Module: tfx_bsl.public.tfxio

Module level imports for tfx_bsl.public.tfxio.

TFXIO defines a common in-memory data representation shared by all TFX libraries and components, as well as an I/O abstraction layer to produce such representations. See the RFC for details: https://github.com/tensorflow/community/blob/master/rfcs/20191017-tfx-standardized-inputs.md

Classes

class BeamRecordCsvTFXIO: TFXIO implementation for CSV records in pcoll[bytes].

class CsvTFXIO: TFXIO implementation for CSV.

class RecordBatchToExamplesEncoder: Encodes pa.RecordBatch as a list of serialized tf.Examples.

class RecordBatchesOptions: Options for TFXIO's RecordBatches.

class TFExampleBeamRecord: TFXIO implementation for serialized tf.Examples in pcoll[bytes].

class TFExampleRecord: TFXIO implementation for tf.Example on TFRecord.

class TFGraphRecordDecoder: Base class for decoders that turns a list of bytes to (composite) tensors.

class TFSequenceExampleBeamRecord: TFXIO implementation for serialized tf.SequenceExamples in pcoll[bytes].

class TFSequenceExampleRecord: TFXIO implementation for tf.SequenceExample on TFRecord.

class TFXIO: Abstract basic class of all TFXIO API implementations.

class TensorAdapter: A TensorAdapter converts a RecordBatch to a collection of TF Tensors.

class TensorAdapterConfig: Config to a TensorAdapter.

class TensorFlowDatasetOptions: Options for TFXIO's TensorFlowDataset.

Type Aliases

TensorRepresentations