{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Reading results .csv"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First, we open the csv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = pd.read_csv(\"./experiment_stages.csv\", index_col=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "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>arm_name</th>\n",
       "      <th>metric_name</th>\n",
       "      <th>mean</th>\n",
       "      <th>sem</th>\n",
       "      <th>trial_index</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0_0</td>\n",
       "      <td>accuracy</td>\n",
       "      <td>0.795562</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0_0</td>\n",
       "      <td>weight</td>\n",
       "      <td>0.701442</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0_0</td>\n",
       "      <td>ram</td>\n",
       "      <td>0.703689</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0_0</td>\n",
       "      <td>latency</td>\n",
       "      <td>0.600640</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1_0</td>\n",
       "      <td>accuracy</td>\n",
       "      <td>0.706125</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3515</th>\n",
       "      <td>898_0</td>\n",
       "      <td>latency</td>\n",
       "      <td>0.327915</td>\n",
       "      <td>0.0</td>\n",
       "      <td>898</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3516</th>\n",
       "      <td>899_0</td>\n",
       "      <td>accuracy</td>\n",
       "      <td>0.598378</td>\n",
       "      <td>0.0</td>\n",
       "      <td>899</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3517</th>\n",
       "      <td>899_0</td>\n",
       "      <td>weight</td>\n",
       "      <td>0.581171</td>\n",
       "      <td>0.0</td>\n",
       "      <td>899</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3518</th>\n",
       "      <td>899_0</td>\n",
       "      <td>ram</td>\n",
       "      <td>0.546563</td>\n",
       "      <td>0.0</td>\n",
       "      <td>899</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3519</th>\n",
       "      <td>899_0</td>\n",
       "      <td>latency</td>\n",
       "      <td>0.335387</td>\n",
       "      <td>0.0</td>\n",
       "      <td>899</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>3520 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "     arm_name metric_name      mean  sem  trial_index\n",
       "0         0_0    accuracy  0.795562  0.0            0\n",
       "1         0_0      weight  0.701442  0.0            0\n",
       "2         0_0         ram  0.703689  0.0            0\n",
       "3         0_0     latency  0.600640  0.0            0\n",
       "4         1_0    accuracy  0.706125  0.0            1\n",
       "...       ...         ...       ...  ...          ...\n",
       "3515    898_0     latency  0.327915  0.0          898\n",
       "3516    899_0    accuracy  0.598378  0.0          899\n",
       "3517    899_0      weight  0.581171  0.0          899\n",
       "3518    899_0         ram  0.546563  0.0          899\n",
       "3519    899_0     latency  0.335387  0.0          899\n",
       "\n",
       "[3520 rows x 5 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Reoder by arm"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We need to reorder it to be able to group different performances in one row and by arm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_use = data.pivot(columns='metric_name', values='mean',\n",
    "                         index='arm_name')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "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>metric_name</th>\n",
       "      <th>accuracy</th>\n",
       "      <th>latency</th>\n",
       "      <th>ram</th>\n",
       "      <th>weight</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>arm_name</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0_0</th>\n",
       "      <td>0.795562</td>\n",
       "      <td>0.600640</td>\n",
       "      <td>0.703689</td>\n",
       "      <td>0.701442</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100_0</th>\n",
       "      <td>0.661357</td>\n",
       "      <td>0.452087</td>\n",
       "      <td>0.556115</td>\n",
       "      <td>0.591458</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>101_0</th>\n",
       "      <td>0.754984</td>\n",
       "      <td>0.572030</td>\n",
       "      <td>0.680436</td>\n",
       "      <td>0.683750</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>102_0</th>\n",
       "      <td>0.759577</td>\n",
       "      <td>0.667570</td>\n",
       "      <td>0.684806</td>\n",
       "      <td>0.715898</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>103_0</th>\n",
       "      <td>0.757547</td>\n",
       "      <td>0.528259</td>\n",
       "      <td>0.619019</td>\n",
