提交 d40881cb 编写于 作者: Y Yu Yang

Loose atol in test_pe_test_wwhile_train

test=develop
上级 36e26a53
......@@ -88,7 +88,7 @@ class ParallelExecutorTestingDuringTraining(unittest.TestCase):
self.assertTrue(
np.allclose(
train_loss, test_loss, atol=1e-8),
train_loss, test_loss, atol=1e-2),
"Train loss: " + str(train_loss) + "\n Test loss:" +
str(test_loss))
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册