Skip to content

Add HiFi-optimized Softmax kernels for Xtensa - #3729

Open
unmeshna017 wants to merge 2 commits into
tensorflow:mainfrom
unmeshna017:xtensa-softmax-hifi
Open

unmeshna017 wants to merge 2 commits into
tensorflow:mainfrom
unmeshna017:xtensa-softmax-hifi

Conversation

@unmeshna017

Copy link
Copy Markdown
Contributor

Adds NNLib-optimized int8, int16 and float32 Softmax variants using the LUT-based softmax APIs.

Depends on the NNLib 09_03_2026 release (PR#3703).

@cad-audio @joshih-cad

Adds HiFi5 NNLib-optimized int8, int16 and float32 Softmax variants using the LUT-based softmax APIs. Guards the SOFTMAX_INT8/INT16 registrations under XTENSA and adds Register_SOFTMAX_FLOAT32(). Depends on the NNLib 09_03_2026 release (PR#3703).
@unmeshna017
unmeshna017 requested a review from a team as a code owner September 17, 2026 07:24
Add HIFI_IQ to the include guard so xa_nnlib_api.h and xa_nnlib_standards.h are pulled in for the hifi_iq build.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant