NSAI on Sort-of-CLEVR.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Neuro-Symbolic AI\n",
    "## Sort-of-CLEVR Dataset\n",
    "### Contents\n",
    "- Generate Dataset\n",
    "- Perception Module\n",
    "    - Train\n",
    "    - Inference\n",
    "- Semantic Parser\n",
    "    - Preprocess Data\n",
    "    - Train Seq2Seq Model\n",
    "    - Inference\n",
    "- Program Executor\n",
    "- Plugging Everything Together"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Install Pytorch 1.7 and torchtext 0.8\n",
    "# !pip install torch==1.7.0+cu101 torchvision==0.8.1+cu101 torchtext==0.8.0 -f https://download.pytorch.org/whl/torch_stable.html"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# !git clone https://github.com/nerdimite/neuro-symbolic-ai-soc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# cd neuro-symbolic-ai-soc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "from data_generator import *\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import os\n",
    "import dlib\n",
    "import cv2\n",
    "from skimage.io import imshow\n",
    "from sklearn.metrics.pairwise import cosine_similarity\n",
    "import torch\n",
    "from torch import nn\n",
    "from torchvision.datasets import ImageFolder\n",
    "from torch.utils.data import DataLoader\n",
    "from torchvision import transforms as tfms\n",
    "from PIL import Image"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Dataset Generation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Breakdown"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Generate a sample i.e.\n",
    "# a tuple of 1 image, 6 objects' properties, 20 QA vectors (10 relational and 10 non-relational)\n",
    "sample = build_sample()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Break down the sample\n",
    "img, objects, queries, programs, answers = convert_sample(sample)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "img = (img * 255).astype('uint8')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x77c9e00aae60>"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "imshow(cv2.cvtColor(img, cv2.COLOR_BGR2RGB))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[(0, array([ 24, 140]), 'c', (9, 125, 39, 155)),\n",
       " (1, array([30, 79]), 'r', (15, 64, 45, 94)),\n",
       " (2, array([128, 138]), 'c', (113, 123, 143, 153)),\n",
       " (3, array([171, 172]), 'r', (156, 157, 186, 187)),\n",
       " (4, array([128,  23]), 'c', (113, 8, 143, 38)),\n",
       " (5, array([88, 15]), 'c', (73, 0, 103, 30))]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "objects"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('What is the shape of the closest object to the orange object?',\n",
       "  'circle',\n",
       "  'filter orange <nxt> relate closest <nxt> query shape'),\n",
       " ('What is the shape of the farthest object to the yellow object?',\n",
       "  'rectangle',\n",
       "  'filter yellow <nxt> relate farthest <nxt> query shape'),\n",
       " ('What is the shape of the farthest object to the yellow object?',\n",
       "  'rectangle',\n",
       "  'filter yellow <nxt> relate farthest <nxt> query shape'),\n",
       " ('What is the shape of the farthest object to the green object?',\n",
       "  'rectangle',\n",
       "  'filter green <nxt> relate farthest <nxt> query shape'),\n",
       " ('What is the shape of the farthest object to the green object?',\n",
       "  'rectangle',\n",
       "  'filter green <nxt> relate farthest <nxt> query shape'),\n",
       " ('What is the shape of the farthest object to the gray object?',\n",
       "  'circle',\n",
       "  'filter gray <nxt> relate farthest <nxt> query shape'),\n",
       " ('How many objects of the shape as that of the red are there?',\n",
       "  '4',\n",
       "  'filter red <nxt> query shape <nxt> filter <nxt> count'),\n",
       " ('What is the shape of the closest object to the red object?',\n",
       "  'rectangle',\n",
       "  'filter red <nxt> relate closest <nxt> query shape'),\n",
       " ('What is the shape of the closest object to the green object?',\n",
       "  'circle',\n",
       "  'filter green <nxt> relate closest <nxt> query shape'),\n",
       " ('How many objects of the shape as that of the blue are there?',\n",
       "  '4',\n",
       "  'filter blue <nxt> query shape <nxt> filter <nxt> count'),\n",
       " ('What is the shape of the yellow of the object?',\n",
       "  'circle',\n",
       "  'filter yellow <nxt> query shape'),\n",
       " ('Is there a red object on the top?',\n",
       "  'no',\n",
       "  'filter red <nxt> query position <nxt> isTop'),\n",
       " ('Is there a orange object on the left?',\n",
       "  'no',\n",
       "  'filter orange <nxt> query positions <nxt> isLeft'),\n",
