main.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "c313e40b-44c2-4927-b17c-d53f1328420d",
   "metadata": {},
   "source": [
    "## Rand_d-Rand_HB: a mockup version of BOHB\n",
    "\n",
    "This program implements a mockup version of BOHB. Features:\n",
    "1. num of parameters:  hyperparameters (>=4), in the continuous range [0,1]\n",
    "2. switch the BO role by a random samples with euclidian distance > \"d\"\n",
    "3. randomly cut samples during SH, no evaluation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "8bce3607-8e5c-4d11-a7d7-8cfd064b1731",
   "metadata": {},
   "outputs": [],
   "source": [
    "# imports\n",
    "import math\n",
    "from rHB import RandomHB"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "87b8e659-fc2f-4736-bb02-6552ab09c00e",
   "metadata": {},
   "outputs": [],
   "source": [
    "# algorithm for search -> in this program the budget is an integer all along for simplification\n",
    "\n",
    "sample_size = 8     # number of initial samples for SH\n",
    "halving_factor = 2  # rate = 1/halving_factor \n",
    "total_budget = 4*80   # amount of \"computing\" resourses, iterations, etc\n",
    "min_budget = 1  # minimal amount of resources\n",
    "max_budget = total_budget \n",
    "\n",
    "hb_halving_factor = 2 # controll the amount of initial samples for each hyperband iteration\n",
    "\n",
    "# model sampling parameters\n",
    "num_hyperparameters = 2 # number of hyperparameters\n",
    "min_sampling_distance = 0.00001 # minimal distance between any two randomly selected points "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "2a4c1193-26bf-487f-9ba0-b07aa7c7c903",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "\n",
      " HB step **  0  **      Step budget: 80     Number of samples: 8\n",
      "    number of rounds:  4\n",
      "    budget per round:  20\n",
      "    sample budget per round:  [2, 5, 10, 20]\n",
      "    round: 0     indexes: [0, 1, 2, 3, 4, 5, 6, 7]\n",
      "        >>>> Start iterations for Object:  0\n",
      "        iteration  0 1     <<<<\n",
      "        >>>> Start iterations for Object:  1\n",
      "        iteration  0 1     <<<<\n",
      "        >>>> Start iterations for Object:  2\n",
      "        iteration  0 1     <<<<\n",
      "        >>>> Start iterations for Object:  3\n",
      "        iteration  0 1     <<<<\n",
      "        >>>> Start iterations for Object:  4\n",
      "        iteration  0 1     <<<<\n",
      "        >>>> Start iterations for Object:  5\n",
      "        iteration  0 1     <<<<\n",
      "        >>>> Start iterations for Object:  6\n",
      "        iteration  0 1     <<<<\n",
      "        >>>> Start iterations for Object:  7\n",
      "        iteration  0 1     <<<<\n",
      "    round: 1     indexes: [0, 3, 5, 7]\n",
      "        >>>> Start iterations for Object:  0\n",
      "        iteration  0 1 2 3 4     <<<<\n",
      "        >>>> Start iterations for Object:  3\n",
      "        iteration  0 1 2 3 4     <<<<\n",
      "        >>>> Start iterations for Object:  5\n",
      "        iteration  0 1 2 3 4     <<<<\n",
      "        >>>> Start iterations for Object:  7\n",
      "        iteration  0 1 2 3 4     <<<<\n",
      "    round: 2     indexes: [0, 3]\n",
      "        >>>> Start iterations for Object:  0\n",
      "        iteration  0 1 2 3 4 5 6 7 8 9     <<<<\n",
      "        >>>> Start iterations for Object:  3\n",
      "        iteration  0 1 2 3 4 5 6 7 8 9     <<<<\n",
      "    round: 3     indexes: [3]\n",
      "        >>>> Start iterations for Object:  3\n",
      "        iteration  0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19     <<<<\n",
