Tf.data API
The tf.data API is TensorFlow’s API for constructing input pipelines that read, transform, batch, and iterate over data efficiently.
Deep Learning with Python, Third Edition
TensorFlow's input पाइപ്പ്-line API for building efficient data preprocessing and loading pipelines.
- How it is used
- Builds input pipelines that supply preprocessed batches to model training without loading an entire large image dataset into memory.
- How deep it goes
- Defines Dataset and works through iteration, batching, mapping, interleaving, and prefetching; focuses on pipeline construction rather than implementation internals.
- Examples
- Uses batched NumPy vectors, COCO images, IMDb reviews, translation pairs, and sharded text for language-model pretraining.
- What you can do after
- Build it from scratch with tf.data.Dataset operations, or apply Keras directory loaders to feed datasets into model.fit().
Read how Deep Learning with Python, Third Edition explains tf.data API →
Deep Learning with R, Third Edition
The tf.data API is TensorFlow's API for constructing input pipelines that process and preprocess data efficiently.
- How it is used
- Builds input pipelines that load and preprocess training data without requiring large datasets to fit in memory.
- How deep it goes
- Works through dataset construction, mapping, batching, prefetching, caching, and interleaving, but does not explain tf.data internals.
- Examples
- Shows COCO image batches, IMDb bag-of-words reviews, English-to-Spanish sentence pairs, and interleaved text shards.
- What you can do after
- Build it from scratch with R/tfdatasets operations to stream, transform, batch, and prefetch model inputs.
Read how Deep Learning with R, Third Edition explains tf.data API →
Related concepts
- TensorFlow Datasetused by · TensorFlow Dataset is an iterable abstraction for representing, transforming, batching, and supplying sequences of data elements to machine-learning models.
- prefetchconfigured by · Prefetching is the proactive retrieval or preparation of anticipated data or resources before they are needed.
- TensorFlowused by · TensorFlow is an open-source software platform for numerical computation and machine learning, with tools for building, training, and deploying models.
- pipelineused by · A pipeline is an ordered or coordinated sequence of connected processing stages or tasks that passes data, work, or results toward an outcome.
- dataset splitconfigures · A dataset split is the division of a dataset into distinct subsets assigned different purposes, such as training, validation, or testing.
- tupleproduces · A tuple is a finite, ordered collection of elements treated as a single composite value or record.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.