def __init__(self, data_path, batch_size, num_thread, crop, rali_cpu=True):
super(valPipeline, self).__init__(batch_size,
num_thread,
rali_cpu=rali_cpu)
world_size = 1
local_rank = 0
self.input = ops.FileReader(file_root=data_path,
shard_id=local_rank,
num_shards=world_size,
random_shuffle=True)
rali_device = 'cpu' if rali_cpu else 'gpu'
decoder_device = 'cpu' if rali_cpu else 'mixed'
device_memory_padding = 211025920 if decoder_device == 'mixed' else 0
host_memory_padding = 140544512 if decoder_device == 'mixed' else 0
self.decode = ops.ImageDecoder(device=decoder_device,
output_type=types.RGB)
self.res = ops.Resize(device=rali_device, resize_x=256, resize_y=256)
self.centrecrop = ops.CentreCrop(crop=(crop, crop))
self.cmnp = ops.CropMirrorNormalize(
device="gpu",
output_dtype=types.FLOAT,
output_layout=types.NCHW,
crop=(crop, crop),
image_type=types.RGB,
mean=[0.485 * 255, 0.456 * 255, 0.406 * 255],
std=[0.229 * 255, 0.224 * 255, 0.225 * 255])
print('rali "{0}" variant'.format(rali_device))
def __init__(self, data_path, batch_size, num_thread, crop, rali_cpu=True):
super(trainPipeline, self).__init__(batch_size,
num_thread,
rali_cpu=rali_cpu)
world_size = 1
local_rank = 0
self.input = ops.FileReader(file_root=data_path,
shard_id=local_rank,
num_shards=world_size,
random_shuffle=True)
rali_device = 'cpu' if rali_cpu else 'gpu'
decoder_device = 'cpu' if rali_cpu else 'mixed'
device_memory_padding = 211025920 if decoder_device == 'mixed' else 0
host_memory_padding = 140544512 if decoder_device == 'mixed' else 0
self.decode = ops.ImageDecoderRandomCrop(
device=decoder_device,
output_type=types.RGB,
device_memory_padding=device_memory_padding,
host_memory_padding=host_memory_padding,
random_aspect_ratio=[0.8, 1.25],
random_area=[0.1, 1.0],
num_attempts=100)
self.res = ops.Resize(device=rali_device, resize_x=crop, resize_y=crop)
self.cmnp = ops.CropMirrorNormalize(
device="gpu",
output_dtype=types.FLOAT,
output_layout=types.NCHW,
crop=(crop, crop),
image_type=types.RGB,
mean=[0.485 * 255, 0.456 * 255, 0.406 * 255],
std=[0.229 * 255, 0.224 * 255, 0.225 * 255])
self.coin = ops.CoinFlip(probability=0.5)
print('rali "{0}" variant'.format(rali_device))
def __init__(self, batch_size, num_threads, device_id, data_dir, crop, rali_cpu = True,rali_type=True ):
super(HybridTrainPipe, self).__init__(batch_size, num_threads, device_id, seed=12 + device_id,rali_cpu=rali_cpu)
self.box = False if rali_type else True
self.input = ops.CaffeReader(path = data_dir, bbox=self.box,random_shuffle=True)
self.rali_type = rali_type
rali_device = 'cpu' if rali_cpu else 'gpu'
decoder_device = 'cpu' if rali_cpu else 'mixed'
device_memory_padding = 211025920 if decoder_device == 'mixed' else 0
host_memory_padding = 140544512 if decoder_device == 'mixed' else 0
self.decode = ops.ImageDecoderRandomCrop(device=decoder_device, output_type=types.RGB,
device_memory_padding=device_memory_padding,
host_memory_padding=host_memory_padding,
random_aspect_ratio=[0.8, 1.25],
random_area=[0.1, 1.0],
num_attempts=100)
self.res = ops.Resize(device=rali_device, resize_x=crop, resize_y=crop)
self.cmnp = ops.CropMirrorNormalize(device="gpu",
output_dtype=types.FLOAT,
output_layout=types.NCHW,
crop=(crop, crop),
image_type=types.RGB,
mean=[0.485 * 255,0.456 * 255,0.406 * 255],
std=[0.229 * 255,0.224 * 255,0.225 * 255])
self.one_hot_labels = ops.OneHot(num_classes=1000)
self.coin = ops.CoinFlip(probability=0.5)
print('rali "{0}" variant'.format(rali_device))
def __init__(self, batch_size, num_threads, device_id, data_dir,ann_dir, crop, rali_cpu = True):
super(COCOPipeline, self).__init__(batch_size, num_threads, device_id, seed=12 + device_id,rali_cpu=rali_cpu)
