jfzhang95/pytorch-deeplab-xception issues and pull requests

#86 - how is the ResNet model trained, on pascal train, or pascal aug, or with coco pretrain?

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

#85 - inference code

Issue - State: closed - Opened by albertchristianto over 5 years ago - 12 comments

#84 - reproduced result is low on ResNet backbone

Issue - State: open - Opened by youngwanLEE over 5 years ago - 22 comments

#83 - Pretrained Model not Loading

Issue - State: open - Opened by eriche7133 over 5 years ago - 7 comments

#82 - About data transform

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

#81 - About dilation ratio in Mobilenet

Issue - State: open - Opened by 666zz666 over 5 years ago

#80 - 'DeepLab' object has no attribute 'module'

Issue - State: closed - Opened by abhimanyuchadha96 over 5 years ago - 8 comments

#79 - inference speed of MobileNet backbone

Issue - State: open - Opened by Ruoqianqian over 5 years ago - 1 comment

#78 - What‘s the batch size with xception backbone?

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

#77 - question about training efficiency(speed)

Issue - State: open - Opened by JongMokKim over 5 years ago

#76 - Do we really need the `FixScaleCrop` in the `transform_val`?

Issue - State: open - Opened by ghost over 5 years ago - 2 comments

#75 - pretrained model is error，failed to decompression

Issue - State: closed - Opened by ChaosCen over 5 years ago - 3 comments

#74 - AttributeError: module 'torch.nn.functional' has no attribute 'interpolate'

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

#73 - train.py backbone=resnet/mobilenet

Issue - State: open - Opened by 100sby over 5 years ago - 3 comments

#72 - CUDA error: out of memory

Issue - State: closed - Opened by Ruoqianqian over 5 years ago - 3 comments

#71 - train.py argument's description fix

Pull Request - State: open - Opened by ShotaroKataoka over 5 years ago

#70 - Expected tensor for argument #1 'input' to have the same device as tensor for argument #2 'weight'; but device 1 does not equal 0

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

#69 - Question about validation

Issue - State: open - Opened by CeciliaPYY over 5 years ago - 3 comments

#68 - Bug fixes

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

#67 - We got left correct and right error when using pretrain model

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

#66 - How can I evaluate with MS and Flip?

Issue - State: open - Opened by kaixinxiao over 5 years ago

#65 - low mIOU in train with coco and ResNet as backbone

Issue - State: open - Opened by zwtburger over 5 years ago - 3 comments

#64 - Update train_voc.sh

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

#63 - Transfer Learning with trained Xception-65 backbone

Issue - State: closed - Opened by wamiq-reyaz over 5 years ago

#62 - CityScapes Pretrained model

Issue - State: open - Opened by vineet-g over 5 years ago - 3 comments

#61 - Sync Batchnorm in inference

Issue - State: open - Opened by AlanStark over 5 years ago - 2 comments

#60 - why not test 'OS=8' ?

Issue - State: open - Opened by idealwei over 5 years ago

#59 - Update README.md to FIX a typo.

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

#58 - _load_pretrained_models(self) method Problem

Issue - State: open - Opened by xd17 over 5 years ago - 1 comment

#57 - Resume Training Problem

Issue - State: open - Opened by herleeyandi over 5 years ago - 3 comments

#56 - Pretrained xception model

Issue - State: open - Opened by lxtGH over 5 years ago

#55 - Batch average over Loss

Issue - State: open - Opened by AlanStark over 5 years ago - 1 comment

#54 - VOC 2012 TEST results

Issue - State: open - Opened by shaibagon over 5 years ago - 17 comments

#53 - provided pretrain models

Issue - State: closed - Opened by ymli39 over 5 years ago

#52 - DRN model cannot be downloaded

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

#51 - Validation is only estimated on part of the images

Issue - State: open - Opened by shaibagon over 5 years ago - 2 comments

#50 - Wrong padding of "label"

Issue - State: open - Opened by shaibagon over 5 years ago

#49 - Unable to reproduce result on VOC with small batch size

Issue - State: open - Opened by xinario over 5 years ago - 7 comments

#48 - A problem about load weight

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

#47 - How to visualize the train porcess

Issue - State: open - Opened by zxshi almost 6 years ago - 2 comments

#46 - Program does not run

Issue - State: open - Opened by zxshi almost 6 years ago - 5 comments

#45 - have your reproduced the result with xception as backbone on cityscapes?

