visual.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "from torch.utils.data import Dataset\n",
    "from torch.utils.data.dataloader import DataLoader\n",
    "import torch.optim as optim\n",
    "from utils.utils import set_seed, load_config\n",
    "from utils.saver import Saver\n",
    "from bunny_GPT2 import GPT2\n",
    "from demo import SortDataset\n",
    "set_seed(42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_dataset = SortDataset(n_samples=1000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "args = load_config('config.yaml')\n",
    "\n",
    "vocab_size = test_dataset.vocab_size\n",
    "block_size = test_dataset.block_size\n",
    "embed_dim = args.embed_dim\n",
    "num_heads = args.num_heads\n",
    "num_blocks = args.num_blocks\n",
    "batch_size = args.batch_size"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<All keys matched successfully>"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sort_model = GPT2(vocab_size=vocab_size, block_size=block_size, embed_dim=embed_dim, num_heads=num_heads, num_blocks=num_blocks)\n",
    "ckpt = torch.load(r'out_logs\\model_789.pt', weights_only=True)\n",
    "sort_model.load_state_dict(ckpt['model'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "input: [0 2 1 2 2]\n",
      "output: [0 1 2 2 2] \n",
      "gt: [0 1 2 2 2]\n"
     ]
    }
   ],
   "source": [
    "x, gt = test_dataset[0]\n",
    "x_ = x.clone()\n",
    "x = x[: test_dataset.n_seq]\n",
    "y = sort_model.generate(x, test_dataset.n_seq)\n",
    "x = x.detach().numpy()\n",
    "gt = gt[test_dataset.n_seq-1:].detach().numpy()\n",
    "print(f\"input: {x}\\noutput: {y['input_ids'][0]} \\ngt: {gt}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 300x1500 with 5 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# draw probability distribution\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "figure, axis = plt.subplots(5, 1, figsize=(3, 15))\n",
    "prev_token = []\n",
    "\n",
    "# Iterate and plot probabilities\n",
    "for i, prob in enumerate(y['probs']):\n",
    "    labels = [0, 1, 2]\n",
    "    prob_i = prob[0]\n",
    "    \n",
    "    # Create a bar plot for each probability distribution\n",
    "    axis[i].bar(labels, prob_i, color=['#ff9999','#66b3ff','#99ff99'], edgecolor='black')\n",
    "    \n",
    "    # Adding title and formatting it\n",
    "    cur_token = y['input_ids'][0][i]\n",
    "    axis[i].set_title(f\"{x} + {prev_token} -> {cur_token}\", fontsize=10, color='blue', fontweight='bold')\n",
    "    \n",
    "    # Adding y-axis label\n",
    "    axis[i].set_ylabel('Probability', fontsize=10)\n",
    "    \n",
    "    # Adding some more styling\n",
    "    axis[i].set_xticks(labels)\n",
    "    axis[i].set_ylim(0, 1)\n",
    "    \n",
    "    prev_token.append(cur_token)\n",
    "\n",
    "# Overall plot title and layout adjustments\n",
    "figure.suptitle('Token Prediction Probabilities', fontsize=16, fontweight='bold')\n",
    "plt.tight_layout(rect=[0, 0, 1, 0.98])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([1, 3, 9, 9])\n",
      "torch.Size([1, 3, 9, 9])\n",
      "torch.Size([1, 3, 9, 9])\n"
     ]
    }
   ],
   "source": [
    "_, attns = sort_model.forward_w_att(x_.unsqueeze(0))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "tokens = x_.detach().numpy().tolist()\n",
    "tokens = [str(t) for t in tokens]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<script src=\"https://cdnjs.cloudflare.com/ajax/libs/require.js/2.3.6/require.min.js\"></script>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "      \n",
       "        <div id=\"bertviz-7e6577d3f3db45f0af7076a248c133ab\" style=\"font-family:'Helvetica Neue', Helvetica, Arial, sans-serif;\">\n",
       "            <span style=\"user-select:none\">\n",
       "                Layer: <select id=\"layer\"></select>\n",
       "                \n",
       "            </span>\n",
       "            <div id='vis'></div>\n",
       "        </div>\n",
       "    "
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/javascript": "/**\n * @fileoverview Transformer Visualization D3 javascript code.\n *\n *\n *  Based on: https://github.com/tensorflow/tensor2tensor/blob/master/tensor2tensor/visualization/attention.js\n *\n * Change log:\n *\n * 12/19/18  Jesse Vig   Assorted cleanup. Changed orientation of attention matrices.\n * 12/29/20  Jesse Vig   Significant refactor.\n * 12/31/20  Jesse Vig   Support multiple visualizations in single notebook.