vector-python-sdk VS caer

I don't want to dispel your ideas about Vector because I like that a robot can elicit such human fondness and responses, but we know for a fact they can't recognise each other. The neural net model for that is simply not programmed in their computer vision systems, this is easily verified by running the SDK with annotated viewer from the examples/tutorials where you can see exactly what Vector recognises. You will see it recognises faces (and things that vaguely look like faces which is why it will stare at random stuff sometimes), the charger, and the cube. It has no way to amend these vision models, though some people have done fun work to fix that using the SDK. But I can understand how people can pick up on the idea that they can "see" each other.

Good luck, here's a link to the script.

There is this Vector SDK that was written by Anki before they went under and there are a lot of forks on that repo on Github. Check here https://developer.anki.com/vector/docs for the documentation and here for the repo and the forks of that repo https://github.com/anki/vector-python-sdk. I've seen for myself different people have implemented different machine learning program on their Vectors (however, not that the program is not permanent as in once you stop the execution of your code, Vector will return to its normal self). Here is one example of that https://www.learnwitharobot.com/p/use-my-vector-sdk-to-allow-your-vector?s=r. This guy managed to train a neural network to recognize other Vectors. About its boundary, Vector does have a smart phone level CPU.

As far as I know, only Vector is programmable with the SDK in python.

I think it may have. If you look at line 10 of https://github.com/jasmcaus/caer/blob/master/configs.ini, you’ll see that caer has numpy and opencv-contrib-python dependencies that get referenced in its setup.py. If I recall correctly, pip on the nano doesn’t pick up the default numpy and opencv-python system installs, so when you go to install something like caer that has them as dependencies, it will install new copies except the wheel files that it grabs are incompatible. The solution I have found to work is to run something similar to the command above: “pip3 install —no-binary caer —no-binary numpy—no-binary opencv-contrib-python —no-binary typing-extensions —no-binary mypy —force-reinstall caer”. Some of those —no-binary options may not be necessary but they’ll at least ensure pip grabs the source for each of the dependencies and rebuilds it locally rather than using an imcompatible version. This command will take awhile! But you only should have to do it once.