How to run inference using the inference APIs?

[akshay-k2]

Hi,

I'm able to run the test script with the trained model and obtain the inference results for BDD100K. However, I'm unable to figure out how to run inference for a single image in the dataset.

I see there are inference APIs available in qdtrack/apis/inference.py. I'm able to initialize the model by calling init_model, however, I'm unsure what to pass as frame_id for calling inference_model(model, imgs, frame_id)

Any help will be highly appreciated. Thank you.

[ghost]

@OceanPang can you please clarify the usage of the inference API?

[jbohnslav]

I'm not involved in the development of this project. However, I believe the frame_id is just the frame number for those imgs. I assumed this on my data with good results.