\n * 02/06/21  Jesse Vig   Move require config from separate jupyter notebook step\n * 05/03/21  Jesse Vig   Adjust height of visualization dynamically\n * 07/25/21  Jesse Vig   Support layer filtering\n * 03/23/22  Daniel SC   Update requirement URLs for d3 and jQuery (source of bug not allowing end result to be displayed on browsers)\n **/\n\nrequire.config({\n  paths: {\n      d3: 'https://cdnjs.cloudflare.com/ajax/libs/d3/5.7.0/d3.min',\n    jquery: 'https://cdnjs.cloudflare.com/ajax/libs/jquery/2.0.0/jquery.min',\n  }\n});\n\nrequirejs(['jquery', 'd3'], function ($, d3) {\n\n    const params = {\"attention\": [{\"name\": null, \"attn\": [[[[1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.11583246290683746, 0.884167492389679, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.20939210057258606, 0.2683132588863373, 0.5222946405410767, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.10234597325325012, 0.3366968035697937, 0.3305542767047882, 0.23040297627449036, 0.0, 0.0, 0.0, 0.0, 0.0], [0.05344945564866066, 0.2648983895778656, 0.29641348123550415, 0.2662043869495392, 0.11903432011604309, 0.0, 0.0, 0.0, 0.0], [0.12044446915388107, 0.13451862335205078, 0.20863918960094452, 0.208304300904274, 0.2544945478439331, 0.07359886914491653, 0.0, 0.0, 0.0], [0.06161512807011604, 0.13893286883831024, 0.29715830087661743, 0.19295872747898102, 0.19412259757518768, 0.043250877410173416, 0.07196151465177536, 0.0, 0.0], [0.033144399523735046, 0.20205403864383698, 0.14036297798156738, 0.18987803161144257, 0.13004235923290253, 0.037369295954704285, 0.027896219864487648, 0.23925261199474335, 0.0], [0.017139436677098274, 0.1695355772972107, 0.10050744563341141, 0.20553725957870483, 0.23546375334262848, 0.016382450237870216, 0.018699975684285164, 0.17180640995502472, 0.06492765247821808]], [[1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.5842100381851196, 0.415789932012558, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.17077818512916565, 0.46969273686408997, 0.3595291078090668, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.3512921631336212, 0.1725083291530609, 0.30179762840270996, 0.17440193891525269, 0.0, 0.0, 0.0, 0.0, 0.0], [0.19855765998363495, 0.02586420811712742, 0.08008670061826706, 0.03495143726468086, 0.6605400443077087, 0.0, 0.0, 0.0, 0.0], [0.1573960930109024, 0.26111674308776855, 0.12008215487003326, 0.19006095826625824, 0.2085576355457306, 0.06278634071350098, 0.0, 0.0, 0.0], [0.11011838912963867, 0.19789858162403107, 0.1670699417591095, 0.17725665867328644, 0.10202071070671082, 0.12795332074165344, 0.11768238991498947, 0.0, 0.0], [0.09984728693962097, 0.1015898808836937, 0.09873709082603455, 0.10283883661031723, 0.11371424049139023, 0.19432424008846283, 0.19374461472034454, 0.09520384669303894, 0.0], [0.04460426792502403, 0.005554715637117624, 0.021227383986115456, 0.006664355751127005, 0.027150459587574005, 0.7710481286048889, 0.07064949721097946, 0.02276136539876461, 0.03033989667892456]], [[1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.20813648402690887, 0.7918635606765747, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.22725418210029602, 0.29173892736434937, 0.4810069501399994, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.1585603952407837, 0.29885709285736084, 0.22812871634960175, 0.31445378065109253, 0.0, 0.0, 0.0, 0.0, 0.0], [0.2788330018520355, 0.12160271406173706, 0.29260411858558655, 0.1389254480600357, 0.16803473234176636, 0.0, 0.0, 0.0, 0.0], [0.1867043524980545, 0.09899254888296127, 0.2651490569114685, 0.10916479676961899, 0.28130125999450684, 0.058687951415777206, 0.0, 0.0, 0.0], [0.10783901810646057, 0.11740631610155106, 0.19695693254470825, 0.15104907751083374, 0.3262477219104767, 0.03380884975194931, 0.06669214367866516, 0.0, 0.0], [0.03384784981608391, 0.23243722319602966, 0.13953465223312378, 0.19310569763183594, 0.20496727526187897, 0.016339682042598724, 0.07289475202560425, 0.10687284916639328, 0.0], [0.23444011807441711, 0.07891810685396194, 0.1519857794046402, 0.06546825915575027, 0.0652051568031311, 0.10796913504600525, 0.17505589127540588, 0.0456203818321228, 0.07533720135688782]]], [[[1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.8090201616287231, 0.19097986817359924, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.1832382082939148, 0.3397786617279053, 0.47698307037353516, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.3805065453052521, 0.13347622752189636, 0.30092647671699524, 0.1850907951593399, 0.0, 0.0, 0.0, 0.0, 0.0], [0.5122894644737244, 0.07694385200738907, 0.256824791431427, 0.08846534043550491, 0.06547650694847107, 0.0, 0.0, 0.0, 0.0], [0.22970852255821228, 0.13582248985767365, 0.20937861502170563, 0.17073699831962585, 0.11098608374595642, 0.14336727559566498, 0.0, 0.0, 0.0], [0.09341931343078613, 0.14378610253334045, 0.1972556859254837, 0.17643100023269653, 0.2930387854576111, 0.0408804751932621, 0.055188633501529694, 0.0, 0.0], [0.11508242040872574, 0.12171899527311325, 0.12208487093448639, 0.1762368232011795, 0.2284460961818695, 0.08813095092773438, 0.057572413235902786, 0.09072752296924591, 0.0], [0.12489533424377441, 0.12223611772060394, 0.07339038699865341, 0.163949117064476, 0.21940051019191742, 0.10057976841926575, 0.07395884394645691, 0.07117773592472076, 0.05041223391890526]], [[1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.630396842956543, 0.3696031868457794, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.21211498975753784, 0.22383400797843933, 0.5640510320663452, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.16418185830116272, 0.22461986541748047, 0.4036988317966461, 