YOLO Tao train and validate on several datasets located in different places

[chijco]

Hi. I am training on YOLO Tao in Jupyter notebook
And I have a following problem. I have many datasets and for each training experimant I want to make some combination of those dataset folders which are located in different places on my harddrive. What we were doing untill now sounds to me not as the best practice: for each experiment we were copying all the needed images from all the datasets into the Docker defined
workspce/tao_experimants/data
This practice very soon left me without space on my harddrive, since most of the experiments use overlapping datasets
Thus:
I want to train and validate while:

1. I have several training and validation directories
2. And as it follows from the mentioned above I want to train and validate while datasets directories are are outside of what was defined as a % env docker local_project_dir =which is actually called workspce/tao_experimants/data
   Is there a way to set “real” path to several datasets, and if yes how and where?

Thank you

[Morganh]

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

@chijco

Is there a way to set “real” path to several datasets, and if yes how and where?

Yes, you can check the ~/.tao_mounts.json file. It is mapping your local directories into the docker.