get_initializer
Function
Obtains the initialization operator (Operation) of tensorflow.data.Iterator. This operator needs to use sess.run () to initialize Iterator.
Prototype
1 2 | from mx_rec.util.initialize import ConfigInitializer ConfigInitializer.get_instance().train_params_config.get_initializer(is_training) |
Parameters
Parameter |
Type |
Mandatory/Optional |
Description |
|---|---|---|---|
is_training |
bool |
Mandatory |
Whether the training mode is enabled.
|
Return Value
- Success: TensorFlow operator (tf.Operation) for initializing the iterator
- Failure: An exception is thrown.
Example
1 2 3 4 5 6 7 8 9 10 11 | import tensorflow as tf from mx_rec.util.initialize import ConfigInitializer from mx_rec.graph.modifier import modify_graph_and_start_emb_cache # In train mode, automatic graph modification needs to be enabled. # In train mode, automatic graph modification must be performed after gradient calculation. # Calculate the gradient. modify_graph_and_start_emb_cache(dump_graph=True) with tf.compat.v1.Session() as sess: # Ensure that the modify_graph_and_start_emb_cache() API has been called. initializer = ConfigInitializer.get_instance().train_params_config.get_initializer(True) sess.run(initializer) |
See Also
For details about the API call sequence and example, see Automatic Graph Modification.
Parent topic: Automatic Graph Modification