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

.github/backend-matrix.yml

@@ -860,6 +860,19 @@ include:
     dockerfile: "./backend/Dockerfile.golang"
     context: "./"
     ubuntu-version: '2404'
+  - build-type: 'cublas'
+    cuda-major-version: "12"
+    cuda-minor-version: "8"
+    platforms: 'linux/amd64'
+    tag-latest: 'auto'
+    tag-suffix: '-gpu-nvidia-cuda-12-nemo-speech-cpp'
+    runs-on: 'ubuntu-latest'
+    base-image: "ubuntu:24.04"
+    skip-drivers: 'false'
+    backend: "nemo-speech-cpp"
+    dockerfile: "./backend/Dockerfile.golang"
+    context: "./"
+    ubuntu-version: '2404'
   - build-type: 'cublas'
     cuda-major-version: "12"
     cuda-minor-version: "8"
@@ -1911,6 +1924,19 @@ include:
     dockerfile: "./backend/Dockerfile.golang"
     context: "./"
     ubuntu-version: '2404'
+  - build-type: 'cublas'
+    cuda-major-version: "13"
+    cuda-minor-version: "0"
+    platforms: 'linux/amd64'
+    tag-latest: 'auto'
+    tag-suffix: '-gpu-nvidia-cuda-13-nemo-speech-cpp'
+    runs-on: 'ubuntu-latest'
+    base-image: "ubuntu:24.04"
+    skip-drivers: 'false'
+    backend: "nemo-speech-cpp"
+    dockerfile: "./backend/Dockerfile.golang"
+    context: "./"
+    ubuntu-version: '2404'
   - build-type: 'cublas'
     cuda-major-version: "13"
     cuda-minor-version: "0"
@@ -1963,6 +1989,24 @@ include:
     backend: "parakeet-cpp"
     dockerfile: "./backend/Dockerfile.golang"
     context: "./"
+  # The CUDA-13 counterpart to the JetPack r36.4.0 row in the nemo-speech-cpp
+  # block below. A Jetson whose CUDA 13 runtime is present reports the
+  # nvidia-l4t-cuda-13 capability, and pointing that key at the JetPack image
+  # would hand it a ggml linked against CUDA 12 whose libcudart.so.12 is not
+  # there to dlopen. Same base and runner as the parakeet-cpp row above.
+  - build-type: 'cublas'
+    cuda-major-version: "13"
+    cuda-minor-version: "0"
+    platforms: 'linux/arm64'
+    skip-drivers: 'false'
+    tag-latest: 'auto'
+    tag-suffix: '-nvidia-l4t-cuda-13-arm64-nemo-speech-cpp'
+    base-image: "ubuntu:24.04"
+    ubuntu-version: '2404'
+    runs-on: 'ubuntu-24.04-arm'
+    backend: "nemo-speech-cpp"
+    dockerfile: "./backend/Dockerfile.golang"
+    context: "./"
   - build-type: 'cublas'
     cuda-major-version: "13"
     cuda-minor-version: "0"
@@ -4183,6 +4227,86 @@ include:
     dockerfile: "./backend/Dockerfile.golang"
     context: "./"
     ubuntu-version: '2404'
+  # nemo-speech-cpp
+  #
+  # No hipblas and no sycl rows, unlike the parakeet-cpp block above: upstream
+  # NeMo-Speech.cpp builds ggml with CUDA, Vulkan or Metal only, so a ROCm or
+  # SYCL image would be a CPU build wearing a GPU tag.
+  #
+  # cpu and vulkan are per-arch pairs sharing one tag-suffix, so
