search

facebookresearch / ov-seg

This is the official PyTorch implementation of the paper Open-Vocabulary Semantic Segmentation with Mask-adapted CLIP.
Other
676 stars 61 forks source link

d2.engine.defaults]: No evaluator found. Use `DefaultTrainer.test(evaluators=)`, or implement its `build_evaluator` method. #13

Closed Ako-r closed 1 year ago

Ako-r commented 1 year ago

And how to evaluate the model on cityscapes?

Jeff-LiangF commented 1 year ago

Hi @Ako-r ,

We don't have the evaluation codes of cityscapes but I believe our model is able to do so.

Some hints about how to write the cityscapes evaluator: 1. register the cityscapes dataset, like pascal context, you may need to define the class names, etc. 2. migrating the data loading scripts of cityscapes from d2.

Btw, if you successfully evaluate on cityscapes, feel free to create a merge request. :)