tutorials/Cupy Example.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true,
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "# Compute on GPU: Cupy Vectors and Operators\n",
    "\n",
    "@Author: Francesco Picetti - picettifrancesco@gmail.com\n",
    "\n",
    "In this notebook we will show how to define a new `vector` class based on [Cupy](https://cupy.chainer.org/) operations.\n",
    "\n",
    "First we need to install `cupy` properly for our CUDA version. See its [guide](https://docs.cupy.dev/en/stable/install.html).\n",
    "\n",
    "1. `pip install cupy-cuda111` (change the version!\n",
    "2. `conda install -c rapidsai cusignal` (for operators)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Import modules"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "pycharm": {
     "is_executing": true,
     "name": "#%%\n"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING! DATAPATH not found. The folder /tmp will be used to write binary files\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/nas/home/fpicetti/miniconda3/envs/occd/lib/python3.10/site-packages/dask_jobqueue/core.py:20: FutureWarning: tmpfile is deprecated and will be removed in a future release. Please use dask.utils.tmpfile instead.\n",
      "  from distributed.utils import tmpfile\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import occamypy as o\n",
    "\n",
    "# Plotting\n",
    "from matplotlib import rcParams\n",
    "from mpl_toolkits.axes_grid1 import make_axes_locatable\n",
    "import matplotlib.pyplot as plt\n",
    "rcParams.update({\n",
    "    'image.cmap'     : 'gray',\n",
    "    'image.aspect'   : 'auto',\n",
    "    'image.interpolation': None,\n",
    "    'axes.grid'      : False,\n",
    "    'figure.figsize' : (10, 6),\n",
    "    'savefig.dpi'    : 300,\n",
    "    'axes.labelsize' : 14,\n",
    "    'axes.titlesize' : 16,\n",
    "    'font.size'      : 14,\n",
    "    'legend.fontsize': 14,\n",
    "    'xtick.labelsize': 14,\n",
    "    'ytick.labelsize': 14,\n",
    "    'text.usetex'    : True,\n",
    "    'font.family'    : 'serif',\n",
    "    'font.serif'     : 'Latin Modern Roman',\n",
    "})"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2D Phantom Deconvolution\n",
    "\n",
    "In this example we reconstruct a phantom CT image starting from a blurred acquisition.\n",
    "Again, we regularize the inversion by imposing the first derivative to be sparse.\n",
    "\n",
    "**Note**: as the vector is instantiated on a GPU, we use the method `plot` for getting a plottable array"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "GPU 0 - Tesla V100-PCIE-16GB\n"
     ]
    }
   ],
   "source": [
    "x = o.VectorCupy(np.load('./data/shepp_logan_phantom.npy', allow_pickle=True).astype(np.float32),\n",
    "                 device=0).scale(1 / 255.)\n",
    "print(x.deviceName)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 792x288 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "G = o.GaussianFilter(x, (3, 3))\n",
    "d = G * x\n",
    "\n",
    "plt.figure(figsize=(11, 4))\n",
    "plt.subplot(121), plt.title('Model')\n",
    "plt.imshow(x.plot(), cmap='gray', vmin=0, vmax=1), plt.colorbar()\n",
    "plt.subplot(122), plt.title('Data')\n",
    "plt.imshow(d.plot(), cmap='gray', vmin=0, vmax=1), plt.colorbar()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "pycharm": {
     "name": "#%% md\n"
    }
   },
   "source": [
    "As regularizer, we compute the gradient. If you need some signal processing operations, check out [cuSIGNAL](https://github.com/rapidsai/cusignal)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAApQAAAEMCAYAAAB6EiaPAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAAAyJklEQVR4nO3dy25cx9ku4LeUAwKBhBjBfxBpEoSeBwGj/wpCA7kAyrmCkHcgY48kjgLqDrpzBb+tUZDB3iB9BZY4z8AMNoKQQLag0CGtzPjtQa8iq4u1zqeqr94HIEx2N9mrl7tffXVcRkRARERERNTVg7kPgIiIiIjSxoKSiIiIiHphQUlEREREvbCgJCIiIqJeWFASERERUS8sKImIiIioFxaUAzHG7Bhjjo0xxyP9/YUxZjHG3x7C2K+/5bFsGWO+MsaId3vU55CI5jdUTjATKTcsKAciIqcAxvxgflV8TcIY86LN44vXfzTS4bQiIpci8jxw16DnsO05IqK4lHyGB8kJZiLl5odzH4Ayl2P9YRE5Getvl/jviZ9vdCOcQ3XniCgz9z7DM2TtbJiJNCT2UNI9xphJW9XGmK0pn28IU58jIhpWzJ9hZiKliD2UIzDG7ALYwqq19q2ILJ37tgDsAzgrbtoWkdfO7x0BeAvgGMDvAfxP8dgjABCRz4wxOwD+hNUQ+4fiefZF5KcNnmOn+FuXuBui/8weZ3EM2wB2iuGLS3v8xe8+dl7bcdsWrjFmG8ABgG/s+QFwVtxuX+fa6xeRN02euwi0b4tzcundZ183ROQz5zz9L+dYjkXkpM85IqJhGWP2cDd0/FxETovb/gTgSxE5qPgsl2XqJQKf4QY5cQlgS0TeFPcxE5mJZIkIvwb6ArAL4F9YFXDA6oMu3mPeYRVI9uc9AAvn532sAmALwA6AHedvHzvfbzu/8y2AvRbPsVf8TvA4i/u/Cry+d97z/Cvw+o9rzpH7vLsA3nn3l73+uuc+to8tft4OnPu147PPEzpvXc8Rv/jFr+G/yrLC+b7qs1yWKWU55+fEOy9bvm2RS8xEfmXzxSHv4X0QkTNgNREauBu+KFrVt7cX378BsO8McXywjxGRU1lN7C59jmKF3onctZibPMdlcXvwOCs8t89jdRiaeYziNWLV+vZ/v+z1lz530Xreds+VfW1livN06Z4nrHoBPi++v3T/TotzREQDk1XP23bxWbejOV8W39d9lptkapDtefN+53mTXGqBmUgqcMh7eFUf2m3cBYfrsrjPfvgrA8+GRBF2uyLyaYfnqAyXEh/scIbz+4/RbjHSEqsW92ushnm+CDwmWERXPPezlscArM6FPYfWV1g/L13OERGN449YZcYBVsWSHd5t8lluXER67LDzLa+4ZCYSFVhQTusMqw+8bwstP6hFq/ArAL8tft52/s4gz+H8XdsqfQendW6MuT0Wr1Vb5VsAdm7TH1v8XulzY/W6thv+HesMq6Gc3qscvXNERONYAvhbMS/QL3J6fZYrPsN12cJMDGAm5olD3sMLFXMAboeet+yHDbgdZnjTIkSsr7AKH9ty3QVw1uI5So8T68Mu2yJyZieAO+G15Tx+F839Bqsh+5Omr7nuuYsAtMfo/k6p4jxdeudp22udtzpHTV4LEXVT5MVbrOaDnzi3N/ksh9R+hp1suf1bxYblO8zEe5iJmWNBOZDiw/oFVvN89osrE9iViUfOh/Q3AA6MMXvGmH2sPnjPi7+xi9WQx44x5oU3H+YAwLPib+9jNaRxVvydFwCOnDCqeo7a4yxC6sz+LnA7zPNlcVy7xfP/ofhbZ/4xVpyqd1j1Mvyr+HpnH1/2+uueu/i7vwXw++I176IIPrO6EsRWyfH91jlPe8V5Oul6johodEdYLTbxlX2Wg5kC