├── daal ├── web-usage │ ├── knn-web-usage.pdf │ ├── pca-web-usage.pdf │ ├── kmeans-web-usage.pdf │ ├── log-reg-web-usage.pdf │ └── decision-forest-web-usage.pdf ├── k-nearest-neighbors-example.ipynb ├── pca-example.ipynb ├── kmeans-example.ipynb ├── decision-forest-example.ipynb └── logistic-regression-example.ipynb ├── README.md └── LICENSE /daal/web-usage/knn-web-usage.pdf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/intel/aws-sagemaker-marketplace/master/daal/web-usage/knn-web-usage.pdf -------------------------------------------------------------------------------- /daal/web-usage/pca-web-usage.pdf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/intel/aws-sagemaker-marketplace/master/daal/web-usage/pca-web-usage.pdf -------------------------------------------------------------------------------- /daal/web-usage/kmeans-web-usage.pdf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/intel/aws-sagemaker-marketplace/master/daal/web-usage/kmeans-web-usage.pdf -------------------------------------------------------------------------------- /daal/web-usage/log-reg-web-usage.pdf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/intel/aws-sagemaker-marketplace/master/daal/web-usage/log-reg-web-usage.pdf -------------------------------------------------------------------------------- /daal/web-usage/decision-forest-web-usage.pdf: -------------------------------------------------------------------------------- https://raw.githubusercontent.com/intel/aws-sagemaker-marketplace/master/daal/web-usage/decision-forest-web-usage.pdf -------------------------------------------------------------------------------- /README.md: -------------------------------------------------------------------------------- 1 | # DISCONTINUATION OF PROJECT # 2 | This project will no longer be maintained by Intel. 3 | Intel has ceased development and contributions including, but not limited to, maintenance, bug fixes, new releases, or updates, to this project. 4 | Intel no longer accepts patches to this project. 5 | 6 | 7 | # Intel AI Software on AWS SageMaker Marketplace 8 | 9 | This repo contain example notebooks with instructions on using Intel AI Software listed in AWS SageMaker Marketplace. 10 | 11 | Intel and [AWS](https://aws.amazon.com/), a subsidiary of Amazon that provides on-demand cloud computing services, have worked [together for over a decade](https://aws.amazon.com/intel/) to ensure AWS services run on a platform optimized for customer workloads at the best value. At the annual [AWS re:Invent](https://reinvent.awsevents.com/) conference from November 26-30, AWS is highlighting advancements in this ongoing collaboration, including the new [AWS Marketplace for Machine Learning](https://aws.amazon.com/mp/ai/), developed by AWS for its Machine Learning as a Service (MLaaS) platform, [Amazon SageMaker](https://aws.amazon.com/sagemaker/). Read more at the [Intel AI Blog](https://ai.intel.com/intel-software-development-tools-on-the-new-aws-marketplace-for-machine-learning/). 12 | 13 | Intel has listed the following AI software tools and libraries on [AWS Marketplace](https://aws.amazon.com/marketplace/search/results?page=1&filters=vendor_id,fulfillment_options&vendor_id=a35b6cd1-6ad5-47c4-ac34-3e7b30c6d3a9&fulfillment_options=SAGEMAKER) (more coming soon): 14 | - [Intel® Data Analytics Acceleration Library (DAAL)](https://software.intel.com/en-us/intel-daal) 15 | - Principal Component Analysis (PCA) - [SDK usage](daal/pca-example.ipynb), [AWS Portal usage](daal/web-usage/pca-web-usage.pdf) 16 | - Logistic Regression - [SDK usage](daal/logistic-regression-example.ipynb), [AWS Portal usage](daal/web-usage/log-reg-web-usage.pdf) 17 | - Decision Forest Regression - [SDK usage](daal/decision-forest-example.ipynb), [AWS Portal usage](daal/web-usage/decision-forest-web-usage.pdf) 18 | - Decision Forest Classification - [SDK usage](daal/decision-forest-example.ipynb), [AWS Portal usage](daal/web-usage/decision-forest-web-usage.pdf) 19 | - K-Means- [SDK usage](daal/kmeans-example.ipynb), [AWS Portal usage](daal/web-usage/kmeans-web-usage.pdf) 20 | - k-Nearest Neighbors (kNN) Classifier - [SDK usage](daal/k-nearest-neighbors-example.ipynb), [AWS Portal usage](daal/web-usage/knn-web-usage.pdf) 21 | 22 | - [Intel® Optimized Deep Learning Libraries](https://www.intel.ai/framework-optimizations/) 23 | - Intel Optimized MXNet ResNet50 Inference 24 | - [BigDL](https://software.intel.com/en-us/ai-academy/frameworks/bigdl), a distributed deep learning library for Apache Spark. 25 | - BigDL Text Classifier on Analytics Zoo 26 | 27 | 28 | -------------------------------------------------------------------------------- /LICENSE: -------------------------------------------------------------------------------- 1 | Apache License 2 | Version 2.0, January 2004 3 | http://www.apache.org/licenses/ 4 | 5 | TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION 6 | 7 | 1. Definitions. 8 | 9 | "License" shall mean the terms and conditions for use, reproduction, 10 | and distribution as defined by Sections 1 through 9 of this document. 11 | 12 | "Licensor" shall mean the copyright owner or entity authorized by 13 | the copyright owner that is granting the License. 14 | 15 | "Legal Entity" shall mean the union of the acting entity and all 16 | other entities that control, are controlled by, or are under common 17 | control with that entity. For the purposes of this definition, 18 | "control" means (i) the power, direct or indirect, to cause the 19 | direction or management of such entity, whether by contract or 20 | otherwise, or (ii) ownership of fifty percent (50%) or more of the 21 | outstanding shares, or (iii) beneficial ownership of such entity. 22 | 23 | "You" (or "Your") shall mean an individual or Legal Entity 24 | exercising permissions granted by this License. 25 | 26 | "Source" form shall mean the preferred form for making modifications, 27 | including but not limited to software source code, documentation 28 | source, and configuration files. 29 | 30 | "Object" form shall mean any form resulting from mechanical 31 | transformation or translation of a Source form, including but 32 | not limited to compiled object code, generated documentation, 33 | and conversions to other media types. 34 | 35 | "Work" shall mean the work of authorship, whether in Source or 36 | Object form, made available under the License, as indicated by a 37 | copyright notice that is included in or attached to the work 38 | (an example is provided in the Appendix below). 39 | 40 | "Derivative Works" shall mean any work, whether in Source or Object 41 | form, that is based on (or derived from) the Work and for which the 42 | editorial revisions, annotations, elaborations, or other modifications 43 | represent, as a whole, an original work of authorship. For the purposes 44 | of this License, Derivative Works shall not include works that remain 45 | separable from, or merely link (or bind by name) to the interfaces of, 46 | the Work and Derivative Works thereof. 47 | 48 | "Contribution" shall mean any work of authorship, including 49 | the original version of the Work and any modifications or additions 50 | to that Work or Derivative Works thereof, that is intentionally 51 | submitted to Licensor for inclusion in the Work by the copyright owner 52 | or by an individual or Legal Entity authorized to submit on behalf of 53 | the copyright owner. For the purposes of this definition, "submitted" 54 | means any form of electronic, verbal, or written communication sent 55 | to the Licensor or its representatives, including but not limited to 56 | communication on electronic mailing lists, source code control systems, 57 | and issue tracking systems that are managed by, or on behalf of, the 58 | Licensor for the purpose of discussing and improving the Work, but 59 | excluding communication that is conspicuously marked or otherwise 60 | designated in writing by the copyright owner as "Not a Contribution." 61 | 62 | "Contributor" shall mean Licensor and any individual or Legal Entity 63 | on behalf of whom a Contribution has been received by Licensor and 64 | subsequently incorporated within the Work. 65 | 66 | 2. Grant of Copyright License. Subject to the terms and conditions of 67 | this License, each Contributor hereby grants to You a perpetual, 68 | worldwide, non-exclusive, no-charge, royalty-free, irrevocable 69 | copyright license to reproduce, prepare Derivative Works of, 70 | publicly display, publicly perform, sublicense, and distribute the 71 | Work and such Derivative Works in Source or Object form. 72 | 73 | 3. Grant of Patent License. Subject to the terms and conditions of 74 | this License, each Contributor hereby grants to You a perpetual, 75 | worldwide, non-exclusive, no-charge, royalty-free, irrevocable 76 | (except as stated in this section) patent license to make, have made, 77 | use, offer to sell, sell, import, and otherwise transfer the Work, 78 | where such license applies only to those patent claims licensable 79 | by such Contributor that are necessarily infringed by their 80 | Contribution(s) alone or by combination of their Contribution(s) 81 | with the Work to which such Contribution(s) was submitted. If You 82 | institute patent litigation against any entity (including a 83 | cross-claim or counterclaim in a lawsuit) alleging that the Work 84 | or a Contribution incorporated within the Work constitutes direct 85 | or contributory patent infringement, then any patent licenses 86 | granted to You under this License for that Work shall terminate 87 | as of the date such litigation is filed. 88 | 89 | 4. Redistribution. You may reproduce and distribute copies of the 90 | Work or Derivative Works thereof in any medium, with or without 91 | modifications, and in Source or Object form, provided that You 92 | meet the following conditions: 93 | 94 | (a) You must give any other recipients of the Work or 95 | Derivative Works a copy of this License; and 96 | 97 | (b) You must cause any modified files to carry prominent notices 98 | stating that You changed the files; and 99 | 100 | (c) You must retain, in the Source form of any Derivative Works 101 | that You distribute, all copyright, patent, trademark, and 102 | attribution notices from the Source form of the Work, 103 | excluding those notices that do not pertain to any part of 104 | the Derivative Works; and 105 | 106 | (d) If the Work includes a "NOTICE" text file as part of its 107 | distribution, then any Derivative Works that You distribute must 108 | include a readable copy of the attribution notices contained 109 | within such NOTICE file, excluding those notices that do not 110 | pertain to any part of the Derivative Works, in at least one 111 | of the following places: within a NOTICE text file distributed 112 | as part of the Derivative Works; within the Source form or 113 | documentation, if provided along with the Derivative Works; or, 114 | within a display generated by the Derivative Works, if and 115 | wherever such third-party notices normally appear. The contents 116 | of the NOTICE file are for informational purposes only and 117 | do not modify the License. You may add Your own attribution 118 | notices within Derivative Works that You distribute, alongside 119 | or as an addendum to the NOTICE text from the Work, provided 120 | that such additional attribution notices cannot be construed 121 | as modifying the License. 122 | 123 | You may add Your own copyright statement to Your modifications and 124 | may provide additional or different license terms and conditions 125 | for use, reproduction, or distribution of Your modifications, or 126 | for any such Derivative Works as a whole, provided Your use, 127 | reproduction, and distribution of the Work otherwise complies with 128 | the conditions stated in this License. 129 | 130 | 5. Submission of Contributions. Unless You explicitly state otherwise, 131 | any Contribution intentionally submitted for inclusion in the Work 132 | by You to the Licensor shall be under the terms and conditions of 133 | this License, without any additional terms or conditions. 134 | Notwithstanding the above, nothing herein shall supersede or modify 135 | the terms of any separate license agreement you may have executed 136 | with Licensor regarding such Contributions. 137 | 138 | 6. Trademarks. This License does not grant permission to use the trade 139 | names, trademarks, service marks, or product names of the Licensor, 140 | except as required for reasonable and customary use in describing the 141 | origin of the Work and reproducing the content of the NOTICE file. 142 | 143 | 7. Disclaimer of Warranty. Unless required by applicable law or 144 | agreed to in writing, Licensor provides the Work (and each 145 | Contributor provides its Contributions) on an "AS IS" BASIS, 146 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or 147 | implied, including, without limitation, any warranties or conditions 148 | of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A 149 | PARTICULAR PURPOSE. You are solely responsible for determining the 150 | appropriateness of using or redistributing the Work and assume any 151 | risks associated with Your exercise of permissions under this License. 152 | 153 | 8. Limitation of Liability. In no event and under no legal theory, 154 | whether in tort (including negligence), contract, or otherwise, 155 | unless required by applicable law (such as deliberate and grossly 156 | negligent acts) or agreed to in writing, shall any Contributor be 157 | liable to You for damages, including any direct, indirect, special, 158 | incidental, or consequential damages of any character arising as a 159 | result of this License or out of the use or inability to use the 160 | Work (including but not limited to damages for loss of goodwill, 161 | work stoppage, computer failure or malfunction, or any and all 162 | other commercial damages or losses), even if such Contributor 163 | has been advised of the possibility of such damages. 164 | 165 | 9. Accepting Warranty or Additional Liability. While redistributing 166 | the Work or Derivative Works thereof, You may choose to offer, 167 | and charge a fee for, acceptance of support, warranty, indemnity, 168 | or other liability obligations and/or rights consistent with this 169 | License. However, in accepting such obligations, You may act only 170 | on Your own behalf and on Your sole responsibility, not on behalf 171 | of any other Contributor, and only if You agree to indemnify, 172 | defend, and hold each Contributor harmless for any liability 173 | incurred by, or claims asserted against, such Contributor by reason 174 | of your accepting any such warranty or additional liability. 175 | 176 | END OF TERMS AND CONDITIONS 177 | 178 | APPENDIX: How to apply the Apache License to your work. 179 | 180 | To apply the Apache License to your work, attach the following 181 | boilerplate notice, with the fields enclosed by brackets "[]" 182 | replaced with your own identifying information. (Don't include 183 | the brackets!) The text should be enclosed in the appropriate 184 | comment syntax for the file format. We also recommend that a 185 | file or class name and description of purpose be included on the 186 | same "printed page" as the copyright notice for easier 187 | identification within third-party archives. 188 | 189 | Copyright [yyyy] [name of copyright owner] 190 | 191 | Licensed under the Apache License, Version 2.0 (the "License"); 192 | you may not use this file except in compliance with the License. 193 | You may obtain a copy of the License at 194 | 195 | http://www.apache.org/licenses/LICENSE-2.0 196 | 197 | Unless required by applicable law or agreed to in writing, software 198 | distributed under the License is distributed on an "AS IS" BASIS, 199 | WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 200 | See the License for the specific language governing permissions and 201 | limitations under the License. 202 | -------------------------------------------------------------------------------- /daal/k-nearest-neighbors-example.ipynb: -------------------------------------------------------------------------------- 1 | { 2 | "cells": [ 3 | { 4 | "cell_type": "markdown", 5 | "metadata": {}, 6 | "source": [ 7 | "# k-Nearest Neighbors (kNN) Classifier with Intel® Data Analytics Acceleration Library in Amazon SageMaker\n", 8 | "\n", 9 | "## Introduction\n", 10 | "\n", 11 | "Intel® Data Analytics Acceleration Library (Intel® DAAL) is the library of Intel® architecture optimized building blocks covering all stages of data analytics: data acquisition from a data source, preprocessing, transformation, data mining, modeling, validation, and decision making. One of its algorithms is kNN.\n", 12 | "\n", 13 | "k-Nearest Neighbors (kNN) classification is a non-parametric classification algorithm. The model of the kNN classifier is based on feature vectors and class labels from the training data set. This classifier induces the class of the query vector from the labels of the feature vectors in the training data set to which the query vector is similar. A similarity between feature vectors is determined by the type of distance (for example, Euclidian) in a multidimensional feature space.\n", 14 | "\n", 15 | "Intel® DAAL developer guide: https://software.intel.com/en-us/daal-programming-guide\n", 16 | "\n", 17 | "Intel® DAAL documentation for kNN: https://software.intel.com/en-us/daal-programming-guide-k-nearest-neighbors-knn-classifier" 18 | ] 19 | }, 20 | { 21 | "cell_type": "markdown", 22 | "metadata": {}, 23 | "source": [ 24 | "## kNN Usage with SageMaker Estimator\n", 25 | "Firstly, you need to import SageMaker package, get execution role and create session." 26 | ] 27 | }, 28 | { 29 | "cell_type": "code", 30 | "execution_count": 1, 31 | "metadata": {}, 32 | "outputs": [], 33 | "source": [ 34 | "import sagemaker\n", 35 | "\n", 36 | "role = sagemaker.get_execution_role()\n", 37 | "sess = sagemaker.Session()" 38 | ] 39 | }, 40 | { 41 | "cell_type": "markdown", 42 | "metadata": {}, 43 | "source": [ 44 | "Secondly, you can specify parameters of kNN.\n", 45 | "#### Hyperparameters\n", 46 | "\n", 47 | " \n", 48 | " \n", 49 | " \n", 50 | " \n", 51 | " \n", 52 | " \n", 53 | " \n", 54 | " \n", 55 | " \n", 56 | " \n", 57 | " \n", 58 | " \n", 59 | " \n", 60 | " \n", 61 | " \n", 62 | " \n", 63 | " \n", 64 | " \n", 65 | " \n", 66 | " \n", 67 | " \n", 68 | " \n", 69 | " \n", 70 | " \n", 71 | " \n", 72 | " \n", 73 | " \n", 74 | " \n", 75 | " \n", 76 | " \n", 77 | " \n", 78 | " \n", 79 | " \n", 80 | " \n", 81 | " \n", 82 | " \n", 83 | " \n", 84 | " \n", 85 | " \n", 86 | " \n", 87 | " \n", 88 | " \n", 89 | "
Parameter nameTypeDefault valueDescription
nClassesint2Number of classes in data
fptypestr\"double\"The floating-point type that the algorithm uses for intermediate computations. Can be \"float\" or \"double\"
methodstr\"defaultDense\"The computation method used by the K-D tree based kNN classification. The only training method supported so far is the default dense method.
kint1The number of neighbors
dataUseInModelstr\"doNotUse\"A parameter to enable/disable use of the input data set in the kNN model. Possible values:
\"doNotUse\" - the algorithm does not include the input data and labels in the trained kNN model but creates a copy of the input data set
\"doUse\" - the algorithm includes the input data and labels in the trained kNN model
seedint777Seed for random number generator engine that is used internally to perform sampling needed to choose dimensions and cut-points for the K-D tree.
\n", 90 | "\n", 91 | "Example of hyperparameters dictionary:" 92 | ] 93 | }, 94 | { 95 | "cell_type": "code", 96 | "execution_count": 3, 97 | "metadata": {}, 98 | "outputs": [], 99 | "source": [ 100 | "knn_params = {\n", 101 | " \"nClasses\":2,\n", 102 | " \"fptype\":\"double\",\n", 103 | " \"method\":\"defaultDense\",\n", 104 | " \"dataUseInModel\":\"doNotUse\",\n", 105 | " \"seed\": 777,\n", 106 | " \"k\":1\n", 107 | "}" 108 | ] 109 | }, 110 | { 111 | "cell_type": "markdown", 112 | "metadata": {}, 113 | "source": [ 114 | "Then, you need to create SageMaker Estimator instance with following parameters:\n", 115 | "\n", 116 | " \n", 117 | " \n", 118 | " \n", 119 | " \n", 120 | " \n", 121 | " \n", 122 | " \n", 123 | " \n", 124 | " \n", 125 | " \n", 126 | " \n", 127 | " \n", 128 | " \n", 129 | " \n", 130 | " \n", 131 | " \n", 132 | " \n", 133 | " \n", 134 | " \n", 135 | " \n", 136 | " \n", 137 | " \n", 138 | " \n", 139 | " \n", 140 | " \n", 141 | " \n", 142 | " \n", 143 | " \n", 144 | " \n", 145 | " \n", 146 | " \n", 147 | " \n", 148 | " \n", 149 | " \n", 150 | " \n", 151 | " \n", 152 | "
Parameter nameDescription
image_nameThe container image to use for training
roleAn AWS IAM role. The SageMaker training jobs and APIs that create SageMaker endpoints use this role to access training data and models
train_instance_countNumber of Amazon EC2 instances to use for training. Should be 1, because it is not distributed version of algorithm
train_instance_typeType of EC2 instance to use for training. See available types on Amazon Marketplace page of algorithm
input_modeThe input mode that the algorithm supports. May be \"File\" or \"Pipe\"
output_pathS3 location for saving the trainig result (model artifacts and output files)
sagemaker_sessionSession object which manages interactions with Amazon SageMaker APIs and any other AWS services needed
hyperparametersDictionary containing the hyperparameters to initialize this estimator with
\n", 153 | "Full SageMaker Estimator documentation: https://sagemaker.readthedocs.io/en/latest/estimators.html" 154 | ] 155 | }, 156 | { 157 | "cell_type": "code", 158 | "execution_count": 6, 159 | "metadata": {}, 160 | "outputs": [], 161 | "source": [ 162 | "daal_knn_arn = \"\" # you can find it on algorithm page in your subscriptions\n", 163 | "\n", 164 | "daal_knn = sagemaker.algorithm.AlgorithmEstimator(\n", 165 | " algorithm_arn=daal_knn_arn,\n", 166 | " role=role,\n", 167 | " base_job_name=\"\",\n", 168 | " train_instance_count=1,\n", 169 | " train_instance_type='ml.m4.xlarge',\n", 170 | " input_mode=\"File\",\n", 171 | " output_path=\"s3:///\",\n", 172 | " sagemaker_session=sess,\n", 173 | " hyperparameters=knn_params\n", 174 | ")" 175 | ] 176 | }, 177 | { 178 | "cell_type": "markdown", 179 | "metadata": {}, 180 | "source": [ 181 | "### Training stage\n", 182 | "On training stage, kNN algorithm consume input data from S3 location.\n", 183 | "This container supports only .csv (\"comma-separated values\") files." 184 | ] 185 | }, 186 | { 187 | "cell_type": "code", 188 | "execution_count": 7, 189 | "metadata": { 190 | "scrolled": false 191 | }, 192 | "outputs": [ 193 | { 194 | "name": "stderr", 195 | "output_type": "stream", 196 | "text": [ 197 | "INFO:sagemaker:Creating training-job with name: daal-knn-sm-2019-02-24-15-31-09-721\n" 198 | ] 199 | }, 200 | { 201 | "name": "stdout", 202 | "output_type": "stream", 203 | "text": [ 204 | "2019-02-24 15:31:09 Starting - Starting the training job...\n", 205 | "2019-02-24 15:31:11 Starting - Launching requested ML instances......\n", 206 | "2019-02-24 15:32:12 Starting - Preparing the instances for training...\n", 207 | "2019-02-24 15:33:02 Downloading - Downloading input data\n", 208 | "2019-02-24 15:33:02 Training - Downloading the training image......\n", 209 | "2019-02-24 15:34:09 Uploading - Uploading generated training model\n", 210 | "\u001b[31m2019-02-24 15:34:04 INFO Container setup completed, In Docker entrypoint - train... \u001b[0m\n", 211 | "\u001b[31m2019-02-24 15:34:04 INFO Default Hyperparameters loaded: \u001b[0m\n", 212 | "\u001b[31m2019-02-24 15:34:04 INFO \u001b[0m\n", 213 | "\u001b[31m{'dataUseInModel': 'doNotUse',\n", 214 | " 'fptype': 'double',\n", 215 | " 'k': '1',\n", 216 | " 'method': 'defaultDense',\n", 217 | " 'nClasses': '2'}\u001b[0m\n", 218 | "\u001b[31m2019-02-24 15:34:04 INFO Updated with user hyperparameters, Final Hyperparameters: \u001b[0m\n", 219 | "\u001b[31m2019-02-24 15:34:04 INFO \u001b[0m\n", 220 | "\u001b[31m{'dataUseInModel': 'doNotUse',\n", 221 | " 'fptype': 'double',\n", 222 | " 'k': '1',\n", 223 | " 'method': 'defaultDense',\n", 224 | " 'nClasses': '2',\n", 225 | " 'seed': '777'}\u001b[0m\n", 226 | "\u001b[31m2019-02-24 15:34:04 INFO Reading training data... \u001b[0m\n", 227 | "\u001b[31m2019-02-24 15:34:04 INFO Train data shape: (20000, 6)\u001b[0m\n", 228 | "\u001b[31m2019-02-24 15:34:04 INFO Files loading time: 0.024965763092041016\u001b[0m\n", 229 | "\u001b[31m2019-02-24 15:34:04 INFO Training Data Shape: (20000, 5)\u001b[0m\n", 230 | "\u001b[31m2019-02-24 15:34:04 INFO Starting DAAL KNN Kdtree training...\u001b[0m\n", 231 | "\u001b[31m2019-02-24 15:34:04 INFO Training time in sec = 0.024128198623657227\u001b[0m\n", 232 | "\u001b[31m2019-02-24 15:34:04 INFO Training complete.\u001b[0m\n", 233 | "\u001b[31m2019-02-24 15:34:04 INFO Saving model results...\u001b[0m\n", 234 | "\u001b[31m2019-02-24 15:34:04 INFO Parameters saved at /opt/ml/model/parameters.json\u001b[0m\n", 235 | "\n", 236 | "2019-02-24 15:34:15 Completed - Training job completed\n", 237 | "Billable seconds: 80\n" 238 | ] 239 | } 240 | ], 241 | "source": [ 242 | "daal_knn.fit({\"training\": \"s3:///\"})" 243 | ] 244 | }, 245 | { 246 | "cell_type": "markdown", 247 | "metadata": {}, 248 | "source": [ 249 | "### Real-time prediction\n", 250 | "Firstly, you need to deploy SageMaker endpoint that consumes data." 251 | ] 252 | }, 253 | { 254 | "cell_type": "code", 255 | "execution_count": 8, 256 | "metadata": { 257 | "scrolled": true 258 | }, 259 | "outputs": [ 260 | { 261 | "name": "stderr", 262 | "output_type": "stream", 263 | "text": [ 264 | "INFO:sagemaker:Creating model package with name: daal-knn-2019-02-24-15-35-14-387\n" 265 | ] 266 | }, 267 | { 268 | "name": "stdout", 269 | "output_type": "stream", 270 | "text": [ 271 | ".........." 272 | ] 273 | }, 274 | { 275 | "name": "stderr", 276 | "output_type": "stream", 277 | "text": [ 278 | "INFO:sagemaker:Creating model with name: daal-knn-2019-02-24-15-35-14-387-2019-02-24-15-35-59-854\n" 279 | ] 280 | }, 281 | { 282 | "name": "stdout", 283 | "output_type": "stream", 284 | "text": [ 285 | "\n" 286 | ] 287 | }, 288 | { 289 | "name": "stderr", 290 | "output_type": "stream", 291 | "text": [ 292 | "INFO:sagemaker:Creating endpoint with name daal-knn-sm-2019-02-24-15-31-09-721\n" 293 | ] 294 | }, 295 | { 296 | "name": "stdout", 297 | "output_type": "stream", 298 | "text": [ 299 | "--------------------------------------------------------------------------!" 300 | ] 301 | } 302 | ], 303 | "source": [ 304 | "predictor = daal_knn.deploy(1, \"ml.m4.xlarge\", serializer=sagemaker.predictor.csv_serializer)" 305 | ] 306 | }, 307 | { 308 | "cell_type": "markdown", 309 | "metadata": {}, 310 | "source": [ 311 | "Secondly, you should pass data as numpy array to endpoint and get predictions.\n", 312 | "\n", 313 | "In this example we are passing random data, but you can use any numpy 2D array" 314 | ] 315 | }, 316 | { 317 | "cell_type": "code", 318 | "execution_count": 11, 319 | "metadata": {}, 320 | "outputs": [ 321 | { 322 | "name": "stdout", 323 | "output_type": "stream", 324 | "text": [ 325 | "1\n", 326 | "1\n", 327 | "2\n", 328 | "4\n", 329 | "4\n", 330 | "3\n", 331 | "3\n", 332 | "1\n", 333 | "2\n", 334 | "4\n", 335 | "\n" 336 | ] 337 | } 338 | ], 339 | "source": [ 340 | "import numpy as np\n", 341 | "\n", 342 | "predict_data = np.random.random(size=(10,5))\n", 343 | "print(predictor.predict(predict_data).decode(\"utf-8\"))" 344 | ] 345 | }, 346 | { 347 | "cell_type": "markdown", 348 | "metadata": {}, 349 | "source": [ 350 | "Don't forget to delete endpoint if you don't need it anymore." 351 | ] 352 | }, 353 | { 354 | "cell_type": "code", 355 | "execution_count": 12, 356 | "metadata": {}, 357 | "outputs": [ 358 | { 359 | "name": "stderr", 360 | "output_type": "stream", 361 | "text": [ 362 | "INFO:sagemaker:Deleting endpoint with name: daal-knn-sm-2019-02-24-15-31-09-721\n" 363 | ] 364 | } 365 | ], 366 | "source": [ 367 | "sess.delete_endpoint(predictor.endpoint)" 368 | ] 369 | }, 370 | { 371 | "cell_type": "markdown", 372 | "metadata": {}, 373 | "source": [ 374 | "### Batch transform job\n", 375 | "If you don't need real-time prediction, you can use transform job. It uses saved model, compute predictions one time and saves it in specified or auto-generated output path.\n", 376 | "\n", 377 | "More about transform jobs: https://docs.aws.amazon.com/sagemaker/latest/dg/how-it-works-batch.html\n", 378 | "\n", 379 | "Transformer API: https://sagemaker.readthedocs.io/en/latest/transformer.html" 380 | ] 381 | }, 382 | { 383 | "cell_type": "code", 384 | "execution_count": 14, 385 | "metadata": {}, 386 | "outputs": [ 387 | { 388 | "name": "stderr", 389 | "output_type": "stream", 390 | "text": [ 391 | "INFO:sagemaker:Creating model package with name: daal-knn-2019-02-24-15-53-00-053\n" 392 | ] 393 | }, 394 | { 395 | "name": "stdout", 396 | "output_type": "stream", 397 | "text": [ 398 | ".........." 399 | ] 400 | }, 401 | { 402 | "name": "stderr", 403 | "output_type": "stream", 404 | "text": [ 405 | "INFO:sagemaker:Creating model with name: daal-knn-2019-02-24-15-53-00-053-2019-02-24-15-53-45-504\n" 406 | ] 407 | }, 408 | { 409 | "name": "stdout", 410 | "output_type": "stream", 411 | "text": [ 412 | "\n" 413 | ] 414 | }, 415 | { 416 | "name": "stderr", 417 | "output_type": "stream", 418 | "text": [ 419 | "INFO:sagemaker:Creating transform job with name: daal-knn-sm-2019-02-24-15-53-45-783\n" 420 | ] 421 | }, 422 | { 423 | "name": "stdout", 424 | "output_type": "stream", 425 | "text": [ 426 | "........................................!\n", 427 | "s3://sagemaker-us-east-2-123123123123/daal-knn-sm-2019-02-24-15-53-45-783\n" 428 | ] 429 | } 430 | ], 431 | "source": [ 432 | "transformer = daal_knn.transformer(1, 'ml.m4.xlarge')\n", 433 | "transformer.transform(\"s3:///\", content_type='text/csv')\n", 434 | "transformer.wait()\n", 435 | "print(transformer.output_path)" 436 | ] 437 | } 438 | ], 439 | "metadata": { 440 | "kernelspec": { 441 | "display_name": "conda_python3", 442 | "language": "python", 443 | "name": "conda_python3" 444 | }, 445 | "language_info": { 446 | "codemirror_mode": { 447 | "name": "ipython", 448 | "version": 3 449 | }, 450 | "file_extension": ".py", 451 | "mimetype": "text/x-python", 452 | "name": "python", 453 | "nbconvert_exporter": "python", 454 | "pygments_lexer": "ipython3", 455 | "version": "3.6.5" 456 | } 457 | }, 458 | "nbformat": 4, 459 | "nbformat_minor": 2 460 | } 461 | -------------------------------------------------------------------------------- /daal/pca-example.ipynb: -------------------------------------------------------------------------------- 1 | { 2 | "cells": [ 3 | { 4 | "cell_type": "markdown", 5 | "metadata": {}, 6 | "source": [ 7 | "# Principal Component Analysis with Intel® Data Analytics Acceleration Library in Amazon SageMaker\n", 8 | "\n", 9 | "## Introduction\n", 10 | "\n", 11 | "Intel® Data Analytics Acceleration Library (Intel® DAAL) is the library of Intel® architecture optimized building blocks covering all stages of data analytics: data acquisition from a data source, preprocessing, transformation, data mining, modeling, validation, and decision making. One of its algorithms is PCA.\n", 12 | "\n", 13 | "Principal Component Analysis (PCA) is a method for exploratory data analysis. PCA transforms a set of observations of possibly correlated variables to a new set of uncorrelated variables, called principal components. Principal components are the directions of the largest variance, that is, the directions where the data is mostly spread out.\n", 14 | "\n", 15 | "Because all principal components are orthogonal to each other, there is no redundant information. This is a way of replacing a group of variables with a smaller set of new variables. PCA is one of powerful techniques for dimension reduction.\n", 16 | "\n", 17 | "Intel® DAAL developer guide: https://software.intel.com/en-us/daal-programming-guide\n", 18 | "\n", 19 | "Intel® DAAL documentation for PCA: https://software.intel.com/en-us/daal-programming-guide-principal-component-analysis" 20 | ] 21 | }, 22 | { 23 | "cell_type": "markdown", 24 | "metadata": {}, 25 | "source": [ 26 | "## PCA Usage with SageMaker Estimator\n", 27 | "Firstly, you need to import SageMaker package, get execution role and create session." 28 | ] 29 | }, 30 | { 31 | "cell_type": "code", 32 | "execution_count": 13, 33 | "metadata": {}, 34 | "outputs": [], 35 | "source": [ 36 | "import sagemaker\n", 37 | "\n", 38 | "role = sagemaker.get_execution_role()\n", 39 | "sess = sagemaker.Session()" 40 | ] 41 | }, 42 | { 43 | "cell_type": "markdown", 44 | "metadata": {}, 45 | "source": [ 46 | "Secondly, you can specify parameters of PCA.\n", 47 | "#### Hyperparameters\n", 48 | "All hyperparameters of PCA algorithm are optional.\n", 49 | "\n", 50 | " \n", 51 | " \n", 52 | " \n", 53 | " \n", 54 | " \n", 55 | " \n", 56 | " \n", 57 | " \n", 58 | " \n", 59 | " \n", 60 | " \n", 61 | " \n", 62 | " \n", 63 | " \n", 64 | " \n", 65 | " \n", 66 | " \n", 67 | " \n", 68 | " \n", 69 | " \n", 70 | " \n", 71 | " \n", 72 | " \n", 73 | " \n", 74 | " \n", 75 | " \n", 76 | " \n", 77 | " \n", 78 | " \n", 79 | " \n", 80 | " \n", 81 | " \n", 82 | " \n", 83 | " \n", 84 | " \n", 85 | " \n", 86 | " \n", 87 | " \n", 88 | " \n", 89 | " \n", 90 | " \n", 91 | " \n", 92 | "
Parameter nameTypeDefault valueDescription
fptypestr\"double\"The floating-point type that the algorithm uses for intermediate computations. Can be \"float\" or \"double\"
methodstr\"correlationDense\"Available methods for PCA computation: \"correlationDense\" (\"defaultDense\") or \"svdDense\"
resultsToComputestr\"none\"Provide one of the following values to request a single characteristic or use bitwise OR to request a combination of the characteristics: mean, variance, eigenvalue. For example, combination of all is \"mean|variance|eigenvalue\"
nComponentsint0Number of principal components.