0.2074994444847107, 0.0, 0.0, 0.0, 0.0, 0.0], [0.43109169602394104, 0.054113633930683136, 0.39762797951698303, 0.053372230380773544, 0.06379447877407074, 0.0, 0.0, 0.0, 0.0], [0.19197046756744385, 0.15701176226139069, 0.15235097706317902, 0.15820172429084778, 0.21662917733192444, 0.12383583188056946, 0.0, 0.0, 0.0], [0.08732523769140244, 0.19746766984462738, 0.27495747804641724, 0.16601130366325378, 0.16353462636470795, 0.027989856898784637, 0.08271379768848419, 0.0, 0.0], [0.030109217390418053, 0.262539267539978, 0.19642876088619232, 0.23985430598258972, 0.12096359580755234, 0.007939137518405914, 0.0693647712469101, 0.07280096411705017, 0.0], [0.01843511126935482, 0.2516745626926422, 0.09083835780620575, 0.2360735982656479, 0.11003099381923676, 0.010639237239956856, 0.06030162423849106, 0.08971142768859863, 0.13229505717754364]], [[1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.5870454907417297, 0.41295450925827026, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.4823109805583954, 0.3582899272441864, 0.1593991369009018, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.43300753831863403, 0.1984279900789261, 0.21754881739616394, 0.15101568400859833, 0.0, 0.0, 0.0, 0.0, 0.0], [0.10235285758972168, 0.280404269695282, 0.18251553177833557, 0.2058979868888855, 0.22882941365242004, 0.0, 0.0, 0.0, 0.0], [0.20861537754535675, 0.19118784368038177, 0.10737765580415726, 0.18525347113609314, 0.14353443682193756, 0.16403119266033173, 0.0, 0.0, 0.0], [0.16503557562828064, 0.09813898801803589, 0.031133856624364853, 0.0675446018576622, 0.07803245633840561, 0.5139825940132141, 0.04613194614648819, 0.0, 0.0], [0.1584610491991043, 0.09127534180879593, 0.06304670125246048, 0.06900466233491898, 0.06389173865318298, 0.26928526163101196, 0.14328324794769287, 0.1417519897222519, 0.0], [0.25161445140838623, 0.029239315539598465, 0.0714222639799118, 0.01879066228866577, 0.024595780298113823, 0.3585717976093292, 0.10119058936834335, 0.07269758731126785, 0.07187756896018982]]], [[[1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.9372128248214722, 0.06278713047504425, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.34250253438949585, 0.5668380856513977, 0.09065934270620346, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.6290720701217651, 0.10822197794914246, 0.15049612522125244, 0.11220986396074295, 0.0, 0.0, 0.0, 0.0, 0.0], [0.970376193523407, 0.0021214366424828768, 0.015626508742570877, 0.0013014038559049368, 0.010574529878795147, 0.0, 0.0, 0.0, 0.0], [0.09130971878767014, 0.06638772040605545, 0.5523999929428101, 0.07101372629404068, 0.09935696423053741, 0.11953180283308029, 0.0, 0.0, 0.0], [0.021298406645655632, 0.26837092638015747, 0.14979466795921326, 0.3500915467739105, 0.11609084904193878, 0.020701348781585693, 0.07365228235721588, 0.0, 0.0], [0.008047293871641159, 0.20200438797473907, 0.10277985781431198, 0.36328285932540894, 0.06489216536283493, 0.008051928132772446, 0.061552878469228745, 0.1893887221813202, 0.0], [0.0011043534614145756, 0.1891116201877594, 0.03197162225842476, 0.3477092683315277, 0.04089963808655739, 0.001434291829355061, 0.021473770961165428, 0.1605575680732727, 0.20573781430721283]], [[1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.884773850440979, 0.1152261346578598, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.564574122428894, 0.2776378393173218, 0.15778802335262299, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.5638299584388733, 0.15842780470848083, 0.17728263139724731, 0.10045959800481796, 0.0, 0.0, 0.0, 0.0, 0.0], [0.8080127239227295, 0.03030439093708992, 0.1265506148338318, 0.014734375290572643, 0.020397968590259552, 0.0, 0.0, 0.0, 0.0], [0.20254802703857422, 0.07231037318706512, 0.536925733089447, 0.07049007713794708, 0.0683068335056305, 0.049418967217206955, 0.0, 0.0, 0.0], [0.11969973146915436, 0.1455283761024475, 0.13576865196228027, 0.1641465425491333, 0.3096204400062561, 0.08562075346708298, 0.03961546719074249, 0.0, 0.0], [0.09068313241004944, 0.19025327265262604, 0.12843848764896393, 0.19580651819705963, 0.18051451444625854, 0.08536171913146973, 0.05424069985747337, 0.07470159977674484, 0.0], [0.021546080708503723, 0.2296065092086792, 0.047056280076503754, 0.22095566987991333, 0.2430584728717804, 0.0647224709391594, 0.030332297086715698, 0.07678408920764923, 0.06593817472457886]], [[1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.9199612736701965, 0.08003874123096466, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.5722076892852783, 0.24290668964385986, 0.1848856806755066, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.5284797549247742, 0.0987781509757042, 0.26617011427879333, 0.10657191276550293, 0.0, 0.0, 0.0, 0.0, 0.0], [0.8895872831344604, 0.01390649750828743, 0.04050413519144058, 0.012708736583590508, 0.04329325258731842, 0.0, 0.0, 0.0, 0.0], [0.16464175283908844, 0.13341635465621948, 0.26562222838401794, 0.1666385382413864, 0.1700018048286438, 0.09967926144599915, 0.0, 0.0, 0.0], [0.029521524906158447, 0.20171837508678436, 0.22596462070941925, 0.3160381615161896, 