td>\n",
       "      <td>1.104887</td>\n",
       "      <td>0.524714</td>\n",
       "      <td>[18, 15, 10, 10, 9, 9, 8, 7, 7, 7]</td>\n",
       "      <td>B-VAE train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2280</td>\n",
       "      <td>0.999999</td>\n",
       "      <td>0.999976</td>\n",
       "      <td>0.997793</td>\n",
       "      <td>1.034952</td>\n",
       "      <td>0.235625</td>\n",
       "      <td>[6, 5, 4, 3, 3, 3, 3, 3, 2, 2]</td>\n",
       "      <td>VDSH test</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2280</td>\n",
       "      <td>0.999999</td>\n",
       "      <td>0.999984</td>\n",
       "      <td>0.998598</td>\n",
       "      <td>1.018767</td>\n",
       "      <td>0.190493</td>\n",
       "      <td>[5, 5, 4, 4, 3, 3, 2, 2, 2, 2]</td>\n",
       "      <td>B-VAE test</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   #data  #empty cells(%)  Gini cells dis  H cells dis  mean #data x cell  \\\n",
       "0  10060         0.999998        0.999980     0.994205           1.098373   \n",
       "1  10060         0.999998        0.999975     0.993307           1.104887   \n",
       "2   2280         0.999999        0.999976     0.997793           1.034952   \n",
       "3   2280         0.999999        0.999984     0.998598           1.018767   \n",
       "\n",
       "   std #data x cell                               top10         Type  \n",
       "0          0.469900    [15, 13, 8, 7, 7, 7, 7, 7, 7, 7]   VDSH train  \n",
       "1          0.524714  [18, 15, 10, 10, 9, 9, 8, 7, 7, 7]  B-VAE train  \n",
       "2          0.235625      [6, 5, 4, 3, 3, 3, 3, 3, 2, 2]    VDSH test  \n",
       "3          0.190493      [5, 5, 4, 4, 3, 3, 2, 2, 2, 2]   B-VAE test  "
      ]
     },
     "execution_count": 76,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#plot together--\n",
    "print(name_dat)\n",
    "compare_cells_dist(train_hash_, train_hash_B, test_hash_, test_hash_B)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MNIST\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>#data</th>\n",
       "      <th>#empty cells(%)</th>\n",
       "      <th>Gini cells dis</th>\n",
       "      <th>H cells dis</th>\n",
       "      <th>mean #data x cell</th>\n",
       "      <th>std #data x cell</th>\n",
       "      <th>top10</th>\n",
       "      <th>Type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>69000</td>\n",
       "      <td>0.999985</td>\n",
       "      <td>0.999993</td>\n",
       "      <td>0.993477</td>\n",
       "      <td>1.092256</td>\n",
       "      <td>0.708676</td>\n",
       "      <td>[63, 61, 42, 35, 30, 28, 25, 24, 22, 21]</td>\n",
       "      <td>VDSH train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>69000</td>\n",
       "      <td>0.999989</td>\n",
       "      <td>0.999860</td>\n",
       "      <td>0.952822</td>\n",
       "      <td>1.490442</td>\n",
       "      <td>3.793286</td>\n",
       "      <td>[244, 233, 227, 200, 158, 145, 131, 126, 124, ...</td>\n",
       "      <td>B-VAE train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.999996</td>\n",
       "      <td>0.999777</td>\n",
       "      <td>1.004016</td>\n",
       "      <td>0.063245</td>\n",
       "      <td>[2, 2, 2, 2, 1, 1, 1, 1, 1, 1]</td>\n",
       "      <td>VDSH test</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.999888</td>\n",
       "      <td>0.995497</td>\n",
       "      <td>1.050420</td>\n",
       "      <td>0.342353</td>\n",
       "      <td>[7, 5, 4, 4, 4, 3, 3, 3, 2, 2]</td>\n",
       "      <td>B-VAE test</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   #data  #empty cells(%)  Gini cells dis  H cells dis  mean #data x cell  \\\n",
       "0  69000         0.999985        0.999993     0.993477           1.092256   \n",
       "1  69000         0.999989        0.999860     0.952822           1.490442   \n",
