diff options
author | Geoffrey Irving <geoffreyi@google.com> | 2016-02-08 12:02:44 -0800 |
---|---|---|
committer | Vijay Vasudevan <vrv@google.com> | 2016-02-08 18:49:25 -0800 |
commit | 3e33d444c6c0861d448755627982284736ed01a0 (patch) | |
tree | 5eef06eeaa394760ffb6834de8c7074abcf19492 /RELEASE.md | |
parent | ff4e52b28dac0533d613973fe58a4b5b2888d349 (diff) |
Generalize tf.image.random_crop to dimension-independent tf.random_crop
The C++ 3-D-only RandomCrop op is now deprecated at GraphDef version 8, replaced
with a python tf.random_crop that works for any dimension. This will allow
random_crop to be used for other purposes.
Unfortunately, tf.image.random_crop took 2 sizes rather than 3 for 3-D tensors.
The new tf.random_crop always takes n sizes for rank n tensors; pass 3 as the
last element if you want to not crop a last dimension of size 3.
Change: 114135451
Diffstat (limited to 'RELEASE.md')
-rw-r--r-- | RELEASE.md | 4 |
1 files changed, 4 insertions, 0 deletions
diff --git a/RELEASE.md b/RELEASE.md index fd46bcdff3..b8146610c8 100644 --- a/RELEASE.md +++ b/RELEASE.md @@ -33,6 +33,10 @@ maintained for short-term compatibility but will be removed. * The non-public `nn.rnn` and the various `nn.seq2seq` methods now return just the final state instead of the list of all states. +* `tf.image.random_crop(image, [height, width])` is now + `tf.random_crop(image, [height, width, depth])`, and `tf.random_crop` works + for any rank (not just 3-D images). The C++ `RandomCrop` op has been replaced + with pure Python. ## Bug fixes |