keras-unet-collection
keras-unet-collection copied to clipboard
The Tensorflow, Keras implementation of U-net, V-net, U-net++, UNET 3+, Attention U-net, R2U-net, ResUnet-a, U^2-Net, TransUNET, and Swin-UNET with optional ImageNet-trained backbones.
I am getting this error when I added custom loss input shape(512,512,3) model=models.unet_2d(input_size, filter_num, n_labels=1, stack_num_down=2, stack_num_up=2, activation='ReLU', output_activation=None, batch_norm=True, pool='max', unpool='nearest', backbone='VGG16', weights='imagenet', freeze_backbone=True, freeze_batch_norm=True, name='unet') //////// selected_layers =...
I have a 6 class semantic segmentation problem. I am trying to use SwinUNet as follows `models.swin_unet_2d((512, 512, 3), filter_num_begin=64, n_labels=6, depth=4, stack_num_down=2, stack_num_up=2, patch_size=(2, 2), num_heads=[4, 8, 8, 8],...
Does anyone else have any problem running these models on TPU? How do we make the library work in TPU environment? Thanks.
I have 6 classes semantic segmentation problem. How attention u-net in this package considers a particular class? How to give more attention to a particular class?
The code for the third exemple in the user guide (attention-Unet for single target regression) does not work for me. When I train the model, I get this error :...
I downloaded the COCO dataset and tested it as in the example. I create a dataset and an error occurs in the train code. ```python ... record = np.mean(keras.losses.categorical_crossentropy(valid_label, y_pred))...
Hi, In TransUNet, `filter_num` shows the number of filters for down and upsampling levels, right? However, If I use 3 filters `filter_num=[64, 128, 128]` instead of the default `filter_num=[64, 128,...
I was using your Image segmentation using customized UNET 3+ and hybrid loss and user_guide_models notebook " to train with Unet_plus_2d model on COCO. but during training,I got the following...
hello A question I have trained swin net model for segmentation and my image size was 64*64*3 ; Can I use (prediction) this model (model and it's weights) when I...
I am trying to use the UNET 3plus with deep supervision and cgm for nodule segmentation from lung ct scans. When I train the model on a 1000 image subset...