feat(python-example): learn it button with dropdown

TODO.md

@@ -8,6 +8,8 @@
 
 ## IN PROGRESS :
 
+- [ ] Add default arguments to python bindings
+- [ ] Add verbose argument for progess bar
 - [ ] Interactive Python example
 - [ ] Python tests
 - [ ] Optimize `Catch2`'s build

examples/train-predict-MNIST/guess_it.py

@@ -10,8 +10,8 @@
 margin = s(10)
 screen_width = s(640)
 screen_height = s(480)
-guess_button_width = s(100)
-guess_button_height = s(50)
+button_width = s(100)
+button_height = s(50)
 guess_text_width = s(200)
 guess_text_height = s(75)
 drawing_surface_width = drawing_surface_height = s(300)
@@ -34,24 +34,27 @@
 drawing_surface.set_clip(None)
 
 manager = pygame_gui.UIManager((screen_width, screen_height))
-guess_button_rect = pygame.Rect(0, 0, guess_button_width, guess_button_height)
+guess_button_rect = pygame.Rect(0, 0, button_width, button_height)
 guess_button_rect.bottomright = (-margin, -margin)
 guess_button = pygame_gui.elements.UIButton(relative_rect=guess_button_rect, text="Guess it", manager=manager, anchors={'right': 'right', 'bottom': 'bottom'})
+clear_button_rect = pygame.Rect((-(margin + (button_width * 2))), -(margin + button_height), button_width, button_height)
+clear_button = pygame_gui.elements.UIButton(relative_rect=clear_button_rect, text="Clear", manager=manager, anchors={'right': 'right', 'bottom': 'bottom'})
 guess_text_rect = pygame.Rect(0, margin, guess_text_width, guess_text_height)
 guess_text = pygame_gui.elements.UITextBox(html_text="", relative_rect=guess_text_rect, manager=manager, anchors={'centerx': 'centerx'})
-learn_button_rect = pygame.Rect(0, 0, guess_button_width, guess_button_height)
-learn_button_rect.bottomleft = (margin, -margin)
-learn_button = pygame_gui.elements.UIButton(relative_rect=learn_button_rect, text="Learn It", manager=manager, anchors={'left': 'left', 'bottom': 'bottom'})
-dropdown_rect = pygame.Rect(0,-(margin + guess_button_height), dropdown_width, dropdown_height)
-dropdown = pygame_gui.elements.UIDropDownMenu(DROPDOWN_OPTIONS, DROPDOWN_OPTIONS[0], dropdown_rect, manager, anchors={'centerx': 'centerx', 'bottom': 'bottom'})
+learn_button_rect = pygame.Rect(margin, margin, button_width, button_height)
+learn_button = pygame_gui.elements.UIButton(relative_rect=learn_button_rect, text="Learn It", manager=manager, anchors={'left': 'left', 'top': 'top'})
+dropdown_rect = pygame.Rect(0, 0, dropdown_width, dropdown_height)
+dropdown_rect.topleft = (margin, margin + button_height)
+dropdown = pygame_gui.elements.UIDropDownMenu(DROPDOWN_OPTIONS, DROPDOWN_OPTIONS[0], dropdown_rect, manager=manager, anchors={'left': 'left', 'top': 'top'})
 
 # This dict will be passed to event handlers
 ui_elements = {
   "guess_button": guess_button,
   "guess_text": guess_text,
   "drawing_surface": drawing_surface,
   "learn_button": learn_button,
-  "dropdown": dropdown
+  "dropdown": dropdown,
+  'clear_button': clear_button
 }
 
 # Fill the main_window with white

examples/train-predict-MNIST/helpers/event_handlers.py

@@ -12,21 +12,53 @@
 
 NNP.models.Model.load_from_file("model.bin", network)
 
+network.setup(optimizer=NNP.optimizers.SGD(0.001), loss=NNP.LOSS.MCE)
+
 drawing = False # Track drawing
 erasing = False # Track erasing
 drawing_color = COLORS["black"]
 erasing_color = COLORS["white"]
 
