Tf-keras-vis

Major Feature
---

* Add new methods (ScoreCAM and Faster-ScoreCAM)


Breaking Changes
---
* `tf_keras_vis.utils.print_gpus()` was deleted.
* `tf_keras_vis.utils.losses.SmoothedLoss` was deleted.

BugFix and Other Changes
---
* Fixes a problem that ActivationMaximization couldn't calculate and normalize losses correctly when multiple samples.
* Fixes a problem that the output of ActivationMaximization was restricted when input_ranges contain None.
* Improve to raise ValueError when input shape doesn't have channels dimension in GradCAM.
* Fixes a problem that GradCAM++ couldn't calculate scores correctly when multiple samples.
* Add `tf_keras_vis.utils. num_of_gpus()` instead of `tf_keras_vis.utils. print_gpus()`.

Major Feature
---

* Add GradCAM++ algorism implemented as GradcamPlusPlus class

Bugfixes
--

* The noise of SmoothGrad is too strong.
* Rotation in InputModifier across channles-dim.

Improvements
---

* Refactoring and Vectorizing some processing in `Gradcam` and `Saliency`
* Remove `.travis.yml` because there is Github Action as CI tools

This release includes the patch for improving Issue 15 that is **serious bugs**. So We highly recommend to update tf-keras-vis.

BugFix and Other Changes
---

* Fixes bug that tf_keras_vis.utils.normalize() can't batchwisely do normalization.
* Add `expand_cam` argument to GradCAM__call__() for returning non-interpolated cam values.
* In setup.py, rename `development` of extras_require to `develop`.

We re-uploaded tf-keras-vis to PyPI as v0.3.1 because wrong library was uploaded.

Bug Fixes
---

* Fixes a lot of bugs in ActivationMaximization and Gradcam

Breaking Changes
---

* `tf_keras_vis.activation_maximization.ActivationMaximization.__call__()`'s `regularizers` argument no longer accepts `dict` object.
* Rename `tf_keras_vis.gradcam.Gradcam.__call__()`'s `seek_penultimate_layer` argument to `seek_penultimate_conv_layer `.
* Rename `tf_keras_vis.utils.listify`'s `empty_list_if_none` argument to `return_empty_list_if_none`.