Deep learning frequently asked questions

Find answers to common questions about deep learning.

• What license do I need for the deep learning tools?

• Do I have to install all the deep learning libraries to run the deep learning tools?

• I have other versions of deep learning libraries installed. Will they work with the current version of AllSource?

• What are the GPU requirements for running deep learning tools?

• I have an older GPU that is incompatible with the software, or I have low GPU memory. What are the requirements?

• How do you monitor how much GPU memory is being used?

• How can I speed up the inferencing tools?

• Why am I seeing conda or jupyter notebook not recognized as internal or external command when trying to install the libraries manually?

• What do I do if I see a conda broken package error, or a verification error when trying to install libraries manually?

• After training, how do I know how well my model has performed?

• After running inferencing tools, how do I know how well my model performed?

• What library versions are required for ArcGIS Pro 3.1?

What license do I need for the deep learning tools?

All deep learning geoprocessing tools and the Label Objects for Deep Learning pane, used for deep learning with imagery in 2D, require the ArcGIS Image Analyst extension. Some of the tools are also available with the ArcGIS Spatial Analyst extension.

The Classification (Deep Learning) toolset, used for classifying point cloud data, requires the ArcGIS 3D Analyst extension.

The interactive Object Detection tool for imagery in a 3D scene requires either an ArcGIS Pro Advanced license or the ArcGIS Image Analyst extension.

Do I have to install all the deep learning libraries to run the deep learning tools?

Yes, you need to follow the instructions listed in Install deep learning frameworks for ArcGIS.

I have other versions of deep learning libraries installed. Will they work with the current version of AllSource?

No, each version of AllSource requires specific versions of the deep learning libraries. You must uninstall existing packages and libraries and install the versions listed in the install instructions.

What are the GPU requirements for running deep learning tools?

The recommended VRAM for running training and inferencing deep learning tools in AllSource is 8GB. If you are only performing inferencing (detection or classification with a pretrained model), 4GB is the minimum required VRAM, but 8GB is recommended.

I have an older GPU that is incompatible with the software, or I have low GPU memory. What are the requirements?

If you do not have the required 4–8GB VRAM, you can run the tools on the CPU, though the processing time will be longer.

How do you monitor how much GPU memory is being used?

Use nvidia-smi, which is command line utility that is installed with your Nvidia drivers.

1. Open a Windows Command Prompt window.
2. Type nvidia-smi.
3. Press the Enter key.

   Note:

   If nvidia-smi is not found, you will need to change to the proper directory in the Command Prompt window before you execute the command. Use the Windows search bar to locate nvidia-smi.

Under the Memory-Usage section, you can see whether GPU memory is being used.

If you want to monitor the continuous usage of your GPU during the execution of the tools, you can run nvidia-smi -l 10. You can use this to determine what the batch size should be when running the deep learning tools . If you see there is some memory not being used, you can increase your batch size during execution. If you see that the memory usage is at its maximum and the tool fails, decreasing your batch size should help.

How can I speed up the inferencing tools?

If you are not already using the GPU, set the Processor Type to GPU in the Environment settings of the tool. You may also try increasing the batch size for optimal GPU utilization. If the batch size is too high, you may see a CUDA_OUT_MEMORY error, so you will need to experiment with the batch size to find the right size for your mode.

Why am I seeing conda or jupyter notebook not recognized as internal or external command when trying to install the libraries manually?

You may see the errors above if you are using the standard Windows command prompt instead of AllSource Python Command Prompt. You can access AllSource Python Command Prompt from the Start menu by searching for Python Command Prompt,or you can find it in the AllSource install location. AllSource Python Command Prompt allows you to access the standard tools and libraries that come with conda or jupyter.

What do I do if I see a conda broken package error, or a verification error when trying to install libraries manually?

Clean your local cache using conda clean –t.

After training, how do I know how well my model has performed?

The output folder for the trained model contains a file named model_metrics.html. This file contains information about your trained model such as the learning rate, training and validation loss, and the average precision score.

After running inferencing tools, how do I know how well my model performed?

There are various methods for verifying the results of your deep learning models. For more information, see Review results.

What library versions are required for ArcGIS Pro 3.1?

For the most recent list of libraries required, see Deep learning packages. The library versions required for previous versions of AllSource are listed in the manual installation guides (PDFs) for each version.