{"cells":[{"metadata":{"_cell_guid":"522ce330-d4f9-4fbe-9684-4b65fd684cca","_uuid":"174484daa5f084ce4970f5048d3e05d2c4429787"},"cell_type":"markdown","source":"# Overview\n\n The notebook is modified from \"QuickDraw Baseline LSTM Torch \"https://www.kaggle.com/kmader/quickdraw-baseline-lstm-torch, by chaning LSTM layer to Bidirectional LSTM layer and add Bach Normalization layer to prevent from Gradient Vanishing/Exploding problem. \n \n \"QuickDraw Baseline LSTM Torch\" kernel uses stroke-based LSTM, it means not from original image data, but from processed stroke vector data. And 'preprocesses' uses 1D convolutions and then uses two stacked LSTMs followed by two dense layers to make the classification. The model can be thought to 'read' the drawing stroke by stroke."},{"metadata":{"_uuid":"d8ccded02a5f2c4a9d9ee2f7688114bcd2e1f11a"},"cell_type":"markdown","source":"### Model Parameters\nHere we keep track of the relevant parameters for the data preprocessing, model construction and training"},{"metadata":{"trusted":true,"_uuid":"8b08fbab2000a563b388f126eac74362641e497c"},"cell_type":"code","source":"batch_size = 2048\nSTROKE_COUNT = 196\nTRAIN_SAMPLES = 750\nVALID_SAMPLES = 75\nTEST_SAMPLES = 50\nuse_gpu = True\nLEARNING_RATE = 1e-3\nEPOCHS = 50","execution_count":226,"outputs":[]},{"metadata":{"_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","trusted":true},"cell_type":"code","source":"%matplotlib inline\nimport os\nimport numpy as np\nimport matplotlib.pyplot as plt\nfrom keras.utils.np_utils import to_categorical\nfrom keras.preprocessing.sequence import pad_sequences\nfrom sklearn.preprocessing import LabelEncoder\nimport pandas as pd\nfrom glob import glob\nimport torch\nimport torch.utils.data as data\nimport torch.nn as nn\nfrom torch.nn import functional as F\nimport gc\ngc.enable()","execution_count":227,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"4caa9ac827c954e17a0f0a9f8e1d102cdf4d04ca"},"cell_type":"code","source":"base_dir = os.path.join('..', 'input')\ntest_path = os.path.join(base_dir, 'test_simplified.csv')","execution_count":228,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"296579e90531a3616919e2edc393836e59a80662"},"cell_type":"code","source":"device = torch.device(\"cuda\" if use_gpu else \"cpu\")\ntorch.manual_seed(42) # try and make the results more reproducible","execution_count":229,"outputs":[{"output_type":"execute_result","execution_count":229,"data":{"text/plain":"<torch._C.Generator at 0x7f50df90a1b0>"},"metadata":{}}]},{"metadata":{"_uuid":"5c6290dcb4437dfa9c3927efb4e5203d360d8264"},"cell_type":"markdown","source":"## Import the Data\nRead and process the stroke data"},{"metadata":{"trusted":true,"_uuid":"7acacf8e960084782425ef1a1a3fd532a240ad48"},"cell_type":"code","source":"from ast import literal_eval\nALL_TRAIN_PATHS = glob(os.path.join(base_dir, 'train_simplified', '*.csv'))\nCOL_NAMES = ['countrycode', 'drawing', 'key_id', 'recognized', 'timestamp', 'word']\n\ndef _stack_it(raw_strokes):\n    \"\"\"preprocess the string and make \n    a standard Nx3 stroke vector\"\"\"\n    stroke_vec = literal_eval(raw_strokes) # string->list\n    # unwrap the list\n    in_strokes = [(xi,yi,i)  \n     for i,(x,y) in enumerate(stroke_vec) \n     for xi,yi in zip(x,y)]\n    c_strokes = np.stack(in_strokes)\n    # replace stroke id with 1 for continue, 2 for new\n    c_strokes[:,2] = [1]+np.diff(c_strokes[:,2]).tolist()\n    c_strokes[:,2] += 1 # since 0 is no stroke\n    # pad the strokes with zeros\n    return pad_sequences(c_strokes.swapaxes(0, 1), \n                         maxlen=STROKE_COUNT, \n                         padding='post').swapaxes(0, 1)\ndef read_batch(samples=5, \n               start_row=0,\n               max_rows = 1000):\n    \"\"\"\n    load and process the csv files\n    this function is horribly inefficient but simple\n    \"\"\"\n    out_df_list = []\n    for c_path in ALL_TRAIN_PATHS:\n        c_df = pd.read_csv(c_path, nrows=max_rows, skiprows=start_row)\n        c_df.columns=COL_NAMES\n        out_df_list += [c_df.sample(samples)[['drawing', 'word']]]\n    full_df = pd.concat(out_df_list)\n    full_df['drawing'] = full_df['drawing'].