diff --git a/modules/deeplearning/srcnn_l1b_l2.py b/modules/deeplearning/srcnn_l1b_l2.py
index b3c0d9ed6de04f724d4aa058d08d5a46bcb4c072..eb65e815bcff46060d0dd12c84faa078b5165948 100644
--- a/modules/deeplearning/srcnn_l1b_l2.py
+++ b/modules/deeplearning/srcnn_l1b_l2.py
@@ -2,7 +2,7 @@ import glob
 import tensorflow as tf
 from util.setup import logdir, modeldir, cachepath, now, ancillary_path
 from util.util import EarlyStop, normalize, denormalize, resample, resample_2d_linear, resample_one,\
-    resample_2d_linear_one, get_grid_values_all, add_noise
+    resample_2d_linear_one, get_grid_values_all, add_noise, smooth_2d, smooth_2d_single
 import os, datetime
 import numpy as np
 import pickle
@@ -266,7 +266,10 @@ class SRCNN:
         # # tmp = resample_2d_linear(x_2, y_2, tmp, t, s)
         # data_norm.append(tmp)
         # --------
-        tmp = input_data[:, label_idx, slc_y_2, slc_x_2]
+        #tmp = input_data[:, label_idx, slc_y_2, slc_x_2]
+        tmp = input_data[:, label_idx, :, :]
+        tmp = smooth_2d(tmp, sigma=1.5)
+        tmp = tmp[:, slc_y_2, slc_x_2]
         if label_param != 'cloud_probability':
             tmp = normalize(tmp, label_param, mean_std_dct)
             if DO_ADD_NOISE:
@@ -285,7 +288,11 @@ class SRCNN:
         data = data.astype(np.float32)
         # -----------------------------------------------------
         # -----------------------------------------------------
-        label = input_data[:, label_idx, y_128, x_128]
+        #label = input_data[:, label_idx, y_128, x_128]
+        label = input_data[:, label_idx, :, :]
+        label = smooth_2d(label, sigma=1.5)
+        label = label[:, y_128, x_128]
+
         if label_param != 'cloud_probability':
             label = normalize(label, label_param, mean_std_dct)
         else:
@@ -755,7 +762,7 @@ def run_evaluate_static(in_file, out_file, ckpt_dir):
     # grd_b = normalize(grd_b, 'refl_0_65um_nom', mean_std_dct)
 
     grd_c = get_grid_values_all(h5f, label_param)
-    # grd_c = gaussian_filter(grd_c, sigma=1.0)
+    grd_c = gaussian_filter(grd_c, sigma=1.5)
     grd_c = grd_c[y_0:y_0+sub_y, x_0:x_0+sub_x]
     hr_grd_c = grd_c.copy()
     hr_grd_c = hr_grd_c[y_128, x_128]