dlpy.model.Model.load_weights

Model.load_weights(path, labels=False, data_spec=None, label_file_name=None, label_length=None, use_gpu=False, embedding_dim=None)

Load the weights from a data file specified by ‘path’

Parameters

pathstring

Specifies the server-side directory of the file that contains the weight table.

labelsbool

Specifies whether to apply user-defined classification labels Default: False

data_speclist of DataSpec, optional

data specification for input and output layer(s) Default: None

label_file_namestring, optional

Fully qualified path to CSV file containing user-defined classification labels. If not specified, ImageNet labels assumed. Default: None

label_lengthint, optional

Length of the classification labels (in characters). Default: None

use_gpu: boolean, optional

GPU processing of model required (or not) Default: False

embedding_dimint, optional

Specifies text embedding dimension. You must specify the data_spec parameter or this parameter is ignored. Default: None

Notes

Currently support HDF5 and sashdat files.