If it is zero, the algorithm will compute the result for number of principal components = number of features.
Remember that number of components must be equal or less than number of features for PCA algorithm
isDeterministicboolFalseIf True, the algorithm applies the \"sign flip\" technique to the results
transformOnTrainboolFalseIf True, training data will be transformed and saved in model package on training stage
\n", 93 | "\n", 94 | "Example of hyperparameters dictionary:" 95 | ] 96 | }, 97 | { 98 | "cell_type": "code", 99 | "execution_count": 14, 100 | "metadata": {}, 101 | "outputs": [], 102 | "source": [ 103 | "pca_params = {\n", 104 | " \"fptype\": \"float\",\n", 105 | " \"method\": \"svdDense\",\n", 106 | " \"resultsToCompute\": \"mean|eigenvalue\",\n", 107 | " \"nComponents\": 4,\n", 108 | " \"isDeterministic\": True\n", 109 | "}" 110 | ] 111 | }, 112 | { 113 | "cell_type": "markdown", 114 | "metadata": {}, 115 | "source": [ 116 | "Then, you need to create SageMaker Estimator instance with following parameters:\n", 117 | "\n", 118 | " \n", 119 | " \n", 120 | " \n", 121 | " \n", 122 | " \n", 123 | " \n", 124 | " \n", 125 | " \n", 126 | " \n", 127 | " \n", 128 | " \n", 129 | " \n", 130 | " \n", 131 | " \n", 132 | " \n", 133 | " \n", 134 | " \n", 135 | " \n", 136 | " \n", 137 | " \n", 138 | " \n", 139 | " \n", 140 | " \n", 141 | " \n", 142 | " \n", 143 | " \n", 144 | " \n", 145 | " \n", 146 | " \n", 147 | " \n", 148 | " \n", 149 | " \n", 150 | " \n", 151 | " \n", 152 | " \n", 153 | " \n", 154 | "
Parameter nameDescription
image_nameThe container image to use for training
roleAn AWS IAM role. The SageMaker training jobs and APIs that create SageMaker endpoints use this role to access training data and models
train_instance_countNumber of Amazon EC2 instances to use for training. Should be 1, because it is not distributed version of algorithm
train_instance_typeType of EC2 instance to use for training. See available types on Amazon Marketplace page of algorithm
input_modeThe input mode that the algorithm supports. May be \"File\" or \"Pipe\"
output_pathS3 location for saving the trainig result (model artifacts and output files)
sagemaker_sessionSession object which manages interactions with Amazon SageMaker APIs and any other AWS services needed
hyperparametersDictionary containing the hyperparameters to initialize this estimator with
\n", 155 | "Full SageMaker Estimator documentation: https://sagemaker.readthedocs.io/en/latest/estimators.html" 156 | ] 157 | }, 158 | { 159 | "cell_type": "code", 160 | "execution_count": 9, 161 | "metadata": {}, 162 | "outputs": [], 163 | "source": [ 164 | "daal_pca_arn = \"\" # you can find it on algorithm page in your subscriptions\n", 165 | "\n", 166 | "daal_pca = sagemaker.algorithm.AlgorithmEstimator(\n", 167 | " algorithm_arn=daal_pca_arn,\n", 168 | " role=role,\n", 169 | " base_job_name=\"\",\n", 170 | " train_instance_count=1,\n", 171 | " train_instance_type='ml.m4.xlarge',\n", 172 | " input_mode=\"File\",\n", 173 | " output_path=\"s3:///\",\n", 174 | " sagemaker_session=sess,\n", 175 | " hyperparameters=pca_params\n", 176 | ")" 177 | ] 178 | }, 179 | { 180 | "cell_type": "markdown", 181 | "metadata": {}, 182 | "source": [ 183 | "### Training stage\n", 184 | "On training stage, PCA algorithm consume input data from S3 location and computes eigen vectors and other results (if they are specified in \"resultsToCompute\" parameter).\n", 185 | "This container supports only .csv (\"comma-separated values\") files." 186 | ] 187 | }, 188 | { 189 | "cell_type": "code", 190 | "execution_count": 10, 191 | "metadata": {}, 192 | "outputs": [ 193 | { 194 | "name": "stderr", 195 | "output_type": "stream", 196 | "text": [ 197 | "INFO:sagemaker:Creating training-job with name: daal-pca-alg-test-2018-11-30-06-50-06-484\n" 198 | ] 199 | }, 200 | { 201 | "name": "stdout", 202 | "output_type": "stream", 203 | "text": [ 204 | "2018-11-30 06:50:06 Starting - Starting the training job...\n", 205 | "2018-11-30 06:50:07 Starting - Launching requested ML instances......\n", 206 | "2018-11-30 06:51:15 Starting - Preparing the instances for training...\n", 207 | "2018-11-30 06:52:07 Downloading - Downloading input data...\n", 208 | "2018-11-30 06:52:30 Training - Downloading the training image...\n", 209 | "2018-11-30 06:53:01 Uploading - Uploading generated training model\n", 210 | "2018-11-30 06:53:01 Completed - Training job completed\n", 211 | "\n", 212 | "\u001b[31m2018-11-30 06:52:49 INFO Training stage started\u001b[0m\n", 213 | "\u001b[31m2018-11-30 06:52:49 INFO Default Paramaters:\u001b[0m\n", 214 | "\u001b[31m2018-11-30 06:52:49 INFO {'fptype': 'double', 'method': 'correlationDense', 'resultsToCompute': '', 'nComponents': '0', 'isDeterministic': 'False', 'transformOnTrain': 'False'}\u001b[0m\n", 215 | "\u001b[31m2018-11-30 06:52:49 INFO Updated with user hyperparameters, uncorrect parameters changed or deleted\u001b[0m\n", 216 | "\u001b[31m2018-11-30 06:52:49 INFO Final Hyperparameters:\u001b[0m\n", 217 | "\u001b[31m2018-11-30 06:52:49 INFO {'fptype': 'float', 'method': 'defaultDense', 'resultsToCompute': 'mean|eigenvalue', 'nComponents': 3, 'isDeterministic': True, 'transformOnTrain': False}\u001b[0m\n", 218 | "\u001b[31m2018-11-30 06:52:54 INFO Train data shape: (1600000, 10)\u001b[0m\n", 219 | "\u001b[31m2018-11-30 06:52:54 INFO Files loading time: 4.69575047492981\u001b[0m\n", 220 | "\u001b[31m2018-11-30 06:52:54 INFO Starting DAAL PCA algorithm...\u001b[0m\n", 221 | "\u001b[31m2018-11-30 06:52:54 INFO Training time, sec: 0.10286879539489746\u001b[0m\n", 222 | "\u001b[31m2018-11-30 06:52:54 INFO Means table saved at /opt/ml/model/mean.csv\u001b[0m\n", 223 | "\u001b[31m2018-11-30 06:52:54 INFO Eigenvalues table saved at /opt/ml/model/eigenvalue.csv\u001b[0m\n", 224 | "\u001b[31m2018-11-30 06:52:54 INFO Eigen vectors saved at /opt/ml/model/vectors.csv\u001b[0m\n", 225 | "\u001b[31m2018-11-30 06:52:54 INFO Parameters saved at /opt/ml/model/parameters.json\u001b[0m\n", 226 | "Billable seconds: 54\n" 227 | ] 228 | } 229 | ], 230 | "source": [ 231 | "daal_pca.fit({\"training\": \"s3:///\"})" 232 | ] 233 | }, 234 | { 235 | "cell_type": "markdown", 236 | "metadata": {}, 237 | "source": [ 238 | "### Real-time prediction\n", 239 | "On prediction stage, PCA algorithm transforms input data using previously computed results.\n", 240 | "Firstly, you need to deploy SageMaker endpoint that consumes data." 241 | ] 242 | }, 243 | { 244 | "cell_type": "code", 245 | "execution_count": 6, 246 | "metadata": {}, 247 | "outputs": [ 248 | { 249 | "name": "stderr", 250 | "output_type": "stream", 251 | "text": [ 252 | "INFO:sagemaker:Creating model package with name: intel-daal-pca1542385402-d0d25e75ca6ef4-2018-11-29-15-37-47-871\n" 253 | ] 254 | }, 255 | { 256 | "name": "stdout", 257 | "output_type": "stream", 258 | "text": [ 259 | ".........." 260 | ] 261 | }, 262 | { 263 | "name": "stderr", 264 | "output_type": "stream", 265 | "text": [ 266 | "INFO:sagemaker:Creating model with name: intel-daal-pca1542385402-d0d25e75ca6ef4-2018-11-29-15-38-33-468\n" 267 | ] 268 | }, 269 | { 270 | "name": "stdout", 271 | "output_type": "stream", 272 | "text": [ 273 | "\n" 274 | ] 275 | }, 276 | { 277 | "name": "stderr", 278 | "output_type": "stream", 279 | "text": [ 280 | "INFO:sagemaker:Creating endpoint with name intel-daal-pca1542385402-d0d25e75ca6ef4-2018-11-29-15-31-41-921\n" 281 | ] 282 | }, 283 | { 284 | "name": "stdout", 285 | "output_type": "stream", 286 | "text": [ 287 | "-------------------------------------------------------------!" 288 | ] 289 | } 290 | ], 291 | "source": [ 292 | "predictor = daal_pca.deploy(1, \"ml.m4.xlarge\", serializer=sagemaker.predictor.csv_serializer)" 293 | ] 294 | }, 295 | { 296 | "cell_type": "markdown", 297 | "metadata": {}, 298 | "source": [ 299 | "Secondly, you should pass data as numpy array to predictor instance and get transformed data as space-separated values.\n", 300 | "\n", 301 | "In this example we are passing random data, but you can use any numpy 2D array with one specific condition for PCA: training data and data to transform must have same numbers of features." 302 | ] 303 | }, 304 | { 305 | "cell_type": "code", 306 | "execution_count": 7, 307 | "metadata": {}, 308 | "outputs": [ 309 | { 310 | "name": "stdout", 311 | "output_type": "stream", 312 | "text": [ 313 | "1.185592651367187500e+00 2.620933353900909424e-01 3.085311949253082275e-01\n", 314 | "6.714667081832885742e-01 8.762556910514831543e-01 -8.037568628787994385e-02\n", 315 | "1.111641526222229004e+00 3.375906124711036682e-02 -1.244278624653816223e-01\n", 316 | "1.294844508171081543e+00 3.067855909466743469e-02 -9.089314937591552734e-02\n", 317 | "8.781186342239379883e-01 2.732140570878982544e-02 -5.232766270637512207e-01\n", 318 | "5.412490963935852051e-01 3.470270335674285889e-01 -8.399704098701477051e-02\n", 319 | "1.048457860946655273e+00 -1.020107120275497437e-01 6.852779537439346313e-02\n", 320 | "8.287011384963989258e-01 2.592234015464782715e-01 -1.914716511964797974e-01\n", 321 | "1.426655769348144531e+00 1.917291432619094849e-02 -1.039648428559303284e-01\n", 322 | "1.183746933937072754e+00 -8.203709125518798828e-02 3.905816972255706787e-01\n", 323 | "\n" 324 | ] 325 | } 326 | ], 327 | "source": [ 328 | "import numpy as np\n", 329 | "\n", 330 | "predict_data = np.random.random(size=(10,10))\n", 331 | "print(predictor.predict(predict_data).decode(\"utf-8\"))" 332 | ] 333 | }, 334 | { 335 | "cell_type": "markdown", 336 | "metadata": {}, 337 | "source": [ 338 | "Don't forget to delete endpoint if you don't need it anymore." 339 | ] 340 | }, 341 | { 342 | "cell_type": "code", 343 | "execution_count": 8, 344 | "metadata": {}, 345 | "outputs": [ 346 | { 347 | "name": "stderr", 348 | "output_type": "stream", 349 | "text": [ 350 | "INFO:sagemaker:Deleting endpoint with name: intel-daal-pca1542385402-d0d25e75ca6ef4-2018-11-29-15-31-41-921\n" 351 | ] 352 | } 353 | ], 354 | "source": [ 355 | "sess.delete_endpoint(predictor.endpoint)" 356 | ] 357 | }, 358 | { 359 | "cell_type": "markdown", 360 | "metadata": {}, 361 | "source": [ 362 | "### Batch transform job\n", 363 | "If you don't need real-time prediction, you can use transform job. It uses saved model, compute transformed data one time and saves it in specified or auto-generated output path.\n", 364 | "\n", 365 | "More about transform jobs: https://docs.aws.amazon.com/sagemaker/latest/dg/how-it-works-batch.html\n", 366 | "\n", 367 | "Transformer API: https://sagemaker.readthedocs.io/en/latest/transformer.html" 368 | ] 369 | }, 370 | { 371 | "cell_type": "code", 372 | "execution_count": 12, 373 | "metadata": {}, 374 | "outputs": [ 375 | { 376 | "name": "stderr", 377 | "output_type": "stream", 378 | "text": [ 379 | "INFO:sagemaker:Creating model package with name: intel-daal-pca1542385402-d0d25e75ca6ef4-2018-11-30-07-42-48-772\n" 380 | ] 381 | }, 382 | { 383 | "name": "stdout", 384 | "output_type": "stream", 385 | "text": [ 386 | ".........." 387 | ] 388 | }, 389 | { 390 | "name": "stderr", 391 | "output_type": "stream", 392 | "text": [ 393 | "INFO:sagemaker:Creating model with name: intel-daal-pca1542385402-d0d25e75ca6ef4-2018-11-30-07-43-34-517\n" 394 | ] 395 | }, 396 | { 397 | "name": "stdout", 398 | "output_type": "stream", 399 | "text": [ 400 | "\n" 401 | ] 402 | }, 403 | { 404 | "name": "stderr", 405 | "output_type": "stream", 406 | "text": [ 407 | "INFO:sagemaker:Creating transform job with name: daal-pca-alg-test-2018-11-30-07-43-35-199\n" 408 | ] 409 | }, 410 | { 411 | "name": "stdout", 412 | "output_type": "stream", 413 | "text": [ 414 | "......................................!\n", 415 | "s3://\n" 416 | ] 417 | } 418 | ], 419 | "source": [ 420 | "transformer = daal_pca.transformer(1, 'ml.m4.xlarge')\n", 421 | "transformer.transform(\"s3:///\", content_type='text/csv')\n", 422 | "transformer.wait()\n", 423 | "print(transformer.output_path)" 424 | ] 425 | } 426 | ], 427 | "metadata": { 428 | "kernelspec": { 429 | "display_name": "Python 3", 430 | "language": "python", 431 | "name": "python3" 432 | }, 433 | "language_info": { 434 | "codemirror_mode": { 435 | "name": "ipython", 436 | "version": 3 437 | }, 438 | "file_extension": ".py", 439 | "mimetype": "text/x-python", 440 | "name": "python", 441 | "nbconvert_exporter": "python", 442 | "pygments_lexer": "ipython3", 443 | "version": "3.6.5" 444 | } 445 | }, 446 | "nbformat": 4, 447 | "nbformat_minor": 2 448 | } 449 | -------------------------------------------------------------------------------- /daal/kmeans-example.ipynb: -------------------------------------------------------------------------------- 1 | { 2 | "cells": [ 3 | { 4 | "cell_type": "markdown", 5 | "metadata": {}, 6 | "source": [ 7 | "# K-Means with Intel® Data Analytics Acceleration Library in Amazon SageMaker\n", 8 | "\n", 9 | "## Introduction\n", 10 | "\n", 11 | "Intel® Data Analytics Acceleration Library (Intel® DAAL) is the library of Intel® architecture optimized building blocks covering all stages of data analytics: data acquisition from a data source, preprocessing, transformation, data mining, modeling, validation, and decision making. One of its algorithms is K-Means.