tutorials/PythonColabTutorial_Merced.ipynb

{
 "nbformat": 4,
 "nbformat_minor": 0,
 "metadata": {
  "colab": {
   "name": "PythonColabTutorial_Merced.ipynb",
   "provenance": [],
   "collapsed_sections": [],
   "toc_visible": true,
   "include_colab_link": true
  },
  "kernelspec": {
   "name": "python3",
   "display_name": "Python 3"
  },
  "accelerator": "GPU"
 },
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "view-in-github",
    "colab_type": "text"
   },
   "source": [
    "<a href=\"https://colab.research.google.com/github/spaceml-org/Self-Supervised-Learner/blob/simsiam/tutorials/PythonColabTutorial_Merced.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Hk6INtB3dO_5"
   },
   "source": [
    "# Tutorial on Using SSL-Curator Repo\r\n",
    "In this tutorial you'll see: \r\n",
    "- The commands to train an SSL model\r\n",
    "-Load the SSL model, to run your own custom inference script\r\n",
    "-Load the SSL model, to run your own dataset for inference with the DALI pipeline\r\n",
    "-The command to train a finetuner model based on your previously trained SSL Model\r\n",
    "-Running custom and DALI inference on the trained model\r\n",
    "\r\n"
   ]
  },
  {
   "cell_type": "code",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "sveuC0AmV6Ye",
    "outputId": "61468f66-1ad9-4f56-a472-e7b7ce99bfbc"
   },
   "source": [
    "# installs\n",
    "!pip install -q split-folders\n",
    "!pip install -q torch==1.7.1 torchvision==0.8.2 torchaudio==0.7.2 torchtext==0.6.0\n",
    "!pip install -q pytorch-lightning==1.1.8\n",
    "!pip install -q pytorch-lightning-bolts\n",
    "!pip install -q --extra-index-url https://developer.download.nvidia.com/compute/redist nvidia-dali-cuda100\n",
    "!pip install -q wandb\n",
    "\n",
    "# download data\n",
    "!gdown http://weegee.vision.ucmerced.edu/datasets/UCMerced_LandUse.zip\n",
    "!unzip -qq UCMerced_LandUse.zip"
   ],
   "execution_count": 1,
   "outputs": [
    {
     "output_type": "stream",
     "text": [
      "\u001b[K     |████████████████████████████████| 776.8MB 23kB/s \n",
      "\u001b[K     |████████████████████████████████| 12.8MB 248kB/s \n",
      "\u001b[K     |████████████████████████████████| 7.6MB 55.2MB/s \n",
      "\u001b[K     |████████████████████████████████| 71kB 9.7MB/s \n",
      "\u001b[K     |████████████████████████████████| 1.2MB 53.8MB/s \n",
      "\u001b[K     |████████████████████████████████| 696kB 8.1MB/s \n",
      "\u001b[K     |████████████████████████████████| 276kB 12.6MB/s \n",
      "\u001b[K     |████████████████████████████████| 112kB 22.3MB/s \n",
      "\u001b[K     |████████████████████████████████| 829kB 18.5MB/s \n",
      "\u001b[K     |████████████████████████████████| 1.3MB 30.0MB/s \n",
      "\u001b[K     |████████████████████████████████| 296kB 25.9MB/s \n",
      "\u001b[K     |████████████████████████████████| 143kB 36.9MB/s \n",
      "\u001b[?25h  Building wheel for PyYAML (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
      "  Building wheel for future (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
      "\u001b[K     |████████████████████████████████| 256kB 8.8MB/s \n",
