Versions Compared

Old Version 1

changes.mady.by.user Vadim Belyakov

Saved on

compared with

New Version 2

changes.mady.by.user Alena Kniazeva

Saved on

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

It may take several minutes to launch neuroanalytics algorithms on an NVIDIA GPU after Server reboot. During  this timeMeanwhile, the neuromodels are optimized for the current GPU type.

You can use the caching function to do ensure that this operation is performed only once. Cashed optimization models are saved  Caching saves the optimization results on the hard disk drive and used for next  runs of analytics.uses it for the subsequent analytics runs. 

To activate the caching, create the GPU_CACHE_DIR system variable  variable and specify in its value the existing folder as its value, where the optimization models result will be stored (see Appendix 10. Creating system variable).

The cache size depends on the number of neural networks used and their type, the minimum size is 70 MB.

Note
titleAttention!

This function works in beta mode for all detection tools which use neuroanalytics (see General information on Neural Analytics), except Face detection. To optimize Face detection operation using the GPU_CACHE_DIR system variable, you need to perform additional actions (see Optimization of Face detection on GPU).