apple/axlearn

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

2372 stars · 406 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

AXLearn is a Python deep learning library built on JAX and XLA for developing and training large-scale models, with an object-oriented configuration system for composing reusable model components. It scales to hundreds of billions of parameters across thousands of accelerators and supports NLP, vision, and speech workloads on public clouds.

Use cases

• train large language models on thousands of accelerators
• build deep learning models with a composable configuration system
• run large-scale training jobs on public clouds
• train speech recognition or computer vision models at scale
• integrate Flax or Hugging Face transformers models into a JAX training pipeline
• manage and deploy distributed training jobs and data

When to choose

• you need to train models with tens or hundreds of billions of parameters
• you want JAX/XLA-based training with high accelerator utilization
• you need cloud tooling for deploying and managing large training jobs
• you prefer an object-oriented, modular approach to model composition

When to avoid

• you need a stable API, since it is under active development and subject to change
• you want a lightweight library for small models or quick experiments
• you prefer PyTorch or TensorFlow ecosystems
• you lack access to large accelerator clusters

Facets

library · maturity active

deep-learning llm-training machine-learning cli deep-learning machine-learning large-language-models computer-vision speech-processing gpu-computing python cloud jax xla gspmd large-scale-training distributed-training model-training flax huggingface natural-language-processing gpu linux

Member repositories

For agents

markdown · JSON · MCP: product_card(name="apple/axlearn")