Datasets:

---

bigcode

/

jupyter-parsed

like 4

Follow

BigCode 1.34k

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import requests", "_____no_output_____" ], [ "requests.get(\"http://127.0.0.1:5000/API/all_data\").json()", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "Data can be found: https://data-seattlecitygis.opendata.arcgis.com/search?collection=Dataset&modified=2021-01-01%2C2021-10-13", "_____no_output_____" ] ], [ [ "import geopandas as gpd", "_____no_output_____" ], [ "gpd.datasets.available", "_____no_output_____" ], [ "world = gpd.read_file(\n gpd.datasets.get_path('naturalearth_lowres')\n)", "_____no_output_____" ], [ "seattle_zoning_2035 = gpd.read_file('Future_Land_Use__2035.shp')\nseattle_zoning_2035.head()", "_____no_output_____" ], [ "education_centers = gpd.read_file('data/Environmental_Education_Centers.shp')\neducation_centers", "_____no_output_____" ] ] ]

[ "markdown", "code" ]

[ [ "markdown" ], [ "code", "code", "code", "code", "code" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "# Advanced Logistic Regression in TensorFlow 2.0 \n\n\n\n## Learning Objectives\n\n1. Load a CSV file using Pandas\n2. Create train, validation, and test sets\n3. Define and train a model using Keras (including setting class weights)\n4. Evaluate the model using various metrics (including precision and recall)\n5. Try common techniques for dealing with imbalanced data:\n Class weighting and\n Oversampling\n\n\n\n## Introduction \nThis lab how to classify a highly imbalanced dataset in which the number of examples in one class greatly outnumbers the examples in another. You will work with the [Credit Card Fraud Detection](https://www.kaggle.com/mlg-ulb/creditcardfraud) dataset hosted on Kaggle. The aim is to detect a mere 492 fraudulent transactions from 284,807 transactions in total. You will use [Keras](../../guide/keras/overview.ipynb) to define the model and [class weights](https://www.tensorflow.org/versions/r2.0/api_docs/python/tf/keras/Model) to help the model learn from the imbalanced data. \n\nPENDING LINK UPDATE: Each learning objective will correspond to a __#TODO__ in the [student lab notebook](https://training-data-analyst/courses/machine_learning/deepdive2/image_classification/labs/5_fashion_mnist_class.ipynb) -- try to complete that notebook first before reviewing this solution notebook.", "_____no_output_____" ], [ "Start by importing the necessary libraries for this lab.", "_____no_output_____" ] ], [ [ "import tensorflow as tf\nfrom tensorflow import keras\n\nimport os\nimport tempfile\n\nimport matplotlib as mpl\nimport matplotlib.pyplot as plt\nimport numpy as np\nimport pandas as pd\nimport seaborn as sns\n\nimport sklearn\nfrom sklearn.metrics import confusion_matrix\nfrom sklearn.model_selection import train_test_split\nfrom sklearn.preprocessing import StandardScaler\n\nprint(\"TensorFlow version: \",tf.version.VERSION)", "TensorFlow version: 2.1.0\n" ] ], [ [ "In the next cell, we're going to customize our Matplot lib visualization figure size and colors. Note that each time Matplotlib loads, it defines a runtime configuration (rc) containing the default styles for every plot element we create. This configuration can be adjusted at any time using the plt.rc convenience routine. ", "_____no_output_____" ] ], [ [ "mpl.rcParams['figure.figsize'] = (12, 10)\ncolors = plt.rcParams['axes.prop_cycle'].by_key()['color']", "_____no_output_____" ] ], [ [ "## Data processing and exploration", "_____no_output_____" ], [ "### Download the Kaggle Credit Card Fraud data set\n\nPandas is a Python library with many helpful utilities for loading and working with structured data and can be used to download CSVs into a dataframe.\n\nNote: This dataset has been collected and analysed during a research collaboration of Worldline and the [Machine Learning Group](http://mlg.ulb.ac.be) of ULB (Université Libre de Bruxelles) on big data mining and fraud detection. More details on current and past projects on related topics are available [here](https://www.researchgate.net/project/Fraud-detection-5) and the page of the [DefeatFraud](https://mlg.ulb.ac.be/wordpress/portfolio_page/defeatfraud-assessment-and-validation-of-deep-feature-engineering-and-learning-solutions-for-fraud-detection/) project", "_____no_output_____" ] ], [ [ "file = tf.keras.utils\nraw_df = pd.read_csv('https://storage.googleapis.com/download.tensorflow.org/data/creditcard.csv')\nraw_df.head()", "_____no_output_____" ] ], [ [ "Now, let's view the statistics of the raw dataframe.", "_____no_output_____" ] ], [ [ "raw_df[['Time', 'V1', 'V2', 'V3', 'V4', 'V5', 'V26', 'V27', 'V28', 'Amount', 'Class']].describe()", "_____no_output_____" ] ], [ [ "### Examine the class label imbalance\n\nLet's look at the dataset imbalance:", "_____no_output_____" ] ], [ [ "neg, pos = np.bincount(raw_df['Class'])\ntotal = neg + pos\nprint('Examples:\\n Total: {}\\n Positive: {} ({:.2f}% of total)\\n'.format(\n total, pos, 100 * pos / total))", "Examples:\n Total: 284807\n Positive: 492 (0.17% of total)\n\n" ] ], [ [ "This shows the small fraction of positive samples.", "_____no_output_____" ], [ "### Clean, split and normalize the data\n\nThe raw data has a few issues. First the `Time` and `Amount` columns are too variable to use directly. Drop the `Time` column (since it's not clear what it means) and take the log of the `Amount` column to reduce its range.", "_____no_output_____" ] ], [ [ "cleaned_df = raw_df.copy()\n\n# You don't want the `Time` column.\ncleaned_df.pop('Time')\n\n# The `Amount` column covers a huge range. Convert to log-space.\neps=0.001 # 0 => 0.1¢\ncleaned_df['Log Ammount'] = np.log(cleaned_df.pop('Amount')+eps)", "_____no_output_____" ] ], [ [ "Split the dataset into train, validation, and test sets. The validation set is used during the model fitting to evaluate the loss and any metrics, however the model is not fit with this data. The test set is completely unused during the training phase and is only used at the end to evaluate how well the model generalizes to new data. This is especially important with imbalanced datasets where [overfitting](https://developers.google.com/machine-learning/crash-course/generalization/peril-of-overfitting) is a significant concern from the lack of training data.", "_____no_output_____" ] ], [ [ "# Use a utility from sklearn to split and shuffle our dataset.\ntrain_df, test_df = train_test_split(cleaned_df, test_size=0.2)\ntrain_df, val_df = train_test_split(train_df, test_size=0.2)\n\n# Form np arrays of labels and features.\ntrain_labels = np.array(train_df.pop('Class'))\nbool_train_labels = train_labels != 0\nval_labels = np.array(val_df.pop('Class'))\ntest_labels = np.array(test_df.pop('Class'))\n\ntrain_features = np.array(train_df)\nval_features = np.array(val_df)\ntest_features = np.array(test_df)", "_____no_output_____" ] ], [ [ "Normalize the input features using the sklearn StandardScaler.\nThis will set the mean to 0 and standard deviation to 1.\n\nNote: The `StandardScaler` is only fit using the `train_features` to be sure the model is not peeking at the validation or test sets. ", "_____no_output_____" ] ], [ [ "scaler = StandardScaler()\ntrain_features = scaler.fit_transform(train_features)\n\nval_features = scaler.transform(val_features)\ntest_features = scaler.transform(test_features)\n\ntrain_features = np.clip(train_features, -5, 5)\nval_features = np.clip(val_features, -5, 5)\ntest_features = np.clip(test_features, -5, 5)\n\n\nprint('Training labels shape:', train_labels.shape)\nprint('Validation labels shape:', val_labels.shape)\nprint('Test labels shape:', test_labels.shape)\n\nprint('Training features shape:', train_features.shape)\nprint('Validation features shape:', val_features.shape)\nprint('Test features shape:', test_features.shape)\n", "Training labels shape: (182276,)\nValidation labels shape: (45569,)\nTest labels shape: (56962,)\nTraining features shape: (182276, 29)\nValidation features shape: (45569, 29)\nTest features shape: (56962, 29)\n" ] ], [ [ "Caution: If you want to deploy a model, it's critical that you preserve the preprocessing calculations. The easiest way to implement them as layers, and attach them to your model before export.\n", "_____no_output_____" ], [ "### Look at the data distribution\n\nNext compare the distributions of the positive and negative examples over a few features. Good questions to ask yourself at this point are:\n\n* Do these distributions make sense? \n * Yes. You've normalized the input and these are mostly concentrated in the `+/- 2` range.\n* Can you see the difference between the ditributions?\n * Yes the positive examples contain a much higher rate of extreme values.", "_____no_output_____" ] ], [ [ "pos_df = pd.DataFrame(train_features[ bool_train_labels], columns = train_df.columns)\nneg_df = pd.DataFrame(train_features[~bool_train_labels], columns = train_df.columns)\n\nsns.jointplot(pos_df['V5'], pos_df['V6'],\n kind='hex', xlim = (-5,5), ylim = (-5,5))\nplt.suptitle(\"Positive distribution\")\n\nsns.jointplot(neg_df['V5'], neg_df['V6'],\n kind='hex', xlim = (-5,5), ylim = (-5,5))\n_ = plt.suptitle(\"Negative distribution\")", "_____no_output_____" ] ], [ [ "## Define the model and metrics\n\nDefine a function that creates a simple neural network with a densly connected hidden layer, a [dropout](https://developers.google.com/machine-learning/glossary/#dropout_regularization) layer to reduce overfitting, and an output sigmoid layer that returns the probability of a transaction being fraudulent: ", "_____no_output_____" ] ], [ [ "METRICS = [\n keras.metrics.TruePositives(name='tp'),\n keras.metrics.FalsePositives(name='fp'),\n keras.metrics.TrueNegatives(name='tn'),\n keras.metrics.FalseNegatives(name='fn'), \n keras.metrics.BinaryAccuracy(name='accuracy'),\n keras.metrics.Precision(name='precision'),\n keras.metrics.Recall(name='recall'),\n keras.metrics.AUC(name='auc'),\n]\n\ndef make_model(metrics = METRICS, output_bias=None):\n if output_bias is not None:\n output_bias = tf.keras.initializers.Constant(output_bias)\n model = keras.Sequential([\n keras.layers.Dense(\n 16, activation='relu',\n input_shape=(train_features.shape[-1],)),\n keras.layers.Dropout(0.5),\n keras.layers.Dense(1, activation='sigmoid',\n bias_initializer=output_bias),\n ])\n\n model.compile(\n optimizer=keras.optimizers.Adam(lr=1e-3),\n loss=keras.losses.BinaryCrossentropy(),\n metrics=metrics)\n\n return model", "_____no_output_____" ] ], [ [ "### Understanding useful metrics\n\nNotice that there are a few metrics defined above that can be computed by the model that will be helpful when evaluating the performance.\n\n\n\n* **False** negatives and **false** positives are samples that were **incorrectly** classified\n* **True** negatives and **true** positives are samples that were **correctly** classified\n* **Accuracy** is the percentage of examples correctly classified\n> $\\frac{\\text{true samples}}{\\text{total samples}}$\n* **Precision** is the percentage of **predicted** positives that were correctly classified\n> $\\frac{\\text{true positives}}{\\text{true positives + false positives}}$\n* **Recall** is the percentage of **actual** positives that were correctly classified\n> $\\frac{\\text{true positives}}{\\text{true positives + false negatives}}$\n* **AUC** refers to the Area Under the Curve of a Receiver Operating Characteristic curve (ROC-AUC). This metric is equal to the probability that a classifier will rank a random positive sample higher than than a random negative sample.\n\nNote: Accuracy is not a helpful metric for this task. You can 99.8%+ accuracy on this task by predicting False all the time. \n\nRead more:\n* [True vs. False and Positive vs. Negative](https://developers.google.com/machine-learning/crash-course/classification/true-false-positive-negative)\n* [Accuracy](https://developers.google.com/machine-learning/crash-course/classification/accuracy)\n* [Precision and Recall](https://developers.google.com/machine-learning/crash-course/classification/precision-and-recall)\n* [ROC-AUC](https://developers.google.com/machine-learning/crash-course/classification/roc-and-auc)", "_____no_output_____" ], [ "## Baseline model", "_____no_output_____" ], [ "### Build the model\n\nNow create and train your model using the function that was defined earlier. Notice that the model is fit using a larger than default batch size of 2048, this is important to ensure that each batch has a decent chance of containing a few positive samples. If the batch size was too small, they would likely have no fraudulent transactions to learn from.\n\n\nNote: this model will not handle the class imbalance well. You will improve it later in this tutorial.", "_____no_output_____" ] ], [ [ "EPOCHS = 100\nBATCH_SIZE = 2048\n\nearly_stopping = tf.keras.callbacks.EarlyStopping(\n monitor='val_auc', \n verbose=1,\n patience=10,\n mode='max',\n restore_best_weights=True)", "_____no_output_____" ], [ "model = make_model()\nmodel.summary()", "Model: \"sequential_8\"\n_________________________________________________________________\nLayer (type) Output Shape Param # \n=================================================================\ndense_16 (Dense) (None, 16) 480 \n_________________________________________________________________\ndropout_8 (Dropout) (None, 16) 0 \n_________________________________________________________________\ndense_17 (Dense) (None, 1) 17 \n=================================================================\nTotal params: 497\nTrainable params: 497\nNon-trainable params: 0\n_________________________________________________________________\n" ] ], [ [ "Test run the model:", "_____no_output_____" ] ], [ [ "model.predict(train_features[:10])", "_____no_output_____" ] ], [ [ "### Optional: Set the correct initial bias.", "_____no_output_____" ], [ "These are initial guesses are not great. You know the dataset is imbalanced. Set the output layer's bias to reflect that (See: [A Recipe for Training Neural Networks: \"init well\"](http://karpathy.github.io/2019/04/25/recipe/#2-set-up-the-end-to-end-trainingevaluation-skeleton--get-dumb-baselines)). This can help with initial convergence.", "_____no_output_____" ], [ "With the default bias initialization the loss should be about `math.log(2) = 0.69314` ", "_____no_output_____" ] ], [ [ "results = model.evaluate(train_features, train_labels, batch_size=BATCH_SIZE, verbose=0)\nprint(\"Loss: {:0.4f}\".format(results[0]))", "Loss: 1.7441\n" ] ], [ [ "The correct bias to set can be derived from:\n\n$$ p_0 = pos/(pos + neg) = 1/(1+e^{-b_0}) $$\n$$ b_0 = -log_e(1/p_0 - 1) $$\n$$ b_0 = log_e(pos/neg)$$", "_____no_output_____" ] ], [ [ "initial_bias = np.log([pos/neg])\ninitial_bias", "_____no_output_____" ] ], [ [ "Set that as the initial bias, and the model will give much more reasonable initial guesses. \n\nIt should be near: `pos/total = 0.0018`", "_____no_output_____" ] ], [ [ "model = make_model(output_bias = initial_bias)\nmodel.predict(train_features[:10])", "_____no_output_____" ] ], [ [ "With this initialization the initial loss should be approximately:\n\n$$-p_0log(p_0)-(1-p_0)log(1-p_0) = 0.01317$$", "_____no_output_____" ] ], [ [ "results = model.evaluate(train_features, train_labels, batch_size=BATCH_SIZE, verbose=0)\nprint(\"Loss: {:0.4f}\".format(results[0]))", "Loss: 0.0275\n" ] ], [ [ "This initial loss is about 50 times less than if would have been with naive initilization.\n\nThis way the model doesn't need to spend the first few epochs just learning that positive examples are unlikely. This also makes it easier to read plots of the loss during training.", "_____no_output_____" ], [ "### Checkpoint the initial weights\n\nTo make the various training runs more comparable, keep this initial model's weights in a checkpoint file, and load them into each model before training.", "_____no_output_____" ] ], [ [ "initial_weights = os.path.join(tempfile.mkdtemp(),'initial_weights')\nmodel.save_weights(initial_weights)", "_____no_output_____" ] ], [ [ "### Confirm that the bias fix helps\n\nBefore moving on, confirm quick that the careful bias initialization actually helped.\n\nTrain the model for 20 epochs, with and without this careful initialization, and compare the losses: ", "_____no_output_____" ] ], [ [ "model = make_model()\nmodel.load_weights(initial_weights)\nmodel.layers[-1].bias.assign([0.0])\nzero_bias_history = model.fit(\n train_features,\n train_labels,\n batch_size=BATCH_SIZE,\n epochs=20,\n validation_data=(val_features, val_labels), \n verbose=0)", "_____no_output_____" ], [ "model = make_model()\nmodel.load_weights(initial_weights)\ncareful_bias_history = model.fit(\n train_features,\n train_labels,\n batch_size=BATCH_SIZE,\n epochs=20,\n validation_data=(val_features, val_labels), \n verbose=0)", "_____no_output_____" ], [ "def plot_loss(history, label, n):\n # Use a log scale to show the wide range of values.\n plt.semilogy(history.epoch, history.history['loss'],\n color=colors[n], label='Train '+label)\n plt.semilogy(history.epoch, history.history['val_loss'],\n color=colors[n], label='Val '+label,\n linestyle=\"--\")\n plt.xlabel('Epoch')\n plt.ylabel('Loss')\n \n plt.legend()", "_____no_output_____" ], [ "plot_loss(zero_bias_history, \"Zero Bias\", 0)\nplot_loss(careful_bias_history, \"Careful Bias\", 1)", "_____no_output_____" ] ], [ [ "The above figure makes it clear: In terms of validation loss, on this problem, this careful initialization gives a clear advantage. ", "_____no_output_____" ], [ "### Train the model", "_____no_output_____" ] ], [ [ "model = make_model()\nmodel.load_weights(initial_weights)\nbaseline_history = model.fit(\n train_features,\n train_labels,\n batch_size=BATCH_SIZE,\n epochs=EPOCHS,\n callbacks = [early_stopping],\n validation_data=(val_features, val_labels))", "Train on 182276 samples, validate on 45569 samples\nEpoch 1/100\n182276/182276 [==============================] - 3s 16us/sample - loss: 0.0256 - tp: 64.0000 - fp: 745.0000 - tn: 181227.0000 - fn: 240.0000 - accuracy: 0.9946 - precision: 0.0791 - recall: 0.2105 - auc: 0.8031 - val_loss: 0.0079 - val_tp: 17.0000 - val_fp: 7.0000 - val_tn: 45479.0000 - val_fn: 66.0000 - val_accuracy: 0.9984 - val_precision: 0.7083 - val_recall: 0.2048 - val_auc: 0.9377\nEpoch 2/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.0100 - tp: 111.0000 - fp: 131.0000 - tn: 181841.0000 - fn: 193.0000 - accuracy: 0.9982 - precision: 0.4587 - recall: 0.3651 - auc: 0.8758 - val_loss: 0.0056 - val_tp: 40.0000 - val_fp: 7.0000 - val_tn: 45479.0000 - val_fn: 43.0000 - val_accuracy: 0.9989 - val_precision: 0.8511 - val_recall: 0.4819 - val_auc: 0.9422\nEpoch 3/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.0075 - tp: 148.0000 - fp: 57.0000 - tn: 181915.0000 - fn: 156.0000 - accuracy: 0.9988 - precision: 0.7220 - recall: 0.4868 - auc: 0.9206 - val_loss: 0.0048 - val_tp: 52.0000 - val_fp: 7.0000 - val_tn: 45479.0000 - val_fn: 31.0000 - val_accuracy: 0.9992 - val_precision: 0.8814 - val_recall: 0.6265 - val_auc: 0.9382\nEpoch 4/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.0065 - tp: 157.0000 - fp: 48.0000 - tn: 181924.0000 - fn: 147.0000 - accuracy: 0.9989 - precision: 0.7659 - recall: 0.5164 - auc: 0.9210 - val_loss: 0.0045 - val_tp: 52.0000 - val_fp: 7.0000 - val_tn: 45479.0000 - val_fn: 31.0000 - val_accuracy: 0.9992 - val_precision: 0.8814 - val_recall: 0.6265 - val_auc: 0.9387\nEpoch 5/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.0058 - tp: 172.0000 - fp: 43.0000 - tn: 181929.0000 - fn: 132.0000 - accuracy: 0.9990 - precision: 0.8000 - recall: 0.5658 - auc: 0.9246 - val_loss: 0.0042 - val_tp: 51.0000 - val_fp: 7.0000 - val_tn: 45479.0000 - val_fn: 32.0000 - val_accuracy: 0.9991 - val_precision: 0.8793 - val_recall: 0.6145 - val_auc: 0.9390\nEpoch 6/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.0054 - tp: 169.0000 - fp: 28.0000 - tn: 181944.0000 - fn: 135.0000 - accuracy: 0.9991 - precision: 0.8579 - recall: 0.5559 - auc: 0.9210 - val_loss: 0.0039 - val_tp: 56.0000 - val_fp: 7.0000 - val_tn: 45479.0000 - val_fn: 27.0000 - val_accuracy: 0.9993 - val_precision: 0.8889 - val_recall: 0.6747 - val_auc: 0.9391\nEpoch 7/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.0054 - tp: 167.0000 - fp: 33.0000 - tn: 181939.0000 - fn: 137.0000 - accuracy: 0.9991 - precision: 0.8350 - recall: 0.5493 - auc: 0.9224 - val_loss: 0.0038 - val_tp: 60.0000 - val_fp: 7.0000 - val_tn: 45479.0000 - val_fn: 23.0000 - val_accuracy: 0.9993 - val_precision: 0.8955 - val_recall: 0.7229 - val_auc: 0.9392\nEpoch 8/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.0050 - tp: 182.0000 - fp: 28.0000 - tn: 181944.0000 - fn: 122.0000 - accuracy: 0.9992 - precision: 0.8667 - recall: 0.5987 - auc: 0.9215 - val_loss: 0.0038 - val_tp: 62.0000 - val_fp: 7.0000 - val_tn: 45479.0000 - val_fn: 21.0000 - val_accuracy: 0.9994 - val_precision: 0.8986 - val_recall: 0.7470 - val_auc: 0.9332\nEpoch 9/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.0047 - tp: 186.0000 - fp: 36.0000 - tn: 181936.0000 - fn: 118.0000 - accuracy: 0.9992 - precision: 0.8378 - recall: 0.6118 - auc: 0.9238 - val_loss: 0.0036 - val_tp: 63.0000 - val_fp: 7.0000 - val_tn: 45479.0000 - val_fn: 20.0000 - val_accuracy: 0.9994 - val_precision: 0.9000 - val_recall: 0.7590 - val_auc: 0.9332\nEpoch 10/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.0048 - tp: 176.0000 - fp: 33.0000 - tn: 181939.0000 - fn: 128.0000 - accuracy: 0.9991 - precision: 0.8421 - recall: 0.5789 - auc: 0.9208 - val_loss: 0.0036 - val_tp: 63.0000 - val_fp: 7.0000 - val_tn: 45479.0000 - val_fn: 20.0000 - val_accuracy: 0.9994 - val_precision: 0.9000 - val_recall: 0.7590 - val_auc: 0.9332\nEpoch 11/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.0045 - tp: 180.0000 - fp: 32.0000 - tn: 181940.0000 - fn: 124.0000 - accuracy: 0.9991 - precision: 0.8491 - recall: 0.5921 - auc: 0.9341 - val_loss: 0.0035 - val_tp: 64.0000 - val_fp: 7.0000 - val_tn: 45479.0000 - val_fn: 19.0000 - val_accuracy: 0.9994 - val_precision: 0.9014 - val_recall: 0.7711 - val_auc: 0.9331\nEpoch 12/100\n169984/182276 [==========================>...] - ETA: 0s - loss: 0.0045 - tp: 175.0000 - fp: 30.0000 - tn: 169674.0000 - fn: 105.0000 - accuracy: 0.9992 - precision: 0.8537 - recall: 0.6250 - auc: 0.9306Restoring model weights from the end of the best epoch.\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.0045 - tp: 188.0000 - fp: 31.0000 - tn: 181941.0000 - fn: 116.0000 - accuracy: 0.9992 - precision: 0.8584 - recall: 0.6184 - auc: 0.9326 - val_loss: 0.0034 - val_tp: 63.0000 - val_fp: 6.0000 - val_tn: 45480.0000 - val_fn: 20.0000 - val_accuracy: 0.9994 - val_precision: 0.9130 - val_recall: 0.7590 - val_auc: 0.9332\nEpoch 00012: early stopping\n" ] ], [ [ "### Check training history\nIn this section, you will produce plots of your model's accuracy and loss on the training and validation set. These are useful to check for overfitting, which you can learn more about in this [tutorial](https://www.tensorflow.org/tutorials/keras/overfit_and_underfit).\n\nAdditionally, you can produce these plots for any of the metrics you created above. False negatives are included as an example.", "_____no_output_____" ] ], [ [ "def plot_metrics(history):\n metrics = ['loss', 'auc', 'precision', 'recall']\n for n, metric in enumerate(metrics):\n name = metric.replace(\"_\",\" \").capitalize()\n plt.subplot(2,2,n+1)\n plt.plot(history.epoch, history.history[metric], color=colors[0], label='Train')\n plt.plot(history.epoch, history.history['val_'+metric],\n color=colors[0], linestyle=\"--\", label='Val')\n plt.xlabel('Epoch')\n plt.ylabel(name)\n if metric == 'loss':\n plt.ylim([0, plt.ylim()[1]])\n elif metric == 'auc':\n plt.ylim([0.8,1])\n else:\n plt.ylim([0,1])\n\n plt.legend()\n", "_____no_output_____" ], [ "plot_metrics(baseline_history)", "_____no_output_____" ] ], [ [ "Note: That the validation curve generally performs better than the training curve. This is mainly caused by the fact that the dropout layer is not active when evaluating the model.", "_____no_output_____" ], [ "### Evaluate metrics\n\nYou can use a [confusion matrix](https://developers.google.com/machine-learning/glossary/#confusion_matrix) to summarize the actual vs. predicted labels where the X axis is the predicted label and the Y axis is the actual label.", "_____no_output_____" ] ], [ [ "train_predictions_baseline = model.predict(train_features, batch_size=BATCH_SIZE)\ntest_predictions_baseline = model.predict(test_features, batch_size=BATCH_SIZE)", "_____no_output_____" ], [ "def plot_cm(labels, predictions, p=0.5):\n cm = confusion_matrix(labels, predictions > p)\n plt.figure(figsize=(5,5))\n sns.heatmap(cm, annot=True, fmt=\"d\")\n plt.title('Confusion matrix @{:.2f}'.format(p))\n plt.ylabel('Actual label')\n plt.xlabel('Predicted label')\n\n print('Legitimate Transactions Detected (True Negatives): ', cm[0][0])\n print('Legitimate Transactions Incorrectly Detected (False Positives): ', cm[0][1])\n print('Fraudulent Transactions Missed (False Negatives): ', cm[1][0])\n print('Fraudulent Transactions Detected (True Positives): ', cm[1][1])\n print('Total Fraudulent Transactions: ', np.sum(cm[1]))", "_____no_output_____" ] ], [ [ "Evaluate your model on the test dataset and display the results for the metrics you created above.", "_____no_output_____" ] ], [ [ "baseline_results = model.evaluate(test_features, test_labels,\n batch_size=BATCH_SIZE, verbose=0)\nfor name, value in zip(model.metrics_names, baseline_results):\n print(name, ': ', value)\nprint()\n\nplot_cm(test_labels, test_predictions_baseline)", "loss : 0.005941324691873794\ntp : 55.0\nfp : 12.0\ntn : 56845.0\nfn : 50.0\naccuracy : 0.99891156\nprecision : 0.8208955\nrecall : 0.52380955\nauc : 0.9390888\n\nLegitimate Transactions Detected (True Negatives): 56845\nLegitimate Transactions Incorrectly Detected (False Positives): 12\nFraudulent Transactions Missed (False Negatives): 50\nFraudulent Transactions Detected (True Positives): 55\nTotal Fraudulent Transactions: 105\n" ] ], [ [ "If the model had predicted everything perfectly, this would be a [diagonal matrix](https://en.wikipedia.org/wiki/Diagonal_matrix) where values off the main diagonal, indicating incorrect predictions, would be zero. In this case the matrix shows that you have relatively few false positives, meaning that there were relatively few legitimate transactions that were incorrectly flagged. However, you would likely want to have even fewer false negatives despite the cost of increasing the number of false positives. This trade off may be preferable because false negatives would allow fraudulent transactions to go through, whereas false positives may cause an email to be sent to a customer to ask them to verify their card activity.", "_____no_output_____" ], [ "### Plot the ROC\n\nNow plot the [ROC](https://developers.google.com/machine-learning/glossary#ROC). This plot is useful because it shows, at a glance, the range of performance the model can reach just by tuning the output threshold.", "_____no_output_____" ] ], [ [ "def plot_roc(name, labels, predictions, **kwargs):\n fp, tp, _ = sklearn.metrics.roc_curve(labels, predictions)\n\n plt.plot(100*fp, 100*tp, label=name, linewidth=2, **kwargs)\n plt.xlabel('False positives [%]')\n plt.ylabel('True positives [%]')\n plt.xlim([-0.5,20])\n plt.ylim([80,100.5])\n plt.grid(True)\n ax = plt.gca()\n ax.set_aspect('equal')", "_____no_output_____" ], [ "plot_roc(\"Train Baseline\", train_labels, train_predictions_baseline, color=colors[0])\nplot_roc(\"Test Baseline\", test_labels, test_predictions_baseline, color=colors[0], linestyle='--')\nplt.legend(loc='lower right')", "_____no_output_____" ] ], [ [ "It looks like the precision is relatively high, but the recall and the area under the ROC curve (AUC) aren't as high as you might like. Classifiers often face challenges when trying to maximize both precision and recall, which is especially true when working with imbalanced datasets. It is important to consider the costs of different types of errors in the context of the problem you care about. In this example, a false negative (a fraudulent transaction is missed) may have a financial cost, while a false positive (a transaction is incorrectly flagged as fraudulent) may decrease user happiness.", "_____no_output_____" ], [ "## Class weights", "_____no_output_____" ], [ "### Calculate class weights\n\nThe goal is to identify fradulent transactions, but you don't have very many of those positive samples to work with, so you would want to have the classifier heavily weight the few examples that are available. You can do this by passing Keras weights for each class through a parameter. These will cause the model to \"pay more attention\" to examples from an under-represented class.", "_____no_output_____" ] ], [ [ "# Scaling by total/2 helps keep the loss to a similar magnitude.\n# The sum of the weights of all examples stays the same.\nweight_for_0 = (1 / neg)*(total)/2.0 \nweight_for_1 = (1 / pos)*(total)/2.0\n\nclass_weight = {0: weight_for_0, 1: weight_for_1}\n\nprint('Weight for class 0: {:.2f}'.format(weight_for_0))\nprint('Weight for class 1: {:.2f}'.format(weight_for_1))", "Weight for class 0: 0.50\nWeight for class 1: 289.44\n" ] ], [ [ "### Train a model with class weights\n\nNow try re-training and evaluating the model with class weights to see how that affects the predictions.\n\nNote: Using `class_weights` changes the range of the loss. This may affect the stability of the training depending on the optimizer. Optimizers whose step size is dependent on the magnitude of the gradient, like `optimizers.SGD`, may fail. The optimizer used here, `optimizers.Adam`, is unaffected by the scaling change. Also note that because of the weighting, the total losses are not comparable between the two models.", "_____no_output_____" ] ], [ [ "weighted_model = make_model()\nweighted_model.load_weights(initial_weights)\n\nweighted_history = weighted_model.fit(\n train_features,\n train_labels,\n batch_size=BATCH_SIZE,\n epochs=EPOCHS,\n callbacks = [early_stopping],\n validation_data=(val_features, val_labels),\n # The class weights go here\n class_weight=class_weight) ", "WARNING:tensorflow:sample_weight modes were coerced from\n ...\n to \n ['...']\nWARNING:tensorflow:sample_weight modes were coerced from\n ...\n to \n ['...']\nTrain on 182276 samples, validate on 45569 samples\nEpoch 1/100\n182276/182276 [==============================] - 3s 19us/sample - loss: 1.0524 - tp: 138.0000 - fp: 2726.0000 - tn: 179246.0000 - fn: 166.0000 - accuracy: 0.9841 - precision: 0.0482 - recall: 0.4539 - auc: 0.8321 - val_loss: 0.4515 - val_tp: 59.0000 - val_fp: 432.0000 - val_tn: 45054.0000 - val_fn: 24.0000 - val_accuracy: 0.9900 - val_precision: 0.1202 - val_recall: 0.7108 - val_auc: 0.9492\nEpoch 2/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.5537 - tp: 216.0000 - fp: 3783.0000 - tn: 178189.0000 - fn: 88.0000 - accuracy: 0.9788 - precision: 0.0540 - recall: 0.7105 - auc: 0.9033 - val_loss: 0.3285 - val_tp: 69.0000 - val_fp: 514.0000 - val_tn: 44972.0000 - val_fn: 14.0000 - val_accuracy: 0.9884 - val_precision: 0.1184 - val_recall: 0.8313 - val_auc: 0.9605\nEpoch 3/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.4178 - tp: 238.0000 - fp: 4540.0000 - tn: 177432.0000 - fn: 66.0000 - accuracy: 0.9747 - precision: 0.0498 - recall: 0.7829 - auc: 0.9237 - val_loss: 0.2840 - val_tp: 69.0000 - val_fp: 570.0000 - val_tn: 44916.0000 - val_fn: 14.0000 - val_accuracy: 0.9872 - val_precision: 0.1080 - val_recall: 0.8313 - val_auc: 0.9669\nEpoch 4/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.3848 - tp: 247.0000 - fp: 5309.0000 - tn: 176663.0000 - fn: 57.0000 - accuracy: 0.9706 - precision: 0.0445 - recall: 0.8125 - auc: 0.9292 - val_loss: 0.2539 - val_tp: 71.0000 - val_fp: 622.0000 - val_tn: 44864.0000 - val_fn: 12.0000 - val_accuracy: 0.9861 - val_precision: 0.1025 - val_recall: 0.8554 - val_auc: 0.9709\nEpoch 5/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.3596 - tp: 254.0000 - fp: 6018.0000 - tn: 175954.0000 - fn: 50.0000 - accuracy: 0.9667 - precision: 0.0405 - recall: 0.8355 - auc: 0.9323 - val_loss: 0.2363 - val_tp: 72.0000 - val_fp: 713.0000 - val_tn: 44773.0000 - val_fn: 11.0000 - val_accuracy: 0.9841 - val_precision: 0.0917 - val_recall: 0.8675 - val_auc: 0.9725\nEpoch 6/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.3115 - tp: 255.0000 - fp: 6366.0000 - tn: 175606.0000 - fn: 49.0000 - accuracy: 0.9648 - precision: 0.0385 - recall: 0.8388 - auc: 0.9477 - val_loss: 0.2243 - val_tp: 72.0000 - val_fp: 768.0000 - val_tn: 44718.0000 - val_fn: 11.0000 - val_accuracy: 0.9829 - val_precision: 0.0857 - val_recall: 0.8675 - val_auc: 0.9728\nEpoch 7/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.3179 - tp: 258.0000 - fp: 6804.0000 - tn: 175168.0000 - fn: 46.0000 - accuracy: 0.9624 - precision: 0.0365 - recall: 0.8487 - auc: 0.9435 - val_loss: 0.2165 - val_tp: 72.0000 - val_fp: 812.0000 - val_tn: 44674.0000 - val_fn: 11.0000 - val_accuracy: 0.9819 - val_precision: 0.0814 - val_recall: 0.8675 - val_auc: 0.9739\nEpoch 8/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.2880 - tp: 260.0000 - fp: 6669.0000 - tn: 175303.0000 - fn: 44.0000 - accuracy: 0.9632 - precision: 0.0375 - recall: 0.8553 - auc: 0.9530 - val_loss: 0.2122 - val_tp: 72.0000 - val_fp: 783.0000 - val_tn: 44703.0000 - val_fn: 11.0000 - val_accuracy: 0.9826 - val_precision: 0.0842 - val_recall: 0.8675 - val_auc: 0.9769\nEpoch 9/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.2676 - tp: 262.0000 - fp: 6904.0000 - tn: 175068.0000 - fn: 42.0000 - accuracy: 0.9619 - precision: 0.0366 - recall: 0.8618 - auc: 0.9594 - val_loss: 0.2056 - val_tp: 72.0000 - val_fp: 855.0000 - val_tn: 44631.0000 - val_fn: 11.0000 - val_accuracy: 0.9810 - val_precision: 0.0777 - val_recall: 0.8675 - val_auc: 0.9750\nEpoch 10/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.2498 - tp: 266.0000 - fp: 6833.0000 - tn: 175139.0000 - fn: 38.0000 - accuracy: 0.9623 - precision: 0.0375 - recall: 0.8750 - auc: 0.9593 - val_loss: 0.2001 - val_tp: 73.0000 - val_fp: 840.0000 - val_tn: 44646.0000 - val_fn: 10.0000 - val_accuracy: 0.9813 - val_precision: 0.0800 - val_recall: 0.8795 - val_auc: 0.9761\nEpoch 11/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.2681 - tp: 262.0000 - fp: 6845.0000 - tn: 175127.0000 - fn: 42.0000 - accuracy: 0.9622 - precision: 0.0369 - recall: 0.8618 - auc: 0.9559 - val_loss: 0.1964 - val_tp: 73.0000 - val_fp: 865.0000 - val_tn: 44621.0000 - val_fn: 10.0000 - val_accuracy: 0.9808 - val_precision: 0.0778 - val_recall: 0.8795 - val_auc: 0.9768\nEpoch 12/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.2406 - tp: 268.0000 - fp: 7070.0000 - tn: 174902.0000 - fn: 36.0000 - accuracy: 0.9610 - precision: 0.0365 - recall: 0.8816 - auc: 0.9646 - val_loss: 0.1940 - val_tp: 73.0000 - val_fp: 848.0000 - val_tn: 44638.0000 - val_fn: 10.0000 - val_accuracy: 0.9812 - val_precision: 0.0793 - val_recall: 0.8795 - val_auc: 0.9771\nEpoch 13/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.2285 - tp: 269.0000 - fp: 6976.0000 - tn: 174996.0000 - fn: 35.0000 - accuracy: 0.9615 - precision: 0.0371 - recall: 0.8849 - auc: 0.9680 - val_loss: 0.1930 - val_tp: 73.0000 - val_fp: 857.0000 - val_tn: 44629.0000 - val_fn: 10.0000 - val_accuracy: 0.9810 - val_precision: 0.0785 - val_recall: 0.8795 - val_auc: 0.9772\nEpoch 14/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.2322 - tp: 268.0000 - fp: 6718.0000 - tn: 175254.0000 - fn: 36.0000 - accuracy: 0.9629 - precision: 0.0384 - recall: 0.8816 - auc: 0.9644 - val_loss: 0.1915 - val_tp: 73.0000 - val_fp: 808.0000 - val_tn: 44678.0000 - val_fn: 10.0000 - val_accuracy: 0.9820 - val_precision: 0.0829 - val_recall: 0.8795 - val_auc: 0.9781\nEpoch 15/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.2631 - tp: 267.0000 - fp: 6578.0000 - tn: 175394.0000 - fn: 37.0000 - accuracy: 0.9637 - precision: 0.0390 - recall: 0.8783 - auc: 0.9551 - val_loss: 0.1900 - val_tp: 73.0000 - val_fp: 803.0000 - val_tn: 44683.0000 - val_fn: 10.0000 - val_accuracy: 0.9822 - val_precision: 0.0833 - val_recall: 0.8795 - val_auc: 0.9781\nEpoch 16/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.2314 - tp: 266.0000 - fp: 6644.0000 - tn: 175328.0000 - fn: 38.0000 - accuracy: 0.9633 - precision: 0.0385 - recall: 0.8750 - auc: 0.9672 - val_loss: 0.1882 - val_tp: 73.0000 - val_fp: 806.0000 - val_tn: 44680.0000 - val_fn: 10.0000 - val_accuracy: 0.9821 - val_precision: 0.0830 - val_recall: 0.8795 - val_auc: 0.9784\nEpoch 17/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.2152 - tp: 271.0000 - fp: 6663.0000 - tn: 175309.0000 - fn: 33.0000 - accuracy: 0.9633 - precision: 0.0391 - recall: 0.8914 - auc: 0.9687 - val_loss: 0.1895 - val_tp: 73.0000 - val_fp: 754.0000 - val_tn: 44732.0000 - val_fn: 10.0000 - val_accuracy: 0.9832 - val_precision: 0.0883 - val_recall: 0.8795 - val_auc: 0.9785\nEpoch 18/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.2420 - tp: 264.0000 - fp: 6535.0000 - tn: 175437.0000 - fn: 40.0000 - accuracy: 0.9639 - precision: 0.0388 - recall: 0.8684 - auc: 0.9610 - val_loss: 0.1895 - val_tp: 73.0000 - val_fp: 749.0000 - val_tn: 44737.0000 - val_fn: 10.0000 - val_accuracy: 0.9833 - val_precision: 0.0888 - val_recall: 0.8795 - val_auc: 0.9786\nEpoch 19/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.2279 - tp: 268.0000 - fp: 6443.0000 - tn: 175529.0000 - fn: 36.0000 - accuracy: 0.9645 - precision: 0.0399 - recall: 0.8816 - auc: 0.9672 - val_loss: 0.1895 - val_tp: 73.0000 - val_fp: 763.0000 - val_tn: 44723.0000 - val_fn: 10.0000 - val_accuracy: 0.9830 - val_precision: 0.0873 - val_recall: 0.8795 - val_auc: 0.9788\nEpoch 20/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.2247 - tp: 267.0000 - fp: 6596.0000 - tn: 175376.0000 - fn: 37.0000 - accuracy: 0.9636 - precision: 0.0389 - recall: 0.8783 - auc: 0.9684 - val_loss: 0.1896 - val_tp: 73.0000 - val_fp: 760.0000 - val_tn: 44726.0000 - val_fn: 10.0000 - val_accuracy: 0.9831 - val_precision: 0.0876 - val_recall: 0.8795 - val_auc: 0.9797\nEpoch 21/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.2296 - tp: 269.0000 - fp: 6562.0000 - tn: 175410.0000 - fn: 35.0000 - accuracy: 0.9638 - precision: 0.0394 - recall: 0.8849 - auc: 0.9656 - val_loss: 0.1889 - val_tp: 73.0000 - val_fp: 750.0000 - val_tn: 44736.0000 - val_fn: 10.0000 - val_accuracy: 0.9833 - val_precision: 0.0887 - val_recall: 0.8795 - val_auc: 0.9797\nEpoch 22/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.1982 - tp: 271.0000 - fp: 6583.0000 - tn: 175389.0000 - fn: 33.0000 - accuracy: 0.9637 - precision: 0.0395 - recall: 0.8914 - auc: 0.9756 - val_loss: 0.1879 - val_tp: 73.0000 - val_fp: 764.0000 - val_tn: 44722.0000 - val_fn: 10.0000 - val_accuracy: 0.9830 - val_precision: 0.0872 - val_recall: 0.8795 - val_auc: 0.9777\nEpoch 23/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.2154 - tp: 273.0000 - fp: 6552.0000 - tn: 175420.0000 - fn: 31.0000 - accuracy: 0.9639 - precision: 0.0400 - recall: 0.8980 - auc: 0.9682 - val_loss: 0.1882 - val_tp: 73.0000 - val_fp: 762.0000 - val_tn: 44724.0000 - val_fn: 10.0000 - val_accuracy: 0.9831 - val_precision: 0.0874 - val_recall: 0.8795 - val_auc: 0.9779\nEpoch 24/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.1861 - tp: 272.0000 - fp: 6248.0000 - tn: 175724.0000 - fn: 32.0000 - accuracy: 0.9655 - precision: 0.0417 - recall: 0.8947 - auc: 0.9779 - val_loss: 0.1885 - val_tp: 73.0000 - val_fp: 772.0000 - val_tn: 44714.0000 - val_fn: 10.0000 - val_accuracy: 0.9828 - val_precision: 0.0864 - val_recall: 0.8795 - val_auc: 0.9785\nEpoch 25/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.1953 - tp: 270.0000 - fp: 6501.0000 - tn: 175471.0000 - fn: 34.0000 - accuracy: 0.9641 - precision: 0.0399 - recall: 0.8882 - auc: 0.9751 - val_loss: 0.1877 - val_tp: 73.0000 - val_fp: 768.0000 - val_tn: 44718.0000 - val_fn: 10.0000 - val_accuracy: 0.9829 - val_precision: 0.0868 - val_recall: 0.8795 - val_auc: 0.9786\nEpoch 26/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.1704 - tp: 277.0000 - fp: 6215.0000 - tn: 175757.0000 - fn: 27.0000 - accuracy: 0.9658 - precision: 0.0427 - recall: 0.9112 - auc: 0.9808 - val_loss: 0.1903 - val_tp: 73.0000 - val_fp: 698.0000 - val_tn: 44788.0000 - val_fn: 10.0000 - val_accuracy: 0.9845 - val_precision: 0.0947 - val_recall: 0.8795 - val_auc: 0.9788\nEpoch 27/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.1946 - tp: 271.0000 - fp: 6036.0000 - tn: 175936.0000 - fn: 33.0000 - accuracy: 0.9667 - precision: 0.0430 - recall: 0.8914 - auc: 0.9748 - val_loss: 0.1908 - val_tp: 73.0000 - val_fp: 692.0000 - val_tn: 44794.0000 - val_fn: 10.0000 - val_accuracy: 0.9846 - val_precision: 0.0954 - val_recall: 0.8795 - val_auc: 0.9786\nEpoch 28/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.2115 - tp: 271.0000 - fp: 5873.0000 - tn: 176099.0000 - fn: 33.0000 - accuracy: 0.9676 - precision: 0.0441 - recall: 0.8914 - auc: 0.9688 - val_loss: 0.1914 - val_tp: 73.0000 - val_fp: 691.0000 - val_tn: 44795.0000 - val_fn: 10.0000 - val_accuracy: 0.9846 - val_precision: 0.0955 - val_recall: 0.8795 - val_auc: 0.9785\nEpoch 29/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.2237 - tp: 266.0000 - fp: 6047.0000 - tn: 175925.0000 - fn: 38.0000 - accuracy: 0.9666 - precision: 0.0421 - recall: 0.8750 - auc: 0.9672 - val_loss: 0.1909 - val_tp: 73.0000 - val_fp: 698.0000 - val_tn: 44788.0000 - val_fn: 10.0000 - val_accuracy: 0.9845 - val_precision: 0.0947 - val_recall: 0.8795 - val_auc: 0.9784\nEpoch 30/100\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.2232 - tp: 272.0000 - fp: 5990.0000 - tn: 175982.0000 - fn: 32.0000 - accuracy: 0.9670 - precision: 0.0434 - recall: 0.8947 - auc: 0.9668 - val_loss: 0.1919 - val_tp: 73.0000 - val_fp: 642.0000 - val_tn: 44844.0000 - val_fn: 10.0000 - val_accuracy: 0.9857 - val_precision: 0.1021 - val_recall: 0.8795 - val_auc: 0.9785\nEpoch 31/100\n178176/182276 [============================>.] - ETA: 0s - loss: 0.2022 - tp: 273.0000 - fp: 5659.0000 - tn: 172216.0000 - fn: 28.0000 - accuracy: 0.9681 - precision: 0.0460 - recall: 0.9070 - auc: 0.9705Restoring model weights from the end of the best epoch.\n182276/182276 [==============================] - 1s 4us/sample - loss: 0.1989 - tp: 276.0000 - fp: 5796.0000 - tn: 176176.0000 - fn: 28.0000 - accuracy: 0.9680 - precision: 0.0455 - recall: 0.9079 - auc: 0.9708 - val_loss: 0.1920 - val_tp: 73.0000 - val_fp: 626.0000 - val_tn: 44860.0000 - val_fn: 10.0000 - val_accuracy: 0.9860 - val_precision: 0.1044 - val_recall: 0.8795 - val_auc: 0.9788\nEpoch 00031: early stopping\n" ] ], [ [ "### Check training history", "_____no_output_____" ] ], [ [ "plot_metrics(weighted_history)", "_____no_output_____" ] ], [ [ "### Evaluate metrics", "_____no_output_____" ] ], [ [ "train_predictions_weighted = weighted_model.predict(train_features, batch_size=BATCH_SIZE)\ntest_predictions_weighted = weighted_model.predict(test_features, batch_size=BATCH_SIZE)", "_____no_output_____" ], [ "weighted_results = weighted_model.evaluate(test_features, test_labels,\n batch_size=BATCH_SIZE, verbose=0)\nfor name, value in zip(weighted_model.metrics_names, weighted_results):\n print(name, ': ', value)\nprint()\n\nplot_cm(test_labels, test_predictions_weighted)", "loss : 0.06950428275801711\ntp : 94.0\nfp : 905.0\ntn : 55952.0\nfn : 11.0\naccuracy : 0.9839191\nprecision : 0.0940941\nrecall : 0.8952381\nauc : 0.9844724\n\nLegitimate Transactions Detected (True Negatives): 55952\nLegitimate Transactions Incorrectly Detected (False Positives): 905\nFraudulent Transactions Missed (False Negatives): 11\nFraudulent Transactions Detected (True Positives): 94\nTotal Fraudulent Transactions: 105\n" ] ], [ [ "Here you can see that with class weights the accuracy and precision are lower because there are more false positives, but conversely the recall and AUC are higher because the model also found more true positives. Despite having lower accuracy, this model has higher recall (and identifies more fraudulent transactions). Of course, there is a cost to both types of error (you wouldn't want to bug users by flagging too many legitimate transactions as fraudulent, either). Carefully consider the trade offs between these different types of errors for your application.", "_____no_output_____" ], [ "### Plot the ROC", "_____no_output_____" ] ], [ [ "plot_roc(\"Train Baseline\", train_labels, train_predictions_baseline, color=colors[0])\nplot_roc(\"Test Baseline\", test_labels, test_predictions_baseline, color=colors[0], linestyle='--')\n\nplot_roc(\"Train Weighted\", train_labels, train_predictions_weighted, color=colors[1])\nplot_roc(\"Test Weighted\", test_labels, test_predictions_weighted, color=colors[1], linestyle='--')\n\n\nplt.legend(loc='lower right')", "_____no_output_____" ] ], [ [ "## Oversampling", "_____no_output_____" ], [ "### Oversample the minority class\n\nA related approach would be to resample the dataset by oversampling the minority class.", "_____no_output_____" ] ], [ [ "pos_features = train_features[bool_train_labels]\nneg_features = train_features[~bool_train_labels]\n\npos_labels = train_labels[bool_train_labels]\nneg_labels = train_labels[~bool_train_labels]", "_____no_output_____" ] ], [ [ "#### Using NumPy\n\nYou can balance the dataset manually by choosing the right number of random \nindices from the positive examples:", "_____no_output_____" ] ], [ [ "ids = np.arange(len(pos_features))\nchoices = np.random.choice(ids, len(neg_features))\n\nres_pos_features = pos_features[choices]\nres_pos_labels = pos_labels[choices]\n\nres_pos_features.shape", "_____no_output_____" ], [ "resampled_features = np.concatenate([res_pos_features, neg_features], axis=0)\nresampled_labels = np.concatenate([res_pos_labels, neg_labels], axis=0)\n\norder = np.arange(len(resampled_labels))\nnp.random.shuffle(order)\nresampled_features = resampled_features[order]\nresampled_labels = resampled_labels[order]\n\nresampled_features.shape", "_____no_output_____" ] ], [ [ "#### Using `tf.data`", "_____no_output_____" ], [ "If you're using `tf.data` the easiest way to produce balanced examples is to start with a `positive` and a `negative` dataset, and merge them. See [the tf.data guide](../../guide/data.ipynb) for more examples.", "_____no_output_____" ] ], [ [ "BUFFER_SIZE = 100000\n\ndef make_ds(features, labels):\n ds = tf.data.Dataset.from_tensor_slices((features, labels))#.cache()\n ds = ds.shuffle(BUFFER_SIZE).repeat()\n return ds\n\npos_ds = make_ds(pos_features, pos_labels)\nneg_ds = make_ds(neg_features, neg_labels)", "_____no_output_____" ] ], [ [ "Each dataset provides `(feature, label)` pairs:", "_____no_output_____" ] ], [ [ "for features, label in pos_ds.take(1):\n print(\"Features:\\n\", features.numpy())\n print()\n print(\"Label: \", label.numpy())", "Features:\n [-2.46955933 3.42534191 -4.42937043 3.70651659 -3.17895499 -1.30458304\n -5. 2.86676917 -4.9308611 -5. 3.58555137 -5.\n 1.51535494 -5. 0.01049775 -5. -5. -5.\n 2.02380731 0.36595419 1.61836304 -1.16743779 0.31324117 -0.35515978\n -0.62579636 -0.55952005 0.51255883 1.15454727 0.87478003]\n\nLabel: 1\n" ] ], [ [ "Merge the two together using `experimental.sample_from_datasets`:", "_____no_output_____" ] ], [ [ "resampled_ds = tf.data.experimental.sample_from_datasets([pos_ds, neg_ds], weights=[0.5, 0.5])\nresampled_ds = resampled_ds.batch(BATCH_SIZE).prefetch(2)", "_____no_output_____" ], [ "for features, label in resampled_ds.take(1):\n print(label.numpy().mean())", "0.48974609375\n" ] ], [ [ "To use this dataset, you'll need the number of steps per epoch.\n\nThe definition of \"epoch\" in this case is less clear. Say it's the number of batches required to see each negative example once:", "_____no_output_____" ] ], [ [ "resampled_steps_per_epoch = np.ceil(2.0*neg/BATCH_SIZE)\nresampled_steps_per_epoch", "_____no_output_____" ] ], [ [ "### Train on the oversampled data\n\nNow try training the model with the resampled data set instead of using class weights to see how these methods compare.\n\nNote: Because the data was balanced by replicating the positive examples, the total dataset size is larger, and each epoch runs for more training steps. ", "_____no_output_____" ] ], [ [ "resampled_model = make_model()\nresampled_model.load_weights(initial_weights)\n\n# Reset the bias to zero, since this dataset is balanced.\noutput_layer = resampled_model.layers[-1] \noutput_layer.bias.assign([0])\n\nval_ds = tf.data.Dataset.from_tensor_slices((val_features, val_labels)).cache()\nval_ds = val_ds.batch(BATCH_SIZE).prefetch(2) \n\nresampled_history = resampled_model.fit(\n resampled_ds,\n epochs=EPOCHS,\n steps_per_epoch=resampled_steps_per_epoch,\n callbacks = [early_stopping],\n validation_data=val_ds)", "Train for 278.0 steps, validate for 23 steps\nEpoch 1/100\n278/278 [==============================] - 13s 48ms/step - loss: 0.4624 - tp: 267186.0000 - fp: 124224.0000 - tn: 160439.0000 - fn: 17495.0000 - accuracy: 0.7511 - precision: 0.6826 - recall: 0.9385 - auc: 0.9268 - val_loss: 0.3299 - val_tp: 79.0000 - val_fp: 2825.0000 - val_tn: 42661.0000 - val_fn: 4.0000 - val_accuracy: 0.9379 - val_precision: 0.0272 - val_recall: 0.9518 - val_auc: 0.9799\nEpoch 2/100\n278/278 [==============================] - 11s 39ms/step - loss: 0.2362 - tp: 264077.0000 - fp: 26654.0000 - tn: 257570.0000 - fn: 21043.0000 - accuracy: 0.9162 - precision: 0.9083 - recall: 0.9262 - auc: 0.9708 - val_loss: 0.1926 - val_tp: 75.0000 - val_fp: 1187.0000 - val_tn: 44299.0000 - val_fn: 8.0000 - val_accuracy: 0.9738 - val_precision: 0.0594 - val_recall: 0.9036 - val_auc: 0.9779\nEpoch 3/100\n278/278 [==============================] - 11s 40ms/step - loss: 0.1887 - tp: 263490.0000 - fp: 12935.0000 - tn: 271381.0000 - fn: 21538.0000 - accuracy: 0.9395 - precision: 0.9532 - recall: 0.9244 - auc: 0.9804 - val_loss: 0.1373 - val_tp: 75.0000 - val_fp: 1064.0000 - val_tn: 44422.0000 - val_fn: 8.0000 - val_accuracy: 0.9765 - val_precision: 0.0658 - val_recall: 0.9036 - val_auc: 0.9778\nEpoch 4/100\n278/278 [==============================] - 11s 41ms/step - loss: 0.1605 - tp: 263933.0000 - fp: 10513.0000 - tn: 274505.0000 - fn: 20393.0000 - accuracy: 0.9457 - precision: 0.9617 - recall: 0.9283 - auc: 0.9866 - val_loss: 0.1078 - val_tp: 75.0000 - val_fp: 1070.0000 - val_tn: 44416.0000 - val_fn: 8.0000 - val_accuracy: 0.9763 - val_precision: 0.0655 - val_recall: 0.9036 - val_auc: 0.9783\nEpoch 5/100\n278/278 [==============================] - 11s 39ms/step - loss: 0.1423 - tp: 265715.0000 - fp: 9592.0000 - tn: 275145.0000 - fn: 18892.0000 - accuracy: 0.9500 - precision: 0.9652 - recall: 0.9336 - auc: 0.9901 - val_loss: 0.0928 - val_tp: 75.0000 - val_fp: 1051.0000 - val_tn: 44435.0000 - val_fn: 8.0000 - val_accuracy: 0.9768 - val_precision: 0.0666 - val_recall: 0.9036 - val_auc: 0.9762\nEpoch 6/100\n278/278 [==============================] - 11s 40ms/step - loss: 0.1297 - tp: 267181.0000 - fp: 8944.0000 - tn: 275445.0000 - fn: 17774.0000 - accuracy: 0.9531 - precision: 0.9676 - recall: 0.9376 - auc: 0.9920 - val_loss: 0.0847 - val_tp: 75.0000 - val_fp: 1077.0000 - val_tn: 44409.0000 - val_fn: 8.0000 - val_accuracy: 0.9762 - val_precision: 0.0651 - val_recall: 0.9036 - val_auc: 0.9748\nEpoch 7/100\n278/278 [==============================] - 11s 39ms/step - loss: 0.1203 - tp: 267440.0000 - fp: 8606.0000 - tn: 276459.0000 - fn: 16839.0000 - accuracy: 0.9553 - precision: 0.9688 - recall: 0.9408 - auc: 0.9933 - val_loss: 0.0775 - val_tp: 75.0000 - val_fp: 1003.0000 - val_tn: 44483.0000 - val_fn: 8.0000 - val_accuracy: 0.9778 - val_precision: 0.0696 - val_recall: 0.9036 - val_auc: 0.9742\nEpoch 8/100\n278/278 [==============================] - 11s 40ms/step - loss: 0.1132 - tp: 268799.0000 - fp: 8165.0000 - tn: 276260.0000 - fn: 16120.0000 - accuracy: 0.9573 - precision: 0.9705 - recall: 0.9434 - auc: 0.9941 - val_loss: 0.0716 - val_tp: 75.0000 - val_fp: 927.0000 - val_tn: 44559.0000 - val_fn: 8.0000 - val_accuracy: 0.9795 - val_precision: 0.0749 - val_recall: 0.9036 - val_auc: 0.9713\nEpoch 9/100\n278/278 [==============================] - 11s 40ms/step - loss: 0.1074 - tp: 269627.0000 - fp: 7971.0000 - tn: 276559.0000 - fn: 15187.0000 - accuracy: 0.9593 - precision: 0.9713 - recall: 0.9467 - auc: 0.9947 - val_loss: 0.0670 - val_tp: 75.0000 - val_fp: 880.0000 - val_tn: 44606.0000 - val_fn: 8.0000 - val_accuracy: 0.9805 - val_precision: 0.0785 - val_recall: 0.9036 - val_auc: 0.9713\nEpoch 10/100\n278/278 [==============================] - 11s 39ms/step - loss: 0.1017 - tp: 270359.0000 - fp: 7590.0000 - tn: 277311.0000 - fn: 14084.0000 - accuracy: 0.9619 - precision: 0.9727 - recall: 0.9505 - auc: 0.9952 - val_loss: 0.0629 - val_tp: 75.0000 - val_fp: 848.0000 - val_tn: 44638.0000 - val_fn: 8.0000 - val_accuracy: 0.9812 - val_precision: 0.0813 - val_recall: 0.9036 - val_auc: 0.9717\nEpoch 11/100\n276/278 [============================>.] - ETA: 0s - loss: 0.0977 - tp: 269672.0000 - fp: 7408.0000 - tn: 274621.0000 - fn: 13547.0000 - accuracy: 0.9629 - precision: 0.9733 - recall: 0.9522 - auc: 0.9955Restoring model weights from the end of the best epoch.\n278/278 [==============================] - 11s 39ms/step - loss: 0.0978 - tp: 271609.0000 - fp: 7474.0000 - tn: 276625.0000 - fn: 13636.0000 - accuracy: 0.9629 - precision: 0.9732 - recall: 0.9522 - auc: 0.9955 - val_loss: 0.0615 - val_tp: 75.0000 - val_fp: 841.0000 - val_tn: 44645.0000 - val_fn: 8.0000 - val_accuracy: 0.9814 - val_precision: 0.0819 - val_recall: 0.9036 - val_auc: 0.9637\nEpoch 00011: early stopping\n" ] ], [ [ "If the training process were considering the whole dataset on each gradient update, this oversampling would be basically identical to the class weighting.\n\nBut when training the model batch-wise, as you did here, the oversampled data provides a smoother gradient signal: Instead of each positive example being shown in one batch with a large weight, they're shown in many different batches each time with a small weight. \n\nThis smoother gradient signal makes it easier to train the model.", "_____no_output_____" ], [ "### Check training history\n\nNote that the distributions of metrics will be different here, because the training data has a totally different distribution from the validation and test data. ", "_____no_output_____" ] ], [ [ "plot_metrics(resampled_history )", "_____no_output_____" ] ], [ [ "### Re-train\n", "_____no_output_____" ], [ "Because training is easier on the balanced data, the above training procedure may overfit quickly. \n\nSo break up the epochs to give the `callbacks.EarlyStopping` finer control over when to stop training.", "_____no_output_____" ] ], [ [ "resampled_model = make_model()\nresampled_model.load_weights(initial_weights)\n\n# Reset the bias to zero, since this dataset is balanced.\noutput_layer = resampled_model.layers[-1] \noutput_layer.bias.assign([0])\n\nresampled_history = resampled_model.fit(\n resampled_ds,\n # These are not real epochs\n steps_per_epoch = 20,\n epochs=10*EPOCHS,\n callbacks = [early_stopping],\n validation_data=(val_ds))", "Train for 20 steps, validate for 23 steps\nEpoch 1/1000\n20/20 [==============================] - 4s 181ms/step - loss: 0.8800 - tp: 18783.0000 - fp: 16378.0000 - tn: 4036.0000 - fn: 1763.0000 - accuracy: 0.5571 - precision: 0.5342 - recall: 0.9142 - auc: 0.7752 - val_loss: 1.3661 - val_tp: 83.0000 - val_fp: 40065.0000 - val_tn: 5421.0000 - val_fn: 0.0000e+00 - val_accuracy: 0.1208 - val_precision: 0.0021 - val_recall: 1.0000 - val_auc: 0.9425\nEpoch 2/1000\n20/20 [==============================] - 1s 35ms/step - loss: 0.7378 - tp: 19613.0000 - fp: 15282.0000 - tn: 5187.0000 - fn: 878.0000 - accuracy: 0.6055 - precision: 0.5621 - recall: 0.9572 - auc: 0.8680 - val_loss: 1.1629 - val_tp: 83.0000 - val_fp: 36851.0000 - val_tn: 8635.0000 - val_fn: 0.0000e+00 - val_accuracy: 0.1913 - val_precision: 0.0022 - val_recall: 1.0000 - val_auc: 0.9580\nEpoch 3/1000\n20/20 [==============================] - 1s 39ms/step - loss: 0.6431 - tp: 19522.0000 - fp: 13990.0000 - tn: 6558.0000 - fn: 890.0000 - accuracy: 0.6367 - precision: 0.5825 - recall: 0.9564 - auc: 0.8950 - val_loss: 0.9853 - val_tp: 82.0000 - val_fp: 32268.0000 - val_tn: 13218.0000 - val_fn: 1.0000 - val_accuracy: 0.2919 - val_precision: 0.0025 - val_recall: 0.9880 - val_auc: 0.9660\nEpoch 4/1000\n20/20 [==============================] - 1s 39ms/step - loss: 0.5563 - tp: 19488.0000 - fp: 12475.0000 - tn: 8032.0000 - fn: 965.0000 - accuracy: 0.6719 - precision: 0.6097 - recall: 0.9528 - auc: 0.9135 - val_loss: 0.8430 - val_tp: 82.0000 - val_fp: 26633.0000 - val_tn: 18853.0000 - val_fn: 1.0000 - val_accuracy: 0.4155 - val_precision: 0.0031 - val_recall: 0.9880 - val_auc: 0.9713\nEpoch 5/1000\n20/20 [==============================] - 1s 37ms/step - loss: 0.4984 - tp: 19489.0000 - fp: 11049.0000 - tn: 9377.0000 - fn: 1045.0000 - accuracy: 0.7047 - precision: 0.6382 - recall: 0.9491 - auc: 0.9242 - val_loss: 0.7307 - val_tp: 82.0000 - val_fp: 20850.0000 - val_tn: 24636.0000 - val_fn: 1.0000 - val_accuracy: 0.5424 - val_precision: 0.0039 - val_recall: 0.9880 - val_auc: 0.9753\nEpoch 6/1000\n20/20 [==============================] - 1s 39ms/step - loss: 0.4463 - tp: 19305.0000 - fp: 9622.0000 - tn: 10895.0000 - fn: 1138.0000 - accuracy: 0.7373 - precision: 0.6674 - recall: 0.9443 - auc: 0.9336 - val_loss: 0.6405 - val_tp: 82.0000 - val_fp: 15843.0000 - val_tn: 29643.0000 - val_fn: 1.0000 - val_accuracy: 0.6523 - val_precision: 0.0051 - val_recall: 0.9880 - val_auc: 0.9773\nEpoch 7/1000\n20/20 [==============================] - 1s 40ms/step - loss: 0.4121 - tp: 19365.0000 - fp: 8524.0000 - tn: 11931.0000 - fn: 1140.0000 - accuracy: 0.7641 - precision: 0.6944 - recall: 0.9444 - auc: 0.9411 - val_loss: 0.5691 - val_tp: 82.0000 - val_fp: 11981.0000 - val_tn: 33505.0000 - val_fn: 1.0000 - val_accuracy: 0.7371 - val_precision: 0.0068 - val_recall: 0.9880 - val_auc: 0.9787\nEpoch 8/1000\n20/20 [==============================] - 1s 39ms/step - loss: 0.3784 - tp: 19242.0000 - fp: 7375.0000 - tn: 13072.0000 - fn: 1271.0000 - accuracy: 0.7889 - precision: 0.7229 - recall: 0.9380 - auc: 0.9461 - val_loss: 0.5120 - val_tp: 80.0000 - val_fp: 9309.0000 - val_tn: 36177.0000 - val_fn: 3.0000 - val_accuracy: 0.7957 - val_precision: 0.0085 - val_recall: 0.9639 - val_auc: 0.9794\nEpoch 9/1000\n20/20 [==============================] - 1s 45ms/step - loss: 0.3551 - tp: 19106.0000 - fp: 6529.0000 - tn: 13989.0000 - fn: 1336.0000 - accuracy: 0.8080 - precision: 0.7453 - recall: 0.9346 - auc: 0.9495 - val_loss: 0.4657 - val_tp: 80.0000 - val_fp: 7354.0000 - val_tn: 38132.0000 - val_fn: 3.0000 - val_accuracy: 0.8386 - val_precision: 0.0108 - val_recall: 0.9639 - val_auc: 0.9799\nEpoch 10/1000\n20/20 [==============================] - 1s 38ms/step - loss: 0.3350 - tp: 19149.0000 - fp: 5794.0000 - tn: 14698.0000 - fn: 1319.0000 - accuracy: 0.8263 - precision: 0.7677 - recall: 0.9356 - auc: 0.9535 - val_loss: 0.4275 - val_tp: 80.0000 - val_fp: 5832.0000 - val_tn: 39654.0000 - val_fn: 3.0000 - val_accuracy: 0.8720 - val_precision: 0.0135 - val_recall: 0.9639 - val_auc: 0.9802\nEpoch 11/1000\n20/20 [==============================] - 1s 40ms/step - loss: 0.3168 - tp: 19224.0000 - fp: 5013.0000 - tn: 15322.0000 - fn: 1401.0000 - accuracy: 0.8434 - precision: 0.7932 - recall: 0.9321 - auc: 0.9552 - val_loss: 0.3969 - val_tp: 80.0000 - val_fp: 4730.0000 - val_tn: 40756.0000 - val_fn: 3.0000 - val_accuracy: 0.8961 - val_precision: 0.0166 - val_recall: 0.9639 - val_auc: 0.9805\nEpoch 12/1000\n20/20 [==============================] - 1s 40ms/step - loss: 0.3077 - tp: 19028.0000 - fp: 4564.0000 - tn: 16058.0000 - fn: 1310.0000 - accuracy: 0.8566 - precision: 0.8065 - recall: 0.9356 - auc: 0.9593 - val_loss: 0.3695 - val_tp: 80.0000 - val_fp: 3819.0000 - val_tn: 41667.0000 - val_fn: 3.0000 - val_accuracy: 0.9161 - val_precision: 0.0205 - val_recall: 0.9639 - val_auc: 0.9804\nEpoch 13/1000\n20/20 [==============================] - 1s 40ms/step - loss: 0.2936 - tp: 19047.0000 - fp: 4028.0000 - tn: 16444.0000 - fn: 1441.0000 - accuracy: 0.8665 - precision: 0.8254 - recall: 0.9297 - auc: 0.9597 - val_loss: 0.3461 - val_tp: 79.0000 - val_fp: 3149.0000 - val_tn: 42337.0000 - val_fn: 4.0000 - val_accuracy: 0.9308 - val_precision: 0.0245 - val_recall: 0.9518 - val_auc: 0.9802\nEpoch 14/1000\n20/20 [==============================] - 1s 38ms/step - loss: 0.2829 - tp: 19087.0000 - fp: 3596.0000 - tn: 16855.0000 - fn: 1422.0000 - accuracy: 0.8775 - precision: 0.8415 - recall: 0.9307 - auc: 0.9619 - val_loss: 0.3266 - val_tp: 79.0000 - val_fp: 2691.0000 - val_tn: 42795.0000 - val_fn: 4.0000 - val_accuracy: 0.9409 - val_precision: 0.0285 - val_recall: 0.9518 - val_auc: 0.9803\nEpoch 15/1000\n20/20 [==============================] - 1s 39ms/step - loss: 0.2748 - tp: 19020.0000 - fp: 3174.0000 - tn: 17283.0000 - fn: 1483.0000 - accuracy: 0.8863 - precision: 0.8570 - recall: 0.9277 - auc: 0.9627 - val_loss: 0.3095 - val_tp: 79.0000 - val_fp: 2360.0000 - val_tn: 43126.0000 - val_fn: 4.0000 - val_accuracy: 0.9481 - val_precision: 0.0324 - val_recall: 0.9518 - val_auc: 0.9797\nEpoch 16/1000\n20/20 [==============================] - 1s 40ms/step - loss: 0.2666 - tp: 18890.0000 - fp: 2889.0000 - tn: 17757.0000 - fn: 1424.0000 - accuracy: 0.8947 - precision: 0.8673 - recall: 0.9299 - auc: 0.9653 - val_loss: 0.2945 - val_tp: 78.0000 - val_fp: 2101.0000 - val_tn: 43385.0000 - val_fn: 5.0000 - val_accuracy: 0.9538 - val_precision: 0.0358 - val_recall: 0.9398 - val_auc: 0.9796\nEpoch 17/1000\n20/20 [==============================] - 1s 38ms/step - loss: 0.2583 - tp: 18959.0000 - fp: 2517.0000 - tn: 17973.0000 - fn: 1511.0000 - accuracy: 0.9017 - precision: 0.8828 - recall: 0.9262 - auc: 0.9657 - val_loss: 0.2817 - val_tp: 78.0000 - val_fp: 1929.0000 - val_tn: 43557.0000 - val_fn: 5.0000 - val_accuracy: 0.9576 - val_precision: 0.0389 - val_recall: 0.9398 - val_auc: 0.9794\nEpoch 18/1000\n20/20 [==============================] - 1s 46ms/step - loss: 0.2511 - tp: 19104.0000 - fp: 2344.0000 - tn: 18043.0000 - fn: 1469.0000 - accuracy: 0.9069 - precision: 0.8907 - recall: 0.9286 - auc: 0.9678 - val_loss: 0.2704 - val_tp: 78.0000 - val_fp: 1787.0000 - val_tn: 43699.0000 - val_fn: 5.0000 - val_accuracy: 0.9607 - val_precision: 0.0418 - val_recall: 0.9398 - val_auc: 0.9793\nEpoch 19/1000\n20/20 [==============================] - 1s 40ms/step - loss: 0.2445 - tp: 19183.0000 - fp: 2087.0000 - tn: 18215.0000 - fn: 1475.0000 - accuracy: 0.9130 - precision: 0.9019 - recall: 0.9286 - auc: 0.9693 - val_loss: 0.2598 - val_tp: 78.0000 - val_fp: 1665.0000 - val_tn: 43821.0000 - val_fn: 5.0000 - val_accuracy: 0.9634 - val_precision: 0.0448 - val_recall: 0.9398 - val_auc: 0.9791\nEpoch 20/1000\n20/20 [==============================] - 1s 39ms/step - loss: 0.2373 - tp: 18995.0000 - fp: 1906.0000 - tn: 18602.0000 - fn: 1457.0000 - accuracy: 0.9179 - precision: 0.9088 - recall: 0.9288 - auc: 0.9712 - val_loss: 0.2500 - val_tp: 78.0000 - val_fp: 1587.0000 - val_tn: 43899.0000 - val_fn: 5.0000 - val_accuracy: 0.9651 - val_precision: 0.0468 - val_recall: 0.9398 - val_auc: 0.9788\nEpoch 21/1000\n19/20 [===========================>..] - ETA: 0s - loss: 0.2378 - tp: 18121.0000 - fp: 1821.0000 - tn: 17599.0000 - fn: 1371.0000 - accuracy: 0.9180 - precision: 0.9087 - recall: 0.9297 - auc: 0.9714Restoring model weights from the end of the best epoch.\n20/20 [==============================] - 1s 40ms/step - loss: 0.2376 - tp: 19083.0000 - fp: 1918.0000 - tn: 18513.0000 - fn: 1446.0000 - accuracy: 0.9179 - precision: 0.9087 - recall: 0.9296 - auc: 0.9714 - val_loss: 0.2401 - val_tp: 78.0000 - val_fp: 1485.0000 - val_tn: 44001.0000 - val_fn: 5.0000 - val_accuracy: 0.9673 - val_precision: 0.0499 - val_recall: 0.9398 - val_auc: 0.9785\nEpoch 00021: early stopping\n" ] ], [ [ "### Re-check training history", "_____no_output_____" ] ], [ [ "plot_metrics(resampled_history)", "_____no_output_____" ] ], [ [ "### Evaluate metrics", "_____no_output_____" ] ], [ [ "train_predictions_resampled = resampled_model.predict(train_features, batch_size=BATCH_SIZE)\ntest_predictions_resampled = resampled_model.predict(test_features, batch_size=BATCH_SIZE)", "_____no_output_____" ], [ "resampled_results = resampled_model.evaluate(test_features, test_labels,\n batch_size=BATCH_SIZE, verbose=0)\nfor name, value in zip(resampled_model.metrics_names, resampled_results):\n print(name, ': ', value)\nprint()\n\nplot_cm(test_labels, test_predictions_resampled)", "loss : 0.3960801533448772\ntp : 99.0\nfp : 5892.0\ntn : 50965.0\nfn : 6.0\naccuracy : 0.8964573\nprecision : 0.016524788\nrecall : 0.94285715\nauc : 0.9804354\n\nLegitimate Transactions Detected (True Negatives): 50965\nLegitimate Transactions Incorrectly Detected (False Positives): 5892\nFraudulent Transactions Missed (False Negatives): 6\nFraudulent Transactions Detected (True Positives): 99\nTotal Fraudulent Transactions: 105\n" ] ], [ [ "### Plot the ROC", "_____no_output_____" ] ], [ [ "plot_roc(\"Train Baseline\", train_labels, train_predictions_baseline, color=colors[0])\nplot_roc(\"Test Baseline\", test_labels, test_predictions_baseline, color=colors[0], linestyle='--')\n\nplot_roc(\"Train Weighted\", train_labels, train_predictions_weighted, color=colors[1])\nplot_roc(\"Test Weighted\", test_labels, test_predictions_weighted, color=colors[1], linestyle='--')\n\nplot_roc(\"Train Resampled\", train_labels, train_predictions_resampled, color=colors[2])\nplot_roc(\"Test Resampled\", test_labels, test_predictions_resampled, color=colors[2], linestyle='--')\nplt.legend(loc='lower right')", "_____no_output_____" ] ], [ [ "## Applying this tutorial to your problem\n\nImbalanced data classification is an inherantly difficult task since there are so few samples to learn from. You should always start with the data first and do your best to collect as many samples as possible and give substantial thought to what features may be relevant so the model can get the most out of your minority class. At some point your model may struggle to improve and yield the results you want, so it is important to keep in mind the context of your problem and the trade offs between different types of errors.", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown" ]