\n", 12 | "\n", 13 | "K-Means is among the most popular and simplest clustering methods. It is intended to partition a data set into a small number of clusters such that feature vectors within a cluster have greater similarity with one another than with feature vectors from other clusters. Each cluster is characterized by a representative point, called a centroid, and a cluster radius.\n", 14 | "\n", 15 | "In other words, the clustering methods enable reducing the problem of analysis of the entire data set to the analysis of clusters.\n", 16 | "\n", 17 | "There are numerous ways to define the measure of similarity and centroids. For K-Means, the centroid is defined as the mean of feature vectors within the cluster.\n", 18 | "\n", 19 | "Intel® DAAL developer guide: https://software.intel.com/en-us/daal-programming-guide\n", 20 | "\n", 21 | "Intel® DAAL documentation for K-Means: https://software.intel.com/en-us/daal-programming-guide-k-means-clustering " 22 | ] 23 | }, 24 | { 25 | "cell_type": "markdown", 26 | "metadata": {}, 27 | "source": [ 28 | "## K-Means Usage with SageMaker Estimator\n", 29 | "Firstly, you need to import SageMaker package, get execution role and create session." 30 | ] 31 | }, 32 | { 33 | "cell_type": "code", 34 | "execution_count": 1, 35 | "metadata": {}, 36 | "outputs": [], 37 | "source": [ 38 | "import sagemaker\n", 39 | "\n", 40 | "role = sagemaker.get_execution_role()\n", 41 | "sess = sagemaker.Session()" 42 | ] 43 | }, 44 | { 45 | "cell_type": "markdown", 46 | "metadata": {}, 47 | "source": [ 48 | "Secondly, you can specify parameters of K-Means.\n", 49 | "#### Hyperparameters\n", 50 | "\"nClusters\" and \"maxIterations\" hyperparameters of K-Means algorithm are required, all other - optional.\n", 51 | "\n", 52 | " \n", 53 | " \n", 54 | " \n", 55 | " \n", 56 | " \n", 57 | " \n", 58 | " \n", 59 | " \n", 60 | " \n", 61 | " \n", 62 | " \n", 63 | " \n", 64 | " \n", 65 | " \n", 66 | " \n", 67 | " \n", 68 | " \n", 69 | " \n", 70 | " \n", 71 | " \n", 72 | " \n", 73 | " \n", 74 | " \n", 75 | " \n", 76 | " \n", 77 | " \n", 78 | " \n", 79 | " \n", 80 | " \n", 81 | " \n", 82 | " \n", 83 | " \n", 84 | " \n", 85 | " \n", 86 | " \n", 87 | " \n", 88 | " \n", 89 | " \n", 90 | " \n", 91 | " \n", 92 | " \n", 93 | " \n", 94 | " \n", 95 | " \n", 96 | " \n", 97 | " \n", 98 | " \n", 99 | " \n", 100 | " \n", 101 | " \n", 102 | " \n", 103 | " \n", 104 | " \n", 105 | " \n", 106 | " \n", 107 | " \n", 108 | " \n", 109 | " \n", 110 | " \n", 111 | " \n", 112 | " \n", 113 | " \n", 114 | " \n", 115 | " \n", 116 | " \n", 117 | " \n", 118 | " \n", 119 | " \n", 120 | " \n", 121 | " \n", 122 | " \n", 123 | " \n", 124 | " \n", 125 | " \n", 126 | " \n", 127 | " \n", 128 | " \n", 129 | " \n", 130 | "
Parameter nameTypeDefault valueDescription
fptypestr\"double\"The floating-point type that the algorithm uses for intermediate computations. Can be \"float\" or \"double\"
nClustersint2The number of clusters
initMethodstr\"defaultDense\"Available initialization methods for K-Means clustering: defaultDense - uses first nClusters points as initial clusters, randomDense - uses random nClusters points as initial clusters, plusPlusDense - uses K-Means++ algorithm; parallelPlusDense - uses parallel K-Means++ algorithm
oversamplingFactorfloat0.5A fraction of nClusters in each of nRounds of parallel K-Means++. L=nClusters*oversamplingFactor points are sampled in a round
nRoundsint5The number of rounds for parallel K-Means++. (L*nRounds) must be greater than nClusters
seedint777The seed for random number generator
methodstr\"lloydDense\"Computation method for K-Means clustering
maxIterationsint100The number of iterations
accuracyThresholdfloat0The threshold for termination of the algorithm
gammafloat1The weight to be used in distance calculation for binary categorical features
distanceTypestr\"euclidean\"The measure of closeness between points (observations) being clustered. The only distance type supported so far is the Euclidian distance
assignFlagboolTrueA flag that enables computation of assignments, that is, assigning cluster indices to respective observations
\n", 131 | "\n", 132 | "Example of hyperparameters dictionary:" 133 | ] 134 | }, 135 | { 136 | "cell_type": "code", 137 | "execution_count": 3, 138 | "metadata": {}, 139 | "outputs": [], 140 | "source": [ 141 | "kmeans_params = {\n", 142 | " \"fptype\": \"float\",\n", 143 | " \"nClusters\": 5,\n", 144 | " \"initMethod\": \"plusPlusDense\",\n", 145 | " \"maxIterations\": 1000\n", 146 | "}" 147 | ] 148 | }, 149 | { 150 | "cell_type": "markdown", 151 | "metadata": {}, 152 | "source": [ 153 | "Then, you need to create SageMaker Estimator instance with following parameters:\n", 154 | "\n", 155 | " \n", 156 | " \n", 157 | " \n", 158 | " \n", 159 | " \n", 160 | " \n", 161 | " \n", 162 | " \n", 163 | " \n", 164 | " \n", 165 | " \n", 166 | " \n", 167 | " \n", 168 | " \n", 169 | " \n", 170 | " \n", 171 | " \n", 172 | " \n", 173 | " \n", 174 | " \n", 175 | " \n", 176 | " \n", 177 | " \n", 178 | " \n", 179 | " \n", 180 | " \n", 181 | " \n", 182 | " \n", 183 | " \n", 184 | " \n", 185 | " \n", 186 | " \n", 187 | " \n", 188 | " \n", 189 | " \n", 190 | " \n", 191 | "
Parameter nameDescription
image_nameThe container image to use for training
roleAn AWS IAM role. The SageMaker training jobs and APIs that create SageMaker endpoints use this role to access training data and models
train_instance_countNumber of Amazon EC2 instances to use for training. Should be 1, because it is not distributed version of algorithm
train_instance_typeType of EC2 instance to use for training. See available types on Amazon Marketplace page of algorithm
input_modeThe input mode that the algorithm supports. May be \"File\" or \"Pipe\"
output_pathS3 location for saving the trainig result (model artifacts and output files)
sagemaker_sessionSession object which manages interactions with Amazon SageMaker APIs and any other AWS services needed
hyperparametersDictionary containing the hyperparameters to initialize this estimator with
\n", 192 | "Full SageMaker Estimator documentation: https://sagemaker.readthedocs.io/en/latest/estimators.html" 193 | ] 194 | }, 195 | { 196 | "cell_type": "code", 197 | "execution_count": 4, 198 | "metadata": {}, 199 | "outputs": [], 200 | "source": [ 201 | "daal_kmeans_arn = \"\" # you can find it on algorithm page in your subscriptions\n", 202 | "\n", 203 | "daal_kmeans = sagemaker.algorithm.AlgorithmEstimator(\n", 204 | " algorithm_arn=daal_kmeans_arn,\n", 205 | " role=role,\n", 206 | " base_job_name=\"\",\n", 207 | " train_instance_count=1,\n", 208 | " train_instance_type='ml.m4.xlarge',\n", 209 | " input_mode=\"File\",\n", 210 | " output_path=\"s3:///\",\n", 211 | " sagemaker_session=sess,\n", 212 | " hyperparameters=kmeans_params\n", 213 | ")" 214 | ] 215 | }, 216 | { 217 | "cell_type": "markdown", 218 | "metadata": {}, 219 | "source": [ 220 | "### Training stage\n", 221 | "On training stage, K-Means algorithm consume input data from S3 location and computes centroids.\n", 222 | "This container supports only .csv (\"comma-separated values\") files." 223 | ] 224 | }, 225 | { 226 | "cell_type": "code", 227 | "execution_count": 5, 228 | "metadata": {}, 229 | "outputs": [ 230 | { 231 | "name": "stderr", 232 | "output_type": "stream", 233 | "text": [ 234 | "INFO:sagemaker:Creating training-job with name: daal-kmeans-test-2019-02-15-15-47-15-619\n" 235 | ] 236 | }, 237 | { 238 | "name": "stdout", 239 | "output_type": "stream", 240 | "text": [ 241 | "2019-02-15 15:47:15 Starting - Starting the training job...\n", 242 | "2019-02-15 15:47:17 Starting - Launching requested ML instances......\n", 243 | "2019-02-15 15:48:17 Starting - Preparing the instances for training...\n", 244 | "2019-02-15 15:49:09 Downloading - Downloading input data\n", 245 | "2019-02-15 15:49:09 Training - Downloading the training image......\n", 246 | "2019-02-15 15:50:14 Uploading - Uploading generated training model\n", 247 | "2019-02-15 15:50:14 Completed - Training job completed\n", 248 | "\n", 249 | "\u001b[31m2019-02-15 15:50:04 INFO Training stage started\u001b[0m\n", 250 | "\u001b[31m2019-02-15 15:50:04 INFO Final Hyperparameters:\u001b[0m\n", 251 | "\u001b[31m2019-02-15 15:50:04 INFO {'fptype': 'float', 'initMethod': 'plusPlusDense', 'seed': '777', 'oversamplingFactor': '0.5', 'nRounds': '5', 'kmeansMethod': 'lloydDense', 'accuracyThreshold': '0', 'gamma': '1', 'distanceType': 'euclidean', 'assignFlag': True, 'maxIterations': '1000', 'method': 'lloydDense', 'assignFlag ': 'True', 'nClusters': '5'}\u001b[0m\n", 252 | "\u001b[31m2019-02-15 15:50:04 INFO Train data shape: (30000, 10)\u001b[0m\n", 253 | "\u001b[31m2019-02-15 15:50:04 INFO Files loading time: 0.10956525802612305\u001b[0m\n", 254 | "\u001b[31m2019-02-15 15:50:04 INFO Starting DAAL k-Means algorithm...\u001b[0m\n", 255 | "\u001b[31m2019-02-15 15:50:04 INFO Training time, sec: 0.34311580657958984\u001b[0m\n", 256 | "\u001b[31m2019-02-15 15:50:04 INFO Parameters saved at /opt/ml/model/parameters.json\u001b[0m\n", 257 | "\u001b[31m2019-02-15 15:50:04 INFO Centroids saved at /opt/ml/model/centroids.csv\u001b[0m\n", 258 | "\u001b[31m2019-02-15 15:50:05 INFO Assignments saved at /opt/ml/model/assignments.csv\u001b[0m\n", 259 | "\u001b[31m2019-02-15 15:50:05 INFO Number of computed iterations: 1000\u001b[0m\n", 260 | "\u001b[31m2019-02-15 15:50:05 INFO Objective function value: 286906.5\u001b[0m\n", 261 | "\u001b[31m2019-02-15 15:50:05 INFO Number of computed iterations and objective function value saved at /opt/ml/model/other.txt\u001b[0m\n", 262 | "Billable seconds: 75\n" 263 | ] 264 | } 265 | ], 266 | "source": [ 267 | "daal_kmeans.fit({\"training\": \"s3:///\"})" 268 | ] 269 | }, 270 | { 271 | "cell_type": "markdown", 272 | "metadata": {}, 273 | "source": [ 274 | "### Real-time prediction\n", 275 | "On prediction stage, K-Means algorithm determines assignments for input data using previously computed centroids.\n", 276 | "Firstly, you need to deploy SageMaker endpoint that consumes data." 277 | ] 278 | }, 279 | { 280 | "cell_type": "code", 281 | "execution_count": 6, 282 | "metadata": {}, 283 | "outputs": [ 284 | { 285 | "name": "stderr", 286 | "output_type": "stream", 287 | "text": [ 288 | "INFO:sagemaker:Creating model package with name: daal-kmeans-new-2019-02-15-15-56-26-951\n" 289 | ] 290 | }, 291 | { 292 | "name": "stdout", 293 | "output_type": "stream", 294 | "text": [ 295 | ".........." 296 | ] 297 | }, 298 | { 299 | "name": "stderr", 300 | "output_type": "stream", 301 | "text": [ 302 | "INFO:sagemaker:Creating model with name: daal-kmeans-new-2019-02-15-15-56-26-951-2019-02-15-15-57-12-443\n" 303 | ] 304 | }, 305 | { 306 | "name": "stdout", 307 | "output_type": "stream", 308 | "text": [ 309 | "\n" 310 | ] 311 | }, 312 | { 313 | "name": "stderr", 314 | "output_type": "stream", 315 | "text": [ 316 | "INFO:sagemaker:Creating endpoint with name daal-kmeans-test-2019-02-15-15-47-15-619\n" 317 | ] 318 | }, 319 | { 320 | "name": "stdout", 321 | "output_type": "stream", 322 | "text": [ 323 | "------------------------------------------------------------------!" 324 | ] 325 | } 326 | ], 327 | "source": [ 328 | "predictor = daal_kmeans.deploy(1, \"ml.m4.xlarge\", serializer=sagemaker.predictor.csv_serializer)" 329 | ] 330 | }, 331 | { 332 | "cell_type": "markdown", 333 | "metadata": {}, 334 | "source": [ 335 | "Secondly, you should pass data as numpy array to predictor instance and get assignments.\n", 336 | "\n", 337 | "In this example we are passing random data, but you can use any numpy 2D array." 338 | ] 339 | }, 340 | { 341 | "cell_type": "code", 342 | "execution_count": 8, 343 | "metadata": { 344 | "scrolled": true 345 | }, 346 | "outputs": [ 347 | { 348 | "name": "stdout", 349 | "output_type": "stream", 350 | "text": [ 351 | "1\n", 352 | "0\n", 353 | "2\n", 354 | "3\n", 355 | "3\n", 356 | "4\n", 357 | "1\n", 358 | "2\n", 359 | "3\n", 360 | "0\n" 361 | ] 362 | } 363 | ], 364 | "source": [ 365 | "import numpy as np\n", 366 | "\n", 367 | "predict_data = np.random.random(size=(10,10))\n", 368 | "print(predictor.predict(predict_data).decode(\"utf-8\"))" 369 | ] 370 | }, 371 | { 372 | "cell_type": "markdown", 373 | "metadata": {}, 374 | "source": [ 375 | "Don't forget to delete endpoint if you don't need it anymore." 