Minor fixes

k2-fsa · Jan 8, 2024 · 0d36d48 · 0d36d48
1 parent 28c5b0c
commit 0d36d48
Show file tree

Hide file tree

Showing 3 changed files with 106 additions and 62 deletions.
diff --git a/android/SherpaOnnxKws/app/src/main/java/com/k2fsa/sherpa/onnx/SherpaOnnx.kt b/android/SherpaOnnxKws/app/src/main/java/com/k2fsa/sherpa/onnx/SherpaOnnx.kt
@@ -117,33 +117,11 @@ to add your own. (It should be straightforward to add a new model
 by following the code)
 
 @param type
-0 - sherpa-onnx-streaming-zipformer-bilingual-zh-en-2023-02-20 (Bilingual, Chinese + English)
-    https://k2-fsa.github.io/sherpa/onnx/pretrained_models/zipformer-transducer-models.html#sherpa-onnx-streaming-zipformer-bilingual-zh-en-2023-02-20-bilingual-chinese-english
+0 - sherpa-onnx-kws-zipformer-wenetspeech-3.3M-2024-01-01 (Chinese)
+    https://www.modelscope.cn/models/pkufool/sherpa-onnx-kws-zipformer-wenetspeech-3.3M-2024-01-01/summary
 
-1 - csukuangfj/sherpa-onnx-lstm-zh-2023-02-20 (Chinese)
-
-    https://k2-fsa.github.io/sherpa/onnx/pretrained_models/lstm-transducer-models.html#csukuangfj-sherpa-onnx-lstm-zh-2023-02-20-chinese
-
-2 - csukuangfj/sherpa-onnx-lstm-en-2023-02-17 (English)
-    https://k2-fsa.github.io/sherpa/onnx/pretrained_models/lstm-transducer-models.html#csukuangfj-sherpa-onnx-lstm-en-2023-02-17-english
-
-3,4 - pkufool/icefall-asr-zipformer-streaming-wenetspeech-20230615
-    https://huggingface.co/pkufool/icefall-asr-zipformer-streaming-wenetspeech-20230615
-    3 - int8 encoder
-    4 - float32 encoder
-
-5 - csukuangfj/sherpa-onnx-streaming-paraformer-bilingual-zh-en
-    https://huggingface.co/csukuangfj/sherpa-onnx-streaming-paraformer-bilingual-zh-en
-
-6 - sherpa-onnx-streaming-zipformer-en-2023-06-26
-    https://huggingface.co/csukuangfj/sherpa-onnx-streaming-zipformer-en-2023-06-26
-
-7 - shaojieli/sherpa-onnx-streaming-zipformer-fr-2023-04-14 (French)
-    https://huggingface.co/shaojieli/sherpa-onnx-streaming-zipformer-fr-2023-04-14
-
-8 - csukuangfj/sherpa-onnx-streaming-zipformer-bilingual-zh-en-2023-02-20 (Bilingual, Chinese + English)
-    https://huggingface.co/csukuangfj/sherpa-onnx-streaming-zipformer-bilingual-zh-en-2023-02-20
-    encoder int8, decoder/joiner float32
+1 - sherpa-onnx-kws-zipformer-gigaspeech-3.3M-2024-01-01 (English)
+    https://www.modelscope.cn/models/pkufool/sherpa-onnx-kws-zipformer-gigaspeech-3.3M-2024-01-01/summary
 
  */
 fun getModelConfig(type: Int): OnlineModelConfig? {
@@ -162,14 +140,14 @@ fun getModelConfig(type: Int): OnlineModelConfig? {
         }
 
