SSD mobilenet_v1 training in transfer learning toolkit

Accelerated Computing Intelligent Video Analytics TAO Toolkit
1

In transfer learning toolkit (TLTK)

I can see under object detection support for following 3 meta-architectures.

  1. DetectNet V2
  2. SSD
  3. Faster RCNN

and everything is fine tuned with resnet18 (pretrained) in sample notebooks available in TLTK.

I want to train the SSD + mobilenet_v1

Q1) How can I train SSD+mobilenet_v1 with TLTK?
Q2) if yes for Q1, Which config files to use to do that?

2

Hi rohit167,
Currently, SSD only supports resnet10 and resnet18 backbones.
Next TLT release will support mobilenet_v1 backbone.