get_asc_insert_func
Function
Obtains the data preprocessing function.
Prototype
1 | def get_asc_insert_func(tgt_key_specs=None, args_index_list=None, table_names=None, **kwargs) |
Parameters
Parameter |
Type |
Mandatory/Optional |
Description |
|---|---|---|---|
tgt_key_specs |
|
Mandatory. Parameters can be passed in two modes. |
Feature object, feature object list, or feature object tuple. The default value is None. |
args_index_list |
list[int] |
List of parameter indexes. The default value is None. The value range is[1, 2^31 – 1]. |
|
table_names |
list[str] |
List of table names. The default value is None. The value range is[1, 2^31 – 1]. |
The parameters can be passed in either of the following modes:
- Pass only tgt_key_specs.
- Pass both args_index_list and table_names.
**kwargs Parameters
Parameter |
Type |
Mandatory/Optional |
Description |
|---|---|---|---|
is_training |
bool |
Optional |
Whether to enable the training mode. The default value is True. Value:
|
dump_graph |
bool |
Optional |
Whether to save the model graph. The default value is False. Value:
|
- is_training and dump_graph of **kwargs are for internal use. You are not advised to pass these two parameters through kwargs.
- If kwargs is used to pass other parameters that are not described, Rec SDK does not use these parameters.
Return Value
- Success: data preprocessing function.
- Failure: An exception is thrown.
Example
1 2 3 4 5 | import tensorflow as tf from mx_rec.core.asc.helper import get_asc_insert_func dataset = tf.data.TFRecordDataset(data_path) # data_path indicates the dataset path. dataset = dataset.map(get_asc_insert_func(tgt_key_specs=feature_spec_list, is_training=True)) # Elements in feature_spec_list are FeatureSpec objects. |
See Also
For details about the API call sequence and example, see Porting and Training.