My tflite file cannot run in your YOLOv9

[lucky-baby-dog]

Hello, I changed v9 in your file to my model file, but when I recompile and run it again, it doesn't open. What's going on

[lucky-baby-dog]

My tflite file can be opened in Netron software and displayed

[surendramaran]

Any error it is showing in logcat when you try to run?
If not then in which way the model is behaving?

[lucky-baby-dog]

Brother, the problem lies here. How can we solve it

[surendramaran]

May I know the output and input shape. visit https://netron.app/ and get properties.

[lucky-baby-dog]

May I know the output and input shape. visit https://netron.app/ and get properties.

[surendramaran]

Yeah, this YOLOv9 model is probably exported and trained from here https://github.com/WongKinYiu/yolov9

While I took models from here https://docs.ultralytics.com/models/yolov9/

So, you may also use ultralytics's environment and use this repo.

I will also try to interpret your model, if successful I will publish codes.

[lucky-baby-dog]

Yeah, this YOLOv9 model is probably exported and trained from here https://github.com/WongKinYiu/yolov9

While I took models from here https://docs.ultralytics.com/models/yolov9/

So, you may also use ultralytics's environment and use this repo.

I will also try to interpret your model, if successful I will publish codes.
Yes, brother. I trained my data based on the YOLOv9 model and obtained a weight file. I converted this file into a PB file using onnx and then into a TFLite file

[surendramaran]

You mind sharing your tflite model, with labels and sample images? [email protected]

[lucky-baby-dog]

You mind sharing your tflite model, with labels and sample images? [email protected]

Hello, I have packaged the items and placed them in the data file，My email address is [email protected]

[yg-dickson]

Does it support yolo11?

[lucky-baby-dog]

Does it support yolo11?

I haven't tried it before, you can give it a try

[Woo-wdy]

yolo11模型部署到yolov8上直接闪退，我用的是yolo11x的模型转成了yolo11_float16.tflite,把他放在yolov8或者yolov9的android里并调用会直接闪退

[lucky-baby-dog]

yolo11模型部署到yolov8上直接闪退，我用的是yolo11x的模型转成了yolo11_float16.tflite,把他放在yolov8或者yolov9的android里并调用会直接闪退

Yes, it did indeed crash directly. I am also looking for the reason behind it

[Woo-wdy]

我找到了原因 把GPU加速关掉就好了 但是识别速度会很慢

[lucky-baby-dog]

我找到了原因 把GPU加速关掉就好了 但是识别速度会很慢

我找到了原因 把GPU加速关掉就好了 但是识别速度会很慢

I switched to my YOLOV9 model and ran it inside YOLOV9, but it crashed and the software couldn't open