Which of the following is not a way for the TensorFlow program to read data?

Question

Options

• APreload data
• BFeeding data
• CRead from the file
• DWrite a file format reader

• A
  8% (3)
• B
  5% (2)
• C
  14% (5)
• D
  73% (27)

Explanation

Option D is correct because TensorFlow does not require users to write a custom file format reader - it already has built-in support for reading standard file formats (like TFRecord, CSV, etc.), so writing one is not a standard data-input mechanism.

Options A, B, and C are all legitimate TensorFlow data input methods: Preloading data (A) means loading the entire dataset into memory (or into constants/variables) before training; Feeding data (B) refers to using feed_dict to supply data at runtime through placeholders; and Reading from a file (C) uses TensorFlow's file I/O pipeline (e.g., tf.data, queue-based readers) to stream data from disk.

Memory tip: Think of the three valid methods as "where the data lives" - already in memory (Preload), pushed in at runtime (Feed), or pulled from disk (File). "Writing a file format reader" is a software development task, not a data ingestion mechanism - if you're building a reader, you're not yet reading data.

No community discussion yet for this question.