#258 related. Callback now depends on model context and receive pipel…

…ines (epoch-related) via args.

arekit/common/experiment/callback.py

@@ -11,7 +11,7 @@ def __init__(self):
     def set_experiment(self, experiment):
         self._experiment = experiment
 
-    def on_initialized(self, model):
+    def on_initialized(self, model_ctx):
         # Do nothing by default.
         pass
 
@@ -23,7 +23,7 @@ def on_fit_started(self, operation_cancel):
         # Do nothing by default.
         pass
 
-    def on_epoch_finished(self, epoch_index, operation_cancel):
+    def on_epoch_finished(self, pipeline, operation_cancel):
         # Do nothing by default.
         pass
 

arekit/contrib/networks/core/callback/utils_hidden_states.py

@@ -5,7 +5,7 @@
 
 from arekit.common.utils import create_dir_if_not_exists
 from arekit.contrib.networks.core.ctx_predict_log import NetworkInputDependentVariables
-from arekit.contrib.networks.core.model import BaseTensorflowModel
+from arekit.contrib.networks.core.model_ctx import TensorflowModelContext
 
 logger = logging.getLogger(__name__)
 logging.basicConfig(level=logging.INFO)
@@ -18,11 +18,11 @@
 __input_dependent_params_template = 'idparams_{data}_e{epoch_index}'
 
 
-def save_model_hidden_values(log_dir, model, epoch_index):
+def save_model_hidden_values(log_dir, model_ctx, epoch_index):
     assert(isinstance(log_dir, str))
-    assert(isinstance(model, BaseTensorflowModel))
+    assert(isinstance(model_ctx, TensorflowModelContext))
 
-    names, values = model.Context.get_hidden_parameters()
+    names, values = model_ctx.get_hidden_parameters()
 
     assert (isinstance(names, list))
     assert (isinstance(values, list))

arekit/contrib/networks/core/model.py

@@ -1,4 +1,3 @@
-import collections
 import logging
 
 from arekit.common.model.base import BaseModel
@@ -11,6 +10,7 @@
 from arekit.contrib.networks.core.network_callback import NetworkCallback
 from arekit.contrib.networks.core.params import NeuralNetworkModelParams
 from arekit.contrib.networks.core.pipeline_epoch import EpochHandlingPipelineItem
+from arekit.contrib.networks.core.utils import get_item_from_pipeline
 from arekit.contrib.networks.tf_helpers.nn_states import TensorflowNetworkStatesProvider
 
 logger = logging.getLogger(__name__)
@@ -81,15 +81,6 @@ def __try_load_state(self):
             self.__states_provider.load_model(sess=self.__context.Session,
                                               path_tf_prefix=self.__context.IO.get_model_source_path_tf_prefix())
 
-    @staticmethod
-    def __get_item_from_pipeline(pipeline, item_type):
-        assert(isinstance(pipeline, list))
-        assert(issubclass(item_type, EpochHandlingPipelineItem))
-
-        for item in pipeline:
-            if isinstance(item, item_type):
-                return item
-
     def __fit(self, epochs_count):
         assert(isinstance(epochs_count, int))
         assert(self.__context.Session is not None)
@@ -109,7 +100,7 @@ def __fit(self, epochs_count):
             self.__run_epoch_pipeline(pipeline=self.__fit_pipeline, data_type=DataType.Train, prefix="Training")
 
             if self.__callback is not None:
-                self.__callback.on_epoch_finished(epoch_index=epoch_index,
+                self.__callback.on_epoch_finished(pipeline=self.__fit_pipeline,
                                                   operation_cancel=operation_cancel)
 
             if BaseTensorflowModel.SaveTensorflowModelStateOnFit:
@@ -125,19 +116,19 @@ def fit(self, model_params, seed):
         assert(isinstance(model_params, NeuralNetworkModelParams))
         self.__context.Network.compile(self.__context.Config, reset_graph=True, graph_seed=seed)
         self.__context.set_optimiser()
-        self.__callback.on_initialized(self)
+        self.__callback.on_initialized(self.__context)
         self.__context.initialize_session()
         self.__try_load_state()
         self.__fit(epochs_count=model_params.EpochsCount)
 
         self.__context.dispose_session()
 
     def predict(self, data_type=DataType.Test, do_compile=False, graph_seed=0):
-        """ Fills the related labeling collection.
-        """
+
         # Optionally perform network compilation
         if do_compile:
             self.__context.Network.compile(config=self.__context.Config, reset_graph=True, graph_seed=graph_seed)
+
         self.__context.initialize_session()
         self.__try_load_state()
         self.__run_epoch_pipeline(pipeline=self.__predict_pipeline,
@@ -146,8 +137,8 @@ def predict(self, data_type=DataType.Test, do_compile=False, graph_seed=0):
 
     def from_fitted(self, item_type):
         assert(issubclass(item_type, EpochHandlingPipelineItem))
-        return self.__get_item_from_pipeline(pipeline=self.__fit_pipeline, item_type=item_type)
+        return get_item_from_pipeline(pipeline=self.__fit_pipeline, item_type=item_type)
 
     def from_predicted(self, item_type):
         assert (issubclass(item_type, EpochHandlingPipelineItem))
-        return self.__get_item_from_pipeline(pipeline=self.__predict_pipeline, item_type=item_type)
+        return get_item_from_pipeline(pipeline=self.__predict_pipeline, item_type=item_type)

arekit/contrib/networks/core/network_callback.py

@@ -5,7 +5,9 @@
 from arekit.common.experiment.callback import Callback
 from arekit.common.utils import progress_bar_defined
 from arekit.contrib.networks.core.cancellation import OperationCancellation
+from arekit.contrib.networks.core.model_ctx import TensorflowModelContext
 from arekit.contrib.networks.core.pipeline_fit import MinibatchFittingPipelineItem
+from arekit.contrib.networks.core.utils import get_item_from_pipeline
 
 logger = logging.getLogger(__name__)
 logger.setLevel(logging.INFO)
@@ -19,12 +21,13 @@ class NetworkCallback(Callback):
 
     def __init__(self):
         super(NetworkCallback, self).__init__()
-        self.__training_epochs_passed = 0
-        self._model = None
+        self._training_epochs_passed = 0
+        self._model_ctx = None
 
