Sixkplus
Sixkplus
Similar results for me. I modified the network a little to use MobileNet and only got 0.482 mIoU on cityscapes. I'm very curious about the implementation in BiSeNet called Xception39,...
@18605973470 Hi! Anyway, after changing the parameters and some implementation details in this suit, Now I can achieve higher than 61% mIoU based on mobileNet. However, there is still a...
666 666 Thank the author for pointing out all the tricks.
Thanks! @hellochick I'll try to fix it! :D
Hey! @hellochick I found this problem is caused by the fact you forget to pass the parameter `img_mean` in mode configuration block to the function `_infer_preprocess` in `model.py`. I've fixed...
@hellochick BTW, Could you please tell me which model you use to get mIoU 67.3? I use the `train_30k` and get 64.2
@hellochick It turns out that during `inference`, you should also implement zero-padding (make it from `[1024, 2048]` to `[1025, 2049]` ) and this will result in the same result as...
Sorry to disturb again, @hellochick ! Would you mind sharing your configurations for training `cityscapes` using `train_30k` mode to achieve over 67%mIoU. I tried to continue training while the accuracy...