Sign In

Deep Learning Toolkit (System) by Ngene - Toolkit for LabVIEW Download

Dep Learning Toolkit for LabVIEW

* 1 ↓14
 logo
Version1.4.0.57
ReleasedJun 14, 2018
Publisher Ngene
LicenseNgene Custom
LabVIEW VersionLabVIEW>=0
Operating System Windows
Project links Homepage  

Description

Empower LabVIEW with Deep Learning
Deep Learning Toolkit for LabVIEW is a high-level API providing possibilities to build, configure, visualize, train, analyze and deploy Deep Learning based systems. The toolkit is mainly implemented in LabVIEW and does not have any outer dependencies, which simplifies the installation, development, deployment and distribution of toolkit based applications and systems (particularly, can be easily deployed on NI's Real Time targets).

Note: This installer requires LabVIEW and VI Package Manager to be run with Administrator privileges.

Main Features:
Create, configure, train and deploy Deep Neural Networks in LabVIEW
Visualize network topology and display common metrics (memory footprint, FLOPs)
API to debug and analyze networks
Save trained networks (configuration and weights) and load for deployment
NI's Real Time target support for deployment
Ready to run real world examples

Supported Layers:
Input (1D, 3D)
Augmentation - Implemented in Input3D
Fully Connected - FC
Convolutional - Conv3D
Batch Normalization
Activation
Pooling (MaxPool, AvgPool)
DropOut (1D, 3D)
SoftMax
Region (for YOLO v2)

Activation types:
Sigmoid
Hyperbolic Tangent
ReLU
Leaky ReLU

Solver (Optimization Algorithm):
Stochastic Gradient Descend (SGD) based Backpropagation algorithm with Momentum

Loss Functions:
MSE - Mean Squared Error
Cross Entropy (LogLoss)

Examples:
Examples are available to demonstrate the use of the toolkit on:
1. MNIST_Classifier_MLP(Train).vi - training the deep neural network for image classification task in handwritten digit recognition problem (based on MNIST database) using MLP (Multilayer Perceptron) architecture
2. MNIST_Classifier_CNN(Train).vi - training the deep neural network for image classification task in handwritten digit recognition problem using CNN (Convolutional Neural Network) architecture
3.MNIST_Classifier(Deploy).vi deploying pretrained network by automatically loading network configuration and weights files generated from the examples above.
4. YOLO_Object_Detection(Cam).vi automatically building and loading pretrained network for object detection based on YOLO (You Only Look Once) architecture.

Development License Activation Instructions
For Toolkit's developemnt license activation please call Third Party Add-on Activation Wizard (Help>>Add-Ons..) from 32 bit version of LabVIEW, as there is a known issue in 64 bit LabVIEW 2017 version

Release Notes

1.4.0.57 (Jun 14, 2018)

V1.4.0.57
1.Added augmentation functionality for 1-dimensional data implemented in Input1D layer.
2.Improved Conv3D and FC layer's performance.
3.Fixed augmentation errors in Input3D layer.

Install

Note, you must have the VIPM Desktop app installed for this button to work.

Versions

  Post an Idea   Post a Resource

Recent Posts