snapshot...

modules/deeplearning/icing_cnn.py

@@ -651,7 +651,6 @@ class IcingIntensityNN:
         best_test_f1 = 0
         best_test_mcc = 0
 
-
         if EARLY_STOP:
             es = EarlyStop()
 
@@ -801,9 +800,7 @@ class IcingIntensityNN:
 
         self.test_preds = preds
 
-        self.h5f_tst.close()
-
-    def do_evaluate(self, ckpt_dir, ll, cc, prob_thresh=0.5):
+    def do_evaluate(self, ckpt_dir, prob_thresh=0.5):
 
         ckpt = tf.train.Checkpoint(step=tf.Variable(1), model=self.model)
         ckpt_manager = tf.train.CheckpointManager(ckpt, ckpt_dir, max_to_keep=3)
@@ -819,32 +816,14 @@ class IcingIntensityNN:
                 pred_s.append(pred)
 
         preds = np.concatenate(pred_s)
+        preds = preds[:,0]
+        self.test_probs = preds
 
         if NumClasses == 2:
             preds = np.where(preds > prob_thresh, 1, 0)
         else:
             preds = np.argmax(preds, axis=1)
-        print(preds.shape[0], np.sum(preds == 1))
-        preds = preds[:,0]
-
-        cc = np.array(cc)
-        ll = np.array(ll)
-        ice_mask = preds == 1
-        print(cc.shape, ll.shape, ice_mask.shape)
-        ice_cc = cc[ice_mask]
-        ice_ll = ll[ice_mask]
-
-        nav = GEOSNavigation(sub_lon=-75.0, CFAC=5.6E-05, COFF=-0.101332, LFAC=-5.6E-05, LOFF=0.128212, num_elems=2500,
-                             num_lines=1500)
-
-        ice_lons = []
-        ice_lats = []
-        for k in range(ice_cc.shape[0]):
-            lon, lat = nav.lc_to_earth(ice_cc[k], ice_ll[k])
-            ice_lons.append(lon)
-            ice_lats.append(lat)
-
-        return ice_lons, ice_lats
+        self.test_preds = preds
 
     def run(self, filename_trn, filename_tst):
         with tf.device('/device:GPU:'+str(self.gpu_device)):
@@ -860,6 +839,7 @@ class IcingIntensityNN:
         self.build_training()
         self.build_evaluation()
         self.restore(ckpt_dir)
+        self.h5f_tst.close()
 
     def run_evaluate(self, filename, ckpt_dir):
         data_dct, ll, cc = make_for_full_domain_predict(filename, name_list=train_params)
@@ -867,8 +847,7 @@ class IcingIntensityNN:
         self.build_model()
         self.build_training()
         self.build_evaluation()
-        ice_lons, ice_lats = self.do_evaluate(ckpt_dir, ll, cc)
-        return filename, ice_lons, ice_lats
+        self.do_evaluate(ckpt_dir)
 
 
 def run_restore_static(filename_tst, ckpt_dir_s_path):
@@ -890,8 +869,51 @@ def run_restore_static(filename_tst, ckpt_dir_s_path):
     return cm_avg
 
 
-def run_evaluate_static(filename, ckpt_dir_s):
-    nn = IcingIntensityNN()
+def run_evaluate_static(filename, ckpt_dir_s_path, prob_thresh=0.5):
+    data_dct, ll, cc = make_for_full_domain_predict(filename, name_list=train_params)
+    ckpt_dir_s = os.listdir(ckpt_dir_s_path)
+    prob_s = []
+    for ckpt in ckpt_dir_s:
+        ckpt_dir = ckpt_dir_s_path + ckpt
+        if not os.path.isdir(ckpt_dir):
+            continue
+        nn = IcingIntensityNN()
+        nn.setup_eval_pipeline(data_dct, len(ll))
+        nn.build_model()
+        nn.build_training()
+        nn.build_evaluation()
+        nn.do_evaluate(ckpt_dir, ll, cc)
+        prob_s.append(nn.test_probs)
+    num = len(prob_s)
+    prob_avg = prob_s[0]
+    for k in range(num-1):
+        prob_avg += prob_s[k+1]
+    prob_avg /= num
+    probs = prob_avg
+
+    if NumClasses == 2:
+        preds = np.where(probs > prob_thresh, 1, 0)
+    else:
+        preds = np.argmax(probs, axis=1)
+
+    cc = np.array(cc)
+    ll = np.array(ll)
+    ice_mask = preds == 1
+    print(cc.shape, ll.shape, ice_mask.shape)
+    ice_cc = cc[ice_mask]
+    ice_ll = ll[ice_mask]
+
+    nav = GEOSNavigation(sub_lon=-75.0, CFAC=5.6E-05, COFF=-0.101332, LFAC=-5.6E-05, LOFF=0.128212, num_elems=2500,
+                         num_lines=1500)
+
+    ice_lons = []
+    ice_lats = []
+    for k in range(ice_cc.shape[0]):
+        lon, lat = nav.lc_to_earth(ice_cc[k], ice_ll[k])
+        ice_lons.append(lon)
+        ice_lats.append(lat)
+
+    return filename, ice_lons, ice_lats
 
 
 if __name__ == "__main__":