PROSAGA码农传奇-机器学习-亚马逊机器学习和SageMaker算法

<div class =“post-text”itemprop =“text”>
  
    我不确定Amazon ML，但SageMaker使用此处列出的docker容器进行内置培训：
    <a href="https://docs.aws.amazon.com/sagemaker/latest/dg/sagemaker-algo-docker-registry-paths.html" rel="noreferrer">
      https://docs.aws.amazon.com/sagemaker/latest/dg/sagemaker-algo-docker-registry-paths.html
    </A>
  
  
    所以，一般来说，你可以用Amazon ML做的任何东西你应该能够用SageMaker做什么（尽管Amazon ML有一个很漂亮的架构编辑器）。
  
  
    您可以查看每个容器，深入了解它们的工作原理。
  
  
    您可以在此处找到SageMaker中可用算法的详尽列表：

<a href="https://docs.aws.amazon.com/sagemaker/latest/dg/algos.html" rel="noreferrer">
      https://docs.aws.amazon.com/sagemaker/latest/dg/algos.html
    </A>
  
  
    目前，截至2017年12月，这些算法都可用：
  
  <UL>
    <LI>
      <a href="https://docs.aws.amazon.com/sagemaker/latest/dg/linear-learner.html" rel="noreferrer">
        线性学习者
      </A>
    </LI>
    <LI>
      <a href="https://docs.aws.amazon.com/sagemaker/latest/dg/fact-machines.html" rel="noreferrer">
        分解机器
      </A>
    </LI>
    <LI>
      <a href="https://docs.aws.amazon.com/sagemaker/latest/dg/xgboost.html" rel="noreferrer">
        XGBoost算法
      </A>
    </LI>
    <LI>
      <a href="https://docs.aws.amazon.com/sagemaker/latest/dg/image-classification.html" rel="noreferrer">
        图像分类算法
      </A>
    </LI>
    <LI>
      <a href="https://docs.aws.amazon.com/sagemaker/latest/dg/seq-2-seq.html" rel="noreferrer">
        Amazon SageMaker Sequence2Sequence
      </A>
    </LI>
    <LI>
      <a href="https://docs.aws.amazon.com/sagemaker/latest/dg/k-means.html" rel="noreferrer">
        K均值算法
      </A>
    </LI>
    <LI>
      <a href="https://docs.aws.amazon.com/sagemaker/latest/dg/pca.html" rel="noreferrer">
        主成分分析（PCA）
      </A>
    </LI>
    <LI>
      <a href="https://docs.aws.amazon.com/sagemaker/latest/dg/lda.html" rel="noreferrer">
        潜在Dirichlet分配（LDA）
      </A>
    </LI>
    <LI>
      <a href="https://docs.aws.amazon.com/sagemaker/latest/dg/ntm.html" rel="noreferrer">
        神经主题模型（NTM）
      </A>
    </LI>
  </UL>
  
    这些算法的一般SageMaker SDK接口如下所示：
  
   <pre>
 <code>
 from sagemaker import KMeans
kmeans = KMeans(role="SageMakerRole",
 train_instance_count=2,
 train_instance_type='ml.c4.8xlarge',
 data_location="s3://training_data/",
 output_path="s3://model_artifacts/",
 k=10)

</code>
 </pre>
  
    这里的图书馆：
    <a href="https://github.com/awslabs/amazon-sagemaker-examples" rel="noreferrer">
      https://github.com/awslabs/amazon-sagemaker-examples
    </A>
    
和这里：
    <a href="https://github.com/aws/sagemaker-python-sdk" rel="noreferrer">
      https://github.com/aws/sagemaker-python-sdk
    </A>
     对于使用SageMaker特别有用。
  
  
    您也可以在这里使用Spark与SageMaker Spark库：
    <a href="https://github.com/aws/sagemaker-spark" rel="noreferrer">
      https://github.com/aws/sagemaker-spark
    </A>
  
</DIV>