[ [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code", "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import numpy as np\nimport matplotlib.pyplot as plt\n\n### Just some matplotlib tweaks\nimport matplotlib as mpl\n\nmpl.rcParams[\"xtick.direction\"] = \"in\"\nmpl.rcParams[\"ytick.direction\"] = \"in\"\nmpl.rcParams[\"lines.markeredgecolor\"] = \"k\"\nmpl.rcParams[\"lines.markeredgewidth\"] = 1.5\nmpl.rcParams[\"figure.dpi\"] = 200\nfrom matplotlib import rc\nrc('font', family='serif')\nrc('text', usetex=True)\nrc('xtick', labelsize='medium')\nrc('ytick', labelsize='medium')\nrc(\"axes\", labelsize = \"large\")\ndef cm2inch(value):\n return value/2.54", "_____no_output_____" ], [ "z = np.linspace(10e-9, 5e-6, 10000)\na = 1.5e-6\nD0 = 4e-21 / (6*np.pi * 0.001 * a)\n#taking alpha = 1\nv_noise = 2*D0 * a * (2*a**2 + 12 *a * z + 21 * z** 2) / (2*a**2 + 9*a*z + z**2)**2\n", "_____no_output_____" ], [ "plt.figure(figsize=( cm2inch(16),cm2inch(8)))\nplt.plot(z*1e6, v_noise*1e6)", "_____no_output_____" ], [ "def eta_z(z):\n return 0.001 * (6*z**2 + 9 * a * z + 2 * a**2)/(6*z**2 + 2*a*z)\n\ndef gamma(z):\n return 6 * np.pi * eta_z(z) * a\n\nlb = 500e-9\nld = 50e-9\nB = 4 #kt unit\ndef F_z(z):\n return - 4e-21 * (-1/ld * B *np.exp(-z/ld) + 1/lb)", "_____no_output_____" ], [ "v_deterministic = 1/gamma(z) * F_z(z)", "_____no_output_____" ], [ "plt.figure(figsize=( cm2inch(16),cm2inch(8)))\nplt.plot(z*1e6, v_deterministic*1e6, label = \"$v_\\mathrm{d}$\")\nplt.plot(z*1e6, v_noise*1e6, label=\"$v_\\mathrm{noise}$\")\nplt.plot(z*1e6, v_noise*1e6 + v_deterministic*1e6, color = \"black\", label = \"$\\\\bar{v}_\\mathrm{d}$\")\nplt.ylabel(\"$v$ ($\\\\mathrm{\\\\mu m.s^{-1}}$)\")\nplt.xlabel(\"$z$ ($\\\\mathrm{\\\\mu m}$)\")\nplt.legend(frameon=False)\nplt.tight_layout()", "_____no_output_____" ], [ "vtot = v_noise*1e6 + v_deterministic*1e6", "_____no_output_____" ], [ "vtot_gradient =np.abs(1/vtot* np.gradient(vtot, np.mean(np.diff(z))))", "_____no_output_____" ], [ "np.mean(np.diff(z))", "_____no_output_____" ], [ "D_z = 4e-21 / gamma(z)\nD_z_gradient = np.abs( 1/ D_z * np.gradient(D_z, np.mean(np.diff(z))))", "_____no_output_____" ], [ "plt.semilogx(z*1e6, vtot_gradient, label = \"Drifts gradient\")\nplt.semilogx(z*1e6, D_z_gradient, label = \"Diffusion gradient\")\nplt.ylabel(\"gradients\")\nplt.xlabel(\"$z$ ($\\mu$m)\")\nplt.legend()", "_____no_output_____" ], [ "D0 = 4e-21 / (6 * np.pi * 0.001 * 1.5e-6)\na = 1.5e-6\ndef tau_max(ld, B, z):\n lb = 500e-9\n return a / (2*D0) * np.power((1/(B/ld + 1/lb) + z),2) / z\n\ndef min_tau_max(ld, B):\n lb = 500e-9\n return 2* a / D0 / (B/ld - 1/lb)\n", "_____no_output_____" ], [ "lb = 500e-9\nB = 10\nz = np.linspace(1e-9, 100e-9, 10000)\n\nfor i in [20e-9, 30e-9, 75e-9, 100e-9]:\n plt.semilogx(z*1e6, tau_max(i, B, z), label= \"$\\ell_\\mathrm{D} =$ \" + str(np.round(i*1e9))[:-2] + \" nm\")\n plt.plot(1/(B/i - 1/lb)* 1e6, min_tau_max(i, B), \"o\", markersize = 2, color = \"b\")\n\nlds = np.linspace(20e-9, 300e-9)\nplt.plot(1/(B/lds - 1/lb)* 1e6, min_tau_max(lds, B), color = \"black\")\n\nplt.ylabel(\"$\\\\tau_\\mathrm{max}$\")\nplt.xlabel(\"$z$ ($\\mu$m)\")\n\n\nplt.legend(frameon = False)\n\n", "_____no_output_____" ], [ "min_tau_max(i, B)", "_____no_output_____" ], [ "fig = plt.figure(figsize = (cm2inch(16),cm2inch(9)))\ngs = fig.add_gridspec(1, 2)\n\nfig.add_subplot(gs[0, 0])\n\nplt.loglog(z*1e6, vtot_gradient, label = \"$\\\\frac{1}{\\\\bar{v}_\\\\mathrm{d}} \\\\frac{\\\\partial \\\\bar{v}_\\\\mathrm{d} }{\\partial z}$\")\nplt.plot(z*1e6, D_z_gradient, label = \"$\\\\frac{1}{D_\\\\bot} \\\\frac{\\\\partial D_\\\\bot }{\\partial z}$\")\nplt.ylabel(\"relative variations (m$^{-1}$)\")\nplt.xlabel(\"$z$ ($\\mu$m)\")\nplt.legend(frameon = False)\n\n#plt.text(0.05, -1e6, \"a)\", fontsize=20)\n\nfig.add_subplot(gs[0, 1])\n\nld = 500e-9\nB = 10\nz = np.linspace(1e-9, 100e-9, 10000)\n\nfor i in [20e-9, 30e-9, 75e-9, 100e-9]:\n plt.semilogx(z*1e6, tau_max(i, B, z), label= \"$\\ell_\\mathrm{D} =$ \" + str(np.round(i*1e9))[:-2] + \" nm\")\n plt.plot(1/(B/i - 1/lb)* 1e6, min_tau_max(i, B), \"o\", markersize = 2, color = \"b\")\n\nlds = np.linspace(20e-9, 100e-9)\nplt.plot(1/(B/lds- 1/lb)* 1e6, min_tau_max(lds, B), color = \"black\")\n\nplt.ylabel(\"$\\\\tau_\\mathrm{max}$\")\nplt.xlabel(\"$z$ ($\\mu$m)\")\n\nplt.text(0.05, 0.06, \"b)\", fontsize=20)\n\nplt.legend(frameon = False)\n\nplt.tight_layout()\n\nplt.savefig(\"maximal_tau.pdf\")\n", "_____no_output_____" ], [ "min_tau_max(20e-9, B)", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "<img src=\"http://cfs22.simplicdn.net/ice9/new_logo.svgz \"/>\n\n# Assignment 01: Evaluate the FAA Dataset\n\n*The comments/sections provided are your cues to perform the assignment. You don't need to limit yourself to the number of rows/cells provided. You can add additional rows in each section to add more lines of code.*\n\n*If at any point in time you need help on solving this assignment, view our demo video to understand the different steps of the code.*\n\n**Happy coding!**\n\n* * *", "_____no_output_____" ], [ "#### 1: VIew and import the dataset", "_____no_output_____" ] ], [ [ "#Import necessary libraries\nimport pandas as pd", "_____no_output_____" ], [ "#Import the FAA (Federal Aviation Authority) dataset\ndf_faa_dataset = pd.read_csv(\"D:/COURSES/Artificial Intellegence Engineer/Data Analytics With Python/Analyse the Federal Aviation Authority Dataset using Pandas/WORK DONE/faa_ai_prelim.csv\")", "_____no_output_____" ] ], [ [ "#### 2: View and understand the dataset", "_____no_output_____" ] ], [ [ "#View the dataset shape\ndf_faa_dataset.shape", "_____no_output_____" ], [ "#View the first five observations\ndf_faa_dataset.head()", "_____no_output_____" ], [ "#View all the columns present in the dataset\ndf_faa_dataset.columns", "_____no_output_____" ] ], [ [ "#### 3: Extract the following attributes from the dataset:\n1. Aircraft make name\n2. State name\n3. Aircraft model name\n4. Text information\n5. Flight phase\n6. Event description type\n7. Fatal flag", "_____no_output_____" ] ], [ [ "#Create a new dataframe with only the required columns\ndf_analyze_dataset = df_faa_dataset[['LOC_STATE_NAME', 'RMK_TEXT', 'EVENT_TYPE_DESC', 'ACFT_MAKE_NAME',\n 'ACFT_MODEL_NAME', 'FLT_PHASE', 'FATAL_FLAG']]", "_____no_output_____" ], [ "#View the type of the object\ntype(df_analyze_dataset)", "_____no_output_____" ], [ "#Check if the dataframe contains all the required attributes\ndf_analyze_dataset.head()", "_____no_output_____" ] ], [ [ "#### 4. Clean the dataset and replace the fatal flag NaN with “No”", "_____no_output_____" ] ], [ [ "#Replace all Fatal Flag missing values with the required output\ndf_analyze_dataset['FATAL_FLAG'].fillna(value=\"No\",inplace=True)", "C:\\Users\\amalp\\AppData\\Local\\Programs\\Python\\Python310\\lib\\site-packages\\pandas\\core\\generic.py:6392: SettingWithCopyWarning: \nA value is trying to be set on a copy of a slice from a DataFrame\n\nSee the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n return self._update_inplace(result)\n" ], [ "#Verify if the missing values are replaced\ndf_analyze_dataset.head()", "_____no_output_____" ], [ "#Check the number of observations\ndf_analyze_dataset.shape", "_____no_output_____" ] ], [ [ "#### 5. Remove all the observations where aircraft names are not available", "_____no_output_____" ] ], [ [ "#Drop the unwanted values/observations from the dataset\ndf_final_dataset = df_analyze_dataset.dropna(subset=['ACFT_MAKE_NAME'])", "_____no_output_____" ] ], [ [ "#### 6. Find the aircraft types and their occurrences in the dataset", "_____no_output_____" ] ], [ [ "#Check the number of observations now to compare it with the original dataset and see how many values have been dropped\ndf_final_dataset.shape", "_____no_output_____" ], [ "#Group the dataset by aircraft name\naircraftType = df_final_dataset.groupby('ACFT_MAKE_NAME')", "_____no_output_____" ], [ "#View the number of times each aircraft type appears in the dataset (Hint: use the size() method)\naircraftType.size()", "_____no_output_____" ] ], [ [ "#### 7: Display the observations where fatal flag is “Yes”", "_____no_output_____" ] ], [ [ "#Group the dataset by fatal flag\nfatalAccedents = df_final_dataset.groupby('FATAL_FLAG')", "_____no_output_____" ], [ "#View the total number of fatal and non-fatal accidents\nfatalAccedents.size()", "_____no_output_____" ], [ "#Create a new dataframe to view only the fatal accidents (Fatal Flag values = Yes)\naccidents_with_fatality = fatalAccedents.get_group('Yes')", "_____no_output_____" ], [ "accidents_with_fatality.head()", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "<center>\n <img src=\"https://s3-api.us-geo.objectstorage.softlayer.net/cf-courses-data/CognitiveClass/Logos/organization_logo/organization_logo.png\" width=\"300\" alt=\"cognitiveclass.ai logo\" />\n</center>\n\n# 1D Numpy in Python\n\nEstimated time needed: **30** minutes\n\n## Objectives\n\nAfter completing this lab you will be able to:\n\n- Import and use `numpy` library\n- Perform operations with `numpy`\n", "_____no_output_____" ], [ "<h2>Table of Contents</h2>\n<div class=\"alert alert-block alert-info\" style=\"margin-top: 20px\">\n <ul>\n <li><a href=\"pre\">Preparation</a></li>\n <li>\n <a href=\"numpy\">What is Numpy?</a>\n <ul>\n <li><a href=\"type\">Type</a></li>\n <li><a href=\"val\">Assign Value</a></li>\n <li><a href=\"slice\">Slicing</a></li>\n <li><a href=\"list\">Assign Value with List</a></li>\n <li><a href=\"other\">Other Attributes</a></li>\n </ul>\n </li>\n <li>\n <a href=\"op\">Numpy Array Operations</a>\n <ul>\n <li><a href=\"add\">Array Addition</a></li>\n <li><a href=\"multi\">Array Multiplication</a></li>\n <li><a href=\"prod\">Product of Two Numpy Arrays</a></li>\n <li><a href=\"dot\">Dot Product</a></li>\n <li><a href=\"cons\">Adding Constant to a Numpy Array</a></li>\n </ul>\n </li>\n <li><a href=\"math\">Mathematical Functions</a></li>\n <li><a href=\"lin\">Linspace</a></li>\n </ul>\n\n</div>\n\n<hr>\n", "_____no_output_____" ], [ "<h2 id=\"pre\">Preparation</h2>\n", "_____no_output_____" ] ], [ [ "# Import the libraries\n\nimport time \nimport sys\nimport numpy as np \n\nimport matplotlib.pyplot as plt\n%matplotlib inline ", "_____no_output_____" ], [ "# Plotting functions\n\ndef Plotvec1(u, z, v):\n \n ax = plt.axes()\n ax.arrow(0, 0, *u, head_width=0.05, color='r', head_length=0.1)\n plt.text(*(u + 0.1), 'u')\n \n ax.arrow(0, 0, *v, head_width=0.05, color='b', head_length=0.1)\n plt.text(*(v + 0.1), 'v')\n ax.arrow(0, 0, *z, head_width=0.05, head_length=0.1)\n plt.text(*(z + 0.1), 'z')\n plt.ylim(-2, 2)\n plt.xlim(-2, 2)\n\ndef Plotvec2(a,b):\n ax = plt.axes()\n ax.arrow(0, 0, *a, head_width=0.05, color ='r', head_length=0.1)\n plt.text(*(a + 0.1), 'a')\n ax.arrow(0, 0, *b, head_width=0.05, color ='b', head_length=0.1)\n plt.text(*(b + 0.1), 'b')\n plt.ylim(-2, 2)\n plt.xlim(-2, 2)", "_____no_output_____" ] ], [ [ "Create a Python List as follows:\n", "_____no_output_____" ] ], [ [ "# Create a python list\n\na = [\"0\", 1, \"two\", \"3\", 4]", "_____no_output_____" ] ], [ [ "We can access the data via an index:\n", "_____no_output_____" ], [ "<img src=\"https://s3-api.us-geo.objectstorage.softlayer.net/cf-courses-data/CognitiveClass/PY0101EN/Chapter%205/Images/NumOneList.png\" width=\"660\" />\n", "_____no_output_____" ], [ "We can access each element using a square bracket as follows: \n", "_____no_output_____" ] ], [ [ "# Print each element\n\nprint(\"a[0]:\", a[0])\nprint(\"a[1]:\", a[1])\nprint(\"a[2]:\", a[2])\nprint(\"a[3]:\", a[3])\nprint(\"a[4]:\", a[4])", "a[0]: 0\na[1]: 1\na[2]: two\na[3]: 3\na[4]: 4\n" ] ], [ [ "<hr>\n", "_____no_output_____" ], [ "<h2 id=\"numpy\">What is Numpy?</h2>\n", "_____no_output_____" ], [ "A numpy array is similar to a list. It's usually fixed in size and each element is of the same type. We can cast a list to a numpy array by first importing numpy: \n", "_____no_output_____" ] ], [ [ "# import numpy library\n\nimport numpy as np ", "_____no_output_____" ] ], [ [ " We then cast the list as follows:\n", "_____no_output_____" ] ], [ [ "# Create a numpy array\n\na = np.array([0, 1, 2, 3, 4])\na", "_____no_output_____" ] ], [ [ "Each element is of the same type, in this case integers: \n", "_____no_output_____" ], [ "<img src=\"https://s3-api.us-geo.objectstorage.softlayer.net/cf-courses-data/CognitiveClass/PY0101EN/Chapter%205/Images/NumOneNp.png\" width=\"500\" />\n", "_____no_output_____" ], [ " As with lists, we can access each element via a square bracket:\n", "_____no_output_____" ] ], [ [ "# Print each element\n\nprint(\"a[0]:\", a[0])\nprint(\"a[1]:\", a[1])\nprint(\"a[2]:\", a[2])\nprint(\"a[3]:\", a[3])\nprint(\"a[4]:\", a[4])", "a[0]: 0\na[1]: 1\na[2]: 2\na[3]: 3\na[4]: 4\n" ] ], [ [ "<h3 id=\"type\">Type</h3>\n", "_____no_output_____" ], [ "If we check the type of the array we get <b>numpy.ndarray</b>:\n", "_____no_output_____" ] ], [ [ "# Check the type of the array\n\ntype(a)", "_____no_output_____" ] ], [ [ "As numpy arrays contain data of the same type, we can use the attribute \"dtype\" to obtain the Data-type of the array’s elements. In this case a 64-bit integer: \n", "_____no_output_____" ] ], [ [ "# Check the type of the values stored in numpy array\n\na.dtype", "_____no_output_____" ] ], [ [ "We can create a numpy array with real numbers:\n", "_____no_output_____" ] ], [ [ "# Create a numpy array\n\nb = np.array([3.1, 11.02, 6.2, 213.2, 5.2])", "_____no_output_____" ] ], [ [ "When we check the type of the array we get <b>numpy.ndarray</b>:\n", "_____no_output_____" ] ], [ [ "# Check the type of array\n\ntype(b)", "_____no_output_____" ] ], [ [ "If we examine the attribute <code>dtype</code> we see float 64, as the elements are not integers: \n", "_____no_output_____" ] ], [ [ "# Check the value type\n\nb.dtype", "_____no_output_____" ] ], [ [ "<h3 id=\"val\">Assign value</h3>\n", "_____no_output_____" ], [ "We can change the value of the array, consider the array <code>c</code>:\n", "_____no_output_____" ] ], [ [ "# Create numpy array\n\nc = np.array([20, 1, 2, 3, 4])\nc", "_____no_output_____" ] ], [ [ "We can change the first element of the array to 100 as follows:\n", "_____no_output_____" ] ], [ [ "# Assign the first element to 100\n\nc[0] = 100\nc", "_____no_output_____" ] ], [ [ "We can change the 5th element of the array to 0 as follows:\n", "_____no_output_____" ] ], [ [ "# Assign the 5th element to 0\n\nc[4] = 0\nc", "_____no_output_____" ] ], [ [ "<h3 id=\"slice\">Slicing</h3>\n", "_____no_output_____" ], [ "Like lists, we can slice the numpy array, and we can select the elements from 1 to 3 and assign it to a new numpy array <code>d</code> as follows:\n", "_____no_output_____" ] ], [ [ "# Slicing the numpy array\n\nd = c[1:4]\nd", "_____no_output_____" ] ], [ [ "We can assign the corresponding indexes to new values as follows: \n", "_____no_output_____" ] ], [ [ "# Set the fourth element and fifth element to 300 and 400\n\nc[3:5] = 300, 400\nc", "_____no_output_____" ] ], [ [ "<h3 id=\"list\">Assign Value with List</h3>\n", "_____no_output_____" ], [ "Similarly, we can use a list to select a specific index.\nThe list ' select ' contains several values:\n", "_____no_output_____" ] ], [ [ "# Create the index list\n\nselect = [0, 2, 3]", "_____no_output_____" ] ], [ [ "We can use the list as an argument in the brackets. The output is the elements corresponding to the particular index:\n", "_____no_output_____" ] ], [ [ "# Use List to select elements\n\nd = c[select]\nd", "_____no_output_____" ] ], [ [ "We can assign the specified elements to a new value. For example, we can assign the values to 100 000 as follows:\n", "_____no_output_____" ] ], [ [ "# Assign the specified elements to new value\n\nc[select] = 100000\nc", "_____no_output_____" ] ], [ [ "<h3 id=\"other\">Other Attributes</h3>\n", "_____no_output_____" ], [ "Let's review some basic array attributes using the array <code>a</code>:\n", "_____no_output_____" ] ], [ [ "# Create a numpy array\n\na = np.array([0, 1, 2, 3, 4])\na", "_____no_output_____" ] ], [ [ "The attribute <code>size</code> is the number of elements in the array:\n", "_____no_output_____" ] ], [ [ "# Get the size of numpy array\n\na.size", "_____no_output_____" ] ], [ [ "The next two attributes will make more sense when we get to higher dimensions but let's review them. The attribute <code>ndim</code> represents the number of array dimensions or the rank of the array, in this case, one:\n", "_____no_output_____" ] ], [ [ "# Get the number of dimensions of numpy array\n\na.ndim", "_____no_output_____" ] ], [ [ "The attribute <code>shape</code> is a tuple of integers indicating the size of the array in each dimension:\n", "_____no_output_____" ] ], [ [ "# Get the shape/size of numpy array\n\na.shape", "_____no_output_____" ], [ "# Create a numpy array\n\na = np.array([1, -1, 1, -1])", "_____no_output_____" ], [ "# Get the mean of numpy array\n\nmean = a.mean()\nmean", "_____no_output_____" ], [ "# Get the standard deviation of numpy array\n\nstandard_deviation=a.std()\nstandard_deviation", "_____no_output_____" ], [ "# Create a numpy array\n\nb = np.array([-1, 2, 3, 4, 5])\nb", "_____no_output_____" ], [ "# Get the biggest value in the numpy array\n\nmax_b = b.max()\nmax_b", "_____no_output_____" ], [ "# Get the smallest value in the numpy array\n\nmin_b = b.min()\nmin_b", "_____no_output_____" ] ], [ [ "<hr>\n", "_____no_output_____" ], [ "<h2 id=\"op\">Numpy Array Operations</h2>\n", "_____no_output_____" ], [ "<h3 id=\"add\">Array Addition</h3>\n", "_____no_output_____" ], [ "Consider the numpy array <code>u</code>:\n", "_____no_output_____" ] ], [ [ "u = np.array([1, 0])\nu", "_____no_output_____" ] ], [ [ "Consider the numpy array <code>v</code>:\n", "_____no_output_____" ] ], [ [ "v = np.array([0, 1])\nv", "_____no_output_____" ] ], [ [ "We can add the two arrays and assign it to z:\n", "_____no_output_____" ] ], [ [ "# Numpy Array Addition\n\nz = u + v\nz", "_____no_output_____" ] ], [ [ " The operation is equivalent to vector addition:\n", "_____no_output_____" ] ], [ [ "# Plot numpy arrays\n\nPlotvec1(u, z, v)", "_____no_output_____" ] ], [ [ "<h3 id=\"multi\">Array Multiplication</h3>\n", "_____no_output_____" ], [ "Consider the vector numpy array <code>y</code>:\n", "_____no_output_____" ] ], [ [ "# Create a numpy array\n\ny = np.array([1, 2])\ny", "_____no_output_____" ] ], [ [ "We can multiply every element in the array by 2:\n", "_____no_output_____" ] ], [ [ "# Numpy Array Multiplication\n\nz = 2 * y\nz", "_____no_output_____" ] ], [ [ " This is equivalent to multiplying a vector by a scaler: \n", "_____no_output_____" ], [ "<h3 id=\"prod\">Product of Two Numpy Arrays</h3>\n", "_____no_output_____" ], [ "Consider the following array <code>u</code>:\n", "_____no_output_____" ] ], [ [ "# Create a numpy array\n\nu = np.array([1, 2])\nu", "_____no_output_____" ] ], [ [ "Consider the following array <code>v</code>:\n", "_____no_output_____" ] ], [ [ "# Create a numpy array\n\nv = np.array([3, 2])\nv", "_____no_output_____" ] ], [ [ " The product of the two numpy arrays <code>u</code> and <code>v</code> is given by:\n", "_____no_output_____" ] ], [ [ "# Calculate the production of two numpy arrays\n\nz = u * v\nz", "_____no_output_____" ] ], [ [ "<h3 id=\"dot\">Dot Product</h3>\n", "_____no_output_____" ], [ "The dot product of the two numpy arrays <code>u</code> and <code>v</code> is given by:\n", "_____no_output_____" ] ], [ [ "# Calculate the dot product\n\nnp.dot(u, v)", "_____no_output_____" ] ], [ [ "<h3 id=\"cons\">Adding Constant to a Numpy Array</h3>\n", "_____no_output_____" ], [ "Consider the following array: \n", "_____no_output_____" ] ], [ [ "# Create a constant to numpy array\n\nu = np.array([1, 2, 3, -1]) \nu", "_____no_output_____" ] ], [ [ "Adding the constant 1 to each element in the array:\n", "_____no_output_____" ] ], [ [ "# Add the constant to array\n\nu + 1", "_____no_output_____" ] ], [ [ " The process is summarised in the following animation:\n", "_____no_output_____" ], [ "<img src=\"https://s3-api.us-geo.objectstorage.softlayer.net/cf-courses-data/CognitiveClass/PY0101EN/Chapter%205/Images/NumOneAdd.gif\" width=\"500\" />\n", "_____no_output_____" ], [ "<hr>\n", "_____no_output_____" ], [ "<h2 id=\"math\">Mathematical Functions</h2>\n", "_____no_output_____" ], [ " We can access the value of <code>pi</code> in numpy as follows :\n", "_____no_output_____" ] ], [ [ "# The value of pi\n\nnp.pi", "_____no_output_____" ] ], [ [ " We can create the following numpy array in Radians:\n", "_____no_output_____" ] ], [ [ "# Create the numpy array in radians\n\nx = np.array([0, np.pi/2 , np.pi])\nx", "_____no_output_____" ] ], [ [ "We can apply the function <code>sin</code> to the array <code>x</code> and assign the values to the array <code>y</code>; this applies the sine function to each element in the array: \n", "_____no_output_____" ] ], [ [ "# Calculate the sin of each elements\n\ny = np.sin(x)\ny", "_____no_output_____" ] ], [ [ "<hr>\n", "_____no_output_____" ], [ "<h2 id=\"lin\">Linspace</h2>\n", "_____no_output_____" ], [ " A useful function for plotting mathematical functions is <code>linspace</code>. Linspace returns evenly spaced numbers over a specified interval. We specify the starting point of the sequence and the ending point of the sequence. The parameter \"num\" indicates the Number of samples to generate, in this case 5:\n", "_____no_output_____" ] ], [ [ "# Makeup a numpy array within [-2, 2] and 5 elements\n\nnp.linspace(-2, 2, num=5)", "_____no_output_____" ] ], [ [ "If we change the parameter <code>num</code> to 9, we get 9 evenly spaced numbers over the interval from -2 to 2: \n", "_____no_output_____" ] ], [ [ "# Makeup a numpy array within [-2, 2] and 9 elements\n\nnp.linspace(-2, 2, num=9)", "_____no_output_____" ] ], [ [ "We can use the function <code>linspace</code> to generate 100 evenly spaced samples from the interval 0 to 2π: \n", "_____no_output_____" ] ], [ [ "# Makeup a numpy array within [0, 2π] and 100 elements \n\nx = np.linspace(0, 2*np.pi, num=100)", "_____no_output_____" ] ], [ [ "We can apply the sine function to each element in the array <code>x</code> and assign it to the array <code>y</code>: \n", "_____no_output_____" ] ], [ [ "# Calculate the sine of x list\n\ny = np.sin(x)", "_____no_output_____" ], [ "# Plot the result\n\nplt.plot(x, y)", "_____no_output_____" ] ], [ [ "<hr>\n", "_____no_output_____" ], [ "<h2 id=\"quiz\">Quiz on 1D Numpy Array</h2>\n", "_____no_output_____" ], [ "Implement the following vector subtraction in numpy: u-v\n", "_____no_output_____" ] ], [ [ "# Write your code below and press Shift+Enter to execute\n\nu = np.array([1, 0])\nv = np.array([0, 1])\nu-v", "_____no_output_____" ] ], [ [ "<details><summary>Click here for the solution</summary>\n\n```python\nu - v\n```\n\n</details>\n", "_____no_output_____" ], [ "<hr>\n", "_____no_output_____" ], [ "Multiply the numpy array z with -2:\n", "_____no_output_____" ] ], [ [ "# Write your code below and press Shift+Enter to execute\n\nz = np.array([2, 4])\n-2*z", "_____no_output_____" ] ], [ [ "<details><summary>Click here for the solution</summary>\n\n```python\n-2 * z\n```\n\n</details>\n", "_____no_output_____" ], [ "<hr>\n", "_____no_output_____" ], [ "Consider the list <code>[1, 2, 3, 4, 5]</code> and <code>[1, 0, 1, 0, 1]</code>, and cast both lists to a numpy array then multiply them together:\n", "_____no_output_____" ] ], [ [ "# Write your code below and press Shift+Enter to execute\na = np.array([1, 2, 3, 4, 5])\nb = np.array([1, 0, 1, 0, 1])\na*b", "_____no_output_____" ] ], [ [ "<details><summary>Click here for the solution</summary>\n\n```python\na = np.array([1, 2, 3, 4, 5])\nb = np.array([1, 0, 1, 0, 1])\na * b\n```\n\n</details>\n", "_____no_output_____" ], [ "<hr>\n", "_____no_output_____" ], [ "Convert the list <code>[-1, 1]</code> and <code>[1, 1]</code> to numpy arrays <code>a</code> and <code>b</code>. Then, plot the arrays as vectors using the fuction <code>Plotvec2</code> and find the dot product:\n", "_____no_output_____" ] ], [ [ "# Write your code below and press Shift+Enter to execute\na = np.array([-1, 1])\nb = np.array([1, 1])\nPlotvec2(a, b)\nprint(\"The dot product is\", np.dot(a,b))", "The dot product is 0\n" ] ], [ [ "<details><summary>Click here for the solution</summary>\n\n```python\na = np.array([-1, 1])\nb = np.array([1, 1])\nPlotvec2(a, b)\nprint(\"The dot product is\", np.dot(a,b))\n\n```\n\n</details>\n", "_____no_output_____" ], [ "<hr>\n", "_____no_output_____" ], [ "Convert the list <code>[1, 0]</code> and <code>[0, 1]</code> to numpy arrays <code>a</code> and <code>b</code>. Then, plot the arrays as vectors using the function <code>Plotvec2</code> and find the dot product:\n", "_____no_output_____" ] ], [ [ "# Write your code below and press Shift+Enter to execute\na = np.array([1, 0])\nb = np.array([0, 1])\nPlotvec2(a, b)\nprint(\"The dot product is\", np.dot(a,b))", "The dot product is 0\n" ] ], [ [ "<details><summary>Click here for the solution</summary>\n\n```python\na = np.array([1, 0])\nb = np.array([0, 1])\nPlotvec2(a, b)\nprint(\"The dot product is\", np.dot(a, b))\n\n```\n\n</details>\n", "_____no_output_____" ], [ "<hr>\n", "_____no_output_____" ], [ "Convert the list <code>[1, 1]</code> and <code>[0, 1]</code> to numpy arrays <code>a</code> and <code>b</code>. Then plot the arrays as vectors using the fuction <code>Plotvec2</code> and find the dot product:\n", "_____no_output_____" ] ], [ [ "# Write your code below and press Shift+Enter to execute\na = np.array([1, 1])\nb = np.array([0, 1])\nPlotvec2(a, b)\nprint(\"The dot product is\", np.dot(a,b))", "The dot product is 1\n" ] ], [ [ "<details><summary>Click here for the solution</summary>\n\n```python\na = np.array([1, 1])\nb = np.array([0, 1])\nPlotvec2(a, b)\nprint(\"The dot product is\", np.dot(a, b))\nprint(\"The dot product is\", np.dot(a, b))\n\n```\n\n</details>\n \n", "_____no_output_____" ], [ "<hr>\n", "_____no_output_____" ], [ "Why are the results of the dot product for <code>[-1, 1]</code> and <code>[1, 1]</code> and the dot product for <code>[1, 0]</code> and <code>[0, 1]</code> zero, but not zero for the dot product for <code>[1, 1]</code> and <code>[0, 1]</code>? <p><i>Hint: Study the corresponding figures, pay attention to the direction the arrows are pointing to.</i></p>\n", "_____no_output_____" ] ], [ [ "# Write your code below and press Shift+Enter to execute\n", "_____no_output_____" ] ], [ [ "<details><summary>Click here for the solution</summary>\n\n```python\nThe vectors used for question 4 and 5 are perpendicular. As a result, the dot product is zero. \n\n```\n\n</details>\n \n", "_____no_output_____" ], [ "<hr>\n<h2>The last exercise!</h2>\n<p>Congratulations, you have completed your first lesson and hands-on lab in Python. However, there is one more thing you need to do. The Data Science community encourages sharing work. The best way to share and showcase your work is to share it on GitHub. By sharing your notebook on GitHub you are not only building your reputation with fellow data scientists, but you can also show it off when applying for a job. Even though this was your first piece of work, it is never too early to start building good habits. So, please read and follow <a href=\"https://cognitiveclass.ai/blog/data-scientists-stand-out-by-sharing-your-notebooks/\" target=\"_blank\">this article</a> to learn how to share your work.\n<hr>\n", "_____no_output_____" ], [ "## Author\n\n<a href=\"https://www.linkedin.com/in/joseph-s-50398b136/\" target=\"_blank\">Joseph Santarcangelo</a>\n\n## Other contributors\n\n<a href=\"www.linkedin.com/in/jiahui-mavis-zhou-a4537814a\">Mavis Zhou</a>\n\n## Change Log\n\n| Date (YYYY-MM-DD) | Version | Changed By | Change Description |\n| ----------------- | ------- | ---------- | ---------------------------------- |\n| 2020-08-26 | 2.0 | Lavanya | Moved lab to course repo in GitLab |\n| | | | |\n| | | | |\n\n<hr/>\n\n## <h3 align=\"center\"> © IBM Corporation 2020. All rights reserved. <h3/>\n", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown" ]

[ [ "markdown", "markdown", "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code" ], [ "markdown", "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Semi-Monocoque Theory: corrective solutions", "_____no_output_____" ] ], [ [ "from pint import UnitRegistry\nimport sympy\nimport networkx as nx\nimport numpy as np\nimport matplotlib.pyplot as plt\nimport sys\n%matplotlib inline\nfrom IPython.display import display", "_____no_output_____" ] ], [ [ "Import **Section** class, which contains all calculations", "_____no_output_____" ] ], [ [ "from Section import Section", "_____no_output_____" ] ], [ [ "Initialization of **sympy** symbolic tool and **pint** for dimension analysis (not really implemented rn as not directly compatible with sympy)", "_____no_output_____" ] ], [ [ "ureg = UnitRegistry()\nsympy.init_printing()", "_____no_output_____" ] ], [ [ "Define **sympy** parameters used for geometric description of sections", "_____no_output_____" ] ], [ [ "A, A0, t, t0, a, b, h, L, E, G = sympy.symbols('A A_0 t t_0 a b h L E G', positive=True)", "_____no_output_____" ] ], [ [ "We also define numerical values for each **symbol** in order to plot scaled section and perform calculations", "_____no_output_____" ] ], [ [ "values = [(A, 150 * ureg.millimeter**2),(A0, 250 * ureg.millimeter**2),(a, 80 * ureg.millimeter), \\\n (b, 20 * ureg.millimeter),(h, 35 * ureg.millimeter),(L, 2000 * ureg.millimeter), \\\n (t, 0.8 *ureg.millimeter),(E, 72e3 * ureg.MPa), (G, 27e3 * ureg.MPa)]\ndatav = [(v[0],v[1].magnitude) for v in values]", "_____no_output_____" ] ], [ [ "# First example: Simple rectangular symmetric section", "_____no_output_____" ], [ "Define graph describing the section:\n\n1) **stringers** are **nodes** with parameters:\n- **x** coordinate\n- **y** coordinate\n- **Area**\n\n2) **panels** are **oriented edges** with parameters:\n- **thickness**\n- **lenght** which is automatically calculated", "_____no_output_____" ] ], [ [ "stringers = {1:[(2*a,h),A],\n 2:[(a,h),A],\n 3:[(sympy.Integer(0),h),A],\n 4:[(sympy.Integer(0),sympy.Integer(0)),A],\n 5:[(2*a,sympy.Integer(0)),A]}\n #5:[(sympy.Rational(1,2)*a,h),A]}\n\npanels = {(1,2):t,\n (2,3):t,\n (3,4):t,\n (4,5):t,\n (5,1):t}", "_____no_output_____" ] ], [ [ "Define section and perform first calculations", "_____no_output_____" ] ], [ [ "S1 = Section(stringers, panels)", "_____no_output_____" ], [ "S1.cycles", "_____no_output_____" ] ], [ [ "## Plot of **S1** section in original reference frame", "_____no_output_____" ], [ "Define a dictionary of coordinates used by **Networkx** to plot section as a Directed graph.\nNote that arrows are actually just thicker stubs", "_____no_output_____" ] ], [ [ "start_pos={ii: [float(S1.g.node[ii]['ip'][i].subs(datav)) for i in range(2)] for ii in S1.g.nodes() }", "_____no_output_____" ], [ "plt.figure(figsize=(12,8),dpi=300)\nnx.draw(S1.g,with_labels=True, arrows= True, pos=start_pos)\nplt.arrow(0,0,20,0)\nplt.arrow(0,0,0,20)\n#plt.text(0,0, 'CG', fontsize=24)\nplt.axis('equal')\nplt.title(\"Section in starting reference Frame\",fontsize=16);", "_____no_output_____" ] ], [ [ "## Plot of **S1** section in inertial reference Frame", "_____no_output_____" ], [ "Section is plotted wrt **center of gravity** and rotated (if necessary) so that *x* and *y* are principal axes.\n**Center of Gravity** and **Shear Center** are drawn", "_____no_output_____" ] ], [ [ "positions={ii: [float(S1.g.node[ii]['pos'][i].subs(datav)) for i in range(2)] for ii in S1.g.nodes() }", "_____no_output_____" ], [ "x_ct, y_ct = S1.ct.subs(datav)\n\nplt.figure(figsize=(12,8),dpi=300)\nnx.draw(S1.g,with_labels=True, pos=positions)\nplt.plot([0],[0],'o',ms=12,label='CG')\nplt.plot([x_ct],[y_ct],'^',ms=12, label='SC')\n#plt.text(0,0, 'CG', fontsize=24)\n#plt.text(x_ct,y_ct, 'SC', fontsize=24)\nplt.legend(loc='lower right', shadow=True)\nplt.axis('equal')\nplt.title(\"Section in pricipal reference Frame\",fontsize=16);", "_____no_output_____" ] ], [ [ "Compute **L** matrix: with 5 nodes we expect 2 **dofs**, one with _symmetric load_ and one with _antisymmetric load_", "_____no_output_____" ] ], [ [ "S1.compute_L()", "_____no_output_____" ], [ "S1.L", "_____no_output_____" ] ], [ [ "Compute **H** matrix", "_____no_output_____" ] ], [ [ "S1.compute_H()", "_____no_output_____" ], [ "S1.H", "_____no_output_____" ] ], [ [ "Compute $\\tilde{K}$ and $\\tilde{M}$ as:\n\n$$\\tilde{K} = L^T \\cdot \\left[ \\frac{A}{A_0} \\right] \\cdot L$$\n$$\\tilde{M} = H^T \\cdot \\left[ \\frac{l}{l_0}\\frac{t_0}{t} \\right] \\cdot L$$", "_____no_output_____" ] ], [ [ "S1.compute_KM(A,h,t)", "_____no_output_____" ], [ "S1.Ktilde", "_____no_output_____" ], [ "S1.Mtilde", "_____no_output_____" ] ], [ [ "Compute **eigenvalues** and **eigenvectors** as:\n\n$$\\left| \\mathbf{I} \\cdot \\beta^2 - \\mathbf{\\tilde{K}}^{-1} \\cdot \\mathbf{\\tilde{M}} \\right| = 0$$\n\nWe substitute some numerical values to simplify the expressions", "_____no_output_____" ] ], [ [ "sol_data = (S1.Ktilde.inv()*(S1.Mtilde.subs(datav))).eigenvects()", "_____no_output_____" ] ], [ [ "**Eigenvalues** correspond to $\\beta^2$", "_____no_output_____" ] ], [ [ "β2 = [sol[0] for sol in sol_data]\nβ2", "_____no_output_____" ] ], [ [ "**Eigenvectors** are orthogonal as expected", "_____no_output_____" ] ], [ [ "X = [sol[2][0] for sol in sol_data]\nX", "_____no_output_____" ] ], [ [ "From $\\beta_i^2$ we compute:\n$$\\lambda_i = \\sqrt{\\frac{E A_0 l_0}{G t_0} \\beta_i^2}$$\n\nsubstuting numerical values", "_____no_output_____" ] ], [ [ "λ = [sympy.N(sympy.sqrt(E*A*h/(G*t)*βi).subs(datav)) for βi in β2]\nλ", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import os\nimport json\n\nimport numpy as np\nimport pandas as pd\nfrom pandas.io.json import json_normalize \n\nfile_path = \"../../data/raw/yp_competitors.json\"\nif os.path.exists(file_path): \n dataset = pd.read_json(file_path, orient='columns') \nelse: \n print(\"file not yet created\")\n \ndataset.head()", "file not yet created\n" ], [ "dataset2 = pd.read_csv('../../data/raw/yp_competitors.csv')\ndataset2[dataset2.location_state == 'HI'].alias.to_csv('../../data/raw/hi_competitors.csv', index=False)", "_____no_output_____" ], [ "# restaurants, 'seafood', 'desserts', 'vegetarian', \n# if i['alias'] in ['restaurants', 'seafood', 'desserts', 'vegetarian']\ncateg = dataset.categories.apply(lambda x: [i['alias'] for i in x])\n\ncateg2 = categ.apply(\n lambda x: \n x \n if len([i for i in x if i in ['restaurants', 'seafood', 'desserts', 'vegetarian']]) > 0 \n else []\n)\n\ncateg2[:10]", "_____no_output_____" ], [ "len([i for i in categ2 if len(i) > 0])", "_____no_output_____" ], [ "# fixing columns\n# transactions \ndataset.transactions = dataset.transactions.apply(lambda x: ','.join(x))\n\n# category\npd_categories_alias = dataset.categories.apply(lambda x: ', '.join([i['alias'] for i in x]))\npd_categories_title = dataset.categories.apply(lambda x: ', '.join([i['title'] for i in x]))\npd_categories_alias.name = 'category_' + 'alias'\npd_categories_title.name = 'category_' + 'title'", "_____no_output_____" ], [ "# expanding json columns\n# coordinate\npd_coordinates = dataset.coordinates.apply(pd.Series)\npd_coordinates.columns = 'coordinate_' + pd_coordinates.columns\n\n# location\npd_location = dataset.location.apply(pd.Series)\npd_location.display_address = pd_location.display_address.apply(lambda x: ','.join(x))\npd_location.columns = 'location_' + pd_location.columns", "_____no_output_____" ], [ "# merging all in one\ndataset2 = pd.concat(\n [dataset, pd_categories_alias, pd_categories_title, pd_coordinates, pd_location], \n axis=1\n).drop(['categories', 'location', 'coordinates'], axis=1)\ndataset2.columns\ndataset = dataset2", "_____no_output_____" ], [ "# from JSON to CSV\ndataset2.to_csv('../../data/raw/yp_competitors.csv', index=False)\npd.read_csv('../../data/raw/yp_competitors.csv').head()", "_____no_output_____" ] ], [ [ "## Examining Total Category List", "_____no_output_____" ] ], [ [ "# restaurants categories in the dataset: all\nprint(sorted(set(', '.join(dataset.category_alias).split(', '))))", "['acaibowls', 'accessories', 'active', 'afghani', 'airportlounges', 'amateursportsteams', 'amusementparks', 'antiques', 'aquariums', 'aquariumservices', 'arabian', 'arcades', 'argentine', 'artclasses', 'artmuseums', 'artsandcrafts', 'asianfusion', 'australian', 'bagels', 'bakeries', 'banks', 'bars', 'baseballfields', 'basketballcourts', 'basque', 'bbq', 'beachequipmentrental', 'beaches', 'beer_and_wine', 'beerbar', 'beergardens', 'bike_repair_maintenance', 'bikerentals', 'bikes', 'biketours', 'boatcharters', 'boating', 'boattours', 'bookstores', 'bootcamps', 'bowling', 'brasseries', 'brazilian', 'breakfast_brunch', 'breweries', 'brewpubs', 'british', 'bubbletea', 'buffets', 'burgers', 'burmese', 'butcher', 'cafes', 'cafeteria', 'cajun', 'cakeshop', 'campgrounds', 'candy', 'cantonese', 'caribbean', 'casinos', 'catering', 'cheese', 'cheesesteaks', 'chicken_wings', 'chickenshop', 'childrensmuseums', 'chinese', 'chocolate', 'cideries', 'cigarbars', 'climbing', 'clubcrawl', 'cocktailbars', 'coffee', 'coffeeroasteries', 'collegeuniv', 'colombian', 'comedyclubs', 'comfortfood', 'convenience', 'conveyorsushi', 'cookingclasses', 'cosmetics', 'countryclubs', 'countrydancehalls', 'couriers', 'creperies', 'cuban', 'cupcakes', 'customcakes', 'czech', 'danceclubs', 'daycamps', 'delis', 'desserts', 'dimsum', 'diners', 'dinnertheater', 'discgolf', 'divebars', 'diving', 'diyfood', 'dog_parks', 'donuts', 'drugstores', 'education', 'educationservices', 'empanadas', 'eventplanning', 'fabricstores', 'falafel', 'farmersmarket', 'festivals', 'filipino', 'fishing', 'fishnchips', 'fitness', 'fleamarkets', 'flowers', 'fondue', 'food', 'food_court', 'fooddeliveryservices', 'foodstands', 'foodtours', 'foodtrucks', 'football', 'french', 'galleries', 'gardens', 'gastropubs', 'gaybars', 'gelato', 'german', 'giftshops', 'gluten_free', 'golf', 'golfequipment', 'golflessons', 'gourmet', 'greek', 'grocery', 'halal', 'hauntedhouses', 'hawaiian', 'headshops', 'healthmarkets', 'healthtrainers', 'herbsandspices', 'hiking', 'himalayan', 'hindu_temples', 'hkcafe', 'hobbyshops', 'homestaging', 'honey', 'hookah_bars', 'horsebackriding', 'hostels', 'hotdog', 'hotdogs', 'hotels', 'hotpot', 'icecream', 'importedfood', 'indpak', 'internetcafe', 'intlgrocery', 'irish', 'irish_pubs', 'italian', 'izakaya', 'japacurry', 'japanese', 'jazzandblues', 'jewelry', 'juicebars', 'karaoke', 'kebab', 'kids_activities', 'kombucha', 'korean', 'kosher', 'lakes', 'landmarks', 'laotian', 'lasertag', 'latin', 'lebanese', 'localflavor', 'localservices', 'lounges', 'macarons', 'markets', 'massage', 'massage_therapy', 'matchmakers', 'meats', 'medcenters', 'media', 'meditationcenters', 'mediterranean', 'menscloth', 'mexican', 'mideastern', 'modern_european', 'mongolian', 'mountainbiking', 'museums', 'musicvenues', 'newamerican', 'newmexican', 'nightlife', 'nonprofit', 'noodles', 'oliveoil', 'organic_stores', 'outdoorgear', 'outdoormovies', 'paddleboarding', 'paintyourownpottery', 'pakistani', 'panasian', 'parks', 'partysupplies', 'pastashops', 'patiocoverings', 'persian', 'personalchefs', 'peruvian', 'pets', 'pettingzoos', 'pizza', 'playgrounds', 'poke', 'polish', 'polynesian', 'poolhalls', 'poolservice', 'popcorn', 'popuprestaurants', 'portuguese', 'publicart', 'publicservicesgovt', 'pubs', 'puertorican', 'rafting', 'ramen', 'raw_food', 'recreation', 'religiousitems', 'resorts', 'restaurants', 'rock_climbing', 'russian', 'salad', 'salvadoran', 'sandwiches', 'scottish', 'scuba', 'seafood', 'seafoodmarkets', 'servicestations', 'shanghainese', 'shavedice', 'shopping', 'shoppingcenters', 'sicilian', 'singaporean', 'skate_parks', 'skiresorts', 'sledding', 'smokehouse', 'snorkeling', 'social_clubs', 'soulfood', 'soup', 'southern', 'souvenirs', 'spanish', 'speakeasies', 'specialtyschools', 'spiritual_shop', 'sportgoods', 'sports_clubs', 'sportsbars', 'sportswear', 'srilankan', 'stationery', 'steak', 'streetvendors', 'summer_camps', 'surfing', 'surfshop', 'sushi', 'swimmingpools', 'szechuan', 'tacos', 'taiwanese', 'tapas', 'tapasmallplates', 'tcm', 'tea', 'tennis', 'teppanyaki', 'tex-mex', 'thai', 'theater', 'themedcafes', 'tikibars', 'tobaccoshops', 'tours', 'toys', 'tradamerican', 'trampoline', 'travelagents', 'travelservices', 'tubing', 'turkish', 'ukrainian', 'unofficialyelpevents', 'vacation_rentals', 'vapeshops', 'vegan', 'vegetarian', 'venues', 'vietnamese', 'vitaminssupplements', 'waffles', 'walkingtours', 'waterstores', 'wedding_planning', 'whiskeybars', 'wholesale_stores', 'wine_bars', 'womenscloth', 'wraps', 'yelpevents', 'yoga', 'zipline', 'zoos']\n" ], [ "yelp_categories = pd.read_json(\"../../data/raw/categories.json\")\n\n# select only one parent categories\nyelp_categories = yelp_categories[yelp_categories.parents.apply(lambda x: len(x) == 1) == True]\nres_list = yelp_categories[yelp_categories.parents.apply(lambda x: len([i for i in x if i in 'restaurants']) > 0)]\nres_list = res_list.alias\nprint(set(res_list))", "{'newamerican', 'nightfood', 'salad', 'bistros', 'beergarden', 'popuprestaurants', 'freiduria', 'schnitzel', 'eastern_european', 'beisl', 'british', 'korean', 'guamanian', 'russian', 'turkish', 'uzbek', 'pakistani', 'delis', 'hungarian', 'singaporean', 'tex-mex', 'latin', 'honduran', 'portuguese', 'scottish', 'african', 'diners', 'brasseries', 'czechslovakian', 'breakfast_brunch', 'yugoslav', 'cafeteria', 'international', 'newmexican', 'trattorie', 'cajun', 'signature_cuisine', 'swabian', 'cafes', 'norwegian', 'tapas', 'flatbread', 'cheesesteaks', 'iberian', 'kurdish', 'fischbroetchen', 'raw_food', 'parma', 'halal', 'peruvian', 'comfortfood', 'israeli', 'argentine', 'taiwanese', 'danish', 'waffles', 'dumplings', 'irish', 'french', 'asianfusion', 'filipino', 'tapasmallplates', 'fishnchips', 'bavarian', 'greek', 'japanese', 'himalayan', 'vegetarian', 'newzealand', 'foodstands', 'ukrainian', 'malaysian', 'hkcafe', 'moroccan', 'catalan', 'galician', 'meatballs', 'somali', 'traditional_swedish', 'bbq', 'newcanadian', 'persian', 'gamemeat', 'australian', 'laos', 'potatoes', 'fondue', 'gluten_free', 'cambodian', 'lyonnais', 'sandwiches', 'kopitiam', 'seafood', 'soulfood', 'wraps', 'milkbars', 'venison', 'giblets', 'beerhall', 'vegan', 'hawaiian', 'italian', 'swedish', 'tradamerican', 'currysausage', 'asturian', 'hotdog', 'arabian', 'jewish', 'steak', 'pizza', 'nicaraguan', 'slovakian', 'hotpot', 'tavolacalda', 'panasian', 'eritrean', 'pfcomercial', 'czech', 'dinnertheater', 'brazilian', 'burmese', 'creperies', 'burgers', 'kosher', 'corsican', 'polynesian', 'canteen', 'rotisserie_chicken', 'serbocroatian', 'georgian', 'wok', 'food_court', 'srilankan', 'ethiopian', 'sud_ouest', 'vietnamese', 'blacksea', 'caribbean', 'german', 'kebab', 'southern', 'cuban', 'cypriot', 'mideastern', 'laotian', 'belgian', 'nikkei', 'noodles', 'bangladeshi', 'bulgarian', 'gastropubs', 'hotdogs', 'opensandwiches', 'chicken_wings', 'indonesian', 'basque', 'sushi', 'thai', 'norcinerie', 'oriental', 'swissfood', 'austrian', 'mongolian', 'romanian', 'heuriger', 'soup', 'chinese', 'poutineries', 'armenian', 'island_pub', 'baguettes', 'chickenshop', 'mediterranean', 'supperclubs', 'buffets', 'syrian', 'modern_australian', 'afghani', 'chilean', 'indpak', 'riceshop', 'pita', 'spanish', 'andalusian', 'tabernas', 'mexican', 'pubfood', 'scandinavian', 'modern_european', 'polish'}\n" ], [ "'arabian' in res_list.values", "_____no_output_____" ], [ "dataset.category_alias.head()", "_____no_output_____" ], [ "df_res_list = dataset[dataset.category_alias.apply(lambda x: len([i for i in x.split(', ') if i in res_list.values]) > 0)]\nlen(df_res_list)", "_____no_output_____" ], [ "df_res_list.head()", "_____no_output_____" ], [ "from itertools import chain\nset(chain(*[i.split(',') for i in set(dataset.transactions)]))", "_____no_output_____" ], [ "df_transactions = dataset[dataset.transactions.apply(lambda x: len(x) > 0)]\nlen(df_transactions)", "_____no_output_____" ], [ "df_transactions_not = dataset[dataset.transactions.apply(lambda x: len(x) == 0)]\nlen(df_transactions_not)", "_____no_output_____" ], [ "# restaurants? [BUSINESSES with TRANSACTION]\nprint(set(', '.join(df_transactions.category_alias).split(', ')))", "{'newamerican', 'salad', 'breweries', 'british', 'korean', 'turkish', 'russian', 'pakistani', 'delis', 'singaporean', 'tex-mex', 'latin', 'cupcakes', 'seafoodmarkets', 'teppanyaki', 'bagels', 'diners', 'lounges', 'brasseries', 'oliveoil', 'breakfast_brunch', 'poke', 'newmexican', 'healthmarkets', 'shavedice', 'bakeries', 'cajun', 'tea', 'cafes', 'tapas', 'gelato', 'cocktailbars', 'cheesesteaks', 'raw_food', 'peruvian', 'halal', 'comfortfood', 'argentine', 'meats', 'taiwanese', 'icecream', 'waffles', 'wine_bars', 'izakaya', 'irish', 'beer_and_wine', 'french', 'asianfusion', 'tapasmallplates', 'bubbletea', 'sportsbars', 'fishnchips', 'greek', 'catering', 'japanese', 'vegetarian', 'himalayan', 'divebars', 'foodstands', 'vapeshops', 'hkcafe', 'szechuan', 'bbq', 'persian', 'australian', 'poolhalls', 'customcakes', 'gluten_free', 'sandwiches', 'seafood', 'wraps', 'vegan', 'italian', 'hawaiian', 'tradamerican', 'tacos', 'donuts', 'acaibowls', 'steak', 'pizza', 'lebanese', 'cakeshop', 'sicilian', 'hotpot', 'japacurry', 'panasian', 'dinnertheater', 'brazilian', 'creperies', 'burgers', 'grocery', 'beerbar', 'salvadoran', 'vitaminssupplements', 'desserts', 'diyfood', 'vietnamese', 'german', 'caribbean', 'southern', 'cuban', 'irish_pubs', 'mideastern', 'foodtrucks', 'laotian', 'pastashops', 'empanadas', 'countrydancehalls', 'noodles', 'tikibars', 'gastropubs', 'hotdogs', 'chicken_wings', 'sushi', 'basque', 'thai', 'ramen', 'hookah_bars', 'venues', 'musicvenues', 'pubs', 'candy', 'whiskeybars', 'internetcafe', 'coffee', 'falafel', 'bars', 'soup', 'beergardens', 'chinese', 'chickenshop', 'mediterranean', 'buffets', 'dimsum', 'afghani', 'indpak', 'importedfood', 'spanish', 'mexican', 'juicebars', 'modern_european', 'polish', 'eventplanning', 'organic_stores'}\n" ], [ "# not restaurants? [BUSINESSES w/o TRANSACTION]\nprint(set(', '.join(df_transactions_not.category_alias).split(', ')))", "{'newamerican', 'poolservice', 'salad', 'museums', 'breweries', 'golflessons', 'stationery', 'countryclubs', 'yoga', 'education', 'airportlounges', 'playgrounds', 'waterstores', 'drugstores', 'popuprestaurants', 'specialtyschools', 'arcades', 'meditationcenters', 'beachequipmentrental', 'rafting', 'toys', 'british', 'markets', 'korean', 'turkish', 'trampoline', 'russian', 'pakistani', 'delis', 'healthtrainers', 'summer_camps', 'outdoormovies', 'tex-mex', 'nightlife', 'herbsandspices', 'localservices', 'artsandcrafts', 'latin', 'travelagents', 'cupcakes', 'daycamps', 'seafoodmarkets', 'shanghainese', 'portuguese', 'gaybars', 'horsebackriding', 'teppanyaki', 'travelservices', 'mountainbiking', 'outdoorgear', 'scottish', 'active', 'bagels', 'lounges', 'diners', 'sledding', 'yelpevents', 'breakfast_brunch', 'sports_clubs', 'chocolate', 'pets', 'cafeteria', 'fooddeliveryservices', 'beaches', 'poke', 'discgolf', 'newmexican', 'souvenirs', 'healthmarkets', 'shavedice', 'bakeries', 'giftshops', 'cajun', 'hobbyshops', 'unofficialyelpevents', 'basque', 'medcenters', 'menscloth', 'popcorn', 'tea', 'intlgrocery', 'cafes', 'cideries', 'gelato', 'tapas', 'cocktailbars', 'cheesesteaks', 'publicservicesgovt', 'shoppingcenters', 'hindu_temples', 'raw_food', 'peruvian', 'halal', 'biketours', 'comfortfood', 'scuba', 'gardens', 'massage', 'argentine', 'meats', 'tours', 'taiwanese', 'icecream', 'waffles', 'wine_bars', 'karaoke', 'foodtours', 'casinos', 'bike_repair_maintenance', 'artmuseums', 'izakaya', 'irish', 'bikes', 'womenscloth', 'beer_and_wine', 'theater', 'french', 'asianfusion', 'filipino', 'tapasmallplates', 'butcher', 'bubbletea', 'sportsbars', 'restaurants', 'greek', 'catering', 'japanese', 'vegetarian', 'fishnchips', 'divebars', 'foodstands', 'dog_parks', 'fishing', 'cantonese', 'ukrainian', 'szechuan', 'campgrounds', 'smokehouse', 'social_clubs', 'bbq', 'golf', 'amusementparks', 'football', 'food', 'cheese', 'persian', 'lasertag', 'conveyorsushi', 'australian', 'boattours', 'poolhalls', 'fondue', 'customcakes', 'gluten_free', 'sandwiches', 'recreation', 'seafood', 'comedyclubs', 'parks', 'soulfood', 'rock_climbing', 'boatcharters', 'wraps', 'galleries', 'media', 'skiresorts', 'sportswear', 'vegan', 'hawaiian', 'tacos', 'tradamerican', 'italian', 'donuts', 'skate_parks', 'coffeeroasteries', 'spiritual_shop', 'educationservices', 'hotdog', 'acaibowls', 'arabian', 'steak', 'vacation_rentals', 'pizza', 'streetvendors', 'cigarbars', 'lakes', 'cakeshop', 'flowers', 'colombian', 'macarons', 'hotpot', 'homestaging', 'sicilian', 'surfshop', 'festivals', 'japacurry', 'swimmingpools', 'servicestations', 'cookingclasses', 'brewpubs', 'jewelry', 'partysupplies', 'panasian', 'fleamarkets', 'dinnertheater', 'golfequipment', 'czech', 'resorts', 'creperies', 'burgers', 'brazilian', 'kosher', 'polynesian', 'sportgoods', 'headshops', 'burmese', 'grocery', 'paddleboarding', 'clubcrawl', 'hotels', 'surfing', 'publicart', 'gourmet', 'food_court', 'srilankan', 'beerbar', 'themedcafes', 'fabricstores', 'salvadoran', 'hostels', 'fitness', 'walkingtours', 'bootcamps', 'aquariumservices', 'desserts', 'diyfood', 'vietnamese', 'german', 'caribbean', 'southern', 'kebab', 'religiousitems', 'cuban', 'irish_pubs', 'mideastern', 'foodtrucks', 'pastashops', 'tcm', 'eventplanning', 'countrydancehalls', 'snorkeling', 'localflavor', 'bookstores', 'noodles', 'tikibars', 'danceclubs', 'gastropubs', 'hotdogs', 'farmersmarket', 'bikerentals', 'banks', 'landmarks', 'chicken_wings', 'sushi', 'couriers', 'tubing', 'matchmakers', 'thai', 'boating', 'kombucha', 'ramen', 'diving', 'speakeasies', 'patiocoverings', 'wholesale_stores', 'honey', 'wedding_planning', 'hookah_bars', 'bowling', 'venues', 'mongolian', 'zoos', 'musicvenues', 'candy', 'whiskeybars', 'pubs', 'artclasses', 'antiques', 'paintyourownpottery', 'internetcafe', 'shopping', 'coffee', 'falafel', 'baseballfields', 'accessories', 'kids_activities', 'bars', 'massage_therapy', 'hiking', 'soup', 'beergardens', 'chinese', 'amateursportsteams', 'nonprofit', 'chickenshop', 'mediterranean', 'aquariums', 'puertorican', 'buffets', 'cosmetics', 'jazzandblues', 'zipline', 'dimsum', 'indpak', 'tobaccoshops', 'climbing', 'hauntedhouses', 'importedfood', 'childrensmuseums', 'convenience', 'basketballcourts', 'spanish', 'mexican', 'tennis', 'juicebars', 'modern_european', 'collegeuniv', 'pettingzoos', 'organic_stores', 'personalchefs'}\n" ], [ "print(f\"{len(dataset)}\")\nprint(f\"{len(dataset.alias.unique())}\")", "6664\n6664\n" ], [ "dataset[dataset.alias == \"kimos-maui-lahaina\"]\ndataset[(dataset.dist_to_alias == \"kimos-maui-lahaina\") & (dataset.distance < 50)]\n", "_____no_output_____" ], [ "pd_location.columns = 'loc_' + pd_location.columns\npd_location.columns", "_____no_output_____" ], [ "import matplotlib.pyplot as plt\ndataset.review_count.plot()\nplt.show()", "_____no_output_____" ], [ "yelp_branches = [\n 'kimos-maui-lahaina',\n 'sunnyside-tahoe-city-2',\n 'dukes-huntington-beach-huntington-beach-2',\n 'dukes-la-jolla-la-jolla',\n 'dukes-malibu-malibu-2',\n 'dukes-beach-house-lahaina',\n 'dukes-kauai-lihue-3',\n 'dukes-waikiki-honolulu-2',\n 'hula-grill-waikiki-honolulu-3',\n 'hula-grill-kaanapali-lahaina-2',\n 'keokis-paradise-koloa',\n 'leilanis-lahaina-2'\n]\n[i for i in dataset.alias.values if i in yelp_branches]", "_____no_output_____" ] ], [ [ "## Exploratory Data Analysis", "_____no_output_____" ] ], [ [ "print(dataset.loc[dataset.alias.isin(yelp_branches)].rating.sum())\nprint(dataset.loc[dataset.alias.isin(yelp_branches)].rating.mean())", "47.5\n3.9583333333333335\n" ], [ "len(dataset)", "_____no_output_____" ], [ "dataset.is_closed[dataset.is_closed == True].count()", "_____no_output_____" ], [ "dataset.price.value_counts()", "_____no_output_____" ], [ "print(f\"sum : {dataset.review_count.sum()}\")\nprint(f\"mean: {dataset.review_count.mean()}\")", "sum : 1416734\nmean: 212.5951380552221\n" ], [ "print(f\"sum : {dataset.rating.sum()}\")\nprint(f\"mean: {dataset.rating.mean()}\")", "sum : 25611.5\nmean: 3.8432623049219687\n" ], [ "dataset.loc[dataset.alias.isin(yelp_branches)].price.value_counts()", "_____no_output_____" ], [ "print(dataset.loc[dataset.alias.isin(yelp_branches)].review_count.sum())\nprint(dataset.loc[dataset.alias.isin(yelp_branches)].review_count.mean())", "26365\n2197.0833333333335\n" ], [ "import math\n\n\ndef distance(origin, destination):\n \"\"\"\n Calculate the Haversine distance.\n\n Parameters\n ----------\n origin : tuple of float\n (lat, long)\n destination : tuple of float\n (lat, long)\n\n Returns\n -------\n distance_in_km : float\n\n Examples\n --------\n >>> origin = (48.1372, 11.5756) # Munich\n >>> destination = (52.5186, 13.4083) # Berlin\n >>> round(distance(origin, destination), 1)\n 504.2\n \"\"\"\n lat1, lon1 = origin\n lat2, lon2 = destination\n radius = 6371 # km\n\n dlat = math.radians(lat2 - lat1)\n dlon = math.radians(lon2 - lon1)\n a = (math.sin(dlat / 2) * math.sin(dlat / 2) +\n math.cos(math.radians(lat1)) * math.cos(math.radians(lat2)) *\n math.sin(dlon / 2) * math.sin(dlon / 2))\n c = 2 * math.atan2(math.sqrt(a), math.sqrt(1 - a))\n d = radius * c\n return d", "_____no_output_____" ], [ "origin = dataset.iloc[0].coordinate_latitude, dataset.iloc[0].coordinate_longitude\ndestination = dataset.iloc[4].coordinate_latitude, dataset.iloc[4].coordinate_longitude\ndistance(origin, destination) * 1000", "_____no_output_____" ], [ "bins = [0, 100, 500, 1000, 2000, 3000, 5000, 10000, 20000]\nlbls = [1, 2, 3, 4, 5, 6, 7, 8, 9]\npd_bins = pd.cut(dataset.review_count, bins, lbls).value_counts()\npd_bins.plot(title='Binned Review Count').tick_params(axis='x', labelrotation=45)", "_____no_output_____" ], [ "bins = [0, 0.5, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5]\npd_bins = pd.cut(dataset.rating, bins).value_counts()\npd_bins.plot(title='Binned Rating').tick_params(axis='x', labelrotation=45)", "_____no_output_____" ], [ "pd_bins", "_____no_output_____" ] ] ]

[ "code", "markdown", "code", "markdown", "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Introduction to Natural Language Processing", "_____no_output_____" ], [ "## Review Questions\n\n1. Vector space model\n\n2. TF-IDF \n\n3. Use of `nltk` library \n\n4. Classification and sentiment analysis", "_____no_output_____" ] ] ]

[ "markdown" ]

[ [ "markdown", "markdown" ] ]

[ "BSD-3-Clause" ]

[ "BSD-3-Clause" ]

[ "BSD-3-Clause" ]

[ [ [ "# Nate Brunacini, [email protected]\n# Supervisor: Kelly A. Douglass\n# This file includes methods to find the gradient (slope of the trend line) of the 3D (or \"R\") metallicities of \n# each spaxel in a MaNGA galaxy and to create a scatter plot of those gradient values.", "_____no_output_____" ], [ "# Import packages\nfrom astropy.io import fits\nimport deproject_spaxel as dps\nimport numpy as np\nimport math\nimport matplotlib.pyplot as plt\n%matplotlib inline\nfrom astropy.table import Table\nfrom scipy.stats import linregress\n\nimport marvin\nfrom marvin.tools.maps import Maps", "/home/nbrunaci/.local/lib/python3.9/site-packages/marvin/core/exceptions.py:50: UserWarning: cannot initiate Sentry error reporting: [Errno 6] No such device or address.\n warnings.warn('cannot initiate Sentry error reporting: {0}.'.format(str(ee)),\n\u001b[0;34m[INFO]: \u001b[0mNo release version set. Setting default to DR15\n\u001b[1;33m[WARNING]: \u001b[0m\u001b[0;39mpath /home/nbrunaci/sas/mangawork/manga/spectro/redux/v2_4_3/drpall-v2_4_3.fits cannot be found. Setting drpall to None.\u001b[0m \u001b[0;36m(MarvinUserWarning)\u001b[0m\n\u001b[1;33m[WARNING]: \u001b[0m\u001b[0;39mpath /home/nbrunaci/sas/mangawork/manga/spectro/analysis/v2_4_3/2.2.1/dapall-v2_4_3-2.2.1.fits cannot be found. Setting dapall to None.\u001b[0m \u001b[0;36m(MarvinUserWarning)\u001b[0m\n" ], [ "# Takes in plateifu and table of kinematic center data, returns coordinates of kinematic center of galaxy\ndef getKinematicCenter(plateifu,c_table):\n plate, ifu = plateifu.split('-')\n bool_index = np.logical_and(c_table['MaNGA_plate'] == int(plate), c_table['MaNGA_IFU'] == int(ifu))\n x_coord = c_table['x0_map'][bool_index].data[0]\n y_coord = c_table['y0_map'][bool_index].data[0]\n return (y_coord,x_coord)\n \n# x0_map,y0_map: pass in as (y,x); same as (row,column)\n\n# Returns coordinates of photometric center of the galaxy with the given plateifu\ndef getPhotometricCenter(plateifu):\n maps = Maps(plateifu)\n# print(maps.datamodel)\n gfluxmap = maps['spx_mflux']\n center = np.unravel_index(np.argmax(gfluxmap.data),gfluxmap.shape)\n return center", "_____no_output_____" ], [ "# Takes in plateifu, data from drpall file, and table of kinematic centers, generates lists of normalized radius from galactic center and metallicity values, and outputs them in a dictionary\ndef radius_lists(plateifu,drp,c_table):\n with fits.open('MetallicityFITS_Pilyugin/Pilyugin_'+plateifu+'.fits', mode='update') as hdul:\n index = np.where(drp['PLATEIFU'] == plateifu)[0][0]# Index of galaxy with the given plateifu; there is only one value but it is nested, hence the [0][0]\n rot_angle = drp['NSA_ELPETRO_PHI'][index] * math.pi/180# Rotation angle; converted from degrees to radians\n inc_angle = np.arccos(drp['NSA_ELPETRO_BA'][index])#math.pi/2.0 - math.asin(drp['NSA_ELPETRO_BA'][index])# Inclination angle; converted from axis ratio to angle in radians\n re = drp['NSA_ELPETRO_TH50_R'][index]# 50% light radius in SDSS r-band (in arcsec)\n \n # Get the kinematic center of the galaxy; if there is none in the data file, use photometric center\n center = getKinematicCenter(plateifu,c_table)\n if center == -99.0:# No kinematic center if value is -99\n center = getPhotometricCenter(plateifu)\n \n #Arrays of values to be plotted\n radii_R = []# List of normalized radii between each spaxel and the galactic center for spaxels with R metallicity values\n R = []# List of R metallicity values excluding those at masked spaxels\n # Add points to lists\n for row in range(hdul[1].shape[1]):\n for col in range(hdul[1].shape[0]):\n # Calcuate deprojected radius for the spaxel\n coords = (row,col)\n rad_spax,_ = dps.deproject_spaxel(coords,center,rot_angle,inc_angle)#Radius in units of spaxels\n rad_arcsec = rad_spax * 0.5# Radius in arcseconds\n rad_normalized = rad_arcsec/re\n # Add normalized radius and metallicity values to lists if not masked at that spaxel\n if not hdul[3].data[row][col]:# Removes masked values\n radii_R.append(rad_normalized)\n R.append(hdul[1].data[row][col])\n return {\n 'radii_R': radii_R,\n 'R': R,\n 'r50':re\n }", "_____no_output_____" ], [ "# Takes in dictionary of radius and metallicity lists such as that output by the radius_lists function and outputs the parameters of the line of best fit\ndef calculate_fits(r_lists):\n # Not sure whether the r, p, and se values are needed. There is also an intercept_stderr value but that must be \n # accessed as an attribute of the returned objected (as in results = linregress(x,y) then results.intercept_stderr)\n# slope_N2, intercept_N2, r_N2, p_N2, se_N2 = linregress(r_lists['radii_N2'], r_lists['N2'])\n# slope_O3N2, intercept_O3N2, r_N2, p_N2, se_N2 = linregress(r_lists['radii_O3N2'], r_lists['O3N2'])\n# slope_N2O2, intercept_N2O2, r_N2, p_N2, se_N2 = linregress(r_lists['radii_N2O2'], r_lists['N2O2'])\n R_params = linregress(r_lists['radii_R'], r_lists['R'])\n return {\n # To access individual paramters, use (for example) N2_params.slope, .intercept, .rvalue, .pvalue, .stderr,\n # .intercept_stderr\n 'R_params': R_params,\n 'r50':r_lists['r50']\n }", "_____no_output_____" ], [ "# Takes in output from radius_lists and calculate_fits functions as well as plateifu and plots scatter plots (metallicity \n# versus normalized radius) with lines of best fit\ndef scatterplots(r_lists,fit_params,plateifu):\n fig, plots = plt.subplots(1)\n fig.set_figheight(5)\n fig.set_figwidth(5)\n plots.plot(r_lists['radii_R'],r_lists['R'],'.')\n plots.set_title('3D Metallicity vs. Normalized Radius')\n plots.set_ylabel('Metallicity')\n plots.set_xlabel('r / r_e')\n x_R = np.linspace(min(r_lists['radii_R']),max(r_lists['radii_R']))#(0.0,1.6)\n y_R = fit_params['R_params'].slope * x_R + fit_params['R_params'].intercept\n plots.plot(x_R,y_R,'-r')\n plt.savefig('Pilyugin_Galaxy_ScatterPlots/'+plateifu+'ScatterPlot_R')\n plt.close()", "_____no_output_____" ], [ "# Wrapper function to call the above functions all at once. Takes in plateifu, data from drpall file, and table of kinematic \n# centers, calculates the parameters of the line of best fit of the normalized radius versus metallicity \n# data, and creates scatter plots\ndef find_gradient(plateifu,drp,c_table):\n r_lists = radius_lists(plateifu,drp,c_table)\n trend = calculate_fits(r_lists)\n scatterplots(r_lists,trend,plateifu)\n return trend", "_____no_output_____" ], [ "# # Calling the functions\n# with fits.open('drpall-v2_4_3.fits', memmap=True) as drpall:\n# c_table = Table.read('DRP-master_file_vflag_BB_smooth1p85_mapFit_N2O2_HIdr2_noWords_v5.txt',format='ascii.commented_header')\n# find_gradient('9487-12701',drpall[1].data,c_table)#('9487-12701',drpall[1].data,c_table)#('8335-12701')#('7443-12705')\n# # plt.savefig('PosterMaps/Scatter_8335-12701')", "\u001b[1;33m[WARNING]: \u001b[0m\u001b[0;39mOverflowError converting to FloatType in column avg_alpha, possibly resulting in degraded precision.\u001b[0m \u001b[0;36m(AstropyWarning)\u001b[0m\n\u001b[1;33m[WARNING]: \u001b[0m\u001b[0;39mOverflowError converting to FloatType in column pos_alpha, possibly resulting in degraded precision.\u001b[0m \u001b[0;36m(AstropyWarning)\u001b[0m\n\u001b[1;33m[WARNING]: \u001b[0m\u001b[0;39mOverflowError converting to FloatType in column neg_alpha, possibly resulting in degraded precision.\u001b[0m \u001b[0;36m(AstropyWarning)\u001b[0m\n" ], [ "# with fits.open('MetallicityFITS/Brown_7992-12705.fits', mode='update') as hdul:\n# print(hdul.info())", "Filename: MetallicityFITS/Brown_7992-12705.fits\nNo. Name Ver Type Cards Dimensions Format\n 0 PRIMARY 1 PrimaryHDU 5 (0,) \n 1 N2_METALLICITY 1 ImageHDU 8 (74, 74) float64 \n 2 O3N2_METALLICITY 1 ImageHDU 8 (74, 74) float64 \n 3 N2O2_METALLICITY 1 ImageHDU 8 (74, 74) float64 \n 4 N2_IVAR 1 ImageHDU 8 (74, 74) float64 \n 5 O3N2_IVAR 1 ImageHDU 8 (74, 74) float64 \n 6 N2O2_IVAR 1 ImageHDU 8 (74, 74) float64 \n 7 N2_MASK 1 ImageHDU 8 (74, 74) int32 \n 8 O3N2_MASK 1 ImageHDU 8 (74, 74) int32 \n 9 N2O2_MASK 1 ImageHDU 8 (74, 74) int32 \nNone\n" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Scikit-learn Tutorial", "_____no_output_____" ], [ "# Introduction to Machine Learning in Python", "_____no_output_____" ], [ "## What is Machine Learning?", "_____no_output_____" ], [ "Machine learning is the process of extracting knowledge from data automatically, usually with the goal of making predictions on new, unseen data. A classical example is a spam filter, for which the user keeps labeling incoming mails as either spam or not spam. A machine learning algorithm then \"learns\" a predictive model from data that distinguishes spam from normal emails, a model which can predict for new emails whether they are spam or not. \n\nCentral to machine learning is the concept of **automating decision making** from data **without the user specifying explicit rules** how this decision should be made.\n\nFor the case of emails, the user doesn't provide a list of words or characteristics that make an email spam. Instead, the user provides examples of spam and non-spam emails that are labeled as such.\n\nThe second central concept is **generalization**. The goal of a machine learning model is to predict on new, previously unseen data. In a real-world application, we are not interested in marking an already labeled email as spam or not. Instead, we want to make the user's life easier by automatically classifying new incoming mail.", "_____no_output_____" ], [ "<img src=\"figures/supervised_workflow.svg\" width=\"100%\">", "_____no_output_____" ], [ "The data is presented to the algorithm usually as a two-dimensional array (or matrix) of numbers. Each data point (also known as a *sample* or *training instance*) that we want to either learn from or make a decision on is represented as a list of numbers, a so-called feature vector, and its containing features represent the properties of this point. \n\nLater, we will work with a popular dataset called *Iris* -- among many other datasets. Iris, a classic benchmark dataset in the field of machine learning, contains the measurements of 150 iris flowers from 3 different species: Iris-Setosa, Iris-Versicolor, and Iris-Virginica. \n\n\n\n<table style=\"width:100%\">\n <tr>\n <th>Species</th>\n <th>Image</th>\n </tr>\n <tr>\n <td>Iris Setosa</td>\n <td><img src=\"figures/iris_setosa.jpg\" width=\"80%\"></td>\n </tr>\n <tr>\n <td>Iris Versicolor</td>\n <td><img src=\"figures/iris_versicolor.jpg\" width=\"80%\"></td>\n </tr>\n <tr>\n <td>Iris Virginica</td>\n <td><img src=\"figures/iris_virginica.jpg\" width=\"80%\"></td>\n </tr>\n</table>\n\n\n\n\n\nWe represent each flower sample as one row in our data array, and the columns (features) represent the flower measurements in centimeters. For instance, we can represent this Iris dataset, consisting of 150 samples and 4 features, a 2-dimensional array or matrix $\\mathbb{R}^{150 \\times 4}$ in the following format:\n\n\n$$\\mathbf{X} = \\begin{bmatrix}\n x_{1}^{(1)} & x_{2}^{(1)} & x_{3}^{(1)} & \\dots & x_{4}^{(1)} \\\\\n x_{1}^{(2)} & x_{2}^{(2)} & x_{3}^{(2)} & \\dots & x_{4}^{(2)} \\\\\n \\vdots & \\vdots & \\vdots & \\ddots & \\vdots \\\\\n x_{1}^{(150)} & x_{2}^{(150)} & x_{3}^{(150)} & \\dots & x_{4}^{(150)}\n\\end{bmatrix}.\n$$\n\n(The superscript denotes the *i*th row, and the subscript denotes the *j*th feature, respectively.", "_____no_output_____" ], [ "There are two kinds of machine learning we will talk about today: ***supervised learning*** and ***unsupervised learning***.", "_____no_output_____" ], [ "### Supervised Learning: Classification and regression\n\nIn **Supervised Learning**, we have a dataset consisting of both input features and a desired output, such as in the spam / no-spam example.\nThe task is to construct a model (or program) which is able to predict the desired output of an unseen object\ngiven the set of features.\n\nSome more complicated examples are:\n\n- Given a multicolor image of an object through a telescope, determine\n whether that object is a star, a quasar, or a galaxy.\n- Given a photograph of a person, identify the person in the photo.\n- Given a list of movies a person has watched and their personal rating\n of the movie, recommend a list of movies they would like.\n- Given a persons age, education and position, infer their salary\n\nWhat these tasks have in common is that there is one or more unknown\nquantities associated with the object which needs to be determined from other\nobserved quantities.\n\nSupervised learning is further broken down into two categories, **classification** and **regression**:\n\n- **In classification, the label is discrete**, such as \"spam\" or \"no spam\". In other words, it provides a clear-cut distinction between categories. Furthermore, it is important to note that class labels are nominal, not ordinal variables. Nominal and ordinal variables are both subcategories of categorical variable. Ordinal variables imply an order, for example, T-shirt sizes \"XL > L > M > S\". On the contrary, nominal variables don't imply an order, for example, we (usually) can't assume \"orange > blue > green\".\n\n\n- **In regression, the label is continuous**, that is a float output. For example,\nin astronomy, the task of determining whether an object is a star, a galaxy, or a quasar is a\nclassification problem: the label is from three distinct categories. On the other hand, we might\nwish to estimate the age of an object based on such observations: this would be a regression problem,\nbecause the label (age) is a continuous quantity.\n\nIn supervised learning, there is always a distinction between a **training set** for which the desired outcome is given, and a **test set** for which the desired outcome needs to be inferred. The learning model fits the predictive model to the training set, and we use the test set to evaluate its generalization performance.\n", "_____no_output_____" ], [ "### Unsupervised Learning\n\nIn **Unsupervised Learning** there is no desired output associated with the data.\nInstead, we are interested in extracting some form of knowledge or model from the given data.\nIn a sense, you can think of unsupervised learning as a means of discovering labels from the data itself.\nUnsupervised learning is often harder to understand and to evaluate.\n\nUnsupervised learning comprises tasks such as *dimensionality reduction*, *clustering*, and\n*density estimation*. For example, in the iris data discussed above, we can used unsupervised\nmethods to determine combinations of the measurements which best display the structure of the\ndata. As we’ll see below, such a projection of the data can be used to visualize the\nfour-dimensional dataset in two dimensions. Some more involved unsupervised learning problems are:\n\n- Given detailed observations of distant galaxies, determine which features or combinations of\n features summarize best the information.\n- Given a mixture of two sound sources (for example, a person talking over some music),\n separate the two (this is called the [blind source separation](http://en.wikipedia.org/wiki/Blind_signal_separation) problem).\n- Given a video, isolate a moving object and categorize in relation to other moving objects which have been seen.\n- Given a large collection of news articles, find recurring topics inside these articles.\n- Given a collection of images, cluster similar images together (for example to group them when visualizing a collection)\n\nSometimes the two may even be combined: e.g. unsupervised learning can be used to find useful\nfeatures in heterogeneous data, and then these features can be used within a supervised\nframework.", "_____no_output_____" ], [ "### (simplified) Machine learning taxonomy\n\n<img src=\"figures/ml_taxonomy.png\" width=\"80%\">", "_____no_output_____" ] ] ]

[ "markdown" ]

[ [ "markdown", "markdown", "markdown", "markdown", "markdown", "markdown", "markdown", "markdown", "markdown", "markdown" ] ]

[ "BSD-3-Clause" ]

[ "BSD-3-Clause" ]

[ "BSD-3-Clause" ]

[ [ [ "columns_to_encode = ['protocol', 'application_name', 'application_category_name', 'content_type']", "_____no_output_____" ], [ "additional_columns = ['udps.num_pkts_up_to_128_bytes', 'udps.num_pkts_128_to_256_bytes',\n 'udps.num_pkts_256_to_512_bytes', 'udps.num_pkts_512_to_1024_bytes',\n 'udps.num_pkts_1024_to_1514_bytes', 'udps.min_ttl', 'udps.max_ttl',\n 'udps.min_ip_pkt_len', 'udps.max_ip_pkt_len', 'udps.src2dst_flags',\n 'udps.dst2src_flags', 'udps.tcp_flags', 'udps.tcp_win_max_in',\n 'udps.tcp_win_max_out', 'udps.icmp_type', 'udps.icmp_v4_type',\n 'udps.dns_query_id', 'udps.dns_query_type', 'udps.dns_ttl_answer',\n 'udps.ftp_command_ret_code', 'udps.retransmitted_in_packets',\n 'udps.retransmitted_out_packets', 'udps.retransmitted_in_bytes',\n 'udps.retransmitted_out_bytes', 'udps.src_to_dst_second_bytes',\n 'udps.dst_to_src_second_bytes', 'udps.src_to_dst_avg_throughput',\n 'udps.dst_to_src_avg_throughput', 'udps.src_to_dst_second_bytes2',\n 'udps.dst_to_src_second_bytes2', 'udps.src_to_dst_avg_throughput2',\n 'udps.dst_to_src_avg_throughput2']", "_____no_output_____" ], [ "columns_to_delete = ['udps.bidirectional_pkts']", "_____no_output_____" ], [ "import pandas as pd\nfrom sklearn import preprocessing\nfrom sklearn.ensemble import ExtraTreesClassifier\nfrom joblib import load\nimport os", "_____no_output_____" ] ], [ [ "# With additional features", "_____no_output_____" ] ], [ [ "models_path = \"/home/soufiane.oualil/lustre/data_sec-um6p-st-sccs-6sevvl76uja/IDS/prod_ai_models/ML/full_binary_0/\"\n\n", "_____no_output_____" ], [ "dataset = \"UNSW-NB15\"", "_____no_output_____" ], [ "dataset_path = f\"/home/soufiane.oualil/lustre/data_sec-um6p-st-sccs-6sevvl76uja/IDS/preprocessed_datasets/{dataset}/flow_features/multi/\"\n\n", "_____no_output_____" ], [ "data = pd.read_pickle(f'{dataset_path}/test.p')", "_____no_output_____" ], [ "data.head()", "_____no_output_____" ], [ "y = data['Attack'].values", "_____no_output_____" ], [ "data = data.drop(columns_to_delete + ['Attack'], axis = 1)", "_____no_output_____" ], [ "data.head()", "_____no_output_____" ], [ "for column in columns_to_encode:\n le = load(f'{models_path}encoders/{column}.joblib') \n data[column] = le.transform(data[column])\n ", "/home/soufiane.oualil/.conda/envs/atlas/lib/python3.8/site-packages/sklearn/base.py:329: UserWarning: Trying to unpickle estimator LabelEncoder from version 0.23.2 when using version 1.0.2. This might lead to breaking code or invalid results. Use at your own risk. For more info please refer to:\nhttps://scikit-learn.org/stable/modules/model_persistence.html#security-maintainability-limitations\n warnings.warn(\n" ], [ "le_labels = load(f'{models_path}/encoders/attack_encoder.joblib')\n", "_____no_output_____" ], [ "data.head()", "_____no_output_____" ], [ "clf = load(f'{models_path}/clf_model.joblib') ", "/home/soufiane.oualil/.conda/envs/atlas/lib/python3.8/site-packages/sklearn/base.py:329: UserWarning: Trying to unpickle estimator ExtraTreeClassifier from version 0.23.2 when using version 1.0.2. This might lead to breaking code or invalid results. Use at your own risk. For more info please refer to:\nhttps://scikit-learn.org/stable/modules/model_persistence.html#security-maintainability-limitations\n warnings.warn(\n/home/soufiane.oualil/.conda/envs/atlas/lib/python3.8/site-packages/sklearn/base.py:329: UserWarning: Trying to unpickle estimator ExtraTreesClassifier from version 0.23.2 when using version 1.0.2. This might lead to breaking code or invalid results. Use at your own risk. For more info please refer to:\nhttps://scikit-learn.org/stable/modules/model_persistence.html#security-maintainability-limitations\n warnings.warn(\n" ], [ "preds = clf.predict(data.values)", "_____no_output_____" ], [ "preds_labels = le_labels.inverse_transform(preds)", "_____no_output_____" ], [ "print(preds_labels[:20])", "['Benign' 'Benign' 'Benign' 'Malign' 'Benign' 'Benign' 'Benign' 'Benign'\n 'Benign' 'Benign' 'Benign' 'Benign' 'Benign' 'Benign' 'Benign' 'Benign'\n 'Benign' 'Benign' 'Benign' 'Benign']\n" ] ], [ [ "# Without additional features", "_____no_output_____" ] ], [ [ "models_path = \"/home/abdellah.elmekki/lustre/data_sec-um6p-st-sccs-6sevvl76uja/IDS/prod_ai_models/ML/full_binary_1/\"\n", "_____no_output_____" ], [ "data = pd.read_pickle(f'{dataset_path}/test.p')", "_____no_output_____" ], [ "y = data['Attack'].values", "_____no_output_____" ], [ "data = data.drop(columns_to_delete + additional_columns + ['Attack'], axis = 1)", "_____no_output_____" ], [ "for column in columns_to_encode:\n le = load(f'{models_path}/encoders/{column}.joblib') \n data[column] = le.transform(data[column])\n ", "_____no_output_____" ], [ "le_labels = load(f'{models_path}/encoders/attack_encoder.joblib')\n", "_____no_output_____" ], [ "clf = load(f'{models_path}/clf_model.joblib') ", "_____no_output_____" ], [ "preds = clf.predict(data.values)", "_____no_output_____" ], [ "print(preds_labels[:20])", "['Benign' 'Benign' 'Benign' 'Malign' 'Benign' 'Benign' 'Benign' 'Benign'\n 'Benign' 'Benign' 'Benign' 'Benign' 'Benign' 'Benign' 'Benign' 'Benign'\n 'Benign' 'Benign' 'Benign' 'Benign']\n" ] ] ]

[ "code", "markdown", "code", "markdown", "code" ]

[ [ "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "FTL" ]

[ "FTL" ]

[ "FTL" ]

