[New feature] Support services for vector similarity retrieval

jingkl commented 3 years ago

Feature Request

If this is a feature request, please fill out the following form in full:

Describe the problem the feature is intended to solve

Tesnorflow Serving is a model serving system for machine learning that trains and acquires models and integrates with Tensorflow models. After integrating the models, it is often possible to get vectors, which represent various features (can be features of text, images, audio, video), and Milvus can then provide vector storage and similarity retrieval services, so it is hoped that Milvus and Tensorflow serving can be deeply combined to provide more services to users

Describe the solution

Hopefully, the following things can be done in tensorflow serving to help add vector retrieval service to tensorflow serving:

Provide the introduction of Milvus
Provide the Milvus tool folder Milvus implements vector import and vector query service, we will provide a file milvus_tool, the script under this folder provides two functions to store and recall vectors using Milvus
Pull request to the tensorflow service
Merge into the tensorflow serving GitHub project
Add tutorials or use cases

Describe alternatives you've considered

Might be able to contribute Milvus source code

Additional context

What is Milvus : https://www.milvus.io/docs/overview.md Milvus Github: https://github.com/milvus-io/milvus/

I hope we can communicate further and look forward to further cooperation with the tensorflow serving team.

windmaple commented 3 years ago

For vector similarity search, I think you can maintain your own repo (w/o being coupled to Serving), right? It's how Google's ScaNN has done it anyway.

jingkl commented 3 years ago

For vector similarity search, I think you can maintain your own repo (w/o being coupled to Serving), right? It's how Google's ScaNN has done it anyway.

We would like to make it easier for users in the tf community to use the vector similarity retrieval tool Milvus, and for users in the Milvus community to use tensorflow serving. We have already cooperated with paddlepaddle, and hope to cooperate with tensorflow serving community in depth.

xiaofan-luan commented 2 years ago

any progress on this proposal? If there is a way to have a connector that seems to be another solution?

singhniraj08 commented 1 year ago

@jingkl,

Are you still looking for a resolution? We are planning on prioritising the feature issues based on the community interests. Please let us know if this issue still persists with the latest TF Serving 2.12.1 release so that we can work on it's implementation. Meanwhile, Please feel free to submit a PR if you are willing to contribute this feature implementation. Thank you for your contributions.

github-actions[bot] commented 1 year ago

This issue has been marked stale because it has no recent activity since 7 days. It will be closed if no further activity occurs. Thank you.

github-actions[bot] commented 1 year ago

This issue was closed due to lack of activity after being marked stale for past 7 days.

tensorflow / serving