Qiule Sun

@rohitgirdhar The visualization of ‘action words’ in section 4.7 is meaningful and helpful for understanding the action word. Could you provide more details about the implementation of visualization of ‘action...

I got the error,'无法打开要写入的文件 "E:\project\STN\Pedestrian_Alignment-master\dataset\bounding_box_train_256\0002_c1s1_000451_03.jpg"',when run the imwrite(im,url256) in prepare_data.

I find the reson. The dictiontory, bounding_box_train_256, is not created.

@InstantWindy The voc server gives the directory structure of submission and also provides an example submission structure for the VOC2010 challenge.

I didn't meet that error. Your folder fomat looks correct. The directory structure of submission is compressed by a single archive file (xxx.tar.gz). The evaluation log is as follow. **Evaluation...

@txfs1926 May be another answer can help you. See #179 . Question1 For obtaining mIoU with backgroud , computing mIoU with backgroud, i,e,. 60 classes for PContext, fisrt need to...

@zhanghang1989 I have noticed that GCU (NeurIPS2018, paper link: https://papers.nips.cc/paper/8135-beyond-grids-learning-graph-representations-for-visual-recognition.pdf) creates visualization of the assignment matrix in their method. It is meaningful to illustrate the proposed method. Do you consider...

@zhanghang1989 GC = tmp.sum(0).sum(0) in https://github.com/zhanghang1989/PyTorch-Encoding/blob/master/encoding/lib/cpu/encoding_cpu.cpp#L44 should be GC = -tmp.sum(0).sum(0) ?

@zhanghang1989 When args.test_val=Ture （https://github.com/zhanghang1989/PyTorch-Encoding/blob/master/experiments/segmentation/test.py#L103), dst (https://github.com/zhanghang1989/PyTorch-Encoding/blob/master/experiments/segmentation/test.py#L103) is the image name rather than the mask. Why is the image name obtained at this time?

@luistelmocosta You can drop all FC layers and append Encoding module to the end of last conv. layer.