Adding hyperopt notebook

yngtodd · Mar 11, 2019 · 4c52d73 · 4c52d73
1 parent f1e8321
commit 4c52d73
Show file tree

Hide file tree

Showing 2 changed files with 564 additions and 0 deletions.
diff --git a/notebooks/.ipynb_checkpoints/hyperopt-checkpoint.ipynb b/notebooks/.ipynb_checkpoints/hyperopt-checkpoint.ipynb
@@ -0,0 +1,282 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 33,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import time \n",
+    "import pprint\n",
+    "import pickle\n",
+    "import numpy as np\n",
+    "\n",
+    "from hyperopt import fmin, tpe, hp, STATUS_OK, Trials\n",
+    "from hyperspace.benchmarks import StyblinskiTang\n",
+    "from skopt import gp_minimize\n",
+    "\n",
+    "from scipy.optimize import OptimizeResult\n",
+    "\n",
+    "\n",
+    "pp = pprint.PrettyPrinter(indent=4)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 31,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "100%|██████████| 100/100 [00:00<00:00, 312.22it/s, best loss: -77.18972901241136]\n",
+      "{'0': -2.729885415079413, '1': -2.702710636326808}\n"
+     ]
+    }
+   ],
+   "source": [
+    "objective = StyblinskiTang(dims=2)\n",
+    "\n",
+    "trials = Trials()\n",
+    "\n",
+    "space = [\n",
+    "    hp.uniform('0', -5, 5),\n",
+    "    hp.uniform('1', -5, 5),\n",
+    "]\n",
+    "\n",
+    "best = fmin(objective,\n",
+    "    space=space,\n",
+    "    algo=tpe.suggest,\n",
+    "    trials=trials,\n",
+    "    max_evals=100)\n",
+    "\n",
+    "print(best)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "{   'book_time': datetime.datetime(2019, 3, 11, 17, 56, 33, 649000),\n",
+      "    'exp_key': None,\n",
+      "    'misc': {   'cmd': ('domain_attachment', 'FMinIter_Domain'),\n",
+      "                'idxs': {'x': [0], 'y': [0], 'z': [0]},\n",
+      "                'tid': 0,\n",
+      "                'vals': {   'x': [1.9185834595646973],\n",
+      "                            'y': [-7.441043187200959],\n",
+      "                            'z': [-7.64457512965655]},\n",
+      "                'workdir': None},\n",
+      "    'owner': None,\n",
+      "    'refresh_time': datetime.datetime(2019, 3, 11, 17, 56, 33, 650000),\n",
+      "    'result': {'loss': 2274.399434544481, 'status': 'ok'},\n",
+      "    'spec': None,\n",
+      "    'state': 2,\n",
+      "    'tid': 0,\n",
+      "    'version': 0}\n"
+     ]
+    }
+   ],
+   "source": [
+    "pp.pprint(trials.trials[0])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "100"
+      ]
+     },
+     "execution_count": 11,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "len(trials.trials)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "for i in range(10):\n",
+    "    print(f'\\nIteration {i}')\n",
+    "    pp.pprint(trials.trials[i])"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Comparison with Bayesian Optimization"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 27,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "bounds = np.tile((-5., 5.), (2, 1))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 28,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "res = gp_minimize(objective, bounds, n_calls=100)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 29,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "-181.64915626952801342"
+      ]
+     },
+     "execution_count": 29,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "res.fun"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Create better format for results"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 64,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def create_result(trials, maximize=False):\n",
+    "    \"\"\"Create a Scipy OptimizeResult object for hyperopt.\"\"\"\n",
+    "    res = OptimizeResult()\n",
+    "    \n",
+    "    x_iters = []\n",
+    "    func_vals = []\n",
+    "    for i in range(len(trials)):\n",
+    "        trial = trials.trials[i]\n",
+    "        x_iters.append(trial['misc']['vals'])\n",
+    "        func_vals.append(trial['result']['loss'])\n",
+    "    \n",
+    "    res.x_iters = x_iters\n",
+    "    res.func_vals = func_vals\n",
+    "    \n",
+    "    if maximize:\n",
+    "        res.fun = np.max(func_vals)\n",
+    "        max_idx = np.argmax(func_vals)\n",
+    "        res.x = res.x_iters[max_idx]\n",
+    "    else:\n",
+    "        res.fun = np.min(func_vals)\n",
+    "        min_idx = np.argmin(func_vals)\n",
+    "        res.x = res.x_iters[min_idx]\n",
+    "        \n",
+    "    return res"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 71,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "res = create_result(trials, maximize=False)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 72,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "-77.18972901241136"
+      ]
+     },
+     "execution_count": 72,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "res.fun"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 73,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "{'0': [-2.729885415079413], '1': [-2.702710636326808]}"
+      ]
+     },
+     "execution_count": 73,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "res.x"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.6.5"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}