SIGN IN SIGN UP
apache / mxnet UNCLAIMED

Lightweight, Portable, Flexible Distributed/Mobile Deep Learning with Dynamic, Mutation-aware Dataflow Dep Scheduler; for Python, R, Julia, Scala, Go, Javascript and more

0 0 25 C++

[FEATURE] Add _npi_power_scalar and _npi_multiply_scalar fuse (#20976)

* [FEATURE] Add _npi_power_scalar and _npi_multiply_scalar fuse

* Merge _npi_power_scalar implementation with implementation of this fuse

* Fix clang

* Fix CI

* Fix review and simplify the implementation

* Add checks for the amount of inputs and outputs

* Fix CI

* Add Reset() function

* Fix DNNLPowMulScalarShape and Type functions

* Fix DNNLPowMulScalarType

* Fix DNNLPowMulScalarType

* Add generic implementation for sq_pow_mul_scalar operator

* Fix sanity

* Fix req

* Add Filter method to property

* Add new line

* Fix gpu CI

* Add '_sg_pow_mul_scalar' to symbol_fp16.py

* Fix CI on MacOS

* Fix SupportDNNL*

* Make PowMulScalarCompute more readable

* Fix PowMulScalarCompute

* Fix memory usage

* Fix build
B
bartekkuncer committed
5abdc77f3c191bc771a5e97a9cecf30fd832de96
Parent: 9745d36
Committed by GitHub <noreply@github.com> on 7/5/2022, 7:39:50 AM