Is it possible to accelerate the training of Deep Learning Model?

Evgeniy · 发表于 2021-12-10 14:58:56

Hi, colleagues!

My question is - Is it possible to accelerate the training of Deep Learning Model.

My DL classifier model is Enhanced.hdl, my source dataset images has size of 1548x677.

My videocard used only 7% of all resources. Is it possible to use videocard by maximum?

Maybe is any settings for that.

vipzms · 发表于 2021-12-14 16:52:19

When the graphics card driver is installed, right-click the desktop or other methods to enter nvida control panel, and set the graphics card to focus on performance.

tojoin · 发表于 2022-1-4 15:33:20

set_dl_model_param_max_gpu_batch_size (DLModelHandle, 50)