Files
2020-02-19 20:05:14 -03:00

445 lines
22 KiB
Python

"""
Created on Fri May 10 15:10:46 2019
@author: dlsaavedra
"""
from keras.optimizers import Adam, SGD
from keras.callbacks import ModelCheckpoint, LearningRateScheduler, TerminateOnNaN, CSVLogger
from keras import backend as K
from keras.models import load_model
from math import ceil
import numpy as np
from matplotlib import pyplot as plt
from models.keras_ssd512 import ssd_512
from models.keras_ssd300 import ssd_300
from keras_loss_function.keras_ssd_loss import SSDLoss
from keras_layers.keras_layer_AnchorBoxes import AnchorBoxes
from keras_layers.keras_layer_DecodeDetections import DecodeDetections
from keras_layers.keras_layer_DecodeDetectionsFast import DecodeDetectionsFast
from keras_layers.keras_layer_L2Normalization import L2Normalization
from ssd_encoder_decoder.ssd_input_encoder import SSDInputEncoder
from ssd_encoder_decoder.ssd_output_decoder import decode_detections, decode_detections_fast
from data_generator.object_detection_2d_data_generator import DataGenerator
from data_generator.object_detection_2d_geometric_ops import Resize
from data_generator.object_detection_2d_photometric_ops import ConvertTo3Channels
from data_generator.data_augmentation_chain_original_ssd import SSDDataAugmentation
from data_generator.object_detection_2d_misc_utils import apply_inverse_transforms
import json
import os
import argparse
K.tensorflow_backend._get_available_gpus()
def lr_schedule(epoch):
if epoch < 80:
return 0.001
elif epoch < 100:
return 0.0001
else:
return 0.00001
def _main_(args):
config_path = args.conf
with open(config_path) as config_buffer:
config = json.loads(config_buffer.read())
###############################
# Parse the annotations
###############################
path_imgs_training = config['train']['train_image_folder']
path_anns_training = config['train']['train_annot_folder']
path_imgs_val = config['valid']['valid_image_folder']
path_anns_val = config['valid']['valid_annot_folder']
labels = config['model']['labels']
categories = {}
#categories = {"Razor": 1, "Gun": 2, "Knife": 3, "Shuriken": 4} #la categoría 0 es la background
for i in range(len(labels)): categories[labels[i]] = i+1
print('\nTraining on: \t' + str(categories) + '\n')
####################################
# Parameters
###################################
#%%
img_height = config['model']['input'] # Height of the model input images
img_width = config['model']['input'] # Width of the model input images
img_channels = 3 # Number of color channels of the model input images
mean_color = [123, 117, 104] # The per-channel mean of the images in the dataset. Do not change this value if you're using any of the pre-trained weights.
swap_channels = [2, 1, 0] # The color channel order in the original SSD is BGR, so we'll have the model reverse the color channel order of the input images.
n_classes = len(labels) # Number of positive classes, e.g. 20 for Pascal VOC, 80 for MS COCO
scales_pascal = [0.1, 0.2, 0.37, 0.54, 0.71, 0.88, 1.05] # The anchor box scaling factors used in the original SSD300 for the Pascal VOC datasets
#scales_coco = [0.07, 0.15, 0.33, 0.51, 0.69, 0.87, 1.05] # The anchor box scaling factors used in the original SSD300 for the MS COCO datasets
scales = scales_pascal
aspect_ratios = [[1.0, 2.0, 0.5],
[1.0, 2.0, 0.5, 3.0, 1.0/3.0],
[1.0, 2.0, 0.5, 3.0, 1.0/3.0],
[1.0, 2.0, 0.5, 3.0, 1.0/3.0],
[1.0, 2.0, 0.5],
[1.0, 2.0, 0.5]] # The anchor box aspect ratios used in the original SSD300; the order matters
two_boxes_for_ar1 = True
steps = [8, 16, 32, 64, 100, 300] # The space between two adjacent anchor box center points for each predictor layer.
offsets = [0.5, 0.5, 0.5, 0.5, 0.5, 0.5] # The offsets of the first anchor box center points from the top and left borders of the image as a fraction of the step size for each predictor layer.
clip_boxes = False # Whether or not to clip the anchor boxes to lie entirely within the image boundaries
variances = [0.1, 0.1, 0.2, 0.2] # The variances by which the encoded target coordinates are divided as in the original implementation
normalize_coords = True
K.clear_session() # Clear previous models from memory.
