Ecosyste.ms: Issues

An open API service for providing issue and pull request metadata for open source projects.

GitHub / qubvel-org/segmentation_models.pytorch issues and pull requests

#776 - Support Timm ViTs

Pull Request - State: closed - Opened by isaaccorley over 1 year ago - 8 comments
Labels: Stale

#758 - Adding optional class_weights to dice_loss

Pull Request - State: closed - Opened by rishiswethan over 1 year ago - 3 comments
Labels: Stale

#757 - integrate SAM (segment anything) encoder with Unet

Pull Request - State: open - Opened by Rusteam over 1 year ago - 62 comments

#689 - Softmax activation function throws deprecation warning

Issue - State: closed - Opened by vdplasthijs almost 2 years ago - 3 comments
Labels: Stale

#424 - cannot import name 'container_abcs' from 'torch._six'

Issue - State: closed - Opened by lifeischaotic over 3 years ago - 29 comments

#238 - DeeplabV3plus

Issue - State: closed - Opened by jing-tian-2019 about 4 years ago - 12 comments
Labels: Stale

#169 - UserWarning: Implicit dimension choice for softmax has been deprecated.

Issue - State: closed - Opened by gireeshkbogu over 4 years ago - 3 comments

#100 - New features and refactoring

Pull Request - State: closed - Opened by qubvel almost 5 years ago

#99 - Xception encoder

Pull Request - State: closed - Opened by chlxydl almost 5 years ago - 4 comments

#98 - Update link to Catalyst training

Pull Request - State: closed - Opened by TezRomacH almost 5 years ago

#97 - I need HRNet

Issue - State: closed - Opened by ShihuaiXu almost 5 years ago - 8 comments
Labels: Stale

#96 - Amend the layers of encoder ?

Issue - State: closed - Opened by ghost almost 5 years ago - 2 comments

#95 - Is there a plan to update required packages?

Issue - State: closed - Opened by sooperset almost 5 years ago - 1 comment

#94 - What is the purpose of parameter "n" in AverageValueMeter

Issue - State: closed - Opened by VSydorskyy almost 5 years ago - 3 comments
Labels: Stale

#92 - if i use resnet18 as the backbone,the input channels must be 3?

Issue - State: closed - Opened by AMSTLHX almost 5 years ago - 8 comments
Labels: Stale

#91 - Why output of unet with activation 'sigmoid' not range from 0 to 1

Issue - State: closed - Opened by cswwp almost 5 years ago - 3 comments

#90 - Code Quality improvements

Pull Request - State: closed - Opened by vfdev-5 almost 5 years ago - 5 comments

#89 - Fixes ImportError from efficientnet_pytorch.utils

Pull Request - State: closed - Opened by vfdev-5 almost 5 years ago - 1 comment

#88 - Add inceptionv4 encoder

Pull Request - State: closed - Opened by khornlund almost 5 years ago - 1 comment

#87 - Frequent 'Connection time out' while downloading encoders like 'se_resnext101_32x4d'

Issue - State: closed - Opened by anuragtr almost 5 years ago - 3 comments
Labels: Stale

#86 - How to install it with source code

Issue - State: closed - Opened by Hesene almost 5 years ago - 1 comment

#85 - Memory efficient swish for efficientnet encoder

Pull Request - State: closed - Opened by qubvel almost 5 years ago - 1 comment

#84 - can't use efficientnet offline

Issue - State: closed - Opened by mobassir94 almost 5 years ago - 3 comments

#83 - EfficientNet-PyTorch eating Memory very fast

Issue - State: closed - Opened by mobassir94 almost 5 years ago - 3 comments

#82 - Add ability to use concatenation instead of addition in FPN

Pull Request - State: closed - Opened by SiarheiFedartsou almost 5 years ago - 1 comment

#81 - Add ability to use Activated BatchNorm in decoder.