         1 -> {
-            val modelDir = "icefall-asr-zipformer-streaming-wenetspeech-20230615"
+            val modelDir = "sherpa-onnx-kws-zipformer-wenetspeech-3.3M-2024-01-01"
             return OnlineModelConfig(
                 transducer = OnlineTransducerModelConfig(
-                    encoder = "$modelDir/exp/encoder-epoch-12-avg-4-chunk-16-left-128.onnx",
-                    decoder = "$modelDir/exp/decoder-epoch-12-avg-4-chunk-16-left-128.onnx",
-                    joiner = "$modelDir/exp/joiner-epoch-12-avg-4-chunk-16-left-128.onnx",
+                    encoder = "$modelDir/encoder-epoch-12-avg-2-chunk-16-left-64.onnx",
+                    decoder = "$modelDir/decoder-epoch-12-avg-2-chunk-16-left-64.onnx",
+                    joiner = "$modelDir/joiner-epoch-12-avg-2-chunk-16-left-64.onnx",
                 ),
-                tokens = "$modelDir/data/lang_char/tokens.txt",
+                tokens = "$modelDir/tokens.txt",
                 modelType = "zipformer2",
             )
         }

diff --git a/build-kws-apk.sh b/build-kws-apk.sh
@@ -21,44 +21,103 @@ log "Building keyword spotting APK for sherpa-onnx v${SHERPA_ONNX_VERSION}"
 log "====================arm64-v8a================="
 ./build-android-arm64-v8a.sh
 log "====================armv7-eabi================"
-#./build-android-armv7-eabi.sh
+./build-android-armv7-eabi.sh
 log "====================x86-64===================="
-#./build-android-x86-64.sh
+./build-android-x86-64.sh
 log "====================x86===================="
-#./build-android-x86.sh
-
+./build-android-x86.sh
 
 mkdir -p apks
 
-log "https://www.modelscope.cn/pkufool/sherpa-onnx-kws-zipformer-wenetspeech-3.3M-2024-01-01.git"
-
 # Download the model
-repo_url=https://www.modelscope.cn/pkufool/sherpa-onnx-kws-zipformer-wenetspeech-3.3M-2024-01-01.git
-#log "Start testing ${repo_url}"
-##repo=$(basename $repo_url)
-#repo=sherpa-onnx-kws-zipformer-wenetspeech-3.3M-2024-01-01
-#log "Download pretrained model and test-data from $repo_url"
-#GIT_LFS_SKIP_SMUDGE=1 git clone $repo_url
-#pushd $repo
-#git lfs pull --include "*.onnx"
-#
-## remove .git to save spaces
-#rm -rf .git
-#rm README.md
-#rm -rfv test_wavs
-#ls -lh
-#popd
-#
-#mv -v $repo ./android/SherpaOnnxKws/app/src/main/assets/
+repo=sherpa-onnx-kws-zipformer-wenetspeech-3.3M-2024-01-01
+
+if [ ! -d $repo ]; then
+  repo_url=https://www.modelscope.cn/pkufool/sherpa-onnx-kws-zipformer-wenetspeech-3.3M-2024-01-01.git
+  log "Start testing ${repo_url}"
+  log "Download pretrained model and test-data from $repo_url"
+  GIT_LFS_SKIP_SMUDGE=1 git clone $repo_url
+  pushd $repo
+  git lfs pull --include "*.onnx"
+
+  # remove .git to save spaces
+  rm -rf .git
+  rm README.md configuration.json .gitattribute
+  rm -rfv test_wavs
+  ls -lh
+  popd
+
+  mv -v $repo ./android/SherpaOnnxKws/app/src/main/assets/
+fi
+
+pushd ./android/SherpaOnnxKws/app/src/main/assets/
+cp $repo/keywords.txt .
+popd
+
+tree ./android/SherpaOnnxKws/app/src/main/assets/
+
+for arch in arm64-v8a armeabi-v7a x86_64 x86; do
+  log "------------------------------------------------------------"
+  log "build apk for $arch"
+  log "------------------------------------------------------------"
+  src_arch=$arch
+  if [ $arch == "armeabi-v7a" ]; then
+    src_arch=armv7-eabi
+  elif [ $arch == "x86_64" ]; then
+    src_arch=x86-64
+  fi
+
+  ls -lh ./build-android-$src_arch/install/lib/*.so
+
+  cp -v ./build-android-$src_arch/install/lib/*.so ./android/SherpaOnnxKws/app/src/main/jniLibs/$arch/
+
+  pushd ./android/SherpaOnnxKws
+  ./gradlew build
+  popd
+
+  mv android/SherpaOnnxKws/app/build/outputs/apk/debug/app-debug.apk ./apks/sherpa-onnx-kws-${SHERPA_ONNX_VERSION}-$arch.apk
+  ls -lh apks
+  rm -v ./android/SherpaOnnxKws/app/src/main/jniLibs/$arch/*.so
+done
+
+git checkout .
+
+rm -rf ./android/SherpaOnnxKws/app/src/main/assets/$repo
+
+# English model
+repo=sherpa-onnx-kws-zipformer-gigaspeech-3.3M-2024-01-01
+
+if [ ! -d $repo ]; then
+  repo_url=https://www.modelscope.cn/pkufool/sherpa-onnx-kws-zipformer-gigaspeech-3.3M-2024-01-01.git
+  log "Start testing ${repo_url}"
+  log "Download pretrained model and test-data from $repo_url"
+  GIT_LFS_SKIP_SMUDGE=1 git clone $repo_url
+  pushd $repo
+  git lfs pull --include "*.onnx"
+
+  # remove .git to save spaces
+  rm -rf .git
+  rm *.int8.onnx
+  rm README.md configuration.json .gitattribute
+  rm -rfv test_wavs
+  ls -lh
+  popd
+
+  mv -v $repo ./android/SherpaOnnxKws/app/src/main/assets/
+fi
+
+pushd ./android/SherpaOnnxKws/app/src/main/assets/
+cp $repo/keywords.txt .
+popd
+
 tree ./android/SherpaOnnxKws/app/src/main/assets/
 