       " ('What is the shape of the red of the object?',\n",
       "  'circle',\n",
       "  'filter red <nxt> query shape'),\n",
       " ('Is there a yellow object on the left?',\n",
       "  'yes',\n",
       "  'filter yellow <nxt> query positions <nxt> isLeft'),\n",
       " ('Is there a red object on the left?',\n",
       "  'yes',\n",
       "  'filter red <nxt> query positions <nxt> isLeft'),\n",
       " ('Is there a green object on the top?',\n",
       "  'yes',\n",
       "  'filter green <nxt> query position <nxt> isTop'),\n",
       " ('Is there a blue object on the left?',\n",
       "  'no',\n",
       "  'filter blue <nxt> query positions <nxt> isLeft'),\n",
       " ('Is there a orange object on the left?',\n",
       "  'no',\n",
       "  'filter orange <nxt> query positions <nxt> isLeft'),\n",
       " ('Is there a yellow object on the left?',\n",
       "  'yes',\n",
       "  'filter yellow <nxt> query positions <nxt> isLeft')]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "list(zip(queries, answers, programs))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Generate the full dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 100/100 [00:03<00:00, 26.08it/s]\n",
      "100%|██████████| 10/10 [00:00<00:00, 33.20it/s]\n"
     ]
    }
   ],
   "source": [
    "# Generate the dataset\n",
    "n_train_imgs = 100\n",
    "n_test_imgs = 10\n",
    "\n",
    "build_dataset(n_train_imgs, data_dir='data/train', prefix='train')\n",
    "build_dataset(n_test_imgs, data_dir='data/test', prefix='test')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Perception Module"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Train the Object Detector"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "from perception import *"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "def load_image_data(csv_file, img_dir):\n",
    "    '''Load the image detection data for training the object detector'''\n",
    "    # Annotations\n",
    "    annot = pd.read_csv(csv_file)\n",
    "    \n",
    "    data = {}\n",
    "    for i in range(len(os.listdir(img_dir))):\n",
    "        # Get all rows belonging to the same scene\n",
    "        scene = annot[annot['filename'] == f'{i}.jpg']\n",
    "        # Read the image\n",
    "        img = cv2.imread(os.path.join(img_dir, f'{i}.jpg'))\n",
    "        \n",
    "        # Add bounding box information for dlib\n",
    "        bboxes = []\n",
    "        for row in scene.values:\n",
    "            x1, y1, x2, y2 = row[-4:]\n",
    "            dlib_box = dlib.rectangle(left=x1 , top=y1, right=x2, bottom=y2)\n",
    "            bboxes.append(dlib_box)\n",
    "        \n",
    "        data[i] = (img, bboxes)\n",
    "    \n",
    "    return data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load the Image Detection Data\n",
    "det_data = load_image_data('data/train/train_img_det.csv', 'data/train/images')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Results: precision: 1, recall: 1, average precision: 1\n",
      "Saved the model to models/detector.svm\n",
      "CPU times: user 22.8 s, sys: 208 ms, total: 23 s\n",
      "Wall time: 8.25 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "# Train the Model\n",
    "train_detector(det_data, 'models/detector.svm')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Shape Classifier"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "class Binarize():\n",
    "    '''PyTorch Transforms Object'''\n",
    "    def __init__(self):\n",
    "        '''Converts Grayscale to Binary (except white every other color is zeroed)'''\n",
    "        pass\n",
    "    \n",
    "    def __call__(self, img_tensor):\n",
    "        '''\n",
    "        Args:\n",
    "            img_tensor (tensor): 0-1 scaled tensor with 1 channel\n",
    "        Returns:\n",
    "            tensor\n",
    "        '''\n",
    "        return (img_tensor > 0.95).float()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load the images\n",
    "shapes_data = ImageFolder('data/shapes_data/', transform=tfms.Compose([tfms.Grayscale(), \n",
    "                                                                       tfms.Resize((40, 40)), \n",
    "                                                                       tfms.ToTensor(),\n",
    "                                                                       Binarize()]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'circle': 0, 'rectangle': 1}"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "shapes_data.class_to_idx"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create the dataloader\n",
    "shapes_loader = DataLoader(shapes_data, batch_size=20, shuffle=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Model\n",
