lygztq commented on a change in pull request #8376:
URL: https://github.com/apache/tvm/pull/8376#discussion_r661943104
##########
File path: tests/python/relay/test_op_level10.py
##########
@@ -380,10 +382,31 @@ def verify_dynamic_batch_matmul(x_shape, y_shape,
out_shape, dtype="float32"):
# TODO(mbrookhart): enable once VM supports heterogenous execution
# @tvm.testing.uses_gpu
Review comment:
I'm not sure, I think [this
line](https://github.com/apache/tvm/blob/84359a97bdb896ddd2a74c65631a00da1ec5ec70/python/tvm/topi/nn/batch_matmul.py#L65-L65)
(which is called by topi.cuda
[here](https://github.com/apache/tvm/blob/90dce48de6ca74a439ed24fc3a3d80d57b305dd0/python/tvm/topi/cuda/batch_matmul.py#L32-L32))
will fail if `x.shape[2]` is static and `y.shape[2]` is dynamic.
By the way, the unit test in ci reported a strange error. I’m not sure if
the error is related to this PR.
--
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: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
