Boris Ginsburg

Boris Ginsburg is an applied research scientist at NVIDIA. His primary research is fast and scalable algorithms for training of neural networks.
Avatar photo

Posts by Boris Ginsburg

Conversational AI

Develop Smaller Speech Recognition Models with the NVIDIA NeMo Framework

As computers and other personal devices have become increasingly prevalent, interest in conversational AI has grown due to its multitude of potential... 7 MIN READ
Conversational AI

Pretraining BERT with Layer-wise Adaptive Learning Rates

Training with larger batches is a straightforward way to scale training of deep neural networks to larger numbers of accelerators and reduce the training time.... 10 MIN READ
Conversational AI

Mixed Precision Training for NLP and Speech Recognition with OpenSeq2Seq

The success of neural networks thus far has been built on bigger datasets, better theoretical models, and reduced training time. Sequential models, in... 11 MIN READ