Oneflow-Inc/oneflow Alternatives

choose when you want the largest ecosystem, most active community, and mature tooling in deep learning

Tensors and Dynamic neural networks in Python with strong GPU acceleration

choose when you want a library from Apple for training deep neural networks, though it is less mature than OneFlow

CoreNet: A library for training deep neural networks

choose when you need a mature production ecosystem with TPU support and extensive deployment tools

An Open Source Machine Learning Framework for Everyone

choose when you need a standalone C++ ML library without a Python binding layer

A C++ standalone library for machine learning

choose when you need a deep learning framework that integrates natively into JVM applications

An Engine-Agnostic Deep Learning Framework in Java

choose when you already use JAX and want a simple neural network module library

JAX-based neural network library

choose if you program in Nim and need a tensor/DL library for CPU, GPU, or embedded targets

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

choose when you need to train or deploy models directly in the browser or Node.js

A WebGL accelerated JavaScript library for training and deploying ML models.

choose when you want a pure Rust deep learning framework with high portability

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

choose when you work in Julia and want a differentiable programming ecosystem

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

choose when you need compile-time shape checking in Rust, but be aware it has not received updates in over two years

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

choose only if you need a historical Chinese deep learning framework, but it is largely unmaintained and stale

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

choose only for a stale research prototype of distributed GPU orchestration, not production-ready

Fault-tolerant, highly scalable GPU orchestration, and a machine learning framework designed for training models with billions to trillions of parameters