Lack of modules and datasets to train a VAE model

[krank10]

I'm trying to train a VAE encoder. It seems that the collision_image_generator.py file lacks some modules:
`from verifiable_learning.DepthToLatent.datasets.depth_collision_image_dataset import IMAGES_PER_TF_RECORD, DepthCollisionImageDataset, collate_batch

from verifiable_learning.DepthToLatent.datasets.depth_semantic_image_dataset import IMAGES_PER_TF_RECORD, DepthSemanticImageDataset, collate_batch

from verifiable_learning.DepthToLatent.networks.VAE.vae import VAE`

further more, the dept image datasets are also unavailable.