pendulum_param_inference.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [],
   "source": [
    "import os\n",
    "import sys"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Current root dir is /home/framos/Code/bayes_sim\n"
     ]
    }
   ],
   "source": [
    "cur_root_dir = os.getcwd()\n",
    "print(\"Current root dir is {}\".format(cur_root_dir))\n",
    "if cur_root_dir not in sys.path:\n",
    "    sys.path.append(cur_root_dir)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [],
   "source": [
    "import warnings\n",
    "warnings.filterwarnings('ignore')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [
    {
     "ename": "ImportError",
     "evalue": "No module named 'src.models.MDN'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mImportError\u001b[0m                               Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-7-e34abc1b3854>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0;32mfrom\u001b[0m \u001b[0msrc\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmodels\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mMDN\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mMDNN\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mMDRFF\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mMDLSTM\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mMDRFFLSTM\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      2\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0msrc\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmodels\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mBayesSim\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mBayesSim\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0msrc\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mPendulumDataGenerator\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mPendulumDataGenerator\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0msrc\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mutils\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mparam_inference\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mpickle\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mImportError\u001b[0m: No module named 'src.models.MDN'"
     ]
    }
   ],
   "source": [
    "from src.models.MDN import MDNN, MDRFF, MDLSTM, MDRFFLSTM\n",
    "from src.models.BayesSim import BayesSim\n",
    "from src.data.PendulumDataGenerator import PendulumDataGenerator\n",
    "from src.utils.param_inference import *\n",
    "import pickle\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {}
   },
   "source": [
    "# Inference on Pendulum Mass and length"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {}
   },
   "source": [
    "### Loading simulation data from file"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {}
   },
   "source": [
    "The FrankaDataGenerator is responsible for running the forward model with priors sampled from an uniform distribution with the believed range of the actual values."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {}
