Merge pull request #108 from PaddlePaddle/fix_squad

Tiny fixes in squad

BERT/run_squad.py

@@ -241,7 +241,7 @@ def train(args):
             data_path=args.train_file,
             batch_size=args.batch_size,
             phase='train',
-            shuffle=False,
+            shuffle=True,
             dev_count=dev_count,
             version_2_with_negative=args.version_2_with_negative,
             epoch=args.epoch)
@@ -396,7 +396,7 @@ def train(args):
                     total_cost, total_num_seqs = [], []
                     time_begin = time.time()
 
-                if steps % args.save_steps == 0:
+                if steps % args.save_steps == 0 or steps == max_train_steps:
                     save_path = os.path.join(args.checkpoints,
                                              "step_" + str(steps))
                     fluid.io.save_persistables(exe, save_path, train_program)