      "\u001b[K     |████████████████████████████████| 389.9MB 45kB/s \n",
      "\u001b[K     |████████████████████████████████| 2.0MB 5.6MB/s \n",
      "\u001b[K     |████████████████████████████████| 102kB 11.7MB/s \n",
      "\u001b[K     |████████████████████████████████| 133kB 36.0MB/s \n",
      "\u001b[K     |████████████████████████████████| 163kB 34.7MB/s \n",
      "\u001b[K     |████████████████████████████████| 71kB 10.8MB/s \n",
      "\u001b[?25h  Building wheel for subprocess32 (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
      "  Building wheel for pathtools (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
      "Downloading...\n",
      "From: http://weegee.vision.ucmerced.edu/datasets/UCMerced_LandUse.zip\n",
      "To: /content/UCMerced_LandUse.zip\n",
      "100% 332M/332M [00:04<00:00, 69.4MB/s]\n"
     ],
     "name": "stdout"
    }
   ]
  },
  {
   "cell_type": "code",
   "metadata": {
    "id": "95y1Skf1fZsg"
   },
   "source": [
    "import os\n",
    "import itertools\n",
    "import shutil\n",
    "import PIL\n",
    "import matplotlib.pyplot as plt\n",
    "import torch\n",
    "from torch import nn\n",
    "from tqdm.notebook import tqdm\n",
    "from torchvision import transforms\n",
    "from torchvision.datasets import ImageFolder\n",
    "from torch.utils.data import DataLoader\n",
    "import numpy as np\n",
    "import PIL.Image as Image\n",
    "import time\n",
    "\n",
    "# not logging on wandb in this demo\n",
    "os.environ[\"WANDB_MODE\"] = \"disabled\""
   ],
   "execution_count": 2,
   "outputs": []
  },
  {
   "cell_type": "code",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "qTmsIw8GWOAC",
    "outputId": "ef27c46d-24ea-4ae0-fcf2-6304caf5a537"
   },
   "source": [
    "!rm -rf SSL\n",
    "!git clone --branch simsiam https://github.com/spaceml-org/Self-Supervised-Learner.git\n",
    "!mv Self-Supervised-Learner SSL"
   ],
   "execution_count": 3,
   "outputs": [
    {
     "output_type": "stream",
     "text": [
      "Cloning into 'Self-Supervised-Learner'...\n",
      "remote: Enumerating objects: 10, done.\u001b[K\n",
      "remote: Counting objects: 100% (10/10), done.\u001b[K\n",
      "remote: Compressing objects: 100% (10/10), done.\u001b[K\n",
      "remote: Total 2537 (delta 2), reused 0 (delta 0), pack-reused 2527\u001b[K\n",
      "Receiving objects: 100% (2537/2537), 11.88 MiB | 34.94 MiB/s, done.\n",
      "Resolving deltas: 100% (1592/1592), done.\n"
     ],
     "name": "stdout"
    }
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "hPScN0CxQLGH"
   },
   "source": [
    "# Training Self-Supervised Learning Model"
   ]
  },
  {
   "cell_type": "code",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "-F8Lyk4eWPkF",
    "outputId": "546c60aa-c4c1-4a79-9ff9-5ae24125bdad"
   },
   "source": [
    "!python SSL/train.py --technique SIMCLR --DATA_PATH /content/UCMerced_LandUse/Images --model minicnn32 --log_name ssl --epochs 5 --batch_size 32 --image_size 256"
   ],
   "execution_count": 4,
   "outputs": [
    {
     "output_type": "stream",
     "text": [
      "\u001b[34mAutomatically splitting data into train and validation data...\u001b[0m\n",
      "Copying files: 2100 files [00:01, 2001.68 files/s]\n",