3ObK2me4IieeF1mwh6KQYyYyE2mdEZH6RxENoAilXSm2MCpu2wLwNYA/tJksT0SUOmYiacIeSprSM9yf4H6J1TAWEVFumImkxmg9lMUQ7x5Wq9N2ACw7zBMkZYrhkC3crUDcAnA6xERwIg2YnXlhJpIWYxaUx3K3+/42gC9E5GCUJyMiUoLZSUQpGmXI210lBtxuHfB5ycOJiAjMTiJK11hzKHcQ2FTVD0siIlrD7CSiJI21sXlor6oPuH9JqTUPHz6Ura3KhxDRCC4vL/Hx40fT9vd+97vfyfv37xs99t27d/9HRH7X+uDy0jo7mZtE87m4uHgvIv/V9vc0ZufsV8opJiTvA8CjR49wcMCpQkRTWywWnX7v/fv3+Oabbxo99sGDB590ehK6h7lJFIdXr1793y6/pzE7xxryDrWog9c3FZGliDwTkWcPHz4c6XCIaCwi0uiLGmmUncxNovRpy86xeihPERi64aWYiHQREdzc3Mx9GJowO4kyoDE7Rykoi2s/3/5cTCj/coznIqJ5pdSCjh2zkygf2rJzzDmUz4trTJ8C2OE+akQ6aQvFCDA7iTKgLTtHKyiLIRp7fVLu+E+klLZQnBuzMx8vX768d9vh4eEMR0Jz0Jads6/yprS5gegG4atXr9b+S3ppC0WiPkJF4hi/z8IzfdqykwUlBdlQqwutsvubFpLu41h8pkfjxHKipvoWj6H8bPo3yx7HQjMNGrOTBSUFi8epQilURLLITIu2VjZRyFQ9h01/v+x4OIyeDm3ZyYIyI2XD07GFTVkRyWH0OGkLRSKrroicMzvLnjt0zGXZT/PSlp0sKDPQdPg6dqGCksXl/LSFIuUt9aHkumF0Lf8eaKAtO1lQKqW5RVpWULK4nF5qV3Ig8lX1QmrJTvd12Nfrv24trzUVGrOTBSUR9aItFImIpqAtO1lQKuH3SObS2vR7JdljOT1tKxUpH2MvYFksFoP9LevgoN8+9/b1+a+dQ+HT05adLCgVGDMI3EDsG2Su/f19AMByuRzsbwKrItKda8miclwah21It6GHescoGrs8X9t89l+3PxTOwnJcGrOTBWWihp4jWVY4DllEusoKSVtoVj2mDudYTktbKJJOQ/VGtikgx8jPsuf3b+9aYIbmWLK4HIe27GRBmZAhP+B+ATlW4diWX0T2LTBDBSULy2ENFYrGmG0AeyiuYQ1gKSKXDX5vwetdU5m+PZJ1BeTU2Vn2fP5xdi0wQ0PiL1++ZFE5Am3ZaWKqkJ8+fSqxFDax6VtMjjV0PbWhCkwWlesWiwXOz89N29/79a9/LcfHx40e+7Of/eydiDwru98YcywinxXfbwP4oi7sjDE7AN6JSOtj14K5GTZWIZnKua4qhNu+Bq4IL/fq1avKXCujMTvZQxm5PvNZtBSRLreI7DIPk3tZDmuoy4cVIej+3TNjzOcA6t642wAuex8AqdK1AR4qwlLNTv+43dfW9t+Gw8ND9lgOTGN2sqCMVJ8eSRsWqQZhU7aQ7NJr6S/cYVHZ3UCjHDsIhJsxZltEzkK/YIzZE5E3xpg/DXEAlL4uuZl6T2RT7usJFZd1r9ffy5JFZX/aspMFZYT6hqK2IKwT6rX0by/jF5QsLNsbKBQfB277AGAr9OCiVR4MS8pPlyFZTb2RbYWKyzb/htgeSy7c6UdbdrKgjASLyGF0GRLnwp1+WoTiJ8aYt87PSxHpum/Ujoi86fi7pETbQjLnIrKMff1th8Q5DN6ftuxkQUlEvbQIxfcVE8tDLerHCA/l7AI4afqkREQx0padLCgj0LZ3Mpc5kn11mWPJIfB2Btyc9xSBoZuyOUAAPjfmdnHiljFmH8BJxeNJma65aTE/11XNsazqpbTcIXD2VNbTmJ3cNmhmbT6AHOLup82q8NxWgXfdNuhXv/qV/OUvf2n02F/84hd1W1+8E5HfFN+vbX1R/PwhtLeaMUa4bVBeedA0N1lE9tfm350c51R23TZIY3ayh3ImbT547JEcRpseS64Cb27ARulzY8wLFJvzevuoHQE4BnD7P8wYswVgv/j+CMCCPZS6dclNi/nZzcHBwb2FO3U9llwF3oy27GQP5QxYTM6v7TB46HtN+vRQ/vnPf2702F/+8pedWvJUjbl5H3NzHOytvK9PD6W27Hww9wHkxh2qqfqQLRYLhuKIlstlsMcyJIeCsis7D6jJF1FXzM04uJfprbskpT+/ktZpzE4OeU+k61ANQ3Fcy+US+/v7tfMrOQReLqXAo7Q0zU0Ob0/LDoPX/VvFIfBq2rKTBeUEmoQii8j5tNm7kqvA79MWihSHtrkJMDun5M6tBOpXg3MV+H3aspND3iNrMlTDYjIe7jB42VA4h8DX3dzcNPoiaqpLbjI7p2fPuzsMXjYUziHw+7RlJ3soiaiz1Ob4EBHFQGN2sodyJE27990J5Gxhx6Hs2uAuDnvf0TaxnOLA3ExL2cborrpFVbnRlp3soZwJh7nj5g99N9kMPVcpBR7Fra4RztyMmz/0XTenMvfiUlt2sodyBHWTyRmK6aibU+mv/s6RtlY2zaNuXh1zMx1N5lQC6yN5OdKWnSwoB1Y3mZxDNempGwLPeZGOiKibWE7TY27qUzcEnvsiHY3ZySHvgbBXUre6IfCc96lMqQVN8WJu6lM3BO7uU5kjbdnJHsoJMBT1aLKtUG60DdvQdJpueM3cTFvd/z87pzI32rKTBeUAcrlmKdXLcU6ltlCkaTSZM8khbj1CV9fx5TanUlt2sqDsiXN/8sM5leu0hSKNr2lukk6cU7miLTs5h7IjzpnMG+dUrqQWeBSPnHNzY2Oj8v7r6+uJjmRanFN5R2N2sqAkol5SWoVIRBQLbdnJIe8R5NTKtl8hV1dXt19acZGOvmEbGk+ui3DcrKzrnezy+NRwkc6KtuxkD2UHuS7CcYPt+vq6dlhmc3Nz7We3sPTv08od8tY87E3UV9WVVVLTd0jb/33/Zw1D4naRjv0+R9qykz2UA9O0CMdvIdsismuYbW5u3n4BUNN72eTa31o1bWFrC05qz/ZMal+EEyom3exskp91j9fWa1m2SEdzL6XG7GQPZUtl15rVNsztFpFj2dzcXCsqU+61rFqk4/dQauupTCnwaB5lhYGm3PRHcIbk/z1/KDzVHssm1/6uu757yrRlJ3soW9BeTLo9kn16Ittwey019FiWzanUvJWQtlY2jaOqINCQm9YUuRkqMFNW9v9fYxHp0pad7KEkAOO2rpvS1GOZE20rFWlYZYtwNMyZnDM33efzpyalqGxOpbuVkLYCU1t2soeyAXf3/rLeyZTnTPpzJOfkFpGp9laWzanUOOytcR4QURMx5aZfXFL8NGYneygzFkOvZIi/aIc9lXFLKfBoOlVz31LunYw1N+2xuNOWUuPOqQxteq6tl1JbdvYqKI0xRwC+BfAlgM8BfBCRN8V92wD2AJwC2AGwFJHLXkdLRNHRFopTYHYSkbbsHGLIewHgbwC2bCDa20XktYicAHgD4GiA55pc2Z6T7kXuU2xpx9rKdqW8vdByuVxboOPyL8uYOm3DNhNSnZ0hbm6mLtbcjPW42tLyPqmiLTv7FpTfiIgRkZ+KyGt7Y9HCviUiZ1i1wlXQsqqbxldWVGohIri5uWn0RWtUZ2fVnpNAurk55Q4YfVxfXyd9lZ2qVd/umoaUaczOQRblGGN2vJt2AFwGHrft3xYzdw6QpnkbMU0mr6NhkU6Ipm2EtLWyp6Q1O0PsvLhUi0malrvqWytt2dm3oNw2xuwCODPGHDnh+Djw2A8Atno+3+w0rOqmaWm/3re2UJyI2uzUtnDCSm2hi+0wSLWXsozt4NHSS6kpO3styvGGahYAjgF82uZvGGP2AewDwKNHj/oczmC0Xqs7hXmTIVpXfftzKVPtqUwp8GLRNztjzM0qKa/qpvlov963tuzs1UNpjNmy3xdzfeywTKhF/RiBoRwRWYrIMxF59vDhwz6HMzrNb2wal+brfWtrZU+hb3bGmJt2bpumRjiQ9jY8AFT2UgI6rvWtLTs7F5TFcM3XJXefIjB0UwRn1Ng7Ga+UV31XSXnVt8bNecemNTvLsHeS+rDTy7TNp9SYnX2GvN8C+KP9wRizh9UWFxCRM2MMnPu2sdpvLUlc1U1D8OdSur2WKUtpFWIkVGanpga4lXLPpMvtpdTweqzU33PasrNzQSkil8aYM2PMC6yGYz4VkefOQ54X950C2BERVmIz0RQk9nrfAFTNp0zZUC3oNht6F4tYdosf/xvAH8oeGxuN2Rka6uYUIRqS7aX0309VV2SKnbbs7Lso5xSrFxC67wyAnXh+0ud5iCheAw7JLETkMwAwxpxhtaH3vWqkmH/4zC5sKXr4vgbwm6EOZGzMTiLSlp2D7EOpRWjfSQ5309BC2wilui/lUPOAWm7o/QzAF87PJwB23IUuREPQMKpjpbDvcE40ZmevHkoiooFa2aUbevsLUkTkxBjjDhFvF7ff+30aX9lwNxvgNLTQNkL2vZfiDgPaspM9lMDapZz863UD3MSchhfaRshd5a2xl7JGqw29iyFj6/e4GyImGoTGrXYAva8rRdqykz2UimnYKihE60bnqWqxUvETY8xb5+eliPRa6l4M1ezY+UM0vVCv0FwN8PPz88r7nz59OtGR0FiqrvOdGm3ZyR7KCJ2fn99+hTx58gRPnjyZ+KiI7ms5D+i9FJtxF19uIDa+GILnCMDzmsfQCNyRHdcc+wVW5WXocU0eq6kR7tLyukLvs5Q2OteYndn3UIY2Mp9jIY4bcHWt6IuLCwBYKyrtbZSOsn0p/Usxxj78PdA8oNYbehdb63xRbMOzxTmU+fELw7rsdB9/fn7OHkualbbszL6gjEGbYtLlFpZPnjxhUUmzGCIU6zb0Ln7+YIPPbgbuBOEuis3BaRox7TvZNDfdx9meShaVaQrtS5nasLe27OSQtyfFbYL8wpLSEtpGKCUDTSwHig29i0sT7nkbeh+h2AqjCMivAHxrjBFjjBT300RiGFq0xWDXgtD+XpthcIpf2VSMGGnLTvZQElFnIjLY5cOkYkNvca4kUzzOgIgoURqzM+uCMoZLNtlWcd9hF3e42/ZS2ksUUnr8OZQxz6McaB4QJW7KEZ2hehOfPn3KeZUJS2UUsYy27My6oJxT13mTTVxcXODJkydrQzpEY9EWilQvho3Mh8rN0LzKIf8+jatso/MUaMtOzqF0aNrI3O2xZDDGL7TReSoGnAdE1MhYmcaspClpy04WlETUi7ZQpGqhBQ9Tbq829oiLXejDkZ00lHUApbAwR1t2ZllQll1qcSpjDnf7z2Ofi6u/05PCXpQtN+clSsbV1ZXKorLL3HqN52FuGrMzy4IyF9fX17i+vsb5+Tk2NzfVbCtkL7kIoNVlF1lcj+Pm5qbRF+lQtv/kFPpsE9TluYBpekVj426lZF977Nsr+e/DFOZSastOFpRYvRHn3JR3Cm4I5FJQ+eFn/4Gwi5ZiK7CXy+XanpSp0NbKJgJWjVWNcyrLGuFVBaN/HmItKlOjLTu5yptUaTqdgFcZGk5KgUfdlU0T0toIt9yeSo0FJtDsEpb+avjYlM2jjLmnUlt2ZtdDGbp2dy4uLi6C1wHXwt3uQ1Pwx7wfpcZ5QNSMO7KjlT/fMMZCqil3qpDLz80m2RnroqWU3o8as5M9lETUS0qBR0QUC23ZmX1BmeK1u9u6vr7GxsYGNjY2AKzPIbQ/p8JtYdu5QH1Wzcd4LpbL5dp1vd09KmOkLRQpLLcRHV+svXJdTbXbyJxif89qy87sC0pKWw6hGLMhr0dLaZm6AT7HHEZ/AYtbVKaWN2WLcbq+jhhff0qdQhqzM7s5lHOLZZ5f6NrfRF1omwdE96WwSfRY6uYepiA0F3SIf4Ni3kbIcvedjo227MyqoHRXKcbeFT606+vr2+/doe+UFum4e09ubm4OVpynUFy7i3FSXZhD1EcsPWJzdwb0FXsBmBNt2ckhb0oOh7njklLgUTehzcxTGl7sww4VX11drQ0b2+Hv2DPIHvcYuRnja/f3lI6580hbdmbVQ+lz33g5hKO9cg5w10sJ3N+TMTbukJM7dBNjmA3FXYgT+ybn2lrZRGW6XLZwLmVD9ZpzMzXaspM9lBnyV327w+Gx8Vd1pxToOdA4sZzq5dAA94XyJ4VFOrZ3MtbjG1oq702N2Zl1DyXdSWEeYSwLmmidtlY2xSmGXLJFZewNWzvMXbaym+KgLTvZQ0lEvaQUeEREsdCWneyhnJh7Xdg5V9uFVn3Hxl/VTXHS1sqmdbFuuTI3m092xCSW1dNlWwTlLsb3sbbszKKgdPehinnF19T8RTqxbCPkDinZ+T+xz1Oagnst71i2DtJ4PVqqNtf1kmO4ihWw3sCNaejbz83cxX5db43ZmeWQdw6XW2zDX6Qzp9ClFXNlV3rbyzDGegnGlAKP2mMjPMyfTznnNkJVl6TNuRHuivF9rC07sywo6T5bVAJ3E7qfPHkyaa9AqHXNPSfjp22lIlVjI/yOu/K7TW9lXWO5zd9ir2RYCu9TbdnJgpJuucPfTbbocQMs9Njvvvvu9vtHjx6V/h32SqZNWyubqA1/43N/A/Syxzf9myH+7f7f5LzJNGjLThaUdM/19fVaONrbXE3mED169Oi2qPzuu+/uFZV1LWsO2cQvtTk+RGOxeekXc6F8a9NYr/s9NsLTpDE7WVBSkB3q9sOq7ybo7I3UR1soEnUV2pGiy8UZ3Me0LSLZO5kObdnJgpIa8S/ZuLGx0aq4/P777/HgwYNGRSTnTaZFWygS9eFfPccvCN2pQHX8UZ3YGuF2N5BYVuCnRlt2sqCkUhcXF7eB4S/Qubq6alRQ2kB88CDeHarcLZIYjO1pm1hO8UqxgKmba16lbj6mi1OE0qMtO1lQElFnGucBERGNTWN2sqCcgTskktp8F3fou8mwd9s5kxzuTo+2UKQwu39vCtuxzK0qu6p2vPDFtHF6SOw9xYvFIur3q7bsjHccciTc1JxS429wHhttV3ugeLlX80qBfznGNoty2gx3A3eXgKR0aMvO7ApKaqfscox+T2UIV3TnQVso0op7yVpg1QBnI3wYtmD0i8vQbU3Ech3x2Pjv19iu560tO9UXlO4bKMZLL6XMXgvcvXSjG5R2C40uk8rZ0i7nXs87hmt6awtFoqGEciy0rZBfSLbJzbm4HQzUjbbs5BxKGoTtsYw9BH0prhqNiYgMtlLRGLMNYA/AKYAdAEsRuez7WOom5gb41JeF7SM0T75s5XfX/GQDvJmY3tMas5MF5UxsAJyfnyexMMduIcQCjHwDtqAXIvIZABhjzgAcASgbY23zWKJo9W2Ez/XvB/8N6E9bdqof8qb42aIaYEs7RUMM2xStZvdvngH4vO9jSaeUihlmGpXRlp21BaUxZscY8y50YMaYF8aY3eK/W03uIyJdBpoHtAPg0r/RD8AOj50Ns5Mst9E8tDkKVs6fHIa27Kwc8jbG7AL4UByEr6rblMNRRBkYcNL448BtHwBs9XzsLJidRFRFY3ZW9lCKyImInPq3V3WbcjhKL3eYKfUWKueDDqdFK/sTY8xb5yu+TTUHwuwcX+oZ1NeYvZ40DW3Z2XVRTlW3ael9RUCSw141J9c5hLwyTvparFR8LyLPSu4LtZIfI5AlLR8bG2ZnhrRlGxvhw9CWnV0Lyqpu0+iHo4hoOAMN25wikB0lhVSbx8aG2TmQ3IsabUVqjrRl5+yrvI0x+7Yb9+PHj3MfDhG10HTIpi44/UAreuy+dH+2C1TqHpsD5uaKO3UlBakPUad2vmOmMTu7FpRV3aatulRFZCkiz0Tk2cOHDzseTtrclmbsgVN2KcaUuMedey/HEAZaqQgAz+0KZwB7IuIuRjnC+nzCqsfGbJDsZG7mLfZ/J6gZbdnZdci7tNvUGIOy+zo+FymV67xRbQYatrEZ8br48cS773nTx0aO2Tmg1BqEoazb2Ni4/d5ecSxWqZ3v2GnLzk49lFXdphyOylPoWt6Uh5ubm0ZflFZ2xnKteI1sXrrFpHt7He492U1s72dt2dlkH8qd4vsjAMciYiva58aYFyiuB+l1m1bdN6nDw0O8fPkSAPDy5UssFou5DkWdzc3N28IxdC3vIa5ROzQNWwXZUIwhHAfcS00VDdmZkpSu7e0XjG6vpL3P/jf2HkvqTmN2VhaURQCe4K571L2vqos11eEoImpJWygOgdlJRHW0ZWfXOZQ0MDuEYfekjH1eoe1xrGpBuz2YV1dX0fRS0rC0hSKtxNADrtE//vEPPHr0KHifzdMmw95TSqkHuEps72lt2Tn7tkGUJrugJqVtJDQMd8dowJWKRJ1cXFwkkUMbGxv4/vvvax93fX2N6+vrYGHJFd56aMtOFpSR8XsqY+NOGm+y8GZzc/O2Z9J9PFd466EtFInGlGLmsRE+Dm3ZyYKSiDoTEXUrFSlNKRQ9bY8x9PipC9IUen5TpDE7WVBSK3YoBmi2ctvdQmiuOZTusHwK/+ikRlsrm8L29/exv78/92FUSmkKDo0v9vertuzMrqA8OLjbhSPWLYRSunIOjc+G4nK5xHK5nPlo7tMWikRjcRcqVrEN8TkXMtrinI3w8WjLTq7ypk78XspQ6MSwDyUvszi+lAKPuouxMeNzLwsb8+edF36YRuzvWW3ZyYIyUilsI1QXitwmSL/UWtBEc9vc3MRf//pXAOXzIefOztgLcg00ZicLSurM3aojxgDivMlpaAtFSp/Nplg/+7ZgnLtwpHlpy04WlETUS0qrEImIYqEtO7NblJMad+ibKEbaJpbTutiuLqLB06dPo5zGBMQ52jSEGN/H2rIzix7Kw8NDAMDLly/x8uXLaFd30zC0LsRxAzGWcEwt8Cgfmj77pI/G7MyioCSi8WgLRVoXS+OlC629bWPQPuc8xvextuzkkHcCYr8cY0y09k7GTNuwDVWLfbNoHzc7r6fx/KTwPtWWnVn2UNrNzReLBRaLxdpm57F6+vTpbTEZ6zZCc9PWwnYDMeb91LRNLCc93J0oqJqW3EyJtuzMsqAkomGk1oKm/mJu3ISkstn5XLSel9jfpxqzk0PeCXFXBnLo+w6v1T0vbcM2dF+M88+on1ymArx69Sra96+27MyqoPRXe1P6cpkzGeMKb0tbKJJOWoa/h3gNueRm7LRlJ4e8iaiXlAKPiCgW2rKTBWWC7AIdO+yd4gId9zV0PX4Oc8dBWyjSfX6v+P7+fvRz1Kia1ty0ixnt+zO2ER2XtuzMasjb56/2JopFKiu8RQQ3NzeNvojmZoe9Ux767lsIal2EkxqN2Zl1QZkyt1cvtwU6XIQTF23zgCjMXdwQcyOnqTmLyr5Fbdvftc+XciHd1HK5TKJ3EtCXndkVlHZhDoDkF+bMveq778Tuthu220C8uLi4/cqBDcVYVytqC0XSTUt2NC0Q/cdoeO1aaMtOzqGkWTXZsJ0rEuOWUuBRd6F5lEC6vZV2+HuOTOn7nO6KdX+0JlRk5pKb/tzeGBvgLm3ZmV0Ppc+9Sk6q8yhTvzSj7Wl1i0tgPSi19Co04f5DHfs/1k1b2NqCk/RIdSjYz8Sy15FLbqZGY3ZmX1BqoWVOpfs6GIRp0BaKVM2dRxl7g6eOX5BNYejn8QtLtwGeW4am1DsJ6MvOLIe8/Q3OU+2Z9Pk9lWNuJ8RFMeOLeTNzV0qrEIl87lBxyiugUz3unGnLTvZQKpT6EDilRVsrm6qVzaVMnV9Yjv08NCz/fRhzI9zSlp1Z9lAS0TBSCzwiohhozE72UGK1MMfd5FwDLXMqc7K/v5/kytmpW9nGmG1jzAtjzG7x362Kx+4Uj3lhjPmq6rFEY8+pTHHxD41HW3ayoFRsjH0quak4+WYYtlmIyGsROQHwBsBR6EFFAD4rHvsawP8A+HrIA8mVO5yYUuOnCXcxCwvAdLibmacw3A3oy86sC0p/cY5WnFOZnlQW5IhMe/kwY8y29/xnAD4vefgzAF84P58A2GEv5fC0zKN0DTmnMvUFPzFzR3ZSojE7sy4ofZqv7T3EEDg3GB9PKtfuDpm4lb0D4NK/0Q/L4rhOADx3btoubr/3+9ROzI2cITXZ57EOezmnxR7KUqNnJxflZKSsqGyyvRCHuanMxBPLHwdu+wBgK/RgETl1fvw9gNcjHFOW3EuCptYIasMf/m7SsOblDqfj7z2ZSjEJ6MvO7AvKw8PD2+FuTXtS1gkVl35hyR7Jebj/UKegRSh+Yox56/y8FJFJKpFiqGZHRD6b4vlylOKCsjZC8yqb9D4yO8fjX2oxNdqyM/uC0ndwcHBbVC4Wi7VLM2pV1kPJIJxG6v8QtwjF9yLyLHSHMWYfwKcVv3tcDMOEWtSPERjK8RxhfQiHekqlwTO0JivBmZ3zSO09qS07WVASUWdDzfFp0do+RWDopphgHmSMeQHgCxG5NMZscQ7lsOwwY6oNoj5YOM4r9eFubdnJRTlE1MuUKxX98CsmlH/p/uyuRDTG7AF44wTh7iAHQkTUk7bsZEGJ1TxKdwshojm4LewUW9oT7qX23G7OC2BPRNx5KUcotsIoAvMrAN8aY8QYIyjZd42Gk+IWLpQeDe8zbdnJIe8Af/ugHOZR0vRS3irINfFKRdvStisOT7z7nnuPMxMeWpZsQyilRhDpkWIj3NKWneyhdOSy0TnFJ5WNzH1NW9hTByfNZ7lcqug9ojjZjcxTboQDOrOTPZQV7Ipvu+qbPZU0FC29k8D0rWyKT0qNINIl5feetuxkQUlEvQw1aZzSlstG5zSvlFd2+7RlJ4e8PXbYG+ACHZqG+w9xiuGobdiGhpHqNZYpXtreT9qykwVlDXeYO5er6NC43I3MU+/J0TgPiLrh4hyaUurvNY3ZWVtQGmN2jDHvArcfGWP2jTFbxX/3nPu27dL04r9bAx/3qLiNEE0h1W2