-
Is it possible to export ONNX with dynamic batch size in yolov4-large?
Is there any performance disadvantage to doing so?
-
**Describe the feature**
Add Cross-Iteration Batch Normalization in: https://arxiv.org/abs/2002.05712
And Accumulate Gradient for training: https://github.com/WongKinYiu/ScaledYOLOv4/blob/yolov4-l…
-
In repository @AlexeyAB mentioned for detecting small & large objects use this [https://raw.githubusercontent.com/AlexeyAB/darknet/master/cfg/yolov4-custom.cfg](url) custom.cfg file
but i think we…
-
Hi,
Thanks a lot for the wonderful paper and the github repo. Also, congrats for getting the highest accuracy.
In the paper, you mentioned: "_The time used for training YOLOv4-tiny is 600 epoch…
-
I have successfully trained "yolov4-tiny" on the vehicle data set in "bdd 100k" (training: 69133 vehicle images, verification: 9902 vehicle images).
In order to compare and detect I use [NanoDet-m ](…
KKDDD updated
3 years ago
-
### Experimentation Details
* Trained a yolov4 model M1 and extracted the first 137 layers of it using the partial command. Lets call the first 137 layers as yolov4.conv.137.custom
* Trained another…
-
這幾天嘗試透過yolov4-csp的branch進行COCO 2017 dataset訓練,為了讓yolov4-tiny能過正常訓練,我在[parse_config.py](https://github.com/WongKinYiu/ScaledYOLOv4/blob/yolov4-csp/utils/parse_config.py)中的[supported](https://github.com…
-
- [yolov4-csp-x-swish.cfg](https://raw.githubusercontent.com/AlexeyAB/darknet/master/cfg/yolov4-csp-x-swish.cfg) - 640x640 - **69.9% mAP@0.5 (51.5% AP@0.5:0.95) - 23(R) FPS / 50(V) FPS** - 221 BFlops …
-
Thank you for your repo, it's a great work.
The COCO mAP in YOLOV4 paper is 43.5, but you can achieve 48.5, why is performance improved so much? the large image size?
-
In the YOLOv4 Darknet, there were different config files for small and large objects in an image. https://github.com/AlexeyAB/darknet#how-to-improve-object-detection this link shows that
Is it some…