search

rezazad68 / transdeeplab

TransDeepLab: Convolution-Free Transformer-based DeepLab v3+ for Medical Image Segmentation
68 stars 14 forks source link

Train my own data set #2

Closed xdz7799 closed 1 year ago

xdz7799 commented 1 year ago

Hello, author, I feel very good after reading your model. I have reproduced your model, but I feel that there is a big fluctuation with the results in the paper. May I ask the author whether this model can be used to train his own data set? Thank you!

YUCOCOCS commented 1 year ago

Hello, author, I feel very good after reading your model. I have reproduced your model, but I feel that there is a big fluctuation with the results in the paper. May I ask the author whether this model can be used to train his own data set? Thank you!

你好，请问你有经过处理之后的Synapse Multi-Organ Segmentation 的数据集嘛

rezazad68 commented 1 year ago

Hello, author, I feel very good after reading your model. I have reproduced your model, but I feel that there is a big fluctuation with the results in the paper. May I ask the author whether this model can be used to train his own data set? Thank you!

Thanks for your interest. Please follows our git guideline to produce the results. Yes, you can use our network for your own dataset but you need to use your own data loader with appropriate data pre-processing steps. Best