hetong007 commented on a change in pull request #11027: Add standard ResNet
data augmentation for ImageRecordIter
URL: https://github.com/apache/incubator-mxnet/pull/11027#discussion_r194187113
##########
File path: example/image-classification/common/data.py
##########
@@ -63,6 +65,20 @@ def add_data_aug_args(parser):
help='max ratio to scale')
aug.add_argument('--min-random-scale', type=float, default=1,
help='min ratio to scale, should >= img_size/input_shape.
otherwise use --pad-size')
+ aug.add_argument('--max-random-area', type=float, default=1,
+ help='max area to crop in random resized crop, whose
range is [0, 1]')
+ aug.add_argument('--min-random-area', type=float, default=1,
+ help='min area to crop in random resized crop, whose
range is [0, 1]')
+ aug.add_argument('--brightness', type=float, default=0,
Review comment:
My opinion is that by setting default values to perform no augmentation,
users know exactly what kind of augmentation has been applied to the pipeline.
It is transparent and will cause less confusion.
On the other hand, not every model uses the same augmentation as ResNet to
train on ImageNet. If later another work introduces a better augmentation
pipeline with different parameters, then we are unable to justify our default
parameters.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