    "shape_classifier = ShapeClassifier()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 1:   0%|          | 0/6 [00:00<?, ?it/s]/home/m0nku/.local/lib/python3.10/site-packages/torch/autograd/graph.py:744: UserWarning: Plan failed with a cudnnException: CUDNN_BACKEND_EXECUTION_PLAN_DESCRIPTOR: cudnnFinalize Descriptor Failed cudnn_status: CUDNN_STATUS_NOT_SUPPORTED (Triggered internally at ../aten/src/ATen/native/cudnn/Conv_v8.cpp:919.)\n",
      "  return Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass\n",
      "Epoch 1: 100%|██████████| 6/6 [00:00<00:00, 18.73it/s, Loss=0.478]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1: Mean Loss = 0.535299281279246\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 2: 100%|██████████| 6/6 [00:00<00:00, 47.07it/s, Loss=0.362]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 2: Mean Loss = 0.38747653365135193\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 3: 100%|██████████| 6/6 [00:00<00:00, 83.44it/s, Loss=0.199]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 3: Mean Loss = 0.27511777232090634\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 4: 100%|██████████| 6/6 [00:00<00:00, 91.95it/s, Loss=0.131]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 4: Mean Loss = 0.18886899451414743\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 5: 100%|██████████| 6/6 [00:00<00:00, 88.79it/s, Loss=0.203]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 5: Mean Loss = 0.1536424309015274\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 6: 100%|██████████| 6/6 [00:00<00:00, 68.96it/s, Loss=0.0775]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 6: Mean Loss = 0.10653669014573097\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 7: 100%|██████████| 6/6 [00:00<00:00, 65.26it/s, Loss=0.0554]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 7: Mean Loss = 0.08505254052579403\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 8: 100%|██████████| 6/6 [00:00<00:00, 91.18it/s, Loss=0.101] \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 8: Mean Loss = 0.07060479124387105\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 9: 100%|██████████| 6/6 [00:00<00:00, 89.09it/s, Loss=0.0299]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 9: Mean Loss = 0.05089210905134678\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 10: 100%|██████████| 6/6 [00:00<00:00, 91.48it/s, Loss=0.0309]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 10: Mean Loss = 0.045339081436395645\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 11: 100%|██████████| 6/6 [00:00<00:00, 94.86it/s, Loss=0.0313]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 11: Mean Loss = 0.03856289821366469\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 12: 100%|██████████| 6/6 [00:00<00:00, 89.57it/s, Loss=0.118] \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 12: Mean Loss = 0.046273721382021904\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 13: 100%|██████████| 6/6 [00:00<00:00, 88.14it/s, Loss=0.0465]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 13: Mean Loss = 0.03418318678935369\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 14: 100%|██████████| 6/6 [00:00<00:00, 77.19it/s, Loss=0.0411]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 14: Mean Loss = 0.027523216015348833\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 15: 100%|██████████| 6/6 [00:00<00:00, 76.18it/s, Loss=0.075] \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 15: Mean Loss = 0.028648322137693565\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 16: 100%|██████████| 6/6 [00:00<00:00, 87.05it/s, Loss=0.0115]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 16: Mean Loss = 0.017349074439456064\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 17: 100%|██████████| 6/6 [00:00<00:00, 85.15it/s, Loss=0.0101]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 17: Mean Loss = 0.016121252439916134\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 18: 100%|██████████| 6/6 [00:00<00:00, 85.94it/s, Loss=0.0163]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 18: Mean Loss = 0.014613977478196224\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 19: 100%|██████████| 6/6 [00:00<00:00, 87.69it/s, Loss=0.0237] \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 19: Mean Loss = 0.014673334856828054\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 20: 100%|██████████| 6/6 [00:00<00:00, 87.81it/s, Loss=0.0422] \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 20: Mean Loss = 0.01725327953075369\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 21: 100%|██████████| 6/6 [00:00<00:00, 79.73it/s, Loss=0.0236] \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 21: Mean Loss = 0.013442980435987314\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 22: 100%|██████████| 6/6 [00:00<00:00, 