vis_demo.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "64510d3a-f8ac-44bc-a909-20517a936fae",
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch.nn.functional as F\n",
    "from torch.utils.data import DataLoader\n",
    "import numpy as np\n",
    "from rot_learner.dataset_trans import TransDataset\n",
    "from rot_learner.util import retrieve_data\n",
    "from rot_learner import main_mis\n",
    "from rot_learner.dataset import RotDataset\n",
    "import copy\n",
    "from rot_learner.data_util import preprocess_config\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "58a098f9-87af-459f-9156-7506e2f4eb57",
   "metadata": {},
   "outputs": [],
   "source": [
    "config_path = './snapshots/cifar10_resnet18/cifar10_resnet18_dp_baseline_epoch_299.pt'\n",
    "PRED_ROTS = [0, 90, 180, 270]\n",
    "PRED_TASK_NUM = len(PRED_ROTS)\n",
    "device = 'cuda:0'\n",
    "task_name = 'Rotation'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "248b98cf-99ec-4f7c-b69a-92ce1a7a60a5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "using model resnet18_dp\n",
      "model weight loaded!\n",
      "Files already downloaded and verified\n",
      "Files already downloaded and verified\n"
     ]
    }
   ],
   "source": [
    "model, features_extractor, dloader,\\\n",
    "trainset_data, trainset_targets,\\\n",
    "valset_data, valset_targets,\\\n",
    "dataset, class_num = preprocess_config(config_path, None, device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "a14f7ab8-5011-4268-87b3-18cf1c7ea94c",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_ds = RotDataset(trainset_data, PRED_ROTS, trainset_targets, dataset=dataset, train_mode=True)\n",
    "val_ds = RotDataset(valset_data, PRED_ROTS, valset_targets, dataset=dataset, train_mode=False)\n",
    "test_ds = RotDataset(dloader.all_test_data, PRED_ROTS, dloader.all_test_targets, dataset=dataset, train_mode=False)\n",
    "\n",
    "trainloader = DataLoader(train_ds, batch_size=300, shuffle=True)\n",
    "valloader = DataLoader(val_ds, batch_size=300, shuffle=False)\n",
    "testloader = DataLoader(test_ds, batch_size=300, shuffle=False, pin_memory=False, num_workers=3)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "c5cbccf3-329e-409d-bf13-c098d1148419",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 1/3: 100%|██████████| 134/134 [00:49<00:00,  2.72it/s, aug_acc=48.97%]\n",
      "Epoch 2/3: 100%|██████████| 134/134 [00:48<00:00,  2.76it/s, aug_acc=52.92%]\n",
      "Epoch 3/3: 100%|██████████| 134/134 [00:47<00:00,  2.82it/s, aug_acc=53.78%]\n",
