Triton inference server cannot find model in 'Disaster Risk Monitoring Using Satellite Imagery' course

I’m following the ‘Disaster Risk Monitoring Using Satellite Imagery’ course.

Everything worked fine until the part where a triton inference server is involved.
(chapter 03_model_deployment_for_inference)

My file struture is correct and the document says that a triton inference server is running in polling mode, to update model definitions every 30s.

It seems that the triton server is running (I get http_200’s if I curl it) but somehow it doesn’t pick up my changes.
When I curl -v triton:8000/v2/models/flood_segmentation_model, it doesn’t find the model

How could I know why it doesn’t pick up models changes and could I restart it in polling mode from the console ?

Thanks for your help