       "      <td>0.657079</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>96_0</th>\n",
       "      <td>0.790015</td>\n",
       "      <td>0.639732</td>\n",
       "      <td>0.720727</td>\n",
       "      <td>0.722822</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>97_0</th>\n",
       "      <td>0.767722</td>\n",
       "      <td>0.732089</td>\n",
       "      <td>0.689055</td>\n",
       "      <td>0.712722</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>98_0</th>\n",
       "      <td>0.722385</td>\n",
       "      <td>0.493155</td>\n",
       "      <td>0.607247</td>\n",
       "      <td>0.639543</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99_0</th>\n",
       "      <td>0.758130</td>\n",
       "      <td>0.652381</td>\n",
       "      <td>0.708979</td>\n",
       "      <td>0.715937</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9_0</th>\n",
       "      <td>0.838108</td>\n",
       "      <td>0.573182</td>\n",
       "      <td>0.686754</td>\n",
       "      <td>0.689862</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>880 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "metric_name  accuracy   latency       ram    weight\n",
       "arm_name                                           \n",
       "0_0          0.795562  0.600640  0.703689  0.701442\n",
       "100_0        0.661357  0.452087  0.556115  0.591458\n",
       "101_0        0.754984  0.572030  0.680436  0.683750\n",
       "102_0        0.759577  0.667570  0.684806  0.715898\n",
       "103_0        0.757547  0.528259  0.619019  0.657079\n",
       "...               ...       ...       ...       ...\n",
       "96_0         0.790015  0.639732  0.720727  0.722822\n",
       "97_0         0.767722  0.732089  0.689055  0.712722\n",
       "98_0         0.722385  0.493155  0.607247  0.639543\n",
       "99_0         0.758130  0.652381  0.708979  0.715937\n",
       "9_0          0.838108  0.573182  0.686754  0.689862\n",
       "\n",
       "[880 rows x 4 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_use"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## DeStandardize"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As we can see results are grouped by arm but we have a problem, they are still in relative values. SpArSeMoD uses sntardardizers internally to place the performance metrics in the same range (approx [0, 1]). To ungroup them we must know the standardization procedure, which is \n",
    "\n",
    "$$\n",
    "standardized value = \\frac{\\log(real value)}{\\log(standardization value)}\n",
    "$$\n",
    "\n",
    "where the standardization value depends on the metric.\n",
    "\n",
    "+ Model Size: $10^8$\n",
    "+ Working memory $10^8$\n",
    "+ Latency $10^4$\n",
    "\n",
    "Hence to recover the original value we have to isolate the real value in the last equation and we will obtain real values for all these metrics. Accuracy has no standardization. With this we will recover the number of parameters.\n",
    "\n",
    "Then depending that we consider if 8 bit data types or 32-float, we will use a different conversion.\n",
    "\n",
    "In the case or model size ('weight') and working memory ('ram'), we can proceed to suppose that they are 8-bit parameters and convert them to kilobytes, by dividing by 1024.\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_use['real_weight'] = np.exp(data_use['weight']*np.log(10**8))*8/8/1024\n",
    "data_use['real_ram'] = np.exp(data_use['ram']*np.log(10**8))*8/8/1024\n",
    "data_use['real_latency'] = np.exp(data_use['latency']*np.log(10**4) + 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "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>metric_name</th>\n",
       "      <th>accuracy</th>\n",
       "      <th>latency</th>\n",
       "      <th>ram</th>\n",
       "      <th>weight</th>\n",
       "      <th>real_weight</th>\n",
       "      <th>real_ram</th>\n",
       "      <th>real_latency</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>arm_name</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0_0</th>\n",
       "      <td>0.795562</td>\n",
       "      <td>0.600640</td>\n",
       "      <td>0.703689</td>\n",
       "      <td>0.701442</td>\n",
       "      <td>399.244141</td>\n",
       "      <td>416.117188</td>\n",
       "      <td>686.835201</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100_0</th>\n",
       "      <td>0.661357</td>\n",
       "      <td>0.452087</td>\n",
       "      <td>0.556115</td>\n",
       "      <td>0.591458</td>\n",
       "      <td>52.645508</td>\n",
       "      <td>27.455078</td>\n",
       "      <td>174.840703</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>101_0</th>\n",
       "      <td>0.754984</td>\n",
       "      <td>0.572030</td>\n",
       "      <td>0.680436</td>\n",
       "      <td>0.683750</td>\n",