CfNpCcSg5ZqfLNpi09SrRPDQsxPFpy87KgrLYqwgAdkoesgDwNwBbIvLGvV1EXsvqcj9vkPjebwcHB2tbCRF14Q4BagpGbaE4hJyzM/WeI4qbpp5wbdlZuSinCDUYE9yO6BsRuXdHsSGm+zfOjDGfA+ASaSKFUgq8qTA7iaiOtuzsPYfSGOO3wHcQuNi4H5Yp4L6UNJZU9530iciklw/TRHN2ujjsTX1o2XfSpzE7+xSU28WwzlkxJ8iG472LjwP4AGCrx3NFgcPe1JWmfSd92oZtJqA+OzUNS1I8tL2vtGVn54LSzvOR1YXDF1hd97G1YlL6W2PM248fP3Y9nNFwGyEaWurbBPm0heLYhsjO2HPTx15K6sL2TGprhFvasrNzQemuPpTVdR/tsEyoRf0YgaGc4neXIvJMRJ49fPiw6+FMxr/ON1EdTdsEhWgLxbENkZ0p5GaoN4l7U1JT/vtESwPcpS07OxWUxXDN1yV3nyIwdFMEZ5L8bYS4NyU15RaTWrYJ8mkLxTHllp0ujY0pGpf294y27Ox66cW3AP5ofyj2UXsD3K5MhHPfNoAvexwjEUUqtcCLQHbZ6fZU2gJB4yKLsWxsbNx+f319PeORTMd9f2hshAM6s7N2H0pjzIvi+yO7t1ox9+es2Hh3H8B/i8hz51ef2815AeyJiIptL/xeSi7SaW5jYwMbGxu4urq6/dLM33NSy6ruEG0rFYfA7LxP2/t+Cm4xGfpZO+3vGW3Z2WQfyhMArwP3nWI1RBP6vTPnd056HmNUDg8P17YRYjFZzW9db25uAsBtUWl/1kxrC9vS1soeArNznT//zS7S0dZLWVfwhXoY63og3du0F5ShCz9ozU1AX3Z2HfImIgKgLxSJiKagLTt7b2yeO42rvu3wdEjZkHVuQ9pVNO856Ws6qVxbcFJ72ld9N+k9rBvC9n/2eyw1z6HMYVW3S2N2sqDswN+bUsuqb7eQLAuuzc3NtWFr/3f8+3PjbxGkfbgb0LdSkaahsbF1fX1d+QWEi0b3/pxp3VqtjLbs5JA3dbK5uYmrq6vSkKR8pDRpnOblb+qvadV3k4Kwrscx16LS//+vvXfS0padLCg7KrvOtx36dnstNbK9jywkV/xhbs2run0ptaApHqHh75SLyjYjMu4ojrswMZeFilZomDsn2rKTBWVP/qpvYH0+ZUqFZZvWsQ09t4dyY2MDFxcX9x6jXe7FpLZQpPHVzadMubCs4+dijlODgLxzE9CZnSwoB2CLSgD3thPKobcSuCsscykiLf8fwBzmTPq0hSJNo2wroVR1yT47dajsb2jvwfQbDjnlJqAvO1lQ0qjYg6mftlCk6WieU9lUXQ5q7MHMdc6kT1t2sqAcSO5zKuto68HMfbjGpS0UaR7a5lQOoa4HMzW5z5n0actOFpQD0zSnksJYTN4REXUrFWl6Oc+prJN6EWkxN9dpzE4WlETUi7ZWNhHRFLRlJzc2H4HWjc+peuPyHFvZgL7NeWke/ufH7ZVMebEO3c9NK+fcBPRlJ3soJ+JfopFD32nJ6XKKbaUUeBS3qkU6nFOZnrLczL2QtLRlJ3soR3J4eLi2UMfSeO1v7UKh6AZi7sGorZVNcfC3FAJ0Xftbu7rcJH3ZyYJyZKHV3xwCT0doqCb3yeSupoGYUijS/EJTSTgEng7mZj2N2cmCcgL+nEpgVVS6vZUUl7IVppwzed/NzU2jL6I2QgWIW6CwqIwPc7MdbdnJgnIiHAJPB4e429HWyqZ4lH3uOAQeH+Zme9qykwXlxDgEHjcO1bQ3dSgaY7aNMS+MMbvFf7ca/h4/XIniEHjcmJvdaMtOE1P1+/TpU8ll9bPbS+kOiXMD9HlUrUZ0/6vVYrHA+fm5aft7P/7xj+XnP/95o8f+/e9/fyciz1ofnMcYcywinxXfbwP4QkQqPzDGmB0A70Sk9WuMXU65WVakcAP0eeSemwDw6tWrTrmmMTvZQ0lEvUzZyi5C0H3uMwCfN/jVbQCXgxwEEdEAtGUnC8qZcE5lPDj3p5+JJ5bvIBBuflh69+2JyJuhDoDmwzmV8WBu9qctO1lQzoxzKufFuT/9TTwP6HHgtg8AtkIPLsLybKgnpzg0mVPJwnJ49rwyN4ehLTt5pZwIHB4e3haTL1++xOHhYbCozGWe1NiqrnrD1nU7LQPvE2PMW+fnpYiMPelth72TOlX1htnPOK+uMxy/QHcvPWsxN5vTmJ0sKCMR2qvS3n5wcLA2BM7CsruqPdL876mZFqH4vmxiuTFmH8CnFb97LCInCLeoHyM8lLML4KTpwVF6/IIyVFhywU5/dbkZ+pnqactOrvKOVGgVuDv8zfPUXt2Guznrusr7Rz/6kTx+HBpJue+f//xn75WKxTDMVyLyG+e2f4nITwOP3cVqQrm1AHAA4KSYkK4Cc/NOqHEY6lmjelXnjbl5p+sqb43ZyR7KSPlzK8uGwQEWl1XqtrXwv6f2pmyUisiZMXd1bxGSX3o/fxCRy6JV7j52McEwEc0oNLfScofBLRaX94XmnnJ4exzaspMFZeRCK8E5x7Ia50hOR0TmuDTYc2PMCwCnWM3zcd/4RwCOAdz+jy82790vvj8CsNDUQ0n3NZlfab9nUbnSpEey7GdqT2N2csg7IWWboQPssbRYTHbTdcj7hz/8oTx69KjRYz98+DDI5ry0jrlZraoY4nB4+TlgEdlM1yFvjdnJHsqE1C3csXLrsWxSRPrf03BiapQS+Zos3Al9D+gsMMu2UyorJMtuo/60ZScLSiLqRVsoEhFNQVt2sqBMVKi30t7m90xq67H0eyTZKzkvbaFIOoV6KKt6K/2fU+6trNrknUPc89GWnZxDqUTV/Eor9XmWTfaTYzHZTdc5lD/4wQ9kY2Oj0WP//e9/JzEPKDXMze7qhnebFGIx6lJAWszNdrrOodSYnSwoldJQYLKAnE6fgvLhw4eNHnt9fZ1EKKaGudlPVW40LS6tOYrMtsfFQnJYfQpKbdnJIW+lqobErbJ/hKYeIi8bVioLZxaRcYmpUUrUVtVQb1W+hAq5uuKuS8HZ5ZrkVdOAqm6jaWnLTvZQZqRJr2Wdrr2aQ1z+jNv+jKdPD+VPfvKTRo/9+PFjEq3s1DA3x1OXNXVzMKfQpHhscz+106eHUlt2sqCk0h7MubAHcnpdC8oHDx40DsX//Oc/SYRiapib02iaRW17Nes0GampwxwdT9eCUmN2csibgoXkVEVmqHhk+KUlpkYp0ViaDhs3nZPZ5febYH6mQ1t2sqCkoKaFZNkwetPikOGXvhkuH0YUhb4FYZ/8Y3amT1t2sqAkol60tbKJiKagLTtZUFIvZT2ZbD3nQUTUhSJRH317KJmdedCYnSwoiagXbaFINBUWj3nTlp0sKImoF22hSEQ0BW3ZyYKSiHrRFopERFPQlp1R7UNpjLkC8Ne5jyMCnwB4P/dBRIDnYbpz8AsR+a+2v2SM+d9YHWMT70Xkd22fg6oZY/4fgO/BzwrAzAB4Dixm58RiKyjfprB559h4HlZ4HngOqBm+T1Z4HngOLJ6H6T2Y+wCIiIiIKG0sKImIiIiol9gKyvBFS/PD87DC88BzQM3wfbLC88BzYPE8TCyqOZRERERElJ7YeiiJiIiIKDFRFJTGmG1jzAtjzG7x3625j2kKxpgjY8y+MWar+O+ec5/ac2KM2THGvAvcXvqaNZ6PivOQ5fuC2sv1/ZDrZ4TZucLsjJS9nuScXwCOne+3ASzmPqaJXvcRAAHwLwAvcjgnAHYB7Kzees3fB9rOR815yO59wa9uX7m+H3L8jDA7G52H7N4XMX3N3kNpjNl2fxaRMwCfz3Q4U/tGRIyI/FREXtsbNZ8TETkRkVP/9qrXrPF8lJ2HQnbvC2ov8/dDdp8RZucKszNesxeUWLU0Lv0b/TeAZsaYHe+mHM9J1WvO8XzwfUF1sn8/8DMCgNl5D98X84ihoHwcuO0DgK2Jj2MO28aYXQBnxdwP+yHI8ZxUvebczgffF9REzu8HfkbuMDvv8H0xox/OfQA587rkFwCOAXw63xFRDPi+IKrGzwiF8H0xrxh6KEOthMcIdE9r464yK+Z02O73HM9J1WvO6nzwfUENZft+4GdkDbOzwPfFvGIoKE8R6I4u3gxqFd3yX5fcneM5qXrN2ZwPvi+ohSzfD/yM3MPsBN8XMZi9oPT/hxaTZL+c6XCm9BbAH+0PxX5Zb4A8z0nVa87sfPB9QY1k/H7gZ8TB7LzF98XMorj0YvE/dw+rVsSOOw9Cs2LC8C5W3e6fisgXzn0qz0nRitzBar+w11jtDXZS3Ff6mrWdj5rzkN37grrJ9f2Q42eE2bnC7IxXFAUlEREREaVr9iFvIiIiIkobC0oiIiIi6oUFJRERERH1woKSiIiIiHphQUlEREREvbCgJCIiIqJeWFASERERUS8sKImIiIioFxaURERERNTL/wdrRd8ReX1GCAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 792x288 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "D = o.Gradient(x)\n",
    "Dx = D * x\n",
    "\n",
    "plt.figure(figsize=(11, 4))\n",
    "plt.subplot(121), plt.title('horizontal gradient')\n",
    "plt.imshow(Dx.vecs[1].plot(), cmap='gray'), plt.colorbar()\n",
    "plt.subplot(122), plt.title('vertical gradient')\n",
    "plt.imshow(Dx.vecs[0].plot(), cmap='gray'), plt.colorbar()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Problem and solver:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "##########################################################################################\n",
      "            SPLIT-BREGMAN Solver\n",
      "    Restart folder: /tmp/restart_2022-04-22T01-47-39.563843/\n",
      "    Inner iterations: 3\n",
      "    Solver iterations: 10\n",
      "    L1 Regularizer weight: 1.00e-03\n",
      "    Bregman update weight: 1.00e+00\n",
      "    Using warm start option for inner problem\n",
