Skip to content

Slides and notebook for the workshop on serving bert models in production

Notifications You must be signed in to change notification settings

npatta01/pytorch-serving-workshop

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

24 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Readme

Overview

This repo contains notebooks for Pytorch Serving Workshop.

Note: We do not need a GPU runtime

Setup

If you came to this repo, during a workshop visit this custom jupyter hub with all the dependencies already set up.

Otherwise, consider using Binder

Contents

There are five notebooks.

a. 00_prepare_dataset.ipynb

Notebook that prepares the e-comeerce dataset and saves it.

b. 01_train.ipynb

Trains a DistilBert model

c. 02_inference_review.ipynb

Notebook that shows how to use the HuggingFace ecosystem. Also shows how to use the trained model from previous notebook.

d. 03_optimizing_model.ipynb

Notebook that shows impact of Quantization and TorschScript

e. 04_packaging.ipynb

Notebook that shows how to use TorchServe to serve models

Slides

Watch the video

Video

PyData Video

References

Pydata 2021 Slides

Pydata 2021 Conference Page

Libraries

This repro uses HuggingFace transformers and dataset pacakge.

The dataset used is Amazon Berkeley Objects (ABO) Dataset created by Amazon and UC Berkeley. For more reference, refer to this paper

Contact

For help or feedback, please reach out to :

About

Slides and notebook for the workshop on serving bert models in production

Topics

Resources

Stars

Watchers

Forks

Packages

No packages published

Languages