[ [ [ "## Install python libraries\n\nThe following cell can be used to ensure that the python libraries used\nin this test notebook are installed.\nThese may be pre-installed in future notebook images.\nOnce this cell has been run, it need not be re-run unless you have restarted your jupyter server.", "_____no_output_____" ] ], [ [ "# Install the library dependencies used in this notebook\n# (comment this out if you prefer to not re-run this cell)\n%pip install trino python-dotenv", "Requirement already satisfied: trino in /opt/app-root/lib/python3.8/site-packages (0.306.0)\nRequirement already satisfied: python-dotenv in /opt/app-root/lib/python3.8/site-packages (0.19.1)\nRequirement already satisfied: requests in /opt/app-root/lib/python3.8/site-packages (from trino) (2.25.1)\nRequirement already satisfied: chardet<5,>=3.0.2 in /opt/app-root/lib/python3.8/site-packages (from requests->trino) (4.0.0)\nRequirement already satisfied: urllib3<1.27,>=1.21.1 in /opt/app-root/lib/python3.8/site-packages (from requests->trino) (1.26.4)\nRequirement already satisfied: certifi>=2017.4.17 in /opt/app-root/lib/python3.8/site-packages (from requests->trino) (2020.12.5)\nRequirement already satisfied: idna<3,>=2.5 in /opt/app-root/lib/python3.8/site-packages (from requests->trino) (2.10)\n\u001b[33mWARNING: You are using pip version 21.1; however, version 21.3 is available.\nYou should consider upgrading via the '/opt/app-root/bin/python3.8 -m pip install --upgrade pip' command.\u001b[0m\nNote: you may need to restart the kernel to use updated packages.\n" ] ], [ [ "## Loading credentials\n\nThe following cell finds a `credentials.env` file at the jupyter \"home\" (top level) directory.\n\nValues in this `dotenv` file are loaded into the `os.environ` table,\nas if they were regular environment variables.\n\nCredentials are stored in `dotenv` files so that they can be referred to by standard\nenvironment variable names, and do not appear in notebooks or other code,\nwhich would be a security leak.", "_____no_output_____" ] ], [ [ "from dotenv import dotenv_values, load_dotenv\nimport os\nimport pathlib\n\ndotenv_dir = os.environ.get('CREDENTIAL_DOTENV_DIR', os.environ.get('PWD', '/opt/app-root/src'))\ndotenv_path = pathlib.Path(dotenv_dir) / 'credentials.env'\nif os.path.exists(dotenv_path):\n load_dotenv(dotenv_path=dotenv_path,override=True)", "_____no_output_____" ] ], [ [ "## Connect to trino\n\nThe following cell creates a trino api connection.\n\nIt assumes that your `credentials.env` file has been edited so that\n`TRINO_PASSWD` has a JWT token obtained from:\nhttps://das-odh-trino.apps.odh-cl1.apps.os-climate.org/\n\nYour `TRINO_USER` value should be your github username.", "_____no_output_____" ] ], [ [ "import trino\nconn = trino.dbapi.connect(\n host=os.environ['TRINO_HOST'],\n port=int(os.environ['TRINO_PORT']),\n user=os.environ['TRINO_USER'],\n http_scheme='https',\n auth=trino.auth.JWTAuthentication(os.environ['TRINO_PASSWD']),\n verify=True,\n)\ncur = conn.cursor()", "_____no_output_____" ] ], [ [ "## Test your trino connection\n\nThis cell shows all the catalogs visible to you.\nIf your trino api connection initialized correctly above,\nthis `show catalogs` command should always succeed.", "_____no_output_____" ] ], [ [ "cur.execute('show catalogs')\ncur.fetchall()", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "Gaussian discriminant analysis con stessa matrice di covarianza per le distribuzioni delle due classi e conseguente separatore lineare. Implementata in scikit-learn. Valutazione con cross validation. ", "_____no_output_____" ] ], [ [ "import warnings\nwarnings.filterwarnings('ignore')\n\n%matplotlib inline", "_____no_output_____" ], [ "import pandas as pd\nimport numpy as np\nimport scipy.stats as st\nfrom sklearn.discriminant_analysis import LinearDiscriminantAnalysis\nfrom sklearn.model_selection import cross_val_score\nimport sklearn.metrics as mt", "_____no_output_____" ], [ "import matplotlib.pyplot as plt\nimport matplotlib.colors as mcolors\n\nplt.style.use('fivethirtyeight')\n\nplt.rcParams['font.family'] = 'sans-serif'\nplt.rcParams['font.serif'] = 'Ubuntu'\nplt.rcParams['font.monospace'] = 'Ubuntu Mono'\nplt.rcParams['font.size'] = 10\nplt.rcParams['axes.labelsize'] = 10\nplt.rcParams['axes.labelweight'] = 'bold'\nplt.rcParams['axes.titlesize'] = 10\nplt.rcParams['xtick.labelsize'] = 8\nplt.rcParams['ytick.labelsize'] = 8\nplt.rcParams['legend.fontsize'] = 10\nplt.rcParams['figure.titlesize'] = 12\nplt.rcParams['image.cmap'] = 'jet'\nplt.rcParams['image.interpolation'] = 'none'\nplt.rcParams['figure.figsize'] = (16, 8)\nplt.rcParams['lines.linewidth'] = 2\nplt.rcParams['lines.markersize'] = 8\n\ncolors = ['#008fd5', '#fc4f30', '#e5ae38', '#6d904f', '#8b8b8b', '#810f7c', \n'#137e6d', '#be0119', '#3b638c', '#af6f09', '#008fd5', '#fc4f30', '#e5ae38', '#6d904f', '#8b8b8b', \n'#810f7c', '#137e6d', '#be0119', '#3b638c', '#af6f09']\n\ncmap = mcolors.LinearSegmentedColormap.from_list(\"\", [\"#82cafc\", \"#069af3\", \"#0485d1\", colors[0], colors[8]])", "_____no_output_____" ] ], [ [ "Leggiamo i dati da un file csv in un dataframe pandas. I dati hanno 3 valori: i primi due corrispondono alle features e sono assegnati alle colonne x1 e x2 del dataframe; il terzo è il valore target, assegnato alla colonna t. Vengono poi creati una matrice X delle features e un vettore target t", "_____no_output_____" ] ], [ [ "# legge i dati in dataframe pandas\ndata = pd.read_csv(\"../../data/ex2data1.txt\", header= None,delimiter=',', names=['x1','x2','t'])\n\n# calcola dimensione dei dati\nn = len(data)\nn0 = len(data[data.t==0])\n\n# calcola dimensionalità delle features\nfeatures = data.columns\nnfeatures = len(features)-1\n\nX = np.array(data[features[:-1]])\nt = np.array(data['t'])\n", "_____no_output_____" ] ], [ [ "Visualizza il dataset.", "_____no_output_____" ] ], [ [ "fig = plt.figure(figsize=(16,8))\nax = fig.gca()\nax.scatter(data[data.t==0].x1, data[data.t==0].x2, s=40, color=colors[0], alpha=.7)\nax.scatter(data[data.t==1].x1, data[data.t==1].x2, s=40,c=colors[1], alpha=.7)\nplt.xlabel('$x_1$', fontsize=12)\nplt.ylabel('$x_2$', fontsize=12)\nplt.xticks(fontsize=10)\nplt.yticks(fontsize=10)\nplt.title('Dataset', fontsize=12)\nplt.show()", "_____no_output_____" ] ], [ [ "Definisce un classificatore basato su GDA quadratica ed effettua il training sul dataset.", "_____no_output_____" ] ], [ [ "clf = LinearDiscriminantAnalysis(store_covariance=True)\nclf.fit(X, t)", "_____no_output_____" ] ], [ [ "Definiamo la griglia 100x100 da utilizzare per la visualizzazione delle varie distribuzioni.", "_____no_output_____" ] ], [ [ "# insieme delle ascisse dei punti\nu = np.linspace(min(X[:,0]), max(X[:,0]), 100)\n# insieme delle ordinate dei punti\nv = np.linspace(min(X[:,1]), max(X[:,1]), 100)\n# deriva i punti della griglia: il punto in posizione i,j nella griglia ha ascissa U(i,j) e ordinata V(i,j)\nU, V = np.meshgrid(u, v)", "_____no_output_____" ] ], [ [ "Calcola sui punti della griglia le probabilità delle classi $p(x|C_0), p(x|C_1)$ e le probabilità a posteriori delle classi $p(C_0|x), p(C_1|x)$", "_____no_output_____" ] ], [ [ "# probabilità a posteriori delle due distribuzioni sulla griglia\nZ = clf.predict_proba(np.c_[U.ravel(), V.ravel()])\npp0 = Z[:, 0].reshape(U.shape)\npp1 = Z[:, 1].reshape(V.shape)\n# rapporto tra le probabilità a posteriori delle classi per tutti i punti della griglia\nz=pp0/pp1 \n\n# probabilità per le due classi sulla griglia\nmu0 = clf.means_[0]\nmu1 = clf.means_[1]\nsigma = clf.covariance_\nvf0=np.vectorize(lambda x,y:st.multivariate_normal.pdf([x,y],mu0,sigma))\nvf1=np.vectorize(lambda x,y:st.multivariate_normal.pdf([x,y],mu1,sigma))\np0=vf0(U,V)\np1=vf1(U,V)", "_____no_output_____" ] ], [ [ "Visualizzazione della distribuzione di $p(x|C_0)$", "_____no_output_____" ] ], [ [ "fig = plt.figure(figsize=(16,8))\nax = fig.gca()\n# inserisce una rappresentazione della probabilità della classe C0 sotto forma di heatmap\nimshow_handle = plt.imshow(p0, origin='lower', extent=(min(X[:,0]), max(X[:,0]), min(X[:,1]), max(X[:,1])), alpha=.7)\nplt.contour(U, V, p0, linewidths=[.7], colors=[colors[6]])\n# rappresenta i punti del dataset\nax.scatter(data[data.t==0].x1, data[data.t==0].x2, s=40, c=colors[0], alpha=.7)\nax.scatter(data[data.t==1].x1, data[data.t==1].x2, s=40,c=colors[1], alpha=.7)\n# rappresenta la media della distribuzione\nax.scatter(mu0[0], mu0[1], s=150,c=colors[3], marker='*', alpha=1)\n# inserisce titoli, etc.\nplt.xlabel('$x_1$', fontsize=12)\nplt.ylabel('$x_2$', fontsize=12)\nplt.xticks(fontsize=10)\nplt.yticks(fontsize=10)\nplt.xlim(u.min(), u.max())\nplt.ylim(v.min(), v.max())\nplt.title('Distribuzione di $p(x|C_0)$', fontsize=12)\nplt.show()", "_____no_output_____" ] ], [ [ "Visualizzazione della distribuzione di $p(x|C1)$", "_____no_output_____" ] ], [ [ "fig = plt.figure(figsize=(16,8))\nax = fig.gca()\n# inserisce una rappresentazione della probabilità della classe C0 sotto forma di heatmap\nimshow_handle = plt.imshow(p1, origin='lower', extent=(min(X[:,0]), max(X[:,0]), min(X[:,1]), max(X[:,1])), alpha=.7)\nplt.contour(U, V, p1, linewidths=[.7], colors=[colors[6]])\n# rappresenta i punti del dataset\nax.scatter(data[data.t==0].x1, data[data.t==0].x2, s=40, c=colors[0], alpha=.7)\nax.scatter(data[data.t==1].x1, data[data.t==1].x2, s=40,c=colors[1], alpha=.7)\n# rappresenta la media della distribuzione\nax.scatter(mu1[0], mu1[1], s=150,c=colors[3], marker='*', alpha=1)\n# inserisce titoli, etc.\nplt.xlabel('$x_1$', fontsize=12)\nplt.ylabel('$x_2$', fontsize=12)\nplt.xticks(fontsize=10)\nplt.yticks(fontsize=10)\nplt.xlim(u.min(), u.max())\nplt.ylim(v.min(), v.max())\nplt.title('Distribuzione di $p(x|C_1)$', fontsize=12)\nplt.show()", "_____no_output_____" ] ], [ [ "Visualizzazione di $p(C_0|x)$", "_____no_output_____" ] ], [ [ "fig = plt.figure(figsize=(8,8))\nax = fig.gca()\nimshow_handle = plt.imshow(pp0, origin='lower', extent=(min(X[:,0]), max(X[:,0]), min(X[:,1]), max(X[:,1])), alpha=.7)\nax.scatter(data[data.t==0].x1, data[data.t==0].x2, s=40, c=colors[0], alpha=.7)\nax.scatter(data[data.t==1].x1, data[data.t==1].x2, s=40,c=colors[1], alpha=.7)\nplt.contour(U, V, z, [1.0], colors=[colors[7]],linewidths=[1])\nplt.xlabel('$x_1$', fontsize=12)\nplt.ylabel('$x_2$', fontsize=12)\nplt.xticks(fontsize=10)\nplt.yticks(fontsize=10)\nplt.xlim(u.min(), u.max())\nplt.ylim(v.min(), v.max())\nplt.title(\"Distribuzione di $p(C_0|x)$\", fontsize=12)\nplt.show()", "_____no_output_____" ] ], [ [ "Visualizzazione di $p(C_1|x)$", "_____no_output_____" ] ], [ [ "fig = plt.figure(figsize=(8,8))\nax = fig.gca()\nimshow_handle = plt.imshow(pp1, origin='lower', extent=(min(X[:,0]), max(X[:,0]), min(X[:,1]), max(X[:,1])), alpha=.7)\nax.scatter(data[data.t==0].x1, data[data.t==0].x2, s=40, c=colors[0], alpha=.7)\nax.scatter(data[data.t==1].x1, data[data.t==1].x2, s=40,c=colors[1], alpha=.7)\nplt.contour(U, V, z, [1.0], colors=[colors[7]],linewidths=[1])\nplt.xlabel('$x_1$', fontsize=12)\nplt.ylabel('$x_2$', fontsize=12)\nplt.xticks(fontsize=10)\nplt.yticks(fontsize=10)\nplt.xlim(u.min(), u.max())\nplt.ylim(v.min(), v.max())\nplt.title(\"Distribuzione di $p(C_1|x)$\", fontsize=12)\nplt.show()", "_____no_output_____" ] ], [ [ "Applica la cross validation (5-fold) per calcolare l'accuracy effettuando la media sui 5 valori restituiti.", "_____no_output_____" ] ], [ [ "print(\"Accuracy: {0:5.3f}\".format(cross_val_score(clf, X, t, cv=5, scoring='accuracy').mean()))", "Accuracy: 0.870\n" ] ] ]

[ "raw", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "raw" ], [ "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "empty" ] ] ]

[ "empty" ]

[ [ "empty" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "from sympy import *", "_____no_output_____" ], [ "import pylatex as p", "_____no_output_____" ] ], [ [ "$$\n w_{t+1} = (1 + r_{t+1}) s(w_t) + y_{t+1}\n$$ (my_other_label)", "_____no_output_____" ], [ "- A link to an equation directive: {eq}`my_label`\n- A link to a dollar math block: {eq}`my_other_label`\n", "_____no_output_____" ], [ "::::{important}\n:::{note}\nThis text is **standard** _Markdown_\n:::\n::::\n", "_____no_output_____" ] ] ]

[ "code", "markdown" ]

[ [ "code", "code" ], [ "markdown", "markdown", "markdown" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Parte 04\n\nNessa parte os modelos criados anteriormente serão utilizados para realizar predições. Para isso, eles devem ser\nregistrados no TFX. Para efetuar as predições, os dados utilizados no treinamento desses modelos serão inseridos\nno SAVIME, o qual ficará encarregado de enviar e receber os dados para/de TFX. ", "_____no_output_____" ] ], [ [ "import os\nimport sys\n\n# Necessário mudar o diretório de trabalho para o nível mais acima\nif not 'notebooks' in os.listdir('.'):\n current_dir = os.path.abspath(os.getcwd())\n parent_dir = os.path.dirname(current_dir)\n os.chdir(parent_dir)\n\n# Inserir aqui o caminho do arquivo de dados: um json contendo informações a respeito \n# da partição de x e y utilizada na parte 01.\ndata_fp = 'saved_models_arima/data.json'\n\n# Configuração do host e porta em que o SAVIME está escutando\nsavime_host = '127.0.0.1'\nsavime_port = 65000\n\n# Configuração TFX\ntfx_host = 'localhost'\ntfx_port = 8501\n\n# Diretório de dados\ndata_dir = 'data'\n\n# Local do array de temperaturas\ndataset_path = os.path.join(data_dir, 'tiny-dataset.hdf5')", "_____no_output_____" ], [ "%load_ext autoreload\n%autoreload 2\n%matplotlib agg\n\nfrom IPython.display import HTML\n\n\nimport json\nimport h5py\nimport numpy as np\nimport seaborn as sns\nimport tensorflow as tf\n\nfrom src.animation import animate_heat_map\nfrom src.predictor_consumer import PredictionConsumer\n\n# Savime imports\nimport pysavime\nfrom pysavime.util.converter import DataVariableBlockConverter\n\nsns.set_context('notebook')\nsns.set_style('whitegrid')\nsns.set_palette(sns.color_palette(\"Paired\"))\n\ntf.get_logger().setLevel('ERROR')\n\nwith open(data_fp, 'r') as _in:\n data = json.load(_in)", "_____no_output_____" ] ], [ [ "A primeira etapa a ser realizada é converter os dados para um formato processável para o SAVIME.", "_____no_output_____" ] ], [ [ "with h5py.File(dataset_path, 'r') as in_:\n array = in_['real'][...]\n \n# Especificar dimensões\ntime_series = ('time_series', range(array.shape[0]))\ntime_step = ('time_step', range(array.shape[1]))\npos_x = ('pos_x', range(array.shape[2]))\npos_y = ('pos_y', range(array.shape[3]))\n\n# Remover última dimensão espúria\nsqueezed_array = np.squeeze(array, axis=-1)\n\n# Salvar array\ntemperatura_data_fp = os.path.join(data_dir, 'temperatura.data')\nsqueezed_array.ravel().astype('float64').tofile(temperatura_data_fp)", "_____no_output_____" ] ], [ [ "Também é nessário fazer a divisão do conjunto de dados de entrada em x e y. Como dito na parte anterior, cada série temporal possuí 10 instantes de tempo. Além disso, os modelos foram treinados a prever o décimo instante de tempo a partir dos nove anteriores. A critério de exemplo, selecionamos abaixo um grupo de séries temporais para realizar a predição de temperatura.", "_____no_output_____" ] ], [ [ "# Seleciona-se apenas um grupo para predição. \nchosen_model_name = data['model']\nchosen_group_ix = 0\nx = squeezed_array[[chosen_group_ix], :-1]\ny = squeezed_array[[chosen_group_ix], 1:]\n\npc = PredictionConsumer(host=tfx_host, port=tfx_port, model_name=chosen_model_name)\ny_hat = pc.predict(x)", "_____no_output_____" ], [ "anim_y = animate_heat_map(np.squeeze(y,axis=0))\nanim_y_html = anim_y.to_html5_video()\nanim_yhat = animate_heat_map(np.squeeze(y_hat,axis=0))\nanim_yhat_html = anim_yhat.to_html5_video()\nHTML(f'<div style=\"float: left;\"> {anim_y_html} </div><div style=\"float: left;\"> {anim_yhat_html} </div>')", "_____no_output_____" ], [ "num_models = 25\nnum_groups, num_time_steps, num_pos_x, num_pos_y = squeezed_array.shape \n\n# Define o dataset com as temperaturas a ser registrado no SAVIME.\ndataset = pysavime.define.file_dataset('temperature_data', temperatura_data_fp, 'double')\nprint('- Dataset CREATE query:', dataset.create_query_str())\n\n# Define o esquema do tar\ngroup_dim = pysavime.define.implicit_tar_dimension('group', 'int32', 0, num_groups - 1)\ntime_step_dim = pysavime.define.implicit_tar_dimension('time_step', 'int32', 0, num_time_steps - 1)\npos_x_dim = pysavime.define.implicit_tar_dimension('pos_x', 'int32', 0, num_pos_x - 1)\npos_y_dim = pysavime.define.implicit_tar_dimension('pos_y', 'int32', 0, num_pos_y - 1)\ntemperature = pysavime.define.tar_attribute('temperature', 'double')\n\ndims = [group_dim, time_step_dim, pos_x_dim, pos_y_dim]\nattributes = [temperature]\ntar = pysavime.define.tar('temperatures_tar', dims, attributes)\nprint('- Tar CREATE query:', tar.create_query_str())\n\n# Define o subtar único responsável por registrar o dataset no tar criado anteriormente.\ngroup_dim_sub = pysavime.define.ordered_subtar_dimension(group_dim, 0, num_groups - 1, True)\ntime_step_dim_sub = pysavime.define.ordered_subtar_dimension(time_step_dim, 0, num_time_steps - 1, True)\npos_x_dim_sub = pysavime.define.ordered_subtar_dimension(pos_x_dim, 0, num_pos_x - 1, True)\npos_y_dim_sub = pysavime.define.ordered_subtar_dimension(pos_y_dim, 0, num_pos_y - 1, True)\ntemperature_sub = pysavime.define.subtar_attribute(temperature, dataset)\n\nsubtar_dims = [group_dim_sub, time_step_dim_sub, pos_x_dim_sub, pos_y_dim_sub]\nsubtar_attrs = [temperature_sub]\nsubtar = pysavime.define.subtar(tar, subtar_dims, subtar_attrs)\nprint('- SubTar LOAD query', subtar.load_query_str())", "- Dataset CREATE query: CREATE_DATASET(\"temperature_data:double:1\", \"@data/temperatura.data\");\n- Tar CREATE query: CREATE_TAR(\"temperatures_tar\", \"*\", \"implicit, group, int32, 0, 399, 1 | implicit, time_step, int32, 0, 9, 1 | implicit, pos_x, int32, 0, 34, 1 | implicit, pos_y, int32, 0, 39, 1\", \"temperature, double: 1\");\n- SubTar LOAD query LOAD_SUBTAR(\"temperatures_tar\", \"ordered, group, 0,399 | ordered, time_step, 0,9 | ordered, pos_x, 0,34 | ordered, pos_y, 0,39\", \"temperature, temperature_data\")\n" ], [ "with pysavime.Client(host='127.0.0.1', port=65000, raise_silent_error=True) as client:\n client.execute(pysavime.operator.create(dataset))\n client.execute(pysavime.operator.create(tar))\n client.execute(pysavime.operator.load(subtar))", "_____no_output_____" ] ], [ [ "Abaixo verificamos se os dados foram corretamente registrados no SAVIME.", "_____no_output_____" ] ], [ [ "with pysavime.Client(host=savime_host, port=savime_port, raise_silent_error=True) as client:\n response = client.execute(pysavime.operator.select(tar))[0]\n \nis_the_same = np.isclose(response.attrs['temperature'].reshape(squeezed_array.shape),squeezed_array).all()\nprint('Checagem:', is_the_same)", "Checagem: True\n" ] ], [ [ "O próximo passo é executar o comando PREDICT.", "_____no_output_____" ] ], [ [ "# Vamos selecionar apenas os 9 primeiros instantes de tempo\ncmd = pysavime.operator.subset(tar, time_step_dim.name, 0, 8)\n\n# Definir as dimensões de entrada e saída do nosso modelo\ninput_dims_spec = [(group_dim.name, num_groups),\n (time_step_dim.name, num_time_steps - 1),\n (pos_x_dim.name, num_pos_x),\n (pos_y_dim.name, num_pos_y)]\n\noutput_dims_spec = [(\"time\", 9)]\n\nregister_cmd = pysavime.operator.register_model(model_identifier=chosen_model_name, \n input_dim_specification=input_dims_spec, \n output_dim_specification=output_dims_spec,\n attribute_specification=[temperature.name])\n\npredict_cmd = pysavime.operator.predict(tar=cmd, model_identifier=chosen_model_name)\n\nprint(register_cmd)\nprint(predict_cmd)", "REGISTER_MODEL(arima_25, \"group-400|time_step-9|pos_x-35|pos_y-40\", \"time-9\", \"temperature\")\nPREDICT(SUBSET(temperatures_tar, time_step, 0, 8), arima_25)\n" ], [ "with pysavime.Client(host=savime_host, port=savime_port, raise_silent_error=True) as client:\n client.execute(register_cmd) \n response = client.execute(predict_cmd)[0]", "_____no_output_____" ], [ "pandas_converter = DataVariableBlockConverter('pandas')\npandas_converter(response)", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Convolutional Neural Networks: Application\n\nWelcome to Course 4's second assignment! In this notebook, you will:\n\n- Implement helper functions that you will use when implementing a TensorFlow model\n- Implement a fully functioning ConvNet using TensorFlow \n\n**After this assignment you will be able to:**\n\n- Build and train a ConvNet in TensorFlow for a classification problem \n\nWe assume here that you are already familiar with TensorFlow. If you are not, please refer the *TensorFlow Tutorial* of the third week of Course 2 (\"*Improving deep neural networks*\").", "_____no_output_____" ], [ "### <font color='darkblue'> Updates to Assignment <font>\n\n#### If you were working on a previous version\n* The current notebook filename is version \"1a\". \n* You can find your work in the file directory as version \"1\".\n* To view the file directory, go to the menu \"File->Open\", and this will open a new tab that shows the file directory.\n\n#### List of Updates\n* `initialize_parameters`: added details about tf.get_variable, `eval`. Clarified test case.\n* Added explanations for the kernel (filter) stride values, max pooling, and flatten functions.\n* Added details about softmax cross entropy with logits.\n* Added instructions for creating the Adam Optimizer.\n* Added explanation of how to evaluate tensors (optimizer and cost).\n* `forward_propagation`: clarified instructions, use \"F\" to store \"flatten\" layer.\n* Updated print statements and 'expected output' for easier visual comparisons.\n* Many thanks to Kevin P. Brown (mentor for the deep learning specialization) for his suggestions on the assignments in this course!", "_____no_output_____" ], [ "## 1.0 - TensorFlow model\n\nIn the previous assignment, you built helper functions using numpy to understand the mechanics behind convolutional neural networks. Most practical applications of deep learning today are built using programming frameworks, which have many built-in functions you can simply call. \n\nAs usual, we will start by loading in the packages. ", "_____no_output_____" ] ], [ [ "import math\nimport numpy as np\nimport h5py\nimport matplotlib.pyplot as plt\nimport scipy\nfrom PIL import Image\nfrom scipy import ndimage\nimport tensorflow as tf\nfrom tensorflow.python.framework import ops\nfrom cnn_utils import *\n\n%matplotlib inline\nnp.random.seed(1)", "_____no_output_____" ] ], [ [ "Run the next cell to load the \"SIGNS\" dataset you are going to use.", "_____no_output_____" ] ], [ [ "# Loading the data (signs)\nX_train_orig, Y_train_orig, X_test_orig, Y_test_orig, classes = load_dataset()", "_____no_output_____" ] ], [ [ "As a reminder, the SIGNS dataset is a collection of 6 signs representing numbers from 0 to 5.\n\n<img src=\"images/SIGNS.png\" style=\"width:800px;height:300px;\">\n\nThe next cell will show you an example of a labelled image in the dataset. Feel free to change the value of `index` below and re-run to see different examples. ", "_____no_output_____" ] ], [ [ "# Example of a picture\nindex = 6\nplt.imshow(X_train_orig[index])\nprint (\"y = \" + str(np.squeeze(Y_train_orig[:, index])))", "y = 2\n" ] ], [ [ "In Course 2, you had built a fully-connected network for this dataset. But since this is an image dataset, it is more natural to apply a ConvNet to it.\n\nTo get started, let's examine the shapes of your data. ", "_____no_output_____" ] ], [ [ "X_train = X_train_orig/255.\nX_test = X_test_orig/255.\nY_train = convert_to_one_hot(Y_train_orig, 6).T\nY_test = convert_to_one_hot(Y_test_orig, 6).T\nprint (\"number of training examples = \" + str(X_train.shape[0]))\nprint (\"number of test examples = \" + str(X_test.shape[0]))\nprint (\"X_train shape: \" + str(X_train.shape))\nprint (\"Y_train shape: \" + str(Y_train.shape))\nprint (\"X_test shape: \" + str(X_test.shape))\nprint (\"Y_test shape: \" + str(Y_test.shape))\nconv_layers = {}", "number of training examples = 1080\nnumber of test examples = 120\nX_train shape: (1080, 64, 64, 3)\nY_train shape: (1080, 6)\nX_test shape: (120, 64, 64, 3)\nY_test shape: (120, 6)\n" ] ], [ [ "### 1.1 - Create placeholders\n\nTensorFlow requires that you create placeholders for the input data that will be fed into the model when running the session.\n\n**Exercise**: Implement the function below to create placeholders for the input image X and the output Y. You should not define the number of training examples for the moment. To do so, you could use \"None\" as the batch size, it will give you the flexibility to choose it later. Hence X should be of dimension **[None, n_H0, n_W0, n_C0]** and Y should be of dimension **[None, n_y]**. [Hint: search for the tf.placeholder documentation\"](https://www.tensorflow.org/api_docs/python/tf/placeholder).", "_____no_output_____" ] ], [ [ "# GRADED FUNCTION: create_placeholders\n\ndef create_placeholders(n_H0, n_W0, n_C0, n_y):\n \"\"\"\n Creates the placeholders for the tensorflow session.\n \n Arguments:\n n_H0 -- scalar, height of an input image\n n_W0 -- scalar, width of an input image\n n_C0 -- scalar, number of channels of the input\n n_y -- scalar, number of classes\n \n Returns:\n X -- placeholder for the data input, of shape [None, n_H0, n_W0, n_C0] and dtype \"float\"\n Y -- placeholder for the input labels, of shape [None, n_y] and dtype \"float\"\n \"\"\"\n\n ### START CODE HERE ### (≈2 lines)\n X = tf.placeholder(tf.float32, [None, n_H0, n_W0, n_C0])\n Y = tf.placeholder(tf.float32, [None, n_y])\n ### END CODE HERE ###\n \n return X, Y", "_____no_output_____" ], [ "X, Y = create_placeholders(64, 64, 3, 6)\nprint (\"X = \" + str(X))\nprint (\"Y = \" + str(Y))", "X = Tensor(\"Placeholder:0\", shape=(?, 64, 64, 3), dtype=float32)\nY = Tensor(\"Placeholder_1:0\", shape=(?, 6), dtype=float32)\n" ] ], [ [ "**Expected Output**\n\n<table> \n<tr>\n<td>\n X = Tensor(\"Placeholder:0\", shape=(?, 64, 64, 3), dtype=float32)\n\n</td>\n</tr>\n<tr>\n<td>\n Y = Tensor(\"Placeholder_1:0\", shape=(?, 6), dtype=float32)\n\n</td>\n</tr>\n</table>", "_____no_output_____" ], [ "### 1.2 - Initialize parameters\n\nYou will initialize weights/filters $W1$ and $W2$ using `tf.contrib.layers.xavier_initializer(seed = 0)`. You don't need to worry about bias variables as you will soon see that TensorFlow functions take care of the bias. Note also that you will only initialize the weights/filters for the conv2d functions. TensorFlow initializes the layers for the fully connected part automatically. We will talk more about that later in this assignment.\n\n**Exercise:** Implement initialize_parameters(). The dimensions for each group of filters are provided below. Reminder - to initialize a parameter $W$ of shape [1,2,3,4] in Tensorflow, use:\n```python\nW = tf.get_variable(\"W\", [1,2,3,4], initializer = ...)\n```\n#### tf.get_variable()\n[Search for the tf.get_variable documentation](https://www.tensorflow.org/api_docs/python/tf/get_variable). Notice that the documentation says:\n```\nGets an existing variable with these parameters or create a new one.\n```\nSo we can use this function to create a tensorflow variable with the specified name, but if the variables already exist, it will get the existing variable with that same name.\n", "_____no_output_____" ] ], [ [ "# GRADED FUNCTION: initialize_parameters\n\ndef initialize_parameters():\n \"\"\"\n Initializes weight parameters to build a neural network with tensorflow. The shapes are:\n W1 : [4, 4, 3, 8]\n W2 : [2, 2, 8, 16]\n Note that we will hard code the shape values in the function to make the grading simpler.\n Normally, functions should take values as inputs rather than hard coding.\n Returns:\n parameters -- a dictionary of tensors containing W1, W2\n \"\"\"\n \n tf.set_random_seed(1) # so that your \"random\" numbers match ours\n \n ### START CODE HERE ### (approx. 2 lines of code)\n W1 = tf.get_variable(\"W1\", [4, 4, 3, 8], initializer=tf.contrib.layers.xavier_initializer(seed=0))\n W2 = tf.get_variable(\"W2\", [2, 2, 8, 16], initializer=tf.contrib.layers.xavier_initializer(seed=0))\n ### END CODE HERE ###\n\n parameters = {\"W1\": W1,\n \"W2\": W2}\n \n return parameters", "_____no_output_____" ], [ "tf.reset_default_graph()\nwith tf.Session() as sess_test:\n parameters = initialize_parameters()\n init = tf.global_variables_initializer()\n sess_test.run(init)\n print(\"W1[1,1,1] = \\n\" + str(parameters[\"W1\"].eval()[1,1,1]))\n print(\"W1.shape: \" + str(parameters[\"W1\"].shape))\n print(\"\\n\")\n print(\"W2[1,1,1] = \\n\" + str(parameters[\"W2\"].eval()[1,1,1]))\n print(\"W2.shape: \" + str(parameters[\"W2\"].shape))", "W1[1,1,1] = \n[ 0.00131723 0.14176141 -0.04434952 0.09197326 0.14984085 -0.03514394\n -0.06847463 0.05245192]\nW1.shape: (4, 4, 3, 8)\n\n\nW2[1,1,1] = \n[-0.08566415 0.17750949 0.11974221 0.16773748 -0.0830943 -0.08058\n -0.00577033 -0.14643836 0.24162132 -0.05857408 -0.19055021 0.1345228\n -0.22779644 -0.1601823 -0.16117483 -0.10286498]\nW2.shape: (2, 2, 8, 16)\n" ] ], [ [ "** Expected Output:**\n\n```\nW1[1,1,1] = \n[ 0.00131723 0.14176141 -0.04434952 0.09197326 0.14984085 -0.03514394\n -0.06847463 0.05245192]\nW1.shape: (4, 4, 3, 8)\n\n\nW2[1,1,1] = \n[-0.08566415 0.17750949 0.11974221 0.16773748 -0.0830943 -0.08058\n -0.00577033 -0.14643836 0.24162132 -0.05857408 -0.19055021 0.1345228\n -0.22779644 -0.1601823 -0.16117483 -0.10286498]\nW2.shape: (2, 2, 8, 16)\n```", "_____no_output_____" ], [ "### 1.3 - Forward propagation\n\nIn TensorFlow, there are built-in functions that implement the convolution steps for you.\n\n- **tf.nn.conv2d(X,W, strides = [1,s,s,1], padding = 'SAME'):** given an input $X$ and a group of filters $W$, this function convolves $W$'s filters on X. The third parameter ([1,s,s,1]) represents the strides for each dimension of the input (m, n_H_prev, n_W_prev, n_C_prev). Normally, you'll choose a stride of 1 for the number of examples (the first value) and for the channels (the fourth value), which is why we wrote the value as `[1,s,s,1]`. You can read the full documentation on [conv2d](https://www.tensorflow.org/api_docs/python/tf/nn/conv2d).\n\n- **tf.nn.max_pool(A, ksize = [1,f,f,1], strides = [1,s,s,1], padding = 'SAME'):** given an input A, this function uses a window of size (f, f) and strides of size (s, s) to carry out max pooling over each window. For max pooling, we usually operate on a single example at a time and a single channel at a time. So the first and fourth value in `[1,f,f,1]` are both 1. You can read the full documentation on [max_pool](https://www.tensorflow.org/api_docs/python/tf/nn/max_pool).\n\n- **tf.nn.relu(Z):** computes the elementwise ReLU of Z (which can be any shape). You can read the full documentation on [relu](https://www.tensorflow.org/api_docs/python/tf/nn/relu).\n\n- **tf.contrib.layers.flatten(P)**: given a tensor \"P\", this function takes each training (or test) example in the batch and flattens it into a 1D vector. \n * If a tensor P has the shape (m,h,w,c), where m is the number of examples (the batch size), it returns a flattened tensor with shape (batch_size, k), where $k=h \\times w \\times c$. \"k\" equals the product of all the dimension sizes other than the first dimension.\n * For example, given a tensor with dimensions [100,2,3,4], it flattens the tensor to be of shape [100, 24], where 24 = 2 * 3 * 4. You can read the full documentation on [flatten](https://www.tensorflow.org/api_docs/python/tf/contrib/layers/flatten).\n\n- **tf.contrib.layers.fully_connected(F, num_outputs):** given the flattened input F, it returns the output computed using a fully connected layer. You can read the full documentation on [full_connected](https://www.tensorflow.org/api_docs/python/tf/contrib/layers/fully_connected).\n\nIn the last function above (`tf.contrib.layers.fully_connected`), the fully connected layer automatically initializes weights in the graph and keeps on training them as you train the model. Hence, you did not need to initialize those weights when initializing the parameters.\n\n\n#### Window, kernel, filter\nThe words \"window\", \"kernel\", and \"filter\" are used to refer to the same thing. This is why the parameter `ksize` refers to \"kernel size\", and we use `(f,f)` to refer to the filter size. Both \"kernel\" and \"filter\" refer to the \"window.\"", "_____no_output_____" ], [ "**Exercise**\n\nImplement the `forward_propagation` function below to build the following model: `CONV2D -> RELU -> MAXPOOL -> CONV2D -> RELU -> MAXPOOL -> FLATTEN -> FULLYCONNECTED`. You should use the functions above. \n\nIn detail, we will use the following parameters for all the steps:\n - Conv2D: stride 1, padding is \"SAME\"\n - ReLU\n - Max pool: Use an 8 by 8 filter size and an 8 by 8 stride, padding is \"SAME\"\n - Conv2D: stride 1, padding is \"SAME\"\n - ReLU\n - Max pool: Use a 4 by 4 filter size and a 4 by 4 stride, padding is \"SAME\"\n - Flatten the previous output.\n - FULLYCONNECTED (FC) layer: Apply a fully connected layer without an non-linear activation function. Do not call the softmax here. This will result in 6 neurons in the output layer, which then get passed later to a softmax. In TensorFlow, the softmax and cost function are lumped together into a single function, which you'll call in a different function when computing the cost. ", "_____no_output_____" ] ], [ [ "# GRADED FUNCTION: forward_propagation\n\ndef forward_propagation(X, parameters):\n \"\"\"\n Implements the forward propagation for the model:\n CONV2D -> RELU -> MAXPOOL -> CONV2D -> RELU -> MAXPOOL -> FLATTEN -> FULLYCONNECTED\n \n Note that for simplicity and grading purposes, we'll hard-code some values\n such as the stride and kernel (filter) sizes. \n Normally, functions should take these values as function parameters.\n \n Arguments:\n X -- input dataset placeholder, of shape (input size, number of examples)\n parameters -- python dictionary containing your parameters \"W1\", \"W2\"\n the shapes are given in initialize_parameters\n\n Returns:\n Z3 -- the output of the last LINEAR unit\n \"\"\"\n \n # Retrieve the parameters from the dictionary \"parameters\" \n W1 = parameters['W1']\n W2 = parameters['W2']\n \n ### START CODE HERE ###\n # CONV2D: stride of 1, padding 'SAME'\n Z1 = tf.nn.conv2d(X, W1, strides=[1, 1, 1, 1], padding='SAME')\n # RELU\n A1 = tf.nn.relu(Z1)\n # MAXPOOL: window 8x8, stride 8, padding 'SAME'\n P1 = tf.nn.max_pool(A1, ksize=[1, 8, 8, 1], strides=[1, 8, 8, 1], padding='SAME')\n # CONV2D: filters W2, stride 1, padding 'SAME'\n Z2 = tf.nn.conv2d(P1, W2, strides=[1, 1, 1, 1], padding='SAME')\n # RELU\n A2 = tf.nn.relu(Z2)\n # MAXPOOL: window 4x4, stride 4, padding 'SAME'\n P2 = tf.nn.max_pool(A2, ksize=[1, 4, 4, 1], strides=[1, 4, 4, 1], padding='SAME')\n # FLATTEN\n F = tf.contrib.layers.flatten(P2)\n # FULLY-CONNECTED without non-linear activation function (not not call softmax).\n # 6 neurons in output layer. Hint: one of the arguments should be \"activation_fn=None\" \n Z3 = tf.contrib.layers.fully_connected(F, 6, activation_fn=None)\n ### END CODE HERE ###\n\n return Z3", "_____no_output_____" ], [ "tf.reset_default_graph()\n\nwith tf.Session() as sess:\n np.random.seed(1)\n X, Y = create_placeholders(64, 64, 3, 6)\n parameters = initialize_parameters()\n Z3 = forward_propagation(X, parameters)\n init = tf.global_variables_initializer()\n sess.run(init)\n a = sess.run(Z3, {X: np.random.randn(2,64,64,3), Y: np.random.randn(2,6)})\n print(\"Z3 = \\n\" + str(a))", "Z3 = \n[[-0.44670227 -1.57208765 -1.53049231 -2.31013036 -1.29104376 0.46852064]\n [-0.17601591 -1.57972014 -1.4737016 -2.61672091 -1.00810647 0.5747785 ]]\n" ] ], [ [ "**Expected Output**:\n\n```\nZ3 = \n[[-0.44670227 -1.57208765 -1.53049231 -2.31013036 -1.29104376 0.46852064]\n [-0.17601591 -1.57972014 -1.4737016 -2.61672091 -1.00810647 0.5747785 ]]\n```", "_____no_output_____" ], [ "### 1.4 - Compute cost\n\nImplement the compute cost function below. Remember that the cost function helps the neural network see how much the model's predictions differ from the correct labels. By adjusting the weights of the network to reduce the cost, the neural network can improve its predictions.\n\nYou might find these two functions helpful: \n\n- **tf.nn.softmax_cross_entropy_with_logits(logits = Z, labels = Y):** computes the softmax entropy loss. This function both computes the softmax activation function as well as the resulting loss. You can check the full documentation [softmax_cross_entropy_with_logits](https://www.tensorflow.org/api_docs/python/tf/nn/softmax_cross_entropy_with_logits).\n- **tf.reduce_mean:** computes the mean of elements across dimensions of a tensor. Use this to calculate the sum of the losses over all the examples to get the overall cost. You can check the full documentation [reduce_mean](https://www.tensorflow.org/api_docs/python/tf/reduce_mean).\n\n#### Details on softmax_cross_entropy_with_logits (optional reading)\n* Softmax is used to format outputs so that they can be used for classification. It assigns a value between 0 and 1 for each category, where the sum of all prediction values (across all possible categories) equals 1.\n* Cross Entropy is compares the model's predicted classifications with the actual labels and results in a numerical value representing the \"loss\" of the model's predictions.\n* \"Logits\" are the result of multiplying the weights and adding the biases. Logits are passed through an activation function (such as a relu), and the result is called the \"activation.\"\n* The function is named `softmax_cross_entropy_with_logits` takes logits as input (and not activations); then uses the model to predict using softmax, and then compares the predictions with the true labels using cross entropy. These are done with a single function to optimize the calculations.\n\n** Exercise**: Compute the cost below using the function above.", "_____no_output_____" ] ], [ [ "# GRADED FUNCTION: compute_cost \n\ndef compute_cost(Z3, Y):\n \"\"\"\n Computes the cost\n \n Arguments:\n Z3 -- output of forward propagation (output of the last LINEAR unit), of shape (number of examples, 6)\n Y -- \"true\" labels vector placeholder, same shape as Z3\n \n Returns:\n cost - Tensor of the cost function\n \"\"\"\n \n ### START CODE HERE ### (1 line of code)\n cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=Z3, labels=Y))\n ### END CODE HERE ###\n \n return cost", "_____no_output_____" ], [ "tf.reset_default_graph()\n\nwith tf.Session() as sess:\n np.random.seed(1)\n X, Y = create_placeholders(64, 64, 3, 6)\n parameters = initialize_parameters()\n Z3 = forward_propagation(X, parameters)\n cost = compute_cost(Z3, Y)\n init = tf.global_variables_initializer()\n sess.run(init)\n a = sess.run(cost, {X: np.random.randn(4,64,64,3), Y: np.random.randn(4,6)})\n print(\"cost = \" + str(a))", "cost = 2.91034\n" ] ], [ [ "**Expected Output**: \n```\ncost = 2.91034\n```", "_____no_output_____" ], [ "## 1.5 Model \n\nFinally you will merge the helper functions you implemented above to build a model. You will train it on the SIGNS dataset. \n\n**Exercise**: Complete the function below. \n\nThe model below should:\n\n- create placeholders\n- initialize parameters\n- forward propagate\n- compute the cost\n- create an optimizer\n\nFinally you will create a session and run a for loop for num_epochs, get the mini-batches, and then for each mini-batch you will optimize the function. [Hint for initializing the variables](https://www.tensorflow.org/api_docs/python/tf/global_variables_initializer)", "_____no_output_____" ], [ "#### Adam Optimizer\nYou can use `tf.train.AdamOptimizer(learning_rate = ...)` to create the optimizer. The optimizer has a `minimize(loss=...)` function that you'll call to set the cost function that the optimizer will minimize.\n\nFor details, check out the documentation for [Adam Optimizer](https://www.tensorflow.org/api_docs/python/tf/train/AdamOptimizer)", "_____no_output_____" ], [ "#### Random mini batches\nIf you took course 2 of the deep learning specialization, you implemented `random_mini_batches()` in the \"Optimization\" programming assignment. This function returns a list of mini-batches. It is already implemented in the `cnn_utils.py` file and imported here, so you can call it like this:\n```Python\nminibatches = random_mini_batches(X, Y, mini_batch_size = 64, seed = 0)\n```\n(You will want to choose the correct variable names when you use it in your code).", "_____no_output_____" ], [ "#### Evaluating the optimizer and cost\n\nWithin a loop, for each mini-batch, you'll use the `tf.Session` object (named `sess`) to feed a mini-batch of inputs and labels into the neural network and evaluate the tensors for the optimizer as well as the cost. Remember that we built a graph data structure and need to feed it inputs and labels and use `sess.run()` in order to get values for the optimizer and cost.\n\nYou'll use this kind of syntax:\n```\noutput_for_var1, output_for_var2 = sess.run(\n fetches=[var1, var2],\n feed_dict={var_inputs: the_batch_of_inputs,\n var_labels: the_batch_of_labels}\n )\n```\n* Notice that `sess.run` takes its first argument `fetches` as a list of objects that you want it to evaluate (in this case, we want to evaluate the optimizer and the cost). \n* It also takes a dictionary for the `feed_dict` parameter. \n* The keys are the `tf.placeholder` variables that we created in the `create_placeholders` function above. \n* The values are the variables holding the actual numpy arrays for each mini-batch. \n* The sess.run outputs a tuple of the evaluated tensors, in the same order as the list given to `fetches`. \n\nFor more information on how to use sess.run, see the documentation [tf.Sesssion#run](https://www.tensorflow.org/api_docs/python/tf/Session#run) documentation.", "_____no_output_____" ] ], [ [ "# GRADED FUNCTION: model\n\ndef model(X_train, Y_train, X_test, Y_test, learning_rate = 0.009,\n num_epochs = 100, minibatch_size = 64, print_cost = True):\n \"\"\"\n Implements a three-layer ConvNet in Tensorflow:\n CONV2D -> RELU -> MAXPOOL -> CONV2D -> RELU -> MAXPOOL -> FLATTEN -> FULLYCONNECTED\n \n Arguments:\n X_train -- training set, of shape (None, 64, 64, 3)\n Y_train -- test set, of shape (None, n_y = 6)\n X_test -- training set, of shape (None, 64, 64, 3)\n Y_test -- test set, of shape (None, n_y = 6)\n learning_rate -- learning rate of the optimization\n num_epochs -- number of epochs of the optimization loop\n minibatch_size -- size of a minibatch\n print_cost -- True to print the cost every 100 epochs\n \n Returns:\n train_accuracy -- real number, accuracy on the train set (X_train)\n test_accuracy -- real number, testing accuracy on the test set (X_test)\n parameters -- parameters learnt by the model. They can then be used to predict.\n \"\"\"\n \n ops.reset_default_graph() # to be able to rerun the model without overwriting tf variables\n tf.set_random_seed(1) # to keep results consistent (tensorflow seed)\n seed = 3 # to keep results consistent (numpy seed)\n (m, n_H0, n_W0, n_C0) = X_train.shape \n n_y = Y_train.shape[1] \n costs = [] # To keep track of the cost\n \n # Create Placeholders of the correct shape\n ### START CODE HERE ### (1 line)\n X, Y = create_placeholders(n_H0, n_W0, n_C0, n_y)\n ### END CODE HERE ###\n\n # Initialize parameters\n ### START CODE HERE ### (1 line)\n parameters = initialize_parameters()\n ### END CODE HERE ###\n \n # Forward propagation: Build the forward propagation in the tensorflow graph\n ### START CODE HERE ### (1 line)\n Z3 = forward_propagation(X, parameters)\n ### END CODE HERE ###\n \n # Cost function: Add cost function to tensorflow graph\n ### START CODE HERE ### (1 line)\n cost = compute_cost(Z3, Y)\n ### END CODE HERE ###\n \n # Backpropagation: Define the tensorflow optimizer. Use an AdamOptimizer that minimizes the cost.\n ### START CODE HERE ### (1 line)\n optimizer = tf.train.AdamOptimizer(learning_rate=learning_rate).minimize(cost)\n ### END CODE HERE ###\n \n # Initialize all the variables globally\n init = tf.global_variables_initializer()\n \n # Start the session to compute the tensorflow graph\n with tf.Session() as sess:\n \n # Run the initialization\n sess.run(init)\n \n # Do the training loop\n for epoch in range(num_epochs):\n\n minibatch_cost = 0.\n num_minibatches = int(m / minibatch_size) # number of minibatches of size minibatch_size in the train set\n seed = seed + 1\n minibatches = random_mini_batches(X_train, Y_train, minibatch_size, seed)\n\n for minibatch in minibatches:\n\n # Select a minibatch\n (minibatch_X, minibatch_Y) = minibatch\n \"\"\"\n # IMPORTANT: The line that runs the graph on a minibatch.\n # Run the session to execute the optimizer and the cost.\n # The feedict should contain a minibatch for (X,Y).\n \"\"\"\n ### START CODE HERE ### (1 line)\n _ , temp_cost = sess.run([optimizer, cost], feed_dict={X:minibatch_X, Y:minibatch_Y})\n ### END CODE HERE ###\n \n minibatch_cost += temp_cost / num_minibatches\n \n\n # Print the cost every epoch\n if print_cost == True and epoch % 5 == 0:\n print (\"Cost after epoch %i: %f\" % (epoch, minibatch_cost))\n if print_cost == True and epoch % 1 == 0:\n costs.append(minibatch_cost)\n \n \n # plot the cost\n plt.plot(np.squeeze(costs))\n plt.ylabel('cost')\n plt.xlabel('iterations (per tens)')\n plt.title(\"Learning rate =\" + str(learning_rate))\n plt.show()\n\n # Calculate the correct predictions\n predict_op = tf.argmax(Z3, 1)\n correct_prediction = tf.equal(predict_op, tf.argmax(Y, 1))\n \n # Calculate accuracy on the test set\n accuracy = tf.reduce_mean(tf.cast(correct_prediction, \"float\"))\n print(accuracy)\n train_accuracy = accuracy.eval({X: X_train, Y: Y_train})\n test_accuracy = accuracy.eval({X: X_test, Y: Y_test})\n print(\"Train Accuracy:\", train_accuracy)\n print(\"Test Accuracy:\", test_accuracy)\n \n return train_accuracy, test_accuracy, parameters", "_____no_output_____" ] ], [ [ "Run the following cell to train your model for 100 epochs. Check if your cost after epoch 0 and 5 matches our output. If not, stop the cell and go back to your code!", "_____no_output_____" ] ], [ [ "_, _, parameters = model(X_train, Y_train, X_test, Y_test)", "Cost after epoch 0: 1.917929\nCost after epoch 5: 1.506757\nCost after epoch 10: 0.955359\nCost after epoch 15: 0.845802\nCost after epoch 20: 0.701174\nCost after epoch 25: 0.571977\nCost after epoch 30: 0.518435\nCost after epoch 35: 0.495806\nCost after epoch 40: 0.429827\nCost after epoch 45: 0.407291\nCost after epoch 50: 0.366394\nCost after epoch 55: 0.376922\nCost after epoch 60: 0.299491\nCost after epoch 65: 0.338870\nCost after epoch 70: 0.316400\nCost after epoch 75: 0.310413\nCost after epoch 80: 0.249549\nCost after epoch 85: 0.243457\nCost after epoch 90: 0.200031\nCost after epoch 95: 0.175452\n" ] ], [ [ "**Expected output**: although it may not match perfectly, your expected output should be close to ours and your cost value should decrease.\n\n<table> \n<tr>\n <td> \n **Cost after epoch 0 =**\n </td>\n\n <td> \n 1.917929\n </td> \n</tr>\n<tr>\n <td> \n **Cost after epoch 5 =**\n </td>\n\n <td> \n 1.506757\n </td> \n</tr>\n<tr>\n <td> \n **Train Accuracy =**\n </td>\n\n <td> \n 0.940741\n </td> \n</tr> \n\n<tr>\n <td> \n **Test Accuracy =**\n </td>\n\n <td> \n 0.783333\n </td> \n</tr> \n</table>", "_____no_output_____" ], [ "Congratulations! You have finished the assignment and built a model that recognizes SIGN language with almost 80% accuracy on the test set. If you wish, feel free to play around with this dataset further. You can actually improve its accuracy by spending more time tuning the hyperparameters, or using regularization (as this model clearly has a high variance). \n\nOnce again, here's a thumbs up for your work! ", "_____no_output_____" ] ], [ [ "fname = \"images/thumbs_up.jpg\"\nimage = np.array(ndimage.imread(fname, flatten=False))\nmy_image = scipy.misc.imresize(image, size=(64,64))\nplt.imshow(my_image)", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown", "markdown", "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown", "markdown", "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "from sqlalchemy import create_engine, MetaData, Table\nfrom sqlalchemy.orm import mapper, sessionmaker", "_____no_output_____" ], [ "class Story(object):\n pass\n\nclass PosTagEntry(object):\n pass\n \ndef loadSession():\n dbPath = '../corpus/stories.sqlite'\n engine = create_engine('sqlite:///%s' % dbPath, echo=True)\n \n metadata = MetaData(engine)\n\n bookmarks = Table('stories', metadata, autoload=True)\n mapper(Story, bookmarks)\n \n bookmarks = Table('pos_tags', metadata, autoload=True)\n mapper(PosTagEntry, bookmarks)\n \n Session = sessionmaker(bind=engine)\n session = Session()\n return session\n\nsession = loadSession()", "2017-11-22 15:17:33,669 INFO sqlalchemy.engine.base.Engine SELECT CAST('test plain returns' AS VARCHAR(60)) AS anon_1\n2017-11-22 15:17:33,669 INFO sqlalchemy.engine.base.Engine ()\n2017-11-22 15:17:33,671 INFO sqlalchemy.engine.base.Engine SELECT CAST('test unicode returns' AS VARCHAR(60)) AS anon_1\n2017-11-22 15:17:33,672 INFO sqlalchemy.engine.base.Engine ()\n2017-11-22 15:17:33,673 INFO sqlalchemy.engine.base.Engine PRAGMA table_info(\"stories\")\n2017-11-22 15:17:33,673 INFO sqlalchemy.engine.base.Engine ()\n2017-11-22 15:17:33,676 INFO sqlalchemy.engine.base.Engine SELECT sql FROM (SELECT * FROM sqlite_master UNION ALL SELECT * FROM sqlite_temp_master) WHERE name = 'stories' AND type = 'table'\n2017-11-22 15:17:33,678 INFO sqlalchemy.engine.base.Engine ()\n2017-11-22 15:17:33,680 INFO sqlalchemy.engine.base.Engine PRAGMA foreign_key_list(\"stories\")\n2017-11-22 15:17:33,681 INFO sqlalchemy.engine.base.Engine ()\n2017-11-22 15:17:33,682 INFO sqlalchemy.engine.base.Engine SELECT sql FROM (SELECT * FROM sqlite_master UNION ALL SELECT * FROM sqlite_temp_master) WHERE name = 'stories' AND type = 'table'\n2017-11-22 15:17:33,682 INFO sqlalchemy.engine.base.Engine ()\n2017-11-22 15:17:33,685 INFO sqlalchemy.engine.base.Engine PRAGMA index_list(\"stories\")\n2017-11-22 15:17:33,685 INFO sqlalchemy.engine.base.Engine ()\n2017-11-22 15:17:33,686 INFO sqlalchemy.engine.base.Engine PRAGMA index_list(\"stories\")\n2017-11-22 15:17:33,686 INFO sqlalchemy.engine.base.Engine ()\n2017-11-22 15:17:33,687 INFO sqlalchemy.engine.base.Engine SELECT sql FROM (SELECT * FROM sqlite_master UNION ALL SELECT * FROM sqlite_temp_master) WHERE name = 'stories' AND type = 'table'\n2017-11-22 15:17:33,688 INFO sqlalchemy.engine.base.Engine ()\n2017-11-22 15:17:33,693 INFO sqlalchemy.engine.base.Engine PRAGMA table_info(\"pos_tags\")\n2017-11-22 15:17:33,694 INFO sqlalchemy.engine.base.Engine ()\n2017-11-22 15:17:33,695 INFO sqlalchemy.engine.base.Engine SELECT sql FROM (SELECT * FROM sqlite_master UNION ALL SELECT * FROM sqlite_temp_master) WHERE name = 'pos_tags' AND type = 'table'\n2017-11-22 15:17:33,696 INFO sqlalchemy.engine.base.Engine ()\n2017-11-22 15:17:33,697 INFO sqlalchemy.engine.base.Engine PRAGMA foreign_key_list(\"pos_tags\")\n2017-11-22 15:17:33,697 INFO sqlalchemy.engine.base.Engine ()\n2017-11-22 15:17:33,699 INFO sqlalchemy.engine.base.Engine SELECT sql FROM (SELECT * FROM sqlite_master UNION ALL SELECT * FROM sqlite_temp_master) WHERE name = 'pos_tags' AND type = 'table'\n2017-11-22 15:17:33,699 INFO sqlalchemy.engine.base.Engine ()\n2017-11-22 15:17:33,700 INFO sqlalchemy.engine.base.Engine PRAGMA index_list(\"pos_tags\")\n2017-11-22 15:17:33,700 INFO sqlalchemy.engine.base.Engine ()\n2017-11-22 15:17:33,701 INFO sqlalchemy.engine.base.Engine PRAGMA index_list(\"pos_tags\")\n2017-11-22 15:17:33,702 INFO sqlalchemy.engine.base.Engine ()\n2017-11-22 15:17:33,704 INFO sqlalchemy.engine.base.Engine PRAGMA index_info(\"sqlite_autoindex_pos_tags_1\")\n2017-11-22 15:17:33,705 INFO sqlalchemy.engine.base.Engine ()\n2017-11-22 15:17:33,707 INFO sqlalchemy.engine.base.Engine SELECT sql FROM (SELECT * FROM sqlite_master UNION ALL SELECT * FROM sqlite_temp_master) WHERE name = 'pos_tags' AND type = 'table'\n2017-11-22 15:17:33,708 INFO sqlalchemy.engine.base.Engine ()\n" ], [ "stories = session.query(Story).all()\nprint(len(stories))\nprint(dir(stories[0]))", "2017-11-22 15:17:33,809 INFO sqlalchemy.engine.base.Engine BEGIN (implicit)\n2017-11-22 15:17:33,811 INFO sqlalchemy.engine.base.Engine SELECT stories.id AS stories_id, stories.title AS stories_title, stories.published AS stories_published, stories.tags AS stories_tags, stories.text AS stories_text, stories.likes AS stories_likes, stories.hrefs AS stories_hrefs, stories.url AS stories_url \nFROM stories\n2017-11-22 15:17:33,815 INFO sqlalchemy.engine.base.Engine ()\n23558\n['__class__', '__delattr__', '__dict__', '__dir__', '__doc__', '__eq__', '__format__', '__ge__', '__getattribute__', '__gt__', '__hash__', '__init__', '__init_subclass__', '__le__', '__lt__', '__module__', '__ne__', '__new__', '__reduce__', '__reduce_ex__', '__repr__', '__setattr__', '__sizeof__', '__str__', '__subclasshook__', '__weakref__', '_sa_class_manager', '_sa_instance_state', 'hrefs', 'id', 'likes', 'published', 'tags', 'text', 'title', 'url']\n" ], [ "print(stories[0].text)", "Работаю в провинциальном городе в магазине отделочных материалов и сантехники.Заходит к нам на днях надменная пергидролевая дева, покрытая слоем штукатурки толщиной в палец. Собирается с мыслями, напускает на себя важный вид и обращается ко мне:Дева, медленно и с видом опытного сантехника: Молодой человек, у вас ванны железные есть?Я: Нет, у нас только акрил. Металлических нет.Дева: Молодой человек, я не спрашиваю металлические, я спрашиваю железные!Я: Извините, железных тоже нет.Дева презрительно смотрит на меня, бурчит что-то себе под нос, и, виляя бедрами, уходит. Смотрим в окно. Выходит. Подходит к побитой жизнью шестерке, деловито садится на переднее сиденье, подзывает торопливо курящего поодаль водителя.Дева, возмущенно: Понабрали крестьян, металлические ванны от железных не отличают!Водитель, тяжело вздохнув, затаптывает окурок, занимает свое место, и экипаж отправляется дальше, на поиски волшебной неметаллической ванны из железа.\n" ], [ "# too slow\n#pos_tags = session.query(PosTagEntry).all()\n#print(len(pos_tags))\n#print(dir(pos_tags[0]))", "_____no_output_____" ], [ "import scripts.get_tagged_text as get_tagged", "_____no_output_____" ], [ "print(get_tagged.get_tagged_story(2, session, Story, PosTagEntry))", "2017-11-22 15:17:34,735 INFO sqlalchemy.engine.base.Engine SELECT pos_tags.id AS pos_tags_id, pos_tags.story_id AS pos_tags_story_id, pos_tags.tag AS pos_tags_tag, pos_tags.start AS pos_tags_start, pos_tags.\"end\" AS pos_tags_end \nFROM pos_tags \nWHERE pos_tags.story_id = ?\n2017-11-22 15:17:34,736 INFO sqlalchemy.engine.base.Engine (2,)\nРаботаю<V> в<PR> провинциальном<A=m> городе<S> в<PR> магазине<S> отделочных<A=pl> материалов<S> и<CONJ> сантехники<S>.Заходит<V> к<PR> нам<S-PRO> на<PR> днях<S> надменная<A=f> пергидролевая<A=f> дева<S>, покрытая<V> слоем<S> штукатурки<S> толщиной<S> в<PR> палец<S>. Собирается<V> с<PR> мыслями<S>, напускает<V> на<PR> себя<S-PRO=acc> важный<A=m> вид<S> и<CONJ> обращается<V> ко<PR> мне<S-PRO>:Дева<S>, медленно<ADV> и<CONJ> с<PR> видом<S> опытного<A=m> сантехника<S>: Молодой<A=m> человек<S>, у<PR> вас<S-PRO> ванны<A=pl> железные<A=pl> есть<V>?Я<S-PRO>: Нет<PART>, у<PR> нас<S-PRO> только<PART> акрил<V>. Металлических<A=pl> нет<PRAEDIC>.Дева<S>: Молодой<A=sg> человек<S>, я<S-PRO> не<PART> спрашиваю<V> металлические<A=pl>, я<S-PRO> спрашиваю<V> железные<A=pl>!Я<S-PRO>: Извините<V>, железных<A=pl> тоже<PART> нет<PRAEDIC>.Дева<S> презрительно<ADV> смотрит<V> на<PR> меня<S-PRO>, бурчит<V> что<CONJ>-то<S-PRO> себе<S-PRO=dat> под<PR> нос<S>, и<CONJ>, виляя<V> бедрами<S>, уходит<V>. Смотрим<V> в<PR> окно<S>. Выходит<V>. Подходит<V> к<PR> побитой<A=f> жизнью<S> шестерке<S>, деловито<ADV> садится<V> на<PR> переднее<A=n> сиденье<S>, подзывает<V> торопливо<ADV> курящего<V> поодаль<S> водителя<S>.Дева<S>, возмущенно<ADV>: Понабрали<V> крестьян<S>, металлические<A=pl> ванны<S> от<PR> железных<A=pl> не<PART> отличают<V>!Водитель<S>, тяжело<ADV> вздохнув<V>, затаптывает<V> окурок<S>, занимает<V> свое<A=n> место<S>, и<CONJ> экипаж<S> отправляется<V> дальше<ADV=comp>, на<PR> поиски<S> волшебной<A=f> неметаллической<A=f> ванны<S> из<PR> железа<S>.\n" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Imports", "_____no_output_____" ] ], [ [ "import requests\nimport getpass\nimport pickle\nimport io\nimport time\nimport pandas as pd\nimport matplotlib.pyplot as plt\nimport seaborn as sns\nimport numpy as np\nimport itertools", "_____no_output_____" ] ], [ [ "# Login\nhttps://www.statistikdaten.bayern.de/genesis/online?Menu=Anmeldung#abreadcrumb", "_____no_output_____" ] ], [ [ "username = input()", "_____no_output_____" ], [ "password = getpass.getpass()", "_____no_output_____" ] ], [ [ "## Test login", "_____no_output_____" ] ], [ [ "class GenesisApi:\n \n def __init__(self, username, password, polling_rate=5):\n self.username = username\n self.password = password\n self.polling_rate = polling_rate\n \n self.__base_url = 'https://www.statistikdaten.bayern.de/genesisWS/rest/2020/'\n \n self.__base_params = {\n 'username': username,\n 'password': password,\n 'language': 'de'\n }\n \n self.__default_table_params = self.__base_params.copy()\n self.__default_table_params.update({\n 'name': '',\n 'area': 'all',\n 'compress': 'false',\n 'transpose': 'false',\n 'startyear': '',\n 'endyear': '',\n 'timeslices': '',\n 'regionalvariable': '',\n 'regionalkey': '',\n 'classifyingkey1': '',\n 'classifyingvariable2': '',\n 'classifyingkey2': '',\n 'classifyingvariable3': '',\n 'classifyingkey3': '',\n 'job': 'true'\n })\n \n self.__default_jobs_params = self.__base_params.copy()\n self.__default_jobs_params.update({\n 'selection': '',\n 'searchcriterion': 'code',\n 'sortcriterion': 'code',\n 'type': 'all',\n 'area': 'all',\n 'pagelength': '100'\n })\n \n self.__default_result_params = self.__base_params.copy()\n self.__default_result_params.update({\n 'name': '',\n 'area': 'all',\n 'compress': 'false'\n })\n \n def check_login(self):\n response = requests.get(self.__base_url + 'helloworld/logincheck', params=self.__base_params)\n b'{\"Status\":\"Sie wurden erfolgreich an- und abgemeldet!\",\"Username\":\"GB3U65P838\"}'\n try:\n return response.json()['Status'] == 'Sie wurden erfolgreich an- und abgemeldet!'\n except Exception as e:\n return False\n\n\n def get_table(self, name, startyear=''):\n startyear = str(startyear)\n \n params = self.__default_table_params.copy()\n params['name'] = name\n params['startyear'] = startyear\n \n response = requests.get(self.__base_url + 'data/table', params=params)\n \n data = response.json()\n code = data['Status']['Code']\n if (code == 0): # Success\n return data\n elif (code == 99): # Table is too big a job has been created\n print('Table is too big, created a job.')\n result_name = data['Status']['Content'].split(':', 1)[1][1:]\n return self.get_job_result(result_name)\n else:\n params['password'] = '***'\n print('Error requesting ' + name + ' with params:', params, 'response:', data)\n return data\n \n def is_job_ready(self, name):\n params = self.__default_jobs_params.copy()\n params['selection'] = 'Werteabruf ' + name\n \n response = requests.get(self.__base_url + 'catalogue/jobs', params=params)\n try:\n return response.json()['List'][0]['State'] == 'Fertig'\n except Exception as e:\n return False\n \n def delete_job_result(self, name):\n params = self.__default_result_params.copy()\n params['name'] = name\n response = requests.get(self.__base_url + 'profile/removeResult', params=params)\n return response \n \n def get_job_result(self, name):\n params = self.__default_result_params.copy()\n params['name'] = name\n \n while(not self.is_job_ready(name)):\n print('Data is not ready waiting ' + str(self.polling_rate) + ' seconds longer.')\n time.sleep(self.polling_rate)\n \n response = requests.get(self.__base_url + 'data/result', params=params)\n self.delete_job_result(name)\n return response.json()\n", "_____no_output_____" ], [ "genesis = GenesisApi(username, password)\ngenesis.check_login()", "_____no_output_____" ] ], [ [ "# Download data\n\nNote: This takes a long time", "_____no_output_____" ] ], [ [ "responses_demographic = {}\n\nfor year in range(1980, 2020 + 1):\n print('Requesting table for the year ' + str(year))\n response = genesis.get_table('12411-003r', year)\n print('Got data')\n responses_demographic[str(year)] = response", "_____no_output_____" ], [ "responses_area = {}\n\n# 33111-201r 1980, 1984, 1988, 1992, 1996, 2000, 2004, 2008, 2009, 2010, 2011, 2012, 2013\n# 33111-101r 2011 - 2015\n# 33111-001r 2014 - 2020\n\nfor year in [1980, 1984, 1988, 1992, 1996, 2000, 2004, 2008, 2009, 2010, 2011, 2012, 2013]:\n print('Requesting table for the year ' + str(year))\n response = genesis.get_table('33111-201r', year)\n print('Got data')\n responses_area[str(year)] = response\n\nfor year in range(2014, 2020 + 1):\n print('Requesting table for the year ' + str(year))\n response = genesis.get_table('33111-001r', year)\n print('Got data')\n responses_area[str(year)] = response", "_____no_output_____" ] ], [ [ "# Convert to DataFrame", "_____no_output_____" ] ], [ [ "def convert_to_dataframe(response, start_at_line, date_line, header_line):\n raw_content = response['Object']['Content']\n content = raw_content.split('\\n', start_at_line)\n date = content[date_line].split(';',1)[0]\n csv = io.StringIO(content[header_line] + '\\n' + content[start_at_line].split('\\n__________', 1)[0])\n df = pd.read_csv(csv, ';')\n df['date'] = pd.to_datetime(date, format='%d.%m.%Y')\n return df", "_____no_output_____" ] ], [ [ "## Demographic", "_____no_output_____" ] ], [ [ "dfs = list()\nfor year, response in responses_demographic.items():\n df = convert_to_dataframe(response, start_at_line=6, date_line=4, header_line=5)\n dfs.append(df)\n\ndf_demographic = pd.concat(dfs, axis=0, ignore_index=True)\n\ncolumn_names = df_demographic.columns.values\ncolumn_names[0] = 'AGS'\ncolumn_names[1] = 'Gemeinde'\ndf_demographic.columns = column_names\n\ndf_demographic['Gemeinde'] = df_demographic['Gemeinde'].str.strip()\ndf_demographic['Insgesamt'] = pd.to_numeric(df_demographic['Insgesamt'], errors='coerce')\ndf_demographic['männlich'] = pd.to_numeric(df_demographic['männlich'], errors='coerce')\ndf_demographic['weiblich'] = pd.to_numeric(df_demographic['weiblich'], errors='coerce')", "_____no_output_____" ], [ "df_demographic\n# TODO Filter regierungsbezirke\n# TODO Filter male and female", "_____no_output_____" ] ], [ [ "## Area", "_____no_output_____" ] ], [ [ "dfs = list()\nfor year, response in responses_area.items():\n df = convert_to_dataframe(response, start_at_line=10, date_line=5, header_line=8)\n\n column_names = df.columns.values\n column_names[0] = 'AGS'\n column_names[1] = 'Gemeinde'\n df.columns = column_names\n\n for column_name in column_names[2: len(column_names) - 1]:\n df[column_name] = pd.to_numeric(df[column_name].str.replace(',', '.'), errors='coerce')\n\n df['Gemeinde'] = df['Gemeinde'].str.strip()\n \n dfs.append(df)\n\ndf_area = pd.concat(dfs, axis=0, ignore_index=True)", "_____no_output_____" ], [ "df_area", "_____no_output_____" ], [ "# TODO Map old area codes to new ones\n# TODO Map area codes to sealed and non-sealed\n# TODO Filter regierungsbezirke", "_____no_output_____" ] ], [ [ "## Combined", "_____no_output_____" ] ], [ [ "df_all = pd.merge(df_area, df_demographic, how='left', on=['AGS', 'Gemeinde', 'date'])\ndf_all.rename(columns={'Insgesamt_x':'Insgesamt Fläche', 'Insgesamt_y':'Insgesamt Bewohner'}, inplace=True)\ndf_all", "_____no_output_____" ] ], [ [ "# Save and load data", "_____no_output_____" ] ], [ [ "df_demographic.to_pickle('df_demographic.pickle')\ndf_area.to_pickle('df_area.pickle')\n\nwith open('responses_demographic.pickle', 'wb') as f:\n pickle.dump(responses_demographic, f, pickle.HIGHEST_PROTOCOL)\n\nwith open('responses_area.pickle', 'wb') as f:\n pickle.dump(responses_area, f, pickle.HIGHEST_PROTOCOL)", "_____no_output_____" ], [ "df_demographic = pd.read_pickle('df_demographic.pickle')\ndf_area = pd.read_pickle('df_area.pickle')\n\nwith open('responses_demographic.pickle', 'rb') as f:\n responses_demographic = pickle.load(f)\n \nwith open('responses_area.pickle', 'rb') as f:\n responses_area = pickle.load(f)", "_____no_output_____" ] ], [ [ "## Categorize", "_____no_output_____" ] ], [ [ "categories = {\n \"living\": [\n \"Wohnen\",\n \"11000 Wohnbaufläche\",\n ],\n\n \"industry\": [\n \"Gewerbe, Industrie\",\n \"Betriebsfläche (ohne Abbauland)\",\n \"Abbauland\",\n \"12100 Industrie und Gewerbe\",\n \"12200 Handel und Dienstleistung\",\n \"12300 Versorgungsanlage\",\n \"12400 Entsorgung\",\n \"13000 Halde\",\n \"14000 Bergbaubetrieb\",\n \"15000 Tagebau, Grube, Steinbruch\",\n ],\n\n \"transport_infrastructure\": [\n \"Straße, Weg, Platz\",\n \"sonstige Verkehrsfläche\",\n \"21000 Straßenverkehr\",\n \"22000 Weg\",\n \"23000 Platz\",\n \"24000 Bahnverkehr\",\n \"25000 Flugverkehr\",\n \"26000 Schiffsverkehr\",\n \"42000 Hafenbecken\",\n ],\n\n \"nature_and_water\": [\n \"Moor\",\n \"Landwirtschaftsfläche (ohne Moor, Heide)\",\n \"Grünanlage\",\n \"Heide\",\n \"Waldfläche\",\n \"Wasserfläche\",\n \"Unland\",\n \"18400 Grünanlage\",\n \"31100 Ackerland\",\n \"31200 Grünland\",\n \"31300 Gartenland\",\n \"31400 Weingarten\",\n \"31500 Obstplantage\",\n \"32000 Wald\",\n \"33000 Gehölz\",\n \"34000 Heide\",\n \"35000 Moor\",\n \"36000 Sumpf\",\n \"37000 Unland, Vegetationslose Fläche\",\n \"41000 Fließgewässer\",\n \"43000 Stehendes Gewässer\",\n ],\n\n \"miscellaneous\": [\n \"Flächen anderer Nutzung (ohne Unland, Friedhof)\",\n \"sonstige Erholungsfläche\",\n \"sonstige Gebäude- und Freifläche\",\n \"Friedhof\",\n \"16000 Fläche gemischter Nutzung\",\n \"17000 Fläche besonderer funktionaler Prägung\",\n \"18100 Sportanlage\",\n \"18200 Freizeitanlage\",\n \"19000 Friedhof\",\n \"18300 Erholungsfläche\",\n ]\n}", "_____no_output_____" ], [ "# Check if we classified all columns and used each only once\nall_columns = set(df_area.columns)\n\nfor l in categories.values():\n all_columns = all_columns - set(l)\n \nall_columns = all_columns - set(['AGS', 'Gemeinde', 'Insgesamt', 'date'])\n\nif (len(all_columns) != 0):\n print (\"The categories\", all_columns, \"have not yet been categorized.\")\n\nfor ((name1, l1), (name2, l2)) in itertools.combinations(categories.items(), 2):\n if (not set(l1).isdisjoint(l2)):\n print(name1, \"and\", name2, \"contain the same category.\")", "_____no_output_____" ], [ "for (name, category) in categories.items():\n df_area[name] = df_area.loc[:,category].sum(axis=1)\n df_area.drop(category, axis=1, inplace=True)\n df_area[name + '_percent'] = df_area[name] / df_area['Insgesamt']", "_____no_output_____" ], [ "used_areas = [\n \"living\",\n \"industry\",\n \"transport_infrastructure\"\n]\ndf_area['used_area'] = 0\nfor name in used_areas:\n df_area['used_area'] = df_area['used_area'] + df_area.loc[:,used_areas].sum(axis=1)\n\ndf_area['used_area_percent'] = df_area['used_area'] / df_area['Insgesamt']", "_____no_output_____" ] ], [ [ "## Rename columns", "_____no_output_____" ] ], [ [ "df_area.rename(columns={\"Insgesamt\": \"total\", \"Gemeinde\": \"municipality\"}, inplace=True)", "_____no_output_____" ] ], [ [ "## Filter unused municipalities", "_____no_output_____" ] ], [ [ "df_area = df_area[df_area[\"AGS\"] <= 9999]", "_____no_output_____" ] ], [ [ "## Merge demographic data", "_____no_output_____" ] ], [ [ "df_area", "_____no_output_____" ], [ "df_demographic.drop([\"männlich\", \"weiblich\"], axis=1, inplace=True)\ndf_demographic.rename(columns={\"Gemeinde\": \"municipality\", \"Insgesamt\": \"demographic\"}, inplace=True)", "_____no_output_____" ], [ "df_area = pd.merge(df_area, df_demographic, how='left', on=['AGS', 'municipality', 'date'])", "_____no_output_____" ] ], [ [ "## Export to JSON", "_____no_output_____" ] ], [ [ "df_area", "_____no_output_____" ], [ "df_export = df_area.copy()\ndf_export['date'] = df_export['date'].dt.strftime('%d.%m.%Y')\n\nwith open(\"data.json\", \"w\", encoding=\"utf-8\") as f:\n df_export.to_json(f, orient=\"records\", force_ascii=False)", "_____no_output_____" ] ], [ [ "# Basic graphs", "_____no_output_____" ] ], [ [ "f, ax = plt.subplots(figsize=(7, 7))\nax.set(yscale=\"log\")\ng = sns.lineplot(data=df_demographic[(df_demographic['Gemeinde']=='Friedberg, St') | (df_demographic['Gemeinde']=='Augsburg (Krfr.St)') | (df_demographic['Gemeinde']=='Garmisch-Partenkirchen, M')], style='Gemeinde', x='date', y='Insgesamt', ax=ax)\ng.set_title('Einwohner')\ng.set(ylim=(1, None))\ng\n#sns.lineplot(data=df_demographic, style='Gemeinde', x='date', y='Insgesamt', ax=ax)#, ylim=(0,300000))", "_____no_output_____" ], [ "f, ax = plt.subplots(figsize=(7, 7))\n#ax.set(yscale=\"log\")\ng = sns.lineplot(data=df_area[(df_area['municipality']=='Friedberg, St') | (df_area['municipality']=='Augsburg (Krfr.St)') | (df_area['municipality']=='Garmisch-Partenkirchen, M')], style='municipality', x='date', y='nature_and_water_percent', ax=ax)\ng.set_title('Natur und Wasserflächen')\n#g.set(ylim=(0, None))\ng", "_____no_output_____" ], [ "gem = ['Bayern']#, 'Oberbayern', 'Schwaben']\nsize = 10\nf, axs = plt.subplots(len(gem), 1, figsize=(size*3, len(gem)*size*3))\n\n#df_area_2 = df_area_2[df_area_2['date'] > pd.to_datetime(\"1.1.2010\", format='%d.%m.%Y')]\n\nfor i in range(0, len(gem)):\n g = df_area[(df_area['municipality']==gem[i])].plot.area(\n x='date', \n y=['living_percent', 'industry_percent', 'transport_infrastructure_percent', 'nature_and_water_percent', 'miscellaneous_percent'], \n stacked=True, \n ax=(axs if len(gem) == 1 else axs[i]))\n g.set_title('Flächen in ' + gem[i])\n g.set(ylim=(0, None))\n\nplt.savefig('flächen.jpg')", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "%matplotlib inline\nimport matplotlib.pyplot as plt\nimport numpy as np\nimport seaborn as sns\nimport tensorflow as tf\nfrom edward.models import Categorical, Normal\nimport edward as ed\nimport pandas as pd\n\nimport warnings\nwarnings.filterwarnings('ignore')", "C:\\ProgramData\\Anaconda3\\lib\\site-packages\\h5py\\__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.\n from ._conv import register_converters as _register_converters\n" ] ], [ [ "## Import data", "_____no_output_____" ] ], [ [ "pd.set_option('display.max_rows', 500)\npd.set_option('display.max_columns', 38)", "_____no_output_____" ], [ "df = pd.read_csv('zar_dataset.csv')", "_____no_output_____" ], [ "df.info()", "<class 'pandas.core.frame.DataFrame'>\nRangeIndex: 4986 entries, 0 to 4985\nData columns (total 37 columns):\nRSI 4986 non-null float64\nTSI 4986 non-null float64\nATR 4986 non-null float64\nBHBI 4986 non-null float64\nBBL 4986 non-null float64\nBBH 4986 non-null float64\nBLBI 4986 non-null float64\nBBMAVG 4986 non-null float64\nDCH 4986 non-null float64\nDCHI 4986 non-null float64\nDCL 4986 non-null float64\nDCLI 4986 non-null float64\nKCC 4986 non-null float64\nKCH 4986 non-null float64\nKCL 4986 non-null float64\nADX 4986 non-null float64\nADXI 4986 non-null int64\nADXN 4986 non-null float64\nADXP 4986 non-null float64\nCCI 4986 non-null float64\nDPO 4986 non-null float64\nSEMA 4986 non-null float64\nLEMA 4986 non-null float64\nIchimoku 4986 non-null float64\nIchimoku_b 4986 non-null float64\nKST 4986 non-null float64\nKST_SIG 4986 non-null float64\nMACD 4986 non-null float64\nMACD_DIFF 4986 non-null float64\nMACD_SIG 4986 non-null float64\nMI 4986 non-null float64\nTRIX 4986 non-null float64\nVIN 4986 non-null float64\nVIP 4986 non-null float64\nCR 4986 non-null float64\nDR 4986 non-null float64\ntarget_return 4986 non-null float64\ndtypes: float64(36), int64(1)\nmemory usage: 1.4 MB\n" ], [ "df.head()", "_____no_output_____" ], [ "df.describe()", "_____no_output_____" ] ], [ [ "## Create Labels For Data, Classification and Regression Labels", "_____no_output_____" ] ], [ [ "from sklearn.preprocessing import StandardScaler\n\nfrom sklearn.model_selection import train_test_split", "_____no_output_____" ], [ "scaler = StandardScaler()", "_____no_output_____" ], [ "df['target_clf'] = df['target_return'].apply(lambda x: float(x/abs(x)) if x!=0 else -1)", "_____no_output_____" ], [ "df['target_clf'] = df['target_clf'].apply(lambda x: x if x==1 else 0)", "_____no_output_____" ], [ "df.rename(columns = {'target_return':'target_reg'}, inplace = True)", "_____no_output_____" ], [ "df.head()", "_____no_output_____" ], [ "df.describe()", "_____no_output_____" ], [ "df.columns", "_____no_output_____" ], [ "df.head()", "_____no_output_____" ], [ "X = df[['RSI', 'TSI', 'ATR', 'BHBI', 'BBL', 'BBH', 'BLBI', 'BBMAVG', 'DCH',\n 'DCHI', 'DCL', 'DCLI', 'KCC', 'KCH', 'KCL', 'ADX', 'ADXI', 'ADXN',\n 'ADXP', 'CCI', 'DPO', 'SEMA', 'LEMA', 'Ichimoku', 'Ichimoku_b', 'KST',\n 'KST_SIG', 'MACD', 'MACD_DIFF', 'MACD_SIG', 'MI', 'TRIX', 'VIN', 'VIP',\n 'CR', 'DR']].as_matrix()", "C:\\Users\\Spare\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:5: FutureWarning: Method .as_matrix will be removed in a future version. Use .values instead.\n \"\"\"\n" ], [ "y_cl = df['target_clf'].as_matrix()", "C:\\Users\\Spare\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:1: FutureWarning: Method .as_matrix will be removed in a future version. Use .values instead.\n \"\"\"Entry point for launching an IPython kernel.\n" ], [ "X_train, X_test, y_train, y_test = train_test_split(X, y_cl, test_size=0.20)", "_____no_output_____" ] ], [ [ "## Function for Feeding Data In Batches", "_____no_output_____" ] ], [ [ "def next_batch(num, data, labels):\n '''\n Return a total of `num` random samples and labels. \n '''\n idx = np.arange(0 , len(data))\n np.random.shuffle(idx)\n idx = idx[:num]\n data_shuffle = [data[ i] for i in idx]\n labels_shuffle = [labels[ i] for i in idx]\n\n return np.asarray(data_shuffle), np.asarray(labels_shuffle)", "_____no_output_____" ], [ "X.shape", "_____no_output_____" ] ], [ [ "## Build Bayesian Model", "_____no_output_____" ] ], [ [ "N = 100 # number of rows in a minibatch.\nD = 36 # number of features.\nK = 2 # number of classes.", "_____no_output_____" ], [ "# Create a placeholder to hold the data (in minibatches) in a TensorFlow graph.\nx = tf.placeholder(tf.float32, [None, D])\n# Normal(0,1) priors for the variables. Note that the syntax assumes TensorFlow 1.1.\nw = Normal(loc=tf.zeros([D, K]), scale=tf.ones([D, K]))\nb = Normal(loc=tf.zeros(K), scale=tf.ones(K))\n# Categorical likelihood for classication.\ny = Categorical(tf.matmul(x,w)+b)", "_____no_output_____" ], [ "# Contruct the q(w) and q(b). in this case we assume Normal distributions.\nqw = Normal(loc=tf.Variable(tf.random_normal([D, K])),\n scale=tf.nn.softplus(tf.Variable(tf.random_normal([D, K]))))\nqb = Normal(loc=tf.Variable(tf.random_normal([K])),\n scale=tf.nn.softplus(tf.Variable(tf.random_normal([K]))))", "_____no_output_____" ], [ "# We use a placeholder for the labels in anticipation of the traning data.\ny_ph = tf.placeholder(tf.int32, [N])\n# Define the VI inference technique, ie. minimise the KL divergence between q and p.\ninference = ed.KLqp({w: qw, b: qb}, data={y:y_ph})", "_____no_output_____" ], [ "# Initialse the infernce variables\ninference.initialize(n_iter=5000, n_print=100, scale={y: float(X_train.shape[0]) / N})", "_____no_output_____" ], [ "# We will use an interactive session.\nsess = tf.InteractiveSession()\n# Initialise all the vairables in the session.\ntf.global_variables_initializer().run()", "_____no_output_____" ], [ "# Let the training begin. We load the data in minibatches and update the VI infernce using each new batch.\nfor _ in range(inference.n_iter):\n X_batch, Y_batch = next_batch(N, X_train, y_train)\n # TensorFlow method gives the label data in a one hot vetor format. We convert that into a single label.\n #Y_batch = np.argmax(Y_batch,axis=1)\n info_dict = inference.update(feed_dict={x: X_batch, y_ph: Y_batch})\n inference.print_progress(info_dict)", "5000/5000 [100%] ██████████████████████████████ Elapsed: 13s | Loss: 2932.863\n" ], [ "X_test = X_test.astype(np.float32)", "_____no_output_____" ], [ "X_test.dtype", "_____no_output_____" ], [ "# Generate samples the posterior and store them.\nn_samples = 200\nprob_lst = []\nsamples = []\nw_samples = []\nb_samples = []\nfor _ in range(n_samples):\n w_samp = qw.sample()\n b_samp = qb.sample()\n w_samples.append(w_samp)\n b_samples.append(b_samp)\n # Also compue the probabiliy of each class for each (w,b) sample.\n prob = tf.nn.softmax(tf.matmul( X_test,w_samp ) + b_samp)\n prob_lst.append(prob.eval())\n sample = tf.concat([tf.reshape(w_samp,[-1]),b_samp],0)\n samples.append(sample.eval())", "_____no_output_____" ] ], [ [ "## Compute The Accuracy Distribution For The Bayesian Neural Net", "_____no_output_____" ] ], [ [ "# Compute the accuracy of the model. \n# For each sample we compute the predicted class and compare with the test labels.\n# Predicted class is defined as the one which as maximum proability.\n# We perform this test for each (w,b) in the posterior giving us a set of accuracies\n# Finally we make a histogram of accuracies for the test data.\nfig, axes = plt.subplots(figsize = (15, 8))\naccy_test = []\nfor prob in prob_lst:\n y_trn_prd = np.argmax(prob,axis=1).astype(np.float32)\n acc = (y_trn_prd == y_test).mean()*100\n accy_test.append(acc)\n\naxes.hist(accy_test)\naxes.set_title(\"Histogram of prediction accuracies on the test data\")\naxes.set_xlabel(\"Accuracy\")# Compute the accuracy of the model. \n# For each sample we compute the predicted class and compare with the test labels.\n# Predicted class is defined as the one which as maximum proability.\n# We perform this test for each (w,b) in the posterior giving us a set of accuracies\n# Finally we make a histogram of accuracies for the test data.\nfig, axes = plt.subplots(figsize = (15, 8))\naccy_test = []\nfor prob in prob_lst:\n y_trn_prd = np.argmax(prob,axis=1).astype(np.float32)\n acc = (y_trn_prd == y_test).mean()*100\n accy_test.append(acc)\n\naxes.hist(accy_test)\naxes.set_title(\"Histogram of prediction accuracies on the test data\")\naxes.set_xlabel(\"Accuracy\")\naxes.set_ylabel(\"Frequency\")\n\nfig.savefig('accuracy_plot.png')\naxes.set_ylabel(\"Frequency\")\n\nfig.savefig('accuracy_plot.png')", "_____no_output_____" ], [ "# Here we compute the mean of probabilties for each class for all the (w,b) samples.\n# We then use the class with maximum of the mean proabilities as the prediction. \n# In other words, we have used (w,b) samples to construct a set of models and\n# used their combined outputs to make the predcitions.\nY_pred = np.argmax(np.mean(prob_lst,axis=0),axis=1)\nprint(\"accuracy in predicting the test data = \", (Y_pred == y_test).mean()*100)", "accuracy in predicting the test data = 91.28256513026052\n" ], [ "# Load the first row from the test data and its label.\ntest_row = X_test[2]\ntest_label = y_test[2]\nprint('truth = ',test_label)", "truth = 0.0\n" ], [ "# Now the check what the model perdicts for each (w,b) sample from the posterior. This may take a few seconds...\nsing_img_probs = []\nfor w_samp,b_samp in zip(w_samples,b_samples):\n prob = tf.nn.softmax(tf.matmul( X_test[2:3],w_samp ) + b_samp)\n sing_img_probs.append(prob.eval())", "_____no_output_____" ], [ "# Create a histogram of these predictions.\nfig, axes = plt.subplots(figsize = (15, 8))\naxes.hist(np.argmax(sing_img_probs,axis=2),bins = range(3))\naxes.set_xticks(np.arange(0,2))\naxes.set_xlim(0,2)\naxes.set_xlabel(\"Accuracy of the prediction of the test row\")\naxes.set_ylabel(\"Frequency\")", "_____no_output_____" ], [ "y_test.mean()", "_____no_output_____" ] ], [ [ "# Test The Model On Brazil Data, We Should Get A wider Distribution Of Accuracies", "_____no_output_____" ], [ "## Brazil", "_____no_output_____" ] ], [ [ "brazil = pd.read_csv('brazil_clean.csv')", "_____no_output_____" ], [ "brazil.head()", "_____no_output_____" ], [ "brazil.info()", "<class 'pandas.core.frame.DataFrame'>\nRangeIndex: 7616 entries, 0 to 7615\nData columns (total 39 columns):\nDate 7616 non-null object\nRSI 7616 non-null float64\nTSI 7616 non-null float64\nATR 7616 non-null float64\nBBH 7616 non-null float64\nBHBI 7616 non-null float64\nBBL 7616 non-null float64\nBLBI 7616 non-null float64\nBBMAVG 7616 non-null float64\nDCH 7616 non-null float64\nDCHI 7616 non-null float64\nDCL 7616 non-null float64\nDCLI 7616 non-null float64\nKCC 7616 non-null float64\nKCH 7616 non-null float64\nKCL 7616 non-null float64\nADX 7616 non-null float64\nADXI 7616 non-null int64\nADXN 7616 non-null float64\nADXP 7616 non-null float64\nCCI 7616 non-null float64\nDPO 7616 non-null float64\nSEMA 7616 non-null float64\nLEMA 7616 non-null float64\nIchimoku 7616 non-null float64\nIchimoku_b 7616 non-null float64\nKST 7616 non-null float64\nKST_SIG 7616 non-null float64\nMACD 7616 non-null float64\nMACD_DIFF 7616 non-null float64\nMACD_SIG 7616 non-null float64\nMI 7616 non-null float64\nTRIX 7616 non-null float64\nVIN 7616 non-null float64\nVIP 7616 non-null float64\nCR 7616 non-null float64\nDR 7616 non-null float64\ntarget_return 7616 non-null float64\ntarget_cl 7616 non-null float64\ndtypes: float64(37), int64(1), object(1)\nmemory usage: 2.3+ MB\n" ], [ "brazil['Date'] = pd.to_datetime(brazil['Date'])", "_____no_output_____" ], [ "brazil.head()", "_____no_output_____" ], [ "brazil.info()", "<class 'pandas.core.frame.DataFrame'>\nRangeIndex: 7616 entries, 0 to 7615\nData columns (total 39 columns):\nDate 7616 non-null datetime64[ns]\nRSI 7616 non-null float64\nTSI 7616 non-null float64\nATR 7616 non-null float64\nBBH 7616 non-null float64\nBHBI 7616 non-null float64\nBBL 7616 non-null float64\nBLBI 7616 non-null float64\nBBMAVG 7616 non-null float64\nDCH 7616 non-null float64\nDCHI 7616 non-null float64\nDCL 7616 non-null float64\nDCLI 7616 non-null float64\nKCC 7616 non-null float64\nKCH 7616 non-null float64\nKCL 7616 non-null float64\nADX 7616 non-null float64\nADXI 7616 non-null int64\nADXN 7616 non-null float64\nADXP 7616 non-null float64\nCCI 7616 non-null float64\nDPO 7616 non-null float64\nSEMA 7616 non-null float64\nLEMA 7616 non-null float64\nIchimoku 7616 non-null float64\nIchimoku_b 7616 non-null float64\nKST 7616 non-null float64\nKST_SIG 7616 non-null float64\nMACD 7616 non-null float64\nMACD_DIFF 7616 non-null float64\nMACD_SIG 7616 non-null float64\nMI 7616 non-null float64\nTRIX 7616 non-null float64\nVIN 7616 non-null float64\nVIP 7616 non-null float64\nCR 7616 non-null float64\nDR 7616 non-null float64\ntarget_return 7616 non-null float64\ntarget_cl 7616 non-null float64\ndtypes: datetime64[ns](1), float64(37), int64(1)\nmemory usage: 2.3 MB\n" ], [ "brazil['target_cl'] = brazil['target_cl'].apply(lambda x: x if x==1 else 0)", "_____no_output_____" ], [ "brazil.rename(columns = {'target_return':'target_reg'}, inplace = True)", "_____no_output_____" ], [ "brazil.head()", "_____no_output_____" ], [ "brazil.describe()", "_____no_output_____" ], [ "brazil.set_index('Date', inplace = True)", "_____no_output_____" ], [ "brazil.head()", "_____no_output_____" ], [ "X_br = brazil[['RSI', 'TSI', 'ATR', 'BHBI', 'BBL', 'BBH', 'BLBI', 'BBMAVG', 'DCH',\n 'DCHI', 'DCL', 'DCLI', 'KCC', 'KCH', 'KCL', 'ADX', 'ADXI', 'ADXN',\n 'ADXP', 'CCI', 'DPO', 'SEMA', 'LEMA', 'Ichimoku', 'Ichimoku_b', 'KST',\n 'KST_SIG', 'MACD', 'MACD_DIFF', 'MACD_SIG', 'MI', 'TRIX', 'VIN', 'VIP',\n 'CR', 'DR']].as_matrix()", "C:\\Users\\Spare\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:5: FutureWarning: Method .as_matrix will be removed in a future version. Use .values instead.\n \"\"\"\n" ], [ "y_cl_br = brazil['target_cl'].as_matrix()", "C:\\Users\\Spare\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:1: FutureWarning: Method .as_matrix will be removed in a future version. Use .values instead.\n \"\"\"Entry point for launching an IPython kernel.\n" ], [ "X_br = X_br.astype(np.float32)", "_____no_output_____" ], [ "prob_lst_bra = []\n\nfor w_samp, b_samp in zip(w_samples, b_samples):\n # Also compue the probabiliy of each class for each (w,b) sample.\n prob = tf.nn.softmax(tf.matmul( X_br,w_samp ) + b_samp)\n prob_lst_bra.append(prob.eval())", "_____no_output_____" ], [ "# Compute the accuracy of the model. \n# For each sample we compute the predicted class and compare with the test labels.\n# Predicted class is defined as the one which as maximum proability.\n# We perform this test for each (w,b) in the posterior giving us a set of accuracies\n# Finally we make a histogram of accuracies for the test data.\nfig, axes = plt.subplots(figsize = (15, 8))\naccy_test = []\nfor prob in prob_lst_bra:\n y_trn_prd = np.argmax(prob,axis=1).astype(np.float32)\n acc = (y_trn_prd == y_cl_br).mean()*100\n accy_test.append(acc)\n\naxes.hist(accy_test)\naxes.set_title(\"Histogram of prediction accuracies on the brazil data\")\naxes.set_xlabel(\"Accuracy\")# Compute the accuracy of the model. \naxes.set_ylabel(\"Frequency\")\n\nfig.savefig('accuracy_plot_bra.png')", "_____no_output_____" ], [ "# For each sample we compute the predicted class and compare with the test labels.\n# Predicted class is defined as the one which as maximum proability.\n# We perform this test for each (w,b) in the posterior giving us a set of accuracies\n# Finally we make a histogram of accuracies for the test data.\nfig, axes = plt.subplots(figsize = (15, 8))\naccy_test = []\nfor prob in prob_lst:\n y_trn_prd = np.argmax(prob,axis=1).astype(np.float32)\n acc = (y_trn_prd == y_test).mean()*100\n accy_test.append(acc)\n\naxes.hist(accy_test)\naxes.set_title(\"Histogram of prediction accuracies on the test data\")\naxes.set_xlabel(\"Accuracy\")\naxes.set_ylabel(\"Frequency\")\n", "_____no_output_____" ] ], [ [ "## Conclusion:\n\n<p>We got good results considering the fact that this model</p>", "_____no_output_____" ] ] ]