Issue - State: open - Opened by CoinCheung almost 6 years ago

#44 - Single GPU

Issue - State: closed - Opened by yyangmei almost 6 years ago - 4 comments

#43 - high resolution input image

Issue - State: open - Opened by gnool almost 6 years ago - 1 comment

#42 - Models are not in tar format

Issue - State: closed - Opened by lutzroeder almost 6 years ago - 2 comments

#41 - ValueError: Expected more than 1 value per channel when training, got input size [1, 256, 1, 1]

Issue - State: closed - Opened by zxshi almost 6 years ago - 6 comments

#40 - a simple problem: why input size 513*513

Issue - State: closed - Opened by zxshi almost 6 years ago - 5 comments

#39 - CuDNN error: CUDNN_STATUS_EXECUTION_FAILED

Issue - State: open - Opened by taomrzhang almost 6 years ago - 10 comments

#38 - A problem about encoding mask

Issue - State: closed - Opened by yyangmei almost 6 years ago - 2 comments

#37 - Can you provide the test and visualization?

Issue - State: closed - Opened by zxshi almost 6 years ago - 2 comments

#36 - mIoU error using xception as backbone

Issue - State: open - Opened by frazerlin almost 6 years ago - 2 comments

#35 - Low mIoU on Cityscapes

Issue - State: open - Opened by fabvio almost 6 years ago - 21 comments

#34 - Why DW is not followed by BN in Xception?

Issue - State: closed - Opened by lionel3 almost 6 years ago - 3 comments

#33 - test

Issue - State: closed - Opened by lucheng07082221 almost 6 years ago - 1 comment

#32 - Hello, Could you provide pretrained models ?

Issue - State: closed - Opened by TreB1eN almost 6 years ago - 6 comments

#31 - xception backbone mIou low

Issue - State: open - Opened by ghy200692162 almost 6 years ago - 3 comments

#30 - decode support the h and w is different of input

Pull Request - State: closed - Opened by WenmuZhou almost 6 years ago - 1 comment

#29 - Performance on Cityscape

Issue - State: closed - Opened by lxtGH almost 6 years ago - 5 comments

#28 - How to output result image from output?

Issue - State: closed - Opened by OOXXXXOO almost 6 years ago - 8 comments

#27 - using your codes train the voc datasets from scratch got very large loss value?

Issue - State: closed - Opened by monk42 almost 6 years ago - 2 comments

#26 - Skip Connection Middle FLow

Issue - State: closed - Opened by fuzailpalnak almost 6 years ago

#25 - what datasets did you use in pretained resnet-101 and xception?sorry, I want to know but i can't find.

Issue - State: closed - Opened by yangzhengzheng almost 6 years ago - 4 comments

#24 - mIOU is low

Issue - State: closed - Opened by ghost almost 6 years ago - 3 comments

#23 - about loss and mIou

Issue - State: closed - Opened by vincennnnt almost 6 years ago - 3 comments

#22 - Hi What is

Issue - State: closed - Opened by AlgoHunt almost 6 years ago - 4 comments

#21 - What is a well lr for train the voc datasets?

Issue - State: closed - Opened by lingtengqiu almost 6 years ago - 2 comments

#20 - A problem about transform code

Issue - State: closed - Opened by Gaoyiminggithub almost 6 years ago - 7 comments

#19 - What's evaluation results?

Issue - State: closed - Opened by WillSuen almost 6 years ago - 1 comment

#18 - Could u offer your pretrained model on backbone?

Issue - State: closed - Opened by ThorinGondor almost 6 years ago

#17 - pretrained models?

Issue - State: closed - Opened by skrish13 almost 6 years ago - 1 comment

#16 - miou calculate

Issue - State: closed - Opened by cuishuhao about 6 years ago - 3 comments

#15 - Result Miou

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

#14 - GPU memory consumption

Issue - State: closed - Opened by JamesKasperYu about 6 years ago - 1 comment

#13 - block20 lack of one sep conv [1024,1024,3,3]

Issue - State: closed - Opened by ghy200692162 about 6 years ago - 1 comment

#12 - something wrong with metric mIoU seemingly (with 63% mIoU using original fcn metric)

Issue - State: open - Opened by OneForward about 6 years ago - 10 comments

#11 - SGD Batch_size = 1 problem

Issue - State: closed - Opened by xiaosean about 6 years ago - 1 comment

#10 - something about pre-trained xception model

Issue - State: closed - Opened by DaPenggg about 6 years ago - 1 comment

#9 - Why you set the initial lr as low as 1e-7 in trian.py?

Issue - State: closed - Opened by minzhouGithub about 6 years ago - 1 comment

#8 - Error in Block with stride>2

Pull Request - State: closed - Opened by Fettpet about 6 years ago

#7 - Missing ReLU activation function

Issue - State: closed - Opened by dannyhung1128 about 6 years ago - 2 comments

GitHub / jfzhang95/pytorch-deeplab-xception issues and pull requests