cfg and weights file

[Micky-123]

Hi @DoranLyong , Your work looks very interesting. I would like to experiment with few images. Can you please point to the place where the darknet .cfg file and the weights are present.

Thanks

[DoranLyong]

Hi @Micky-123, you seem to want to know where the configurations are for Darknet setup.

I recommend referring to this link, which I totally followed. The .cfg file is only for training, so it is not included in this repository.

The weight file is located here, which was converted from .weights to. This is because after training my darknet yolov4-tiny model, I employed it for tensorflow lite version. Check this repository.

[Micky-123]

Thanks @DoranLyong .

I am looking for the Darknet weights and the darknet cfg file. You would have used to generate the TFLite model. Can you please point to that if you have them

BTW - I am getting the following error while executing "python run_person_detector.py --image ./data/persons.jpg --score 0.75": \yolov4-tiny-tflite-for-person-detection\core\utils.py", line 152, in draw_bbox for i in range(num_boxes): TypeError: only integer scalar arrays can be converted to a scalar index Did you come across this any time?

Thanks

[DoranLyong]

Hi, @Micky-123. I found out some code block makes wrong in building bounding boxes.
I'm gonna fix it soon.

I involved the link for both configuration setups:

• yolov4-tiny-custom.cfg
• yolov4-tiny-custom_last.weights

You could employ them for your own model.

[Micky-123]

Hi, @Micky-123. I found out some code block makes wrong in building bounding boxes.

I'm gonna fix it soon.

I involved the link for both configuration setups:

• yolov4-tiny-custom.cfg
• yolov4-tiny-custom_last.weights

You could employ them for your own model.

Thanks @DoranLyong! Appreciate it. Looking forward for the bug fix.

[DoranLyong]

@Micky-123, I updated the training methods in this repository

You could train your model with your custom person dataset, following the instructions :)

[Micky-123]

@Micky-123, I updated the training methods in this repository

You could train your model with your custom person dataset, following the instructions :)

Thanks DoranLyong!