      "step: 0  results:  [[0.10130168350611268, 0.5609032234744546, 0.11043711556125725], [0.7249915141211845, 0.8071886397277713, 1.2410659246176543], [0.6490052290403441, 0.6412649034544097, 1.0451188073207507], [0.4703499080498196, 0.4122851527511666, 0.017895919771886686], [0.9702645595487708, 0.26333695994638806, 0.9992172307910788], [0.680814704457053, 0.8009376461156598, 0.7087175558466419], [0.47420538451049543, 0.3717941167014659, 0.6852595959816887], [0.48279890630450795, 0.1933230507537962, 0.32338703608292]]\n",
      "\n",
      "\n",
      " HB step **  1  **      Step budget: 80     Number of samples: 4\n",
      "    number of rounds:  3\n",
      "    budget per round:  26\n",
      "    sample budget per round:  [6, 13, 26]\n",
      "    round: 0     indexes: [0, 1, 2, 3]\n",
      "        >>>> Start iterations for Object:  0\n",
      "        iteration  0 1 2 3 4 5     <<<<\n",
      "        >>>> Start iterations for Object:  1\n",
      "        iteration  0 1 2 3 4 5     <<<<\n",
      "        >>>> Start iterations for Object:  2\n",
      "        iteration  0 1 2 3 4 5     <<<<\n",
      "        >>>> Start iterations for Object:  3\n",
      "        iteration  0 1 2 3 4 5     <<<<\n",
      "    round: 1     indexes: [0, 3]\n",
      "        >>>> Start iterations for Object:  0\n",
      "        iteration  0 1 2 3 4 5 6 7 8 9 10 11 12     <<<<\n",
      "        >>>> Start iterations for Object:  3\n",
      "        iteration  0 1 2 3 4 5 6 7 8 9 10 11 12     <<<<\n",
      "    round: 2     indexes: [3]\n",
      "        >>>> Start iterations for Object:  3\n",
      "        iteration  0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25     <<<<\n",
      "step: 1  results:  [[0.10130168350611268, 0.5609032234744546, 0.11043711556125725], [0.7249915141211845, 0.8071886397277713, 1.2410659246176543], [0.6490052290403441, 0.6412649034544097, 1.0451188073207507], [0.4703499080498196, 0.4122851527511666, 0.017895919771886686], [0.9702645595487708, 0.26333695994638806, 0.9992172307910788], [0.680814704457053, 0.8009376461156598, 0.7087175558466419], [0.47420538451049543, 0.3717941167014659, 0.6852595959816887], [0.48279890630450795, 0.1933230507537962, 0.32338703608292], [0.849272609483536, 0.5219958575782, 0.18523803641211575], [0.6177184126045602, 0.611083563490417, 0.6530357509778908], [0.9950303558933128, 0.036716832652060716, 0.5483127576277419], [0.7097071051522538, 0.654156736406843, 0.011903753920959964]]\n",
      "\n",
      "\n",
      " HB step **  2  **      Step budget: 80     Number of samples: 2\n",
      "    number of rounds:  2\n",
      "    budget per round:  40\n",
      "    sample budget per round:  [20, 40]\n",
      "    round: 0     indexes: [0, 1]\n",
      "        >>>> Start iterations for Object:  0\n",
      "        iteration  0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19     <<<<\n",
      "        >>>> Start iterations for Object:  1\n",
      "        iteration  0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19     <<<<\n",
      "    round: 1     indexes: [0]\n",
      "        >>>> Start iterations for Object:  0\n",
      "        iteration  0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39     <<<<\n",
      "step: 2  results:  [[0.10130168350611268, 0.5609032234744546, 0.11043711556125725], [0.7249915141211845, 0.8071886397277713, 1.2410659246176543], [0.6490052290403441, 0.6412649034544097, 1.0451188073207507], [0.4703499080498196, 0.4122851527511666, 0.017895919771886686], [0.9702645595487708, 0.26333695994638806, 0.9992172307910788], [0.680814704457053, 0.8009376461156598, 0.7087175558466419], [0.47420538451049543, 0.3717941167014659, 0.6852595959816887], [0.48279890630450795, 0.1933230507537962, 0.32338703608292], [0.849272609483536, 0.5219958575782, 0.18523803641211575], [0.6177184126045602, 0.611083563490417, 0.6530357509778908], [0.9950303558933128, 0.036716832652060716, 0.5483127576277419], [0.7097071051522538, 0.654156736406843, 0.011903753920959964], [0.871489896266444, 0.01387848327166108, 0.0015910160972485274], [0.9184845531189675, 0.053176520176408504, 0.11813130268713383]]\n",
