Please provide the following information when requesting support.
• Hardware :- RTX A4000
• Network Type :- OCD
• TLT Version (Please run “tlt info --verbose” and share “docker_tag” here) nvidia/tao/tao-toolkit: 5.0.0-tf2.11.0
• Training spec file(If have, please share here) default tao notabook of ocd
I have a ship container dataset. I want to train an OCD model for text detection in containers.
How to annotate and train in the Tao model
You can refer to https://docs.nvidia.com/tao/tao-toolkit/text/object_detection/ocd.html#preparing-the-dataset and https://github.com/NVIDIA/tao_tutorials/tree/main/notebooks/tao_launcher_starter_kit/ocdnet.
hi,
morganh I’ve already read it as well as tried it.
cat ICDAR2015/test/gt/gt_img_14.txt
268,82,335,93,332,164,267,164,the
344,94,433,112,427,159,336,163,Future
208,191,374,184,371,213,208,241,Communications
370,176,420,176,416,204,373,213,###
1,57,261,76,261,187,0,190,venting
1,208,203,200,203,241,3,294,intelligence.
How did annotated this format? in my ship container images
Do you mean how to generate a label file?
Actually the format follows ICDAR2015 as below mentioned in https://rrc.cvc.uab.es/?ch=4&com=tasks
x1, y1, x2, y2, x3, y3, x4, y4, transcription
yes how to generate label file for custom datasets on x1, y1, x2, y2, x3, y3, x4, y4, transcription this format ?
To generate label file, firstly you need to label the object. You can use some label tools. For example, “labelme”.
Then draw bbox and save the coordinates and the transcription. The labelme will save to a json file, then you can do some conversion to change to expected format.
ok i will try it
Thanks issue has been resolve
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.