Support spoken language identification with whisper (#694)

.github/workflows/build-wheels-linux.yaml

@@ -82,7 +82,6 @@ jobs:
         env:
           HF_TOKEN: ${{ secrets.HF_TOKEN }}
         uses: nick-fields/retry@v3
-        shell: bash
         with:
           max_attempts: 20
           timeout_seconds: 200

.github/workflows/build-wheels-macos-arm64.yaml

@@ -21,27 +21,12 @@ jobs:
       fail-fast: false
       matrix:
         os: [macos-latest]
-        python-version: ["cp37", "cp38", "cp39", "cp310", "cp311", "cp312"]
+        python-version: ["cp38", "cp39", "cp310", "cp311", "cp312"]
 
     steps:
       - uses: actions/checkout@v4
 
-      # see https://cibuildwheel.readthedocs.io/en/stable/changelog/
-      # for a list of versions
       - name: Build wheels
-        if: matrix.python-version == 'cp37'
-        uses: pypa/cibuildwheel@v2.11.4
-        env:
-          CIBW_BUILD: "${{ matrix.python-version}}-* "
-          CIBW_ENVIRONMENT: SHERPA_ONNX_CMAKE_ARGS="-DCMAKE_OSX_ARCHITECTURES='arm64'"
-          CIBW_ARCHS: "arm64"
-          CIBW_BUILD_VERBOSITY: 3
-
-          #  Don't repair macOS wheels
-          CIBW_REPAIR_WHEEL_COMMAND_MACOS: ""
-
-      - name: Build wheels
-        if: matrix.python-version != 'cp37'
         uses: pypa/cibuildwheel@v2.15.0
         env:
           CIBW_BUILD: "${{ matrix.python-version}}-* "

.github/workflows/linux-gpu.yaml

@@ -92,6 +92,14 @@ jobs:
           file build/bin/sherpa-onnx
           readelf -d build/bin/sherpa-onnx
 
+      - name: Test spoken language identification
+        shell: bash
+        run: |
+          export PATH=$PWD/build/bin:$PATH
+          export EXE=sherpa-onnx-offline-language-identification
+
+          .github/scripts/test-spoken-language-identification.sh
+
       - name: Test online CTC
         shell: bash
         run: |
@@ -116,6 +124,7 @@ jobs:
 
           .github/scripts/test-online-paraformer.sh
 
+
       - name: Test offline Whisper
         shell: bash
         run: |

.github/workflows/linux.yaml

@@ -123,6 +123,15 @@ jobs:
           name: release-${{ matrix.build_type }}-${{ matrix.shared_lib }}
           path: build/bin/*
 
+      - name: Test spoken language identification
+        if: matrix.build_type != 'Debug'
+        shell: bash
+        run: |
+          export PATH=$PWD/build/bin:$PATH
+          export EXE=sherpa-onnx-offline-language-identification
+
+          .github/scripts/test-spoken-language-identification.sh
+
       - name: Test transducer kws
         shell: bash
         run: |
@@ -140,6 +149,7 @@ jobs:
           .github/scripts/test-online-ctc.sh
 
       - name: Test offline Whisper
+        if: matrix.build_type != 'Debug'
         shell: bash
         run: |
           export PATH=$PWD/build/bin:$PATH

.github/workflows/macos.yaml

@@ -102,6 +102,15 @@ jobs:
           otool -L build/bin/sherpa-onnx
           otool -l build/bin/sherpa-onnx
 
+      - name: Test spoken language identification
+        if: matrix.build_type != 'Debug'
+        shell: bash
+        run: |
+          export PATH=$PWD/build/bin:$PATH
+          export EXE=sherpa-onnx-offline-language-identification
+
+          .github/scripts/test-spoken-language-identification.sh
+
       - name: Test transducer kws
         shell: bash
         run: |
@@ -135,6 +144,7 @@ jobs:
           .github/scripts/test-online-paraformer.sh
 
       - name: Test offline Whisper
+        if: matrix.build_type != 'Debug'
         shell: bash
         run: |
           export PATH=$PWD/build/bin:$PATH

.github/workflows/windows-x64-cuda.yaml

@@ -68,6 +68,14 @@ jobs:
 
           ls -lh ./bin/Release/sherpa-onnx.exe
 
+      - name: Test spoken language identification
+        shell: bash
+        run: |
+          export PATH=$PWD/build/bin/Release:$PATH
+          export EXE=sherpa-onnx-offline-language-identification.exe
+
+          .github/scripts/test-spoken-language-identification.sh
+
       - name: Test online CTC
         shell: bash
         run: |

.github/workflows/windows-x64.yaml

@@ -68,6 +68,14 @@ jobs:
 
           ls -lh ./bin/Release/sherpa-onnx.exe
 
+      - name: Test spoken language identification
+        shell: bash
+        run: |
+          export PATH=$PWD/build/bin/Release:$PATH
+          export EXE=sherpa-onnx-offline-language-identification.exe
+
+          .github/scripts/test-spoken-language-identification.sh
+
       - name: Test online CTC
         shell: bash
         run: |

.github/workflows/windows-x86.yaml

@@ -69,6 +69,14 @@ jobs:
 
           ls -lh ./bin/Release/sherpa-onnx.exe
 
+      # - name: Test spoken language identification
+      #   shell: bash
+      #   run: |
+      #     export PATH=$PWD/build/bin/Release:$PATH
+      #     export EXE=sherpa-onnx-offline-language-identification.exe
+      #
+      #     .github/scripts/test-spoken-language-identification.sh
+
       - name: Test online CTC
         shell: bash
         run: |