0.06304743885993958, 0.04469815641641617, 0.11901167035102844, 0.0, 0.0], [0.018640773370862007, 0.14681920409202576, 0.1846427470445633, 0.16796568036079407, 0.03296377509832382, 0.019411159679293633, 0.1117875948548317, 0.3177691102027893, 0.0], [0.00827923882752657, 0.1063661128282547, 0.11433600634336472, 0.11846795678138733, 0.026303714141249657, 0.009230039082467556, 0.05439882352948189, 0.21334347128868103, 0.349274605512619]]]], \"left_text\": [\"0\", \"2\", \"1\", \"2\", \"2\", \"0\", \"1\", \"2\", \"2\"], \"right_text\": [\"0\", \"2\", \"1\", \"2\", \"2\", \"0\", \"1\", \"2\", \"2\"]}], \"default_filter\": \"0\", \"root_div_id\": \"bertviz-7e6577d3f3db45f0af7076a248c133ab\", \"layer\": null, \"heads\": null, \"include_layers\": [0, 1, 2]}; // HACK: {\"attention\": [{\"name\": null, \"attn\": [[[[1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.11583246290683746, 0.884167492389679, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.20939210057258606, 0.2683132588863373, 0.5222946405410767, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.10234597325325012, 0.3366968035697937, 0.3305542767047882, 0.23040297627449036, 0.0, 0.0, 0.0, 0.0, 0.0], [0.05344945564866066, 0.2648983895778656, 0.29641348123550415, 0.2662043869495392, 0.11903432011604309, 0.0, 0.0, 0.0, 0.0], [0.12044446915388107, 0.13451862335205078, 0.20863918960094452, 0.208304300904274, 0.2544945478439331, 0.07359886914491653, 0.0, 0.0, 0.0], [0.06161512807011604, 0.13893286883831024, 0.29715830087661743, 0.19295872747898102, 0.19412259757518768, 0.043250877410173416, 0.07196151465177536, 0.0, 0.0], [0.033144399523735046, 0.20205403864383698, 0.14036297798156738, 0.18987803161144257, 0.13004235923290253, 0.037369295954704285, 0.027896219864487648, 0.23925261199474335, 0.0], [0.017139436677098274, 0.1695355772972107, 0.10050744563341141, 0.20553725957870483, 0.23546375334262848, 0.016382450237870216, 0.018699975684285164, 0.17180640995502472, 0.06492765247821808]], [[1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.5842100381851196, 0.415789932012558, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.17077818512916565, 0.46969273686408997, 0.3595291078090668, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.3512921631336212, 0.1725083291530609, 0.30179762840270996, 0.17440193891525269, 0.0, 0.0, 0.0, 0.0, 0.0], [0.19855765998363495, 0.02586420811712742, 0.08008670061826706, 0.03495143726468086, 0.6605400443077087, 0.0, 0.0, 0.0, 0.0], [0.1573960930109024, 0.26111674308776855, 0.12008215487003326, 0.19006095826625824, 0.2085576355457306, 0.06278634071350098, 0.0, 0.0, 0.0], [0.11011838912963867, 0.19789858162403107, 0.1670699417591095, 0.17725665867328644, 0.10202071070671082, 0.12795332074165344, 0.11768238991498947, 0.0, 0.0], [0.09984728693962097, 0.1015898808836937, 0.09873709082603455, 0.10283883661031723, 0.11371424049139023, 0.19432424008846283, 0.19374461472034454, 0.09520384669303894, 0.0], [0.04460426792502403, 0.005554715637117624, 0.021227383986115456, 0.006664355751127005, 0.027150459587574005, 0.7710481286048889, 0.07064949721097946, 0.02276136539876461, 0.03033989667892456]], [[1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.20813648402690887, 0.7918635606765747, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.22725418210029602, 0.29173892736434937, 0.4810069501399994, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.1585603952407837, 0.29885709285736084, 0.22812871634960175, 0.31445378065109253, 0.0, 0.0, 0.0, 0.0, 0.0], [0.2788330018520355, 0.12160271406173706, 0.29260411858558655, 0.1389254480600357, 0.16803473234176636, 0.0, 0.0, 0.0, 0.0], [0.1867043524980545, 0.09899254888296127, 0.2651490569114685, 0.10916479676961899, 0.28130125999450684, 0.058687951415777206, 0.0, 0.0, 0.0], [0.10783901810646057, 0.11740631610155106, 0.19695693254470825, 0.15104907751083374, 0.3262477219104767, 0.03380884975194931, 0.06669214367866516, 0.0, 0.0], [0.03384784981608391, 0.23243722319602966, 0.13953465223312378, 0.19310569763183594, 0.20496727526187897, 0.016339682042598724, 0.07289475202560425, 0.10687284916639328, 0.0], [0.23444011807441711, 0.07891810685396194, 0.1519857794046402, 0.06546825915575027, 0.0652051568031311, 0.10796913504600525, 0.17505589127540588, 0.0456203818321228, 0.07533720135688782]]], [[[1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.8090201616287231, 0.19097986817359924, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.1832382082939148, 0.3397786617279053, 0.47698307037353516, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.3805065453052521, 0.13347622752189636, 0.30092647671699524, 0.1850907951593399, 0.0, 0.0, 0.0, 0.0, 0.0], [0.5122894644737244, 0.07694385200738907, 0.256824791431427, 0.08846534043550491, 0.06547650694847107, 0.0, 0.0, 0.0, 0.0], [0.22970852255821228, 0.13582248985767365, 0.20937861502170563, 0.17073699831962585, 0.11098608374595642, 0.14336727559566498, 0.0, 0.0, 0.0], [0.09341931343078613, 0.14378610253334045, 0.1972556859254837, 0.17643100023269653, 0.2930387854576111, 0.0408804751932621, 