CPU Inference not working with optimized onnx cpu model

[MarcelCode]

I want to create an optimized model for ONNX CPU.

My code for exporting the model:

model_path = "./model/raw"

model = ModelManager.get(model_path)

runtime = RuntimeType.ONNX_CPU

optimized_model = model.export(runtime_type=runtime,
                               image_size=640,
                               device="cpu",
                               out_dir="./model/cpu",
                               overwrite=True)

For inference I am using the following code:

infer_model = InferModel(model_dir="./model/cpu",
                             runtime_type=RuntimeType.ONNX_CPU)

result = infer_model.infer("./image/raw/20230703_102957_38_2490-2093-1352-12.png",
                           annotate=True)

I am getting the following error:
RuntimeError: Found no NVIDIA driver on your system. Please check that you have an NVIDIA GPU

Am I doing sth wrong?

I saw that in infer/infer_model.py line 178 the device is hardcoded to cuda. If I change it to "cpu" it works.

Can you confirm that this is a bug or have I initialized my onnx cpu model in a wrong way?

[CuriousDolphin]

Hello! yes i confirm this is a bug, will be fixed in the next release

[CuriousDolphin]

Hi @MarcelCode , checkout new release (0.19.1): https://github.com/FocoosAI/focoos/releases/tag/v0.19.1
let us know if the bug is now fixed.

[MarcelCode]

Works thanks!