[ "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown" ]

[ [ "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code" ], [ "markdown", "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ], [ "markdown" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "# Train your first model\n\nThis is the second of our [beginner tutorial series](https://github.com/deepjavalibrary/djl/tree/master/jupyter/tutorial) that will take you through creating, training, and running inference on a neural network. In this tutorial, you will learn how to train an image classification model that can recognize handwritten digits.\n\n## Preparation\n\nThis tutorial requires the installation of the Java Jupyter Kernel. To install the kernel, see the [Jupyter README](https://github.com/deepjavalibrary/djl/blob/master/jupyter/README.md).", "_____no_output_____" ] ], [ [ "// Add the snapshot repository to get the DJL snapshot artifacts\n// %mavenRepo snapshots https://oss.sonatype.org/content/repositories/snapshots/\n\n// Add the maven dependencies\n%maven ai.djl:api:0.17.0\n%maven ai.djl:basicdataset:0.17.0\n%maven ai.djl:model-zoo:0.17.0\n%maven ai.djl.mxnet:mxnet-engine:0.17.0\n%maven org.slf4j:slf4j-simple:1.7.32", "_____no_output_____" ], [ "import java.nio.file.*;\n\nimport ai.djl.*;\nimport ai.djl.basicdataset.cv.classification.Mnist;\nimport ai.djl.ndarray.types.*;\nimport ai.djl.training.*;\nimport ai.djl.training.dataset.*;\nimport ai.djl.training.initializer.*;\nimport ai.djl.training.loss.*;\nimport ai.djl.training.listener.*;\nimport ai.djl.training.evaluator.*;\nimport ai.djl.training.optimizer.*;\nimport ai.djl.training.util.*;\nimport ai.djl.basicmodelzoo.cv.classification.*;\nimport ai.djl.basicmodelzoo.basic.*;", "_____no_output_____" ] ], [ [ "# Step 1: Prepare MNIST dataset for training\n\nIn order to train, you must create a [Dataset class](https://javadoc.io/static/ai.djl/api/0.17.0/index.html?ai/djl/training/dataset/Dataset.html) to contain your training data. A dataset is a collection of sample input/output pairs for the function represented by your neural network. Each single input/output is represented by a [Record](https://javadoc.io/static/ai.djl/api/0.17.0/index.html?ai/djl/training/dataset/Record.html). Each record could have multiple arrays of inputs or outputs such as an image question and answer dataset where the input is both an image and a question about the image while the output is the answer to the question.\n\nBecause data learning is highly parallelizable, training is often done not with a single record at a time, but a [Batch](https://javadoc.io/static/ai.djl/api/0.17.0/index.html?ai/djl/training/dataset/Batch.html). This can lead to significant performance gains, especially when working with images\n\n## Sampler\n\nThen, we must decide the parameters for loading data from the dataset. The only parameter we need for MNIST is the choice of [Sampler](https://javadoc.io/static/ai.djl/api/0.17.0/index.html?ai/djl/training/dataset/Sampler.html). The sampler decides which and how many element from datasets are part of each batch when iterating through it. We will have it randomly shuffle the elements for the batch and use a batchSize of 32. The batchSize is usually the largest power of 2 that fits within memory.", "_____no_output_____" ] ], [ [ "int batchSize = 32;\nMnist mnist = Mnist.builder().setSampling(batchSize, true).build();\nmnist.prepare(new ProgressBar());", "_____no_output_____" ] ], [ [ "# Step 2: Create your Model\n\nNext we will build a model. A [Model](https://javadoc.io/static/ai.djl/api/0.17.0/index.html?ai/djl/Model.html) contains a neural network [Block](https://javadoc.io/static/ai.djl/api/0.17.0/index.html?ai/djl/nn/Block.html) along with additional artifacts used for the training process. It possesses additional information about the inputs, outputs, shapes, and data types you will use. Generally, you will use the Model once you have fully completed your Block.\n\nIn this part of the tutorial, we will use the built-in Multilayer Perceptron Block from the Model Zoo. To learn how to build it from scratch, see the previous tutorial: [Create Your First Network](01_create_your_first_network.ipynb).\n\nBecause images in the MNIST dataset are 28x28 grayscale images, we will create an MLP block with 28 x 28 input. The output will be 10 because there are 10 possible classes (0 to 9) each image could be. For the hidden layers, we have chosen `new int[] {128, 64}` by experimenting with different values.", "_____no_output_____" ] ], [ [ "Model model = Model.newInstance(\"mlp\");\nmodel.setBlock(new Mlp(28 * 28, 10, new int[] {128, 64}));", "_____no_output_____" ] ], [ [ "# Step 3: Create a Trainer\n\nNow, you can create a [`Trainer`](https://javadoc.io/static/ai.djl/api/0.17.0/index.html?ai/djl/training/Trainer.html) to train your model. The trainer is the main class to orchestrate the training process. Usually, they will be opened using a try-with-resources and closed after training is over.\n\nThe trainer takes an existing model and attempts to optimize the parameters inside the model's Block to best match the dataset. Most optimization is based upon [Stochastic Gradient Descent](https://en.wikipedia.org/wiki/Stochastic_gradient_descent) (SGD).\n\n## Step 3.1: Setup your training configurations\n\nBefore you create your trainer, we we will need a [training configuration](https://javadoc.io/static/ai.djl/api/0.17.0/index.html?ai/djl/training/DefaultTrainingConfig.html) that describes how to train your model.\n\nThe following are a few common items you may need to configure your training:\n\n* **REQUIRED** [`Loss`](https://javadoc.io/static/ai.djl/api/0.17.0/index.html?ai/djl/training/loss/Loss.html) function: A loss function is used to measure how well our model matches the dataset. Because the lower value of the function is better, it's called the \"loss\" function. The Loss is the only required argument to the model\n* [`Evaluator`](https://javadoc.io/static/ai.djl/api/0.17.0/index.html?ai/djl/training/evaluator/Evaluator.html) function: An evaluator function is also used to measure how well our model matches the dataset. Unlike the loss, they are only there for people to look at and are not used for optimizing the model. Since many losses are not as intuitive, adding other evaluators such as Accuracy can help to understand how your model is doing. If you know of any useful evaluators, we recommend adding them.\n* [`Training Listeners`](https://javadoc.io/static/ai.djl/api/0.17.0/index.html?ai/djl/repository/zoo/ZooModel.html): The training listener adds additional functionality to the training process through a listener interface. This can include showing training progress, stopping early if training becomes undefined, or recording performance metrics. We offer several easy sets of [default listeners](https://javadoc.io/static/ai.djl/api/0.17.0/index.html?ai/djl/repository/zoo/ZooModel.html).\n\nYou can also configure other options such as the Device, Initializer, and Optimizer. See [more details](https://javadoc.io/static/ai.djl/api/0.17.0/index.html?ai/djl/training/TrainingConfig.html).", "_____no_output_____" ] ], [ [ "DefaultTrainingConfig config = new DefaultTrainingConfig(Loss.softmaxCrossEntropyLoss())\n //softmaxCrossEntropyLoss is a standard loss for classification problems\n .addEvaluator(new Accuracy()) // Use accuracy so we humans can understand how accurate the model is\n .addTrainingListeners(TrainingListener.Defaults.logging());\n\n// Now that we have our training configuration, we should create a new trainer for our model\nTrainer trainer = model.newTrainer(config);", "_____no_output_____" ] ], [ [ "# Step 5: Initialize Training\n\nBefore training your model, you have to initialize all of the parameters with starting values. You can use the trainer for this initialization by passing in the input shape.\n\n* The first axis of the input shape is the batch size. This won't impact the parameter initialization, so you can use 1 here.\n* The second axis of the input shape of the MLP - the number of pixels in the input image.", "_____no_output_____" ] ], [ [ "trainer.initialize(new Shape(1, 28 * 28));", "_____no_output_____" ] ], [ [ "# Step 6: Train your model\n\nNow, we can train the model.\n\nWhen training, it is usually organized into epochs where each epoch trains the model on each item in the dataset once. It is slightly faster than training randomly.\n\nThen, we will use the EasyTrain to, as the name promises, make the training easy. If you want to see more details about how the training loop works, see [the EasyTrain class](https://github.com/deepjavalibrary/djl/blob/0.9/api/src/main/java/ai/djl/training/EasyTrain.java) or [read our Dive into Deep Learning book](https://d2l.djl.ai).", "_____no_output_____" ] ], [ [ "// Deep learning is typically trained in epochs where each epoch trains the model on each item in the dataset once.\nint epoch = 2;\n\nEasyTrain.fit(trainer, epoch, mnist, null);", "_____no_output_____" ] ], [ [ "# Step 7: Save your model\n\nOnce your model is trained, you should save it so that it can be reloaded later. You can also add metadata to it such as training accuracy, number of epochs trained, etc that can be used when loading the model or when examining it.", "_____no_output_____" ] ], [ [ "Path modelDir = Paths.get(\"build/mlp\");\nFiles.createDirectories(modelDir);\n\nmodel.setProperty(\"Epoch\", String.valueOf(epoch));\n\nmodel.save(modelDir, \"mlp\");\n\nmodel", "_____no_output_____" ] ], [ [ "# Summary\n\nNow, you've successfully trained a model that can recognize handwritten digits. You'll learn how to apply this model in the next chapter: [Run image classification with your model](03_image_classification_with_your_model.ipynb).\n\nYou can find the complete source code for this tutorial in the [examples project](https://github.com/deepjavalibrary/djl/blob/master/examples/src/main/java/ai/djl/examples/training/TrainMnist.java).", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown" ]

[ [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import os\nimport cv2\nimport numpy as np\nimport mrcnn.model as modellib\nimport matplotlib.pyplot as plt\nfrom PIL import Image\nfrom skimage.io import imread\nfrom skimage.color import gray2rgb\nfrom mrcnn import utils\nfrom sklearn.ensemble import RandomForestClassifier\nfrom samples.coco import coco\n%matplotlib inline ", "Using TensorFlow backend.\n" ], [ "# where model and COCO weights are stored\nMODEL_PATH = \"./logs\"\nMODEL_WEIGHTS_PATH = \"./mask_rcnn_coco.h5\"\n# where data set is\ndata_root = \"../CNNs/CUB_200_2011/CUB_200_2011/\"\n# root dir to save images after background removal\ndest = \"./res\"\n\nif not os.path.exists(MODEL_WEIGHTS_PATH):\n utils.download_trained_weights(MODEL_WEIGHTS_PATH)\n\n \nclass InferenceConfig(coco.CocoConfig):\n GPU_COUNT = 1\n IMAGES_PER_GPU = 1 ", "_____no_output_____" ], [ "config = InferenceConfig()\n# Create MaskCNN model in inference mode and load weights\nmodel = modellib.MaskRCNN(mode=\"inference\", model_dir=MODEL_PATH, config=config)\nmodel.load_weights(MODEL_WEIGHTS_PATH, by_name=True)", "_____no_output_____" ], [ "# remove background from image\nmode = [\"random_forest_train\", \"random_forest_test\"]\n\n# class id for bird in COCO\nbird_id = 15\n\nfor m in mode:\n if not os.path.exists(os.path.join(dest, m)):\n os.makedirs(os.path.join(dest, m))\n\nfor m in mode:\n src_dir = os.path.join(data_root, m)\n for img in os.listdir(src_dir):\n file_name = img.split('.')[0]\n # read in one image\n image_input = imread(os.path.join(src_dir, img))\n # convert to RGB\n if image_input.ndim == 2:\n image_input = gray2rgb(image_input)\n results = model.detect([image_input])\n\n\n # unpack inference results \n result = results[0]\n masks = result['masks'].astype(np.uint8)\n scores = result['scores']\n rois = result['rois']\n \n # get result index for all regions identified as bird\n idxes = [i for i in range(len(result['class_ids'])) if result['class_ids'][i] == bird_id]\n \n # keep the one with highest confidence\n if len(idxes) > 0:\n idx = idxes[0]\n max_score = scores[idx]\n\n for index in idxes:\n if scores[index] > max_score:\n max_score = scores[index]\n idx = index\n \n # bounding box for the max\n (y1, x1, y2, x2) = rois[idx] \n width = x2 - x1\n height = y2 - y1\n \n # get mask for the region containing a bird\n bitmap = masks[:,:,idx] \n bitmap[bitmap > 0] = 255 \n bitmap = np.tile(bitmap[:, :, None], [1, 1, 3])\n \n # save result\n path_output_image = f'{dest}/{m}/{img}'\n image_output = Image.fromarray(np.bitwise_and(image_input, bitmap))\n image_output.save(path_output_image)", "_____no_output_____" ], [ "# extraact RGB or HSV color feature from image\ndef extract_hist_feature(img, rgb=True, verbose=False):\n src = cv2.cvtColor(img, cv2.COLOR_BGR2HSV) if not rgb else img\n chans = cv2.split(src)\n colors = (\"b\", \"g\", \"r\")\n name = ['Blue', 'Green', 'Red'] if rgb else [\"Hue\", \"Saturation\", \"Value\"]\n\n features = []\n \n \n # loop over the image channels\n i = 1\n for (chan, color) in zip(chans, colors):\n # create a histogram for the current channel and\n # concatenate the resulting histograms for each\n # channel\n hist = cv2.calcHist([chan], [0], None, [10], [10, 210])\n hist = hist.reshape(-1)\n hist += 1e-5\n features.extend(hist / hist.sum())\n if verbose:\n plt.figure(figsize=(10, 3)) \n plt.subplot(1, 3, i)\n plt.title(f'{name[i - 1]} Histogram')\n plt.xlabel(\"Value\")\n plt.ylabel(\"# of Pixels\")\n plt.xticks([10 + x * 22 for x in range(10)])\n plt.bar([21 + x * 22 for x in range(10)], hist, width=22, color = color)\n i += 1\n plt.tight_layout()\n return features", "_____no_output_____" ], [ "# mapping from image_id to image class\nimg_id2class = {}\nclass_file = os.path.join(data_root, 'image_class_labels.txt')\nwith open(class_file, 'r', encoding='utf-8') as f:\n for line in f:\n img_id, class_id = map(int, line.rstrip().split(' '))\n img_id2class[img_id] = class_id", "_____no_output_____" ] ], [ [ "# HSV feature", "_____no_output_____" ] ], [ [ "train_x = []\ntrain_y = []\ntrain_dir = os.path.join(dest, \"random_forest_train\")\nfor img in os.listdir(train_dir):\n img_id = int(img.split(\".\")[0])\n train_y.append(int(img_id2class[img_id]))\n train_x.append(extract_hist_feature(cv2.imread(os.path.join(train_dir, img)), rgb=False))", "_____no_output_____" ], [ "classifier = RandomForestClassifier()\nclassifier.fit(train_x, train_y)", "_____no_output_____" ], [ "test_x = []\ntest_y = []\ntest_dir = os.path.join(dest, \"random_forest_test\")\nfor img in os.listdir(test_dir):\n img_id = int(img.split(\".\")[0])\n test_y.append(int(img_id2class[img_id]))\n test_x.append(extract_hist_feature(cv2.imread(os.path.join(test_dir, img)), rgb=False))", "_____no_output_____" ], [ "classifier.score(test_x, test_y)", "_____no_output_____" ] ], [ [ "# RGB feature", "_____no_output_____" ] ], [ [ "train_x = []\ntrain_y = []\ntrain_dir = os.path.join(dest, \"random_forest_train\")\nfor img in os.listdir(train_dir):\n img_id = int(img.split(\".\")[0])\n train_y.append(int(img_id2class[img_id]))\n train_x.append(extract_hist_feature(cv2.imread(os.path.join(train_dir, img)), rgb=True))", "_____no_output_____" ], [ "classifier = RandomForestClassifier()\nclassifier.fit(train_x, train_y)", "_____no_output_____" ], [ "test_x = []\ntest_y = []\ntest_dir = os.path.join(dest, \"random_forest_test\")\nfor img in os.listdir(test_dir):\n img_id = int(img.split(\".\")[0])\n test_y.append(int(img_id2class[img_id]))\n test_x.append(extract_hist_feature(cv2.imread(os.path.join(test_dir, img)), rgb=True))", "_____no_output_____" ], [ "classifier.score(test_x, test_y)", "_____no_output_____" ] ] ]

[ "code", "markdown", "code", "markdown", "code" ]

[ [ "code", "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import pandas as pd\npd.set_option(\"display.max_columns\", None)\nimport numpy as np\nimport matplotlib.pyplot as plt", "_____no_output_____" ], [ "df_nov_dec = pd.read_csv(\"data/flights_2018_nov_dec_raw.csv\")\ndf_jan = pd.read_csv(\"data/flights_2018_jan_raw.csv\")", "/Users/louisrossi/opt/anaconda3/envs/ml/lib/python3.8/site-packages/IPython/core/interactiveshell.py:3441: DtypeWarning: Columns (25) have mixed types.Specify dtype option on import or set low_memory=False.\n exec(code_obj, self.user_global_ns, self.user_ns)\n" ], [ "df = pd.concat([df_nov_dec, df_jan]).reset_index().drop(columns=[\"index\"])", "_____no_output_____" ], [ "df_ = df.sample(frac = 0.05)", "_____no_output_____" ], [ "df.dtypes", "_____no_output_____" ], [ "def missing(x):\n n_missing = x.isnull().sum().sort_values(ascending=False)\n p_missing = (x.isnull().sum()/x.isnull().count()).sort_values(ascending=False)\n missing_ = pd.concat([n_missing, p_missing],axis=1, keys = ['number','percent'])\n return missing_", "_____no_output_____" ], [ "missing(df_)", "_____no_output_____" ], [ "from scipy import stats\nimport seaborn as sns\narr_delay = df_.arr_delay", "_____no_output_____" ], [ "stats.kstest(arr_delay,stats.norm.cdf)", "_____no_output_____" ], [ "stats.shapiro(arr_delay)", "/Users/louisrossi/opt/anaconda3/envs/ml/lib/python3.8/site-packages/scipy/stats/morestats.py:1760: UserWarning: p-value may not be accurate for N > 5000.\n warnings.warn(\"p-value may not be accurate for N > 5000.\")\n" ], [ "sample = df.sample(frac=0.05)\nstats.shapiro(sample['arr_delay'])\n\n#fail to reject the null hypothesis that data is normally dist", "_____no_output_____" ], [ "sns.histplot(arr_delay)\nplt.xlim(-300, 300)", "_____no_output_____" ], [ "import datetime as dt\nfrom datetime import date\nfrom datetime import time", "_____no_output_____" ], [ "df['fl_date'] = pd.to_datetime(df['fl_date'])", "_____no_output_____" ], [ "type(df.fl_date[0])", "_____no_output_____" ], [ "df['month'] = df['fl_date'].dt.month\ndf['month'].head()", "_____no_output_____" ], [ "monthly_count = df.groupby(['month'])['arr_delay'].count()", "_____no_output_____" ], [ "monthly_count = pd.DataFrame(monthly_count)\nmonthly_count", "_____no_output_____" ], [ "sns.barplot(x= monthly_count.index,y=monthly_count['arr_delay'])", "_____no_output_____" ], [ "monthly_avg = df.groupby(['month'])['arr_delay'].mean()\nsns.barplot(x=df['month'],y=df['arr_delay'])", "_____no_output_____" ], [ "task4a = df.groupby(['dep_time'])['taxi_out'].count()\nsns.histplot(task4a)", "_____no_output_____" ], [ "task4a_count = pd.DataFrame(task4a)\nsns.barplot(x=task4a_count.index,y=task4a_count['taxi_out'])", "_____no_output_____" ], [ "sns.barplot(x=df['dep_time'],y=df['taxi_out'], ci=None)", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "BSD-3-Clause" ]

[ "BSD-3-Clause" ]

[ "BSD-3-Clause" ]

[ [ [ "# How to cite us\n\nIf you use JOOMMF in your research, apart from acknowledging OOMMF (http://math.nist.gov/oommf/oommf_cites.html) please acknowledge our interface JOOMMF by citing the following paper:\n\nBeg, M., Pepper, R. A., & Fangohr, H. (2017). User interfaces for computational science: A domain specific language for OOMMF embedded in Python. *AIP Advances* **7**, 56025. https://doi.org/10.1063/1.4977225", "_____no_output_____" ] ] ]

[ "markdown" ]

[ [ "markdown" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Transfer Learning\n\nMost of the time you won't want to train a whole convolutional network yourself. Modern ConvNets training on huge datasets like ImageNet take weeks on multiple GPUs. Instead, most people use a pretrained network either as a fixed feature extractor, or as an initial network to fine tune. In this notebook, you'll be using [VGGNet](https://arxiv.org/pdf/1409.1556.pdf) trained on the [ImageNet dataset](http://www.image-net.org/) as a feature extractor. Below is a diagram of the VGGNet architecture.\n\n<img src=\"assets/cnnarchitecture.jpg\" width=700px>\n\nVGGNet is great because it's simple and has great performance, coming in second in the ImageNet competition. The idea here is that we keep all the convolutional layers, but replace the final fully connected layers with our own classifier. This way we can use VGGNet as a feature extractor for our images then easily train a simple classifier on top of that. What we'll do is take the first fully connected layer with 4096 units, including thresholding with ReLUs. We can use those values as a code for each image, then build a classifier on top of those codes.\n\nYou can read more about transfer learning from [the CS231n course notes](http://cs231n.github.io/transfer-learning/#tf).\n\n## Pretrained VGGNet\n\nWe'll be using a pretrained network from https://github.com/machrisaa/tensorflow-vgg. \n\nThis is a really nice implementation of VGGNet, quite easy to work with. The network has already been trained and the parameters are available from this link. ", "_____no_output_____" ] ], [ [ "from tensorflow.python.client import device_lib\nprint(device_lib.list_local_devices())\n\nprint(device_lib)", "[name: \"/device:CPU:0\"\ndevice_type: \"CPU\"\nmemory_limit: 268435456\nlocality {\n}\nincarnation: 16373191140287572975\n, name: \"/device:GPU:0\"\ndevice_type: \"GPU\"\nmemory_limit: 6740156088\nlocality {\n bus_id: 1\n}\nincarnation: 5624907406584000031\nphysical_device_desc: \"device: 0, name: GeForce GTX 1080, pci bus id: 0000:01:00.0, compute capability: 6.1\"\n]\n<module 'tensorflow.python.client.device_lib' from 'C:\\\\Program Files\\\\Anaconda3\\\\envs\\\\trs\\\\lib\\\\site-packages\\\\tensorflow\\\\python\\\\client\\\\device_lib.py'>\n" ], [ "from urllib.request import urlretrieve\nfrom os.path import isfile, isdir\nfrom tqdm import tqdm\n\nvgg_dir = 'tensorflow_vgg/'\n# Make sure vgg exists\nif not isdir(vgg_dir):\n raise Exception(\"VGG directory doesn't exist!\")\n\nclass DLProgress(tqdm):\n last_block = 0\n\n def hook(self, block_num=1, block_size=1, total_size=None):\n self.total = total_size\n self.update((block_num - self.last_block) * block_size)\n self.last_block = block_num\n\nif not isfile(vgg_dir + \"vgg16.npy\"):\n with DLProgress(unit='B', unit_scale=True, miniters=1, desc='VGG16 Parameters') as pbar:\n urlretrieve(\n 'https://s3.amazonaws.com/content.udacity-data.com/nd101/vgg16.npy',\n vgg_dir + 'vgg16.npy',\n pbar.hook)\nelse:\n print(\"Parameter file already exists!\")", "Parameter file already exists!\n" ] ], [ [ "## Flower power\n\nHere we'll be using VGGNet to classify images of flowers. To get the flower dataset, run the cell below. This dataset comes from the [TensorFlow inception tutorial](https://www.tensorflow.org/tutorials/image_retraining).", "_____no_output_____" ] ], [ [ "import tarfile\n\ndataset_folder_path = 'flower_photos'\n\nclass DLProgress(tqdm):\n last_block = 0\n\n def hook(self, block_num=1, block_size=1, total_size=None):\n self.total = total_size\n self.update((block_num - self.last_block) * block_size)\n self.last_block = block_num\n\nif not isfile('flower_photos.tar.gz'):\n with DLProgress(unit='B', unit_scale=True, miniters=1, desc='Flowers Dataset') as pbar:\n urlretrieve(\n 'http://download.tensorflow.org/example_images/flower_photos.tgz',\n 'flower_photos.tar.gz',\n pbar.hook)\n\nif not isdir(dataset_folder_path):\n with tarfile.open('flower_photos.tar.gz') as tar:\n tar.extractall()\n tar.close()", "_____no_output_____" ] ], [ [ "## ConvNet Codes\n\nBelow, we'll run through all the images in our dataset and get codes for each of them. That is, we'll run the images through the VGGNet convolutional layers and record the values of the first fully connected layer. We can then write these to a file for later when we build our own classifier.\n\nHere we're using the `vgg16` module from `tensorflow_vgg`. The network takes images of size $244 \\times 224 \\times 3$ as input. Then it has 5 sets of convolutional layers. The network implemented here has this structure (copied from [the source code](https://github.com/machrisaa/tensorflow-vgg/blob/master/vgg16.py):\n\n```\nself.conv1_1 = self.conv_layer(bgr, \"conv1_1\")\nself.conv1_2 = self.conv_layer(self.conv1_1, \"conv1_2\")\nself.pool1 = self.max_pool(self.conv1_2, 'pool1')\n\nself.conv2_1 = self.conv_layer(self.pool1, \"conv2_1\")\nself.conv2_2 = self.conv_layer(self.conv2_1, \"conv2_2\")\nself.pool2 = self.max_pool(self.conv2_2, 'pool2')\n\nself.conv3_1 = self.conv_layer(self.pool2, \"conv3_1\")\nself.conv3_2 = self.conv_layer(self.conv3_1, \"conv3_2\")\nself.conv3_3 = self.conv_layer(self.conv3_2, \"conv3_3\")\nself.pool3 = self.max_pool(self.conv3_3, 'pool3')\n\nself.conv4_1 = self.conv_layer(self.pool3, \"conv4_1\")\nself.conv4_2 = self.conv_layer(self.conv4_1, \"conv4_2\")\nself.conv4_3 = self.conv_layer(self.conv4_2, \"conv4_3\")\nself.pool4 = self.max_pool(self.conv4_3, 'pool4')\n\nself.conv5_1 = self.conv_layer(self.pool4, \"conv5_1\")\nself.conv5_2 = self.conv_layer(self.conv5_1, \"conv5_2\")\nself.conv5_3 = self.conv_layer(self.conv5_2, \"conv5_3\")\nself.pool5 = self.max_pool(self.conv5_3, 'pool5')\n\nself.fc6 = self.fc_layer(self.pool5, \"fc6\")\nself.relu6 = tf.nn.relu(self.fc6)\n```\n\nSo what we want are the values of the first fully connected layer, after being ReLUd (`self.relu6`). To build the network, we use\n\n```\nwith tf.Session() as sess:\n vgg = vgg16.Vgg16()\n input_ = tf.placeholder(tf.float32, [None, 224, 224, 3])\n with tf.name_scope(\"content_vgg\"):\n vgg.build(input_)\n```\n\nThis creates the `vgg` object, then builds the graph with `vgg.build(input_)`. Then to get the values from the layer,\n\n```\nfeed_dict = {input_: images}\ncodes = sess.run(vgg.relu6, feed_dict=feed_dict)\n```", "_____no_output_____" ] ], [ [ "import os\n\nimport numpy as np\nimport tensorflow as tf\n\nfrom tensorflow_vgg import vgg16\nfrom tensorflow_vgg import utils", "_____no_output_____" ], [ "data_dir = 'flower_photos/'\ncontents = os.listdir(data_dir)\nclasses = [each for each in contents if os.path.isdir(data_dir + each)]", "_____no_output_____" ] ], [ [ "Below I'm running images through the VGG network in batches.", "_____no_output_____" ] ], [ [ "# Set the batch size higher if you can fit in in your GPU memory\nbatch_size = 10\ncodes_list = []\nlabels = []\nbatch = []\n\ncodes = None\n\nwith tf.Session() as sess:\n vgg = vgg16.Vgg16()\n input_ = tf.placeholder(tf.float32, [None, 224, 224, 3])\n with tf.name_scope(\"content_vgg\"):\n vgg.build(input_)\n\n for each in classes:\n print(\"Starting {} images\".format(each))\n class_path = data_dir + each\n files = os.listdir(class_path)\n for ii, file in enumerate(files, 1):\n # Add images to the current batch\n # utils.load_image crops the input images for us, from the center\n img = utils.load_image(os.path.join(class_path, file))\n print(img.shape)\n batch.append(img.reshape((1, 224, 224, 3)))\n labels.append(each)\n \n # Running the batch through the network to get the codes\n if ii % batch_size == 0 or ii == len(files):\n images = np.concatenate(batch)\n \n feed_dict = {input_: images}\n print(images.shape)\n codes_batch = sess.run(vgg.relu6, feed_dict=feed_dict)\n \n # Here I'm building an array of the codes\n if codes is None:\n codes = codes_batch\n else:\n codes = np.concatenate((codes, codes_batch))\n print(codes.shape)\n # Reset to start building the next batch\n batch = []\n print('{} images processed'.format(ii))", "C:\\Users\\freed\\havardclass\\Udacity Nano\\deep-learning\\transfer-learning\\tensorflow_vgg\\vgg16.npy\nnpy file loaded\nbuild model started\nbuild model finished: 0s\nStarting daisy images\n(224, 224, 3)\n(224, 224, 3)\n(224, 224, 3)\n(224, 224, 3)\n(224, 224, 3)\n(224, 224, 3)\n(224, 224, 3)\n(224, 224, 3)\n(224, 224, 3)\n(224, 224, 3)\n(10, 224, 224, 3)\n" ], [ "codes.shape", "_____no_output_____" ], [ "# write codes to file\nwith open('codes', 'w') as f:\n codes.tofile(f)\n \n# write labels to file\nimport csv\nwith open('labels', 'w') as f:\n writer = csv.writer(f, delimiter='\\n')\n writer.writerow(labels)", "_____no_output_____" ] ], [ [ "## Building the Classifier\n\nNow that we have codes for all the images, we can build a simple classifier on top of them. The codes behave just like normal input into a simple neural network. Below I'm going to have you do most of the work.", "_____no_output_____" ] ], [ [ "# read codes and labels from file\nimport csv\n\nwith open('labels') as f:\n reader = csv.reader(f, delimiter='\\n')\n labels = np.array([each for each in reader if len(each) > 0]).squeeze()\nwith open('codes') as f:\n codes = np.fromfile(f, dtype=np.float32)\n codes = codes.reshape((len(labels), -1))", "_____no_output_____" ] ], [ [ "### Data prep\n\nAs usual, now we need to one-hot encode our labels and create validation/test sets. First up, creating our labels!\n\n> **Exercise:** From scikit-learn, use [LabelBinarizer](http://scikit-learn.org/stable/modules/generated/sklearn.preprocessing.LabelBinarizer.html) to create one-hot encoded vectors from the labels. ", "_____no_output_____" ] ], [ [ "codes", "_____no_output_____" ], [ "from sklearn.preprocessing import LabelBinarizer\n\nlb = LabelBinarizer()\nlb.fit(labels)\n\nlabels_vecs = lb.transform(labels)", "_____no_output_____" ], [ "labels_vecs", "_____no_output_____" ] ], [ [ "Now you'll want to create your training, validation, and test sets. An important thing to note here is that our labels and data aren't randomized yet. We'll want to shuffle our data so the validation and test sets contain data from all classes. Otherwise, you could end up with testing sets that are all one class. Typically, you'll also want to make sure that each smaller set has the same the distribution of classes as it is for the whole data set. The easiest way to accomplish both these goals is to use [`StratifiedShuffleSplit`](http://scikit-learn.org/stable/modules/generated/sklearn.model_selection.StratifiedShuffleSplit.html) from scikit-learn.\n\nYou can create the splitter like so:\n```\nss = StratifiedShuffleSplit(n_splits=1, test_size=0.2)\n```\nThen split the data with \n```\nsplitter = ss.split(x, y)\n```\n\n`ss.split` returns a generator of indices. You can pass the indices into the arrays to get the split sets. The fact that it's a generator means you either need to iterate over it, or use `next(splitter)` to get the indices. Be sure to read the [documentation](http://scikit-learn.org/stable/modules/generated/sklearn.model_selection.StratifiedShuffleSplit.html) and the [user guide](http://scikit-learn.org/stable/modules/cross_validation.html#random-permutations-cross-validation-a-k-a-shuffle-split).\n\n> **Exercise:** Use StratifiedShuffleSplit to split the codes and labels into training, validation, and test sets.", "_____no_output_____" ] ], [ [ "from sklearn.model_selection import StratifiedShuffleSplit\n\nss = StratifiedShuffleSplit(n_splits=1, test_size=0.2)\n\ntrain_idx, val_idx = next(ss.split(codes, labels_vecs))\n\nhalf_val_len = int(len(val_idx)/2)\nval_idx, test_idx = val_idx[:half_val_len], val_idx[half_val_len:]\n\ntrain_x, train_y = codes[train_idx], labels_vecs[train_idx]\nval_x, val_y = codes[val_idx], labels_vecs[val_idx]\ntest_x, test_y = codes[test_idx], labels_vecs[test_idx]", "_____no_output_____" ], [ "print(\"Train shapes (x, y):\", train_x.shape, train_y.shape)\nprint(\"Validation shapes (x, y):\", val_x.shape, val_y.shape)\nprint(\"Test shapes (x, y):\", test_x.shape, test_y.shape)", "Train shapes (x, y): (2936, 4096) (2936, 5)\nValidation shapes (x, y): (367, 4096) (367, 5)\nTest shapes (x, y): (367, 4096) (367, 5)\n" ] ], [ [ "If you did it right, you should see these sizes for the training sets:\n\n```\nTrain shapes (x, y): (2936, 4096) (2936, 5)\nValidation shapes (x, y): (367, 4096) (367, 5)\nTest shapes (x, y): (367, 4096) (367, 5)\n```", "_____no_output_____" ], [ "### Classifier layers\n\nOnce you have the convolutional codes, you just need to build a classfier from some fully connected layers. You use the codes as the inputs and the image labels as targets. Otherwise the classifier is a typical neural network.\n\n> **Exercise:** With the codes and labels loaded, build the classifier. Consider the codes as your inputs, each of them are 4096D vectors. You'll want to use a hidden layer and an output layer as your classifier. Remember that the output layer needs to have one unit for each class and a softmax activation function. Use the cross entropy to calculate the cost.", "_____no_output_____" ] ], [ [ "inputs_ = tf.placeholder(tf.float32, shape=[None, codes.shape[1]])\nlabels_ = tf.placeholder(tf.int64, shape=[None, labels_vecs.shape[1]])\n\nfc = tf.contrib.layers.fully_connected(inputs_, 256)\n \nlogits = tf.contrib.layers.fully_connected(fc, labels_vecs.shape[1], activation_fn=None)\ncross_entropy = tf.nn.softmax_cross_entropy_with_logits(labels=labels_, logits=logits)\ncost = tf.reduce_mean(cross_entropy)\n\noptimizer = tf.train.AdamOptimizer().minimize(cost)\n\npredicted = tf.nn.softmax(logits)\ncorrect_pred = tf.equal(tf.argmax(predicted, 1), tf.argmax(labels_, 1))\naccuracy = tf.reduce_mean(tf.cast(correct_pred, tf.float32))", "_____no_output_____" ] ], [ [ "### Batches!\n\nHere is just a simple way to do batches. I've written it so that it includes all the data. Sometimes you'll throw out some data at the end to make sure you have full batches. Here I just extend the last batch to include the remaining data.", "_____no_output_____" ] ], [ [ "def get_batches(x, y, n_batches=10):\n \"\"\" Return a generator that yields batches from arrays x and y. \"\"\"\n batch_size = len(x)//n_batches\n \n for ii in range(0, n_batches*batch_size, batch_size):\n # If we're not on the last batch, grab data with size batch_size\n if ii != (n_batches-1)*batch_size:\n X, Y = x[ii: ii+batch_size], y[ii: ii+batch_size] \n # On the last batch, grab the rest of the data\n else:\n X, Y = x[ii:], y[ii:]\n # I love generators\n yield X, Y", "_____no_output_____" ] ], [ [ "### Training\n\nHere, we'll train the network.\n\n> **Exercise:** So far we've been providing the training code for you. Here, I'm going to give you a bit more of a challenge and have you write the code to train the network. Of course, you'll be able to see my solution if you need help.", "_____no_output_____" ] ], [ [ "epochs = 10\niteration = 0\nsaver = tf.train.Saver()\nwith tf.Session() as sess:\n \n sess.run(tf.global_variables_initializer())\n for e in range(epochs):\n for x, y in get_batches(train_x, train_y):\n feed = {inputs_: x,\n labels_: y}\n loss, _ = sess.run([cost, optimizer], feed_dict=feed)\n print(\"Epoch: {}/{}\".format(e+1, epochs),\n \"Iteration: {}\".format(iteration),\n \"Training loss: {:.5f}\".format(loss))\n iteration += 1\n \n if iteration % 5 == 0:\n feed = {inputs_: val_x,\n labels_: val_y}\n val_acc = sess.run(accuracy, feed_dict=feed)\n print(\"Epoch: {}/{}\".format(e, epochs),\n \"Iteration: {}\".format(iteration),\n \"Validation Acc: {:.4f}\".format(val_acc))\n saver.save(sess, \"checkpoints/flowers.ckpt\")", "_____no_output_____" ] ], [ [ "### Testing\n\nBelow you see the test accuracy. You can also see the predictions returned for images.", "_____no_output_____" ] ], [ [ "with tf.Session() as sess:\n saver.restore(sess, tf.train.latest_checkpoint('checkpoints'))\n \n feed = {inputs_: test_x,\n labels_: test_y}\n test_acc = sess.run(accuracy, feed_dict=feed)\n print(\"Test accuracy: {:.4f}\".format(test_acc))", "_____no_output_____" ], [ "%matplotlib inline\n\nimport matplotlib.pyplot as plt\nfrom scipy.ndimage import imread", "_____no_output_____" ] ], [ [ "Below, feel free to choose images and see how the trained classifier predicts the flowers in them.", "_____no_output_____" ] ], [ [ "test_img_path = 'flower_photos/roses/10894627425_ec76bbc757_n.jpg'\ntest_img = imread(test_img_path)\nplt.imshow(test_img)", "_____no_output_____" ], [ "# Run this cell if you don't have a vgg graph built\nwith tf.Session() as sess:\n input_ = tf.placeholder(tf.float32, [None, 224, 224, 3])\n vgg = vgg16.Vgg16()\n vgg.build(input_)", "_____no_output_____" ], [ "with tf.Session() as sess:\n img = utils.load_image(test_img_path)\n img = img.reshape((1, 224, 224, 3))\n\n feed_dict = {input_: img}\n code = sess.run(vgg.relu6, feed_dict=feed_dict)\n \nsaver = tf.train.Saver()\nwith tf.Session() as sess:\n saver.restore(sess, tf.train.latest_checkpoint('checkpoints'))\n \n feed = {inputs_: code}\n prediction = sess.run(predicted, feed_dict=feed).squeeze()", "_____no_output_____" ], [ "plt.imshow(test_img)", "_____no_output_____" ], [ "plt.barh(np.arange(5), prediction)\n_ = plt.yticks(np.arange(5), lb.classes_)", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code" ] ]

[ "MIT-0" ]

[ "MIT-0" ]

[ "MIT-0" ]

[ [ [ "### Set up AWS Panorama development environment on SageMaker Notebook\n\nThis notebook installs dependencies required for AWS Panorama application development. Run following cells just once, before starting labs.", "_____no_output_____" ] ], [ [ "%pip install panoramacli\n%pip install mxnet\n%pip install gluoncv", "_____no_output_____" ], [ "!./scripts/install-docker.sh", "_____no_output_____" ], [ "# for CPU build\n!./scripts/install-dlr.sh\n\n# for p2/p3/g4 instance, we could use pre-built package to skip long building time\n#%pip install https://neo-ai-dlr-release.s3-us-west-2.amazonaws.com/v1.10.0/gpu/dlr-1.10.0-py3-none-any.whl\n", "_____no_output_____" ], [ "!./scripts/install-glibc-sm.sh", "_____no_output_____" ], [ "!./scripts/create-opt-aws-panorama.sh", "_____no_output_____" ], [ "!./scripts/install-videos.sh", "_____no_output_____" ] ], [ [ "#### Environment setting up has completed\n\nOpen a notebooks under \"labs/\".", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown" ]

