fix the ips of Bert in the static graph (PaddlePaddle#1518)

* fix_ips and data sampler * fix train_batch_sampler * fix train_batch_sampler * fix batch_sampler * fix bug Co-authored-by: Zeyu Chen <[email protected]>
liu-nlper · Dec 30, 2021 · e1842d7 · e1842d7
1 parent a824222
commit e1842d7
Show file tree

Hide file tree

Showing 2 changed files with 3 additions and 3 deletions.
diff --git a/examples/language_model/bert/static/dataset.py b/examples/language_model/bert/static/dataset.py
@@ -28,7 +28,7 @@ def create_pretraining_dataset(input_file,
                                places=None):
     train_data = PretrainingDataset(
         input_file=input_file, max_pred_length=max_pred_length)
-    train_batch_sampler = paddle.io.DistributedBatchSampler(
+    train_batch_sampler = paddle.io.BatchSampler(
         train_data, batch_size=args.batch_size, shuffle=True)
 
     def _collate_data(data, stack_fn=Stack()):

diff --git a/examples/language_model/bert/static/run_pretrain.py b/examples/language_model/bert/static/run_pretrain.py
@@ -392,8 +392,8 @@ def do_train(args):
                         % (global_step, epoch, step, loss_return[0],
                            reader_cost_avg.get_average(),
                            train_cost_avg.get_average(),
-                           total_samples / args.logging_steps, total_samples /
-                           (args.logging_steps * train_cost_avg.get_average())))
+                           total_samples / args.logging_steps, args.batch_size /
+                           (reader_cost_avg.get_average() + train_cost_avg.get_average())))
                     total_samples = 0
                     train_cost_avg.reset()
                     reader_cost_avg.reset()