upload all for tf2.0

Author: SAAS R7 User1

Diff for: .gitignore

@@ -0,0 +1,3 @@
+scripts/*
+history/*
+data/*

Diff for: LICENSE

@@ -0,0 +1,50 @@
+# Explainable-Neural-Networks
+
+## Installation 
+```shell
+pip install git+https://github.com/ajzhanghku/XNN.git
+# with token
+pip install git+https://[email protected]/ajzhanghku/XNN.git
+```
+
+
+## Usage
+- xNN
+
+```python
+from xnn.xNN import xNN
+from xnn.SOSxNN import SOSxNN
+from xnn.visualizer import visualize_xnn, visualize_sosxnn
+
+model = xNN(input_num = train_x.shape[1], input_dummy_num = 0, subnet_num = 10, subnet_layers = [10, 6], \
+        task = "Regression", activation = tf.tanh, batch_size = min(1000, int(np.floor(train_x.shape[0]*0.20))), \
+        training_epochs = 10000, beta_threshold = 0.05, tune_epochs = 500, lr_BP = 0.001, \
+        l1_proj = 0.001, l1_subnet = 0.01, verbose = False, \
+        val_ratio = 0.2, early_stop_thres = 2500)
+pred_train, tr_x, tr_y, pred_val, val_x, val_y = model.fit(train_x, train_y); 
+pred_test = model.predict(test_x)
+visualize_xnn(simu_dir, "Demo_XNN", model, tr_x, dummy_name = None)
+```
+
+
+- SOSxNN
+```python
+model = SOSxNN(input_num = train_x.shape[1], input_dummy_num = 0, subnet_num = 10, subnet_layers = [10, 6], \
+        task = "Regression", activation = tf.tanh, batch_size = min(1000, int(np.floor(train_x.shape[0]*0.20))), \
+        training_epochs = 10000, beta_threshold = 0.05, tune_epochs = 500, lr_BP = 0.001, lr_CL= 0.1, \
+        l1_proj = 0.001, l1_subnet = 0.01, smooth_lambda = 10**(-6), verbose = False, \
+        val_ratio = 0.2, early_stop_thres = 2500)
+pred_train, tr_x, tr_y, pred_val, val_x, val_y = model.fit(train_x, train_y); 
+pred_test = model.predict(test_x)
+visualize_sosxnn(simu_dir, "Demo_SOSxNN", model, tr_x, dummy_name = None)
+```
+
+References
+----------
+J. Vaughan, A. Sudjianto, E. Brahimi, J. Chen, and V. N. Nair, "Explainable
+neural networks based on additive index models," The RMA
+Journal, pp. 40-49, October 2018.
+
+Yang, Zebin, Aijun Zhang, and Agus Sudjianto. "Enhancing Explainability of
+Neural Networks through Architecture Constraints." 
+arXiv preprint arXiv:1901.03838 (2019).