      "100%|██████████| 34/34 [00:07<00:00,  4.82it/s, acc=92.98%, aug_acc=52.86%]\n"
     ]
    }
   ],
   "source": [
    "rot_net = main_mis.train(features_extractor, PRED_TASK_NUM, 512, 3, trainloader, device, dataset)\n",
    "cls_res, rot_res = main_mis.test(rot_net, model, testloader, class_num, PRED_TASK_NUM, device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "6a994527-0695-43e2-a322-7ab7f144239b",
   "metadata": {},
   "outputs": [],
   "source": [
    "pred_scores, pred_corrects, gt_labels, pred_labels, pred_scores_all, pred_scores_for_gts, pred_features, pred_logits, pred_loss_all = cls_res\n",
    "rot_scores, rot_accs, rot_scores_full, rot_scores_all, rot_features, rot_logits, all_rot_corrects, rot_loss_all = rot_res"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "f341afb7-17a7-4729-800b-70fa6775ef6d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 252x252 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import rot_learner\n",
    "from importlib import reload\n",
    "reload(rot_learner.util)\n",
    "from rot_learner.util import show_ssl_cls, show_dist\n",
    "\n",
    "\n",
    "show_ssl_cls(np.array(pred_scores), pred_scores_for_gts, np.array(pred_corrects), np.array(rot_scores_full)[:, 0], rot_accs, gt_labels, 0, None, f'Rotation', \n",
    "            config={'title': f'{dataset}\\n{task_name}','xlabel':'Probing Confidence', 'alpha':1}\n",
    "        )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "05276c70-0734-411e-9b3e-20ffb321ccd1",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "findfont: Font family ['normal'] not found. Falling back to DejaVu Sans.\n",
      "findfont: Font family ['normal'] not found. Falling back to DejaVu Sans.\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAU4AAADMCAYAAAAcXEoOAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAAAtTElEQVR4nO3dd1gUV/s38C8dKQFBQRR9NIgoUYqAsoAFBQJGRSyxoMYafBBiiZRgbASxRhTE9hBJsSUqooixxWjsgg2TgIIJIkUUEQUWlgXO+4ev88uG4iy4y6L357q8LmbmzJn7YLxzZubMOUqMMQZCCCG8Kbd0AIQQ0tpQ4iSEEClR4iSEEClR4iSEEClR4iSEEClR4iSEEClR4iQKZcqUKVi8eLHEvq+//hpOTk6wsLBAQkJCC0VGyP9RonGcRJGUlJRAVVUVOjo6AIDbt2/j448/RmxsLKytraGrqwtNTc0m1//jjz/i2LFjyMjIQElJCXbv3g17e/s65RISErB9+3bk5eXB1NQU/v7+GDlyZJOvS94u1OMkCkVfX59LmgCQnZ0NZWVluLm5oX379k1OmlVVVQCAiooKODo6YtGiRQ2WPX36NBYvXoyJEyfi8OHDGDduHEJCQnDu3LkmXZu8fVRbOgDybtq9ezd2796NnJwc6Orqwt7eHjExMZgyZQq6dOmClStXIjQ0FIcOHQIAWFhYAADu3r2LP/74A1FRUfjjjz9QWVmJ999/H/PmzcPAgQO5+ocMGYIRI0bg+fPn+Pnnn9GlSxfs378f06ZNAwDk5uY2GFtcXByGDRvGlTUzM8Pt27cRFxeHQYMGyeYXQloV6nESuYuOjsb69esxadIkJCUlIS4uDpaWlnXKLV68GGFhYVBRUcGFCxdw4cIFAEBZWRmGDRuG77//HgkJCXBxcYG/vz/+/vtvifN/+OEHGBoaYt++