 pushd android/SherpaOnnxKws/app/src/main/java/com/k2fsa/sherpa/onnx
-#sed -i.bak s/"type = 0"/"type = 6"/ ./MainActivity.kt
-#git diff
+sed -i.bak s/"type = 0"/"type = 1"/ ./MainActivity.kt
+git diff
 popd
 
-#for arch in arm64-v8a armeabi-v7a x86_64 x86; do
-for arch in arm64-v8a; do
+for arch in arm64-v8a armeabi-v7a x86_64 x86; do
   log "------------------------------------------------------------"
   log "build apk for $arch"
   log "------------------------------------------------------------"
@@ -77,11 +136,11 @@ for arch in arm64-v8a; do
   ./gradlew build
   popd
 
-  mv android/SherpaOnnxKws/app/build/outputs/apk/debug/app-debug.apk ./apks/sherpa-onnx-${SHERPA_ONNX_VERSION}-$arch.apk
+  mv android/SherpaOnnxKws/app/build/outputs/apk/debug/app-debug.apk ./apks/sherpa-onnx-kws-${SHERPA_ONNX_VERSION}-$arch-en.apk
   ls -lh apks
-  #rm -v ./android/SherpaOnnxKws/app/src/main/jniLibs/$arch/*.so
+  rm -v ./android/SherpaOnnxKws/app/src/main/jniLibs/$arch/*.so
 done
 
-#git checkout .
+git checkout .
 
-#rm -rf ./android/SherpaOnnxKws/app/src/main/assets/$repo
+rm -rf ./android/SherpaOnnxKws/app/src/main/assets/$repo
diff --git a/sherpa-onnx/csrc/keyword-spotter-transducer-impl.h b/sherpa-onnx/csrc/keyword-spotter-transducer-impl.h
@@ -38,11 +38,18 @@ static KeywordResult Convert(const TransducerKeywordsResult &src,
   r.tokens.reserve(src.tokens.size());
   r.timestamps.reserve(src.tokens.size());
   r.keyword = src.keyword;
+  bool from_tokens = src.keyword.size() == 0;
 
   for (auto i : src.tokens) {
     auto sym = sym_table[i];
+    if (from_tokens) {
+      r.keyword.append(sym);
+    }
     r.tokens.push_back(std::move(sym));
   }
+  if (from_tokens && r.keyword.size()) {
+    r.keyword = r.keyword.substr(1);
+  }
 
   float frame_shift_s = frame_shift_ms / 1000. * subsampling_factor;
   for (auto t : src.timestamps) {