Replies: 2 comments 1 reply
-
从原理上看
|
Beta Was this translation helpful? Give feedback.
1 reply
-
@qelk123 请问您是要省略tensor expression,直接使用TensorIR function吗? |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
在第六讲中,借助BlockBuilder构造Tensor IR,按我的理解是通过写call_function_map 手工指定torch算子与TE ComputeOP之间的关系进行转换,再通过create_prim_func实现到Tensor IR 的lowering的。请问,TE 这一层次的Compute声明是否可以省略呢?特别是对于某些用TE这套DSL难以描述的算子或者是我已给出的高效Tensor IR实现的情形,是否能够绕过TE这一层呢?
Beta Was this translation helpful? Give feedback.
All reactions