Mapillary Vistas Dataloader

meetps / pytorch-semseg

Semantic Segmentation Architectures Implemented in PyTorch

https://meetshah.dev/semantic-segmentation/deep-learning/pytorch/visdom/2017/06/01/semantic-segmentation-over-the-years.html

MIT License

3.38k stars 799 forks source link

Mapillary Vistas Dataloader #55

Open lucasbrynte opened 6 years ago

lucasbrynte commented 6 years ago

Implement dataloader for the Mapillary Vistas dataset: https://blog.mapillary.com/product/2017/05/03/mapillary-vistas-dataset.html

lucasbrynte commented 6 years ago

@meetshah1995 I will take a look into this, but have som doubts. Taking a look at the existing dataloaders, there seems to be a mix of using either torch.utils.data.Dataset or torch.utils.data.DataLoader.

Which should I go for?
Any other considerations when implementing a dataloader?

meetps commented 6 years ago

data.Dataset is just a class representing a Dataset. and data.DataLoader combines a dataset and a sampler, and provides single- or multi-process iterators over the dataset. More here

You could simply use the cityscapes loader as a starting point and modify that for Mapillary (Cool dataset btw ! ).