       "2   1000         1.000000        0.999996     0.999777           1.004016   \n",
       "3   1000         1.000000        0.999888     0.995497           1.050420   \n",
       "\n",
       "   std #data x cell                                              top10  \\\n",
       "0          0.708676           [63, 61, 42, 35, 30, 28, 25, 24, 22, 21]   \n",
       "1          3.793286  [244, 233, 227, 200, 158, 145, 131, 126, 124, ...   \n",
       "2          0.063245                     [2, 2, 2, 2, 1, 1, 1, 1, 1, 1]   \n",
       "3          0.342353                     [7, 5, 4, 4, 4, 3, 3, 3, 2, 2]   \n",
       "\n",
       "          Type  \n",
       "0   VDSH train  \n",
       "1  B-VAE train  \n",
       "2    VDSH test  \n",
       "3   B-VAE test  "
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#plot together--\n",
    "print(name_dat)\n",
    "compare_cells_dist(train_hash_, train_hash_B, test_hash_, test_hash_B)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CIFAR-10\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>#data</th>\n",
       "      <th>#empty cells(%)</th>\n",
       "      <th>Gini cells dis</th>\n",
       "      <th>H cells dis</th>\n",
       "      <th>mean #data x cell</th>\n",
       "      <th>std #data x cell</th>\n",
       "      <th>top10</th>\n",
       "      <th>Type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>58101</td>\n",
       "      <td>0.999987</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.999457</td>\n",
       "      <td>1.010294</td>\n",
       "      <td>0.145406</td>\n",
       "      <td>[11, 9, 8, 8, 8, 8, 8, 6, 5, 5]</td>\n",
       "      <td>VDSH train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>58101</td>\n",
       "      <td>0.999987</td>\n",
       "      <td>0.999994</td>\n",
       "      <td>0.996104</td>\n",
       "      <td>1.037370</td>\n",
       "      <td>0.622376</td>\n",
       "      <td>[63, 63, 46, 30, 29, 28, 26, 22, 21, 21]</td>\n",
       "      <td>B-VAE train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1899</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.999999</td>\n",
       "      <td>0.999946</td>\n",
       "      <td>1.001054</td>\n",
       "      <td>0.032453</td>\n",
       "      <td>[2, 2, 1, 1, 1, 1, 1, 1, 1, 1]</td>\n",
       "      <td>VDSH test</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1899</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.999994</td>\n",
       "      <td>0.999584</td>\n",
       "      <td>1.004762</td>\n",
       "      <td>0.105302</td>\n",
       "      <td>[5, 2, 2, 2, 2, 2, 1, 1, 1, 1]</td>\n",
       "      <td>B-VAE test</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   #data  #empty cells(%)  Gini cells dis  H cells dis  mean #data x cell  \\\n",
       "0  58101         0.999987        1.000000     0.999457           1.010294   \n",
       "1  58101         0.999987        0.999994     0.996104           1.037370   \n",
       "2   1899         1.000000        0.999999     0.999946           1.001054   \n",
       "3   1899         1.000000        0.999994     0.999584           1.004762   \n",
       "\n",
       "   std #data x cell                                     top10         Type  \n",
       "0          0.145406           [11, 9, 8, 8, 8, 8, 8, 6, 5, 5]   VDSH train  \n",
       "1          0.622376  [63, 63, 46, 30, 29, 28, 26, 22, 21, 21]  B-VAE train  \n",
       "2          0.032453            [2, 2, 1, 1, 1, 1, 1, 1, 1, 1]    VDSH test  \n",
       "3          0.105302            [5, 2, 2, 2, 2, 2, 1, 1, 1, 1]   B-VAE test  "
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#plot together--\n",
    "print(name_dat)\n",
    "compare_cells_dist(train_hash_, train_hash_B, test_hash_, test_hash_B)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Nus-Wide\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>#data</th>\n",
       "      <th>#empty cells(%)</th>\n",