93.89it/s, Loss=0.00995]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 22: Mean Loss = 0.009661665884777904\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 23: 100%|██████████| 6/6 [00:00<00:00, 96.19it/s, Loss=0.0114] \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 23: Mean Loss = 0.009871084398279587\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 24: 100%|██████████| 6/6 [00:00<00:00, 88.05it/s, Loss=0.00999]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 24: Mean Loss = 0.007236357933531205\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 25: 100%|██████████| 6/6 [00:00<00:00, 90.52it/s, Loss=0.00361]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 25: Mean Loss = 0.005719191317136089\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 26: 100%|██████████| 6/6 [00:00<00:00, 96.57it/s, Loss=0.00527]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 26: Mean Loss = 0.0055531235023712116\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 27: 100%|██████████| 6/6 [00:00<00:00, 92.90it/s, Loss=0.00173] \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 27: Mean Loss = 0.004187676822766662\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 28: 100%|██████████| 6/6 [00:00<00:00, 81.50it/s, Loss=0.0028]  \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 28: Mean Loss = 0.004547716778082152\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 29: 100%|██████████| 6/6 [00:00<00:00, 74.10it/s, Loss=0.00126]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 29: Mean Loss = 0.003732759039849043\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 30: 100%|██████████| 6/6 [00:00<00:00, 93.16it/s, Loss=0.0049] \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 30: Mean Loss = 0.003866467314461867\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 31: 100%|██████████| 6/6 [00:00<00:00, 94.95it/s, Loss=0.00108]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 31: Mean Loss = 0.00411093842315798\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 32: 100%|██████████| 6/6 [00:00<00:00, 95.40it/s, Loss=0.0494] \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 32: Mean Loss = 0.010862537058225522\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 33: 100%|██████████| 6/6 [00:00<00:00, 49.42it/s, Loss=0.00438]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 33: Mean Loss = 0.0038346213210994997\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 34: 100%|██████████| 6/6 [00:00<00:00, 77.28it/s, Loss=0.00146] \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 34: Mean Loss = 0.0036049613748521856\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 35: 100%|██████████| 6/6 [00:00<00:00, 78.98it/s, Loss=0.00781] \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 35: Mean Loss = 0.003437642522233849\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 36: 100%|██████████| 6/6 [00:00<00:00, 91.04it/s, Loss=0.000605]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 36: Mean Loss = 0.0038979382855662457\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 37: 100%|██████████| 6/6 [00:00<00:00, 71.55it/s, Loss=0.0274]  \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 37: Mean Loss = 0.006558644769635673\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 38: 100%|██████████| 6/6 [00:00<00:00, 75.75it/s, Loss=0.000693]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 38: Mean Loss = 0.002215169632108882\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 39: 100%|██████████| 6/6 [00:00<00:00, 85.62it/s, Loss=0.000671]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 39: Mean Loss = 0.002866969366247455\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 40: 100%|██████████| 6/6 [00:00<00:00, 91.42it/s, Loss=0.000282]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 40: Mean Loss = 0.0018007522351884593\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 41: 100%|██████████| 6/6 [00:00<00:00, 94.09it/s, Loss=0.00044] \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 41: Mean Loss = 0.0017834966032144923\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 42: 100%|██████████| 6/6 [00:00<00:00, 93.44it/s, Loss=0.000501]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 42: Mean Loss = 0.001346442445840997\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "Epoch 43: 100%|██████████| 6/6 [00:00<00:00, 82.36it/s, Loss=0.00437] \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 43: Mean Loss = 0.0021167072360791885\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 44: 100%|██████████| 6/6 [00:00<00:00, 90.15it/s, Loss=0.000147]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 44: Mean Loss = 0.0010680181488472347\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 45: 100%|██████████| 6/6 [00:00<00:00, 