BentoML 0.8.4 release!

[parano]

BentoML 0.8.4 has just arrived with tons of new features and improvements!

https://github.com/bentoml/BentoML/releases/tag/v0.8.4

Thank you all for your contributions to this release! @bojiang @joshuacwnewton @guy4261 @Sharathmk99 @co42 @jackyzha0 @Korusuke @akainth015 @omrihar @yubozhao

A few quick questions for the community:

• What is your favorite feature in 0.8.4?
• Any questions about the new features or the breaking changes in this release?
• What are the features in the issues list that you want the most and why?

[omrihar]

Good job! I especially like the separation of training and bentoml service creation :)

I'm really excited about the future of BentoML!

[yubozhao]

Love the Ngrok integration, super easy to demo in a meetup now.

[parano]

Also when working with Google Colab or a remote Jupyter Notebook environment, this makes it possible to try out the API server easily!

[withsmilo]

Great job!
I have been waiting a long time for bentoml containerize command to be released!

[parano]

We have a few more model management related features coming in next release! The goal is to make it easy to push locally saved bundles to a remote YataiService, pulling down saved bundles from remote YataiService to local Yatai, and allow the user to prune locally saved bundles to free up space on the dev machine.

This should make collaborating and sharing models more easily in a team setting, and simplify building model training pipelines and CI/CD pipelines with BentoML.

[HenryDashwood]

You guys are on fire right now!