376 | ] 377 | }, 378 | { 379 | "cell_type": "code", 380 | "execution_count": 9, 381 | "metadata": {}, 382 | "outputs": [ 383 | { 384 | "name": "stderr", 385 | "output_type": "stream", 386 | "text": [ 387 | "INFO:sagemaker:Deleting endpoint with name: daal-kmeans-test-2019-02-15-15-47-15-619\n" 388 | ] 389 | } 390 | ], 391 | "source": [ 392 | "sess.delete_endpoint(predictor.endpoint)" 393 | ] 394 | }, 395 | { 396 | "cell_type": "markdown", 397 | "metadata": {}, 398 | "source": [ 399 | "### Batch transform job\n", 400 | "If you don't need real-time prediction, you can use transform job. It uses saved model with centroids, compute assignments one time and saves it in specified or auto-generated output path.\n", 401 | "\n", 402 | "More about transform jobs: https://docs.aws.amazon.com/sagemaker/latest/dg/how-it-works-batch.html\n", 403 | "\n", 404 | "Transformer API: https://sagemaker.readthedocs.io/en/latest/transformer.html" 405 | ] 406 | }, 407 | { 408 | "cell_type": "code", 409 | "execution_count": 10, 410 | "metadata": {}, 411 | "outputs": [ 412 | { 413 | "name": "stderr", 414 | "output_type": "stream", 415 | "text": [ 416 | "INFO:sagemaker:Creating model package with name: daal-kmeans-new-2019-02-15-16-06-34-857\n" 417 | ] 418 | }, 419 | { 420 | "name": "stdout", 421 | "output_type": "stream", 422 | "text": [ 423 | ".........." 424 | ] 425 | }, 426 | { 427 | "name": "stderr", 428 | "output_type": "stream", 429 | "text": [ 430 | "INFO:sagemaker:Creating model with name: daal-kmeans-new-2019-02-15-16-06-34-857-2019-02-15-16-07-20-327\n" 431 | ] 432 | }, 433 | { 434 | "name": "stdout", 435 | "output_type": "stream", 436 | "text": [ 437 | "\n" 438 | ] 439 | }, 440 | { 441 | "name": "stderr", 442 | "output_type": "stream", 443 | "text": [ 444 | "INFO:sagemaker:Creating transform job with name: daal-kmeans-test-2019-02-15-16-07-20-877\n" 445 | ] 446 | }, 447 | { 448 | "name": "stdout", 449 | "output_type": "stream", 450 | "text": [ 451 | "........................................!\n", 452 | "s3://sagemaker-us-east-2-123123123123/daal-kmeans-test-2019-02-15-16-07-20-877\n" 453 | ] 454 | } 455 | ], 456 | "source": [ 457 | "transformer = daal_kmeans.transformer(1, 'ml.m4.xlarge')\n", 458 | "transformer.transform(\"s3:///\", content_type='text/csv')\n", 459 | "transformer.wait()\n", 460 | "print(transformer.output_path)" 461 | ] 462 | }, 463 | { 464 | "cell_type": "code", 465 | "execution_count": null, 466 | "metadata": {}, 467 | "outputs": [], 468 | "source": [] 469 | } 470 | ], 471 | "metadata": { 472 | "kernelspec": { 473 | "display_name": "Python 3", 474 | "language": "python", 475 | "name": "python3" 476 | }, 477 | "language_info": { 478 | "codemirror_mode": { 479 | "name": "ipython", 480 | "version": 3 481 | }, 482 | "file_extension": ".py", 483 | "mimetype": "text/x-python", 484 | "name": "python", 485 | "nbconvert_exporter": "python", 486 | "pygments_lexer": "ipython3", 487 | "version": "3.6.5" 488 | } 489 | }, 490 | "nbformat": 4, 491 | "nbformat_minor": 2 492 | } 493 | -------------------------------------------------------------------------------- /daal/decision-forest-example.ipynb: -------------------------------------------------------------------------------- 1 | { 2 | "cells": [ 3 | { 4 | "cell_type": "markdown", 5 | "metadata": {}, 6 | "source": [ 7 | "# Decision Forest Classification and Regression with Intel® Data Analytics Acceleration Library in Amazon SageMaker\n", 8 | "\n", 9 | "## Introduction\n", 10 | "\n", 11 | "Intel® Data Analytics Acceleration Library (Intel® DAAL) is the library of Intel® architecture optimized building blocks covering all stages of data analytics: data acquisition from a data source, preprocessing, transformation, data mining, modeling, validation, and decision making. One of its algorithms is Decision Forest.\n", 12 | "\n", 13 | "The library provides decision forest classification and regression algorithms based on an ensemble of tree-structured classifiers (decision trees) built using the general technique of bootstrap aggregation (bagging) and random choice of features. Decision tree is a binary tree graph. Its internal (split) nodes represent a decision function used to select the following (child) node at the prediction stage. Its leaf (terminal) nodes represent the corresponding response values, which are the result of the prediction from the tree.\n", 14 | "\n", 15 | "Intel® DAAL developer guide: https://software.intel.com/en-us/daal-programming-guide\n", 16 | "\n", 17 | "Intel® DAAL documentation for Decision Forest: https://software.intel.com/en-us/daal-programming-guide-decision-forest" 18 | ] 19 | }, 20 | { 21 | "cell_type": "markdown", 22 | "metadata": {}, 23 | "source": [ 24 | "## Decision Forest Usage with SageMaker Estimator\n", 25 | "Firstly, you need to import SageMaker package, get execution role and create session." 26 | ] 27 | }, 28 | { 29 | "cell_type": "code", 30 | "execution_count": 1, 31 | "metadata": {}, 32 | "outputs": [], 33 | "source": [ 34 | "import sagemaker\n", 35 | "\n", 36 | "role = sagemaker.get_execution_role()\n", 37 | "sess = sagemaker.Session()" 38 | ] 39 | }, 40 | { 41 | "cell_type": "markdown", 42 | "metadata": {}, 43 | "source": [ 44 | "Secondly, you can specify parameters of Decision Forest.\n", 45 | "#### Hyperparameters\n", 46 | "\n", 47 | " \n", 48 | " \n", 49 | " \n", 50 | " \n", 51 | " \n", 52 | " \n", 53 | " \n", 54 | " \n", 55 | " \n", 56 | " \n", 57 | " \n", 58 | " \n", 59 | " \n", 60 | " \n", 61 | " \n", 62 | " \n", 63 | " \n", 64 | " \n", 65 | " \n", 66 | " \n", 67 | " \n", 68 | " \n", 69 | " \n", 70 | " \n", 71 | " \n", 72 | " \n", 73 | " \n", 74 | " \n", 75 | " \n", 76 | " \n", 77 | " \n", 78 | " \n", 79 | " \n", 80 | " \n", 81 | " \n", 82 | " \n", 83 | " \n", 84 | " \n", 85 | " \n", 86 | " \n", 87 | " \n", 88 | " \n", 89 | " \n", 90 | " \n", 91 | " \n", 92 | " \n", 93 | " \n", 94 | " \n", 95 | " \n", 96 | " \n", 97 | " \n", 98 | " \n", 99 | " \n", 100 | " \n", 101 | " \n", 102 | " \n", 103 | " \n", 104 | " \n", 105 | " \n", 106 | " \n", 107 | " \n", 108 | " \n", 109 | " \n", 110 | " \n", 111 | " \n", 112 | " \n", 113 | " \n", 114 | " \n", 115 | " \n", 116 | " \n", 117 | " \n", 118 | " \n", 119 | " \n", 120 | " \n", 121 | " \n", 122 | " \n", 123 | " \n", 124 | " \n", 125 | " \n", 126 | " \n", 127 | " \n", 128 | " \n", 129 | " \n", 130 | " \n", 131 | " \n", 132 | " \n", 133 | " \n", 134 | " \n", 135 | " \n", 136 | " \n", 137 | "
Parameter nameTypeDefault valueDescription
nClassesint2Number of classes in data (only for classification)
fptypestr\"double\"The floating-point type that the algorithm uses for intermediate computations. Can be \"float\" or \"double\"
methodstr\"defaultDense\"The only training method supported so far is the default dense method
nTreesint100The number of trees in the forest
observationsPerTreeFractionint1Fraction of the training set S used to form the bootstrap set for a single tree training, observationsPerTreeFraction in (0, 1]. The observations are sampled randomly with replacement
featuresPerNodeint0The number of features tried as possible splits per node. If the parameter is set to 0, the library uses the square root of the number of features for classification and (the number of features)/3 for regression
maxTreeDepthint0Maximal tree depth. Default is 0 (unlimited).
minObservationsInLeafNodeint1 for classification, 5 for regressionThe number of neighbors
seedint777The seed for random number generator, which is used to choose the bootstrap set, split features in every split node in a tree, and generate permutation required in computations of MDA variable importance
impurityThresholdfloat0The threshold value used as stopping criteria: if the impurity value in the node is smaller than the threshold, the node is not split anymore
varImportancestr\"None\"The variable importance computation mode. Possible values:
none – variable importance is not calculated
MDI - Mean Decrease of Impurity, also known as the Gini importance or Mean Decrease Gini
MDA_Raw - Mean Decrease of Accuracy (permutation importance)
MDA_Scaled - the MDA_Raw value scaled by its standard deviation
resultsToComputestr\"None\"Provide one of the following values to request a single characteristic or use bitwise OR to request a combination of the characteristics:
computeOutOfBagError, computeOutOfBagErrorPerObservation
memorySavingModeboolFalseIf True, memory saving mode is enabled
bootstrapboolFalse for classification, True for regressionIf True, bootstrap is enabled
\n", 138 | "\n", 139 | "Example of hyperparameters dictionary:" 140 | ] 141 | }, 142 | { 143 | "cell_type": "code", 144 | "execution_count": 2, 145 | "metadata": {}, 146 | "outputs": [], 147 | "source": [ 148 | "decision_forest_params = {\n", 149 | " \"nClasses\": 5,\n", 150 | " \"fptype\":\"double\",\n", 151 | " \"method\":\"defaultDense\",\n", 152 | " \"nTrees\":\"100\",\n", 153 | " \"observationsPerTreeFraction\":\"1\",\n", 154 | " \"featuresPerNode\":\"0\",\n", 155 | " \"maxTreeDepth\":\"0\",\n", 156 | " \"minObservationsInLeafNode\":\"1\",\n", 157 | " \"seed\":\"777\",\n", 158 | " \"impurityThreshold\":\"0\",\n", 159 | " \"varImportance\":\"None\",\n", 160 | " \"resultsToCompute\":\"None\",\n", 161 | " \"memorySavingMode\":\"False\",\n", 162 | " \"bootstrap\":\"False\"\n", 163 | "}" 164 | ] 165 | }, 166 | { 167 | "cell_type": "markdown", 168 | "metadata": {}, 169 | "source": [ 170 | "Then, you need to create SageMaker Estimator instance with following parameters:\n", 171 | "\n", 172 | " \n", 173 | " \n", 174 | " \n", 175 | " \n", 176 | " \n", 177 | " \n", 178 | " \n", 179 | " \n", 180 | " \n", 181 | " \n", 182 | " \n", 183 | " \n", 184 | " \n", 185 | " \n", 186 | " \n", 187 | " \n", 188 | " \n", 189 | " \n", 190 | " \n", 191 | " \n", 192 | " \n", 193 | " \n", 194 | " \n", 195 | " \n", 196 | " \n", 197 | " \n", 198 | " \n", 199 | " \n", 200 | " \n", 201 | " \n", 202 | " \n", 203 | " \n", 204 | " \n", 205 | " \n", 206 | " \n", 207 | " \n", 208 | "
Parameter nameDescription
image_nameThe container image to use for training
roleAn AWS IAM role. The SageMaker training jobs and APIs that create SageMaker endpoints use this role to access training data and models
train_instance_countNumber of Amazon EC2 instances to use for training. Should be 1, because it is not distributed version of algorithm
train_instance_typeType of EC2 instance to use for training. See available types on Amazon Marketplace page of algorithm
input_modeThe input mode that the algorithm supports. May be \"File\" or \"Pipe\"
output_pathS3 location for saving the trainig result (model artifacts and output files)
sagemaker_sessionSession object which manages interactions with Amazon SageMaker APIs and any other AWS services needed
hyperparametersDictionary containing the hyperparameters to initialize this estimator with
