Researcher.YuanYuhui

Could you share your detailed training settings ?

1

Hi, I hope to reproduce your training results. So could you share your training details? Regards

ImageNet pretrained Aligned Xception model

11

Thanks for sharing the repo. I notice that another [repo](https://github.com/jfzhang95/pytorch-deeplab-xception) mentions that there exist no ImageNet pre-trained Aligned Xception model, and the performance is worse than the ResNet-101. In other...

RuntimeWarning: Negative areas founds: 3 warnings.warn("Negative areas founds: %d" % neg_area_idx.size, RuntimeWarning))

5

I add a non-local block to the backbone and train the network following "e2e_mask_rcnn_R-50-RPN_1x.yaml". However, I am not sure how to set the learning rate of the extra introduced non-local...

Small concerns on the experiments in Table-4

2

Really nice work and super impressive results on the low-data regime (shown in Table 4, also pasted as following)!!  We have some small concerns about the red circle marked...

4 X P40 is not enough?

1

I have tried to reproduce your results with 4X P100 but meet such errors. I am wondering whether anyone could give me some hints. ``` loss_rpn_bbox_fpn6: 0.014951 INFO train_net_step.py: 443:...

Fail to Reproduce the FCN baseline on Pascal-Context & Cityscapes with openseg.pytorch.

7

Really excellent work! In fact, we have tried to integerate the ResNeSt backbone in our own code-base [openseg.pytorch](https://github.com/openseg-group/openseg.pytorch). Here is the backbone configuration [resnest_models.py](https://github.com/openseg-group/openseg.pytorch/blob/master/lib/models/backbones/resnet/resnest_models.py) and the running script [run_r_101_d_8_fcn.sh](https://github.com/openseg-group/openseg.pytorch/blob/master/scripts/cityscapes/resnest/run_r_101_d_8_fcn.sh). (Here...

ImageNet pretrained Aligned Xception model

1

Thanks for sharing the repo. I notice that you mentioned that there exist no ImageNet pretrained Aligned Xception model, and the performance is worse than the ResNet-101. I am wondering...

Nan error caused by “N X C X 1 X 1” input features

4

**Update!** I have found the error is related to the input shape of the SW. When the input shape is NXCX1X1, the output of the SW will become NAN. The...

Ask for the ImageNet pretrained ResNet-152 or ResNext-101/ResNext-152

1

@hangzhaomit Really great work! We find that the provided pretrained models all replace the 7x7 convolutions with three 3x3 convolutions and achieve really good performance. I just have tried to...

Will you release the results on CityScapes?

8

Great work! I am wondering the reproduced performance on the CityScapes. It would be great if you could share the related results.