Davout 0.1.1.dev167__tar.gz → 0.1.1.dev168__tar.gz

{davout-0.1.1.dev167/source/Davout.egg-info → davout-0.1.1.dev168}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: Davout
-Version: 0.1.1.dev167
+Version: 0.1.1.dev168
 Summary: Scientific computing utilities combining FEM, ANN and tooling
 Author: Matheus Janczkowski
 Author-email: Matheus Janczkowski <matheusj2009@hotmail.com>

{davout-0.1.1.dev167 → davout-0.1.1.dev168}/pyproject.toml

@@ -63,7 +63,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "Davout"
-version = "0.1.1.dev167"
+version = "0.1.1.dev168"
 description = "Scientific computing utilities combining FEM, ANN and tooling"
 readme = "README.md"
 authors = [

{davout-0.1.1.dev167 → davout-0.1.1.dev168}/source/Davout/DeepMech/tool_box/training_tools.py

@@ -526,8 +526,13 @@ class ModelCustomTraining:
                 # Gets the gradient from the second position of the call
                 # method
 
-                gradient_value = tf.norm(self.loss_class(
-                self.model_parameters)[1]).numpy()
+                evaluated_gradient = self.loss_class(
+                self.model_parameters)[1]
+
+                gradient_value = tf.norm(evaluated_gradient).numpy()
+
+                gradient_maximum_component = tf.reduce_max(
+                tf.abs(evaluated_gradient)).numpy()
 
                 if gradient_value<=self.gradient_tolerance:
 
@@ -541,12 +546,14 @@ class ModelCustomTraining:
 
                 print("Iteration group "+integer_toString(i, max_digits)
                 +": loss="+format(loss_value.numpy(), '.5e')+", gradie"+
-                "nt norm: "+format(gradient_value, '.5e')+", gradient "+
-                "norm divided by the square root of the number of trai"+
-                "nable parameters: "+format(gradient_value/
-                self.number_trainable_parameters, '.5e')+"\nThis cycle"+
-                " took "+str(time.time()-self.initial_time)+" to compl"+
-                "ete\n")
+                "nt norm: "+format(gradient_value, '.5e')+",\ngradient"+
+                " norm divided by the square root of the number of tra"+
+                "inable parameters: "+format(gradient_value/
+                self.number_trainable_parameters, '.5e')+"\nThe maximu"+
+                "m component of the gradient in absolute value is: "+
+                format(gradient_maximum_component, '.5e')+"\nThis cycl"+
+                "e took "+str(time.time()-self.initial_time)+" to comp"+
+                "lete\n")
 
         # Otherwise, just call the training function
 

{davout-0.1.1.dev167 → davout-0.1.1.dev168/source/Davout.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: Davout
-Version: 0.1.1.dev167
+Version: 0.1.1.dev168
 Summary: Scientific computing utilities combining FEM, ANN and tooling
 Author: Matheus Janczkowski
 Author-email: Matheus Janczkowski <matheusj2009@hotmail.com>