upgrade ade dgu load api (#4659)

* fix_ade_readme

* fix_load_api

PaddleNLP/dialogue_system/auto_dialogue_evaluation/inference_model.py

@@ -71,9 +71,9 @@ def do_save_inference_model(args):
     assert (args.init_from_params) or (args.init_from_pretrain_model)
 
     if args.init_from_params:
-        fluid.load(test_prog, args.init_from_params)
+        fluid.load(test_prog, args.init_from_params, executor=exe)
     elif args.init_from_pretrain_model:
-        fluid.load(test_prog, args.init_from_pretrain_model)
+        fluid.load(test_prog, args.init_from_pretrain_model, executor=exe)
 
     # saving inference model
     fluid.io.save_inference_model(

PaddleNLP/dialogue_system/dialogue_general_understanding/inference_model.py

@@ -99,7 +99,7 @@ def do_save_inference_model(args):
     assert (args.init_from_params)
 
     if args.init_from_params:
-        fluid.load(test_prog, args.init_from_params)
+        fluid.load(test_prog, args.init_from_params, executor=exe)
 
     # saving inference model
     fluid.io.save_inference_model(

PaddleNLP/dialogue_system/dialogue_general_understanding/predict.py

@@ -109,7 +109,7 @@ def do_predict(args):
     assert (args.init_from_params)
 
     if args.init_from_params:
-        fluid.load(test_prog, args.init_from_params)
+        fluid.load(test_prog, args.init_from_params, executor=exe)
 
     compiled_test_prog = fluid.CompiledProgram(test_prog)