zxybazh commented on code in PR #14324:
URL: https://github.com/apache/tvm/pull/14324#discussion_r1140580956
##########
tests/python/relax/test_op_index.py:
##########
@@ -376,22 +409,22 @@ def test_strided_slice_infer_struct_info_shape_symbolic():
_check_inference(
bb,
relax.op.strided_slice(x0, axes=[0], begin=[1], end=[3]),
- relax.TensorStructInfo((2, n), "float32"),
+ relax.TensorStructInfo((tir.min(3, m) - tir.min(1, m) + 1 - 1, n),
"float32"),
Review Comment:
Thanks for the comments. In this case, I think even when `m < 3` it's still
a valid case there's no need to raise an error. And this seems to be a simple
case where the input shape has a simple `m` variable. If the input is another
symbolic shape we may not be able to point out some invalid range like `m < 3`.
On the other hand, indeed in most cases the result is the same as static
version as `(2, n)` so I'm wondering if this could cause any side effect if we
keep the symbolic shape generated here.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscr...@tvm.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
