-
I am trying to do training on custom data. After training i got file yolo-obj_last.weights.
After this i typed command :
darknet.exe detector test data/obj.data cfg/yolov4-obj.cfg yolov4-obj_last.we…
-
### Search before asking
- [X] I have searched the YOLOv3 [issues](https://github.com/ultralytics/yolov3/issues) and [discussions](https://github.com/ultralytics/yolov3/discussions) and found no simi…
-
-
Can you explain why have you used the following constants? I have inspected a few different yolov3 implementation but none had a similar operation.
https://github.com/ultralytics/yolov3/blob/a284fc…
-
I tried to transform keras model to onnx, and this is my code,
```
onnx_model = keras2onnx.convert_keras(self.model, self.model.name)
temp_model_file = './mymodel.onnx'
onnx.save_model(onnx_model,…
-
For yolov4, @AlexeyAB suggests to do the following modification in order to detect objects smaller 16px:
"
for training for small objects (smaller than 16x16 after the image is resized to 416x416) -…
-
### Search before asking
- [X] I have searched the YOLOv8 [issues](https://github.com/ultralytics/ultralytics/issues) and [discussions](https://github.com/ultralytics/ultralytics/discussions) and fou…
-
Command used for training with single GPU:
python3 train.py --batch-size 16 --img 416 416 --data BRA-Dataset.yaml --cfg yolov4-p5.yaml --weights yolov4-p5_.pt --device 0 --name yolov4-p5BRA-Dataset…
-
### Describe the problem you are having
Awesome job getting OpenVINO into Frigate @NateMeyer! I just got a thin client with a J5005, and OpenVINO support was perfectly timed!
I tried using the ``…
-
This error comes out when training with multiple gpus. Go to `build_target` function in the file `utils/general.py`.
Add `anchors = anchors.to(targets.device)` right below the line
`anchors = model.…