78.15it/s, Loss=0.000754]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 45: Mean Loss = 0.0011836101233105485\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 46: 100%|██████████| 6/6 [00:00<00:00, 76.27it/s, Loss=0.000451]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 46: Mean Loss = 0.0010031750716734678\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 47: 100%|██████████| 6/6 [00:00<00:00, 73.88it/s, Loss=0.000201]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 47: Mean Loss = 0.0007156923578198379\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 48: 100%|██████████| 6/6 [00:00<00:00, 77.85it/s, Loss=0.0029]  \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 48: Mean Loss = 0.000852493164226568\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 49: 100%|██████████| 6/6 [00:00<00:00, 80.20it/s, Loss=0.00133] \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 49: Mean Loss = 0.0007244507226763138\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 50: 100%|██████████| 6/6 [00:00<00:00, 93.32it/s, Loss=0.000351]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 50: Mean Loss = 0.0006824038161236482\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "# Train\n",
    "shape_classifier.train_classifier(shapes_loader, lr=0.0001, epochs=50, filename='models/classifier.pth')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Perception Pipeline\n",
    "`detector -> attribute (shape and color) extraction -> structural scene representation`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load the Model\n",
    "perceiver = PerceptionPipe('models/detector.svm', 'models/classifier.pth', device='cuda')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Breakdown"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[(array([[[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         ...,\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [  0, 156, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]]], dtype=uint8),\n",
       "  (171, 172)),\n",
       " (array([[[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         ...,\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]]], dtype=uint8),\n",
       "  (23, 140)),\n",
       " (array([[[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         ...,\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]]], dtype=uint8),\n",
       "  (127, 20)),\n",
       " (array([[[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         ...,\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]]], dtype=uint8),\n",
       "  (127, 136)),\n",
       " (array([[[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [  0, 255,   0],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         ...,\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]]], dtype=uint8),\n",
       "  (31, 80)),\n",
       " (array([[[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         ...,\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]],\n",
       "  \n",
       "         [[255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          ...,\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255],\n",
       "          [255, 255, 255]]], dtype=uint8),\n",
       "  (87, 17))]"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Inference with Detector\n",
    "objects = perceiver.detect(img)\n",
    "objects"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "# View the isolated objects (this will not work in colab as its headless)\n",
    "# for obj, _ in objects:\n",
    "#     cv2.imshow('frame', obj)\n",
    "#     cv2.waitKey(0)\n",
    "#     cv2.destroyAllWindows()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape = circle\n",
      "Color = blue\n"
     ]
    }
   ],
   "source": [
    "# Inference with Attribute Extractor (Shape Classifier, Color Extractor)\n",
    "shape, color = perceiver.extract_attributes(objects[3][0])\n",
    "print(f'Shape = {shape}\\nColor = {color}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Pipeline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "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>shape</th>\n",
       "      <th>color</th>\n",
       "      <th>position</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>rectangle</td>\n",
       "      <td>orange</td>\n",
       "      <td>(171, 172)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>circle</td>\n",
       "      <td>red</td>\n",
       "      <td>(23, 140)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>circle</td>\n",
       "      <td>gray</td>\n",
       "      <td>(127, 20)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>circle</td>\n",
       "      <td>blue</td>\n",
       "      <td>(127, 136)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>rectangle</td>\n",
       "      <td>green</td>\n",