fVi1ahWv2KqqqnDnzh0MGDBAYv+AAQNw69Yt1NTUNLHV5G1CPU4iV0KhEHFxcZg3bx4mT57M7f/ggw/qlNXV1YWuri4AoH379tz+/v37S5RbsGABfv31Vxw/fhz//e9/uf19+vRBYGCgVPE9e/YM1dXVaNeuncT+9u3bo6qqCs+fP4eBgYFUdZK3DyVOIldZWVkQiURwdnZuch3FxcWIjo7GlStXUFRUhJqaGohEIuTn50uUs7Kyam64hNSLEidpdUJDQ1FQUICgoCCYmppCU1MTCxYsgFgslijXpk0bqetu27YtVFVVUVRUJLG/qKgI6urq0NPTa1bs5O1AzziJXJmZmUFDQwMXL15sch0pKSmYOHEihg4dCgsLC7Rv377Rlz3SUFdXR58+fXD+/HmJ/efPn4eNjQ1UVFTeyHVI60Y9TiJX2tramD59OjZv3gxNTU04OTlBJBLh3Llz8PPz41VHt27dkJSUBDs7O9TW1mLTpk28X9o8efIERUVFePz4MQAgJycH2traaNeuHfccddasWfjss89gZWWFAQMG4Ny5czh16hS2bt3atEaTtw4lTiJ38+fPh4GBAb7//nusWrUK7733Xr2D0BuyatUqLFu2DOPGjUO7du0wc+ZMVFZW8jp337592Lx5M7f9xRdfAAACAgK4F0lubm6IiIjA9u3bsW7dOnTq1AmrVq2ioUiEQ18OEUKIlOgZJyGESIkSJyGESIkSJyGESIkSJyGESIkSJyGESKlVDUe6fv16S4dACHkL2NnZNev8VpU4geY3WFGlp6ejV69eLR2GTFDbWqe3tW1vogMm91v17Oxs9OnTR2Ii2aSkJLi6usLGxgb+/v4oKSmRd1iEEMKb3BNneHg4+vTpw21nZmZi6dKlWLt2LS5evIg2bdpgxYoV8g6LEEJ4k2viTE5Ohq6uLgQCAbcvKSkJQ4YMgYODA7S1tTFv3jycOnUKZWVl8gyNEEJ4k1viLCsrQ3R0NPdt8CuZmZncsggA0KVLF6ipqSE7O1teoRFCiFTk9nJo48aNGDNmDDp06CCxXygUcrN8v6Kjo4Py8vJ660lPT5dZjC2psrKS2tYKUdsU17F7L+rdP9i0+WlPLokzPT0dly9f5hbe+ictLa06t+VlZWXQ1taut6638S0f8Pa+wQSoba1Va2/bzRc5DRx50uy65ZI4r169iry8PLi6ugJ42cusqamBj48PBgwYgIyMDK7sw4cPIRaL0bVrV3mERgghUpNL4hw/fjw++ugjbnvnzp3Iy8vD8uXL8fTpU4wfPx6pqamwtLTEpk2b4O7uLrG2NiGEKBK5JM42bdpIrP+ipaUFdXV1GBgYwMDAACtWrMCiRYtQUlICgUDAeylXQghpCS3y5dC/l2wdMWIERowY0RKhEEKI1FrdJ5d87bna0IPhN2NS/y68y6ampmL9+vXIzMyEiooK3n//fYSFhdHytYS0Um9t4lQUZWVlmDNnDpYvXw4vLy+IxWKkpqZCXV29pUMjhDQRTSsnY3///TcAYPjw4VBRUYGmpiZcXFzQs2dPxMTESHyzn5ubCwsLC1RXVwMASkpK8MUXX8DFxQUODg7w9/fnyp4+fRre3t7o27cv3Nzc8NtvvwEASktLERYWBhcXFwwYMABRUVHcCpAPHjzA5MmTYWdnh/79+2P+/PkAAMYYIiMjIRAI0LdvX4wYMQL37t0DAFRVVWHNmjUYPHgwnJycsHTpUm5htOLiYvj5+cHe3h79+vXDpEmTUFtbK9tfKCEKgHqcMtatWzeoqKggJCQEw4YNg42NDfT09HidGxwcDC0tLSQnJ0NLSws3b94EAKSlpSEkJATR0dEQCAR48uQJNxY2NDQUhoaGOHnyJCoqKuDn5wcTExNMmDABmzZtgrOzM77//nuIxWLcuXMHAHDhwgWkpqbixIkT0NXVxV9//cV9lLB+/Xrk5OQgMTERqqqqWLRoEWJjY/H5558jPj4exsbGuHz5MgDg9u3bUFJSetO/QkIUDvU4ZUxHRwd79uyBkpISlixZAoFAgDlz5qCoqKjR8x4/fozffvsNK1asgJ6eHtTU1NCvXz8AwIEDBzBmzBg4OztDWVkZxsbGMDMzQ1FREc6dO4ewsDBoaWnB0NAQ06ZNQ3JyMgBAVVUV+fn5ePz4MTQ0NLgleVVVVVFeXo6//voLjDGYmZnByMgIjDH89NNPCAsLg76+PnR0dODn5ydR35MnT5Cfnw81NTXY29tT4iTvBOpxyoGZmRlWr14NALh//z6CgoIQGRmJbt26NXjOo0ePoKenV2/vtKCgoN41vvPz81FdXQ0XFxduX21tLUxMTAAAQUFB2LRpE8aOHQs9PT1Mnz4dY8eOhUAggK+vL8LDw5GXlwcPDw+EhIRAJBKhoqICo0eP5upjjHG34zNnzsTmzZsxY8YMAC/H63766adN+A0R0rpQ4pQzMzMzjB49Gj/++CMsLS2554UAJHqhHTp0wPPnz/HixQu89957EnWYmJggJ6fuqIEOHTpAXV0dV65cgapq3b/a9u3bIyIiAsDLN/3Tp0+Hg4MD/vOf/2Dq1KmYOnUqnj59ivnz5yMuLg6fffYZNDU1kZycDGNj4