Pull Request - State: closed - Opened by SiarheiFedartsou almost 5 years ago - 6 comments

#80 - Model request: Add encoders trained on WSL images

Issue - State: closed - Opened by ternaus almost 5 years ago - 1 comment

#79 - Feature request: option to freeze stages

Issue - State: closed - Opened by ternaus almost 5 years ago - 7 comments
Labels: Stale

#78 - Feature request: Add an option to use concatenation in FPN

Issue - State: closed - Opened by ternaus almost 5 years ago - 1 comment

#77 - Add Conv2d Weight Standartization to fpn segmentation block

Pull Request - State: closed - Opened by valyukov almost 5 years ago - 3 comments
Labels: Stale

#76 - PspNet Aux Output not added during validation

Issue - State: closed - Opened by jaideep11061982 almost 5 years ago - 1 comment

#75 - How to override input 3x3 param for Encoders

Issue - State: closed - Opened by jaideep11061982 almost 5 years ago - 5 comments
Labels: Stale

#74 - Add mock library as dependency in Dockerfile.dev

Pull Request - State: closed - Opened by SiarheiFedartsou almost 5 years ago - 1 comment

#73 - Add EfficientNet encoder

Pull Request - State: closed - Opened by SiarheiFedartsou almost 5 years ago - 1 comment

#72 - Add optional final upsampling

Pull Request - State: closed - Opened by qubvel almost 5 years ago

#71 - Feature request: make optional second output in the middle of the network.

Issue - State: closed - Opened by ternaus almost 5 years ago - 7 comments
Labels: Stale

#70 - Update decoder initialization

Pull Request - State: closed - Opened by qubvel almost 5 years ago

#69 - Fix package deps

Pull Request - State: closed - Opened by qubvel almost 5 years ago

#68 - Deeplabv3 implementation

Issue - State: closed - Opened by 421psh almost 5 years ago - 2 comments
Labels: Stale

#67 - add description of dependency

Issue - State: closed - Opened by trytolose almost 5 years ago - 2 comments
Labels: Stale

#66 - a new backbone

Issue - State: closed - Opened by xiaoyang-coder almost 5 years ago - 3 comments
Labels: Stale

#65 - Make upsampling optional in FPNDecoder.

Issue - State: closed - Opened by ternaus about 5 years ago - 1 comment

#64 - TypeError: __init__() got an unexpected keyword argument 'groups'

Issue - State: closed - Opened by andriilitvynchuk about 5 years ago - 2 comments

#63 - remove torchnet from requirements

Pull Request - State: closed - Opened by qubvel about 5 years ago

#62 - fixed requirements

Pull Request - State: closed - Opened by phillies about 5 years ago - 1 comment

#61 - Request. Mobilenet as encoder

Issue - State: closed - Opened by Vozf about 5 years ago - 1 comment

#60 - cannot import name 'cfg'

Issue - State: closed - Opened by ZinuoCai about 5 years ago - 1 comment

#59 - Add LogSoftmax to the list of activations.

Issue - State: closed - Opened by ternaus about 5 years ago - 1 comment

#58 - torchvision 0.4

Issue - State: closed - Opened by ternaus about 5 years ago - 4 comments

#57 - Image size issues

Issue - State: closed - Opened by psinger about 5 years ago - 6 comments

#56 - KeyError: 'resnext50_32x4d'

Issue - State: closed - Opened by hktxt about 5 years ago - 1 comment

#55 - resnext model

Pull Request - State: closed - Opened by laol777 about 5 years ago

#54 - How to handle the multispectral image (NOT RGB image)?

Issue - State: closed - Opened by yifaliu about 5 years ago - 5 comments

#53 - scSE module for unet

Pull Request - State: closed - Opened by laol777 about 5 years ago - 1 comment

#52 - Which dataset do you use in example.ipynb?

Issue - State: closed - Opened by heartInsert about 5 years ago - 2 comments

#51 - how can i unfreeze the layers of vgg16/vgg11 encoder used with unet decoder?

Issue - State: closed - Opened by mobassir94 about 5 years ago - 1 comment

#50 - Question about activation in example

Issue - State: closed - Opened by VSydorskyy about 5 years ago - 2 comments

#49 - confused by BCEDiceLoss

Issue - State: closed - Opened by VSydorskyy about 5 years ago - 2 comments

#48 - confused by loss function forward params

Issue - State: closed - Opened by cswwp about 5 years ago - 3 comments

#47 - cannot import name 'cfg' from 'torchvision.models.vgg'

Issue - State: closed - Opened by JimWang97 about 5 years ago - 1 comment

#46 - Request: Add segmentation models to Kaggle docker.

Issue - State: closed - Opened by ternaus about 5 years ago - 3 comments
Labels: Stale

#45 - Add preprocessing_params

Pull Request - State: closed - Opened by qubvel about 5 years ago

#44 - Any difference between sigmoid and softmax activation when using f_score and IOU metrics?

