-
-
Hello ,Did you train the IcNet network? I trained it on the VOC2012 dataset and found the result was no better than the paper.I trained it for 15k iterations.
-
I think the your ICNet implementation is not accurate.
First, there is no PPM in the head. Second, no shared computation. In the implementation, forward twice lead to slower speed.
https://github.…
lxtGH updated
4 years ago
-
icnet中的infer.py中 `img = paddle.dataset.image.to_chw(image)[np.newaxis, :]` , paddle读取到的img.shape =(1, 3, 1710, 3384) , 读取之后进行cv2.resize , 报错,请问 paddle读取的img 如何resize
-
Getting an error when trying to train using model icnet, here is the full output:
python train.py --config configs/bdd.yml
RUNDIR: runs/bdd/32825
3500
/usr/local/lib/python2.7/dist-packages/scip…
-
-
Hi!
First of all, thanks for sharing your very helpful code.
Following the _README.md_, I have downloaded _icnet.npy_ from Google Drive link to the _model/cityscapes_ folder.
However, when I…
-
Hi, icnet returned a tuple when training.... but when calculating loss, it directly get size from tuple and got this error:
```
Traceback (most recent call last):
File "train.py", line 230, in
…
-
ICNet for Real-Time Semantic Segmentation on High-Resolution Images
https://hszhao.github.io/projects/icnet/
-
Thanks too much for your sharing code!
I trained the icnet on cityscapes, but best mIoU I got was around 58%
My config:
aux=True,
backbone='resnet50',
base_size=520,
batch_size=4,
crop_size…