zr16ejoIDQ0FKGhobh37x4++eQT9OnTR2L2K0LeRnSrLmP379/Hzp078ejRIwAve4tHjx6FtbU1evXqhZSUFOTn56O8vBzbt2/nzjMyMsLAgQOxYsUKPH/+HGKxGCkpKQCAsWPHIiEhAZcvX0ZtbS0KCwtx//59GBkZwdnZGatXr0ZZWRlqa2uRk5ODa9euAQB+/vlnLg49PT0oKSlBWVkZaWlpuH37NsRiMdq0aQN1dXUoKytDWVkZ48aNQ2RkJJ4+fQoAKCwsxPnz5wEAv/76Kx48eADGGHR1daGiokK36uSdQD1OGdPR0cHt27cRHx+P0tJS6OrqwtXVFcHBwdDR0cGwYcMwcuRIaGtrY+7cuThz5gx37tq1a7Fq1SpuGFP//v3h4OAAKysrrFq1CpGRkcjNzUW7du2wdOlSmJmZYe3atVi/fj2GDRuG8vJydO7cGbNnzwYA3LlzB5GRkSgrK4OhoSEWL16Mzp07Izc3l6tLXV0dLi4umDlzJoCXt/exsbH4+OOP8ezZMxgbG2PixIkYMGAAHjx4gK+++grFxcV47733MHHiRDg6OrbI75kQeVJijLGWDoKv69ev05pDrRC1rXVq7W1r6CMYC9Unzc4jvG7Vvb298e233772TTAhhLwLeCXOuXPnIjU1FW5ubpg1axaSkpIgEolkHRshhCgkXonTw8MDmzdvxtmzZzF06FDs2bMHzs7O+OKLL7jBz4QQ8q6Q6uWQvr4+fHx8oKWlhbi4OJw8eRKpqalQVlbGsmXL4OTkJKs4CSFEYfBKnIwxXLhwAYcPH8bZs2dhY2ODTz/9FO7u7tDU1MSJEycQFBSEixcvyjpeQghpcbwSp4uLC9q2bQtvb28EBQXVGQz94YcfYteuXTIJkBBCFA2vxLlt2zb06dOn0TI//PDDGwmIEEIUHa+XQ68GQ/8bfVpHCHkX8UqcYrG43n009yIh5F3U6K36pEmToKSkhKqqKvj6+koce/ToEWxtbWUaXLOkxsu2fvvpvIoNGTIERUVFUFFR4fb5+Phg6dKlsoqMECJjjSbOcePGgTGGO3fuYOzYsdx+JSUlGBoa0nfJPG3btu21Q7VezdL+733/TLivI215QkjTNJo4fXx8AADW1tYwMzOTS0DvioSEBPz000+wsrLC4cOH4e7ujurqamhoaCA/Px8pKSnYsmULjI2NsXz5cqSnp8PY2BgLFy7E0KFDAbyc7f3f5cViMdasWYOCggLo6Ohg2rRpDT6jJoQ0TYOJMzExEaNGjQIA3Lx5k1u24d/+2RMl0klLS8NHH32Eixcv4vfff8e+fftw9OhR7NixA9u3b4dQKISPjw/GjBmDb775BtevX4e/vz8OHjyI999/HwAkyovFYgwdOhQbN26Evb09nj9/jtzc3BZuJSFvnwYTZ3JyMpc4Dx8+XG8ZJSUlSpw8zJ07V+IWOjg4GKqqqjAyMsKUKVMAABoaGgCAoUOHcjO3ZGRkQCgU4tNPP4WysjIEAgFcXV2RnJzMrU3/z/IaGhpQVVVFVlYWevbs2eAM8oSQ5mkwcf7vf//jfn4TYzQXLVqEK1euQCgUon379pg1axbGjRsHALh8+TJWrFiBgoICWFlZYfXq1ejUqVOzr6koYmNj6zzjTEhIQIcOHeqUfbXMBfBy3aEOHTpAWfn/Bj907NgRhYWF9ZYHgOjoaGzduhVff/01LCws8Pnnnyv2SzxCWiFew5GKi4tRXl4O4OULiIMHDyIxMVGq4Uh+fn44c+YMbty4gS1btmDjxo34/fffUVxcjICAAMybNw/Xrl1D7969sWDBgqa1ppV53WzpRkZGePTokcTvuaCgoN5lLF6xsrLC1q1bcenSJbi5uXFLABNC3hxeidPPzw8PHjwAAGzYsAE7d+5EfHw8twAZH+bm5lBXVwfwMmEoKSkhJycHp06dgrm5Oby8vKChoYHAwEBkZGTg/v37TWjO28XKygqampqIi4uDWCzG1atXcebMGQwbNqze8lVVVThy5AhKS0uhpqYGbW1tid4qIeTN4PXJZXZ2NjcTdFJSEvbt2wctLS0MHz4cYWFhvC+2fPlyHDp0CJWVlbC0tMSgQYMQFRUFCwsLroyWlha6dOmCrKyset/kp6en87qW/v9fW0dWSnjGUVVVBT8/P4kEZm1tjX79+qGiooJrT2VlJUpKSqCqqirRxuDgYGzfvh1bt26FgYEBPvvsM1RVVSE9Pb1OebFYjD179mD58uWora1Fp06dEBAQwPt3JiuVlZUtHoOsUNsUV8GjF/XutzBt/opBvGpQVlaGWCzG33//DV1dXXTs2BG1tbXc7Ttfy5cvx5IlS3Dz5k1cu3YN6urqEAqFMDAwkCino6PTYN28p/KX8ZT/Jq8vAgC4cOFCg8cCAgK4n9PT07Ft27Y6ZXr16oUPP/yw3vPrK79v3z6ekclPa1+CoTHUNsV180X9S2cAT5pdN6/EOXDgQMybNw8lJSXcbWJWVlajz9oaoqKiAnt7exw5cgR79+6FlpYWysrKJMqUl5dDW1tb6roJIUQeeCXOlStX4tChQ1BVVYW3tzcA4NmzZ9yQmKaoqalBTk4OzM3NcejQIW6/UChETk