WebAssembly exmaple for speaker diarization

.github/workflows/wasm-simd-hf-space-de-tts.yaml

@@ -29,7 +29,7 @@ jobs:
       - name: Install emsdk
         uses: mymindstorm/setup-emsdk@v14
         with:
-          version: 3.1.51
+          version: 3.1.53
           actions-cache-folder: 'emsdk-cache'
 
       - name: View emsdk version

.github/workflows/wasm-simd-hf-space-en-asr-zipformer.yaml

@@ -28,7 +28,7 @@ jobs:
       - name: Install emsdk
         uses: mymindstorm/setup-emsdk@v14
         with:
-          version: 3.1.51
+          version: 3.1.53
           actions-cache-folder: 'emsdk-cache'
 
       - name: View emsdk version

.github/workflows/wasm-simd-hf-space-en-tts.yaml

@@ -29,7 +29,7 @@ jobs:
       - name: Install emsdk
         uses: mymindstorm/setup-emsdk@v14
         with:
-          version: 3.1.51
+          version: 3.1.53
           actions-cache-folder: 'emsdk-cache'
 
       - name: View emsdk version

.github/workflows/wasm-simd-hf-space-silero-vad.yaml

@@ -29,7 +29,7 @@ jobs:
       - name: Install emsdk
         uses: mymindstorm/setup-emsdk@v14
         with:
-          version: 3.1.51
+          version: 3.1.53
           actions-cache-folder: 'emsdk-cache'
 