\\\n        map(_stack_it)\n    \n    return full_df","execution_count":230,"outputs":[]},{"metadata":{"_uuid":"5ec1854b21b36cd2fd7f7d0717aaa8da32506a6a"},"cell_type":"markdown","source":"# Reading and Parsing\nSince it is too much data (23GB) to read in at once, we just take a portion of it for training, validation and hold-out testing. This should give us an idea about how well the model works, but leaves lots of room for improvement later"},{"metadata":{"_cell_guid":"79c7e3d0-c299-4dcb-8224-4455121ee9b0","_uuid":"d629ff2d2480ee46fbb7e2d37f6b5fab8052498a","trusted":true},"cell_type":"code","source":"train_args = dict(samples=TRAIN_SAMPLES, \n                  start_row=0, \n                  max_rows=int(TRAIN_SAMPLES*1.5))\nvalid_args = dict(samples=VALID_SAMPLES, \n                  start_row=train_args['max_rows']+1, \n                  max_rows=VALID_SAMPLES+25)\ntest_args = dict(samples=TEST_SAMPLES, \n                 start_row=valid_args['max_rows']+train_args['max_rows']+1, \n                 max_rows=TEST_SAMPLES+25)\ntrain_df = read_batch(**train_args)\nvalid_df = read_batch(**valid_args)\ntest_df = read_batch(**test_args)\nword_encoder = LabelEncoder()\nword_encoder.fit(train_df['word'])\nprint('words', len(word_encoder.classes_), '=>', ', '.join([x for x in word_encoder.classes_]))","execution_count":231,"outputs":[{"output_type":"stream","text":"words 340 => The Eiffel Tower, The Great Wall of China, The Mona Lisa, airplane, alarm clock, ambulance, angel, animal migration, ant, anvil, apple, arm, asparagus, axe, backpack, banana, bandage, barn, baseball, baseball bat, basket, basketball, bat, bathtub, beach, bear, beard, bed, bee, belt, bench, bicycle, binoculars, bird, birthday cake, blackberry, blueberry, book, boomerang, bottlecap, bowtie, bracelet, brain, bread, bridge, broccoli, broom, bucket, bulldozer, bus, bush, butterfly, cactus, cake, calculator, calendar, camel, camera, camouflage, campfire, candle, cannon, canoe, car, carrot, castle, cat, ceiling fan, cell phone, cello, chair, chandelier, church, circle, clarinet, clock, cloud, coffee cup, compass, computer, cookie, cooler, couch, cow, crab, crayon, crocodile, crown, cruise ship, cup, diamond, dishwasher, diving board, dog, dolphin, donut, door, dragon, dresser, drill, drums, duck, dumbbell, ear, elbow, elephant, envelope, eraser, eye, eyeglasses, face, fan, feather, fence, finger, fire hydrant, fireplace, firetruck, fish, flamingo, flashlight, flip flops, floor lamp, flower, flying saucer, foot, fork, frog, frying pan, garden, garden hose, giraffe, goatee, golf club, grapes, grass, guitar, hamburger, hammer, hand, harp, hat, headphones, hedgehog, helicopter, helmet, hexagon, hockey puck, hockey stick, horse, hospital, hot air balloon, hot dog, hot tub, hourglass, house, house plant, hurricane, ice cream, jacket, jail, kangaroo, key, keyboard, knee, ladder, lantern, laptop, leaf, leg, light bulb, lighthouse, lightning, line, lion, lipstick, lobster, lollipop, mailbox, map, marker, matches, megaphone, mermaid, microphone, microwave, monkey, moon, mosquito, motorbike, mountain, mouse, moustache, mouth, mug, mushroom, nail, necklace, nose, ocean, octagon, octopus, onion, oven, owl, paint can, paintbrush, palm tree, panda, pants, paper clip, parachute, parrot, passport, peanut, pear, peas, pencil, penguin, piano, pickup truck, picture frame, pig, pillow, pineapple, pizza, pliers, police car, pond, pool, popsicle, postcard, potato, power outlet, purse, rabbit, raccoon, radio, rain, rainbow, rake, remote control, rhinoceros, river, roller coaster, rollerskates, sailboat, sandwich, saw, saxophone, school bus, scissors, scorpion, screwdriver, sea turtle, see saw, shark, sheep, shoe, shorts, shovel, sink, skateboard, skull, skyscraper, sleeping bag, smiley face, snail, snake, snorkel, snowflake, snowman, soccer ball, sock, speedboat, spider, spoon, spreadsheet, square, squiggle, squirrel, stairs, star, steak, stereo, stethoscope, stitches, stop sign, stove, strawberry, streetlight, string bean, submarine, suitcase, sun, swan, sweater, swing set, sword, t-shirt, table, teapot, teddy-bear, telephone, television, tennis racquet, tent, tiger, toaster, toe, toilet, tooth, toothbrush, toothpaste, tornado, tractor, traffic light, train, tree, triangle, trombone, truck, trumpet, umbrella, underwear, van, vase, violin, washing machine, watermelon, waterslide, whale, wheel, windmill, wine bottle, wine glass, wristwatch, yoga, zebra, zigzag\n","name":"stdout"}]},{"metadata":{"_cell_guid":"6d29237e-ece3-4dfd-9095-475296f4a608","_uuid":"8bae16a4973a215861fbb536a602c4f5abf3b4bf"},"cell_type":"markdown","source":"# Stroke-based Classification\nHere we use the stroke information to train a model and see if the strokes give us a better idea of what the shape could be. "},{"metadata":{"_cell_guid":"ff5ddced-d77e-473f-899d-82cf11ad2bd9","_uuid":"409468f1d5abd17b819482473a4f354a61f8d7ef","trusted":true},"cell_type":"code","source":"def get_Xy(in_df):\n    X = np.stack(in_df['drawing'], 0)\n    y = to_categorical(word_encoder.transform(in_df['word'].values))\n    return X, y\ntrain_X, train_y = get_Xy(train_df)\nvalid_X, valid_y = get_Xy(valid_df)\ntest_X, test_y = get_Xy(test_df)\nprint(train_X.shape)","execution_count":244,"outputs":[{"output_type":"stream","text":"(255000, 196, 3)\n","name":"stdout"}]},{"metadata":{"_cell_guid":"56240ed9-42b0-4f62-b3d1-f92017f04e30","_uuid":"5cc79204a0a1da048d1d58ba8dfdafd0af3ebcb8","trusted":true},"cell_type":"code","source":"fig, m_axs = plt.subplots(3,3, figsize = (16, 16))\nrand_idxs = np.random.choice(range(train_X.shape[0]), size = 9)\nfor c_id, c_ax in zip(rand_idxs, m_axs.flatten()):\n    test_arr = train_X[c_id]\n    test_arr = test_arr[test_arr[:,2]>0, :] # only keep valid points\n    lab_idx = np.cumsum(test_arr[:,2]-1)\n    for i in np.unique(lab_idx):\n        c_ax.plot(test_arr[lab_idx==i,0], \n                np.max(test_arr[:,1])-test_arr[lab_idx==i,1], '.-')\n    c_ax.axis('off')\n    c_ax.set_title(word_encoder.classes_[np.argmax(train_y[c_id])])","execution_count":245,"outputs":[{"output_type":"display_data","data":{"text/plain":"<Figure size 1152x1152 with 9 Axes>","image/png":"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\n"},"metadata":{}}]},{"metadata":{"_cell_guid":"5e1d5bba-0fb4-432c-bd0b-ad69be0ef9ac","_uuid":"b4a087a17798c2ec8eb520bc916bcad38d4ebff2","collapsed":true},"cell_type":"markdown","source":"# LSTM to Parse Strokes\nThe model suggeted from the tutorial is\n\n![Suggested Model](https://www.tensorflow.org/versions/master/images/quickdraw_model.png)"},{"metadata":{"trusted":true,"_uuid":"e87449fb9c115250b912df2d4a4cac1b6d138791"},"cell_type":"code","source":"class LSTMNet(nn.Module):\n    def __init__(self, in_channels, classes):\n        \"\"\"Define the components of a LSTM Stroke Model\"\"\"\n        super(LSTMNet, self).