def get_pipeline(folder="train", custom_reader=None):
pipe = Pipeline(batch_size=64, num_threads=1, device_id=1)
if custom_reader:
raw_files, labels = custom_reader
else:
raw_files, labels = fn.file_reader(file_root="%s" % folder,
random_shuffle=True)
decode = fn.image_decoder(raw_files, device="mixed", output_type=types.GRAY)
resize = fn.resize(decode, device="gpu", image_type=types.RGB,
interp_type=types.INTERP_LINEAR, resize_x=WIDTH, resize_y=HEIGHT)
hsv = fn.hsv(resize, hue=fn.uniform(range=(-10, 10)), saturation=fn.uniform(range=(-.5, .5)),
value=fn.uniform(range=(0.9, 1.2)), device="gpu", dtype=types.UINT8)
bc = fn.brightness_contrast(hsv, device="gpu", brightness=fn.uniform(range=(.9, 1.1)))
cmn = fn.crop_mirror_normalize(bc, device="gpu", output_dtype=types.FLOAT,
output_layout=types.NHWC,
image_type=types.GRAY,
mean=[255 // 2],
std=[255 // 2])
rot = fn.rotate(cmn, angle=fn.uniform(range=(-40, 40)), device="gpu", keep_size=True)
tpose = fn.transpose(rot, perm=(2, 0, 1), device="gpu") # Reshaping to a format PyTorch likes
pipe.set_outputs(tpose, labels)
pipe.build()
dali_iter = DALIClassificationIterator([pipe], -1)
return dali_iter
def create_coco_pipeline(file_root,
annotations_file,
batch_size=1,
device_id=0,
num_threads=4,
local_rank=0,
world_size=1):
pipeline = Pipeline(batch_size, num_threads,
local_rank, seed=42 + device_id)
with pipeline:
images, bboxes, labels = fn.coco_reader(file_root=file_root,
annotations_file=annotations_file,
skip_empty=True,
shard_id=local_rank,
num_shards=world_size,
ratio=True,
ltrb=True,
random_shuffle=False,
shuffle_after_epoch=True,
name="Reader")
crop_begin, crop_size, bboxes, labels = fn.random_bbox_crop(bboxes, labels,
device="cpu",
aspect_ratio=[0.5, 2.0],
thresholds=[0, 0.1, 0.3, 0.5, 0.7, 0.9],
scaling=[0.3, 1.0],
bbox_layout="xyXY",
allow_no_crop=True,
num_attempts=50)
images = fn.image_decoder_slice(images, crop_begin, crop_size, device="mixed", output_type=types.RGB)
flip_coin = fn.coin_flip(probability=0.5)
images = fn.resize(images,
resize_x=300,
resize_y=300,
min_filter=types.DALIInterpType.INTERP_TRIANGULAR)
# use float to avoid clipping and quantizing the intermediate result
images = fn.hsv(images, dtype=types.FLOAT, hue=fn.uniform(range=[-0.5, 0.5]),
saturation=fn.uniform(range=[0.5, 1.5]))
images = fn.brightness_contrast(images,
contrast_center = 128, # input is in float, but in 0..255 range
dtype = types.UINT8,
brightness = fn.uniform(range=[0.875, 1.125]),
contrast = fn.uniform(range=[0.5, 1.5]))
bboxes = fn.bb_flip(bboxes, ltrb=True, horizontal=flip_coin)
images = fn.crop_mirror_normalize(images,
mean=[104., 117., 123.],
std=[1., 1., 1.],
mirror=flip_coin,
dtype=types.FLOAT,
output_layout="CHW",
pad_output=False)
pipeline.set_outputs(images, bboxes, labels)
return pipeline
def create_coco_pipeline(default_boxes, args):
try:
shard_id = torch.distributed.get_rank()
num_shards = torch.distributed.get_world_size()
except RuntimeError:
shard_id = 0
num_shards = 1
images, bboxes, labels = fn.readers.coco(
file_root=args.train_coco_root,
annotations_file=args.train_annotate,
skip_empty=True,
shard_id=shard_id,
num_shards=num_shards,
ratio=True,
ltrb=True,
random_shuffle=False,
shuffle_after_epoch=True,
name="Reader")
crop_begin, crop_size, bboxes, labels = fn.random_bbox_crop(
bboxes,
labels,
device="cpu",
aspect_ratio=[0.5, 2.0],
thresholds=[0, 0.1, 0.3, 0.5, 0.7, 0.9],
scaling=[0.3, 1.0],
bbox_layout="xyXY",
