Skip to main content

dense_embeddings

bge_m3_embedding_server::handler::dense

Function dense_embeddings

pub async fn dense_embeddings(
    __arg0: State<Arc<AppState>>,
    __arg1: Json<DenseRequest>,
) -> Result<Json<DenseResponse>, AppError>

Expand description

Handles POST /v1/embeddings — returns dense (float32) embeddings.

§Errors

AppError::ServiceUnavailable if the model is not ready or no workers are live.
AppError::InvalidRequest if the batch is empty, exceeds max_batch, or any text exceeds the per-string character limit.
AppError::Internal if the embedding pool returns an inference error.

§Panics

Panics if the request semaphore has been closed — should not occur in normal operation.