search

MaybeShewill-CV / bisenetv2-tensorflow

Unofficial tensorflow implementation of real-time scene image segmentation model "BiSeNet V2: Bilateral Network with Guided Aggregation for Real-time Semantic Segmentation"
https://maybeshewill-cv.github.io/bisenetv2-tensorflow/
MIT License
224 stars 59 forks source link

about training new custom data set #29

Closed zjj321 closed 3 years ago

zjj321 commented 3 years ago

I have trained new data set it worked and successful and all imag is resize 960*540。

but when i try to train smaller dataset: every img is 896 544, it is always print error:slice index 262144 of dimension 0 out of bounds。and i changed even smaller image size: 512320, it print same error。

somewhere should i need to do?

MaybeShewill-CV commented 3 years ago

@zjj321 Change the config here https://github.com/MaybeShewill-CV/bisenetv2-tensorflow/blob/710db8646ceb505999b9283c0837c0b5cf67876d/config/cityscapes/cityscapes_bisenetv2.yaml#L80-L83 You may change it into 0.5 896 544 or 0.25 896 544

zjj321 commented 3 years ago

best wishs,thank you so much