Abstract
Developing AI model is becoming more and more complex. Researchers need to manage
heterogeneous computing resources to run their training tasks. We propose a system to manage
all the heterogeneous computing resources and schedule AI training tasks. This system can
greatly reduce researchers’ effort on managing resources and training tasks. With this system
researchers can focus on developing AI models with lower cost and higher efficiency.
Creative Commons License
This work is licensed under a Creative Commons Attribution-Share Alike 4.0 License.
Recommended Citation
INC, HP, "A METHOD TO AUTOMATICALLY MANAGE AI TRAINING TASK ON HYBRID COMPUTING PLATFORM", Technical Disclosure Commons, (August 30, 2021)
https://www.tdcommons.org/dpubs_series/4555