       "      <td>288.205078</td>\n",
       "      <td>271.136719</td>\n",
       "      <td>527.732875</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>102_0</th>\n",
       "      <td>0.759577</td>\n",
       "      <td>0.667570</td>\n",
       "      <td>0.684806</td>\n",
       "      <td>0.715898</td>\n",
       "      <td>521.052734</td>\n",
       "      <td>293.867187</td>\n",
       "      <td>1272.259462</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>103_0</th>\n",
       "      <td>0.757547</td>\n",
       "      <td>0.528259</td>\n",
       "      <td>0.619019</td>\n",
       "      <td>0.657079</td>\n",
       "      <td>176.331055</td>\n",
       "      <td>87.467773</td>\n",
       "      <td>352.640357</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>96_0</th>\n",
       "      <td>0.790015</td>\n",
       "      <td>0.639732</td>\n",
       "      <td>0.720727</td>\n",
       "      <td>0.722822</td>\n",
       "      <td>591.936523</td>\n",
       "      <td>569.524414</td>\n",
       "      <td>984.514618</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>97_0</th>\n",
       "      <td>0.767722</td>\n",
       "      <td>0.732089</td>\n",
       "      <td>0.689055</td>\n",
       "      <td>0.712722</td>\n",
       "      <td>491.443359</td>\n",
       "      <td>317.789063</td>\n",
       "      <td>2304.890761</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>98_0</th>\n",
       "      <td>0.722385</td>\n",
       "      <td>0.493155</td>\n",
       "      <td>0.607247</td>\n",
       "      <td>0.639543</td>\n",
       "      <td>127.656250</td>\n",
       "      <td>70.416992</td>\n",
       "      <td>255.220262</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99_0</th>\n",
       "      <td>0.758130</td>\n",
       "      <td>0.652381</td>\n",
       "      <td>0.708979</td>\n",
       "      <td>0.715937</td>\n",
       "      <td>521.431641</td>\n",
       "      <td>458.701172</td>\n",
       "      <td>1106.158681</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9_0</th>\n",
       "      <td>0.838108</td>\n",
       "      <td>0.573182</td>\n",
       "      <td>0.686754</td>\n",
       "      <td>0.689862</td>\n",
       "      <td>322.550781</td>\n",
       "      <td>304.603516</td>\n",
       "      <td>533.361417</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>880 rows × 7 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "metric_name  accuracy   latency       ram    weight  real_weight    real_ram  \\\n",
       "arm_name                                                                       \n",
       "0_0          0.795562  0.600640  0.703689  0.701442   399.244141  416.117188   \n",
       "100_0        0.661357  0.452087  0.556115  0.591458    52.645508   27.455078   \n",
       "101_0        0.754984  0.572030  0.680436  0.683750   288.205078  271.136719   \n",
       "102_0        0.759577  0.667570  0.684806  0.715898   521.052734  293.867187   \n",
       "103_0        0.757547  0.528259  0.619019  0.657079   176.331055   87.467773   \n",
       "...               ...       ...       ...       ...          ...         ...   \n",
       "96_0         0.790015  0.639732  0.720727  0.722822   591.936523  569.524414   \n",
       "97_0         0.767722  0.732089  0.689055  0.712722   491.443359  317.789063   \n",
       "98_0         0.722385  0.493155  0.607247  0.639543   127.656250   70.416992   \n",
       "99_0         0.758130  0.652381  0.708979  0.715937   521.431641  458.701172   \n",
       "9_0          0.838108  0.573182  0.686754  0.689862   322.550781  304.603516   \n",
       "\n",
       "metric_name  real_latency  \n",
       "arm_name                   \n",
       "0_0            686.835201  \n",
       "100_0          174.840703  \n",
       "101_0          527.732875  \n",
       "102_0         1272.259462  \n",
       "103_0          352.640357  \n",
       "...                   ...  \n",
       "96_0           984.514618  \n",
       "97_0          2304.890761  \n",
       "98_0           255.220262  \n",
       "99_0          1106.158681  \n",
       "9_0            533.361417  \n",
       "\n",
       "[880 rows x 7 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_use"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Pareto frontier"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "However, now we have all the result most of them do not interest us, since we only want the Pareto front. For this we have thw following function, that returns the pareto front"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sparsemod.heir import is_pareto_efficient"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "pareto_mask = is_pareto_efficient(data_use.values)\n",
    "pareto_arms = data_use.index[pareto_mask].values"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We then sort the vlaues, for example, by accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "output_type": "error",