\n", 209 | "Full SageMaker Estimator documentation: https://sagemaker.readthedocs.io/en/latest/estimators.html" 210 | ] 211 | }, 212 | { 213 | "cell_type": "code", 214 | "execution_count": 3, 215 | "metadata": {}, 216 | "outputs": [], 217 | "source": [ 218 | "daal_decision_forest_arn = \"\" # you can find it on algorithm page in your subscriptions\n", 219 | "\n", 220 | "daal_decision_forest = sagemaker.algorithm.AlgorithmEstimator(\n", 221 | " algorithm_arn=daal_decision_forest_arn,\n", 222 | " role=role,\n", 223 | " base_job_name=\"\",\n", 224 | " train_instance_count=1,\n", 225 | " train_instance_type='ml.m4.xlarge',\n", 226 | " input_mode=\"File\",\n", 227 | " output_path=\"s3:///\",\n", 228 | " sagemaker_session=sess,\n", 229 | " hyperparameters=decision_forest_params\n", 230 | ")" 231 | ] 232 | }, 233 | { 234 | "cell_type": "markdown", 235 | "metadata": {}, 236 | "source": [ 237 | "### Training stage\n", 238 | "On training stage, Decision Forest algorithm consume input data from S3 location.\n", 239 | "This container supports only .csv (\"comma-separated values\") files." 240 | ] 241 | }, 242 | { 243 | "cell_type": "code", 244 | "execution_count": 5, 245 | "metadata": { 246 | "scrolled": false 247 | }, 248 | "outputs": [ 249 | { 250 | "name": "stderr", 251 | "output_type": "stream", 252 | "text": [ 253 | "INFO:sagemaker:Creating training-job with name: testdaaldfcls2\n" 254 | ] 255 | }, 256 | { 257 | "name": "stdout", 258 | "output_type": "stream", 259 | "text": [ 260 | "2019-02-24 15:09:50 Starting - Starting the training job...\n", 261 | "2019-02-24 15:09:51 Starting - Launching requested ML instances......\n", 262 | "2019-02-24 15:11:18 Starting - Preparing the instances for training...\n", 263 | "2019-02-24 15:11:49 Downloading - Downloading input data...\n", 264 | "2019-02-24 15:11:56 Training - Downloading the training image.....\n", 265 | "\u001b[31m2019-02-24 15:12:57 INFO Container setup completed, In Docker entrypoint - train... \u001b[0m\n", 266 | "\u001b[31m2019-02-24 15:12:57 INFO Default Hyperparameters loaded: \u001b[0m\n", 267 | "\u001b[31m2019-02-24 15:12:57 INFO {'nClasses': '2', 'task': 'none', 'fptype': 'double', 'method': 'defaultDense', 'nTrees': '100', 'observationsPerTreeFraction': '1', 'featuresPerNode': '0', 'maxTreeDepth': '0', 'minObservationsInLeafNode': '0', 'seed': '777', 'impurityThreshold': '0', 'varImportance': 'none', 'resultsToCompute': '', 'memorySavingMode': 'False', 'bootstrap': 'False'}\u001b[0m\n", 268 | "\u001b[31m2019-02-24 15:12:57 INFO classification\u001b[0m\n", 269 | "\u001b[31m2019-02-24 15:12:57 INFO Updated with user hyperparameters, Final Hyperparameters: \u001b[0m\n", 270 | "\u001b[31m2019-02-24 15:12:57 INFO {'nClasses': '5', 'task': 'classification', 'fptype': 'double', 'method': 'defaultDense', 'nTrees': '100', 'observationsPerTreeFraction': '1.0', 'featuresPerNode': '0', 'maxTreeDepth': '0', 'minObservationsInLeafNode': '1', 'seed': '777', 'impurityThreshold': '0.0', 'varImportance': 'none', 'resultsToCompute': 'none', 'memorySavingMode': 'False', 'bootstrap': 'False'}\u001b[0m\n", 271 | "\u001b[31m2019-02-24 15:12:57 INFO Reading training data... \u001b[0m\n", 272 | "\u001b[31m2019-02-24 15:12:57 INFO Train data shape: (16000, 6)\u001b[0m\n", 273 | "\u001b[31m2019-02-24 15:12:57 INFO Files loading time: 0.021045207977294922\u001b[0m\n", 274 | "\u001b[31m2019-02-24 15:12:57 INFO Starting DAAL Decision Forest training...\u001b[0m\n", 275 | "\u001b[31m2019-02-24 15:12:58 INFO Training time, sec: 0.35150837898254395\u001b[0m\n", 276 | "\u001b[31m2019-02-24 15:12:58 INFO Saving model results...\u001b[0m\n", 277 | "\u001b[31m2019-02-24 15:12:58 INFO To get outOfBagErrorPerObservation use the method 'outOfBagErrorPerObservation' on the training model object\u001b[0m\n", 278 | "\u001b[31m2019-02-24 15:12:58 INFO Parameters saved at /opt/ml/model/parameters.json\u001b[0m\n", 279 | "\n", 280 | "2019-02-24 15:13:08 Uploading - Uploading generated training model\n", 281 | "2019-02-24 15:13:08 Completed - Training job completed\n", 282 | "Billable seconds: 79\n" 283 | ] 284 | } 285 | ], 286 | "source": [ 287 | "daal_decision_forest.fit({\"training\": \"s3:///\"})" 288 | ] 289 | }, 290 | { 291 | "cell_type": "markdown", 292 | "metadata": {}, 293 | "source": [ 294 | "### Real-time prediction\n", 295 | "Firstly, you need to deploy SageMaker endpoint that consumes data." 296 | ] 297 | }, 298 | { 299 | "cell_type": "code", 300 | "execution_count": 6, 301 | "metadata": { 302 | "scrolled": true 303 | }, 304 | "outputs": [ 305 | { 306 | "name": "stderr", 307 | "output_type": "stream", 308 | "text": [ 309 | "INFO:sagemaker:Creating model package with name: daal-df-cls-2019-02-24-15-13-43-176\n" 310 | ] 311 | }, 312 | { 313 | "name": "stdout", 314 | "output_type": "stream", 315 | "text": [ 316 | ".........." 317 | ] 318 | }, 319 | { 320 | "name": "stderr", 321 | "output_type": "stream", 322 | "text": [ 323 | "INFO:sagemaker:Creating model with name: daal-df-cls-2019-02-24-15-13-43-176-2019-02-24-15-14-28-639\n" 324 | ] 325 | }, 326 | { 327 | "name": "stdout", 328 | "output_type": "stream", 329 | "text": [ 330 | "\n" 331 | ] 332 | }, 333 | { 334 | "name": "stderr", 335 | "output_type": "stream", 336 | "text": [ 337 | "INFO:sagemaker:Creating endpoint with name testdaaldfcls2\n" 338 | ] 339 | }, 340 | { 341 | "name": "stdout", 342 | "output_type": "stream", 343 | "text": [ 344 | "--------------------------------------------------------------!" 345 | ] 346 | } 347 | ], 348 | "source": [ 349 | "predictor = daal_decision_forest.deploy(1, \"ml.m4.xlarge\", serializer=sagemaker.predictor.csv_serializer)" 350 | ] 351 | }, 352 | { 353 | "cell_type": "markdown", 354 | "metadata": {}, 355 | "source": [ 356 | "Secondly, you should pass data as numpy array to endpoint and get predictions.\n", 357 | "\n", 358 | "In this example we are passing random data, but you can use any numpy 2D array" 359 | ] 360 | }, 361 | { 362 | "cell_type": "code", 363 | "execution_count": 7, 364 | "metadata": {}, 365 | "outputs": [ 366 | { 367 | "name": "stdout", 368 | "output_type": "stream", 369 | "text": [ 370 | "4\n", 371 | "2\n", 372 | "2\n", 373 | "1\n", 374 | "3\n", 375 | "2\n", 376 | "3\n", 377 | "3\n", 378 | "3\n", 379 | "3\n", 380 | "\n" 381 | ] 382 | } 383 | ], 384 | "source": [ 385 | "import numpy as np\n", 386 | "\n", 387 | "predict_data = np.random.random(size=(10,10))\n", 388 | "print(predictor.predict(predict_data).decode(\"utf-8\"))" 389 | ] 390 | }, 391 | { 392 | "cell_type": "markdown", 393 | "metadata": {}, 394 | "source": [ 395 | "Don't forget to delete endpoint if you don't need it anymore." 396 | ] 397 | }, 398 | { 399 | "cell_type": "code", 400 | "execution_count": 8, 401 | "metadata": {}, 402 | "outputs": [ 403 | { 404 | "name": "stderr", 405 | "output_type": "stream", 406 | "text": [ 407 | "INFO:sagemaker:Deleting endpoint with name: testdaaldfcls2\n" 408 | ] 409 | } 410 | ], 411 | "source": [ 412 | "sess.delete_endpoint(predictor.endpoint)" 413 | ] 414 | }, 415 | { 416 | "cell_type": "markdown", 417 | "metadata": {}, 418 | "source": [ 419 | "### Batch transform job\n", 420 | "If you don't need real-time prediction, you can use transform job. It uses saved model, compute predictions one time and saves it in specified or auto-generated output path.\n", 421 | "\n", 422 | "More about transform jobs: https://docs.aws.amazon.com/sagemaker/latest/dg/how-it-works-batch.html\n", 423 | "\n", 424 | "Transformer API: https://sagemaker.readthedocs.io/en/latest/transformer.html" 425 | ] 426 | }, 427 | { 428 | "cell_type": "code", 429 | "execution_count": 10, 430 | "metadata": {}, 431 | "outputs": [ 432 | { 433 | "name": "stderr", 434 | "output_type": "stream", 435 | "text": [ 436 | "INFO:sagemaker:Creating model package with name: daal-df-cls-2019-02-24-15-22-38-863\n" 437 | ] 438 | }, 439 | { 440 | "name": "stdout", 441 | "output_type": "stream", 442 | "text": [ 443 | ".........." 444 | ] 445 | }, 446 | { 447 | "name": "stderr", 448 | "output_type": "stream", 449 | "text": [ 450 | "INFO:sagemaker:Creating model with name: daal-df-cls-2019-02-24-15-22-38-863-2019-02-24-15-23-24-300\n" 451 | ] 452 | }, 453 | { 454 | "name": "stdout", 455 | "output_type": "stream", 456 | "text": [ 457 | "\n" 458 | ] 459 | }, 460 | { 461 | "name": "stderr", 462 | "output_type": "stream", 463 | "text": [ 464 | "INFO:sagemaker:Creating transform job with name: testdfclst1\n" 465 | ] 466 | }, 467 | { 468 | "name": "stdout", 469 | "output_type": "stream", 470 | "text": [ 471 | "........................................!\n", 472 | "s3://sagemaker-us-east-2-123123123123/testdfclst1\n" 473 | ] 474 | } 475 | ], 476 | "source": [ 477 | "transformer = daal_decision_forest.transformer(1, 'ml.m4.xlarge')\n", 478 | "transformer.transform(\"s3:///\", content_type='text/csv', job_name=\"\")\n", 479 | "transformer.wait()\n", 480 | "print(transformer.output_path)" 481 | ] 482 | } 483 | ], 484 | "metadata": { 485 | "kernelspec": { 486 | "display_name": "conda_python3", 487 | "language": "python", 488 | "name": "conda_python3" 489 | }, 490 | "language_info": { 491 | "codemirror_mode": { 492 | "name": "ipython", 493 | "version": 3 494 | }, 495 | "file_extension": ".py", 496 | "mimetype": "text/x-python", 497 | "name": "python", 498 | "nbconvert_exporter": "python", 499 | "pygments_lexer": "ipython3", 500 | "version": "3.6.5" 501 | } 502 | }, 503 | "nbformat": 4, 504 | "nbformat_minor": 2 505 | } 506 | -------------------------------------------------------------------------------- /daal/logistic-regression-example.ipynb: -------------------------------------------------------------------------------- 1 | { 2 | "cells": [ 3 | { 4 | "cell_type": "markdown", 5 | "metadata": {}, 6 | "source": [ 7 | "# Logistic Regression with Intel® Data Analytics Acceleration Library in Amazon SageMaker\n", 8 | "\n", 9 | "# Introduction\n", 10 | "\n", 11 | "Intel® Data Analytics Acceleration Library (Intel® DAAL) is the library of Intel® architecture optimized building blocks covering all stages of data analytics: data acquisition from a data source, preprocessing, transformation, data mining, modeling, validation, and decision making. One of its classification algorithms is Logistic Regression.\n", 12 | "\n", 13 | "Logistic Regression is a method for modeling the relationships between one or more explanatory variables and a categorical variable by expressing the posterior statistical distribution of the categorical variable via linear functions on observed data. If the categorical variable is binary, that is it takes only two values, \"0\" and \"1\", the Logistic Regression is simple, otherwise, it is multinomial.\n", 14 | "DAAL Logistic Regression algorithm support L1 and L2 regularizations.\n", 15 | "\n", 16 | "\n", 17 | "Intel® DAAL developer guide: https://software.intel.com/en-us/daal-programming-guide\n", 18 | "\n", 19 | "Intel® DAAL documentation for Logistic Regression: https://software.intel.com/en-us/daal-programming-guide-logistic-regression" 20 | ] 21 | }, 22 | { 23 | "cell_type": "markdown", 24 | "metadata": {}, 25 | "source": [ 26 | "* [Hyperparameters description](#1-bullet)\n", 27 | "* [Usage of the algorithm](#2-bullet)\n", 28 | " * [Upload the data for training](#3-bullet)\n", 29 | " * [Creating Training Job using Algorithm ARN](#4-bullet)\n", 30 | " * [Live Inference Endpoint for Prediction stage](#5-bullet)\n", 31 | " * [Batch transform job](#6-bullet)" 32 | ] 33 | }, 34 | { 35 | "cell_type": "markdown", 36 | "metadata": {}, 37 | "source": [ 38 | "# Hyperparameters description: \n", 39 | "\n", 40 | "\n", 41 | " \n", 42 | " \n", 43 | " \n", 44 | " \n", 45 | " \n", 46 | " \n", 47 | " \n", 48 | " \n", 49 | " \n", 50 | " \n", 51 | " \n", 52 | " \n", 53 | "
Required parametersTypeDefault valueDescription
nClassesintegerNoneNumber of classes in training dataset
\n", 54 | "\n", 55 | "\n", 56 | "\n", 57 | "\n", 58 | " \n", 59 | " \n", 60 | " \n", 61 | " \n", 62 | " \n", 63 | " \n", 64 | " \n", 65 | " \n", 66 | " \n", 67 | " \n", 68 | " \n", 69 | " \n", 70 | " \n", 71 | " \n", 72 | " \n", 73 | " \n", 74 | " \n", 75 | " \n", 76 | " \n", 77 | " \n", 78 | " \n", 79 | " \n", 80 | " \n", 81 | " \n", 82 | " \n", 83 | " \n", 84 | " \n", 85 | " \n", 86 | " \n", 87 | " \n", 88 | " \n", 89 | " \n", 90 | " \n", 91 | " \n", 92 | " \n", 95 | " \n", 96 | " \n", 97 | " \n", 98 | " \n", 99 | " \n", 100 | " \n", 101 | " \n", 102 | " \n", 103 | " \n", 104 | " \n", 105 | " \n", 106 | " \n", 108 | " \n", 109 | " \n", 110 | " \n", 111 | " \n", 112 | " \n", 113 | " \n", 114 | " \n", 115 | " \n", 116 | " \n", 117 | " \n", 118 | " \n", 119 | " \n", 120 | " \n", 121 | " \n", 122 | " \n", 123 | " \n", 124 | " \n", 125 | " \n", 126 | " \n", 127 | " \n", 128 | " \n", 129 | " \n", 130 | " \n", 131 | " \n", 132 | " \n", 133 | " \n", 134 | " \n", 135 | " \n", 136 | " \n", 137 | " \n", 140 | " \n", 141 | "
Optional parametersTypeDefault valueDescription
penaltyL1float0Penalty coefficient for L1 regularization
penaltyL2float0Penalty coefficient for L2 regularization
interceptFlagboolFalseA flag that indicates a need to compute θ0j
solverNamestr'sgd'Name of solver that will be used for training stage
available values: 'lbfgs', 'adagrad', 'saga', 'sgd'
solverMethodstr'defaultDense'Method of the solver. Available values for 'sgd':
\n", 93 | " 'momentum', 'minibatch', 'defaultDense'
\n", 94 | " available values for others solver: 'defaultDense'
solverMaxIterationsinteger100Max number of iterations for training stage
solverAccuracyThresholdfloat1.0-e4Accuracy of the algorithm. The algorithm terminates when
\n", 107 | " this accuracy is achieved.
solverBatchSizeintegernumber of rows
in training dataset		Number of batch indices to compute the stochastic gradient.
solverLearningRatefloat1.0-e3learning rate for optimization problem
applicable for 'sgd','adagrad', 'saga' only
solverStepLengthfloat1.0-e3step size for optimization problem
applicable for 'lbfgs' only
solverCorrectionPairBatchSizeintegernumber of rows
in training dataset		Number of batch indices to compute Hessian aproximation.