      "\u001b[34mModel architecture successfully loaded\u001b[0m\n",
      "GPU available: True, used: True\n",
      "TPU available: None, using: 0 TPU cores\n",
      "LOCAL_RANK: 0 - CUDA_VISIBLE_DEVICES: [0]\n",
      "read 1680 files from 21 directories\n",
      "read 420 files from 21 directories\n",
      "\n",
      "  | Name       | Type       | Params\n",
      "------------------------------------------\n",
      "0 | projection | Projection | 20.9 K\n",
      "1 | encoder    | miniCNN    | 128 K \n",
      "------------------------------------------\n",
      "149 K     Trainable params\n",
      "0         Non-trainable params\n",
      "149 K     Total params\n",
      "Validation sanity check:   0% 0/2 [00:00<?, ?it/s]/usr/local/lib/python3.7/dist-packages/pytorch_lightning/core/step_result.py:145: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).\n",
      "  value = torch.tensor(value, device=device, dtype=torch.float)\n",
      "Epoch 0:  92% 60/65 [00:02<00:00, 24.32it/s, loss=3.45, v_num=2oez]\n",
      "Validating: 0it [00:00, ?it/s]\u001b[A\n",
      "Validating:   0% 0/13 [00:00<?, ?it/s]\u001b[A\n",
      "Epoch 0: 100% 65/65 [00:02<00:00, 24.42it/s, loss=3.2, v_num=2oez] \n",
      "Epoch 1:  92% 60/65 [00:02<00:00, 24.51it/s, loss=2.51, v_num=2oez]\n",
      "Validating: 0it [00:00, ?it/s]\u001b[A\n",
      "Validating:   0% 0/13 [00:00<?, ?it/s]\u001b[A\n",
      "Epoch 1: 100% 65/65 [00:02<00:00, 24.65it/s, loss=2.34, v_num=2oez]\n",
      "Epoch 2:  92% 60/65 [00:02<00:00, 24.80it/s, loss=2.07, v_num=2oez]\n",
      "Validating: 0it [00:00, ?it/s]\u001b[A\n",
      "Validating:   0% 0/13 [00:00<?, ?it/s]\u001b[A\n",
      "Epoch 2: 100% 65/65 [00:02<00:00, 25.04it/s, loss=2.04, v_num=2oez]\n",
      "Epoch 3:  92% 60/65 [00:02<00:00, 24.78it/s, loss=1.85, v_num=2oez]\n",
      "Validating: 0it [00:00, ?it/s]\u001b[A\n",
      "Validating:   0% 0/13 [00:00<?, ?it/s]\u001b[A\n",
      "Epoch 3: 100% 65/65 [00:02<00:00, 24.84it/s, loss=1.82, v_num=2oez]\n",
      "Epoch 4:  92% 60/65 [00:02<00:00, 24.59it/s, loss=1.84, v_num=2oez]\n",
      "Validating: 0it [00:00, ?it/s]\u001b[A\n",
      "Validating:   0% 0/13 [00:00<?, ?it/s]\u001b[A\n",
      "Epoch 4: 100% 65/65 [00:02<00:00, 24.68it/s, loss=1.77, v_num=2oez]\n",
      "Epoch 4: 100% 65/65 [00:02<00:00, 24.67it/s, loss=1.77, v_num=2oez]\n",
      "\u001b[34mYOUR MODEL CAN BE ACCESSED AT: \u001b[0m ./models/SIMCLR_ssl.ckpt\n"
     ],
     "name": "stdout"
    }
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "shZCiS7okFEa"
   },
   "source": [
    "# Fine Tuning Model"
   ]
  },
  {
   "cell_type": "code",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "t-se1ogwWb7K",
    "outputId": "4df82c9a-40ab-4324-87db-1822d782df3b"
   },
   "source": [
    "!python SSL/train.py --technique CLASSIFIER --DATA_PATH /content/UCMerced_LandUse/Images --model ./models/SIMCLR_ssl.ckpt --log_name ft --epochs 20 --batch_size 32 --image_size 256"
   ],
   "execution_count": 5,
   "outputs": [
    {
     "output_type": "stream",
     "text": [
      "\u001b[34mAutomatically splitting data into train and validation data...\u001b[0m\n",
      "Copying files: 2100 files [00:01, 1961.93 files/s]\n",
