General class outlining the training environment of a model. More...
Inheritance diagram for nabu.neuralnetworks.trainers.trainer.Trainer:
Public Member Functions | |
| def | __init__ (self, conf, dataconf, modelconf, evaluatorconf, expdir, server, task_index) |
| NnetTrainer constructor, creates the training graph. More... | |
| def | train (self, testing=False) |
| train the model More... | |
| def | chief_only_hooks (self, outputs) |
| add hooks only for the chief worker More... | |
| def | hooks (self, outputs) |
| add hooks for the session More... | |
| def | aditional_loss (self) |
| add an aditional loss More... | |
Public Attributes | |
| conf | |
| dataconf | |
| evaluatorconf | |
| expdir | |
| server | |
| task_index | |
| model | |
Detailed Description
General class outlining the training environment of a model.
Constructor & Destructor Documentation
§ __init__()
| def nabu.neuralnetworks.trainers.trainer.Trainer.__init__ | ( | self, | |
| conf, | |||
| dataconf, | |||
| modelconf, | |||
| evaluatorconf, | |||
| expdir, | |||
| server, | |||
| task_index | |||
| ) |
NnetTrainer constructor, creates the training graph.
- Parameters
-
conf the trainer config as a ConfigParser dataconf the data configuration as a ConfigParser modelconf the neural net model configuration evaluatorconf the evaluator configuration for evaluating if None no evaluation will be done expdir directory where the summaries will be written server optional server to be used for distributed training task_index optional index of the worker task in the cluster
Member Function Documentation
§ aditional_loss()
| def nabu.neuralnetworks.trainers.trainer.Trainer.aditional_loss | ( | self | ) |
add an aditional loss
- Returns
- the aditional loss or None
§ chief_only_hooks()
| def nabu.neuralnetworks.trainers.trainer.Trainer.chief_only_hooks | ( | self, | |
| outputs | |||
| ) |
add hooks only for the chief worker
- Parameters
-
outputs the outputs generated by the create graph method
- Returns
- a list of hooks
§ hooks()
| def nabu.neuralnetworks.trainers.trainer.Trainer.hooks | ( | self, | |
| outputs | |||
| ) |
add hooks for the session
- Parameters
-
outputs the outputs generated by the create graph method
- Returns
- a list of hooks
§ train()
| def nabu.neuralnetworks.trainers.trainer.Trainer.train | ( | self, | |
testing = False |
|||
| ) |
train the model
- Parameters
-
testing if true only the graph will be created for debugging purposes
The documentation for this class was generated from the following file:
