LSTM_REG.ipynb

{"nbformat":4,"nbformat_minor":0,"metadata":{"colab":{"name":"LSTM_REG.ipynb","provenance":[],"collapsed_sections":[],"mount_file_id":"1bUrb-ePhoo5sRQnIlRQzYwef8FRsLu2O","authorship_tag":"ABX9TyPdupu3eClBJ2vzDIOzO4ze"},"kernelspec":{"name":"python3","display_name":"Python 3"}},"cells":[{"cell_type":"code","metadata":{"id":"5zn9maoCzLyT","colab_type":"code","outputId":"38aa0e1c-31ca-413a-efa2-fafcf0347a56","executionInfo":{"status":"ok","timestamp":1582065306850,"user_tz":0,"elapsed":3228,"user":{"displayName":"Andy Wang","photoUrl":"","userId":"02776860930356410397"}},"colab":{"base_uri":"https://localhost:8080/","height":34}},"source":["import torch\n","import torch.nn as nn\n","import numpy as np\n","from torchvision import datasets, transforms\n","\n","import seaborn as sns\n","import matplotlib.pyplot as plt\n","\n","GPU = True\n","device_idx = 0\n","if GPU:\n","    device = torch.device(\"cuda:\" + str(device_idx) if torch.cuda.is_available() else \"cpu\")\n","else:\n","    device = torch.device(\"cpu\")\n","print(device)\n","\n","# Set default dtype for model weights\n","torch.set_default_dtype(torch.double)"],"execution_count":1,"outputs":[{"output_type":"stream","text":["cpu\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"Lhv5MZ3JyH59","colab_type":"code","colab":{}},"source":["# Load the data\n","import os\n","os.chdir(\"/content/drive/My Drive/Colab Notebooks/NLP/coursework\")\n","import pickle\n","with open(\"en_vectors.pk\", \"rb\") as f:\n","    en_sentences_vectors = pickle.load(f)\n","with open(\"de_vectors.pk\", \"rb\") as f:\n","    de_sentences_vectors = pickle.load(f)\n","with open(\"scores.pk\", \"rb\") as f:\n","    scores = pickle.load(f)"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"JTm0ONuPB7Z2","colab_type":"code","outputId":"96531310-3a2e-4bbb-f324-919c90f1bc06","executionInfo":{"status":"ok","timestamp":1581872486610,"user_tz":0,"elapsed":2183,"user":{"displayName":"Andy Wang","photoUrl":"","userId":"02776860930356410397"}},"colab":{"base_uri":"https://localhost:8080/","height":545}},"source":["en_lengths = np.array([len(en_sentences_vectors[i]) for i in range(len(en_sentences_vectors))])\n","de_lengths = np.array([len(de_sentences_vectors[i]) for i in range(len(de_sentences_vectors))])\n","scores = np.array(scores)\n","\n","# Plotting the lengths of sentences to see where we should pad\n","sns.distplot(en_lengths)\n","plt.title(\"Lengths of sentences EN\")\n","plt.show()\n","sns.distplot(de_lengths)\n","plt.title(\"Lengths of sentences DE\")\n","plt.show()\n"],"execution_count":0,"outputs":[{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAXoAAAEICAYAAABRSj9aAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nj7OgN8YYj7OgN8YYj7OgN8YYj/v/Notu2NRd+XcAAAAASUVORK5CYII=\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[]}},{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAXoAAAEICAYAAABRSj9aAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n/w/0yBA5NxKE6gAAAABJRU5ErkJggg==\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[]}}]},{"cell_type":"code","metadata":{"id":"r-pmaN-XE-4o","colab_type":"code","outputId":"d1cc30b0-c247-4f2f-98c9-cef9ea162d7c","executionInfo":{"status":"ok","timestamp":1581872487509,"user_tz":0,"elapsed":3070,"user":{"displayName":"Andy Wang","photoUrl":"","userId":"02776860930356410397"}},"colab":{"base_uri":"https://localhost:8080/","height":809}},"source":["# Split into 2 group so that inputs are not too sparse\n","# Check whether the proposed splits have the same distribution of scores\n","sns.distplot(scores)\n","plt.title(\"Full Distribution of scores\")\n","plt.show()\n","\n","sns.distplot(scores[en_lengths < 17])\n","plt.title(\"EN_SENT, len < 17\")\n","plt.show()\n","\n","sns.distplot(scores[en_lengths >= 