+  # backend-merge-jobs assembles a multi-arch manifest from the two digests.
+  # The arm64 legs are not redundant with the Jetson image below: an ARM server
+  # with no NVIDIA GPU reports the "default" capability and would otherwise pull
+  # an amd64-only manifest.
+  - build-type: ''
+    cuda-major-version: ""
+    cuda-minor-version: ""
+    platforms: 'linux/amd64'
+    platform-tag: 'amd64'
+    tag-latest: 'auto'
+    tag-suffix: '-cpu-nemo-speech-cpp'
+    runs-on: 'ubuntu-latest'
+    base-image: "ubuntu:24.04"
+    skip-drivers: 'false'
+    backend: "nemo-speech-cpp"
+    dockerfile: "./backend/Dockerfile.golang"
+    context: "./"
+    ubuntu-version: '2404'
+  - build-type: ''
+    cuda-major-version: ""
+    cuda-minor-version: ""
+    platforms: 'linux/arm64'
+    platform-tag: 'arm64'
+    tag-latest: 'auto'
+    tag-suffix: '-cpu-nemo-speech-cpp'
+    runs-on: 'ubuntu-24.04-arm'
+    base-image: "ubuntu:24.04"
+    skip-drivers: 'false'
+    backend: "nemo-speech-cpp"
+    dockerfile: "./backend/Dockerfile.golang"
+    context: "./"
+    ubuntu-version: '2404'
+  - build-type: 'vulkan'
+    cuda-major-version: ""
+    cuda-minor-version: ""
+    platforms: 'linux/amd64'
+    platform-tag: 'amd64'
+    tag-latest: 'auto'
+    tag-suffix: '-gpu-vulkan-nemo-speech-cpp'
+    runs-on: 'ubuntu-latest'
+    base-image: "ubuntu:24.04"
+    skip-drivers: 'false'
+    backend: "nemo-speech-cpp"
+    dockerfile: "./backend/Dockerfile.golang"
+    context: "./"
+    ubuntu-version: '2404'
+  - build-type: 'vulkan'
+    cuda-major-version: ""
+    cuda-minor-version: ""
+    platforms: 'linux/arm64'
+    platform-tag: 'arm64'
+    tag-latest: 'auto'
+    tag-suffix: '-gpu-vulkan-nemo-speech-cpp'
+    runs-on: 'ubuntu-24.04-arm'
+    base-image: "ubuntu:24.04"
+    skip-drivers: 'false'
+    backend: "nemo-speech-cpp"
+    dockerfile: "./backend/Dockerfile.golang"
+    context: "./"
+    ubuntu-version: '2404'
+  - build-type: 'cublas'
+    cuda-major-version: "12"
+    cuda-minor-version: "0"
+    platforms: 'linux/arm64'
+    skip-drivers: 'false'
+    tag-latest: 'auto'
+    tag-suffix: '-nvidia-l4t-arm64-nemo-speech-cpp'
+    base-image: "nvcr.io/nvidia/l4t-jetpack:r36.4.0"
+    runs-on: 'ubuntu-24.04-arm'
+    backend: "nemo-speech-cpp"
+    dockerfile: "./backend/Dockerfile.golang"
+    context: "./"
+    ubuntu-version: '2204'
   # moss-transcribe-cpp
   - build-type: ''
     cuda-major-version: ""
@@ -6226,6 +6350,10 @@ includeDarwin:
     tag-suffix: "-metal-darwin-arm64-moss-transcribe-cpp"
     build-type: "metal"
     lang: "go"
+  - backend: "nemo-speech-cpp"
+    tag-suffix: "-metal-darwin-arm64-nemo-speech-cpp"
+    build-type: "metal"
+    lang: "go"
   - backend: "ced"
     tag-suffix: "-metal-darwin-arm64-ced"
     build-type: "metal"