4Ounfv3uS6CSFElni9OVBXV8f48eMxZswYqKq+zLX9+/fHRx99xOsiT58+RXJyMsrLy1FTU4Pz588jOTkZAoEA7u7uyMzMxIkTJyASiRAbGwsLCwv6UokQorB49ThLSkqwc+dOpKenQygUShzbvXv3a89XUlLC3r17sWzZMu6lRVhYGPfpYExMDMLDwxEUFARra2ts2LChCU0hhBD54JU4P//8c1RVVcHLywtt2rSR+iIGBgaNzhDv5OSE48ePS10vIYS0BF6J8+bNm7hy5Qo3DpMQQt5lvJ5xWlhY4JGMx0USQkhrwavH6ejoiFmzZmH06NFo166dxDGa5IMQ8q7hlThTU1NhbGxcZ/lfmh2JEPIu4pU4aQVLQgj5P7xngHj27BkSExMRFxcHACgsLKTnnoSQdxKvxHnt2jV4enoiKSkJsbGxAIAHDx5g+fLlsoyNEEIUEq/EGRkZiY0bN+Kbb77hvhyytrZGWlqaTIMjhBBFxCtx5uXlQSAQAPi/yXfV1NTqXZmREELedrwSp5mZGc6fPy+x79KlS+jRo4dMgiKEEEXG6616aGgo/Pz8MHjwYFRWVmLp0qU4c+YMtmzZIuv4CCFE4fDqcdrY2ODIkSPo3r07xowZA1NTUxw4cABWVlayjo8QQhQO7znkjY2NMXv2bFnGQgghrcJrE2dWVhZiYmJw/fp1lJSUQF9fH/b29ggICKDJhgkh76RGb9Wzs7Px8ccfQyQSYcGCBdi6dSvmz5+PiooKfPzxx/jrr7/kFSchhCiMRnuc27dvh7e3N5YtWyaxf+zYsfjqq6/wv//9D6tWrZJpgIQQomga7XGmpKRgxowZ9R6bPn06rl69KpOgCCFEkTWaOIuLi2FqalrvsY4dO+LZs2cyCYoQQhTZa4cjvfpSqM6JysoNHiOEkLdZo884Kysr4evrW+8xxhhEIpFMgiKEEEXWaOJcuXJloyePGzfujQZDCCGtQaOJ08fHR15xEEJIq8F7ImNCCCEvySVxVlVVISwsDK6urrC1tYW3tzfOnTvHHb98+TI8PT1hbW2NKVOmIC8vTx5hEUJIk8glcVZXV8PExAQ//PADrl+/jvnz52P+/PnIzc1FcXExAgICMG/ePFy7dg29e/fGggUL5BEWIYQ0Ce9JPppDS0sLgYGB3LarqytMTU3xxx9/oKSkBObm5vDy8gIABAYGwtHREffv34eZmZk8wiOEEKnwSpxVVVWIjY3F0aNHUVJSguvXr+PChQvIzs7G5MmTpb5oUVERsrOz0b17d+zduxcWFhbcMS0tLXTp0gVZWVn1Js709HSpr9caVFZWUttaIWqb4ip49KLe/Ramze8v8qohMjIShYWFWL9+PTe1nLm5OVatWiV14hSLxVi0aBF8fHxgZmYGoVAIAwMDiTI6OjooLy+v9/xevXpJdb3WIj09ndrWClHbFNfNFzkNHHnS7Lp5Jc7Tp0/j5MmT0NLSgrLyy8eixsbGKCwslOpitbW1CA4OhpqaGpYsWQLgZQ+zrKxMolx5eTm0tbWlqpsQQuSF18uh+hZmKy4uhr6+Pu8LMcawePFiFBUVISYmBmpqagBe9lwzMjK4ckKhEDk5OTTXJyFEYfFKnJ6enggJCcHDhw8BAI8fP0Z4eDg++ugj3hdatmwZ7t+/j23btkFTU5Pb7+7ujszMTJw4cQIikQixsbGwsLCgF0OEEIXFK3EuWLAApqamGDlyJF68eIEPP/wQRkZGmDt3Lq+L5OXl4ccff0R6ejpcXFxga2sLW1tbHDlyBAYGBoiJiUFUVBQcHByQlpaGDRs2NKtRhBAiS7yecaqrqyMsLAxhYWEoLi5G27ZtpZoZqVOnTrh7926Dx52cnHD8+HHe9RFCSEvi1eP09/fHzz//jKqqKhgYGNB0coSQdxqvxNmvXz988803EAgECAkJwfnz51FbWyvr2AghRCHxSpzTpk3DgQMHcPDgQXTu3BmRkZEYMGAAIiIiZB0fIYQoHKm+Ve/atSsCAgIQFRUFCwsL7N69W1ZxEUKIwuL97VFOTg6OHj2K5ORkFBcXw9PTE/7+/rKMjRBCFBKvxDlmzBhkZ2dj6NChCA4OhrOzM1RV5TI/CCGEKBxe2W/mzJkYMmSIxMB1Qgh5VzWYOBlj3LAjT09PAKj3Tfqrb9cJIeRd0WDitLOzw40bNwAAlpaWdcZuvkqsrXnaKUIIaYoGE2dycjL38y+//CKXYAghpDVo8D7bxMSE+/n48ePo1KlTnT8nT56US5CEEKJIeD2gjI2NrXf/1q1b32gwhBDSGjT6Vv3y5csAXr4UunLlChhj3LHc3FyabJgQ8k5qNHEuXrwYACASiRAWFsbtV1JSQvv27fHll1/KNjpCCFFAjSbOM2fOAACCg4Oxdu1auQRECCGKjtczTkqahBDyf3h9OVRWVoaYmBikpKTg2bNnEs86z549K6vYCCFEIfHqcS5fvhx//vkn/P39UVJSgi+//BImJiaYNm2ajMMjhBDFw6vHefHiRRw7dgxt27aFiooK3Nzc0KdPH8yZM4eSJyHkncOrx1lbWwtdXV0AL9dBLy0tRfv27fHgwQOZBkcIIYqIV4+zZ8+eSElJgUAggL29PZYvXw5tbW107dpVxuERQoji4ZU4IyIiuBdCixcvxoYNG/DixQt6204IUVhmOfvrP/D+4GbXzStxdu7cmfvZ0NAQK1eubPaFCSGktWowcR44cIBXBWPHjuVVbteuXUhISMC9e/cwfPhwrF69mjt2+fJlrFixAgUFBbCyssLq1avRqVMnXvUSQoi8NZg4Dx8+/NqTlZSUeCdOIyMj+Pv74/z58xCJRNz+4uJiBAQEICIiAkOGDMHGjRuxYMEC/PTTT7zqJYQQeWswcf7www9v9EIeHh4AgDt37qCwsJDbf+rUKZibm8PLywsAEBgYCEdHR9y/fx9mZmZvNAZCCHkTeK978ezZMyQmJiIuLg4AUFhYiEePHjU7gMzMTFhYWHDbWlpa6NKlC7KysppdNyGEyAKvl0PXrl1DYGAgevfujRs3bmDWrFl48OABdu7ciW3btjUrAKFQCAMDA4l9Ojo6KC8vr7f827pUR2VlJbWtFaK2Ka7nL57Xu9/wDdTNK3FGRkZi48aNEAgEcHBwAABYW1sjLS2t2QFoaWmhrKxMYl95eXmDc3326tWr2ddUROnp6dS2Vojaprhe/K4ns7p53arn5eVBIBAAALdom5qaGmpqapodgLm5OTIyMrhtoVCInJwcdO/evdl1E0KILPBKnGZmZjh//rzEvkuXLqFHjx68L1RdXQ2RSITa2lrU1NRAJBKhuroa7u7uyMzMxIkTJyASiRAbGwsLCwt6MUQIUVi8btVDQ0Ph5+eHwYMHo7KyEkuXLsWZM2ewZcsW3hfaunUrNm/ezG0fOXIEAQEBCAwMRExMDMLDwxEUFARra2ts2LBB+pYQQoic8EqcNjY2OHLkCI4cOYIxY8bAxMQEBw4cQIcOHXhfKDAwEIGBgfUec3JywvHjx3nXRQghLYlX4gQAY2NjzJ49m9u+e/cuIiMjER0dLZPACCFEUTWaOCsqKrB9+3ZkZGTgP//5DwIDA/Hs2TOsXr0aly5dwqhRo+QUJiGEKI5GE2d4eDj+/PNPuLi44LfffsO9e/fw119/YdSoUfjqq6/qjL8khJB3QaOJ8/z58zh8+DAMDQ0xZcoUDB48GLt27YK9vb284iOEEIXT6HAkoVAIQ8OX4+w7dOgALS0tSpqEkHdeoz3OmpoaXLlyRWJVy39vvxoYTwgh74pGE6ehoSHCwsK4bX19fYltJSUl/PLLL7KLjhBCFFCjifPMmTPyioMQQloN3tPKEUIIeYkSJyGESIkSJyGESIkSJyGESIkSJyGESIkSJyGESIkSJyGESIn3tHKEEKKQUuPlfknqcRJCiJQocRJCiJToVp0Qovha4Ha8MdTjJIQQKVGPkxAiGw31Eu2nS3+OgqEeJyGESIl6nISQ12tK7/EtRomTvJ2acsv3JpNAY9eX13UU9fqt5Ha8MQqTOEtKSrB48WJcvHgRbdu2xcKFCzFixIiWDou8S+TVq2rpxMHz+vqPHgHlV2QcTPNd/btY7tdUmMQZHh4ONTU1XLx4Eenp6fDz80PPnj1hbm7e0qERRSWvBNTIdVpLciFvlkIkTqFQiJMnTyIpKQna2tqwt7fHkCFDcPjwYSxatKilwyPS+leikUguDfXeWroXRuSqJXqJb5JCJM7s7GyoqKigW7du3L6ePXsiJSWlTtnr16/LMzS5emvapmQluW1ihfxXPzfUxn+f01r8s21vGxm2TfV9GVUsJwqROIVCIXR0dCT26erqory8XGKfnZ2dPMMihJB6KcQ4Ti0tLZSVlUnsKysrg7a2dgtFRAghDVOIxNm1a1fU1NQgOzub25eRkYHu3bu3XFCEENIAhUicWlpacHd3R3R0NIRCIa5fv45ffvkF3t7eLR0aIYTUoRCJEwCWLVuG0tJS2NnZwdfXF6qqqsjIyKi3LGMM69atQ//+/dG/f3+sW7cOjDE5RyydkpISzJ07FzY2NnB1dUVSUlK95eLi4jB8+HDY2tpiyJAhiIuLk3Ok0uPbtleqqqrg5eWFgQMHyinCppOmbX/88Qd8fX1ha2sLJycnfPfdd3KMVHp821ZVVYWlS5fCyckJ/fr1w5w5c1BYWCjnaPnbtWsXRo8ejd69eyM0NLTRst9++y2cnZ3Rt29ffPHFF6iqquJ3EaZAFixYwObNm8fKyspYSkoK69u3L7t3716dcnv37mUeHh6soKCAPXr0iHl5ebE9e/a0QMT88W3bjh072O+//87EYjG7f/8+Gzx4MDt69GgLRMwf37a9smXLFjZp0iQ2YMAAOUbZNHzb9vTpU+bo6MgOHz7MRCIRKy0tZVlZWS0QMX/S/Dc5YsQI9uTJE1ZZWcmCgoLY3LlzWyBifk6cOMFOnTrFli5dykJCQhos99tvvzGBQMDu3bvHSkpK2OTJk9m6det4XUNhepyvxnLOmzevzljOf0tMTMSMGTPQoUMHGBsbY/r06Th06FALRM2PNG2bPXs2PvjgA6iqquL999/H0KFDcePGjRaImh9p2gYADx8+xJEjR/Dpp5/KOVLpSdO2b7/9Fi4uLhg5ciTU1dWho6MDMzOzFoiaH2nalpubCxcXF7Rr1w4aGhoYNmwYMjMzWyBqfjw8PODm5gZ9ff