0.055188633501529694, 0.0, 0.0], [0.11508242040872574, 0.12171899527311325, 0.12208487093448639, 0.1762368232011795, 0.2284460961818695, 0.08813095092773438, 0.057572413235902786, 0.09072752296924591, 0.0], [0.12489533424377441, 0.12223611772060394, 0.07339038699865341, 0.163949117064476, 0.21940051019191742, 0.10057976841926575, 0.07395884394645691, 0.07117773592472076, 0.05041223391890526]], [[1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.630396842956543, 0.3696031868457794, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.21211498975753784, 0.22383400797843933, 0.5640510320663452, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.16418185830116272, 0.22461986541748047, 0.4036988317966461, 0.2074994444847107, 0.0, 0.0, 0.0, 0.0, 0.0], [0.43109169602394104, 0.054113633930683136, 0.39762797951698303, 0.053372230380773544, 0.06379447877407074, 0.0, 0.0, 0.0, 0.0], [0.19197046756744385, 0.15701176226139069, 0.15235097706317902, 0.15820172429084778, 0.21662917733192444, 0.12383583188056946, 0.0, 0.0, 0.0], [0.08732523769140244, 0.19746766984462738, 0.27495747804641724, 0.16601130366325378, 0.16353462636470795, 0.027989856898784637, 0.08271379768848419, 0.0, 0.0], [0.030109217390418053, 0.262539267539978, 0.19642876088619232, 0.23985430598258972, 0.12096359580755234, 0.007939137518405914, 0.0693647712469101, 0.07280096411705017, 0.0], [0.01843511126935482, 0.2516745626926422, 0.09083835780620575, 0.2360735982656479, 0.11003099381923676, 0.010639237239956856, 0.06030162423849106, 0.08971142768859863, 0.13229505717754364]], [[1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.5870454907417297, 0.41295450925827026, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.4823109805583954, 0.3582899272441864, 0.1593991369009018, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.43300753831863403, 0.1984279900789261, 0.21754881739616394, 0.15101568400859833, 0.0, 0.0, 0.0, 0.0, 0.0], [0.10235285758972168, 0.280404269695282, 0.18251553177833557, 0.2058979868888855, 0.22882941365242004, 0.0, 0.0, 0.0, 0.0], [0.20861537754535675, 0.19118784368038177, 0.10737765580415726, 0.18525347113609314, 0.14353443682193756, 0.16403119266033173, 0.0, 0.0, 0.0], [0.16503557562828064, 0.09813898801803589, 0.031133856624364853, 0.0675446018576622, 0.07803245633840561, 0.5139825940132141, 0.04613194614648819, 0.0, 0.0], [0.1584610491991043, 0.09127534180879593, 0.06304670125246048, 0.06900466233491898, 0.06389173865318298, 0.26928526163101196, 0.14328324794769287, 0.1417519897222519, 0.0], [0.25161445140838623, 0.029239315539598465, 0.0714222639799118, 0.01879066228866577, 0.024595780298113823, 0.3585717976093292, 0.10119058936834335, 0.07269758731126785, 0.07187756896018982]]], [[[1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.9372128248214722, 0.06278713047504425, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.34250253438949585, 0.5668380856513977, 0.09065934270620346, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.6290720701217651, 0.10822197794914246, 0.15049612522125244, 0.11220986396074295, 0.0, 0.0, 0.0, 0.0, 0.0], [0.970376193523407, 0.0021214366424828768, 0.015626508742570877, 0.0013014038559049368, 0.010574529878795147, 0.0, 0.0, 0.0, 0.0], [0.09130971878767014, 0.06638772040605545, 0.5523999929428101, 0.07101372629404068, 0.09935696423053741, 0.11953180283308029, 0.0, 0.0, 0.0], [0.021298406645655632, 0.26837092638015747, 0.14979466795921326, 0.3500915467739105, 0.11609084904193878, 0.020701348781585693, 0.07365228235721588, 0.0, 0.0], [0.008047293871641159, 0.20200438797473907, 0.10277985781431198, 0.36328285932540894, 0.06489216536283493, 0.008051928132772446, 0.061552878469228745, 0.1893887221813202, 0.0], [0.0011043534614145756, 0.1891116201877594, 0.03197162225842476, 0.3477092683315277, 0.04089963808655739, 0.001434291829355061, 0.021473770961165428, 0.1605575680732727, 0.20573781430721283]], [[1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.884773850440979, 0.1152261346578598, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.564574122428894, 0.2776378393173218, 0.15778802335262299, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.5638299584388733, 0.15842780470848083, 0.17728263139724731, 0.10045959800481796, 0.0, 0.0, 0.0, 0.0, 0.0], [0.8080127239227295, 0.03030439093708992, 0.1265506148338318, 0.014734375290572643, 0.020397968590259552, 0.0, 0.0, 0.0, 0.0], [0.20254802703857422, 0.07231037318706512, 0.536925733089447, 0.07049007713794708, 0.0683068335056305, 0.049418967217206955, 0.0, 0.0, 0.0], [0.11969973146915436, 0.1455283761024475, 0.13576865196228027, 0.1641465425491333, 0.3096204400062561, 0.08562075346708298, 0.03961546719074249, 0.0, 0.0], [0.09068313241004944, 0.19025327265262604, 0.12843848764896393, 0.19580651819705963, 0.18051451444625854, 0.08536171913146973, 0.05424069985747337, 0.07470159977674484, 0.0], [0.021546080708503723, 0.2296065092086792, 0.047056280076503754, 0.22095566987991333, 0.2430584728717804, 0.0647224709391594, 0.030332297086715698, 0.07678408920764923, 