def prepare_dataloader(tfrecord_dir, batch_size, subset="train"):
anchorobj = anchor.Anchor(img_size=550,
feature_map_size=[69, 35, 18, 9, 5],
aspect_ratio=[1, 0.5, 2],
scale=[24, 48, 96, 192, 384])
parser = yolact_parser.Parser(output_size=550,
anchor_instance=anchorobj,
match_threshold=0.5,
unmatched_threshold=0.5,
mode=subset)
files = tf.io.matching_files(os.path.join(tfrecord_dir, "coco_%s.*" % subset))
num_shards = tf.cast(tf.shape(files)[0], tf.int64)
shards = tf.data.Dataset.from_tensor_slices(files)
shards = shards.shuffle(num_shards)
shards = shards.repeat()
dataset = shards.interleave(tf.data.TFRecordDataset,
cycle_length=num_shards,
num_parallel_calls=tf.data.experimental.AUTOTUNE)
dataset = dataset.shuffle(buffer_size=2048)
dataset = dataset.map(map_func=parser, num_parallel_calls=tf.data.experimental.AUTOTUNE)
dataset = dataset.batch(batch_size)
dataset = dataset.prefetch(buffer_size=tf.data.experimental.AUTOTUNE)
return dataset
def prepare_dataloader(img_h, img_w, feature_map_size, protonet_out_size, aspect_ratio, scale, tfrecord_dir, batch_size, subset="train"):
anchorobj = anchor.Anchor(img_size_h=img_h,img_size_w=img_w,
feature_map_size=feature_map_size,
aspect_ratio=aspect_ratio,
scale=scale)
parser = yolact_parser.Parser(output_size=[img_h, img_w], # (h,w)
anchor_instance=anchorobj,
match_threshold=0.5,
unmatched_threshold=0.5,
mode=subset,
proto_output_size=[int(protonet_out_size[0]), int(protonet_out_size[1])])
files = tf.io.matching_files(os.path.join(tfrecord_dir, "%s.*" % subset))
num_shards = tf.cast(tf.shape(files)[0], tf.int64)
shards = tf.data.Dataset.from_tensor_slices(files)
shards = shards.shuffle(num_shards)
shards = shards.repeat()
dataset = shards.interleave(tf.data.TFRecordDataset,
cycle_length=num_shards,
num_parallel_calls=tf.data.experimental.AUTOTUNE)
dataset = dataset.shuffle(buffer_size=2048)
dataset = dataset.map(map_func=parser, num_parallel_calls=tf.data.experimental.AUTOTUNE)
dataset = dataset.batch(batch_size)
dataset = dataset.prefetch(buffer_size=tf.data.experimental.AUTOTUNE)
return dataset
def __init__(self, img_h, img_w, fpn_channels, num_class, num_mask, aspect_ratio, scales):
super(Yolact, self).__init__()
out = ['conv3_block4_out', 'conv4_block6_out', 'conv5_block3_out']
# use pre-trained ResNet50
# Keras BatchNormalization problem
# https://github.com/keras-team/keras/pull/9965#issuecomment-501933060
tf.keras.layers.BatchNormalization = FrozenBatchNormalization
base_model = tf.keras.applications.ResNet50(input_shape=(img_h, img_w, 3),
include_top=False,
layers=tf.keras.layers,
weights='imagenet')
# extract certain feature maps for FPN
self.backbone_resnet = tf.keras.Model(inputs=base_model.input,
outputs=[base_model.get_layer(x).output for x in out])
# Calculating feature map size
# https://stackoverflow.com/a/44242277/4582711
# https://github.com/tensorflow/tensorflow/issues/4297#issuecomment-246080982
self.feature_map_size = np.array([list(base_model.get_layer(x).output.shape[1:3]) for x in out])
out_height_p6 = np.ceil((self.feature_map_size[-1, 0]).astype(np.float32) / float(2))
out_width_p6 = np.ceil((self.feature_map_size[-1, 1]).astype(np.float32) / float(2))
out_height_p7 = np.ceil(out_height_p6 / float(2))
out_width_p7 = np.ceil(out_width_p6/ float(2))
self.feature_map_size = np.concatenate((self.feature_map_size, [[out_height_p6, out_width_p6], [out_height_p7, out_width_p7]]), axis=0)
self.protonet_out_size = self.feature_map_size[0]*2 # Only one upsampling on p3
self.backbone_fpn = FeaturePyramidNeck(fpn_channels)
self.protonet = ProtoNet(num_mask)
# semantic segmentation branch to boost feature richness
self.semantic_segmentation = tf.keras.layers.Conv2D(num_class-1, (1, 1), 1, padding="same",
kernel_initializer=tf.keras.initializers.glorot_uniform())
anchorobj = anchor.Anchor(img_size_h=img_h,img_size_w=img_w,
feature_map_size=self.feature_map_size,
aspect_ratio=aspect_ratio,
scale=scales)
