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
557 stars 85 forks source link

The result of fiery static #32

Closed DFLyan closed 2 years ago

DFLyan commented 2 years ago

If I want to get the result of Fiery Static of Setting2 in Table I of your paper, should I use the config in "configs/single_timeframe.yml"? When I train the network using this config file from scratch, the IOU is 39.2 when I use the "evaluate.py". However, in the paper, the result is 35.8. Is there another parameter needed to be modified, when I want the network to be one frame as input and the segmentation result of the present frame as output?

anthonyhu commented 2 years ago

Hi,

If you want to reproduce the results of FIERY Static in the lift-splat-shoot evaluation setting, you should train the config https://github.com/wayveai/fiery/blob/master/fiery/configs/literature/static_lss_setting.yml

The configs/single_timeframe.yml config additionally masks vehicle labels that are not visible from the camera, which is why the IoU is larger. This masking was not present in the original lift-splat-shoot evaluation, so for a fairer comparison it was removed.

DFLyan commented 2 years ago

OK, thanks!