17])\n","plt.title(\"EN_SENT, len >= 17\")\n","plt.show()\n","\n","# sns.distplot(scores[de_lengths < 17])\n","# plt.title(\"DE_SENT, len < 17\")\n","# plt.show()\n","\n","# sns.distplot(scores[de_lengths >= 17])\n","# plt.title(\"DE_SENT, len >= 17\")\n","# plt.show()"],"execution_count":0,"outputs":[{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAXQAAAEICAYAAABPgw/pAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n0EVEmoQCXUSkSSjQRUSahAJdRKRJ/H/K9cc5Tbdd8AAAAABJRU5ErkJggg==\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[]}},{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAXQAAAEICAYAAABPgw/pAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAAElFTkSuQmCC\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[]}},{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAXQAAAEICAYAAABPgw/pAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQmCC\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[]}}]},{"cell_type":"code","metadata":{"id":"K5q6FF8aXXfr","colab_type":"code","colab":{}},"source":["# Split up into two seperate sets and pad\n","pad = [np.zeros(300,)]\n","\n","scores_set_1 = scores[en_lengths < 17]\n","scores_set_2 = scores[en_lengths >= 17]\n","\n","en_set_1 = []\n","en_set_2 = []\n","idx_set_1 = []\n","idx_set_2 = []\n","\n","max_len_en = max(len(sent) for sent in en_sentences_vectors)\n","\n","for i, en_sentence in enumerate(en_sentences_vectors):\n","    sent_len = len(en_sentence)\n","    if sent_len < 17:\n","        en_set_1.append(en_sentence + pad * (16 - sent_len))\n","        idx_set_1.append(i)\n","    else:\n","        en_set_2.append(en_sentence + pad * (max_len_en - sent_len))\n","        idx_set_2.append(i)\n","\n","\n","de_set_1 = []\n","de_set_2 = []\n","max_len_de_set_1 = max(de_lengths[idx_set_1])\n","max_len_de_set_2 = max(len(sent) for sent in de_sentences_vectors)\n","\n","for idx in idx_set_1:\n","    de_set_1.append(de_sentences_vectors[idx] + pad * (max_len_de_set_1 - len(de_sentences_vectors[idx])))\n","for idx in idx_set_2:\n","    de_set_2.append(de_sentences_vectors[idx] + pad * (max_len_de_set_2 - len(de_sentences_vectors[idx])))"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"q43POV9QytX_","colab_type":"code","colab":{}},"source":["class Sentences(torch.utils.data.Dataset):\n","    def __init__(self, en_sentences, de_sentences, scores):\n","        super(Sentences, self).__init__()\n","        self.en_sentences_vectors = torch.tensor(en_sentences, device=device)\n","        self.de_sentences_vectors = torch.tensor(de_sentences, device=device)\n","        self.scores = torch.tensor(scores, device=device)\n","\n","    def __len__(self):\n","        return len(self.scores)\n","\n","    def __getitem__(self, idx):\n","        if torch.is_tensor(idx):\n","            idx = idx.tolist()\n","        \n","        en, de, scores = self.en_sentences_vectors[idx], self.de_sentences_vectors[idx], self.scores[idx]\n","        return en, de, scores"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"KOd75PS_fGoA","colab_type":"code","colab":{}},"source":["# Create train and test sets\n","train_pct = 0.8\n","first_set_train_len = int(train_pct * len(en_set_1))\n","second_set_train_len = int(train_pct * len(en_set_2))\n","\n","first_set = Sentences(en_set_1, de_set_1, scores_set_1)\n","second_set = Sentences(en_set_2, de_set_2, scores_set_2)\n","\n","first_set_train, first_set_test = torch.utils.data.random_split(first_set, [first_set_train_len, len(en_set_1) - first_set_train_len])\n","second_set_train, second_set_test = torch.utils.data.random_split(second_set, [second_set_train_len, len(en_set_2) - second_set_train_len])"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"fkw5SZx1fTZZ","colab_type":"code","colab":{}},"source":["# Create Dataloaders\n","batch_size = 32\n","\n","loader_first_train = torch.utils.data.DataLoader(first_set_train, batch_size=batch_size, shuffle=True)\n","loader_first_test = torch.utils.data.DataLoader(first_set_test, batch_size=batch_size, shuffle=False)\n","\n","# Ensure same number of batches for second set\n","num_train_batches = len(loader_first_train)\n","num_test_batches = len(loader_first_test)\n","\n","loader_second_train = torch.utils.data.DataLoader(second_set_train, batch_size=batch_size, shuffle=True)\n","loader_second_test = torch.utils.data.DataLoader(second_set_test, batch_size=batch_size, shuffle=False)"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"O7keA4QQ0x0t","colab_type":"code","colab":{}},"source":["# https://stackoverflow.com/questions/53010465/bidirectional-lstm-output-question-in-pytorch\n","# Hyper Parameters\n","embed_dim = 300\n","lstm_hidden_dim = 256\n","num_layers = 2\n","reg_inputs = 4 * lstm_hidden_dim\n","reg_hid_dim1 = 2 * lstm_hidden_dim\n","reg_hid_dim2 = lstm_hidden_dim\n","learning_rate = 1e-4\n","\n","class LSTM_REG(torch.nn.Module):\n","    def __init__(self):\n","        super(LSTM_REG, self).__init__()\n","        self.en_lstm = torch.nn.GRU(embed_dim, lstm_hidden_dim, num_layers, bidirectional=True, batch_first=True)\n","        self.de_lstm = torch.nn.GRU(embed_dim, lstm_hidden_dim, num_layers, bidirectional=True, batch_first=True)\n","        self.reg = torch.nn.Sequential(\n","                    nn.Dropout(),\n","                    nn.Linear(reg_inputs, reg_hid_dim1),\n","                    nn.ReLU(),\n","                    nn.Linear(reg_hid_dim1, reg_hid_dim2),\n","                    nn.ReLU(),\n","                    nn.Linear(reg_hid_dim2, 1)\n","        )\n","\n","    def forward(self, en_batch, de_batch):\n","        # inputs should be 3D, BATCH, NUM_WORDS, lstm_hidden_dim\n","        en_all_hids, en_last_hid = self.en_lstm(en_batch)\n","        de_all_hids, de_last_hid = self.de_lstm(de_batch)\n","\n","        # using last state of last layer in each direction\n","        en_last_hid_resh = en_last_hid.view(num_layers, 2, en_batch.size(0), lstm_hidden_dim)\n","        en_reg_input_1 = en_last_hid_resh[-1, 0, :, :]\n","        en_reg_input_2 = en_last_hid_resh[-1, 1, :, :]\n","\n","        de_last_hid_resh = de_last_hid.view(num_layers, 2, de_batch.size(0), lstm_hidden_dim)\n","        de_reg_input_1 = de_last_hid_resh[-1, 0, :, :]\n","        de_reg_input_2 = de_last_hid_resh[-1, 1, :, :]\n","\n","        reg_input = torch.cat((en_reg_input_1, en_reg_input_2, de_reg_input_1, de_reg_input_2), dim=-1)\n","\n","        out = self.reg(reg_input)\n","        return out\n","\n"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"tFgnl4bonSGg","colab_type":"code","colab":{}},"source":["# Create model\n","model = LSTM_REG()\n","model = model.to(device)"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"uDNWHcEh0lxP","colab_type":"code","colab":{}},"source":["# Optimiser\n","optimizer = torch.optim.Adam(model.parameters(), lr=learning_rate)"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"M3IepQDap5AW","colab_type":"code","colab":{}},"source":["# Loss function\n","def calc_loss(output, target):\n","  loss = torch.nn.functional.mse_loss(output.squeeze(), target, reduction=\"mean\")\n","  return loss\n","\n","# Test function\n","def eval_target(output, target):\n","    acc = torch.nn.functional.l1_loss(output.squeeze(), target, reduction=\"mean\")\n","    return acc"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"VieQzJimED7Y","colab_type":"code","outputId":"709dd321-8ae3-465c-9b27-54d2a85b474d","executionInfo":{"status":"error","timestamp":1581874536360,"user_tz":0,"elapsed":1123873,"user":{"displayName":"Andy