Dataset API (tf.data) Machine Learning; Data Engineering
noun phrase
Definition: A high-level TensorFlow API for building machine-learning input pipelines. The tf.data API introduces the tf.data.Dataset abstraction, which represents a potentially large sequence of elements, where each element consists of one or more tensor components; it supports reading data from sources and applying transformations such as mapping, batching, shuffling, and prefetching for model training and inference [TensorFlow Documentation].
Example in context: “TensorFlow, a popular DL framework, provides the tf.data API for users to build input data pipelines from the primitive dataset operators…” [Han et al. 2025].
Synonyms: tf.data API; TensorFlow Dataset API
Related terms:tf.data.Dataset; input pipeline; data preprocessing pipeline; data loader; prefetching