      "Trying to return model encoder only...\n",
      "Namespace(DATA_PATH='./split_data_CLASSIFIER_ft/train', VAL_PATH='./split_data_CLASSIFIER_ft/val', batch_size=32, checkpoint_path='./models/SIMCLR_ssl.ckpt', cpus=1, decay_epochs=[60, 80], dropout=0.1, epochs=20, final_lr=1e-06, gamma=0.1, gpus=1, hidden_dim=128, image_size=256, learning_rate=0.001, linear_lr=0.1, log_name='ft', model='./models/SIMCLR_ssl.ckpt', momentum=0.9, nesterov=False, patience=-1, resize=False, scheduler_type='cosine', seed=1729, technique='SIMCLR', val_split=0.2, weight_decay=1e-06, withhold_split=0)\n",
      "\u001b[34mSuccessfully found previous model SIMCLR\u001b[0m\n",
      "\u001b[34mModel architecture successfully loaded\u001b[0m\n",
      "GPU available: True, used: True\n",
      "TPU available: None, using: 0 TPU cores\n",
      "LOCAL_RANK: 0 - CUDA_VISIBLE_DEVICES: [0]\n",
      "read 1680 files from 21 directories\n",
      "read 420 files from 21 directories\n",
      "\n",
      "  | Name         | Type         | Params\n",
      "----------------------------------------------\n",
      "0 | linear_layer | SSLEvaluator | 7.1 K \n",
      "1 | train_acc    | Accuracy     | 0     \n",
      "2 | val_acc      | Accuracy     | 0     \n",
      "3 | encoder      | miniCNN      | 128 K \n",
      "----------------------------------------------\n",
      "135 K     Trainable params\n",
      "0         Non-trainable params\n",
      "135 K     Total params\n",
      "Validation sanity check:   0% 0/2 [00:00<?, ?it/s]/usr/local/lib/python3.7/dist-packages/pytorch_lightning/core/step_result.py:145: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor).\n",
      "  value = torch.tensor(value, device=device, dtype=torch.float)\n",
      "Epoch 0:  92% 60/65 [00:01<00:00, 45.21it/s, loss=3.2, v_num=tY9b, val_loss=4.46, val_acc_epoch=0.0781, tloss=2.84, tastep=0.0938]\n",
      "Validating: 0it [00:00, ?it/s]\u001b[A\n",
      "Validating:   0% 0/13 [00:00<?, ?it/s]\u001b[A\n",
      "Epoch 0: 100% 65/65 [00:01<00:00, 43.85it/s, loss=3.1, v_num=tY9b, val_loss=3.16, val_acc_epoch=0.0841, tloss=3.2, tastep=0.0938] \n",
      "Epoch 1:  92% 60/65 [00:01<00:00, 45.15it/s, loss=2.91, v_num=tY9b, val_loss=3.16, val_acc_epoch=0.0841, tloss=2.57, tastep=0.188]\n",
      "Validating: 0it [00:00, ?it/s]\u001b[A\n",
      "Validating:   0% 0/13 [00:00<?, ?it/s]\u001b[A\n",
      "Epoch 1: 100% 65/65 [00:01<00:00, 43.65it/s, loss=2.85, v_num=tY9b, val_loss=2.74, val_acc_epoch=0.137, tloss=2.88, tastep=0.0938]\n",
      "Epoch 2:  92% 60/65 [00:01<00:00, 44.98it/s, loss=2.82, v_num=tY9b, val_loss=2.74, val_acc_epoch=0.137, tloss=2.84, tastep=0.125]\n",
      "Validating: 0it [00:00, ?it/s]\u001b[A\n",
      "Validating:   0% 0/13 [00:00<?, ?it/s]\u001b[A\n",
      "Epoch 2: 100% 65/65 [00:01<00:00, 43.84it/s, loss=2.89, v_num=tY9b, val_loss=2.62, val_acc_epoch=0.166, tloss=2.86, tastep=0.188]\n",
      "Epoch 3:  92% 60/65 [00:01<00:00, 44.84it/s, loss=2.8, v_num=tY9b, val_loss=2.62, val_acc_epoch=0.166, tloss=2.72, tastep=0.188]\n",
      "Validating: 0it [00:00, ?it/s]\u001b[A\n",
      "Validating:   0% 0/13 [00:00<?, ?it/s]\u001b[A\n",
