read_results.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import os\n",
    "import os.path as osp\n",
    "import pandas as pd\n",
    "import glob\n",
    "import json\n",
    "import pickle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "results = {}\n",
    "\n",
    "POSEC3D_RES = sorted(glob.glob('./results/posec3d*'))\n",
    "for dir_ in POSEC3D_RES:\n",
    "    name = osp.basename(dir_)\n",
    "    res_dir = sorted(glob.glob(osp.join(dir_, '*/')))[-1]\n",
    "    res_file = glob.glob(osp.join(res_dir, '*.json'))[-1]\n",
    "    with open(res_file) as f:\n",
    "        res = json.load(f)\n",
    "    results[name] = res['acc/top1']\n",
    "\n",
    "STGCN_RES = sorted(glob.glob('./results/stgcn*'))\n",
    "for dir_ in STGCN_RES:\n",
    "    name = osp.basename(dir_)\n",
    "    res_dir = sorted(glob.glob(osp.join(dir_, '*/')))[-1]\n",
    "    res_file = glob.glob(osp.join(res_dir, '*.json'))[-1]\n",
    "    with open(res_file) as f:\n",
    "        res = json.load(f)\n",
    "    results[name] = res['acc/top1']\n",
    "\n",
    "STGCN_RES = sorted(glob.glob('./results/infogcn*'))\n",
    "for dir_ in STGCN_RES:\n",
    "    name = osp.basename(dir_)\n",
    "    res_file = glob.glob(osp.join(dir_, 'log.txt'))[-1]\n",
    "    with open(res_file) as f:\n",
    "        lines = f.readlines()\n",
    "    acc_line = [l for l in lines if 'Best accuracy:' in l][0].strip()\n",
    "    acc = float(acc_line.split(' ')[-1])\n",
    "    results[name] = acc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'posec3d_infact_plus_2d_primitive': 0.7105263157894737,\n",
       " 'posec3d_infact_plus_2d_primitive_FT': 0.6973684210526315,\n",
       " 'posec3d_infact_plus_2d_transitional': 0.5,\n",
       " 'posec3d_infact_plus_2d_transitional_FT': 0.6875,\n",
       " 'stgcn_infact_plus_2d_primitive': 0.5657894736842105,\n",
       " 'stgcn_infact_plus_2d_primitive_FT': 0.618421052631579,\n",
       " 'stgcn_infact_plus_2d_transitional': 0.375,\n",
       " 'stgcn_infact_plus_2d_transitional_FT': 0.5625,\n",
       " 'stgcn_infact_plus_3d_primitive': 0.6710526315789473,\n",
       " 'stgcn_infact_plus_3d_primitive_FT': 0.6973684210526315,\n",
       " 'stgcn_infact_plus_3d_transitional': 0.5625,\n",
       " 'stgcn_infact_plus_3d_transitional_FT': 0.65625,\n",
       " 'infogcn_Data_InfAct_plus_2d_primitive_InfAct_plus': 0.6973684210526315,\n",
       " 'infogcn_Data_InfAct_plus_2d_primitive_InfAct_plus_FT': 0.6842105263157895,\n",
       " 'infogcn_Data_InfAct_plus_2d_transitional_InfAct_plus': 0.46875,\n",
       " 'infogcn_Data_InfAct_plus_2d_transitional_InfAct_plus_FT': 0.5,\n",
       " 'infogcn_Data_InfAct_plus_3d_primitive_InfAct_plus': 0.6973684210526315,\n",
       " 'infogcn_Data_InfAct_plus_3d_primitive_InfAct_plus_FT': 0.7368421052631579,\n",
       " 'infogcn_Data_InfAct_plus_3d_transitional_InfAct_plus': 0.4375,\n",
       " 'infogcn_Data_InfAct_plus_3d_transitional_InfAct_plus_FT': 0.625,\n",
       " 'infogcn_Data_NTU_ntu60_2d': 0.9136289197549584,\n",
       " 'infogcn_Data_NTU_ntu60_3d_h36m': 0.8496391096015042}"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def read_preds(pth):\n",
    "    with open(pth, 'rb') as f:\n",
    "        preds_file = pickle.load(f)\n",
    "\n",
    "    gts = np.array([i['gt_labels']['item'].item() for i in preds_file])\n",
    "    preds = np.array([i['pred_labels']['item'].item() for i in preds_file])\n",
    "\n",
    "    return gts, preds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import confusion_matrix, ConfusionMatrixDisplay\n",
    "import matplotlib.pyplot as plt\n",
    "POSTURE_LABEL2IDS = {'Supine': 0, 'Prone': 1, 'Sitting': 2, 'Standing': 3, 'All-fours': 4}\n",
    "POSTURE_IDS2LABEL = {v:k for k,v in POSTURE_LABEL2IDS.items()}\n",
    "\n",
    "def plot_cm(gts, preds, ax=None):\n",
    "    labels = list(POSTURE_IDS2LABEL.values())\n",
    "    cm = ConfusionMatrixDisplay.from_predictions(y_true=gts, y_pred=preds, normalize='true', display_labels=labels, ax=ax)\n",
    "    return cm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "MODEL = 'posec3d'\n",
    "DIM = '2d'\n",
    "FOLD = 'primitive'\n",
    "gts, preds = read_preds(f'./results_test/{MODEL}_infact_plus_{DIM}_{FOLD}/preds.pkl')\n",
    "\n",
    "with open(f'./results_test/{MODEL}_infact_plus_{DIM}_{FOLD}/preds.pkl', 'rb') as f:\n",
    "    a = pickle.load(f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(512,)"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "a[0]['pred_feats']['item'].detach().numpy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'model: stgcn, dim: 3d, actions: primitive, accuracy: 45.39%')"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "MODEL = 'stgcn'\n",
    "DIM = '3d'\n",
    "FOLD = 'primitive'\n",
    "gts, preds = read_preds(f'./results_test/{MODEL}_infact_plus_{DIM}_{FOLD}/preds.pkl')\n",
    "acc = sum(gts==preds)/len(preds)\n",
    "fig, ax = plt.subplots()\n",
    "plot_cm(gts, preds, ax)\n",
    "ax.set_title(f'model: {MODEL}, dim: {DIM}, actions: {FOLD}, accuracy: {acc*100:.2f}%')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "DeepLearning39",
   "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": "3.9.16"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}