Sandeep Kumar Panda sandeeppanda22

Useful materials:

	def display_yolo(file, model, score_threshold, iou_threshold):
	'''
	Display predictions from YOLO model.

	Parameters
	----------
	- file : string list : list of images path.
	- model : YOLO model.
	- score_threshold : threshold used for filtering predicted bounding boxes.
	- iou_threshold : threshold used for non max suppression.

	def train(epochs, model, train_dataset, val_dataset, steps_per_epoch_train, steps_per_epoch_val, train_name = 'train'):
	'''
	Train YOLO model for n epochs.
	Eval loss on training and validation dataset.
	Log training loss and validation loss for tensorboard.
	Save best weights during training (according to validation loss).

	Parameters
	----------
	- epochs : integer, number of epochs to train the model.

	def yolov2_loss(detector_mask, matching_true_boxes, class_one_hot, true_boxes_grid, y_pred, info=False):
	'''
	Calculate YOLO V2 loss from prediction (y_pred) and ground truth tensors (detector_mask,
	matching_true_boxes, class_one_hot, true_boxes_grid,)

	Parameters
	----------
	- detector_mask : tensor, shape (batch, size, GRID_W, GRID_H, anchors_count, 1)
	1 if bounding box detected by grid cell, else 0
	- matching_true_boxes : tensor, shape (batch_size, GRID_W, GRID_H, anchors_count, 5)

	def augmentation_generator(yolo_dataset):
	'''
	Augmented batch generator from a yolo dataset

	Parameters
	----------
	- YOLO dataset

	Returns
	-------

	def parse_function(img_obj, true_boxes):
	x_img_string = tf.io.read_file(img_obj)
	x_img = tf.image.decode_png(x_img_string, channels=3) # dtype=tf.uint8
	x_img = tf.image.convert_image_dtype(x_img, tf.float32) # pixel value /255, dtype=tf.float32, channels : RGB
	return x_img, true_boxes

	def get_dataset(img_dir, ann_dir, labels, batch_size):
	'''
	Create a YOLO dataset

	val_dataset = None
	val_dataset= get_dataset(val_image_folder, val_annot_folder, LABELS, VAL_BATCH_SIZE)

	train_dataset = None
	train_dataset= get_dataset(train_image_folder, train_annot_folder, LABELS, TRAIN_BATCH_SIZE)

	class WeightReader:
	def __init__(self, weight_file):
	self.offset = 4
	self.all_weights = np.fromfile(weight_file, dtype='float32')

	def read_bytes(self, size):
	self.offset = self.offset + size
	return self.all_weights[self.offset-size:self.offset]

	def reset(self):

	# Yolo model (thanks to https://github.com/experiencor/keras-yolo2)

	input_image = tf.keras.layers.Input((IMAGE_H, IMAGE_W, 3), dtype='float32')

	# Layer 1
	x = Conv2D(32, (3,3), strides=(1,1), padding='same', name='conv_1', use_bias=False)(input_image)
	x = BatchNormalization(name='norm_1')(x)
	x = LeakyReLU(alpha=0.1)(x)
	x = MaxPooling2D(pool_size=(2, 2))(x)