« blib/lib/Paws/SageMaker/CategoricalParameterRangeSpecification.pm
↑ Summary
blib/lib/Paws/SageMaker/ChannelSpecification.pm »
| 1 | # Generated by default/object.tt | ||
| 2 | package Paws::SageMaker::Channel; | ||
| 3 | use Moose; | ||
| 4 | has ChannelName => (is => 'ro', isa => 'Str', required => 1); | ||
| 5 | has CompressionType => (is => 'ro', isa => 'Str'); | ||
| 6 | has ContentType => (is => 'ro', isa => 'Str'); | ||
| 7 | has DataSource => (is => 'ro', isa => 'Paws::SageMaker::DataSource', required => 1); | ||
| 8 | has InputMode => (is => 'ro', isa => 'Str'); | ||
| 9 | has RecordWrapperType => (is => 'ro', isa => 'Str'); | ||
| 10 | has ShuffleConfig => (is => 'ro', isa => 'Paws::SageMaker::ShuffleConfig'); | ||
| 11 | |||
| 12 | 1; | ||
| 13 | |||
| 14 | ### main pod documentation begin ### | ||
| 15 | |||
| 16 | =head1 NAME | ||
| 17 | |||
| 18 | Paws::SageMaker::Channel | ||
| 19 | |||
| 20 | =head1 USAGE | ||
| 21 | |||
| 22 | This class represents one of two things: | ||
| 23 | |||
| 24 | =head3 Arguments in a call to a service | ||
| 25 | |||
| 26 | Use the attributes of this class as arguments to methods. You shouldn't make instances of this class. | ||
| 27 | Each attribute should be used as a named argument in the calls that expect this type of object. | ||
| 28 | |||
| 29 | As an example, if Att1 is expected to be a Paws::SageMaker::Channel object: | ||
| 30 | |||
| 31 | $service_obj->Method(Att1 => { ChannelName => $value, ..., ShuffleConfig => $value }); | ||
| 32 | |||
| 33 | =head3 Results returned from an API call | ||
| 34 | |||
| 35 | Use accessors for each attribute. If Att1 is expected to be an Paws::SageMaker::Channel object: | ||
| 36 | |||
| 37 | $result = $service_obj->Method(...); | ||
| 38 | $result->Att1->ChannelName | ||
| 39 | |||
| 40 | =head1 DESCRIPTION | ||
| 41 | |||
| 42 | A channel is a named input source that training algorithms can consume. | ||
| 43 | |||
| 44 | =head1 ATTRIBUTES | ||
| 45 | |||
| 46 | |||
| 47 | =head2 B<REQUIRED> ChannelName => Str | ||
| 48 | |||
| 49 | The name of the channel. | ||
| 50 | |||
| 51 | |||
| 52 | =head2 CompressionType => Str | ||
| 53 | |||
| 54 | If training data is compressed, the compression type. The default value | ||
| 55 | is C<None>. C<CompressionType> is used only in Pipe input mode. In File | ||
| 56 | mode, leave this field unset or set it to None. | ||
| 57 | |||
| 58 | |||
| 59 | =head2 ContentType => Str | ||
| 60 | |||
| 61 | The MIME type of the data. | ||
| 62 | |||
| 63 | |||
| 64 | =head2 B<REQUIRED> DataSource => L<Paws::SageMaker::DataSource> | ||
| 65 | |||
| 66 | The location of the channel data. | ||
| 67 | |||
| 68 | |||
| 69 | =head2 InputMode => Str | ||
| 70 | |||
| 71 | (Optional) The input mode to use for the data channel in a training | ||
| 72 | job. If you don't set a value for C<InputMode>, Amazon SageMaker uses | ||
| 73 | the value set for C<TrainingInputMode>. Use this parameter to override | ||
| 74 | the C<TrainingInputMode> setting in a AlgorithmSpecification request | ||
| 75 | when you have a channel that needs a different input mode from the | ||
| 76 | training job's general setting. To download the data from Amazon Simple | ||
| 77 | Storage Service (Amazon S3) to the provisioned ML storage volume, and | ||
| 78 | mount the directory to a Docker volume, use C<File> input mode. To | ||
| 79 | stream data directly from Amazon S3 to the container, choose C<Pipe> | ||
| 80 | input mode. | ||
| 81 | |||
| 82 | To use a model for incremental training, choose C<File> input model. | ||
| 83 | |||
| 84 | |||
| 85 | =head2 RecordWrapperType => Str | ||
| 86 | |||
| 87 | Specify RecordIO as the value when input data is in raw format but the | ||
| 88 | training algorithm requires the RecordIO format. In this case, Amazon | ||
| 89 | SageMaker wraps each individual S3 object in a RecordIO record. If the | ||
| 90 | input data is already in RecordIO format, you don't need to set this | ||
| 91 | attribute. For more information, see Create a Dataset Using RecordIO | ||
| 92 | (https://mxnet.apache.org/api/architecture/note_data_loading#data-format). | ||
| 93 | |||
| 94 | In File mode, leave this field unset or set it to None. | ||
| 95 | |||
| 96 | |||
| 97 | =head2 ShuffleConfig => L<Paws::SageMaker::ShuffleConfig> | ||
| 98 | |||
| 99 | A configuration for a shuffle option for input data in a channel. If | ||
| 100 | you use C<S3Prefix> for C<S3DataType>, this shuffles the results of the | ||
| 101 | S3 key prefix matches. If you use C<ManifestFile>, the order of the S3 | ||
| 102 | object references in the C<ManifestFile> is shuffled. If you use | ||
| 103 | C<AugmentedManifestFile>, the order of the JSON lines in the | ||
| 104 | C<AugmentedManifestFile> is shuffled. The shuffling order is determined | ||
| 105 | using the C<Seed> value. | ||
| 106 | |||
| 107 | For Pipe input mode, shuffling is done at the start of every epoch. | ||
| 108 | With large datasets this ensures that the order of the training data is | ||
| 109 | different for each epoch, it helps reduce bias and possible | ||
| 110 | overfitting. In a multi-node training job when ShuffleConfig is | ||
| 111 | combined with C<S3DataDistributionType> of C<ShardedByS3Key>, the data | ||
| 112 | is shuffled across nodes so that the content sent to a particular node | ||
| 113 | on the first epoch might be sent to a different node on the second | ||
| 114 | epoch. | ||
| 115 | |||
| 116 | |||
| 117 | |||
| 118 | =head1 SEE ALSO | ||
| 119 | |||
| 120 | This class forms part of L<Paws>, describing an object used in L<Paws::SageMaker> | ||
| 121 | |||
| 122 | =head1 BUGS and CONTRIBUTIONS | ||
| 123 | |||
| 124 | The source code is located here: L<https://github.com/pplu/aws-sdk-perl> | ||
| 125 | |||
| 126 | Please report bugs to: L<https://github.com/pplu/aws-sdk-perl/issues> | ||
| 127 | |||
| 128 | =cut | ||
| 129 |