clear_samples_class_gmmT_clear_samples_class_gmmClearSamplesClassGmmClearSamplesClassGmmclear_samples_class_gmm (算子)

名称

clear_samples_class_gmmT_clear_samples_class_gmmClearSamplesClassGmmClearSamplesClassGmmclear_samples_class_gmm — 清除高斯混合模型的训练数据。

签名

clear_samples_class_gmm( : : GMMHandle : )

Herror T_clear_samples_class_gmm(const Htuple GMMHandle)

void ClearSamplesClassGmm(const HTuple& GMMHandle)

static void HClassGmm::ClearSamplesClassGmm(const HClassGmmArray& GMMHandle)

void HClassGmm::ClearSamplesClassGmm() const

static void HOperatorSet.ClearSamplesClassGmm(HTuple GMMHandle)

static void HClassGmm.ClearSamplesClassGmm(HClassGmm[] GMMHandle)

void HClassGmm.ClearSamplesClassGmm()

def clear_samples_class_gmm(gmmhandle: MaybeSequence[HHandle]) -> None

描述

clear_samples_class_gmmclear_samples_class_gmmClearSamplesClassGmmClearSamplesClassGmmClearSamplesClassGmmclear_samples_class_gmm clears all training samples that have been stored in the Gaussian Mixture Model (GMM) GMMHandleGMMHandleGMMHandleGMMHandleGMMHandlegmmhandle. clear_samples_class_gmmclear_samples_class_gmmClearSamplesClassGmmClearSamplesClassGmmClearSamplesClassGmmclear_samples_class_gmm should only be used if the GMM is trained in the same process that uses the GMM for evaluation with evaluate_class_gmmevaluate_class_gmmEvaluateClassGmmEvaluateClassGmmEvaluateClassGmmevaluate_class_gmm or for classification with classify_class_gmmclassify_class_gmmClassifyClassGmmClassifyClassGmmClassifyClassGmmclassify_class_gmm。In this case, the memory required for the training samples can be freed with clear_samples_class_gmmclear_samples_class_gmmClearSamplesClassGmmClearSamplesClassGmmClearSamplesClassGmmclear_samples_class_gmm, and hence memory can be saved. In the normal usage, in which the GMM is trained offline and written to a file with write_class_gmmwrite_class_gmmWriteClassGmmWriteClassGmmWriteClassGmmwrite_class_gmm, it is typically unnecessary to call clear_samples_class_gmmclear_samples_class_gmmClearSamplesClassGmmClearSamplesClassGmmClearSamplesClassGmmclear_samples_class_gmm because write_class_gmmwrite_class_gmmWriteClassGmmWriteClassGmmWriteClassGmmwrite_class_gmm does not save the training samples, and hence the online process, which reads the GMM with read_class_gmmread_class_gmmReadClassGmmReadClassGmmReadClassGmmread_class_gmm, requires no memory for the training samples.

执行信息

此算子修改后续输入参数的状态:

在执行此算子时,若该参数值需在多个线程间使用,则必须对其访问进行同步。

参数

GMMHandleGMMHandleGMMHandleGMMHandleGMMHandlegmmhandle (输入控制,状态被修改)  class_gmm(-array) HClassGmm, HTupleMaybeSequence[HHandle]HTupleHtuple (handle) (IntPtr) (HHandle) (handle)

GMM 句柄。

结果

如果参数有效,算子 clear_samples_class_gmmclear_samples_class_gmmClearSamplesClassGmmClearSamplesClassGmmClearSamplesClassGmmclear_samples_class_gmm 返回值 2 ( H_MSG_TRUE )。如有必要,则抛出异常。

可能的前趋

train_class_gmmtrain_class_gmmTrainClassGmmTrainClassGmmTrainClassGmmtrain_class_gmm, write_samples_class_gmmwrite_samples_class_gmmWriteSamplesClassGmmWriteSamplesClassGmmWriteSamplesClassGmmwrite_samples_class_gmm

另见

create_class_gmmcreate_class_gmmCreateClassGmmCreateClassGmmCreateClassGmmcreate_class_gmm, clear_class_gmmclear_class_gmmClearClassGmmClearClassGmmClearClassGmmclear_class_gmm, add_sample_class_gmmadd_sample_class_gmmAddSampleClassGmmAddSampleClassGmmAddSampleClassGmmadd_sample_class_gmm, read_samples_class_gmmread_samples_class_gmmReadSamplesClassGmmReadSamplesClassGmmReadSamplesClassGmmread_samples_class_gmm

模块

基础