applicable for 'lbfgs' only
solverLinteger1The number of iterations between
calculations of the curvature estimates
\n", 138 | " applicable for 'lbfgs' only\n", 139 | "
\n" 142 | ] 143 | }, 144 | { 145 | "cell_type": "markdown", 146 | "metadata": {}, 147 | "source": [ 148 | " For more detailes please visit:\n", 149 | " https://software.intel.com/en-us/daal-programming-guide\n", 150 | " \n", 151 | " All parameters that start from 'solver' have a name without 'solver' prefix in DAAL documentation" 152 | ] 153 | }, 154 | { 155 | "cell_type": "markdown", 156 | "metadata": {}, 157 | "source": [ 158 | "# Usage of the algorithm \n", 159 | "At the first we need to import SageMaker Python package, get execution role and create session." 160 | ] 161 | }, 162 | { 163 | "cell_type": "code", 164 | "execution_count": 1, 165 | "metadata": {}, 166 | "outputs": [], 167 | "source": [ 168 | "import numpy as np\n", 169 | "import pandas as pd\n", 170 | "from sagemaker import get_execution_role\n", 171 | "import sagemaker as sage\n", 172 | "\n", 173 | "role = get_execution_role()\n", 174 | "sess = sage.Session()" 175 | ] 176 | }, 177 | { 178 | "cell_type": "markdown", 179 | "metadata": {}, 180 | "source": [ 181 | "## Upload the data for training \n", 182 | "When training large models with huge amounts of data, you'll typically use big data tools, like Amazon Athena, AWS Glue, or Amazon EMR, to create your data in S3. For the purposes of this example, we're using some the classic Iris dataset\n", 183 | "\n", 184 | "We can use use the tools provided by the SageMaker Python SDK to upload the data to a default bucket." 185 | ] 186 | }, 187 | { 188 | "cell_type": "markdown", 189 | "metadata": {}, 190 | "source": [ 191 | "At first download iris-data via sklearn and split it into 'training' and 'test' data" 192 | ] 193 | }, 194 | { 195 | "cell_type": "code", 196 | "execution_count": 2, 197 | "metadata": {}, 198 | "outputs": [], 199 | "source": [ 200 | "from sklearn import datasets\n", 201 | "iris = datasets.load_iris()\n", 202 | "\n", 203 | "from sklearn.model_selection import train_test_split\n", 204 | "X_train, X_test, y_train, y_test = train_test_split(iris.data, iris.target, test_size=0.33, random_state=777)" 205 | ] 206 | }, 207 | { 208 | "cell_type": "markdown", 209 | "metadata": {}, 210 | "source": [ 211 | "Upload the training data to S3" 212 | ] 213 | }, 214 | { 215 | "cell_type": "code", 216 | "execution_count": 31, 217 | "metadata": {}, 218 | "outputs": [ 219 | { 220 | "name": "stdout", 221 | "output_type": "stream", 222 | "text": [ 223 | "Training artifacts will be uploaded at: s3://daal-log-reg-test/log_reg_iris_data_test\n", 224 | "And data_location will be a parameter for fit method (see training stage below).\n" 225 | ] 226 | }, 227 | { 228 | "data": { 229 | "text/plain": [ 230 | "'s3://daal-log-reg-test/log_reg_iris_data_test/training_data.csv'" 231 | ] 232 | }, 233 | "execution_count": 31, 234 | "metadata": {}, 235 | "output_type": "execute_result" 236 | } 237 | ], 238 | "source": [ 239 | "import numpy as np\n", 240 | "\n", 241 | "reshaped_Y_train = y_train.reshape(y_train.shape[0],1)\n", 242 | "\n", 243 | "#Last column in training dataset is labels. \n", 244 | "training_data = np.concatenate((X_train,reshaped_Y_train),axis=1)\n", 245 | "\n", 246 | "#save the training data\n", 247 | "train_data_file = 'training_data.csv'\n", 248 | "# NO comma at the end of each line in training data.\n", 249 | "np.savetxt(train_data_file,training_data,delimiter=',')\n", 250 | "\n", 251 | "# S3 prefix\n", 252 | "bucket_name = 'daal-log-reg-test'\n", 253 | "data_key = 'log_reg_iris_data_test'\n", 254 | "\n", 255 | "output_location = 's3://{}/{}'.format(bucket_name, 'output')\n", 256 | "data_location = output_location = 's3://{}/{}'.format(bucket_name, data_key)\n", 257 | "print (\"Training artifacts will be uploaded at: \" + output_location)\n", 258 | "print (\"And data_location will be a parameter for fit method (see training stage below).\")\n", 259 | "\n", 260 | "sess.upload_data(train_data_file, bucket=bucket_name, key_prefix=data_key)" 261 | ] 262 | }, 263 | { 264 | "cell_type": "markdown", 265 | "metadata": {}, 266 | "source": [ 267 | "Example of hyperparameters list:" 268 | ] 269 | }, 270 | { 271 | "cell_type": "code", 272 | "execution_count": 32, 273 | "metadata": {}, 274 | "outputs": [], 275 | "source": [ 276 | "hyperparameters={\"nClasses\": 3,\n", 277 | " \"penaltyL1\": 0,\n", 278 | " \"penaltyL2\": 0,\n", 279 | " \"interceptFlag\": False,\n", 280 | " \"solverBatchSize\":100, #as training data has 100 samples only, but default value is 150\n", 281 | " #\"solverName\": \"lbfgs\", \n", 282 | " #\"solverMaxIterations\": 1000,\n", 283 | " #\"solverAccuracyThreshold\": 0.0001,\n", 284 | " #\"solverL\": 1\n", 285 | " }" 286 | ] 287 | }, 288 | { 289 | "cell_type": "markdown", 290 | "metadata": {}, 291 | "source": [ 292 | "Then, you need to create SageMaker Estimator instance with following parameters:\n", 293 | "\n", 294 | " \n", 295 | " \n", 296 | " \n", 297 | " \n", 298 | " \n", 299 | " \n", 300 | " \n", 301 | " \n", 302 | " \n", 303 | " \n", 304 | " \n", 305 | " \n", 306 | " \n", 307 | " \n", 308 | " \n", 309 | " \n", 310 | " \n", 311 | " \n", 312 | " \n", 313 | " \n", 314 | " \n", 315 | " \n", 316 | " \n", 317 | " \n", 318 | " \n", 319 | " \n", 320 | " \n", 321 | " \n", 322 | " \n", 323 | " \n", 324 | " \n", 325 | " \n", 326 | " \n", 327 | " \n", 328 | " \n", 329 | " \n", 330 | "
Parameter nameDescription
image_nameThe container image to use for training
roleAn AWS IAM role. The SageMaker training jobs and APIs that create SageMaker endpoints use this role to access training data and models
train_instance_countNumber of Amazon EC2 instances to use for training. Should be 1, because it is not distributed version of algorithm
train_instance_typeType of EC2 instance to use for training. See available types on Amazon Marketplace page of algorithm
input_modeThe input mode that the algorithm supports. May be \"File\" or \"Pipe\"
output_pathS3 location for saving the trainig result (model artifacts and output files)
sagemaker_sessionSession object which manages interactions with Amazon SageMaker APIs and any other AWS services needed
hyperparametersDictionary containing the hyperparameters to initialize this estimator with
\n", 331 | "SageMaker Estimator documentation: https://sagemaker.readthedocs.io/en/latest/estimators.html" 332 | ] 333 | }, 334 | { 335 | "cell_type": "markdown", 336 | "metadata": {}, 337 | "source": [ 338 | "## Creating Training Job using Algorithm ARN\n", 339 | "Please put in the algorithm arn you want to use below. This can either be an AWS Marketplace algorithm you subscribed to (or) one of the algorithms you created in your own account.\n", 340 | "The algorithm arn listed below belongs to the Intel DAAL Logistic Regression." 341 | ] 342 | }, 343 | { 344 | "cell_type": "code", 345 | "execution_count": 33, 346 | "metadata": {}, 347 | "outputs": [], 348 | "source": [ 349 | "daal_log_reg_arn = \"arn:aws:sagemaker:us-east-2:057799348421:algorithm/intel-daal-logistic-regression-ce8a1f38da2f8a234e4205a356021dbf\" # you can find it on algorithm page in your subscriptions\n", 350 | "#daal_log_reg_arn = \"\"\n", 351 | "daal_log_reg = sage.algorithm.AlgorithmEstimator(\n", 352 | " algorithm_arn=daal_log_reg_arn,\n", 353 | " base_job_name=\"daal-log-reg-alg-test\",\n", 354 | " role=role,\n", 355 | " train_instance_count=1,\n", 356 | " train_instance_type='ml.m4.xlarge',\n", 357 | " input_mode=\"File\",\n", 358 | " output_path=output_location,\n", 359 | " sagemaker_session=sess,\n", 360 | " hyperparameters=hyperparameters\n", 361 | ")" 362 | ] 363 | }, 364 | { 365 | "cell_type": "markdown", 366 | "metadata": {}, 367 | "source": [ 368 | "## Run training stage\n", 369 | "On training stage, Logistic Rergession algorithm consume input data from S3 location and train the model." 370 | ] 371 | }, 372 | { 373 | "cell_type": "code", 374 | "execution_count": 35, 375 | "metadata": {}, 376 | "outputs": [ 377 | { 378 | "name": "stderr", 379 | "output_type": "stream", 380 | "text": [ 381 | "INFO:sagemaker:Creating training-job with name: daal-log-reg-alg-test-2018-11-30-10-58-20-329\n" 382 | ] 383 | }, 384 | { 385 | "name": "stdout", 386 | "output_type": "stream", 387 | "text": [ 388 | "2018-11-30 10:58:20 Starting - Starting the training job...\n", 389 | "2018-11-30 10:58:25 Starting - Launching requested ML instances......\n", 390 | "2018-11-30 10:59:21 Starting - Preparing the instances for training......\n", 391 | "2018-11-30 11:00:40 Downloading - Downloading input data\n", 392 | "2018-11-30 11:00:40 Training - Downloading the training image..\n", 393 | "\u001b[31m2018-11-30 11:00:59 INFO Container setup completed, In Docker entrypoint - train... \u001b[0m\n", 394 | "\u001b[31m2018-11-30 11:00:59 INFO Default Hyperparameters loaded: \u001b[0m\n", 395 | "\u001b[31m2018-11-30 11:00:59 INFO {'dtype': 'float',\n", 396 | " 'interceptFlag': True,\n", 397 | " 'nClasses': 0,\n", 398 | " 'penaltyL1': 0,\n", 399 | " 'penaltyL2': 0}\u001b[0m\n", 400 | "\u001b[31m2018-11-30 11:00:59 INFO Reading training data... \u001b[0m\n", 401 | "\u001b[31m2018-11-30 11:00:59 INFO Training data with labels shape: (100, 5)\u001b[0m\n", 402 | "\u001b[31m2018-11-30 11:00:59 INFO Updated with user hyperparameters, Final Hyperparameters: \u001b[0m\n", 403 | "\u001b[31m2018-11-30 11:00:59 INFO {'dtype': 'float',\n", 404 | " 'interceptFlag': 'False',\n", 405 | " 'nClasses': '3',\n", 406 | " 'penaltyL1': '0.0',\n", 407 | " 'penaltyL2': '0.0'}\u001b[0m\n", 408 | "\u001b[31m2018-11-30 11:00:59 INFO {'solverAccuracyThreshold': '0.0001',\n", 409 | " 'solverBatchSize': '100',\n", 410 | " 'solverCorrectionPairBatchSize': 100,\n", 411 | " 'solverL': 1,\n", 412 | " 'solverLearningRate': '0.001',\n", 413 | " 'solverMaxIterations': '1000',\n", 414 | " 'solverMethod': 'momentum',\n", 415 | " 'solverName': 'sgd',\n", 416 | " 'solverStepLength': 0.001}\u001b[0m\n", 417 | "\u001b[31m2018-11-30 11:00:59 INFO If optional parameters were not specified default values will be used.\u001b[0m\n", 418 | "\u001b[31m2018-11-30 11:00:59 INFO Training Data Shape: (100, 4)\u001b[0m\n", 419 | "\u001b[31m2018-11-30 11:00:59 INFO Training Labels Shape: (100, 1)\u001b[0m\n", 420 | "\u001b[31m2018-11-30 11:00:59 INFO Starting DAAL Logistic Regression training...\u001b[0m\n", 421 | "\u001b[31m2018-11-30 11:00:59 INFO Training time in sec = 0.040670156478881836\u001b[0m\n", 422 | "\u001b[31m2018-11-30 11:00:59 INFO number of classes saved at /opt/ml/model/daal-log-reg-train-features-classes.csv\u001b[0m\n", 423 | "\u001b[31m2018-11-30 11:00:59 INFO dtype saved at /opt/ml/model/daal-log-reg-dtype.txt\u001b[0m\n", 424 | "\u001b[31m2018-11-30 11:00:59 INFO Model saved at <_io.BufferedWriter name='/opt/ml/model/daal-log-reg-train-model.pkl'>\u001b[0m\n", 425 | "\n", 426 | "2018-11-30 11:01:05 Uploading - Uploading generated training model\n", 427 | "2018-11-30 11:01:05 Completed - Training job completed\n", 428 | "Billable seconds: 43\n" 429 | ] 430 | } 431 | ], 432 | "source": [ 433 | "daal_log_reg.fit({\"training\": data_location})" 434 | ] 435 | }, 436 | { 437 | "cell_type": "markdown", 438 | "metadata": {}, 439 | "source": [ 440 | "## Live Inference Endpoint for Prediction stage\n", 441 | "On prediction stage, Logistic Regression algorithm compute probabulity of classes and retern the class for each input samples.\n", 442 | "Firstly, you need to deploy SageMaker endpoint that consumes data.\n" 443 | ] 444 | }, 445 | { 446 | "cell_type": "code", 447 | "execution_count": 23, 448 | "metadata": {}, 449 | "outputs": [ 450 | { 451 | "name": "stderr", 452 | "output_type": "stream", 453 | "text": [ 454 | "INFO:sagemaker:Creating model package with name: intel-daal-logistic-regression-ce8a1f38-2018-11-30-10-05-13-703\n" 455 | ] 456 | }, 457 | { 458 | "name": "stdout", 459 | "output_type": "stream", 460 | "text": [ 461 | ".........." 462 | ] 463 | }, 464 | { 465 | "name": "stderr", 466 | "output_type": "stream", 467 | "text": [ 468 | "INFO:sagemaker:Creating model with name: intel-daal-logistic-regression-ce8a1f38-2018-11-30-10-05-59-280\n" 469 | ] 470 | }, 471 | { 472 | "name": "stdout", 473 | "output_type": "stream", 474 | "text": [ 475 | "\n" 476 | ] 477 | }, 478 | { 479 | "name": "stderr", 480 | "output_type": "stream", 481 | "text": [ 482 | "INFO:sagemaker:Creating endpoint with name daal-log-reg-alg-test-2018-11-30-10-00-36-782\n" 483 | ] 484 | }, 485 | { 486 | "name": "stdout", 487 | "output_type": "stream", 488 | "text": [ 489 | "--------------------------------------------------------------------------!" 