Convert onnx to hef using hailomz

[ruslan]

Hi, i tried to convert onnx file trained using custom dataset, but encounter this error when im using the hailomz command:

raise NegativeSlopeExponentNonFixable(
hailo_model_optimization.acceleras.utils.acceleras_exceptions.NegativeSlopeExponentNonFixable: Quantization failed in layer yolov8s/conv42 due to unsupported required slope. Desired shift is 9.0, but op has only 8 data bits. This error raises when the data or weight range are not balanced. Mostly happens when using random calibration-set/weights, the calibration-set is not normalized properly or batch-normalization was not used during training.`Preformatted text`

[Nadav]

Please check this topic:

Problem With Model Optimization

I might have a solution, but I’m not sure if it’s general, but I think that it worth checking. It seems that on datasets that have low number of classes and/or low number of pictures, some nodes on the YOLOv8 are getting almost nullified. I’ve used this command to force a wider dynamic range on the outputs: quantization_param([conv42, conv53, conv63], force_range_out=[0.0, 1.0]) This is relevant for YOLOV8s trained from our retraining dockers. These are the end nodes before the NMS layer.

[ruslan]

Problem solved, Thank you for your response.

During the training i noticed that the compilation doesn’t utilize the GPU which i verified using the “nvidia-smi”. Any workaround to utilize the GPU?

[info] Starting Model Optimization
[warning] Reducing optimization level to 0 (the accuracy won't be optimized and compression won't be used) because there's less data than the recommended amount (1024), and there's no available GPU

[Nadav]

Hi @ruslan, the compiler doesn’t make use of the GPU, only the quantizer\optimizer.

[nina-vilela]

@ruslan not identifying the GPU is usually caused by incompatible CUDA versions.

You can find the DFC system requirements in its guide: