pytorch-segmentation-toolbox

pytorch-segmentation-toolbox copied to clipboard

This PR fixes the predict_sliding function of `evaluate.py` not giving correct values. Currently instead of adding the full batch of obtained probabilities, it only adds the ones from the first...

ironcadiz

The program got stuck and I use V100.

1

The program got stuck just like former issues, however, I use 4 V100s, which should have enough memory. I also tried to change bs to 4 and image size to...

hzhupku

Hi, may I ask why you add [nn.Dropout2d(0.1)](https://github.com/speedinghzl/pytorch-segmentation-toolbox/blob/master/networks/pspnet.py#L74) in the PSP module, which is different from the original [implementation](https://github.com/hszhao/semseg/blob/master/model/pspnet.py#L8)

erichhhhho

About cuda version.

1

Hi, Thanks to provide the newer code of pytorch 1.1, but my GPU is RTX 2070 and it don't support cuda 9.0, I want to know if this code can...

chccgiven

How can I use this model on my own data set?

3

Thank you for doing such an outstanding job!How can I use this model on my own data set?

jlhou

How to use this code in pytorch1.0

2

Because the inplace_abn(the third party libs that this project used) have some bugs in 8 GPUs pytorch0.4, if we want to train in 8GPUs, we should update the newest bn...

mingrui-xie

Thanks for your code. I run it well, but I find a porblem. There are 30 classes in their official introduction while you set num_classes to 19. I do not...

DRosemei

I try using ohem in my model, argument i set probility is 0.7 and k is 100000,but i find it not work and the result is become worse. i experiment...

forrestsz

the program is stucked before anything.

1

File "atrous_concat.py", line 159, in out = model(img) File "/home/tensor/anaconda2/envs/py35/lib/python3.5/site-packages/torch/nn/modules/module.py", line 477, in __call__ result = self.forward(*input, **kwargs) File "atrous_concat.py", line 118, in forward c2,c3,c4,c5 = self.backbone(x) File "/home/tensor/anaconda2/envs/py35/lib/python3.5/site-packages/torch/nn/modules/module.py", line...

Anikily

question about maxpool

1

Thanks for your excellent work. I notice that the self.maxpool in your arch is redefined, where the ceiling mode is used. Can you explain why you use this mode and...

chenxiaoyu523