cocoercivity.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([100, 200])\n"
     ]
    }
   ],
   "source": [
    "import torch\n",
    "\n",
    "array = torch.load('vanila_inf_20240504-140634.pt')\n",
    "print(array.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0 tensor(0.0337)\n",
      "1 tensor(0.0151)\n",
      "2 tensor(0.0028)\n",
      "3 tensor(0.0323)\n",
      "4 tensor(-0.0588)\n",
      "5 tensor(0.0032)\n",
      "6 tensor(0.0223)\n",
      "7 tensor(-0.0221)\n",
      "8 tensor(0.0032)\n",
      "9 tensor(0.0066)\n",
      "10 tensor(0.0050)\n",
      "11 tensor(0.0375)\n",
      "12 tensor(0.0193)\n",
      "13 tensor(-0.6570)\n",
      "14 tensor(-0.0467)\n",
      "15 tensor(0.0057)\n",
      "16 tensor(0.0102)\n",
      "17 tensor(0.0091)\n",
      "18 tensor(0.0113)\n",
      "19 tensor(-0.0150)\n",
      "20 tensor(-0.0158)\n",
      "21 tensor(0.0075)\n",
      "22 tensor(0.0032)\n",
      "23 tensor(0.0263)\n",
      "24 tensor(0.0003)\n",
      "25 tensor(-0.0596)\n",
      "26 tensor(-0.0889)\n",
      "27 tensor(-0.1599)\n",
      "28 tensor(0.0061)\n",
      "29 tensor(0.0329)\n",
      "30 tensor(0.0427)\n",
      "31 tensor(-0.0096)\n",
      "32 tensor(0.0029)\n",
      "33 tensor(0.0017)\n",
      "34 tensor(-0.1011)\n",
      "35 tensor(0.0647)\n",
      "36 tensor(0.0100)\n",
      "37 tensor(-0.2765)\n",
      "38 tensor(0.0339)\n",
      "39 tensor(-0.0398)\n",
      "40 tensor(-0.4150)\n",
      "41 tensor(-0.0717)\n",
      "42 tensor(0.0360)\n",
      "43 tensor(-0.0173)\n",
      "44 tensor(-0.2108)\n",
      "45 tensor(0.0325)\n",
      "46 tensor(-0.2570)\n",
      "47 tensor(-0.2246)\n",
      "48 tensor(-0.4357)\n",
      "49 tensor(0.0177)\n",
      "50 tensor(-0.0763)\n",
      "51 tensor(-0.0117)\n",
      "52 tensor(-0.8868)\n",
      "53 tensor(-0.0305)\n",
      "54 tensor(0.0415)\n",
      "55 tensor(0.0042)\n",
      "56 tensor(-0.2226)\n",
      "57 tensor(0.0746)\n",
      "58 tensor(-0.0002)\n",
      "59 tensor(0.0430)\n",
      "60 tensor(-0.0709)\n",
      "61 tensor(0.0198)\n",
      "62 tensor(0.0130)\n",
      "63 tensor(-0.0003)\n",
      "64 tensor(0.0109)\n",
      "65 tensor(0.0005)\n",
      "66 tensor(-0.1993)\n",
      "67 tensor(0.0203)\n",
      "68 tensor(0.0478)\n",
      "69 tensor(0.0228)\n",
      "70 tensor(0.0023)\n",
      "71 tensor(-0.0293)\n",
      "72 tensor(0.0005)\n",
      "73 tensor(-0.0485)\n",
      "74 tensor(-0.1078)\n",
      "75 tensor(-0.1263)\n",
      "76 tensor(-0.1298)\n",
      "77 tensor(0.0764)\n",
      "78 tensor(0.0033)\n",
      "79 tensor(-0.2162)\n",
      "80 tensor(-0.0605)\n",
      "81 tensor(0.0653)\n",
      "82 tensor(0.0055)\n",
      "83 tensor(0.0042)\n",
      "84 tensor(-0.0627)\n",
      "85 tensor(0.0318)\n",
      "86 tensor(-0.2348)\n",
      "87 tensor(0.0004)\n",
      "88 tensor(-0.4075)\n",
      "89 tensor(0.0153)\n",
      "90 tensor(0.0041)\n",
      "91 tensor(0.0064)\n",
      "92 tensor(0.0245)\n",
      "93 tensor(0.0435)\n",
      "94 tensor(-0.5617)\n",
      "95 tensor(0.0086)\n",
      "96 tensor(0.0568)\n",