      "Epoch 3: 100% 65/65 [00:01<00:00, 43.92it/s, loss=2.8, v_num=tY9b, val_loss=2.62, val_acc_epoch=0.161, tloss=3.08, tastep=0.0312]\n",
      "Epoch 4:  92% 60/65 [00:01<00:00, 45.23it/s, loss=2.7, v_num=tY9b, val_loss=2.62, val_acc_epoch=0.161, tloss=2.8, tastep=0.0938]\n",
      "Validating: 0it [00:00, ?it/s]\u001b[A\n",
      "Validating:   0% 0/13 [00:00<?, ?it/s]\u001b[A\n",
      "Epoch 4: 100% 65/65 [00:01<00:00, 44.16it/s, loss=2.72, v_num=tY9b, val_loss=3.01, val_acc_epoch=0.12, tloss=2.65, tastep=0.125]\n",
      "Epoch 5:  92% 60/65 [00:01<00:00, 45.14it/s, loss=2.73, v_num=tY9b, val_loss=3.01, val_acc_epoch=0.12, tloss=2.63, tastep=0.0625]\n",
      "Validating: 0it [00:00, ?it/s]\u001b[A\n",
      "Validating:   0% 0/13 [00:00<?, ?it/s]\u001b[A\n",
      "Epoch 5: 100% 65/65 [00:01<00:00, 43.98it/s, loss=2.66, v_num=tY9b, val_loss=3.54, val_acc_epoch=0.0865, tloss=2.5, tastep=0.188]\n",
      "Epoch 6:  92% 60/65 [00:01<00:00, 45.22it/s, loss=2.68, v_num=tY9b, val_loss=3.54, val_acc_epoch=0.0865, tloss=2.75, tastep=0.0938]\n",
      "Validating: 0it [00:00, ?it/s]\u001b[A\n",
      "Validating:   0% 0/13 [00:00<?, ?it/s]\u001b[A\n",
      "Epoch 6: 100% 65/65 [00:01<00:00, 44.40it/s, loss=2.66, v_num=tY9b, val_loss=2.67, val_acc_epoch=0.18, tloss=2.43, tastep=0.312]   \n",
      "Epoch 7:  92% 60/65 [00:01<00:00, 45.35it/s, loss=2.65, v_num=tY9b, val_loss=2.67, val_acc_epoch=0.18, tloss=2.75, tastep=0.0938]\n",
      "Validating: 0it [00:00, ?it/s]\u001b[A\n",
      "Validating:   0% 0/13 [00:00<?, ?it/s]\u001b[A\n",
      "Epoch 7: 100% 65/65 [00:01<00:00, 44.37it/s, loss=2.62, v_num=tY9b, val_loss=2.7, val_acc_epoch=0.168, tloss=2.42, tastep=0.188] \n",
      "Epoch 8:  92% 60/65 [00:01<00:00, 45.79it/s, loss=2.59, v_num=tY9b, val_loss=2.7, val_acc_epoch=0.168, tloss=2.77, tastep=0.125]\n",
      "Validating: 0it [00:00, ?it/s]\u001b[A\n",
      "Validating:   0% 0/13 [00:00<?, ?it/s]\u001b[A\n",
      "Epoch 8: 100% 65/65 [00:01<00:00, 44.81it/s, loss=2.57, v_num=tY9b, val_loss=2.76, val_acc_epoch=0.147, tloss=2.45, tastep=0.281]\n",
      "Epoch 9:  92% 60/65 [00:01<00:00, 44.45it/s, loss=2.59, v_num=tY9b, val_loss=2.76, val_acc_epoch=0.147, tloss=2.52, tastep=0.156]\n",
      "Validating: 0it [00:00, ?it/s]\u001b[A\n",
      "Validating:   0% 0/13 [00:00<?, ?it/s]\u001b[A\n",
      "Epoch 9: 100% 65/65 [00:01<00:00, 43.08it/s, loss=2.57, v_num=tY9b, val_loss=2.45, val_acc_epoch=0.2, tloss=2.44, tastep=0.156]  \n",
      "Epoch 10:  92% 60/65 [00:01<00:00, 45.22it/s, loss=2.51, v_num=tY9b, val_loss=2.45, val_acc_epoch=0.2, tloss=2.58, tastep=0.188]\n",
      "Validating: 0it [00:00, ?it/s]\u001b[A\n",
      "Validating:   0% 0/13 [00:00<?, ?it/s]\u001b[A\n",
      "Epoch 10: 100% 65/65 [00:01<00:00, 44.13it/s, loss=2.45, v_num=tY9b, val_loss=2.22, val_acc_epoch=0.255, tloss=2.56, tastep=0.156]\n",
      "Epoch 11:  92% 60/65 [00:01<00:00, 45.76it/s, loss=2.41, v_num=tY9b, val_loss=2.22, val_acc_epoch=0.255, tloss=2.27, tastep=0.281]\n",
      "Validating: 0it [00:00, ?it/s]\u001b[A\n",
      "Validating:   0% 0/13 [00:00<?, ?it/s]\u001b[A\n",
      "Epoch 11: 100% 65/65 [00:01<00:00, 44.64it/s, loss=2.44, v_num=tY9b, val_loss=2.34, val_acc_epoch=0.252, tloss=2.38, tastep=0.156]\n",
