search

wayveai / fiery

PyTorch code for the paper "FIERY: Future Instance Segmentation in Bird's-Eye view from Surround Monocular Cameras"
https://wayve.ai/blog/fiery-future-instance-prediction-birds-eye-view
MIT License
556 stars 85 forks source link

Using custom data #25

Open hemanth-v25 opened 2 years ago

hemanth-v25 commented 2 years ago

How do I use this model for my custom data? My model contains 3 frontal camera videos. It doesn't contain any map data.

anthonyhu commented 2 years ago

Hello,

The data loaded as a PyTorch Dataset (https://github.com/wayveai/fiery/blob/master/fiery/data.py) currently assumes that it comes from either NuScenes or Lyft Dataset. It is possible to create a new custom dataset and run Fiery on it - to do so I'd suggest you download the NuScenes "mini" dataset and get familiar with the inputs returned by FuturePredictionDataset and adapt your custom dataset accordingly.