disable_old_mem_opt_in_transformer_model, test=develop (#2767)

PaddleNLP/neural_machine_translation/transformer/train.py

@@ -4,6 +4,10 @@ import copy
 import logging
 import multiprocessing
 import os
+
+if os.environ.get('FLAGS_eager_delete_tensor_gb', None) is None:
+    os.environ['FLAGS_eager_delete_tensor_gb'] = '0'
+
 import six
 import sys
 sys.path.append("../../")
@@ -720,9 +724,6 @@ def train(args):
                 optimizer = fluid.optimizer.SGD(0.003)
             optimizer.minimize(avg_cost)
 
-    if args.use_mem_opt:
-        fluid.memory_optimize(train_prog)
-
     if args.local:
         logging.info("local start_up:")
         train_loop(exe, train_prog, startup_prog, dev_count, sum_cost, avg_cost,
@@ -806,4 +807,4 @@ if __name__ == "__main__":
     logging.getLogger().setLevel(logging.INFO)
 
     args = parse_args()
-    train(args)
\ No newline at end of file
+    train(args)