model_path = config['train']['saved_weights_name']
# 3: Instantiate an optimizer and the SSD loss function and compile the model.
# If you want to follow the original Caffe implementation, use the preset SGD
# optimizer, otherwise I'd recommend the commented-out Adam optimizer.
if config['model']['backend'] == 'ssd512':
aspect_ratios = [[1.0, 2.0, 0.5],
[1.0, 2.0, 0.5, 3.0, 1.0/3.0],
[1.0, 2.0, 0.5, 3.0, 1.0/3.0],
[1.0, 2.0, 0.5, 3.0, 1.0/3.0],
[1.0, 2.0, 0.5, 3.0, 1.0/3.0],
[1.0, 2.0, 0.5],
[1.0, 2.0, 0.5]]
steps = [8, 16, 32, 64, 100, 200, 300] # The space between two adjacent anchor box center points for each predictor layer.
offsets = [0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5]
scales = [0.07, 0.15, 0.3, 0.45, 0.6, 0.75, 0.9, 1.05]
elif config['model']['backend'] == 'ssd7':
#weights_path = 'VGG_ILSVRC_16_layers_fc_reduced.h5'
scales = [0.08, 0.16, 0.32, 0.64, 0.96] # An explicit list of anchor box scaling factors. If this is passed, it will override `min_scale` and `max_scale`.
aspect_ratios = [0.5 ,1.0, 2.0] # The list of aspect ratios for the anchor boxes
two_boxes_for_ar1 = True # Whether or not you want to generate two anchor boxes for aspect ratio 1
steps = None # In case you'd like to set the step sizes for the anchor box grids manually; not recommended
offsets = None
if os.path.exists(model_path):
print("\nLoading pretrained weights.\n")
# We need to create an SSDLoss object in order to pass that to the model loader.
ssd_loss = SSDLoss(neg_pos_ratio=3, alpha=1.0)
K.clear_session() # Clear previous models from memory.
model = load_model(model_path, custom_objects={'AnchorBoxes': AnchorBoxes,
'L2Normalization': L2Normalization,
'compute_loss': ssd_loss.compute_loss})
else:
####################################
# Build the Keras model.
###################################
if config['model']['backend'] == 'ssd300':
#weights_path = 'VGG_VOC0712Plus_SSD_300x300_ft_iter_160000.h5'
from models.keras_ssd300 import ssd_300 as ssd
model = ssd_300(image_size=(img_height, img_width, img_channels),
n_classes=n_classes,
mode='training',
l2_regularization=0.0005,
scales=scales,
aspect_ratios_per_layer=aspect_ratios,
two_boxes_for_ar1=two_boxes_for_ar1,
steps=steps,
offsets=offsets,
clip_boxes=clip_boxes,
variances=variances,
normalize_coords=normalize_coords,
subtract_mean=mean_color,
swap_channels=swap_channels)
elif config['model']['backend'] == 'ssd512':
#weights_path = 'VGG_VOC0712Plus_SSD_512x512_ft_iter_160000.h5'
from models.keras_ssd512 import ssd_512 as ssd
# 2: Load some weights into the model.
model = ssd(image_size=(img_height, img_width, img_channels),
n_classes=n_classes,
mode='training',
l2_regularization=0.0005,
scales=scales,
aspect_ratios_per_layer=aspect_ratios,
two_boxes_for_ar1=two_boxes_for_ar1,
steps=steps,
offsets=offsets,
clip_boxes=clip_boxes,
variances=variances,
normalize_coords=normalize_coords,
swap_channels=swap_channels)
elif config['model']['backend'] == 'ssd7':
#weights_path = 'VGG_ILSVRC_16_layers_fc_reduced.h5'
from models.keras_ssd7 import build_model as ssd
scales = [0.08, 0.16, 0.32, 0.64, 0.96] # An explicit list of anchor box scaling factors. If this is passed, it will override `min_scale` and `max_scale`.