__init__()   \n        self.bn1 = nn.BatchNorm1d(num_features=48)\n        self.relu = nn.ReLU(inplace=True)\n        self.conv_1 = nn.Conv1d(in_channels=in_channels, out_channels=48, kernel_size=5)\n        self.conv_2 = nn.Conv1d(in_channels=48, out_channels=64, kernel_size=5)\n        self.conv_3 = nn.Conv1d(in_channels=64, out_channels=96, kernel_size=3)\n        \n        self.lstm = nn.LSTM(input_size=96, hidden_size=128, \n                            batch_first=False,\n                            dropout=0.3, num_layers=2, bidirectional=True)\n        self.dropout = nn.Dropout(0.3)\n        self.feat_vec = nn.Linear(256, 512)\n        self.pred = nn.Linear(512, classes)\n\n    def forward(self,x):\n        \"\"\"Input x is expected to be a 3d tensor (N x seq_length x C)\n           N - Number of images in minibatch\n           seq_length - Length of the sequen\n           C - number of channels\"\"\"\n        x = self.bn1(self.conv_1(x))\n        #x = self.relu(x)\n        x = self.dropout(x)\n        x = self.relu(self.conv_2(x))\n        x = self.dropout(x)\n        x = self.relu(self.conv_3(x))\n        x = self.dropout(x)\n        x = x.permute(2, 0, 1) # prepare for the LSTM\n        output, (h_0, c_0) = self.lstm(x)\n        #print(\"output shape : \", output.shape) #[186,2048,256]\n        x = self.relu(self.feat_vec(output[-1]))\n        return self.pred(x)","execution_count":246,"outputs":[]},{"metadata":{"trusted":true,"scrolled":true,"_uuid":"859fd4467047861fe4e179b298ee086ffd33bff6"},"cell_type":"code","source":"model = LSTMNet(classes=len(word_encoder.classes_), in_channels = train_X.shape[2]).to(device)\nmodel","execution_count":247,"outputs":[{"output_type":"execute_result","execution_count":247,"data":{"text/plain":"LSTMNet(\n  (bn1): BatchNorm1d(48, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n  (relu): ReLU(inplace)\n  (conv_1): Conv1d(3, 48, kernel_size=(5,), stride=(1,))\n  (conv_2): Conv1d(48, 64, kernel_size=(5,), stride=(1,))\n  (conv_3): Conv1d(64, 96, kernel_size=(3,), stride=(1,))\n  (lstm): LSTM(96, 128, num_layers=2, dropout=0.3, bidirectional=True)\n  (dropout): Dropout(p=0.3)\n  (feat_vec): Linear(in_features=256, out_features=512, bias=True)\n  (pred): Linear(in_features=512, out_features=340, bias=True)\n)"},"metadata":{}}]},{"metadata":{"trusted":true,"_uuid":"200b87f3492e5bbb42a81b0ccff63515e4f771d5"},"cell_type":"code","source":"criterion = nn.CrossEntropyLoss() #Use cross entropy loss\noptimizer = torch.optim.Adam(model.parameters(), lr=LEARNING_RATE)","execution_count":248,"outputs":[]},{"metadata":{"_uuid":"3db0f526d5ebdf3e89fca1480c45d895901281e4"},"cell_type":"markdown","source":"## Prepare Training Data"},{"metadata":{"trusted":true,"_uuid":"f61761f43f2cfe46fd7d8cb6253e66c0df9abbc8"},"cell_type":"code","source":"train_loader = torch.utils.data.DataLoader(list(zip(train_X.swapaxes(1, 2).astype('float32'), # pytorch and tf/keras have different orders\n                                                    np.argmax(train_y, 1)\n                                                   )),\n                                           shuffle=True, batch_size=batch_size, \n                                           pin_memory=False)\ntest_loader = torch.utils.data.DataLoader(list(zip(valid_X.swapaxes(1, 2).astype('float32'), \n                                                   np.argmax(valid_y, 1)\n                                                  )),\n                                           shuffle=True, batch_size=batch_size, \n                                           pin_memory=False)","execution_count":249,"outputs":[]},{"metadata":{"_cell_guid":"825b3af8-9451-487b-a1e1-538f2f1489e1","_uuid":"ed2fc26af74aed1a93bbc253d61b72db5a81f5cc","trusted":true,"scrolled":false},"cell_type":"code","source":"# nice wait bars\nfrom tqdm import tqdm_notebook, tnrange\nfrom collections import defaultdict\nfrom IPython.display import clear_output, display\ntrain_results = defaultdict(list)\ntrain_iter, test_iter, best_acc = 0,0,0\nfig, ((ax1, ax2), (ax3, ax4)) = plt.subplots(2, 2, figsize = (10, 10))\nax1.set_title('Train Loss')\nax2.set_title('Train Accuracy')\nax3.set_title('Test Loss')\nax4.set_title('Test Accuracy')\n\nfor i in tnrange(EPOCHS, desc='Epochs'):\n    clear_output(wait=True)\n    display(fig)\n    print(\"Epoch \",i)\n    ## Train Phase\n    #Model switches to train phase\n    model.train() \n    \n    # Running through all mini batches in the dataset\n    count, loss_val, correct, total = train_iter, 0, 0, 0\n    for data, target in tqdm_notebook(train_loader, desc='Training'):    \n        if use_gpu: #Using GPU & Cuda\n            data, target = data.to(device), target.to(device)\n\n        output = model(data) #FWD prop\n        loss = criterion(output, target) #Cross entropy loss\n        c_loss = loss.data.item()\n        ax1.plot(count, c_loss, 'r.')\n        loss_val += c_loss\n\n        optimizer.zero_grad() #Zero out any cached gradients\n        loss.backward() #Backward pass\n        optimizer.step() #Update the weights\n\n        #Compute accuracy\n        predicted = output.data.max(1)[1] #get index of max\n        total += target.size(0) #total samples in mini batch\n        c_acc = (predicted == target).sum().item()\n        ax2.plot(count, c_acc/target.size(0), 'r.')\n        correct += c_acc\n        count +=1\n    train_loss_val, train_iter, train_acc = loss_val/len(train_loader.dataset), count, correct/float(total)\n    \n    print(\"Training loss: \", train_loss_val, \" train acc: \",train_acc)    \n    ## Test Phase\n    \n    #Model switches to test phase\n    model.eval()\n\n    #Running through all mini batches in the dataset\n    count, correct, total, lost_val = test_iter, 0, 0, 0\n    for data, target in tqdm_notebook(test_loader, desc='Testing'):\n        if use_gpu: #Using GPU & Cuda\n            data, target = data.to(device), target.to(device)\n        output = model(data)\n        loss = criterion(output, target) #Cross entropy loss\n        c_loss = loss.data.item()\n        ax3.plot(count, c_loss, 'b.')\n        loss_val += c_loss\n        #Compute accuracy\n        predicted = output.data.max(1)[1] #get index of max\n        total += target.size(0) #total samples in mini batch\n        c_acc = (predicted == target).sum().item()\n        ax4.plot(count, c_acc/target.size(0), 'b.')\n        correct += c_acc\n        count += 1\n\n    #Accuracy over entire dataset\n    test_acc, test_iter, test_loss_val = correct/float(total), count, loss_val/len(test_loader.dataset)\n    print(\"Epoch: \",i,\" test set accuracy: \",test_acc)\n    \n    train_results['epoch'].append(i)\n    train_results['train_loss'].append(train_loss_val)\n    train_results['train_acc'].append(train_acc)\n    train_results['train_iter'].append(train_iter)\n    \n    train_results['test_loss'].append(test_loss_val)\n    train_results['test_acc'].append(test_acc)\n    train_results['test_iter'].append(test_iter)\n    \n    #Save model with best accuracy\n    if test_acc > best_acc:\n        best_acc = test_acc\n        torch.save(model.state_dict(), 'best_model.pth') \n# plt.show()\n# fig.savefig('train.png')","execution_count":250,"outputs":[{"output_type":"display_data","data":{"text/plain":"<Figure size 720x720 with 4 Axes>","image/png":"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\n"},"metadata":{}},{"output_type":"stream","text":"Epoch  0\n","name":"stdout"},{"output_type":"display_data","data":{"text/plain":"HBox(children=(IntProgress(value=0, description='Training', max=125, style=ProgressStyle(description_width='in…","application/vnd.jupyter.widget-view+json":{"version_major":2,"version_minor":0,"model_id":"3a22f5559fd54e4185e1897e34b9a8c1"}},"metadata":{}},{"output_type":"error","ename":"RuntimeError","evalue":"CUDA