未验证 提交 1681a2dd 编写于 作者: W WangXi 提交者: GitHub

update fp16 gray_list for tensor parallel (#33660)

上级 f91dfe15
...@@ -146,6 +146,8 @@ gray_list = { ...@@ -146,6 +146,8 @@ gray_list = {
'cast', 'cast',
'fused_bn_add_activation', 'fused_bn_add_activation',
'c_identity', 'c_identity',
'c_concat',
'c_allreduce_sum',
} }
# The set of ops that don't support fp16 calculation # The set of ops that don't support fp16 calculation
......
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