-    def on_initialized(self, model):
-        super(NetworkCallback, self).on_initialized(model)
-        self._model = model
+    def on_initialized(self, model_ctx):
+        assert(isinstance(model_ctx, TensorflowModelContext))
+        super(NetworkCallback, self).on_initialized(model_ctx)
+        self._model_ctx = model_ctx
 
     @staticmethod
     def __create_epoch_stat(epoch_index, avg_fit_cost, avg_fit_acc):
@@ -38,22 +41,22 @@ def __create_epoch_stat(epoch_index, avg_fit_cost, avg_fit_acc):
         return u"{time}: {epochs}: {avg_fc}, {avg_ac}".format(
             time=time, epochs=epochs, avg_fc=avg_fc, avg_ac=avg_ac)
 
-    def on_epoch_finished(self, epoch_index, operation_cancel):
-        assert(isinstance(epoch_index, int))
+    def on_epoch_finished(self, pipeline, operation_cancel):
         assert(isinstance(operation_cancel, OperationCancellation))
 
-        item = self._model.from_fitted(MinibatchFittingPipelineItem)
-        self.__training_epochs_passed += 1
+        item = get_item_from_pipeline(pipeline=pipeline, item_type=MinibatchFittingPipelineItem)
+
+        self._training_epochs_passed += 1
 
         if item is None:
             return
 
         assert(isinstance(item, MinibatchFittingPipelineItem))
 
-        super(NetworkCallback, self).on_epoch_finished(epoch_index=epoch_index,
+        super(NetworkCallback, self).on_epoch_finished(pipeline=pipeline,
                                                        operation_cancel=operation_cancel)
 
-        message = self.__create_epoch_stat(epoch_index=epoch_index,
+        message = self.__create_epoch_stat(epoch_index=self._training_epochs_passed,
                                            avg_fit_cost=item.TotalFitCost,
                                            avg_fit_acc=item.TotalFitAccuracy)
 
@@ -66,5 +69,5 @@ def handle_batches_iter(self, batches_iter, total, prefix, unit='mbs'):
         assert(isinstance(batches_iter, collections.Iterable))
         assert(isinstance(unit, str))
         assert(isinstance(prefix, str))
-        desc = "{prefix} e={epoch}".format(prefix=prefix, epoch=self.__training_epochs_passed)
+        desc = "{prefix} e={epoch}".format(prefix=prefix, epoch=self._training_epochs_passed)
         return progress_bar_defined(iterable=batches_iter, unit=unit, total=total, desc=desc)

arekit/contrib/networks/core/utils.py

@@ -0,0 +1,10 @@
+from arekit.contrib.networks.core.pipeline_epoch import EpochHandlingPipelineItem
+
+
+def get_item_from_pipeline(pipeline, item_type):
+    assert (isinstance(pipeline, list))
+    assert (issubclass(item_type, EpochHandlingPipelineItem))
+
+    for item in pipeline:
+        if isinstance(item, item_type):
+            return item

examples/rusentrel/callback.py

@@ -5,6 +5,7 @@
 from arekit.contrib.networks.core.cancellation import OperationCancellation
 from arekit.contrib.networks.core.network_callback import NetworkCallback
 from arekit.contrib.networks.core.pipeline_fit import MinibatchFittingPipelineItem
+from arekit.contrib.networks.core.utils import get_item_from_pipeline
 
 logger = logging.getLogger(__name__)
 logger.setLevel(logging.INFO)
@@ -29,13 +30,12 @@ def __is_cancel_needed(self, avg_fit_acc):
 
     # endregion
 
-    def on_epoch_finished(self, epoch_index, operation_cancel):
-        assert(isinstance(epoch_index, int))
+    def on_epoch_finished(self, pipeline, operation_cancel):
         assert(isinstance(operation_cancel, OperationCancellation))
-        super(TrainingCallback, self).on_epoch_finished(epoch_index=epoch_index,
+        super(TrainingCallback, self).on_epoch_finished(pipeline=pipeline,
                                                         operation_cancel=operation_cancel)
 
-        item = self._model.from_predicted(MinibatchFittingPipelineItem)
+        item = get_item_from_pipeline(pipeline=pipeline, item_type=MinibatchFittingPipelineItem)
 
         if item is None:
             return
@@ -49,4 +49,6 @@ def on_epoch_finished(self, epoch_index, operation_cancel):
             return
 
         # Saving model hidden values using the related numpy utils.
-        save_model_hidden_values(log_dir=self.__log_dir, epoch_index=epoch_index, model=self._model)
+        save_model_hidden_values(log_dir=self.__log_dir,
+                                 epoch_index=self._training_epochs_passed,
+                                 model_ctx=self._model_ctx)