1GyyUmJmLs2LEwNzeHnp4e/P39eecRhUmcDY3lzMrKqlM2MzMTPXv2lCinyH+R0rTtnxhjSE1NVeiXZNK2LSIiAgsXLoSmpqa8Qmwyadp269Yt6OnpYcKECRAIBJgzZw7y8xV3hKc0bRs7dixu3LiBwsJCVFRUICkpqVU8Znmdf+cRCwsLFBUV4dmzZ689V2ESJ9+xnPWV1dXVhVAoVNjnnNK07Z9iYmJQW1uLMWPGyDK8ZpGmbadOnUJNTQ3c3d3lFV6zSNO2wsJCJCYmIiwsDGfPnoWpqSkWLlwor1ClJk3bunbtChMTEwwcOBB2dna4f/8+5s6dK69QZaa+PALgtf8uAQVKnNKM5dTS0pJoXFlZGbS0tKCkpCTzOJuiKeNUd+3ahcTEROzYsQPq6uqyDrHJ+LZNKBRi3bp1+PLLL+UZXrNI8/emoaEBd3d3WFlZQUNDA3PnzsXNmzdRWloqr3ClIk3bVqxYgaqqKly9ehW3bt2Cu7s7Zs+eLa9QZebfv4NXP/MZP64wiVOasZzm5uYSb9wzMjIUejIQacepHjhwADt27MB3332HDh06yCnKpuHbtgcPHiAvLw++vr5wdnZGYGAgnjx5AmdnZ+Tm5so5an6k+XuzsLCQ2FbU/4m/Ik3bMjIy4OPjA319fairq2PKlClIS0tDcXHr/t7c3Nwcd+/e5bYzMjLQrl07tG3b9rXnKkzilGYsp7e3N+Lj41FYWIjCwkLEx8fDx8enBaLmR5q2HTlyBFFRUYiPj0fnzp1bIFrp8G2bubk5zp49i8TERCQmJiIiIgKGhoZITEyEiYlJC0XfOGn+3kaPHo3Tp08jPT0dYrEYW7ZsgZ2dHXf7p2ikaVufPn1w+PBhlJaWQiwWY8+ePTAyMoKBgUELRP561dXVEIlEqK2tRU1NDUQiEaqrq+uU8/b2xoEDB5CVlYUXL15g69at/PPImxkA8GY8e/aM/fe//2XW1tZs0KBB7MiRI4wxxlJSUpiNjQ1Xrra2lq1Zs4Y5ODgwBwcHtmbNGlZbW9tSYfPCt22urq7M0tKS2djYcH+WLFnSUmHzwrdt/3TlypVWMRxJmrbt3r2bubi4MHt7e+bn58fy8/NbImTe+LatuLiYLVy4kDk6OjI7Ozs2YcIEdvv27ZYK+7Wio6NZjx49JP5ER0ezvLw8ZmNjw/Ly8riyO3fuZAKBgNna2rLQ0FAmEol4XUOJMQV9o0IIIQpKYW7VCSGktaDESQghUqLESQghUqLESQghUqLESQghUqLESQghUqLESRRKVFQU+vfvD2dn55YOhZAGUeJ8yw0ZMgS9e/eu83ncqFGjYGFhIZPPHffv3w9PT09uQt/Zs2fX+S66Pvn5+YiPj8exY8dw8eJFJCQkYOLEiW88vqaoqqrC6tWrMXDgQG6S6ZUrV7Z0WKSFKMQql0S2OnXqhOTkZEyZMgUAcPfuXVRUVMjkWteuXUNUVBTi4uJgaWmJkpIS/Prrr7zOzc/Ph76+PgwNDWUSW3Ps2LEDv//+O/bv3w8jIyPk5eUhNTX1jV6juroaqqr0T7I1oB7nO8Db2xuJiYncdmJiIkaNGiVR5uzZsxg1ahT69u2LQYMGISYmhjt27NgxDBkyhOs1njt3Ds7OzvVO8nDnzh3Y2NjA0tISAKCvrw8fHx9u+q7S0lIEBwfD0dERrq6u2LJlC2pra3Hp0iXMmDEDjx8/hq2tLebPn49ly5bh1q1bsLW1hb29PQAgNDQUy5cvx6xZs2Bra4sJEybgyZMnWLlyJRwcHODp6Yk///yTi2fHjh1wc3ODra0thg0bhlOnTnHHli1bhsDAQG573bp1+OSTT+qdnvDOnTtwc3ODsbExlJSUYGpqKvE7LCgoQEBAABwdHdG/f3+Eh4cDAGpra7Flyxa4urpCIBAgODiYmzEpNzcXFhYW2L9/PwYPHoxPPvkEwMtJXry8vODg4ICZM2ciLy8PwMv5WSMjIyEQCNC3b1+MGDEC9+7dq++vnMiaLL4VJYrD1dWVXbx4kXl4eLCsrCxWXV3NBgwYwHJzc1mPHj3Yw4cPGWMvvx3PyMhgNTU1LD09nQkEAnbq1CmunoULF7KQkBBWXFzMnJ2d2ZkzZ+q9XkpKCuvTpw/btGkTS01NrfPtb1BQEJszZw4rLS1lDx8+ZB4eHuynn37iYvjn9+sHDx5kEyZMkDg/JCSE9evXj925c4dVVlayKVOmMFdXV3bo0CFWXV3NNmzYwCZPnsyVP3bsGHv06BGrqalhycnJzNramhUWFjLGGBMKhczDw4MdPHiQpaSksH79+rGCgoJ62xUbG8sGDRrEdu3axTIyMiTmRqiurmYjRoxgK1euZOXl5ayyspKlpKQwxhjbv38/c3NzYzk5OaysrIzNnTuXLVq0iDHG2MOHD1mPHj1YUFAQKy8vZxUVFezUqVPMzc2NZWVlMbFYzGJjY9n48eMZYy+XevDx8WHPnz9ntbW1LCsri2sLkS9KnG+5V4kzNjaWrV+/np07d45NmzaNicViicT5bxEREWzlypXc9vPnz9mgQYPY8OHDXzvpyNmzZ5mfnx+zs7NjNjY2LDIyklVXV7Pq6mr2wQcfsMzMTK