Wang","photoUrl":"","userId":"02776860930356410397"}},"colab":{"base_uri":"https://localhost:8080/","height":1000}},"source":["num_epochs = 100\n","\n","verbose_print = False\n","\n","train_losses = []\n","test_losses = []\n","test_acc = []\n","test_pearson = []\n","\n","for epoch in range(num_epochs):\n","    model.train()\n","    epoch_loss = []\n","    total_steps = len(loader_first_train) + len(loader_second_train)\n","    for idx_1, (en, de, labels) in enumerate(loader_first_train):\n","        model.zero_grad()\n","        en = en.to(device)\n","        de = de.to(device)\n","        labels = labels.to(device)\n","        \n","        output = model(en, de)\n","        loss = calc_loss(output, labels)\n","        loss.backward()\n","        avg_batch_loss = loss.item()\n","        optimizer.step()\n","        epoch_loss.append(avg_batch_loss)\n","        if verbose_print:\n","            print(f\"Epoch {epoch}, Batch {idx_1 + 1} Train Loss: {avg_batch_loss:.4f}\")\n","\n","    for idx_2, (en, de, labels) in enumerate(loader_second_train):\n","        model.zero_grad()\n","        en = en.to(device)\n","        de = de.to(device)\n","        labels = labels.to(device)\n","        \n","        output = model(en, de)\n","        loss = calc_loss(output, labels)\n","        loss.backward()\n","        avg_batch_loss = loss.item()\n","        optimizer.step()\n","        epoch_loss.append(avg_batch_loss)\n","        if verbose_print:\n","            print(f\"Epoch {epoch}, Batch {idx_1 + idx_2 + 2} Train Loss: {avg_batch_loss:.4f}\")\n","    \n","    avg_epoch_loss = sum(epoch_loss) / total_steps\n","    train_losses.append(avg_epoch_loss)\n","    print(f\"Average Train Loss in Epoch {epoch}: {avg_epoch_loss:.4f}\")\n","\n","    # Test on test set\n","\n","    model.eval()\n","    epoch_loss = []\n","    epoch_acc = []\n","\n","    all_outputs = []\n","    all_labels = []\n","\n","    total_steps = len(loader_first_test) + len(loader_second_test)\n","    with torch.no_grad():\n","        for idx_1, (en, de, labels) in enumerate(loader_first_test):\n","            # Record for Pearson\n","            all_labels.extend(labels.tolist())\n","\n","            en = en.to(device)\n","            de = de.to(device)\n","            labels = labels.to(device)\n","            \n","            output = model(en, de)\n","            # Record for Pearson\n","            all_outputs.extend(output.squeeze().tolist())\n","\n","            loss = calc_loss(output, labels)\n","            avg_batch_loss = loss.item()\n","            acc = eval_target(output, labels)\n","            avg_batch_acc= acc.item()\n","            epoch_loss.append(avg_batch_loss)\n","            epoch_acc.append(avg_batch_acc)\n","            if verbose_print:\n","                print(f\"Epoch {epoch}, Batch {idx_1 + 1} Test Loss: {avg_batch_loss:.4f}, Test Acc: {avg_batch_acc:.4f}\")\n","\n","        for idx_2, (en, de, labels) in enumerate(loader_second_test):\n","            en = en.to(device)\n","            de = de.to(device)\n","            labels = labels.to(device)\n","            \n","            output = model(en, de)\n","            loss = calc_loss(output, labels)\n","            avg_batch_loss = loss.item()\n","            acc = eval_target(output, labels)\n","            avg_batch_acc= acc.item()\n","            epoch_loss.append(avg_batch_loss)\n","            epoch_acc.append(avg_batch_acc)\n","            if verbose_print:\n","                print(f\"Epoch {epoch}, Batch {idx_1 + idx_2 + 2} Test Loss: {avg_batch_loss:.4f}, Test Acc: {avg_batch_acc:.4f}\")\n","    \n","    avg_epoch_loss = sum(epoch_loss) / total_steps\n","    avg_epoch_acc = sum(epoch_acc) / total_steps\n","    test_losses.append(avg_epoch_loss)\n","    test_acc.append(avg_epoch_acc)\n","    print(f\"Average Test Loss in Epoch {epoch}: {avg_epoch_loss:.4f}\")\n","    print(f\"Average Test Acc in Epoch {epoch}: {avg_epoch_acc:.4f}\")\n","\n","    # Calc Pearson\n","    pearson = np.corrcoef(all_outputs, all_labels)[0, 1]\n","    test_pearson.append(pearson)\n","    print(f\"Pearson Coeff in Epoch {epoch}: {pearson}\")"],"execution_count":0,"outputs":[{"output_type":"stream","text":["Average Train Loss in Epoch 0: 0.2690\n","Average Test Loss in Epoch 0: 0.7136\n","Average Test Acc in Epoch 0: 0.5206\n","Pearson Coeff: -0.01208018209332845\n","Average Train Loss in Epoch 1: 0.2559\n","Average Test Loss in Epoch 1: 0.7513\n","Average Test Acc in Epoch 1: 0.5242\n","Pearson Coeff: -0.014037274885986096\n","Average Train Loss in Epoch 2: 0.2375\n","Average Test Loss in Epoch 2: 0.7637\n","Average Test Acc in Epoch 2: 0.5307\n","Pearson Coeff: -0.013333832990121417\n","Average Train Loss in Epoch 3: 0.2249\n","Average Test Loss in Epoch 3: 0.7991\n","Average Test Acc in Epoch 3: 0.5414\n","Pearson Coeff: -0.016848432261950708\n","Average Train Loss in Epoch 4: 0.2173\n","Average Test Loss in Epoch 4: 0.7994\n","Average Test Acc in Epoch 4: 0.5454\n","Pearson Coeff: -0.026447537163266872\n","Average Train Loss in Epoch 5: 0.2063\n","Average Test Loss in Epoch 5: 0.8754\n","Average Test Acc in Epoch 5: 0.5898\n","Pearson Coeff: -0.017139690356882415\n","Average Train Loss in Epoch 6: 0.1945\n","Average Test Loss in Epoch 6: 0.7957\n","Average Test Acc in Epoch 6: 0.5410\n","Pearson Coeff: -0.01853516115846188\n","Average Train Loss in Epoch 7: 0.1871\n","Average Test Loss in Epoch 7: 0.9355\n","Average Test Acc in Epoch 7: 0.6226\n","Pearson Coeff: -0.014093725437208068\n","Average Train Loss in Epoch 8: 0.1829\n","Average Test Loss in Epoch 8: 0.8238\n","Average Test Acc in Epoch 8: 0.5677\n","Pearson Coeff: -0.01604121885415707\n","Average Train Loss in Epoch 9: 0.1674\n","Average Test Loss in Epoch 9: 0.8189\n","Average Test Acc in Epoch 9: 0.5718\n","Pearson Coeff: -0.011355667091977497\n","Average Train Loss in Epoch 10: 0.1615\n","Average Test Loss in Epoch 10: 0.7745\n","Average Test Acc in Epoch 10: 0.5505\n","Pearson Coeff: -0.008329896889208035\n","Average Train Loss in Epoch 11: 0.1513\n","Average Test Loss in Epoch 11: 0.7987\n","Average Test Acc in Epoch 11: 0.5516\n","Pearson Coeff: -0.017556706908327738\n","Average Train Loss in Epoch 12: 0.1489\n","Average Test Loss in Epoch 12: 0.7881\n","Average Test Acc in Epoch 12: 0.5565\n","Pearson Coeff: -0.003880381677590462\n","Average Train Loss in Epoch 13: 0.1409\n","Average Test Loss in Epoch 13: 0.8571\n","Average Test Acc in Epoch 13: 0.5778\n","Pearson Coeff: -0.005713221776814097\n","Average Train Loss in Epoch 14: 0.1309\n","Average Test Loss in Epoch 14: 0.7921\n","Average Test Acc in Epoch 14: 0.5487\n","Pearson Coeff: -0.00501958803724375\n","Average Train Loss in Epoch 15: 0.1307\n","Average Test Loss in Epoch 15: 0.8275\n","Average Test Acc in Epoch 15: 0.5650\n","Pearson Coeff: -0.023926045029101766\n","Average Train Loss in Epoch 16: 0.1211\n","Average Test Loss in Epoch 16: 0.8229\n","Average Test Acc in Epoch 16: 0.5650\n","Pearson Coeff: -0.020596314231323434\n","Average Train Loss in Epoch 17: 0.1094\n","Average Test Loss in Epoch 17: 0.8146\n","Average Test Acc in Epoch 17: 0.5623\n","Pearson Coeff: -0.010429241405483387\n","Average Train Loss in Epoch 18: 0.1050\n","Average Test Loss in Epoch 18: 0.8349\n","Average Test Acc in Epoch 18: 0.5729\n","Pearson Coeff: -0.0204639270659814\n","Average Train Loss in Epoch 19: 0.0996\n","Average Test Loss in Epoch 19: 0.8547\n","Average Test Acc in Epoch 19: 0.5764\n","Pearson Coeff: -0.01901544690943162\n","Average Train Loss in Epoch 20: 0.0920\n","Average Test Loss in Epoch 20: 0.9082\n","Average Test Acc in Epoch 20: 