.github/workflows/bump_deps.yaml

@@ -62,6 +62,10 @@ jobs:
             variable: "MOSS_VERSION"
             branch: "master"
             file: "backend/go/moss-transcribe-cpp/Makefile"
+          - repository: "NVIDIA/NeMo-Speech.cpp"
+            variable: "NEMO_SPEECH_VERSION"
+            branch: "main"
+            file: "backend/go/nemo-speech-cpp/Makefile"
           - repository: "localai-org/ced.cpp"
             variable: "CED_VERSION"
             branch: "main"

.github/workflows/test-extra.yml

@@ -50,6 +50,7 @@ jobs:
       sherpa-onnx: ${{ steps.detect.outputs.sherpa-onnx }}
       whisper: ${{ steps.detect.outputs.whisper }}
       parakeet-cpp: ${{ steps.detect.outputs.parakeet-cpp }}
+      nemo-speech-cpp: ${{ steps.detect.outputs.nemo-speech-cpp }}
     steps:
       - name: Checkout repository
         uses: actions/checkout@v7
@@ -900,6 +901,57 @@ jobs:
       - name: Test magpie-tts-cpp
         run: |
           make --jobs=5 --output-sync=target -C backend/go/magpie-tts-cpp test
+  # Per-backend unit suite for nemo-speech-cpp. This job exists for one reason
+  # above all: abi_test.go asserts the size and field offsets of every Go mirror
+  # struct against the C ABI it is dlopened into. Those assertions are the only
+  # thing standing between a purego symbol rename or an upstream header change
+  # and silent memory corruption at run time, and they are worthless unless
+  # something executes them. `make -C backend/go/nemo-speech-cpp test` sets
+  # NEMO_SPEECH_REQUIRE_LIBS=1, which turns "library missing" from a skip into a
+  # failure, so this job cannot report green having checked nothing.
+  #
+  # The backend Makefile's `test` target depends on `stage-libs`, so it clones
+  # upstream at the pinned SHA and builds the native runtime itself. There is no
+  # separate build step for that reason, and no model download: the specs are
+  # ABI and pure-Go only.
+  #
+  # WITH_NORM=OFF skips the Sparrowhawk/OpenFST inverse-text-normalization
+  # stack, which is the single most expensive leg of the build and needs a gcc-12
+  # pin because OpenFST's templates ICE on gcc-13/14. It costs no coverage here:
+  # nothing in include/nemo_speech/{asr,tts,diar,nmt}.h is conditional on it (the
+  # only preprocessor conditionals in those headers are include guards,
+  # __cplusplus and the _WIN32 export macros), so every struct layout this suite
+  # checks is identical either way. The shipped images still build WITH_NORM=ON;
+  # that path is covered by the backend image build in backend_pr.yml.
+  tests-nemo-speech-cpp:
+    needs: detect-changes
+    if: needs.detect-changes.outputs.nemo-speech-cpp == 'true' || needs.detect-changes.outputs.run-all == 'true'
+    runs-on: ubuntu-latest
+    timeout-minutes: 90
+    steps:
+      - name: Clone
+        uses: actions/checkout@v7
+        with:
+          submodules: true
+      - name: Dependencies
+        run: |
+          sudo apt-get update
+          sudo apt-get install -y build-essential cmake ninja-build curl libopenblas-dev ffmpeg
+      - name: Setup Go
+        uses: actions/setup-go@v5
+      - name: Display Go version
+        run: go version
+      - name: Proto Dependencies
+        run: |
+          curl -L -s https://github.com/protocolbuffers/protobuf/releases/download/v26.1/protoc-26.1-linux-x86_64.zip -o protoc.zip && \
+          unzip -j -d /usr/local/bin protoc.zip bin/protoc && \
+          rm protoc.zip
+          go install google.golang.org/protobuf/cmd/protoc-gen-go@v1.34.2
+          go install google.golang.org/grpc/cmd/protoc-gen-go-grpc@1958fcbe2ca8bd93af633f11e97d44e567e945af
+          PATH="$PATH:$HOME/go/bin" make protogen-go
+      - name: Test nemo-speech-cpp
+        run: |
+          make --jobs=5 --output-sync=target -C backend/go/nemo-speech-cpp WITH_NORM=OFF test
   # Per-backend smoke for rfdetr-cpp: builds the .so + Go binary and runs
   # `make -C backend/go/rfdetr-cpp test`. test.sh fetches the small (~20 MB)
   # rfdetr-nano-q8_0 GGUF from the published mudler/rfdetr-cpp-nano HF repo

