liveBook · Manning

JaxSeq

As explained in Deep Learning with JAX

A JAX-based library built on Transformers for training large language models and supporting several pretrained model families.

JaxSeq is a JAX-based library built on Hugging Face’s Transformers library for training large language models. It currently supports GPT-2, GPT-J, T5, and OPT. section 12.1.2

JaxSeq is useful when you need to:

  • Train language models using JAX while building on Transformers. section 12.1.2
  • Work with GPT-2, GPT-J, T5, or OPT model families. section 12.1.2

The surrounding JAX LLM ecosystem changes quickly, so model coverage and project capabilities should be checked before committing to JaxSeq. section 12.1.2

How Deep Learning with JAX covers it

How deep it goes
Identifies its foundation, training role, and supported models, but does not detail APIs, workflows, or implementation internals.
Examples
Names GPT-2, GPT-J, T5, and OPT as supported model families.
What you can do after
Evaluate whether JaxSeq fits JAX-based LLM training involving GPT-2, GPT-J, T5, or OPT.
How it connects
Connects JAX with Hugging Face Transformers and the GPT-2, GPT-J, T5, and OPT model families.
Cover of Deep Learning with JAX
liveBookmanning.com
Read Deep Learning with JAX on liveBook Get the book at manning.com

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Deep Learning with JAX. The section links point to the parts of the book it draws on.