test_override_clustering_index_binary.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import faiss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "8"
      ]
     },
     "execution_count": 2,
     "metadata": {
      "bento_obj_id": "139916685704704"
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "faiss.get_num_gpus()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "d = 256\n",
    "quantizer = faiss.IndexBinaryFlat(d)\n",
    "index = faiss.IndexBinaryIVF(quantizer, d, 4096)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "xt = faiss.randint((100000, 256 // 8)).astype('uint8')\n",
    "xt.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 1min 35s, sys: 15.8 s, total: 1min 51s\n",
      "Wall time: 11.2 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "index.train(xt)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "quantizer2 = faiss.IndexBinaryFlat(d)\n",
    "index2 = faiss.IndexBinaryIVF(quantizer2, d, 4096)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "clustering_index = faiss.index_cpu_to_all_gpus(faiss.IndexFlatL2(d))\n",
    "index2.clustering_index = clustering_index"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 1.33 s, sys: 697 ms, total: 2.03 s\n",
      "Wall time: 690 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "index2.train(xt)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "bento_stylesheets": {
   "bento/extensions/flow/main.css": true,
   "bento/extensions/kernel_selector/main.css": true,
   "bento/extensions/kernel_ui/main.css": true,
   "bento/extensions/new_kernel/main.css": true,
   "bento/extensions/system_usage/main.css": true,
   "bento/extensions/theme/main.css": true
  },
  "kernelspec": {
   "display_name": "pytorch",
   "language": "python",
   "name": "bento_kernel_pytorch"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.3rc1+"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}

Very nice of you !

BTW: provided trained index, would you recommend to cpu or gpu for search?