[ [ "markdown" ], [ "code", "code", "code", "code", "code", "code" ], [ "markdown" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Transfer learning with Tensorflow\n在这个Notebook当中，我们将介绍如何实用Tensorflow框架实现迁移学习。简单的来说，迁移学习就是利用已经训练好的模型中学习到的特征（Features），再根据用户需要添加额外的网络层，进行快速的针对新的特定的数据集的模型训练。由于这样生成的模型大部分的模型参数已经训练好并且已经学习到一定数量的hidden features，在提供新的数据集的时候再进行训练就能有效利用已学习到的知识来进行预测。\n\n本notebook所使用的预训练好的模型是MobileNet V2，其具体的原理就留给负责这部分的同学在后续具体介绍。我们当前只需要了解其大致的网络结构即可（结构如下图）。\n\n\n\n\n\nMobileNet论文：https://arxiv.org/abs/1801.04381\n\n我们使用的是“猫猫狗狗”数据集，具体的样例我们会在下文的数据部分展示。\n\n下面是大纲：\n\n1. 数据集 & 构造模型输入流\n2. 组合模型\n - 载入预训练好的模型及其参数\n - 在预训练好的模型的基础上添加额外的分类层\n3. 训练模型\n4. 测试模型\n\n参考文献：https://tensorflow.google.cn/api_docs/python/tf/keras/preprocessing/image_dataset_from_directory\n", "_____no_output_____" ] ], [ [ "import matplotlib.pyplot as plt\nimport numpy as np\nimport os\nimport tensorflow as tf\n\nfrom tensorflow.keras.preprocessing import image_dataset_from_directory", "_____no_output_____" ] ], [ [ "## 1. 数据预处理\n### 数据集分组\n在这里我们数据集已经按照路径分为Train集和Validation集。", "_____no_output_____" ] ], [ [ "PATH = os.path.join('./data', 'cats_and_dogs_filtered')\n\ntrain_dir = os.path.join(PATH, 'train')\nvalidation_dir = os.path.join(PATH, 'validation')\n\nBATCH_SIZE = 32\nIMG_SIZE = (160, 160)\n\ntrain_dataset = image_dataset_from_directory(train_dir,\n shuffle=True,\n batch_size=BATCH_SIZE,\n image_size=IMG_SIZE)\nvalidation_dataset = image_dataset_from_directory(validation_dir,\n shuffle=True,\n batch_size=BATCH_SIZE,\n image_size=IMG_SIZE)\n", "Found 2000 files belonging to 2 classes.\nFound 1000 files belonging to 2 classes.\n" ] ], [ [ "### 查看数据集中的样本", "_____no_output_____" ] ], [ [ "class_names = train_dataset.class_names\n\nplt.figure(figsize=(10, 10))\nfor images, labels in train_dataset.take(1):\n for i in range(9):\n ax = plt.subplot(3, 3, i+1)\n plt.imshow(images[i].numpy().astype(\"uint8\"))\n plt.title(class_names[labels[i]])\n plt.axis(\"off\")", "_____no_output_____" ] ], [ [ "### 将图片像素值归一化", "_____no_output_____" ] ], [ [ "rescale_input = tf.keras.layers.experimental.preprocessing.Rescaling(1./255, offset=0)", "_____no_output_____" ] ], [ [ "## 2.组合模型\n### 载入预训练好的模型（MobileNet-v2）", "_____no_output_____" ] ], [ [ "IMG_SHAPE = IMG_SIZE + (3,)\nbase_model = tf.keras.applications.MobileNetV2(input_shape=IMG_SHAPE,\n include_top=False,\n weights='imagenet')", "_____no_output_____" ], [ "image_batch, label_batch = next(iter(train_dataset))\nfeature_batch = base_model(image_batch)\nprint(feature_batch.shape)", "(32, 5, 5, 1280)\n" ] ], [ [ "### 特征提取（Feature Extraction）", "_____no_output_____" ] ], [ [ "# Freeze the MobileNet\nbase_model.trainable = False\nbase_model.summary()", "Model: \"mobilenetv2_1.00_160\"\n__________________________________________________________________________________________________\nLayer (type) Output Shape Param # Connected to \n==================================================================================================\ninput_1 (InputLayer) [(None, 160, 160, 3) 0 \n__________________________________________________________________________________________________\nConv1 (Conv2D) (None, 80, 80, 32) 864 input_1[0][0] \n__________________________________________________________________________________________________\nbn_Conv1 (BatchNormalization) (None, 80, 80, 32) 128 Conv1[0][0] \n__________________________________________________________________________________________________\nConv1_relu (ReLU) (None, 80, 80, 32) 0 bn_Conv1[0][0] \n__________________________________________________________________________________________________\nexpanded_conv_depthwise (Depthw (None, 80, 80, 32) 288 Conv1_relu[0][0] \n__________________________________________________________________________________________________\nexpanded_conv_depthwise_BN (Bat (None, 80, 80, 32) 128 expanded_conv_depthwise[0][0] \n__________________________________________________________________________________________________\nexpanded_conv_depthwise_relu (R (None, 80, 80, 32) 0 expanded_conv_depthwise_BN[0][0] \n__________________________________________________________________________________________________\nexpanded_conv_project (Conv2D) (None, 80, 80, 16) 512 expanded_conv_depthwise_relu[0][0\n__________________________________________________________________________________________________\nexpanded_conv_project_BN (Batch (None, 80, 80, 16) 64 expanded_conv_project[0][0] \n__________________________________________________________________________________________________\nblock_1_expand (Conv2D) (None, 80, 80, 96) 1536 expanded_conv_project_BN[0][0] \n__________________________________________________________________________________________________\nblock_1_expand_BN (BatchNormali (None, 80, 80, 96) 384 block_1_expand[0][0] \n__________________________________________________________________________________________________\nblock_1_expand_relu (ReLU) (None, 80, 80, 96) 0 block_1_expand_BN[0][0] \n__________________________________________________________________________________________________\nblock_1_pad (ZeroPadding2D) (None, 81, 81, 96) 0 block_1_expand_relu[0][0] \n__________________________________________________________________________________________________\nblock_1_depthwise (DepthwiseCon (None, 40, 40, 96) 864 block_1_pad[0][0] \n__________________________________________________________________________________________________\nblock_1_depthwise_BN (BatchNorm (None, 40, 40, 96) 384 block_1_depthwise[0][0] \n__________________________________________________________________________________________________\nblock_1_depthwise_relu (ReLU) (None, 40, 40, 96) 0 block_1_depthwise_BN[0][0] \n__________________________________________________________________________________________________\nblock_1_project (Conv2D) (None, 40, 40, 24) 2304 block_1_depthwise_relu[0][0] \n__________________________________________________________________________________________________\nblock_1_project_BN (BatchNormal (None, 40, 40, 24) 96 block_1_project[0][0] \n__________________________________________________________________________________________________\nblock_2_expand (Conv2D) (None, 40, 40, 144) 3456 block_1_project_BN[0][0] \n__________________________________________________________________________________________________\nblock_2_expand_BN (BatchNormali (None, 40, 40, 144) 576 block_2_expand[0][0] \n__________________________________________________________________________________________________\nblock_2_expand_relu (ReLU) (None, 40, 40, 144) 0 block_2_expand_BN[0][0] \n__________________________________________________________________________________________________\nblock_2_depthwise (DepthwiseCon (None, 40, 40, 144) 1296 block_2_expand_relu[0][0] \n__________________________________________________________________________________________________\nblock_2_depthwise_BN (BatchNorm (None, 40, 40, 144) 576 block_2_depthwise[0][0] \n__________________________________________________________________________________________________\nblock_2_depthwise_relu (ReLU) (None, 40, 40, 144) 0 block_2_depthwise_BN[0][0] \n__________________________________________________________________________________________________\nblock_2_project (Conv2D) (None, 40, 40, 24) 3456 block_2_depthwise_relu[0][0] \n__________________________________________________________________________________________________\nblock_2_project_BN (BatchNormal (None, 40, 40, 24) 96 block_2_project[0][0] \n__________________________________________________________________________________________________\nblock_2_add (Add) (None, 40, 40, 24) 0 block_1_project_BN[0][0] \n block_2_project_BN[0][0] \n__________________________________________________________________________________________________\nblock_3_expand (Conv2D) (None, 40, 40, 144) 3456 block_2_add[0][0] \n__________________________________________________________________________________________________\nblock_3_expand_BN (BatchNormali (None, 40, 40, 144) 576 block_3_expand[0][0] \n__________________________________________________________________________________________________\nblock_3_expand_relu (ReLU) (None, 40, 40, 144) 0 block_3_expand_BN[0][0] \n__________________________________________________________________________________________________\nblock_3_pad (ZeroPadding2D) (None, 41, 41, 144) 0 block_3_expand_relu[0][0] \n__________________________________________________________________________________________________\nblock_3_depthwise (DepthwiseCon (None, 20, 20, 144) 1296 block_3_pad[0][0] \n__________________________________________________________________________________________________\nblock_3_depthwise_BN (BatchNorm (None, 20, 20, 144) 576 block_3_depthwise[0][0] \n__________________________________________________________________________________________________\nblock_3_depthwise_relu (ReLU) (None, 20, 20, 144) 0 block_3_depthwise_BN[0][0] \n__________________________________________________________________________________________________\nblock_3_project (Conv2D) (None, 20, 20, 32) 4608 block_3_depthwise_relu[0][0] \n__________________________________________________________________________________________________\nblock_3_project_BN (BatchNormal (None, 20, 20, 32) 128 block_3_project[0][0] \n__________________________________________________________________________________________________\nblock_4_expand (Conv2D) (None, 20, 20, 192) 6144 block_3_project_BN[0][0] \n__________________________________________________________________________________________________\nblock_4_expand_BN (BatchNormali (None, 20, 20, 192) 768 block_4_expand[0][0] \n__________________________________________________________________________________________________\nblock_4_expand_relu (ReLU) (None, 20, 20, 192) 0 block_4_expand_BN[0][0] \n__________________________________________________________________________________________________\nblock_4_depthwise (DepthwiseCon (None, 20, 20, 192) 1728 block_4_expand_relu[0][0] \n__________________________________________________________________________________________________\nblock_4_depthwise_BN (BatchNorm (None, 20, 20, 192) 768 block_4_depthwise[0][0] \n__________________________________________________________________________________________________\nblock_4_depthwise_relu (ReLU) (None, 20, 20, 192) 0 block_4_depthwise_BN[0][0] \n__________________________________________________________________________________________________\nblock_4_project (Conv2D) (None, 20, 20, 32) 6144 block_4_depthwise_relu[0][0] \n__________________________________________________________________________________________________\nblock_4_project_BN (BatchNormal (None, 20, 20, 32) 128 block_4_project[0][0] \n__________________________________________________________________________________________________\nblock_4_add (Add) (None, 20, 20, 32) 0 block_3_project_BN[0][0] \n block_4_project_BN[0][0] \n__________________________________________________________________________________________________\nblock_5_expand (Conv2D) (None, 20, 20, 192) 6144 block_4_add[0][0] \n__________________________________________________________________________________________________\nblock_5_expand_BN (BatchNormali (None, 20, 20, 192) 768 block_5_expand[0][0] \n__________________________________________________________________________________________________\nblock_5_expand_relu (ReLU) (None, 20, 20, 192) 0 block_5_expand_BN[0][0] \n__________________________________________________________________________________________________\nblock_5_depthwise (DepthwiseCon (None, 20, 20, 192) 1728 block_5_expand_relu[0][0] \n__________________________________________________________________________________________________\nblock_5_depthwise_BN (BatchNorm (None, 20, 20, 192) 768 block_5_depthwise[0][0] \n__________________________________________________________________________________________________\nblock_5_depthwise_relu (ReLU) (None, 20, 20, 192) 0 block_5_depthwise_BN[0][0] \n__________________________________________________________________________________________________\nblock_5_project (Conv2D) (None, 20, 20, 32) 6144 block_5_depthwise_relu[0][0] \n__________________________________________________________________________________________________\nblock_5_project_BN (BatchNormal (None, 20, 20, 32) 128 block_5_project[0][0] \n__________________________________________________________________________________________________\nblock_5_add (Add) (None, 20, 20, 32) 0 block_4_add[0][0] \n block_5_project_BN[0][0] \n__________________________________________________________________________________________________\nblock_6_expand (Conv2D) (None, 20, 20, 192) 6144 block_5_add[0][0] \n__________________________________________________________________________________________________\nblock_6_expand_BN (BatchNormali (None, 20, 20, 192) 768 block_6_expand[0][0] \n__________________________________________________________________________________________________\nblock_6_expand_relu (ReLU) (None, 20, 20, 192) 0 block_6_expand_BN[0][0] \n__________________________________________________________________________________________________\nblock_6_pad (ZeroPadding2D) (None, 21, 21, 192) 0 block_6_expand_relu[0][0] \n__________________________________________________________________________________________________\nblock_6_depthwise (DepthwiseCon (None, 10, 10, 192) 1728 block_6_pad[0][0] \n__________________________________________________________________________________________________\nblock_6_depthwise_BN (BatchNorm (None, 10, 10, 192) 768 block_6_depthwise[0][0] \n__________________________________________________________________________________________________\nblock_6_depthwise_relu (ReLU) (None, 10, 10, 192) 0 block_6_depthwise_BN[0][0] \n__________________________________________________________________________________________________\nblock_6_project (Conv2D) (None, 10, 10, 64) 12288 block_6_depthwise_relu[0][0] \n__________________________________________________________________________________________________\nblock_6_project_BN (BatchNormal (None, 10, 10, 64) 256 block_6_project[0][0] \n__________________________________________________________________________________________________\nblock_7_expand (Conv2D) (None, 10, 10, 384) 24576 block_6_project_BN[0][0] \n__________________________________________________________________________________________________\nblock_7_expand_BN (BatchNormali (None, 10, 10, 384) 1536 block_7_expand[0][0] \n__________________________________________________________________________________________________\nblock_7_expand_relu (ReLU) (None, 10, 10, 384) 0 block_7_expand_BN[0][0] \n__________________________________________________________________________________________________\nblock_7_depthwise (DepthwiseCon (None, 10, 10, 384) 3456 block_7_expand_relu[0][0] \n__________________________________________________________________________________________________\nblock_7_depthwise_BN (BatchNorm (None, 10, 10, 384) 1536 block_7_depthwise[0][0] \n__________________________________________________________________________________________________\nblock_7_depthwise_relu (ReLU) (None, 10, 10, 384) 0 block_7_depthwise_BN[0][0] \n__________________________________________________________________________________________________\nblock_7_project (Conv2D) (None, 10, 10, 64) 24576 block_7_depthwise_relu[0][0] \n__________________________________________________________________________________________________\nblock_7_project_BN (BatchNormal (None, 10, 10, 64) 256 block_7_project[0][0] \n__________________________________________________________________________________________________\nblock_7_add (Add) (None, 10, 10, 64) 0 block_6_project_BN[0][0] \n block_7_project_BN[0][0] \n__________________________________________________________________________________________________\nblock_8_expand (Conv2D) (None, 10, 10, 384) 24576 block_7_add[0][0] \n__________________________________________________________________________________________________\nblock_8_expand_BN (BatchNormali (None, 10, 10, 384) 1536 block_8_expand[0][0] \n__________________________________________________________________________________________________\nblock_8_expand_relu (ReLU) (None, 10, 10, 384) 0 block_8_expand_BN[0][0] \n__________________________________________________________________________________________________\nblock_8_depthwise (DepthwiseCon (None, 10, 10, 384) 3456 block_8_expand_relu[0][0] \n__________________________________________________________________________________________________\nblock_8_depthwise_BN (BatchNorm (None, 10, 10, 384) 1536 block_8_depthwise[0][0] \n__________________________________________________________________________________________________\nblock_8_depthwise_relu (ReLU) (None, 10, 10, 384) 0 block_8_depthwise_BN[0][0] \n__________________________________________________________________________________________________\nblock_8_project (Conv2D) (None, 10, 10, 64) 24576 block_8_depthwise_relu[0][0] \n__________________________________________________________________________________________________\nblock_8_project_BN (BatchNormal (None, 10, 10, 64) 256 block_8_project[0][0] \n__________________________________________________________________________________________________\nblock_8_add (Add) (None, 10, 10, 64) 0 block_7_add[0][0] \n block_8_project_BN[0][0] \n__________________________________________________________________________________________________\nblock_9_expand (Conv2D) (None, 10, 10, 384) 24576 block_8_add[0][0] \n__________________________________________________________________________________________________\nblock_9_expand_BN (BatchNormali (None, 10, 10, 384) 1536 block_9_expand[0][0] \n__________________________________________________________________________________________________\nblock_9_expand_relu (ReLU) (None, 10, 10, 384) 0 block_9_expand_BN[0][0] \n__________________________________________________________________________________________________\nblock_9_depthwise (DepthwiseCon (None, 10, 10, 384) 3456 block_9_expand_relu[0][0] \n__________________________________________________________________________________________________\nblock_9_depthwise_BN (BatchNorm (None, 10, 10, 384) 1536 block_9_depthwise[0][0] \n__________________________________________________________________________________________________\nblock_9_depthwise_relu (ReLU) (None, 10, 10, 384) 0 block_9_depthwise_BN[0][0] \n__________________________________________________________________________________________________\nblock_9_project (Conv2D) (None, 10, 10, 64) 24576 block_9_depthwise_relu[0][0] \n__________________________________________________________________________________________________\nblock_9_project_BN (BatchNormal (None, 10, 10, 64) 256 block_9_project[0][0] \n__________________________________________________________________________________________________\nblock_9_add (Add) (None, 10, 10, 64) 0 block_8_add[0][0] \n block_9_project_BN[0][0] \n__________________________________________________________________________________________________\nblock_10_expand (Conv2D) (None, 10, 10, 384) 24576 block_9_add[0][0] \n__________________________________________________________________________________________________\nblock_10_expand_BN (BatchNormal (None, 10, 10, 384) 1536 block_10_expand[0][0] \n__________________________________________________________________________________________________\nblock_10_expand_relu (ReLU) (None, 10, 10, 384) 0 block_10_expand_BN[0][0] \n__________________________________________________________________________________________________\nblock_10_depthwise (DepthwiseCo (None, 10, 10, 384) 3456 block_10_expand_relu[0][0] \n__________________________________________________________________________________________________\nblock_10_depthwise_BN (BatchNor (None, 10, 10, 384) 1536 block_10_depthwise[0][0] \n__________________________________________________________________________________________________\nblock_10_depthwise_relu (ReLU) (None, 10, 10, 384) 0 block_10_depthwise_BN[0][0] \n__________________________________________________________________________________________________\nblock_10_project (Conv2D) (None, 10, 10, 96) 36864 block_10_depthwise_relu[0][0] \n__________________________________________________________________________________________________\nblock_10_project_BN (BatchNorma (None, 10, 10, 96) 384 block_10_project[0][0] \n__________________________________________________________________________________________________\nblock_11_expand (Conv2D) (None, 10, 10, 576) 55296 block_10_project_BN[0][0] \n__________________________________________________________________________________________________\nblock_11_expand_BN (BatchNormal (None, 10, 10, 576) 2304 block_11_expand[0][0] \n__________________________________________________________________________________________________\nblock_11_expand_relu (ReLU) (None, 10, 10, 576) 0 block_11_expand_BN[0][0] \n__________________________________________________________________________________________________\nblock_11_depthwise (DepthwiseCo (None, 10, 10, 576) 5184 block_11_expand_relu[0][0] \n__________________________________________________________________________________________________\nblock_11_depthwise_BN (BatchNor (None, 10, 10, 576) 2304 block_11_depthwise[0][0] \n__________________________________________________________________________________________________\nblock_11_depthwise_relu (ReLU) (None, 10, 10, 576) 0 block_11_depthwise_BN[0][0] \n__________________________________________________________________________________________________\nblock_11_project (Conv2D) (None, 10, 10, 96) 55296 block_11_depthwise_relu[0][0] \n__________________________________________________________________________________________________\nblock_11_project_BN (BatchNorma (None, 10, 10, 96) 384 block_11_project[0][0] \n__________________________________________________________________________________________________\nblock_11_add (Add) (None, 10, 10, 96) 0 block_10_project_BN[0][0] \n block_11_project_BN[0][0] \n__________________________________________________________________________________________________\nblock_12_expand (Conv2D) (None, 10, 10, 576) 55296 block_11_add[0][0] \n__________________________________________________________________________________________________\nblock_12_expand_BN (BatchNormal (None, 10, 10, 576) 2304 block_12_expand[0][0] \n__________________________________________________________________________________________________\nblock_12_expand_relu (ReLU) (None, 10, 10, 576) 0 block_12_expand_BN[0][0] \n__________________________________________________________________________________________________\nblock_12_depthwise (DepthwiseCo (None, 10, 10, 576) 5184 block_12_expand_relu[0][0] \n__________________________________________________________________________________________________\nblock_12_depthwise_BN (BatchNor (None, 10, 10, 576) 2304 block_12_depthwise[0][0] \n__________________________________________________________________________________________________\nblock_12_depthwise_relu (ReLU) (None, 10, 10, 576) 0 block_12_depthwise_BN[0][0] \n__________________________________________________________________________________________________\nblock_12_project (Conv2D) (None, 10, 10, 96) 55296 block_12_depthwise_relu[0][0] \n__________________________________________________________________________________________________\nblock_12_project_BN (BatchNorma (None, 10, 10, 96) 384 block_12_project[0][0] \n__________________________________________________________________________________________________\nblock_12_add (Add) (None, 10, 10, 96) 0 block_11_add[0][0] \n block_12_project_BN[0][0] \n__________________________________________________________________________________________________\nblock_13_expand (Conv2D) (None, 10, 10, 576) 55296 block_12_add[0][0] \n__________________________________________________________________________________________________\nblock_13_expand_BN (BatchNormal (None, 10, 10, 576) 2304 block_13_expand[0][0] \n__________________________________________________________________________________________________\nblock_13_expand_relu (ReLU) (None, 10, 10, 576) 0 block_13_expand_BN[0][0] \n__________________________________________________________________________________________________\nblock_13_pad (ZeroPadding2D) (None, 11, 11, 576) 0 block_13_expand_relu[0][0] \n__________________________________________________________________________________________________\nblock_13_depthwise (DepthwiseCo (None, 5, 5, 576) 5184 block_13_pad[0][0] \n__________________________________________________________________________________________________\nblock_13_depthwise_BN (BatchNor (None, 5, 5, 576) 2304 block_13_depthwise[0][0] \n__________________________________________________________________________________________________\nblock_13_depthwise_relu (ReLU) (None, 5, 5, 576) 0 block_13_depthwise_BN[0][0] \n__________________________________________________________________________________________________\nblock_13_project (Conv2D) (None, 5, 5, 160) 92160 block_13_depthwise_relu[0][0] \n__________________________________________________________________________________________________\nblock_13_project_BN (BatchNorma (None, 5, 5, 160) 640 block_13_project[0][0] \n__________________________________________________________________________________________________\nblock_14_expand (Conv2D) (None, 5, 5, 960) 153600 block_13_project_BN[0][0] \n__________________________________________________________________________________________________\nblock_14_expand_BN (BatchNormal (None, 5, 5, 960) 3840 block_14_expand[0][0] \n__________________________________________________________________________________________________\nblock_14_expand_relu (ReLU) (None, 5, 5, 960) 0 block_14_expand_BN[0][0] \n__________________________________________________________________________________________________\nblock_14_depthwise (DepthwiseCo (None, 5, 5, 960) 8640 block_14_expand_relu[0][0] \n__________________________________________________________________________________________________\nblock_14_depthwise_BN (BatchNor (None, 5, 5, 960) 3840 block_14_depthwise[0][0] \n__________________________________________________________________________________________________\nblock_14_depthwise_relu (ReLU) (None, 5, 5, 960) 0 block_14_depthwise_BN[0][0] \n__________________________________________________________________________________________________\nblock_14_project (Conv2D) (None, 5, 5, 160) 153600 block_14_depthwise_relu[0][0] \n__________________________________________________________________________________________________\nblock_14_project_BN (BatchNorma (None, 5, 5, 160) 640 block_14_project[0][0] \n__________________________________________________________________________________________________\nblock_14_add (Add) (None, 5, 5, 160) 0 block_13_project_BN[0][0] \n block_14_project_BN[0][0] \n__________________________________________________________________________________________________\nblock_15_expand (Conv2D) (None, 5, 5, 960) 153600 block_14_add[0][0] \n__________________________________________________________________________________________________\nblock_15_expand_BN (BatchNormal (None, 5, 5, 960) 3840 block_15_expand[0][0] \n__________________________________________________________________________________________________\nblock_15_expand_relu (ReLU) (None, 5, 5, 960) 0 block_15_expand_BN[0][0] \n__________________________________________________________________________________________________\nblock_15_depthwise (DepthwiseCo (None, 5, 5, 960) 8640 block_15_expand_relu[0][0] \n__________________________________________________________________________________________________\nblock_15_depthwise_BN (BatchNor (None, 5, 5, 960) 3840 block_15_depthwise[0][0] \n__________________________________________________________________________________________________\nblock_15_depthwise_relu (ReLU) (None, 5, 5, 960) 0 block_15_depthwise_BN[0][0] \n__________________________________________________________________________________________________\nblock_15_project (Conv2D) (None, 5, 5, 160) 153600 block_15_depthwise_relu[0][0] \n__________________________________________________________________________________________________\nblock_15_project_BN (BatchNorma (None, 5, 5, 160) 640 block_15_project[0][0] \n__________________________________________________________________________________________________\nblock_15_add (Add) (None, 5, 5, 160) 0 block_14_add[0][0] \n block_15_project_BN[0][0] \n__________________________________________________________________________________________________\nblock_16_expand (Conv2D) (None, 5, 5, 960) 153600 block_15_add[0][0] \n__________________________________________________________________________________________________\nblock_16_expand_BN (BatchNormal (None, 5, 5, 960) 3840 block_16_expand[0][0] \n__________________________________________________________________________________________________\nblock_16_expand_relu (ReLU) (None, 5, 5, 960) 0 block_16_expand_BN[0][0] \n__________________________________________________________________________________________________\nblock_16_depthwise (DepthwiseCo (None, 5, 5, 960) 8640 block_16_expand_relu[0][0] \n__________________________________________________________________________________________________\nblock_16_depthwise_BN (BatchNor (None, 5, 5, 960) 3840 block_16_depthwise[0][0] \n__________________________________________________________________________________________________\nblock_16_depthwise_relu (ReLU) (None, 5, 5, 960) 0 block_16_depthwise_BN[0][0] \n__________________________________________________________________________________________________\nblock_16_project (Conv2D) (None, 5, 5, 320) 307200 block_16_depthwise_relu[0][0] \n__________________________________________________________________________________________________\nblock_16_project_BN (BatchNorma (None, 5, 5, 320) 1280 block_16_project[0][0] \n__________________________________________________________________________________________________\nConv_1 (Conv2D) (None, 5, 5, 1280) 409600 block_16_project_BN[0][0] \n__________________________________________________________________________________________________\nConv_1_bn (BatchNormalization) (None, 5, 5, 1280) 5120 Conv_1[0][0] \n__________________________________________________________________________________________________\nout_relu (ReLU) (None, 5, 5, 1280) 0 Conv_1_bn[0][0] \n==================================================================================================\nTotal params: 2,257,984\nTrainable params: 0\nNon-trainable params: 2,257,984\n__________________________________________________________________________________________________\n" ], [ "global_average_layer = tf.keras.layers.GlobalAveragePooling2D()\nfeature_batch_average = global_average_layer(feature_batch)\nprint(feature_batch_average.shape)", "(32, 1280)\n" ], [ "prediction_layer = tf.keras.layers.Dense(1)\nprediction_batch = prediction_layer(feature_batch_average)\nprint(prediction_batch.shape)", "(32, 1)\n" ] ], [ [ "### 将上述的各个层集合成新的 Model", "_____no_output_____" ] ], [ [ "inputs = tf.keras.Input(shape=(160, 160, 3))\nx = rescale_input(inputs)\nx = base_model(x, training=False)\nx = tf.keras.layers.Dropout(0.2)(x)\noutputs = prediction_layer(x)\nmodel = tf.keras.Model(inputs, outputs)", "_____no_output_____" ] ], [ [ "### 编译模型", "_____no_output_____" ] ], [ [ "lr = 0.0001\nmodel.compile(optimizer=tf.keras.optimizers.Adam(lr=lr),\n loss=tf.keras.losses.BinaryCrossentropy(from_logits=True),\n metrics='accuracy')", "_____no_output_____" ], [ "model.summary()", "Model: \"model_1\"\n_________________________________________________________________\nLayer (type) Output Shape Param # \n=================================================================\ninput_3 (InputLayer) [(None, 160, 160, 3)] 0 \n_________________________________________________________________\nrescaling (Rescaling) (None, 160, 160, 3) 0 \n_________________________________________________________________\nmobilenetv2_1.00_160 (Functi (None, 5, 5, 1280) 2257984 \n_________________________________________________________________\ndropout_1 (Dropout) (None, 5, 5, 1280) 0 \n_________________________________________________________________\ndense_3 (Dense) (None, 5, 5, 1) 1281 \n=================================================================\nTotal params: 2,259,265\nTrainable params: 1,281\nNon-trainable params: 2,257,984\n_________________________________________________________________\n" ] ], [ [ "## 3.训练模型", "_____no_output_____" ], [ "### 迁移学习后的Validation准确率", "_____no_output_____" ] ], [ [ "epochs = 10\nhistory = model.fit(train_dataset,\n epochs=initial_epochs,\n validation_data=validation_dataset)", "Epoch 1/10\n63/63 [==============================] - 33s 501ms/step - loss: 0.6727 - accuracy: 0.5925 - val_loss: 0.5018 - val_accuracy: 0.7060\nEpoch 2/10\n63/63 [==============================] - 23s 360ms/step - loss: 0.4419 - accuracy: 0.7635 - val_loss: 0.3507 - val_accuracy: 0.8390\nEpoch 3/10\n63/63 [==============================] - 19s 294ms/step - loss: 0.3292 - accuracy: 0.8435 - val_loss: 0.2682 - val_accuracy: 0.9110\nEpoch 4/10\n63/63 [==============================] - 19s 295ms/step - loss: 0.2573 - accuracy: 0.9000 - val_loss: 0.2208 - val_accuracy: 0.9310\nEpoch 5/10\n63/63 [==============================] - 19s 296ms/step - loss: 0.2187 - accuracy: 0.9180 - val_loss: 0.1895 - val_accuracy: 0.9430\nEpoch 6/10\n63/63 [==============================] - 20s 315ms/step - loss: 0.1942 - accuracy: 0.9280 - val_loss: 0.1679 - val_accuracy: 0.9510\nEpoch 7/10\n63/63 [==============================] - 19s 307ms/step - loss: 0.1683 - accuracy: 0.9360 - val_loss: 0.1523 - val_accuracy: 0.9540\nEpoch 8/10\n63/63 [==============================] - 19s 292ms/step - loss: 0.1510 - accuracy: 0.9485 - val_loss: 0.1406 - val_accuracy: 0.9550\nEpoch 9/10\n63/63 [==============================] - 19s 300ms/step - loss: 0.1393 - accuracy: 0.9495 - val_loss: 0.1314 - val_accuracy: 0.9590\nEpoch 10/10\n63/63 [==============================] - 18s 290ms/step - loss: 0.1318 - accuracy: 0.9570 - val_loss: 0.1239 - val_accuracy: 0.9600\n" ] ], [ [ "### 学习曲线", "_____no_output_____" ] ], [ [ "acc = history.history['accuracy']\nval_acc = history.history['val_accuracy']\n\nloss = history.history['loss']\nval_loss = history.history['val_loss']\n\nplt.figure(figsize=(8, 8))\nplt.subplot(2, 1, 1)\nplt.plot(acc, label='Training Accuracy')\nplt.plot(val_acc, label='Validation Accuracy')\nplt.legend(loc='lower right')\nplt.ylabel('Accuracy')\nplt.ylim([min(plt.ylim()),1])\nplt.title('Training and Validation Accuracy')\n\nplt.subplot(2, 1, 2)\nplt.plot(loss, label='Training Loss')\nplt.plot(val_loss, label='Validation Loss')\nplt.legend(loc='upper right')\nplt.ylabel('Cross Entropy')\nplt.ylim([0,1.0])\nplt.title('Training and Validation Loss')\nplt.xlabel('epoch')\nplt.show()", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import pandas as pd", "_____no_output_____" ], [ "import numpy as np", "_____no_output_____" ], [ "import matplotlib.pyplot as plt", "_____no_output_____" ], [ "from sklearn.model_selection import train_test_split", "_____no_output_____" ], [ "import seaborn as sns", "_____no_output_____" ], [ "%matplotlib inline", "_____no_output_____" ], [ "train = pd.read_csv(\"data/train.csv\", header = 0)", "_____no_output_____" ], [ "test = pd.read_csv(\"data/test.csv\", header= 0)", "_____no_output_____" ], [ "train.info()", "<class 'pandas.core.frame.DataFrame'>\nRangeIndex: 891 entries, 0 to 890\nData columns (total 12 columns):\nPassengerId 891 non-null int64\nSurvived 891 non-null int64\nPclass 891 non-null int64\nName 891 non-null object\nSex 891 non-null object\nAge 714 non-null float64\nSibSp 891 non-null int64\nParch 891 non-null int64\nTicket 891 non-null object\nFare 891 non-null float64\nCabin 204 non-null object\nEmbarked 889 non-null object\ndtypes: float64(2), int64(5), object(5)\nmemory usage: 83.6+ KB\n" ], [ "train.head(5)", "_____no_output_____" ], [ "train.Embarked = train.Embarked.replace ([\"C\", \"Q\", \"S\"], [0, 1, 2])", "_____no_output_____" ], [ "train[\"Embarked\"].fillna(train.Embarked.mean(), inplace=True)", "_____no_output_____" ], [ "train.Sex = train.Sex.replace(['male', 'female'], [0, 1])", "_____no_output_____" ], [ "train[\"Age\"].fillna(train.Age.mean(), inplace=True)", "_____no_output_____" ], [ "train", "_____no_output_____" ], [ "corrmat = train.corr()", "_____no_output_____" ], [ "corrmat", "_____no_output_____" ], [ "f, ax = plt.subplots(figsize=(12, 9))", "_____no_output_____" ], [ "sns.heatmap(corrmat, vmax=0.8, square=True)", "_____no_output_____" ], [ "split_data = []", "_____no_output_____" ], [ "for survived in [0,1]:\n split_data.append(train[train.Survived == survived])", "_____no_output_____" ], [ "split_data", "_____no_output_____" ], [ "temp = [i[\"Pclass\"].dropna() for i in split_data]", "_____no_output_____" ], [ "print(temp)", "[0 3\n4 3\n5 3\n6 1\n7 3\n12 3\n13 3\n14 3\n16 3\n18 3\n20 2\n24 3\n26 3\n27 1\n29 3\n30 1\n33 2\n34 1\n35 1\n37 3\n38 3\n40 3\n41 2\n42 3\n45 3\n46 3\n48 3\n49 3\n50 3\n51 3\n ..\n844 3\n845 3\n846 3\n847 3\n848 2\n850 3\n851 3\n852 3\n854 2\n859 3\n860 3\n861 2\n863 3\n864 2\n867 1\n868 3\n870 3\n872 1\n873 3\n876 3\n877 3\n878 3\n881 3\n882 3\n883 2\n884 3\n885 3\n886 2\n888 3\n890 3\nName: Pclass, Length: 549, dtype: int64, 1 1\n2 3\n3 1\n8 3\n9 2\n10 3\n11 1\n15 2\n17 2\n19 3\n21 2\n22 3\n23 1\n25 3\n28 3\n31 1\n32 3\n36 3\n39 3\n43 2\n44 3\n47 3\n52 1\n53 2\n55 1\n56 2\n58 2\n61 1\n65 3\n66 2\n ..\n809 1\n820 1\n821 3\n823 3\n827 2\n828 3\n829 1\n830 3\n831 2\n835 1\n838 3\n839 1\n842 1\n849 1\n853 1\n855 3\n856 1\n857 1\n858 3\n862 1\n865 2\n866 2\n869 3\n871 1\n874 2\n875 3\n879 1\n880 2\n887 1\n889 1\nName: Pclass, Length: 342, dtype: int64]\n" ], [ "plt.hist(temp, histtype=\"barstacked\", bins=3)", "/Users/yamadahikaru/.pyenv/versions/anaconda3-5.0.1/lib/python3.6/site-packages/numpy/core/fromnumeric.py:57: FutureWarning: reshape is deprecated and will raise in a subsequent release. Please use .values.reshape(...) instead\n return getattr(obj, method)(*args, **kwds)\n" ], [ "plt.show()", "_____no_output_____" ], [ "temp = [i[\"Age\"].dropna() for i in split_data]", "_____no_output_____" ], [ "plt.hist(temp, histtype=\"barstacked\", bins=16)", "/Users/yamadahikaru/.pyenv/versions/anaconda3-5.0.1/lib/python3.6/site-packages/numpy/core/fromnumeric.py:57: FutureWarning: reshape is deprecated and will raise in a subsequent release. Please use .values.reshape(...) instead\n return getattr(obj, method)(*args, **kwds)\n" ], [ "plt.show()", "_____no_output_____" ], [ "train", "_____no_output_____" ], [ "train.dtypes", "_____no_output_____" ], [ "train = train.replace(\"male\", 0).replace(\"female\", 1)", "_____no_output_____" ], [ "train.dtypes", "_____no_output_____" ], [ "train", "_____no_output_____" ], [ "corrmat = train.corr()", "_____no_output_____" ], [ "corrmat", "_____no_output_____" ], [ "sns.heatmap(corrmat, vmax=0.8, square=True)", "_____no_output_____" ], [ "combine_salutation = [train]", "_____no_output_____" ], [ "combine_salutation", "_____no_output_____" ], [ "for train in combine_salutation: \n train['Salutation'] = train.Name.str.extract(' ([A-Za-z]+).', expand=False) \nfor train in combine_salutation: \n train['Salutation'] = train['Salutation'].replace(['Lady', 'Countess','Capt', 'Col','Don', 'Dr', 'Major', 'Rev', 'Sir', 'Jonkheer', 'Dona'], 'Rare')\n train['Salutation'] = train['Salutation'].replace('Mlle', 'Miss')\n train['Salutation'] = train['Salutation'].replace('Ms', 'Miss')\n train['Salutation'] = train['Salutation'].replace('Mme', 'Mrs')\n del train['Name']\nSalutation_mapping = {\"Mr\": 1, \"Miss\": 2, \"Mrs\": 3, \"Master\": 4, \"Rare\": 5} \nfor train in combine_salutation: \n train['Salutation'] = train['Salutation'].map(Salutation_mapping) \n train['Salutation'] = train['Salutation'].fillna(0)", "_____no_output_____" ], [ "combine_salutation", "_____no_output_____" ], [ "combine_tickets = combine_salutation", "_____no_output_____" ], [ "combine_tickets", "_____no_output_____" ], [ "for train in combine_tickets:\n train['Ticket_Lett'] = train['Ticket'].apply(lambda x: str(x)[0])\n train['Ticket_Lett'] = train['Ticket_Lett'].apply(lambda x: str(x)) \n train['Ticket_Lett'] = np.where((train['Ticket_Lett']).isin(['1', '2', '3', 'S', 'P', 'C', 'A']), train['Ticket_Lett'], np.where((train['Ticket_Lett']).isin(['W', '4', '7', '6', 'L', '5', '8']), '0','0')) \n train['Ticket_Len'] = train['Ticket'].apply(lambda x: len(x)) \n del train['Ticket'] \ntrain['Ticket_Lett']=train['Ticket_Lett'].replace(\"1\",1).replace(\"2\",2).replace(\"3\",3).replace(\"0\",0).replace(\"S\",3).replace(\"P\",1).replace(\"C\",2).replace(\"A\",0)\n", "_____no_output_____" ], [ "combine_tickets", "_____no_output_____" ], [ "combine_cabin = combine_tickets", "_____no_output_____" ], [ "for train in combine_cabin: \n train['Cabin_Lett'] = train['Cabin'].apply(lambda x: str(x)[0]) \n train['Cabin_Lett'] = train['Cabin_Lett'].apply(lambda x: str(x)) \n train['Cabin_Lett'] = np.where((train['Cabin_Lett']).isin([ 'F', 'E', 'D', 'C', 'B', 'A']),train['Cabin_Lett'], np.where((train['Cabin_Lett']).isin(['W', '4', '7', '6', 'L', '5', '8']), '0','0'))\ndel train['Cabin'] \ntrain['Cabin_Lett']=train['Cabin_Lett'].replace(\"A\",1).replace(\"B\",2).replace(\"C\",1).replace(\"0\",0).replace(\"D\",2).replace(\"E\",2).replace(\"F\",1)", "_____no_output_____" ], [ "train[\"Familysize\"] = train[\"SibSp\"] + train[\"Parch\"] + 1", "_____no_output_____" ], [ "train_data = train.values\nxs = train_data[:, 2:] # Pclass以降の変数\ny = train_data[:, 1] # 正解データ", "_____no_output_____" ], [ "xs", "_____no_output_____" ], [ "y", "_____no_output_____" ], [ "test.info()", "<class 'pandas.core.frame.DataFrame'>\nRangeIndex: 418 entries, 0 to 417\nData columns (total 11 columns):\nPassengerId 418 non-null int64\nPclass 418 non-null int64\nName 418 non-null object\nSex 418 non-null object\nAge 332 non-null float64\nSibSp 418 non-null int64\nParch 418 non-null int64\nTicket 418 non-null object\nFare 417 non-null float64\nCabin 91 non-null object\nEmbarked 418 non-null object\ndtypes: float64(2), int64(4), object(5)\nmemory usage: 36.0+ KB\n" ], [ "test['Age'].fillna(train.Age.mean(), inplace = True)", "_____no_output_____" ], [ "test[\"Fare\"].fillna(train.Fare.mean(), inplace = True)", "_____no_output_____" ], [ "test.Name = test.Name.replace(\"male\",0).replace(\"female\",1)", "_____no_output_____" ], [ "test.Embarked = test.Embarked.replace ([\"C\", \"Q\", \"S\"], [0, 1, 2])", "_____no_output_____" ], [ "test.Sex = test.Sex.replace(['male', 'female'], [0, 1])", "_____no_output_____" ], [ "test", "_____no_output_____" ], [ "combine = [test]\nfor test in combine:\n test['Salutation'] = test.Name.str.extract(' ([A-Za-z]+)\\.', expand=False)\nfor test in combine:\n test['Salutation'] = test['Salutation'].replace(['Lady', 'Countess','Capt', 'Col',\\\n 'Don', 'Dr', 'Major', 'Rev', 'Sir', 'Jonkheer', 'Dona'], 'Rare')\n\n test['Salutation'] = test['Salutation'].replace('Mlle', 'Miss')\n test['Salutation'] = test['Salutation'].replace('Ms', 'Miss')\n test['Salutation'] = test['Salutation'].replace('Mme', 'Mrs')\n del test['Name']\nSalutation_mapping = {\"Mr\": 1, \"Miss\": 2, \"Mrs\": 3, \"Master\": 4, \"Rare\": 5}\n\nfor test in combine:\n test['Salutation'] = test['Salutation'].map(Salutation_mapping)\n test['Salutation'] = test['Salutation'].fillna(0)\n\nfor test in combine:\n test['Ticket_Lett'] = test['Ticket'].apply(lambda x: str(x)[0])\n test['Ticket_Lett'] = test['Ticket_Lett'].apply(lambda x: str(x))\n test['Ticket_Lett'] = np.where((test['Ticket_Lett']).isin(['1', '2', '3', 'S', 'P', 'C', 'A']), test['Ticket_Lett'],\n np.where((test['Ticket_Lett']).isin(['W', '4', '7', '6', 'L', '5', '8']),\n '0', '0'))\n test['Ticket_Len'] = test['Ticket'].apply(lambda x: len(x))\n del test['Ticket']\ntest['Ticket_Lett']=test['Ticket_Lett'].replace(\"1\",1).replace(\"2\",2).replace(\"3\",3).replace(\"0\",0).replace(\"S\",3).replace(\"P\",1).replace(\"C\",2).replace(\"A\",0) \n\nfor test in combine:\n test['Cabin_Lett'] = test['Cabin'].apply(lambda x: str(x)[0])\n test['Cabin_Lett'] = test['Cabin_Lett'].apply(lambda x: str(x))\n test['Cabin_Lett'] = np.where((test['Cabin_Lett']).isin(['T', 'H', 'G', 'F', 'E', 'D', 'C', 'B', 'A']),test['Cabin_Lett'],\n np.where((test['Cabin_Lett']).isin(['W', '4', '7', '6', 'L', '5', '8']),\n '0','0')) \n del test['Cabin']\ntest['Cabin_Lett']=test['Cabin_Lett'].replace(\"A\",1).replace(\"B\",2).replace(\"C\",1).replace(\"0\",0).replace(\"D\",2).replace(\"E\",2).replace(\"F\",1).replace(\"G\",1) \n\ntest[\"FamilySize\"] = train[\"SibSp\"] + train[\"Parch\"] + 1\n \ntest_data = test.values\nxs_test = test_data[:, 1:]", "_____no_output_____" ], [ "train.head(3)", "_____no_output_____" ], [ "test.head(3)", "_____no_output_____" ], [ "from sklearn.ensemble import RandomForestClassifier", "_____no_output_____" ], [ "model = RandomForestClassifier()", "_____no_output_____" ], [ "model.fit(xs, y)", "_____no_output_____" ], [ "Y_pred = model.predict(xs_test)", "_____no_output_____" ], [ "import csv\nwith open(\"predict_result_data.csv\", \"w\") as f:\n writer = csv.writer(f, lineterminator='\\n')\n writer.writerow([\"PassengerId\", \"Survived\"])\n for pid, survived in zip(test_data[:,0].astype(int), Y_pred.astype(int)):\n writer.writerow([pid, survived])", "_____no_output_____" ], [ "jupyter nbconvert --to python test.ipynb", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "# 基本程序设计\n- 一切代码输入，请使用英文输入法", "_____no_output_____" ], [ "## 编写一个简单的程序\n- 圆公式面积： area = radius \\* radius \\* 3.1415", "_____no_output_____" ], [ "### 在Python里面不需要定义数据的类型", "_____no_output_____" ], [ "## 控制台的读取与输入\n- input 输入进去的是字符串\n- eval", "_____no_output_____" ], [ "- 在jupyter用shift + tab 键可以跳出解释文档", "_____no_output_____" ], [ "## 变量命名的规范\n- 由字母、数字、下划线构成\n- 不能以数字开头 \\*\n- 标识符不能是关键词(实际上是可以强制改变的，但是对于代码规范而言是极其不适合)\n- 可以是任意长度\n- 驼峰式命名", "_____no_output_____" ], [ "## 变量、赋值语句和赋值表达式\n- 变量: 通俗理解为可以变化的量\n- x = 2 \\* x + 1 在数学中是一个方程，而在语言中它是一个表达式\n- test = test + 1 \\* 变量在赋值之前必须有值", "_____no_output_____" ], [ "## 同时赋值\nvar1, var2,var3... = exp1,exp2,exp3...", "_____no_output_____" ], [ "## 定义常量\n- 常量：表示一种定值标识符，适合于多次使用的场景。比如PI\n- 注意：在其他低级语言中如果定义了常量，那么，该常量是不可以被改变的，但是在Python中一切皆对象，常量也是可以被改变的", "_____no_output_____" ], [ "## 数值数据类型和运算符\n- 在Python中有两种数值类型（int 和 float）适用于加减乘除、模、幂次\n<img src = \"../Photo/01.jpg\"></img>", "_____no_output_____" ], [ "## 运算符 /、//、**", "_____no_output_____" ], [ "## 运算符 %", "_____no_output_____" ], [ "## EP：\n- 25/4 多少，如果要将其转变为整数该怎么改写\n- 输入一个数字判断是奇数还是偶数\n- 进阶: 输入一个秒数，写一个程序将其转换成分和秒：例如500秒等于8分20秒\n- 进阶: 如果今天是星期六，那么10天以后是星期几？ 提示：每个星期的第0天是星期天", "_____no_output_____" ], [ "## 科学计数法\n- 1.234e+2\n- 1.234e-2", "_____no_output_____" ], [ "## 计算表达式和运算优先级\n<img src = \"../Photo/02.png\"></img>\n<img src = \"../Photo/03.png\"></img>", "_____no_output_____" ], [ "## 增强型赋值运算\n<img src = \"../Photo/04.png\"></img>", "_____no_output_____" ], [ "## 类型转换\n- float -> int\n- 四舍五入 round", "_____no_output_____" ], [ "## EP:\n- 如果一个年营业税为0.06%，那么对于197.55e+2的年收入，需要交税为多少？(结果保留2为小数)\n- 必须使用科学计数法", "_____no_output_____" ], [ "# Project\n- 用Python写一个贷款计算器程序：输入的是月供(monthlyPayment) 输出的是总还款数(totalpayment)\n", "_____no_output_____" ], [ "# Homework\n- 1\n<img src=\"../Photo/06.png\"></img>", "_____no_output_____" ] ], [ [ "C=input()\nF=float((9/5))*float(C)+32\nprint(F)", "43\n109.4\n" ] ], [ [ "- 2\n<img src=\"../Photo/07.png\"></img>", "_____no_output_____" ] ], [ [ "r=input()\nh=input()\nS=float(r)**2*3.14\nV=float(S)*float(h)\nprint('底面积为：%.2f'%S)\nprint('体积为：%.2f'%V)", "5.5\n12\n底面积为：94.98\n体积为：1139.82\n" ] ], [ [ "- 3\n<img src=\"../Photo/08.png\"></img>", "_____no_output_____" ] ], [ [ "feet=input()\nmeters=float(feet)*0.305\nprint(meters)", "16.5\n5.0325\n" ] ], [ [ "- 4\n<img src=\"../Photo/10.png\"></img>", "_____no_output_____" ] ], [ [ "k=input()\nt1=input()\nt2=input()\nq=float(k)*(float(t2)-float(t1))*4184\nprint(q)", "55.5\n3.5\n10.5\n1625484.0\n" ] ], [ [ "- 5\n<img src=\"../Photo/11.png\"></img>", "_____no_output_____" ] ], [ [ "ce=input()\nnll=input()\nlx=float(ce)*(float(nll)/1200)\nprint('%.5f'%lx)", "1000\n3.5\n2.91667\n" ] ], [ [ "- 6\n<img src=\"../Photo/12.png\"></img>", "_____no_output_____" ] ], [ [ "v0=input()\nv1=input()\nt=input()\npj=(float(v1)-float(v0))/float(t)\nprint('%.4f'%pj)", "5.5\n50.9\n4.5\n10.0889\n" ] ], [ [ "- 7 进阶\n<img src=\"../Photo/13.png\"></img>", "_____no_output_____" ] ], [ [ "amout=input()\nmoney=0\nfor i in range(6):\n money=(float(amout)+float(money))*(1+0.00417)\nprint('%.2f'%money)\n ", "100\n608.82\n" ] ], [ [ "- 8 进阶\n<img src=\"../Photo/14.png\"></img>", "_____no_output_____" ] ], [ [ "number=int(input())\nif (number>1000)or (number<=0):\n print('false')\nelse:\n gewei=number%10\n shiwei=(number//10)%10\n baiwei=number//100\n sum=gewei+shiwei+baiwei\n print(sum)\n", "999\n27\n" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown", "markdown", "markdown", "markdown", "markdown", "markdown", "markdown", "markdown", "markdown", "markdown", "markdown", "markdown", "markdown", "markdown", "markdown", "markdown", "markdown", "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# PTN Template\nThis notebook serves as a template for single dataset PTN experiments \nIt can be run on its own by setting STANDALONE to True (do a find for \"STANDALONE\" to see where) \nBut it is intended to be executed as part of a *papermill.py script. See any of the \nexperimentes with a papermill script to get started with that workflow. ", "_____no_output_____" ] ], [ [ "%load_ext autoreload\n%autoreload 2\n%matplotlib inline\n\n \nimport os, json, sys, time, random\nimport numpy as np\nimport torch\nfrom torch.optim import Adam\nfrom easydict import EasyDict\nimport matplotlib.pyplot as plt\n\nfrom steves_models.steves_ptn import Steves_Prototypical_Network\n\nfrom steves_utils.lazy_iterable_wrapper import Lazy_Iterable_Wrapper\nfrom steves_utils.iterable_aggregator import Iterable_Aggregator\nfrom steves_utils.ptn_train_eval_test_jig import PTN_Train_Eval_Test_Jig\nfrom steves_utils.torch_sequential_builder import build_sequential\nfrom steves_utils.torch_utils import get_dataset_metrics, ptn_confusion_by_domain_over_dataloader\nfrom steves_utils.utils_v2 import (per_domain_accuracy_from_confusion, get_datasets_base_path)\nfrom steves_utils.PTN.utils import independent_accuracy_assesment\n\nfrom steves_utils.stratified_dataset.episodic_accessor import Episodic_Accessor_Factory\n\nfrom steves_utils.ptn_do_report import (\n get_loss_curve,\n get_results_table,\n get_parameters_table,\n get_domain_accuracies,\n)\n\nfrom steves_utils.transforms import get_chained_transform", "_____no_output_____" ] ], [ [ "# Required Parameters\nThese are allowed parameters, not defaults\nEach of these values need to be present in the injected parameters (the notebook will raise an exception if they are not present)\n\nPapermill uses the cell tag \"parameters\" to inject the real parameters below this cell.\nEnable tags to see what I mean", "_____no_output_____" ] ], [ [ "required_parameters = {\n \"experiment_name\",\n \"lr\",\n \"device\",\n \"seed\",\n \"dataset_seed\",\n \"labels_source\",\n \"labels_target\",\n \"domains_source\",\n \"domains_target\",\n \"num_examples_per_domain_per_label_source\",\n \"num_examples_per_domain_per_label_target\",\n \"n_shot\",\n \"n_way\",\n \"n_query\",\n \"train_k_factor\",\n \"val_k_factor\",\n \"test_k_factor\",\n \"n_epoch\",\n \"patience\",\n \"criteria_for_best\",\n \"x_transforms_source\",\n \"x_transforms_target\",\n \"episode_transforms_source\",\n \"episode_transforms_target\",\n \"pickle_name\",\n \"x_net\",\n \"NUM_LOGS_PER_EPOCH\",\n \"BEST_MODEL_PATH\",\n \"torch_default_dtype\"\n}", "_____no_output_____" ], [ "\n\nstandalone_parameters = {}\nstandalone_parameters[\"experiment_name\"] = \"STANDALONE PTN\"\nstandalone_parameters[\"lr\"] = 0.0001\nstandalone_parameters[\"device\"] = \"cuda\"\n\nstandalone_parameters[\"seed\"] = 1337\nstandalone_parameters[\"dataset_seed\"] = 1337\n\n\nstandalone_parameters[\"num_examples_per_domain_per_label_source\"]=100\nstandalone_parameters[\"num_examples_per_domain_per_label_target\"]=100\n\nstandalone_parameters[\"n_shot\"] = 3\nstandalone_parameters[\"n_query\"] = 2\nstandalone_parameters[\"train_k_factor\"] = 1\nstandalone_parameters[\"val_k_factor\"] = 2\nstandalone_parameters[\"test_k_factor\"] = 2\n\n\nstandalone_parameters[\"n_epoch\"] = 100\n\nstandalone_parameters[\"patience\"] = 10\nstandalone_parameters[\"criteria_for_best\"] = \"target_accuracy\"\n\nstandalone_parameters[\"x_transforms_source\"] = [\"unit_power\"]\nstandalone_parameters[\"x_transforms_target\"] = [\"unit_power\"]\nstandalone_parameters[\"episode_transforms_source\"] = []\nstandalone_parameters[\"episode_transforms_target\"] = []\n\nstandalone_parameters[\"torch_default_dtype\"] = \"torch.float32\" \n\n\n\nstandalone_parameters[\"x_net\"] = [\n {\"class\": \"nnReshape\", \"kargs\": {\"shape\":[-1, 1, 2, 256]}},\n {\"class\": \"Conv2d\", \"kargs\": { \"in_channels\":1, \"out_channels\":256, \"kernel_size\":(1,7), \"bias\":False, \"padding\":(0,3), },},\n {\"class\": \"ReLU\", \"kargs\": {\"inplace\": True}},\n {\"class\": \"BatchNorm2d\", \"kargs\": {\"num_features\":256}},\n\n {\"class\": \"Conv2d\", \"kargs\": { \"in_channels\":256, \"out_channels\":80, \"kernel_size\":(2,7), \"bias\":True, \"padding\":(0,3), },},\n {\"class\": \"ReLU\", \"kargs\": {\"inplace\": True}},\n {\"class\": \"BatchNorm2d\", \"kargs\": {\"num_features\":80}},\n {\"class\": \"Flatten\", \"kargs\": {}},\n\n {\"class\": \"Linear\", \"kargs\": {\"in_features\": 80*256, \"out_features\": 256}}, # 80 units per IQ pair\n {\"class\": \"ReLU\", \"kargs\": {\"inplace\": True}},\n {\"class\": \"BatchNorm1d\", \"kargs\": {\"num_features\":256}},\n\n {\"class\": \"Linear\", \"kargs\": {\"in_features\": 256, \"out_features\": 256}},\n]\n\n# Parameters relevant to results\n# These parameters will basically never need to change\nstandalone_parameters[\"NUM_LOGS_PER_EPOCH\"] = 10\nstandalone_parameters[\"BEST_MODEL_PATH\"] = \"./best_model.pth\"\n\n# uncomment for CORES dataset\nfrom steves_utils.CORES.utils import (\n ALL_NODES,\n ALL_NODES_MINIMUM_1000_EXAMPLES,\n ALL_DAYS\n)\n\n\nstandalone_parameters[\"labels_source\"] = ALL_NODES\nstandalone_parameters[\"labels_target\"] = ALL_NODES\n\nstandalone_parameters[\"domains_source\"] = [1]\nstandalone_parameters[\"domains_target\"] = [2,3,4,5]\n\nstandalone_parameters[\"pickle_name\"] = \"cores.stratified_ds.2022A.pkl\"\n\n\n# Uncomment these for ORACLE dataset\n# from steves_utils.ORACLE.utils_v2 import (\n# ALL_DISTANCES_FEET,\n# ALL_RUNS,\n# ALL_SERIAL_NUMBERS,\n# )\n# standalone_parameters[\"labels_source\"] = ALL_SERIAL_NUMBERS\n# standalone_parameters[\"labels_target\"] = ALL_SERIAL_NUMBERS\n# standalone_parameters[\"domains_source\"] = [8,20, 38,50]\n# standalone_parameters[\"domains_target\"] = [14, 26, 32, 44, 56]\n# standalone_parameters[\"pickle_name\"] = \"oracle.frame_indexed.stratified_ds.2022A.pkl\"\n# standalone_parameters[\"num_examples_per_domain_per_label_source\"]=1000\n# standalone_parameters[\"num_examples_per_domain_per_label_target\"]=1000\n\n# Uncomment these for Metahan dataset\n# standalone_parameters[\"labels_source\"] = list(range(19))\n# standalone_parameters[\"labels_target\"] = list(range(19))\n# standalone_parameters[\"domains_source\"] = [0]\n# standalone_parameters[\"domains_target\"] = [1]\n# standalone_parameters[\"pickle_name\"] = \"metehan.stratified_ds.2022A.pkl\"\n# standalone_parameters[\"n_way\"] = len(standalone_parameters[\"labels_source\"])\n# standalone_parameters[\"num_examples_per_domain_per_label_source\"]=200\n# standalone_parameters[\"num_examples_per_domain_per_label_target\"]=100\n\n\nstandalone_parameters[\"n_way\"] = len(standalone_parameters[\"labels_source\"])", "_____no_output_____" ], [ "# Parameters\nparameters = {\n \"experiment_name\": \"tuned_1v2:oracle.run1\",\n \"device\": \"cuda\",\n \"lr\": 0.0001,\n \"labels_source\": [\n \"3123D52\",\n \"3123D65\",\n \"3123D79\",\n \"3123D80\",\n \"3123D54\",\n \"3123D70\",\n \"3123D7B\",\n \"3123D89\",\n \"3123D58\",\n \"3123D76\",\n \"3123D7D\",\n \"3123EFE\",\n \"3123D64\",\n \"3123D78\",\n \"3123D7E\",\n \"3124E4A\",\n ],\n \"labels_target\": [\n \"3123D52\",\n \"3123D65\",\n \"3123D79\",\n \"3123D80\",\n \"3123D54\",\n \"3123D70\",\n \"3123D7B\",\n \"3123D89\",\n \"3123D58\",\n \"3123D76\",\n \"3123D7D\",\n \"3123EFE\",\n \"3123D64\",\n \"3123D78\",\n \"3123D7E\",\n \"3124E4A\",\n ],\n \"episode_transforms_source\": [],\n \"episode_transforms_target\": [],\n \"domains_source\": [8, 32, 50],\n \"domains_target\": [14, 20, 26, 38, 44],\n \"num_examples_per_domain_per_label_source\": -1,\n \"num_examples_per_domain_per_label_target\": -1,\n \"n_shot\": 3,\n \"n_way\": 16,\n \"n_query\": 2,\n \"train_k_factor\": 3,\n \"val_k_factor\": 2,\n \"test_k_factor\": 2,\n \"torch_default_dtype\": \"torch.float32\",\n \"n_epoch\": 50,\n \"patience\": 3,\n \"criteria_for_best\": \"target_accuracy\",\n \"x_net\": [\n {\"class\": \"nnReshape\", \"kargs\": {\"shape\": [-1, 1, 2, 256]}},\n {\n \"class\": \"Conv2d\",\n \"kargs\": {\n \"in_channels\": 1,\n \"out_channels\": 256,\n \"kernel_size\": [1, 7],\n \"bias\": False,\n \"padding\": [0, 3],\n },\n },\n {\"class\": \"ReLU\", \"kargs\": {\"inplace\": True}},\n {\"class\": \"BatchNorm2d\", \"kargs\": {\"num_features\": 256}},\n {\n \"class\": \"Conv2d\",\n \"kargs\": {\n \"in_channels\": 256,\n \"out_channels\": 80,\n \"kernel_size\": [2, 7],\n \"bias\": True,\n \"padding\": [0, 3],\n },\n },\n {\"class\": \"ReLU\", \"kargs\": {\"inplace\": True}},\n {\"class\": \"BatchNorm2d\", \"kargs\": {\"num_features\": 80}},\n {\"class\": \"Flatten\", \"kargs\": {}},\n {\"class\": \"Linear\", \"kargs\": {\"in_features\": 20480, \"out_features\": 256}},\n {\"class\": \"ReLU\", \"kargs\": {\"inplace\": True}},\n {\"class\": \"BatchNorm1d\", \"kargs\": {\"num_features\": 256}},\n {\"class\": \"Linear\", \"kargs\": {\"in_features\": 256, \"out_features\": 256}},\n ],\n \"NUM_LOGS_PER_EPOCH\": 10,\n \"BEST_MODEL_PATH\": \"./best_model.pth\",\n \"pickle_name\": \"oracle.Run1_10kExamples_stratified_ds.2022A.pkl\",\n \"x_transforms_source\": [\"unit_mag\"],\n \"x_transforms_target\": [\"unit_mag\"],\n \"dataset_seed\": 1337,\n \"seed\": 1337,\n}\n", "_____no_output_____" ], [ "# Set this to True if you want to run this template directly\nSTANDALONE = False\nif STANDALONE:\n print(\"parameters not injected, running with standalone_parameters\")\n parameters = standalone_parameters\n\nif not 'parameters' in locals() and not 'parameters' in globals():\n raise Exception(\"Parameter injection failed\")\n\n#Use an easy dict for all the parameters\np = EasyDict(parameters)\n\nsupplied_keys = set(p.keys())\n\nif supplied_keys != required_parameters:\n print(\"Parameters are incorrect\")\n if len(supplied_keys - required_parameters)>0: print(\"Shouldn't have:\", str(supplied_keys - required_parameters))\n if len(required_parameters - supplied_keys)>0: print(\"Need to have:\", str(required_parameters - supplied_keys))\n raise RuntimeError(\"Parameters are incorrect\")\n\n", "_____no_output_____" ], [ "###################################\n# Set the RNGs and make it all deterministic\n###################################\nnp.random.seed(p.seed)\nrandom.seed(p.seed)\ntorch.manual_seed(p.seed)\n\ntorch.use_deterministic_algorithms(True) ", "_____no_output_____" ], [ "###########################################\n# The stratified datasets honor this\n###########################################\ntorch.set_default_dtype(eval(p.torch_default_dtype))", "_____no_output_____" ], [ "###################################\n# Build the network(s)\n# Note: It's critical to do this AFTER setting the RNG\n# (This is due to the randomized initial weights)\n###################################\nx_net = build_sequential(p.x_net)", "_____no_output_____" ], [ "start_time_secs = time.time()", "_____no_output_____" ], [ "###################################\n# Build the dataset\n###################################\n\nif p.x_transforms_source == []: x_transform_source = None\nelse: x_transform_source = get_chained_transform(p.x_transforms_source) \n\nif p.x_transforms_target == []: x_transform_target = None\nelse: x_transform_target = get_chained_transform(p.x_transforms_target)\n\nif p.episode_transforms_source == []: episode_transform_source = None\nelse: raise Exception(\"episode_transform_source not implemented\")\n\nif p.episode_transforms_target == []: episode_transform_target = None\nelse: raise Exception(\"episode_transform_target not implemented\")\n\n\neaf_source = Episodic_Accessor_Factory(\n labels=p.labels_source,\n domains=p.domains_source,\n num_examples_per_domain_per_label=p.num_examples_per_domain_per_label_source,\n iterator_seed=p.seed,\n dataset_seed=p.dataset_seed,\n n_shot=p.n_shot,\n n_way=p.n_way,\n n_query=p.n_query,\n train_val_test_k_factors=(p.train_k_factor,p.val_k_factor,p.test_k_factor),\n pickle_path=os.path.join(get_datasets_base_path(), p.pickle_name),\n x_transform_func=x_transform_source,\n example_transform_func=episode_transform_source,\n \n)\ntrain_original_source, val_original_source, test_original_source = eaf_source.get_train(), eaf_source.get_val(), eaf_source.get_test()\n\n\neaf_target = Episodic_Accessor_Factory(\n labels=p.labels_target,\n domains=p.domains_target,\n num_examples_per_domain_per_label=p.num_examples_per_domain_per_label_target,\n iterator_seed=p.seed,\n dataset_seed=p.dataset_seed,\n n_shot=p.n_shot,\n n_way=p.n_way,\n n_query=p.n_query,\n train_val_test_k_factors=(p.train_k_factor,p.val_k_factor,p.test_k_factor),\n pickle_path=os.path.join(get_datasets_base_path(), p.pickle_name),\n x_transform_func=x_transform_target,\n example_transform_func=episode_transform_target,\n)\ntrain_original_target, val_original_target, test_original_target = eaf_target.get_train(), eaf_target.get_val(), eaf_target.get_test()\n\n\ntransform_lambda = lambda ex: ex[1] # Original is (<domain>, <episode>) so we strip down to episode only\n\ntrain_processed_source = Lazy_Iterable_Wrapper(train_original_source, transform_lambda)\nval_processed_source = Lazy_Iterable_Wrapper(val_original_source, transform_lambda)\ntest_processed_source = Lazy_Iterable_Wrapper(test_original_source, transform_lambda)\n\ntrain_processed_target = Lazy_Iterable_Wrapper(train_original_target, transform_lambda)\nval_processed_target = Lazy_Iterable_Wrapper(val_original_target, transform_lambda)\ntest_processed_target = Lazy_Iterable_Wrapper(test_original_target, transform_lambda)\n\ndatasets = EasyDict({\n \"source\": {\n \"original\": {\"train\":train_original_source, \"val\":val_original_source, \"test\":test_original_source},\n \"processed\": {\"train\":train_processed_source, \"val\":val_processed_source, \"test\":test_processed_source}\n },\n \"target\": {\n \"original\": {\"train\":train_original_target, \"val\":val_original_target, \"test\":test_original_target},\n \"processed\": {\"train\":train_processed_target, \"val\":val_processed_target, \"test\":test_processed_target}\n },\n})", "_____no_output_____" ], [ "# Some quick unit tests on the data\nfrom steves_utils.transforms import get_average_power, get_average_magnitude\n\nq_x, q_y, s_x, s_y, truth = next(iter(train_processed_source))\n\nassert q_x.dtype == eval(p.torch_default_dtype)\nassert s_x.dtype == eval(p.torch_default_dtype)\n\nprint(\"Visually inspect these to see if they line up with expected values given the transforms\")\nprint('x_transforms_source', p.x_transforms_source)\nprint('x_transforms_target', p.x_transforms_target)\nprint(\"Average magnitude, source:\", get_average_magnitude(q_x[0].numpy()))\nprint(\"Average power, source:\", get_average_power(q_x[0].numpy()))\n\nq_x, q_y, s_x, s_y, truth = next(iter(train_processed_target))\nprint(\"Average magnitude, target:\", get_average_magnitude(q_x[0].numpy()))\nprint(\"Average power, target:\", get_average_power(q_x[0].numpy()))\n", "Visually inspect these to see if they line up with expected values given the transforms\nx_transforms_source ['unit_mag']\nx_transforms_target ['unit_mag']\nAverage magnitude, source: 1.0\nAverage power, source: 1.3161097\n" ], [ "###################################\n# Build the model\n###################################\nmodel = Steves_Prototypical_Network(x_net, device=p.device, x_shape=(2,256))\noptimizer = Adam(params=model.parameters(), lr=p.lr)", "(2, 256)\n" ], [ "###################################\n# train\n###################################\njig = PTN_Train_Eval_Test_Jig(model, p.BEST_MODEL_PATH, p.device)\n\njig.train(\n train_iterable=datasets.source.processed.train,\n source_val_iterable=datasets.source.processed.val,\n target_val_iterable=datasets.target.processed.val,\n num_epochs=p.n_epoch,\n num_logs_per_epoch=p.NUM_LOGS_PER_EPOCH,\n patience=p.patience,\n optimizer=optimizer,\n criteria_for_best=p.criteria_for_best,\n)", "epoch: 1, [batch: 1 / 12600], examples_per_second: 25.5044, train_label_loss: 2.8021, \n" ], [ "total_experiment_time_secs = time.time() - start_time_secs", "_____no_output_____" ], [ "###################################\n# Evaluate the model\n###################################\nsource_test_label_accuracy, source_test_label_loss = jig.test(datasets.source.processed.test)\ntarget_test_label_accuracy, target_test_label_loss = jig.test(datasets.target.processed.test)\n\nsource_val_label_accuracy, source_val_label_loss = jig.test(datasets.source.processed.val)\ntarget_val_label_accuracy, target_val_label_loss = jig.test(datasets.target.processed.val)\n\nhistory = jig.get_history()\n\ntotal_epochs_trained = len(history[\"epoch_indices\"])\n\nval_dl = Iterable_Aggregator((datasets.source.original.val,datasets.target.original.val))\n\nconfusion = ptn_confusion_by_domain_over_dataloader(model, p.device, val_dl)\nper_domain_accuracy = per_domain_accuracy_from_confusion(confusion)\n\n# Add a key to per_domain_accuracy for if it was a source domain\nfor domain, accuracy in per_domain_accuracy.items():\n per_domain_accuracy[domain] = {\n \"accuracy\": accuracy,\n \"source?\": domain in p.domains_source\n }\n\n# Do an independent accuracy assesment JUST TO BE SURE!\n# _source_test_label_accuracy = independent_accuracy_assesment(model, datasets.source.processed.test, p.device)\n# _target_test_label_accuracy = independent_accuracy_assesment(model, datasets.target.processed.test, p.device)\n# _source_val_label_accuracy = independent_accuracy_assesment(model, datasets.source.processed.val, p.device)\n# _target_val_label_accuracy = independent_accuracy_assesment(model, datasets.target.processed.val, p.device)\n\n# assert(_source_test_label_accuracy == source_test_label_accuracy)\n# assert(_target_test_label_accuracy == target_test_label_accuracy)\n# assert(_source_val_label_accuracy == source_val_label_accuracy)\n# assert(_target_val_label_accuracy == target_val_label_accuracy)\n\nexperiment = {\n \"experiment_name\": p.experiment_name,\n \"parameters\": dict(p),\n \"results\": {\n \"source_test_label_accuracy\": source_test_label_accuracy,\n \"source_test_label_loss\": source_test_label_loss,\n \"target_test_label_accuracy\": target_test_label_accuracy,\n \"target_test_label_loss\": target_test_label_loss,\n \"source_val_label_accuracy\": source_val_label_accuracy,\n \"source_val_label_loss\": source_val_label_loss,\n \"target_val_label_accuracy\": target_val_label_accuracy,\n \"target_val_label_loss\": target_val_label_loss,\n \"total_epochs_trained\": total_epochs_trained,\n \"total_experiment_time_secs\": total_experiment_time_secs,\n \"confusion\": confusion,\n \"per_domain_accuracy\": per_domain_accuracy,\n },\n \"history\": history,\n \"dataset_metrics\": get_dataset_metrics(datasets, \"ptn\"),\n}", "_____no_output_____" ], [ "ax = get_loss_curve(experiment)\nplt.show()", "_____no_output_____" ], [ "get_results_table(experiment)", "_____no_output_____" ], [ "get_domain_accuracies(experiment)", "_____no_output_____" ], [ "print(\"Source Test Label Accuracy:\", experiment[\"results\"][\"source_test_label_accuracy\"], \"Target Test Label Accuracy:\", experiment[\"results\"][\"target_test_label_accuracy\"])\nprint(\"Source Val Label Accuracy:\", experiment[\"results\"][\"source_val_label_accuracy\"], \"Target Val Label Accuracy:\", experiment[\"results\"][\"target_val_label_accuracy\"])", "Source Test Label Accuracy: 0.6930208333333333 Target Test Label Accuracy: 0.6151145833333334\nSource Val Label Accuracy: 0.69234375 Target Val Label Accuracy: 0.6133020833333334\n" ], [ "json.dumps(experiment)", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Dependencies\nimport tweepy\nimport json\nimport numpy as np\nfrom config2 import consumer_key, consumer_secret, access_token, access_token_secret", "_____no_output_____" ], [ "\n", "_____no_output_____" ], [ "# Twitter API Keys\nconsumer_key = consumer_key\nconsumer_secret = consumer_secret\naccess_token = access_token\naccess_token_secret = access_token_secret", "_____no_output_____" ], [ "# Setup Tweepy API Authentication\nauth = tweepy.OAuthHandler(consumer_key, consumer_secret)\nauth.set_access_token(access_token, access_token_secret)\napi = tweepy.API(auth, parser=tweepy.parsers.JSONParser())", "_____no_output_____" ], [ "target_term = '@facebook'", "_____no_output_____" ], [ "# Lists to hold sentiments\n# Variables for holding sentiments\ncompound_list = []\npositive_list = []\nnegative_list = []\nneutral_list = []\n\nfor x in range(1, 20):\n\n # Get all tweets from home feed\n public_tweets = api.user_timeline(target_term, page=x)\n\n # Loop through all tweets\n for tweet in public_tweets:\n\n # Run Vader Analysis on each tweet\n results = analyzer.polarity_scores(tweet[\"text\"])\n compound = results[\"compound\"]\n pos = results[\"pos\"]\n neu = results[\"neu\"]\n neg = results[\"neg\"]\n\n # Add each value to the appropriate list\n compound_list.append(compound)\n positive_list.append(pos)\n negative_list.append(neg)\n neutral_list.append(neu)", "_____no_output_____" ], [ "print(\nsum(compound_list)/len(compound_list),\nsum(positive_list)/len(positive_list),\nsum(negative_list)/len(negative_list),\nsum(neutral_list)/len(neutral_list))", "0.15938236842105247 0.1086210526315789 0.042015789473684235 0.8493921052631583\n" ], [ "sentiments = {\n 'Average Compounded': sum(compound_list) / len(compound_list),\n 'Average Negative': sum(negative_list) / len(negative_list),\n 'Average Positive': sum(positive_list) / len(positive_list),\n 'Average Neutral': sum(neutral_list) / len(neutral_list)\n}\n", "_____no_output_____" ], [ "sentiments", "_____no_output_____" ], [ "len(compound_list)", "_____no_output_____" ] ], [ [ "## mentions of facebook", "_____no_output_____" ] ], [ [ "# Search for all tweets\n# public_tweets = api.search(target_term, count=300, result_type=\"recent\")", "_____no_output_____" ], [ "# Twitter API Keys\nconsumer_key = consumer_key\nconsumer_secret = consumer_secret\naccess_token = access_token\naccess_token_secret = access_token_secret\n\n# Setup Tweepy API Authentication\nauth = tweepy.OAuthHandler(consumer_key, consumer_secret)\nauth.set_access_token(access_token, access_token_secret)\napi = tweepy.API(auth, parser=tweepy.parsers.JSONParser())\n\n", "_____no_output_____" ], [ "print(target_term)", "@facebook\n" ], [ "# Loop through all public_tweets\n\n\nfb_tweets = []\n\ndate = []\noldest_tweet = None\nfor x in range(1,20):\n public_tweets = api.search(target_term, count=100, result_type=\"recent\", max_id=oldest_tweet)\n\n for tweet in public_tweets['statuses']:\n tweet_id = tweet[\"id\"]\n tweet_author = tweet[\"user\"][\"screen_name\"]\n tweet_text = tweet[\"text\"]\n\n fb_tweets.append(tweet['text'])\n date.append(tweet['created_at'])\n \n oldest_tweet = tweet_id - 1\n\nprint(len(fb_tweets))", "1892\n" ], [ "compound_list = []\npositive_list = []\nnegative_list = []\nneutral_list = []\n\nfor tweet in fb_tweets:\n\n # Run Vader Analysis on each tweet\n results = analyzer.polarity_scores(tweet)\n compound = results[\"compound\"]\n pos = results[\"pos\"]\n neu = results[\"neu\"]\n neg = results[\"neg\"]\n\n # Add each value to the appropriate list\n compound_list.append(compound)\n positive_list.append(pos)\n negative_list.append(neg)\n neutral_list.append(neu)\n", "_____no_output_____" ], [ "sentiments = {\n 'Average Compounded': sum(compound_list) / len(compound_list),\n 'Average Negative': sum(negative_list) / len(negative_list),\n 'Average Positive': sum(positive_list) / len(positive_list),\n 'Average Neutral': sum(neutral_list) / len(neutral_list)\n}\n", "_____no_output_____" ], [ "sentiments", "_____no_output_____" ], [ "len(fb_tweets)", "_____no_output_____" ], [ "date[1891]", "_____no_output_____" ], [ "june_14 = {\n 'Text': fb_tweets,\n 'Compounded': compound_list,\n 'Negative': negative_list,\n 'Positive': positive_list,\n 'Neutral': neutral_list,\n 'Date': date\n}", "_____no_output_____" ], [ "import pandas as pd", "_____no_output_____" ], [ "tweets_june_14_df = pd.DataFrame(june_14)", "_____no_output_____" ], [ "tweets_june_14_df.head()", "_____no_output_____" ], [ "tweets_june_14_df.to_csv('tweets_june_14.csv')", "_____no_output_____" ], [ "print(date[0])\nprint(date[1891])", "Fri Jun 15 00:10:34 +0000 2018\nThu Jun 14 18:50:53 +0000 2018\n" ] ] ]

[ "code", "markdown", "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "### 1.Importing the Relevant Libraries", "_____no_output_____" ] ], [ [ "%matplotlib inline\nimport numpy as np\nimport pandas as pd \nimport matplotlib.pyplot as plt\n\n\nfrom sklearn.linear_model import Ridge, Lasso, LinearRegression\nfrom sklearn.tree import DecisionTreeRegressor\nfrom sklearn.neighbors import KNeighborsRegressor\nfrom sklearn import metrics\nfrom sklearn.ensemble import RandomForestRegressor\n\nfrom sklearn.model_selection import train_test_split\nfrom sklearn.preprocessing import StandardScaler", "_____no_output_____" ] ], [ [ "### 2.Reading Data", "_____no_output_____" ] ], [ [ "df = pd.read_csv(\"WindTunnel.csv\")\ndf", "_____no_output_____" ], [ "df = pd.read_csv(\"WindTunnel.csv\")\nplt.xlabel('Freqency')\nplt.ylabel('Velocity')\nplt.plot(df.Freqency,df.Velocity)\n", "_____no_output_____" ], [ "reg = linear_model.LinearRegression()\nreg.fit(df[[\"Freqency\"]],df.Velocity)\n", "_____no_output_____" ], [ "reg.predict([[65.0]])", "_____no_output_____" ], [ "L=0.2\nr=0.1\nv=3.45\nA=0.0323\n\n\nCL = (2*L)/(r*(v**2)*A)\nCL", "_____no_output_____" ], [ "df = pd.read_csv(\"WindTunnel1.csv\")", "_____no_output_____" ], [ "df", "_____no_output_____" ], [ "df1 = df.copy()", "_____no_output_____" ], [ "df1[\"Coef_lift\"] = (2*df['Lift'])/(df['Dencity']*df['Area']*(df['Velocity']**2))", "_____no_output_____" ], [ "df1", "_____no_output_____" ], [ "X = df1.drop(['Coef_lift'],axis='columns')", "_____no_output_____" ], [ "y = df1.Coef_lift", "_____no_output_____" ] ], [ [ "### Data Pre-process ", "_____no_output_____" ] ], [ [ "\nX_train, X_valid, y_train, y_valid = train_test_split(X,y,test_size = 0.2,random_state = 10)", "_____no_output_____" ], [ "\n\nsc = StandardScaler()\nX_train = sc.fit_transform(X_train)\nX_test = sc.transform(X_valid)\n", "_____no_output_____" ], [ "algos = [LinearRegression(), Ridge(), Lasso(),\n KNeighborsRegressor(), DecisionTreeRegressor(),RandomForestRegressor()]\n\nnames = ['Linear Regression', 'Ridge Regression', 'Lasso Regression',\n 'K Neighbors Regressor', 'Decision Tree Regressor', 'RandomForestRegressor']\n\nrmse_list = []\n", "_____no_output_____" ], [ "for name in algos:\n model = name\n model.fit(X_train,y_train)\n y_pred = model.predict(X_valid)\n MSE= metrics.mean_squared_error(y_valid,y_pred)\n rmse = np.sqrt(MSE)\n rmse_list.append(rmse)", "C:\\Users\\hp\\.conda\\envs\\new\\lib\\site-packages\\sklearn\\ensemble\\forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n" ], [ "evaluation = pd.DataFrame({'Model': names,\n 'RMSE': rmse_list})\n\nevaluation", "_____no_output_____" ] ], [ [ "### Building Model\n ", "_____no_output_____" ] ], [ [ "\nclf = Ridge()\nclf.fit(X_train,y_train)", "_____no_output_____" ], [ "clf.score(X_test,y_test)", "_____no_output_____" ], [ "clf.predict([[6.8,0.74,0.0323,0.27]])", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "doc = nlp(u\"Shubhangi visited the Taj Mahal after taking a SpiceJet flight from Pune.\")", "_____no_output_____" ], [ "for ent in doc.ents:\n print(ent.text, ent.label_)\n", "_____no_output_____" ], [ "doc1 = nlp(u\"Shubhangi Hora visited the Taj Mahal after taking a SpiceJet flight from Pune.\")", "_____no_output_____" ], [ "for ent in doc1.ents:\n print(ent.text, ent.label_)\n", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code" ] ]

[ "BSD-3-Clause" ]

[ "BSD-3-Clause" ]

[ "BSD-3-Clause" ]

[ [ [ "from google.colab import drive\ndrive.mount('/content/drive')", "Mounted at /content/drive\n" ], [ "!wget -c https://repo.continuum.io/miniconda/Miniconda3-latest-Linux-x86_64.sh\n!chmod +x Miniconda3-latest-Linux-x86_64.sh\n!bash ./Miniconda3-latest-Linux-x86_64.sh -b -f -p /usr/local\n!conda install -q -y -c rdkit rdkit python=3.7", "--2022-01-26 18:33:03-- https://repo.continuum.io/miniconda/Miniconda3-latest-Linux-x86_64.sh\nResolving repo.continuum.io (repo.continuum.io)... 104.18.200.79, 104.18.201.79, 2606:4700::6812:c94f, ...\nConnecting to repo.continuum.io (repo.continuum.io)|104.18.200.79|:443... connected.\nHTTP request sent, awaiting response... 301 Moved Permanently\nLocation: https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh [following]\n--2022-01-26 18:33:04-- https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh\nResolving repo.anaconda.com (repo.anaconda.com)... 104.16.131.3, 104.16.130.3, 2606:4700::6810:8203, ...\nConnecting to repo.anaconda.com (repo.anaconda.com)|104.16.131.3|:443... connected.\nHTTP request sent, awaiting response... 200 OK\nLength: 66709754 (64M) [application/x-sh]\nSaving to: ‘Miniconda3-latest-Linux-x86_64.sh’\n\nMiniconda3-latest-L 100%[===================>] 63.62M 75.1MB/s in 0.8s \n\n2022-01-26 18:33:04 (75.1 MB/s) - ‘Miniconda3-latest-Linux-x86_64.sh’ saved [66709754/66709754]\n\nPREFIX=/usr/local\nUnpacking payload ...\nCollecting package metadata (current_repodata.json): - \b\b\\ \b\b| \b\bdone\nSolving environment: - \b\b\\ \b\b| \b\bdone\n\n## Package Plan ##\n\n environment location: /usr/local\n\n added / updated specs:\n - _libgcc_mutex==0.1=main\n - _openmp_mutex==4.5=1_gnu\n - brotlipy==0.7.0=py39h27cfd23_1003\n - ca-certificates==2021.7.5=h06a4308_1\n - certifi==2021.5.30=py39h06a4308_0\n - cffi==1.14.6=py39h400218f_0\n - chardet==4.0.0=py39h06a4308_1003\n - conda-package-handling==1.7.3=py39h27cfd23_1\n - conda==4.10.3=py39h06a4308_0\n - cryptography==3.4.7=py39hd23ed53_0\n - idna==2.10=pyhd3eb1b0_0\n - ld_impl_linux-64==2.35.1=h7274673_9\n - libffi==3.3=he6710b0_2\n - libgcc-ng==9.3.0=h5101ec6_17\n - libgomp==9.3.0=h5101ec6_17\n - libstdcxx-ng==9.3.0=hd4cf53a_17\n - ncurses==6.2=he6710b0_1\n - openssl==1.1.1k=h27cfd23_0\n - pip==21.1.3=py39h06a4308_0\n - pycosat==0.6.3=py39h27cfd23_0\n - pycparser==2.20=py_2\n - pyopenssl==20.0.1=pyhd3eb1b0_1\n - pysocks==1.7.1=py39h06a4308_0\n - python==3.9.5=h12debd9_4\n - readline==8.1=h27cfd23_0\n - requests==2.25.1=pyhd3eb1b0_0\n - ruamel_yaml==0.15.100=py39h27cfd23_0\n - setuptools==52.0.0=py39h06a4308_0\n - six==1.16.0=pyhd3eb1b0_0\n - sqlite==3.36.0=hc218d9a_0\n - tk==8.6.10=hbc83047_0\n - tqdm==4.61.2=pyhd3eb1b0_1\n - tzdata==2021a=h52ac0ba_0\n - urllib3==1.26.6=pyhd3eb1b0_1\n - wheel==0.36.2=pyhd3eb1b0_0\n - xz==5.2.5=h7b6447c_0\n - yaml==0.2.5=h7b6447c_0\n - zlib==1.2.11=h7b6447c_3\n\n\nThe following NEW packages will be INSTALLED:\n\n _libgcc_mutex pkgs/main/linux-64::_libgcc_mutex-0.1-main\n _openmp_mutex pkgs/main/linux-64::_openmp_mutex-4.5-1_gnu\n brotlipy pkgs/main/linux-64::brotlipy-0.7.0-py39h27cfd23_1003\n ca-certificates pkgs/main/linux-64::ca-certificates-2021.7.5-h06a4308_1\n certifi pkgs/main/linux-64::certifi-2021.5.30-py39h06a4308_0\n cffi pkgs/main/linux-64::cffi-1.14.6-py39h400218f_0\n chardet pkgs/main/linux-64::chardet-4.0.0-py39h06a4308_1003\n conda pkgs/main/linux-64::conda-4.10.3-py39h06a4308_0\n conda-package-han~ pkgs/main/linux-64::conda-package-handling-1.7.3-py39h27cfd23_1\n cryptography pkgs/main/linux-64::cryptography-3.4.7-py39hd23ed53_0\n idna pkgs/main/noarch::idna-2.10-pyhd3eb1b0_0\n ld_impl_linux-64 pkgs/main/linux-64::ld_impl_linux-64-2.35.1-h7274673_9\n libffi pkgs/main/linux-64::libffi-3.3-he6710b0_2\n libgcc-ng pkgs/main/linux-64::libgcc-ng-9.3.0-h5101ec6_17\n libgomp pkgs/main/linux-64::libgomp-9.3.0-h5101ec6_17\n libstdcxx-ng pkgs/main/linux-64::libstdcxx-ng-9.3.0-hd4cf53a_17\n ncurses pkgs/main/linux-64::ncurses-6.2-he6710b0_1\n openssl pkgs/main/linux-64::openssl-1.1.1k-h27cfd23_0\n pip pkgs/main/linux-64::pip-21.1.3-py39h06a4308_0\n pycosat pkgs/main/linux-64::pycosat-0.6.3-py39h27cfd23_0\n pycparser pkgs/main/noarch::pycparser-2.20-py_2\n pyopenssl pkgs/main/noarch::pyopenssl-20.0.1-pyhd3eb1b0_1\n pysocks pkgs/main/linux-64::pysocks-1.7.1-py39h06a4308_0\n python pkgs/main/linux-64::python-3.9.5-h12debd9_4\n readline pkgs/main/linux-64::readline-8.1-h27cfd23_0\n requests pkgs/main/noarch::requests-2.25.1-pyhd3eb1b0_0\n ruamel_yaml pkgs/main/linux-64::ruamel_yaml-0.15.100-py39h27cfd23_0\n setuptools pkgs/main/linux-64::setuptools-52.0.0-py39h06a4308_0\n six pkgs/main/noarch::six-1.16.0-pyhd3eb1b0_0\n sqlite pkgs/main/linux-64::sqlite-3.36.0-hc218d9a_0\n tk pkgs/main/linux-64::tk-8.6.10-hbc83047_0\n tqdm pkgs/main/noarch::tqdm-4.61.2-pyhd3eb1b0_1\n tzdata pkgs/main/noarch::tzdata-2021a-h52ac0ba_0\n urllib3 pkgs/main/noarch::urllib3-1.26.6-pyhd3eb1b0_1\n wheel pkgs/main/noarch::wheel-0.36.2-pyhd3eb1b0_0\n xz pkgs/main/linux-64::xz-5.2.5-h7b6447c_0\n yaml pkgs/main/linux-64::yaml-0.2.5-h7b6447c_0\n zlib pkgs/main/linux-64::zlib-1.2.11-h7b6447c_3\n\n\nPreparing transaction: - \b\b\\ \b\b| \b\b/ \b\bdone\nExecuting transaction: \\ \b\b| \b\b/ \b\b- \b\b\\ \b\b| \b\b/ \b\b- \b\b\\ \b\b| \b\b/ \b\b- \b\b\\ \b\b| \b\bdone\ninstallation finished.\nWARNING:\n You currently have a PYTHONPATH environment variable set. This may cause\n unexpected behavior when running the Python interpreter in Miniconda3.\n For best results, please verify that your PYTHONPATH only points to\n directories of packages that are compatible with the Python interpreter\n in Miniconda3: /usr/local\nCollecting package metadata (current_repodata.json): ...working... done\nSolving environment: ...working... done\n\n## Package Plan ##\n\n environment location: /usr/local\n\n added / updated specs:\n - python=3.7\n - rdkit\n\n\nThe following packages will be downloaded:\n\n package | build\n ---------------------------|-----------------\n blas-1.0 | mkl 6 KB\n bottleneck-1.3.2 | py37heb32a55_1 125 KB\n brotlipy-0.7.0 |py37h27cfd23_1003 320 KB\n bzip2-1.0.8 | h7b6447c_0 78 KB\n ca-certificates-2021.10.26 | h06a4308_2 115 KB\n cairo-1.16.0 | hf32fb01_1 1.0 MB\n certifi-2021.10.8 | py37h06a4308_2 151 KB\n cffi-1.15.0 | py37hd667e15_1 222 KB\n chardet-4.0.0 |py37h06a4308_1003 195 KB\n conda-4.11.0 | py37h06a4308_0 14.4 MB\n conda-package-handling-1.7.3| py37h27cfd23_1 881 KB\n cryptography-36.0.0 | py37h9ce1e76_0 1.3 MB\n fontconfig-2.13.1 | h6c09931_0 250 KB\n freetype-2.11.0 | h70c0345_0 618 KB\n giflib-5.2.1 | h7b6447c_0 78 KB\n glib-2.69.1 | h4ff587b_1 1.7 MB\n icu-58.2 | he6710b0_3 10.5 MB\n intel-openmp-2021.4.0 | h06a4308_3561 4.2 MB\n jpeg-9d | h7f8727e_0 232 KB\n lcms2-2.12 | h3be6417_0 312 KB\n libboost-1.73.0 | h3ff78a5_11 13.9 MB\n libpng-1.6.37 | hbc83047_0 278 KB\n libtiff-4.2.0 | h85742a9_0 502 KB\n libuuid-1.0.3 | h7f8727e_2 17 KB\n libwebp-1.2.0 | h89dd481_0 493 KB\n libwebp-base-1.2.0 | h27cfd23_0 437 KB\n libxcb-1.14 | h7b6447c_0 505 KB\n libxml2-2.9.12 | h03d6c58_0 1.2 MB\n lz4-c-1.9.3 | h295c915_1 185 KB\n mkl-2021.4.0 | h06a4308_640 142.6 MB\n mkl-service-2.4.0 | py37h7f8727e_0 56 KB\n mkl_fft-1.3.1 | py37hd3c417c_0 172 KB\n mkl_random-1.2.2 | py37h51133e4_0 287 KB\n numexpr-2.8.1 | py37h6abb31d_0 123 KB\n numpy-1.21.2 | py37h20f2e39_0 23 KB\n numpy-base-1.21.2 | py37h79a1101_0 4.8 MB\n olefile-0.46 | py37_0 50 KB\n openssl-1.1.1m | h7f8727e_0 2.5 MB\n packaging-21.3 | pyhd3eb1b0_0 36 KB\n pandas-1.3.5 | py37h8c16a72_0 9.3 MB\n pcre-8.45 | h295c915_0 207 KB\n pillow-8.4.0 | py37h5aabda8_0 644 KB\n pip-21.2.2 | py37h06a4308_0 1.8 MB\n pixman-0.40.0 | h7f8727e_1 373 KB\n py-boost-1.73.0 | py37ha9443f7_11 204 KB\n pycosat-0.6.3 | py37h27cfd23_0 81 KB\n pyparsing-3.0.4 | pyhd3eb1b0_0 81 KB\n pysocks-1.7.1 | py37_1 27 KB\n python-3.7.11 | h12debd9_0 45.3 MB\n python-dateutil-2.8.2 | pyhd3eb1b0_0 233 KB\n pytz-2021.3 | pyhd3eb1b0_0 171 KB\n rdkit-2020.09.1.0 | py37hd50e099_1 25.8 MB rdkit\n ruamel_yaml-0.15.100 | py37h27cfd23_0 253 KB\n setuptools-58.0.4 | py37h06a4308_0 775 KB\n zstd-1.4.9 | haebb681_0 480 KB\n ------------------------------------------------------------\n Total: 290.5 MB\n\nThe following NEW packages will be INSTALLED:\n\n blas pkgs/main/linux-64::blas-1.0-mkl\n bottleneck pkgs/main/linux-64::bottleneck-1.3.2-py37heb32a55_1\n bzip2 pkgs/main/linux-64::bzip2-1.0.8-h7b6447c_0\n cairo pkgs/main/linux-64::cairo-1.16.0-hf32fb01_1\n fontconfig pkgs/main/linux-64::fontconfig-2.13.1-h6c09931_0\n freetype pkgs/main/linux-64::freetype-2.11.0-h70c0345_0\n giflib pkgs/main/linux-64::giflib-5.2.1-h7b6447c_0\n glib pkgs/main/linux-64::glib-2.69.1-h4ff587b_1\n icu pkgs/main/linux-64::icu-58.2-he6710b0_3\n intel-openmp pkgs/main/linux-64::intel-openmp-2021.4.0-h06a4308_3561\n jpeg pkgs/main/linux-64::jpeg-9d-h7f8727e_0\n lcms2 pkgs/main/linux-64::lcms2-2.12-h3be6417_0\n libboost pkgs/main/linux-64::libboost-1.73.0-h3ff78a5_11\n libpng pkgs/main/linux-64::libpng-1.6.37-hbc83047_0\n libtiff pkgs/main/linux-64::libtiff-4.2.0-h85742a9_0\n libuuid pkgs/main/linux-64::libuuid-1.0.3-h7f8727e_2\n libwebp pkgs/main/linux-64::libwebp-1.2.0-h89dd481_0\n libwebp-base pkgs/main/linux-64::libwebp-base-1.2.0-h27cfd23_0\n libxcb pkgs/main/linux-64::libxcb-1.14-h7b6447c_0\n libxml2 pkgs/main/linux-64::libxml2-2.9.12-h03d6c58_0\n lz4-c pkgs/main/linux-64::lz4-c-1.9.3-h295c915_1\n mkl pkgs/main/linux-64::mkl-2021.4.0-h06a4308_640\n mkl-service pkgs/main/linux-64::mkl-service-2.4.0-py37h7f8727e_0\n mkl_fft pkgs/main/linux-64::mkl_fft-1.3.1-py37hd3c417c_0\n mkl_random pkgs/main/linux-64::mkl_random-1.2.2-py37h51133e4_0\n numexpr pkgs/main/linux-64::numexpr-2.8.1-py37h6abb31d_0\n numpy pkgs/main/linux-64::numpy-1.21.2-py37h20f2e39_0\n numpy-base pkgs/main/linux-64::numpy-base-1.21.2-py37h79a1101_0\n olefile pkgs/main/linux-64::olefile-0.46-py37_0\n packaging pkgs/main/noarch::packaging-21.3-pyhd3eb1b0_0\n pandas pkgs/main/linux-64::pandas-1.3.5-py37h8c16a72_0\n pcre pkgs/main/linux-64::pcre-8.45-h295c915_0\n pillow pkgs/main/linux-64::pillow-8.4.0-py37h5aabda8_0\n pixman pkgs/main/linux-64::pixman-0.40.0-h7f8727e_1\n py-boost pkgs/main/linux-64::py-boost-1.73.0-py37ha9443f7_11\n pyparsing pkgs/main/noarch::pyparsing-3.0.4-pyhd3eb1b0_0\n python-dateutil pkgs/main/noarch::python-dateutil-2.8.2-pyhd3eb1b0_0\n pytz pkgs/main/noarch::pytz-2021.3-pyhd3eb1b0_0\n rdkit rdkit/linux-64::rdkit-2020.09.1.0-py37hd50e099_1\n zstd pkgs/main/linux-64::zstd-1.4.9-haebb681_0\n\nThe following packages will be UPDATED:\n\n ca-certificates 2021.7.5-h06a4308_1 --> 2021.10.26-h06a4308_2\n certifi 2021.5.30-py39h06a4308_0 --> 2021.10.8-py37h06a4308_2\n cffi 1.14.6-py39h400218f_0 --> 1.15.0-py37hd667e15_1\n conda 4.10.3-py39h06a4308_0 --> 4.11.0-py37h06a4308_0\n cryptography 3.4.7-py39hd23ed53_0 --> 36.0.0-py37h9ce1e76_0\n openssl 1.1.1k-h27cfd23_0 --> 1.1.1m-h7f8727e_0\n pip 21.1.3-py39h06a4308_0 --> 21.2.2-py37h06a4308_0\n pysocks 1.7.1-py39h06a4308_0 --> 1.7.1-py37_1\n setuptools 52.0.0-py39h06a4308_0 --> 58.0.4-py37h06a4308_0\n\nThe following packages will be DOWNGRADED:\n\n brotlipy 0.7.0-py39h27cfd23_1003 --> 0.7.0-py37h27cfd23_1003\n chardet 4.0.0-py39h06a4308_1003 --> 4.0.0-py37h06a4308_1003\n conda-package-han~ 1.7.3-py39h27cfd23_1 --> 1.7.3-py37h27cfd23_1\n pycosat 0.6.3-py39h27cfd23_0 --> 0.6.3-py37h27cfd23_0\n python 3.9.5-h12debd9_4 --> 3.7.11-h12debd9_0\n ruamel_yaml 0.15.100-py39h27cfd23_0 --> 0.15.100-py37h27cfd23_0\n\n\nPreparing transaction: ...working... done\nVerifying transaction: ...working... done\nExecuting transaction: ...working... done\n" ], [ "!python -c \"import site; print (site.getsitepackages())\"", "['/usr/local/lib/python3.7/site-packages']\n" ], [ "import sys\nimport pprint\npprint.pprint(sys.path)\nsys.path.append('/usr/local/lib/python3.7/site-packages/')\nfrom rdkit import rdBase\nprint(rdBase.rdkitVersion)", "['',\n '/content',\n '/env/python',\n '/usr/lib/python37.zip',\n '/usr/lib/python3.7',\n '/usr/lib/python3.7/lib-dynload',\n '/usr/local/lib/python3.7/dist-packages',\n '/usr/lib/python3/dist-packages',\n '/usr/local/lib/python3.7/dist-packages/IPython/extensions',\n '/root/.ipython']\n2020.09.1\n" ], [ "import pandas as pd\nimport matplotlib.pyplot as plt\nfrom pandas.plotting import scatter_matrix\n\ndata = pd.read_csv(\"/content/drive/MyDrive/Colab Notebooks/lab/BDE_t.csv\")\ndata.head()", "_____no_output_____" ], [ "from rdkit import Chem\nfrom rdkit.Chem import Draw\nmols = [Chem.MolFromSmiles(smile) for smile in data['SMILES']]", "_____no_output_____" ], [ "from rdkit.Chem import AllChem\nimport numpy as np\nfingerprints = []\nsafe = []\nfor mol_idx, mol in enumerate(mols):\n try:\n fingerprint = [x for x in AllChem.GetMorganFingerprintAsBitVect(mol, 2, 2**11)]\n fingerprints.append(fingerprint)\n safe.append(mol_idx)\n except:\n print(\"Error\", mol_idx)\n continue\nfingerprints = np.array(fingerprints)\nprint(fingerprints.shape)\npd.DataFrame(fingerprints).head()", "(3567, 2048)\n" ], [ "data_y=data\ndata_y=data_y.drop(\"SMILES\", axis=1)\ndata_y=data_y.drop(\"IP\", axis=1)\ndata_y=data_y.drop(\"PDE\", axis=1)\ndata_y=data_y.drop(\"BDE\", axis=1)\ndata_y=data_y.drop(\"ETE\", axis=1)", "_____no_output_____" ], [ "data_y", "_____no_output_____" ], [ "X=fingerprints\nX = np.array(X, dtype = np.float32)\nnp.save(\"/content/drive/MyDrive/Colab Notebooks/lab/X_TCI_2_PA.npy\", X)\nprint(X)", "[[0. 1. 0. ... 0. 0. 0.]\n [0. 1. 0. ... 0. 0. 0.]\n [0. 1. 0. ... 0. 0. 0.]\n ...\n [0. 0. 0. ... 0. 0. 0.]\n [0. 0. 0. ... 0. 0. 0.]\n [0. 0. 0. ... 0. 0. 0.]]\n" ], [ "\nfrom rdkit import Chem\nfrom rdkit.Chem.Draw import IPythonConsole\nimport numpy as np\nfrom sklearn.preprocessing import StandardScaler\nfrom sklearn import model_selection\nfrom keras.models import Sequential\nfrom keras.layers import Dense, Activation\nY=data_y\n\nY = np.array(Y, dtype = np.float32)\n\nAve=Y.mean(axis=0)\nStd=Y.std(ddof=1,axis=0)\n\nY=(Y - Ave) / Std\nnp.save(\"/content/drive/MyDrive/Colab Notebooks/lab/Y_TCI_2_PA.npy\", Y)\nprint(Y)\nprint(Ave)\nprint(Std)", "[[-0.44942018]\n [-0.4200536 ]\n [-0.79657614]\n ...\n [ 1.6851792 ]\n [ 2.1974144 ]\n [-0.84641206]]\n[272.58444]\n[17.763977]\n" ], [ "X_train, X_test, y_train, y_test = model_selection.train_test_split(X, Y, test_size=0.20, random_state=42)\nnp.save(\"/content/drive/MyDrive/Colab Notebooks/lab/X_train_TCI_2_PA.npy\", X_train)\nnp.save(\"/content/drive/MyDrive/Colab Notebooks/lab/X_test_TCI_2_PA.npy\", X_test)\nnp.save(\"/content/drive/MyDrive/Colab Notebooks/lab/y_train_TCI_2_PA.npy\", y_train)\nnp.save(\"/content/drive/MyDrive/Colab Notebooks/lab/y_test_TCI_2_PA.npy\", y_test)", "_____no_output_____" ], [ "import numpy as np\nfrom sklearn import model_selection\nfrom sklearn.metrics import mean_squared_error\nfrom sklearn.metrics import r2_score\nfrom keras.models import Sequential\nfrom keras.layers import Activation, Dense\nfrom keras.wrappers.scikit_learn import KerasRegressor\nfrom tensorflow import keras\nfrom tensorflow.keras.optimizers import Adam\nX_train = np.load(\"/content/drive/MyDrive/Colab Notebooks/lab/X_train_TCI_2_PA.npy\")\ny_train = np.load(\"/content/drive/MyDrive/Colab Notebooks/lab/y_train_TCI_2_PA.npy\")\nX_test = np.load(\"/content/drive/MyDrive/Colab Notebooks/lab/X_test_TCI_2_PA.npy\")\ny_test = np.load(\"/content/drive/MyDrive/Colab Notebooks/lab/y_test_TCI_2_PA.npy\")\n\nfrom keras.layers import Input, Dense, Dropout\nfrom keras.models import Model\n\ninputs = Input(shape=(2**11,))\n\nx = Dense(2**10, activation='tanh', kernel_regularizer=keras.regularizers.l2(0.001))(inputs)\nx = Dropout(.2)(x)\nx = Dense(2**5, activation='tanh', kernel_regularizer=keras.regularizers.l2(0.001))(x)\nx = Dropout(.2)(x)\nx = Dense(2**2, activation='tanh', kernel_regularizer=keras.regularizers.l2(0.001))(x)\nx = Dropout(.2)(x)\npredictions = Dense(1, activation=\"linear\")(x)\n\nearly_stop = keras.callbacks.EarlyStopping(monitor='val_loss', patience=10)\n\nmodel = Model(inputs=inputs, outputs=predictions)\n\nmodel.compile(Adam(lr=1e-3), loss=\"mean_squared_error\")\n\nhistory = model.fit(X_train, y_train, batch_size=128, epochs=2000, validation_data=(X_test, y_test),callbacks=[early_stop])", "/usr/local/lib/python3.7/dist-packages/keras/optimizer_v2/adam.py:105: UserWarning: The `lr` argument is deprecated, use `learning_rate` instead.\n super(Adam, self).__init__(name, **kwargs)\n" ], [ "import matplotlib.pyplot as plt\nprint('2乗誤差の平均',model.evaluate(X_test, y_test))\n\ntrain_acc = history.history['loss']\ntest_acc = history.history['val_loss']\n\nx = np.arange(len(train_acc))\nplt.plot(x, train_acc, label = 'train mse')\nplt.plot(x, test_acc, label = 'test mse')\nplt.ylim(0, 1)\nplt.legend()", "23/23 [==============================] - 0s 5ms/step - loss: 0.3172\n2乗誤差の平均 0.3171910047531128\n" ], [ "y_pred = model.predict(X_test)\ny_train_pred = model.predict(X_train)\n\nYYY=y_test.transpose()\nyyy=y_pred.transpose()\nplt.scatter(YYY[0], yyy[0], c='r', marker='s',label=\"ALL\")\nplt.legend()\nplt.plot([-3,3],[-3,3])", "_____no_output_____" ], [ "model.save('/content/drive/MyDrive/Colab Notebooks/lab/model_TCI_2_PA.h5')", "_____no_output_____" ], [ "from keras.layers import Input, Dense, Dropout\nfrom keras.models import Model\nfrom keras.models import load_model\nmodel =load_model('/content/drive/MyDrive/Colab Notebooks/lab/model_TCI_2_PA.h5')", "_____no_output_____" ], [ "import numpy as np\nfrom sklearn import model_selection\nfrom sklearn.metrics import mean_squared_error\nfrom sklearn.metrics import r2_score\nfrom keras.models import Sequential\nfrom keras.layers import Activation, Dense\nfrom keras.wrappers.scikit_learn import KerasRegressor\nfrom tensorflow import keras\nfrom tensorflow.keras.optimizers import Adam\nX_train = np.load(\"/content/drive/MyDrive/Colab Notebooks/lab/X_train_TCI_2_PA.npy\")\ny_train = np.load(\"/content/drive/MyDrive/Colab Notebooks/lab/y_train_TCI_2_PA.npy\")\nX_test = np.load(\"/content/drive/MyDrive/Colab Notebooks/lab/X_test_TCI_2_PA.npy\")\ny_test = np.load(\"/content/drive/MyDrive/Colab Notebooks/lab/y_test_TCI_2_PA.npy\")", "_____no_output_____" ], [ "import matplotlib.pyplot as plt\ny_pred = model.predict(X_test)\ny_pred_t = model.predict(X_train)\n#Y_test=y_test*Std+Ave\n#Y_pred=y_pred*Std+Ave\nplt.scatter(y_test, y_pred, c='r', marker='s',label=\"ALL\")\nplt.scatter(y_train, y_pred_t, c='b', marker='s',label=\"ALL\")\nplt.legend()\nplt.plot([-4,4],[-4,4])", "_____no_output_____" ], [ "Ave=Ave.reshape(1,1)\nStd=Std.reshape(1,1)\n\nYYY=y_test*Std+Ave\nyyy=y_pred*Std+Ave\nYYYY=y_train*Std+Ave\nyyyy=y_pred_t*Std+Ave\n\nYYY=YYY.transpose()\nyyy=yyy.transpose()\nYYYY=YYYY.transpose()\nyyyy=yyyy.transpose()", "_____no_output_____" ], [ "plt.scatter(YYY[0], yyy[0], c='r', marker='s',label=\"test\")\nplt.scatter(YYYY[0], yyyy[0], c='b', marker='s',label=\"train\")\nplt.xlabel('calculated Value (kcal/mol)')\nplt.ylabel('predicted Value (kcal/mol)')\nplt.rcParams['xtick.direction'] = 'out'\nplt.rcParams['ytick.direction'] = 'out'\nplt.plot([60,110],[60,110])\nplt.legend()\n\nplt.savefig('/content/drive/MyDrive/Colab Notebooks/lab/PA.png', transparent=True)", "_____no_output_____" ], [ "import tensorflow as tf\nfrom tensorflow import keras\nfrom keras.models import load_model\nbase_model =load_model('/content/drive/MyDrive/Colab Notebooks/lab/model_TCI_2_PA.h5')", "_____no_output_____" ], [ "base_model.summary()", "Model: \"model\"\n_________________________________________________________________\n Layer (type) Output Shape Param # \n=================================================================\n input_1 (InputLayer) [(None, 2048)] 0 \n \n dense (Dense) (None, 1024) 2098176 \n \n dropout (Dropout) (None, 1024) 0 \n \n dense_1 (Dense) (None, 32) 32800 \n \n dropout_1 (Dropout) (None, 32) 0 \n \n dense_2 (Dense) (None, 4) 132 \n \n dropout_2 (Dropout) (None, 4) 0 \n \n dense_3 (Dense) (None, 1) 5 \n \n=================================================================\nTotal params: 2,131,113\nTrainable params: 2,131,113\nNon-trainable params: 0\n_________________________________________________________________\n" ], [ "import pandas as pd\nimport matplotlib.pyplot as plt\nfrom pandas.plotting import scatter_matrix\n\ndt = pd.read_csv(\"/content/drive/MyDrive/Colab Notebooks/lab/H_ORAC.csv\")\ndt.head()\n\nfrom rdkit import Chem\nfrom rdkit.Chem import Draw\nmols = [Chem.MolFromSmiles(smile) for smile in dt['SMILES']]", "_____no_output_____" ], [ "from rdkit.Chem import AllChem\nimport numpy as np\nfingerprints = []\nsafe = []\nfor mol_idx, mol in enumerate(mols):\n try:\n fingerprint = [x for x in AllChem.GetMorganFingerprintAsBitVect(mol, 2, 2**11)]\n fingerprints.append(fingerprint)\n safe.append(mol_idx)\n except:\n print(\"Error\", mol_idx)\n continue\nfingerprints = np.array(fingerprints)\nprint(fingerprints.shape)\npd.DataFrame(fingerprints).head()", "(70, 2048)\n" ], [ "X2=fingerprints\nX2 = np.array(X2, dtype = np.float32)\nnp.save(\"/content/drive/MyDrive/Colab Notebooks/lab/X_2d_ORAC.npy\", X2)\nprint(X2)", "[[0. 0. 0. ... 0. 0. 0.]\n [0. 0. 0. ... 0. 0. 0.]\n [0. 0. 0. ... 0. 0. 0.]\n ...\n [0. 1. 0. ... 0. 0. 0.]\n [0. 1. 0. ... 0. 0. 0.]\n [0. 0. 0. ... 0. 0. 0.]]\n" ], [ "\nfrom rdkit import Chem\nfrom rdkit.Chem.Draw import IPythonConsole\nimport numpy as np\nfrom sklearn.preprocessing import StandardScaler\nfrom sklearn import model_selection\nfrom keras.models import Sequential\nfrom keras.layers import Dense, Activation, Input, Dropout\nY2=[y for y in dt['H_ORAC']]\n\nY2 = np.array(Y2, dtype = np.float32)\n\nAve=Y2.mean()\nStd=Y2.std(ddof=1)\n\nY2=(Y2 - Ave )/ Std\n\nnp.save(\"/content/drive/MyDrive/Colab Notebooks/lab/Y_2d_ORAC.npy\", Y2)\nprint(Y2)", "[-0.5434038 -0.1577336 -0.45749253 -0.2201499 -0.53286856 0.21718943\n 0.517175 0.3992576 0.96539456 -0.06432275 4.858367 0.71591735\n 0.86406463 -0.3478773 0.12073161 0.54230046 0.29801032 0.2953972\n 0.8034653 -0.92614245 -0.235798 0.6265299 -0.80607396 -0.5244356\n -0.61346394 -0.62144846 -0.61873865 -0.31283364 -1.0512096 -1.0507561\n -1.0488456 0.26605016 0.05566086 -0.92539907 -0.03932004 -0.99443716\n -0.28263178 -0.77516484 -0.559626 -0.6305682 -0.45141768 -0.3848276\n -0.42543682 -0.5228378 2.7227314 -0.78508466 -1.0499865 -0.02662846\n -0.7496714 1.2863741 1.7340133 -0.47654933 1.5442306 -0.13065857\n -0.8465681 0.31667623 -0.7957791 -0.66367626 0.6136809 1.1964297\n 1.6072878 1.611745 0.02019791 -0.14124039 -0.8637594 -0.9887596\n 0.99622774 -0.4986037 -0.91868335 -0.13418403]\n" ], [ "from keras.models import Model\nfrom tensorflow import keras\nimport numpy as np\nfrom sklearn import model_selection\nfrom sklearn.metrics import mean_squared_error\nfrom sklearn.metrics import r2_score\nfrom keras.layers import Activation, Dense, Dropout\nfrom keras.wrappers.scikit_learn import KerasRegressor\nfrom tensorflow.keras.optimizers import Adam\n\nxx = base_model.output\nxx = Dense(2**5, activation='tanh', kernel_regularizer=keras.regularizers.l2(0.001))(xx)\nxx = Dropout(.2)(xx)\nxx = Dense(2**2, activation='tanh', kernel_regularizer=keras.regularizers.l2(0.001))(xx)\nxx = Dropout(.2)(xx)\n\npredictions = Dense(1, activation='linear')(xx)\n\nmodel2 = Model(inputs=base_model.input, outputs=predictions)\nmodel2.layers.pop(-9)\n\nmodel2.compile(Adam(lr=1e-3), loss=\"mean_squared_error\")\n\nmodel2.summary()", "Model: \"model_1\"\n_________________________________________________________________\n Layer (type) Output Shape Param # \n=================================================================\n input_1 (InputLayer) [(None, 2048)] 0 \n \n dense (Dense) (None, 1024) 2098176 \n \n dropout (Dropout) (None, 1024) 0 \n \n dense_1 (Dense) (None, 32) 32800 \n \n dropout_1 (Dropout) (None, 32) 0 \n \n dense_2 (Dense) (None, 4) 132 \n \n dropout_2 (Dropout) (None, 4) 0 \n \n dense_3 (Dense) (None, 1) 5 \n \n dense_4 (Dense) (None, 32) 64 \n \n dropout_3 (Dropout) (None, 32) 0 \n \n dense_5 (Dense) (None, 4) 132 \n \n dropout_4 (Dropout) (None, 4) 0 \n \n dense_6 (Dense) (None, 1) 5 \n \n=================================================================\nTotal params: 2,131,314\nTrainable params: 2,131,314\nNon-trainable params: 0\n_________________________________________________________________\n" ], [ "X2_train, X2_test, y2_train, y2_test = model_selection.train_test_split(X2, Y2, test_size=0.20, random_state=20211208)\nnp.save(\"/content/drive/MyDrive/Colab Notebooks/lab/X2_train.npy\", X2_train)\nnp.save(\"/content/drive/MyDrive/Colab Notebooks/lab/X2_test.npy\", X2_test)\nnp.save(\"/content/drive/MyDrive/Colab Notebooks/lab/y2_train.npy\", y2_train)\nnp.save(\"/content/drive/MyDrive/Colab Notebooks/lab/y2_test.npy\", y2_test)\nearly_stop = keras.callbacks.EarlyStopping(monitor='val_loss', patience=10)\n\nhistory2 = model2.fit(X2_train, y2_train, batch_size=128, epochs=2000, validation_data=(X2_test, y2_test),callbacks=[early_stop])\nimport matplotlib.pyplot as plt\nprint('2乗誤差の平均',model2.evaluate(X2_test, y2_test))\n\ntrain_acc = history2.history['loss']\ntest_acc = history2.history['val_loss']\n\nx = np.arange(len(train_acc))\nplt.plot(x, train_acc, label = 'train mse')\nplt.plot(x, test_acc, label = 'val mse')\nplt.ylim(0, 3)\nplt.legend()", "Epoch 1/2000\n1/1 [==============================] - 1s 915ms/step - loss: 1.2907 - val_loss: 0.5613\nEpoch 2/2000\n1/1 [==============================] - 0s 31ms/step - loss: 1.2758 - val_loss: 0.5125\nEpoch 3/2000\n1/1 [==============================] - 0s 37ms/step - loss: 1.1736 - val_loss: 0.4677\nEpoch 4/2000\n1/1 [==============================] - 0s 50ms/step - loss: 1.0804 - val_loss: 0.4314\nEpoch 5/2000\n1/1 [==============================] - 0s 29ms/step - loss: 1.0882 - val_loss: 0.3815\nEpoch 6/2000\n1/1 [==============================] - 0s 33ms/step - loss: 0.9999 - val_loss: 0.3417\nEpoch 7/2000\n1/1 [==============================] - 0s 30ms/step - loss: 1.0112 - val_loss: 0.2949\nEpoch 8/2000\n1/1 [==============================] - 0s 35ms/step - loss: 0.9857 - val_loss: 0.2556\nEpoch 9/2000\n1/1 [==============================] - 0s 45ms/step - loss: 0.9533 - val_loss: 0.2202\nEpoch 10/2000\n1/1 [==============================] - 0s 28ms/step - loss: 0.7460 - val_loss: 0.2054\nEpoch 11/2000\n1/1 [==============================] - 0s 29ms/step - loss: 0.7974 - val_loss: 0.2107\nEpoch 12/2000\n1/1 [==============================] - 0s 30ms/step - loss: 0.8484 - val_loss: 0.2287\nEpoch 13/2000\n1/1 [==============================] - 0s 33ms/step - loss: 0.7074 - val_loss: 0.2658\nEpoch 14/2000\n1/1 [==============================] - 0s 29ms/step - loss: 0.8329 - val_loss: 0.2851\nEpoch 15/2000\n1/1 [==============================] - 0s 27ms/step - loss: 0.7247 - val_loss: 0.3062\nEpoch 16/2000\n1/1 [==============================] - 0s 26ms/step - loss: 0.5538 - val_loss: 0.3363\nEpoch 17/2000\n1/1 [==============================] - 0s 30ms/step - loss: 0.9251 - val_loss: 0.3391\nEpoch 18/2000\n1/1 [==============================] - 0s 26ms/step - loss: 0.6963 - val_loss: 0.3446\nEpoch 19/2000\n1/1 [==============================] - 0s 27ms/step - loss: 0.6832 - val_loss: 0.3702\nEpoch 20/2000\n1/1 [==============================] - 0s 27ms/step - loss: 0.5793 - val_loss: 0.4033\n1/1 [==============================] - 0s 19ms/step - loss: 0.4033\n2乗誤差の平均 0.403340607881546\n" ], [ "\nY2_pred_t = model2.predict(X2_test)\nY2_train_pred_t = model2.predict(X2_train)\nplt.scatter(y2_test, Y2_pred_t, c='r', marker='s',label=\"test\")\nplt.scatter(y2_train, Y2_train_pred_t, c='b', marker='s',label=\"train\")\nplt.legend()\nplt.plot([-1,5],[-1,5])", "_____no_output_____" ], [ "y3_test=y2_test*Std+Ave\nY3_train_pred_t=Y2_train_pred_t*Std+Ave\ny3_train=y2_train*Std+Ave\nY3_pred_t=Y2_pred_t*Std+Ave\nplt.scatter(y3_test, Y3_pred_t, c='r', marker='s',label=\"test\")\nplt.scatter(y3_train, Y3_train_pred_t, c='b', marker='s',label=\"train\")\nplt.xlabel('Ground Truth (mol(TE)/mol)')\nplt.ylabel('predicted Value (mol(TE)/mol)')\nplt.rcParams['xtick.direction'] = 'out'\nplt.rcParams['ytick.direction'] = 'out'\nplt.plot([0,12],[0,12])\nplt.legend()", "_____no_output_____" ], [ "from sklearn.metrics import mean_squared_error\nfrom math import sqrt\nprint(mean_squared_error(y3_test, Y3_pred_t))\nprint(mean_squared_error(y3_train, Y3_train_pred_t))", "1.3653971\n1.5537233\n" ], [ "", "_____no_output_____" ], [ "", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "<a href=\"https://colab.research.google.com/github/akhila-sakinala/akhila-sakinala.github.io/blob/master/makaan_webscraping.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>", "_____no_output_____" ], [ "## Import packages", "_____no_output_____" ] ], [ [ "# import packages\nimport requests\nfrom bs4 import BeautifulSoup\n", "_____no_output_____" ], [ "url = \"https://www.makaan.com/hyderabad-residential-property/rent-property-in-hyderabad-city\"\n\nresponse = requests.get(url)\nsoup = BeautifulSoup(response.text,\"html.parser\")", "_____no_output_____" ], [ "s_tag = soup.find_all('span',attrs={'class' : 'seller-type'})", "_____no_output_____" ], [ "for each_owner in s_tag:\n print(each_owner.text)", "OWNER\nOWNER\nOWNER\nOWNER\nOWNER\nOWNER\nOWNER\nOWNER\nOWNER\nOWNER\nOWNER\nOWNER\nOWNER\nOWNER\nOWNER\nOWNER\nOWNER\nOWNER\nOWNER\nOWNER\n" ], [ "s_val = soup.find_all('a',attrs={'class' : 'typelink'})", "_____no_output_____" ], [ "for price in s_val:\n print(price.span.text)", "2 \n1 \n2 \n3 \n2 \n1 \n3 \n1 \n2 \n4 \n2 \n2 \n2 \n6 \n3 \n3 \n1 \n2 \n1 \n2 \n" ], [ "", "_____no_output_____" ] ] ]