       "      <td>(31, 80)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>circle</td>\n",
       "      <td>yellow</td>\n",
       "      <td>(87, 17)</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       shape   color    position\n",
       "0  rectangle  orange  (171, 172)\n",
       "0     circle     red   (23, 140)\n",
       "0     circle    gray   (127, 20)\n",
       "0     circle    blue  (127, 136)\n",
       "0  rectangle   green    (31, 80)\n",
       "0     circle  yellow    (87, 17)"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scene = perceiver.scene_repr(img)\n",
    "scene"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Semantic Parser"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Preprocess Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "from semantic_parser import *\n",
    "from torchtext.data import BucketIterator"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create the preprocessor object and preprocess\n",
    "preproc = Preprocessor('data/train/train_que_to_prog.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Get the dataset object\n",
    "train_data = preproc.train_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "defaultdict(<bound method Vocab._default_unk_index of <torchtext.vocab.Vocab object at 0x77c9cadabfa0>>,\n",
       "            {'<unk>': 0,\n",
       "             '<pad>': 1,\n",
       "             '<sos>': 2,\n",
       "             '<eos>': 3,\n",
       "             '<nxt>': 4,\n",
       "             'filter': 5,\n",
       "             'query': 6,\n",
       "             'shape': 7,\n",
       "             'relate': 8,\n",
       "             'orange': 9,\n",
       "             'gray': 10,\n",
       "             'count': 11,\n",
       "             'isleft': 12,\n",
       "             'positions': 13,\n",
       "             'istop': 14,\n",
       "             'position': 15,\n",
       "             'closest': 16,\n",
       "             'red': 17,\n",
       "             'yellow': 18,\n",
       "             'farthest': 19,\n",
       "             'blue': 20,\n",
       "             'green': 21})"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Looking at the Vocabulary\n",
    "preproc.prog_f.vocab.stoi"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "defaultdict(<bound method Vocab._default_unk_index of <torchtext.vocab.Vocab object at 0x77c9cadabfd0>>,\n",
       "            {'<unk>': 0,\n",
       "             '<pad>': 1,\n",
       "             '<sos>': 2,\n",
       "             '<eos>': 3,\n",
       "             'the': 4,\n",
       "             'object': 5,\n",
       "             '?': 6,\n",
       "             'of': 7,\n",
       "             'is': 8,\n",
       "             'shape': 9,\n",
       "             'there': 10,\n",
       "             'what': 11,\n",
       "             'a': 12,\n",
       "             'on': 13,\n",
       "             'to': 14,\n",
       "             'orange': 15,\n",
       "             'gray': 16,\n",
       "             'are': 17,\n",
       "             'as': 18,\n",
       "             'how': 19,\n",
       "             'many': 20,\n",
       "             'objects': 21,\n",
       "             'that': 22,\n",
       "             'left': 23,\n",
       "             'top': 24,\n",
       "             'closest': 25,\n",
       "             'red': 26,\n",
       "             'yellow': 27,\n",
       "             'farthest': 28,\n",
       "             'blue': 29,\n",
       "             'green': 30})"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Looking at the Vocabulary\n",
    "preproc.que_f.vocab.stoi"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Training hyperparameters\n",
    "num_epochs = 3\n",
    "learning_rate = 3e-4\n",
    "batch_size = 8\n",
    "num_steps = len(train_data) / batch_size\n",
    "\n",
    "# Model hyperparameters\n",
    "config = {\n",
    "    'que_vocab_size': len(preproc.que_f.vocab),\n",
    "    'prog_vocab_size': len(preproc.prog_f.vocab),\n",
    "    'embedding_dim': 256,\n",
    "    'num_heads': 8,\n",
    "    'num_encoder_layers': 3,\n",
    "    'num_decoder_layers': 3,\n",
    "    'dropout': 0.10,\n",
    "    'max_len': 20,\n",
    "    'forward_expansion': 4,\n",
    "    'que_pad_idx': preproc.que_f.vocab.stoi[\"<pad>\"]\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Save the config as a json file\n",
    "import json\n",
    "with open('config.json', 'w') as f:\n",
    "    json.dump(config, f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create Training Generator\n",
    "train_loader = BucketIterator.splits((train_data,),\n",
    "                                     batch_size=batch_size,\n",
    "                                     sort_within_batch=True,\n",