aspect_ratios = [0.5 ,1.0, 2.0] # The list of aspect ratios for the anchor boxes
two_boxes_for_ar1 = True # Whether or not you want to generate two anchor boxes for aspect ratio 1
steps = None # In case you'd like to set the step sizes for the anchor box grids manually; not recommended
offsets = None
model = ssd(image_size=(img_height, img_width, img_channels),
n_classes=n_classes,
mode='training',
l2_regularization=0.0005,
scales=scales,
aspect_ratios_global=aspect_ratios,
aspect_ratios_per_layer=None,
two_boxes_for_ar1=two_boxes_for_ar1,
steps=steps,
offsets=offsets,
clip_boxes=clip_boxes,
variances=variances,
normalize_coords=normalize_coords,
subtract_mean=None,
divide_by_stddev=None)
else :
print('Wrong Backend')
print('OK create model')
#sgd = SGD(lr=config['train']['learning_rate'], momentum=0.9, decay=0.0, nesterov=False)
# TODO: Set the path to the weights you want to load. only for ssd300 or ssd512
weights_path = 'VGG_ILSVRC_16_layers_fc_reduced.h5'
print("\nLoading pretrained weights VGG.\n")
model.load_weights(weights_path, by_name=True)
# 3: Instantiate an optimizer and the SSD loss function and compile the model.
# If you want to follow the original Caffe implementation, use the preset SGD
# optimizer, otherwise I'd recommend the commented-out Adam optimizer.
#adam = Adam(lr=0.001, beta_1=0.9, beta_2=0.999, epsilon=1e-08, decay=0.0)
#sgd = SGD(lr=0.001, momentum=0.9, decay=0.0, nesterov=False)
optimizer = Adam(lr=config['train']['learning_rate'], beta_1=0.9, beta_2=0.999, epsilon=1e-08, decay=0.0)
ssd_loss = SSDLoss(neg_pos_ratio=3, alpha=1.0)
model.compile(optimizer=optimizer, loss=ssd_loss.compute_loss)
model.summary()
#####################################################################
# Instantiate two `DataGenerator` objects: One for training, one for validation.
######################################################################
# Optional: If you have enough memory, consider loading the images into memory for the reasons explained above.
train_dataset = DataGenerator(load_images_into_memory=False, hdf5_dataset_path=None)
val_dataset = DataGenerator(load_images_into_memory=False, hdf5_dataset_path=None)
# 2: Parse the image and label lists for the training and validation datasets. This can take a while.
# The XML parser needs to now what object class names to look for and in which order to map them to integers.
classes = ['background'] + labels
train_dataset.parse_xml(images_dirs= [config['train']['train_image_folder']],
image_set_filenames=[config['train']['train_image_set_filename']],
annotations_dirs=[config['train']['train_annot_folder']],
classes=classes,
include_classes='all',
exclude_truncated=False,
exclude_difficult=False,
ret=False)
val_dataset.parse_xml(images_dirs= [config['valid']['valid_image_folder']],
image_set_filenames=[config['valid']['valid_image_set_filename']],
annotations_dirs=[config['valid']['valid_annot_folder']],
classes=classes,
include_classes='all',
exclude_truncated=False,
exclude_difficult=False,
ret=False)
#########################
# 3: Set the batch size.
#########################
batch_size = config['train']['batch_size'] # Change the batch size if you like, or if you run into GPU memory issues.
##########################
# 4: Set the image transformations for pre-processing and data augmentation options.
##########################
# For the training generator:
# For the validation generator:
convert_to_3_channels = ConvertTo3Channels()
resize = Resize(height=img_height, width=img_width)