0.06593817472457886]], [[1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.9199612736701965, 0.08003874123096466, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.5722076892852783, 0.24290668964385986, 0.1848856806755066, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], [0.5284797549247742, 0.0987781509757042, 0.26617011427879333, 0.10657191276550293, 0.0, 0.0, 0.0, 0.0, 0.0], [0.8895872831344604, 0.01390649750828743, 0.04050413519144058, 0.012708736583590508, 0.04329325258731842, 0.0, 0.0, 0.0, 0.0], [0.16464175283908844, 0.13341635465621948, 0.26562222838401794, 0.1666385382413864, 0.1700018048286438, 0.09967926144599915, 0.0, 0.0, 0.0], [0.029521524906158447, 0.20171837508678436, 0.22596462070941925, 0.3160381615161896, 0.06304743885993958, 0.04469815641641617, 0.11901167035102844, 0.0, 0.0], [0.018640773370862007, 0.14681920409202576, 0.1846427470445633, 0.16796568036079407, 0.03296377509832382, 0.019411159679293633, 0.1117875948548317, 0.3177691102027893, 0.0], [0.00827923882752657, 0.1063661128282547, 0.11433600634336472, 0.11846795678138733, 0.026303714141249657, 0.009230039082467556, 0.05439882352948189, 0.21334347128868103, 0.349274605512619]]]], \"left_text\": [\"0\", \"2\", \"1\", \"2\", \"2\", \"0\", \"1\", \"2\", \"2\"], \"right_text\": [\"0\", \"2\", \"1\", \"2\", \"2\", \"0\", \"1\", \"2\", \"2\"]}], \"default_filter\": \"0\", \"root_div_id\": \"bertviz-7e6577d3f3db45f0af7076a248c133ab\", \"layer\": null, \"heads\": null, \"include_layers\": [0, 1, 2]} is a template marker that is replaced by actual params.\n    const TEXT_SIZE = 15;\n    const BOXWIDTH = 110;\n    const BOXHEIGHT = 22.5;\n    const MATRIX_WIDTH = 115;\n    const CHECKBOX_SIZE = 20;\n    const TEXT_TOP = 30;\n\n    console.log(\"d3 version\", d3.version)\n    let headColors;\n    try {\n        headColors = d3.scaleOrdinal(d3.schemeCategory10);\n    } catch (err) {\n        console.log('Older d3 version')\n        headColors = d3.scale.category10();\n    }\n    let config = {};\n    initialize();\n    renderVis();\n\n    function initialize() {\n        config.attention = params['attention'];\n        config.filter = params['default_filter'];\n        config.rootDivId = params['root_div_id'];\n        config.nLayers = config.attention[config.filter]['attn'].length;\n        config.nHeads = config.attention[config.filter]['attn'][0].length;\n        config.layers = params['include_layers']\n\n        if (params['heads']) {\n            config.headVis = new Array(config.nHeads).fill(false);\n            params['heads'].forEach(x => config.headVis[x] = true);\n        } else {\n            config.headVis = new Array(config.nHeads).fill(true);\n        }\n        config.initialTextLength = config.attention[config.filter].right_text.length;\n        config.layer_seq = (params['layer'] == null ? 0 : config.layers.findIndex(layer => params['layer'] === layer));\n        config.layer = config.layers[config.layer_seq]\n\n        let layerEl = $(`#${config.rootDivId} #layer`);\n        for (const layer of config.layers) {\n            layerEl.append($(\"<option />\").val(layer).text(layer));\n        }\n        layerEl.val(config.layer).change();\n        layerEl.on('change', function (e) {\n            config.layer = +e.currentTarget.value;\n            config.layer_seq = config.layers.findIndex(layer => config.layer === layer);\n            renderVis();\n        });\n\n        $(`#${config.rootDivId} #filter`).on('change', function (e) {\n            config.filter = e.currentTarget.value;\n            renderVis();\n        });\n    }\n\n    function renderVis() {\n\n        // Load parameters\n        const attnData = config.attention[config.filter];\n        const leftText = attnData.left_text;\n        const rightText = attnData.right_text;\n\n        // Select attention for given layer\n        const layerAttention = attnData.attn[config.layer_seq];\n\n        // Clear vis\n        $(`#${config.rootDivId} #vis`).empty();\n\n        // Determine size of visualization\n        const height = Math.max(leftText.length, rightText.length) * BOXHEIGHT + TEXT_TOP;\n        const svg = d3.select(`#${config.rootDivId} #vis`)\n            .append('svg')\n            .attr(\"width\", \"100%\")\n            .attr(\"height\", height + \"px\");\n\n        // Display tokens on left and right side of visualization\n        renderText(svg, leftText, true, layerAttention, 0);\n        renderText(svg, rightText, false, layerAttention, MATRIX_WIDTH + BOXWIDTH);\n\n        // Render attention arcs\n        renderAttention(svg, layerAttention);\n\n        // Draw squares at top of visualization, one for each head\n        drawCheckboxes(0, svg, layerAttention);\n    }\n\n    function renderText(svg, text, isLeft, attention, leftPos) {\n\n        const textContainer = svg.append(\"svg:g\")\n            .attr(\"id\", isLeft ? \"left\" : \"right\");\n\n        // Add attention highlights superimposed over words\n        textContainer.append(\"g\")\n            .classed(\"attentionBoxes\", true)\n            .selectAll(\"g\")\n            .data(attention)\n            .enter()\n            .append(\"g\")\n            .attr(\"head-index\", (d, i) => i)\n            .selectAll(\"rect\")\n            .data(d => isLeft ? d : transpose(d)) // if right text, transpose attention to get right-to-left weights\n            .enter()\n            .append(\"rect\")\n            .attr(\"x\", function () {\n                var headIndex = +this.parentNode.getAttribute(\"head-index\");\n                return leftPos + boxOffsets(headIndex);\n            })\n            .attr(\"y\", (+1) * BOXHEIGHT)\n            .attr(\"width\", BOXWIDTH / activeHeads())\n            .attr(\"height\", BOXHEIGHT)\n            .attr(\"fill\", function () {\n                return headColors(+this.parentNode.getAttribute(\"head-index\"))\n            })\n            .style(\"opacity\", 0.0);\n\n        const tokenContainer = textContainer.append(\"g\").selectAll(\"g\")\n            .data(text)\n            .enter()\n            .append(\"g\");\n\n        // Add gray background that appears when hovering over text\n        tokenContainer.append(\"rect\")\n            .classed(\"background\", true)\n            .style(\"opacity\", 0.0)\n            .attr(\"fill\", \"lightgray\")\n            .attr(\"x\", leftPos)\n            .attr(\"y\", (d, i) => TEXT_TOP + i * BOXHEIGHT)\n            .attr(\"width\", BOXWIDTH)\n            .attr(\"height\", BOXHEIGHT);\n\n        // Add token text\n        const textEl = tokenContainer.append(\"text\")\n            .text(d => d)\n            .attr(\"font-size\", TEXT_SIZE + \"px\")\n            .style(\"cursor\", \"default\")\n            .style(\"-webkit-user-select\", \"none\")\n            .attr(\"x\", leftPos)\n            .attr(\"y\", (d, i) => TEXT_TOP + i * BOXHEIGHT);\n\n        if (isLeft) {\n            textEl.style(\"text-anchor\", \"end\")\n                .attr(\"dx\", BOXWIDTH - 0.5 * TEXT_SIZE)\n                .attr(\"dy\", TEXT_SIZE);\n        } else {\n            textEl.style(\"text-anchor\", \"start\")\n                .attr(\"dx\", +0.5 * TEXT_SIZE)\n                .attr(\"dy\", TEXT_SIZE);\n        }\n\n        tokenContainer.on(\"mouseover\", function (d, index) {\n\n            // Show gray background for moused-over token\n            textContainer.selectAll(\".background\")\n                .style(\"opacity\", (d, i) => i === index ? 1.0 : 0.0)\n\n            // Reset visibility attribute for any previously highlighted attention arcs\n            svg.select(\"#attention\")\n                .selectAll(\"line[visibility='visible']\")\n                .attr(\"visibility\", null)\n\n            // Hide group containing attention arcs\n            svg.select(\"#attention\").attr(\"visibility\", \"hidden\");\n\n            // Set to visible appropriate attention arcs to be highlighted\n            if (isLeft) {\n                svg.select(\"#attention\").selectAll(\"line[left-token-index='\" + index + \"']\").attr(\"visibility\", \"visible\");\n            } else {\n                svg.select(\"#attention\").selectAll(\"line[right-token-index='\" + index + \"']\").attr(\"visibility\", \"visible\");\n            }\n\n            // Update color boxes superimposed over tokens\n            const id = isLeft ? \"right\" : \"left\";\n            const leftPos = isLeft ? MATRIX_WIDTH + BOXWIDTH : 0;\n            svg.select(\"#\" + id)\n                .selectAll(\".attentionBoxes\")\n                .selectAll(\"g\")\n                .attr(\"head-index\", (d, i) => i)\n                .selectAll(\"rect\")\n                .attr(\"x\", function () {\n                    const headIndex = +this.parentNode.getAttribute(\"head-index\");\n                    return leftPos + boxOffsets(headIndex);\n                })\n                .attr(\"y\", (d, i) => TEXT_TOP + i * BOXHEIGHT)\n                .attr(\"width\", BOXWIDTH / activeHeads())\n                .attr(\"height\", BOXHEIGHT)\n                .style(\"opacity\", function (d) {\n                    const headIndex = +this.parentNode.getAttribute(\"head-index\");\n                    if (config.headVis[headIndex])\n                        if (d) {\n                            return d[index];\n                        } else {\n                            return 0.0;\n                        }\n                    else\n                        return 0.0;\n                });\n        });\n\n        textContainer.on(\"mouseleave\", function () {\n\n            // Unhighlight selected token\n            d3.select(this).selectAll(\".background\")\n                .style(\"opacity\", 0.0);\n\n            // Reset visibility attributes for previously selected lines\n            svg.select(\"#attention\")\n                .selectAll(\"line[visibility='visible']\")\n                .attr(\"visibility\", null) ;\n            svg.select(\"#attention\").attr(\"visibility\", \"visible\");\n\n            // Reset highlights superimposed over tokens\n            svg.selectAll(\".attentionBoxes\")\n                .selectAll(\"g\")\n                .selectAll(\"rect\")\n                .style(\"opacity\", 0.0);\n        });\n    }\n\n    function renderAttention(svg, attention) {\n\n        // Remove previous dom elements\n        svg.select(\"#attention\").remove();\n\n        // Add new elements\n        svg.append(\"g\")\n            .attr(\"id\", \"attention\") // Container for all attention arcs\n            .selectAll(\".headAttention\")\n            .data(attention)\n            .enter()\n            .append(\"g\")\n            .classed(\"headAttention\", true) // Group attention arcs by head\n            .attr(\"head-index\", (d, i) => i)\n            .selectAll(\".tokenAttention\")\n            .data(d => d)\n            .enter()\n            .append(\"g\")\n            .classed(\"tokenAttention\", true) // Group attention arcs by left token\n            .attr(\"left-token-index\", (d, i) => i)\n            .selectAll(\"line\")\n            .data(d => d)\n            .enter()\n            .append(\"line\")\n            .attr(\"x1\", BOXWIDTH)\n            .attr(\"y1\", function () {\n                const leftTokenIndex = +this.parentNode.getAttribute(\"left-token-index\")\n                return TEXT_TOP + leftTokenIndex * BOXHEIGHT + (BOXHEIGHT / 2)\n            })\n            .attr(\"x2\", BOXWIDTH + MATRIX_WIDTH)\n            .attr(\"y2\", (d, rightTokenIndex) => TEXT_TOP + rightTokenIndex * BOXHEIGHT + (BOXHEIGHT / 2))\n            .attr(\"stroke-width\", 2)\n            .attr(\"stroke\", function () {\n                const headIndex = +this.parentNode.parentNode.getAttribute(\"head-index\");\n                return headColors(headIndex)\n            })\n            .attr(\"left-token-index\", function () {\n                return +this.parentNode.getAttribute(\"left-token-index\")\n            })\n            .attr(\"right-token-index\", (d, i) => i)\n        ;\n        updateAttention(svg)\n    }\n\n    function updateAttention(svg) {\n        svg.select(\"#attention\")\n            .selectAll(\"line\")\n            .attr(\"stroke-opacity\", function (d) {\n                const headIndex = +this.parentNode.parentNode.getAttribute(\"head-index\");\n                // If head is selected\n                if (config.headVis[headIndex]) {\n                    // Set opacity to attention weight divided by number of active heads\n                    return d / activeHeads()\n                } else {\n                    return 0.0;\n                }\n            })\n    }\n\n    function boxOffsets(i) {\n        const numHeadsAbove = config.headVis.reduce(\n            function (acc, val, cur) {\n                return val && cur < i ? acc + 1 : acc;\n            }, 0);\n        return numHeadsAbove * (BOXWIDTH / activeHeads());\n    }\n\n    function activeHeads() {\n        return config.headVis.reduce(function (acc, val) {\n            return val ? acc + 1 : acc;\n        }, 0);\n    }\n\n    function drawCheckboxes(top, svg) {\n        const checkboxContainer = svg.append(\"g\");\n        const checkbox = checkboxContainer.selectAll(\"rect\")\n            .data(config.headVis)\n            .enter()\n            .append(\"rect\")\n            .attr(\"fill\", (d, i) => headColors(i))\n            .attr(\"x\", (d, i) => i * CHECKBOX_SIZE)\n            .attr(\"y\", top)\n            .attr(\"width\", CHECKBOX_SIZE)\n            .attr(\"height\", CHECKBOX_SIZE);\n\n        function updateCheckboxes() {\n            checkboxContainer.selectAll(\"rect\")\n                .data(config.headVis)\n                .attr(\"fill\", (d, i) => d ? headColors(i): lighten(headColors(i)));\n        }\n\n        updateCheckboxes();\n\n        checkbox.on(\"click\", function (d, i) {\n            if (config.headVis[i] && activeHeads() === 1) return;\n            config.headVis[i] = !config.headVis[i];\n            updateCheckboxes();\n            updateAttention(svg);\n        });\n\n        checkbox.on(\"dblclick\", function (d, i) {\n            // If we double click on the only active head then reset\n            if (config.headVis[i] && activeHeads() === 1) {\n                config.headVis = new Array(config.nHeads).fill(true);\n            } else {\n                config.headVis = new Array(config.nHeads).fill(false);\n                config.headVis[i] = true;\n            }\n            updateCheckboxes();\n            updateAttention(svg);\n        });\n    }\n\n    function lighten(color) {\n        const c = d3.hsl(color);\n        const increment = (1 - c.l) * 0.6;\n        c.l += increment;\n        c.s -= increment;\n        return c;\n    }\n\n    function transpose(mat) {\n        return mat[0].map(function (col, i) {\n            return mat.map(function (row) {\n                return row[i];\n            });\n        });\n    }\n\n});",
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "from bertviz import head_view\n",
    "head_view(attns,tokens)"
   ]
  }
 ],
 "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": "3.10.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}