search

Amshaker / unetr_plus_plus

[IEEE TMI-2024] UNETR++: Delving into Efficient and Accurate 3D Medical Image Segmentation
Apache License 2.0
340 stars 32 forks source link

Regarding finetuning #70

Open shweta2266 opened 6 months ago

shweta2266 commented 6 months ago

Dear Authors,

Thank you so much for providing this amazing work. I could reproduce the results for lung tumor segmentation on the provided preprocessed dataset. However, I want to finetune the lung tumor segmentation model on my data. I preprocessed the data according to the given code. When I try to fine-tune the model using the provided pre-trained weights, the dice score is zero consistently and not improving at all. Can you please help me with this?

Regards Shweta

WangXiaoXv commented 5 months ago

Dear Authors,

Thank you so much for providing this amazing work. I could reproduce the results for lung tumor segmentation on the provided preprocessed dataset. However, I want to finetune the lung tumor segmentation model on my data. I preprocessed the data according to the given code. When I try to fine-tune the model using the provided pre-trained weights, the dice score is zero consistently and not improving at all. Can you please help me with this?

Regards Shweta

Hello, I apologize for any inconvenience caused. When running the training_scripts/run_training_acdc.sh script, I encountered a negative train_loss. I am eagerly awaiting your response! 屏幕截图 2024-03-22 202334 图片1