      "Epoch 12:  92% 60/65 [00:01<00:00, 44.55it/s, loss=2.44, v_num=tY9b, val_loss=2.34, val_acc_epoch=0.252, tloss=2.2, tastep=0.344]\n",
      "Validating: 0it [00:00, ?it/s]\u001b[A\n",
      "Validating:   0% 0/13 [00:00<?, ?it/s]\u001b[A\n",
      "Epoch 12: 100% 65/65 [00:01<00:00, 43.78it/s, loss=2.36, v_num=tY9b, val_loss=2.23, val_acc_epoch=0.276, tloss=2.26, tastep=0.312]\n",
      "Epoch 13:  92% 60/65 [00:01<00:00, 44.83it/s, loss=2.38, v_num=tY9b, val_loss=2.23, val_acc_epoch=0.276, tloss=2.51, tastep=0.219]\n",
      "Validating: 0it [00:00, ?it/s]\u001b[A\n",
      "Validating:   0% 0/13 [00:00<?, ?it/s]\u001b[A\n",
      "Epoch 13: 100% 65/65 [00:01<00:00, 43.52it/s, loss=2.39, v_num=tY9b, val_loss=2.18, val_acc_epoch=0.312, tloss=2.31, tastep=0.344]\n",
      "Epoch 14:  92% 60/65 [00:01<00:00, 45.48it/s, loss=2.38, v_num=tY9b, val_loss=2.18, val_acc_epoch=0.312, tloss=2.65, tastep=0.125]\n",
      "Validating: 0it [00:00, ?it/s]\u001b[A\n",
      "Validating:   0% 0/13 [00:00<?, ?it/s]\u001b[A\n",
      "Epoch 14: 100% 65/65 [00:01<00:00, 44.38it/s, loss=2.35, v_num=tY9b, val_loss=2.26, val_acc_epoch=0.276, tloss=2.29, tastep=0.25] \n",
      "Epoch 15:  92% 60/65 [00:01<00:00, 44.46it/s, loss=2.34, v_num=tY9b, val_loss=2.26, val_acc_epoch=0.276, tloss=2.46, tastep=0.219]\n",
      "Validating: 0it [00:00, ?it/s]\u001b[A\n",
      "Validating:   0% 0/13 [00:00<?, ?it/s]\u001b[A\n",
      "Epoch 15: 100% 65/65 [00:01<00:00, 43.37it/s, loss=2.44, v_num=tY9b, val_loss=2.22, val_acc_epoch=0.337, tloss=2.72, tastep=0.312]\n",
      "Epoch 16:  92% 60/65 [00:01<00:00, 44.31it/s, loss=2.3, v_num=tY9b, val_loss=2.22, val_acc_epoch=0.337, tloss=2.3, tastep=0.25]\n",
      "Validating: 0it [00:00, ?it/s]\u001b[A\n",
      "Validating:   0% 0/13 [00:00<?, ?it/s]\u001b[A\n",
      "Epoch 16: 100% 65/65 [00:01<00:00, 43.59it/s, loss=2.34, v_num=tY9b, val_loss=2.45, val_acc_epoch=0.24, tloss=2.17, tastep=0.25]\n",
      "Epoch 17:  92% 60/65 [00:01<00:00, 45.76it/s, loss=2.33, v_num=tY9b, val_loss=2.45, val_acc_epoch=0.24, tloss=2.22, tastep=0.188]\n",
      "Validating: 0it [00:00, ?it/s]\u001b[A\n",
      "Validating:   0% 0/13 [00:00<?, ?it/s]\u001b[A\n",
      "Epoch 17: 100% 65/65 [00:01<00:00, 44.85it/s, loss=2.25, v_num=tY9b, val_loss=2.19, val_acc_epoch=0.293, tloss=2.04, tastep=0.438]\n",
      "Epoch 18:  92% 60/65 [00:01<00:00, 43.67it/s, loss=2.29, v_num=tY9b, val_loss=2.19, val_acc_epoch=0.293, tloss=1.92, tastep=0.5]\n",
      "Validating: 0it [00:00, ?it/s]\u001b[A\n",
      "Validating:   0% 0/13 [00:00<?, ?it/s]\u001b[A\n",
      "Epoch 18: 100% 65/65 [00:01<00:00, 42.54it/s, loss=2.27, v_num=tY9b, val_loss=2.14, val_acc_epoch=0.332, tloss=2.21, tastep=0.375]\n",
      "Epoch 19:  92% 60/65 [00:01<00:00, 45.45it/s, loss=2.26, v_num=tY9b, val_loss=2.14, val_acc_epoch=0.332, tloss=1.98, tastep=0.438]\n",
      "Validating: 0it [00:00, ?it/s]\u001b[A\n",
      "Validating:   0% 0/13 [00:00<?, ?it/s]\u001b[A\n",
      "Epoch 19: 100% 65/65 [00:01<00:00, 44.02it/s, loss=2.23, v_num=tY9b, val_loss=2.08, val_acc_epoch=0.358, tloss=2.15, tastep=0.219]\n",
      "Epoch 19: 100% 65/65 [00:01<00:00, 43.99it/s, loss=2.23, v_num=tY9b, val_loss=2.08, val_acc_epoch=0.358, tloss=2.15, tastep=0.219]\n",