0.6088\n","Pearson Coeff: -0.02622327972034507\n","Average Train Loss in Epoch 21: 0.0935\n","Average Test Loss in Epoch 21: 0.8776\n","Average Test Acc in Epoch 21: 0.5976\n","Pearson Coeff: -0.01675181788188873\n","Average Train Loss in Epoch 22: 0.0838\n","Average Test Loss in Epoch 22: 0.8761\n","Average Test Acc in Epoch 22: 0.5968\n","Pearson Coeff: -0.03190381975984247\n","Average Train Loss in Epoch 23: 0.0787\n","Average Test Loss in Epoch 23: 0.8382\n","Average Test Acc in Epoch 23: 0.5853\n","Pearson Coeff: -0.022063239628905395\n","Average Train Loss in Epoch 24: 0.0741\n","Average Test Loss in Epoch 24: 0.9045\n","Average Test Acc in Epoch 24: 0.6039\n","Pearson Coeff: -0.01854973142800878\n","Average Train Loss in Epoch 25: 0.0717\n","Average Test Loss in Epoch 25: 0.8609\n","Average Test Acc in Epoch 25: 0.5972\n","Pearson Coeff: -0.01687700939723388\n","Average Train Loss in Epoch 26: 0.0637\n","Average Test Loss in Epoch 26: 0.8870\n","Average Test Acc in Epoch 26: 0.6139\n","Pearson Coeff: -0.02378837960305664\n","Average Train Loss in Epoch 27: 0.0629\n","Average Test Loss in Epoch 27: 0.8457\n","Average Test Acc in Epoch 27: 0.5856\n","Pearson Coeff: -0.034309769784961655\n","Average Train Loss in Epoch 28: 0.0638\n","Average Test Loss in Epoch 28: 0.9125\n","Average Test Acc in Epoch 28: 0.6183\n","Pearson Coeff: -0.029752787276147673\n","Average Train Loss in Epoch 29: 0.0578\n","Average Test Loss in Epoch 29: 0.8921\n","Average Test Acc in Epoch 29: 0.6114\n","Pearson Coeff: -0.022304642193334532\n","Average Train Loss in Epoch 30: 0.0553\n","Average Test Loss in Epoch 30: 0.8868\n","Average Test Acc in Epoch 30: 0.6040\n","Pearson Coeff: -0.027390578163451976\n","Average Train Loss in Epoch 31: 0.0539\n","Average Test Loss in Epoch 31: 0.8558\n","Average Test Acc in Epoch 31: 0.6176\n","Pearson Coeff: -0.014179041241344429\n","Average Train Loss in Epoch 32: 0.0482\n","Average Test Loss in Epoch 32: 0.9083\n","Average Test Acc in Epoch 32: 0.6237\n","Pearson Coeff: -0.027167569601999727\n","Average Train Loss in Epoch 33: 0.0502\n","Average Test Loss in Epoch 33: 0.8870\n","Average Test Acc in Epoch 33: 0.6193\n","Pearson Coeff: -0.017429815512140817\n","Average Train Loss in Epoch 34: 0.0483\n","Average Test Loss in Epoch 34: 0.8662\n","Average Test Acc in Epoch 34: 0.6139\n","Pearson Coeff: -0.01623911005137501\n","Average Train Loss in Epoch 35: 0.0453\n","Average Test Loss in Epoch 35: 0.8848\n","Average Test Acc in Epoch 35: 0.6119\n","Pearson Coeff: -0.03112387984537785\n","Average Train Loss in Epoch 36: 0.0421\n","Average Test Loss in Epoch 36: 0.9268\n","Average Test Acc in Epoch 36: 0.6283\n","Pearson Coeff: -0.027083975064787838\n","Average Train Loss in Epoch 37: 0.0422\n","Average Test Loss in Epoch 37: 0.8710\n","Average Test Acc in Epoch 37: 0.6129\n","Pearson Coeff: -0.03464512705192923\n","Average Train Loss in Epoch 38: 0.0418\n","Average Test Loss in Epoch 38: 0.8755\n","Average Test Acc in Epoch 38: 0.6152\n","Pearson Coeff: -0.037708141820807395\n","Average Train Loss in Epoch 39: 0.0386\n","Average Test Loss in Epoch 39: 0.8680\n","Average Test Acc in Epoch 39: 0.6086\n","Pearson Coeff: -0.022985572295723674\n","Average Train Loss in Epoch 40: 0.0371\n","Average Test Loss in Epoch 40: 0.8996\n","Average Test Acc in Epoch 40: 0.6363\n","Pearson Coeff: -0.022588057477465104\n","Average Train Loss in Epoch 41: 0.0349\n","Average Test Loss in Epoch 41: 0.8588\n","Average Test Acc in Epoch 41: 0.6099\n","Pearson Coeff: -0.03136384672355856\n"],"name":"stdout"},{"output_type":"error","ename":"KeyboardInterrupt","evalue":"ignored","traceback":["\u001b[0;31m---------------------------------------------------------------------------\u001b[0m","\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)","\u001b[0;32m<ipython-input-28-9f678d04bb12>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     