7s3r17uUTHN3EuXryY2/7++++Zp6cnt52RkcHs7OwajG3kyJES/0O4desWc3BwYIMHD2ZJSUkNnlddXc127drFxo8fzz744APm7OzMEhISGGOM3bhxg/Xv35+JxeI6502dOpXt2rWL275//z6ztLRkYrGYS5w5OTnc8ZkzZ3L/I2GMsZqaGmZlZcVyc3PZpUuXmIeHB7t58yarqalpMFYie/RA5R3h7e2NyZMnIzc3t96pw27fvo3169cjMzMTYrEYVVVV8PT05I6/99578PT0RHx8PKKjoxu91qBBgzBo0CDU1tbi6tWrmDdvHrp16wY3NzeIxWJ07NiRK9uxY0epV0z85zNQTU1NtGvXTmJbKBRy24mJiYiPj+dud4VCocTSCNbW1jA1NUVxcTG8vLwavKaKigp8fX3h6+uLyspKHDx4EGFhYbCyskJBQQE6duxY7/PJx48fo1OnTtx2p06dUF1djadPn3L7/jnnan5+PiIjI7FmzRpuH2MMhYWFEAgE8PX1RXh4OPLy8uDh4YGQkJA6M7kT2aNnnO+ITp06wdTUFOfOnYOHh0ed459//jmGDh2Kc+fO4fr165gwYYLEs7709HQcPHgQw4cPR0REBK9rKisrQyAQwNHREZmZmWjbti3U1NQk1uIpKCiAsbFxvec3dzLgvLw8fPnll1iyZAmuXr2K1NTUOhNe7969G2KxGEZGRryXYtbU1ISvry/ee+89ZGVlwcTEBAUFBfXO+fjqRdIr+fn5UFVVlUj+/2yniYkJVqxYgdTUVO5PWloa+vbtCwCYOnUqEhIScOzYMWRnZ7eK5aPfRpQ43yErV67Ed999By0trTrHysvLoaenBw0NDaSlpeHo0aPcMZFIhKCgICxYsACrVq3C48ePsXv37nqvcfr0aSQnJ+P58+dgjCEtLQ3Xrl2DtbU1VFRU4OnpiaioKJSVlSEvLw/x8fEYOXJkvXUZGhqisLCQ/1rX/1JRUQElJSVuwt2DBw9KLOr3999/Y+PGjVi3bh3Wrl2LuLg4pKen11vXt99+i6tXr6KyshLV1dU4dOgQysvLYWlpCSsrK7Rv3x5ff/01hEIhRCIRrl+/DgAYPnw4vvvuOzx8+BDl5eWIioqCl5dXg2/PJ0yYgB07dnBxlpaW4ueffwYApKWl4fbt2xCLxWjTpg3U1dWhrEz/hFsC3aq/Q7p06dLgsWXLlmHNmjUIDw9Hv3794OXlhRcvXgAAvv76a3To0AGTJk0C8PLt89SpU+Hs7IyuXbtK1KOnp4fNmzcjPDwcVVVVMDIywsyZM7nkuGTJEnz11Vdwc3ODhoYGxo0b1+BidI6OjujevTtcXFygpKSEq1evStXe7t27Y8aMGZgwYQKUlJS4UQPAy6E/QUFBmD17NrfS4YIFCxAcHIyDBw/WWeepTZs2WLNmDR48eAAlJSV07doVMTEx3Cz927ZtQ0REBFxdXQEAI0aMgJ2dHcaMGYPCwkJMnjwZIpEILi4uWLJkSYMxu7u7o7y8HAsXLkReXh50dXXh5OQELy8vlJeXIzIyErm5uVBXV4eLiwtmzpwp1e+EvBk0kTEhhEiJ+vmEECIlSpyEECIlSpyEECIlSpyEECIlSpyEECIlSpyEECIlSpyEECIlSpyEECIlSpyEECKl/wf627Nk4EkICgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 345.6x216 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "draw_config = {'kde': False, 'norm_hist': True, 'bins':30}\n",
    "\n",
    "show_dist(pred_scores, pred_corrects, None, 'Max Softmax Scores', draw_config, {'title': f'{dataset}', 'xlabel': 'Max Softmax Scores'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "076c2224-5e3f-4536-b937-3098322b8c75",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 345.6x216 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_dist(np.array(rot_scores_full)[:, 0], pred_corrects, None, 'Probing Confidence', draw_config, {'title': f'{dataset}', 'xlabel': 'Probing Confidence'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "189fd73f-796c-41c5-9ff7-40fafe8d4de9",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "conf2",
   "language": "python",
   "name": "conf2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}