I'm running Habana's models and I don't see the same level of performance as what is published on the GitHub and Developer site

Greg_S · June 30, 2021, 6:07am

Please refer to Model Optimization Guide for a list of optimizations. Common methods include tuning batchsize, using mixed precision instead of float32, and implementing data pipelines with prefetching to device. Also please make sure ops are landing on HPU instead of running on the Host CPU. Users can review this section of the Optimization guide for more detials.

Topic		Replies	Views
How do I ensure that all the operators in my model are running on the HPU, instead of the Host CPU? FAQ models	0	590	June 30, 2021
What is the best way to ensure that my model is taking advantage of BF16 and mixed mode training FAQ performance	0	579	June 30, 2021
What should peak utilization of a core be? Training	2	434	January 4, 2023
How to measure the TPC run time? FAQ performance	1	422	March 14, 2023
I'm trying to run my model and it does not compile, it crashes, what should I do? FAQ debug	0	596	June 30, 2021

I'm running Habana's models and I don't see the same level of performance as what is published on the GitHub and Developer site

Related Topics