TorchServe: Access different models with "--models all" option

deployment
1

I’d like to create a TorchServe handler that draws on 3 models, but I’m not clear on how to access/identify the different models in the script, when starting TorchServe with the --models all option.

Any help appreciated.

2

Hi @jbm we built our workflow feature to handle such use cases, let me know if this works for you but you’ll be able to define how the 3 models interact in a YAML script https://github.com/pytorch/serve/tree/master/examples/Workflows

3

Ah, okay, I’ll have a look. Thanks!