Name: Sam Ade Jacobs
Type: User
Company: @LLNL @microsoft
Bio: - PhD Computer Science, Texas A&M Uni
- Computer Scientist at Microsoft
- R&D at the intersection of HPC and AI (aka large scale learning)
Location: San Francisco Bay Area
Sam Ade Jacobs's Projects
Caffe: a fast open framework for deep learning.
:zap: Caffe real-time data augmentation on-the-fly!!
ECP-CANDLE Benchmarks
Differentiable architecture search for convolutional and recurrent networks
Deep Learning papers reading roadmap for anyone who are eager to learn this amazing tech!
DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective.
Example models using DeepSpeed
Basic GAN template for the Exalearn project: code for training a DCGAN on slices of N-body dark matter simulations and performing simple analysis tasks.
TF2.0 version of ExaGAN, with experimental spectral constraints and regularization
Fast and memory-efficient exact attention
Cyclical consistency between forward and inverse surrogates in Inertial Confinement Fusion (ICF)
Livermore Big Artificial Neural Network Toolkit
Ongoing research training transformer language models at scale, including: BERT & GPT-2
Molecular Sets (MOSES): A Benchmarking Platform for Molecular Generation Models
Blogging
Trax — your path to advanced deep learning
Explainable Neural Architecture Search (XNAS)