self.num_anchors = anchorobj.num_anchors
self.priors = anchorobj.anchors
# print("prior shape:", self.priors.shape)
# print("num anchor per feature map: ", self.num_anchor)
# shared prediction head
# Here, len(aspect_ratio) is passed as during prior calculations, individula scale is selected for each layer.
# So, when scale are [24, 48, 96, 130, 192] that means 24 is for p3; 48 is for p4 and so on.
# So, number of priors for that layer will be HxWxlen(aspect_ratio)
# Hence, passing len(aspect_ratio)
# This implementation differs from the original used in yolact
self.predictionHead = PredictionModule(256, len(aspect_ratio), num_class, num_mask)
# post-processing for evaluation
self.detect = Detect(num_class, bkg_label=0, top_k=200,
conf_thresh=0.05, nms_thresh=0.5)
self.max_output_size = 300
import numpy as np
import tensorflow as tf
from data import anchor
test_bbox = tf.convert_to_tensor(
(np.array([[204.044, 253.8351, 487.8226, 427.06363],
[0, 140.01741, 550, 290.21936],
[40.005028, 117.37102, 255.7913, 205.13097],
[263.31314, 67.0434, 514.04736, 124.48139],
[0, 503.79834, 487.0279, 550]])),
dtype=tf.float32)
test_labels = tf.convert_to_tensor((np.array([[1], [2], [3], [4], [5]])),
dtype=tf.float32)
anchorobj = anchor.Anchor(img_size=550,
feature_map_size=[69, 35, 18, 9, 5],
aspect_ratio=[1, 0.5, 2],
scale=[24, 48, 96, 192, 384])
print(anchorobj.get_anchors())
target_cls, target_loc, max_id_for_anchors, match_positiveness = anchorobj.matching(
threshold_pos=0.5,
threshold_neg=0.4,
gt_bbox=test_bbox,
gt_labels=test_labels)
print(target_loc)
aspect_ratio=[1, 0.5, 2],
scales=[24, 48, 96, 192, 384])
model = YOLACT.gen()
ckpt_dir = "checkpoints-SGD"
latest = tf.train.latest_checkpoint(ckpt_dir)
checkpoint = tf.train.Checkpoint(optimizer=optimizer, model=model)
status = checkpoint.restore(tf.train.latest_checkpoint(ckpt_dir))
print("Restore Ckpt Sucessfully!!")
# Load Validation Images and do Detection
# -----------------------------------------------------------------------------------------------
# Need default anchor
anchorobj = anchor.Anchor(img_size=256, feature_map_size=[32, 16, 8, 4, 2], aspect_ratio=[1, 0.5, 2], scale=[24, 48, 96, 192, 384])
valid_dataset = dataset_coco.prepare_evalloader(img_size=256,
tfrecord_dir='data/obj_tfrecord_256x256_20200921',
subset='val')
anchors = anchorobj.get_anchors()
detect_layer = Detect(num_cls=13, label_background=0, top_k=200, conf_threshold=0.3, nms_threshold=0.5, anchors=anchors)
remapping = [
'Background',
'Face',
'Body',
'Bicycle',
'Car',
'Motorbike',
'Airplane',
'Ship',