updated constraint on custom layers

RELEASE.md

@@ -2,7 +2,18 @@
 
 **Functionality**
 
-* EEGdataset can preload the entire dataset.
+- **dataloading module**:
+    - EEGdataset can preload the entire dataset.
+- **models module**:
+    - custom layers were moved in a new models.layer submodule
+    - layer constraints now include MaxNorm, MinMaxNorm, UnitNorm, with axis selection like in Keras.
+    - added Conv1d layer with norm constraint
+
+**Maintenance**
+
+* fixed typos on model module unittest.
+* Added new tests for novel functionalities.
+
 
 # Version 0.1.1 (latest)
 

docs/api/selfeeg.models.layers.ConstrainedConv1d.rst

@@ -0,0 +1,20 @@
+ConstrainedConv1d
+=================
+
+.. currentmodule:: selfeeg.models.layers
+
+.. autoclass:: ConstrainedConv1d
+   :show-inheritance:
+   :noindex:
+
+   .. rubric:: Methods Summary
+
+   .. autosummary::
+
+      ~ConstrainedConv1d.forward
+      ~ConstrainedConv1d.scale_norm
+
+   .. rubric:: Methods Documentation
+
+   .. automethod:: forward
+   .. automethod:: scale_norm

docs/api/selfeeg.models.zoo.ConstrainedConv2d.rst → docs/api/selfeeg.models.layers.ConstrainedConv2d.rst

@@ -1,7 +1,7 @@
 ConstrainedConv2d
 =================
 
-.. currentmodule:: selfeeg.models.zoo
+.. currentmodule:: selfeeg.models.layers
 
 .. autoclass:: ConstrainedConv2d
    :show-inheritance:

docs/api/selfeeg.models.zoo.ConstrainedDense.rst → docs/api/selfeeg.models.layers.ConstrainedDense.rst

@@ -1,7 +1,7 @@
 ConstrainedDense
 ================
 
-.. currentmodule:: selfeeg.models.zoo
+.. currentmodule:: selfeeg.models.layers
 
 .. autoclass:: ConstrainedDense
    :show-inheritance:

docs/api/selfeeg.models.zoo.DepthwiseConv2d.rst → docs/api/selfeeg.models.layers.DepthwiseConv2d.rst

@@ -1,7 +1,7 @@
 DepthwiseConv2d
 ===============
 
-.. currentmodule:: selfeeg.models.zoo
+.. currentmodule:: selfeeg.models.layers
 
 .. autoclass:: DepthwiseConv2d
    :show-inheritance:

docs/api/selfeeg.models.zoo.SeparableConv2d.rst → docs/api/selfeeg.models.layers.SeparableConv2d.rst

@@ -1,7 +1,7 @@
 SeparableConv2d
 ===============
 
-.. currentmodule:: selfeeg.models.zoo
+.. currentmodule:: selfeeg.models.layers
 
 .. autoclass:: SeparableConv2d
    :show-inheritance:

docs/selfeeg.models.rst

@@ -1,9 +1,20 @@
 selfeeg.models
 ==============
 
-This module collects various Deep Learning models proposed for EEG applications.
-In addition, it implements some layers not directly available in the PyTorch nn.Module, such as:
-a Depthwise Conv2d layer, a Separable Conv2d layer, Conv2d with max norm constraint, Linear layer with max norm constraint.
+This module collects various Deep Learning models and custom layers.
+It is divided in two submodules:
+
+- **layers**: a collection custom layers with the possibility to add norm constraints.
+- **zoo**: a collection of deep learning models proposed for EEG applications.
+
+models.layers module
+--------------------
+
+.. automodapi:: selfeeg.models.layers
+  :no-inheritance-diagram:
+  :no-main-docstr:
+  :noindex:
+  :no-heading:
 
 models.zoo module
 -----------------

selfeeg/dataloading/load.py

@@ -1,3 +1,5 @@
+from __future__ import annotations
+
 import glob
 import math
 import os

selfeeg/models/__init__.py

@@ -1,5 +1,6 @@
 from .zoo import (
     BasicBlock1,
+    ConstrainedConv1d,
     ConstrainedConv2d,
     ConstrainedDense,
     DeepConvNet,
@@ -23,3 +24,11 @@
     TinySleepNet,
     TinySleepNetEncoder,
 )
+
+from .layers import (
+    ConstrainedConv1d,
+    ConstrainedConv2d,
+    ConstrainedDense,
+    DepthwiseConv2d,
+    SeparableConv2d,
+)