Beast code in Giters

Kenta Izumi's repositories

Ongoing research training transformer models at scale

NOASSERTION000

Language:PythonMIT000

Language:PythonApache-2.0000

Language:Jupyter NotebookMIT1000

Understand and test language model architectures on synthetic tasks.

000

Simple and efficient pytorch-native transformer text generation in <1000 LOC of python.

BSD-3-Clause000

Apache-2.0000

Language:Python000

Language:TypeScriptMIT000

JAX implementation of the Llama 2 model

CC0-1.0000

Large language models (LLMs) made easy, EasyLM is a one stop solution for pre-training, finetuning, evaluating and serving LLMs in JAX/Flax.

Apache-2.0100

Flax is a neural network library for JAX that is designed for flexibility.

Language:PythonApache-2.0000

Model parallel transformers in JAX and Haiku

Apache-2.0000