[tvm] branch main updated: [Relay, Op] Add conv2d generic layout op strategy when meta schedule is enabled (#12104)

[ma...@apache.org]

This is an automated email from the ASF dual-hosted git repository.

masahi pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/tvm.git


The following commit(s) were added to refs/heads/main by this push:
     new e0a0e67e55 [Relay, Op] Add conv2d generic layout op strategy when meta schedule is enabled (#12104)
e0a0e67e55 is described below

commit e0a0e67e55c036077fd07728ac28ecd7d2c9d72b
Author: Wuwei Lin <wu...@apache.org>
AuthorDate: Fri Jul 15 00:39:56 2022 -0700

    [Relay, Op] Add conv2d generic layout op strategy when meta schedule is enabled (#12104)
---
 python/tvm/relay/op/strategy/cuda.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/python/tvm/relay/op/strategy/cuda.py b/python/tvm/relay/op/strategy/cuda.py
index e3c74e15c2..9bedfe8cb0 100644
--- a/python/tvm/relay/op/strategy/cuda.py
+++ b/python/tvm/relay/op/strategy/cuda.py
@@ -320,7 +320,7 @@ def conv2d_strategy_cuda(attrs, inputs, out_type, target):
                 wrap_topi_schedule(topi.cuda.schedule_conv2d_NCHWc_int8),
                 name="conv2d_NCHWc_int8.cuda",
             )
-        elif is_auto_scheduler_enabled():
+        elif is_auto_scheduler_enabled() or is_meta_schedule_enabled():
             strategy.add_implementation(
                 wrap_compute_conv2d(
                     topi.nn.conv, need_data_layout=True, need_kernel_layout=True, has_groups=True