Closed LeiWang1999 closed 1 month ago
I think the best solution is to implement a new tl prim_func pass, such as tl.Simplify
, similar to tir.Specialize
found in 3rdparty/tvm/src/tir/ir/specialize.cc.”
Thanks. PR #216 has been created to implement tl.transform.Simplify
pass and add test case for it.
I think the best solution is to implement a new tl prim_func pass, such as
tl.Simplify
, similar totir.Specialize
found in 3rdparty/tvm/src/tir/ir/specialize.cc.”
Resloved.
Ref to PR #214.
The scheduling template in TL is quite complex because, for different tensor arguments, we need to implement separate TL templates. This complexity arises because TL cannot automatically remove unused tensor arguments, and the tensor arguments cannot be extended as flexibly as needed.
Expected Behavior:
If
with_scale
is set toFalse
, theScale
argument should be removed from the function's argument list.