add method to scale data

modules/deeplearning/unet_l1b_l2.py

@@ -60,7 +60,8 @@ l2_params = ['cloud_fraction', 'cld_temp_acha', 'cld_press_acha']
 zero_out_params = ['cld_reff_dcomp', 'cld_opd_dcomp', 'iwc_dcomp', 'lwc_dcomp']
 DO_ZERO_OUT = False
 
-label_param = l2_params[1]
+label_idx = 1
+label_param = l2_params[label_idx]
 
 
 def build_conv2d_block(conv, num_filters, activation, block_name, padding='SAME'):
@@ -231,7 +232,7 @@ class UNET:
 
             f = self.test_label_files
             nda = np.load(f)
-            label = nda[idxs, 0, :, :]
+            label = nda[idxs, label_idx, :, :]
             label = np.expand_dims(label, axis=3)
 
         data = data.astype(np.float32)