TAO Spec File Documentation

Accelerated Computing Intelligent Video Analytics TAO Toolkit
, ,
1

There are a ton of fields inside of the experiment training spec file. Is there any documentation on all of the available fields and their meanings? i.e. i see augmentation_config has field spatial_augmentation which has zoom_min and zoom_max – what are the units?

• Hardware (T4/V100/Xavier/Nano/etc): N/A
• Network Type: Peoplenet
• TLT Version (Please run “tlt info --verbose” and share “docker_tag” here): N/A
• Training spec file(If have, please share here):
detectnet_v2_retrain_resnet34_kitti.txt (6.2 KB)

• How to reproduce the issue ? (This is for errors. Please share the command line and the detailed log here.): N/A

2

Please refer to DetectNet_v2 - NVIDIA Docs. For zoom_in, it is the minimum zoom scale of the input image. It is a scale and it has not units.

A zooming ratio of 1.0 will not affect the image, while values higher than 1 will result in ‘zooming out’ (image gets rendered smaller than the canvas), and vice versa for values below 1.0.
The source code is in tao_tensorflow1_backend/nvidia_tao_tf1/core/processors/augment/spatial.py at 2ec95cbbe0d74d6a180ea6e989f64d2d97d97712 · NVIDIA/tao_tensorflow1_backend · GitHub. You can also refer to it.

1 Like
3

awesome thanks for linking the source code! Does an equivelant exist in the tao tf2 repo? I looked but wasnt able to find anything

4

No, detectnet_v2 exists in the tao tf1 repo. So tf2 repo does have it.

5

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.