lhutton1 commented on code in PR #13890:
URL: https://github.com/apache/tvm/pull/13890#discussion_r1097287518
##########
python/tvm/relay/backend/contrib/ethosu/op/binary_elementwise.py:
##########
@@ -193,6 +202,12 @@ def ethosu_binary_elementwise(
The layout of the Input Feature Map tensor 2. Can be "NHWC" or
"NHCWB16".
ofm_layout : str, optional
The layout of the Output Feature Map tensor. Can be "NHWC" or
"NHCWB16".
+ use_rescale : bool, optional
+ True if use explicit scaling.
Review Comment:
Nit: should this be "Use explicit scaling if True."?
##########
python/tvm/relay/backend/contrib/ethosu/op/binary_elementwise.py:
##########
@@ -117,6 +123,9 @@ def ethosu_binary_elementwise(
ifm_layout: Optional[str] = "NHWC",
ifm2_layout: Optional[str] = "NHWC",
ofm_layout: Optional[str] = "NHWC",
+ use_rescale: bool = False,
Review Comment:
Nit: should there be `Optional[...]` type hints?
##########
python/tvm/relay/backend/contrib/ethosu/te/binary_elementwise.py:
##########
@@ -121,6 +124,12 @@ def binary_elementwise_compute(
{int32}->{int8, uint8, int32}, any pairing"
SHL:
{int32}->{int32} only
+ use_rescale : bool
+ True if use explicit scaling.
Review Comment:
same as above
--
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]