      "##########################################################################################\n",
      "\n",
      "iter = 000, obj = 6.57986e+02, df_obj = 6.58e+02, reg_obj = 0.00e+00, rnorm = 3.63e+01\n",
      "iter = 001, obj = 1.56466e+00, df_obj = 2.16e-02, reg_obj = 1.54e+00, rnorm = 1.26e+01\n",
      "iter = 002, obj = 1.59651e+00, df_obj = 7.26e-03, reg_obj = 1.59e+00, rnorm = 1.33e+01\n",
      "iter = 003, obj = 1.60710e+00, df_obj = 4.03e-03, reg_obj = 1.60e+00, rnorm = 1.37e+01\n",
      "iter = 004, obj = 1.61852e+00, df_obj = 2.70e-03, reg_obj = 1.62e+00, rnorm = 1.39e+01\n",
      "iter = 005, obj = 1.62172e+00, df_obj = 1.99e-03, reg_obj = 1.62e+00, rnorm = 1.41e+01\n",
      "iter = 006, obj = 1.62525e+00, df_obj = 1.55e-03, reg_obj = 1.62e+00, rnorm = 1.42e+01\n",
      "iter = 007, obj = 1.62453e+00, df_obj = 1.27e-03, reg_obj = 1.62e+00, rnorm = 1.43e+01\n",
      "iter = 008, obj = 1.62431e+00, df_obj = 1.06e-03, reg_obj = 1.62e+00, rnorm = 1.44e+01\n",
      "iter = 009, obj = 1.62156e+00, df_obj = 9.12e-04, reg_obj = 1.62e+00, rnorm = 1.45e+01\n",
      "iter = 010, obj = 1.61953e+00, df_obj = 7.95e-04, reg_obj = 1.62e+00, rnorm = 1.46e+01\n",
      "iter = 011, obj = 1.61583e+00, df_obj = 7.05e-04, reg_obj = 1.62e+00, rnorm = 1.46e+01\n",
      "iter = 012, obj = 1.61266e+00, df_obj = 6.31e-04, reg_obj = 1.61e+00, rnorm = 1.47e+01\n",
      "iter = 013, obj = 1.60840e+00, df_obj = 5.71e-04, reg_obj = 1.61e+00, rnorm = 1.47e+01\n",
      "iter = 014, obj = 1.60475e+00, df_obj = 5.20e-04, reg_obj = 1.60e+00, rnorm = 1.48e+01\n",
      "iter = 015, obj = 1.60033e+00, df_obj = 4.78e-04, reg_obj = 1.60e+00, rnorm = 1.48e+01\n",
      "iter = 016, obj = 1.59643e+00, df_obj = 4.42e-04, reg_obj = 1.60e+00, rnorm = 1.49e+01\n",
      "iter = 017, obj = 1.59186e+00, df_obj = 4.11e-04, reg_obj = 1.59e+00, rnorm = 1.49e+01\n",
      "iter = 018, obj = 1.58772e+00, df_obj = 3.84e-04, reg_obj = 1.59e+00, rnorm = 1.49e+01\n",
      "iter = 019, obj = 1.58311e+00, df_obj = 3.60e-04, reg_obj = 1.58e+00, rnorm = 1.50e+01\n",
      "iter = 020, obj = 1.57901e+00, df_obj = 3.39e-04, reg_obj = 1.58e+00, rnorm = 1.50e+01\n",
      "iter = 021, obj = 1.57454e+00, df_obj = 3.21e-04, reg_obj = 1.57e+00, rnorm = 1.50e+01\n",
      "iter = 022, obj = 1.57051e+00, df_obj = 3.04e-04, reg_obj = 1.57e+00, rnorm = 1.50e+01\n",
      "iter = 023, obj = 1.56604e+00, df_obj = 2.89e-04, reg_obj = 1.57e+00, rnorm = 1.51e+01\n",
      "iter = 024, obj = 1.56200e+00, df_obj = 2.76e-04, reg_obj = 1.56e+00, rnorm = 1.51e+01\n",
      "iter = 025, obj = 1.55766e+00, df_obj = 2.63e-04, reg_obj = 1.56e+00, rnorm = 1.51e+01\n",
      "iter = 026, obj = 1.55372e+00, df_obj = 2.52e-04, reg_obj = 1.55e+00, rnorm = 1.51e+01\n",
      "iter = 027, obj = 1.54949e+00, df_obj = 2.42e-04, reg_obj = 1.55e+00, rnorm = 1.52e+01\n",
      "iter = 028, obj = 1.54560e+00, df_obj = 2.33e-04, reg_obj = 1.55e+00, rnorm = 1.52e+01\n",
      "iter = 029, obj = 1.54148e+00, df_obj = 2.24e-04, reg_obj = 1.54e+00, rnorm = 1.52e+01\n",
      "iter = 030, obj = 1.53773e+00, df_obj = 2.16e-04, reg_obj = 1.54e+00, rnorm = 1.52e+01\n",
      "iter = 031, obj = 1.53377e+00, df_obj = 2.09e-04, reg_obj = 1.53e+00, rnorm = 1.52e+01\n",
      "iter = 032, obj = 1.53010e+00, df_obj = 2.02e-04, reg_obj = 1.53e+00, rnorm = 1.53e+01\n",
      "iter = 033, obj = 1.52622e+00, df_obj = 1.95e-04, reg_obj = 1.53e+00, rnorm = 1.53e+01\n",
      "iter = 034, obj = 1.52267e+00, df_obj = 1.89e-04, reg_obj = 1.52e+00, rnorm = 1.53e+01\n",
      "iter = 035, obj = 1.51894e+00, df_obj = 1.84e-04, reg_obj = 1.52e+00, rnorm = 1.53e+01\n",
      "iter = 036, obj = 1.51552e+00, df_obj = 1.79e-04, reg_obj = 1.52e+00, rnorm = 1.53e+01\n",
      "iter = 037, obj = 1.51192e+00, df_obj = 1.74e-04, reg_obj = 1.51e+00, rnorm = 1.54e+01\n",
      "iter = 038, obj = 1.50863e+00, df_obj = 1.69e-04, reg_obj = 1.51e+00, rnorm = 1.54e+01\n",
      "iter = 039, obj = 1.50513e+00, df_obj = 1.65e-04, reg_obj = 1.50e+00, rnorm = 1.54e+01\n",
      "iter = 040, obj = 1.50192e+00, df_obj = 1.61e-04, reg_obj = 1.50e+00, rnorm = 1.54e+01\n",
      "iter = 041, obj = 1.49856e+00, df_obj = 1.57e-04, reg_obj = 1.50e+00, rnorm = 1.54e+01\n",
      "iter = 042, obj = 1.49549e+00, df_obj = 1.53e-04, reg_obj = 1.50e+00, rnorm = 1.54e+01\n",
      "iter = 043, obj = 1.49227e+00, df_obj = 1.50e-04, reg_obj = 1.49e+00, rnorm = 1.54e+01\n",
      "iter = 044, obj = 1.48930e+00, df_obj = 1.46e-04, reg_obj = 1.49e+00, rnorm = 1.55e+01\n",
      "iter = 045, obj = 1.48619e+00, df_obj = 1.43e-04, reg_obj = 1.49e+00, rnorm = 1.55e+01\n",
      "iter = 046, obj = 1.48331e+00, df_obj = 1.40e-04, reg_obj = 1.48e+00, rnorm = 1.55e+01\n",
      "iter = 047, obj = 1.48033e+00, df_obj = 1.38e-04, reg_obj = 1.48e+00, rnorm = 1.55e+01\n",
      "iter = 048, obj = 1.47757e+00, df_obj = 1.35e-04, reg_obj = 1.48e+00, rnorm = 1.55e+01\n",
      "iter = 049, obj = 1.47468e+00, df_obj = 1.32e-04, reg_obj = 1.47e+00, rnorm = 1.55e+01\n",
      "iter = 050, obj = 1.47200e+00, df_obj = 1.30e-04, reg_obj = 1.47e+00, rnorm = 1.55e+01\n",
      "iter = 051, obj = 1.46921e+00, df_obj = 1.27e-04, reg_obj = 1.47e+00, rnorm = 1.56e+01\n",
      "iter = 052, obj = 1.46662e+00, df_obj = 1.25e-04, reg_obj = 1.47e+00, rnorm = 1.56e+01\n",
      "iter = 053, obj = 1.46392e+00, df_obj = 1.23e-04, reg_obj = 1.46e+00, rnorm = 1.56e+01\n",
