4uiiurz1/pytorch-nested-unet

Issues

• TypeError: 'NoneType' object is not subscriptable

  #13 opened 5 years ago by manvirvirk
  31

• training speed gets slower when the number of classes increases

  #59 opened 7 months ago by AlphMax
  0

• this repo is unable to run!

  #58 opened a year ago by MorleyOlsen
  2

• Why is the iou result of your code 0.843, but the iou result of the original author of UNet++ for DSB2018 data set 0.92?

  #54 opened 2 years ago by 4dirveduck
  2

• albumentations

  #51 opened 3 years ago by leonard-cooper
  4

• is there a pre-trained model available?

  #57 opened 2 years ago by santoshcoder23
  0

• What is the correct format for training on custom data?

  #52 opened 3 years ago by Callum232310
  1

• How can I predict segmentation results after training?

  #53 opened 2 years ago by simi-ck
  12

• Tensor size mismatch

  #56 opened 2 years ago by liuzhihuilzh
  0

• How do I train on different image dimenstions e.g 192 x 192 instead of 96 x 96 pixels?

  #55 opened 2 years ago by mornedev
  1

• Why is the official IOU 0.92 and your IOU only 0.84

  #42 opened 4 years ago by zhouyi-git
  1

• 请问这个mask目录下的0和1应该怎么分？我的mask标签是一张图啊，类别是两种灰度值

  #32 opened 4 years ago by qiaolegeke
  2

• Question about "ValueError: With n_samples=0, test_size=0.2 and train_size=None, the resulting train set will be empty. Adjust any of the aforementioned parameters".

  #48 opened 3 years ago by wxyWendy
  3

• how to outputs segmentation of all the 4 layers with deep_supervision ?

  #40 opened 4 years ago by zhou-rui1
  1

• When I run the train.py,the results are "loss nan - iou 0.0000 - val_loss nan - val_iou 0.0000"?

  #49 opened 3 years ago by wxyWendy
  2

• hellow I prepare my own data according your suggestion , but it's fault.

  #27 opened 5 years ago by jackft2
  8

• crime detection project

  #50 opened 3 years ago by VIYYAPUVSDSAIVAMSI
  1

• early stopping , could not train after 10 epoch

  #16 opened 5 years ago by manvirvirk
  2

• Where does the code reflect that "the global loss is weighted and accumulated by the loss functions of each branch"?

  #47 opened 3 years ago by wxyWendy
  1

• error when run test.py: CUDNN_STATUS_EXECUTION_FAILED

  #11 opened 3 years ago by godlovehill
  3

• deploy on opencv

  #46 opened 3 years ago by godlovehill
  1

• Is the mask label of the data trained by myself an 8-bit color label graph, or an 8-bit grayscale graph?

  #45 opened 3 years ago by ushio20
  0

• about deep supervision

  #44 opened 3 years ago by suh1030
  0

• 请问masks目录下的0,1....是类别数吗，我只有背景和目标两类，标签在同一张图上，所以我跳过了dataset中的i

  #31 opened 4 years ago by qiaolegeke
  5

• 训练自己的数据集疑问

  #25 opened 5 years ago by wuzuowuyou
  4

• Question about training the model

  #43 opened 4 years ago by mpegia
  0

• Pre training model？？

  #41 opened 4 years ago by xiongzhu666
  0

• IoU only 0.843? How can we achieve 92.37%±0.98 IoU showed in UNet++ paper?

  #12 opened 4 years ago by LiuFG
  5

• How to train own data with multiple classes

  #34 opened 4 years ago by sunny0414
  1

• All bias in conv layers before the batch norm is unnecessary.

  #39 opened 4 years ago by nzw0301
  0

• Incorrect IOU definition

  #36 opened 4 years ago by nzw0301
  2

• what kind of image get best perfomance?

  #38 opened 4 years ago by zhou-rui1
  0

• 怎么模型不见了？

  #37 opened 4 years ago by Rayzlx
  0

• 用自己的数据集，为什么会出现 loss为0 iou为0 val为0 val_iou为0

  #35 opened 4 years ago by computerxiaobai-ls
  0

• How

  #33 opened 4 years ago by sunny0414
  0

• Help!! Unable to predict CNN model after training: How to resolve incompatible layers?

  #30 opened 5 years ago by JacobOfCorns
  0

• how to train on my own database?

  #10 opened 5 years ago by manvirvirk
  4

• Getting "ValueError: With n_samples=0... the resulting train set will be empty." even after seemingly appending datasets properly.

  #29 opened 5 years ago by JacobOfCorns
  1

• I prun the model,but time increase,can you help me?

  #28 opened 5 years ago by shaonian2018
  0

• Instance segmentation

  #26 opened 5 years ago by Ostyk
  0

• Identical folders in the masks?

  #24 opened 5 years ago by cvengneer
  0

• how to find confusion matrix of this model?

  #23 opened 5 years ago by manvirvirk
  0

• Please change the validation code.

  #19 opened 5 years ago by MinchangSung0223
  1

• Recommendation of multi-categories medical dataset for semantic segmentation

  #22 opened 5 years ago by super233
  0

• If num_class default 1, mean mask contain of two color: white and black ?

  #21 opened 5 years ago by hungbnb-0923
  0

• Why transpose of image and mask?

  #20 opened 5 years ago by hungbnb-0923
  0

• model output

  #18 opened 5 years ago by manvirvirk
  1

• model output

  #17 opened 5 years ago by manvirvirk
  0

• could not get segmented image after running val.py file

  #15 opened 5 years ago by manvirvirk
  5

• Do we get segmented image at output?

  #14 opened 5 years ago by manvirvirk
  0