MLP regression (prediction)

更新时间:
复制 MD 格式

The MLP Regression (Infer) component runs inference on new data using a trained multilayer perceptron (MLP) regression model. During inference, the component loads the model, preprocesses the input data, performs a forward propagation pass, and writes the prediction results to an Object Storage Service (OSS) path.

Supported computing resources

Deep Learning Containers (DLC)

Inputs and outputs

Input ports

The component accepts two inputs: inference data and a trained model.

Inference data

  • Connect the Read File Data component to read inference data from an OSS path.

  • Alternatively, specify an OSS path directly in the Infer Data Oss path parameter.

Input model

  • Connect the Read File Data component to read a model file from an OSS path.

  • The model file can come from the MLP Regression (Training) component.

Output port

The component writes a .csv result file to the OSS path specified in the Output File Oss Dir parameter.

Configure the component

In Machine Learning Designer, add the MLP Regression (Infer) component to your pipeline and set the following parameters.

Tab Parameter Required Description Default
Field Settings Infer Data Oss path No OSS path to an inference data file. Required only when no upstream data is connected. The file must be a headerless .csv containing only numerical feature columns — each row is one sample and each column is one feature. Example: test_features.csv. None
Output File Oss Dir Yes OSS directory where the inference result .csv is saved. None
Pretrained Model Oss Path No OSS path to a pre-trained model file. If left blank, no pre-trained model is loaded. None
Parameter Settings MLP Layer Num Yes Total number of MLP layers, excluding the output layer. 3
List of Hidden Layer Size Yes Number of output channels for each hidden layer, specified as a comma-separated list. The i-th value corresponds to the i-th hidden layer. If you enter a single number, that value is applied to all hidden layers. 64,32,16
List of Dropout Ratio Yes Dropout rate for each dropout layer, specified as a comma-separated list. If you enter a single value, it is applied to all dropout layers. 0.5
Execution Tuning Select Resource Group Public Resource Group No The instance type (CPU or GPU) and virtual private cloud (VPC) to use. None
Dedicated resource group No The number of CPU cores, memory, shared memory, and GPUs to use. None
Maximum Running Duration (seconds) No Maximum time the component is allowed to run. The job is terminated if this limit is exceeded. None