    "                                     sort_key=lambda x: len(x.query),\n",
    "                                     device=device)[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/m0nku/.local/lib/python3.10/site-packages/torch/nn/modules/transformer.py:306: UserWarning: enable_nested_tensor is True, but self.use_nested_tensor is False because encoder_layer.self_attn.batch_first was not True(use batch_first for better inference performance)\n",
      "  warnings.warn(f\"enable_nested_tensor is True, but self.use_nested_tensor is False because {why_not_sparsity_fast_path}\")\n"
     ]
    }
   ],
   "source": [
    "# Create Model\n",
    "seq2seq = Seq2Seq(config)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/m0nku/.local/lib/python3.10/site-packages/torch/optim/lr_scheduler.py:28: UserWarning: The verbose parameter is deprecated. Please use get_last_lr() to access the learning rate.\n",
      "  warnings.warn(\"The verbose parameter is deprecated. Please use get_last_lr() \"\n",
      "Epoch 1: 100%|██████████| 250/250.0 [00:09<00:00, 26.50it/s, Loss=0.0042] \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1: Mean Loss = 0.2506533559206873\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 2: 100%|██████████| 250/250.0 [00:08<00:00, 28.24it/s, Loss=0.00143]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 2: Mean Loss = 0.0021417034552432598\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Epoch 3: 100%|██████████| 250/250.0 [00:08<00:00, 28.11it/s, Loss=0.00072] \n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 3: Mean Loss = 0.0009672136448789388\n",
      "\n",
      "CPU times: user 1min 5s, sys: 439 ms, total: 1min 5s\n",
      "Wall time: 27.2 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "# Train Model\n",
    "seq2seq.train_model(train_loader, num_epochs, num_steps, filename='models/semantic_parser.pth')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Inference on Queries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "sem_parser = SemanticParser(preproc, config)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['filter green', 'query shape']"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "program = sem_parser.predict('What is the furthest shape from the green object?')\n",
    "program"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Program Executor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "from program_executor import ProgramExecutor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "executor = ProgramExecutor()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Plugging Everything Together"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "from program_executor import *\n",
    "from perception import *\n",
    "from semantic_parser import *\n",
    "import torch\n",
    "from skimage.io import imshow\n",
    "import cv2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load the config file for transformers\n",
    "import json\n",
    "with open('config.json', 'r') as f:\n",
    "    config = json.load(f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "class NSAIPipeline():\n",
    "    '''End-to-End Pipeline of Neuro-Symbolic AI on Sort-of-CLEVR dataset'''\n",
    "    def __init__(self, \n",
    "                 config,\n",
    "                 detector='models/detector.svm',\n",
    "                 classifier='models/classifier.pth',\n",
    "                 sem_parser='models/semantic_parser.pth',\n",
    "                 train_csv='data/train/train_que_to_prog.csv',\n",
    "                 device=None):\n",
    "        \n",
    "        if device is None:\n",
    "            device = 'cuda' if torch.cuda.is_available() else 'cpu'\n",
    "        self.device = device\n",
    "        \n",
    "        # Perception Module\n",
    "        self.perceiver = PerceptionPipe(detector, classifier, self.device)\n",
    "        \n",
    "        # Semantic Parser\n",
    "        self.preproc = Preprocessor(train_csv)\n",
    "        self.sem_parser = SemanticParser(self.preproc, config, filename=sem_parser, device=self.device)\n",
    "        \n",
    "        # Program Executor\n",
    "        self.executor = ProgramExecutor()\n",
    "        \n",
    "    def predict(self, img, query):\n",
    "        '''\n",
    "        Make Prediction on a single image and question pair\n",
    "        \n",
    "        Args:\n",
    "            img (str/array): pixel values should be in 0-255 range\n",
    "                             of dtype uint8 in BGR color format or\n",
    "                             file path of the image\n",
    "            query (str): question about the image\n",
    "            \n",
    "        Returns:\n",
    "            str: answer of the query\n",
    "        '''\n",
    "        # Load img if it's a path\n",
