merge: upstream main through release-0.4 (Higgs Audio v3 TTS 4B)

Author: owlab-developer

.devops/cpu.Dockerfile

@@ -13,7 +13,6 @@ ARG GCC_VERSION=14
 FROM docker.io/ubuntu:$UBUNTU_VERSION AS build
 
 ARG GCC_VERSION=14
-ARG ENGINE_ENABLE_NATIVE_CPU=ON
 
 # Install build toolchain
 RUN apt-get update && \
@@ -27,13 +26,22 @@ ENV CC=gcc-${GCC_VERSION} CXX=g++-${GCC_VERSION}
 WORKDIR /app
 COPY . .
 
+# Validate architecture (CPU backend is only supported on amd64 and arm64)
+ARG TARGETARCH
+RUN if [ "$TARGETARCH" = "amd64" ] || [ "$TARGETARCH" = "arm64" ]; then \
+        echo "Building for $TARGETARCH"; \
+    else \
+        echo "Unsupported architecture: $TARGETARCH"; \
+        exit 1; \
+    fi
+
 # Configure and build
 RUN cmake -S . -B build \
         -DCMAKE_BUILD_TYPE=Release \
+        -DENGINE_ENABLE_CPU_ALL_VARIANTS=ON \
         -DENGINE_ENABLE_CUDA=OFF \
         -DENGINE_ENABLE_VULKAN=OFF \
         -DENGINE_ENABLE_OPENMP=ON \
-        -DENGINE_ENABLE_NATIVE_CPU=$ENGINE_ENABLE_NATIVE_CPU \
         -DENGINE_BUILD_EXAMPLES=OFF \
         -DENGINE_BUILD_TESTS=OFF \
         -DENGINE_BUILD_WARMBENCH=OFF && \
@@ -43,7 +51,11 @@ RUN cmake -S . -B build \
         --target model_perf \
         --target miocodec_wavlm_parity
 
-# Collect all binaries + multiplexer into /app/full
+# Collect shared libraries
+RUN mkdir -p /app/lib && \
+    find build -name "*.so*" -exec cp -P {} /app/lib \;
+
+# Collect binaries + multiplexer into /app/full
 RUN mkdir -p /app/full && \
     cp build/bin/audiocpp_cli build/bin/audiocpp_server \
        build/bin/model_perf build/bin/miocodec_wavlm_parity /app/full/ && \
@@ -77,6 +89,8 @@ RUN apt-get update && \
     find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete && \
     find /var/cache -type f -delete
 
+COPY --from=build /app/lib/ /app
+
 WORKDIR /app
 
 # ── FULL: All binaries + entrypoint.sh multiplexer ────────────────────────────

.devops/cuda.Dockerfile

@@ -17,7 +17,6 @@ ARG BASE_CUDA_RUN_CONTAINER=docker.io/nvidia/cuda:${CUDA_VERSION}-runtime-ubuntu
 FROM ${BASE_CUDA_DEV_CONTAINER} AS build
 
 ARG GCC_VERSION=14
-ARG ENGINE_ENABLE_NATIVE_CPU=ON
 
 # Install build toolchain
 RUN apt-get update && \
@@ -34,21 +33,26 @@ COPY . .
 # Configure and build
 RUN cmake -S . -B build \
         -DCMAKE_BUILD_TYPE=Release \
+        -DENGINE_ENABLE_CPU_ALL_VARIANTS=ON \
         -DENGINE_ENABLE_CUDA=ON \
         -DENGINE_ENABLE_CUDA_GRAPHS=ON \
         -DENGINE_ENABLE_VULKAN=OFF \
         -DENGINE_ENABLE_OPENMP=ON \
-        -DENGINE_ENABLE_NATIVE_CPU=$ENGINE_ENABLE_NATIVE_CPU \
         -DENGINE_BUILD_EXAMPLES=OFF \
         -DENGINE_BUILD_TESTS=OFF \
-        -DENGINE_BUILD_WARMBENCH=OFF && \
+        -DENGINE_BUILD_WARMBENCH=OFF \
+        -DCMAKE_EXE_LINKER_FLAGS=-Wl,--allow-shlib-undefined && \
     cmake --build build --parallel $(nproc) \
         --target audiocpp_cli \
         --target audiocpp_server \
         --target model_perf \
         --target miocodec_wavlm_parity
 
-# Collect all binaries + multiplexer into /app/full
+# Collect shared libraries
+RUN mkdir -p /app/lib && \
+    find build -name "*.so*" -exec cp -P {} /app/lib \;
+
+# Collect binaries + multiplexer into /app/full
 RUN mkdir -p /app/full && \
     cp build/bin/audiocpp_cli build/bin/audiocpp_server \
        build/bin/model_perf build/bin/miocodec_wavlm_parity /app/full/ && \
@@ -82,6 +86,8 @@ RUN apt-get update && \
     find /var/cache/apt/archives /var/lib/apt/lists -not -name lock -type f -delete && \
     find /var/cache -type f -delete
 
+COPY --from=build /app/lib/ /app
+
 WORKDIR /app
 
 # ── FULL: All binaries + entrypoint.sh multiplexer ────────────────────────────

.github/workflows/docker.yml

@@ -181,7 +181,6 @@ jobs:
             APP_REVISION=${{ github.sha }}
             IMAGE_URL=${{ github.server_url }}/${{ github.repository }}
             IMAGE_SOURCE=${{ github.server_url }}/${{ github.repository }}
-            ENGINE_ENABLE_NATIVE_CPU=OFF
             ${{ matrix.cuda_version && format('CUDA_VERSION={0}', matrix.cuda_version) || '' }}
           cache-from: type=registry,ref=ghcr.io/${{ needs.metadata.outputs.image_repo }}:cache-${{ matrix.arch }}-${{ matrix.tag }}
           cache-to: type=registry,ref=ghcr.io/${{ needs.metadata.outputs.image_repo }}:cache-${{ matrix.arch }}-${{ matrix.tag }},mode=max

.gitignore

@@ -28,3 +28,18 @@ __pycache__/
 /*.wav
 /flutter/
 /.cache/
+
+# --- portable bundle: distributed manually, never committed ---
+/audiocpp-portable/
+
+# --- WebUI runtime artifacts: keep the source + empty dirs, drop generated files ---
+/webui/output/*
+!/webui/output/.gitkeep
+/webui/logs/*
+!/webui/logs/.gitkeep
+/webui/third_party/
+/webui/llm_api_key.txt
+# written by the in-UI language picker; per-machine, not a project setting
+/webui/configs/ui_language.json
+# personal voice recording — stays local, repo is public
+/webui/voice/my-record.wav