TransUNet
TransUNet copied to clipboard
Beckschen
This repository includes the official project of TransUNet, presented in our paper: TransUNet: Transformers Make Strong Encoders for Medical Image Segmentation.
Hi!Sir. I'm having some problems,which *.zip should i take for training/testing, there three .zip files in Abdomen dir, i.e., RawData.zip, Reg-Training-Testing.zip and Reg-Training-Training.zip, please tell me. Thank you.
Hi, thanks for the great work. I just wonder that you have the pretrained models of R50+ViT_16. In your article, you said all R50 and ViT are pretrained on ImageNet...
Hi All, Is there difference in performance consistency between loading saved pretrained weights using PyTorch load_state_dict() and using your defined function of load_from ?
Hi It appears there is no validation set for the synapse dataset. Could you tell me how you selected the best model during training? and Have you saved just one...
Firstly, thank you for making the code public, it's neatly written and quite understandable. I think I am a bit confused about (partly because of my lack of knowledge about...
the model i trianed got loss value about 0.032, i want to konw the loss value from official experiment is how large
thanks for your great work! I was using pretrained model R50-ViT-B_16 and trained it on some opensource dataset for lung segmentation. During training, the total loss stayed above 0.7 and...
When running the test file, tracing back to class DecoderCup under the vit_seg_modeling file, the variable 'self.config.skip_channels' was not found.  
当我运行test.py时，报出如下错误： `python test.py --dataset Synapse --vit_name R50-ViT-B_16 Namespace(Dataset=, base_lr=0.01, batch_size=24, dataset='Synapse', deterministic=1, exp='TU_Synapse224', img_size=224, is_pretrain=True, is_savenii=False, list_dir='./lists/lists_Synapse', max_epochs=150, max_iterations=20000, n_skip=3, num_classes=9, seed=1234, test_save_dir='../predictions', vit_name='R50-ViT-B_16', vit_patches_size=16, volume_path='../data/Synapse/test_vol_h5', z_spacing=1) TU_pretrain_R50-ViT-B_16_skip3_epo150_bs24_224 12 test...
