seldon-od-transformer

Chart to deploy an outlier detector as a transformer in an inference graph.

Usage

使用本chart，首先要添加 seldonio Helm 仓库：

helm repo add seldonio https://storage.googleapis.com/seldon-charts
helm repo update

一旦完成，你可以按照如下使用预估图模板：

helm template $MY_MODEL_NAME seldonio/seldon-od-transformer --namespace $MODELS_NAMESPACE

注意你也可以直接部署预估图到集群中： 使用：

helm install $MY_MODEL_NAME seldonio/seldon-od-transformer --namespace $MODELS_NAMESPACE

主页: https://github.com/SeldonIO/seldon-core

源码

• https://github.com/SeldonIO/seldon-core

• https://github.com/SeldonIO/seldon-core/tree/master/helm-charts/seldon-od-transformer

设置值

键	类型	默认值	描述
model.image.name	string	"seldonio/mock_classifier:1.0"
model.name	string	"classifier"
name	string	"seldon-od-transformer"
outlierDetection.enabled	bool	true
outlierDetection.isolationforest.image.name	string	"seldonio/outlier-if-tranformer:0.1"
outlierDetection.isolationforest.load_path	string	"./models/"
outlierDetection.isolationforest.model_name	string	"if"
outlierDetection.isolationforest.threshold	int	0
outlierDetection.mahalanobis.image.name	string	"seldonio/outlier-mahalanobis-tranformer:0.1"
outlierDetection.mahalanobis.max_n	int	-1
outlierDetection.mahalanobis.n_components	int	3
outlierDetection.mahalanobis.n_stdev	int	3
outlierDetection.mahalanobis.start_clip	int	50
outlierDetection.mahalanobis.threshold	int	25
outlierDetection.name	string	"outlier-detector"
outlierDetection.parameterTypes.load_path	string	"STRING"
outlierDetection.parameterTypes.max_n	string	"INT"
outlierDetection.parameterTypes.model_name	string	"STRING"
outlierDetection.parameterTypes.n_components	string	"INT"
outlierDetection.parameterTypes.n_stdev	string	"FLOAT"
outlierDetection.parameterTypes.reservoir_size	string	"INT"
outlierDetection.parameterTypes.start_clip	string	"INT"
outlierDetection.parameterTypes.threshold	string	"FLOAT"
outlierDetection.seq2seq.image.name	string	"seldonio/outlier-s2s-lstm-tranformer:0.1"
outlierDetection.seq2seq.load_path	string	"./models/"
outlierDetection.seq2seq.model_name	string	"seq2seq"
outlierDetection.seq2seq.reservoir_size	int	50000
outlierDetection.seq2seq.threshold	float	0.003
outlierDetection.type	string	"vae"	Type of outlier detector. Valid values are: vae, mahalanobis, seq2seq and isolationforest.
outlierDetection.vae.image.name	string	"seldonio/outlier-vae-tranformer:0.1"
outlierDetection.vae.load_path	string	"./models/"
outlierDetection.vae.model_name	string	"vae"
outlierDetection.vae.reservoir_size	int	50000
outlierDetection.vae.threshold	int	10
predictorLabels.fluentd	string	"true"
predictorLabels.version	string	"v1"
replicas	int	1
sdepLabels.app	string	"seldon"