######################################3
# 5: Instantiate an encoder that can encode ground truth labels into the format needed by the SSD loss function.
#########################################
# The encoder constructor needs the spatial dimensions of the model's predictor layers to create the anchor boxes.
if config['model']['backend'] == 'ssd512':
predictor_sizes = [model.get_layer('conv4_3_norm_mbox_conf').output_shape[1:3],
model.get_layer('fc7_mbox_conf').output_shape[1:3],
model.get_layer('conv6_2_mbox_conf').output_shape[1:3],
model.get_layer('conv7_2_mbox_conf').output_shape[1:3],
model.get_layer('conv8_2_mbox_conf').output_shape[1:3],
model.get_layer('conv9_2_mbox_conf').output_shape[1:3],
model.get_layer('conv10_2_mbox_conf').output_shape[1:3]]
ssd_input_encoder = SSDInputEncoder(img_height=img_height,
img_width=img_width,
n_classes=n_classes,
predictor_sizes=predictor_sizes,
scales=scales,
aspect_ratios_per_layer=aspect_ratios,
two_boxes_for_ar1=two_boxes_for_ar1,
steps=steps,
offsets=offsets,
clip_boxes=clip_boxes,
variances=variances,
matching_type='multi',
pos_iou_threshold=0.5,
neg_iou_limit=0.5,
normalize_coords=normalize_coords)
elif config['model']['backend'] == 'ssd300':
predictor_sizes = [model.get_layer('conv4_3_norm_mbox_conf').output_shape[1:3],
model.get_layer('fc7_mbox_conf').output_shape[1:3],
model.get_layer('conv6_2_mbox_conf').output_shape[1:3],
model.get_layer('conv7_2_mbox_conf').output_shape[1:3],
model.get_layer('conv8_2_mbox_conf').output_shape[1:3],
model.get_layer('conv9_2_mbox_conf').output_shape[1:3]]
ssd_input_encoder = SSDInputEncoder(img_height=img_height,
img_width=img_width,
n_classes=n_classes,
predictor_sizes=predictor_sizes,
scales=scales,
aspect_ratios_per_layer=aspect_ratios,
two_boxes_for_ar1=two_boxes_for_ar1,
steps=steps,
offsets=offsets,
clip_boxes=clip_boxes,
variances=variances,
matching_type='multi',
pos_iou_threshold=0.5,
neg_iou_limit=0.5,
normalize_coords=normalize_coords)
elif config['model']['backend'] == 'ssd7':
predictor_sizes = [model.get_layer('classes4').output_shape[1:3],
model.get_layer('classes5').output_shape[1:3],
model.get_layer('classes6').output_shape[1:3],
model.get_layer('classes7').output_shape[1:3]]
ssd_input_encoder = SSDInputEncoder(img_height=img_height,
img_width=img_width,
n_classes=n_classes,
predictor_sizes=predictor_sizes,
scales=scales,
aspect_ratios_global=aspect_ratios,
two_boxes_for_ar1=two_boxes_for_ar1,
steps=steps,
offsets=offsets,
clip_boxes=clip_boxes,
variances=variances,
matching_type='multi',
pos_iou_threshold=0.5,
neg_iou_limit=0.3,
normalize_coords=normalize_coords)
#######################
# 6: Create the generator handles that will be passed to Keras' `fit_generator()` function.
#######################
train_generator = train_dataset.generate(batch_size=batch_size,
shuffle=True,
transformations= [SSDDataAugmentation(img_height=img_height,img_width=img_width)],
label_encoder=ssd_input_encoder,
returns={'processed_images',
'encoded_labels'},
keep_images_without_gt=False)
val_generator = val_dataset.generate(batch_size=batch_size,
shuffle=False,
transformations=[convert_to_3_channels,
resize],
label_encoder=ssd_input_encoder,
returns={'processed_images',
'encoded_labels'},
keep_images_without_gt=False)
# Get the number of samples in the training and validations datasets.
train_dataset_size = train_dataset.get_dataset_size()
val_dataset_size = val_dataset.get_dataset_size()
print("Number of images in the training dataset:\t{:>6}".format(train_dataset_size))
print("Number of images in the validation dataset:\t{:>6}".format(val_dataset_size))
##########################
# Define model callbacks.
#########################
# TODO: Set the filepath under which you want to save the model.
model_checkpoint = ModelCheckpoint(filepath= config['train']['saved_weights_name'],
monitor='val_loss',
verbose=1,
save_best_only=True,
save_weights_only=False,
mode='auto',
period=1)
#model_checkpoint.best =
csv_logger = CSVLogger(filename='log.csv',
separator=',',
append=True)
learning_rate_scheduler = LearningRateScheduler(schedule=lr_schedule,
verbose=1)
terminate_on_nan = TerminateOnNaN()
callbacks = [model_checkpoint,
csv_logger,
learning_rate_scheduler,
terminate_on_nan]
#print(model.summary())
batch_images, batch_labels = next(train_generator)
# i = 0 # Which batch item to look at
#
# print("Image:", batch_filenames[i])
# print()
# print("Ground truth boxes:\n")
# print(batch_labels[i])
initial_epoch = 0
final_epoch = config['train']['nb_epochs']
#final_epoch = 20
steps_per_epoch = 500
history = model.fit_generator(generator=train_generator,
steps_per_epoch=steps_per_epoch,
epochs=final_epoch,
callbacks=callbacks,
validation_data=val_generator,
validation_steps=ceil(val_dataset_size/batch_size),
initial_epoch=initial_epoch,
verbose = 1 if config['train']['debug'] else 2)
if __name__ == '__main__':
argparser = argparse.ArgumentParser(description='train and evaluate ssd model on any dataset')
argparser.add_argument('-c', '--conf', help='path to configuration file')
args = argparser.parse_args()
_main_(args)