      "iter = 054, obj = 1.46142e+00, df_obj = 1.21e-04, reg_obj = 1.46e+00, rnorm = 1.56e+01\n",
      "iter = 055, obj = 1.45883e+00, df_obj = 1.19e-04, reg_obj = 1.46e+00, rnorm = 1.56e+01\n",
      "iter = 056, obj = 1.45643e+00, df_obj = 1.17e-04, reg_obj = 1.46e+00, rnorm = 1.56e+01\n",
      "iter = 057, obj = 1.45396e+00, df_obj = 1.15e-04, reg_obj = 1.45e+00, rnorm = 1.56e+01\n",
      "iter = 058, obj = 1.45165e+00, df_obj = 1.13e-04, reg_obj = 1.45e+00, rnorm = 1.57e+01\n",
      "iter = 059, obj = 1.44924e+00, df_obj = 1.12e-04, reg_obj = 1.45e+00, rnorm = 1.57e+01\n",
      "iter = 060, obj = 1.44699e+00, df_obj = 1.10e-04, reg_obj = 1.45e+00, rnorm = 1.57e+01\n",
      "iter = 061, obj = 1.44467e+00, df_obj = 1.08e-04, reg_obj = 1.44e+00, rnorm = 1.57e+01\n",
      "iter = 062, obj = 1.44252e+00, df_obj = 1.07e-04, reg_obj = 1.44e+00, rnorm = 1.57e+01\n",
      "iter = 063, obj = 1.44032e+00, df_obj = 1.05e-04, reg_obj = 1.44e+00, rnorm = 1.57e+01\n",
      "iter = 064, obj = 1.43823e+00, df_obj = 1.04e-04, reg_obj = 1.44e+00, rnorm = 1.57e+01\n",
      "iter = 065, obj = 1.43611e+00, df_obj = 1.02e-04, reg_obj = 1.44e+00, rnorm = 1.57e+01\n",
      "iter = 066, obj = 1.43408e+00, df_obj = 1.01e-04, reg_obj = 1.43e+00, rnorm = 1.57e+01\n",
      "iter = 067, obj = 1.43201e+00, df_obj = 9.98e-05, reg_obj = 1.43e+00, rnorm = 1.58e+01\n",
      "iter = 068, obj = 1.43005e+00, df_obj = 9.85e-05, reg_obj = 1.43e+00, rnorm = 1.58e+01\n",
      "iter = 069, obj = 1.42804e+00, df_obj = 9.74e-05, reg_obj = 1.43e+00, rnorm = 1.58e+01\n",
      "iter = 070, obj = 1.42612e+00, df_obj = 9.62e-05, reg_obj = 1.43e+00, rnorm = 1.58e+01\n",
      "iter = 071, obj = 1.42417e+00, df_obj = 9.51e-05, reg_obj = 1.42e+00, rnorm = 1.58e+01\n",
      "iter = 072, obj = 1.42232e+00, df_obj = 9.40e-05, reg_obj = 1.42e+00, rnorm = 1.58e+01\n",
      "iter = 073, obj = 1.42044e+00, df_obj = 9.30e-05, reg_obj = 1.42e+00, rnorm = 1.58e+01\n",
      "iter = 074, obj = 1.41864e+00, df_obj = 9.19e-05, reg_obj = 1.42e+00, rnorm = 1.58e+01\n",
      "iter = 075, obj = 1.41682e+00, df_obj = 9.10e-05, reg_obj = 1.42e+00, rnorm = 1.58e+01\n",
      "iter = 076, obj = 1.41508e+00, df_obj = 9.00e-05, reg_obj = 1.41e+00, rnorm = 1.59e+01\n",
      "iter = 077, obj = 1.41331e+00, df_obj = 8.91e-05, reg_obj = 1.41e+00, rnorm = 1.59e+01\n",
      "iter = 078, obj = 1.41162e+00, df_obj = 8.82e-05, reg_obj = 1.41e+00, rnorm = 1.59e+01\n",
      "iter = 079, obj = 1.40991e+00, df_obj = 8.73e-05, reg_obj = 1.41e+00, rnorm = 1.59e+01\n",
      "iter = 080, obj = 1.40828e+00, df_obj = 8.64e-05, reg_obj = 1.41e+00, rnorm = 1.59e+01\n",
      "iter = 081, obj = 1.40662e+00, df_obj = 8.56e-05, reg_obj = 1.41e+00, rnorm = 1.59e+01\n",
      "iter = 082, obj = 1.40504e+00, df_obj = 8.48e-05, reg_obj = 1.40e+00, rnorm = 1.59e+01\n",
      "iter = 083, obj = 1.40344e+00, df_obj = 8.40e-05, reg_obj = 1.40e+00, rnorm = 1.59e+01\n",
      "iter = 084, obj = 1.40191e+00, df_obj = 8.32e-05, reg_obj = 1.40e+00, rnorm = 1.59e+01\n",
      "iter = 085, obj = 1.40036e+00, df_obj = 8.24e-05, reg_obj = 1.40e+00, rnorm = 1.59e+01\n",
      "iter = 086, obj = 1.39887e+00, df_obj = 8.17e-05, reg_obj = 1.40e+00, rnorm = 1.60e+01\n",
      "iter = 087, obj = 1.39737e+00, df_obj = 8.10e-05, reg_obj = 1.40e+00, rnorm = 1.60e+01\n",
      "iter = 088, obj = 1.39591e+00, df_obj = 8.02e-05, reg_obj = 1.40e+00, rnorm = 1.60e+01\n",
      "iter = 089, obj = 1.39444e+00, df_obj = 7.96e-05, reg_obj = 1.39e+00, rnorm = 1.60e+01\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "iter = 090, obj = 1.39301e+00, df_obj = 7.89e-05, reg_obj = 1.39e+00, rnorm = 1.60e+01\n",
      "iter = 091, obj = 1.39157e+00, df_obj = 7.82e-05, reg_obj = 1.39e+00, rnorm = 1.60e+01\n",
      "iter = 092, obj = 1.39018e+00, df_obj = 7.75e-05, reg_obj = 1.39e+00, rnorm = 1.60e+01\n",
      "iter = 093, obj = 1.38877e+00, df_obj = 7.69e-05, reg_obj = 1.39e+00, rnorm = 1.60e+01\n",
      "iter = 094, obj = 1.38740e+00, df_obj = 7.63e-05, reg_obj = 1.39e+00, rnorm = 1.60e+01\n",
      "iter = 095, obj = 1.38603e+00, df_obj = 7.57e-05, reg_obj = 1.39e+00, rnorm = 1.60e+01\n",
      "iter = 096, obj = 1.38469e+00, df_obj = 7.51e-05, reg_obj = 1.38e+00, rnorm = 1.61e+01\n",
      "iter = 097, obj = 1.38336e+00, df_obj = 7.45e-05, reg_obj = 1.38e+00, rnorm = 1.61e+01\n",
      "iter = 098, obj = 1.38207e+00, df_obj = 7.39e-05, reg_obj = 1.38e+00, rnorm = 1.61e+01\n",
      "iter = 099, obj = 1.38077e+00, df_obj = 7.34e-05, reg_obj = 1.38e+00, rnorm = 1.61e+01\n",
      "iter = 100, obj = 1.37952e+00, df_obj = 7.28e-05, reg_obj = 1.38e+00, rnorm = 1.61e+01\n",
      "iter = 101, obj = 1.37826e+00, df_obj = 7.23e-05, reg_obj = 1.38e+00, rnorm = 1.61e+01\n",
      "iter = 102, obj = 1.37704e+00, df_obj = 7.17e-05, reg_obj = 1.38e+00, rnorm = 1.61e+01\n",
      "iter = 103, obj = 1.37580e+00, df_obj = 7.12e-05, reg_obj = 1.38e+00, rnorm = 1.61e+01\n",
      "iter = 104, obj = 1.37461e+00, df_obj = 7.07e-05, reg_obj = 1.37e+00, rnorm = 1.61e+01\n",
      "iter = 105, obj = 1.37339e+00, df_obj = 7.02e-05, reg_obj = 1.37e+00, rnorm = 1.61e+01\n",
      "iter = 106, obj = 1.37223e+00, df_obj = 6.97e-05, reg_obj = 1.37e+00, rnorm = 1.61e+01\n",
      "iter = 107, obj = 1.37104e+00, df_obj = 6.92e-05, reg_obj = 1.37e+00, rnorm = 1.62e+01\n",
      "iter = 108, obj = 1.36992e+00, df_obj = 6.87e-05, reg_obj = 1.37e+00, rnorm = 1.62e+01\n",
      "iter = 109, obj = 1.36875e+00, df_obj = 6.83e-05, reg_obj = 1.37e+00, rnorm = 1.62e+01\n",
