Commit 79548341 authored by A. Unique TensorFlower's avatar A. Unique TensorFlower Committed by TensorFlower Gardener
Browse files

Improve quantized (uint8/int16) subtraction support. 

Subtraction always took the slow path on uint8, changed to
support the non-4D activation path for non-broadcast.

Added support for int16 subtraction including broadcast and activation fns.

PiperOrigin-RevId: 231328853
parent a39f31a9
Show whitespace changes
Inline Side-by-side
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Please to comment