Files
kserve/test/e2e/data/text_embedding_input_openai_base64.json
Fabian ScheidtandGitHub 19670c3214 Add Support for OpenAI-compatible Embeddings API (#4129)
* Add boilerplate setup for embeddings api

Signed-off-by: Fabian Scheidt <[email protected]>

* Integrate call to model with embedding request

Signed-off-by: Fabian Scheidt <[email protected]>

* Adjust error responses to be consistent with openai

Signed-off-by: Fabian Scheidt <[email protected]>

* Adjust doc strings to match implementation

Signed-off-by: Fabian Scheidt <[email protected]>

* Update openai input validation to use pydantic

Signed-off-by: Fabian Scheidt <[email protected]>

* Refactor embedding object type

Signed-off-by: Fabian Scheidt <[email protected]>

* Add e2e test for OpenAI embedding endpoint

Signed-off-by: Fabian Scheidt <[email protected]>

* Add unit test for OpenAI embeddings

Signed-off-by: Fabian Scheidt <[email protected]>

---------

Signed-off-by: Fabian Scheidt <[email protected]>
2025-01-16 01:23:45 -05:00

6 lines
118 B
JSON

{
"input": ["This is an example sentence."],
"model": "hf-text-embedding-openai",
"encoding_format": "base64"
}