Missing files after training in Python API

[noamco23-meet]

Hi, very new to this package so excuse me if this is a simple question.

I'm trying to run through the documentation to create a script that will take a training slp file, train it, and then infer poses on a raw video. This is supposed to be covered in the interactive and resumable training and interactive and realtime inference pages provided in the documentation if I'm not mistaken. I've implemented the code suggested in the first link using atrain.pkg.slp file provided in the documentation. The trainer finishes the training and returns this folder:

The console also returns this error:

My main question regards the inference stage, where the documentation states that "a raw video and a set of trained top-down ID models" are required. I have the raw video, but the documentation calls for a centroid_model.zip and a centered_instance_id_model.zip which I do not have. I tried following the links provided in the example:

I'm confused on as to where I'm supposed to get these zip files considering that the documentation states "These zip files just have the contents of standard SLEAP model folders that are generated during training" but at the end of the training I only got one model folder. I realised that one of the zip files included "multi class topdown" in the name, while my training name was "baseline top down" so maybe I'm using the wrong kind of training (?) but I'm only getting this from the documentation so I assume one part does connect to the other. I'm also not sure if the tensorflow error has anything to do with this issue considering that the documentation in the first link suggests that only these two things are saved:
INFO:sleap.nn.evals:Saved predictions: models/baseline_model.topdown/labels_pr.train.slp INFO:sleap.nn.evals:Saved metrics: models/baseline_model.topdown/metrics.train.npz

And I get these lines when I train as well.

Any help is greatly appreciated. Thanks in advance :)

[noamco23-meet]

Turns out I was missing another segment. In topdown training, two trainings need to occur: one to identify each animal and another centroid training. I don't know why it didn't appear in the documentation.