Reusing your custom onnx

[Schlockschlock]

Hello,

I'm starting to learn Python and I'm trying to do myself a hcaptcha solver like you, for training. For the "new" challenges (bridge, rooms, horses, etc.) I understand that you're using custom trained models (bridge.onnx, lions.onnx, etc.). So I would like to reuse those custom ONNX file for the specific challenges. That's why I'm trying to read your code to find the 2 most important parts:

• where you're loading those ONNX custom model files
• and where you're performing the object detection with the boxes coordinates

As far as I understand, those 2 functions are located in the /src/services/hcaptcha_challenger/solutions/yolo.py right?
Thank you

[QIN2DIM]

where you're loading those ONNX custom model files

• YOLO:

  hcaptcha-challenger/src/services/hcaptcha_challenger/solutions/yolo.py

  Line 135 in 59ff1af

  self.net = cv2.dnn.readNetFromONNX(self.onnx_model["path"])

• ResNet:

  hcaptcha-challenger/src/services/hcaptcha_challenger/solutions/resnet.py

  Line 41 in 59ff1af

  self.net = cv2.dnn.readNetFromONNX(self.onnx_model["path"])

where you're performing the object detection with the boxes coordinates

• YOLO:

  hcaptcha-challenger/src/services/hcaptcha_challenger/solutions/yolo.py

  Line 189 in 59ff1af

  def solution(self, img_stream: bytes, label: str, **kwargs) -> bool:

• ResNet:

  hcaptcha-challenger/src/services/hcaptcha_challenger/solutions/resnet.py

  Line 95 in 59ff1af

  def solution(self, img_stream, **kwargs) -> bool:

[Schlockschlock]

Thank you very much

[QIN2DIM]

🥂