Tlt-dataset-convert PROBLEMS

Hello everyone, I hope you are well, I am trying to run the mask detector example in DEEPSTREAM 5.1, but I get the following error:
bash: tlt-dataset-convert: command not found

link: GitHub - NVIDIA-AI-IOT/face-mask-detection: Face Mask Detection using NVIDIA Transfer Learning Toolkit (TLT) and DeepStream for COVID-19
Plataform: JETSON NANO
CUDA Version**: 10.2
DEEPSTREAM 5.1
Operating System + Version: UBUNTU 18.04
Python Version (if applicable): 3

I don’t know what the problem can be, I appreciate your help

Moving this topic from deepstream forum into TLT forum.

1 Like

@cbenglenok
Which TLT docker did you run? Do you use
docker pull nvcr.io/nvidia/tlt-streamanalytics:v2.0_py3
mentioned in the GitHub - NVIDIA-AI-IOT/face-mask-detection: Face Mask Detection using NVIDIA Transfer Learning Toolkit (TLT) and DeepStream for COVID-19 ?

Hi, I do not use docker, i follow this method: face-mask-detection/face-mask-detection.ipynb at master · NVIDIA-AI-IOT/face-mask-detection · GitHub
and i this part I have an error

can i use docker? because in the tutorial it only uses it for gpu not for jetson nano

The tlt-dataset-convert is a command in tlt-streamanalytics:v2.0_py3 docker. If not use it, the command will not be available.
The training is recommended to run in gpu.
The inference can be at gpu or at Jetson device.