Input format of Image Classification (TF2)

Accelerated Computing Intelligent Video Analytics DeepStream SDK
#1

Please provide complete information as applicable to your setup.

• Hardware Platform (Jetson / GPU)
GPU
• DeepStream Version
6.1.1
• JetPack Version (valid for Jetson only)
• TensorRT Version
• NVIDIA GPU Driver Version (valid for GPU only)
• Issue Type( questions, new requirements, bugs)
• How to reproduce the issue ? (This is for bugs. Including which sample app is using, the configuration files content, the command line used and other details for reproducing)
• Requirement details( This is for new requirement. Including the module name-for which plugin or for which sample application, the function description)

Hello,

Can you provide information regarding the input format of the newly released TF2 classification? Image Classification (TF2) — TAO Toolkit 4.0 documentation (nvidia.com)

TAO is unable to generate a DS config file for the new TF classification task and there is currently no information regarding the input format available.

net-scale-factor, offsets, model-color-format, etc…

(When will the TAO sample apps be updated for 4.0? A sample DS config file would be greatly appreciated)

Thanks!

/M

#4

@IvensaMDH
Please refer to Deploying to DeepStream for Classification TF1/TF2 — TAO Toolkit 4.0 documentation

net-scale-factor=1.0
offsets=103.939;116.779;123.68
model-color-format=1

1 Like
#5

@Morganh what about for re-identification? It seems like a bit of a treasure hunt to find sample configs on the new TAO4.0 models ;)

/M

#6

We will consider the deepstream sample for re-identification model.

1 Like
#7

Any updates on this?

Can you perhaps provide a brief description of the input format (net-scale-factor, offsets etc.) as well pre- and postprocessing guidelines?

Just a brief description would be helpful and we’ll figure out the rest ourselves!

Thanks,

/M

#8

@Fiona.Chen Can you check up on this? We would like to adopt the model in our pipelines but want to make sure that we feed the network with a correctly preprocessed input.

Thanks!

#9

@IvensaMDH I think the parameters have been provided in the previous post.

For the sample application, we have take it into our roadmap.

#10

My apologies for the confusion - I was looking for the parameters for the Re-Identification model: Re-Identification | NVIDIA NGC

Can you provide this as well?

THanks!

#11

net-scale-factor=0.01735207357279195
offsets=123.675;116.28;103.53
model-color-format=0

1 Like