You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In own benchmark, using same models, Im struggling with general inference at 20-30% degradation regardless of model size. However I am doing mini-batch processing and would like to try the bulk batch methods as paper indicates.
Did you guys open source the benchmark code?
In own benchmark, using same models, Im struggling with general inference at 20-30% degradation regardless of model size. However I am doing mini-batch processing and would like to try the bulk batch methods as paper indicates.
Reference: https://phala.network/posts/confidential-computing-on-nvidia-h100-gpu-a-performance-benchmark-study
The text was updated successfully, but these errors were encountered: