def main():
model = vc.CNN_16()
X_fname = 'X_train_train.npy'
y_fname = 'y_train_train.npy'
X_train = np.load(X_fname)
y_train = np.load(y_fname)
print(X_train.shape)
print(y_train.shape)
print("Training started")
callbacks = []
earlystop_callback = EarlyStopping(monitor='val_loss', patience=5, verbose=0)
batch_print_callback = LambdaCallback(on_batch_begin=lambda batch, logs: print(batch))
epoch_print_callback = LambdaCallback(on_epoch_end=lambda epoch, logs: print("epoch:", epoch))
callbacks.append(earlystop_callback)
callbacks.append(batch_print_callback)
callbacks.append(epoch_print_callback)
batch_size = 512
model.fit(X_train, y_train, nb_epoch=400, \
batch_size=batch_size, \
validation_split=0.2, \
shuffle=True, verbose=0, \
callbacks=callbacks)
model.save_weights('my_model_weights.h5')
scores = model.evaluate(X_train, y_train, verbose=0)
print ("Train loss : %.3f" % scores[0])
print ("Train accuracy : %.3f" % scores[1])
print ("Training finished")
def main():
model = vgg.CNN_16('final_weights.h5')
print ('Image Prediction Mode')
img = preprocessing(cv2.imread('girl2.jpg'))
X = np.expand_dims(img, axis=0)
X = np.expand_dims(X, axis=0)
# predict_classes will enable us to select most probable class
result = model.predict_classes(X)
print(result)
def main():
model = vgg.CNN_16(
'C:/Users/Harsh/PycharmProjects/EMOTION_classifier_project/Model_Training/final_weights.h5'
)
print('Image Prediction Mode')
img = preprocessing(
cv2.imread(
'C:/Users/Harsh/PycharmProjects/EMOTION_classifier_project/girl2.jpg'
))
X = np.expand_dims(img, axis=0)
X = np.expand_dims(X, axis=0)
# predict_classes will enable us to select most probable class
result = model.predict_classes(X)
print(result)
valid_data = training[6000:] # 2000 for validation
train_data = training[:6000] # 6000 for train
print("test set size:", len(test_data))
print("validation set size:", len(valid_data))
print("train set size:", len(train_data))
train_set = CustomDataset(train_data, transform=transforms.ToTensor())
valid_set = CustomDataset(valid_data, transform=transforms.ToTensor())
test_set = CustomDataset(test_data, transform=transforms.ToTensor())
train_dataloader = DataLoader(dataset=train_set, batch_size=100, shuffle=True)
test_dataloader = DataLoader(dataset=test_set, batch_size=1, shuffle=True)
valid_dataloader = DataLoader(dataset=valid_set, batch_size=1, shuffle=True)
model = CNN().to(device)
def train(model, train_dataloader, valid_dataloader, device):
model.train()
criterion = nn.MultiLabelSoftMarginLoss()
optimizer = torch.optim.Adam(model.parameters(), lr=0.001)
for epoch in range(15):
model.train()
for i, (images, labels) in enumerate(train_dataloader):
images = images.to(device)
labels = labels.to(device)
predict = model(images)
optimizer.zero_grad()
loss = criterion(predict, labels)
'training_set_r10',
'training_set_r11',
]
training_set_dict = {}
for training_set in training_set_list:
training_set_dict.update({str(training_set): eval(training_set)})
training_set = torch.utils.data.ConcatDataset(training_set_dict.values())
training_loader = DataLoader(training_set,
batch_size=batch_size,
shuffle=True,
num_workers=in_workers)
# Define the model
model = CNN(as_gray=as_grayscale, use_convcoord=use_convcoord)
model.double().to(device=device)
print(model)
# Loss and optimizer
criterion = nn.SmoothL1Loss()
optimizer = torch.optim.SGD(model.parameters(),
lr=learning_rate,
weight_decay=1e-3)
#################################
# Train the model #
#################################
total_step = len(training_loader)
model_save_name = ''.join([
import sys,cv2
import numpy as np
import CNN_Model.CNN as vc
windowsName = 'Preview Screen'
CASCADE_PATH = "C:/Users/Harsh/PycharmProjects/EMOTION_classifier/Prediction_Video/haarcascade_frontalface_default.xml"
cascade = cv2.CascadeClassifier(CASCADE_PATH)
emotion = ['Angry', 'Fear', 'Happy','Sad', 'Surprise', 'Neutral']
model = vc.CNN_16('C:/Users/Harsh/PycharmProjects/EMOTION_classifier/Model_Training/final_weights.h5')
capture = cv2.VideoCapture("C:/Users/Harsh/PycharmProjects/EMOTION_classifier/Vid2.mp4")
#capture = cv2.VideoCapture(0)
def grayFace(image):
img_gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
img_gray = cv2.equalizeHist(img_gray)
return img_gray
def getFaceCoordinates(image):
img_gray = grayFace(image)
rects = cascade.detectMultiScale(
img_gray,
scaleFactor=1.3,
minNeighbors=2,
#scaleFactor=1.1,
#minNeighbors=3,
minSize=(48, 48)
)
return rects
import sys, cv2
import numpy as np
import CNN_Model.CNN as vc
windowsName = 'Preview Screen'
CASCADE_PATH = "D:/ML/emotiondetector/venv/Prediction_Video/haarcascade_frontalface_default.xml"
cascade = cv2.CascadeClassifier(CASCADE_PATH)
emotion = ['Angry', 'Fear', 'Happy', 'Sad', 'Surprise', 'Neutral']
model = vc.CNN_16('my_model_weights.h5')
capture = cv2.VideoCapture("Vid2.mp4")
def grayFace(image):
img_gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
img_gray = cv2.equalizeHist(img_gray)
return img_gray
def getFaceCoordinates(image):
img_gray = grayFace(image)
rects = cascade.detectMultiScale(
img_gray,
scaleFactor=1.3,
minNeighbors=2,
#scaleFactor=1.1,
#minNeighbors=3,
minSize=(48, 48))
return rects