upgrade save and load interface (#4311)

* upgrade dcn and xdeepfm, change from all old save/load api to fluid.save and fluid.load
* test=develop

PaddleRec/ctr/dcn/infer.py

@@ -45,7 +45,7 @@ def infer():
     startup_program = fluid.framework.Program()
     test_program = fluid.framework.Program()
     cur_model_path = os.path.join(args.model_output_dir,
-                                  'epoch_' + args.test_epoch)
+                                  'epoch_' + args.test_epoch, "checkpoint")
 
     with fluid.scope_guard(inference_scope):
         with fluid.framework.program_guard(test_program, startup_program):
@@ -62,10 +62,9 @@ def infer():
             exe = fluid.Executor(place)
             feeder = fluid.DataFeeder(
                 feed_list=dcn_model.data_list, place=place)
-            fluid.io.load_persistables(
-                executor=exe,
-                dirname=cur_model_path,
-                main_program=fluid.default_main_program())
+
+            exe.run(startup_program)
+            fluid.io.load(fluid.default_main_program(), cur_model_path)
 
             for var in dcn_model.auc_states:  # reset auc states
                 set_zero(var.name, scope=inference_scope, place=place)

PaddleRec/ctr/dcn/local_train.py

@@ -80,13 +80,10 @@ def train(args):
             debug=False,
             print_period=args.print_steps)
         model_dir = os.path.join(args.model_output_dir,
-                                 'epoch_' + str(epoch_id + 1))
+                                 'epoch_' + str(epoch_id + 1), "checkpoint")
         sys.stderr.write('epoch%d is finished and takes %f s\n' % (
             (epoch_id + 1), time.time() - start))
-        fluid.io.save_persistables(
-            executor=exe,
-            dirname=model_dir,
-            main_program=fluid.default_main_program())
+        fluid.save(fluid.default_main_program(), model_dir)
 
 
 if __name__ == '__main__':

PaddleRec/ctr/xdeepfm/infer.py

@@ -36,7 +36,7 @@ def infer():
     startup_program = fluid.framework.Program()
     test_program = fluid.framework.Program()
     cur_model_path = os.path.join(args.model_output_dir,
-                                  'epoch_' + args.test_epoch)
+                                  'epoch_' + args.test_epoch, "checkpoint")
 
     with fluid.scope_guard(inference_scope):
         with fluid.framework.program_guard(test_program, startup_program):
@@ -48,10 +48,9 @@ def infer():
 
             exe = fluid.Executor(place)
             feeder = fluid.DataFeeder(feed_list=data_list, place=place)
-            fluid.io.load_persistables(
-                executor=exe,
-                dirname=cur_model_path,
-                main_program=fluid.default_main_program())
+
+            exe.run(startup_program)
+            fluid.io.load(fluid.default_main_program(), cur_model_path)
 
             for var in auc_states:  # reset auc states
                 set_zero(var.name, scope=inference_scope, place=place)

PaddleRec/ctr/xdeepfm/local_train.py

@@ -55,13 +55,10 @@ def train():
             debug=False,
             print_period=args.print_steps)
         model_dir = os.path.join(args.model_output_dir,
-                                 'epoch_' + str(epoch_id + 1))
+                                 'epoch_' + str(epoch_id + 1), "checkpoint")
         sys.stderr.write('epoch%d is finished and takes %f s\n' % (
             (epoch_id + 1), time.time() - start))
-        fluid.io.save_persistables(
-            executor=exe,
-            dirname=model_dir,
-            main_program=fluid.default_main_program())
+        fluid.io.save_persistables(fluid.default_main_program(), model_dir)
 
 
 if __name__ == '__main__':