diff options
| author | Soumith Chintala <soumith@gmail.com> | 2016-07-04 15:38:55 +0300 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2016-07-04 15:38:55 +0300 |
| commit | 6894a4502a6f692cd6408b780425a2a88f33370b (patch) | |
| tree | 6f6cad0dbfbbd61857a937d88991cb11611f2fcc | |
| parent | 2e792b56428ec18a39ed8c0b56b026fef5eb2822 (diff) | |
| parent | 6d02b954d5e1be3438238a9ab1aeac7614a97615 (diff) | |
Merge pull request #174 from clement-masson/patch-1
Add details about image saving in the doc
| -rw-r--r-- | doc/saveload.md | 1 |
1 files changed, 1 insertions, 0 deletions
diff --git a/doc/saveload.md b/doc/saveload.md index e9e144f..d90fc20 100644 --- a/doc/saveload.md +++ b/doc/saveload.md @@ -35,6 +35,7 @@ local img = image.load(imagefile,1,'byte') Saves Tensor `tensor` to disk at path `filename`. The format to which the image is saved is extrapolated from the `filename`'s extension suffix. The `tensor` should be of size `nChannel x height x width`. +To save with a minimal loss, the tensor values should lie in the range [0, 1] since the tensor is clamped between 0 and 1 before being saved to the disk. <a name="image.decompressJPG"></a> ### [res] image.decompressJPG(tensor, [depth, tensortype]) ### |