       - name: View emsdk version

.github/workflows/wasm-simd-hf-space-speaker-diarization.yaml

@@ -0,0 +1,167 @@
+name: wasm-simd-hf-space-speaker-diarization
+
+on:
+  push:
+    branches:
+      - wasm
+      - wasm-speaker-diarization
+    tags:
+      - 'v[0-9]+.[0-9]+.[0-9]+*'
+
+  workflow_dispatch:
+
+concurrency:
+  group: wasm-simd-hf-space-speaker-diarization-${{ github.ref }}
+  cancel-in-progress: true
+
+jobs:
+  wasm-simd-hf-space-speaker-diarization:
+    runs-on: ${{ matrix.os }}
+    strategy:
+      fail-fast: false
+      matrix:
+        os: [ubuntu-latest]
+
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+
+      - name: Install emsdk
+        uses: mymindstorm/setup-emsdk@v14
+        with:
+          version: 3.1.53
+          actions-cache-folder: 'emsdk-cache'
+
+      - name: View emsdk version
+        shell: bash
+        run: |
+          emcc -v
+          echo "--------------------"
+          emcc --check
+
+      - name: Download model files
+        shell: bash
+        run: |
+          cd wasm/speaker-diarization/assets/
+          ls -lh
+          echo "----------"
+
+          curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/speaker-segmentation-models/sherpa-onnx-pyannote-segmentation-3-0.tar.bz2
+          tar xvf sherpa-onnx-pyannote-segmentation-3-0.tar.bz2
+          rm sherpa-onnx-pyannote-segmentation-3-0.tar.bz2
+          mv sherpa-onnx-pyannote-segmentation-3-0/model.onnx ./segmentation.onnx
+          rm -rf sherpa-onnx-pyannote-segmentation-3-0
+
+          curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/speaker-recongition-models/3dspeaker_speech_eres2net_base_sv_zh-cn_3dspeaker_16k.onnx
+          mv 3dspeaker_speech_eres2net_base_sv_zh-cn_3dspeaker_16k.onnx ./embedding.onnx
+
+          echo "----------"
+
+          ls -lh
+
+      - name: Build sherpa-onnx for WebAssembly
+        shell: bash
+        run: |
+          ./build-wasm-simd-speaker-diarization.sh
+
+      - name: collect files
+        shell: bash
+        run: |
+          SHERPA_ONNX_VERSION=v$(grep "SHERPA_ONNX_VERSION" ./CMakeLists.txt  | cut -d " " -f 2  | cut -d '"' -f 2)
+
+          dst=sherpa-onnx-wasm-simd-${SHERPA_ONNX_VERSION}-speaker-diarization
+          mv build-wasm-simd-speaker-diarization/install/bin/wasm/speaker-diarization $dst
+          ls -lh $dst
+          tar cjfv $dst.tar.bz2 ./$dst
+
+      - name: Upload wasm files
+        uses: actions/upload-artifact@v4
+        with:
+          name: sherpa-onnx-wasm-simd-speaker-diarization
+          path: ./sherpa-onnx-wasm-simd-*.tar.bz2
+
+      - name: Release
+        if: (github.repository_owner == 'csukuangfj' || github.repository_owner == 'k2-fsa') && github.event_name == 'push' && contains(github.ref, 'refs/tags/')
+        uses: svenstaro/upload-release-action@v2
+        with:
+          file_glob: true
+          overwrite: true
+          file: ./*.tar.bz2
+
+      - name: Publish to ModelScope
+        # if: false
+        env:
+          MS_TOKEN: ${{ secrets.MODEL_SCOPE_GIT_TOKEN }}
+        uses: nick-fields/retry@v2
+        with:
+          max_attempts: 20
+          timeout_seconds: 200
+          shell: bash
+          command: |
+            SHERPA_ONNX_VERSION=v$(grep "SHERPA_ONNX_VERSION" ./CMakeLists.txt  | cut -d " " -f 2  | cut -d '"' -f 2)
+
+            git config --global user.email "[email protected]"
+            git config --global user.name "Fangjun Kuang"
+
+            rm -rf ms
+            export GIT_LFS_SKIP_SMUDGE=1
+            export GIT_CLONE_PROTECTION_ACTIVE=false
+
+            git clone https://www.modelscope.cn/studios/csukuangfj/web-assembly-speaker-diarization-sherpa-onnx.git ms
+            cd ms
+            rm -fv *.js
+            rm -fv *.data
+            git fetch
+            git pull
+            git merge -m "merge remote" --ff origin main
+
+            cp -v ../sherpa-onnx-wasm-simd-${SHERPA_ONNX_VERSION}-*/* .
+
+            git status
+            git lfs track "*.data"
+            git lfs track "*.wasm"
+            ls -lh
+
+            git add .
+            git commit -m "update model"
+            git push https://oauth2:${MS_TOKEN}@www.modelscope.cn/studios/csukuangfj/web-assembly-speaker-diarization-sherpa-onnx.git
+
+      - name: Publish to huggingface
+        env:
+          HF_TOKEN: ${{ secrets.HF_TOKEN }}
+        uses: nick-fields/retry@v2
+        with:
+          max_attempts: 20
+          timeout_seconds: 200
+          shell: bash
+          command: |
+            SHERPA_ONNX_VERSION=v$(grep "SHERPA_ONNX_VERSION" ./CMakeLists.txt  | cut -d " " -f 2  | cut -d '"' -f 2)
+
+            git config --global user.email "[email protected]"
+            git config --global user.name "Fangjun Kuang"
+
+            rm -rf huggingface
+            export GIT_LFS_SKIP_SMUDGE=1
+            export GIT_CLONE_PROTECTION_ACTIVE=false
+
+            git clone https://csukuangfj:[email protected]/spaces/k2-fsa/web-assembly-speaker-diarization-sherpa-onnx huggingface
+            ls -lh
+
+            cd huggingface
+            rm -fv *.js
+            rm -fv *.data
+            git fetch
+            git pull
+            git merge -m "merge remote" --ff origin main
+
+            cp -v ../sherpa-onnx-wasm-simd-${SHERPA_ONNX_VERSION}-*/* .
+
+            git status
+            git lfs track "*.data"
+            git lfs track "*.wasm"
+            ls -lh
+
+            git add .
+            git commit -m "update model"
+            git push https://csukuangfj:[email protected]/spaces/k2-fsa/web-assembly-speaker-diarization-sherpa-onnx main

.github/workflows/wasm-simd-hf-space-vad-asr.yaml

@@ -37,7 +37,7 @@ jobs:
       - name: Install emsdk
         uses: mymindstorm/setup-emsdk@v14
         with:
-          version: 3.1.51
+          version: 3.1.53
           actions-cache-folder: 'emsdk-cache'
 
       - name: View emsdk version

.github/workflows/wasm-simd-hf-space-zh-cantonese-en-asr-paraformer.yaml

@@ -29,7 +29,7 @@ jobs:
       - name: Install emsdk
         uses: mymindstorm/setup-emsdk@v14
         with:
-          version: 3.1.51
+          version: 3.1.53
           actions-cache-folder: 'emsdk-cache'
 
       - name: View emsdk version

.github/workflows/wasm-simd-hf-space-zh-en-asr-paraformer.yaml

@@ -29,7 +29,7 @@ jobs:
       - name: Install emsdk
         uses: mymindstorm/setup-emsdk@v14
         with:
-          version: 3.1.51
+          version: 3.1.53
           actions-cache-folder: 'emsdk-cache'
 
