Search Results for author:
Found 4 papers, 2 papers with code
DataStates-LLM: Lazy Asynchronous Checkpointing for Large Language Models
LLMs have seen rapid adoption in all domains.
Understanding The Effectiveness of Lossy Compression in Machine Learning Training Sets
We designed a systematic methodology for evaluating data reduction techniques for ML/AI, and we use it to perform a very comprehensive evaluation with 17 data reduction methods on 7 ML/AI applications to show modern lossy compression methods can achieve a 50-100x compression ratio improvement for a 1% or less loss in quality.
Understanding Patterns of Deep Learning ModelEvolution in Network Architecture Search
We describe how evolutionary patterns appear in distributed settings and opportunities for caching and improved scheduling.
cuSZ: An Efficient GPU-Based Error-Bounded Lossy Compression Framework for Scientific Data
To the best of our knowledge, cuSZ is the first error-bounded lossy compressor on GPUs for scientific data.
Distributed, Parallel, and Cluster Computing
