Xinyang (Young) Geng

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

scalax. A simple library for scaling up JAX programs

CQL. Conservative Q Learning on top of SAC

koala_data_pipeline. The data processing pipeline for the Koala chatbot language model

m3ae_public. Multimodal Masked Autoencoders (M3AE): A JAX/Flax Implementation

JaxCQL. Conservative Q learning in Jax

mlxu. Machine Learning eXperiment Utilities

mintext. Minimal but scalable implementation of large language models in JAX

tpu_pod_commander. TPU pod commander is a package for managing and launching jobs on Google Cloud TPU pods.

gps_superball_public. GPS codebase for NASA Superball tensegrity robot

SimpleSAC. A simple and easy to use implementation of the soft actor-critic algorithm.

MLClass. Coursera machine learning class

promoter_design. Python

cs184_final. CS184 final project: a point cloud to mesh generator

parameterized_model. TensorFlow parameterized model library

mmlu_easylm. MMLU evaluation for language models served with EasyLM

ResumeAnalytics. StatNews Resume Project

brc_cql_example. Python

cuda-training-series. Training materials associated with NVIDIA's CUDA Training Series (www.olcf.ornl.gov/cuda-training-series/)

ml_cheatsheet. Cheatsheet for CS189/289 at UC Berkeley

UVaClient. Command Line Client for UVa Online Judge

d4rl. A benchmark for offline reinforcement learning.

ShuffleX. A smart music player with machine learning based shuffle algorithm.

pytorch-maml-q-rl. Python

cs9e. CS 9E homework

acd_archive. Automatically archive to Amazon cloud drive

dopamine. Dopamine is a research framework for fast prototyping of reinforcement learning algorithms.

iColorTF. iColor implementation in TensorFlow

EvoLib. A framework in Java for genetic algorithm development