TensorFlow’s Experimental NumPy Interface Brings Familiarity to Deep Learning

TensorFlow now supports a subset of NumPy via tf.experimental.numpy, enabling GPU-accelerated NumPy operations. This guide walks through setup, ND arrays, type promotion, broadcasting, indexing, model creation, and NumPy interoperability. While efficient, intermixing may trigger memory copies, so use with care.

Content Overview

• Overview
• Setup
• Enabling NumPy behavior
• TensorFlow NumPy ND array
• Type promotion
• Broadcasting
• Indexing
• Example Model
• TensorFlow NumPy and NumPy
• NumPy interoperability
• Buffer copies
• Operator precedence

Overview

TensorFlow implements a subset of the NumPy API, available as tf.experimental.numpy. This allows running NumPy code, accelerated by TensorFlow, while also allowing access to all of TensorFlow's APIs.

Setup

import matplotlib.pyplot as plt
import numpy as np
import tensorflow as tf
import tensorflow.experimental.numpy as tnp
import timeit

print("Using TensorFlow version %s" % tf.__version__)

2024-08-15 01:31:55.452313: E external/local_xla/xla/stream_executor/cuda/cuda_fft ...