labelling of images

[Rworld-Ravi]

Hi
thanks for sharing the resources. in your readme you said to label the images using json tool but in your code you are loading pickle objects which are not there. am i missing something , its really confusing because in the your git there are some scripts like 'convert_old_format.py' & 'lable_images.py' which are using different name scheme for reading images to make a pickle object.

Any help would be appreciated as i have invested a lot of time on this.
Thanks

[araffin]

Hi,

Are you sure you have the latest version of the repo?

See https://github.com/sergionr2/RacingRobot/releases for a rough changelog.

Because we switched from using .pkl to .json file for the labels.

[Rworld-Ravi]

Hi. Sorry for coming so late. Had to postpone the twiking for some other tasks. Anyways, actually i used your provided image file for the pi, and was trying to train the code for my arena. But I think i will have to install everything from scratch to play with the latest repo. Right??

[araffin]

If you have the image, you can update to the latest version using git pull in the repo folder.

Anyway, i would recommend you doing the labelling and training on your computer, not the raspberry pi (which has limited computing capacity)

But I think i will have to install everything from scratch to play with the latest repo.
I would say that most of dependencies are already installed in the image, you can try to update the repo and if something is missing, update the dependencies (that is easier that installing everything from scratch)

I am working on simplifying dependencies by having a docker image (but that's not ready yet).