    "        if type(img) == str:\n",
    "            img = cv2.imread(img)\n",
    "        \n",
    "        # Structured Scene Representation\n",
    "        scene = self.perceiver.scene_repr(img)\n",
    "        # Synthesize Program from Query\n",
    "        program = self.sem_parser.predict(query)\n",
    "        # Execute Program\n",
    "        answer = self.executor(scene, program)\n",
    "        \n",
    "        return answer, program\n",
    "    \n",
    "    def evaluate(self, csv, img_dir, debug=True):\n",
    "        '''\n",
    "        Evaluate the model on a dataset\n",
    "        \n",
    "        Args:\n",
    "            csv (str): path of the csv containing image filename, answer, query and program\n",
    "            img_dir (str): directory containing the images\n",
    "            debug (bool): View the data points which were wrong\n",
    "        Returns:\n",
    "            int: accuracy of the model\n",
    "        '''\n",
    "        data = pd.read_csv(csv).values\n",
    "        \n",
    "        correct = []\n",
    "        for filename, answer, query, program in tqdm(data):\n",
    "            # Load Image\n",
    "            img_path = os.path.join(img_dir, filename)\n",
    "            img = cv2.imread(img_path)\n",
    "            \n",
    "            # Make prediction\n",
    "            pred, pred_prog = self.predict(img, query)\n",
    "            \n",
    "            # Verify answer\n",
    "            if str(pred) == answer:\n",
    "                correct.append(1)\n",
    "            else:\n",
    "                if debug:\n",
    "                    print(filename, answer, pred, query, pred_prog)\n",
    "                correct.append(0)\n",
    "        \n",
    "        acc = (sum(correct) / len(correct)) * 100\n",
    "        \n",
    "        return acc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "nsai = NSAIPipeline(config)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x77c957907be0>"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "img2 = cv2.imread('data/test/images/1.jpg')\n",
    "imshow(cv2.cvtColor(img2, cv2.COLOR_BGR2RGB))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(6, ['filter gray', 'query shape', 'filter', 'count'])"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nsai.predict('data/test/images/1.jpg', 'How many objects of the same shape as the gray object are there?')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x77c957c377c0>"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "img2 = cv2.imread('data/train/images/2.jpg')\n",
    "imshow(cv2.cvtColor(img2, cv2.COLOR_BGR2RGB))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "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>shape</th>\n",
       "      <th>color</th>\n",
       "      <th>position</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>rectangle</td>\n",
       "      <td>orange</td>\n",
       "      <td>(155, 172)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>rectangle</td>\n",
       "      <td>red</td>\n",
       "      <td>(31, 108)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>circle</td>\n",
       "      <td>gray</td>\n",
       "      <td>(31, 32)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>circle</td>\n",
       "      <td>green</td>\n",
       "      <td>(91, 164)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>circle</td>\n",
       "      <td>blue</td>\n",
       "      <td>(167, 108)</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>circle</td>\n",
       "      <td>yellow</td>\n",
       "      <td>(131, 84)</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       shape   color    position\n",
       "0  rectangle  orange  (155, 172)\n",
       "0  rectangle     red   (31, 108)\n",
       "0     circle    gray    (31, 32)\n",
       "0     circle   green   (91, 164)\n",
       "0     circle    blue  (167, 108)\n",
       "0     circle  yellow   (131, 84)"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nsai.executor.scene = nsai.perceiver.scene_repr(img2)\n",
    "nsai.executor.scene"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['filter yellow', 'query position', 'istop']"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "program = nsai.sem_parser.predict('')\n",
    "program"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nsai.executor.count(nsai.executor.filter_(nsai.executor.query(nsai.executor.filter_('red'), 'shape')))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "('no', ['filter yellow', 'query position', 'istop'])"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nsai.predict(img2, 'Is there a yellow object on the top? ')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.10.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}