6.3_VideoModelTraining.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 6.3 Training Models\n",
    "\n",
    "Module 6 - Video Action Recognition\n",
    "\n",
    "For book, references and training materials, please check this project website [http://activefitness.ai/ai-in-sports-with-python](http://activefitness.ai/ai-in-sports-with-python).\n",
    "\n",
    "Book: [Applied Machine Learning for Health and Fitness](https://www.apress.com/us/book/9781484257715), Chapter 9"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "cuda\n"
     ]
    }
   ],
   "source": [
    "import torch\n",
    "import torchvision\n",
    "import torchvision.models as models    \n",
    "\n",
    "# check if cuda is available\n",
    "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
    "print(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torchvision\n",
    "import torchvision.models as models  \n",
    "\n",
    "# Normalization: Kinetics 400\n",
    "mean = [0.43216, 0.394666, 0.37645]  \n",
    "std = [0.22803, 0.22145, 0.216989] \n",
    "\n",
    "def normalize(video): \n",
    "    return video.permute(3, 0, 1, 2).to(torch.float32) / 255\n",
    "\n",
    "def resize(video, size): \n",
    "    return torch.nn.functional.interpolate(video, size=size, scale_factor=None, mode='bilinear', align_corners=False)\n",
    "\n",
    "def crop(video, output_size): \n",
    "    # center crop    \n",
    "    h, w = video.shape[-2:] \n",
    "    th, tw = output_size \n",
    "    i = int(round((h - th) / 2.)) \n",
    "    j = int(round((w - tw) / 2.)) \n",
    "    return video[..., i:(i + th), j:(j + tw)]\n",
    "\n",
    "def normalize_base(video, mean, std): \n",
    "    shape = (-1,) + (1,) * (video.dim() - 1) \n",
    "    mean = torch.as_tensor(mean).reshape(shape) \n",
    "    std = torch.as_tensor(std).reshape(shape) \n",
    "    return (video - mean) / std"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([195, 720, 1280, 3]) torch.Size([2, 311296]) {'video_fps': 29.97002997002997, 'audio_fps': 48000}\n"
     ]
    }
   ],
   "source": [
    "import torchvision.io \n",
    "video_file = 'media/surfing_backflip.mp4'\n",
    "video, audio, info = torchvision.io.read_video(video_file, pts_unit='sec') \n",
    "print(video.shape, audio.shape, info)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "frames 3, size 195 112\n"
     ]
    }
   ],
   "source": [
    "video = normalize(video) \n",
    "video = resize(video,(128, 171)) \n",
    "video = crop(video,(112, 112)) \n",
    "video = normalize_base(video, mean=mean, std=std)\n",
    "shape = video.shape\n",
    "print(f'frames {shape[0]}, size {shape[1]} {shape[2]}') "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor(337)\n"
     ]
    }
   ],
   "source": [
    "# score the video\n",
    "score = model(video.unsqueeze(0)) \n",
    "# get prediction with max score\n",
    "prediction = score.argmax() \n",
    "print(prediction)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "surfing water\n"
     ]
    }
   ],
   "source": [
    "from utils.kinetics import kinetics\n",
    "classes = kinetics.classes()\n",
    "print(classes[prediction.item()])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torch.nn as nn\n",
    "import torchvision\n",
    "import torchvision.models as models\n",
    "from torch.utils.data import DataLoader as DataLoader\n",
    "from torchvision import transforms\n",
    "from torchvision.datasets.kinetics import Kinetics400\n",
    "from torchvision.datasets.samplers import DistributedSampler, UniformClipSampler, RandomClipSampler\n",
    "import matplotlib.pyplot as plt\n",
    "from pathlib import Path\n",
    "\n",
    "Path.ls = lambda x: [o.name for o in x.iterdir()]\n",
    "from torchvision.io.video import read_video\n",
    "from functools import partial as partial\n",
    "read_video = partial(read_video, pts_unit='sec')\n",
    "torchvision.io.read_video = partial(torchvision.io.read_video, pts_unit = 'sec')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "base_dir = Path('data/kinetics400/')\n",
