* Copyright (c) 2026 Huawei Technologies Co., Ltd.
* This program is free software, you can redistribute it and/or modify it under the terms and conditions of
* CANN Open Software License Agreement Version 2.0 (the "License").
* Please refer to the License for details. You may not use this file except in compliance with the License.
* THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, EITHER EXPRESS OR IMPLIED,
* INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR PURPOSE.
* See LICENSE in the root of the software repository for the full text of the License.
*/
#ifndef OP_API_INC_EMBEDDING_H_
#define OP_API_INC_EMBEDDING_H_
#include "aclnn/aclnn_base.h"
#include "aclnn_util.h"
#ifdef __cplusplus
extern "C" {
#endif
* @brief aclnnEmbedding的第一段接口,根据具体的计算流程,计算workspace大小。
* @domain aclnn_ops_infer
*
* 算子功能:把数据集合映射到向量空间,进而将数据进行量化
*
* api计算的基本路径:
* ```mermaid
* graph LR
* A[(weight)] --> B([l0op::Contiguous])
* B --> C([l0op::GatherV2])
* D[(indices)] --> E([l0op::Contiguous]) --> C
* F((dim=0)) --> G[(dimTensor)] --> C
* C --> H([l0op::ViewCopy]) --> Out[(out)]
* ```
*
* @param [in] weight: npu
* device侧的aclTensor,数据类型支持BFLOAT16,
* FLOAT、FLOAT16、INT64、INT32、INT16、INT8、UINT8、BOOL、DOUBLE、COMPLEX64、
* COMPLEX128,支持非连续Tensor,数据格式支持ND。
* @param [in] indices: npu
* device侧的aclTensor,数据类型支持INT64、INT32,支持非连续Tensor,数据格式支持ND。
* @param [out] out: npu
* device侧的aclTensor,数据类型同weight,数据格式支持ND,比indices的维度多1。
* @param [out] workspace_size: 返回用户需要在npu device侧申请的workspace大小。
* @param [out] executor: 返回op执行器,包含算子计算流程。
* @return aclnnStatus: 返回状态码。
*/
ACLNN_API aclnnStatus aclnnEmbeddingGetWorkspaceSize(const aclTensor* weight, const aclTensor* indices,
const aclTensor* out, uint64_t* workspaceSize,
aclOpExecutor** executor);
* @brief aclnnEmbedding的第二段接口,用于执行计算。
*
* * 算子功能:把数据集合映射到向量空间,进而将数据进行量化
*
* api计算的基本路径:
* ```mermaid
* graph LR
* A[(weight)] --> B([l0op::Contiguous])
* B --> C([l0op::GatherV2])
* D[(indices)] --> E([l0op::Contiguous]) --> C
* F((dim=0)) --> G[(dimTensor)] --> C
* C --> H([l0op::ViewCopy]) --> Out[(out)]
* ```
*
* @param [in] workspace: 在npu device侧申请的workspace内存起址。
* @param [in] workspace_size: 在npu device侧申请的workspace大小,由第一段接口aclnnEmbeddingGetWorkspaceSize获取。
* @param [in] executor: op执行器,包含了算子计算流程。
* @param [in] stream: acl stream流。
* @return aclnnStatus: 返回状态码。
*/
ACLNN_API aclnnStatus aclnnEmbedding(void* workspace, uint64_t workspaceSize, aclOpExecutor* executor,
const aclrtStream stream);
#ifdef __cplusplus
}
#endif
#endif