TensorFlow

TensorFlow is an end-to-end, open source machine learning platform for developers, researchers, data scientists, educators, and organizations building ML and deep learning applications.

Key capabilities include:

• Model building with tf.keras (Keras API) for high-level neural network development and rapid prototyping.
• Efficient numerical computation with tensors and computational graphs, with support for running workloads on CPUs, GPUs, and TPUs.
• Training and workflow tooling, including tf.data input pipelines for preprocessing and TensorBoard for model analysis and debugging.
• Production deployment across environments, including TensorFlow Serving (server inference), TensorFlow.js (browser and JavaScript runtimes), LiteRT/TensorFlow Lite (mobile, edge, and microcontrollers), and TFX for production ML pipelines and ongoing data/model quality checks.

Pricing: Free (open source).

Notable for its Google Brain origins (released in 2015), large open-source ecosystem, and breadth of deployment targets spanning servers, cloud, desktop, mobile, browsers, edge devices, and microcontrollers.