Example of conditional outputs for estimators (partial)

keras_multiout_traditional_partial.py

 # just the fn from model_store that takes:                                                
 model_builder = model_class(model_artifacts, **model_parameters)                          
 model_builder.build_model(inp_placeholder)
 # dict of tensors like {'softmax':softmax_layer, 'oov_code':oov_code}
 tensors = model_builder.give_outputs()                                                                                                               
                                                                                           
 if mode == tensorflow.estimator.ModeKeys.PREDICT:                                         
     return tensorflow.estimator.EstimatorSpec(                                            
         mode=mode,                                                                        
         predictions=tensors,                                                              
         export_outputs={                                                                  
             signature_constants.DEFAULT_SERVING_SIGNATURE_DEF_KEY: PredictOutput(tensors) 
         },                                                                                
     )                                                                                     
                                                                                           
 loss_tensor = tensorflow.losses.softmax_cross_entropy(                                    
     labels, tensors['softmax'])                                                           
 tensorflow.summary.scalar('{0}_loss'.format(mode), loss_tensor)                           
                                                                                           
 train_op = tensorflow.contrib.layers.optimize_loss(                                       
     loss=loss_tensor,                                                                     
     global_step=tensorflow.train.get_global_step(),                                       
     optimizer='sgd',                                  
     learning_rate=0.01
 )                                                                                         
                                                                                           
 # since we're already assuming softmax & cross entropy, might as well just hardcode in    
 # acc as the eval metric                                                                  
 # these need to be converted to non-categorical!                                          
 eval_metrics_ops = {                                                                      
     "accuracy": tensorflow.metrics.accuracy(                                              
         tensorflow.argmax(labels, axis=1),                                                
         tensorflow.argmax(tensors['softmax'], axis=1))                                    
 }                                                                                         
                                                                                           
 tensorflow.summary.scalar('{0}_accuracy'.format(mode), eval_metrics_ops[                  
                           'accuracy'][1])  # get the update op                            
                                                                                           
 if mode in (tensorflow.estimator.ModeKeys.TRAIN,                                          
             tensorflow.estimator.ModeKeys.EVAL):                                          
     return tensorflow.estimator.EstimatorSpec(                                            
         mode=mode,                                                                        
         loss=loss_tensor,                                                                 
         train_op=train_op,                                                                
         eval_metric_ops=eval_metrics_ops,                                                 
     )