This release contains dataset No. 1 to train and test terrain classifiers, slip regressors, and pre-trained ML models in Pytorch.
- datasets: this folder contains datasets for training terrain classifiers in
train
andvalid,
and training slip regressors inslip_models.
In thetest
folder, there are testing subsets used for motion planning environments. - trained_models: this folder contains trained ML models. As a terrain classifier, we adopt a simple Unet. As slip regressors, we use Gaussian process regressors implemented in gpytorch.
You can download datasets.zip
and trained_models.zip
and place them as ./datasets
and ./trained_models
after unzipping them, respectively.