Tool/software:
Dear TI Team,
I hope this message finds you well.
I am currently working on getting the YOLOv8 object detection model running on the TDA4VM. Initially, I faced compilation issues, but after modifying the activation function to ReLU and setting the kernel shape to 3, the model now compiles. However, I am encountering errors during the execution phase. I have attempted compilation using both the TIDL tools and EdgeAI benchmark, but I am receiving errors in both environments.
Previously, I reached out to one of the TI officials regarding this issue but unfortunately did not receive a response. I am attaching the modified YOLOv8 architecture and the logs from both the TIDL tools and EdgeAI benchmark for your reference.
I would greatly appreciate your assistance in resolving these issues.
Thank you in advance for your support.
Best regards,
Ramaseshan
Model artifacts generated using benchmark:
5707.yolov8_model_artifacts.zip
tidl tools compilation error:
benchmark compilation error: (both compilation & it's inference - where you could see accuracy is 0%)
benchmark_logs_run_inference(after compilation).txt
benchmark_logs_run_import(compilation).txt
Architecture changed in yolov8 which made the model to compile:
