Deploying FastSAM using Torchserve

[akshay-iyer]

First of all, let me thank and congratulate the authors for this wok. I feel it is a smart way of combining what SAM and the existing semantic segmentation work has done. I'm looking to deploy FastSAM using Torchserve and for the same, the model needs to be jit scripted or jit traced.

Currently I'm unable to perform as both typically require a single model file with the nn.Module defined and a forward function implemented.

Has anyone deployed or scripted the FastSAM model? Any help will be highly appreciated.