[Fix][Transform-V2] Reduce embedding precision from double to float

[xiaochen-zhou]

Purpose of this pull request

Reduce embedding precision from double to float，close #9611

Does this PR introduce any user-facing change?

no

How was this patch tested?

Exists tests

Check list

• If any new Jar binary package adding in your PR, please add License Notice according
  New License Guide
• If necessary, please update the documentation to describe the new feature. https://github.com/apache/seatunnel/tree/dev/docs
• If you are contributing the connector code, please check that the following files are updated:
  1. Update plugin-mapping.properties and add new connector information in it
  2. Update the pom file of seatunnel-dist
  3. Add ci label in label-scope-conf
  4. Add e2e testcase in seatunnel-e2e
  5. Update connector plugin_config

[xiaochen-zhou]

I think we can start by reducing the embedding precision from double to float. The precision loss isn’t just happening with Zhipu—it’s actually an issue with almost all models where the embedding type is returned as double, like

Qianfan

openai model：

So, as a quick fix, we can switch to float for now and add a note in the docs to let users know. @Hisoka-X

[Hisoka-X]

So, as a quick fix, we can switch to float for now and add a note in the docs to let users know. @Hisoka-X

+1. Next step, we should support double vector type.

[Hisoka-X]

Thanks @xiaochen-zhou . Could you add a test case to cover it?

[xiaochen-zhou]

Thanks @xiaochen-zhou . Could you add a test case to cover it?

OK.

[loupipalien]

So, as a quick fix, we can switch to float for now and add a note in the docs to let users know. @Hisoka-X

+1. Next step, we should support double vector type.

@Hisoka-X @xiaochen-zhou Another question, is there a plan to support multimodal embeddings?https://www.volcengine.com/docs/82379/1523520

[xiaochen-zhou]

So, as a quick fix, we can switch to float for now and add a note in the docs to let users know. @Hisoka-X

+1. Next step, we should support double vector type.

@Hisoka-X @xiaochen-zhou Another question, is there a plan to support multimodal embeddings?https://www.volcengine.com/docs/82379/1523520

I think this suggestion is great, and I would be happy to try implementing it. @Hisoka-X

[Hisoka-X]

So, as a quick fix, we can switch to float for now and add a note in the docs to let users know. @Hisoka-X

+1. Next step, we should support double vector type.

@Hisoka-X @xiaochen-zhou Another question, is there a plan to support multimodal embeddings?https://www.volcengine.com/docs/82379/1523520

+1

[Hisoka-X]

Thanks @xiaochen-zhou