       - name: View emsdk version

.github/workflows/wasm-simd-hf-space-zh-en-asr-zipformer.yaml

@@ -29,7 +29,7 @@ jobs:
       - name: Install emsdk
         uses: mymindstorm/setup-emsdk@v14
         with:
-          version: 3.1.51
+          version: 3.1.53
           actions-cache-folder: 'emsdk-cache'
 
       - name: View emsdk version

CMakeLists.txt

@@ -32,6 +32,7 @@ option(SHERPA_ONNX_ENABLE_WEBSOCKET "Whether to build webscoket server/client" O
 option(SHERPA_ONNX_ENABLE_GPU "Enable ONNX Runtime GPU support" OFF)
 option(SHERPA_ONNX_ENABLE_DIRECTML "Enable ONNX Runtime DirectML support" OFF)
 option(SHERPA_ONNX_ENABLE_WASM "Whether to enable WASM" OFF)
+option(SHERPA_ONNX_ENABLE_WASM_SPEAKER_DIARIZATION "Whether to enable WASM for speaker diarization" OFF)
 option(SHERPA_ONNX_ENABLE_WASM_TTS "Whether to enable WASM for TTS" OFF)
 option(SHERPA_ONNX_ENABLE_WASM_ASR "Whether to enable WASM for ASR" OFF)
 option(SHERPA_ONNX_ENABLE_WASM_KWS "Whether to enable WASM for KWS" OFF)
@@ -135,6 +136,7 @@ message(STATUS "SHERPA_ONNX_ENABLE_C_API ${SHERPA_ONNX_ENABLE_C_API}")
 message(STATUS "SHERPA_ONNX_ENABLE_WEBSOCKET ${SHERPA_ONNX_ENABLE_WEBSOCKET}")
 message(STATUS "SHERPA_ONNX_ENABLE_GPU ${SHERPA_ONNX_ENABLE_GPU}")
 message(STATUS "SHERPA_ONNX_ENABLE_WASM ${SHERPA_ONNX_ENABLE_WASM}")
+message(STATUS "SHERPA_ONNX_ENABLE_WASM_SPEAKER_DIARIZATION ${SHERPA_ONNX_ENABLE_WASM_SPEAKER_DIARIZATION}")
 message(STATUS "SHERPA_ONNX_ENABLE_WASM_TTS ${SHERPA_ONNX_ENABLE_WASM_TTS}")
 message(STATUS "SHERPA_ONNX_ENABLE_WASM_ASR ${SHERPA_ONNX_ENABLE_WASM_ASR}")
 message(STATUS "SHERPA_ONNX_ENABLE_WASM_KWS ${SHERPA_ONNX_ENABLE_WASM_KWS}")
@@ -196,9 +198,19 @@ else()
   add_definitions(-DSHERPA_ONNX_ENABLE_DIRECTML=0)
 endif()
 
+if(SHERPA_ONNX_ENABLE_WASM_SPEAKER_DIARIZATION)
+  if(NOT SHERPA_ONNX_ENABLE_SPEAKER_DIARIZATION)
+    message(FATAL_ERROR "Please set SHERPA_ONNX_ENABLE_SPEAKER_DIARIZATION to ON if you want to build WASM for speaker diarization")
+  endif()
+
+  if(NOT SHERPA_ONNX_ENABLE_WASM)
+    message(FATAL_ERROR "Please set SHERPA_ONNX_ENABLE_WASM to ON if you enable WASM for speaker diarization")
+  endif()
+endif()
+
 if(SHERPA_ONNX_ENABLE_WASM_TTS)
   if(NOT SHERPA_ONNX_ENABLE_TTS)
-    message(FATAL_ERROR "Please set SHERPA_ONNX_ENABLE_TTS to ON if you want to build wasm TTS")
+    message(FATAL_ERROR "Please set SHERPA_ONNX_ENABLE_TTS to ON if you want to build WASM for TTS")
   endif()
 
   if(NOT SHERPA_ONNX_ENABLE_WASM)