    "data_dir = base_dir/'dataset'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "!tree {data_dir/'train'}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|                                                                    | 0/41 [00:00<?, ?it/s]C:\\ProgramData\\Anaconda3\\envs\\rl\\lib\\site-packages\\torchvision\\io\\video.py:106: UserWarning: The pts_unit 'pts' gives wrong results and will be removed in a follow-up version. Please use pts_unit 'sec'.\n",
      "  warnings.warn(\"The pts_unit 'pts' gives wrong results and will be removed in a \" +\n",
      "100%|███████████████████████████████████████████████████████████| 41/41 [18:08<00:00, 26.56s/it]\n"
     ]
    }
   ],
   "source": [
    "data = torchvision.datasets.Kinetics400(\n",
    "            data_dir/'train',\n",
    "            frames_per_clip=32,\n",
    "            step_between_clips=1,\n",
    "            frame_rate=None,\n",
    "            extensions=('mp4',),\n",
    "            num_workers=0\n",
    "        )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([32, 480, 272, 3])\n"
     ]
    }
   ],
   "source": [
    "print((data[0][0]).shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Visualizing dataset\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Length: 153939\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>filepath</th>\n",
       "      <th>frames</th>\n",
       "      <th>fps</th>\n",
       "      <th>clips</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>data\\kinetics400\\dataset\\train\\playing_tennis\\...</td>\n",
       "      <td>300</td>\n",
       "      <td>30.000000</td>\n",
       "      <td>269</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>data\\kinetics400\\dataset\\train\\playing_tennis\\...</td>\n",
       "      <td>140</td>\n",
       "      <td>14.000000</td>\n",
       "      <td>109</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>data\\kinetics400\\dataset\\train\\playing_tennis\\...</td>\n",
       "      <td>100</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>69</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>data\\kinetics400\\dataset\\train\\playing_tennis\\...</td>\n",
       "      <td>100</td>\n",
       "      <td>10.000000</td>\n",
       "      <td>69</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>data\\kinetics400\\dataset\\train\\playing_tennis\\...</td>\n",
       "      <td>89</td>\n",
       "      <td>15.000000</td>\n",
       "      <td>58</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>640</th>\n",
       "      <td>data\\kinetics400\\dataset\\train\\surfing_water\\c...</td>\n",
       "      <td>300</td>\n",
       "      <td>29.970030</td>\n",
       "      <td>269</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>641</th>\n",
       "      <td>data\\kinetics400\\dataset\\train\\surfing_water\\c...</td>\n",
       "      <td>240</td>\n",
       "      <td>23.974427</td>\n",
       "      <td>209</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>642</th>\n",
       "      <td>data\\kinetics400\\dataset\\train\\surfing_water\\c...</td>\n",
       "      <td>300</td>\n",
       "      <td>29.970030</td>\n",
       "      <td>269</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>643</th>\n",
       "      <td>data\\kinetics400\\dataset\\train\\surfing_water\\c...</td>\n",
       "      <td>250</td>\n",
       "      <td>25.000000</td>\n",
       "      <td>219</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>644</th>\n",
       "      <td>data\\kinetics400\\dataset\\train\\surfing_water\\c...</td>\n",
       "      <td>250</td>\n",
       "      <td>25.002519</td>\n",
       "      <td>219</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>645 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                              filepath  frames        fps  \\\n",
       "0    data\\kinetics400\\dataset\\train\\playing_tennis\\...     300  30.000000   \n",
       "1    data\\kinetics400\\dataset\\train\\playing_tennis\\...     140  14.000000   \n",
       "2    data\\kinetics400\\dataset\\train\\playing_tennis\\...     100  10.000000   \n",
       "3    data\\kinetics400\\dataset\\train\\playing_tennis\\...     100  10.000000   \n",