     "ename": "NameError",
     "evalue": "name 'data_use' is not defined",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-4-1c4f97ad16ea>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mdata_use\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mpareto_mask\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msort_values\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mby\u001b[0m \u001b[0;34m=\u001b[0m\u001b[0;34m'error'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m: name 'data_use' is not defined"
     ]
    }
   ],
   "source": [
    "data_use[pareto_mask].sort_values(by ='error')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Where we can see that the best model has a total accuracy of accuracy = (1- error) = 1 - 0.441 = 0.559, a model sie of 112KB in 8-bit vlaues, a RAM of 72 KB and a latency of 221 ms."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Pareto Front and evolution"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can visualize the frontier obtained by Sparse"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "output_type": "error",
     "ename": "NameError",
     "evalue": "name 'data_use' is not defined",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-3-999968b784b3>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfigure\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfigsize\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m10\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m6\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      2\u001b[0m \u001b[0mcm\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcm\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_cmap\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'RdYlBu_r'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 3\u001b[0;31m sc = plt.scatter(1-data_use[pareto_mask].sort_values(by ='accuracy')['accuracy'],\n\u001b[0m\u001b[1;32m      4\u001b[0m          \u001b[0mdata_use\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mpareto_mask\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msort_values\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mby\u001b[0m \u001b[0;34m=\u001b[0m\u001b[0;34m'accuracy'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'real_weight'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m            c=data_use[pareto_mask].sort_values(by ='accuracy')['real_ram'], s=40, cmap=cm)\n",
      "\u001b[0;31mNameError\u001b[0m: name 'data_use' is not defined"
     ]
    }
   ],
   "source": [
    "plt.figure(figsize=(10,6))\n",
    "cm = plt.cm.get_cmap('RdYlBu_r')\n",
    "sc = plt.scatter(1-data_use[pareto_mask].sort_values(by ='accuracy')['accuracy'],\n",
    "         data_use[pareto_mask].sort_values(by ='accuracy')['real_weight'],\n",
    "           c=data_use[pareto_mask].sort_values(by ='accuracy')['real_ram'], s=40, cmap=cm)\n",
    "\n",
    "\n",
    "\n",
    "cb = plt.colorbar(sc)\n",
    "cb.ax.tick_params(labelsize=20)\n",
    "cb.ax.set_ylabel(\" RAM (KB)\", fontsize=25)\n",
    "plt.xticks(fontsize=20)\n",
    "plt.yticks(fontsize=20)\n",
    "plt.ylabel(\"Model Size (KB)\", fontsize=25, fontfamily='serif')\n",
    "plt.xlabel(\"Accuracy\", fontsize=25, fontfamily='serif')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "as well as the evolution"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0, 'Accuracy')"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "data_use['arm'] = data_use.index\n",
    "data_use['arm'] = data_use['arm'].apply(lambda x: int(x.split(\"_\")[0]))\n",
    "data_use.sort_values(by='arm')\n",
    "\n",
    "plt.figure(figsize=(10,6))\n",
    "cm = plt.cm.get_cmap('RdYlBu_r')\n",
    "sc = plt.scatter(1 - data_use.sort_values(by='arm')['accuracy'],\n",
    "         data_use.sort_values(by='arm')['real_weight'],\n",
    "           c=range(len(data_use.sort_values(by='arm'))), s=40, cmap=cm)\n",
    "\n",
    "\n",
    "\n",
    "cb = plt.colorbar(sc)\n",
    "cb.ax.tick_params(labelsize=20)\n",
    "cb.ax.set_ylabel(\" Evaluation Step\", fontsize=25, fontfamily='serif')\n",
    "plt.xticks(fontsize=20)\n",
    "plt.yticks(fontsize=20)\n",
    "plt.ylabel(\"Model Size (KB)\", fontsize=25, fontfamily='serif')\n",
    "plt.xlabel(\"Accuracy\", fontsize=25, fontfamily='serif')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Inspecting an arm"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In the performance lookup we saw that the bet arm was `749_0`. As we want the specific configuration to build the network we need to observe the specific parameters. For this we have to load a Sparse experiment. in `./cnn_cost_example/read_experiment/` a function has been built for that purpose"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[('__call__', <function LevelMapper.__call__ at 0x7fbe43643c20>), ('__init__', <function LevelMapper.__init__ at 0x7fbe43643b90>)]\n",