      "\n",
      "\n",
      " HB step **  3  **      Step budget: 80     Number of samples: 1\n",
      "    number of rounds:  1\n",
      "    budget per round:  80\n",
      "    sample budget per round:  [80]\n",
      "    round: 0     indexes: [0]\n",
      "        >>>> Start iterations for Object:  0\n",
      "        iteration  0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79     <<<<\n",
      "step: 3  results:  [[0.10130168350611268, 0.5609032234744546, 0.11043711556125725], [0.7249915141211845, 0.8071886397277713, 1.2410659246176543], [0.6490052290403441, 0.6412649034544097, 1.0451188073207507], [0.4703499080498196, 0.4122851527511666, 0.017895919771886686], [0.9702645595487708, 0.26333695994638806, 0.9992172307910788], [0.680814704457053, 0.8009376461156598, 0.7087175558466419], [0.47420538451049543, 0.3717941167014659, 0.6852595959816887], [0.48279890630450795, 0.1933230507537962, 0.32338703608292], [0.849272609483536, 0.5219958575782, 0.18523803641211575], [0.6177184126045602, 0.611083563490417, 0.6530357509778908], [0.9950303558933128, 0.036716832652060716, 0.5483127576277419], [0.7097071051522538, 0.654156736406843, 0.011903753920959964], [0.871489896266444, 0.01387848327166108, 0.0015910160972485274], [0.9184845531189675, 0.053176520176408504, 0.11813130268713383], [0.5435909914609759, 0.4674412313611621, 0.0002208847598991772]]\n"
     ]
    }
   ],
   "source": [
    "experiment = RandomHB(sample_size, halving_factor, total_budget, min_budget, max_budget, hb_halving_factor, num_hyperparameters, min_sampling_distance)\n",
    "experiment.start()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "fad78246-0eb0-407f-8711-be7314334918",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[[0.10130168350611268, 0.5609032234744546, 0.11043711556125725],\n",
       " [0.7249915141211845, 0.8071886397277713, 1.2410659246176543],\n",
       " [0.6490052290403441, 0.6412649034544097, 1.0451188073207507],\n",
       " [0.4703499080498196, 0.4122851527511666, 0.017895919771886686],\n",
       " [0.9702645595487708, 0.26333695994638806, 0.9992172307910788],\n",
       " [0.680814704457053, 0.8009376461156598, 0.7087175558466419],\n",
       " [0.47420538451049543, 0.3717941167014659, 0.6852595959816887],\n",
       " [0.48279890630450795, 0.1933230507537962, 0.32338703608292],\n",
       " [0.849272609483536, 0.5219958575782, 0.18523803641211575],\n",
       " [0.6177184126045602, 0.611083563490417, 0.6530357509778908],\n",
       " [0.9950303558933128, 0.036716832652060716, 0.5483127576277419],\n",
       " [0.7097071051522538, 0.654156736406843, 0.011903753920959964],\n",
       " [0.871489896266444, 0.01387848327166108, 0.0015910160972485274],\n",
       " [0.9184845531189675, 0.053176520176408504, 0.11813130268713383],\n",
       " [0.5435909914609759, 0.4674412313611621, 0.0002208847598991772]]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "experiment.pull_results()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "498cd901-7ca2-426b-a12d-bb94bfe918a8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0.5435909914609759, 0.4674412313611621, 0.0002208847598991772]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "experiment.pull_best()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}