What's new in TensorFlow 2.4

NumPy (Numerical Python) is the most popular Python library in the Data Science world. Countless other Python modules are based on NumPy. It provides a powerful API for working with arrays and several other operations, covering branches of mathematics like linear algebra and Fourier transform. This allows running NumPy code on GPU, accelerated by TensorFlow, while also allowing access to all of TensorFlow's APIs. Most TensorFlow models use the float32 data type; however, using lower precision data types, such as float16 can improve model performance by 3x on GPUs and 60% on TPUs.