train_2scheduler.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "d8f44046-6fc5-4ed9-ac6d-173131a649a3",
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "from torch.utils.data import DataLoader\n",
    "from dataset import VitalDataset_fs\n",
    "idx = 3\n",
    "root_dir = f'../data/all_{idx}'\n",
    "trdt  = VitalDataset_fs(root_dir,f'../data/pd_gy/train_{idx}.json')\n",
    "valdt = VitalDataset_fs(root_dir,f'../data/pd_gy/val_{idx}.json') \n",
    "tedt  = VitalDataset_fs(root_dir,f'../data/pd_gy/test_{idx}.json') \n",
    "\n",
    "trdl  = torch.utils.data.DataLoader(trdt,4)\n",
    "valdl  = torch.utils.data.DataLoader(valdt,4)\n",
    "tedl  = torch.utils.data.DataLoader(valdt,4)\n",
    "\n",
    "from models import PrePostNet\n",
    "from train import get_classweight\n",
    "model = PrePostNet()\n",
    "\n",
    "from torch import nn\n",
    "device     = 'cuda' if torch.cuda.is_available() else 'cpu'\n",
    "\n",
    "batch_size = 4\n",
    "\n",
    "class_weight = get_classweight(trdt).to(device)\n",
    "loss       = nn.CrossEntropyLoss(class_weight)\n",
    "# loss       = nn.CrossEntropyLoss()\n",
    "\n",
    "\n",
    "lr         = 0.01\n",
    "params = [p for p in model.parameters() if p.requires_grad]\n",
    "opt       = torch.optim.SGD(params, lr=lr, momentum=0.9)\n",
    "scheduler = torch.optim.lr_scheduler.StepLR(opt, step_size=5, gamma=0.5)\n",
    "# opt  = torch.optim.Adam(params,lr=lr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "0d5fd189-bde2-4ee9-a94a-cc78cbf5a382",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0] acc/loss: 0.81/0.232283\n",
      "[1] acc/loss: 0.83/0.159888\n",
      "[2] acc/loss: 0.83/0.176197\n",
      "[3] acc/loss: 0.78/0.530542\n",
      "[4] acc/loss: 0.68/0.296496\n",
      "[5] acc/loss: 0.73/0.239821\n",
      "[6] acc/loss: 0.68/0.273841\n"
     ]
    }
   ],
   "source": [
    "from train import trainval\n",
    "best_model,val_losses = trainval(trdl,valdl,model,loss,opt,scheduler=scheduler,device=device, exist_acc=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "138063d2-6258-4c32-867e-1402bd19169f",
   "metadata": {},
   "source": [
    "## vis optim lr "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "c3222b1e-13cf-4d11-b63c-9e395c3ef5ff",
   "metadata": {},
   "outputs": [],
   "source": [
    "def visualize_scheduler(optimizer, scheduler, epochs):\n",
    "    lrs = []\n",
    "    for _ in range(epochs):\n",
    "        lrs.append(optimizer.param_groups[0]['lr'])\n",
    "        scheduler.step()\n",
    "        # 학습 과정이 없으므로 Optimizer.step()은 생략해도 됨\n",
    "\n",
    "    plt.plot(lrs)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "722fef3e-733b-42d8-8d83-9e287787bc41",
   "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"
    }
   ],
   "source": [
    "from matplotlib import pyplot as plt\n",
    "from torch import optim\n",
    "epochs = 300\n",
    "optimizer = optim.SGD([torch.tensor(1)], lr=0.1, momentum=0.9)\n",
    "scheduler = optim.lr_scheduler.StepLR(optimizer, step_size=10, gamma=0.5)\n",
    "\n",
    "visualize_scheduler(optimizer, scheduler, epochs)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "53f1025c-8547-430c-bb78-6bc56ec0bd03",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "torch0",
   "language": "python",
   "name": "torch0"
  },
  "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
}