       "4    data\\kinetics400\\dataset\\train\\playing_tennis\\...      89  15.000000   \n",
       "..                                                 ...     ...        ...   \n",
       "640  data\\kinetics400\\dataset\\train\\surfing_water\\c...     300  29.970030   \n",
       "641  data\\kinetics400\\dataset\\train\\surfing_water\\c...     240  23.974427   \n",
       "642  data\\kinetics400\\dataset\\train\\surfing_water\\c...     300  29.970030   \n",
       "643  data\\kinetics400\\dataset\\train\\surfing_water\\c...     250  25.000000   \n",
       "644  data\\kinetics400\\dataset\\train\\surfing_water\\c...     250  25.002519   \n",
       "\n",
       "     clips  \n",
       "0      269  \n",
       "1      109  \n",
       "2       69  \n",
       "3       69  \n",
       "4       58  \n",
       "..     ...  \n",
       "640    269  \n",
       "641    209  \n",
       "642    269  \n",
       "643    219  \n",
       "644    219  \n",
       "\n",
       "[645 rows x 4 columns]"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "from utils.video_classification.helpers import to_dataframe\n",
    "\n",
    "to_dataframe(data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Length: 153939\n"
     ]
    }
   ],
   "source": [
    "VIDEO_NUMBER = 130\n",
    "video_table = to_dataframe(data)\n",
    "video_info = video_table['filepath'][VIDEO_NUMBER]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<video src=\"data\\kinetics400\\dataset\\train\\playing_tennis\\SNbQWJoCvo8.mp4\" controls  width=\"400\" >\n",
       "      Your browser does not support the <code>video</code> element.\n",
       "    </video>"
      ],
      "text/plain": [
       "<IPython.core.display.Video object>"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from IPython.display import Video\n",
    "from IPython.core.interactiveshell import InteractiveShell\n",
    "InteractiveShell.ast_node_interactivity = \"all\"\n",
    "\n",
    "Video(video_info, width=400, embed=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def show_clip_start_end(f):\n",
    "    last = len(f)\n",
    "    plt.imshow(f[0])\n",
    "    plt.title(f'frame: 1')\n",
    "    plt.axis('off')\n",
    "    plt.show()\n",
    "    plt.imshow(f[last-1])\n",
    "    plt.title(f'frame: {last}')\n",
    "    plt.axis('off')\n",
    "    plt.show()\n",
    "\n",
    "show_clip_start_end(data[0][0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Video normalization\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "import utils.video_classification.transforms as T\n",
    "\n",
    "t = torchvision.transforms.Compose([\n",
    "        T.ToFloatTensorInZeroOne(),\n",
    "        T.Resize((128, 171)),\n",
    "        T.RandomHorizontalFlip(),\n",
    "        T.Normalize(mean=[0.43216, 0.394666, 0.37645],\n",
    "                            std=[0.22803, 0.22145, 0.216989]),\n",
    "        T.RandomCrop((112, 112))\n",
    "    ])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████████████████████████████████| 10/10 [00:39<00:00,  3.97s/it]\n"
     ]
    }
   ],
   "source": [
    "train_data = torchvision.datasets.Kinetics400(\n",
    "            data_dir/'train',\n",
    "            frames_per_clip=32,\n",
    "            step_between_clips=1,\n",
    "            frame_rate=None,\n",
    "            transform=t,\n",
    "            extensions=('mp4',),\n",
    "            num_workers=0\n",
    "        )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "from torch_lr_finder import LRFinder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(torch.Size([4, 3, 32, 112, 112]), torch.Size([4]))"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from utils.video_classification.first_clip_sampler import FirstClipSampler\n",
    "from torch.utils.data.dataloader import default_collate\n",
    "\n",
    "def collate_fn(batch):\n",
    "    # remove audio from the batch\n",
    "    batch = [(d[0], d[2]) for d in batch]\n",
    "    return default_collate(batch)\n",
    "\n",
    "train_sampler = FirstClipSampler(train_data.video_clips, 2)\n",