      "97 tensor(-0.0048)\n",
      "98 tensor(0.0290)\n",
      "99 tensor(-0.1785)\n"
     ]
    }
   ],
   "source": [
    "for i in range(100):\n",
    "    print(i, array[i].min())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_1550752/1136347692.py:26: UserWarning: Data has no positive values, and therefore cannot be log-scaled.\n",
      "  plt.axvline(x=percentile, color='blue', linestyle='--')\n",
      "/tmp/ipykernel_1550752/1136347692.py:32: UserWarning: Data has no positive values, and therefore cannot be log-scaled.\n",
      "  plt.tight_layout()\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 400x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "from matplotlib.ticker import FuncFormatter\n",
    "\n",
    "array_for_plot = array[:,:80].flatten()\n",
    "min_value = torch.min(array_for_plot)\n",
    "percentile = np.percentile(array_for_plot, 0.1)\n",
    "logbins = np.logspace(torch.log10(min_value),0,10)\n",
    "# 새로운 Figure 객체 생성하여 크기 지정\n",
    "plt.rcParams.update({'font.size': 14})\n",
    "plt.figure(figsize=(4, 3))\n",
    "plt.hist(array_for_plot, bins=logbins, color='gray', edgecolor='black')\n",
    "#plt.xlabel(r'$\\langle \\mathcal{E} \\mathcal{D} z - \\mathcal{E} \\mathcal{D} z, z - z \\rangle$')\n",
    "#plt.ylabel('Frequency')\n",
    "#plt.title('Histogram')\n",
    "plt.xscale('log')\n",
    "\n",
    "# yticks의 레이블을 지정하는 함수 정의\n",
    "def format_yticks(value, pos):\n",
    "    return '{:.0f}k'.format(value / 1000)\n",
    "\n",
    "# yticks에 함수 적용\n",
    "plt.gca().yaxis.set_major_formatter(FuncFormatter(format_yticks))\n",
    "\n",
    "plt.axvline(x=min_value, color='red', linestyle='-')\n",
    "plt.axvline(x=percentile, color='blue', linestyle='--')\n",
    "plt.text(0.18, 0.80, r'$\\min=$', fontsize=14, transform=plt.gcf().transFigure)\n",
    "plt.text(0.18, 0.73, r'${:.1f}\\times10^{{-5}}$'.format(min_value* 1e5), fontsize=14, transform=plt.gcf().transFigure)\n",
    "plt.text(0.5, 0.80, 'percentile(0.1)=', fontsize=11, transform=plt.gcf().transFigure)\n",
    "plt.text(0.5, 0.73, r'${:.1f}\\times10^{{-3}}$'.format(percentile*1e3), fontsize=13, transform=plt.gcf().transFigure)\n",
    "\n",
    "plt.tight_layout()\n",
    "# plt.savefig('/home/icl2/khlee/InstaFlow/vanila_inf.pdf')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "\n",
    "array = torch.load('beta_adam.pt')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 400x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "\n",
    "array_for_plot = array[:,80:].flatten()\n",
    "min_value = torch.min(array_for_plot)\n",
    "#logbins = np.logspace(torch.log10(min_value),0,10)\n",
    "# 새로운 Figure 객체 생성하여 크기 지정\n",
    "plt.figure(figsize=(4, 3))\n",
    "plt.hist(array_for_plot, bins=100, color='blue', edgecolor='black')\n",
    "#plt.xlabel(r'$\\langle \\mathcal{E} \\mathcal{D} z - \\mathcal{E} \\mathcal{D} z, z - z \\rangle$')\n",
    "#plt.ylabel('Frequency')\n",
    "#plt.title('Histogram')\n",
    "#plt.xscale('log')\n",
    "\n",
    "plt.axvline(x=min_value, color='red', linestyle='-')\n",
