audio: tensorflow: keyword detect - #11092

Draft

singalsu wants to merge 17 commits into

thesofproject:mainfrom

singalsu:tflm_mfcc_development

Commits on Sep 4, 2026

[DNM] Audio: Buffers: Add support for DP-to-DP component binding
singalsu
committed
e653abb
[DNM] audio: stft_process: switch assert include to rtos/panic.h
singalsu
committed
c7f40a1
audio: tensorflow: static build support for HiFi3

lrgirdwo
authored and
singalsu
committed
ed33ff5
audio: tensorflow: add model for banana, strawberry, orange keywords
singalsu
committed
59fa89d
audio: tflm: add wake-word inference component with KPB trigger

lrgirdwo
authored and
singalsu
committed
c681808
audio: mfcc: tune: emit mel40 and mel40_compress configs
singalsu
committed
9d78fee
audio: mfcc: raise max_frames to a full hop in DP mode
singalsu
committed
79f9c0a
topology2: add TFLM Wake-on-Voice branches for HDA and SoundWire

lrgirdwo
authored and
singalsu
committed
511b171
platform: cavs25: enable TFLM Wake-on-Voice on cAVS 2.5

lrgirdwo
authored and
singalsu
committed
b8869d2
docs: tensorflow: add TFLM Wake-on-Voice architecture README

lrgirdwo
authored and
singalsu
committed
c8246e5
audio: tensorflow: tune: TFLM wake-word training pipeline
singalsu
committed
4da0ac6
audio: tensorflow: tune: add single-speaker Piper keyword generator
singalsu
committed
01c6dd7
audio: tflm: soft mel-log AGC for level-robust wake-word detection
singalsu
committed
85381c8
topology2: host-copier: add D0i3 compatibility attributes
singalsu
committed
1aee10e
audio: tensorflow: support loading model from bytes control
singalsu
committed
4f35a49
audio: tensorflow: tune: export model blobs for topology and sof-ctl
singalsu
committed
f0fd1b1
[DNM] platform: cavs25: enable TFLM model loading from bytes control
singalsu
committed
e9305d0

Navigation Menu

Uh oh!

audio: tensorflow: keyword detect - #11092

audio: tensorflow: keyword detect#11092
singalsu wants to merge 17 commits into
thesofproject:mainfrom
singalsu:tflm_mfcc_development

Commits on Sep 4, 2026

[DNM] Audio: Buffers: Add support for DP-to-DP component binding

[DNM] audio: stft_process: switch assert include to rtos/panic.h

audio: tensorflow: static build support for HiFi3

audio: tensorflow: add model for banana, strawberry, orange keywords

audio: tflm: add wake-word inference component with KPB trigger

audio: mfcc: tune: emit mel40 and mel40_compress configs

audio: mfcc: raise max_frames to a full hop in DP mode

topology2: add TFLM Wake-on-Voice branches for HDA and SoundWire

platform: cavs25: enable TFLM Wake-on-Voice on cAVS 2.5

docs: tensorflow: add TFLM Wake-on-Voice architecture README

audio: tensorflow: tune: TFLM wake-word training pipeline

audio: tensorflow: tune: add single-speaker Piper keyword generator

audio: tflm: soft mel-log AGC for level-robust wake-word detection

topology2: host-copier: add D0i3 compatibility attributes

audio: tensorflow: support loading model from bytes control

audio: tensorflow: tune: export model blobs for topology and sof-ctl

[DNM] platform: cavs25: enable TFLM model loading from bytes control