[ "markdown", "code" ]

[ [ "markdown", "markdown" ], [ "code", "code", "code", "code", "code", "code", "code" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "# Enter State Farm", "_____no_output_____" ] ], [ [ "from theano.sandbox import cuda\ncuda.use('gpu0')", "_____no_output_____" ], [ "%matplotlib inline\nfrom __future__ import print_function, division\npath = \"data/state/\"\n#path = \"data/state/sample/\"\nimport utils; reload(utils)\nfrom utils import *\nfrom IPython.display import FileLink", "Using Theano backend.\n" ], [ "batch_size=64", "_____no_output_____" ] ], [ [ "## Setup batches", "_____no_output_____" ] ], [ [ "batches = get_batches(path+'train', batch_size=batch_size)\nval_batches = get_batches(path+'valid', batch_size=batch_size*2, shuffle=False)", "Found 22424 images belonging to 10 classes.\n" ], [ "(val_classes, trn_classes, val_labels, trn_labels, \n val_filenames, filenames, test_filenames) = get_classes(path)", "Found 18946 images belonging to 10 classes.\nFound 3478 images belonging to 10 classes.\nFound 79726 images belonging to 1 classes.\n" ] ], [ [ "Rather than using batches, we could just import all the data into an array to save some processing time. (In most examples I'm using the batches, however - just because that's how I happened to start out.)", "_____no_output_____" ] ], [ [ "trn = get_data(path+'train')\nval = get_data(path+'valid')", "Found 18946 images belonging to 10 classes.\n" ], [ "save_array(path+'results/val.dat', val)\nsave_array(path+'results/trn.dat', trn)", "_____no_output_____" ], [ "val = load_array(path+'results/val.dat')\ntrn = load_array(path+'results/trn.dat')", "_____no_output_____" ] ], [ [ "## Re-run sample experiments on full dataset", "_____no_output_____" ], [ "We should find that everything that worked on the sample (see statefarm-sample.ipynb), works on the full dataset too. Only better! Because now we have more data. So let's see how they go - the models in this section are exact copies of the sample notebook models.", "_____no_output_____" ], [ "### Single conv layer", "_____no_output_____" ] ], [ [ "def conv1(batches):\n model = Sequential([\n BatchNormalization(axis=1, input_shape=(3,224,224)),\n Convolution2D(32,3,3, activation='relu'),\n BatchNormalization(axis=1),\n MaxPooling2D((3,3)),\n Convolution2D(64,3,3, activation='relu'),\n BatchNormalization(axis=1),\n MaxPooling2D((3,3)),\n Flatten(),\n Dense(200, activation='relu'),\n BatchNormalization(),\n Dense(10, activation='softmax')\n ])\n\n model.compile(Adam(lr=1e-4), loss='categorical_crossentropy', metrics=['accuracy'])\n model.fit_generator(batches, batches.nb_sample, nb_epoch=2, validation_data=val_batches, \n nb_val_samples=val_batches.nb_sample)\n model.optimizer.lr = 0.001\n model.fit_generator(batches, batches.nb_sample, nb_epoch=4, validation_data=val_batches, \n nb_val_samples=val_batches.nb_sample)\n return model", "_____no_output_____" ], [ "model = conv1(batches)", "Epoch 1/2\n18946/18946 [==============================] - 114s - loss: 0.2273 - acc: 0.9405 - val_loss: 2.4946 - val_acc: 0.2826\nEpoch 2/2\n18946/18946 [==============================] - 114s - loss: 0.0120 - acc: 0.9990 - val_loss: 1.5872 - val_acc: 0.5253\nEpoch 1/4\n18946/18946 [==============================] - 114s - loss: 0.0093 - acc: 0.9992 - val_loss: 1.4836 - val_acc: 0.5825\nEpoch 2/4\n18946/18946 [==============================] - 114s - loss: 0.0032 - acc: 1.0000 - val_loss: 1.3142 - val_acc: 0.6162\nEpoch 3/4\n18946/18946 [==============================] - 114s - loss: 0.0035 - acc: 0.9996 - val_loss: 1.5061 - val_acc: 0.5771\nEpoch 4/4\n18946/18946 [==============================] - 114s - loss: 0.0036 - acc: 0.9997 - val_loss: 1.4528 - val_acc: 0.5808\n" ] ], [ [ "Interestingly, with no regularization or augmentation we're getting some reasonable results from our simple convolutional model. So with augmentation, we hopefully will see some very good results.", "_____no_output_____" ], [ "### Data augmentation", "_____no_output_____" ] ], [ [ "gen_t = image.ImageDataGenerator(rotation_range=15, height_shift_range=0.05, \n shear_range=0.1, channel_shift_range=20, width_shift_range=0.1)\nbatches = get_batches(path+'train', gen_t, batch_size=batch_size)", "Found 18946 images belonging to 10 classes.\n" ], [ "model = conv1(batches)", "Epoch 1/2\n18946/18946 [==============================] - 114s - loss: 1.2804 - acc: 0.5891 - val_loss: 2.0614 - val_acc: 0.3407\nEpoch 2/2\n18946/18946 [==============================] - 114s - loss: 0.6716 - acc: 0.7916 - val_loss: 1.3377 - val_acc: 0.6208\nEpoch 1/4\n18946/18946 [==============================] - 115s - loss: 0.4787 - acc: 0.8594 - val_loss: 1.2230 - val_acc: 0.6228\nEpoch 2/4\n18946/18946 [==============================] - 114s - loss: 0.3724 - acc: 0.8931 - val_loss: 1.3030 - val_acc: 0.6282\nEpoch 3/4\n18946/18946 [==============================] - 114s - loss: 0.3086 - acc: 0.9162 - val_loss: 1.1986 - val_acc: 0.7119\nEpoch 4/4\n18946/18946 [==============================] - 114s - loss: 0.2612 - acc: 0.9283 - val_loss: 1.4794 - val_acc: 0.5799\n" ], [ "model.optimizer.lr = 0.0001\nmodel.fit_generator(batches, batches.nb_sample, nb_epoch=15, validation_data=val_batches, \n nb_val_samples=val_batches.nb_sample)", "Epoch 1/15\n18946/18946 [==============================] - 114s - loss: 0.2391 - acc: 0.9361 - val_loss: 1.2511 - val_acc: 0.6886\nEpoch 2/15\n18946/18946 [==============================] - 114s - loss: 0.2075 - acc: 0.9430 - val_loss: 1.1327 - val_acc: 0.7294\nEpoch 3/15\n18946/18946 [==============================] - 114s - loss: 0.1800 - acc: 0.9529 - val_loss: 1.1099 - val_acc: 0.7294\nEpoch 4/15\n18946/18946 [==============================] - 114s - loss: 0.1675 - acc: 0.9557 - val_loss: 1.0660 - val_acc: 0.7363\nEpoch 5/15\n18946/18946 [==============================] - 114s - loss: 0.1432 - acc: 0.9625 - val_loss: 1.1585 - val_acc: 0.7073\nEpoch 6/15\n18946/18946 [==============================] - 114s - loss: 0.1358 - acc: 0.9627 - val_loss: 1.1389 - val_acc: 0.6947\nEpoch 7/15\n18946/18946 [==============================] - 114s - loss: 0.1283 - acc: 0.9665 - val_loss: 1.1329 - val_acc: 0.7369\nEpoch 8/15\n18946/18946 [==============================] - 114s - loss: 0.1180 - acc: 0.9686 - val_loss: 1.1817 - val_acc: 0.7194\nEpoch 9/15\n18946/18946 [==============================] - 114s - loss: 0.1137 - acc: 0.9704 - val_loss: 1.0923 - val_acc: 0.7142\nEpoch 10/15\n18946/18946 [==============================] - 114s - loss: 0.1076 - acc: 0.9720 - val_loss: 1.0983 - val_acc: 0.7358\nEpoch 11/15\n18946/18946 [==============================] - 114s - loss: 0.1032 - acc: 0.9736 - val_loss: 1.0206 - val_acc: 0.7458\nEpoch 12/15\n18946/18946 [==============================] - 114s - loss: 0.0956 - acc: 0.9740 - val_loss: 0.9039 - val_acc: 0.7809\nEpoch 13/15\n18946/18946 [==============================] - 114s - loss: 0.0962 - acc: 0.9740 - val_loss: 1.3386 - val_acc: 0.6587\nEpoch 14/15\n18946/18946 [==============================] - 114s - loss: 0.0892 - acc: 0.9777 - val_loss: 1.1150 - val_acc: 0.7470\nEpoch 15/15\n18946/18946 [==============================] - 114s - loss: 0.0886 - acc: 0.9773 - val_loss: 1.9190 - val_acc: 0.5802\n" ] ], [ [ "I'm shocked by *how* good these results are! We're regularly seeing 75-80% accuracy on the validation set, which puts us into the top third or better of the competition. With such a simple model and no dropout or semi-supervised learning, this really speaks to the power of this approach to data augmentation.", "_____no_output_____" ], [ "### Four conv/pooling pairs + dropout", "_____no_output_____" ], [ "Unfortunately, the results are still very unstable - the validation accuracy jumps from epoch to epoch. Perhaps a deeper model with some dropout would help.", "_____no_output_____" ] ], [ [ "gen_t = image.ImageDataGenerator(rotation_range=15, height_shift_range=0.05, \n shear_range=0.1, channel_shift_range=20, width_shift_range=0.1)\nbatches = get_batches(path+'train', gen_t, batch_size=batch_size)", "Found 18946 images belonging to 10 classes.\n" ], [ "model = Sequential([\n BatchNormalization(axis=1, input_shape=(3,224,224)),\n Convolution2D(32,3,3, activation='relu'),\n BatchNormalization(axis=1),\n MaxPooling2D(),\n Convolution2D(64,3,3, activation='relu'),\n BatchNormalization(axis=1),\n MaxPooling2D(),\n Convolution2D(128,3,3, activation='relu'),\n BatchNormalization(axis=1),\n MaxPooling2D(),\n Flatten(),\n Dense(200, activation='relu'),\n BatchNormalization(),\n Dropout(0.5),\n Dense(200, activation='relu'),\n BatchNormalization(),\n Dropout(0.5),\n Dense(10, activation='softmax')\n ])", "_____no_output_____" ], [ "model.compile(Adam(lr=10e-5), loss='categorical_crossentropy', metrics=['accuracy'])", "_____no_output_____" ], [ "model.fit_generator(batches, batches.nb_sample, nb_epoch=2, validation_data=val_batches, \n nb_val_samples=val_batches.nb_sample)", "Epoch 1/2\n18946/18946 [==============================] - 159s - loss: 2.6578 - acc: 0.2492 - val_loss: 1.8681 - val_acc: 0.3844\nEpoch 2/2\n18946/18946 [==============================] - 158s - loss: 1.8098 - acc: 0.4334 - val_loss: 1.3152 - val_acc: 0.5670\n" ], [ "model.optimizer.lr=0.001", "_____no_output_____" ], [ "model.fit_generator(batches, batches.nb_sample, nb_epoch=10, validation_data=val_batches, \n nb_val_samples=val_batches.nb_sample)", "Epoch 1/10\n18946/18946 [==============================] - 159s - loss: 1.4232 - acc: 0.5405 - val_loss: 1.0877 - val_acc: 0.6452\nEpoch 2/10\n18946/18946 [==============================] - 159s - loss: 1.1155 - acc: 0.6346 - val_loss: 1.2730 - val_acc: 0.6878\nEpoch 3/10\n18946/18946 [==============================] - 159s - loss: 0.9043 - acc: 0.7025 - val_loss: 1.1393 - val_acc: 0.6354\nEpoch 4/10\n18946/18946 [==============================] - 159s - loss: 0.7444 - acc: 0.7529 - val_loss: 1.1037 - val_acc: 0.7087\nEpoch 5/10\n18946/18946 [==============================] - 159s - loss: 0.6299 - acc: 0.7955 - val_loss: 0.9123 - val_acc: 0.7455\nEpoch 6/10\n18946/18946 [==============================] - 159s - loss: 0.5220 - acc: 0.8275 - val_loss: 1.0418 - val_acc: 0.7484\nEpoch 7/10\n18946/18946 [==============================] - 159s - loss: 0.4686 - acc: 0.8495 - val_loss: 1.2907 - val_acc: 0.6599\nEpoch 8/10\n18946/18946 [==============================] - 159s - loss: 0.4190 - acc: 0.8653 - val_loss: 1.1321 - val_acc: 0.6906\nEpoch 9/10\n18946/18946 [==============================] - 159s - loss: 0.3735 - acc: 0.8802 - val_loss: 1.1235 - val_acc: 0.7458\nEpoch 10/10\n18946/18946 [==============================] - 159s - loss: 0.3226 - acc: 0.8969 - val_loss: 1.2040 - val_acc: 0.7343\n" ], [ "model.optimizer.lr=0.00001", "_____no_output_____" ], [ "model.fit_generator(batches, batches.nb_sample, nb_epoch=10, validation_data=val_batches, \n nb_val_samples=val_batches.nb_sample)", "Epoch 1/10\n18946/18946 [==============================] - 159s - loss: 0.3183 - acc: 0.8976 - val_loss: 1.0359 - val_acc: 0.7688\nEpoch 2/10\n18946/18946 [==============================] - 158s - loss: 0.2788 - acc: 0.9109 - val_loss: 1.5806 - val_acc: 0.6705\nEpoch 3/10\n18946/18946 [==============================] - 158s - loss: 0.2810 - acc: 0.9124 - val_loss: 0.9836 - val_acc: 0.7887\nEpoch 4/10\n18946/18946 [==============================] - 158s - loss: 0.2403 - acc: 0.9244 - val_loss: 1.1832 - val_acc: 0.7493\nEpoch 5/10\n18946/18946 [==============================] - 159s - loss: 0.2195 - acc: 0.9303 - val_loss: 1.1524 - val_acc: 0.7510\nEpoch 6/10\n18946/18946 [==============================] - 159s - loss: 0.2085 - acc: 0.9359 - val_loss: 1.2245 - val_acc: 0.7415\nEpoch 7/10\n18946/18946 [==============================] - 158s - loss: 0.1961 - acc: 0.9399 - val_loss: 1.1232 - val_acc: 0.7654\nEpoch 8/10\n18946/18946 [==============================] - 158s - loss: 0.1851 - acc: 0.9416 - val_loss: 1.0956 - val_acc: 0.6892\nEpoch 9/10\n18946/18946 [==============================] - 158s - loss: 0.1798 - acc: 0.9451 - val_loss: 1.0586 - val_acc: 0.7740\nEpoch 10/10\n18946/18946 [==============================] - 159s - loss: 0.1669 - acc: 0.9471 - val_loss: 1.4633 - val_acc: 0.6656\n" ] ], [ [ "This is looking quite a bit better - the accuracy is similar, but the stability is higher. There's still some way to go however...", "_____no_output_____" ], [ "### Imagenet conv features", "_____no_output_____" ], [ "Since we have so little data, and it is similar to imagenet images (full color photos), using pre-trained VGG weights is likely to be helpful - in fact it seems likely that we won't need to fine-tune the convolutional layer weights much, if at all. So we can pre-compute the output of the last convolutional layer, as we did in lesson 3 when we experimented with dropout. (However this means that we can't use full data augmentation, since we can't pre-compute something that changes every image.)", "_____no_output_____" ] ], [ [ "vgg = Vgg16()\nmodel=vgg.model\nlast_conv_idx = [i for i,l in enumerate(model.layers) if type(l) is Convolution2D][-1]\nconv_layers = model.layers[:last_conv_idx+1]", "_____no_output_____" ], [ "conv_model = Sequential(conv_layers)", "_____no_output_____" ], [ "(val_classes, trn_classes, val_labels, trn_labels, \n val_filenames, filenames, test_filenames) = get_classes(path)", "Found 18946 images belonging to 10 classes.\nFound 3478 images belonging to 10 classes.\nFound 79726 images belonging to 1 classes.\n" ], [ "conv_feat = conv_model.predict_generator(batches, batches.nb_sample)\nconv_val_feat = conv_model.predict_generator(val_batches, val_batches.nb_sample)\nconv_test_feat = conv_model.predict_generator(test_batches, test_batches.nb_sample)", "_____no_output_____" ], [ "save_array(path+'results/conv_val_feat.dat', conv_val_feat)\nsave_array(path+'results/conv_test_feat.dat', conv_test_feat)\nsave_array(path+'results/conv_feat.dat', conv_feat)", "_____no_output_____" ], [ "conv_feat = load_array(path+'results/conv_feat.dat')\nconv_val_feat = load_array(path+'results/conv_val_feat.dat')\nconv_val_feat.shape", "_____no_output_____" ] ], [ [ "### Batchnorm dense layers on pretrained conv layers", "_____no_output_____" ], [ "Since we've pre-computed the output of the last convolutional layer, we need to create a network that takes that as input, and predicts our 10 classes. Let's try using a simplified version of VGG's dense layers.", "_____no_output_____" ] ], [ [ "def get_bn_layers(p):\n return [\n MaxPooling2D(input_shape=conv_layers[-1].output_shape[1:]),\n Flatten(),\n Dropout(p/2),\n Dense(128, activation='relu'),\n BatchNormalization(),\n Dropout(p/2),\n Dense(128, activation='relu'),\n BatchNormalization(),\n Dropout(p),\n Dense(10, activation='softmax')\n ]", "_____no_output_____" ], [ "p=0.8", "_____no_output_____" ], [ "bn_model = Sequential(get_bn_layers(p))\nbn_model.compile(Adam(lr=0.001), loss='categorical_crossentropy', metrics=['accuracy'])", "_____no_output_____" ], [ "bn_model.fit(conv_feat, trn_labels, batch_size=batch_size, nb_epoch=1, \n validation_data=(conv_val_feat, val_labels))", "Train on 18946 samples, validate on 3478 samples\nEpoch 1/1\n18946/18946 [==============================] - 3s - loss: 1.5894 - acc: 0.5625 - val_loss: 0.7031 - val_acc: 0.7522\n" ], [ "bn_model.optimizer.lr=0.01", "_____no_output_____" ], [ "bn_model.fit(conv_feat, trn_labels, batch_size=batch_size, nb_epoch=2, \n validation_data=(conv_val_feat, val_labels))", "Train on 18946 samples, validate on 3478 samples\nEpoch 1/2\n18946/18946 [==============================] - 3s - loss: 0.2870 - acc: 0.9109 - val_loss: 0.7728 - val_acc: 0.7683\nEpoch 2/2\n18946/18946 [==============================] - 3s - loss: 0.1422 - acc: 0.9594 - val_loss: 0.7576 - val_acc: 0.7936\n" ], [ "bn_model.save_weights(path+'models/conv8.h5')", "_____no_output_____" ] ], [ [ "Looking good! Let's try pre-computing 5 epochs worth of augmented data, so we can experiment with combining dropout and augmentation on the pre-trained model.", "_____no_output_____" ], [ "### Pre-computed data augmentation + dropout", "_____no_output_____" ], [ "We'll use our usual data augmentation parameters:", "_____no_output_____" ] ], [ [ "gen_t = image.ImageDataGenerator(rotation_range=15, height_shift_range=0.05, \n shear_range=0.1, channel_shift_range=20, width_shift_range=0.1)\nda_batches = get_batches(path+'train', gen_t, batch_size=batch_size, shuffle=False)", "Found 18946 images belonging to 10 classes.\n" ] ], [ [ "We use those to create a dataset of convolutional features 5x bigger than the training set.", "_____no_output_____" ] ], [ [ "da_conv_feat = conv_model.predict_generator(da_batches, da_batches.nb_sample*5)", "_____no_output_____" ], [ "save_array(path+'results/da_conv_feat2.dat', da_conv_feat)", "_____no_output_____" ], [ "da_conv_feat = load_array(path+'results/da_conv_feat2.dat')", "_____no_output_____" ] ], [ [ "Let's include the real training data as well in its non-augmented form.", "_____no_output_____" ] ], [ [ "da_conv_feat = np.concatenate([da_conv_feat, conv_feat])", "_____no_output_____" ] ], [ [ "Since we've now got a dataset 6x bigger than before, we'll need to copy our labels 6 times too.", "_____no_output_____" ] ], [ [ "da_trn_labels = np.concatenate([trn_labels]*6)", "_____no_output_____" ] ], [ [ "Based on some experiments the previous model works well, with bigger dense layers.", "_____no_output_____" ] ], [ [ "def get_bn_da_layers(p):\n return [\n MaxPooling2D(input_shape=conv_layers[-1].output_shape[1:]),\n Flatten(),\n Dropout(p),\n Dense(256, activation='relu'),\n BatchNormalization(),\n Dropout(p),\n Dense(256, activation='relu'),\n BatchNormalization(),\n Dropout(p),\n Dense(10, activation='softmax')\n ]", "_____no_output_____" ], [ "p=0.8", "_____no_output_____" ], [ "bn_model = Sequential(get_bn_da_layers(p))\nbn_model.compile(Adam(lr=0.001), loss='categorical_crossentropy', metrics=['accuracy'])", "_____no_output_____" ] ], [ [ "Now we can train the model as usual, with pre-computed augmented data.", "_____no_output_____" ] ], [ [ "bn_model.fit(da_conv_feat, da_trn_labels, batch_size=batch_size, nb_epoch=1, \n validation_data=(conv_val_feat, val_labels))", "Train on 113676 samples, validate on 3478 samples\nEpoch 1/1\n113676/113676 [==============================] - 16s - loss: 1.5848 - acc: 0.5068 - val_loss: 0.6340 - val_acc: 0.8131\n" ], [ "bn_model.optimizer.lr=0.01", "_____no_output_____" ], [ "bn_model.fit(da_conv_feat, da_trn_labels, batch_size=batch_size, nb_epoch=4, \n validation_data=(conv_val_feat, val_labels))", "Train on 113676 samples, validate on 3478 samples\nEpoch 1/4\n113676/113676 [==============================] - 16s - loss: 0.6652 - acc: 0.7785 - val_loss: 0.6343 - val_acc: 0.8082\nEpoch 2/4\n113676/113676 [==============================] - 16s - loss: 0.5247 - acc: 0.8318 - val_loss: 0.6951 - val_acc: 0.8085\nEpoch 3/4\n113676/113676 [==============================] - 16s - loss: 0.4553 - acc: 0.8544 - val_loss: 0.6067 - val_acc: 0.8189\nEpoch 4/4\n113676/113676 [==============================] - 16s - loss: 0.4127 - acc: 0.8686 - val_loss: 0.7701 - val_acc: 0.7915\n" ], [ "bn_model.optimizer.lr=0.0001", "_____no_output_____" ], [ "bn_model.fit(da_conv_feat, da_trn_labels, batch_size=batch_size, nb_epoch=4, \n validation_data=(conv_val_feat, val_labels))", "Train on 113676 samples, validate on 3478 samples\nEpoch 1/4\n113676/113676 [==============================] - 16s - loss: 0.3837 - acc: 0.8775 - val_loss: 0.6904 - val_acc: 0.8197\nEpoch 2/4\n113676/113676 [==============================] - 16s - loss: 0.3576 - acc: 0.8872 - val_loss: 0.6593 - val_acc: 0.8209\nEpoch 3/4\n113676/113676 [==============================] - 16s - loss: 0.3384 - acc: 0.8939 - val_loss: 0.7057 - val_acc: 0.8085\nEpoch 4/4\n113676/113676 [==============================] - 16s - loss: 0.3254 - acc: 0.8977 - val_loss: 0.6867 - val_acc: 0.8128\n" ] ], [ [ "Looks good - let's save those weights.", "_____no_output_____" ] ], [ [ "bn_model.save_weights(path+'models/da_conv8_1.h5')", "_____no_output_____" ] ], [ [ "### Pseudo labeling", "_____no_output_____" ], [ "We're going to try using a combination of [pseudo labeling](http://deeplearning.net/wp-content/uploads/2013/03/pseudo_label_final.pdf) and [knowledge distillation](https://arxiv.org/abs/1503.02531) to allow us to use unlabeled data (i.e. do semi-supervised learning). For our initial experiment we'll use the validation set as the unlabeled data, so that we can see that it is working without using the test set. At a later date we'll try using the test set.", "_____no_output_____" ], [ "To do this, we simply calculate the predictions of our model...", "_____no_output_____" ] ], [ [ "val_pseudo = bn_model.predict(conv_val_feat, batch_size=batch_size)", "_____no_output_____" ] ], [ [ "...concatenate them with our training labels...", "_____no_output_____" ] ], [ [ "comb_pseudo = np.concatenate([da_trn_labels, val_pseudo])", "_____no_output_____" ], [ "comb_feat = np.concatenate([da_conv_feat, conv_val_feat])", "_____no_output_____" ] ], [ [ "...and fine-tune our model using that data.", "_____no_output_____" ] ], [ [ "bn_model.load_weights(path+'models/da_conv8_1.h5')", "_____no_output_____" ], [ "bn_model.fit(comb_feat, comb_pseudo, batch_size=batch_size, nb_epoch=1, \n validation_data=(conv_val_feat, val_labels))", "Train on 117154 samples, validate on 3478 samples\nEpoch 1/1\n117154/117154 [==============================] - 17s - loss: 0.3412 - acc: 0.8948 - val_loss: 0.7653 - val_acc: 0.8191\n" ], [ "bn_model.fit(comb_feat, comb_pseudo, batch_size=batch_size, nb_epoch=4, \n validation_data=(conv_val_feat, val_labels))", "Train on 117154 samples, validate on 3478 samples\nEpoch 1/4\n117154/117154 [==============================] - 17s - loss: 0.3237 - acc: 0.9008 - val_loss: 0.7536 - val_acc: 0.8229\nEpoch 2/4\n117154/117154 [==============================] - 17s - loss: 0.3076 - acc: 0.9050 - val_loss: 0.7572 - val_acc: 0.8235\nEpoch 3/4\n117154/117154 [==============================] - 17s - loss: 0.2984 - acc: 0.9085 - val_loss: 0.7852 - val_acc: 0.8269\nEpoch 4/4\n117154/117154 [==============================] - 17s - loss: 0.2902 - acc: 0.9117 - val_loss: 0.7630 - val_acc: 0.8263\n" ], [ "bn_model.optimizer.lr=0.00001", "_____no_output_____" ], [ "bn_model.fit(comb_feat, comb_pseudo, batch_size=batch_size, nb_epoch=4, \n validation_data=(conv_val_feat, val_labels))", "Train on 117154 samples, validate on 3478 samples\nEpoch 1/4\n117154/117154 [==============================] - 17s - loss: 0.2837 - acc: 0.9134 - val_loss: 0.7901 - val_acc: 0.8200\nEpoch 2/4\n117154/117154 [==============================] - 17s - loss: 0.2760 - acc: 0.9155 - val_loss: 0.7648 - val_acc: 0.8275\nEpoch 3/4\n117154/117154 [==============================] - 17s - loss: 0.2723 - acc: 0.9183 - val_loss: 0.7382 - val_acc: 0.8358\nEpoch 4/4\n117154/117154 [==============================] - 17s - loss: 0.2657 - acc: 0.9191 - val_loss: 0.7227 - val_acc: 0.8329\n" ] ], [ [ "That's a distinct improvement - even although the validation set isn't very big. This looks encouraging for when we try this on the test set.", "_____no_output_____" ] ], [ [ "bn_model.save_weights(path+'models/bn-ps8.h5')", "_____no_output_____" ] ], [ [ "### Submit", "_____no_output_____" ], [ "We'll find a good clipping amount using the validation set, prior to submitting.", "_____no_output_____" ] ], [ [ "def do_clip(arr, mx): return np.clip(arr, (1-mx)/9, mx)", "_____no_output_____" ], [ "keras.metrics.categorical_crossentropy(val_labels, do_clip(val_preds, 0.93)).eval()", "_____no_output_____" ], [ "conv_test_feat = load_array(path+'results/conv_test_feat.dat')", "_____no_output_____" ], [ "preds = bn_model.predict(conv_test_feat, batch_size=batch_size*2)", "_____no_output_____" ], [ "subm = do_clip(preds,0.93)", "_____no_output_____" ], [ "subm_name = path+'results/subm.gz'", "_____no_output_____" ], [ "classes = sorted(batches.class_indices, key=batches.class_indices.get)", "_____no_output_____" ], [ "submission = pd.DataFrame(subm, columns=classes)\nsubmission.insert(0, 'img', [a[4:] for a in test_filenames])\nsubmission.head()", "_____no_output_____" ], [ "submission.to_csv(subm_name, index=False, compression='gzip')", "_____no_output_____" ], [ "FileLink(subm_name)", "_____no_output_____" ] ], [ [ "This gets 0.534 on the leaderboard.", "_____no_output_____" ], [ "## The \"things that didn't really work\" section", "_____no_output_____" ], [ "You can safely ignore everything from here on, because they didn't really help.", "_____no_output_____" ], [ "### Finetune some conv layers too", "_____no_output_____" ] ], [ [ "for l in get_bn_layers(p): conv_model.add(l)", "_____no_output_____" ], [ "for l1,l2 in zip(bn_model.layers, conv_model.layers[last_conv_idx+1:]):\n l2.set_weights(l1.get_weights())", "_____no_output_____" ], [ "for l in conv_model.layers: l.trainable =False", "_____no_output_____" ], [ "for l in conv_model.layers[last_conv_idx+1:]: l.trainable =True", "_____no_output_____" ], [ "comb = np.concatenate([trn, val])", "_____no_output_____" ], [ "gen_t = image.ImageDataGenerator(rotation_range=8, height_shift_range=0.04, \n shear_range=0.03, channel_shift_range=10, width_shift_range=0.08)", "_____no_output_____" ], [ "batches = gen_t.flow(comb, comb_pseudo, batch_size=batch_size)", "_____no_output_____" ], [ "val_batches = get_batches(path+'valid', batch_size=batch_size*2, shuffle=False)", "Found 3478 images belonging to 10 classes.\n" ], [ "conv_model.compile(Adam(lr=0.00001), loss='categorical_crossentropy', metrics=['accuracy'])", "_____no_output_____" ], [ "conv_model.fit_generator(batches, batches.N, nb_epoch=1, validation_data=val_batches, \n nb_val_samples=val_batches.N)", "Epoch 1/1\n22400/22424 [============================>.] - ETA: 0s - loss: 0.4348 - acc: 0.9200" ], [ "conv_model.optimizer.lr = 0.0001", "_____no_output_____" ], [ "conv_model.fit_generator(batches, batches.N, nb_epoch=3, validation_data=val_batches, \n nb_val_samples=val_batches.N)", "_____no_output_____" ], [ "for l in conv_model.layers[16:]: l.trainable =True", "_____no_output_____" ], [ "conv_model.optimizer.lr = 0.00001", "_____no_output_____" ], [ "conv_model.fit_generator(batches, batches.N, nb_epoch=8, validation_data=val_batches, \n nb_val_samples=val_batches.N)", "_____no_output_____" ], [ "conv_model.save_weights(path+'models/conv8_ps.h5')", "_____no_output_____" ], [ "conv_model.load_weights(path+'models/conv8_da.h5')", "_____no_output_____" ], [ "val_pseudo = conv_model.predict(val, batch_size=batch_size*2)", "_____no_output_____" ], [ "save_array(path+'models/pseudo8_da.dat', val_pseudo)", "_____no_output_____" ] ], [ [ "### Ensembling", "_____no_output_____" ] ], [ [ "drivers_ds = pd.read_csv(path+'driver_imgs_list.csv')\ndrivers_ds.head()", "_____no_output_____" ], [ "img2driver = drivers_ds.set_index('img')['subject'].to_dict()", "_____no_output_____" ], [ "driver2imgs = {k: g[\"img\"].tolist() \n for k,g in drivers_ds[['subject', 'img']].groupby(\"subject\")}", "_____no_output_____" ], [ "def get_idx(driver_list):\n return [i for i,f in enumerate(filenames) if img2driver[f[3:]] in driver_list]", "_____no_output_____" ], [ "drivers = driver2imgs.keys()", "_____no_output_____" ], [ "rnd_drivers = np.random.permutation(drivers)", "_____no_output_____" ], [ "ds1 = rnd_drivers[:len(rnd_drivers)//2]\nds2 = rnd_drivers[len(rnd_drivers)//2:]", "_____no_output_____" ], [ "models=[fit_conv([d]) for d in drivers]\nmodels=[m for m in models if m is not None]", "_____no_output_____" ], [ "all_preds = np.stack([m.predict(conv_test_feat, batch_size=128) for m in models])\navg_preds = all_preds.mean(axis=0)\navg_preds = avg_preds/np.expand_dims(avg_preds.sum(axis=1), 1)", "_____no_output_____" ], [ "keras.metrics.categorical_crossentropy(val_labels, np.clip(avg_val_preds,0.01,0.99)).eval()", "_____no_output_____" ], [ "keras.metrics.categorical_accuracy(val_labels, np.clip(avg_val_preds,0.01,0.99)).eval()", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown", "markdown", "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code", "code", "code" ], [ "markdown", "markdown", "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code" ], [ "markdown", "markdown", "markdown" ], [ "code", "code", "code", "code", "code", "code" ], [ "markdown", "markdown" ], [ "code", "code", "code", "code", "code", "code", "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ], [ "markdown", "markdown", "markdown", "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "#Built-in function", "_____no_output_____" ] ], [ [ "print(abs(4.5))\nprint(abs(-4.5))", "4.5\n4.5\n" ] ], [ [ "all - Return True if all elements of the iterable are true (or if the iterable is empty)", "_____no_output_____" ] ], [ [ "a = [2, 3, 4, 5]\nb = [0]\nc = []\nd = [2, 3, 4, 0]\n\nprint(all(a))\nprint(all(b))\nprint(all(c))\nprint(all(d))\n", "True\nFalse\nTrue\nFalse\n" ] ], [ [ "any - Return True if any element of the iterable is true. If the iterable is empty, return False. Equivalent to:", "_____no_output_____" ] ], [ [ "a = [2, 3, 4, 5]\nb = [0]\nc = []\nd = [2, 3, 4, 0]\n\nprint(any(a))\nprint(any(b))\nprint(any(c))\nprint(any(d))", "True\nFalse\nFalse\nTrue\n" ] ], [ [ "ascii", "_____no_output_____" ] ], [ [ "print(ascii(1111))", "1111\n" ], [ "dir()", "_____no_output_____" ], [ "dir(struct)", "_____no_output_____" ], [ "meta = {}\nmeta[item] = []", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Using notebooks we will explore the process of trainign and consuming a model\n\nFirst let's load some packages to maniupulate images\n", "_____no_output_____" ] ], [ [ "#r \"nuget:SixLabors.ImageSharp,1.0.2\"", "_____no_output_____" ] ], [ [ "## Get images\n\nWe can download images from the web, let's create some helper function", "_____no_output_____" ] ], [ [ "using SixLabors.ImageSharp;\nusing SixLabors.ImageSharp.PixelFormats;\nusing System.Net.Http;\n\nImage GetImage(string url)\n{\n var client = new HttpClient();\n var image = client.GetByteArrayAsync(url).Result;\n return Image.Load(image);\n}", "_____no_output_____" ], [ "var image = GetImage(\"https://user-images.githubusercontent.com/2546640/56708992-deee8780-66ec-11e9-9991-eb85abb1d10a.png\");\nimage", "_____no_output_____" ] ], [ [ "## it would be better to see the image, let's use the foramtter api", "_____no_output_____" ] ], [ [ "using System.IO;\nusing SixLabors.ImageSharp.Formats.Png;\nusing Microsoft.DotNet.Interactive.Formatting;\n\nFormatter.Register<Image>((image, writer) =>\n{\n var id = Guid.NewGuid().ToString(\"N\");\n using var stream = new MemoryStream();\n image.Save(stream, new PngEncoder());\n stream.Flush();\n var data = stream.ToArray();\n var imageSource = $\"data:image/png;base64, {Convert.ToBase64String(data)}\";\n PocketView imgTag = PocketViewTags.img[id: id, src: imageSource, height: image.Height, width: image.Width]();\n writer.Write(imgTag);\n}, HtmlFormatter.MimeType);\n", "_____no_output_____" ], [ "image", "_____no_output_____" ] ], [ [ "Good but something smaller would be better", "_____no_output_____" ] ], [ [ "using SixLabors.ImageSharp.Processing;\n\nImage Reduce(Image source, int maxSize = 300){\n var max = Math.Max(source.Width, source.Height);\n var ratio = ((double)(maxSize)) / max;\n return source.Clone(c => c.Resize((int)(source.Width * ratio), (int)(source.Height * ratio)));\n}\n", "_____no_output_____" ], [ "Reduce(image)", "_____no_output_____" ] ], [ [ "Better, now I am interested in bayblade, let's display some", "_____no_output_____" ] ], [ [ "var urls = new string[]{\n \"https://cdn.shopify.com/s/files/1/0016/0674/6186/products/B154_1_1024x1024.jpg?v=1573909023\",\n \"https://i.ytimg.com/vi/yUH2QeluaIU/maxresdefault.jpg\",\n \"https://www.biggerbids.com/members/images/29371/public/8065336_-DSC5628-32467-26524-.jpg\",\n \"https://i.ytimg.com/vi/BT4SwVmnqqQ/maxresdefault.jpg\",\n \"https://cdn.shopify.com/s/files/1/0016/0674/6186/products/B160covercopy2_1200x1200.jpg?v=1585425105\",\n \"https://animeukiyo.com/wp-content/uploads/2020/05/king-helios-zone-1B-1140x570.jpg\",\n \"https://http2.mlstatic.com/beyblade-burn-phoenix-ice-blue-90wf-takara-tomy-frete-pac-D_NQ_NP_19415-MLB20171031427_092014-F.jpg\"\n};\n\nvar beyBlades = urls.Select(url => new { Image = Reduce(GetImage(url))});\n", "_____no_output_____" ], [ "beyBlades", "_____no_output_____" ] ], [ [ "## Enter lobe\nWe will now use lobe and it's .NET Bindings to developa model to classify those images. Let's start lobe and have a look first, then we will proceed with loading the pacakges we need.", "_____no_output_____" ] ], [ [ "#r \"nuget:lobe\"\n#r \"nuget:lobe.ImageSharp\"\n\nusing lobe;\nusing lobe.ImageSharp;", "_____no_output_____" ] ], [ [ "Lobe can be accessed via web api let's use that for fast loops", "_____no_output_____" ] ], [ [ "#r \"nuget:lobe.Http\"\n\nusing lobe.Http;\n", "_____no_output_____" ], [ "var beyblades_start = new Uri(\"http://localhost:38100/predict/3af915df-14b7-4834-afbd-6615deca4e26\");\nvar beyblades = new Uri(\"http://localhost:38100/predict/f56e1050-391e-4cd6-9bb9-ff74dc4d84f5\");\nvar beyblades_2 = new Uri(\"http://localhost:38100/predict/f56e1050-391e-4cd6-9bb9-ff74dc4d84f5\");\nvar beyblades_3 = new Uri(\"http://localhost:38100/predict/a3271b3a-f63b-4c00-9304-beda43375284\");\nvar beyblade_remote = new Uri(\"http://lobe-diego.ngrok.io/predict/2a6a3005-a8cc-4bc1-a71a-a0fe85f258bb\");\n\nvar httpClassifier = new LobeClient(beyblades_3);\n\nhttpClassifier.Classify(beyBlades.First().Image.CloneAs<Rgb24>())\n\n", "_____no_output_____" ], [ "var imageSources = urls.Select(url => Reduce(GetImage(url),800).CloneAs<Rgb24>()).ToList();", "_____no_output_____" ], [ "var classifications = imageSources.Select((img) => {\n var cls = httpClassifier.Classify(img);\n return new {\n Image = Reduce(img),\n Label = cls.Prediction.Label,\n Confidence = cls.Prediction.Confidence\n };\n});\n\nclassifications", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "- Install TPOT within Anancoda - https://anaconda.org/conda-forge/tpot\n- More Details - https://epistasislab.github.io/tpot/using/\n- GitHub - https://github.com/EpistasisLab/tpot/", "_____no_output_____" ] ], [ [ "import pandas as pd\nfrom tpot import TPOTClassifier\nfrom sklearn.model_selection import train_test_split", "_____no_output_____" ], [ "train = pd.read_csv(\"excel_full_train.csv\")\ntest = pd.read_csv(\"excel_test.csv\")", "_____no_output_____" ], [ "X = train.drop(['PassengerId','Survived'], axis = 1)\ny = train['Survived']", "_____no_output_____" ], [ "#### Use Test Train Split to divide into train and test\nimport numpy as np\nfrom sklearn.model_selection import train_test_split\nX_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.30, random_state=21)", "_____no_output_____" ] ], [ [ "### Run AutoML with TPOT", "_____no_output_____" ], [ "**Verbose** \n- How much information TPOT communicates while it is running.\n- 0 = none, 1 = minimal, 2 = high, 3 = all.\n- A setting of 2 or higher will add a progress bar during the optimization procedure.", "_____no_output_____" ] ], [ [ "#Set max time for 10 Minutes\ntpot = TPOTClassifier(verbosity=2, max_time_mins=1)", "_____no_output_____" ], [ "tpot.fit(X_train, y_train)", "_____no_output_____" ], [ "print(f'Train : {tpot.score(X_test, y_test):.3f}')\nprint(f'Test : {tpot.score(X_train, y_train):.3f}')", "Train : 0.840\nTest : 0.867\n" ] ], [ [ "### Export Best Pipeline", "_____no_output_____" ] ], [ [ "tpot.export('Auto_ML_TPOT_titanic_pipeline2.py')", "_____no_output_____" ] ], [ [ "### Prediction of Test", "_____no_output_____" ] ], [ [ "sub_test = test.drop(['PassengerId'], axis = 1)", "_____no_output_____" ], [ "sub_test_pred = tpot.predict(sub_test).astype(int)", "_____no_output_____" ], [ "AllSub = pd.DataFrame({ 'PassengerId': test['PassengerId'],\n 'Survived' : sub_test_pred\n \n})\n\nAllSub.to_csv(\"Auto_ML_TPOT_Titanic_Solution.csv\", index = False)", "_____no_output_____" ], [ "#Kaggle LB Score - 0.78468", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code", "code", "code", "code" ], [ "markdown", "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code", "code" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "# Display Sample Records", "_____no_output_____" ] ], [ [ "import gzip\nimport json\nimport re\nimport os\nimport sys\nimport numpy as np\nimport pandas as pd", "_____no_output_____" ] ], [ [ "**Specify your directory here:**", "_____no_output_____" ] ], [ [ "DIR = './data'", "_____no_output_____" ] ], [ [ "**This function shows how to load datasets**", "_____no_output_____" ] ], [ [ "def load_data(file_name, head = 500):\n '''\n Given a *.json.gz file, returns a list of dictionaries,\n optionally can select the first n records\n '''\n count = 0\n data = []\n with gzip.open(file_name) as fin:\n for l in fin:\n d = json.loads(l)\n count += 1\n data.append(d)\n \n # break if reaches the 500th line\n if (head is not None) and (count >= head):\n break\n return data", "_____no_output_____" ] ], [ [ "**Load and display sample records of books/authors/works/series**", "_____no_output_____" ] ], [ [ "poetry = load_data(os.path.join(DIR, 'goodreads_books_poetry.json.gz'))\n\n# books = load_data(os.path.join(DIR, 'goodreads_books.json.gz'))\n# authors = load_data(os.path.join(DIR, 'goodreads_book_authors.json.gz'))\n# works = load_data(os.path.join(DIR, 'goodreads_book_works.json.gz'))\n# series = load_data(os.path.join(DIR, 'goodreads_book_series.json.gz'))", "_____no_output_____" ], [ "len(poetry)", "_____no_output_____" ], [ "poetry[0]", "_____no_output_____" ], [ "# print(' == sample record (books) ==')\n# display(np.random.choice(books))\n# print(' == sample record (authors) ==')\n# display(np.random.choice(authors))\n# print(' == sample record (works) ==')\n# display(np.random.choice(works))\n# print(' == sample record (series) ==')\n# display(np.random.choice(series))", "_____no_output_____" ] ], [ [ "**Load and display sample records of user-book interactions (shelves)**", "_____no_output_____" ] ], [ [ "interactions = load_data(os.path.join(DIR, 'goodreads_interactions_poetry.json.gz'))\nnp.random.choice(interactions)", "_____no_output_____" ] ], [ [ "**Load and display sample records of book reviews**", "_____no_output_____" ] ], [ [ "reviews = load_data(os.path.join(DIR, 'goodreads_reviews_poetry.json.gz'))\nnp.random.choice(reviews)", "_____no_output_____" ] ], [ [ "**Load and display sample records of book reviews (with spoiler tags)**", "_____no_output_____" ] ], [ [ "spoilers = load_data(os.path.join(DIR, 'goodreads_reviews_spoiler.json.gz'))\nnp.random.choice([s for s in spoilers if s['has_spoiler']])", "_____no_output_____" ], [ "# spoilers = load_data(os.path.join(DIR, 'goodreads_reviews_spoiler_raw.json.gz'))\n# np.random.choice([s for s in spoilers if 'view spoiler' in s['review_text']])", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import matplotlib.pyplot as plt\nimport seaborn as sns\nsns.set_context('poster')\n\nfrom pyplr.CIE import *\n", "_____no_output_____" ], [ "p = sns.color_palette(\"tab10\")\nsns.palplot(p)", "_____no_output_____" ], [ "fig, axs = plt.subplots(1,3, figsize=(12,4))\n\nget_CIE_1924_photopic_vl(asdf=True).plot(ax=axs[0], legend=False)\naxs[0].set_title('CIE 1924 V($\\lambda$)')\n\nget_CIE_CMF(asdf=True).plot(\n ax=axs[1], \n color={'X':p[0],'Y':p[2],'Z':p[3]}, \n legend=False)\naxs[1].set_title('CIE CMFs')\n\nget_CIES026(asdf=True).plot(\n ax=axs[2], \n color={'S':p[0],'M':p[2],'L':p[3],'Rods':p[7],'Mel':p[9]}, \n legend=False)\naxs[2].set_title('CIE S 026')\n\nfor ax in axs:\n #ax.set_xticks([])\n #ax.set_yticks([])\n pass\nplt.tight_layout()\nfig.savefig('../img/CIE.tiff', dpi=300)", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code" ] ]

[ "CC0-1.0" ]

[ "CC0-1.0" ]

[ "CC0-1.0" ]

[ [ [ "## Introduction to the Interstellar Medium\n### Jonathan Williams", "_____no_output_____" ], [ "### Figure 4.2: Extinction curve", "_____no_output_____" ], [ "#### uses extcurve_s16.py and cubicspline.py from https://faun.rc.fas.harvard.edu/eschlafly/apored/extcurve.html", "_____no_output_____" ] ], [ [ "import numpy as np\nimport matplotlib.pyplot as plt\n%matplotlib inline", "_____no_output_____" ], [ "import extcurve_s16", "_____no_output_____" ], [ "fig = plt.figure(figsize=(6,4))\nax1 = fig.add_subplot(1,1,1)\n#ax1.set_xlabel('$\\lambda$ (nm)', fontsize=16)\nax1.set_xlabel('$\\lambda\\ (\\mu m)$', fontsize=16)\nax1.set_ylabel('$A(\\lambda)/A_K$', fontsize=16)\n#ax1.set_xlim(350,2500)\nax1.set_xlim(0.350,2.500)\n#ax1.set_ylim(0,1.3)\nax1.set_ylim(0,15)\n\nlam = np.linspace(500,2500, 100)\nlam_ext = np.linspace(350,500, 10)\noir = np.nonzero((lam > 500) & (lam < 3000))\n\nec = extcurve_s16.extcurve(0.0)\n#f = ec(5420)/ec(5510)\nf = ec(5420)/ec(21900)\nx = np.log10(lam)\ny = f*ec(10*lam)\nw = 500/lam[oir]\nw = lam[oir] * 0 + 1\na,b = np.polyfit(x[oir],np.log10(y[oir]),1,w=w)\nprint(\"R_V = 3.3 power law index = {0:4.2f}\".format(a))\n#ax1.plot(10**x,10**(a*x+b),'r-')\n#ax1.plot(lam, y, 'k-', lw=2)\n#ax1.plot(lam_ext, f*ec(10*lam_ext), 'k:', lw=2)\nax1.plot(lam/1000, y, 'k-', lw=2)\nax1.plot(lam_ext/1000, f*ec(10*lam_ext), 'k:', lw=2)\n\nec = extcurve_s16.extcurve(0.04)\n#f = ec(5420)/ec(5510)\nf = ec(5420)/ec(21900)\ny = f*ec(10*lam)\na,b = np.polyfit(x[oir],np.log10(y[oir]),1,w=w)\nprint(\"R_V = 3.6 power law index = {0:4.2f}\".format(a))\n#ax1.plot(10**x,10**(a*x+b),'r-')\n#ax1.plot(lam, f*ec(10*lam), 'k-', lw=0.5)\n#ax1.plot(lam_ext, f*ec(10*lam_ext), 'k:', lw=0.5)\nax1.plot(lam/1000, f*ec(10*lam), 'k-', lw=0.5)\nax1.plot(lam_ext/1000, f*ec(10*lam_ext), 'k:', lw=0.5)\n\nec = extcurve_s16.extcurve(-0.04)\n#f = ec(5420)/ec(5510)\nf = ec(5420)/ec(21900)\ny = f*ec(10*lam)\na,b = np.polyfit(x[oir],np.log10(y[oir]),1,w=w)\nprint(\"R_V = 3.0 power law index = {0:4.2f}\".format(a))\nax1.plot(lam/1000, f*ec(10*lam), 'k-', lw=0.5)\nax1.plot(lam_ext/1000, f*ec(10*lam_ext), 'k:', lw=0.5)\n\nylab = 3.7\nplt.text(.445, ylab, 'B', fontsize=16, ha='center')\nplt.text(.551, ylab, 'V', fontsize=16, ha='center')\nplt.text(.656, ylab, 'R', fontsize=16, ha='center')\nplt.text(.806, ylab, 'I', fontsize=16, ha='center')\nplt.text(1.220,ylab, 'J', fontsize=16, ha='center')\nplt.text(1.630,ylab, 'H', fontsize=16, ha='center')\nplt.text(2.190,ylab, 'K', fontsize=16, ha='center')\n\nplt.savefig('extinction.pdf')", "R_V = 3.3 power law index = -1.76\nR_V = 3.6 power law index = -1.72\nR_V = 3.0 power law index = -1.79\n" ] ] ]

[ "markdown", "code" ]

[ [ "markdown", "markdown", "markdown" ], [ "code", "code", "code" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "<a href=\"https://colab.research.google.com/github/archana1822/DMDW/blob/main/ASSIGNMENT_7.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>", "_____no_output_____" ] ], [ [ " dataset = [\n ['i1','i2','i5'],\n ['i2', 'i4'],\n ['i2', 'i3'],\n ['i1', 'i2', 'i4'],\n ['i1', 'i3'],\n ['i2', 'i3'],\n ['i1','i3'],\n ['i1', 'i2', 'i3','i5'],\n ['i1', 'i2','i3']]\n \n", "_____no_output_____" ], [ "import pandas as pd\nfrom mlxtend.preprocessing import TransactionEncoder\n\nte = TransactionEncoder()\nte_ary = te.fit(dataset).transform(dataset)\ndf = pd.DataFrame(te_ary, columns=te.columns_)\ndf", "_____no_output_____" ], [ "from mlxtend.frequent_patterns import apriori\n\napriori(df, min_support=0.22)", "_____no_output_____" ], [ "apriori(df, min_support=0.22, use_colnames=True)", "_____no_output_____" ], [ "frequent_itemsets = apriori(df, min_support=0.2, use_colnames=True)\nfrequent_itemsets['length'] = frequent_itemsets['itemsets'].apply(lambda x: len(x))\nfrequent_itemsets", "_____no_output_____" ], [ "frequent_itemsets[ (frequent_itemsets['length'] == 2) &\n (frequent_itemsets['support'] >= 0.2) ]", "_____no_output_____" ], [ "frequent_itemsets[ (frequent_itemsets['length'] == 3) &\n (frequent_itemsets['support'] >= 0.2) ]", "_____no_output_____" ], [ "dataset = [\n ['i1','i2','i5'],\n ['i2', 'i4'],\n ['i1', 'i2', 'i4'],\n ['i1', 'i3'],\n ['i2', 'i3'],\n ['i1','i3'],\n ['i1', 'i2','i3']]", "_____no_output_____" ], [ "print(dataset)", "[['i1', 'i2', 'i5'], ['i2', 'i4'], ['i1', 'i2', 'i4'], ['i1', 'i3'], ['i2', 'i3'], ['i1', 'i3'], ['i1', 'i2', 'i3']]\n" ], [ "frequent_itemsets = apriori(df, min_support=0.22, use_colnames=True)\nfrequent_itemsets['length'] = frequent_itemsets['itemsets'].apply(lambda x: len(x))\nfrequent_itemsets\n", "_____no_output_____" ], [ "dataset = [\n ['i1','i2','i4'],\n ['i1', 'i4'],\n ['i2', 'i3', 'i4'],\n ['i2', 'i3'],\n ['i2', 'i4'],\n ['i1','i5'],\n ['i1', 'i4','i5']]", "_____no_output_____" ], [ "frequent_itemsets = apriori(df, min_support=0.20, use_colnames=True)\nfrequent_itemsets['length'] = frequent_itemsets['itemsets'].apply(lambda x: len(x))\nfrequent_itemsets", "_____no_output_____" ], [ "", "_____no_output_____" ], [ "", "_____no_output_____" ], [ "", "_____no_output_____" ] ] ]

[ "markdown", "code" ]

[ [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT", "BSD-3-Clause" ]

[ "MIT", "BSD-3-Clause" ]

[ "MIT", "BSD-3-Clause" ]

[ [ [ ".. _data_tutorial:\n\n.. currentmodule:: seaborn", "_____no_output_____" ], [ "Data structures accepted by seaborn\n===================================\n\n.. raw:: html\n\n <div class=col-md-9>\n\nAs a data visualization library, seaborn requires that you provide it with data. This chapter explains the various ways to accomplish that task. Seaborn supports several different dataset formats, and most functions accept data represented with objects from the `pandas <https://pandas.pydata.org/>`_ or `numpy <https://numpy.org/>`_ libraries as well as built-in Python types like lists and dictionaries. Understanding the usage patterns associated with these different options will help you quickly create useful visualizations for nearly any dataset.\n\n.. note::\n As of current writing (v0.11.0), the full breadth of options covered here are supported by only a subset of the modules in seaborn (namely, the :ref:`relational <relational_api>` and :ref:`distribution <distribution_api>` modules). The other modules offer much of the same flexibility, but have some exceptions (e.g., :func:`catplot` and :func:`lmplot` are limited to long-form data with named variables). The data-ingest code will be standardized over the next few release cycles, but until that point, be mindful of the specific documentation for each function if it is not doing what you expect with your dataset.", "_____no_output_____" ] ], [ [ "import numpy as np\nimport pandas as pd\nimport seaborn as sns\nsns.set_theme()", "_____no_output_____" ] ], [ [ "Long-form vs. wide-form data\n----------------------------\n\nMost plotting functions in seaborn are oriented towards *vectors* of data. When plotting ``x`` against ``y``, each variable should be a vector. Seaborn accepts data *sets* that have more than one vector organized in some tabular fashion. There is a fundamental distinction between \"long-form\" and \"wide-form\" data tables, and seaborn will treat each differently.\n\nLong-form data\n~~~~~~~~~~~~~~\n\nA long-form data table has the following characteristics:\n\n- Each variable is a column\n- Each observation is a row", "_____no_output_____" ], [ "As a simple example, consider the \"flights\" dataset, which records the number of airline passengers who flew in each month from 1949 to 1960. This dataset has three variables (*year*, *month*, and number of *passengers*):", "_____no_output_____" ] ], [ [ "flights = sns.load_dataset(\"flights\")\nflights.head()", "_____no_output_____" ] ], [ [ "With long-form data, columns in the table are given roles in the plot by explicitly assigning them to one of the variables. For example, making a monthly plot of the number of passengers per year looks like this:", "_____no_output_____" ] ], [ [ "sns.relplot(data=flights, x=\"year\", y=\"passengers\", hue=\"month\", kind=\"line\")", "_____no_output_____" ] ], [ [ "The advantage of long-form data is that it lends itself well to this explicit specification of the plot. It can accommodate datasets of arbitrary complexity, so long as the variables and observations can be clearly defined. But this format takes some getting used to, because it is often not the model of the data that one has in their head.\n\nWide-form data\n~~~~~~~~~~~~~~\n\nFor simple datasets, it is often more intuitive to think about data the way it might be viewed in a spreadsheet, where the columns and rows contain *levels* of different variables. For example, we can convert the flights dataset into a wide-form organization by \"pivoting\" it so that each column has each month's time series over years:", "_____no_output_____" ] ], [ [ "flights_wide = flights.pivot(index=\"year\", columns=\"month\", values=\"passengers\")\nflights_wide.head()", "_____no_output_____" ] ], [ [ "Here we have the same three variables, but they are organized differently. The variables in this dataset are linked to the *dimensions* of the table, rather than to named fields. Each observation is defined by both the value at a cell in the table and the coordinates of that cell with respect to the row and column indices.", "_____no_output_____" ], [ "With long-form data, we can access variables in the dataset by their name. That is not the case with wide-form data. Nevertheless, because there is a clear association between the dimensions of the table and the variable in the dataset, seaborn is able to assign those variables roles in the plot.\n\n.. note::\n Seaborn treats the argument to ``data`` as wide form when neither ``x`` nor ``y`` are assigned.", "_____no_output_____" ] ], [ [ "sns.relplot(data=flights_wide, kind=\"line\")", "_____no_output_____" ] ], [ [ "This plot looks very similar to the one before. Seaborn has assigned the index of the dataframe to ``x``, the values of the dataframe to ``y``, and it has drawn a separate line for each month. There is a notable difference between the two plots, however. When the dataset went through the \"pivot\" operation that converted it from long-form to wide-form, the information about what the values mean was lost. As a result, there is no y axis label. (The lines also have dashes here, because :func:`relplot` has mapped the column variable to both the ``hue`` and ``style`` semantic so that the plot is more accessible. We didn't do that in the long-form case, but we could have by setting ``style=\"month\"``).\n\nThus far, we did much less typing while using wide-form data and made nearly the same plot. This seems easier! But a big advantage of long-form data is that, once you have the data in the correct format, you no longer need to think about its *structure*. You can design your plots by thinking only about the variables contained within it. For example, to draw lines that represent the monthly time series for each year, simply reassign the variables:", "_____no_output_____" ] ], [ [ "sns.relplot(data=flights, x=\"month\", y=\"passengers\", hue=\"year\", kind=\"line\")", "_____no_output_____" ] ], [ [ "To achieve the same remapping with the wide-form dataset, we would need to transpose the table:", "_____no_output_____" ] ], [ [ "sns.relplot(data=flights_wide.transpose(), kind=\"line\")", "_____no_output_____" ] ], [ [ "(This example also illustrates another wrinkle, which is that seaborn currently considers the column variable in a wide-form dataset to be categorical regardless of its datatype, whereas, because the long-form variable is numeric, it is assigned a quantitative color palette and legend. This may change in the future).\n\nThe absence of explicit variable assignments also means that each plot type needs to define a fixed mapping between the dimensions of the wide-form data and the roles in the plot. Because this natural mapping may vary across plot types, the results are less predictable when using wide-form data. For example, the :ref:`categorical <categorical_api>` plots assign the *column* dimension of the table to ``x`` and then aggregate across the rows (ignoring the index):", "_____no_output_____" ] ], [ [ "sns.catplot(data=flights_wide, kind=\"box\")", "_____no_output_____" ] ], [ [ "When using pandas to represent wide-form data, you are limited to just a few variables (no more than three). This is because seaborn does not make use of multi-index information, which is how pandas represents additional variables in a tabular format. The `xarray <http://xarray.pydata.org/en/stable/>`_ project offers labeled N-dimensional array objects, which can be considered a generalization of wide-form data to higher dimensions. At present, seaborn does not directly support objects from ``xarray``, but they can be transformed into a long-form :class:`pandas.DataFrame` using the ``to_pandas`` method and then plotted in seaborn like any other long-form data set.\n\nIn summary, we can think of long-form and wide-form datasets as looking something like this:", "_____no_output_____" ] ], [ [ "import matplotlib.pyplot as plt\nf = plt.figure(figsize=(7, 5))\n\ngs = plt.GridSpec(\n ncols=6, nrows=2, figure=f,\n left=0, right=.35, bottom=0, top=.9,\n height_ratios=(1, 20),\n wspace=.1, hspace=.01\n)\n\ncolors = [c + (.5,) for c in sns.color_palette()]\n\nf.add_subplot(gs[0, :], facecolor=\".8\")\n[\n f.add_subplot(gs[1:, i], facecolor=colors[i])\n for i in range(gs.ncols)\n]\n\ngs = plt.GridSpec(\n ncols=2, nrows=2, figure=f,\n left=.4, right=1, bottom=.2, top=.8,\n height_ratios=(1, 8), width_ratios=(1, 11),\n wspace=.015, hspace=.02\n)\n\nf.add_subplot(gs[0, 1:], facecolor=colors[2])\nf.add_subplot(gs[1:, 0], facecolor=colors[1])\nf.add_subplot(gs[1, 1], facecolor=colors[0])\n\nfor ax in f.axes:\n ax.set(xticks=[], yticks=[])\n\nf.text(.35 / 2, .91, \"Long-form\", ha=\"center\", va=\"bottom\", size=15)\nf.text(.7, .81, \"Wide-form\", ha=\"center\", va=\"bottom\", size=15)", "_____no_output_____" ] ], [ [ "Messy data\n~~~~~~~~~~\n\nMany datasets cannot be clearly interpreted using either long-form or wide-form rules. If datasets that are clearly long-form or wide-form are `\"tidy\" <https://vita.had.co.nz/papers/tidy-data.pdf>`_, we might say that these more ambiguous datasets are \"messy\". In a messy dataset, the variables are neither uniquely defined by the keys nor by the dimensions of the table. This often occurs with *repeated-measures* data, where it is natural to organize a table such that each row corresponds to the *unit* of data collection. Consider this simple dataset from a psychology experiment in which twenty subjects performed a memory task where they studied anagrams while their attention was either divided or focused:", "_____no_output_____" ] ], [ [ "anagrams = sns.load_dataset(\"anagrams\")\nanagrams", "_____no_output_____" ] ], [ [ "The attention variable is *between-subjects*, but there is also a *within-subjects* variable: the number of possible solutions to the anagrams, which varied from 1 to 3. The dependent measure is a score of memory performance. These two variables (number and score) are jointly encoded across several columns. As a result, the whole dataset is neither clearly long-form nor clearly wide-form.\n\nHow might we tell seaborn to plot the average score as a function of attention and number of solutions? We'd first need to coerce the data into one of our two structures. Let's transform it to a tidy long-form table, such that each variable is a column and each row is an observation. We can use the method :meth:`pandas.DataFrame.melt` to accomplish this task:", "_____no_output_____" ] ], [ [ "anagrams_long = anagrams.melt(id_vars=[\"subidr\", \"attnr\"], var_name=\"solutions\", value_name=\"score\")\nanagrams_long.head()", "_____no_output_____" ] ], [ [ "Now we can make the plot that we want:", "_____no_output_____" ] ], [ [ "sns.catplot(data=anagrams_long, x=\"solutions\", y=\"score\", hue=\"attnr\", kind=\"point\")", "_____no_output_____" ] ], [ [ "Further reading and take-home points\n~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~\n\nFor a longer discussion about tabular data structures, you could read the `\"Tidy Data\" <https://vita.had.co.nz/papers/tidy-data.pdf>`_ paper by Hadley Whickham. Note that seaborn uses a slightly different set of concepts than are defined in the paper. While the paper associates tidyness with long-form structure, we have drawn a distinction between \"tidy wide-form\" data, where there is a clear mapping between variables in the dataset and the dimensions of the table, and \"messy data\", where no such mapping exists.\n\nThe long-form structure has clear advantages. It allows you to create figures by explicitly assigning variables in the dataset to roles in plot, and you can do so with more than three variables. When possible, try to represent your data with a long-form structure when embarking on serious analysis. Most of the examples in the seaborn documentation will use long-form data. But in cases where it is more natural to keep the dataset wide, remember that seaborn can remain useful.", "_____no_output_____" ], [ "Options for visualizing long-form data\n--------------------------------------\n\nWhile long-form data has a precise definition, seaborn is fairly flexible in terms of how it is actually organized across the data structures in memory. The examples in the rest of the documentation will typically use :class:`pandas.DataFrame` objects and reference variables in them by assigning names of their columns to the variables in the plot. But it is also possible to store vectors in a Python dictionary or a class that implements that interface:", "_____no_output_____" ] ], [ [ "flights_dict = flights.to_dict()\nsns.relplot(data=flights_dict, x=\"year\", y=\"passengers\", hue=\"month\", kind=\"line\")", "_____no_output_____" ] ], [ [ "Many pandas operations, such as a the split-apply-combine operations of a group-by, will produce a dataframe where information has moved from the columns of the input dataframe to the index of the output. So long as the name is retained, you can still reference the data as normal:", "_____no_output_____" ] ], [ [ "flights_avg = flights.groupby(\"year\").mean()\nsns.relplot(data=flights_avg, x=\"year\", y=\"passengers\", kind=\"line\")", "_____no_output_____" ] ], [ [ "Additionally, it's possible to pass vectors of data directly as arguments to ``x``, ``y``, and other plotting variables. If these vectors are pandas objects, the ``name`` attribute will be used to label the plot:", "_____no_output_____" ] ], [ [ "year = flights_avg.index\npassengers = flights_avg[\"passengers\"]\nsns.relplot(x=year, y=passengers, kind=\"line\")", "_____no_output_____" ] ], [ [ "Numpy arrays and other objects that implement the Python sequence interface work too, but if they don't have names, the plot will not be as informative without further tweaking:", "_____no_output_____" ] ], [ [ "sns.relplot(x=year.to_numpy(), y=passengers.to_list(), kind=\"line\")", "_____no_output_____" ] ], [ [ "Options for visualizing wide-form data\n--------------------------------------\n\nThe options for passing wide-form data are even more flexible. As with long-form data, pandas objects are preferable because the name (and, in some cases, index) information can be used. But in essence, any format that can be viewed as a single vector or a collection of vectors can be passed to ``data``, and a valid plot can usually be constructed.\n\nThe example we saw above used a rectangular :class:`pandas.DataFrame`, which can be thought of as a collection of its columns. A dict or list of pandas objects will also work, but we'll lose the axis labels:", "_____no_output_____" ] ], [ [ "flights_wide_list = [col for _, col in flights_wide.items()]\nsns.relplot(data=flights_wide_list, kind=\"line\")", "_____no_output_____" ] ], [ [ "The vectors in a collection do not need to have the same length. If they have an ``index``, it will be used to align them:", "_____no_output_____" ] ], [ [ "two_series = [flights_wide.loc[:1955, \"Jan\"], flights_wide.loc[1952:, \"Aug\"]]\nsns.relplot(data=two_series, kind=\"line\")", "_____no_output_____" ] ], [ [ "Whereas an ordinal index will be used for numpy arrays or simple Python sequences:", "_____no_output_____" ] ], [ [ "two_arrays = [s.to_numpy() for s in two_series]\nsns.relplot(data=two_arrays, kind=\"line\")", "_____no_output_____" ] ], [ [ "But a dictionary of such vectors will at least use the keys:", "_____no_output_____" ] ], [ [ "two_arrays_dict = {s.name: s.to_numpy() for s in two_series}\nsns.relplot(data=two_arrays_dict, kind=\"line\")", "_____no_output_____" ] ], [ [ "Rectangular numpy arrays are treated just like a dataframe without index information, so they are viewed as a collection of column vectors. Note that this is different from how numpy indexing operations work, where a single indexer will access a row. But it is consistent with how pandas would turn the array into a dataframe or how matplotlib would plot it:", "_____no_output_____" ] ], [ [ "flights_array = flights_wide.to_numpy()\nsns.relplot(data=flights_array, kind=\"line\")", "_____no_output_____" ], [ "# TODO once the categorical module is refactored, its single vectors will get special treatment\n# (they'll look like collection of singletons, rather than a single collection). That should be noted.", "_____no_output_____" ] ] ]

[ "raw", "code", "raw", "code", "raw", "code", "raw", "code", "raw", "code", "raw", "code", "raw", "code", "raw", "code", "raw", "code", "raw", "code", "raw", "code", "raw", "code", "raw", "code", "raw", "code", "raw", "code", "raw", "code", "raw", "code", "raw", "code", "raw", "code", "raw", "code", "raw", "code" ]

[ [ "raw", "raw" ], [ "code" ], [ "raw", "raw" ], [ "code" ], [ "raw" ], [ "code" ], [ "raw" ], [ "code" ], [ "raw", "raw" ], [ "code" ], [ "raw" ], [ "code" ], [ "raw" ], [ "code" ], [ "raw" ], [ "code" ], [ "raw" ], [ "code" ], [ "raw" ], [ "code" ], [ "raw" ], [ "code" ], [ "raw" ], [ "code" ], [ "raw", "raw" ], [ "code" ], [ "raw" ], [ "code" ], [ "raw" ], [ "code" ], [ "raw" ], [ "code" ], [ "raw" ], [ "code" ], [ "raw" ], [ "code" ], [ "raw" ], [ "code" ], [ "raw" ], [ "code" ], [ "raw" ], [ "code", "code" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "# 风格迁移的实现\n\n本文件是集智学园开发的“火炬上的深度学习”课程的配套源代码。我们讲解了Prisma软件实现风格迁移的实现原理\n\n在这节课中，我们将学会玩图像的风格迁移。\n\n\n\n我们需要准备两张图像，一张作为化作风格，一张作为图像内容\n\n同时，在本文件中，我们还展示了如何实用GPU来进行计算 \n\n本文件是集智学园http://campus.swarma.org 出品的“火炬上的深度学习”第IV课的配套源代码", "_____no_output_____" ] ], [ [ "#导入必要的包\nfrom __future__ import print_function\n\nimport torch\nimport torch.nn as nn\nimport torch.optim as optim\n\nfrom PIL import Image\nimport matplotlib.pyplot as plt\n\nimport torchvision.transforms as transforms\nimport torchvision.models as models\n\nimport copy\n\n# 是否用GPU计算，如果检测到有安装好的GPU，则利用它来计算\nuse_cuda = torch.cuda.is_available()\ndtype = torch.cuda.FloatTensor if use_cuda else torch.FloatTensor", "_____no_output_____" ] ], [ [ "## 一、准备输入文件\n\n我们需要准备两张同样大小的文件，一张作为风格，一张作为内容", "_____no_output_____" ] ], [ [ "#风格图像的路径,自行设定\nstyle = 'images/escher.jpg'\n\n#内容图像的路径，自行设定\ncontent = 'images/portrait1.jpg'\n\n#风格损失所占比重\nstyle_weight=1000\n\n#内容损失所占比重\ncontent_weight=1\n\n#希望得到的图片大小（越大越清晰，计算越慢）\nimsize = 128\n\nloader = transforms.Compose([\n transforms.Resize(imsize), # 将加载的图像转变为指定的大小\n transforms.ToTensor()]) # 将图像转化为tensor\n\n#图片加载函数\ndef image_loader(image_name):\n image = Image.open(image_name)\n image = loader(image).clone().detach().requires_grad_(True)\n # 为了适应卷积网络的需要，虚拟一个batch的维度\n image = image.unsqueeze(0)\n return image\n\n#载入图片并检查尺寸\nstyle_img = image_loader(style).type(dtype)\ncontent_img = image_loader(content).type(dtype)\n\nassert style_img.size() == content_img.size(), \\\n \"我们需要输入相同尺寸的风格和内容图像\"\n\n# 绘制图像的函数\ndef imshow(tensor, title=None):\n image = tensor.clone().cpu() # 克隆Tensor防止改变\n image = image.view(3, imsize, imsize) # 删除添加的batch层\n image = unloader(image)\n plt.imshow(image)\n if title is not None:\n plt.title(title)\n plt.pause(0.001) # 停一会以便更新视图\n\n#绘制图片并查看\nunloader = transforms.ToPILImage() # 将其转化为PIL图像（Python Imaging Library） \nplt.ion()\n\nplt.figure()\nimshow(style_img.data, title='Style Image')\n\nplt.figure()\nimshow(content_img.data, title='Content Image')", "_____no_output_____" ] ], [ [ "## 二、风格迁移网络的实现\n\n值得注意的是，风格迁移的实现并没有训练一个神经网络，而是将已训练好的卷积神经网络价格直接迁移过来\n网络的学习过程并不体现为对神经网络权重的训练，而是训练一张输入的图像，让它尽可能地靠近内容图像的内容和风格图像的风格\n\n为了实现风格迁移，我们需要在迁移网络的基础上再构建一个计算图，这样可以加速计算。构建计算图分为两部：\n\n1、加载一个训练好的CNN；\n\n2、在原网络的基础上添加计算风格损失和内容损失的新计算层", "_____no_output_____" ], [ "### 1. 加载已训练好的大型网络VGG", "_____no_output_____" ] ], [ [ "cnn = models.vgg19(pretrained=True).features\n\n# 如果可能就用GPU计算:\nif use_cuda:\n cnn = cnn.cuda()", "_____no_output_____" ] ], [ [ "### 2. 重新定义新的计算模块", "_____no_output_____" ] ], [ [ "#内容损失模块\nclass ContentLoss(nn.Module):\n\n def __init__(self, target, weight):\n super(ContentLoss, self).__init__()\n # 由于网络的权重都是从target上迁移过来，所以在计算梯度的时候，需要把它和原始计算图分离\n self.target = target.detach() * weight\n self.weight = weight\n self.criterion = nn.MSELoss()\n\n def forward(self, input):\n # 输入input为一个特征图\n # 它的功能就是计算误差，误差就是当前计算的内容与target之间的均方误差\n self.loss = self.criterion(input * self.weight, self.target)\n self.output = input\n return self.output\n\n def backward(self, retain_graph=True):\n # 开始进行反向传播算法\n self.loss.backward(retain_graph=retain_graph)\n return self.loss\n\nclass StyleLoss(nn.Module):\n\n # 计算风格损失的神经模块\n def __init__(self, target, weight):\n super(StyleLoss, self).__init__()\n self.target = target.detach() * weight\n self.weight = weight\n #self.gram = GramMatrix()\n self.criterion = nn.MSELoss()\n\n def forward(self, input):\n # 输入input就是一个特征图\n self.output = input.clone()\n # 计算本图像的gram矩阵，并将它与target对比\n input = input.cuda() if use_cuda else input\n self_G = Gram(input)\n self_G.mul_(self.weight)\n # 计算损失函数，即输入特征图的gram矩阵与目标特征图的gram矩阵之间的差异\n self.loss = self.criterion(self_G, self.target)\n return self.output\n\n def backward(self, retain_graph=True):\n # 反向传播算法\n self.loss.backward(retain_graph=retain_graph)\n return self.loss\n\n#定义Gram矩阵\ndef Gram(input):\n # 输入一个特征图，计算gram矩阵\n a, b, c, d = input.size() # a=batch size(=1)\n # b=特征图的数量\n # (c,d)=特征图的图像尺寸 (N=c*d)\n\n features = input.view(a * b, c * d) # 将特征图图像扁平化为一个向量\n\n G = torch.mm(features, features.t()) # 计算任意两个向量之间的乘积\n\n # 我们通过除以特征图中的像素数量来归一化特征图\n return G.div(a * b * c * d)", "_____no_output_____" ], [ "\n# 希望计算的内容或者风格层 :\ncontent_layers = ['conv_4'] #只考虑第四个卷积层的内容\n\n\nstyle_layers = ['conv_1', 'conv_2', 'conv_3', 'conv_4', 'conv_5']\n# 考虑第1、2、3、4、5层的风格损失\n\n\n# 定义列表存储每一个周期的计算损失\ncontent_losses = []\nstyle_losses = []\n\nmodel = nn.Sequential() # 一个新的序贯网络模型\n\n# 如果有GPU就把这些计算挪到GPU上:\nif use_cuda:\n model = model.cuda()\n\n\n \n \n# 接下来要做的操作是：循环vgg的每一层，同时构造一个全新的神经网络model\n# 这个新网络与vgg基本一样，只是多了一些新的层来计算风格损失和内容损失。\n# 将每层卷积核的数据都加载到新的网络模型model上来\ni = 1\nfor layer in list(cnn):\n if isinstance(layer, nn.Conv2d):\n name = \"conv_\" + str(i)\n #将已加载的模块放到model这个新的神经模块中\n model.add_module(name, layer)\n\n if name in content_layers:\n # 如果当前层模型在定义好的要计算内容的层:\n target = model(content_img).clone() #将内容图像当前层的feature信息拷贝到target中\n content_loss = ContentLoss(target, content_weight) #定义content_loss的目标函数\n content_loss = content_loss if use_cuda else content_loss\n model.add_module(\"content_loss_\" + str(i), content_loss) #在新网络上加content_loss层\n content_losses.append(content_loss)\n\n if name in style_layers:\n # 如果当前层在指定的风格层中，进行风格层损失的计算\n target_feature = model(style_img).clone()\n target_feature = target_feature.cuda() if use_cuda else target_feature\n target_feature_gram = Gram(target_feature)\n style_loss = StyleLoss(target_feature_gram, style_weight)\n style_loss = style_loss.cuda() if use_cuda else style_loss\n model.add_module(\"style_loss_\" + str(i), style_loss)\n style_losses.append(style_loss)\n\n if isinstance(layer, nn.ReLU):\n #如果不是卷积层，则做同样处理\n name = \"relu_\" + str(i)\n model.add_module(name, layer)\n\n i += 1\n\n if isinstance(layer, nn.MaxPool2d):\n name = \"pool_\" + str(i)\n model.add_module(name, layer) # ***\n\n", "_____no_output_____" ] ], [ [ "## 二、风格迁移的训练", "_____no_output_____" ], [ "### 1. 首先，我们需要现准备一张原始的图像，可以是一张噪音图或者就是内容图", "_____no_output_____" ] ], [ [ "\n# 如果想从调整一张噪声图像开始，请用下面一行的代码\ninput_img = torch.randn(content_img.data.size())\n\nif use_cuda:\n input_img = input_img.cuda()\n content_img = content_img.cuda()\n style_img = style_img.cuda()\n# 将选中的待调整图打印出来:\nplt.figure()\nimshow(input_img.data, title='Input Image')\n", "_____no_output_____" ] ], [ [ "### 2. 优化输入的图像（训练过程）", "_____no_output_____" ] ], [ [ "# 首先，需要先讲输入图像变成神经网络的参数，这样我们就可以用反向传播算法来调节这个输入图像了\ninput_param = nn.Parameter(input_img.data)\n\n#定义个优化器，采用LBFGS优化算法来优化（试验效果很好，它的特点是可以计算大规模数据的梯度下降）\noptimizer = optim.LBFGS([input_param])\n\n# 迭代步数\nnum_steps=300\n\n\n\"\"\"运行风格迁移的主算法过程.\"\"\"\nprint('正在构造风格迁移模型..')\n\nprint('开始优化..')\nfor i in range(num_steps):\n #每一个训练周期\n \n # 限制输入图像的色彩取值范围在0-1间\n input_param.data.clamp_(0, 1)\n \n # 清空梯度\n optimizer.zero_grad()\n # 将图像输入构造的神经网络中\n model(input_param)\n style_score = 0\n content_score = 0\n \n # 每个损失函数层都开始反向传播算法\n for sl in style_losses:\n style_score += sl.backward()\n for cl in content_losses:\n content_score += cl.backward()\n\n # 每隔50个周期打印一次训练数据\n if i % 50 == 0:\n print(\"运行 {}轮:\".format(i))\n print('风格损失 : {:4f} 内容损失: {:4f}'.format(\n style_score.data.item(), content_score.data.item()))\n print()\n def closure():\n return style_score + content_score\n #一步优化\n optimizer.step(closure)\n\n# 做一些修正，防止数据超界...\noutput = input_param.data.clamp_(0, 1)\n\n# 打印结果图\nplt.figure()\nimshow(output, title='Output Image')\n\nplt.ioff()\nplt.show()", "_____no_output_____" ] ], [ [ "本文件是集智学园http://campus.swarma.org 出品的“火炬上的深度学习”第IV课的配套源代码", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown" ]

[ [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "# Overfitting and regularization (with ``gluon``)\n\nNow that we've built a [regularized logistic regression model from scratch](regularization-scratch.html), let's make this more efficient with ``gluon``. We recommend that you read that section for a description as to why regularization is a good idea. As always, we begin by loading libraries and some data.\n\n[**REFINED DRAFT - RELEASE STAGE: CATFOOD**]", "_____no_output_____" ] ], [ [ "from __future__ import print_function\nimport mxnet as mx\nfrom mxnet import autograd\nfrom mxnet import gluon\nimport mxnet.ndarray as nd\nimport numpy as np\nctx = mx.cpu()\n\n# for plotting purposes\n%matplotlib inline\nimport matplotlib\nimport matplotlib.pyplot as plt", "_____no_output_____" ] ], [ [ "## The MNIST Dataset", "_____no_output_____" ] ], [ [ "mnist = mx.test_utils.get_mnist()\nnum_examples = 1000\nbatch_size = 64\ntrain_data = mx.gluon.data.DataLoader(\n mx.gluon.data.ArrayDataset(mnist[\"train_data\"][:num_examples],\n mnist[\"train_label\"][:num_examples].astype(np.float32)), \n batch_size, shuffle=True)\ntest_data = mx.gluon.data.DataLoader(\n mx.gluon.data.ArrayDataset(mnist[\"test_data\"][:num_examples],\n mnist[\"test_label\"][:num_examples].astype(np.float32)), \n batch_size, shuffle=False)", "_____no_output_____" ] ], [ [ "## Multiclass Logistic Regression", "_____no_output_____" ] ], [ [ "net = gluon.nn.Sequential()\nwith net.name_scope():\n net.add(gluon.nn.Dense(10))", "_____no_output_____" ] ], [ [ "## Parameter initialization\n", "_____no_output_____" ] ], [ [ "net.collect_params().initialize(mx.init.Xavier(magnitude=2.24), ctx=ctx)", "_____no_output_____" ] ], [ [ "## Softmax Cross Entropy Loss", "_____no_output_____" ] ], [ [ "loss = gluon.loss.SoftmaxCrossEntropyLoss()", "_____no_output_____" ] ], [ [ "## Optimizer\n\nBy default ``gluon`` tries to keep the coefficients from diverging by using a *weight decay* penalty. So, to get the real overfitting experience we need to switch it off. We do this by passing `'wd': 0.0'` when we instantiate the trainer. ", "_____no_output_____" ] ], [ [ "trainer = gluon.Trainer(net.collect_params(), 'sgd', {'learning_rate': 0.01, 'wd': 0.0})", "_____no_output_____" ] ], [ [ "## Evaluation Metric", "_____no_output_____" ] ], [ [ "def evaluate_accuracy(data_iterator, net, loss_fun):\n acc = mx.metric.Accuracy()\n loss_avg = 0.\n for i, (data, label) in enumerate(data_iterator):\n data = data.as_in_context(ctx).reshape((-1,784))\n label = label.as_in_context(ctx)\n output = net(data)\n loss = loss_fun(output, label) \n predictions = nd.argmax(output, axis=1)\n acc.update(preds=predictions, labels=label)\n loss_avg = loss_avg*i/(i+1) + nd.mean(loss).asscalar()/(i+1)\n return acc.get()[1], loss_avg\n\ndef plot_learningcurves(loss_tr,loss_ts, acc_tr,acc_ts):\n xs = list(range(len(loss_tr)))\n \n f = plt.figure(figsize=(12,6))\n fg1 = f.add_subplot(121)\n fg2 = f.add_subplot(122)\n \n fg1.set_xlabel('epoch',fontsize=14)\n fg1.set_title('Comparing loss functions')\n fg1.semilogy(xs, loss_tr)\n fg1.semilogy(xs, loss_ts)\n fg1.grid(True,which=\"both\")\n\n fg1.legend(['training loss', 'testing loss'],fontsize=14)\n \n fg2.set_title('Comparing accuracy')\n fg1.set_xlabel('epoch',fontsize=14)\n fg2.plot(xs, acc_tr)\n fg2.plot(xs, acc_ts)\n fg2.grid(True,which=\"both\")\n fg2.legend(['training accuracy', 'testing accuracy'],fontsize=14)\n plt.show()", "_____no_output_____" ] ], [ [ "## Execute training loop", "_____no_output_____" ] ], [ [ "epochs = 700\nmoving_loss = 0.\nniter=0\n\nloss_seq_train = []\nloss_seq_test = []\nacc_seq_train = []\nacc_seq_test = []\n\nfor e in range(epochs):\n for i, (data, label) in enumerate(train_data):\n data = data.as_in_context(ctx).reshape((-1,784))\n label = label.as_in_context(ctx)\n with autograd.record():\n output = net(data)\n cross_entropy = loss(output, label)\n cross_entropy.backward()\n trainer.step(data.shape[0])\n \n ##########################\n # Keep a moving average of the losses\n ##########################\n niter +=1\n moving_loss = .99 * moving_loss + .01 * nd.mean(cross_entropy).asscalar()\n est_loss = moving_loss/(1-0.99**niter)\n \n test_accuracy, test_loss = evaluate_accuracy(test_data, net, loss)\n train_accuracy, train_loss = evaluate_accuracy(train_data, net, loss)\n \n # save them for later\n loss_seq_train.append(train_loss)\n loss_seq_test.append(test_loss)\n acc_seq_train.append(train_accuracy)\n acc_seq_test.append(test_accuracy)\n \n \n if e % 20 == 0:\n print(\"Completed epoch %s. Train Loss: %s, Test Loss %s, Train_acc %s, Test_acc %s\" % \n (e+1, train_loss, test_loss, train_accuracy, test_accuracy)) \n\n## Plotting the learning curves\nplot_learningcurves(loss_seq_train,loss_seq_test,acc_seq_train,acc_seq_test)", "_____no_output_____" ] ], [ [ "## Regularization\n\nNow let's see what this mysterious *weight decay* is all about. We begin with a bit of math. When we add an L2 penalty to the weights we are effectively adding $\\frac{\\lambda}{2} \\|w\\|^2$ to the loss. Hence, every time we compute the gradient it gets an additional $\\lambda w$ term that is added to $g_t$, since this is the very derivative of the L2 penalty. As a result we end up taking a descent step not in the direction $-\\eta g_t$ but rather in the direction $-\\eta (g_t + \\lambda w)$. This effectively shrinks $w$ at each step by $\\eta \\lambda w$, thus the name weight decay. To make this work in practice we just need to set the weight decay to something nonzero.", "_____no_output_____" ] ], [ [ "net.collect_params().initialize(mx.init.Xavier(magnitude=2.24), ctx=ctx, force_reinit=True)\ntrainer = gluon.Trainer(net.collect_params(), 'sgd', {'learning_rate': 0.01, 'wd': 0.001})\n\nmoving_loss = 0.\nniter=0\nloss_seq_train = []\nloss_seq_test = []\nacc_seq_train = []\nacc_seq_test = []\n\nfor e in range(epochs):\n for i, (data, label) in enumerate(train_data):\n data = data.as_in_context(ctx).reshape((-1,784))\n label = label.as_in_context(ctx)\n with autograd.record():\n output = net(data)\n cross_entropy = loss(output, label)\n cross_entropy.backward()\n trainer.step(data.shape[0])\n \n ##########################\n # Keep a moving average of the losses\n ##########################\n niter +=1\n moving_loss = .99 * moving_loss + .01 * nd.mean(cross_entropy).asscalar()\n est_loss = moving_loss/(1-0.99**niter)\n \n test_accuracy, test_loss = evaluate_accuracy(test_data, net,loss)\n train_accuracy, train_loss = evaluate_accuracy(train_data, net, loss)\n \n # save them for later\n loss_seq_train.append(train_loss)\n loss_seq_test.append(test_loss)\n acc_seq_train.append(train_accuracy)\n acc_seq_test.append(test_accuracy)\n \n if e % 20 == 0:\n print(\"Completed epoch %s. Train Loss: %s, Test Loss %s, Train_acc %s, Test_acc %s\" % \n (e+1, train_loss, test_loss, train_accuracy, test_accuracy)) \n \n## Plotting the learning curves\nplot_learningcurves(loss_seq_train,loss_seq_test,acc_seq_train,acc_seq_test)", "_____no_output_____" ] ], [ [ "As we can see, the test accuracy improves a bit. Note that the amount by which it improves actually depends on the amount of weight decay. We recommend that you try and experiment with different extents of weight decay. For instance, a larger weight decay (e.g. $0.01$) will lead to inferior performance, one that's larger still ($0.1$) will lead to terrible results. This is one of the reasons why tuning parameters is quite so important in getting good experimental results in practice.", "_____no_output_____" ], [ "## Next\n[Learning environments](../chapter02_supervised-learning/environment.ipynb)", "_____no_output_____" ], [ "For whinges or inquiries, [open an issue on GitHub.](https://github.com/zackchase/mxnet-the-straight-dope)", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown" ]

[ [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import pandas as pd\nimport numpy as np\nimport matplotlib.pyplot as pp\nimport matplotlib.pyplot as plt\n%matplotlib inline\nimport seaborn as sns\nfrom sklearn.model_selection import train_test_split\nfrom sklearn import linear_model\nfrom sklearn.metrics import mean_squared_error, r2_score\n# Input data files are available in the \"../input/\" directory.\n# For example, running this (by clicking run or pressing Shift+Enter) will list the files in the input directory", "_____no_output_____" ], [ "%matplotlib inline", "_____no_output_____" ], [ "df = pd.read_csv('avocado.csv')\ndf.head()", "_____no_output_____" ], [ "df['Date'] = pd.to_datetime(df['Date'])", "_____no_output_____" ], [ "df.head(5)", "_____no_output_____" ], [ "df = df.drop(['4046','4225','4770','Large Bags','Small Bags','XLarge Bags','Total Volume'],axis=1)", "_____no_output_____" ], [ "US = df.loc[(df['region']) == 'TotalUS'] \nUS.head(5)", "_____no_output_____" ], [ "df['AveragePrice'].plot(kind='hist', rot=70, logx=True, logy=True)\n# Season # 1= Spring #2= Summer #3= Fall #4= Winter", "_____no_output_____" ], [ "# Create bee swarm plot with Seaborn's default settings\n_ = sns.swarmplot(x='Season#', y='AveragePrice', data=US)\n\n# Label the axes\n_ = plt.xlabel('Season')\n_ = plt.ylabel('Average Price')\n\n# Show the plot\nplt.show()", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Introduction\n\n**Prerequisites**\n\n- Python Fundamentals\n\n\n**Outcomes**\n\n- Understand the core pandas objects \n- Series \n- DataFrame \n- Index into particular elements of a Series and DataFrame \n- Understand what `.dtype`/`.dtypes` do \n- Make basic visualizations \n\n\n**Data**\n\n- US regional unemployment data from Bureau of Labor Statistics ", "_____no_output_____" ], [ "## Pandas\n\nThis notebook begins the material on `pandas`\n\nTo start we will import the pandas package and give it the nickname\n`\"pd\"`, which is the conventional way to import pandas", "_____no_output_____" ] ], [ [ "import pandas as pd\n\n# Don't worry about this line for now!\n%matplotlib inline", "_____no_output_____" ] ], [ [ "Sometimes it will be helpful to know which version of pandas we are\nusing\n\nWe can check this by running the code below", "_____no_output_____" ], [ "## Series\n\nThe first main pandas type we will introduce is called Series\n\nA Series is a single column of data, with row labels for each\nobservation\n\nPandas refers to the row labels as the *index* of the Series\n\n<img src=\"https://storage.googleapis.com/ds4e/_static/intro_files/PandasSeries.png\" alt=\"PandasSeries.png\" style=\"\">\n\n \nBelow we create a Series which contains the US unemployment rate every\nother year starting in 1995", "_____no_output_____" ] ], [ [ "values = [5.6, 5.3, 4.3, 4.2, 5.8, 5.3, 4.6, 7.8, 9.1, 8., 5.7]\nyears = list(range(1995, 2017, 2))\n\nunemp = pd.Series(data=values, index=years, name=\"Unemployment\")", "_____no_output_____" ], [ "unemp", "_____no_output_____" ] ], [ [ "We can look at the index and values in our Series", "_____no_output_____" ] ], [ [ "unemp.index", "_____no_output_____" ], [ "unemp.values", "_____no_output_____" ], [ "unemp.", "_____no_output_____" ] ], [ [ "### What can we do with a Series object?", "_____no_output_____" ], [ "#### `.head` and `.tail`\n\nOften our data will have many rows and we won’t want to display it all\nat once\n\nThe methods `.head` and `.tail` show rows at the beginning and end\nof our Series, respectively", "_____no_output_____" ] ], [ [ "unemp.head()", "_____no_output_____" ], [ "unemp.tail()", "_____no_output_____" ] ], [ [ "#### Basic Plotting\n\nWe can also plot data using the `.plot` method\n\nThis is why we needed the `%matplotlib inline` — it tells the notebook\nto display figures inside the notebook itself\n\n*Note*: Pandas can do much more in terms of visualization\n\nWe will talk about more advanced visualization features later", "_____no_output_____" ] ], [ [ "unemp.plot(kind=\"bar\")", "_____no_output_____" ] ], [ [ "#### Unique values\n\nIn this dataset it doesn’t make much sense, but we may want to find the\nunique values in a Series\n\nThis can be done with the `.unique` method", "_____no_output_____" ] ], [ [ "unemp.unique()", "_____no_output_____" ] ], [ [ "#### Indexing\n\nSometimes we will want to select particular elements from a Series\n\nWe can do this using `.loc[index_things]`; where `index_things` is\nan item from the index, or a list of items in the index\n\nWe will see this more in depth in a coming lecture, but for now we\ndemonstrate how to select one or multiple elements of the Series", "_____no_output_____" ] ], [ [ "unemp", "_____no_output_____" ], [ "unemp.loc[[2009, 1995]]", "_____no_output_____" ], [ "unemp.iloc[-1]", "_____no_output_____" ], [ "unemp.loc[[1995, 2005, 2015]]", "_____no_output_____" ] ], [ [ "<blockquote>\n\n**Check for understanding**\n\nFor each of the following exercises, we recommend reading the documentation\nfor help\n\n- Display only the first 2 elements of the Series using the `.head` method \n- Using the `plot` method, make a bar plot \n- Use `.loc` to select the lowest/highest unemployment rate shown in the Series \n- Run the code `unemp.dtype` below. What does it give you? Talk with your neighbor about where it might come from \n\n\n\n</blockquote>", "_____no_output_____" ] ], [ [ "unemp.loc[[unemp.idxmin(), unemp.idxmax()]]", "_____no_output_____" ] ], [ [ "## DataFrame\n\nA DataFrame is how pandas stores one or more columns of data\n\nWe can think a DataFrames a multiple Series stacked side by side as\ncolumns\n\nThis is similar to a sheet in an Excel workbook or a table in a SQL\ndatabase\n\nIn addition to row labels (an index), DataFrames also have column labels\n\nWe refer to these column labels as the columns or column names\n\n<img src=\"https://storage.googleapis.com/ds4e/_static/intro_files/PandasDataFrame.png\" alt=\"PandasDataFrame.png\" style=\"\">\n\n \nBelow we create a DataFrame that contains the unemployment rate every\nother year by region of the US starting in 1995.", "_____no_output_____" ] ], [ [ "data = {\"NorthEast\": [5.9, 5.6, 4.4, 3.8, 5.8, 4.9, 4.3, 7.1, 8.3, 7.9, 5.7],\n \"MidWest\": [4.5, 4.3, 3.6, 4. , 5.7, 5.7, 4.9, 8.1, 8.7, 7.4, 5.1],\n \"South\": [5.3, 5.2, 4.2, 4. , 5.7, 5.2, 4.3, 7.6, 9.1, 7.4, 5.5],\n \"West\": [6.6, 6., 5.2, 4.6, 6.5, 5.5, 4.5, 8.6, 10.7, 8.5, 6.1],\n \"National\": [5.6, 5.3, 4.3, 4.2, 5.8, 5.3, 4.6, 7.8, 9.1, 8., 5.7]}\n\nunemp_region = pd.DataFrame(data, index=years)\nunemp_region", "_____no_output_____" ] ], [ [ "We can retrieve the index and the DataFrame values in the same way we\ndid with a Series", "_____no_output_____" ] ], [ [ "unemp_region.index", "_____no_output_____" ], [ "unemp_region.values", "_____no_output_____" ] ], [ [ "### What can we do with a DataFrame?\n\nPretty much everything we can do with a Series", "_____no_output_____" ], [ "#### `.head` and `.tail`\n\nAs with Series, we can use `.head` and `.tail` to show only the\nfirst or last `n` rows", "_____no_output_____" ] ], [ [ "unemp_region.head()", "_____no_output_____" ], [ "unemp_region.tail(3)", "_____no_output_____" ] ], [ [ "#### Plotting\n\nWe can generate plots with the `.plot` method\n\nNotice we now have a separate line for each column of data", "_____no_output_____" ] ], [ [ "unemp_region.plot()", "_____no_output_____" ] ], [ [ "#### Indexing\n\nWe can also do indexing using `.loc`\n\nHowever, there is a little more to it than before because we can choose\nsubsets of both row and columns", "_____no_output_____" ] ], [ [ "unemp_region.head()", "_____no_output_____" ], [ "unemp_region.loc[1995, \"NorthEast\"]", "_____no_output_____" ], [ "unemp_region.loc[[1995, 2005], \"South\"]", "_____no_output_____" ], [ "unemp_region.loc[1995, [\"NorthEast\", \"National\"]]", "_____no_output_____" ], [ "unemp_region.loc[:, \"NorthEast\"]", "_____no_output_____" ], [ "# `[string]` with no `.loc` extracts a whole column\nunemp_region[\"MidWest\"]", "_____no_output_____" ] ], [ [ "### Computations with columns\n\nPandas can do various computations and mathematical operations on\ncolumns\n\nLet’s take a look at a few of them", "_____no_output_____" ] ], [ [ "# Divide by 100 to move from percent units to a rate\nunemp_region[\"West\"] / 100", "_____no_output_____" ], [ "# Find maximum\nunemp_region[\"West\"].max()", "_____no_output_____" ], [ "unemp_region[\"West\"].iloc[1:5]", "_____no_output_____" ], [ "unemp_region[\"MidWest\"].head(6)", "_____no_output_____" ], [ "# Find the difference between two columns\n# Notice that pandas applies `-` to _all rows_ at one time\n# We'll see more of this throughout these materials\nunemp_region[\"West\"].iloc[1:5] - unemp_region[\"MidWest\"].head(6)", "_____no_output_____" ], [ "# Find correlation between two columns\nunemp_region.West.corr(unemp_region[\"MidWest\"])", "_____no_output_____" ], [ "# find correlation between all column pairs\nunemp_region.corr()", "_____no_output_____" ] ], [ [ "<blockquote>\n\n**Check for understanding**\n\nFor each of the following, we recommend reading the documentation for help\n\n- Use introspection (or google-fu) to find a way to obtain a list with\n all of the column names in `unemp_region` \n- Using the `plot` method, make a bar plot. What does it look like\n now? \n- Use `.loc` to select the the unemployment data for the\n `NorthEast` and `West` for the years 1995, 2005, 2011, and 2015. \n- Run the code `unemp_region.dtypes` below. What does it give you?\n How does this compare with `unemp.dtype`? \n\n\n\n</blockquote>", "_____no_output_____" ], [ "## Data types\n\nWe asked you to run the commands `unemp.dtype` and\n`unemp_region.dtypes` and think about what these methods output\n\nYou might have guessed that they return the type of the values inside\neach column\n\nOccasionally, you might need to investigate what types you have in your\nDataFrame when an operation is not doing what you expect it to", "_____no_output_____" ] ], [ [ "unemp.dtype", "_____no_output_____" ], [ "unemp_region.dtypes", "_____no_output_____" ] ], [ [ "DataFrames will only distinguish between a few types\n\n- Booleans (`bool`) \n- Floating point numbers (`float64`) \n- Integers (`int64`) \n- Dates (`datetime`) — we will learn this soon \n- Categorical data (`categorical`) \n- Everything else, including strings (`object`) \n\n\nIn the future, we will often refer to the type of data stored in a\ncolumn as its `dtype`\n\nLet’s look at an example for when having an incorrect `dtype` can\ncause problems\n\nSuppose that when we imported the data the `South` column was\ninterpreted as a string", "_____no_output_____" ] ], [ [ "str_unemp = unemp_region.copy()\nstr_unemp[\"South\"] = str_unemp[\"South\"].astype(str)\nstr_unemp.dtypes", "_____no_output_____" ] ], [ [ "Everything *looks* ok…", "_____no_output_____" ] ], [ [ "str_unemp.head()", "_____no_output_____" ] ], [ [ "But if we try to do something like compute the sum of all the columns,\nwe get unexpected results…", "_____no_output_____" ] ], [ [ "str_unemp.sum()", "_____no_output_____" ] ], [ [ "This happened because `.sum` effectively calls `+` on all rows in\neach column\n\nRecall that when we apply `+` to two strings, the result is the\nstrings mashed together\n\nSo in this case we saw that the entries in all the rows of the South\ncolumn were stitched together into one long string", "_____no_output_____" ], [ "## Changing DataFrames\n\nWe can change the data inside of a DataFrame in various ways:\n\n- Adding new columns \n- Changing index labels or column names \n- Altering existing data (e.g. doing some arithmetic or making a column\n of strings lowercase) \n\n\nSome of these “mutations” will be topics of future notebooks, so we will\nonly briefly discuss a few of the things we can do below", "_____no_output_____" ], [ "### Creating new columns\n\nWe can create new data by “assigning values to a column” similar to how\nwe assign values to a variable\n\nIn pandas, we create a new column of a DataFrame by writing", "_____no_output_____" ], [ "```python\ndf[\"New Column Name\"] = new_values\n```\n", "_____no_output_____" ], [ "Below we create an unweighted mean of the unemployment rate across the\nfour regions of the US — notice this differs from the national\nunemployment rate", "_____no_output_____" ] ], [ [ "unemp_region[\"UnweightedMean\"] = (unemp_region[\"NorthEast\"] +\n unemp_region[\"MidWest\"] +\n unemp_region[\"South\"] +\n unemp_region[\"West\"])/4", "_____no_output_____" ], [ "unemp_region.head()", "_____no_output_____" ] ], [ [ "### Changing values\n\nChanging the values inside of a DataFrame should be done sparingly\n\nHowever, it can be done by assigning a value to a location in the\nDataFrame\n\n`df.loc[index, column] = value`", "_____no_output_____" ] ], [ [ "unemp_region.loc[1995, \"UnweightedMean\"] = 0.0", "_____no_output_____" ], [ "unemp_region.head()", "_____no_output_____" ] ], [ [ "### Renaming columns\n\nWe can also rename the columns of a DataFrame\n\nThis is helpful because the names that sometimes come with datasets are\nunbearable…\n\nFor example, the original name for the North East unemployment rate\ngiven by the Bureau of Labor Statistics was `LASRD910000000000003`…\n\nThey have their reasons for using these names, but it can make our job\ndifficult since we need to type it sometimes repeatedly\n\nWe can rename columns by passing a dictionary to the `rename` method\n\nThis dictionary contains the old names as the keys and new names as the\nvalues.\n\nSee the example below", "_____no_output_____" ] ], [ [ "names = {\"NorthEast\": \"NE\",\n \"MidWest\": \"MW\",\n \"South\": \"S\",\n \"West\": \"W\"}\nunemp_region.rename(columns=names)", "_____no_output_____" ], [ "unemp_region.head()", "_____no_output_____" ] ], [ [ "We renamed our columns… Why does the DataFrame still show the old\ncolumn names?\n\nMany of the operations that pandas does creates a copy of your data by\ndefault\n\nIt does this in order to protect your data and make sure you don’t\noverwrite information you’d like to keep\n\nWe can make these operations permanent by either\n\n1. Assigning the output back to the variable name\n `df = df.rename(columns=rename_dict)` or \n1. Looking into whether the method has an `inplace` option. For\n example, `df.rename(columns=rename_dict, inplace=True)` \n\n\nThere are times when setting `inplace=True` will make your code faster\n(e.g. if you have a very large DataFrame and you don’t want to copy all\nthe data), but that doesn’t always happen\n\nWe recommend using the first option until you get comfortable with\npandas because operations that don’t alter your data are (usually)\neasier to reason about", "_____no_output_____" ] ], [ [ "names = {\"NorthEast\": \"NE\",\n \"MidWest\": \"MW\",\n \"South\": \"S\",\n \"West\": \"W\"}\n\nunemp_shortname = unemp_region.rename(columns=names)\nunemp_shortname.head()", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown", "markdown", "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ] ]

[ "BSD-3-Clause" ]

[ "BSD-3-Clause" ]

[ "BSD-3-Clause" ]

[ [ [ "<a href=\"https://colab.research.google.com/github/Abudhagir/DeepLearningTutorials/blob/master/Classify_text_with_bert.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>", "_____no_output_____" ], [ "##### Copyright 2020 The TensorFlow Hub Authors.\n", "_____no_output_____" ] ], [ [ "#@title Licensed under the Apache License, Version 2.0 (the \"License\");\n# you may not use this file except in compliance with the License.\n# You may obtain a copy of the License at\n#\n# https://www.apache.org/licenses/LICENSE-2.0\n#\n# Unless required by applicable law or agreed to in writing, software\n# distributed under the License is distributed on an \"AS IS\" BASIS,\n# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n# See the License for the specific language governing permissions and\n# limitations under the License.", "_____no_output_____" ] ], [ [ "<table class=\"tfo-notebook-buttons\" align=\"left\">\n <td>\n <a target=\"_blank\" href=\"https://www.tensorflow.org/text/tutorials/classify_text_with_bert\"><img src=\"https://www.tensorflow.org/images/tf_logo_32px.png\" />View on TensorFlow.org</a>\n </td>\n <td>\n <a target=\"_blank\" href=\"https://colab.research.google.com/github/tensorflow/text/blob/master/docs/tutorials/classify_text_with_bert.ipynb\"><img src=\"https://www.tensorflow.org/images/colab_logo_32px.png\" />Run in Google Colab</a>\n </td>\n <td>\n <a target=\"_blank\" href=\"https://github.com/tensorflow/text/blob/master/docs/tutorials/classify_text_with_bert.ipynb\"><img src=\"https://www.tensorflow.org/images/GitHub-Mark-32px.png\" />View on GitHub</a>\n </td>\n <td>\n <a href=\"https://storage.googleapis.com/tensorflow_docs/text/docs/tutorials/classify_text_with_bert.ipynb\"><img src=\"https://www.tensorflow.org/images/download_logo_32px.png\" />Download notebook</a>\n </td>\n <td>\n <a href=\"https://tfhub.dev/google/collections/bert/1\"><img src=\"https://www.tensorflow.org/images/hub_logo_32px.png\" />See TF Hub model</a>\n </td>\n</table>", "_____no_output_____" ], [ "# Classify text with BERT\n\nThis tutorial contains complete code to fine-tune BERT to perform sentiment analysis on a dataset of plain-text IMDB movie reviews.\nIn addition to training a model, you will learn how to preprocess text into an appropriate format.\n\nIn this notebook, you will:\n\n- Load the IMDB dataset\n- Load a BERT model from TensorFlow Hub\n- Build your own model by combining BERT with a classifier\n- Train your own model, fine-tuning BERT as part of that\n- Save your model and use it to classify sentences\n\nIf you're new to working with the IMDB dataset, please see [Basic text classification](https://www.tensorflow.org/tutorials/keras/text_classification) for more details.", "_____no_output_____" ], [ "## About BERT\n\n[BERT](https://arxiv.org/abs/1810.04805) and other Transformer encoder architectures have been wildly successful on a variety of tasks in NLP (natural language processing). They compute vector-space representations of natural language that are suitable for use in deep learning models. The BERT family of models uses the Transformer encoder architecture to process each token of input text in the full context of all tokens before and after, hence the name: Bidirectional Encoder Representations from Transformers. \n\nBERT models are usually pre-trained on a large corpus of text, then fine-tuned for specific tasks.\n", "_____no_output_____" ], [ "## Setup\n", "_____no_output_____" ], [ "#TensorFlow Text\n\"TensorFlow Text provides a collection of text related classes and ops ready to use with TensorFlow 2.0. The library can perform the preprocessing regularly required by text-based models, and includes other features useful for sequence modeling not provided by core TensorFlow\" (https://www.tensorflow.org/text/guide/tf_text_intro)\n", "_____no_output_____" ] ], [ [ "# A dependency of the preprocessing for BERT inputs\n!pip install -q -U tensorflow-text", "\u001b[K |████████████████████████████████| 4.4 MB 5.3 MB/s \n\u001b[?25h" ] ], [ [ "We will use the AdamW optimizer from [tensorflow/models](https://github.com/tensorflow/models). \"TensorFlow Model Garden is a repository with a number of different implementations of state-of-the-art (SOTA) models and modeling solutions for TensorFlow users.\"\n", "_____no_output_____" ] ], [ [ "!pip install -q tf-models-official", "\u001b[?25l\r\u001b[K |▏ | 10 kB 22.1 MB/s eta 0:00:01\r\u001b[K |▍ | 20 kB 28.4 MB/s eta 0:00:01\r\u001b[K |▋ | 30 kB 26.6 MB/s eta 0:00:01\r\u001b[K |▊ | 40 kB 19.5 MB/s eta 0:00:01\r\u001b[K |█ | 51 kB 5.7 MB/s eta 0:00:01\r\u001b[K |█▏ | 61 kB 6.1 MB/s eta 0:00:01\r\u001b[K |█▎ | 71 kB 5.6 MB/s eta 0:00:01\r\u001b[K |█▌ | 81 kB 6.3 MB/s eta 0:00:01\r\u001b[K |█▊ | 92 kB 6.5 MB/s eta 0:00:01\r\u001b[K |█▉ | 102 kB 5.3 MB/s eta 0:00:01\r\u001b[K |██ | 112 kB 5.3 MB/s eta 0:00:01\r\u001b[K |██▎ | 122 kB 5.3 MB/s eta 0:00:01\r\u001b[K |██▍ | 133 kB 5.3 MB/s eta 0:00:01\r\u001b[K |██▋ | 143 kB 5.3 MB/s eta 0:00:01\r\u001b[K |██▉ | 153 kB 5.3 MB/s eta 0:00:01\r\u001b[K |███ | 163 kB 5.3 MB/s eta 0:00:01\r\u001b[K |███▏ | 174 kB 5.3 MB/s eta 0:00:01\r\u001b[K |███▍ | 184 kB 5.3 MB/s eta 0:00:01\r\u001b[K |███▌ | 194 kB 5.3 MB/s eta 0:00:01\r\u001b[K |███▊ | 204 kB 5.3 MB/s eta 0:00:01\r\u001b[K |████ | 215 kB 5.3 MB/s eta 0:00:01\r\u001b[K |████ | 225 kB 5.3 MB/s eta 0:00:01\r\u001b[K |████▎ | 235 kB 5.3 MB/s eta 0:00:01\r\u001b[K |████▌ | 245 kB 5.3 MB/s eta 0:00:01\r\u001b[K |████▋ | 256 kB 5.3 MB/s eta 0:00:01\r\u001b[K |████▉ | 266 kB 5.3 MB/s eta 0:00:01\r\u001b[K |█████ | 276 kB 5.3 MB/s eta 0:00:01\r\u001b[K |█████▏ | 286 kB 5.3 MB/s eta 0:00:01\r\u001b[K |█████▍ | 296 kB 5.3 MB/s eta 0:00:01\r\u001b[K |█████▋ | 307 kB 5.3 MB/s eta 0:00:01\r\u001b[K |█████▊ | 317 kB 5.3 MB/s eta 0:00:01\r\u001b[K |██████ | 327 kB 5.3 MB/s eta 0:00:01\r\u001b[K |██████▏ | 337 kB 5.3 MB/s eta 0:00:01\r\u001b[K |██████▎ | 348 kB 5.3 MB/s eta 0:00:01\r\u001b[K |██████▌ | 358 kB 5.3 MB/s eta 0:00:01\r\u001b[K |██████▊ | 368 kB 5.3 MB/s eta 0:00:01\r\u001b[K |██████▉ | 378 kB 5.3 MB/s eta 0:00:01\r\u001b[K |███████ | 389 kB 5.3 MB/s eta 0:00:01\r\u001b[K |███████▎ | 399 kB 5.3 MB/s eta 0:00:01\r\u001b[K |███████▍ | 409 kB 5.3 MB/s eta 0:00:01\r\u001b[K |███████▋ | 419 kB 5.3 MB/s eta 0:00:01\r\u001b[K |███████▉ | 430 kB 5.3 MB/s eta 0:00:01\r\u001b[K |████████ | 440 kB 5.3 MB/s eta 0:00:01\r\u001b[K |████████▏ | 450 kB 5.3 MB/s eta 0:00:01\r\u001b[K |████████▍ | 460 kB 5.3 MB/s eta 0:00:01\r\u001b[K |████████▌ | 471 kB 5.3 MB/s eta 0:00:01\r\u001b[K |████████▊ | 481 kB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████ | 491 kB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████ | 501 kB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████▎ | 512 kB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████▌ | 522 kB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████▋ | 532 kB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████▉ | 542 kB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████ | 552 kB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████▏ | 563 kB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████▍ | 573 kB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████▋ | 583 kB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████▊ | 593 kB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████ | 604 kB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████▏ | 614 kB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████▎ | 624 kB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████▌ | 634 kB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████▊ | 645 kB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████▉ | 655 kB 5.3 MB/s eta 0:00:01\r\u001b[K |████████████ | 665 kB 5.3 MB/s eta 0:00:01\r\u001b[K |████████████▎ | 675 kB 5.3 MB/s eta 0:00:01\r\u001b[K |████████████▍ | 686 kB 5.3 MB/s eta 0:00:01\r\u001b[K |████████████▋ | 696 kB 5.3 MB/s eta 0:00:01\r\u001b[K |████████████▉ | 706 kB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████████ | 716 kB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████████▏ | 727 kB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████████▍ | 737 kB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████████▌ | 747 kB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████████▊ | 757 kB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████████ | 768 kB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████████ | 778 kB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████████▎ | 788 kB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████████▌ | 798 kB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████████▋ | 808 kB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████████▉ | 819 kB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████████ | 829 kB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████████▏ | 839 kB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████████▍ | 849 kB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████████▋ | 860 kB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████████▊ | 870 kB 5.3 MB/s eta 0:00:01\r\u001b[K |████████████████ | 880 kB 5.3 MB/s eta 0:00:01\r\u001b[K |████████████████▏ | 890 kB 5.3 MB/s eta 0:00:01\r\u001b[K |████████████████▎ | 901 kB 5.3 MB/s eta 0:00:01\r\u001b[K |████████████████▌ | 911 kB 5.3 MB/s eta 0:00:01\r\u001b[K |████████████████▊ | 921 kB 5.3 MB/s eta 0:00:01\r\u001b[K |████████████████▉ | 931 kB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████████████ | 942 kB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████████████▎ | 952 kB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████████████▍ | 962 kB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████████████▋ | 972 kB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████████████▉ | 983 kB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████████████ | 993 kB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████████████▏ | 1.0 MB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████████████▍ | 1.0 MB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████████████▋ | 1.0 MB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████████████▊ | 1.0 MB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████████████ | 1.0 MB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████████████▏ | 1.1 MB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████████████▎ | 1.1 MB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████████████▌ | 1.1 MB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████████████▊ | 1.1 MB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████████████▉ | 1.1 MB 5.3 MB/s eta 0:00:01\r\u001b[K |████████████████████ | 1.1 MB 5.3 MB/s eta 0:00:01\r\u001b[K |████████████████████▎ | 1.1 MB 5.3 MB/s eta 0:00:01\r\u001b[K |████████████████████▍ | 1.1 MB 5.3 MB/s eta 0:00:01\r\u001b[K |████████████████████▋ | 1.1 MB 5.3 MB/s eta 0:00:01\r\u001b[K |████████████████████▉ | 1.1 MB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████████████████ | 1.2 MB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████████████████▏ | 1.2 MB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████████████████▍ | 1.2 MB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████████████████▌ | 1.2 MB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████████████████▊ | 1.2 MB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████████████████ | 1.2 MB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████████████████ | 1.2 MB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████████████████▎ | 1.2 MB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████████████████▌ | 1.2 MB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████████████████▋ | 1.2 MB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████████████████▉ | 1.3 MB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████████████████ | 1.3 MB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████████████████▏ | 1.3 MB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████████████████▍ | 1.3 MB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████████████████▋ | 1.3 MB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████████████████▊ | 1.3 MB 5.3 MB/s eta 0:00:01\r\u001b[K |████████████████████████ | 1.3 MB 5.3 MB/s eta 0:00:01\r\u001b[K |████████████████████████▏ | 1.3 MB 5.3 MB/s eta 0:00:01\r\u001b[K |████████████████████████▎ | 1.3 MB 5.3 MB/s eta 0:00:01\r\u001b[K |████████████████████████▌ | 1.4 MB 5.3 MB/s eta 0:00:01\r\u001b[K |████████████████████████▊ | 1.4 MB 5.3 MB/s eta 0:00:01\r\u001b[K |████████████████████████▉ | 1.4 MB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████████████████████ | 1.4 MB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████████████████████▎ | 1.4 MB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████████████████████▍ | 1.4 MB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████████████████████▋ | 1.4 MB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████████████████████▉ | 1.4 MB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████████████████████ | 1.4 MB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████████████████████▏ | 1.4 MB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████████████████████▍ | 1.5 MB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████████████████████▌ | 1.5 MB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████████████████████▊ | 1.5 MB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████████████████████ | 1.5 MB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████████████████████ | 1.5 MB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████████████████████▎ | 1.5 MB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████████████████████▌ | 1.5 MB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████████████████████▋ | 1.5 MB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████████████████████▉ | 1.5 MB 5.3 MB/s eta 0:00:01\r\u001b[K |████████████████████████████ | 1.5 MB 5.3 MB/s eta 0:00:01\r\u001b[K |████████████████████████████▏ | 1.6 MB 5.3 MB/s eta 0:00:01\r\u001b[K |████████████████████████████▍ | 1.6 MB 5.3 MB/s eta 0:00:01\r\u001b[K |████████████████████████████▋ | 1.6 MB 5.3 MB/s eta 0:00:01\r\u001b[K |████████████████████████████▊ | 1.6 MB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████████████████████████ | 1.6 MB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████████████████████████▏ | 1.6 MB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████████████████████████▎ | 1.6 MB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████████████████████████▌ | 1.6 MB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████████████████████████▊ | 1.6 MB 5.3 MB/s eta 0:00:01\r\u001b[K |█████████████████████████████▉ | 1.6 MB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████████████████████████ | 1.7 MB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████████████████████████▎ | 1.7 MB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████████████████████████▍ | 1.7 MB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████████████████████████▋ | 1.7 MB 5.3 MB/s eta 0:00:01\r\u001b[K |██████████████████████████████▉ | 1.7 MB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████████████████████████ | 1.7 MB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████████████████████████▏| 1.7 MB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████████████████████████▍| 1.7 MB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████████████████████████▌| 1.7 MB 5.3 MB/s eta 0:00:01\r\u001b[K |███████████████████████████████▊| 1.8 MB 5.3 MB/s eta 0:00:01\r\u001b[K |████████████████████████████████| 1.8 MB 5.3 MB/s eta 0:00:01\r\u001b[K |████████████████████████████████| 1.8 MB 5.3 MB/s \n\u001b[K |████████████████████████████████| 211 kB 57.8 MB/s \n\u001b[K |████████████████████████████████| 90 kB 8.5 MB/s \n\u001b[K |████████████████████████████████| 99 kB 10.8 MB/s \n\u001b[K |████████████████████████████████| 1.2 MB 29.8 MB/s \n\u001b[K |████████████████████████████████| 636 kB 48.8 MB/s \n\u001b[K |████████████████████████████████| 1.1 MB 34.2 MB/s \n\u001b[K |████████████████████████████████| 43 kB 2.3 MB/s \n\u001b[K |████████████████████████████████| 352 kB 45.4 MB/s \n\u001b[K |████████████████████████████████| 37.1 MB 43 kB/s \n\u001b[?25h Building wheel for py-cpuinfo (setup.py) ... \u001b[?25l\u001b[?25hdone\n Building wheel for seqeval (setup.py) ... \u001b[?25l\u001b[?25hdone\n" ], [ "%tensorflow_version 2.x\nimport tensorflow as tf\ndevice_name = tf.test.gpu_device_name()\nif device_name != '/device:GPU:0':\n raise SystemError('GPU device not found')\nprint('Found GPU at: {}'.format(device_name))", "Found GPU at: /device:GPU:0\n" ], [ "import os\nimport shutil\n\nimport tensorflow_hub as hub # TFHub is a repository of trained machine learning models (https://www.tensorflow.org/hub)\nimport tensorflow_text as text\nfrom official.nlp import optimization # to create AdamW optimizer\n\nimport matplotlib.pyplot as plt\n\ntf.get_logger().setLevel('ERROR')", "_____no_output_____" ] ], [ [ "## Sentiment analysis\n\nThis notebook trains a sentiment analysis model to classify movie reviews as *positive* or *negative*, based on the text of the review.\n\nYou'll use the [Large Movie Review Dataset](https://ai.stanford.edu/~amaas/data/sentiment/) that contains the text of 50,000 movie reviews from the [Internet Movie Database](https://www.imdb.com/).", "_____no_output_____" ], [ "#Keras\n\"Keras is the high-level API of TensorFlow 2: an approachable, highly-productive interface for solving machine learning problems, with a focus on modern deep learning. It provides essential abstractions and building blocks for developing and shipping machine learning solutions with high iteration velocity\" (https://keras.io/about/). A useful hands-on colab detailing keras abstractions can be found [here](https://jaredwinick.github.io/what_is_tf_keras/). Additional details can be found [here](https://machinelearningmastery.com/tensorflow-tutorial-deep-learning-with-tf-keras/).", "_____no_output_____" ], [ "### Download the IMDB dataset\n\nLet's download and extract the dataset, then explore the directory structure.\n", "_____no_output_____" ] ], [ [ "url = 'https://ai.stanford.edu/~amaas/data/sentiment/aclImdb_v1.tar.gz'\n\ndataset = tf.keras.utils.get_file('aclImdb_v1.tar.gz', url,\n untar=True, cache_dir='.',\n cache_subdir='')\n\ndataset_dir = os.path.join(os.path.dirname(dataset), 'aclImdb')\nprint(dataset_dir)\n\n!ls ./aclImdb/train\n\ntrain_dir = os.path.join(dataset_dir, 'train')\n\n# remove unused folders to make it easier to load the data\nremove_dir = os.path.join(train_dir, 'unsup')\nshutil.rmtree(remove_dir)", "Downloading data from https://ai.stanford.edu/~amaas/data/sentiment/aclImdb_v1.tar.gz\n84131840/84125825 [==============================] - 5s 0us/step\n84140032/84125825 [==============================] - 5s 0us/step\n./aclImdb\nlabeledBow.feat pos\tunsupBow.feat urls_pos.txt\nneg\t\t unsup\turls_neg.txt urls_unsup.txt\n" ] ], [ [ "Next, you will use the `text_dataset_from_directory` utility to create a labeled `tf.data.Dataset`. The `tf.data.Dataset` API supports writing descriptive and efficient input pipelines. Dataset usage follows a common pattern:\n- Create a source dataset from your input data.\n- Apply dataset transformations to preprocess the data.\n- Iterate over the dataset and process the elements.\nMore details can be found [here](https://www.tensorflow.org/api_docs/python/tf/data/Dataset).\n\nThe IMDB dataset has already been divided into train and test, but it lacks a validation set. Let's create a validation set using an 80:20 split of the training data by using the `validation_split` argument below.\n\nNote: When using the `validation_split` and `subset` arguments, make sure to either specify a random seed, or to pass `shuffle=False`, so that the validation and training splits have no overlap.", "_____no_output_____" ], [ "#Prefetching\nPrefetching overlaps the preprocessing and model execution of a training step. While the model is executing training step s, the input pipeline is reading the data for step s+1. Doing so reduces the step time to the maximum (as opposed to the sum) of the training and the time it takes to extract the data.\n\nThe tf.data API provides the tf.data.Dataset.prefetch transformation. It can be used to decouple the time when data is produced from the time when data is consumed. In particular, the transformation uses a background thread and an internal buffer to prefetch elements from the input dataset ahead of the time they are requested. The number of elements to prefetch should be equal to (or possibly greater than) the number of batches consumed by a single training step. You could either manually tune this value, or set it to tf.data.AUTOTUNE, which will prompt the tf.data runtime to tune the value dynamically at runtime. [Source](https://www.tensorflow.org/guide/data_performance#prefetching)\n", "_____no_output_____" ] ], [ [ "AUTOTUNE = tf.data.AUTOTUNE\nbatch_size = 32\nseed = 42\n\nraw_train_ds = tf.keras.preprocessing.text_dataset_from_directory(\n 'aclImdb/train',\n batch_size=batch_size,\n validation_split=0.2,\n subset='training',\n seed=seed)\n\nclass_names = raw_train_ds.class_names\nprint(class_names)\ntrain_ds = raw_train_ds.cache().prefetch(buffer_size=AUTOTUNE)\n\nval_ds = tf.keras.preprocessing.text_dataset_from_directory(\n 'aclImdb/train',\n batch_size=batch_size,\n validation_split=0.2,\n subset='validation',\n seed=seed)\n\nval_ds = val_ds.cache().prefetch(buffer_size=AUTOTUNE)\n\ntest_ds = tf.keras.preprocessing.text_dataset_from_directory(\n 'aclImdb/test',\n batch_size=batch_size)\n\ntest_ds = test_ds.cache().prefetch(buffer_size=AUTOTUNE)", "Found 25000 files belonging to 2 classes.\nUsing 20000 files for training.\n['neg', 'pos']\nFound 25000 files belonging to 2 classes.\nUsing 5000 files for validation.\nFound 25000 files belonging to 2 classes.\n" ] ], [ [ "Let's take a look at a few reviews.", "_____no_output_____" ] ], [ [ "for text_batch, label_batch in train_ds.take(1):\n for i in range(3):\n print(f'Review: {text_batch.numpy()[i]}')\n label = label_batch.numpy()[i]\n print(f'Label : {label} ({class_names[label]})')", "Review: b'\"Pandemonium\" is a horror movie spoof that comes off more stupid than funny. Believe me when I tell you, I love comedies. Especially comedy spoofs. \"Airplane\", \"The Naked Gun\" trilogy, \"Blazing Saddles\", \"High Anxiety\", and \"Spaceballs\" are some of my favorite comedies that spoof a particular genre. \"Pandemonium\" is not up there with those films. Most of the scenes in this movie had me sitting there in stunned silence because the movie wasn\\'t all that funny. There are a few laughs in the film, but when you watch a comedy, you expect to laugh a lot more than a few times and that\\'s all this film has going for it. Geez, \"Scream\" had more laughs than this film and that was more of a horror film. How bizarre is that?<br /><br />*1/2 (out of four)'\nLabel : 0 (neg)\nReview: b\"David Mamet is a very interesting and a very un-equal director. His first movie 'House of Games' was the one I liked best, and it set a series of films with characters whose perspective of life changes as they get into complicated situations, and so does the perspective of the viewer.<br /><br />So is 'Homicide' which from the title tries to set the mind of the viewer to the usual crime drama. The principal characters are two cops, one Jewish and one Irish who deal with a racially charged area. The murder of an old Jewish shop owner who proves to be an ancient veteran of the Israeli Independence war triggers the Jewish identity in the mind and heart of the Jewish detective.<br /><br />This is were the flaws of the film are the more obvious. The process of awakening is theatrical and hard to believe, the group of Jewish militants is operatic, and the way the detective eventually walks to the final violent confrontation is pathetic. The end of the film itself is Mamet-like smart, but disappoints from a human emotional perspective.<br /><br />Joe Mantegna and William Macy give strong performances, but the flaws of the story are too evident to be easily compensated.\"\nLabel : 0 (neg)\nReview: b'Great documentary about the lives of NY firefighters during the worst terrorist attack of all time.. That reason alone is why this should be a must see collectors item.. What shocked me was not only the attacks, but the\"High Fat Diet\" and physical appearance of some of these firefighters. I think a lot of Doctors would agree with me that,in the physical shape they were in, some of these firefighters would NOT of made it to the 79th floor carrying over 60 lbs of gear. Having said that i now have a greater respect for firefighters and i realize becoming a firefighter is a life altering job. The French have a history of making great documentary\\'s and that is what this is, a Great Documentary.....'\nLabel : 1 (pos)\n" ] ], [ [ "## Loading models from TensorFlow Hub\n\nHere you can choose which BERT model you will load from TensorFlow Hub and fine-tune. There are multiple BERT models available.\n\n - [BERT-Base](https://tfhub.dev/tensorflow/bert_en_uncased_L-12_H-768_A-12/3), [Uncased](https://tfhub.dev/tensorflow/bert_en_uncased_L-12_H-768_A-12/3) and [seven more models](https://tfhub.dev/google/collections/bert/1) with trained weights released by the original BERT authors.\n - [Small BERTs](https://tfhub.dev/google/collections/bert/1) have the same general architecture but fewer and/or smaller Transformer blocks, which lets you explore tradeoffs between speed, size and quality.\n - [ALBERT](https://tfhub.dev/google/collections/albert/1): four different sizes of \"A Lite BERT\" that reduces model size (but not computation time) by sharing parameters between layers.\n\nThe model documentation on TensorFlow Hub has more details and references to the\nresearch literature. Follow the links above, or click on the [`tfhub.dev`](http://tfhub.dev) URL\nprinted after the next cell execution.\n\nThe suggestion is to start with a Small BERT (with fewer parameters) since they are faster to fine-tune. If you like a small model but with higher accuracy, ALBERT might be your next option. If you want even better accuracy, choose\none of the classic BERT sizes or their recent refinements like Electra, Talking Heads, or a BERT Expert.\n\nAside from the models available below, there are [multiple versions](https://tfhub.dev/google/collections/transformer_encoders_text/1) of the models that are larger and can yield even better accuracy, but they are too big to be fine-tuned on a single GPU. You will be able to do that on the [Solve GLUE tasks using BERT on a TPU colab](https://www.tensorflow.org/text/tutorials/bert_glue).\n\nYou'll see in the code below that switching the tfhub.dev URL is enough to try any of these models, because all the differences between them are encapsulated in the SavedModels from TF Hub.", "_____no_output_____" ] ], [ [ "#@title Choose a BERT model to fine-tune\n\nbert_model_name = \"small_bert/bert_en_uncased_L-4_H-512_A-8\" #@param [\"bert_en_uncased_L-12_H-768_A-12\", \"bert_en_cased_L-12_H-768_A-12\", \"bert_multi_cased_L-12_H-768_A-12\", \"small_bert/bert_en_uncased_L-2_H-128_A-2\", \"small_bert/bert_en_uncased_L-2_H-256_A-4\", \"small_bert/bert_en_uncased_L-2_H-512_A-8\", \"small_bert/bert_en_uncased_L-2_H-768_A-12\", \"small_bert/bert_en_uncased_L-4_H-128_A-2\", \"small_bert/bert_en_uncased_L-4_H-256_A-4\", \"small_bert/bert_en_uncased_L-4_H-512_A-8\", \"small_bert/bert_en_uncased_L-4_H-768_A-12\", \"small_bert/bert_en_uncased_L-6_H-128_A-2\", \"small_bert/bert_en_uncased_L-6_H-256_A-4\", \"small_bert/bert_en_uncased_L-6_H-512_A-8\", \"small_bert/bert_en_uncased_L-6_H-768_A-12\", \"small_bert/bert_en_uncased_L-8_H-128_A-2\", \"small_bert/bert_en_uncased_L-8_H-256_A-4\", \"small_bert/bert_en_uncased_L-8_H-512_A-8\", \"small_bert/bert_en_uncased_L-8_H-768_A-12\", \"small_bert/bert_en_uncased_L-10_H-128_A-2\", \"small_bert/bert_en_uncased_L-10_H-256_A-4\", \"small_bert/bert_en_uncased_L-10_H-512_A-8\", \"small_bert/bert_en_uncased_L-10_H-768_A-12\", \"small_bert/bert_en_uncased_L-12_H-128_A-2\", \"small_bert/bert_en_uncased_L-12_H-256_A-4\", \"small_bert/bert_en_uncased_L-12_H-512_A-8\", \"small_bert/bert_en_uncased_L-12_H-768_A-12\", \"albert_en_base\", \"electra_small\", \"electra_base\", \"experts_pubmed\", \"experts_wiki_books\", \"talking-heads_base\"]\n\nmap_name_to_handle = {\n 'bert_en_uncased_L-12_H-768_A-12':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_L-12_H-768_A-12/3',\n 'bert_en_cased_L-12_H-768_A-12':\n 'https://tfhub.dev/tensorflow/bert_en_cased_L-12_H-768_A-12/3',\n 'bert_multi_cased_L-12_H-768_A-12':\n 'https://tfhub.dev/tensorflow/bert_multi_cased_L-12_H-768_A-12/3',\n 'small_bert/bert_en_uncased_L-2_H-128_A-2':\n 'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-2_H-128_A-2/1',\n 'small_bert/bert_en_uncased_L-2_H-256_A-4':\n 'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-2_H-256_A-4/1',\n 'small_bert/bert_en_uncased_L-2_H-512_A-8':\n 'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-2_H-512_A-8/1',\n 'small_bert/bert_en_uncased_L-2_H-768_A-12':\n 'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-2_H-768_A-12/1',\n 'small_bert/bert_en_uncased_L-4_H-128_A-2':\n 'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-4_H-128_A-2/1',\n 'small_bert/bert_en_uncased_L-4_H-256_A-4':\n 'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-4_H-256_A-4/1',\n 'small_bert/bert_en_uncased_L-4_H-512_A-8':\n 'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-4_H-512_A-8/1',\n 'small_bert/bert_en_uncased_L-4_H-768_A-12':\n 'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-4_H-768_A-12/1',\n 'small_bert/bert_en_uncased_L-6_H-128_A-2':\n 'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-6_H-128_A-2/1',\n 'small_bert/bert_en_uncased_L-6_H-256_A-4':\n 'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-6_H-256_A-4/1',\n 'small_bert/bert_en_uncased_L-6_H-512_A-8':\n 'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-6_H-512_A-8/1',\n 'small_bert/bert_en_uncased_L-6_H-768_A-12':\n 'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-6_H-768_A-12/1',\n 'small_bert/bert_en_uncased_L-8_H-128_A-2':\n 'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-8_H-128_A-2/1',\n 'small_bert/bert_en_uncased_L-8_H-256_A-4':\n 'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-8_H-256_A-4/1',\n 'small_bert/bert_en_uncased_L-8_H-512_A-8':\n 'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-8_H-512_A-8/1',\n 'small_bert/bert_en_uncased_L-8_H-768_A-12':\n 'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-8_H-768_A-12/1',\n 'small_bert/bert_en_uncased_L-10_H-128_A-2':\n 'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-10_H-128_A-2/1',\n 'small_bert/bert_en_uncased_L-10_H-256_A-4':\n 'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-10_H-256_A-4/1',\n 'small_bert/bert_en_uncased_L-10_H-512_A-8':\n 'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-10_H-512_A-8/1',\n 'small_bert/bert_en_uncased_L-10_H-768_A-12':\n 'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-10_H-768_A-12/1',\n 'small_bert/bert_en_uncased_L-12_H-128_A-2':\n 'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-12_H-128_A-2/1',\n 'small_bert/bert_en_uncased_L-12_H-256_A-4':\n 'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-12_H-256_A-4/1',\n 'small_bert/bert_en_uncased_L-12_H-512_A-8':\n 'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-12_H-512_A-8/1',\n 'small_bert/bert_en_uncased_L-12_H-768_A-12':\n 'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-12_H-768_A-12/1',\n 'albert_en_base':\n 'https://tfhub.dev/tensorflow/albert_en_base/2',\n 'electra_small':\n 'https://tfhub.dev/google/electra_small/2',\n 'electra_base':\n 'https://tfhub.dev/google/electra_base/2',\n 'experts_pubmed':\n 'https://tfhub.dev/google/experts/bert/pubmed/2',\n 'experts_wiki_books':\n 'https://tfhub.dev/google/experts/bert/wiki_books/2',\n 'talking-heads_base':\n 'https://tfhub.dev/tensorflow/talkheads_ggelu_bert_en_base/1',\n}\n\nmap_model_to_preprocess = {\n 'bert_en_uncased_L-12_H-768_A-12':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n 'bert_en_cased_L-12_H-768_A-12':\n 'https://tfhub.dev/tensorflow/bert_en_cased_preprocess/3',\n 'small_bert/bert_en_uncased_L-2_H-128_A-2':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n 'small_bert/bert_en_uncased_L-2_H-256_A-4':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n 'small_bert/bert_en_uncased_L-2_H-512_A-8':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n 'small_bert/bert_en_uncased_L-2_H-768_A-12':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n 'small_bert/bert_en_uncased_L-4_H-128_A-2':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n 'small_bert/bert_en_uncased_L-4_H-256_A-4':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n 'small_bert/bert_en_uncased_L-4_H-512_A-8':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n 'small_bert/bert_en_uncased_L-4_H-768_A-12':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n 'small_bert/bert_en_uncased_L-6_H-128_A-2':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n 'small_bert/bert_en_uncased_L-6_H-256_A-4':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n 'small_bert/bert_en_uncased_L-6_H-512_A-8':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n 'small_bert/bert_en_uncased_L-6_H-768_A-12':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n 'small_bert/bert_en_uncased_L-8_H-128_A-2':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n 'small_bert/bert_en_uncased_L-8_H-256_A-4':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n 'small_bert/bert_en_uncased_L-8_H-512_A-8':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n 'small_bert/bert_en_uncased_L-8_H-768_A-12':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n 'small_bert/bert_en_uncased_L-10_H-128_A-2':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n 'small_bert/bert_en_uncased_L-10_H-256_A-4':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n 'small_bert/bert_en_uncased_L-10_H-512_A-8':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n 'small_bert/bert_en_uncased_L-10_H-768_A-12':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n 'small_bert/bert_en_uncased_L-12_H-128_A-2':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n 'small_bert/bert_en_uncased_L-12_H-256_A-4':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n 'small_bert/bert_en_uncased_L-12_H-512_A-8':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n 'small_bert/bert_en_uncased_L-12_H-768_A-12':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n 'bert_multi_cased_L-12_H-768_A-12':\n 'https://tfhub.dev/tensorflow/bert_multi_cased_preprocess/3',\n 'albert_en_base':\n 'https://tfhub.dev/tensorflow/albert_en_preprocess/3',\n 'electra_small':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n 'electra_base':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n 'experts_pubmed':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n 'experts_wiki_books':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n 'talking-heads_base':\n 'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',\n}\n\ntfhub_handle_encoder = map_name_to_handle[bert_model_name]\ntfhub_handle_preprocess = map_model_to_preprocess[bert_model_name]\n\nprint(f'BERT model selected : {tfhub_handle_encoder}')\nprint(f'Preprocess model auto-selected: {tfhub_handle_preprocess}')", "BERT model selected : https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-4_H-512_A-8/1\nPreprocess model auto-selected: https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3\n" ] ], [ [ "## The preprocessing model\n\nText inputs need to be transformed to numeric token ids and arranged in several Tensors before being input to BERT. TensorFlow Hub provides a matching preprocessing model for each of the BERT models discussed above, which implements this transformation using TF ops from the TF.text library. It is not necessary to run pure Python code outside your TensorFlow model to preprocess text.\n\nThe preprocessing model must be the one referenced by the documentation of the BERT model, which you can read at the URL printed above. For BERT models from the drop-down above, the preprocessing model is selected automatically.\n\nNote: You will load the preprocessing model into a [hub.KerasLayer](https://www.tensorflow.org/hub/api_docs/python/hub/KerasLayer) to compose your fine-tuned model. More information on Keras layers can be found [here](https://keras.io/api/layers/). \"Layers are the basic building blocks of neural networks in Keras. A layer consists of a tensor-in tensor-out computation function (the layer's call method) and some state, held in TensorFlow variables (the layer's weights).\"\n\n[hub.KerasLayer](https://www.tensorflow.org/hub/api_docs/python/hub/KerasLayer) is the preferred API to load a TF2-style [SavedModel](https://www.tensorflow.org/guide/saved_model) from TF Hub into a Keras model.", "_____no_output_____" ] ], [ [ "bert_preprocess_model = hub.KerasLayer(tfhub_handle_preprocess)", "_____no_output_____" ] ], [ [ "Let's try the preprocessing model on some text and see the output:", "_____no_output_____" ] ], [ [ "text_test = ['this is such an amazing movie!']\ntext_preprocessed = bert_preprocess_model(text_test)\n\nprint(f'Keys : {list(text_preprocessed.keys())}')\nprint(f'Shape : {text_preprocessed[\"input_word_ids\"].shape}')\nprint(f'Word Ids : {text_preprocessed[\"input_word_ids\"][0, :12]}')\nprint(f'Input Mask : {text_preprocessed[\"input_mask\"][0, :12]}')\nprint(f'Type Ids : {text_preprocessed[\"input_type_ids\"][0, :12]}')", "Keys : ['input_type_ids', 'input_mask', 'input_word_ids']\nShape : (1, 128)\nWord Ids : [ 101 2023 2003 2107 2019 6429 3185 999 102 0 0 0]\nInput Mask : [1 1 1 1 1 1 1 1 1 0 0 0]\nType Ids : [0 0 0 0 0 0 0 0 0 0 0 0]\n" ] ], [ [ "As you can see, now you have the 3 outputs from the preprocessing that a BERT model would use (`input_words_id`, `input_mask` and `input_type_ids`).\n\nSome other important points:\n- The input is truncated to 128 tokens. The number of tokens can be customized, and you can see more details on the [Solve GLUE tasks using BERT on a TPU colab](https://www.tensorflow.org/text/tutorials/bert_glue).\n- The `input_type_ids` only have one value (0) because this is a single sentence input. For a multiple sentence input, it would have one number for each input.\n\nSince this text preprocessor is a TensorFlow model, It can be included in your model directly.", "_____no_output_____" ], [ "## Using the BERT model\n\nBefore putting BERT into your own model, let's take a look at its outputs. You will load it from TF Hub and see the returned values.", "_____no_output_____" ] ], [ [ "bert_model = hub.KerasLayer(tfhub_handle_encoder)", "_____no_output_____" ], [ "bert_results = bert_model(text_preprocessed)\n\nprint(f'Loaded BERT: {tfhub_handle_encoder}')\nprint(f'Pooled Outputs Shape:{bert_results[\"pooled_output\"].shape}')\nprint(f'Pooled Outputs Values:{bert_results[\"pooled_output\"][0, :12]}')\nprint(f'Sequence Outputs Shape:{bert_results[\"sequence_output\"].shape}')\nprint(f'Sequence Outputs Values:{bert_results[\"sequence_output\"][0, :12]}')", "Loaded BERT: https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-4_H-512_A-8/1\nPooled Outputs Shape:(1, 512)\nPooled Outputs Values:[ 0.7626282 0.9928099 -0.18611862 0.3667383 0.15233758 0.655044\n 0.9681154 -0.94862705 0.0021616 -0.9877732 0.06842764 -0.97630596]\nSequence Outputs Shape:(1, 128, 512)\nSequence Outputs Values:[[-0.28946292 0.34321183 0.33231512 ... 0.21300802 0.7102092\n -0.05771042]\n [-0.28741995 0.31980985 -0.23018652 ... 0.5845511 -0.21329862\n 0.72692007]\n [-0.6615692 0.68876815 -0.8743301 ... 0.1087728 -0.26173076\n 0.47855455]\n ...\n [-0.22561137 -0.2892573 -0.07064426 ... 0.47566032 0.8327724\n 0.40025347]\n [-0.2982421 -0.27473164 -0.05450544 ... 0.4884972 1.0955367\n 0.18163365]\n [-0.4437818 0.00930662 0.07223704 ... 0.17290089 1.1833239\n 0.07897975]]\n" ] ], [ [ "The BERT models return a map with 3 important keys: `pooled_output`, `sequence_output`, `encoder_outputs`:\n\n- `pooled_output` represents each input sequence as a whole. The shape is `[batch_size, H]`. You can think of this as an embedding for the entire movie review.\n- `sequence_output` represents each input token in the context. The shape is `[batch_size, seq_length, H]`. You can think of this as a contextual embedding for every token in the movie review.\n- `encoder_outputs` are the intermediate activations of the `L` Transformer blocks. `outputs[\"encoder_outputs\"][i]` is a Tensor of shape `[batch_size, seq_length, 1024]` with the outputs of the i-th Transformer block, for `0 <= i < L`. The last value of the list is equal to `sequence_output`.\n\nFor the fine-tuning you are going to use the `pooled_output` array.", "_____no_output_____" ], [ "## Define your model\n\nYou will create a very simple fine-tuned model, with the preprocessing model, the selected BERT model, one Dense and a Dropout layer.\n\nNote: for more information about the base model's input and output you can follow the model's URL for documentation. Here specifically, you don't need to worry about it because the preprocessing model will take care of that for you.\n", "_____no_output_____" ] ], [ [ "def build_classifier_model():\n text_input = tf.keras.layers.Input(shape=(), dtype=tf.string, name='text')\n preprocessing_layer = hub.KerasLayer(tfhub_handle_preprocess, name='preprocessing')\n encoder_inputs = preprocessing_layer(text_input)\n encoder = hub.KerasLayer(tfhub_handle_encoder, trainable=True, name='BERT_encoder')\n outputs = encoder(encoder_inputs)\n net = outputs['pooled_output']\n net = tf.keras.layers.Dropout(0.1)(net)\n net = tf.keras.layers.Dense(1, activation=None, name='classifier')(net)\n return tf.keras.Model(text_input, net)", "_____no_output_____" ] ], [ [ "Let's check that the model runs with the output of the preprocessing model.", "_____no_output_____" ] ], [ [ "classifier_model = build_classifier_model()\nbert_raw_result = classifier_model(tf.constant(text_test))\nprint(tf.sigmoid(bert_raw_result))", "tf.Tensor([[0.5890199]], shape=(1, 1), dtype=float32)\n" ] ], [ [ "The output is meaningless, of course, because the model has not been trained yet.\n\nLet's take a look at the model's structure.", "_____no_output_____" ] ], [ [ "tf.keras.utils.plot_model(classifier_model)", "_____no_output_____" ] ], [ [ "## Model training\n\nYou now have all the pieces to train a model, including the preprocessing module, BERT encoder, data, and classifier.", "_____no_output_____" ], [ "### Loss function\n\nSince this is a binary classification problem and the model outputs a probability (a single-unit layer), you'll use `losses.BinaryCrossentropy` loss function. More information on the BinaryCrossentropy loss can be found [here](https://www.tensorflow.org/api_docs/python/tf/keras/losses/BinaryCrossentropy).\n", "_____no_output_____" ] ], [ [ "loss = tf.keras.losses.BinaryCrossentropy(from_logits=True)\nmetrics = tf.metrics.BinaryAccuracy()", "_____no_output_____" ] ], [ [ "### Optimizer\n\nFor fine-tuning, let's use the same optimizer that BERT was originally trained with: the \"Adaptive Moments\" (Adam). This optimizer minimizes the prediction loss and does regularization by weight decay (not using moments), which is also known as [AdamW](https://arxiv.org/abs/1711.05101).\n\nFor the learning rate (`init_lr`), you will use the same schedule as BERT pre-training: linear decay of a notional initial learning rate, prefixed with a linear warm-up phase over the first 10% of training steps (`num_warmup_steps`). In line with the BERT paper, the initial learning rate is smaller for fine-tuning (best of 5e-5, 3e-5, 2e-5).", "_____no_output_____" ] ], [ [ "epochs = 3\nsteps_per_epoch = tf.data.experimental.cardinality(train_ds).numpy()\nnum_train_steps = steps_per_epoch * epochs\nnum_warmup_steps = int(0.1*num_train_steps)\n\ninit_lr = 3e-5\noptimizer = optimization.create_optimizer(init_lr=init_lr,\n num_train_steps=num_train_steps,\n num_warmup_steps=num_warmup_steps,\n optimizer_type='adamw')", "_____no_output_____" ] ], [ [ "### Loading the BERT model and training\n\nUsing the `classifier_model` you created earlier, you can compile the model with the loss, metric and optimizer.", "_____no_output_____" ] ], [ [ "classifier_model.compile(optimizer=optimizer,\n loss=loss,\n metrics=metrics)", "_____no_output_____" ] ], [ [ "Note: training time will vary depending on the complexity of the BERT model you have selected.", "_____no_output_____" ] ], [ [ "print(f'Training model with {tfhub_handle_encoder}')\nhistory = classifier_model.fit(x=train_ds,\n validation_data=val_ds,\n epochs=epochs)", "Training model with https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-4_H-512_A-8/1\nEpoch 1/3\n625/625 [==============================] - 290s 453ms/step - loss: 0.4603 - binary_accuracy: 0.7601 - val_loss: 0.3787 - val_binary_accuracy: 0.8374\nEpoch 2/3\n625/625 [==============================] - 279s 447ms/step - loss: 0.3213 - binary_accuracy: 0.8560 - val_loss: 0.3574 - val_binary_accuracy: 0.8452\nEpoch 3/3\n625/625 [==============================] - 280s 448ms/step - loss: 0.2537 - binary_accuracy: 0.8921 - val_loss: 0.3812 - val_binary_accuracy: 0.8478\n" ] ], [ [ "### Evaluate the model\n\nLet's see how the model performs. Two values will be returned. Loss (a number which represents the error, lower values are better), and accuracy.", "_____no_output_____" ] ], [ [ "loss, accuracy = classifier_model.evaluate(test_ds)\n\nprint(f'Loss: {loss}')\nprint(f'Accuracy: {accuracy}')", "782/782 [==============================] - 152s 195ms/step - loss: 0.3684 - binary_accuracy: 0.8516\nLoss: 0.3683711290359497\nAccuracy: 0.8515999913215637\n" ] ], [ [ "### Plot the accuracy and loss over time\n\nBased on the `History` object returned by `model.fit()`. You can plot the training and validation loss for comparison, as well as the training and validation accuracy. More information on the History object can be found [here](https://machinelearningmastery.com/display-deep-learning-model-training-history-in-keras/).", "_____no_output_____" ] ], [ [ "history_dict = history.history\nprint(history_dict.keys())\n\nacc = history_dict['binary_accuracy']\nval_acc = history_dict['val_binary_accuracy']\nloss = history_dict['loss']\nval_loss = history_dict['val_loss']\n\nepochs = range(1, len(acc) + 1)\nfig = plt.figure(figsize=(10, 6))\nfig.tight_layout()\n\nplt.subplot(2, 1, 1)\n# \"bo\" is for \"blue dot\"\nplt.plot(epochs, loss, 'r', label='Training loss')\n# b is for \"solid blue line\"\nplt.plot(epochs, val_loss, 'b', label='Validation loss')\nplt.title('Training and validation loss')\n# plt.xlabel('Epochs')\nplt.ylabel('Loss')\nplt.legend()\n\nplt.subplot(2, 1, 2)\nplt.plot(epochs, acc, 'r', label='Training acc')\nplt.plot(epochs, val_acc, 'b', label='Validation acc')\nplt.title('Training and validation accuracy')\nplt.xlabel('Epochs')\nplt.ylabel('Accuracy')\nplt.legend(loc='lower right')", "dict_keys(['loss', 'binary_accuracy', 'val_loss', 'val_binary_accuracy'])\n" ] ], [ [ "In this plot, the red lines represent the training loss and accuracy, and the blue lines are the validation loss and accuracy.", "_____no_output_____" ], [ "## Export for inference\n\nNow you just save your fine-tuned model for later use.", "_____no_output_____" ] ], [ [ "dataset_name = 'imdb'\nsaved_model_path = './{}_bert'.format(dataset_name.replace('/', '_'))\n\nclassifier_model.save(saved_model_path, include_optimizer=False)", "WARNING:absl:Found untraced functions such as restored_function_body, restored_function_body, restored_function_body, restored_function_body, restored_function_body while saving (showing 5 of 310). These functions will not be directly callable after loading.\n" ] ], [ [ "Let's reload the model, so you can try it side by side with the model that is still in memory.", "_____no_output_____" ] ], [ [ "reloaded_model = tf.saved_model.load(saved_model_path)", "_____no_output_____" ] ], [ [ "Here you can test your model on any sentence you want, just add to the examples variable below.", "_____no_output_____" ] ], [ [ "def print_my_examples(inputs, results):\n result_for_printing = \\\n [f'input: {inputs[i]:<30} : score: {results[i][0]:.6f}'\n for i in range(len(inputs))]\n print(*result_for_printing, sep='\\n')\n print()\n\n\nexamples = [\n 'this is such an amazing movie!', # this is the same sentence tried earlier\n 'The movie was great!',\n 'The movie was meh.',\n 'The movie was okish.',\n 'The movie was terrible...'\n]\n\nreloaded_results = tf.sigmoid(reloaded_model(tf.constant(examples)))\noriginal_results = tf.sigmoid(classifier_model(tf.constant(examples)))\n\nprint('Results from the saved model:')\nprint_my_examples(examples, reloaded_results)\nprint('Results from the model in memory:')\nprint_my_examples(examples, original_results)", "Results from the saved model:\ninput: this is such an amazing movie! : score: 0.998838\ninput: The movie was great! : score: 0.993884\ninput: The movie was meh. : score: 0.915520\ninput: The movie was okish. : score: 0.100049\ninput: The movie was terrible... : score: 0.003937\n\nResults from the model in memory:\ninput: this is such an amazing movie! : score: 0.998838\ninput: The movie was great! : score: 0.993884\ninput: The movie was meh. : score: 0.915520\ninput: The movie was okish. : score: 0.100049\ninput: The movie was terrible... : score: 0.003937\n\n" ], [ "", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Develop Model", "_____no_output_____" ], [ "In this noteook, we will go through the steps to load the ResNet152 model, pre-process the images to the required format and call the model to find the top predictions.", "_____no_output_____" ] ], [ [ "import PIL\nimport numpy as np\nimport torch\nimport torch.nn as nn\nimport torchvision\nimport wget\nfrom PIL import Image\nfrom torchvision import models, transforms", "_____no_output_____" ], [ "print(torch.__version__)\nprint(torchvision.__version__)", "0.4.1.post2\n0.2.1\n" ] ], [ [ "We download the synset for the model. This translates the output of the model to a specific label.", "_____no_output_____" ] ], [ [ "!wget \"http://data.dmlc.ml/mxnet/models/imagenet/synset.txt\"", "--2018-10-09 07:00:23-- http://data.dmlc.ml/mxnet/models/imagenet/synset.txt\nResolving data.dmlc.ml... 54.208.175.7\nConnecting to data.dmlc.ml|54.208.175.7|:80... connected.\nHTTP request sent, awaiting response... 200 OK\nLength: 31675 (31K) [text/plain]\nSaving to: ‘synset.txt.3’\n\nsynset.txt.3 100%[===================>] 30.93K --.-KB/s in 0.002s \n\n2018-10-09 07:00:24 (14.7 MB/s) - ‘synset.txt.3’ saved [31675/31675]\n\n" ] ], [ [ "We first load the model which we imported torchvision. This can take about 10s.", "_____no_output_____" ] ], [ [ "%%time\nmodel = models.resnet152(pretrained=True)", "CPU times: user 1.29 s, sys: 450 ms, total: 1.74 s\nWall time: 1.74 s\n" ] ], [ [ "You can print the summary of the model in the below cell. We cleared the output here for brevity. When you run the cell you should see a list of the layers and the size of the model in terms of number of parameters at the bottom of the output.", "_____no_output_____" ] ], [ [ "model=model.cuda()", "_____no_output_____" ], [ "print(model)\nprint('Number of parameters {}'.format(sum([param.view(-1).size()[0] for param in model.parameters()])))", "ResNet(\n (conv1): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)\n (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n (maxpool): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)\n (layer1): Sequential(\n (0): Bottleneck(\n (conv1): Conv2d(64, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n (downsample): Sequential(\n (0): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n )\n )\n (1): Bottleneck(\n (conv1): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (2): Bottleneck(\n (conv1): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(64, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n )\n (layer2): Sequential(\n (0): Bottleneck(\n (conv1): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n (downsample): Sequential(\n (0): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)\n (1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n )\n )\n (1): Bottleneck(\n (conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (2): Bottleneck(\n (conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (3): Bottleneck(\n (conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (4): Bottleneck(\n (conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (5): Bottleneck(\n (conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (6): Bottleneck(\n (conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (7): Bottleneck(\n (conv1): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(128, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n )\n (layer3): Sequential(\n (0): Bottleneck(\n (conv1): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n (downsample): Sequential(\n (0): Conv2d(512, 1024, kernel_size=(1, 1), stride=(2, 2), bias=False)\n (1): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n )\n )\n (1): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (2): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (3): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (4): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (5): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (6): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (7): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (8): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (9): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (10): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (11): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (12): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (13): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (14): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (15): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (16): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (17): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (18): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (19): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (20): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (21): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (22): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (23): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (24): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (25): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (26): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (27): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (28): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (29): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (30): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (31): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (32): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (33): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (34): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (35): Bottleneck(\n (conv1): Conv2d(1024, 256, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(256, 1024, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(1024, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n )\n (layer4): Sequential(\n (0): Bottleneck(\n (conv1): Conv2d(1024, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n (downsample): Sequential(\n (0): Conv2d(1024, 2048, kernel_size=(1, 1), stride=(2, 2), bias=False)\n (1): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n )\n )\n (1): Bottleneck(\n (conv1): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n (2): Bottleneck(\n (conv1): Conv2d(2048, 512, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (conv3): Conv2d(512, 2048, kernel_size=(1, 1), stride=(1, 1), bias=False)\n (bn3): BatchNorm2d(2048, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n (relu): ReLU(inplace)\n )\n )\n (avgpool): AvgPool2d(kernel_size=7, stride=1, padding=0)\n (fc): Linear(in_features=2048, out_features=1000, bias=True)\n)\nNumber of parameters 60192808\n" ] ], [ [ "Let's test our model with an image of a Lynx.", "_____no_output_____" ] ], [ [ "wget.download('https://upload.wikimedia.org/wikipedia/commons/thumb/6/68/Lynx_lynx_poing.jpg/220px-Lynx_lynx_poing.jpg')", "_____no_output_____" ], [ "img_path = '220px-Lynx_lynx_poing.jpg'\nprint(Image.open(img_path).size)\nImage.open(img_path)", "(220, 330)\n" ] ], [ [ "Below, we load the image. Then we compose transformation which resize the image to (224, 224) and then convert it to a PyTorch tensor and normalize the pixel values.", "_____no_output_____" ] ], [ [ "img = Image.open(img_path).convert('RGB')", "_____no_output_____" ], [ "preprocess_input = transforms.Compose([\n torchvision.transforms.Resize((224, 224), interpolation=PIL.Image.BICUBIC),\n transforms.ToTensor(),\n transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])\n])", "_____no_output_____" ], [ "img = Image.open(img_path)\nimg = preprocess_input(img)", "_____no_output_____" ] ], [ [ "Let's make a label look up function to make it easy to lookup the classes from the synset file", "_____no_output_____" ] ], [ [ "def create_label_lookup():\n with open('synset.txt', 'r') as f:\n label_list = [l.rstrip() for l in f]\n def _label_lookup(*label_locks):\n return [label_list[l] for l in label_locks]\n return _label_lookup", "_____no_output_____" ], [ "label_lookup = create_label_lookup()", "_____no_output_____" ] ], [ [ "We will apply softmax to the output of the model to get probabilities for each label", "_____no_output_____" ] ], [ [ "softmax = nn.Softmax(dim=1).cuda()", "_____no_output_____" ] ], [ [ "Now, let's call the model on our image to predict the top 3 labels. This will take a few seconds.", "_____no_output_____" ] ], [ [ "model = model.eval()", "_____no_output_____" ], [ "%%time\nwith torch.no_grad():\n img = img.unsqueeze(0)\n image_gpu = img.type(torch.float).cuda()\n outputs = model(image_gpu)\n probabilities = softmax(outputs)", "CPU times: user 416 ms, sys: 41.9 ms, total: 458 ms\nWall time: 76.6 ms\n" ], [ "label_lookup = create_label_lookup()", "_____no_output_____" ], [ "probabilities_numpy = probabilities.cpu().numpy().squeeze()", "_____no_output_____" ], [ "top_results = np.flip(np.sort(probabilities_numpy), 0)[:3]", "_____no_output_____" ], [ "labels = label_lookup(*np.flip(probabilities_numpy.argsort(),0)[:3])", "_____no_output_____" ], [ "dict(zip(labels, top_results))", "_____no_output_____" ] ], [ [ "The top guess is Lynx with probability about 99%. We can now move on to [developing the model api for our model](01_DevelopModelDriver.ipynb).", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown" ]

