Multiple data sources in dataset_config

[junghyun.hwang]

Please provide the following information when requesting support.

• Hardware (T4/V100/Xavier/Nano/etc)
• Network Type (Detectnet_v2/Faster_rcnn/Yolo_v4/LPRnet/Mask_rcnn/Classification/etc)
• TLT Version (Please run “tlt info --verbose” and share “docker_tag” here)
• Training spec file(If have, please share here)
• How to reproduce the issue ? (This is for errors. Please share the command line and the detailed log here.)

In the following link, there is an example Experiment Spec File.
One of the configs is the data_sources,
and I am wondering what might be the correct syntax for multiple data sources.

would the following be the right syntax?
train_data_sources:
- image_dir: /path/to/images/dataset_a
json_file: /path/to/annotations/dataset_a.json
- image_dir: /path/to/images/dataset_b
json_file: /path/to/annotations/dataset_b.json

Or does it not accept multiple data sources at all?

[Morganh]

@junghyun.hwang

In the following link,

Can you share the link? Thanks.

[junghyun.hwang]

Oh sorry, DeformableDETR - NVIDIA Docs.
It is in the ‘Creating an Experiment Spec File’ section.

[Morganh]

You can refer to below.

  train_data_sources:
    - image_dir: /path/to/1
      json_file: /path/to/1.json
    - image_dir: /path/to/2
      json_file: /path/to/2.json