      "iter = 110, obj = 1.36765e+00, df_obj = 6.78e-05, reg_obj = 1.37e+00, rnorm = 1.62e+01\n",
      "iter = 111, obj = 1.36650e+00, df_obj = 6.74e-05, reg_obj = 1.37e+00, rnorm = 1.62e+01\n",
      "iter = 112, obj = 1.36542e+00, df_obj = 6.69e-05, reg_obj = 1.37e+00, rnorm = 1.62e+01\n",
      "iter = 113, obj = 1.36430e+00, df_obj = 6.65e-05, reg_obj = 1.36e+00, rnorm = 1.62e+01\n",
      "iter = 114, obj = 1.36325e+00, df_obj = 6.61e-05, reg_obj = 1.36e+00, rnorm = 1.62e+01\n",
      "iter = 115, obj = 1.36214e+00, df_obj = 6.57e-05, reg_obj = 1.36e+00, rnorm = 1.62e+01\n",
      "iter = 116, obj = 1.36111e+00, df_obj = 6.53e-05, reg_obj = 1.36e+00, rnorm = 1.62e+01\n",
      "iter = 117, obj = 1.36004e+00, df_obj = 6.49e-05, reg_obj = 1.36e+00, rnorm = 1.62e+01\n",
      "iter = 118, obj = 1.35904e+00, df_obj = 6.45e-05, reg_obj = 1.36e+00, rnorm = 1.63e+01\n",
      "iter = 119, obj = 1.35798e+00, df_obj = 6.41e-05, reg_obj = 1.36e+00, rnorm = 1.63e+01\n",
      "iter = 120, obj = 1.35699e+00, df_obj = 6.37e-05, reg_obj = 1.36e+00, rnorm = 1.63e+01\n",
      "iter = 121, obj = 1.35597e+00, df_obj = 6.33e-05, reg_obj = 1.36e+00, rnorm = 1.63e+01\n",
      "iter = 122, obj = 1.35503e+00, df_obj = 6.30e-05, reg_obj = 1.35e+00, rnorm = 1.63e+01\n",
      "iter = 123, obj = 1.35404e+00, df_obj = 6.26e-05, reg_obj = 1.35e+00, rnorm = 1.63e+01\n",
      "iter = 124, obj = 1.35311e+00, df_obj = 6.22e-05, reg_obj = 1.35e+00, rnorm = 1.63e+01\n",
      "iter = 125, obj = 1.35213e+00, df_obj = 6.19e-05, reg_obj = 1.35e+00, rnorm = 1.63e+01\n",
      "iter = 126, obj = 1.35122e+00, df_obj = 6.15e-05, reg_obj = 1.35e+00, rnorm = 1.63e+01\n",
      "iter = 127, obj = 1.35027e+00, df_obj = 6.12e-05, reg_obj = 1.35e+00, rnorm = 1.63e+01\n",
      "iter = 128, obj = 1.34938e+00, df_obj = 6.09e-05, reg_obj = 1.35e+00, rnorm = 1.63e+01\n",
      "iter = 129, obj = 1.34844e+00, df_obj = 6.06e-05, reg_obj = 1.35e+00, rnorm = 1.63e+01\n",
      "iter = 130, obj = 1.34758e+00, df_obj = 6.02e-05, reg_obj = 1.35e+00, rnorm = 1.64e+01\n",
      "iter = 131, obj = 1.34666e+00, df_obj = 5.99e-05, reg_obj = 1.35e+00, rnorm = 1.64e+01\n",
      "iter = 132, obj = 1.34583e+00, df_obj = 5.96e-05, reg_obj = 1.35e+00, rnorm = 1.64e+01\n",
      "iter = 133, obj = 1.34492e+00, df_obj = 5.93e-05, reg_obj = 1.34e+00, rnorm = 1.64e+01\n",
      "iter = 134, obj = 1.34410e+00, df_obj = 5.90e-05, reg_obj = 1.34e+00, rnorm = 1.64e+01\n",
      "iter = 135, obj = 1.34321e+00, df_obj = 5.87e-05, reg_obj = 1.34e+00, rnorm = 1.64e+01\n",
      "iter = 136, obj = 1.34242e+00, df_obj = 5.84e-05, reg_obj = 1.34e+00, rnorm = 1.64e+01\n",
      "iter = 137, obj = 1.34155e+00, df_obj = 5.81e-05, reg_obj = 1.34e+00, rnorm = 1.64e+01\n",
      "iter = 138, obj = 1.34077e+00, df_obj = 5.78e-05, reg_obj = 1.34e+00, rnorm = 1.64e+01\n",
      "iter = 139, obj = 1.33991e+00, df_obj = 5.75e-05, reg_obj = 1.34e+00, rnorm = 1.64e+01\n",
      "iter = 140, obj = 1.33916e+00, df_obj = 5.72e-05, reg_obj = 1.34e+00, rnorm = 1.64e+01\n",
      "iter = 141, obj = 1.33834e+00, df_obj = 5.69e-05, reg_obj = 1.34e+00, rnorm = 1.64e+01\n",
      "iter = 142, obj = 1.33762e+00, df_obj = 5.66e-05, reg_obj = 1.34e+00, rnorm = 1.65e+01\n",
      "iter = 143, obj = 1.33681e+00, df_obj = 5.64e-05, reg_obj = 1.34e+00, rnorm = 1.65e+01\n",
      "iter = 144, obj = 1.33609e+00, df_obj = 5.61e-05, reg_obj = 1.34e+00, rnorm = 1.65e+01\n",
      "iter = 145, obj = 1.33530e+00, df_obj = 5.58e-05, reg_obj = 1.34e+00, rnorm = 1.65e+01\n",
      "iter = 146, obj = 1.33460e+00, df_obj = 5.55e-05, reg_obj = 1.33e+00, rnorm = 1.65e+01\n",
      "iter = 147, obj = 1.33383e+00, df_obj = 5.53e-05, reg_obj = 1.33e+00, rnorm = 1.65e+01\n",
      "iter = 148, obj = 1.33314e+00, df_obj = 5.50e-05, reg_obj = 1.33e+00, rnorm = 1.65e+01\n",
      "iter = 149, obj = 1.33238e+00, df_obj = 5.48e-05, reg_obj = 1.33e+00, rnorm = 1.65e+01\n",
      "iter = 150, obj = 1.33172e+00, df_obj = 5.45e-05, reg_obj = 1.33e+00, rnorm = 1.65e+01\n",
      "iter = 151, obj = 1.33098e+00, df_obj = 5.43e-05, reg_obj = 1.33e+00, rnorm = 1.65e+01\n",
      "iter = 152, obj = 1.33033e+00, df_obj = 5.40e-05, reg_obj = 1.33e+00, rnorm = 1.65e+01\n",
      "iter = 153, obj = 1.32962e+00, df_obj = 5.38e-05, reg_obj = 1.33e+00, rnorm = 1.65e+01\n",
      "iter = 154, obj = 1.32899e+00, df_obj = 5.35e-05, reg_obj = 1.33e+00, rnorm = 1.65e+01\n",
      "iter = 155, obj = 1.32829e+00, df_obj = 5.33e-05, reg_obj = 1.33e+00, rnorm = 1.66e+01\n",
      "iter = 156, obj = 1.32768e+00, df_obj = 5.31e-05, reg_obj = 1.33e+00, rnorm = 1.66e+01\n",
      "iter = 157, obj = 1.32699e+00, df_obj = 5.28e-05, reg_obj = 1.33e+00, rnorm = 1.66e+01\n",
      "iter = 158, obj = 1.32639e+00, df_obj = 5.26e-05, reg_obj = 1.33e+00, rnorm = 1.66e+01\n",
      "iter = 159, obj = 1.32572e+00, df_obj = 5.24e-05, reg_obj = 1.33e+00, rnorm = 1.66e+01\n",
      "iter = 160, obj = 1.32514e+00, df_obj = 5.21e-05, reg_obj = 1.33e+00, rnorm = 1.66e+01\n",
      "iter = 161, obj = 1.32449e+00, df_obj = 5.19e-05, reg_obj = 1.32e+00, rnorm = 1.66e+01\n",
      "iter = 162, obj = 1.32392e+00, df_obj = 5.17e-05, reg_obj = 1.32e+00, rnorm = 1.66e+01\n",
      "iter = 163, obj = 1.32328e+00, df_obj = 5.15e-05, reg_obj = 1.32e+00, rnorm = 1.66e+01\n",
      "iter = 164, obj = 1.32272e+00, df_obj = 5.13e-05, reg_obj = 1.32e+00, rnorm = 1.66e+01\n",
      "iter = 165, obj = 1.32208e+00, df_obj = 5.11e-05, reg_obj = 1.32e+00, rnorm = 1.66e+01\n",
