first working version

k2-fsa · Oct 13, 2023 · 65d338b · 65d338b
1 parent 222580c
commit 65d338b
Show file tree

Hide file tree

Showing 3 changed files with 71 additions and 2 deletions.
diff --git a/sherpa-onnx/csrc/offline-tts-vits-impl.h b/sherpa-onnx/csrc/offline-tts-vits-impl.h
@@ -4,6 +4,7 @@
 #ifndef SHERPA_ONNX_CSRC_OFFLINE_TTS_VITS_IMPL_H_
 #define SHERPA_ONNX_CSRC_OFFLINE_TTS_VITS_IMPL_H_
 
+#include <fstream>
 #include <string>
 
 #include "sherpa-onnx/csrc/macros.h"
@@ -19,7 +20,36 @@ class OfflineTtsVitsImpl : public OfflineTtsImpl {
     SHERPA_ONNX_LOGE("config: %s\n", config.ToString().c_str());
   }
 
-  GeneratedAudio Generate(const std::string &text) const override { return {}; }
+  GeneratedAudio Generate(const std::string &text) const override {
+    SHERPA_ONNX_LOGE("txt: %s", text.c_str());
+    auto memory_info =
+        Ort::MemoryInfo::CreateCpu(OrtDeviceAllocator, OrtMemTypeDefault);
+
+    std::vector<int64_t> x = {
+        0, 54,  0, 157, 0, 102, 0, 54, 0, 51,  0, 158, 0, 156, 0, 72,
+        0, 56,  0, 83,  0, 3,   0, 16, 0, 157, 0, 43,  0, 135, 0, 85,
+        0, 16,  0, 55,  0, 156, 0, 57, 0, 135, 0, 61,  0, 62,  0, 16,
+        0, 44,  0, 52,  0, 156, 0, 63, 0, 158, 0, 125, 0, 102, 0, 48,
+        0, 83,  0, 54,  0, 16,  0, 72, 0, 56,  0, 46,  0, 16,  0, 54,
+        0, 156, 0, 138, 0, 64,  0, 54, 0, 51,  0, 16,  0, 70,  0, 61,
+        0, 156, 0, 102, 0, 61,  0, 62, 0, 83,  0, 56,  0, 62,  0};
+    std::array<int64_t, 2> x_shape = {1, static_cast<int32_t>(x.size())};
+    Ort::Value x_tensor = Ort::Value::CreateTensor(
+        memory_info, x.data(), x.size(), x_shape.data(), x_shape.size());
+    Ort::Value audio = model_->Run(std::move(x_tensor));
+
+    std::vector<int64_t> audio_shape =
+        audio.GetTensorTypeAndShapeInfo().GetShape();
+
+    SHERPA_ONNX_LOGE("%d, %d", int(audio_shape.size()), int(audio_shape[2]));
+    const float *p = audio.GetTensorData<float>();
+    std::ofstream os("t.pcm", std::ios::binary);
+    os.write(reinterpret_cast<const char *>(p), sizeof(float) * audio_shape[2]);
+
+    // sox -t raw -r 22050 -b 32 -e floating-point -c 1 ./t.pcm ./t.wav
+
+    return {};
+  }
 
  private:
   std::unique_ptr<OfflineTtsVitsModel> model_;

diff --git a/sherpa-onnx/csrc/offline-tts-vits-model.cc b/sherpa-onnx/csrc/offline-tts-vits-model.cc
@@ -25,7 +25,45 @@ class OfflineTtsVitsModel::Impl {
     Init(buf.data(), buf.size());
   }
 
-  Ort::Value Run(Ort::Value x) { return Ort::Value(nullptr); }
+  Ort::Value Run(Ort::Value x) {
+    auto memory_info =
+        Ort::MemoryInfo::CreateCpu(OrtDeviceAllocator, OrtMemTypeDefault);
+
+    std::vector<int64_t> x_shape = x.GetTensorTypeAndShapeInfo().GetShape();
+    if (x_shape[0] != 1) {
+      SHERPA_ONNX_LOGE("Support only batch_size == 1. Given: %d",
+                       static_cast<int32_t>(x_shape[0]));
+      exit(-1);
+    }
+
+    int64_t len = x_shape[1];
+    int64_t len_shape = 1;
+
+    Ort::Value x_length =
+        Ort::Value::CreateTensor(memory_info, &len, 1, &len_shape, 1);
+
+    int64_t scale_shape = 1;
+    float noise_scale = 1;
+    float length_scale = 1;
+    float noise_scale_w = 1;
+
+    Ort::Value noise_scale_tensor =
+        Ort::Value::CreateTensor(memory_info, &noise_scale, 1, &scale_shape, 1);
+    Ort::Value length_scale_tensor = Ort::Value::CreateTensor(
+        memory_info, &length_scale, 1, &scale_shape, 1);
+    Ort::Value noise_scale_w_tensor = Ort::Value::CreateTensor(
+        memory_info, &noise_scale_w, 1, &scale_shape, 1);
+
+    std::array<Ort::Value, 5> inputs = {
+        std::move(x), std::move(x_length), std::move(noise_scale_tensor),
+        std::move(length_scale_tensor), std::move(noise_scale_w_tensor)};
+
+    auto out =
+        sess_->Run({}, input_names_ptr_.data(), inputs.data(), inputs.size(),
+                   output_names_ptr_.data(), output_names_ptr_.size());
+
+    return std::move(out[0]);
+  }
 
   int32_t SampleRate() const { return sample_rate_; }
 

diff --git a/sherpa-onnx/csrc/sherpa-onnx-offline-tts.cc b/sherpa-onnx/csrc/sherpa-onnx-offline-tts.cc
@@ -35,6 +35,7 @@ It will generate a file test.wav.
   }
 
   sherpa_onnx::OfflineTts tts(config);
+  tts.Generate("hello world\n");
 
   return 0;
 }