Name | Modified | Size | Downloads / Week |
---|---|---|---|
Parent folder | |||
Keras SSD v0.9.0.tar.gz | 2018-05-03 | 10.7 MB | |
Keras SSD v0.9.0.zip | 2018-05-03 | 10.7 MB | |
README.md | 2018-05-03 | 1.8 kB | |
Totals: 3 Items | 21.4 MB | 0 |
Release 0.9.0
Breaking Changes
- None
Major Features and Improvements
- Added a new, flexible
Evaluator
class that computes average precision scores. Among other things, it can compute average precisions according to both the Pascal VOC pre-2010 and post-2010 algorithms. - Added two new features to
DataGenerator
:- Convert the dataset into an HDF5 file: This stores the images of a dataset as uncompressed arrays in a contiguous block of memory within an HDF5 file, which requires a lot of disk space but reduces the image loading times during the batch generation.
- Load the entire dataset into memory: This loads all images of a dataset into memory, thereby eliminating image loading times altogether. Requires enough memory to hold the entire dataset.
For several minor other improvements please refer to the commits since the last release (v0.8.0).
Bug Fixes and Other Changes
- Fixed a bug in
DataGenerator.parse_xml()
: Before the fix there were cases in which the XML parser would parse the wrong bounding boxes for some objects. The only known situation in which this bug occurred is for the 'person' class of the Pascal VOC datasets, where the ground truth provides not only a bounding box for the person itself, but also additional bounding boxes for various body parts such. Depending on the order of these ground truth boxes within the XML files, the parser would sometimes parse the bounding box of a body part instead of the bounding box of the person. The parser now loads the correct bounding boxes in these cases. - Provided a better training/validation split for the Udacity traffic dataset. The new split is much more balanced than the old one.
API Changes
- None
Known Issues
- None