   },
   "source": [
    "In order to collect data we assume:\n",
    "    - There is a controller/policy of which you can run your forward model (simulation) with.\n",
    "    - You are able to calculate some sufficient statistics from your trajectories. In our case we use a Reinforcement Learning policy that we calculate cross-correlation\n",
    "    - You have some understanding of the prior parameters of your simulation. For instance, you know that a specific parameter can't go below or over some threshold"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading a model without an environment, this model cannot be trained until it has a valid environment.\n",
      "WARNING:tensorflow:From /home/rafaelpossas/py36/lib/python3.6/site-packages/stable_baselines/common/input.py:26: to_float (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.cast instead.\n",
      "WARNING:tensorflow:From /home/rafaelpossas/py36/lib/python3.6/site-packages/stable_baselines/common/policies.py:436: flatten (from tensorflow.python.layers.core) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use keras.layers.flatten instead.\n",
      "WARNING:tensorflow:From /home/rafaelpossas/py36/lib/python3.6/site-packages/tensorflow/python/framework/op_def_library.py:263: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Colocations handled automatically by placer.\n",
      "INFO:tensorflow:Summary name model/pi_fc0/w:0 is illegal; using model/pi_fc0/w_0 instead.\n",
      "INFO:tensorflow:Summary name model/pi_fc0/b:0 is illegal; using model/pi_fc0/b_0 instead.\n",
      "INFO:tensorflow:Summary name model/vf_fc0/w:0 is illegal; using model/vf_fc0/w_0 instead.\n",
      "INFO:tensorflow:Summary name model/vf_fc0/b:0 is illegal; using model/vf_fc0/b_0 instead.\n",
      "INFO:tensorflow:Summary name model/pi_fc1/w:0 is illegal; using model/pi_fc1/w_0 instead.\n",
      "INFO:tensorflow:Summary name model/pi_fc1/b:0 is illegal; using model/pi_fc1/b_0 instead.\n",
      "INFO:tensorflow:Summary name model/vf_fc1/w:0 is illegal; using model/vf_fc1/w_0 instead.\n",
      "INFO:tensorflow:Summary name model/vf_fc1/b:0 is illegal; using model/vf_fc1/b_0 instead.\n",
      "INFO:tensorflow:Summary name model/vf/w:0 is illegal; using model/vf/w_0 instead.\n",
      "INFO:tensorflow:Summary name model/vf/b:0 is illegal; using model/vf/b_0 instead.\n",
      "INFO:tensorflow:Summary name model/pi/w:0 is illegal; using model/pi/w_0 instead.\n",
      "INFO:tensorflow:Summary name model/pi/b:0 is illegal; using model/pi/b_0 instead.\n",
      "INFO:tensorflow:Summary name model/pi/logstd:0 is illegal; using model/pi/logstd_0 instead.\n",
      "INFO:tensorflow:Summary name model/q/w:0 is illegal; using model/q/w_0 instead.\n",
      "INFO:tensorflow:Summary name model/q/b:0 is illegal; using model/q/b_0 instead.\n",
      "WARNING:tensorflow:From /home/rafaelpossas/py36/lib/python3.6/site-packages/tensorflow/python/ops/math_ops.py:3066: to_int32 (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.cast instead.\n"