    "train_dl = DataLoader(train_data, batch_size=4, sampler=train_sampler, collate_fn=collate_fn, pin_memory=True)\n",
    "x,y = next(iter(train_dl))\n",
    "x.shape, y.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "criterion = nn.CrossEntropyLoss()\n",
    "optimizer = torch.optim.Adam(model.parameters(), lr=1e-6, weight_decay=1e-2)\n",
    "# if you are getting memory problems running this, \n",
    "# try reducing DataLoader batch_size above to 16 or even 4\n",
    "lr_finder = LRFinder(model, optimizer, criterion,device=device)\n",
    "lr_finder.range_test(train_dl, end_lr=10, num_iter=90)\n",
    "lr_finder.plot()\n",
    "lr_finder.reset()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_data = torchvision.datasets.Kinetics400(\n",
    "            data_dir/'train',\n",
    "            frames_per_clip=32,\n",
    "            step_between_clips=1,\n",
    "            frame_rate=None,\n",
    "            transform=t,\n",
    "            extensions=('mp4',),\n",
    "            num_workers=0\n",
    "        )\n",
    "\n",
    "valid_data = torchvision.datasets.Kinetics400(\n",
    "            data_dir/'valid',\n",
    "            frames_per_clip=32,\n",
    "            step_between_clips=1,\n",
    "            frame_rate=None,\n",
    "            transform=t,\n",
    "            extensions=('mp4',),\n",
    "            num_workers=0\n",
    "        )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_sampler = FirstClipSampler(train_data.video_clips, 2)\n",
    "train_dl = DataLoader(train_data, \n",
    "                      batch_size=4, \n",
    "                      sampler=train_sampler, \n",
    "                      collate_fn=collate_fn, \n",
    "                      pin_memory=True)\n",
    "valid_sampler = FirstClipSampler(valid_data.video_clips, 2)\n",
    "valid_dl = DataLoader(valid_data, \n",
    "                      batch_size=4, \n",
    "                      sampler=valid_sampler, \n",
    "                      collate_fn=collate_fn, pin_memory=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "cache_dir = data_dir/'.cache'\n",
    "if not os.path.exists(cache_dir):\n",
    "    os.makedirs(cache_dir)\n",
    "cache_dir.ls()\n",
    "\n",
    "torch.save(train_data, f'{cache_dir}/train')\n",
    "torch.save(valid_data, f'{cache_dir}/valid')\n",
    "train_data = torch.load(cache_dir/'train')\n",
    "valid_data = torch.load(cache_dir/'valid') "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torchvision\n",
    "import torchvision.models as models\n",
    "\n",
    "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
    "print(device)\n",
    "\n",
    "model = models.video.r2plus1d_18()\n",
    "model.cuda()\n",
    "lr = 1e-2\n",
    "criterion = nn.CrossEntropyLoss()\n",
    "optim = torch.optim.Adam(model.parameters(), lr)\n",
    "lr_scheduler = torch.optim.lr_scheduler.OneCycleLR(optim, max_lr=5e-1, steps_per_epoch=len(train_dl), epochs=10)\n",
    "metrics_dir = cache_dir/'train-metrics'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import sys\n",
    "import time\n",
    "import datetime\n",
    "from utils.video_classification.train import train_one_epoch, evaluate\n",
    "\n",
    "start_time = time.time()\n",
    " \n",
    "for epoch in range(10):\n",
    "    train_one_epoch(model, \n",
    "                    criterion, \n",
    "                    optim, \n",
    "                    lr_scheduler, \n",
    "                    train_dl, device, \n",
    "                    epoch, print_freq=100)\n",
    "    evaluate(model, \n",
    "             criterion, \n",
    "             valid_dl, \n",
    "             device)\n",
    "\n",
    "total_time = time.time() - start_time\n",
    "total_time_str = str(datetime.timedelta(seconds=int(total_time)))\n",
    "print('Training time {}'.format(total_time_str))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "SAVED_MODEL_PATH = './videoresnet_action.pth'\n",
    "torch.save(model.state_dict(), SAVED_MODEL_PATH)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "rl",
   "language": "python",
   "name": "rl"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}