Limin Zhang

same error

> > same error > > I had solved it. change your method to install sacrebleu rather than pip. you are linux or windows?

> > same error > > I had solved it. change your method to install sacrebleu rather than pip. I can't solve this error by your method.

@silence1114 I didn't test, beacuse I thought the test during training is same as test solely. Do you have other discoveries?

See ``` model.compile(optimizer=sgd, loss='categorical_crossentropy', metrics=['accuracy']) ``` in `layers_builder.py`

@Parkjh215 is right. Beacuse your validation generator get 0 images. Change the code in `main.py`: ```Python # Create validation generator val_datagen = AugmentingDataGenerator(rescale=1./255) val_generator = val_datagen.flow_from_directory( args.validation, MaskGenerator(512, 512, 3),...

@pyradd I change some hyper paramaters, and the config is : " aux=True, backbone='resnet50', base_size=1024, batch_size=8 crop_size=720, dataset='citys', device='cuda', distributed=False, epochs=200, lr=0.01, model='icnet' " And the mIoU is about 59.5%....

@yangxudong9 Yes, I agree with your view. I have made several changes to ICNet, and I have experimentally verified that these changes can make ICNet achieve the results in the...

@lizhengwei1992 thank you for your reply, and： - From the perspective of technical implementation, do you think it is difficult to test any size input image? - In other words,...

@LovPe Hi, I am not sure about the "deep_stem" param, can you explain more about the param? Additionally, do you change any other hyperprams?