AndrewZhaoLuo commented on code in PR #14008:
URL: https://github.com/apache/tvm/pull/14008#discussion_r1113703994
##########
python/tvm/topi/cuda/scatter_elements.py:
##########
@@ -20,6 +20,188 @@
from tvm import te, tir
from ..utils import ceil_div, get_const_int
from ..math import cast
+from .nms import atomic_add
+
+
+def gen_scatter_add_1d_atomic(data, indices, updates, out, axis, _):
Review Comment:
Ah this is something I think I might have missed. Within indices along a
particular axis, do we expect all indices to be unique? Otherwise I believe
something similar must be done with all operators.
##########
python/tvm/relay/op/_transform.py:
##########
@@ -687,7 +678,6 @@ def argwhere_shape_func(attrs, inputs, out_ndims):
_reg.register_shape_func("scatter", False, elemwise_shape_func)
Review Comment:
Do you plan to deprecate `relay.scatter` too?
--
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]