Makefile

@@ -1,5 +1,5 @@
 # Disable parallel execution for backend builds
-.NOTPARALLEL: backends/diffusers backends/llama-cpp backends/turboquant backends/bonsai backends/outetts backends/piper backends/stablediffusion-ggml backends/trellis2cpp backends/trellis2cpp-darwin backends/whisper backends/crispasr backends/parakeet-cpp backends/moss-transcribe-cpp backends/faster-whisper backends/silero-vad backends/local-store backends/valkey-store backends/cloud-proxy backends/huggingface backends/rfdetr backends/rfdetr-cpp backends/insightface backends/speaker-recognition backends/kitten-tts backends/kokoro backends/chatterbox backends/llama-cpp-darwin backends/neutts build-darwin-python-backend build-darwin-go-backend backends/mlx backends/diffuser-darwin backends/mlx-vlm backends/mlx-audio backends/mlx-distributed backends/stablediffusion-ggml-darwin backends/vllm backends/vllm-omni backends/longcat-video backends/sglang backends/moonshine backends/pocket-tts backends/qwen-tts backends/faster-qwen3-tts backends/qwen-asr backends/nemo backends/voxcpm backends/whisperx backends/ace-step backends/acestep-cpp backends/fish-speech backends/voxtral backends/opus backends/trl backends/llama-cpp-quantization backends/kokoros backends/sam3-cpp backends/qwen3-tts-cpp backends/moss-tts-cpp backends/magpie-tts-cpp backends/vllm-cpp backends/omnivoice-cpp backends/vibevoice-cpp backends/localvqe backends/tinygrad backends/sherpa-onnx backends/ds4 backends/ds4-darwin backends/liquid-audio backends/supertonic backends/depth-anything-cpp backends/privacy-filter backends/privacy-filter-darwin backends/audio-cpp backends/audio-cpp-darwin
+.NOTPARALLEL: backends/diffusers backends/llama-cpp backends/turboquant backends/bonsai backends/outetts backends/piper backends/stablediffusion-ggml backends/trellis2cpp backends/trellis2cpp-darwin backends/whisper backends/crispasr backends/parakeet-cpp backends/moss-transcribe-cpp backends/nemo-speech-cpp backends/faster-whisper backends/silero-vad backends/local-store backends/valkey-store backends/cloud-proxy backends/huggingface backends/rfdetr backends/rfdetr-cpp backends/insightface backends/speaker-recognition backends/kitten-tts backends/kokoro backends/chatterbox backends/llama-cpp-darwin backends/neutts build-darwin-python-backend build-darwin-go-backend backends/mlx backends/diffuser-darwin backends/mlx-vlm backends/mlx-audio backends/mlx-distributed backends/stablediffusion-ggml-darwin backends/vllm backends/vllm-omni backends/longcat-video backends/sglang backends/moonshine backends/pocket-tts backends/qwen-tts backends/faster-qwen3-tts backends/qwen-asr backends/nemo backends/voxcpm backends/whisperx backends/ace-step backends/acestep-cpp backends/fish-speech backends/voxtral backends/opus backends/trl backends/llama-cpp-quantization backends/kokoros backends/sam3-cpp backends/qwen3-tts-cpp backends/moss-tts-cpp backends/magpie-tts-cpp backends/vllm-cpp backends/omnivoice-cpp backends/vibevoice-cpp backends/localvqe backends/tinygrad backends/sherpa-onnx backends/ds4 backends/ds4-darwin backends/liquid-audio backends/supertonic backends/depth-anything-cpp backends/privacy-filter backends/privacy-filter-darwin backends/audio-cpp backends/audio-cpp-darwin
 
 GOCMD=go
 GOTEST=$(GOCMD) test
@@ -654,6 +654,7 @@ test-extra: prepare-test-extra
 	$(MAKE) -C backend/go/depth-anything-cpp test
 	$(MAKE) -C backend/go/supertonic test
 	$(MAKE) -C backend/go/vllm-cpp test
+	$(MAKE) -C backend/go/nemo-speech-cpp test
 	$(MAKE) -C backend/go/trellis2cpp test
 	$(MAKE) -C backend/go/valkey-store test
 