allow_no_crop=True,
num_attempts=50)
images = fn.image_decoder_slice(images,
crop_begin,
crop_size,
device="mixed",
output_type=types.RGB)
flip_coin = fn.random.coin_flip(probability=0.5)
images = fn.resize(images,
resize_x=300,
resize_y=300,
min_filter=types.DALIInterpType.INTERP_TRIANGULAR)
saturation = fn.uniform(range=[0.5, 1.5])
contrast = fn.uniform(range=[0.5, 1.5])
brightness = fn.uniform(range=[0.875, 1.125])
hue = fn.uniform(range=[-0.5, 0.5])
images = fn.hsv(images, dtype=types.FLOAT, hue=hue,
saturation=saturation) # use float to avoid clipping and
# quantizing the intermediate result
images = fn.brightness_contrast(
images,
contrast_center=128, # input is in float, but in 0..255 range
dtype=types.UINT8,
brightness=brightness,
contrast=contrast)
dtype = types.FLOAT16 if args.fp16 else types.FLOAT
bboxes = fn.bb_flip(bboxes, ltrb=True, horizontal=flip_coin)
images = fn.crop_mirror_normalize(
images,
crop=(300, 300),
mean=[0.485 * 255, 0.456 * 255, 0.406 * 255],
std=[0.229 * 255, 0.224 * 255, 0.225 * 255],
mirror=flip_coin,
dtype=dtype,
output_layout="CHW",
pad_output=False)
bboxes, labels = fn.box_encoder(bboxes,
labels,
criteria=0.5,
anchors=default_boxes.as_ltrb_list())
labels = labels.gpu()
bboxes = bboxes.gpu()
return images, bboxes, labels
def train_pipeline(cfg: TrainLoaderConfig):
jpeg, label = fn.readers.file(
file_root=ROOT_DATA_DIR + "/train/",
random_shuffle=True,
shard_id=env_rank(),
num_shards=env_world_size(),
name="Reader",
)
image = fn.decoders.image_random_crop(
jpeg,
device="mixed",
random_aspect_ratio=[0.75, 1.25],
random_area=[cfg.min_area, 1.0],
num_attempts=100,
output_type=types.RGB,
)
image_tr = fn.resize(image,
device="gpu",
size=cfg.image_size,
interp_type=types.INTERP_TRIANGULAR)
if cfg.random_interpolation:
image_cub = fn.resize(image,
device="gpu",
size=cfg.image_size,
interp_type=types.INTERP_CUBIC)
image = mix(fn.random.coin_flip(probability=0.5), image_cub, image_tr)
else:
image = image_tr
if cfg.blur_prob > 0:
blur_image = fn.gaussian_blur(
image,
device="gpu",
window_size=11,
sigma=fn.random.uniform(range=[0.5, 1.1]))
image = mix(
fn.random.coin_flip(probability=cfg.blur_prob, dtype=types.BOOL),
blur_image, image)
if cfg.color_twist_prob > 0:
image_ct = fn.color_twist(
image,
device="gpu",
contrast=fn.random.uniform(range=[0.7, 1.3]),
brightness=fn.random.uniform(range=[0.7, 1.3]),
hue=fn.random.uniform(range=[-20, 20]), # in degrees
saturation=fn.random.uniform(range=[0.7, 1.3]),
)
image = mix(
fn.random.coin_flip(probability=cfg.color_twist_prob,
dtype=types.BOOL), image_ct, image)
if cfg.gray_prob > 0:
grayscale_coin = fn.cast(
fn.random.coin_flip(probability=cfg.gray_prob), dtype=types.FLOAT)
image = fn.hsv(image, device="gpu", saturation=grayscale_coin)
if cfg.re_prob: # random erasing
image_re = fn.erase(
image,
device="gpu",
anchor=fn.random.uniform(range=(0.0, 1), shape=cfg.re_count * 2),
shape=fn.random.uniform(range=(0.05, 0.25),
shape=cfg.re_count * 2),
axis_names="HW",
fill_value=DATA_MEAN,
normalized_anchor=True,
normalized_shape=True,
)
image = mix(
fn.random.coin_flip(probability=cfg.re_prob, dtype=types.BOOL),
image_re, image)
image = fn.crop_mirror_normalize(
image,
device="gpu",
crop=(cfg.image_size, cfg.image_size),
mirror=fn.random.coin_flip(probability=0.5),
mean=DATA_MEAN,
std=DATA_STD,
dtype=types.FLOAT,
output_layout=types.NCHW,
)
label = fn.one_hot(label, num_classes=cfg.num_classes).gpu()
return image, label