    "plt.text(0.2, 0.7, r'$\\beta=\\mathrm{{{:.2e}}}$'.format(min_value), fontsize=12, transform=plt.gcf().transFigure)\n",
    "\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([100, 100])\n"
     ]
    }
   ],
   "source": [
    "# vanila\n",
    "import torch\n",
    "\n",
    "array = torch.load('beta_vanila.pt')\n",
    "print(array.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 400x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "from matplotlib.ticker import FuncFormatter\n",
    "\n",
    "array_for_plot = array[:,:80].flatten()\n",
    "min_value = torch.min(array_for_plot)\n",
    "percentile = np.percentile(array_for_plot, 0.1)\n",
    "logbins = np.logspace(torch.log10(min_value),0,10)\n",
    "# 새로운 Figure 객체 생성하여 크기 지정\n",
    "plt.rcParams.update({'font.size': 14})\n",
    "plt.figure(figsize=(4, 3))\n",
    "plt.hist(array_for_plot, bins=logbins, color='gray', edgecolor='black')\n",
    "#plt.xlabel(r'$\\langle \\mathcal{E} \\mathcal{D} z - \\mathcal{E} \\mathcal{D} z, z - z \\rangle$')\n",
    "#plt.ylabel('Frequency')\n",
    "#plt.title('Histogram')\n",
    "plt.xscale('log')\n",
    "\n",
    "# yticks의 레이블을 지정하는 함수 정의\n",
    "def format_yticks(value, pos):\n",
    "    return '{:.0f}k'.format(value / 1000)\n",
    "\n",
    "# yticks에 함수 적용\n",
    "plt.gca().yaxis.set_major_formatter(FuncFormatter(format_yticks))\n",
    "\n",
    "plt.axvline(x=min_value, color='red', linestyle='-')\n",
    "plt.axvline(x=percentile, color='blue', linestyle='--')\n",
    "plt.text(0.18, 0.80, r'$\\min=$', fontsize=14, transform=plt.gcf().transFigure)\n",
    "plt.text(0.18, 0.73, r'${:.1f}\\times10^{{-5}}$'.format(min_value* 1e5), fontsize=14, transform=plt.gcf().transFigure)\n",
    "plt.text(0.5, 0.80, 'percentile(0.1)=', fontsize=11, transform=plt.gcf().transFigure)\n",
    "plt.text(0.5, 0.73, r'${:.1f}\\times10^{{-3}}$'.format(percentile*1e3), fontsize=13, transform=plt.gcf().transFigure)\n",
    "\n",
    "plt.tight_layout()\n",
    "plt.savefig('/home/icl2/khlee/InstaFlow/vanila.pdf')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "\n",
    "array = torch.load('beta_vanila.pt')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 400x300 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "\n",
    "array_for_plot = array[:,80:].flatten()\n",
    "min_value = torch.min(array_for_plot)\n",
    "#logbins = np.logspace(torch.log10(min_value),0,10)\n",
    "# 새로운 Figure 객체 생성하여 크기 지정\n",
    "plt.figure(figsize=(4, 3))\n",
    "plt.hist(array_for_plot, bins=100, color='blue', edgecolor='black')\n",
    "#plt.xlabel(r'$\\langle \\mathcal{E} \\mathcal{D} z - \\mathcal{E} \\mathcal{D} z, z - z \\rangle$')\n",
    "#plt.ylabel('Frequency')\n",
    "#plt.title('Histogram')\n",
    "#plt.xscale('log')\n",
    "\n",
    "plt.axvline(x=min_value, color='red', linestyle='-')\n",
    "plt.text(0.2, 0.7, r'$\\beta=\\mathrm{{{:.2e}}}$'.format(min_value), fontsize=12, transform=plt.gcf().transFigure)\n",
    "\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "exact_inversion2",
   "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": "3.10.11"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}