490 | ] 491 | } 492 | ], 493 | "source": [ 494 | "from sagemaker.predictor import csv_serializer\n", 495 | "predictor = daal_log_reg.deploy(1, 'ml.m4.xlarge', serializer=csv_serializer)" 496 | ] 497 | }, 498 | { 499 | "cell_type": "markdown", 500 | "metadata": {}, 501 | "source": [ 502 | "Define functions to handle response from predictor instance at first:" 503 | ] 504 | }, 505 | { 506 | "cell_type": "code", 507 | "execution_count": 24, 508 | "metadata": {}, 509 | "outputs": [], 510 | "source": [ 511 | "def output_to_np(prediction, numberOfSamples, nClasses):\n", 512 | " if nClasses == 2:\n", 513 | " return np.fromstring(prediction, dtype=np.float64, sep=' ').reshape(2,numberOfSamples)\n", 514 | " if nClasses > 2:\n", 515 | " return np.fromstring(prediction, dtype=np.float64, sep=' ').reshape(nClasses+1,numberOfSamples)\n", 516 | "\n", 517 | "def output_to_pd(prediction, nClasses):\n", 518 | " C=[]\n", 519 | " C.append('lables')\n", 520 | " if nClasses > 2:\n", 521 | " for i in range(nClasses):\n", 522 | " C.append('probability of class ' + str(i))\n", 523 | " if nClasses == 2:\n", 524 | " C.append('probability of class 1')\n", 525 | " return pd.DataFrame(np.transpose(prediction), columns=C)" 526 | ] 527 | }, 528 | { 529 | "cell_type": "markdown", 530 | "metadata": {}, 531 | "source": [ 532 | "After deployment, you should pass data as numpy array to predictor instance and get predicted lables and probabilities." 533 | ] 534 | }, 535 | { 536 | "cell_type": "code", 537 | "execution_count": 26, 538 | "metadata": {}, 539 | "outputs": [ 540 | { 541 | "name": "stdout", 542 | "output_type": "stream", 543 | "text": [ 544 | "2.0\n", 545 | "0.0\n", 546 | "2.0\n", 547 | "2.0\n", 548 | "1.0\n", 549 | "0.0\n", 550 | "2.0\n", 551 | "2.0\n", 552 | "0.0\n", 553 | "0.0\n", 554 | "2.0\n", 555 | "1.0\n", 556 | "1.0\n", 557 | "2.0\n", 558 | "2.0\n", 559 | "2.0\n", 560 | "0.0\n", 561 | "2.0\n", 562 | "0.0\n", 563 | "1.0\n", 564 | "1.0\n", 565 | "1.0\n", 566 | "2.0\n", 567 | "0.0\n", 568 | "2.0\n", 569 | "0.0\n", 570 | "1.0\n", 571 | "0.0\n", 572 | "2.0\n", 573 | "2.0\n", 574 | "0.0\n", 575 | "2.0\n", 576 | "0.0\n", 577 | "2.0\n", 578 | "1.0\n", 579 | "0.0\n", 580 | "0.0\n", 581 | "0.0\n", 582 | "1.0\n", 583 | "0.0\n", 584 | "0.0\n", 585 | "2.0\n", 586 | "1.0\n", 587 | "1.0\n", 588 | "0.0\n", 589 | "2.0\n", 590 | "2.0\n", 591 | "0.0\n", 592 | "2.0\n", 593 | "1.0\n", 594 | "\n", 595 | "\n" 596 | ] 597 | } 598 | ], 599 | "source": [ 600 | "#usage from slplited data\n", 601 | "payload = X_test\n", 602 | "\n", 603 | "ground_truth = y_test\n", 604 | "prediction = predictor.predict(payload).decode('utf-8')\n", 605 | "print(prediction)\n", 606 | "#np_res = output_to_np(prediction=prediction,numberOfSamples=payload.shape[0],nClasses=3)\n", 607 | "#pd_res = output_to_pd(prediction=np_res, nClasses=3)" 608 | ] 609 | }, 610 | { 611 | "cell_type": "markdown", 612 | "metadata": {}, 613 | "source": [ 614 | "Print the first 5 rows of obtained prediction:" 615 | ] 616 | }, 617 | { 618 | "cell_type": "code", 619 | "execution_count": 32, 620 | "metadata": {}, 621 | "outputs": [ 622 | { 623 | "data": { 624 | "text/html": [ 625 | "
\n", 626 | "\n", 639 | "\n", 640 | " \n", 641 | " \n", 642 | " \n", 643 | " \n", 644 | " \n", 645 | " \n", 646 | " \n", 647 | " \n", 648 | " \n", 649 | " \n", 650 | " \n", 651 | " \n", 652 | " \n", 653 | " \n", 654 | " \n", 655 | " \n", 656 | " \n", 657 | " \n", 658 | " \n", 659 | " \n", 660 | " \n", 661 | " \n", 662 | " \n", 663 | " \n", 664 | " \n", 665 | " \n", 666 | " \n", 667 | " \n", 668 | " \n", 669 | " \n", 670 | " \n", 671 | " \n", 672 | " \n", 673 | " \n", 674 | " \n", 675 | " \n", 676 | " \n", 677 | " \n", 678 | " \n", 679 | " \n", 680 | " \n", 681 | " \n", 682 | " \n", 683 | " \n", 684 | " \n", 685 | " \n", 686 | "
lablesprobability of class 0probability of class 1probability of class 2
02.09.311358e-100.0106599.893406e-01
10.09.948357e-010.0051641.804341e-12
22.02.409668e-090.0187539.812474e-01
32.03.761642e-090.0035499.964515e-01
41.02.466924e-050.8580581.419178e-01
\n", 687 | "
" 688 | ], 689 | "text/plain": [ 690 | " lables probability of class 0 probability of class 1 \\\n", 691 | "0 2.0 9.311358e-10 0.010659 \n", 692 | "1 0.0 9.948357e-01 0.005164 \n", 693 | "2 2.0 2.409668e-09 0.018753 \n", 694 | "3 2.0 3.761642e-09 0.003549 \n", 695 | "4 1.0 2.466924e-05 0.858058 \n", 696 | "\n", 697 | " probability of class 2 \n", 698 | "0 9.893406e-01 \n", 699 | "1 1.804341e-12 \n", 700 | "2 9.812474e-01 \n", 701 | "3 9.964515e-01 \n", 702 | "4 1.419178e-01 " 703 | ] 704 | }, 705 | "execution_count": 32, 706 | "metadata": {}, 707 | "output_type": "execute_result" 708 | } 709 | ], 710 | "source": [ 711 | "pd_res.head()" 712 | ] 713 | }, 714 | { 715 | "cell_type": "markdown", 716 | "metadata": {}, 717 | "source": [ 718 | "Compute the accuracy of trained model on test dataset 'y_test'." 719 | ] 720 | }, 721 | { 722 | "cell_type": "code", 723 | "execution_count": 35, 724 | "metadata": { 725 | "scrolled": false 726 | }, 727 | "outputs": [ 728 | { 729 | "name": "stdout", 730 | "output_type": "stream", 731 | "text": [ 732 | "DAAL Accuracy on Iris Train Set: 1.0\n" 733 | ] 734 | } 735 | ], 736 | "source": [ 737 | "from sklearn.metrics.cluster import v_measure_score\n", 738 | "\n", 739 | "prediction_arr = np_res[0]\n", 740 | "ground_truth = y_test\n", 741 | "#print(prediction_arr)\n", 742 | "#print(ground_truth)\n", 743 | "\n", 744 | "print(\"DAAL Accuracy on Iris Train Set: \", str(v_measure_score(prediction_arr, ground_truth)))" 745 | ] 746 | }, 747 | { 748 | "cell_type": "markdown", 749 | "metadata": {}, 750 | "source": [ 751 | "Don't forget to delete endpoint if you don't need it anymore. Otherwise it will run as a daemon process.\n" 752 | ] 753 | }, 754 | { 755 | "cell_type": "code", 756 | "execution_count": 27, 757 | "metadata": {}, 758 | "outputs": [ 759 | { 760 | "name": "stderr", 761 | "output_type": "stream", 762 | "text": [ 763 | "INFO:sagemaker:Deleting endpoint with name: daal-log-reg-alg-test-2018-11-30-10-00-36-782\n" 764 | ] 765 | } 766 | ], 767 | "source": [ 768 | "sess.delete_endpoint(predictor.endpoint)" 769 | ] 770 | }, 771 | { 772 | "cell_type": "markdown", 773 | "metadata": {}, 774 | "source": [ 775 | "## Batch transform job\n", 776 | "If you don't need real-time prediction, you can use transform job. It uses saved model, compute transformed data one time and saves it in specified or auto-generated output path.\n", 777 | "\n", 778 | "More about transform jobs: https://docs.aws.amazon.com/sagemaker/latest/dg/how-it-works-batch.html\n", 779 | "\n", 780 | "Transformer API: https://sagemaker.readthedocs.io/en/latest/transformer.html" 781 | ] 782 | }, 783 | { 784 | "cell_type": "code", 785 | "execution_count": 28, 786 | "metadata": {}, 787 | "outputs": [ 788 | { 789 | "name": "stderr", 790 | "output_type": "stream", 791 | "text": [ 792 | "INFO:sagemaker:Creating model package with name: intel-daal-logistic-regression-ce8a1f38-2018-11-30-10-25-22-669\n" 793 | ] 794 | }, 795 | { 796 | "name": "stdout", 797 | "output_type": "stream", 798 | "text": [ 799 | ".........." 800 | ] 801 | }, 802 | { 803 | "name": "stderr", 804 | "output_type": "stream", 805 | "text": [ 806 | "INFO:sagemaker:Creating model with name: intel-daal-logistic-regression-ce8a1f38-2018-11-30-10-26-08-303\n" 807 | ] 808 | }, 809 | { 810 | "name": "stdout", 811 | "output_type": "stream", 812 | "text": [ 813 | "\n" 814 | ] 815 | } 816 | ], 817 | "source": [ 818 | "transformer = daal_log_reg.transformer(1, \"ml.m4.xlarge\")" 819 | ] 820 | }, 821 | { 822 | "cell_type": "code", 823 | "execution_count": 29, 824 | "metadata": { 825 | "scrolled": true 826 | }, 827 | "outputs": [ 828 | { 829 | "name": "stderr", 830 | "output_type": "stream", 831 | "text": [ 832 | "INFO:sagemaker:Creating transform job with name: daal-log-reg-alg-test-2018-11-30-10-26-55-276\n" 833 | ] 834 | }, 835 | { 836 | "name": "stdout", 837 | "output_type": "stream", 838 | "text": [ 839 | ".......................................!\n" 840 | ] 841 | } 842 | ], 843 | "source": [ 844 | "transformer.transform(\"s3://daal-log-reg-test/input/data/test_data.csv\", content_type=\"text/csv\")\n", 845 | "transformer.wait()" 846 | ] 847 | }, 848 | { 849 | "cell_type": "code", 850 | "execution_count": 30, 851 | "metadata": {}, 852 | "outputs": [ 853 | { 854 | "name": "stdout", 855 | "output_type": "stream", 856 | "text": [ 857 | "2.0\n", 858 | "1.0\n", 859 | "2.0\n", 860 | "2.0\n", 861 | "1.0\n", 862 | "0.0\n", 863 | "0.0\n", 864 | "0.0\n", 865 | "2.0\n", 866 | "2.0\n", 867 | "1.0\n", 868 | "2.0\n", 869 | "0.0\n", 870 | "0.0\n", 871 | "0.0\n", 872 | "2.0\n", 873 | "1.0\n", 874 | "2.0\n", 875 | "0.0\n", 876 | "0.0\n", 877 | "1.0\n", 878 | "0.0\n", 879 | "2.0\n", 880 | "1.0\n", 881 | "0.0\n", 882 | "2.0\n", 883 | "1.0\n", 884 | "2.0\n", 885 | "0.0\n", 886 | "0.0\n", 887 | "0.0\n", 888 | "1.0\n", 889 | "2.0\n", 890 | "1.0\n", 891 | "1.0\n", 892 | "2.0\n", 893 | "2.0\n", 894 | "2.0\n", 895 | "2.0\n", 896 | "1.0\n", 897 | "1.0\n", 898 | "2.0\n", 899 | "1.0\n", 900 | "0.0\n", 901 | "2.0\n", 902 | "0.0\n", 903 | "0.0\n", 904 | "0.0\n", 905 | "1.0\n", 906 | "1.0\n", 907 | "1.0\n", 908 | "0.0\n", 909 | "1.0\n", 910 | "1.0\n", 911 | "0.0\n", 912 | "2.0\n", 913 | "2.0\n", 914 | "0.0\n", 915 | "0.0\n", 916 | "1.0\n", 917 | "0.0\n", 918 | "0.0\n", 919 | "1.0\n", 920 | "1.0\n", 921 | "2.0\n", 922 | "0.0\n", 923 | "0.0\n", 924 | "2.0\n", 925 | "1.0\n", 926 | "1.0\n", 927 | "1.0\n", 928 | "2.0\n", 929 | "1.0\n", 930 | "1.0\n", 931 | "2.0\n", 932 | "2.0\n", 933 | "1.0\n", 934 | "1.0\n", 935 | "2.0\n", 936 | "1.0\n", 937 | "2.0\n", 938 | "0.0\n", 939 | "2.0\n", 940 | "0.0\n", 941 | "1.0\n", 942 | "2.0\n", 943 | "1.0\n", 944 | "1.0\n", 945 | "1.0\n", 946 | "0.0\n", 947 | "0.0\n", 948 | "0.0\n", 949 | "0.0\n", 950 | "2.0\n", 951 | "2.0\n", 952 | "1.0\n", 953 | "1.0\n", 954 | "1.0\n", 955 | "0.0\n", 956 | "2.0\n", 957 | "\n", 958 | "\n" 959 | ] 960 | } 961 | ], 962 | "source": [ 963 | "from urllib.parse import urlparse\n", 964 | "\n", 965 | "parsed_url = urlparse(transformer.output_path)\n", 966 | "bucket_name = parsed_url.netloc\n", 967 | "file_key = '{}/{}.out'.format(parsed_url.path[1:], \"test_data.csv\") # size of data is equal to 100\n", 968 | "\n", 969 | "s3_client = sess.boto_session.client('s3')\n", 970 | "\n", 971 | "response = s3_client.get_object(Bucket = sess.default_bucket(), Key = file_key)\n", 972 | "response_bytes = response['Body'].read().decode('utf-8')\n", 973 | "print(response_bytes)\n", 974 | "#size_data = 100\n", 975 | "#np_res = output_to_np(prediction=response_bytes,numberOfSamples=size_data,nClasses=3)\n", 976 | "#pd_res = output_to_pd(prediction=np_res, nClasses=3)\n" 977 | ] 978 | }, 979 | { 980 | "cell_type": "code", 981 | "execution_count": 47, 982 | "metadata": {}, 983 | "outputs": [ 984 | { 985 | "data": { 986 | "text/html": [ 987 | "
\n", 988 | "\n", 1001 | "\n", 1002 | " \n", 1003 | " \n", 1004 | " \n", 1005 | " \n", 1006 | " \n", 1007 | " \n", 1008 | " \n", 1009 | " \n", 1010 | " \n", 1011 | " \n", 1012 | " \n", 1013 | " \n", 1014 | " \n", 1015 | " \n", 1016 | " \n", 1017 | " \n", 1018 | " \n", 1019 | " \n", 1020 | " \n", 1021 | " \n", 1022 | " \n", 1023 | " \n", 1024 | " \n", 1025 | " \n", 1026 | " \n", 1027 | " \n", 1028 | " \n", 1029 | " \n", 1030 | " \n", 1031 | " \n", 1032 | " \n", 1033 | " \n", 1034 | " \n", 1035 | " \n", 1036 | " \n", 1037 | " \n", 1038 | " \n", 1039 | " \n", 1040 | " \n", 1041 | " \n", 1042 | " \n", 1043 | " \n", 1044 | " \n", 1045 | " \n", 1046 | " \n", 1047 | " \n", 1048 | "
lablesprobability of class 0probability of class 1probability of class 2
02.02.848256e-100.0080350.991965
11.07.074607e-060.8105790.189414
22.02.953847e-090.0051160.994884
32.08.515568e-090.0144030.985597
41.01.201045e-020.9875500.000439
\n", 1049 | "
" 1050 | ], 1051 | "text/plain": [ 1052 | " lables probability of class 0 probability of class 1 \\\n", 1053 | "0 2.0 2.848256e-10 0.008035 \n", 1054 | "1 1.0 7.074607e-06 0.810579 \n", 1055 | "2 2.0 2.953847e-09 0.005116 \n", 1056 | "3 2.0 8.515568e-09 0.014403 \n", 1057 | "4 1.0 1.201045e-02 0.987550 \n", 1058 | "\n", 1059 | " probability of class 2 \n", 1060 | "0 0.991965 \n", 1061 | "1 0.189414 \n", 1062 | "2 0.994884 \n", 1063 | "3 0.985597 \n", 1064 | "4 0.000439 " 1065 | ] 1066 | }, 1067 | "execution_count": 47, 1068 | "metadata": {}, 1069 | "output_type": "execute_result" 1070 | } 1071 | ], 1072 | "source": [ 1073 | "pd_res.head()" 1074 | ] 1075 | } 1076 | ], 1077 | "metadata": { 1078 | "kernelspec": { 1079 | "display_name": "Python 3", 1080 | "language": "python", 1081 | "name": "python3" 1082 | }, 1083 | "language_info": { 1084 | "codemirror_mode": { 1085 | "name": "ipython", 1086 | "version": 3 1087 | }, 1088 | "file_extension": ".py", 1089 | "mimetype": "text/x-python", 1090 | "name": "python", 1091 | "nbconvert_exporter": "python", 1092 | "pygments_lexer": "ipython3", 1093 | "version": "3.6.5" 1094 | } 1095 | }, 1096 | "nbformat": 4, 1097 | "nbformat_minor": 2 1098 | } 1099 | --------------------------------------------------------------------------------