feat: Cortex API supports /embeddings

[freelerobot]

Related issues:

• feat: Custom OpenAI compatible server support with preserved inference parameters jan#2840

[dan-menlo]

@vansangpfiev @namchuai Can I check if we currently support this in cortex.cpp? (I recall the JS version did it already)

[vansangpfiev]

@vansangpfiev @namchuai Can I check if we currently support this in cortex.cpp? (I recall the JS version did it already)

No - we don't support it yet.

[dan-menlo]

@nguyenhoangthuan99 I am reassigning this to you, as it is fairly similar to /chat/completions. Having Cortex support embeddings will allow us to integrate things like Page Assist Browser extension

[nguyenhoangthuan99]

We need to update documentations for API references and capabilities

[gabrielle-ong]

Thank you @nguyenhoangthuan99! Nicely done. QAed on cortex.llamacpp v0.1.38
✅ QA: /embeddings endpoint

encoding:

• float (default),
• base64
  
  

Inputs:

1. String (above)
2. array of strings
3. array of integers
4. array of arrays

Docs: