-
I exported a 73-label multilabel network. It was made with Keras using sigmoid outputs. I get very different outputs from each backend, and they are all wrong.
| Toolkit | True | False |
|-------…
-
why I run the code,the accuracy is better than the table.
I have nothing changed but resnet50's top1 result is 20.39 better than the table 22.51
-
xiole updated
6 months ago
-
```
class SEResNet_IR(nn.Module):
def __init__(self, num_layers, feature_dim=512, drop_ratio=0.4, mode = 'ir'):
super(SEResNet_IR, self).__init__()
assert num_layers in [50, 10…
-
关注博主很久啦,感谢分享,最近刚好有一个超分的项目正在进行。想问问,有没有实现论文中SRresNet部分的模型。可以分享吗~谢谢!
-
-
Mish is a new novel activation function proposed in this [paper](https://arxiv.org/abs/1908.08681).
It has shown promising results so far and has been adopted in several packages including:
- [Ten…
-
Mish is a new novel activation function proposed in this [paper](https://arxiv.org/abs/1908.08681).
It has shown promising results so far and has been adopted in several packages including:
- [Ten…
-
Traceback (most recent call last):
File "se_resnet.py", line 408, in
model = SEResNet50()
File "se_resnet.py", line 213, in SEResNet50
classes=classes)
File "se_resnet.py", line 13…
-
Mish is a new novel activation function proposed in this [paper](https://arxiv.org/abs/1908.08681).
It has shown promising results so far and has been adopted in several packages including:
- [Ten…