Habitat Pre-Trained Models and Kernel Metadata

Geoffrey X. Yu, Yubo Gao, Pavel Golikov, Gennady PekhimenkoZenodo 2021 · DOI

Abstract

This archive contains the pre-trained models and kernel metadata (registers per thread) that are used by Habitat. Habitat is a tool that predicts a deep neural network's training iteration execution time on a given GPU. Habitat began as a research project in the EcoSystem group at the University of Toronto. The accompanying research paper will appear in the proceedings of USENIX ATC'21.
BibTeX
@inproceedings{yu2021habitat,
  author = {Geoffrey X. Yu and Yubo Gao and Pavel Golikov and Gennady Pekhimenko},
  title = {{Habitat Pre-Trained Models and Kernel Metadata}},
  booktitle = {Zenodo},
  year = {2021},
  doi = {10.5281/zenodo.4876277},
}

Loading related content...

Figure