Google Vertex AI inference integration

New API reference

For the most up-to-date API details, refer to Inference APIs.

Creates an inference endpoint to perform an inference task with the googlevertexai service.

Request ¶

PUT /_inference/<task_type>/<inference_id>

Path parameters ¶

<inference_id>
(Required, string) The unique identifier of the inference endpoint.
<task_type>

(Required, string) The type of the inference task that the model will perform.

Available task types:

  • rerank
  • text_embedding.

Request body ¶

chunking_settings

(Optional, object) Chunking configuration object. Refer to Configuring chunking to learn more about chunking.

max_chunk_size
(Optional, integer) Specifies the maximum size of a chunk in words. Defaults to 250. This value cannot be higher than 300 or lower than 20 (for sentence strategy) or 10 (for word strategy).
overlap
(Optional, integer) Only for word chunking strategy. Specifies the number of overlapping words for chunks. Defaults to 100. This value cannot be higher than the half of max_chunk_size.
sentence_overlap
(Optional, integer) Only for sentence chunking strategy. Specifies the numnber of overlapping sentences for chunks. It can be either 1 or 0. Defaults to 1.
strategy
(Optional, string) Specifies the chunking strategy. It could be either sentence or word.
service
(Required, string) The type of service supported for the specified task type. In this case, googlevertexai.
service_settings

(Required, object) Settings used to install the inference model.

These settings are specific to the googlevertexai service.

service_account_json
(Required, string) A valid service account in json format for the Google Vertex AI API.
model_id
(Required, string) The name of the model to use for the inference task. You can find the supported models at Text embeddings API.
location
(Required, string) The name of the location to use for the inference task. You find the supported locations at Generative AI on Vertex AI locations.
project_id
(Required, string) The name of the project to use for the inference task.
rate_limit
(Optional, object) By default, the googlevertexai service sets the number of requests allowed per minute to 30.000. This helps to minimize the number of rate limit errors returned from Google Vertex AI. To modify this, set the requests_per_minute setting of this object in your service settings:

text "rate_limit": { "requests_per_minute": <<number_of_requests>> }

More information about the rate limits for Google Vertex AI can be found in the Google Vertex AI Quotas docs.

task_settings

(Optional, object) Settings to configure the inference task. These settings are specific to the <task_type> you specified.

`task_settings` for the `rerank` task type
top_n
(optional, boolean) Specifies the number of the top n documents, which should be returned.
`task_settings` for the `text_embedding` task type
auto_truncate
(optional, boolean) Specifies if the API truncates inputs longer than the maximum token length automatically.

Google Vertex AI service example ¶

The following example shows how to create an inference endpoint called google_vertex_ai_embeddings to perform a text_embedding task type.

			PUT _inference/text_embedding/google_vertex_ai_embeddings
			{
    "service": "googlevertexai",
    "service_settings": {
        "service_account_json": "<service_account_json>",
        "model_id": "<model_id>",
        "location": "<location>",
        "project_id": "<project_id>"
    }
}

The next example shows how to create an inference endpoint called google_vertex_ai_rerank to perform a rerank task type.

			PUT _inference/rerank/google_vertex_ai_rerank
			{
    "service": "googlevertexai",
    "service_settings": {
        "service_account_json": "<service_account_json>",
        "project_id": "<project_id>"
    }
}