Comments (7)
求教。使用默认的代码跑LITS,效果很差,跑不出提供的progress.png显示出的效果,在提供的代码上跑nnUNet效果倒是很正常,
from unetplusplus.
我始终无法复现提供的效果的loss图。用nnUNet跑了一遍loss确实和图基本一致,一路降到-4。
但是用UNet++,loss很难收敛,到100个epoch才不到-1.0,和图里的效果相差太大了.....
from unetplusplus.
我尝试只改动了paths.py配置数据集路径,其余的部分和源码一致丝毫未改。但是progress.png显示loss下降速度比图里慢的太多了。真心求教!
from unetplusplus.
同学,我不清楚你问题的答案,但是想和你交流一下。
我按作者在readme里“How to use UNet++”的步骤,结果在“Liver and tumour segmentation in Medical Segmentation Decathlon”上曲线乱飞。batchsize调到4时,占用V100约99%的显存,曲线也不行。想交流看看,你是怎么设置的实验的?
from unetplusplus.
@ChibisukeDragon 基于nnunet的默认plans是batchs_size=2,默认的UNet++用上deep supervision大概要占24G显存。再调大batchsize显存肯定不够。作者在提供的代码里,在trainer部分应该改成output[0], deep_supervision.py里有深度监督的逻辑也被注释了,得把这些都开下来才能跑出正常的曲线。但是速度真是极其慢,极其慢。3090一个epoch大概300多秒
from unetplusplus.
我当时看到的是这个issue:#15
作者如果配置使用3d_fullres,单卡根本开不到12的batchsize。。。
所以没收敛是因为没开deep supervision的问题吗?!我一直以为是batchsize太小才不收敛 @cuihu1998
from unetplusplus.
@ChibisukeDragon 图里loss快速掉到-1然后最终收敛到-3、-4这种情况,是开deep supervision的,因为多个layer的loss全加在一起了。不开监督loss就会没图里那么低。3d_fullres bs开不到那么大的,12是2d的默认batchsize。
from unetplusplus.
Related Issues (20)
- Getting very bad validation metrics HOT 2
- how to train on my designated GPU? HOT 1
- Can this model be trained on multi GPUs ?
- importlib_metadata.PackageNotFoundError: No package metadata was found for nnunet HOT 1
- 周学长您好,我在配置环境中发现个小坑想提醒大家
- RuntimeError: Given transposed=1, weight of size [256, 128, 2, 2], expected input[12, 64, 256, 256] to have 256 channels, but got 64 channels instead HOT 2
- Question : why use " l = weights[0] * self.loss(x[-1], y[0])" in loss_functions/deep_supervision.py ?
- Anyone here who have update this code for tensorflow version 2.x HOT 1
- issue in resize_segmentation HOT 3
- 'unet_final_features' referenced before assignment HOT 1
- Generic_UNetPlusPlus测试报错
- legacy_upsampling2d_support
- How to Train on nnunet 2d mode?
- PermissionError: [Errno 13] Permission denied: '/media/yang/nnUNet_raw_data_base'
- AttributeError: 'Generic_UNetPlusPlus' object has no attribute 'upsample_mode'
- Pre-trained model
- UNetPlusPlus如何基于自定义数据集完成实例分割任务?(而不是语义分割)
- How do I train U-Net++ on my custom 3D data?
- Help
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from unetplusplus.