Xplainer

Major Features and Improvements
- General settings added for all models: number of images to explain, probability threshold.
- Method specific settings added for Integrated Gradients and LIME.

Initial release of xplainer.ai toolkit. Supported: Image recognition, TensorFlow 2+.

Features:
- Model: Simple overview (input, parameters, is flat), summary, plot.
- Explain: Gradient based methods (GradCam, GradientsInput, IntegratedGradients, SmoothGrad, Vanilla Gradient)
and Local (LIME). Explanation of the bet prediction.