out of memory. Tried to allocate 4.02 GiB (GPU 0; 15.90 GiB total capacity; 9.32 GiB already allocated; 2.96 GiB free; 2.91 GiB cached)","traceback":["\u001b[0;31m---------------------------------------------------------------------------\u001b[0m","\u001b[0;31mRuntimeError\u001b[0m                              Traceback (most recent call last)","\u001b[0;32m<ipython-input-250-e3205f0f587a>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     25\u001b[0m             \u001b[0mdata\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtarget\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdata\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mto\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdevice\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtarget\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mto\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdevice\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     26\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 27\u001b[0;31m         \u001b[0moutput\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;31m#FWD prop\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     28\u001b[0m         \u001b[0mloss\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mcriterion\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0moutput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtarget\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;31m#Cross entropy loss\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     29\u001b[0m         \u001b[0mc_loss\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mloss\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mitem\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;32m/opt/conda/lib/python3.6/site-packages/torch/nn/modules/module.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *input, **kwargs)\u001b[0m\n\u001b[1;32m    487\u001b[0m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_slow_forward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    488\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 489\u001b[0;31m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    490\u001b[0m         \u001b[0;32mfor\u001b[0m \u001b[0mhook\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_forward_hooks\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalues\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    491\u001b[0m             \u001b[0mhook_result\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mhook\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mresult\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;32m<ipython-input-246-679e6f71bf93>\u001b[0m in \u001b[0;36mforward\u001b[0;34m(self, x)\u001b[0m\n\u001b[1;32m     29\u001b[0m         \u001b[0mx\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdropout\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     30\u001b[0m         \u001b[0mx\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpermute\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;31m# prepare for the LSTM\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 31\u001b[0;31m         \u001b[0moutput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mh_0\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mc_0\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlstm\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     32\u001b[0m         \u001b[0;31m#print(\"output shape : \", output.shape) #[186,2048,256]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     33\u001b[0m         \u001b[0mx\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrelu\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfeat_vec\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0moutput\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;32m/opt/conda/lib/python3.6/site-packages/torch/nn/modules/module.