-
can you please make a code for these models? thanks in advance
-
Hi everyone,
**Proposal:**
I would like to propose the addition of several other activation functions to the framework:
- [ ] LeakyReLU
- [ ] PReLU
- [ ] ReLU6
- [ ] Tanh
- [ ] Softplus
…
-
I try to change stride in yolov4-mish cfg to detect small object like u guided in yolov4 cfg.
Img_size : (640, 1632), box (about 10 x 10). This is my config file [yolo4-mish.txt](https://github.com/A…
-
Couldn't find activation function mish, going with ReLU
I've been getting this error whenever I try to run the new YoloV4
-
@AlexeyAB
-
The activation function of res_unit (residual_block) In darknet from AlexeyAB's repo is mish, but in this repo is leaky. Is it right here to use leaky?
-
大佬您好,感谢您分享您的代码。
最近把您最新的master分支内的pscsp-x-mish的模型在我们自己的数据集上跑了一遍,但是效果跟u5分支里的yolov4x-mish差了一些。17轮训练的对比效果如下:
yolov4-mish:P=0.412 R=0.6644 mAP=0.5058
pacsp-x-mish:P=0.3575 R=0.6297 mAP=0.446…
-
Dear Wong!
I was trying to reproduce some results from your repository from scratch, but I failed with this :(
Can you give the cue what might have gone wrong?
Here is the detailed algorithm:…
-
### 🐛 Describe the bug
RuntimeError occurs during `torch.allclose` assertion, indicating a size mismatch between tensors at dimension 3 (size 9 vs. size 3) in a custom model's forward method. The err…
-
You could increase SMIM accuracy by using Ranger, which combine state of the art optimizers + gradient centralization
https://github.com/lessw2020/Ranger-Deep-Learning-Optimizer
Hortogonally, you …