      "\u001b[34mYOUR MODEL CAN BE ACCESSED AT: \u001b[0m ./models/CLASSIFIER_ft.ckpt\n"
     ],
     "name": "stdout"
    }
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "9pS1_ifkOXlj"
   },
   "source": [
    "# Loading Trained Models in Python"
   ]
  },
  {
   "cell_type": "code",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "ctJHNKnMUY7w",
    "outputId": "61d6a360-d069-4bb8-829e-b9fd5db75b41"
   },
   "source": [
    "# load model for custom inference script below\n",
    "%cd /content/SSL\n",
    "from models import SIMCLR, CLASSIFIER\n",
    "\n",
    "%cd /content/"
   ],
   "execution_count": 6,
   "outputs": [
    {
     "output_type": "stream",
     "text": [
      "/content/SSL\n",
      "/content\n"
     ],
     "name": "stdout"
    }
   ]
  },
  {
   "cell_type": "code",
   "metadata": {
    "id": "xs8g_Vt2PRBj",
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "outputId": "878dc7b9-ae11-4378-b341-0e848dd98c9a"
   },
   "source": [
    "simclr_model = SIMCLR.SIMCLR.load_from_checkpoint(\"/content/models/SIMCLR_ssl.ckpt\")\n",
    "classifier_model = CLASSIFIER.CLASSIFIER.load_from_checkpoint(\n",
    "    \"/content/models/CLASSIFIER_ft.ckpt\"\n",
    ")\n",
    "\n",
    "simclr_model.eval()\n",
    "classifier_model.eval()\n",
    "\n",
    "simclr_model.cuda()\n",
    "classifier_model.cuda()"
   ],
   "execution_count": 7,
   "outputs": [
    {
     "output_type": "execute_result",
     "data": {
      "text/plain": [
       "CLASSIFIER(\n",
       "  (linear_layer): SSLEvaluator(\n",
       "    (block_forward): Sequential(\n",
       "      (0): Flatten()\n",
       "      (1): Dropout(p=0.1, inplace=False)\n",
       "      (2): Linear(in_features=32, out_features=128, bias=False)\n",
       "      (3): BatchNorm1d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
       "      (4): ReLU(inplace=True)\n",
       "      (5): Dropout(p=0.1, inplace=False)\n",
       "      (6): Linear(in_features=128, out_features=21, bias=True)\n",
       "    )\n",
       "  )\n",
       "  (train_acc): Accuracy()\n",
       "  (val_acc): Accuracy()\n",
       "  (encoder): miniCNN(\n",
       "    (conv1): Conv2d(3, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "    (conv2): Conv2d(16, 32, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))\n",
       "    (conv3): Conv2d(32, 48, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))\n",
       "    (adaptive_pool): AdaptiveAvgPool2d(output_size=(16, 16))\n",
       "    (conv4): Conv2d(48, 64, kernel_size=(5, 5), stride=(2, 2), padding=(2, 2))\n",
       "    (pool): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
       "    (flatten): Flatten(start_dim=1, end_dim=-1)\n",
       "    (fc1): Linear(in_features=1024, out_features=32, bias=True)\n",
       "  )\n",
       ")"
      ]
     },
     "metadata": {
      "tags": []
     },
     "execution_count": 7
    }
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "BKp9nZZcQkBG"
   },
   "source": [
    "To run inference, we need to transform an image from a 244x244x3 image (the shape of UC Merced Data) to a 3x256x256 image, which is what we trained our model on."