     ]
    }
   ],
   "source": [
    "policy_file = os.path.join(cur_root_dir, \"src\",\"models\",\"controllers\",\"PPO\", \"Pendulum-v0.pkl\")\n",
    "g = PendulumDataGenerator(policy_file=policy_file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 1/1 [00:00<00:00, 13.85it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Drawing Parameters and Running simulation...\n",
      "Total data size: 1\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "params, stats = g.gen(1)\n",
    "shapes = {\"params\": params.shape[1], \"data\": stats.shape[1]}\n",
    "\n",
    "print(\"Total data size: {}\".format(params.shape[0]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {}
   },
   "source": [
    "### Training the model using a Mixture of Density network"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {}
   },
   "source": [
    "We train a mixture of density network / mixture of density random fourier features where the inputs are the sufficient statistics calculated from the data and the output are the weights, means and variances of the Mixture of Gaussians"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|          | 3/1000 [00:00<00:39, 25.20it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Drawing Parameters and Running simulation...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 1000/1000 [00:48<00:00, 20.48it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training model....\n",
      "\n",
      "*********************************  End  *********************************\n",
      "\n",
      "Duration: 0:00:57.903148\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "log_mdn, inf_mdn = train(epochs=1000, batch_size=32, params_dim=2, stats_dim=9, num_sampled_points=1000,\n",
    "                         generator=g, model=\"MDRFFLSTM\", n_components=10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {}
   },
   "source": [
    "### Training the Model using a Mixture of Density Random Fourier Features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Data has already been generated, loading from memory...\n",
      "\n",
      "*********************************  End  *********************************\n",
      "\n",
      "Duration: 0:01:02.146599\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "log_rff, inf_rff = train(epochs=1000, batch_size=32, params_dim=2, stats_dim=9, num_sampled_points=1000,\n",
    "                         generator=g, model=\"MDRFF\", n_components=2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {}
   },
   "source": [
    "### Recovering the posterior from a specific stiffness parameter"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {}
   },
   "source": [
    "Now you can either run an experiment in the real robot or collect trajectories for a specific value of your dynamics parameter. You can then input the sufficient statistics of these trajectories in Bayes Sim and it will recover your full posterior over the chosen parameter"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {}