README.md

@@ -116,6 +116,7 @@ We also have spaces built using WebAssembly. They are listed below:
 |VAD + speech recognition (English + Chinese, 及多种中文方言) with Paraformer-small          |[Click me][wasm-hf-vad-asr-zh-en-paraformer-small]| [地址][wasm-ms-vad-asr-zh-en-paraformer-small]|
 |Speech synthesis (English)                                                                  |[Click me][wasm-hf-tts-piper-en]| [地址][wasm-ms-tts-piper-en]|
 |Speech synthesis (German)                                                                   |[Click me][wasm-hf-tts-piper-de]| [地址][wasm-ms-tts-piper-de]|
+|Speaker diarization                                                                         |[Click me][wasm-hf-speaker-diarization]|[地址][wasm-ms-speaker-diarization]|
 
 ### Links for pre-built Android APKs
 
@@ -173,6 +174,7 @@ We also have spaces built using WebAssembly. They are listed below:
 | Speaker identification (Speaker ID)         | [Address][sid-models]                                                                 |
 | Spoken language identification (Language ID)| See multi-lingual [Whisper][Whisper] ASR models from  [Speech recognition][asr-models]|
 | Punctuation                                 | [Address][punct-models]                                                               |
+| Speaker segmentation                        | [Address][speaker-segmentation-models]                                                |
 
 ### Useful links
 
@@ -261,6 +263,8 @@ Video demo in Chinese: [爆了！炫神教你开打字挂！真正影响胜率
 [wasm-ms-tts-piper-en]: https://modelscope.cn/studios/k2-fsa/web-assembly-tts-sherpa-onnx-en
 [wasm-hf-tts-piper-de]: https://huggingface.co/spaces/k2-fsa/web-assembly-tts-sherpa-onnx-de
 [wasm-ms-tts-piper-de]: https://modelscope.cn/studios/k2-fsa/web-assembly-tts-sherpa-onnx-de
+[wasm-hf-speaker-diarization]: https://huggingface.co/spaces/k2-fsa/web-assembly-speaker-diarization-sherpa-onnx
+[wasm-ms-speaker-diarization]: https://www.modelscope.cn/studios/csukuangfj/web-assembly-speaker-diarization-sherpa-onnx
 [apk-streaming-asr]: https://k2-fsa.github.io/sherpa/onnx/android/apk.html
 [apk-streaming-asr-cn]: https://k2-fsa.github.io/sherpa/onnx/android/apk-cn.html
 [apk-tts]: https://k2-fsa.github.io/sherpa/onnx/tts/apk-engine.html
@@ -303,5 +307,6 @@ Video demo in Chinese: [爆了！炫神教你开打字挂！真正影响胜率
 [sid-models]: https://github.com/k2-fsa/sherpa-onnx/releases/tag/speaker-recongition-models
 [slid-models]: https://github.com/k2-fsa/sherpa-onnx/releases/tag/speaker-recongition-models
 [punct-models]: https://github.com/k2-fsa/sherpa-onnx/releases/tag/punctuation-models
+[speaker-segmentation-models]: https://github.com/k2-fsa/sherpa-onnx/releases/tag/speaker-segmentation-models
 [GigaSpeech]: https://github.com/SpeechColab/GigaSpeech
 [WenetSpeech]: https://github.com/wenet-e2e/WenetSpeech

build-wasm-simd-asr.sh

@@ -14,8 +14,8 @@ if [ x"$EMSCRIPTEN" == x"" ]; then
     echo "git clone https://github.com/emscripten-core/emsdk.git"
     echo "cd emsdk"
     echo "git pull"
-    echo "./emsdk install latest"
-    echo "./emsdk activate latest"
+    echo "./emsdk install 3.1.53"
+    echo "./emsdk activate 3.1.53"
     echo "source ./emsdk_env.sh"
     exit 1
   else

build-wasm-simd-kws.sh

@@ -9,8 +9,8 @@ if [ x"$EMSCRIPTEN" == x"" ]; then
     echo "git clone https://github.com/emscripten-core/emsdk.git"
     echo "cd emsdk"
     echo "git pull"
-    echo "./emsdk install latest"
-    echo "./emsdk activate latest"
+    echo "./emsdk install 3.1.53"
+    echo "./emsdk activate 3.1.53"
     echo "source ./emsdk_env.sh"
     exit 1
   else

build-wasm-simd-nodejs.sh

@@ -16,8 +16,8 @@ if [ x"$EMSCRIPTEN" == x"" ]; then
     echo "git clone https://github.com/emscripten-core/emsdk.git"
     echo "cd emsdk"
     echo "git pull"
-    echo "./emsdk install latest"
-    echo "./emsdk activate latest"
+    echo "./emsdk install 3.1.53"
+    echo "./emsdk activate 3.1.53"
     echo "source ./emsdk_env.sh"
     exit 1
   else