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *input, **kwargs)\u001b[0m\n\u001b[1;32m    487\u001b[0m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_slow_forward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    488\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 489\u001b[0;31m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    490\u001b[0m         \u001b[0;32mfor\u001b[0m \u001b[0mhook\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_forward_hooks\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalues\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    491\u001b[0m             \u001b[0mhook_result\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mhook\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mresult\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;32m/opt/conda/lib/python3.6/site-packages/torch/nn/modules/rnn.py\u001b[0m in \u001b[0;36mforward\u001b[0;34m(self, input, hx)\u001b[0m\n\u001b[1;32m    177\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mbatch_sizes\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    178\u001b[0m             result = _impl(input, hx, self._flat_weights, self.bias, self.num_layers,\n\u001b[0;32m--> 179\u001b[0;31m                            self.dropout, self.training, self.bidirectional, self.batch_first)\n\u001b[0m\u001b[1;32m    180\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    181\u001b[0m             result = _impl(input, batch_sizes, hx, self._flat_weights, self.bias,\n","\u001b[0;31mRuntimeError\u001b[0m: CUDA out of memory. Tried to allocate 4.02 GiB (GPU 0; 15.90 GiB total capacity; 9.32 GiB already allocated; 2.96 GiB free; 2.91 GiB cached)"]},{"output_type":"display_data","data":{"text/plain":"<Figure size 720x720 with 4 Axes>","image/png":"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\n"},"metadata":{}}]},{"metadata":{"trusted":true,"_uuid":"8a9f2bf3b23925f2f20c7c92eb346e428ed89097"},"cell_type":"code","source":"train_results_df = pd.DataFrame(train_results)\nfig, (ax1, ax2) = plt.subplots(1, 2, figsize = (10, 5))\ntrain_results_df.plot('epoch', ['train_loss', 'test_loss'], ax=ax1)\nax1.set_title('Loss')\ntrain_results_df.plot('epoch', ['train_acc', 'test_acc'], ax=ax2)\nax2.set_title('Accuracy')\nfig.savefig('epochs.png')","execution_count":null,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"3c56a2474fb24344d245afd99aab26f090b5d7e5"},"cell_type":"code","source":"# load the best model\nmodel.load_state_dict(torch.load('best_model.pth'))","execution_count":null,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"23f4e4a3d0b7080782b48cbef73db27aebeaf7ca"},"cell_type":"code","source":"#Model switches to test phase\ntop_k_count = 5\nmodel.eval()\n#Running through all mini batches in the dataset\ncount, correct, top_k_correct, total, lost_val = test_iter, 0, 0, 0, 0\npred_cat = []\ntest_cat = []\nfor data, target in tqdm_notebook(test_loader, desc='Testing'):\n    if use_gpu: #Using GPU & Cuda\n        data, target = data.to(device), target.to(device)\n    output = model(data)\n    loss = criterion(output, target) #Cross entropy loss\n    c_loss = loss.data.item()\n    loss_val += c_loss\n    #Compute accuracy\n    predicted = output.data.max(1)[1] #get index of max\n    # add outputs for better visualization\n    test_cat.append(output.data.max(1)[1].cpu().numpy())\n    pred_cat.append(target.data.cpu().numpy())\n        \n    total += target.size(0) #total samples in mini batch\n    c_acc = (predicted == target).sum().item()\n    top_k_correct += c_acc\n    for k in range(1, top_k_count+1):\n        top_k_correct += (output.data.argsort(1, descending=True)[:, k] == target).sum().item()\n    correct += c_acc\n    count += 1\n\n#Accuracy over entire dataset\ntest_acc, test_iter, test_loss_val, top_k_acc = correct/float(total), count, loss_val/len(test_loader.dataset), top_k_correct/float(total)\ntest_cat = np.concatenate(test_cat, 0)\npred_cat = np.concatenate(pred_cat, 0)\nprint('Accuracy: %2.1f%%, Top %d Accuracy %2.1f%%' % (100*test_acc, top_k_count, 100*top_k_acc))","execution_count":null,"outputs":[]},{"metadata":{"_cell_guid":"ee75c585-b134-4ea2-b8f3-219e24efd1f1","_uuid":"6b9cdf52d233de60108d72f540db978801b578c1","trusted":true},"cell_type":"code","source":"from sklearn.metrics import confusion_matrix, classification_report\nplt.matshow(confusion_matrix(test_cat, pred_cat))\nprint(classification_report(test_cat, pred_cat, \n                            target_names = [x for x in word_encoder.classes_]))","execution_count":null,"outputs":[]},{"metadata":{"_cell_guid":"db1d371b-4b2c-478f-b6df-76db58a24fbe","_uuid":"bd9a16adcb46e07d7949644e69bf3483f7dce571"},"cell_type":"markdown","source":"# Reading Point by Point"},{"metadata":{"_uuid":"e99b1ed154f26381d12918e2b4e12db807e6535f"},"cell_type":"markdown","source":"# Submission\nWe can create a submission using the model"},{"metadata":{"_cell_guid":"436a4fce-3843-4c84-8eeb-0161fe3c4e04","_uuid":"4f3a40e23f2e917b68171822944491ab348e15b3","trusted":true},"cell_type":"code","source":"sub_df = pd.read_csv(test_path)\nsub_df['drawing'] = sub_df['drawing'].map(_stack_it)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"72825ea87d35ad96b0254e3af5f5aaf64fb9c78f"},"cell_type":"code","source":"sub_vec = np.stack(sub_df['drawing'].values, 0)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"083ad23b727642fa9663b0ea37710a247288fe91"},"cell_type":"code","source":"submission_loader = torch.utils.data.DataLoader(list(zip(sub_vec.swapaxes(1, 2).astype('float32'), \n                                                   sub_df.index.values\n                                                  )),\n                                           shuffle=False, batch_size=batch_size, \n                                           pin_memory=False)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"3e23e5e90d6a354f6725efaf2f54d89983239ad5"},"cell_type":"code","source":"# make the predictions\nmodel.eval()\npred_out = []\nidx_out = []\nfor data, target in tqdm_notebook(submission_loader, desc='Testing'):\n    if use_gpu: #Using GPU & Cuda\n        data = data.to(device) \n    output = model(data)\n    pred_out += [output.data.cpu().numpy()]\n    idx_out += [target.numpy()]\npred_out = np.concatenate(pred_out, 0)\nidx_out = np.concatenate(idx_out, 0)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"639ca8a511e5e1a02b6cd0333cc04213f8497487"},"cell_type":"code","source":"top_3_pred = [word_encoder.classes_[np.argsort(-1*c_pred)[:3]] for c_pred in pred_out]","execution_count":null,"outputs":[]},{"metadata":{"trusted":true,"_uuid":"68dd3629f5e5b30bede2d4b485a6f1dfabc8d5a4"},"cell_type":"code","source":"top_3_pred = [' '.join([col.replace(' ', '_') for col in row]) for row in top_3_pred]\ntop_3_pred[:3]","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"sub_out_df = pd.DataFrame({'key_id': sub_df['key_id'][idx_out].values, \n                          'word': top_3_pred})\nsub_out_df[['key_id', 'word']].to_csv('submission.csv', index=False)\nsub_out_df.sample(5)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"","execution_count":null,"outputs":[]}],"metadata":{"kernelspec":{"display_name":"Python 3","language":"python","name":"python3"},"language_info":{"name":"python","version":"3.6.6","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"}},"nbformat":4,"nbformat_minor":1}