Does the toolkit support GPU acceleration?

Yes, the training and inference can be accelerated on Nvidia GPUs starting from Compute Capability 2.0. Please refer to the link below for complete list of supported GPUs. https://en.wikipedia.org/wiki/CUDA#GPUs_supported

What type of layers are supported by the toolkit?

Version 3.1 toolkit supports the following layers:

1. Fully Connected (with BatchNormalization)
2. Convolutional (with BatchNormalization)
3. Pooling (Max and Average)
4. Upsampling
5. Dropout
6. Flatten
7. SoftMax
8. Region/Detection

Does the toolkit support recurrent neural networks such as LSTM or GRU?

No, the toolkit currently supports only feed forward types of networks. Recurrent neural networks will be supported in near future.

What type of hardware is supported by the toolkit?

The toolkit supports x86 based PC win Windows OS for training and inference, and NI Real-Time targets with Linux RT OS for inference only. With the FPGA Accelerator add-on the inference can be accelerated on specific targets with FPGAs.

Where can I find the documentation for the toolkit?

You can open a specific part of help documentation about specific VI or control by pressing show "Detailed help" from the Context Help. Alternatively you can open the help file from "Help>>Ngene>>Deep_Learning_Toolkit..." from the main menu.

Which version of LabVIEW is supported by the toolkit?

The latest version of the toolkit (v3.1) supports both 32-bit and 64-bit versions of LabVIEW starting from 2016.

Does the toolkit support object detection?

Yes, starting from the version 3.0 the toolkit supports training YOLO v2 based architectures for object detection. Reference example for simple geometrical object detection can be found in the examples list which comes with the installer.

How can I annotate images for training an object detection model?

You can use NNotate software, which comes handy when annotating images and videos for object detection.

Does the toolkit support evaluation?

Yes, you will get 15 days evaluation period after the installation.

How to move already activated license to another computer?

You will need to deactivate the toolkit on the PC where it is already activated and activate the same licnense on the new PC.

How to activate the toolkit?

Before activation you should have received activation keys: "License ID" and "Activation Password" which will be required during activation process.

1. Important: Save and close all currently open projects and VIs.
2. Important: Run LabVIEW with elevated privileges (As Administrator)
3. Open Third Party Add-on Activation Wizard from "Help>>Activate Add-ons.." menu from LabVEIW.
4. Select "Ngene Deep Learning Toolkit x.x.x" from the product list and press "Activate >>" button.
5. Select "Automatically activate through an internet connection" and press "Next >>" button.​
6. Provide the received License ID and Password in the correspoding fields and press "Activate >>" button.​
7. You should receive a message about sucessful activation.​
8. Press "Restart LabVIEW" button.