Change code to support yolov8

[modu82]

Onnx conversion Tensorrt updated support for YOLOv8，How should I modify the export code

[modu82]

I tried to remove the encryption in the code and directly import the converted model, but it was not successful

[eka2122]

hey bro do you have discord? I need help

[modu82]

hey bro do you have discord? I need help

What can I help you with

[modu82]

i try change this code

with open(engine_path, "rb") as f:
            serialized_engine = fw.decrypt(f.read())

        engine = runtime.deserialize_cuda_engine(serialized_engine)

but engine is not Unable to obtain correctly serialized_engine

[eka2122]

hey bro do you have discord? I need help

What can I help you with

about the config, my aimbot can't work? even though everything is running according to the instructions

[modu82]

hey bro do you have discord? I need help

What can I help you with

about the config, my aimbot can't work? even though everything is running according to the instructions

I have some basic debugging methods: Set visualization to true Set Apex to Window Mode Check for normal operation in visualization

[modu82]

hey bro do you have discord? I need help

What can I help you with

about the config, my aimbot can't work? even though everything is running according to the instructions

and If your terminal outputs locking information normally, then the program should be working properly

[eka2122]

hey bro do you have discord? I need help

What can I help you with

about the config, my aimbot can't work? even though everything is running according to the instructions

and If your terminal outputs locking information normally, then the program should be working properly

i am bad in english can you give me your aimbot configs?

[eka2122]

hey bro do you have discord? I need help

What can I help you with

about the config, my aimbot can't work? even though everything is running according to the instructions

and If your terminal outputs locking information normally, then the program should be working properly

i am bad in english can you give me your aimbot configs?

中文会不

I'm from Indonesia

[modu82]

hey bro do you have discord? I need help

What can I help you with

about the config, my aimbot can't work? even though everything is running according to the instructions

and If your terminal outputs locking information normally, then the program should be working properly

i am bad in english can you give me your aimbot configs?

中文会不

I'm from Indonesia

Normally, the configuration file does not need to be modified，Please pay attention to the terminal information output after running Apex.py

[modu82]

hey bro do you have discord? I need help

What can I help you with

about the config, my aimbot can't work? even though everything is running according to the instructions

and If your terminal outputs locking information normally, then the program should be working properly

i am bad in english can you give me your aimbot configs?

More information is needed to know where the problem lies

[eka2122]

hey bro do you have discord? I need help

What can I help you with

about the config, my aimbot can't work? even though everything is running according to the instructions

and If your terminal outputs locking information normally, then the program should be working properly

i am bad in english can you give me your aimbot configs?

More information is needed to know where the problem lies

yes the terminal is not running! what is the button to enable aimbot by default

[PlutoNameless]

To the best of my knowledge, the YOLOv5's ONNX doesn't include NMS (Non-Maximum Suppression), and I'm not entirely sure if there have been any adjustments made in YOLOv8. In YOLOv7, we used NMS End 2 End for ONNX export. If you need to make other adjustments, you might have to consider implementing NMS.

[modu82]

I encountered this problem when using my own model

Traceback (most recent call last):
  File "D:\Python\APEX_AIMBOT\apex.py", line 17, in <module>
    apex = ApexAimBot(config_path='configs/apex.yaml', onnx_path='weights/best_apex.onnx', engine_path='weights/best_apex.trt')
  File "D:\Python\APEX_AIMBOT\apex.py", line 8, in __init__
    super().__init__(config_path, onnx_path, engine_path)
  File "D:\Python\APEX_AIMBOT\AimBot.py", line 38, in __init__
    self.speed_test()
  File "D:\Python\APEX_AIMBOT\AimBot.py", line 106, in speed_test
    self.engine.inference(img)
  File "D:\Python\APEX_AIMBOT\utils\InferenceEngine.py", line 57, in inference
    num, final_boxes, final_scores, final_cls_inds = self.forward(img)
ValueError: not enough values to unpack (expected 4, got 1)

[modu82]

To the best of my knowledge, the YOLOv5's ONNX doesn't include NMS (Non-Maximum Suppression), and I'm not entirely sure if there have been any adjustments made in YOLOv8. In YOLOv7, we used NMS End 2 End for ONNX export. If you need to make other adjustments, you might have to consider implementing NMS.

I encountered this problem when using my own model，Do I need any special operations when building the model

Traceback (most recent call last):
  File "D:\Python\APEX_AIMBOT\apex.py", line 17, in <module>
    apex = ApexAimBot(config_path='configs/apex.yaml', onnx_path='weights/best_apex.onnx', engine_path='weights/best_apex.trt')
  File "D:\Python\APEX_AIMBOT\apex.py", line 8, in __init__
    super().__init__(config_path, onnx_path, engine_path)
  File "D:\Python\APEX_AIMBOT\AimBot.py", line 38, in __init__
    self.speed_test()
  File "D:\Python\APEX_AIMBOT\AimBot.py", line 106, in speed_test
    self.engine.inference(img)
  File "D:\Python\APEX_AIMBOT\utils\InferenceEngine.py", line 57, in inference
    num, final_boxes, final_scores, final_cls_inds = self.forward(img)
ValueError: not enough values to unpack (expected 4, got 1)

[xiazaipw]

hi bro did you solve it? I also encountered this problem.Unable to use own model ONNX, convert TRT

[Tr3yWay996]

Onnx conversion Tensorrt updated support for YOLOv8，How should I modify the export code

So ? anything new ? any compatibility with AMD gpus ?