      "iter = 166, obj = 1.32153e+00, df_obj = 5.08e-05, reg_obj = 1.32e+00, rnorm = 1.66e+01\n",
      "iter = 167, obj = 1.32090e+00, df_obj = 5.06e-05, reg_obj = 1.32e+00, rnorm = 1.66e+01\n",
      "iter = 168, obj = 1.32035e+00, df_obj = 5.04e-05, reg_obj = 1.32e+00, rnorm = 1.66e+01\n",
      "iter = 169, obj = 1.31974e+00, df_obj = 5.02e-05, reg_obj = 1.32e+00, rnorm = 1.66e+01\n",
      "iter = 170, obj = 1.31921e+00, df_obj = 5.00e-05, reg_obj = 1.32e+00, rnorm = 1.67e+01\n",
      "iter = 171, obj = 1.31860e+00, df_obj = 4.98e-05, reg_obj = 1.32e+00, rnorm = 1.67e+01\n",
      "iter = 172, obj = 1.31808e+00, df_obj = 4.96e-05, reg_obj = 1.32e+00, rnorm = 1.67e+01\n",
      "iter = 173, obj = 1.31749e+00, df_obj = 4.94e-05, reg_obj = 1.32e+00, rnorm = 1.67e+01\n",
      "iter = 174, obj = 1.31697e+00, df_obj = 4.92e-05, reg_obj = 1.32e+00, rnorm = 1.67e+01\n",
      "iter = 175, obj = 1.31638e+00, df_obj = 4.91e-05, reg_obj = 1.32e+00, rnorm = 1.67e+01\n",
      "iter = 176, obj = 1.31587e+00, df_obj = 4.89e-05, reg_obj = 1.32e+00, rnorm = 1.67e+01\n",
      "iter = 177, obj = 1.31528e+00, df_obj = 4.87e-05, reg_obj = 1.32e+00, rnorm = 1.67e+01\n",
      "iter = 178, obj = 1.31478e+00, df_obj = 4.85e-05, reg_obj = 1.31e+00, rnorm = 1.67e+01\n",
      "iter = 179, obj = 1.31421e+00, df_obj = 4.83e-05, reg_obj = 1.31e+00, rnorm = 1.67e+01\n",
      "iter = 180, obj = 1.31373e+00, df_obj = 4.81e-05, reg_obj = 1.31e+00, rnorm = 1.67e+01\n",
      "iter = 181, obj = 1.31318e+00, df_obj = 4.80e-05, reg_obj = 1.31e+00, rnorm = 1.67e+01\n",
      "iter = 182, obj = 1.31271e+00, df_obj = 4.78e-05, reg_obj = 1.31e+00, rnorm = 1.67e+01\n",
      "iter = 183, obj = 1.31216e+00, df_obj = 4.77e-05, reg_obj = 1.31e+00, rnorm = 1.67e+01\n",
      "iter = 184, obj = 1.31169e+00, df_obj = 4.75e-05, reg_obj = 1.31e+00, rnorm = 1.67e+01\n",
      "iter = 185, obj = 1.31114e+00, df_obj = 4.73e-05, reg_obj = 1.31e+00, rnorm = 1.68e+01\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "iter = 186, obj = 1.31068e+00, df_obj = 4.71e-05, reg_obj = 1.31e+00, rnorm = 1.68e+01\n",
      "iter = 187, obj = 1.31014e+00, df_obj = 4.70e-05, reg_obj = 1.31e+00, rnorm = 1.68e+01\n",
      "iter = 188, obj = 1.30968e+00, df_obj = 4.68e-05, reg_obj = 1.31e+00, rnorm = 1.68e+01\n",
      "iter = 189, obj = 1.30916e+00, df_obj = 4.67e-05, reg_obj = 1.31e+00, rnorm = 1.68e+01\n",
      "iter = 190, obj = 1.30870e+00, df_obj = 4.65e-05, reg_obj = 1.31e+00, rnorm = 1.68e+01\n",
      "iter = 191, obj = 1.30818e+00, df_obj = 4.64e-05, reg_obj = 1.31e+00, rnorm = 1.68e+01\n",
      "iter = 192, obj = 1.30773e+00, df_obj = 4.62e-05, reg_obj = 1.31e+00, rnorm = 1.68e+01\n",
      "iter = 193, obj = 1.30722e+00, df_obj = 4.60e-05, reg_obj = 1.31e+00, rnorm = 1.68e+01\n",
      "iter = 194, obj = 1.30677e+00, df_obj = 4.59e-05, reg_obj = 1.31e+00, rnorm = 1.68e+01\n",
      "iter = 195, obj = 1.30627e+00, df_obj = 4.57e-05, reg_obj = 1.31e+00, rnorm = 1.68e+01\n",
      "iter = 196, obj = 1.30583e+00, df_obj = 4.56e-05, reg_obj = 1.31e+00, rnorm = 1.68e+01\n",
      "iter = 197, obj = 1.30533e+00, df_obj = 4.54e-05, reg_obj = 1.31e+00, rnorm = 1.68e+01\n",
      "iter = 198, obj = 1.30490e+00, df_obj = 4.53e-05, reg_obj = 1.30e+00, rnorm = 1.68e+01\n",
      "iter = 199, obj = 1.30440e+00, df_obj = 4.51e-05, reg_obj = 1.30e+00, rnorm = 1.68e+01\n",
      "iter = 200, obj = 1.30397e+00, df_obj = 4.50e-05, reg_obj = 1.30e+00, rnorm = 1.69e+01\n",
      "Terminate: maximum number of iterations reached\n",
      "\n",
      "##########################################################################################\n",
      "            SPLIT-BREGMAN Solver end\n",
      "##########################################################################################\n"
     ]
    }
   ],
   "source": [
    "problemSB = o.GeneralizedLasso(x.clone().zero(), d, G, reg=D, eps=1e-3)\n",
    "SB = o.SplitBregman(o.BasicStopper(niter=200), niter_inner=3, niter_solver=10,\n",
    "                    linear_solver='LSQR', breg_weight=1., warm_start=True)\n",
    "SB.setDefaults(save_obj=True)\n",
    "SB.run(problemSB, verbose=True, inner_verbose=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "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": [
    "its = \"%d-%d-%d\" % (len(SB.obj), SB.niter_inner, SB.niter_solver)\n",
    "\n",
    "fig, ax = plt.subplots(1, 2, figsize=(10,6))\n",
    "ax[0].imshow(x.plot(), cmap='bone', clim=(0,1))\n",
    "ax[0].grid(False)\n",
    "ax[0].set_title('Model')\n",
    "ax[1].imshow(problemSB.model.plot(), cmap='bone', clim=(0,1))\n",
    "ax[1].grid(False)\n",
    "ax[1].set_title(r'SB %s, $\\varepsilon=%.e$' % (its, problemSB.eps))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(10,6))\n",
    "ax.semilogy(SB.obj / SB.obj[0], 'k', lw=2, label='Obj')\n",
    "ax.semilogy(SB.obj_terms[:, 0] / SB.obj[0], 'b--', label=r\"$0.5 \\Vert \\mathbf{Gx-d} \\Vert_2^2$\")\n",
    "ax.semilogy(SB.obj_terms[:, 1] / SB.obj[0], 'r--', label=r\"$\\varepsilon \\Vert \\mathbf{Dx}\\Vert_1$\")\n",
    "ax.set_ylim(1e-8, 1), ax.set_xlim(0, SB.stopper.niter)\n",
    "ax.set_xlabel(\"Iterations\")\n",
    "ax.legend(loc=\"right\")\n",
    "ax.grid(True)\n",
    "plt.suptitle(r'Split-Bregman %s, $\\varepsilon=%.e$' % (its, problemSB.eps))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.10.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}