       "      <th>Gini cells dis</th>\n",
       "      <th>H cells dis</th>\n",
       "      <th>mean #data x cell</th>\n",
       "      <th>std #data x cell</th>\n",
       "      <th>top10</th>\n",
       "      <th>Type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>156283</td>\n",
       "      <td>0.999966</td>\n",
       "      <td>0.999996</td>\n",
       "      <td>0.994484</td>\n",
       "      <td>1.062109</td>\n",
       "      <td>0.863122</td>\n",
       "      <td>[152, 116, 104, 76, 56, 52, 51, 40, 38, 37]</td>\n",
       "      <td>VDSH train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>156283</td>\n",
       "      <td>0.999972</td>\n",
       "      <td>0.999250</td>\n",
       "      <td>0.939773</td>\n",
       "      <td>1.317321</td>\n",
       "      <td>12.428037</td>\n",
       "      <td>[3079, 1742, 1457, 1180, 737, 362, 340, 328, 3...</td>\n",
       "      <td>B-VAE train</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2100</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.999990</td>\n",
       "      <td>0.999069</td>\n",
       "      <td>1.015965</td>\n",
       "      <td>0.143346</td>\n",
       "      <td>[4, 3, 3, 2, 2, 2, 2, 2, 2, 2]</td>\n",
       "      <td>VDSH test</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2100</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.997416</td>\n",
       "      <td>0.952817</td>\n",
       "      <td>1.228789</td>\n",
       "      <td>2.581333</td>\n",
       "      <td>[83, 40, 37, 28, 18, 16, 13, 10, 9, 6]</td>\n",
       "      <td>B-VAE test</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    #data  #empty cells(%)  Gini cells dis  H cells dis  mean #data x cell  \\\n",
       "0  156283         0.999966        0.999996     0.994484           1.062109   \n",
       "1  156283         0.999972        0.999250     0.939773           1.317321   \n",
       "2    2100         1.000000        0.999990     0.999069           1.015965   \n",
       "3    2100         1.000000        0.997416     0.952817           1.228789   \n",
       "\n",
       "   std #data x cell                                              top10  \\\n",
       "0          0.863122        [152, 116, 104, 76, 56, 52, 51, 40, 38, 37]   \n",
       "1         12.428037  [3079, 1742, 1457, 1180, 737, 362, 340, 328, 3...   \n",
       "2          0.143346                     [4, 3, 3, 2, 2, 2, 2, 2, 2, 2]   \n",
       "3          2.581333             [83, 40, 37, 28, 18, 16, 13, 10, 9, 6]   \n",
       "\n",
       "          Type  \n",
       "0   VDSH train  \n",
       "1  B-VAE train  \n",
       "2    VDSH test  \n",
       "3   B-VAE test  "
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#plot together--\n",
    "print(name_dat)\n",
    "compare_cells_dist(train_hash_, train_hash_B, test_hash_, test_hash_B)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Entrenamiento----\n",
      "Cantidad de datos a llenar la tabla hash:  14683\n",
      "Cantidad de memorias ocupadas hash1:  13800\n",
      "Cantidad de memorias ocupadas hash2:  12998\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAzIAAAD8CAYAAACsP5F0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8ba31d8588>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Pruebas-----\n",
      "Cantidad de datos a llenar la tabla hash:  3668\n",
      "Cantidad de memorias ocupadas hash1:  3591\n",
      "Cantidad de memorias ocupadas hash2:  3406\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA3QAAAD8CAYAAAA/m+aTAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f8a9c036b70>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "compare_cells_plot(BITS_S,train_hash_B, train_hash_, test_hash_B, test_hash_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python [conda env:tesis]",
   "language": "python",
   "name": "conda-env-tesis-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}