Daniel Bolya

Yeah your mistake was `'class_names': ('zao'),` should be `'class_names': ('zao',),` (just pointing that out in case you found a different mistake). The loss codes are as follows: ``` # Loss...

The numbers you care about there are the top one (0.745 = 74.5 in my eval.py) and the 2 ones under it (94.1 and 90.3). Since the maximum AP is...

Looks like whatever weights file you tried to load there was corrupt. Can you elaborate on what you were trying to do?

There may be an issue either with your training data or the model. YOLACT because of how it's designed can have an issue with overlap, but that's usually not that...

That would be a good test, though I will warn that we've tried VGG16+FPN before and got sub-par performance (though maybe that's to be expected). If you want to change...

Sounds good! If it ends up working well, I might just make an official MobileNet-v2 version for YOLACTv1.1. That seems like it would be nice to have.

Hmm with the mAP starting that low (< 1), it looks likes the behavior we get when we train from scratch. Are you using imagenet pretrained weights in your MobileNet...

Hmm, can you still do a sanity check and change `self.load_state_dict(state_dict, strict=False)` to `self.load_state_dict(state_dict, strict=True)` and see if there's anything unexpected is missing? I think even with one anchor per...

That's a JIT error, maybe try disabling JIT? Run it with ```PYTORCH_JIT=0 python train.py ```

¯\\\_(ツ)\_/¯ Might want to check the code for tensorboard and debug from there. I'm not familiar, so I don't think I'd be able to give much help.