Transfer Learning Toolkit (TLT) user survey

JavaScript isn't enabled in your browser, so this file can't be opened. Enable and reload.

https://developer.nvidia.com/transfer-learning-toolkit

What type of applications are you building Deep Learning models for? (Select all that applies)

Smart city

Smart retail

Automated optical inspection

Smart warehouse

Robotics

Preventive maintenance

Other:

What is the biggest challenge in training your own model?

Low model accuracy

Not enough labeled data

Complex training Frameworks

Takes long time to train

Low inference throughput

Other:

Clear selection

Are you currently using Transfer Learning Toolkit? *

Yes

What are the challenges that’s preventing you from training with TLT? (Select all that applies)

Model architecture (YOLO, MaskRCNN, UNET, other) not supported

Difficult to use

Lack of features

Accuracy issues

Lack of adequate Documentation

Lack of Developer tutorials

No issues

Other:

How do you currently deploy your model?

TensorFlow/PyTorch for training and deployment using DeepStream

TensorFlow/PyTorch for training and deployment using TensorRT

Train and deploy using TensorFlow/PyTorch

TLT with DeepStream

TLT with deployment using TensorRT but no DeepStream

Yet to deploy

Other:

Clear selection

Of the popular models not supported by TLT, which one do you want TLT to support in the future? Visit the product page for full list of supported models. https://developer.nvidia.com/transfer-learning-toolkit

MaskRCNN

YOLO V3

RetinaNet

UNET/DeepLab

Language models/Seq2Seq/BERT

Action recognition

Others

What features will you find useful in TLT?

Quantize Aware Training

Mix Precision training

Easy to use UI

Visualization using TensorBoard

Other:

Submit

Clear form

Never submit passwords through Google Forms.

This form was created inside of NVIDIA Corporation. Report Abuse

Forms