ekalda commented on PR #17048:
URL: https://github.com/apache/tvm/pull/17048#issuecomment-2150432744
Thanks @Anndrey24 and @lhutton1!
--
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
ekalda merged PR #17048:
URL: https://github.com/apache/tvm/pull/17048
--
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.or
lhutton1 commented on code in PR #17048:
URL: https://github.com/apache/tvm/pull/17048#discussion_r1627646069
##
python/tvm/topi/arm_cpu/conv2d.py:
##
@@ -680,6 +681,43 @@ def compute_conv2d_NHWC_hybrid_SME(cfg, data, kernel,
strides, padding, dilation
)
+@autotvm.regi
Anndrey24 commented on code in PR #17048:
URL: https://github.com/apache/tvm/pull/17048#discussion_r1627399170
##
python/tvm/topi/arm_cpu/conv2d.py:
##
@@ -698,6 +737,8 @@ def schedule_conv2d_NHWC_hybrid_TIR(sch: tvm.tir.Schedule):
"A_padded_K",
"A_padded_M",
Anndrey24 commented on code in PR #17048:
URL: https://github.com/apache/tvm/pull/17048#discussion_r1627396909
##
python/tvm/topi/arm_cpu/conv2d.py:
##
@@ -680,6 +681,43 @@ def compute_conv2d_NHWC_hybrid_SME(cfg, data, kernel,
strides, padding, dilation
)
+@autotvm.reg
lhutton1 commented on code in PR #17048:
URL: https://github.com/apache/tvm/pull/17048#discussion_r1626149820
##
python/tvm/relay/op/strategy/arm_cpu.py:
##
@@ -583,11 +584,25 @@ def
conv2d_gemm_without_weight_transform_strategy_arm_cpu(attrs, inputs, out_typ
)
Anndrey24 opened a new pull request, #17048:
URL: https://github.com/apache/tvm/pull/17048
This commit extends the SME conv2d NHWC schedule to support convolutions
with float16 inputs (data and kernel) and a float32 output using the tensor
intrinsics added in #16981.
cc @ekalda @lhut