22\u001b[0m         \u001b[0mloss\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mcalc_loss\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0moutput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlabels\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     23\u001b[0m         \u001b[0mloss\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mbackward\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---> 24\u001b[0;31m         \u001b[0mavg_batch_loss\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mloss\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[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     25\u001b[0m         \u001b[0moptimizer\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mstep\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     26\u001b[0m         \u001b[0mepoch_loss\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mavg_batch_loss\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;31mKeyboardInterrupt\u001b[0m: "]}]},{"cell_type":"code","metadata":{"id":"OV-ZxbuwBw76","colab_type":"code","outputId":"ecafd43b-f529-40fc-bf11-be7890a91bef","executionInfo":{"status":"error","timestamp":1581875259505,"user_tz":0,"elapsed":857,"user":{"displayName":"Andy Wang","photoUrl":"","userId":"02776860930356410397"}},"colab":{"base_uri":"https://localhost:8080/","height":1000}},"source":["# plot stats\n","plt.plot(range(len(train_losses)), train_losses)\n","plt.xlabel(\"Epochs\")\n","plt.title('Train losses')\n","plt.show()\n","\n","plt.plot(range(len(test_losses)), test_losses)\n","plt.xlabel(\"Epochs\")\n","plt.title('Test losses')\n","plt.show()\n","\n","plt.plot(range(len(test_acc)), test_acc)\n","plt.xlabel(\"Epochs\")\n","plt.title('Test acc')\n","plt.show()\n","\n","plt.plot(range(len(test_pearson)), test_pearson)\n","plt.xlabel(\"Epochs\")\n","plt.title('Test pearson')\n","plt.show()\n"],"execution_count":0,"outputs":[{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAXoAAAEWCAYAAABollyxAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[]}},{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAXoAAAEWCAYAAABollyxAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRU5ErkJggg==\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[]}},{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAXoAAAEWCAYAAABollyxAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n/wHJerCbVYBcXAAAAABJRU5ErkJggg==\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[]}},{"output_type":"error","ename":"NameError","evalue":"ignored","traceback":["\u001b[0;31m---------------------------------------------------------------------------\u001b[0m","\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)","\u001b[0;32m<ipython-input-30-b6aad75e9c73>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     14\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshow\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     15\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 16\u001b[0;31m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mplot\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtest_pearson\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtest_pearson\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     17\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mxlabel\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Epochs\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     18\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtitle\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'Test pearson'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;31mNameError\u001b[0m: name 'test_pearson' is not defined"]}]}]}