[ [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code" ], [ "markdown" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "### WHAT IS TORCH.NN REALLY", "_____no_output_____" ] ], [ [ "\"\"\"MINIST data setup\n\"\"\"", "_____no_output_____" ], [ "from pathlib import Path\n\nDATA_PATH = Path(\"./data\")\nPATH = DATA_PATH/\"mnist.pkl\"\n# PATH = DATA_PATH / \"mnist\"\n\n# PATH.mkdir(parents=True, exist_ok=True)\n\n# URL = \"http://deeplearning.net/data/mnist/\"\n# FILENAME = \"mnist.pkl.gz\"\n\n# if not (PATH / FILENAME).exists():\n# content = requests.get(URL + FILENAME).content\n# (PATH / FILENAME).open(\"wb\").write(content)\n\nimport pickle\nwith open(PATH.as_posix(), \"rb\") as f:\n ((x_train, y_train), (x_valid, y_valid), _) = pickle.load(f, encoding=\"latin-1\")", "_____no_output_____" ], [ "from matplotlib import pyplot\nimport numpy as np\n\npyplot.imshow(x_train[0].reshape((28, 28)), cmap=\"gray\")\nprint(x_train.shape)", "_____no_output_____" ], [ "import torch\n\nx_train, y_train, x_valid, y_valid = map(\n torch.tensor, (x_train, y_train, x_valid, y_valid)\n) # https://www.geeksforgeeks.org/python-map-function/\nn, c = x_train.shape\n# x_train, x_train.shape, y_train.min(), y_train.max()\nprint(x_train, y_train)\nprint(x_train.shape)\nprint(y_train.min(), y_train.max())", "_____no_output_____" ], [ "\"\"\"Neural net from scratch (no torch.nn)\n\"\"\"", "_____no_output_____" ], [ "import math\n\nweights = torch.randn(784, 10) / math.sqrt(784)\nweights.requires_grad_()\nbias = torch.zeros(10, requires_grad=True)", "_____no_output_____" ], [ "def log_softmax(x):\n # https://discuss.pytorch.org/t/what-is-the-difference-between-log-softmax-and-softmax/11801\n # https://stackoverflow.com/questions/44790670/torch-sum-a-tensor-along-an-axis\n # https://stackoverflow.com/questions/57237352/what-does-unsqueeze-do-in-pytorch\n # https://pytorch.org/docs/stable/notes/broadcasting.html\n return x - x.exp().sum(-1).log().unsqueeze(-1)\n\ndef model(xb):\n # https://stackoverflow.com/questions/5919530/what-is-the-pythonic-way-to-calculate-dot-product\n return log_softmax(xb @ weights + bias)", "_____no_output_____" ], [ "bs = 64 # batch size\n\nxb = x_train[0:bs] # a mini-batch from x\npreds = model(xb) # predictions\n# preds[0], preds.shape\nprint(preds[0], preds.shape)", "_____no_output_____" ], [ "# negative log-likelihood\n# https://stats.stackexchange.com/questions/198038/cross-entropy-or-log-likelihood-in-output-layer\ndef nll(input, target):\n # https://blog.csdn.net/u010496337/article/details/50574154\n return -input[range(target.shape[0]), target].mean()\n\nloss_func = nll", "_____no_output_____" ], [ "yb = y_train[0:bs]\nprint(loss_func(preds, yb))", "_____no_output_____" ], [ "def accuracy(out, yb):\n preds = torch.argmax(out, dim=1)\n return (preds==yb).float().mean() # Can only calculate the mean of floating types", "_____no_output_____" ], [ "print(accuracy(preds, yb))", "_____no_output_____" ], [ "# from IPython.core.debugger import set_trace\nlr = 0.5 # learning rate\nepochs = 2 # how many epochs to train for\n\nfor epoch in range(epochs):\n for i in range((n - 1) // bs + 1):\n # set_trace()\n start_i = i * bs\n end_i = start_i + bs\n xb = x_train[start_i:end_i]\n yb = y_train[start_i:end_i]\n pred = model(xb)\n loss = loss_func(pred, yb)\n\n loss.backward()\n with torch.no_grad():\n weights -= weights.grad * lr\n bias -= bias.grad * lr\n weights.grad.zero_()\n bias.grad.zero_()", "_____no_output_____" ], [ "print(loss_func(model(xb), yb), accuracy(model(xb), yb))", "_____no_output_____" ], [ "\"\"\"Using torch.nn.functional\n- making our code one or more of: shorter, more understandable, and/or more flexible.\n\"\"\"", "_____no_output_____" ], [ "# This module contains all the functions in the torch.nn library\n# (whereas other parts of the library contain classes)\nimport torch.nn.functional as F\n\nloss_func = F.cross_entropy\n\ndef model(xb):\n return xb @ weights + bias", "_____no_output_____" ], [ "print(loss_func(model(xb), yb), accuracy(model(xb), yb))", "_____no_output_____" ], [ "\"\"\"Refactor using nn.Module\n\"\"\"", "_____no_output_____" ], [ "from torch import nn\n\nclass Mnist_Logistic(nn.Module):\n def __init__(self):\n super().__init__()\n self.weights = nn.Parameter(torch.randn(784, 10) / math.sqrt(784))\n self.bias = nn.Parameter(torch.zeros(10))\n\n def forward(self, xb):\n return xb @ self.weights + self.bias", "_____no_output_____" ], [ "model = Mnist_Logistic()", "_____no_output_____" ], [ "print(loss_func(model(xb), yb))", "_____no_output_____" ], [ "def fit():\n for epoch in range(epochs):\n for i in range((n - 1) // bs + 1):\n start_i = i * bs\n end_i = start_i + bs\n xb = x_train[start_i:end_i]\n yb = y_train[start_i:end_i]\n pred = model(xb)\n loss = loss_func(pred, yb)\n\n loss.backward()\n with torch.no_grad():\n for p in model.parameters():\n p -= p.grad * lr\n model.zero_grad()\n\nfit()", "_____no_output_____" ], [ "print(loss_func(model(xb), yb))", "_____no_output_____" ], [ "\"\"\"Refactor using nn.Linear\n\"\"\"", "_____no_output_____" ], [ "class Mnist_Logistic(nn.Module):\n def __init__(self):\n super().__init__()\n self.lin = nn.Linear(784, 10)\n\n def forward(self, xb):\n return self.lin(xb)", "_____no_output_____" ], [ "model = Mnist_Logistic()\nfit()\n\nprint(loss_func(model(xb), yb))", "_____no_output_____" ], [ "\"\"\"Refactor using optim\n\"\"\"", "_____no_output_____" ], [ "from torch import optim\ndef get_model():\n model = Mnist_Logistic()\n return model, optim.SGD(model.parameters(), lr=lr)\n\nmodel, opt = get_model()\nprint(loss_func(model(xb), yb))\n\nfor epoch in range(epochs):\n for i in range((n - 1) // bs + 1):\n start_i = i * bs\n end_i = start_i + bs\n xb = x_train[start_i:end_i]\n yb = y_train[start_i:end_i]\n pred = model(xb)\n loss = loss_func(pred, yb)\n\n loss.backward()\n opt.step()\n opt.zero_grad()\n\nprint(loss_func(model(xb), yb))", "_____no_output_____" ], [ "\"\"\"Refactor using Dataset\n- PyTorch’s TensorDataset is a Dataset wrapping tensors. \n By defining a length and way of indexing, this also gives us a way to iterate, \n index, and slice along the first dimension of a tensor. \n This will make it easier to access both the independent \n and dependent variables in the same line as we train.\n\"\"\"", "_____no_output_____" ], [ "from torch.utils.data import TensorDataset\ntrain_ds = TensorDataset(x_train, y_train)\nmodel, opt = get_model()\n\nfor epoch in range(epochs):\n for i in range((n - 1) // bs + 1):\n xb, yb = train_ds[i * bs: i * bs + bs]\n pred = model(xb)\n loss = loss_func(pred, yb)\n\n loss.backward()\n opt.step()\n opt.zero_grad()\n\nprint(loss_func(model(xb), yb))", "_____no_output_____" ], [ "\"\"\"Refactor using DataLoader\n-Pytorch’s DataLoader is responsible for managing batches. \n You can create a DataLoader from any Dataset.\n DataLoader makes it easier to iterate over batches. \n Rather than having to use train_ds[i*bs : i*bs+bs], \n the DataLoader gives us each minibatch automatically.\n\"\"\"", "_____no_output_____" ], [ "from torch.utils.data import DataLoader\n\ntrain_ds = TensorDataset(x_train, y_train)\ntrain_dl = DataLoader(train_ds, batch_size=bs)\n\nmodel, opt = get_model()\n\nfor epoch in range(epochs):\n for xb, yb in train_dl:\n pred = model(xb)\n loss = loss_func(pred, yb)\n\n loss.backward()\n opt.step()\n opt.zero_grad()\n\nprint(loss_func(model(xb), yb))", "_____no_output_____" ], [ "\"\"\"Add Validation\n- Shuffling the training data is important to prevent correlation between batches and overfitting\n\"\"\"", "_____no_output_____" ], [ "train_ds = TensorDataset(x_train, y_train)\ntrain_dl = DataLoader(train_ds, batch_size=bs, shuffle=True)\n\nvalid_ds = TensorDataset(x_valid, y_valid)\nvalid_dl = DataLoader(valid_ds, batch_size=bs * 2)\n\nmodel, opt = get_model()\n\nfor epoch in range(epochs):\n # Note that we always call model.train() before training, and model.eval() before inference, \n # because these are used by layers such as nn.BatchNorm2d and nn.Dropout \n # to ensure appropriate behaviour for these different phases\n model.train()\n for xb, yb in train_dl:\n pred = model(xb)\n loss = loss_func(pred, yb)\n\n loss.backward()\n opt.step()\n opt.zero_grad()\n\n model.eval()\n with torch.no_grad():\n valid_loss = sum(loss_func(model(xb), yb) for xb, yb in valid_dl)\n\n print(epoch, valid_loss / len(valid_dl))", "_____no_output_____" ], [ "\"\"\"Create fit() and get_data()\n\"\"\"", "_____no_output_____" ], [ "def loss_batch(model, loss_func, xb, yb, opt=None):\n loss = loss_func(model(xb), yb)\n\n if opt is not None:\n loss.backward()\n opt.step()\n opt.zero_grad()\n\n return loss.item(), len(xb)", "_____no_output_____" ], [ "def get_data(train_ds, valid_ds, bs):\n return (\n DataLoader(train_ds, batch_size=bs, shuffle=True),\n DataLoader(valid_ds, batch_size=bs * 2),\n )", "_____no_output_____" ], [ "import numpy as np\n\ndef fit(epochs, model, loss_func, opt, train_dl, valid_dl):\n for epoch in range(epochs):\n model.train()\n for xb, yb in train_dl:\n loss_batch(model, loss_func, xb, yb, opt)\n\n model.eval()\n with torch.no_grad():\n losses, nums = zip(\n *[loss_batch(model, loss_func, xb, yb) for xb, yb in valid_dl]\n )\n val_loss = np.sum(np.multiply(losses, nums)) / np.sum(nums)\n\n print(epoch, val_loss)", "_____no_output_____" ], [ "train_dl, valid_dl = get_data(train_ds, valid_ds, bs)\nmodel, opt = get_model()\nfit(epochs, model, loss_func, opt, train_dl, valid_dl)", "_____no_output_____" ], [ "\"\"\"Switch to CNN\n\"\"\"", "_____no_output_____" ], [ "class Mnist_CNN(nn.Module):\n def __init__(self):\n super().__init__()\n self.conv1 = nn.Conv2d(1, 16, kernel_size=3, stride=2, padding=1)\n self.conv2 = nn.Conv2d(16, 16, kernel_size=3, stride=2, padding=1)\n self.conv3 = nn.Conv2d(16, 10, kernel_size=3, stride=2, padding=1)\n\n def forward(self, xb):\n xb = xb.view(-1, 1, 28, 28)\n xb = F.relu(self.conv1(xb))\n xb = F.relu(self.conv2(xb))\n xb = F.relu(self.conv3(xb))\n xb = F.avg_pool2d(xb, 4)\n return xb.view(-1, xb.size(1))\n\nlr = 0.1\nmodel = Mnist_CNN()\nopt = optim.SGD(model.parameters(), lr=lr, momentum=0.9)\n\nfit(epochs, model, loss_func, opt, train_dl, valid_dl)", "_____no_output_____" ], [ "\"\"\"nn Sequential\n\"\"\"", "_____no_output_____" ], [ "class Lambda(nn.Module):\n def __init__(self, func):\n super().__init__()\n self.func = func\n\n def forward(self, x):\n return self.func(x)\n\n\ndef preprocess(x):\n return x.view(-1, 1, 28, 28)", "_____no_output_____" ], [ "model = nn.Sequential(\n Lambda(preprocess),\n nn.Conv2d(1, 16, kernel_size=3, stride=2, padding=1),\n nn.ReLU(),\n nn.Conv2d(16, 16, kernel_size=3, stride=2, padding=1),\n nn.ReLU(),\n nn.Conv2d(16, 10, kernel_size=3, stride=2, padding=1),\n nn.ReLU(),\n nn.AvgPool2d(4),\n Lambda(lambda x: x.view(x.size(0), -1)),\n)\n\nopt = optim.SGD(model.parameters(), lr=lr, momentum=0.9)\n\nfit(epochs, model, loss_func, opt, train_dl, valid_dl)", "_____no_output_____" ], [ "\"\"\"Wrapping DataLoader\n\"\"\"", "_____no_output_____" ], [ "def preprocess(x, y):\n return x.view(-1, 1, 28, 28), y\n\n\nclass WrappedDataLoader:\n def __init__(self, dl, func):\n self.dl = dl\n self.func = func\n\n def __len__(self):\n return len(self.dl)\n\n def __iter__(self):\n batches = iter(self.dl)\n for b in batches:\n yield (self.func(*b))\n\ntrain_dl, valid_dl = get_data(train_ds, valid_ds, bs)\ntrain_dl = WrappedDataLoader(train_dl, preprocess)\nvalid_dl = WrappedDataLoader(valid_dl, preprocess)", "_____no_output_____" ], [ "model = nn.Sequential(\n nn.Conv2d(1, 16, kernel_size=3, stride=2, padding=1),\n nn.ReLU(),\n nn.Conv2d(16, 16, kernel_size=3, stride=2, padding=1),\n nn.ReLU(),\n nn.Conv2d(16, 10, kernel_size=3, stride=2, padding=1),\n nn.ReLU(),\n nn.AdaptiveAvgPool2d(1),\n Lambda(lambda x: x.view(x.size(0), -1)),\n)\n\nopt = optim.SGD(model.parameters(), lr=lr, momentum=0.9)\nfit(epochs, model, loss_func, opt, train_dl, valid_dl)", "_____no_output_____" ], [ "\"\"\"Using your GPU\n\"\"\"", "_____no_output_____" ], [ "dev = torch.device(\n \"cuda\") if torch.cuda.is_available() else torch.device(\"cpu\")", "_____no_output_____" ], [ "def preprocess(x, y):\n return x.view(-1, 1, 28, 28).to(dev), y.to(dev)\n\n\ntrain_dl, valid_dl = get_data(train_ds, valid_ds, bs)\ntrain_dl = WrappedDataLoader(train_dl, preprocess)\nvalid_dl = WrappedDataLoader(valid_dl, preprocess)", "_____no_output_____" ], [ "model.to(dev)\nopt = optim.SGD(model.parameters(), lr=lr, momentum=0.9)\nfit(epochs, model, loss_func, opt, train_dl, valid_dl)", "_____no_output_____" ] ] ]

[ "markdown", "code" ]

[ [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Integrated Project #1: Video Game", "_____no_output_____" ], [ "The goal of this project is to:", "_____no_output_____" ] ], [ [ "import pandas as pd\nimport numpy as np\nfrom scipy import stats as st\nimport matplotlib.pyplot as plt\nimport matplotlib.patches as mpatches\nimport re, math", "_____no_output_____" ] ], [ [ "## Project Description", "_____no_output_____" ], [ "You work for the online store Ice, which sells video games all over the world. User and expert reviews, genres, platforms (e.g. Xbox or PlayStation), and historical data on game sales are available from open sources. You need to identify patterns that determine whether a game succeeds or not. This will allow you to spot potential big winners and plan advertising campaigns.\n\nIn front of you is data going back to 2016. Let’s imagine that it’s December 2016 and you’re planning a campaign for 2017.\n\n(The important thing is to get experience working with data. It doesn't really matter whether you're forecasting 2017 sales based on data from 2016 or 2027 sales based on data from 2026.)\n\nThe dataset contains the abbreviation ESRB. The Entertainment Software Rating Board evaluates a game's content and assigns an age rating such as Teen or Mature.", "_____no_output_____" ], [ "## Table of Contents", "_____no_output_____" ], [ "- [The Goal](#goal)\n- [Step 0](#imports): Imports\n- [Step 1](#step1): Open the data file and study the general information\n - [Step 1 conclusion](#step1con)\n- [Step 2](#step2): Prepare the data\n - [Names](#step2name)\n - [Year of Release](#step2year)\n - [Sales](#step2sales)\n - [Score](#step2scores)\n - [Ratings](#step2ratings)\n - [Step 2 conclusion](#step2con)\n- [Step 3](#step3): Analyze the data\n - [Step 3 conclusion](#step3con)\n- [Step 4](#step4): Analyze the data\n - [Step 4 conclusion](#step4con)\n- [Step 5](#step5): Test the hypotheses\n - [Hypothesis 1](#step5h1): The average revenue from users of Ultimate and Surf calling plans differs\n - [Hypothesis 2](#step5h2): The average revenue from users in NY-NJ area is different from that of the users from other regions\n - [Step 5 conclusion](#step5con)\n- [Step 6](#step6): Write an overall conclusion", "_____no_output_____" ], [ "### Step 1. Open the data file and study the general information\n<a id='step1'></a>", "_____no_output_____" ] ], [ [ "raw_games_data = pd.read_csv('/datasets/games.csv')\ngames_data = raw_games_data\ngames_data.info()", "<class 'pandas.core.frame.DataFrame'>\nRangeIndex: 16715 entries, 0 to 16714\nData columns (total 11 columns):\nName 16713 non-null object\nPlatform 16715 non-null object\nYear_of_Release 16446 non-null float64\nGenre 16713 non-null object\nNA_sales 16715 non-null float64\nEU_sales 16715 non-null float64\nJP_sales 16715 non-null float64\nOther_sales 16715 non-null float64\nCritic_Score 8137 non-null float64\nUser_Score 10014 non-null object\nRating 9949 non-null object\ndtypes: float64(6), object(5)\nmemory usage: 1.4+ MB\n" ] ], [ [ "#### Step 1 conclusion\n<a id='step1con'></a>", "_____no_output_____" ], [ "We do have some nulls, and in some columns, such as the Ratings, there are a lot. To work with the information, we will need to replace the column names with lowercase text, and acknowledge the following issues:\n\nName:\n- There are two nulls, and in the information, they are also lacking genres, critic/user scores, and an ESRB rating. Because there are only 2 of 16715 entries, these should be removed.\n\nYear_of_Release:\n- We need to fill in the nulls. Some of the sports games have the year in the name (for example, 'Madden NFL 2004') so we will try to utilize those. Then we will try to fill in games with multiple platforms, but the year is only missing from one of the platforms. Otherwise, we will fill based on the mode.\n- We need to change the types of this column to integers.\n\nGenre:\n- The only nulls are from the same two nulls mentioned in the Name column. so these will be taken care of as well.\n\nSales:\n- The sales has a significant amount of zeros. These may be the result of consoles not sold in some countries, or the game itself not being sold in some countries. We would want to look at this further to see if something is going on here. Specifically for the Other sales, this seems to be the lowest category of purchasers of video games, and the zero seems to be more of an acceptable amount here.\n\nScores:\n- Scores have a large amount of missing data. This will need to be filled in, most likely with averages based on the copies sold. Popular games that sell well will likely be higher rated.\n- Critic score will need to be changed to an integer as it is a 0 to 100 score, and the user score will need to be changed to a float. \n- TBDs in the user score column oddly seem related games that are based off of movies/TV and brands. This may be an issue related to \n\nRating:\n - ESRB rating also has a significant number of missing values. We will likely need to figure out the most common with the mode. Some are more intuitive than others, such as shooters would tend to be more M for Mature.", "_____no_output_____" ], [ "### Step 2. Prepare the data\n<a id='step2'></a>", "_____no_output_____" ], [ "First for the entire dataset, we will need to replace the column names with lowercase text.", "_____no_output_____" ] ], [ [ "games_data.columns = [x.lower() for x in games_data.columns]", "_____no_output_____" ] ], [ [ "#### Name\n<a id='step2name'></a>", "_____no_output_____" ], [ "The two nulls of the set may just be failures in the data gather process, as the name of the game is the principle identifier. Because there are only 2 of 16715 entries, these should be removed.", "_____no_output_____" ] ], [ [ "games_data.drop(games_data[games_data['name'].isnull()].index, inplace=True)", "_____no_output_____" ] ], [ [ "#### Year of Release\n<a id='step2year'></a>", "_____no_output_____" ], [ "The years may be missing because this data seems focused on sales. The data may not prioritize the year then. \n\nFirst, we can attempt to draw information directly from the title. Some of the sports games, such as 'Madden NFL 2004' have the year in the name.", "_____no_output_____" ] ], [ [ "check_years = games_data.query('year_of_release.isnull()')\nfor i, row in check_years.iterrows():\n try:\n year = int(x = re.findall(\"[0-9][0-9][0-9][0-9]\", row['name']))\n except:\n continue\n games_data.loc[i, 'year_of_release'] = year", "_____no_output_____" ], [ "check_years = games_data.query('year_of_release.isnull()')\nfor i, row in check_years.iterrows():\n try:\n year = int(row['name'][-2:])\n except:\n continue\n if year > 80:\n year += 1900\n elif year < 20:\n year += 2000\n else:\n continue\n games_data.loc[i, 'year_of_release'] = year", "_____no_output_____" ] ], [ [ "After that, we can try to see if some years are missing, but the same game but for a different platform has the year.", "_____no_output_____" ] ], [ [ "check_years = games_data.query('year_of_release.isnull()')\ncheck_against = games_data.query('year_of_release.notnull()')", "_____no_output_____" ], [ "for i, row in check_years.iterrows():\n name = row['name']\n multiplatform = check_against.query('name == @name')\n if len(multiplatform):\n year = list(multiplatform['year_of_release'])[0]\n games_data.loc[i, 'year_of_release'] = year", "_____no_output_____" ] ], [ [ "Anything left over we can fill by using the mode based on the platform. As platforms are done in generations, they typically are popular for only a few consecutive years until the next console is released. Therefore, it should be fine to use the mode.", "_____no_output_____" ] ], [ [ "check_years = games_data.query('year_of_release.isnull()')\ncheck_against = games_data.query('year_of_release.notnull()')\n\nkeys = check_against.platform.unique()\nvalues = list(check_against.groupby('platform')['year_of_release'].agg(pd.Series.mode))\nreference = {keys[i]: values[i] for i in range(len(keys))}\n\nfor i,val in check_years.platform.iteritems():\n replace = reference[val]\n if not isinstance(replace, float):\n replace = replace[0]\n games_data.loc[i,'year_of_release'] = replace", "_____no_output_____" ] ], [ [ "Lastly, because they are years, we need to change them to integers.", "_____no_output_____" ] ], [ [ "games_data['year_of_release'] = pd.to_numeric(games_data['year_of_release'], downcast='integer')", "_____no_output_____" ] ], [ [ "#### Sales\n<a id='step2sales'></a>", "_____no_output_____" ], [ "Lets take a look at the sales by platform.", "_____no_output_____" ] ], [ [ "check = games_data[['platform', 'na_sales', 'eu_sales', 'jp_sales']]\nvalues = check.groupby('platform').mean()\nprint(values)", " na_sales eu_sales jp_sales\nplatform \n2600 0.681203 0.041128 0.000000\n3DO 0.000000 0.000000 0.033333\n3DS 0.160558 0.118231 0.193596\nDC 0.104423 0.032500 0.164615\nDS 0.177778 0.087815 0.081623\nGB 1.166531 0.487959 0.868571\nGBA 0.228151 0.091545 0.057579\nGC 0.240036 0.069622 0.038813\nGEN 0.713704 0.204444 0.098889\nGG 0.000000 0.000000 0.040000\nN64 0.435799 0.128715 0.107273\nNES 1.285102 0.215816 1.006633\nNG 0.000000 0.000000 0.120000\nPC 0.097053 0.146242 0.000175\nPCFX 0.000000 0.000000 0.030000\nPS 0.281136 0.178454 0.116809\nPS2 0.270171 0.157006 0.064415\nPS3 0.295635 0.248152 0.060248\nPS4 0.277398 0.359923 0.040714\nPSP 0.090298 0.055153 0.063507\nPSV 0.029256 0.030512 0.050953\nSAT 0.004162 0.003121 0.186474\nSCD 0.166667 0.060000 0.075000\nSNES 0.256192 0.079665 0.487657\nTG16 0.000000 0.000000 0.080000\nWS 0.000000 0.000000 0.236667\nWii 0.376439 0.198644 0.052523\nWiiU 0.259184 0.170952 0.088503\nX360 0.477393 0.214548 0.009849\nXB 0.226566 0.073968 0.001675\nXOne 0.377004 0.208866 0.001377\n" ] ], [ [ "Initially the zeros look like problems with our data, but after some research, it appears that these represent a lack of console based sales. For example, the Atari 2600 shows zero sales for Japan, but the Atari 2600 was not sold in Japan. Instead, a console labelled the Atari 2800 was. Similarly, the Game Gear (Presumably the GG item) was a Japanese based handheld console, which is why there are zero sales in NA and EU.\n\nWe would like to use the total sales later on, so we should add a global sales column.", "_____no_output_____" ] ], [ [ "games_data.insert(loc=8, column='total_sales', value=0.0)\nfor i, row in games_data.iterrows():\n games_data.loc[i,'total_sales'] = row['na_sales'] + row['eu_sales'] + row['jp_sales'] + row['other_sales']\ngames_data.sort_values(['total_sales'], ascending=False)", "_____no_output_____" ] ], [ [ "#### Scores\n<a id='step2scores'></a>", "_____no_output_____" ], [ "Similar to the years, the scores may not be prioritized in the origination of the data. Because the scores have a lot of missing data, filling directly by an average may significantly weight the data and give biased results. We want to localize the information so we will get rolling averages by genre and total sales. Theoretically, the community of gamers likely are based on genre, so gamers interested in racing games would likely pick up more racing games and have a better understanding of what makes a racing game good or bad. Similarly, better scoring games should get better traction in sales, so that will be the other factor.\n\nFirst we will start with the critic scores.", "_____no_output_____" ] ], [ [ "check = games_data.sort_values(['genre', 'total_sales'], ascending=(True, False))\ncheck_critic_null = check.query('critic_score.isnull()')\nfor i, row in check_critic_null.iterrows():\n up, down, new_val = 1, 1, np.nan\n genre = row['genre']\n try:\n while pd.isna(check.loc[i-up, 'critic_score']):\n if check.loc[i-up, 'genre'] != genre:\n up = -1\n break\n up += 1\n except:\n up=-1\n \n try:\n while pd.isna(check.loc[i+down, 'critic_score']):\n if check.loc[i+down, 'genre'] != genre:\n down = -1\n break\n down += 1\n except:\n down=-1\n \n if up != -1 and down != -1:\n new_val = int((check.loc[i-up, 'critic_score'] + check.loc[i+down, 'critic_score'])/2)\n elif up != -1:\n new_val = check.loc[i-up, 'critic_score']\n elif down != -1:\n new_val = check.loc[i+down, 'critic_score']\n elif pd.notna(check.loc[i, 'user_score']) and check.loc[i, 'user_score'] != 'tbd':\n new_val = int(float(check.loc[i, 'user_score'])*10)\n \n games_data.loc[i, 'critic_score'] = new_val\ngames_data.info()", "<class 'pandas.core.frame.DataFrame'>\nInt64Index: 16713 entries, 0 to 16714\nData columns (total 12 columns):\nname 16713 non-null object\nplatform 16713 non-null object\nyear_of_release 16713 non-null int16\ngenre 16713 non-null object\nna_sales 16713 non-null float64\neu_sales 16713 non-null float64\njp_sales 16713 non-null float64\nother_sales 16713 non-null float64\ntotal_sales 16713 non-null float64\ncritic_score 14354 non-null float64\nuser_score 10014 non-null object\nrating 9949 non-null object\ndtypes: float64(6), int16(1), object(5)\nmemory usage: 2.2+ MB\n" ] ], [ [ "Left over NaN values should be because there are no genre specific scores. This is certainly possible with the amount of missing values. Now we should try to base it on the total sales and not have it genre specific. Lastly, if there are still values left, we should use the user value to determine the critic value.", "_____no_output_____" ] ], [ [ "check = games_data.sort_values(['total_sales'], ascending=False)\ncheck_critic_null = check.query('critic_score.isnull()')\nfor i, row in check_critic_null.iterrows():\n up, down, new_val = 1, 1, np.nan\n try:\n while pd.isna(check.loc[i-up, 'critic_score']):\n up += 1\n except:\n up=-1\n \n try:\n while pd.isna(check.loc[i+down, 'critic_score']):\n down += 1\n except:\n down=-1\n \n if up != -1 and down != -1:\n new_val = int((check.loc[i-up, 'critic_score'] + check.loc[i+down, 'critic_score'])/2)\n elif up != -1:\n new_val = check.loc[i-up, 'critic_score']\n elif down != -1:\n new_val = check.loc[i+down, 'critic_score']\n elif pd.notna(check.loc[i, 'user_score']) and check.loc[i, 'user_score'] != 'tbd':\n new_val = int(float(check.loc[i, 'user_score'])*10)\n if new_val != np.nan:\n games_data.loc[i, 'critic_score'] = new_val\ngames_data.info()", "<class 'pandas.core.frame.DataFrame'>\nInt64Index: 16713 entries, 0 to 16714\nData columns (total 12 columns):\nname 16713 non-null object\nplatform 16713 non-null object\nyear_of_release 16713 non-null int16\ngenre 16713 non-null object\nna_sales 16713 non-null float64\neu_sales 16713 non-null float64\njp_sales 16713 non-null float64\nother_sales 16713 non-null float64\ntotal_sales 16713 non-null float64\ncritic_score 16713 non-null float64\nuser_score 10014 non-null object\nrating 9949 non-null object\ndtypes: float64(6), int16(1), object(5)\nmemory usage: 2.2+ MB\n" ] ], [ [ "Now we can repeat the same process with user scores. In user scores, there are TBD values. These values are most likely due to the sample size requirements of the score. Looking at the data, a majority of the TBD values appear to be on low selling games, and therefore are 'waiting' for a certain number of user scores to determine it is an acceptable sized survey. We can treat these the same as if they were NaN values.", "_____no_output_____" ] ], [ [ "check_user_null = check.query('user_score.isnull()')\n\nfor i, row in check_user_null.iterrows():\n up, down, new_val = 1, 1, -1\n genre = row['genre']\n \n try:\n while pd.isna(check.loc[i-up, 'user_score']) or check.loc[i-up, 'user_score'] == 'tbd':\n if check.loc[i-up, 'genre'] != genre:\n up = -1\n break\n up += 1\n except:\n up=-1\n \n try:\n while pd.isna(check.loc[i+down, 'user_score']) or check.loc[i+down, 'user_score'] == 'tbd':\n if check.loc[i+down, 'genre'] != genre:\n down = -1\n break\n down += 1\n except:\n down=-1\n\n if up != -1 and down != -1:\n new_val = (float(check.loc[i-up, 'user_score']) + float(check.loc[i+down, 'user_score']))/2\n elif up != -1:\n new_val = check.loc[i-up, 'user_score']\n elif down != -1:\n new_val = check.loc[i+down, 'user_score']\n if new_val != -1:\n games_data.loc[i, 'user_score'] = round(float(new_val),1)\ngames_data.info()", "<class 'pandas.core.frame.DataFrame'>\nInt64Index: 16713 entries, 0 to 16714\nData columns (total 12 columns):\nname 16713 non-null object\nplatform 16713 non-null object\nyear_of_release 16713 non-null int16\ngenre 16713 non-null object\nna_sales 16713 non-null float64\neu_sales 16713 non-null float64\njp_sales 16713 non-null float64\nother_sales 16713 non-null float64\ntotal_sales 16713 non-null float64\ncritic_score 16713 non-null float64\nuser_score 14426 non-null object\nrating 9949 non-null object\ndtypes: float64(6), int16(1), object(5)\nmemory usage: 2.2+ MB\n" ], [ "for i, row in games_data.iterrows():\n if row['user_score'] == 'tbd' or row['user_score'] is np.nan:\n games_data.loc[i, 'user_score'] = round(row['critic_score']/10, 1)", "_____no_output_____" ] ], [ [ "The critic score are integers on a scale from 1 to 100, and the user scores are floats from 0.0 to 10.0, so we need to cast them as such.", "_____no_output_____" ] ], [ [ "games_data['critic_score'] = pd.to_numeric(games_data['critic_score'], downcast='integer')\ngames_data['user_score'] = pd.to_numeric(games_data['user_score'], downcast='float')\ngames_data.info()", "<class 'pandas.core.frame.DataFrame'>\nInt64Index: 16713 entries, 0 to 16714\nData columns (total 12 columns):\nname 16713 non-null object\nplatform 16713 non-null object\nyear_of_release 16713 non-null int16\ngenre 16713 non-null object\nna_sales 16713 non-null float64\neu_sales 16713 non-null float64\njp_sales 16713 non-null float64\nother_sales 16713 non-null float64\ntotal_sales 16713 non-null float64\ncritic_score 16713 non-null int8\nuser_score 16713 non-null float32\nrating 9949 non-null object\ndtypes: float32(1), float64(5), int16(1), int8(1), object(4)\nmemory usage: 2.0+ MB\n" ] ], [ [ "#### Ratings\n<a id='step2ratings'></a>", "_____no_output_____" ] ], [ [ "check_rating = games_data.query('rating.isnull()')\ncheck_against = games_data.query('rating.notnull()')", "_____no_output_____" ], [ "import sys\nimport warnings\nif not sys.warnoptions:\n warnings.simplefilter(\"ignore\")", "_____no_output_____" ], [ "check_against['keys'] = check_against.platform+\".\"+check_against.genre\nkeys = list(check_against['keys'].unique())", "_____no_output_____" ], [ "values = list(check_against.groupby(['platform', 'genre'])['rating'].agg(pd.Series.mode))\nprint(check_against.groupby(['platform', 'genre'])['rating'].agg(pd.Series.mode))\nreference = {keys[i]: values[i] for i in range(len(values))}\n\nfor i,row in check_rating.iterrows():\n check = row.platform + \".\" + row.genre\n try:\n replace = reference[check]\n games_data.loc[i,'rating'] = replace\n except:\n continue", "platform genre \n3DS Action E10+\n Adventure E10+\n Fighting T\n Misc E\n Platform E\n ... \nXOne Role-Playing M\n Shooter M\n Simulation [E, T]\n Sports E\n Strategy [E, T]\nName: rating, Length: 198, dtype: object\n" ], [ "check_rating = games_data.query('rating.isnull()')\ncheck_against = games_data.query('rating.notnull()')", "_____no_output_____" ], [ "check_rating_null = games_data.query('rating.isnull()')\nfor i, row in check_critic_null.iterrows():\n up, down, new_val = 1, 1, np.nan\n try:\n while pd.isna(games_data.loc[i-up, 'rating']):\n up += 1\n except:\n up=-1\n \n try:\n while pd.isna(games_data.loc[i+down, 'rating']):\n down += 1\n except:\n down=-1\n \n if up != -1 and down != -1:\n if up < down:\n new_val = games_data.loc[i-up, 'rating']\n else:\n new_val = games_data.loc[i+down, 'rating']\n elif up != -1:\n new_val = games_data.loc[i-up, 'rating']\n elif down != -1:\n new_val = games_data.loc[i+down, 'rating']\n else:\n genre = row['genre']\n new_val = games_data.groupby(['genre'])['rating'].agg(pd.Series.mode).loc[genre]\n games_data.loc[i, 'rating'] = new_val\ngames_data.info()", "<class 'pandas.core.frame.DataFrame'>\nInt64Index: 16713 entries, 0 to 16714\nData columns (total 12 columns):\nname 16713 non-null object\nplatform 16713 non-null object\nyear_of_release 16713 non-null int16\ngenre 16713 non-null object\nna_sales 16713 non-null float64\neu_sales 16713 non-null float64\njp_sales 16713 non-null float64\nother_sales 16713 non-null float64\ntotal_sales 16713 non-null float64\ncritic_score 16713 non-null int8\nuser_score 16713 non-null float32\nrating 15855 non-null object\ndtypes: float32(1), float64(5), int16(1), int8(1), object(4)\nmemory usage: 2.0+ MB\n" ], [ "keys = check_against.genre.unique()\nvalues = list(check_against.groupby(['genre'])['rating'].agg(pd.Series.mode))\nreference = {keys[i]: values[i] for i in range(len(values))}\n\nfor i,row in check_rating.iterrows():\n check = row.genre\n try:\n replace = reference[check]\n games_data.loc[i,'rating'] = replace\n except:\n continue", "_____no_output_____" ], [ "games_data.info()", "<class 'pandas.core.frame.DataFrame'>\nInt64Index: 16713 entries, 0 to 16714\nData columns (total 12 columns):\nname 16713 non-null object\nplatform 16713 non-null object\nyear_of_release 16713 non-null int16\ngenre 16713 non-null object\nna_sales 16713 non-null float64\neu_sales 16713 non-null float64\njp_sales 16713 non-null float64\nother_sales 16713 non-null float64\ntotal_sales 16713 non-null float64\ncritic_score 16713 non-null int8\nuser_score 16713 non-null float32\nrating 16713 non-null object\ndtypes: float32(1), float64(5), int16(1), int8(1), object(4)\nmemory usage: 2.0+ MB\n" ] ], [ [ "Lastly, it turns out that K-A was a rating that is the same as E, as K-A is kids through adults, and was later changed to mean E. We should change that in this data as well.", "_____no_output_____" ] ], [ [ "games_data.loc[games_data['rating'] == \"K-A\", \"rating\"] = \"E\"", "_____no_output_____" ], [ "games_data.info()", "<class 'pandas.core.frame.DataFrame'>\nInt64Index: 16713 entries, 0 to 16714\nData columns (total 12 columns):\nname 16713 non-null object\nplatform 16713 non-null object\nyear_of_release 16713 non-null int16\ngenre 16713 non-null object\nna_sales 16713 non-null float64\neu_sales 16713 non-null float64\njp_sales 16713 non-null float64\nother_sales 16713 non-null float64\ntotal_sales 16713 non-null float64\ncritic_score 16713 non-null int8\nuser_score 16713 non-null float32\nrating 16713 non-null object\ndtypes: float32(1), float64(5), int16(1), int8(1), object(4)\nmemory usage: 2.0+ MB\n" ] ], [ [ "All of the ratings are now filled in.", "_____no_output_____" ], [ "#### Step 2 Conclusion\n<a id='step2con'></a>", "_____no_output_____" ], [ " All of the data has been cleaned and filled in. There are no longer any missing values, and there are no more obtuse values such as TBD. All of the characteristics are their correct types, and are adequately downsized to optimized types.", "_____no_output_____" ], [ "### Step 3. Analyze the data\n<a id='step3'></a>", "_____no_output_____" ], [ "- Look at how many games were released in different years. Is the data for every period significant?\n- Look at how sales varied from platform to platform. Choose the platforms with the greatest total sales and build a distribution based on data for each year. Find platforms that used to be popular but now have zero sales. How long does it generally take for new platforms to appear and old ones to fade?\n- Determine what period you should take data for. To do so, look at your answers to the previous questions. The data should allow you to build a prognosis for 2017.\n- Work only with the data that you've decided is relevant. Disregard the data for previous years.\n- Which platforms are leading in sales? Which ones are growing or shrinking? Select several potentially profitable platforms.\n- Build a box plot for the global sales of all games, broken down by platform. Are the differences in sales significant? What about average sales on various platforms? Describe your findings.\n- Take a look at how user and professional reviews affect sales for one popular platform (you choose). Build a scatter plot and calculate the correlation between reviews and sales. Draw conclusions.\n- Keeping your conclusions in mind, compare the sales of the same games on other platforms.\n- Take a look at the general distribution of games by genre. What can we say about the most profitable genres? Can you generalize about genres with high and low sales?", "_____no_output_____" ], [ "First lets look at the total sales by release year.", "_____no_output_____" ] ], [ [ "total_years = games_data.year_of_release.max()-games_data.year_of_release.min()\ngames_data.year_of_release.hist(bins=total_years)\n\nplt.ylabel('Total Sales')\nplt.xlabel('Year')\nplt.title('Distribution of Sales by Year')\nplt.show()", "_____no_output_____" ] ], [ [ "There appears to be an early tail, most likely when gaming had not yet fully joined the ranks of pop culture that we know it has today. This delay is likely due to consumer access and early technology. \n\nIt can be compared to the cell phone we know today. It used to be a large brick that had a large price tag of nearly //$4,000 and was extremely limited battery life of about 30 minutes, as mentioned by [this NBC article](https://www.nbcnews.com/id/wbna7432915).\n\nThis was not seen as something really necessary for anyone but wealthy business leaders. Soon, technology became cheaper and now most citizens of developed countries have a cell phone.\n\nThat being said, lets remove this time period needed for gaming to take off. ", "_____no_output_____" ] ], [ [ "q1 = games_data.year_of_release.quantile(q=.25)\nq3 = games_data.year_of_release.quantile(q=.75)\nIQR = q3-q1\ngames_data = games_data.query('year_of_release > @q1 - @IQR*1.5')", "_____no_output_____" ], [ "total_years = games_data.year_of_release.max()-games_data.year_of_release.min()\ngames_data.year_of_release.hist(bins=total_years)\n\nplt.ylabel('Total Sales')\nplt.xlabel('Year')\nplt.title('Distribution of Sales by Year')\nplt.show()", "_____no_output_____" ] ], [ [ "Now lets try to filter out the less popular platforms. Also, as we are trying to predict near future results, we need to make sure that the consoles are still selling games in the most recent year. Otherwise, they will not be selling games in 2017 either.", "_____no_output_____" ] ], [ [ "grouped_platform_sales = games_data.groupby(['platform', 'year_of_release'])['total_sales'].agg(['sum', 'count'])", "_____no_output_____" ], [ "plats = []\nfor platform, df in grouped_platform_sales.groupby(level=0):\n #print(df.index)\n keep = df.index.isin(['2016'], level='year_of_release')\n #print(df)\n if 1 in keep:\n plats.append(platform)\nprint(plats)", "['3DS', 'PC', 'PS3', 'PS4', 'PSV', 'Wii', 'WiiU', 'X360', 'XOne']\n" ] ], [ [ "To understand each platform's performance, we need to calculate the total number of sales per year, per platform.", "_____no_output_____" ] ], [ [ "usable_platforms = grouped_platform_sales[grouped_platform_sales.index.get_level_values('platform').isin(plats)]\nprint(usable_platforms)\nclean_games_data = games_data.query('platform.isin(@plats)')", " sum count\nplatform year_of_release \n3DS 1993 0.40 1\n 1999 0.47 5\n 2000 0.02 1\n 2010 0.30 1\n 2011 63.20 116\n... ... ...\nX360 2016 1.52 13\nXOne 2013 18.96 19\n 2014 54.07 61\n 2015 60.14 80\n 2016 26.15 87\n\n[101 rows x 2 columns]\n" ] ], [ [ "There are a few games that are highly skewing the results, such as Wii Sports, that may be diamonds in the rough, and can not be used to predict future sales.", "_____no_output_____" ] ], [ [ "q1 = clean_games_data.total_sales.quantile(q=.25)\nq3 = clean_games_data.total_sales.quantile(q=.75)\nIQR = q3-q1\nfiltered_clean_games_data = clean_games_data.query('total_sales < @q3 + @IQR*1.5')\ntotal_years = clean_games_data.year_of_release.max()-clean_games_data.year_of_release.min()\nplat_count = clean_games_data.pivot_table(values= 'total_sales', index='year_of_release', columns='platform', aggfunc='sum', fill_value=0)\nfiltered_plat_count = filtered_clean_games_data.pivot_table(values= 'total_sales', index='year_of_release', columns='platform', aggfunc='sum', fill_value=0)", "_____no_output_____" ], [ "# This is to make sure that colors are different with a large number of different colored bars in our graphs\ndef floatRgb(mag, cmin, cmax):\n \"\"\" Return a tuple of floats between 0 and 1 for R, G, and B. \"\"\"\n # Normalize to 0-1\n try: x = float(mag-cmin)/(cmax-cmin)\n except ZeroDivisionError: x = 0.5 # cmax == cmin\n blue = min((max((4*(0.75-x), 0.)), 1.))\n red = min((max((4*(x-0.25), 0.)), 1.))\n green = min((max((4*math.fabs(x-0.5)-1., 0.)), 1.))\n return red, green, blue\n\ndef rgb(mag, cmin, cmax):\n \"\"\" Return a tuple of integers, as used in AWT/Java plots. \"\"\"\n red, green, blue = floatRgb(mag, cmin, cmax)\n return int(red*255), int(green*255), int(blue*255)\n\ndef strRgb(mag, cmin, cmax):\n \"\"\" Return a hex string, as used in Tk plots. \"\"\"\n return \"#%02x%02x%02x\" % rgb(mag, cmin, cmax)", "_____no_output_____" ], [ "# Plotting\nplots = [clean_games_data, filtered_clean_games_data]\nplot_totals = [plat_count, filtered_plat_count]\n\nfor plot in range(len(plots)):\n plt.figure(figsize=(16,8))\n print(plots[plot].platform.unique())\n color_vals = []\n #rotates through the platforms\n for i in range(len(plots[plot].platform.unique())):\n num = i*1/len(plots[plot].platform.unique())\n color = strRgb(num,0,1)\n color_vals.append(color)\n\n # Creating dictionaries with colors\n colors = {i: color_vals[i] for i in range(len(color_vals))}\n vals = list(plots[plot].platform.unique())\n platforms = {i: vals[i] for i in range(len(vals))}\n\n # Plotting in a loop\n for i in range(len(plot_totals[plot].index)):\n year = plot_totals[plot].index[i]\n year_data = plot_totals[plot].loc[year]\n baseline = 0\n color_index = 0\n for j in year_data:\n plt.bar(x = i, height = j, bottom = baseline, color=colors[color_index])\n baseline += j\n color_index += 1\n\n plt.xticks(np.arange(len(plot_totals[plot].index)), plot_totals[plot].index, rotation = 270);\n\n # Creating legend\n patches = list()\n for i in reversed(range(len(plots[plot].platform.unique()))):\n patch = mpatches.Patch(color = colors[i], label = plot_totals[plot].columns[i])\n patches.append(patch)\n plt.legend(handles=patches, fontsize=12, framealpha=1)\n\n # Some additioanl plot prep\n plt.rcParams['axes.axisbelow'] = True\n plt.grid(color='gray', linestyle='dashed')\n plt.ylabel('Amount of Sales')\n plt.xlabel('Year')\n plt.title('Number of Games by Year and Platform');", "['Wii' 'X360' 'PS3' 'PS4' '3DS' 'PC' 'XOne' 'WiiU' 'PSV']\n['PC' 'Wii' 'PS3' 'XOne' 'X360' 'WiiU' '3DS' 'PS4' 'PSV']\n" ] ], [ [ "<div class=\"alert alert-success\" role=\"alert\">\nReviewer's comment v. 1:\n \nAN excellent graphs, but Ridgeplots can be useful here: https://matplotlib.org/matplotblog/posts/create-ridgeplots-in-matplotlib/ \n</div>", "_____no_output_____" ], [ "We can see a large boost of games sold around 2009 through 2011, primarily for the success of the Wii, PS3, and Xbox 360 consoles. After that burst, the sales drop, and then next gen consoles become popular, but not at the same level and are already falling well below previous years by 2016.\n\nTo make a prediction for the next year, we need to attempt a parabolic trend, as it will be based on the growth or decay of the popularity of the platforms, and it should also represent how quickly the platforms are coming in and out of popularity. \n\nWe can see in the above plots on a single platform basis that there are parabolic trends where there is not enough time yet for game developers to create games for a brand new platform, they get that time to make it, and over time the platform becomes outdated, and developers and consumers both prepare for the new consoles. In particular, consumers may want to save money on video games if they believe a platform is nearing the end of its stride, and would want to be financially ready for the next platform. \n\nThis parabolic trend tends to line up for platforms, as major competing consoles launch at the same time. For example, the playstation series from Sony typically launches around the same time as Microsoft's Xbox line to drive sales with competition.", "_____no_output_____" ] ], [ [ "predicting_2017 = filtered_clean_games_data.query('year_of_release.isin([2014, 2015, 2016])')\ntemp = pd.pivot_table(predicting_2017, values='total_sales', index='platform', columns='year_of_release', aggfunc='sum')", "_____no_output_____" ], [ "def calc_parabola_vertex(x1, y1, x2, y2, x3, y3):\n\n denom = (x1-x2) * (x1-x3) * (x2-x3);\n A = (x3 * (y2-y1) + x2 * (y1-y3) + x1 * (y3-y2)) / denom;\n B = (x3*x3 * (y1-y2) + x2*x2 * (y3-y1) + x1*x1 * (y2-y3)) / denom;\n C = (x2 * x3 * (x2-x3) * y1+x3 * x1 * (x3-x1) * y2+x1 * x2 * (x1-x2) * y3) / denom;\n\n return A,B,C", "_____no_output_____" ], [ "for i, row in temp.iterrows():\n x1, y1 = [2014, row[2014]]\n x2, y2 = [2015, row[2015]]\n x3, y3 = [2016, row[2016]]\n \n a, b, c = calc_parabola_vertex(x1, y1, x2, y2, x3, y3)\n new_val=(a*(2017**2))+(b*2017)+c\n \n if new_val > 0:\n temp.loc[i, 2017] = new_val\n else:\n temp.loc[i, 2017] = 0\ntemp", "_____no_output_____" ] ], [ [ "Now that we have the estimated amounts of sales per platform, we need to integrate it into our sales graph.", "_____no_output_____" ] ], [ [ "filtered_plat_count = filtered_plat_count.append(temp[2017])\nprint(temp[2017].sum())", "32.4200055571273\n" ], [ "# Plotting\nplots = [filtered_clean_games_data]\nplot_totals = [filtered_plat_count]\n\nfor plot in range(len(plots)):\n plt.figure(figsize=(16,8))\n print(plots[plot].platform.unique())\n color_vals = []\n for i in range(len(plots[plot].platform.unique())):\n num = i*1/len(plots[plot].platform.unique())\n color = strRgb(num,0,1)\n color_vals.append(color)\n\n # Creating dictionaries with colors and cancelaltion causes\n colors = {i: color_vals[i] for i in range(len(color_vals))}\n vals = list(plots[plot].platform.unique())\n platforms = {i: vals[i] for i in range(len(vals))}\n\n # Plotting in a loop\n for i in range(len(plot_totals[plot].index)):\n year = plot_totals[plot].index[i]\n year_data = plot_totals[plot].loc[year]\n baseline = 0\n color_index = 0\n for j in year_data:\n plt.bar(x = i, height = j, bottom = baseline, color=colors[color_index])\n baseline += j\n color_index += 1\n #plt.text(x = i, y = plat_count[i] + 0.05, s = round(plat_count[i], 1), \\\n #ha = 'center', fontsize=13)\n# for j in year_data:\n# plt.bar(x = 2017, height = j, bottom = baseline, color=colors[color_index])\n# baseline += j\n# color_index += 1\n \n\n plt.xticks(np.arange(len(plot_totals[plot].index)), plot_totals[plot].index, rotation = 270);\n\n # Creating legend\n patches = list()\n for i in reversed(range(len(plots[plot].platform.unique()))):\n patch = mpatches.Patch(color = colors[i], label = plot_totals[plot].columns[i])\n patches.append(patch)\n plt.legend(handles=patches, fontsize=12, framealpha=1)\n\n # Some additioanl plot prep\n plt.rcParams['axes.axisbelow'] = True\n plt.grid(color='gray', linestyle='dashed')\n plt.ylabel('Amount of Sales')\n plt.xlabel('Year')\n plt.title('Number of Games by Year and Platform');", "['PC' 'Wii' 'PS3' 'XOne' 'X360' 'WiiU' '3DS' 'PS4' 'PSV']\n" ] ], [ [ "We can see that from about 2013 to 2016, it has risen, and began dropping at a faster rate. Our prediction of 2017 at this level of modelling visibly follows that trend. One thing to note is that this indicates, from what we know of the gaming industry, that it would likely be time for a new generation of consoles to come out, restarting the wave of consumer sales.", "_____no_output_____" ], [ "Next, lets look at the distribution of these games by platform.", "_____no_output_____" ] ], [ [ "filtered_clean_games_data.boxplot(column='total_sales', by='platform', figsize=(16,8))\nplt.ylabel('Total Sales')\nplt.xlabel('Platform')\nplt.title('Distribution of Sales by Platform')\nplt.show()\n", "_____no_output_____" ] ], [ [ "It appears that there are a significant number of outliers accross the board. This shows, that a large portion of each platforms market performance is largely based on triple A titles, but there are still a significant number of games that are indie games, less advertised games, or games that just generally did not get the same amount of traction among consumers. \n\nIt looks like overall, the PS3 and Xbox 360 generally had better selling games, as the distribution is spread out to higher sales. The PSV was not know for its popularity, so this explains is low distribution. As for the PC, it is known for having a lot of indie games, as it is more accessible for game makers to distribute games. This accessibility also explains the large number of outliers as well.", "_____no_output_____" ], [ "Now lets take a look at how critic and user reviews affect sales of a single platform. For this example, we will look at the Wii.", "_____no_output_____" ] ], [ [ "wii_data = filtered_clean_games_data[clean_games_data['platform'] == 'Wii']\nfig, axes = plt.subplots(ncols=3, figsize=(16,8))\n\naxes[0].scatter(wii_data.critic_score, wii_data.total_sales, color='orange', alpha=.5)\naxes[1].scatter(wii_data.user_score, wii_data.total_sales, color='blue', alpha=.5)\naxes[2].scatter(wii_data.user_score*10, wii_data.total_sales, color='blue', alpha=.3)\naxes[2].scatter(wii_data.critic_score, wii_data.total_sales, color='orange', alpha=.3)\n\npop_a = mpatches.Patch(color='blue', label='user')\npop_b = mpatches.Patch(color='orange', label='critic')\n\naxes[2].legend(handles=[pop_a,pop_b], loc='upper left')\n\naxes[0].set(title='Critic Score vs. Total Sales', xlabel='Critic Score', ylabel='Total Sales')\naxes[1].set(title='User Score vs. Total Sales', xlabel='User Score')\naxes[2].set(title='Overlapped Critic and User Score vs. Total Sales', xlabel='Critic Score and Equivalent Scale of User Score')\n\nplt.show()", "_____no_output_____" ] ], [ [ "Although critic and user reviews look similar, we can see that barring a few outliers, users tend to be more willing to rate games higher than critics. It also seems that the shape of the critics scoring appears to be more rectangular than the users score. This implies that the amount of sales has less of an impact on the scoring than users do. Users may be more inclined to be influenced by word of mouth and riding the wave of a game's popularity.", "_____no_output_____" ], [ "A lot of games are multiplatform, so lets see if there is much of a difference between platforms.", "_____no_output_____" ] ], [ [ "wii_data = wii_data[['name', 'na_sales', 'eu_sales', 'jp_sales', 'other_sales', 'total_sales']]\nx360_data = filtered_clean_games_data[clean_games_data['platform'] == 'X360']\nx360_data = x360_data[['name', 'na_sales', 'eu_sales', 'jp_sales', 'other_sales', 'total_sales']]\nwii_x360_cross = pd.merge(wii_data, x360_data, on=\"name\", suffixes=(\"Wii\", \"X360\"))\n\nfig = plt.figure()\nax1 = fig.add_subplot(111)\n\nax1.hist(wii_x360_cross['total_salesWii'], bins=30, alpha= 0.5, label='Wii Sales')\nax1.hist(wii_x360_cross['total_salesX360'], bins=30, alpha= 0.5, label='XBox 360 Sales')\nplt.legend(loc='upper right');\nplt.ylabel('Frequency')\nplt.xlabel('Total Sales')\nplt.title('Distribution of Sales by Platform')\nplt.show()", "_____no_output_____" ] ], [ [ "It does appear that there may be a slight bias towards Xbox 360. This does make some sense, as the consoles are very different. The Xbox is primarily a button input, while the wii does have button inputs, but the console was largely popular to it's motion control. Because the Xbox does not have motion controls, motion control games would not be multi platform, and so it does not have that advantage of what area of expertise gave it its popularity.", "_____no_output_____" ], [ "Now lets take a look into the same level of detail for the game genres.", "_____no_output_____" ] ], [ [ "top_plats = filtered_clean_games_data.groupby('genre')['total_sales'].sum()\nfiltered_clean_games_data.boxplot(column='total_sales', by='genre', figsize=(16,8))\nplt.ylabel('Total Sales')\nplt.xlabel('Genre')\nplt.title('Distribution of Sales by Genre')\nplt.show()", "_____no_output_____" ] ], [ [ "The largest genres are Action, Fighting, Platform, Shooter, and Sports. This makes sense as they make up a large part of the triple A title games, including well established franchises such as Zelda, Mortal Combat, Mario, Call of Duty, and Fifa. The lowest are Adventure, Puzzle, and Strategy, games that are typical as indie titles and represent lower volume and pricing.", "_____no_output_____" ] ], [ [ "top_plats.plot('bar', figsize=(16,8))\nplt.ylabel('Total Sales')\nplt.xlabel('Platform')\nplt.title('Total Sales by Genre')\nplt.show()", "_____no_output_____" ] ], [ [ "Similarly to the distribution, the largest amount of sales are in Action and Sports, while the lower sales are in puzzle and strategy. The differences between these total amounts and the distribution is largely in the volume of games in the market place.", "_____no_output_____" ], [ "#### Step 3 Conclusion\n<a id='step3con'></a>", "_____no_output_____" ], [ "After viewing the preliminary data, we saw that the earlier years are not very representative, so all lower outlying years were filtered out. We then filtered for the popular and relevant consoles, based on being more recently selling platforms. We saw a large boost of games sold around 2009 through 2011, primarily for the success of the Wii, PS3, and Xbox 360 consoles - at those years, they were relatively new. After that burst, the success fell, and then next gen consoles came out, but the wave was not as successful and are already falling well below by 2016. Because of this trend, and without the knowledge of new consoles, the trend naturally falls, and we expect sales around 32.4 million USD.\n\nAs expected with the waves of success, consoles such as the PS3 and Xbox 360 had higher distributions of sales, while handhelds and PC games sold typically lower. \n\nWe also found that users and critics scored games very similar, but users may be slightly more biased by the traction in sales and popularity by word of mouth. Consoles were also relatively similar, but small discrepancies can be found between multiplatform games, and this may be due to the strengths and weaknesses of the consoles in relation to the game types. \n\nWe also looked at the distribution of sales based on genre and noticed that more total sales by genre correlated with higher distribution of better selling games. They also are typically the genres of triple A games, so these distributions make sense.", "_____no_output_____" ], [ "### Step 4. Create a user profile for each region\n<a id='step4'></a>", "_____no_output_____" ] ], [ [ "categories = ['platform', 'genre', 'rating']\nfig, axes = plt.subplots(nrows=3, ncols=3, figsize=(16,16))\n\ncolor_vals = ['orange', 'cyan', 'lime']\ncolors = {i: color_vals[i] for i in range(len(color_vals))}\nvals = ['na_sales', 'eu_sales', 'jp_sales']\nlocations = {i: vals[i] for i in range(len(vals))}\n\nfor i in range(len(categories)):\n top_order = filtered_clean_games_data.groupby(categories[i])['na_sales', 'eu_sales', 'jp_sales'].sum()\n na_top_order = top_order.sort_values(by='na_sales', ascending=False)\n eu_top_order = top_order.sort_values(by='eu_sales', ascending=False)\n jp_top_order = top_order.sort_values(by='jp_sales', ascending=False)\n top = [na_top_order, eu_top_order, jp_top_order]\n \n for j in range(len(top)):\n # Plotting in a loop\n for k in range(5):\n platforms = top[j].index[k]\n platforms_data = top[j].loc[platforms]\n baseline = 0\n color_index = 0\n for m in platforms_data:\n axes[i,j].bar(x = k, height = m, bottom = baseline, color=colors[color_index])\n baseline += m\n color_index += 1\n\n title_label = 'Top 5 ' + top_order.index.name.title() + 's for ' + top_order.columns[j][:2].upper()\n axes[i,j].set_title(label=title_label)\n axes[i,j].xaxis.set(ticks=np.arange(5), ticklabels=top[j].index)\n axes[i,0].set_ylabel(ylabel='Sales')\n \n#Creating legend\npatches = list()\nfor i in range(3):\n #patch = mpatches.Patch(color = colors[i], label = cancellation_cause[cancellation_code_per_carrier_pct.columns[i]])\n patch = mpatches.Patch(color = colors[i], label = top_order.columns[i])\n patches.append(patch)\nfig.legend(handles=patches, fontsize=12, framealpha=1, loc='upper left')\nfig.show()", "_____no_output_____" ] ], [ [ "#### Step 4 Conclusion\n<a id='step4con'></a>", "_____no_output_____" ], [ "For the top 5 platforms by location, the Xbox 360, Wii, and PS3 were very popular in North America, but nothing outstanding byond those three. The EU is similar, but PC was preferenced over the Wii, keeping course with tactile, button based platforms. In Japan, PS3 was the largest platform, but handhelds were highly prefered over what was popular for both the EU and NA groups. This deiscrepancy may be largely due to [Japan's significantly higher use of public transport](https://en.wikipedia.org/wiki/List_of_countries_by_rail_usage). This means they may be more inclined to use that time on a train to use a handheld console for convenience. \n\nFor the top 5 genres by location, Action and Sports were very the most popular in North America, and the EU. In Japan, Role-Playing was a close second, which was at the 5th spot in NA and was not even present in the EU's top 5. This may be due to the popularity with sports in the respective countries. For example, some of the two most successful sports games are the Madden NFL american football series and FIFA Soccer (european football) series. Both of these may correlate with the popularity of the sports in the United States and Europe respectively.\n\nFor the top 5 ratings by location, E and T were every groups first and second, respectively. In Japan and the EU, M took precedent over E10+, but the opposite was true in NA. The differences between M and E10+ however, are quite small in all regions, and may be considered negligible.", "_____no_output_____" ], [ "### Step 5. Test the following hypotheses:\n<a id='step5'></a>", "_____no_output_____" ] ], [ [ "# the level of significance\nalpha = .05", "_____no_output_____" ] ], [ [ "#### Average user ratings of the Xbox One and PC platforms are the same.\n<a id='step5h1'></a>", "_____no_output_____" ], [ "A dual sample t-test will be used to determine if the _surf_ plan and _ultimate_ plan generate different monthly revenues per person. We will create the following hypotheses:", "_____no_output_____" ], [ "The null hypothesis, $H_0$: The average score from users of the Xbox One games and PC games are equal. \nThe alternative hypothesis, $H_A$: The average score from users of the Xbox One games and PC games are not equal. ", "_____no_output_____" ] ], [ [ "set1 = filtered_clean_games_data[filtered_clean_games_data.platform == 'XOne']['user_score']\nset2 = filtered_clean_games_data[filtered_clean_games_data.platform == 'PC']['user_score']\n\nresults = st.ttest_ind(\n filtered_clean_games_data[filtered_clean_games_data.platform == 'XOne']['user_score'],\n filtered_clean_games_data[filtered_clean_games_data.platform == 'PC']['user_score'],\n equal_var=False)\n\nprint('p-value: ', results.pvalue)\n\nif results.pvalue > alpha:\n print('We cannot reject the null hypothesis')\nelse:\n print('We can reject the null hypothesis')", "p-value: 0.001260648856892789\nWe can reject the null hypothesis\n" ], [ "fig, axes = plt.subplots(ncols=2, figsize=(16,4))\n\nxbox = filtered_clean_games_data[filtered_clean_games_data.platform == 'XOne']\npc = filtered_clean_games_data[filtered_clean_games_data.platform == 'PC']\n\naxes[0].hist(xbox.user_score, bins=len(xbox.user_score.unique()), color='orange')\naxes[1].hist(pc.user_score, bins=len(pc.user_score.unique()), color='blue')\naxes[0].set(title='Distribution of Xbox User Scores', xlabel='User Score', ylabel='Frequency')\naxes[1].set(title='Distribution of PC User Scores', xlabel='User Score', ylabel='Frequency')\n\nplt.show()\n\npop_a = mpatches.Patch(color='blue', label='PC')\npop_b = mpatches.Patch(color='orange', label='Xbox')\n\nfig, axes = plt.subplots(ncols=1, figsize=(16,8))\naxes.hist([xbox.user_score, pc.user_score], bins=len(xbox.user_score.unique()), color=['orange', 'blue'])\naxes.set(title='Distribution of Xbox and PC User Scores', xlabel='User Score', ylabel='Frequency')\naxes.legend(handles=[pop_a,pop_b], loc='upper left')\nplt.show()", "_____no_output_____" ] ], [ [ "To confirm, it does appear that the distribution of the PC user score is more left skewed than the Xbox user score. the PC scores peak around 6.7, and tail more evenly in both directions, while there seems to be a larger dostribution of high ranked PC games.", "_____no_output_____" ], [ "The variances of the two subsamples are not equal, and therefore the parameter, `equal_var` must be set to False to compare sets with different variances and/or sets of different sizes.", "_____no_output_____" ], [ "The null hypothesis of a dual sample t-test is that the two groups are similar, and the alternative hypothesis is that they are dissimilar. \n\nIn this case, the null hypothesis is that the average score from users of the Xbox One games are similar to the average scores of the PC games. In the results of the t-test, the p-value was below our level of significance and we could reject the null variable and say that the average scores differ between the two groups. From the correlation of user score to sales, as well as the distribution of Xbox One games' higher total sales vs the PC games' lower total sales, this makes sense that they would not be equal.", "_____no_output_____" ], [ "#### Average user ratings for the Action and Sports genres are different.\n<a id='step5h2'></a>", "_____no_output_____" ], [ "The null hypothesis, $H_0$: The average score from users of the Action games and Sports games are equal. \nThe alternative hypothesis, $H_A$: The average score from users of the Action games and Sports games are not equal. ", "_____no_output_____" ] ], [ [ "results = st.ttest_ind(\n filtered_clean_games_data[filtered_clean_games_data.genre == 'Action']['user_score'],\n filtered_clean_games_data[filtered_clean_games_data.genre == 'Sports']['user_score'],\n equal_var=False)\n\nprint('p-value: ', results.pvalue)\n\nif results.pvalue > alpha:\n print('We cannot reject the null hypothesis')\nelse:\n print('We can reject the null hypothesis')", "p-value: 5.279399270185236e-10\nWe can reject the null hypothesis\n" ] ], [ [ "The variances of these two subsamples are also not equal, so the `equal_var` must be set to False.\n\nFor this example, the null hypothesis is that the average score from users of the Action games are similar to the average scores of the Sports games. In the results of the t-test, the p-value was again below our level of significance and we could reject the null variable and say that the average scores differ between the two groups. The distribution of Action games' higher total sales vs the Sports games' lower total sales, once again we can make sense that they would not be equal.", "_____no_output_____" ], [ "#### Step 5 Conclusion\n<a id='step5con'></a>", "_____no_output_____" ], [ "The original hypotheses that we had were that the average user scores of the Xbox One and PC platforms are the same, and that the average user scores for the Action and Sports genres are the same.\n\nWe can conclude, that for the first hypothesis, the two platforms did result in different user scores. This makes sense from our prelimiary visualisation of the data that users typically score games higher for games that have sold better amongst consumers. Because the Xbox tended to sell higher amounts per game as compared to the PC, it would make sense that the scoring would be higher.\n\nAs for the second hypothesis, we were testing the alternative hypothesis, and concluded similarly that the genres Action and Sports do score differently among direct consumers. This is most likely due to the same idea that Action sells more overall than Sports. Per game it may be a lower distribution, but Sports games tend to be repetetive year over year, so ratings may not vary, while action titles may be less consistent and vary much more.", "_____no_output_____" ], [ "### Step 6. Write a general conclusion\n<a id='step6'></a>", "_____no_output_____" ], [ "In this project, the data has been reviewed, filled, and changed into the correct types. The data has been filtered, split by year, and annual performances were tracked, and the next year was predicted. Once done so, locational based sale distributions were made by platform, genre, and ESRB rating. Lastly, some hypothesis testing was conducted to determine results from user scores. \n\nWe initially found that console game performances have cyclical performances based on console generations coming and going. the years before 1993 were uneventful and not telling of current market experiences, as they are minimal in comparison. In 2017, we predicted approximately 32.4 million USD in total sales. This is the case without any new consoles entering the market. This decay in sales seems to be the case after a lack of new market participants. \n\nFor the most part, the EU and North America had largely similar preferences for platforms and Genres, while Japan prefered handheld based platforms and role-playing games. All groups had little distinction from each other when it came to the games' ESRB ratings.\n\nFor the first hypothesis, we used a null hypothesis that these two sets of user scores were the same, based on a 5% significance level. By comparing these sets with the null hypothesis, that they were the same, we concluded that it failed the null hypothesis, so it was rejected. \n\nWe then used the alternative hypothesis that these two sets of user scores were indeed different, based on the same level of significance. By comparing these two sets with the null hypothesis, that they were the same, we concluded that it failed the null hypothesis, so it was rejected.\n\nWe believe that any advertising should continue to be focused on large, triple A titles. They typically meet the criteria for driving sales, while there is a large number of games that tend to minimally affect the sales. Also, sales have gone down, likely in anticipation for the next generation of platforms, to drive a new wave of sales, as the current generation is starting to decay.", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown" ]

[ [ "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown", "markdown", "markdown", "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown", "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code", "code" ], [ "markdown", "markdown", "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown", "markdown", "markdown" ] ]

[ "CC-BY-4.0" ]

[ "CC-BY-4.0" ]

[ "CC-BY-4.0" ]

[ [ [ "# 📝 Exercise M6.03\n\nThe aim of this exercise is to:\n\n* verifying if a random forest or a gradient-boosting decision tree overfit\n if the number of estimators is not properly chosen;\n* use the early-stopping strategy to avoid adding unnecessary trees, to\n get the best generalization performances.\n\nWe will use the California housing dataset to conduct our experiments.", "_____no_output_____" ] ], [ [ "from sklearn.datasets import fetch_california_housing\nfrom sklearn.model_selection import train_test_split\n\ndata, target = fetch_california_housing(return_X_y=True, as_frame=True)\ntarget *= 100 # rescale the target in k$\ndata_train, data_test, target_train, target_test = train_test_split(\n data, target, random_state=0, test_size=0.5)", "_____no_output_____" ] ], [ [ "<div class=\"admonition note alert alert-info\">\n<p class=\"first admonition-title\" style=\"font-weight: bold;\">Note</p>\n<p class=\"last\">If you want a deeper overview regarding this dataset, you can refer to the\nAppendix - Datasets description section at the end of this MOOC.</p>\n</div>", "_____no_output_____" ], [ "Create a gradient boosting decision tree with `max_depth=5` and\n`learning_rate=0.5`.", "_____no_output_____" ] ], [ [ "# Write your code here.", "_____no_output_____" ] ], [ [ "\nAlso create a random forest with fully grown trees by setting `max_depth=None`.", "_____no_output_____" ] ], [ [ "# Write your code here.", "_____no_output_____" ] ], [ [ "\nFor both the gradient-boosting and random forest models, create a validation\ncurve using the training set to assess the impact of the number of trees on\nthe performance of each model. Evaluate the list of parameters `param_range =\n[1, 2, 5, 10, 20, 50, 100]` and use the mean absolute error.", "_____no_output_____" ] ], [ [ "# Write your code here.", "_____no_output_____" ] ], [ [ "Both gradient boosting and random forest models will always improve when\nincreasing the number of trees in the ensemble. However, it will reach a\nplateau where adding new trees will just make fitting and scoring slower.\n\nTo avoid adding new unnecessary tree, unlike random-forest gradient-boosting\noffers an early-stopping option. Internally, the algorithm will use an\nout-of-sample set to compute the generalization performance of the model at\neach addition of a tree. Thus, if the generalization performance is not\nimproving for several iterations, it will stop adding trees.\n\nNow, create a gradient-boosting model with `n_estimators=1_000`. This number\nof trees will be too large. Change the parameter `n_iter_no_change` such\nthat the gradient boosting fitting will stop after adding 5 trees that do not\nimprove the overall generalization performance.", "_____no_output_____" ] ], [ [ "# Write your code here.", "_____no_output_____" ] ], [ [ "Estimate the generalization performance of this model again using\nthe `sklearn.metrics.mean_absolute_error` metric but this time using\nthe test set that we held out at the beginning of the notebook.\nCompare the resulting value with the values observed in the validation\ncurve.", "_____no_output_____" ] ], [ [ "# Write your code here.", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Convolutional Layer\n\nIn this notebook, we visualize four filtered outputs (a.k.a. activation maps) of a convolutional layer. \n\nIn this example, *we* are defining four filters that are applied to an input image by initializing the **weights** of a convolutional layer, but a trained CNN will learn the values of these weights.\n\n<img src='notebook_ims/conv_layer.gif' height=60% width=60% />", "_____no_output_____" ], [ "### Import the image", "_____no_output_____" ] ], [ [ "import cv2\nimport matplotlib.pyplot as plt\n%matplotlib inline\n\n# TODO: Feel free to try out your own images here by changing img_path\n# to a file path to another image on your computer!\nimg_path = 'data/udacity_sdc.png'\n\n# load color image \nbgr_img = cv2.imread(img_path)\n# convert to grayscale\ngray_img = cv2.cvtColor(bgr_img, cv2.COLOR_BGR2GRAY)\n\n# normalize, rescale entries to lie in [0,1]\ngray_img = gray_img.astype(\"float32\")/255\n\n# plot image\nplt.imshow(gray_img, cmap='gray')\nplt.show()", "_____no_output_____" ] ], [ [ "### Define and visualize the filters", "_____no_output_____" ] ], [ [ "import numpy as np\n\n## TODO: Feel free to modify the numbers here, to try out another filter!\nfilter_vals = np.array([[1, 1, 1, -1], [1, 1, -1, 1], [1, -1, 1, 1], [-1, 1, 1, 1]])\n\nprint('Filter shape: ', filter_vals.shape)\n", "Filter shape: (4, 4)\n" ], [ "# Defining four different filters, \n# all of which are linear combinations of the `filter_vals` defined above\n\n# define four filters\nfilter_1 = filter_vals\nfilter_2 = -filter_1\nfilter_3 = filter_1.T\nfilter_4 = -filter_3\nfilters = np.array([filter_1, filter_2, filter_3, filter_4])\n\n# For an example, print out the values of filter 1\nprint('Filter 1: \\n', filter_1)", "Filter 1: \n [[ 1 1 1 -1]\n [ 1 1 -1 1]\n [ 1 -1 1 1]\n [-1 1 1 1]]\n" ], [ "# visualize all four filters\nfig = plt.figure(figsize=(10, 5))\nfor i in range(4):\n ax = fig.add_subplot(1, 4, i+1, xticks=[], yticks=[])\n ax.imshow(filters[i], cmap='gray')\n ax.set_title('Filter %s' % str(i+1))\n width, height = filters[i].shape\n for x in range(width):\n for y in range(height):\n ax.annotate(str(filters[i][x][y]), xy=(y,x),\n horizontalalignment='center',\n verticalalignment='center',\n color='white' if filters[i][x][y]<0 else 'black')", "_____no_output_____" ] ], [ [ "## Define a convolutional layer \n\nThe various layers that make up any neural network are documented, [here](http://pytorch.org/docs/stable/nn.html). For a convolutional neural network, we'll start by defining a:\n* Convolutional layer\n\nInitialize a single convolutional layer so that it contains all your created filters. Note that you are not training this network; you are initializing the weights in a convolutional layer so that you can visualize what happens after a forward pass through this network!\n\n\n#### `__init__` and `forward`\nTo define a neural network in PyTorch, you define the layers of a model in the function `__init__` and define the forward behavior of a network that applyies those initialized layers to an input (`x`) in the function `forward`. In PyTorch we convert all inputs into the Tensor datatype, which is similar to a list data type in Python. \n\nBelow, I define the structure of a class called `Net` that has a convolutional layer that can contain four 3x3 grayscale filters.", "_____no_output_____" ] ], [ [ "import torch\nimport torch.nn as nn\nimport torch.nn.functional as F\n \n# define a neural network with a single convolutional layer with four filters\nclass Net(nn.Module):\n \n def __init__(self, weight):\n super(Net, self).__init__()\n # initializes the weights of the convolutional layer to be the weights of the 4 defined filters\n k_height, k_width = weight.shape[2:]\n # assumes there are 4 grayscale filters\n self.conv = nn.Conv2d(1, 4, kernel_size=(k_height, k_width), bias=False)\n self.conv.weight = torch.nn.Parameter(weight)\n\n def forward(self, x):\n # calculates the output of a convolutional layer\n # pre- and post-activation\n conv_x = self.conv(x)\n activated_x = F.relu(conv_x)\n \n # returns both layers\n return conv_x, activated_x\n \n# instantiate the model and set the weights\nweight = torch.from_numpy(filters).unsqueeze(1).type(torch.FloatTensor)\nmodel = Net(weight)\n\n# print out the layer in the network\nprint(model)", "Net(\n (conv): Conv2d(1, 4, kernel_size=(4, 4), stride=(1, 1), bias=False)\n)\n" ] ], [ [ "### Visualize the output of each filter\n\nFirst, we'll define a helper function, `viz_layer` that takes in a specific layer and number of filters (optional argument), and displays the output of that layer once an image has been passed through.", "_____no_output_____" ] ], [ [ "# helper function for visualizing the output of a given layer\n# default number of filters is 4\ndef viz_layer(layer, n_filters= 4):\n fig = plt.figure(figsize=(20, 20))\n \n for i in range(n_filters):\n ax = fig.add_subplot(1, n_filters, i+1, xticks=[], yticks=[])\n # grab layer outputs\n ax.imshow(np.squeeze(layer[0,i].data.numpy()), cmap='gray')\n ax.set_title('Output %s' % str(i+1))", "_____no_output_____" ] ], [ [ "Let's look at the output of a convolutional layer, before and after a ReLu activation function is applied.", "_____no_output_____" ] ], [ [ "# plot original image\nplt.imshow(gray_img, cmap='gray')\n\n# visualize all filters\nfig = plt.figure(figsize=(12, 6))\nfig.subplots_adjust(left=0, right=1.5, bottom=0.8, top=1, hspace=0.05, wspace=0.05)\nfor i in range(4):\n ax = fig.add_subplot(1, 4, i+1, xticks=[], yticks=[])\n ax.imshow(filters[i], cmap='gray')\n ax.set_title('Filter %s' % str(i+1))\n\n \n# convert the image into an input Tensor\ngray_img_tensor = torch.from_numpy(gray_img).unsqueeze(0).unsqueeze(1)\n\n# get the convolutional layer (pre and post activation)\nconv_layer, activated_layer = model(gray_img_tensor)\n\n# visualize the output of a conv layer\nviz_layer(conv_layer)", "_____no_output_____" ] ], [ [ "#### ReLu activation\n\nIn this model, we've used an activation function that scales the output of the convolutional layer. We've chose a ReLu function to do this, and this function simply turns all negative pixel values in 0's (black). See the equation pictured below for input pixel values, `x`. \n\n<img src='notebook_ims/relu_ex.png' height=50% width=50% />", "_____no_output_____" ] ], [ [ "# after a ReLu is applied\n# visualize the output of an activated conv layer\nviz_layer(activated_layer)", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "#A\nimport numpy as np", "_____no_output_____" ], [ "#B\narr1=np.zeros((2,3,4))\narr1", "_____no_output_____" ], [ "#C\narr2=np.ones((2,3,4))\narr2", "_____no_output_____" ], [ "#D\narr3=np.arange(0,1000)\narr3", "_____no_output_____" ], [ "#E\nli1=[2,3.2,5.5,-6.4,-2.2,2.4]\na=np.array(li1)\nprint(a)\ntype(a)\n", "[ 2. 3.2 5.5 -6.4 -2.2 2.4]\n" ], [ "#F\nprint(a[1])", "3.2\n" ], [ "#G\nprint(a[1:4])", "[ 3.2 5.5 -6.4]\n" ], [ "#H\nli2=[[2,3.2,5.5,-6.4,-2.2,2.4],\n [1,22,4,0.1,5.3,-9],\n [3,1,2.1,21,1.1,-2]]\na=np.array(li2)\na.ndim", "_____no_output_____" ], [ "#I\nprint(a[:,2])\nprint(a[1:4,0:4])\nprint(a[1:,2])", "[5.5 4. 2.1]\n[[ 1. 22. 4. 0.1]\n [ 3. 1. 2.1 21. ]]\n[4. 2.1]\n" ], [ "#J\narr4=np.arange(4)\narr5=np.arange(10,14)\narr6=np.array((arr4,arr5))\nprint(arr6.shape)\nprint(arr6.size)\nprint(arr6.max())\nprint(arr6.min())\n\n\n\n", "(2, 4)\n8\n13\n0\n" ], [ "#K-1\narr6.reshape(2,2,2)", "_____no_output_____" ], [ "#K-2\narr6.T", "_____no_output_____" ], [ "#K-3\narr6.flatten()", "_____no_output_____" ], [ "#K-4\narr6.astype('float')", "_____no_output_____" ], [ "#L\narr7=np.arange(1,21)\narr7", "_____no_output_____" ], [ "#M\nli3=[]\n\nfor i in arr7:\n if i % 3==0 and i > 0 and i < 30:\n li3.append(i)\narr8=np.array([li3])\narr8", "_____no_output_____" ], [ "#N\narr9=np.linspace(0,1,8)\narr9", "_____no_output_____" ], [ "#O\narr10=np.arange(1,9)\nA=arr10.reshape(2,4)\nprint(A)", "[[1 2 3 4]\n [5 6 7 8]]\n" ], [ "#P\narr11=np.array([1,2])\nB=arr11\nprint(B)", "[1 2]\n" ], [ "#Q\nB=B[:,np.newaxis]\nA+B", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "<a href=\"https://colab.research.google.com/github/ricospeloacaso/python_para_investimentos/blob/master/16_CVM_Os_Melhores_e_os_Piores_Fundos_de_Investimento_do_mes_Python_para_Investimentos.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>", "_____no_output_____" ], [ "###Ricos pelo Acaso", "_____no_output_____" ], [ "* Link para o vídeo: https://youtu.be/NHCUUZOvk7k\n---\n* Base de Dados: http://dados.cvm.gov.br/\n\n", "_____no_output_____" ], [ "###Coletando os dados da CVM", "_____no_output_____" ] ], [ [ "import pandas as pd\npd.set_option(\"display.max_colwidth\", 150)\n#pd.options.display.float_format = '{:.2f}'.format", "_____no_output_____" ] ], [ [ ">Funções que buscam dados no site da CVM e retornam um DataFrame Pandas:\n", "_____no_output_____" ] ], [ [ "", "_____no_output_____" ], [ "def busca_informes_cvm(ano, mes):\n url = 'http://dados.cvm.gov.br/dados/FI/DOC/INF_DIARIO/DADOS/inf_diario_fi_{:02d}{:02d}.csv'.format(ano,mes)\n return pd.read_csv(url, sep=';')", "_____no_output_____" ], [ "def busca_cadastro_cvm(ano, mes, dia):\n url = 'http://dados.cvm.gov.br/dados/FI/CAD/DADOS/inf_cadastral_fi_{}{:02d}{:02d}.csv'.format(ano, mes, dia)\n return pd.read_csv(url, sep=';', encoding='ISO-8859-1')", "_____no_output_____" ] ], [ [ ">Buscando dados no site da CVM", "_____no_output_____" ] ], [ [ "informes_diarios = busca_informes_cvm(2020,4)", "_____no_output_____" ], [ "informes_diarios", "_____no_output_____" ], [ "cadastro_cvm = busca_cadastro_cvm(2020,5,1)", "_____no_output_____" ], [ "cadastro_cvm", "_____no_output_____" ] ], [ [ "###Manipulando os dados da CVM", "_____no_output_____" ], [ ">Definindo filtros para os Fundos de Investimento\n", "_____no_output_____" ] ], [ [ "minimo_cotistas = 100", "_____no_output_____" ] ], [ [ ">Manipulando os dados e aplicando filtros", "_____no_output_____" ] ], [ [ "fundos = informes_diarios[informes_diarios['NR_COTST'] >= minimo_cotistas].pivot(index='DT_COMPTC', columns='CNPJ_FUNDO', values=['VL_TOTAL',\t'VL_QUOTA',\t'VL_PATRIM_LIQ',\t'CAPTC_DIA',\t'RESG_DIA'])", "_____no_output_____" ], [ "fundos", "_____no_output_____" ] ], [ [ ">Normalizando os dados de cotas para efeitos comparativos", "_____no_output_____" ] ], [ [ "cotas_normalizadas = fundos['VL_QUOTA'] / fundos['VL_QUOTA'].iloc[0]", "_____no_output_____" ], [ "cotas_normalizadas", "_____no_output_____" ] ], [ [ "###Fundos de Investimento com os melhores desempenhos em Abril de 2020", "_____no_output_____" ] ], [ [ "melhores = pd.DataFrame()\nmelhores['retorno(%)'] = (cotas_normalizadas.iloc[-1].sort_values(ascending=False)[:5] - 1) * 100\nmelhores", "_____no_output_____" ] ], [ [ ">Buscando dados dos Fundos de Investimento pelo CNPJ", "_____no_output_____" ] ], [ [ "for cnpj in melhores.index:\n fundo = cadastro_cvm[cadastro_cvm['CNPJ_FUNDO'] == cnpj]\n melhores.at[cnpj, 'Fundo de Investimento'] = fundo['DENOM_SOCIAL'].values[0]\n melhores.at[cnpj, 'Classe'] = fundo['CLASSE'].values[0]\n melhores.at[cnpj, 'PL'] = fundo['VL_PATRIM_LIQ'].values[0]", "_____no_output_____" ], [ "melhores", "_____no_output_____" ] ], [ [ "###Fundos de Investimento com os piores desempenhos em Abril de 2020", "_____no_output_____" ] ], [ [ "piores = pd.DataFrame()\npiores['retorno(%)'] = (cotas_normalizadas.iloc[-1].sort_values(ascending=True)[:5] - 1) * 100\npiores", "_____no_output_____" ] ], [ [ ">Buscando dados dos Fundos de Investimento pelo CNPJ", "_____no_output_____" ] ], [ [ "for cnpj in piores.index:\n fundo = cadastro_cvm[cadastro_cvm['CNPJ_FUNDO'] == cnpj]\n piores.at[cnpj, 'Fundo de Investimento'] = fundo['DENOM_SOCIAL'].values[0]\n piores.at[cnpj, 'Classe'] = fundo['CLASSE'].values[0]\n piores.at[cnpj, 'PL'] = fundo['VL_PATRIM_LIQ'].values[0]", "_____no_output_____" ], [ "piores", "_____no_output_____" ], [ "", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown", "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import numpy as np\nimport os\nimport re\nimport pickle", "_____no_output_____" ], [ "def clearstring(string):\n string = re.sub('[^\\'\\\"A-Za-z0-9 ]+', '', string)\n string = string.split(' ')\n string = filter(None, string)\n string = [y.strip() for y in string]\n string = [y for y in string if len(y) > 3 and y.find('nbsp') < 0]\n return ' '.join(string)\n\ndef read_data(location):\n list_folder = os.listdir(location)\n label = list_folder\n label.sort()\n outer_string, outer_label = [], []\n for i in range(len(list_folder)):\n list_file = os.listdir('data/' + list_folder[i])\n strings = []\n for x in range(len(list_file)):\n with open('data/' + list_folder[i] + '/' + list_file[x], 'r') as fopen:\n strings += fopen.read().split('\\n')\n strings = list(filter(None, strings))\n for k in range(len(strings)):\n strings[k] = clearstring(strings[k])\n labels = [i] * len(strings)\n outer_string += strings\n outer_label += labels\n \n dataset = np.array([outer_string, outer_label])\n dataset = dataset.T\n np.random.shuffle(dataset)\n \n return dataset", "_____no_output_____" ], [ "dataset = read_data('data/')\ndataset[:5,:]", "_____no_output_____" ], [ "with open('dataset-emotion.p', 'wb') as fopen:\n pickle.dump(dataset, fopen)", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# HSV Color Space, Balloons", "_____no_output_____" ], [ "### Import resources and display image", "_____no_output_____" ] ], [ [ "import numpy as np\nimport matplotlib.pyplot as plt\nimport cv2\n", "_____no_output_____" ], [ "%matplotlib inline\n\n# Read in the image\nimage = cv2.imread('images/water_balloons.jpg')\n\n# Make a copy of the image\nimage_copy = np.copy(image)\n\n# Change color to RGB (from BGR)\nimage = cv2.cvtColor(image_copy, cv2.COLOR_BGR2RGB)\n\nplt.imshow(image)", "_____no_output_____" ] ], [ [ "### Plot color channels", "_____no_output_____" ] ], [ [ "# RGB channels\nr = image[:,:,0]\ng = image[:,:,1]\nb = image[:,:,2]\n\nf, (ax1, ax2, ax3) = plt.subplots(1, 3, figsize=(20,10))\n\nax1.set_title('Red')\nax1.imshow(r, cmap='gray')\n\nax2.set_title('Green')\nax2.imshow(g, cmap='gray')\n\nax3.set_title('Blue')\nax3.imshow(b, cmap='gray')\n", "_____no_output_____" ], [ "# Convert from RGB to HSV\nhsv = cv2.cvtColor(image, cv2.COLOR_RGB2HSV)\n\n# HSV channels\nh = hsv[:,:,0]\ns = hsv[:,:,1]\nv = hsv[:,:,2]\n\nf, (ax1, ax2, ax3) = plt.subplots(1, 3, figsize=(20,10))\n\nax1.set_title('Hue')\nax1.imshow(h, cmap='gray')\n\nax2.set_title('Saturation')\nax2.imshow(s, cmap='gray')\n\nax3.set_title('Value')\nax3.imshow(v, cmap='gray')\n", "_____no_output_____" ] ], [ [ "### Define pink and hue selection thresholds", "_____no_output_____" ] ], [ [ "# Define our color selection criteria in HSV values\nlower_hue = np.array([160,0,0]) \nupper_hue = np.array([180,255,255])\n", "_____no_output_____" ], [ "# Define our color selection criteria in RGB values\nlower_pink = np.array([180,0,100]) \nupper_pink = np.array([255,255,230])", "_____no_output_____" ] ], [ [ "### Mask the image ", "_____no_output_____" ] ], [ [ "# Define the masked area in RGB space\nmask_rgb = cv2.inRange(image, lower_pink, upper_pink)\n\n# mask the image\nmasked_image = np.copy(image)\nmasked_image[mask_rgb==0] = [0,0,0]\n\n# Vizualize the mask\nplt.imshow(masked_image)", "_____no_output_____" ], [ "# Now try HSV!\n\n# Define the masked area in HSV space\nmask_hsv = cv2.inRange(hsv, lower_hue, upper_hue)\n\n# mask the image\nmasked_image = np.copy(image)\nmasked_image[mask_hsv==0] = [0,0,0]\n\n# Vizualize the mask\nplt.imshow(masked_image)", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]