+def get_drawing(context):
+  """
+    Get the drawing from the drawing_surface
+
+    Args:
+      context
+
+    Returns:
+      normalized_img: The normalized drawing 
+  """
+  drawing_image = pygame.surfarray.pixels3d(context["ui_elements"]["drawing_surface"])
+  grayscale_image = format_image_grayscale(drawing_image, (28, 28))
+  normalized_image = normalize_img(numpy.transpose(grayscale_image))
+  return normalized_image
+
 
 def handle_ui_button_pressed(context):
   if context["event"].ui_element == context["ui_elements"]["guess_button"]:
-    drawing_image = pygame.surfarray.pixels3d(context["ui_elements"]["drawing_surface"])
-    grayscale_image = format_image_grayscale(drawing_image, (28, 28))
-    normalized_image = normalize_img(numpy.transpose(grayscale_image))
+    normalized_image = get_drawing(context)
     prediction = find_highest_indexes_in_matrix(network.predict([normalized_image]))
     context["ui_elements"]["guess_text"].append_html_text(f"I'm guessing : {prediction[0]}<br>")
+
   if context["event"].ui_element == context["ui_elements"]["learn_button"]:
-    print("I'm learning")
+    normalized_image = get_drawing(context)
+    target = float(context["ui_elements"]["dropdown"].selected_option)
+    loss = network.train([normalized_image], [target], 1)
+    context["ui_elements"]["guess_text"].append_html_text(f"I'm learning that it's a {int(target)}<br>loss : {loss}")
+
+  if context["event"].ui_element == context["ui_elements"]["dropdown"]:
+    print("dropdown has been clicked")
+
+  if context["event"].ui_element == context["ui_elements"]["clear_button"]:
+    context["ui_elements"]["drawing_surface"].fill(erasing_color)
+  return
+
+def handle_dropdown_change(context):
+  event = context['event']
+  if event.ui_element == context["ui_elements"]["dropdown"]:
+    print("Selected Option ", event.text)
+    context["ui_elements"]["dropdown"].close = True
   return
 
 def handle_mouse_button_down(context):
@@ -64,6 +96,7 @@ def handle_quit(context):
 
 EVENT_HANDLER_MAP = {
   pygame_gui.UI_BUTTON_PRESSED: handle_ui_button_pressed,
+  pygame_gui.UI_DROP_DOWN_MENU_CHANGED: handle_dropdown_change,
   pygame.MOUSEBUTTONDOWN: handle_mouse_button_down,
   pygame.MOUSEBUTTONUP: handle_mouse_button_up,
   pygame.MOUSEMOTION: handle_mouse_motion,

examples/train-predict-MNIST/main.py

@@ -11,7 +11,7 @@
 from helpers.utils import *
 from halo import Halo
 
-NUM_TRAININGS = 60000
+NUM_TRAININGS = 10000
 NUM_PREDICTIONS = 1000
 MNIST_DATASET_FILE = "./dataset/mnist.npz"
 
@@ -35,7 +35,7 @@
 network.addLayer(NNP.layers.Dense(10, NNP.ACTIVATION.SOFTMAX, NNP.WEIGHT_INIT.LECUN))
 
 # Setting up the networks parameters
-network.setup(optimizer=NNP.optimizers.Adam(0.001), loss=NNP.LOSS.MCE)
+network.setup(optimizer=NNP.optimizers.Adam(0.01), loss=NNP.LOSS.MCE)
 
 # combining the data with the labels for later shuffling 
 combined = list(zip(x_train, y_train))

src/bindings/NeuralNetPy.cpp

@@ -417,6 +417,10 @@ PYBIND11_MODULE(NeuralNetPy, m) {
            static_cast<double (Network::*)(
                std::vector<std::vector<double>>, std::vector<double>, int,
                const std::vector<std::shared_ptr<Callback>>)>(&Network::train),
+               py::arg("inputs"),
+               py::arg("targets"),
+               py::arg("epochs"),
+               py::arg("callbacks") = std::vector<std::shared_ptr<Callback>>(),
            R"pbdoc(
             Train the network by passing it 2 dimensional inputs (vectors).
 
@@ -458,6 +462,10 @@ PYBIND11_MODULE(NeuralNetPy, m) {
                std::vector<std::vector<std::vector<double>>>,
                std::vector<double>, int,
                const std::vector<std::shared_ptr<Callback>>)>(&Network::train),
+               py::arg("inputs"),
+               py::arg("targets"),
+               py::arg("epochs"),
+               py::arg("callbacks") = std::vector<std::shared_ptr<Callback>>(),
            R"pbdoc(
         Train the network by passing it a list of 3 dimensional inputs (matrices).