   ]
  },
  {
   "cell_type": "code",
   "metadata": {
    "id": "JPT8hP5qQPIN",
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 286
    },
    "outputId": "3eacce84-b1fc-4c96-9be4-f21df50cef74"
   },
   "source": [
    "# How to run inference manually\n",
    "\n",
    "import numpy as np\n",
    "from PIL import Image\n",
    "import torch\n",
    "import matplotlib.pyplot as plt\n",
    "from torchvision import transforms\n",
    "\n",
    "f = \"/content/UCMerced_LandUse/Images/agricultural/agricultural00.tif\"  # insert image here\n",
    "\n",
    "im = Image.open(f).convert(\"RGB\")\n",
    "plt.imshow(im)\n",
    "\n",
    "\n",
    "def to_tensor(pil):\n",
    "    return torch.tensor(np.array(pil)).permute(2, 0, 1).float()\n",
    "\n",
    "\n",
    "t = transforms.Compose([transforms.Resize((256, 256)), transforms.Lambda(to_tensor)])\n",
    "\n",
    "datapoint = (\n",
    "    t(im).unsqueeze(0).cuda()\n",
    ")  # only a single datapoint so we unsqueeze to add a dimension\n",
    "\n",
    "with torch.no_grad():\n",
    "    embedding = simclr_model(datapoint)  # get embedding\n",
    "    category = classifier_model(datapoint)  # get classification\n",
    "    print(\"Category is: \", torch.argmax(category, 1))"
   ],
   "execution_count": 8,
   "outputs": [
    {
     "output_type": "stream",
     "text": [
      "Category is:  tensor([0], device='cuda:0')\n"
     ],
     "name": "stdout"
    },
    {
     "output_type": "display_data",
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": [],
      "needs_background": "light"
     }
    }
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "ueQuaKEDiTHS"
   },
   "source": [
    "Or below, if you don't want to write the transforms yourself, we will use the built in DALI pipeline to do inference quickly on a large number of datapoints"
   ]
  },
  {
   "cell_type": "code",
   "metadata": {
    "id": "wUgtfzJIRfwY"
   },
   "source": [
    "# We'll pretend the agriculture images are a new dataset we want to run inference on after training on UC Merced Data previously\n",
    "!rm -rf /content/new_images_to_run_inference\n",
    "!mkdir /content/new_images_to_run_inference\n",
    "!cp -r /content/UCMerced_LandUse/Images/agricultural /content/new_images_to_run_inference"
   ],
   "execution_count": 9,
   "outputs": []
  },
  {
   "cell_type": "code",
   "metadata": {
    "id": "8NaqR5kVVTKK"
   },
   "source": [
    "model = CLASSIFIER.CLASSIFIER.load_from_checkpoint(\n",
    "    \"/content/models/CLASSIFIER_ft.ckpt\",\n",
    "    DATA_PATH=\"/content/new_images_to_run_inference\",\n",
    "    batch_size=196,\n",
    ")  # specify new datapath"
   ],
   "execution_count": 10,
   "outputs": []
  },
  {
   "cell_type": "code",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "QSGf4uQfVcfp",
    "outputId": "c26d715d-ee0b-4d85-bc55-c6c47cfc7fb3"
   },
   "source": [
    "model.local_rank = 0\n",
    "model.setup()  # we set up inference with this call to instantiate the DALI data pipeline\n",
    "model.eval()\n",
    "model.cuda()\n",
    "\n",
    "for batch in model.inference_dataloader:\n",
    "    predictions = model(batch)\n",
    "    print(predictions.shape)  # dimensions of predictions are batch_size * num_classes\n",
    "    break"
   ],
   "execution_count": 11,
   "outputs": [
    {
     "output_type": "stream",
     "text": [
      "torch.Size([196, 21])\n"
     ],
     "name": "stdout"
    }
   ]
  },
  {
   "cell_type": "code",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "qZpXWkN-tzMB",
    "outputId": "836d551a-05f4-4460-d8e5-9ec8d7d80425"
   },
   "source": [
    "# This inference pipeline scales quite nicely to large datasets\n",
    "num_datapoints = 100000\n",
    "batch_size = 196\n",
    "\n",
    "start = time.time()\n",
    "\n",
    "for i in range(num_datapoints // batch_size):\n",
    "    data = next(model.inference_dataloader)\n",
    "    batch_predictions = model(data)\n",
    "print(\"Time Taken: \", time.time() - start)"
   ],
   "execution_count": 12,
   "outputs": [
    {
     "output_type": "stream",
     "text": [
      "Time Taken:  30.133768796920776\n"
     ],
     "name": "stdout"
    }
   ]
  },
  {
   "cell_type": "code",
   "metadata": {
    "id": "h_gmwygvVOVn"
   },
   "source": [
    ""
   ],
   "execution_count": 12,
   "outputs": []
  }
 ]
}