feat(nemo-speech-cpp): add the NVIDIA NeMo-Speech.cpp backend - #11406

Draft

localai-bot wants to merge 30 commits into

masterfrom

feat/nemo-speech-cpp-backend

Commits on Aug 6, 2026

feat(nemo-speech-cpp): scaffold the backend and upstream build
mudler
committed
2c5f5bf
fix(nemo-speech-cpp): make 'build' produce the package and bundle the ITN stack
mudler
committed
1c82d44
fix(nemo-speech-cpp): make the closure guard fail closed and give CI its toolchain
mudler
committed
5e91f6b
fix(nemo-speech-cpp): pin protoc for ITN and make the norm stack its own target
mudler
committed
fd8df77
fix(nemo-speech-cpp): move the backend apt gate below the expensive layers
mudler
committed
75162ec
feat(nemo-speech-cpp): parse model options
mudler
committed
472d144
feat(nemo-speech-cpp): guard the empty option value and warn on a bad gpu index
mudler
committed
9d0ac11
feat(nemo-speech-cpp): detect model family and discover TTS assets
mudler
committed
2127a26
feat(nemo-speech-cpp): bind the C ABI with layout assertions
mudler
committed
da51e95
test(nemo-speech-cpp): run the ABI specs in CI and refuse to skip them
mudler
committed
272a219
feat(nemo-speech-cpp): select the family at load and gate RPCs on it
mudler
committed
14bdd6c
fix(nemo-speech-cpp): pin the load ordering, close the engineMu race
mudler
committed
139c57a
feat(nemo-speech-cpp): implement offline transcription
mudler
committed
cf79798
feat(nemo-speech-cpp): implement streaming and live transcription
mudler
committed
4d29504
fix(nemo-speech-cpp): make live deltas concatenate and fill segment words
mudler
committed
55070db
feat(nemo-speech-cpp): implement standalone diarization
mudler
committed
4e553d4
fix(nemo-speech-cpp): pin the diarizer geometry sentinels and cap the segment buffer
mudler
committed
48c9e63
feat(nemo-speech-cpp): implement TTS and streaming TTS
mudler
committed
579cbb0
feat(nemo-speech-cpp): surface NMT translation through Predict
mudler
committed
14ed23e
test(nemo-speech-cpp): pin the three-segment pair tag in an NMT directive
mudler
committed
fdce1dc
feat(nemo-speech-cpp): register the backend and give its specs a CI job
mudler
committed
568ae16
fix(nemo-speech-cpp): build the CUDA-13 Jetson image the l4t-cuda-13 key needs
mudler
committed
6866691
docs(nemo-speech-cpp): document the backend and list it in the importer
mudler
committed
52b5a61
docs(nemo-speech-cpp): correct the translation limits, the macOS gap and the TTS conversion
mudler
committed
b15c010
fix(nemo-speech-cpp): map every C status, not just the NMT one
mudler
committed
6f301d1
fix(nemo-speech-cpp): audit the gosec unsafe and file-inclusion sites
mudler
committed
eb4863e

Navigation Menu

Uh oh!

feat(nemo-speech-cpp): add the NVIDIA NeMo-Speech.cpp backend - #11406

feat(nemo-speech-cpp): add the NVIDIA NeMo-Speech.cpp backend#11406
localai-bot wants to merge 30 commits into
masterfrom
feat/nemo-speech-cpp-backend

Commits on Aug 6, 2026

feat(nemo-speech-cpp): scaffold the backend and upstream build

fix(nemo-speech-cpp): make 'build' produce the package and bundle the ITN stack

fix(nemo-speech-cpp): make the closure guard fail closed and give CI its toolchain

fix(nemo-speech-cpp): pin protoc for ITN and make the norm stack its own target

fix(nemo-speech-cpp): move the backend apt gate below the expensive layers

feat(nemo-speech-cpp): parse model options

feat(nemo-speech-cpp): guard the empty option value and warn on a bad gpu index

feat(nemo-speech-cpp): detect model family and discover TTS assets

feat(nemo-speech-cpp): bind the C ABI with layout assertions

test(nemo-speech-cpp): run the ABI specs in CI and refuse to skip them

feat(nemo-speech-cpp): select the family at load and gate RPCs on it

fix(nemo-speech-cpp): pin the load ordering, close the engineMu race

feat(nemo-speech-cpp): implement offline transcription

feat(nemo-speech-cpp): implement streaming and live transcription

fix(nemo-speech-cpp): make live deltas concatenate and fill segment words

feat(nemo-speech-cpp): implement standalone diarization

fix(nemo-speech-cpp): pin the diarizer geometry sentinels and cap the segment buffer

feat(nemo-speech-cpp): implement TTS and streaming TTS

feat(nemo-speech-cpp): surface NMT translation through Predict

test(nemo-speech-cpp): pin the three-segment pair tag in an NMT directive

feat(nemo-speech-cpp): register the backend and give its specs a CI job

fix(nemo-speech-cpp): build the CUDA-13 Jetson image the l4t-cuda-13 key needs

docs(nemo-speech-cpp): document the backend and list it in the importer

docs(nemo-speech-cpp): correct the translation limits, the macOS gap and the TTS conversion

fix(nemo-speech-cpp): map every C status, not just the NMT one

fix(nemo-speech-cpp): audit the gosec unsafe and file-inclusion sites

Commits on Aug 7, 2026

fix(nemo-speech-cpp): skip the CUDA-only ggml patch series on darwin

fix(nemo-speech-cpp): restore std::binary_function for MeCab on libc++

fix(nemo-speech-cpp): repair OpenFST's FstImpl::operator= for gcc-14

fix(nemo-speech-cpp): install cmake 3.31 on bases that ship less than 3.26