   },
   "source": [
    "### MDN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [],
   "source": [
    "true_obs = np.array([[1.0, 1.5]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "pycharm": {
     "is_executing": false
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Scaled true obs: [[1.  1.5]]\n",
      "Params: [1.  1.5]\n",
      "\n",
      "\n",
      "*********************************  Prediction ends  *********************************\n",
      "\n",
      "\n",
      "Duration: 0:00:00.059814\n",
      "component 1: mixture weight = 0.0117; mean = 0.7231; variance = 0.0006\n",
      "component 2: mixture weight = 0.9883; mean = 0.9973; variance = 0.0002\n",
      "component 1: mixture weight = 0.0117; mean = 1.9251; variance = 0.0152\n",
      "component 2: mixture weight = 0.9883; mean = 1.6194; variance = 0.0365\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "get_results_from_true_obs(env_params=[\"length\", \"mass\"], true_obs=true_obs, generator=g, inf=inf_mdn, shapes=shapes,\n",
    "                          p_lower=[0.1, 0.1], p_upper=[2.0, 2.0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {}
   },
   "source": [
    "### MDRFF"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "pycharm": {}
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Scaled true obs: [[1.  1.5]]\n",
      "Params: [1.  1.5]\n",
      "\n",
      "\n",
      "*********************************  Prediction ends  *********************************\n",
      "\n",
      "\n",
      "Duration: 0:00:00.080045\n",
      "component 1: mixture weight = 0.5205; mean = 0.9960; variance = 0.0000\n",
      "component 2: mixture weight = 0.4795; mean = 0.9818; variance = 0.0000\n",
      "component 1: mixture weight = 0.5205; mean = 1.8484; variance = 0.0147\n",
      "component 2: mixture weight = 0.4795; mean = 1.5046; variance = 0.0005\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAEKCAYAAAACS67iAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAIABJREFUeJzt3Xt0VPW5//H3A+Eeyk1AlogBlGsCkxgjFClUqShyQDhcrHoARSm1FrHHC3rwWl3Fo8fWX2ultChoVUArllqlCEgVLVeLinIJ0FRjLYSbcg+B7++P2RlymUx2LpPMTD6vtVgzs2fP3g97hQ9PvnvPd5tzDhERiX/1arsAERGpHgp0EZEEoUAXEUkQCnQRkQShQBcRSRAKdBGRBKFAFxFJEAp0EZEEoUAXEUkQSTW5s7POOsulpKTU5C4l0WzbFnzs3r126xCpQRs3btzrnGtb3no1GugpKSls2LChJncpiWbw4ODjqlW1WYVIjTKzf/pZT0MuIiIJQoEuIpIgFOgiIgmiRsfQReq6kydPkpuby/Hjx2u7FIlBjRs3pmPHjjRo0KBSn1egi9Sg3NxcmjdvTkpKCmZW2+VIDHHOsW/fPnJzc+ncuXOltqEhF5EadPz4cdq0aaMwl1LMjDZt2lTptzcFukgNU5hLWar6s6FAl7iydy/k59d2FSKxSYEucePYMdj8KXz0cW1XEt/q169PIBAgNTWVsWPHcvTo0Upva9WqVQwfPhyAJUuWMGvWrDLXPXjwIL/+9a8rvI8HH3yQJ554otI1RlLZmgCGDRvGwYMHq7miqlGgS9wo7Mx1gUjVNGnShE2bNrF582YaNmzI7Nmzi73vnOP06dMV3u6IESOYMWNGme9XJTyjpTI1FR6fN998k5YtW1boM9GmQJe4UVAQfNQIdPUZOHAgO3bsICcnh+7duzNhwgRSU1P54osvWLZsGf379ycjI4OxY8dy+PBhAJYuXUqPHj3IyMjgtddeC21r3rx53HrrrQDs3r2bUaNG0bdvX/r27csHH3zAjBkz2LlzJ4FAgDvvvBOAxx9/nIsuuog+ffrwwAMPhLb16KOP0q1bNy655BK2Fc7fU8KkSZOYOnUqmZmZdOvWjTfeeAMInni+4YYbSEtLIz09nXfeeQeATz/9lKysLAKBAH369CE7O9t3TeGOT0pKCnv37gXgySefJDU1ldTUVH7xi1+U+Zlo02WLEjdCY+cJkujTp8OmTdW7zUAAvDwpV0FBAW+99RZXXHEFANnZ2cyfP59+/fqxd+9eHnnkEZYvX06zZs147LHHePLJJ7nrrru4+eabWblyJeeffz7jx48Pu+1p06YxaNAgFi9ezKlTpzh8+DCzZs1i8+bNbPL+0suWLSM7O5t169bhnGPEiBG8++67NGvWjAULFrBp0yYKCgrIyMjgwgsvDLufnJwc1q1bx86dO/nud7/Ljh07ePrppzEzPvnkE7Zu3crll1/O9u3bmT17NrfddhvXXXcd+fn5nDp1yndNnTp1KnZ8itq4cSPPPfcca9euxTnHxRdfzKBBg2jVqlWZn4kWBbrEjcJAT5A8rzXHjh0jEAgAwQ598uTJ/Otf/+K8884LBc+aNWv47LPPGDBgAAD5+fn079+frVu30rlzZy644AIArr/+eubMmVNqHytXruT5558HgmP2LVq04MCBA8XWWbZsGcuWLSM9PR2Aw4cPk52dzaFDhxg1ahRNmzYFgkM5ZRk3bhz16tXjggsuoEuXLmzdupXVq1fz4x//GIAePXpw3nnnsX37dvr378+jjz5Kbm4uo0ePDv0d/NTUqVOnYsenqNWrVzNq1CiaNWsGwOjRo3nvvfcYMWJEmZ+JFgW6xI1E69D9dtLVrXAMvaTCQILgmO/3vvc9Xn755WLrhPtcZTnnuOeee/jBD35QbPkvKnBgSl7mF+myv2uvvZaLL76YP//5zwwbNozf/OY3dOnSxVdNOTk5xY6PX5X5TFVoDF3ixsmTwccEyfOY1q9fP95//3127NgBwJEjR9i+fTs9evQgJyeHnTt3ApQK/EKXXXYZzzzzDACnTp3i66+/pnnz5hw6dCi0ztChQ3n22WdDY/Nffvkle/bs4Tvf+Q6vv/46x44d49ChQ/zpT38qs85XXnmF06dPs3PnTnbt2kX37t0ZOHAgL774IgDbt2/n888/p3v37uzatYsuXbowbdo0Ro4cyccff+y7pkgGDhzI66+/ztGjRzly5AiLFy9m4MCBET8TLerQJW7UwEUC4mnbti3z5s3j+9//PidOnADgkUceoVu3bsyZM4errrqKpk2bMnDgwGKBWOipp55iypQpzJ07l/r16/PMM8/Qv39/BgwYQGpqKldeeSWPP/44W7ZsoX///gAkJyfz+9//noyMDMaPH0/fvn1p164dF110UZl1durUiaysLL755htmz55N48aNueWWW/jhD39IWloaSUlJzJs3j0aNGrFo0SJeeOEFGjRowNlnn829995L69atfdVUv379MmvIyMhg0qRJZGVlAXDTTTeRnp5OTk5OZQ9/pZlzrsZ2lpmZ6XSDC6msjz+G/X0H06ABDMhfVdvlVMqWLVvo2bNnbZeRECZNmsTw4cMZM2ZMbZdSrcL9jJjZRudcZnmf1ZCLxI1Q71FzPYhIXNGQi8SNwkCvwV8qJYbNmzevtkuIOerQJW5oDF0kMgW6xA115iKRKdAlboQ6dF23KBKWAl3ihjp0kch8BbqZtTSzV81sq5ltMbP+ZtbazN42s2zvsVW0i5W6TYFedTk5OaSmphZb5md62g0bNjBt2jQATpw4wZAhQwgEAixcuDBqtVbE7NmzQ1MN1GV+r3J5CljqnBtjZg2BpsC9wArn3CwzmwHMAO6OUp0iOilaizIzM8nMDF4G/fe//x2o2DQAp06divjlnKooKChg6tSpFf5MUlLiXeRXboduZi2A7wBzAZxz+c65g8BIYL632nzg6mgVKQLq0GvC4MGDufvuu8nKyqJbt2689957wJkbWezZs4frr7+e9evXEwgE2LlzJytWrCA9PZ20tDRuvPHG0DdLU1JSuPvuu8nIyOCVV15h8ODB3H777WRmZtKzZ0/Wr18fmiRr5syZYetJTk7m9ttvp3fv3lx22WXk5eWF6pw+fTqZmZk89dRTxX7L2LRpE/369aNPnz6MGjUqNClYyc8kIj//RXUG8oDnzKwvsBG4DWjvnPvKW+ffQPtwHzazKcAUCH5NV6SyEq1Dn750Opv+Xb3z5wbODvCLK6o261dBQQHr1q3jzTff5KGHHmL58uWh99q1a8fvfvc7nnjiCd544w2OHz/O4MGDWbFiBd26dWPChAk888wzTJ8+HYA2bdrw4YcfAsFhkYYNG7JhwwaeeuopRo4cycaNG2ndujVdu3bl9ttvp02bNsVqOXLkCJmZmfz85z/n4Ycf5qGHHuJXv/oVEJwBsvCb5w8++GDoMxMmTOCXv/wlgwYN4v777+ehhx4KTfhV9DOJyM8YehKQATzjnEsHjhAcXglxwfkDwvZPzrk5zrlM51xm27Ztq1qv1GHq0KuurNkIiy4fPXo0ABdeeGG585Fs27aNzp07061bNwAmTpzIu+++G3q/5HzphVPhpqWl0bt3bzp06ECjRo3o0qVL2BtA1KtXL7SN66+/ntWrV5e5bYCvv/6agwcPMmjQIF/1JBo/HXoukOucW+u9fpVgoO82sw7Oua/MrAMQeUoykSpKtECvaiddGW3atCk1L/n+/fvp3Llz6HWjRo2A4DzmBYW3iaqkktPHFm67Xr16oeeFr/3sq+h/PPEwnW1NK7dDd879G/jCzLp7iy4DPgOWABO9ZROBP0alQhFP4ZCLLkOvvOTkZDp06MDKlSuBYJgvXbqUSy65pFLb6969Ozk5OaFpdl944YVQd1wdTp8+zauvvgrASy+9VG6dLVq0oFWrVqGx/+quJ9b5Pc37Y+BF7wqXXcANBP8zWGRmk4F/AuOiU6JIkHPBME+wRr3GPf/88/zoRz/iJz/5CQAPPPAAXbt2rdS2GjduzHPPPcfYsWMpKCjgoosuqvAVJ5E0a9aMdevW8cgjj9CuXTtfl0nOnz+fqVOncvToUbp06cJzzz1XbfXEOk2fK3Fj5Uqod9lgkpLgkpOrarucStH0uRWTnJwcutlEXaHpc6VO0JCLSGQKdIkbiXZSVMpX17rzqlKgS9xItOvQRaqbAl3ihjp0kcgU6BI3FOgikSnQJW5oyEUkMgW6xA116FW3b98+AoEAgUCAs88+m3POOSf0Oj8/v9bqWr58OVdfHd35/VatWkXv3r1L/V0LCgpo2bJlte/vww8/ZOnSpaHXM2fODM0pEy2JN3+kJCzdsajq2rRpE5r29sEHHyQ5OZk77rij2DrOOZxz1KuXWP3e73//e+677z6uueaaGtnfhx9+yObNm7niiitqZH+gDl3iiDr06NmxYwe9evXiuuuuo3fv3nzxxRfFutYFCxZw0003AbB7925Gjx5NZmYmWVlZrFmzptT2MjMz2bZtW+j1JZdcwqZNm1izZg39+/cnPT2dAQMGkJ2dXeqzJTvZHj16kJubCwS/BZqVlUUgEOCWW27hdJhxuGXLlhEIBEhLS+Pmm28mPz+f2bNn89prr3HPPfcwYcKEiMdi1qxZZGVl0adPHx5++OHQ8UlNTWXy5Mn07t2bK6+8kuPHjwOwZs0a+vTpQyAQ4I477iAQCHDs2DEefvhhXnzxRQKBQGj6gk8++YRBgwbRpUsXnn766Yh1VIY6dIkbCRfo06dDBW4S4UsgAJX8tX7r1q08//zzZGZmRpwoa9q0adx1113069ePnJwchg8fzubNm4utM378eBYtWsR9991Hbm4u+/fvJxAI8PXXX/Pee++RlJTE0qVLmTlzpu+7Hm3evJnFixfzwQcfkJSUxJQpU1iwYAHXXnttaJ2jR49y44038te//pWuXbty3XXXMWfOHG699VZWr17NmDFjIg7tvPnmm3z++eesXbsW5xzDhg3jgw8+oF27dmzbto2XX36ZtLQ0Ro8ezeuvv84111zDDTfcEPqPpvC3nSZNmnD//fezefPm0H9OmzZtYvv27axYsYKDBw/Ss2dPpk6dWq03/lCgS9zQSdHo6tq1a+iuRJEsX768WPd94MABjh07RpMmTULLxo0bx3/8x39w3333sXDhQsaOHQvAwYMHmTBhAjt37qxwfcuXL2f9+vWhGo8dO8a5555bbJ0tW7bQrVu30Nw0EyZMYO7cudx6662+9rFs2TLeeust0tPTgeAXm7Zv3067du04//zzSUtLA85MLbx3717y8/PJysoC4Nprry02f3xJw4cPp2HDhrRr147WrVuTl5fH2WefXbEDEYECXeJGwnXoUT5BVlFFp5atV68eRed5KhxegOAY+7p162jYsGGZ2zrvvPNITk7ms88+Y+HChcybNw+A//mf/2Ho0KHccsst7NixI+z4clJSUrGhlMJ9O+e48cYb+elPf1rpv2N5nHPMnDmTyZMnF1u+Y8eOYtP9VnZq4erYRiQaQ5e4oQ695tSrV49WrVqRnZ3N6dOnWbx4cei9IUOGFBv/LeveouPHj+dnP/sZJ06coFevXkDwBhTnnHMOQCjkS0pJSWHjxo0ArFu3LnTjiyFDhrBo0SL27t0LBK/Y+fzzz4t9tmfPnmRnZ7Nr1y4geCK0ItPnDh06lLlz53LkyBEAcnNzQ/sL56yzzqJBgwahuyAtWLAg9F7z5s05dOiQ731XBwW6xI2E69Bj3GOPPcbQoUP59re/TceOHUPLn376ad5//3369OlDr169+O1vfxv282PHjuWll15i3LgzM2vffffd3HnnnWRkZFDWTK9jx45l9+7dpKamMmfOHLp06QIE73L0wAMPMGTIEPr06cPll1/O7t27i322adOmzJ07l9GjR5OWlkajRo24+eabff+dhw0bxpgxY+jXrx9paWmMGzeu3Plknn32WW644QbS09M5fvw4LVq0AODSSy/lo48+Ij09PXRSNNo0fa7EjYULof01g2nQAAbkr6rtcipF0+cmnsOHD5OcnAzAo48+yv79+/m///u/Sm+vKtPnagxd4oamz5VYtGTJEv73f/+XgoICUlJSyhxKqgkKdIkbGnKRWHTttdcWu3SyNmkMXeJGopwUrclhTokvVf3ZUKBL3EiEHGzcuDH79u1TqEspzjn27dtH48aNK70NDblI3EiEDOzYsSO5ubnk5eXVdikSgxo3blzsiqKKUqBL3EiEIZcGDRrQuXPn2i5DEpSvQDezHOAQcAoocM5lmllrYCGQAuQA45xzB6JTpkhidOgi0VSRMfTvOucCRa6FnAGscM5dAKzwXotEjabPFYmsKidFRwLzvefzgejOTi91XqhDV6cuEpbfQHfAMjPbaGZTvGXtnXNfec//DbQP90Ezm2JmG8xsg04ESVVoyEUkMr8nRS9xzn1pZu2At81sa9E3nXPOzML+c3POzQHmQPCr/1WqVuo0DbmIROarQ3fOfek97gEWA1nAbjPrAOA97olWkSKgDl2kPOUGupk1M7Pmhc+By4HNwBJgorfaROCP0SpSBBLjskWRaPIz5NIeWGxmheu/5JxbambrgUVmNhn4JzAuwjZEqkwdukhk5Qa6c24X0DfM8n3AZdEoSiQcBbpIZJrLReKGhlxEIlOgS9xQhy4SmQJd4oZucCESmQJd4oY6dJHIFOgSNxToIpEp0CVu6KSoSGQKdIkb6tBFIlOgS9xQoItEpkCXuKEhF5HIFOgSN9Shi0SmQJe4oelzRSJToEvcUIcuEpkCXeKGAl0kMgW6xA2dFBWJTIEucUM3iRaJTIEucUMdukhkCnSJG4Uduhp0kfAU6BI3dFJUJDIFusSN0JCLgl0kLAW6xA0NuYhE5jvQzay+mf3dzN7wXnc2s7VmtsPMFppZw+iVKaKToiLlqUiHfhuwpcjrx4CfO+fOBw4Ak6uzMJGSdNmiSGS+At3MOgJXAb/zXhtwKfCqt8p84OpoFChSSCdFRSLz26H/ArgLKPyltw1w0DlX4L3OBc6p5tpEiikcclGui4RXbqCb2XBgj3NuY2V2YGZTzGyDmW3Iy8urzCZEAHXoIuXx06EPAEaYWQ6wgOBQy1NASzNL8tbpCHwZ7sPOuTnOuUznXGbbtm2roWSpq3TZokhk5Qa6c+4e51xH51wKcA2w0jl3HfAOMMZbbSLwx6hVKYIuWxQpT1WuQ78b+ImZ7SA4pj63ekoSCU9DLiKRJZW/yhnOuVXAKu/5LiCr+ksSCU9DLiKR6ZuiEjfUoYtEpkCXuKHLFkUiU6BL3FCHLhKZAl3ihsbQRSJToEvc0GWLIpEp0CVuaLZFkcgU6BI3NOQiEpkCXeKGOnSRyBToEjd02aJIZAp0iRvq0EUiU6BL3FCgi0SmQJe4UfSLRfqSkUhpCnSJG0U7dAW6SGkKdIkbCnSRyBToEjcU6CKRKdAlbhQNdJ0gFSlNgS5xQx26SGQKdIkbCnSRyBToEjd02aJIZAp0iRvq0EUiU6BL3FCgi0RWbqCbWWMzW2dmH5nZp2b2kLe8s5mtNbMdZrbQzBpGv1ypi44dg2bN4O23zyzTVS4ipfnp0E8Alzrn+gIB4Aoz6wc8BvzcOXc+cACYHL0ypS7bvh2OHi2+TB26SGnlBroLOuy9bOD9ccClwKve8vnA1VGpUOq8Q4dKL1Ogi5TmawzdzOqb2SZgD/A2sBM46Jwr8FbJBc6JTolS1ynQRfzxFejOuVPOuQDQEcgCevjdgZlNMbMNZrYhLy+vkmVKXXbkSOllGkMXKa1CV7k45w4C7wD9gZZmluS91RH4sozPzHHOZTrnMtu2bVulYqVuys8vvUwdukhpfq5yaWtmLb3nTYDvAVsIBvsYb7WJwB+jVaTUbSdPll6mDl2ktKTyV6EDMN/M6hP8D2CRc+4NM/sMWGBmjwB/B+ZGsU6pw8IFujp0kdLKDXTn3MdAepjluwiOp4tElTp0EX/0TVGJeerQRfxRoEvMU4cu4o8CXWKeOnQRfxToEvPUoYv4o0CXmKdAF/FHgS4xT0MuIv4o0CXmqUMX8UeBLjFPHbqIPwp0iXnq0EX8UaBLzFOHLuKPAl1injp0EX8U6BLz1KGL+KNAl5inDl3EHwW6xDx16CL+KNAl5qlDF/FHgS4xTx26iD8KdIl56tBF/FGgS8w7eRLq1y++TIEuUpoCXWLeyZPQtGnxZRpyESlNgS4x7+RJSCpx91t16CKllXuTaJHadvIkZGRAfj50zoV//EMdukg45XboZnaumb1jZp+Z2admdpu3vLWZvW1m2d5jq+iXK3VRfj60bg3vvgvNk4PL1KGLlOZnyKUA+G/nXC+gH/AjM+sFzABWOOcuAFZ4r0Wq3cmT0KCB98KCD+rQRUorN9Cdc1855z70nh8CtgDnACOB+d5q84Gro1Wk1G0nT0LDhsHnXp6rQxcJo0InRc0sBUgH1gLtnXNfeW/9G2hfrZWJeIqdFFWHLlIm34FuZsnAH4Dpzrlvir7nnHNA2H9iZjbFzDaY2Ya8vLwqFSt1U0FBkSEXjzp0kdJ8BbqZNSAY5i86517zFu82sw7e+x2APeE+65yb45zLdM5ltm3btjpqljqmaKAXDrmoQxcpzc9VLgbMBbY4554s8tYSYKL3fCLwx+ovTyT8kIs6dJHS/FyHPgD4L+ATM9vkLbsXmAUsMrPJwD+BcdEpUeo6degi/pQb6M651Zz5d1TSZdVbjkhpBQXq0EX80Ff/JaY5B6dO6av/In4o0CWmFRQEHzXkIlI+BbrEtMK50DXkIlI+BbrENHXoIv4p0CWmFQa6xtBFyqdAl5hW1pCLOnSR0hToEtPKGnJRhy5SmgJdYpo6dBH/FOgS00p26IXUoYuUpkCXmFbypKiuchEpmwJdYpquQxfxT4EuMU0nRUX8U6BLTNNJURH/FOgS0/TFIhH/FOgS0/TVfxH/FOgS03RSVMQ/BbrENHXoIv4p0CWmlerQPerQRUpToEtMK3VSVFe5iJRJgS4xTdehi/inQJeYpuvQRfwrN9DN7Fkz22Nmm4ssa21mb5tZtvfYKrplSl2lyblE/PPToc8DriixbAawwjl3AbDCey1S7Up26LrKRaRs5Qa6c+5dYH+JxSOB+d7z+cDV1VyXCFD2SVF16CKlVXYMvb1z7ivv+b+B9tVUj0gxOikq4l+VT4o65xxQ5i/AZjbFzDaY2Ya8vLyq7k7qGJ0UFfGvsoG+28w6AHiPe8pa0Tk3xzmX6ZzLbNu2bSV3J3WVToqK+FfZQF8CTPSeTwT+WD3liBSnk6Ii/vm5bPFl4G9AdzPLNbPJwCzge2aWDQzxXotUO50UFfEvqbwVnHPfL+Oty6q5FpFSTpwAM3XoIn7om6IS044fh0aNgqFelDp0kdIU6BLTTpwIBnqIrnIRKZMCXWLaiRPQuPGZ17oOXaRsCnSJaerQRfxToEtMKxXoHnXoIqUp0CWmlQx0DbmIlE2BLjGtrCGXwuvTReQMBbrEtHAdutmZb5CKyBkKdIlp4cbQFegi4SnQJaYVfrGoqHoKdJGwFOgS09Shi/inQJeYFjbQ6ynQRcJRoEtMK/lNUVCHLlIWBbrEtHAdusbQRcJToEtMO3oUmjQpvkwdukh4CnSJWc7B4cPQvHnx5Qp0kfAU6BKzjh4NfsW/VKDrpKhIWAp0iVmHDgUfv/Wt4svVoYuEp0CXmFUY6CU7dJ0UFQlPgS4x65tvgo/hxtA1OZdIaQp0iVl79gQf27UrvtzqBS9nFJHiqhToZnaFmW0zsx1mNqO6ihIB+Oqr4GOHDsWXN0iCAwdqvh6RWJdU2Q+aWX3gaeB7QC6w3syWOOc+q67iSvrDZ39gz5E90dq8xJg/fwFkwpKvoEFecNmIQ/8ivxF82eEZnllfq+VJnGrdpDXjeo/DzMpfOc5UOtCBLGCHc24XgJktAEYCUQn0nIM5jHllTDQ2LbFsONy27MzLnvuA+nD00lu45c1aq0riXGq7VHq3613bZVS7qgT6OcAXRV7nAhdXrZzwet/9Q7JP/wWS4ZwPFtB0z+BSNwkOd8/gcDcSru5lJXdc7fsMt0sfn63Ov1O17zPcLsNs69A3cNPN8MhPzyxvtXwUeXnA44s5uwM0Tw6zMR8SsDkTH461XscXA0eQ/sRV1CtoVqP7XjrxTwzu2yWq+6hKoPtiZlOAKQCdOnWq1DY6Nu/EN/svJOnAYAKtrqRBi2952y65r3D7r/lldbmO6txWy5bw3/8N3yoa2vUb0qE93Puf7cnOLr0NP8L+JyZ1wikbQtL+m8ivf7DG953cJMzdzqtZVQL9S+DcIq87esuKcc7NAeYAZGZmVuqf0l9m3lOZj0mCqmfw6KO1XYXEpybAb2u7iKipylUu64ELzKyzmTUErgGWVE9ZIiJSUZXu0J1zBWZ2K/AXoD7wrHPu02qrTEREKqRKY+jOuTcBXWsgIhID9E1REZEEoUAXEUkQCnQRkQShQBcRSRAKdBGRBGGuBr82Z2Z5wD+Bs4C9Nbbj+KPjE5mOT2Q6PuWLt2N0nnOubXkr1Wigh3ZqtsE5l1njO44TOj6R6fhEpuNTvkQ9RhpyERFJEAp0EZEEUVuBPqeW9hsvdHwi0/GJTMenfAl5jGplDF1ERKqfhlxERBJEVAO9vJtIm1kjM1vovb/WzFKiWU+s8XF8JplZnplt8v7cVBt11gYze9bM9pjZ5jLeNzP7f96x+9jMMmq6xtrm4xgNNrOvi/z83F/TNdYWMzvXzN4xs8/M7FMzuy3MOon3M+Sci8ofglPq7gS6AA2Bj4BeJda5BZjtPb8GWBitemLtj8/jMwn4VW3XWkvH5ztABrC5jPeHAW8BBvQD1tZ2zTF4jAYDb9R2nbV0bDoAGd7z5sD2MP++Eu5nKJodeugm0s65fKDwJtJFjQTme89fBS6zRLwVd3h+jk+d5Zx7F9gfYZWRwPMuaA3Q0sw61Ex1scHHMaqznHNfOec+9J4fArYQvA9yUQn3MxTNQA93E+mSBzS0jnOuAPgaaBPFmmKJn+MD8J/er4Ovmtm5Yd6vq/wev7quv5l9ZGZvmVnJBloYAAAEZklEQVTi3ebeB28oNx1YW+KthPsZ0knR2PYnIMU51wd4mzO/zYj48SHBr4z3BX4JvF7L9dQ4M0sG/gBMd859U9v1RFs0A93PTaRD65hZEtAC2BfFmmJJucfHObfPOXfCe/k74MIaqi0e+LpJeV3mnPvGOXfYe/4m0MDMzqrlsmqMmTUgGOYvOudeC7NKwv0MRTPQ/dxEegkw0Xs+BljpvLMVdUC5x6fEeN4IguOAErQEmOBdqdAP+No591VtFxVLzOzswnNSZpZF8N97nWiYvL/3XGCLc+7JMlZLuJ+hKt1TNBJXxk2kzexhYINzbgnBA/6Cme0geHLnmmjVE2t8Hp9pZjYCKCB4fCbVWsE1zMxeJniVxllmlgs8ADQAcM7NJngv22HADuAocEPtVFp7fByjMcAPzawAOAZcU4capgHAfwGfmNkmb9m9QCdI3J8hfVNURCRB6KSoiEiCUKCLiCQIBbqISIJQoIuIJAgFuohIglCgS60zs8NR2GbAzIYVef2gmd0Rhf3k1KUv60hsU6BLogoQvMZYpM5QoEtMMbM7zWy9NyHZQ96yFDPbYma/9ea2XmZmTbz3LvLW3WRmj5vZZu+btw8D473l473N9zKzVWa2y8ymhdn3VDN7vMjrSWb2K+/562a20dv/lDCfTSk6L7mZ3WFmD3rPu5rZUu/z75lZD2/5WK/ej8zs3eo6hlJ3KdAlZpjZ5cAFBKcWDgAXmtl3vLcvAJ52zvUGDgL/6S1/DviBcy4AnALwpiO+n+D8+gHn3EJv3R7AUG/7D3hzfRT1B2BUkdfjCU5rDHCjc+5CIJPgN3grMivoHODH3ufvAH7tLb8fGOpNnjWiAtsTCStqX/0XqYTLvT9/914nEwzyz4F/OOcKv8K9EUgxs5ZAc+fc37zlLwHDI2z/z95kZyfMbA/QnuCUqQA45/K87r0fkE3wP4D3vbenmVlh2J/r1VXuvCjebH/fBl4pMtV/I+/xfWCemS0Cwk0eJVIhCnSJJQb8zDn3m2ILg/NZnyiy6BTQpBLbL7mNcD//C4BxwFZgsXPOmdlgYAjQ3zl31MxWAY1LfK6A4r/xFr5fDzjo/QZRjHNuqpldDFwFbDSzC51zdWLyLIkODblILPkLcKPX1WJm55hZu7JWds4dBA55oQjFJ3c7RPDWYxW1mOCdbL7PmeGWFsABL8x7ELxdWUm7gXZm1sbMGuH9puDNwf0PMxvr/Z3MzPp6z7s659Y65+4H8ig+latIhSnQJWY455YRHDb5m5l9QvC2hOWF8mTgt96Mes0I3vUK4B2CJ0GLnhT1U8MBgtMUn+ecW+ctXgokmdkWYBawJsznThI8EbuO4M1IthZ5+zpgspl9BHzKmVsNPm5mn3gnUz8geF9ZkUrTbIsS18wsufAmDmY2A+jgnCt1h3eRukBj6BLvrjKzewj+LP+TOjRnvEhJ6tBFRBKExtBFRBKEAl1EJEEo0EVEEoQCXUQkQSjQRUQShAJdRCRB/H8b0eyYJeN/ggAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "get_results_from_true_obs(env_params=[\"length\", \"mass\"], true_obs=true_obs, generator=g, inf=inf_rff, shapes=shapes,\n",
    "                          p_lower=[0.1, 0.1], p_upper=[2.0, 2.0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "pycharm": {}
   },
   "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": "3.5.2"
  },
  "stem_cell": {
   "cell_type": "raw",
   "metadata": {
    "pycharm": {
     "metadata": false
    }
   },
   "source": ""
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}