Run Tao inside docker

Please provide the following information when requesting support.

• Hardware 3090
• Network Type (Yolo_v4)
• TAO
• How to reproduce the issue ? (This is for errors. Please share the command line and the detailed log here.)

Hi I am trying to run tao in a docker container (it will build, train, deploy, from a script so assuming tao is installed etc on the PC is not an option so building a docker container I think is the best way). I have looked at this forum post:

but running:

sudo docker run --gpus all -it -v /workspace/tlt-experiments/:/workspace/tlt-experiments -p 8888:8888 nvcr.io/nvidia/tao/tao-toolkit-pyt:v3.21.11-py3 /bin/bash

followed by:

yolo_v4 train

leads to command not found - how can I install this so it works - as I understand if I do pip install nvidia-tao in the container that will run a container within another container which is not what I want?

Thanks

Can you set correct docker name?
Refer to TAO Toolkit for CV | NVIDIA NGC
For yolov4, You can use nvcr.io/nvidia/tao/tao-toolkit-tf:v3.21.11-tf1.15.5-py3

Hello - thanks! on that one I get a hardware compatibility error but can at least see the yolov4 command

pycuda._driver.LogicError: cuInit failed: forward compatibility was attempted on non supported HW

There is no update from you for a period, assuming this is not an issue any more.
Hence we are closing this topic. If need further support, please open a new one.
Thanks

Could you share the detailed steps?
Which docker did you trigger?
And then how did you trigger tao docker?

More, can you share the "$ nvidia-smi " result of your local PC?

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