Template for deploying any HuggingFace pipeline for supported tasks with Torchserve

Hi! Lately I have been working on my repo that contains a template to deploy any Huggingface model supported by the pipeline, where pipeline is a simple-to-use abstraction provided by HF. The repo also includes copy-paste commands in READMEs for AWS EC2 instance.

Although I have only focused on deploying model with PyTorch backend as I would be adding scripts to deploy Torchscripted & LLM.Int8 pipeline models soon. Moreover, TF models present in an HF repo (example of an HF repo) can also be deployed by changing the framework attribute while initialising pipeline.

I have also tried to make the repo as beginner friendly as possible by including comments, references and compact code. There are also plans to integrate the HuggingFace optimum library that integrates elegantly with pipeline, so by extension, it would integrate well with my repo too with a few short scripts.

My repo could be useful to the open-source community and I believe it would reach a greater audience if added to the News section and/or examples/Huggingface_Transformers.

Thanks.

[msaroufim]

We can split the work into

1. Add support for pipeline in existing handler code, refactor and simplify as needed
2. Docker support for large models with its own sub folder

[msaroufim]

Ok that clarifies things, thank you. please feel free to make your PR directly. I think for now we can focus on improving the support for our HF models and once the PR is in we can work together to publicize the work. Please add @HamidShojanazeri and myself as reviewers for your PRs