tensorflow/tfjs Alternatives

when you need a Python-based DL framework with dynamic computation graphs and broader ecosystem than JS.

Tensors and Dynamic neural networks in Python with strong GPU acceleration

when you want a lightweight C++ ML library for performance-critical, non-web deployment.

A C++ standalone library for machine learning

when you want to build/deploy models in Java/JVM with engine-agnostic API.

An Engine-Agnostic Deep Learning Framework in Java

when you work in Julia and want a differentiable programming approach.

Relax! Flux is the ML library that doesn't make you tensor

when you use JAX and need a neural network library with a familiar object-oriented API.

JAX-based neural network library

when you need a Nim tensor library for CPU/GPU with metaprogramming.

A fast, ergonomic and portable tensor library in Nim with a deep learning focus for CPU, GPU and embedded devices via OpenMP, Cuda and OpenCL backends

when you need the full TensorFlow framework in Python/C++ for server/mobile, not browser.

An Open Source Machine Learning Framework for Everyone

when you prefer a Rust DL framework with high portability and performance.

Burn is a next generation tensor library and Deep Learning Framework that doesn't compromise on flexibility, efficiency and portability.

when you want a distributed DL framework with a different optimization approach (e.g., data parallelism).

OneFlow is a deep learning framework designed to be user-friendly, scalable and efficient.

when you need Apple's library for training large-scale models in Python.

CoreNet: A library for training deep neural networks

when you want a fast, scalable DL framework with automatic differentiation in C++/Python.

MegEngine 是一个快速、可拓展、易于使用且支持自动求导的深度学习框架

when you want a Rust DL library with compile-time tensor shape checking.

Deep learning in Rust, with shape checked tensors and neural networks

when you need a Rust ML library with WASM backends for browser/Node.

Powerful Powerful Machine Learning library with GPU, CPU and WASM backends

when you need a C# deep learning library with CUDA/OpenCL support.

An easy to use C# deep learning library with CUDA/OpenCL support

when you want the high-level Keras API directly in Python (on top of TensorFlow) rather than TF.js.

Deep Learning for humans

when you need to run Keras models in the browser with WebGL but prefer a lighter, older library (though unmaintained).

Run Keras models in the browser, with GPU support using WebGL

an older Node.js binding to native TensorFlow for model inference/training outside browser.

Node.js + TensorFlow