      "[('__call__', <function BalancedPositiveNegativeSampler.__call__ at 0x7fbe435924d0>), ('__init__', <function BalancedPositiveNegativeSampler.__init__ at 0x7fbe43592440>)]\n",
      "[('__init__', <function BoxCoder.__init__ at 0x7fbe43599cb0>), ('decode', <function BoxCoder.decode at 0x7fbe43599e60>), ('decode_single', <function BoxCoder.decode_single at 0x7fbe43599ef0>), ('encode', <function BoxCoder.encode at 0x7fbe43599d40>), ('encode_single', <function BoxCoder.encode_single at 0x7fbe43599dd0>)]\n",
      "[('__call__', <function Matcher.__call__ at 0x7fbe43599c20>), ('__init__', <function Matcher.__init__ at 0x7fbe43599a70>), ('set_low_quality_matches_', <function Matcher.set_low_quality_matches_ at 0x7fbe43599b90>)]\n",
      "[('__init__', <function ImageList.__init__ at 0x7fbe4358f050>), ('to', <function ImageList.to at 0x7fbe4358f0e0>)]\n",
      "[('__init__', <function Timebase.__init__ at 0x7fbe4337b440>)]\n",
      "[('__init__', <function VideoMetaData.__init__ at 0x7fbe4337b5f0>)]\n"
     ]
    }
   ],
   "source": [
    "from read_experiment import load_experiment"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "ename": "AttributeError",
     "evalue": "'SparseExperiment' object has no attribute 'cuda'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-14-4fe35f5be4a0>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mexp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdata\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mload_experiment\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdata_folder\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"../data/data_cost/files\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m~/Documents/projects/sparse/SpArSe/examples/cnn_cost_example/read_experiment.py\u001b[0m in \u001b[0;36mload_experiment\u001b[0;34m(data_folder)\u001b[0m\n\u001b[1;32m     46\u001b[0m         \u001b[0mmodels_path\u001b[0m \u001b[0;34m=\u001b[0m\u001b[0mpath\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mjoin\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m\"ROOT\"\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"models\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     47\u001b[0m     )\n\u001b[0;32m---> 48\u001b[0;31m     \u001b[0mexp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdata\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msparse_exp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcreate_load_experiment\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     49\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0mexp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdata\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     50\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/envs/sparse/lib/python3.7/site-packages/sparsemod/utils_experiment.py\u001b[0m in \u001b[0;36mcreate_load_experiment\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m     59\u001b[0m             \u001b[0mexp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mattach_data\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     60\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 61\u001b[0;31m             \u001b[0mexp\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_experiment\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     62\u001b[0m             \u001b[0mdata\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mData\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     63\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mexp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdata\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/envs/sparse/lib/python3.7/site-packages/sparsemod/utils_experiment.py\u001b[0m in \u001b[0;36mget_experiment\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m     78\u001b[0m                 \u001b[0msplitter\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msplitter\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     79\u001b[0m                 \u001b[0mmodels_path\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmodels_path\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 80\u001b[0;31m                 \u001b[0mcuda\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcuda\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     81\u001b[0m             ),\n\u001b[1;32m     82\u001b[0m             WeightMetric(\n",
      "\u001b[0;31mAttributeError\u001b[0m: 'SparseExperiment' object has no attribute 'cuda'"
     ]
    }
   ],
   "source": [
    "exp, data = load_experiment(data_folder=\"../data/data_cost/files\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "exp.arms_by_name['779_0']"
   ]
  }
 ],
 "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": "3.7.7-final"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}