- 04 4月, 2018 2 次提交
-
-
由 Waleed Abdulla 提交于
This doesn’t change the functionality, other than cleaning up the code to prepare for future changes to support detection on different image sizes.
-
由 Waleed Abdulla 提交于
In a previous commit, the detections tensor was changed to use normalized coordinates [0, 1]. This fixes the notebooks to visualize the updated detections format correctly.
-
- 03 4月, 2018 3 次提交
-
-
由 Waleed Abdulla 提交于
The image_meta tensor stores meta data about the image. This update adds the scale by which the image was resized and it’s final shape.
-
由 Waleed Abdulla 提交于
-
由 Waleed Abdulla 提交于
-
- 02 4月, 2018 6 次提交
-
-
由 Waleed Abdulla 提交于
-
由 Waleed Abdulla 提交于
Make DetectionLayer return normalized coordinates to avoid unnecessary conversion to pixels and back to normalized.
-
由 Waleed Abdulla 提交于
Move code to separate functions and fix off-by-1 error.
-
由 Waleed Abdulla 提交于
Allows using a config variable, TRAIN_NB, rather that updating the BatchNorm class.
-
由 Waleed Abdulla 提交于
-
由 Waleed Abdulla 提交于
skimage.transform.resize() crashes with a floating point exception if the image has 1-pixel width or height and the mode=“reflection”. Changing mode to “constant” (the default) fixes the issue. And, it’s the correct mode to use anyway.
-
- 30 3月, 2018 3 次提交
-
-
由 Waleed Abdulla 提交于
If the mask is a rectangle, its mini-mask pixels become all 1s, and it has no zeros. Scipy imresize fails on that and returns 0s instead. This update fixes the problem.
-
由 Waleed Abdulla 提交于
Requires installing imgaug (pip3 install imgaug) https://github.com/aleju/imgaug List of augmentations: http://imgaug.readthedocs.io/en/latest/source/augmenters.html
-
由 Waleed Abdulla 提交于
This is more reliable than using the batch size to determine worker count because load per batch of images varies depending on image sizes and where they’re loaded from.
-
- 27 3月, 2018 4 次提交
-
-
由 Waleed Abdulla 提交于
-
由 Waleed Abdulla 提交于
-
由 Waleed Abdulla 提交于
-
由 Waleed Abdulla 提交于
-
- 20 3月, 2018 1 次提交
-
-
由 Waleed Abdulla 提交于
-
- 16 3月, 2018 1 次提交
-
-
由 Waleed Abdulla 提交于
-
- 14 3月, 2018 1 次提交
-
-
由 Leo Han 提交于
-
- 13 3月, 2018 2 次提交
-
-
-
由 Leo Han 提交于
-
- 12 2月, 2018 7 次提交
-
-
由 Waleed Abdulla 提交于
-
由 Waleed Abdulla 提交于
-
由 Waleed Abdulla 提交于
-
由 刘定坤 提交于
-
由 Dingkun.Liu 提交于
-
由 Waleed 提交于
Convert Detection Layer from Python to TensorFlow ops.
-
由 Waleed Abdulla 提交于
-
- 05 2月, 2018 1 次提交
-
-
由 German Novikov 提交于
-
- 29 1月, 2018 4 次提交
- 08 1月, 2018 1 次提交
-
-
由 Cory Pruce 提交于
-
- 06 1月, 2018 1 次提交
-
-
由 Cory Pruce 提交于
-
- 05 1月, 2018 3 次提交
-
-
由 Cory Pruce 提交于
shapes more-or-less match now. weights load. current error=InvalidArgumentError (see above for traceback): WhereOp: Unhandled input dimensions: 0
-
由 Shenoy 提交于
-
由 Cory Pruce 提交于
Remove py_func for cross-environment serializability. Convert DetectionLayer/refine_detections to tf
-