Add C# API for SenseVoice models

k2-fsa · Jul 20, 2024 · ef9f074 · ef9f074
1 parent 25f0a10
commit ef9f074
Show file tree

Hide file tree

Showing 5 changed files with 64 additions and 13 deletions.
diff --git a/.github/scripts/test-dot-net.sh b/.github/scripts/test-dot-net.sh
@@ -2,16 +2,8 @@
 
 cd dotnet-examples/
 
-cd ./keyword-spotting-from-files
-./run.sh
-
-cd ../online-decode-files
-./run-transducer-itn.sh
-./run-zipformer2-ctc.sh
-./run-transducer.sh
-./run-paraformer.sh
-
-cd ../offline-decode-files
+cd ./offline-decode-files
+./run-sense-voice-ctc.sh
 ./run-paraformer-itn.sh
 ./run-telespeech-ctc.sh
 ./run-nemo-ctc.sh
@@ -21,6 +13,15 @@ cd ../offline-decode-files
 ./run-whisper.sh
 ./run-tdnn-yesno.sh
 
+cd ../keyword-spotting-from-files
+./run.sh
+
+cd ../online-decode-files
+./run-transducer-itn.sh
+./run-zipformer2-ctc.sh
+./run-transducer.sh
+./run-paraformer.sh
+
 cd ../vad-non-streaming-asr-paraformer
 ./run.sh
 

diff --git a/dotnet-examples/offline-decode-files/Program.cs b/dotnet-examples/offline-decode-files/Program.cs
@@ -61,6 +61,12 @@ class Options
     [Option("telespeech-ctc", Required = false, HelpText = "Path to model.onnx. Used only for TeleSpeech CTC models")]
     public string TeleSpeechCtc { get; set; } = "";
 
+    [Option("sense-voice-model", Required = false, HelpText = "Path to model.onnx. Used only for SenseVoice CTC models")]
+    public string SenseVoiceModel { get; set; } = "";
+
+    [Option("sense-voice-use-itn", Required = false, HelpText = "1 to use inverse text normalization for sense voice.")]
+    public int SenseVoiceUseItn { get; set; } = 1;
+
     [Option("num-threads", Required = false, Default = 1, HelpText = "Number of threads for computation")]
     public int NumThreads { get; set; } = 1;
 
@@ -225,6 +231,11 @@ private static void Run(Options options)
     {
       config.ModelConfig.Tdnn.Model = options.TdnnModel;
     }
+    else if (!String.IsNullOrEmpty(options.SenseVoiceModel))
+    {
+      config.ModelConfig.SenseVoice.Model = options.SenseVoiceModel;
+      config.ModelConfig.SenseVoice.UseInverseTextNormalization = options.SenseVoiceUseItn;
+    }
     else
     {
       Console.WriteLine("Please provide a model");

diff --git a/dotnet-examples/offline-decode-files/run-sense-voice-ctc.sh b/dotnet-examples/offline-decode-files/run-sense-voice-ctc.sh
@@ -0,0 +1,14 @@
+#!/usr/bin/env bash
+
+set -ex
+
+if [ ! -d ./sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17 ]; then
+  curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17.tar.bz2
+  tar xvf sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17.tar.bz2
+  rm sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17.tar.bz2
+fi
+
+dotnet run \
+  --sense-voice-model=./sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17/model.int8.onnx \
+  --tokens=./sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17/tokens.txt \
+  --files ./sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17/test_wavs/zh.wav
diff --git a/scripts/dotnet/OfflineModelConfig.cs b/scripts/dotnet/OfflineModelConfig.cs
@@ -23,6 +23,7 @@ public OfflineModelConfig()
             ModelingUnit = "cjkchar";
             BpeVocab = "";
             TeleSpeechCtc = "";
+            SenseVoice = new OfflineSenseVoiceModelConfig();
         }
         public OfflineTransducerModelConfig Transducer;
         public OfflineParaformerModelConfig Paraformer;
@@ -51,7 +52,7 @@ public OfflineModelConfig()
 
         [MarshalAs(UnmanagedType.LPStr)]
         public string TeleSpeechCtc;
-    }
-
 
-}
+        public OfflineSenseVoiceModelConfig SenseVoice;
+    }
+}
diff --git a/scripts/dotnet/OfflineSenseVoiceModelConfig.cs b/scripts/dotnet/OfflineSenseVoiceModelConfig.cs
@@ -0,0 +1,24 @@
+/// Copyright (c)  2024  Xiaomi Corporation (authors: Fangjun Kuang)
+
+using System.Runtime.InteropServices;
+
+namespace SherpaOnnx
+{
+    [StructLayout(LayoutKind.Sequential)]
+    public struct OfflineSenseVoiceModelConfig
+    {
+        public OfflineSenseVoiceModelConfig()
+        {
+            Model = "";
+            Language = "";
+            UseInverseTextNormalization = 0;
+        }
+        [MarshalAs(UnmanagedType.LPStr)]
+        public string Model;
+
+        [MarshalAs(UnmanagedType.LPStr)]
+        public string Language;
+
+        public int UseInverseTextNormalization;
+    }
+}