« blib/lib/Paws/SageMaker/TransformEnvironmentMap.pm
↑ Summary
blib/lib/Paws/SageMaker/TransformJob.pm »
| 1 | # Generated by default/object.tt | ||
| 2 | package Paws::SageMaker::TransformInput; | ||
| 3 | use Moose; | ||
| 4 | has CompressionType => (is => 'ro', isa => 'Str'); | ||
| 5 | has ContentType => (is => 'ro', isa => 'Str'); | ||
| 6 | has DataSource => (is => 'ro', isa => 'Paws::SageMaker::TransformDataSource', required => 1); | ||
| 7 | has SplitType => (is => 'ro', isa => 'Str'); | ||
| 8 | |||
| 9 | 1; | ||
| 10 | |||
| 11 | ### main pod documentation begin ### | ||
| 12 | |||
| 13 | =head1 NAME | ||
| 14 | |||
| 15 | Paws::SageMaker::TransformInput | ||
| 16 | |||
| 17 | =head1 USAGE | ||
| 18 | |||
| 19 | This class represents one of two things: | ||
| 20 | |||
| 21 | =head3 Arguments in a call to a service | ||
| 22 | |||
| 23 | Use the attributes of this class as arguments to methods. You shouldn't make instances of this class. | ||
| 24 | Each attribute should be used as a named argument in the calls that expect this type of object. | ||
| 25 | |||
| 26 | As an example, if Att1 is expected to be a Paws::SageMaker::TransformInput object: | ||
| 27 | |||
| 28 | $service_obj->Method(Att1 => { CompressionType => $value, ..., SplitType => $value }); | ||
| 29 | |||
| 30 | =head3 Results returned from an API call | ||
| 31 | |||
| 32 | Use accessors for each attribute. If Att1 is expected to be an Paws::SageMaker::TransformInput object: | ||
| 33 | |||
| 34 | $result = $service_obj->Method(...); | ||
| 35 | $result->Att1->CompressionType | ||
| 36 | |||
| 37 | =head1 DESCRIPTION | ||
| 38 | |||
| 39 | Describes the input source of a transform job and the way the transform | ||
| 40 | job consumes it. | ||
| 41 | |||
| 42 | =head1 ATTRIBUTES | ||
| 43 | |||
| 44 | |||
| 45 | =head2 CompressionType => Str | ||
| 46 | |||
| 47 | If your transform data is compressed, specify the compression type. | ||
| 48 | Amazon SageMaker automatically decompresses the data for the transform | ||
| 49 | job accordingly. The default value is C<None>. | ||
| 50 | |||
| 51 | |||
| 52 | =head2 ContentType => Str | ||
| 53 | |||
| 54 | The multipurpose internet mail extension (MIME) type of the data. | ||
| 55 | Amazon SageMaker uses the MIME type with each http call to transfer | ||
| 56 | data to the transform job. | ||
| 57 | |||
| 58 | |||
| 59 | =head2 B<REQUIRED> DataSource => L<Paws::SageMaker::TransformDataSource> | ||
| 60 | |||
| 61 | Describes the location of the channel data, which is, the S3 location | ||
| 62 | of the input data that the model can consume. | ||
| 63 | |||
| 64 | |||
| 65 | =head2 SplitType => Str | ||
| 66 | |||
| 67 | The method to use to split the transform job's data files into smaller | ||
| 68 | batches. Splitting is necessary when the total size of each object is | ||
| 69 | too large to fit in a single request. You can also use data splitting | ||
| 70 | to improve performance by processing multiple concurrent mini-batches. | ||
| 71 | The default value for C<SplitType> is C<None>, which indicates that | ||
| 72 | input data files are not split, and request payloads contain the entire | ||
| 73 | contents of an input object. Set the value of this parameter to C<Line> | ||
| 74 | to split records on a newline character boundary. C<SplitType> also | ||
| 75 | supports a number of record-oriented binary data formats. Currently, | ||
| 76 | the supported record formats are: | ||
| 77 | |||
| 78 | =over | ||
| 79 | |||
| 80 | =item * | ||
| 81 | |||
| 82 | RecordIO | ||
| 83 | |||
| 84 | =item * | ||
| 85 | |||
| 86 | TFRecord | ||
| 87 | |||
| 88 | =back | ||
| 89 | |||
| 90 | When splitting is enabled, the size of a mini-batch depends on the | ||
| 91 | values of the C<BatchStrategy> and C<MaxPayloadInMB> parameters. When | ||
| 92 | the value of C<BatchStrategy> is C<MultiRecord>, Amazon SageMaker sends | ||
| 93 | the maximum number of records in each request, up to the | ||
| 94 | C<MaxPayloadInMB> limit. If the value of C<BatchStrategy> is | ||
| 95 | C<SingleRecord>, Amazon SageMaker sends individual records in each | ||
| 96 | request. | ||
| 97 | |||
| 98 | Some data formats represent a record as a binary payload wrapped with | ||
| 99 | extra padding bytes. When splitting is applied to a binary data format, | ||
| 100 | padding is removed if the value of C<BatchStrategy> is set to | ||
| 101 | C<SingleRecord>. Padding is not removed if the value of | ||
| 102 | C<BatchStrategy> is set to C<MultiRecord>. | ||
| 103 | |||
| 104 | For more information about C<RecordIO>, see Create a Dataset Using | ||
| 105 | RecordIO (https://mxnet.apache.org/api/faq/recordio) in the MXNet | ||
| 106 | documentation. For more information about C<TFRecord>, see Consuming | ||
| 107 | TFRecord data | ||
| 108 | (https://www.tensorflow.org/guide/data#consuming_tfrecord_data) in the | ||
| 109 | TensorFlow documentation. | ||
| 110 | |||
| 111 | |||
| 112 | |||
| 113 | =head1 SEE ALSO | ||
| 114 | |||
| 115 | This class forms part of L<Paws>, describing an object used in L<Paws::SageMaker> | ||
| 116 | |||
| 117 | =head1 BUGS and CONTRIBUTIONS | ||
| 118 | |||
| 119 | The source code is located here: L<https://github.com/pplu/aws-sdk-perl> | ||
| 120 | |||
| 121 | Please report bugs to: L<https://github.com/pplu/aws-sdk-perl/issues> | ||
| 122 | |||
| 123 | =cut | ||
| 124 |