@@ -1298,6 +1299,7 @@ BACKEND_WHISPER = whisper|golang|.|false|true
 BACKEND_CRISPASR = crispasr|golang|.|false|true
 BACKEND_PARAKEET_CPP = parakeet-cpp|golang|.|false|true
 BACKEND_MOSS_TRANSCRIBE_CPP = moss-transcribe-cpp|golang|.|false|true
+BACKEND_NEMO_SPEECH_CPP = nemo-speech-cpp|golang|.|false|true
 BACKEND_DEPTH_ANYTHING_CPP = depth-anything-cpp|golang|.|false|true
 BACKEND_VOXTRAL = voxtral|golang|.|false|true
 BACKEND_ACESTEP_CPP = acestep-cpp|golang|.|false|true
@@ -1400,6 +1402,7 @@ $(eval $(call generate-docker-build-target,$(BACKEND_WHISPER)))
 $(eval $(call generate-docker-build-target,$(BACKEND_CRISPASR)))
 $(eval $(call generate-docker-build-target,$(BACKEND_PARAKEET_CPP)))
 $(eval $(call generate-docker-build-target,$(BACKEND_MOSS_TRANSCRIBE_CPP)))
+$(eval $(call generate-docker-build-target,$(BACKEND_NEMO_SPEECH_CPP)))
 $(eval $(call generate-docker-build-target,$(BACKEND_DEPTH_ANYTHING_CPP)))
 $(eval $(call generate-docker-build-target,$(BACKEND_VOXTRAL)))
 $(eval $(call generate-docker-build-target,$(BACKEND_OPUS)))
@@ -1456,7 +1459,7 @@ $(eval $(call generate-docker-build-target,$(BACKEND_SUPERTONIC)))
 docker-save-%: backend-images
 	docker save local-ai-backend:$* -o backend-images/$*.tar
 
-docker-build-backends: docker-build-llama-cpp docker-build-ik-llama-cpp docker-build-turboquant docker-build-bonsai docker-build-ds4 docker-build-rerankers docker-build-vllm docker-build-vllm-omni docker-build-longcat-video docker-build-sglang docker-build-transformers docker-build-outetts docker-build-diffusers docker-build-kokoro docker-build-faster-whisper docker-build-crispasr docker-build-coqui docker-build-chatterbox docker-build-vibevoice docker-build-liquid-audio docker-build-moonshine docker-build-pocket-tts docker-build-qwen-tts docker-build-fish-speech docker-build-faster-qwen3-tts docker-build-qwen-asr docker-build-nemo docker-build-voxcpm docker-build-whisperx docker-build-ace-step docker-build-acestep-cpp docker-build-voxtral docker-build-mlx-distributed docker-build-trl docker-build-llama-cpp-quantization docker-build-tinygrad docker-build-kokoros docker-build-sam3-cpp docker-build-rfdetr-cpp docker-build-qwen3-tts-cpp docker-build-moss-tts-cpp docker-build-magpie-tts-cpp docker-build-vllm-cpp docker-build-omnivoice-cpp docker-build-vibevoice-cpp docker-build-localvqe docker-build-insightface docker-build-speaker-recognition docker-build-sherpa-onnx docker-build-cloud-proxy docker-build-supertonic docker-build-depth-anything-cpp docker-build-moss-transcribe-cpp docker-build-privacy-filter docker-build-trellis2cpp docker-build-valkey-store docker-build-audio-cpp
+docker-build-backends: docker-build-llama-cpp docker-build-ik-llama-cpp docker-build-turboquant docker-build-bonsai docker-build-ds4 docker-build-rerankers docker-build-vllm docker-build-vllm-omni docker-build-longcat-video docker-build-sglang docker-build-transformers docker-build-outetts docker-build-diffusers docker-build-kokoro docker-build-faster-whisper docker-build-crispasr docker-build-coqui docker-build-chatterbox docker-build-vibevoice docker-build-liquid-audio docker-build-moonshine docker-build-pocket-tts docker-build-qwen-tts docker-build-fish-speech docker-build-faster-qwen3-tts docker-build-qwen-asr docker-build-nemo docker-build-voxcpm docker-build-whisperx docker-build-ace-step docker-build-acestep-cpp docker-build-voxtral docker-build-mlx-distributed docker-build-trl docker-build-llama-cpp-quantization docker-build-tinygrad docker-build-kokoros docker-build-sam3-cpp docker-build-rfdetr-cpp docker-build-qwen3-tts-cpp docker-build-moss-tts-cpp docker-build-magpie-tts-cpp docker-build-vllm-cpp docker-build-omnivoice-cpp docker-build-vibevoice-cpp docker-build-localvqe docker-build-insightface docker-build-speaker-recognition docker-build-sherpa-onnx docker-build-cloud-proxy docker-build-supertonic docker-build-depth-anything-cpp docker-build-moss-transcribe-cpp docker-build-nemo-speech-cpp docker-build-privacy-filter docker-build-trellis2cpp docker-build-valkey-store docker-build-audio-cpp
 
 ########################################################
 ### Mock Backend for E2E Tests