Keras

Health v2 · maintenance only

Adoption not part of the score

64373 stars · 19798 forks observed · 2026-10-19

What it is AI-extracted, prompt v1, taxonomy v1, 2026-09-21, confidence 0.99

Keras is a high‑level, multi‑backend deep learning library for Python. It provides a user‑friendly API that works with TensorFlow, JAX, PyTorch and OpenVINO for both training and inference.

Use cases

• train an image classification model with Keras on TensorFlow or PyTorch
• build a natural‑language processing model using Keras and JAX
• quickly prototype deep learning experiments with a concise, readable API
• run inference on pretrained models from KerasHub (e.g., Llama, Stable Diffusion)
• scale training from a laptop to multi‑GPU or TPU clusters

When to choose

• You need a high‑level, easy‑to‑debug API for rapid model development
• You want the flexibility to switch between TensorFlow, JAX, PyTorch, or OpenVINO backends
• You prefer a Python‑centric ecosystem with extensive documentation and community support

When to avoid

• You require low‑level custom kernel development or non‑Python language bindings
• Your project depends on a framework not supported as a Keras backend
• You need a minimal runtime footprint for embedded or edge devices without Python

Facets

library · maturity active

deep-learning machine-learning artificial-intelligence deep-learning machine-learning data-science python windows multi-backend keras3 tensorflow jax pytorch openvino keras-hub linux macos

Member repositories

For agents

markdown · JSON · MCP: product_card(name="keras-team/keras")