Page History

A utility was added to neural analytics, which allows you to create GPU neural network caches without using Axxon PSIM. The presence of cache speeds up the initialization and optimizes video memory consumption.

Creating GPU neural network caches:

1. Command example:
   

   Code Block
   C:\Program Files (x86)\Axxon PSIM\Modules64\caffewrapper\NeuroAnalyticsGpuCacheGenerator.exe -g 0 -p "<System disk>\<Directory of neural network location>\Neural_network_name.ann"

   where the -g parameter determines the sequence number of a graphics card according to nvidia-smi, and the -p parameter allows you to select the required neural network to create a cache. 
   To create a cache for multiple neural networks, list the paths to the selected neural networks, separated by a space.
   Command example:
   

   Code Block
   C:\Program Files (x86)\Axxon PSIM\Modules64\caffewrapper\NeuroAnalyticsGpuCacheGenerator.exe -g 0 -p "<System disk>\<Directory of neural network location>\Neural_network_name.ann" "C:\Program Files (x86)\Axxon PSIM\Modules64\caffewrapper\Models\poseEstimator_original.ann"

2. -v is a parameter to output the procedure log to the console during cache generation.
   Command example to automatically create caches of four neural networks included in neural analytics with log output:
   

   Code Block
   C:\Program Files (x86)\Axxon PSIM\Modules64\caffewrapper\NeuroAnalyticsGpuCacheGenerator.exe -g 0 -v

   Command example:
   

   Code Block
   C:\Program Files (x86)\Axxon PSIM\Modules64\caffewrapper\NeuroAnalyticsGpuCacheGenerator.exe -g 0 -p "<System disk>\<Directory of neural network location>\Neural_network_name.ann" -v

3. --int8=1 is a parameter to create a quantized version of the cache for those neural networks for which quantization is available. By default, the --int8=0 parameter is disabled.
   Command example:
   

   Code Block
   C:\Program Files (x86)\Axxon PSIM\Modules64\caffewrapper\NeuroAnalyticsGpuCacheGenerator.exe -g 0 --int8=1

   Note
   title Attention!
   The neural networks for which the quantization mode is available are included in neural analytics together with the *.info file.

   The neural networks for which the quantization mode is available:
   

   • GeneralNMCar_v1.0GPU_onnx.ann—vehicles.
   • GeneralNMHuman_v1.0GPU_onnx.ann—human.
   • GeneralNMHumanTopView_v0.8GPU_onnx.ann—human, top view

It can take several minutes to launch neural analytics algorithms on NVIDIA GPU after Server restart. Meanwhile, the neural models are optimized for the current GPU type.

You can use the caching function to ensure that this operation is performed only once. Caching saves the optimization results on the hard drive and uses it for the subsequent analytics runs.

When you install DetectorPack, the GPU_CACHE_DIR system variable is automatically created. Its value specifies the path to the folder with optimization results C:\Program Files (x86)\AxxonPSIM\Modules64\caffewrapper\Networks\cached. 

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

...

...

1. • .