Hi there! I am trying to use one of the SDK supported models with NNComponent in create_nn() (mentioned here [AI models — DepthAI SDK Docs 1.13.1 documentation (luxonis.com)](https://docs-old.luxonis.com/projects/sdk/en/latest/features/ai_models/)) which I have the bin and blob file for locally. Each time I run though, it tries to download them from the internet. Is there a way to run the model without needing the internet every time and instead using the local files?

@"hatcob"#p25868 How are you creating the NNComponent? It should first search through the local files for the model. They are under https://github.com/luxonis/depthai/tree/tof_calib/depthai_sdk/src/depthai_sdk/nn_models Thanks, Jaka

Hi @"hatcob"#p25915 Which model? The function that converts string to find the model is here: https://github.com/luxonis/depthai/blob/main/depthai_sdk/src/depthai_sdk/components/nn_component.py#L242-L247 Thanks, Jaka

Using local .bin .blob files

hatcob

Hi there! I am trying to use one of the SDK supported models with NNComponent in create_nn() (mentioned here AI models — DepthAI SDK Docs 1.13.1 documentation (luxonis.com)) which I have the bin and blob file for locally. Each time I run though, it tries to download them from the internet. Is there a way to run the model without needing the internet every time and instead using the local files?

jakaskerl

hatcob
How are you creating the NNComponent? It should first search through the local files for the model.
They are under luxonis/depthaitree/tof_calib/depthai_sdk/src/depthai_sdk/nn_models

Thanks,
Jaka

hatcob

Thanks @jakaskerl for the response! I was using .create_nn() passing in the string name of the model. The bin and blob file of that name are in the same folder. Should I instead be passing config.json file with path to blob in it?

Thank you!

jakaskerl

Hi hatcob
Which model?
The function that converts string to find the model is here:
luxonis/depthaiblob/main/depthai_sdk/src/depthai_sdk/components/nn_component.py#L242-L247

Thanks,
Jaka