Issue - State: closed - Opened by NIRVANALAN about 5 years ago - 4 comments
Labels: Stale

#43 - Feature request: preprocess_input as a dictionary

Issue - State: closed - Opened by ternaus about 5 years ago - 4 comments
Labels: Stale

#42 - Bug with `import segmentation_models_pytorch as smp`

Issue - State: closed - Opened by ternaus about 5 years ago

#41 - error when using fscore

Issue - State: closed - Opened by NIRVANALAN about 5 years ago - 3 comments

#40 - Can I use this framework to train my own data?

Issue - State: closed - Opened by luostars about 5 years ago - 3 comments
Labels: Stale

#39 - how to transfer a standard pretrained resnet model to 4-channel in your code?

Issue - State: closed - Opened by SonwYang about 5 years ago - 10 comments
Labels: Stale

#38 - example doesn't work

Issue - State: closed - Opened by rohrdr about 5 years ago - 2 comments
Labels: Stale

#37 - No module named 'segmentation_models_pytorch.common.blocks'

Issue - State: closed - Opened by pyaf about 5 years ago

#36 - How to use nn.BCEWithLogitsLoss()?

Issue - State: closed - Opened by jovenwayfarer about 5 years ago - 1 comment

#35 - Correction of an error in the function description

Pull Request - State: closed - Opened by sankovalev about 5 years ago - 1 comment

#34 - how to use multi-gpu?

Issue - State: closed - Opened by hust-kevin about 5 years ago - 10 comments

#33 - IF I have only one class lable，how to do

Issue - State: closed - Opened by mk123qwe about 5 years ago - 1 comment

#32 - confusion with the train and valid operations

Issue - State: closed - Opened by songkq about 5 years ago

#31 - multiclass jaccard loss

Issue - State: closed - Opened by akshay-vaidya about 5 years ago - 1 comment

#30 - Use concatenation for feature pyramid aggregation?

Issue - State: closed - Opened by hyc-xyz about 5 years ago - 3 comments

#29 - multiclass

Issue - State: closed - Opened by HuangQinJian about 5 years ago - 12 comments

#28 - How to use the weights after trainning?

Issue - State: closed - Opened by jianyin2016 about 5 years ago - 2 comments

#27 - How to

Issue - State: closed - Opened by jianyin2016 about 5 years ago

#26 - How to PSPNET train with single channels input?

Issue - State: closed - Opened by parulML about 5 years ago - 1 comment

#25 - Feature/torchvision 0.3

Pull Request - State: closed - Opened by qubvel about 5 years ago

#24 - [Ask for advice] Modify models by adding skip connections

Issue - State: closed - Opened by crossknight about 5 years ago - 4 comments

#23 - train error

Issue - State: closed - Opened by HuangQinJian over 5 years ago

#22 - image shape

Issue - State: closed - Opened by HuangQinJian over 5 years ago

#21 - invalid hash value

Issue - State: closed - Opened by HuangQinJian over 5 years ago - 1 comment

#20 - If I have three class to seg,how to set?Like this?

Issue - State: closed - Opened by HuangQinJian over 5 years ago - 6 comments

#19 - [Request] EfficientNet as a encoder

Issue - State: closed - Opened by crossknight over 5 years ago - 4 comments

#18 - input size

Issue - State: closed - Opened by HuangQinJian over 5 years ago - 2 comments

#17 - IOU metric sometimes bigger than 1

Issue - State: closed - Opened by Diyago over 5 years ago - 6 comments

#16 - incompatible when TorchVison update to 0.3

Issue - State: closed - Opened by seefun over 5 years ago - 3 comments
Labels: enhancement

#15 - Fix pickling `preprocess_input` function

Pull Request - State: closed - Opened by qubvel over 5 years ago - 1 comment

#13 - pr_mask is not correct although the accuracy on CamVid test set is very high

Issue - State: closed - Opened by KleinXin over 5 years ago - 6 comments

#11 - ImportError: No module named pretrainedmodels.models.torchvision_models

Issue - State: closed - Opened by KleinXin over 5 years ago - 2 comments

#10 - Add `None` activation

Pull Request - State: closed - Opened by qubvel over 5 years ago

#9 - None activation for UNet output.

Issue - State: closed - Opened by ternaus over 5 years ago