aclfvSearch

Applicability

Product

Supported

Atlas A3 training products / Atlas A3 inference products

Atlas A2 training products / Atlas A2 inference products

Atlas 200I/500 A2 inference products

Atlas inference products

Atlas training products

Description

Performs 1:N or N:M feature search.

Atlas inference accelerator modules : This API is not supported.

Prototype

aclError aclfvSearch(aclfvSearchType type, aclfvSearchInput *searchInput, aclfvSearchResult *searchRst)

Parameters

Parameter

Input/Output

Description

type

Input

Feature search scenario type.

searchInput

Input

Pointer to the search input information.

Call aclfvCreateSearchInput to create data of the aclfvSearchInput type in advance.

searchRst

Output

Pointer to the search output result.

Call aclfvCreateSearchResult to create data of the aclfvSearchResult type in advance. The queryCnt argument must be the same as that of searchInput.

The number of returned search results is specified by the topk argument passed to the aclfvCreateSearchInput call and is also related to the topk argument passed to the aclfvInit call for initialization. If the topk argument passed to the aclfvCreateSearchInput call is greater than or equal to that passed to the aclfvInit call, the number of returned search results is determined by the topk argument passed to the aclfvInit call; otherwise, the number of returned search results is determined by the topk argument passed to the aclfvCreateSearchInput call.

Returns

0 on success; else, failure. For details, see aclError.

Restrictions

  • In N:M mode, multithreaded adding, search, or deletion is not allowed.
  • In 1:N or N:M mode, if no repository is added, zero results are returned by the aclfvSearch call.
  • This API call automatically creates a stream for synchronous task execution within the API. When the tasks are complete, the stream is automatically destroyed.

    Hardware has restrictions on the total number of streams. For details, see aclrtCreateStream.

See Also

For details about the API call sequence and sample code, see Feature Vector Search.