hgpu.org » Applications » Computer science » Computer vision » Single Server Multi-GPU Training of ConvNets
Single Server Multi-GPU Training of ConvNets
Omry Yadan, Keith Adams, Yaniv Taigman, Marc’Aurelio Ranzato
Facebook AI Group
arXiv:1312.5853 [cs.LG], (20 Dec 2013)
BibTeX
@article{2013arXiv1312.5853Y,
author={Yadan}, O. and {Adams}, K. and {Taigman}, Y. and {Ranzato}, M.},
title={"{Single Server Multi-GPU Training of ConvNets}"},
journal={ArXiv e-prints},
archivePrefix={"arXiv"},
eprint={1312.5853},
primaryClass={"cs.LG"},
keywords={Computer Science – Learning, Computer Science – Neural and Evolutionary Computing},
year={2013},
month={dec},
adsurl={http://adsabs.harvard.edu/abs/2013arXiv1312.5853Y},
adsnote={Provided by the SAO/NASA Astrophysics Data System}
}
In this work we evaluate different approaches to parallelize computation of convolutional neural networks across several GPUs within the same server.
December 23, 2013 by hgpu
No votes yet.
Please wait...
Recent source codes
* * *
Most viewed papers (last 30 days)
- Data-efficient LLM Fine-tuning for Code Generation
- LithOS: An Operating System for Efficient Machine Learning on GPUs
- Large Language Model Powered C-to-CUDA Code Translation: A Novel Auto-Parallelization Framework
- MSCCL++: Rethinking GPU Communication Abstractions for Cutting-edge AI Applications
- GigaAPI for GPU Parallelization
- Scalability Evaluation of HPC Multi-GPU Training for ECG-based LLMs
- A Power-Efficient Scheduling Approach in a Cpu-Gpu Computing System by Thread-Based Parallel Programming
- DeepCompile: A Compiler-Driven Approach to Optimizing Distributed Deep Learning Training
- InteropUnityCUDA: A Tool for Interoperability Between Unity and CUDA
- GPU-centric Communication Schemes for HPC and ML Applications
* * *