snapshot...

modules/deeplearning/cloud_fraction_fcn_abi.py

@@ -73,7 +73,7 @@ print('data_params_full: ', data_params_full)
 print('label_param: ', label_param)
 
 KERNEL_SIZE = 3
-X_LEN = Y_LEN = 64
+X_LEN = Y_LEN = 128
 
 if KERNEL_SIZE == 3:
     slc_x = slice(0, int(X_LEN/4) + 2)
@@ -911,13 +911,14 @@ def run_restore_static(directory, ckpt_dir, out_file=None):
     nn = SRCNN()
     labels, preds, inputs = nn.run_restore(directory, ckpt_dir)
     if out_file is not None:
+        y_hi, x_hi = (Y_LEN // 4) + 1, (X_LEN // 4) + 1
         np.save(out_file,
                 [np.squeeze(labels), preds.argmax(axis=3),
-                 denormalize(inputs[:, 1:17, 1:17, 0], 'temp_11_0um_nom', mean_std_dct),
-                 denormalize(inputs[:, 1:17, 1:17, 1], 'refl_0_65um_nom', mean_std_dct),
-                 denormalize(inputs[:, 1:17, 1:17, 2], 'refl_0_65um_nom', mean_std_dct),
-                 inputs[:, 1:17, 1:17, 3],
-                 inputs[:, 1:17, 1:17, 4]])
+                 denormalize(inputs[:, 1:y_hi, 1:x_hi, 0], 'temp_11_0um_nom', mean_std_dct),
+                 denormalize(inputs[:, 1:y_hi, 1:x_hi, 1], 'refl_0_65um_nom', mean_std_dct),
+                 denormalize(inputs[:, 1:y_hi, 1:x_hi, 2], 'refl_0_65um_nom', mean_std_dct),
+                 inputs[:, 1:y_hi, 1:x_hi, 3],
+                 inputs[:, 1:y_hi, 1:x_hi, 4]])
 
 
 def run_evaluate_static(in_file, out_file, ckpt_dir):