Tempo para computação para classificar pixels usando deep learning?

Question

Alguém já executou a classificação de pixels usando deep learning no arcgispro? Quanto tempo geralmente leva para computar/concluir a operação? Eu tenho conjuntos de dados de 5 GB e estou tentando realizar a classificação LULC usando o algoritmo unet. Sei que muitas coisas impactam o poder de computação, mas tenho um bom laptop com 16 GB de RAM e 6 GB de GPU, e já se passaram mais de 24 horas desde que comecei a operação, mas ainda está mostrando apenas 16% de progresso. Alguma ideia?

MartyRyan · Answer

Update: 3:51PM EST - I'm still running Classify Pixels Using Deep Learning on the same workstation - it's now at 73% - waiting till it ends, but this seems awfully slow. This is the second day in a row it progressed only 5% during my entire workday.

MartyRyan · Answer

I am using a Dell Precision Tower 7810. It has: Intel Xeon E5-2603 v4CPU (6 core at 1.7GHz) / NVIDIA Quadro M4000 (8GB) / Hard drives: (1) 256GB NVMe (OS) and (1) 1 TB Samsung SSD

My image is .5 meter / 3 band / 8 pixel unsigned raster, It is about 593 GB (uncompressed).

"Classify Pixels Using Deep Learning" is showing "Generate Raster" at 41%....I started it Friday 3/12 at 9:45 am. So I am at almost 72 hours now. Just my experience and since this is my first time doing this, I'm not sure if this is normal or if anything is wrong. I think it's running as it should, as my Task Manager shows "high" and "very high" useage.

VinayViswambharan1 · Answer

Can you bring up the task manager and see if your GPU is being used? What is the activity (percent use) of the CPU and GPU?