mirror of
https://github.com/sbrl/research-rainfallradar
synced 2024-11-22 01:12:59 +00:00
finish wiring for --water-size
This commit is contained in:
parent
5dac70aa08
commit
9f3ae96894
2 changed files with 5 additions and 2 deletions
|
@ -46,7 +46,7 @@ def parse_item(metadata, shape_water_desired, water_threshold=0.1):
|
|||
|
||||
return tf.function(parse_item_inner)
|
||||
|
||||
def make_dataset(filepaths, metadata, shape_watch_desired=[100,100], water_threshold=0.1, compression_type="GZIP", parallel_reads_multiplier=1.5, shuffle_buffer_size=128, batch_size=64, prefetch=True, shuffle=True):
|
||||
def make_dataset(filepaths, metadata, shape_water_desired=[100,100], water_threshold=0.1, compression_type="GZIP", parallel_reads_multiplier=1.5, shuffle_buffer_size=128, batch_size=64, prefetch=True, shuffle=True):
|
||||
if "NO_PREFETCH" in os.environ:
|
||||
logger.info("disabling data prefetching.")
|
||||
|
||||
|
@ -56,7 +56,7 @@ def make_dataset(filepaths, metadata, shape_watch_desired=[100,100], water_thres
|
|||
)
|
||||
if shuffle:
|
||||
dataset = dataset.shuffle(shuffle_buffer_size)
|
||||
dataset = dataset.map(parse_item(metadata, shape_water_desired=shape_watch_desired, water_threshold=water_threshold), num_parallel_calls=tf.data.AUTOTUNE)
|
||||
dataset = dataset.map(parse_item(metadata, shape_water_desired=shape_water_desired, water_threshold=water_threshold), num_parallel_calls=tf.data.AUTOTUNE)
|
||||
|
||||
if batch_size != None:
|
||||
dataset = dataset.batch(batch_size, drop_remainder=True)
|
||||
|
|
|
@ -33,6 +33,8 @@ def run(args):
|
|||
args.read_multiplier = 1.5
|
||||
if (not hasattr(args, "water_threshold")) or args.water_threshold == None:
|
||||
args.water_threshold = 1.5
|
||||
if (not hasattr(args, "water_size")) or args.water_size == None:
|
||||
args.water_size = 1.5
|
||||
|
||||
|
||||
# TODO: Validate args here.
|
||||
|
@ -43,6 +45,7 @@ def run(args):
|
|||
dirpath_input=args.input,
|
||||
batch_size=args.batch_size,
|
||||
water_threshold=args.water_threshold,
|
||||
shape_water_desired=[args.water_size, args.water_size]
|
||||
)
|
||||
dataset_metadata = read_metadata(args.input)
|
||||
|
||||
|
|
Loading…
Reference in a new issue