Mlflow roadmap updated and posted!
99 views
Skip to first unread message
Ben Wilson
Jun 21, 2022, 2:06:05 PM6/21/22
to mlflow-users
Hello MLflow community!
Late last year we went through the process of identifying long-outstanding feature requests that we felt would bring real value to MLflow if they were implemented. After presenting this list of items, you, the community of MLflow users and developers, came out in force and implemented every single one of them.
These new features that were built helped to create a more functional and robust MLOps tool ecosystem. On behalf of the wider MLflow community, we'd like to thank you for those incredible contributions.
Because of the resounding success of the last roadmap, we wanted to open up a new roadmap with a broadened scope, continuing to focus on the features that the community has asked for most.
The roadmap's items can be found here.
The features for this roadmap milestone are broader and some are more enthusiastically ambitious than our last go-around. They range from simple implementations that are geared towards those who wish to get started with open source software contribution to complex design-required features that will dramatically improve the scalability and production-capable deployment of MLflow.
We are incredibly excited to see the implementations that you're going to come up with and look forward to collaborating with you to help build a more robust MLOps tool that everyone can freely use to simplify and enhance their ML development experience.
Some noteworthy items on this roadmap are:
- Adding tag-based search functionality
- Support an 'or' search operator
- Container images for MLflow servers
- Swagger docs (OpenAPI standards) for MLflow APIs
- Support for custom container images in serving
- Support NoSQL as a backend for MLflow Tracking
- An overhaul to metrics and parameters displays in run comparison UI
- End-to-end getting started guides
We may add a few items to this list over the next few months if we feel that there are significant new features that we want to prioritize getting merged. Some tasks may be taken on by maintainers if there are no community volunteers willing to work on them.
For interested first-time contributors, please read through the contributing guide. If you're interested in getting an MLflow development environment setup fairly quickly (and painlessly), see the development environment setup script guide.
Late last year we went through the process of identifying long-outstanding feature requests that we felt would bring real value to MLflow if they were implemented. After presenting this list of items, you, the community of MLflow users and developers, came out in force and implemented every single one of them.
These new features that were built helped to create a more functional and robust MLOps tool ecosystem. On behalf of the wider MLflow community, we'd like to thank you for those incredible contributions.
Because of the resounding success of the last roadmap, we wanted to open up a new roadmap with a broadened scope, continuing to focus on the features that the community has asked for most.
The roadmap's items can be found here.
The features for this roadmap milestone are broader and some are more enthusiastically ambitious than our last go-around. They range from simple implementations that are geared towards those who wish to get started with open source software contribution to complex design-required features that will dramatically improve the scalability and production-capable deployment of MLflow.
We are incredibly excited to see the implementations that you're going to come up with and look forward to collaborating with you to help build a more robust MLOps tool that everyone can freely use to simplify and enhance their ML development experience.
Some noteworthy items on this roadmap are:
- Adding tag-based search functionality
- Support an 'or' search operator
- Container images for MLflow servers
- Swagger docs (OpenAPI standards) for MLflow APIs
- Support for custom container images in serving
- Support NoSQL as a backend for MLflow Tracking
- An overhaul to metrics and parameters displays in run comparison UI
- End-to-end getting started guides
We may add a few items to this list over the next few months if we feel that there are significant new features that we want to prioritize getting merged. Some tasks may be taken on by maintainers if there are no community volunteers willing to work on them.
For interested first-time contributors, please read through the contributing guide. If you're interested in getting an MLflow development environment setup fairly quickly (and painlessly), see the development environment setup script guide.
We're looking forward to collaborating!
Reply all
Reply to author
Forward
0 new messages