self.input = ops.COCOReader(file_root = data_dir, annotations_file = ann_dir)
rali_device = 'cpu' if rali_cpu else 'gpu'
decoder_device = 'cpu' if rali_cpu else 'mixed'
# device_memory_padding = 211025920 if decoder_device == 'mixed' else 0
# host_memory_padding = 140544512 if decoder_device == 'mixed' else 0
# self.decode = ops.ImageDecoderRandomCrop(device=decoder_device, output_type=types.RGB,
# device_memory_padding=device_memory_padding,
# host_memory_padding=host_memory_padding,
# random_aspect_ratio=[0.8, 1.25],
# random_area=[0.1, 1.0],
# num_attempts=100)
self.decode = ops.ImageDecoder(device=decoder_device, output_type=types.RGB)
self.crop = ops.SSDRandomCrop(num_attempts=5)
self.res = ops.Resize(device=rali_device, resize_x=crop, resize_y=crop)
self.twist = ops.ColorTwist(device=rali_device)
self.cmnp = ops.CropMirrorNormalize(device="gpu",
output_dtype=types.FLOAT,
output_layout=types.NCHW,
crop=(crop, crop),
image_type=types.RGB,
mirror=0,
mean=[0.485 * 255,0.456 * 255,0.406 * 255],
std=[0.229 * 255,0.224 * 255,0.225 * 255])
# Random variables
self.rng1 = ops.Uniform(range=[0.5, 1.5])
self.rng2 = ops.Uniform(range=[0.875, 1.125])
self.rng3 = ops.Uniform(range=[-0.5, 0.5])
print('rali "{0}" variant'.format(rali_device))
def __init__(self,
batch_size,
num_threads,
device_id,
data_dir,
crop,
rali_cpu=True):
super(HybridPipe, self).__init__(batch_size,
num_threads,
device_id,
seed=12 + device_id,
rali_cpu=rali_cpu)
self.input = ops.TFRecordReader(path=data_dir,
index_path="",
features={
'image/encoded':
tf.FixedLenFeature((), tf.string,
""),
'image/class/label':
tf.FixedLenFeature([1], tf.int64,
-1),
'image/class/text':
tf.FixedLenFeature([], tf.string,
''),
'image/object/bbox/xmin':
tf.VarLenFeature(dtype=tf.float32),
'image/object/bbox/ymin':
tf.VarLenFeature(dtype=tf.float32),
'image/object/bbox/xmax':
tf.VarLenFeature(dtype=tf.float32),
'image/object/bbox/ymax':
tf.VarLenFeature(dtype=tf.float32)
})
rali_device = 'cpu' if rali_cpu else 'gpu'
decoder_device = 'cpu' if rali_cpu else 'mixed'
device_memory_padding = 211025920 if decoder_device == 'mixed' else 0
host_memory_padding = 140544512 if decoder_device == 'mixed' else 0
self.decode = ops.ImageDecoderRandomCrop(
device=decoder_device,
output_type=types.RGB,
device_memory_padding=device_memory_padding,
host_memory_padding=host_memory_padding,
random_aspect_ratio=[0.8, 1.25],
random_area=[0.1, 1.0],
num_attempts=100)
self.cmnp = ops.CropMirrorNormalize(
device="cpu",
output_dtype=types.FLOAT,
output_layout=types.NCHW,
crop=(crop, crop),
image_type=types.RGB,
mean=[0.485 * 255, 0.456 * 255, 0.406 * 255],
std=[0.229 * 255, 0.224 * 255, 0.225 * 255])
self.coin = ops.CoinFlip(probability=0.5)
print('rali "{0}" variant'.format(rali_device))
def __init__(self,
batch_size,
num_threads,
device_id,
data_dir,
crop,
rali_cpu=True):
super(HybridTrainPipe, self).__init__(batch_size,
num_threads,
device_id,
seed=12 + device_id,
rali_cpu=rali_cpu)
world_size = 1
local_rank = 0
self.input = ops.FileReader(file_root=data_dir,
shard_id=local_rank,
num_shards=world_size,
random_shuffle=True)
rali_device = 'cpu' if rali_cpu else 'gpu'
decoder_device = 'cpu' if rali_cpu else 'mixed'
device_memory_padding = 211025920 if decoder_device == 'mixed' else 0
host_memory_padding = 140544512 if decoder_device == 'mixed' else 0
self.decode = ops.ImageDecoderRandomCrop(
device=decoder_device,
output_type=types.RGB,
device_memory_padding=device_memory_padding,
host_memory_padding=host_memory_padding,
random_aspect_ratio=[0.8, 1.25],
random_area=[0.1, 1.0],
num_attempts=100)
self.res = ops.Resize(device=rali_device, resize_x=crop, resize_y=crop)
# #self.res = ops.Crop(crop=(crop, crop))
# self.rain = ops.Rain(rain=0.5)
# self.blur = ops.Blur(blur=0.5)
# self.jitter = ops.Jitter()
# self.contrast =ops.Rotate(angle=20)
# self.hue = ops.Hue()
# self.blend = ops.Blend(blend=0.5)
# self.snp = ops.SnPNoise(snpNoise = 0.5)
# self.ving = ops.Vignette(vignette = 0.2)
# self.exp = ops.Exposure(exposure = 0.2)
# #self.wf = ops.WarpAffine()
# self.sat = ops.Saturation()
self.cmnp = ops.CropMirrorNormalize(
device="gpu",
output_dtype=types.FLOAT,
output_layout=types.NCHW,
crop=(crop, crop),
image_type=types.RGB,
mean=[0.485 * 255, 0.456 * 255, 0.406 * 255],
std=[0.229 * 255, 0.224 * 255, 0.225 * 255])
self.coin = ops.CoinFlip(probability=0.5)
print('rali "{0}" variant'.format(rali_device))
def __init__(self,
feature_key_map,
tfrecordreader_type,
batch_size,
num_threads,
device_id,
data_dir,
crop,
rali_cpu=True):
super(HybridPipe, self).__init__(batch_size,
num_threads,
device_id,
seed=12 + device_id,
rali_cpu=rali_cpu)
self.input = ops.TFRecordReader(path=data_dir,
index_path="",
reader_type=tfrecordreader_type,
user_feature_key_map=feature_key_map,
features={
'image/encoded':
tf.FixedLenFeature((), tf.string,
""),
'image/class/label':
tf.FixedLenFeature([1], tf.int64,
-1),
'image/filename':
tf.FixedLenFeature((), tf.string,
"")
})
rali_device = 'cpu' if rali_cpu else 'gpu'
decoder_device = 'cpu' if rali_cpu else 'mixed'
self.decode = ops.ImageDecoder(user_feature_key_map=feature_key_map,
device=decoder_device,
output_type=types.RGB)
self.res = ops.Resize(device=rali_device,
resize_x=crop[0],
resize_y=crop[1])
self.cmnp = ops.CropMirrorNormalize(device="cpu",
output_dtype=types.FLOAT,
output_layout=types.NCHW,
crop=crop,
image_type=types.RGB,
mean=[0, 0, 0],
std=[255, 255, 255])
self.coin = ops.CoinFlip(probability=0.5)
print('rali "{0}" variant'.format(rali_device))
def __init__(self,
batch_size,
num_threads,
device_id,
data_dir,
ann_dir,
default_boxes,
crop,
rali_cpu=True):
super(COCOPipeline, self).__init__(batch_size,
num_threads,
device_id,
seed=12 + device_id,
rali_cpu=rali_cpu)
self.input = ops.COCOReader(file_root=data_dir,
annotations_file=ann_dir)
rali_device = 'cpu' if rali_cpu else 'gpu'
decoder_device = 'cpu' if rali_cpu else 'mixed'
# device_memory_padding = 211025920 if decoder_device == 'mixed' else 0
# host_memory_padding = 140544512 if decoder_device == 'mixed' else 0
# self.decode = ops.ImageDecoderRandomCrop(device=decoder_device, output_type=types.RGB,
# device_memory_padding=device_memory_padding,
# host_memory_padding=host_memory_padding,
# random_aspect_ratio=[0.8, 1.25],
# random_area=[0.1, 1.0],
# num_attempts=100)
self.decode = ops.ImageDecoder(device=decoder_device,
output_type=types.RGB)
self.crop = ops.SSDRandomCrop(num_attempts=5)
self.decode_slice = ops.ImageDecoderSlice(device=decoder_device,
output_type=types.RGB)
self.random_bbox_crop = ops.RandomBBoxCrop(
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],
ltrb=True,
allow_no_crop=True,
num_attempts=1)
self.res = ops.Resize(device=rali_device, resize_x=crop, resize_y=crop)
self.twist = ops.ColorTwist(device=rali_device)
self.bbflip = ops.BBFlip(device=rali_device, ltrb=True)
self.cmnp = ops.CropMirrorNormalize(
device="gpu",
output_dtype=types.FLOAT,
output_layout=types.NCHW,
crop=(crop, crop),
image_type=types.RGB,
mirror=0,
mean=[0.485 * 255, 0.456 * 255, 0.406 * 255],
std=[0.229 * 255, 0.224 * 255, 0.225 * 255])
self.boxEncoder = ops.BoxEncoder(device=rali_device,
criteria=0.5,
anchors=default_boxes)
self.cast = ops.Cast(device=rali_device, dtype=types.FLOAT)
# Random variables
self.rng1 = ops.Uniform(range=[0.5, 1.5])
self.rng2 = ops.Uniform(range=[0.875, 1.125])
self.rng3 = ops.Uniform(range=[-0.5, 0.5])
self.coin_flip = ops.CoinFlip(probability=0.5)
print('rali "{0}" variant'.format(rali_device))
