Closed mauritsreitsma closed 1 year ago
Looks like great progress. Let us know how things develop on the keypoint detection front. Would make sense to start a log describing the experiments you have run and what the results were. Then slowly, we should be able to start narrowing down the issue and eventually fix it.
One big commit, that should probably be multiple but I kinda missed commiting each step individually. Hope that is not a too big problem.
This is about making the training modular such that I can run training on a different machine and then validate on images locally with the trained model. Also adds visualization of the detected key points to the images.