Add Go API for TTS (#377)

go-api-examples/non-streaming-tts/go.mod

@@ -0,0 +1,3 @@
+module non-streaming-tts
+
+go 1.12

go-api-examples/non-streaming-tts/main.go

@@ -0,0 +1,61 @@
+package main
+
+import (
+	sherpa "github.com/k2-fsa/sherpa-onnx-go/sherpa_onnx"
+	flag "github.com/spf13/pflag"
+	"log"
+)
+
+func main() {
+	log.SetFlags(log.LstdFlags | log.Lmicroseconds)
+
+	config := sherpa.OfflineTtsConfig{}
+	sid := 0
+	filename := "./generated.wav"
+
+	flag.StringVar(&config.Model.Vits.Model, "vits-model", "", "Path to the vits ONNX model")
+	flag.StringVar(&config.Model.Vits.Lexicon, "vits-lexicon", "", "Path to lexicon.txt")
+	flag.StringVar(&config.Model.Vits.Tokens, "vits-tokens", "", "Path to tokens.txt")
+
+	flag.Float32Var(&config.Model.Vits.NoiseScale, "vits-noise-scale", 0.667, "noise_scale for VITS")
+	flag.Float32Var(&config.Model.Vits.NoiseScaleW, "vits-noise-scale-w", 0.8, "noise_scale_w for VITS")
+	flag.Float32Var(&config.Model.Vits.LengthScale, "vits-length-scale", 1.0, "length_scale for VITS. small -> faster in speech speed; large -> slower")
+
+	flag.IntVar(&config.Model.NumThreads, "num-threads", 1, "Number of threads for computing")
+	flag.IntVar(&config.Model.Debug, "debug", 0, "Whether to show debug message")
+	flag.StringVar(&config.Model.Provider, "provider", "cpu", "Provider to use")
+
+	flag.IntVar(&sid, "sid", 0, "Speaker ID. Used only for multi-speaker models")
+	flag.StringVar(&filename, "output-filename", "./generated.wav", "Filename to save the generated audio")
+
+	flag.Parse()
+
+	if len(flag.Args()) != 1 {
+		log.Fatalf("Please provide the text to generate audios")
+	}
+
+	text := flag.Arg(0)
+
+	log.Println("Input text:", text)
+	log.Println("Speaker ID:", sid)
+	log.Println("Output filename:", filename)
+
+	log.Println("Initializing model (may take several seconds)")
+
+	tts := sherpa.NewOfflineTts(&config)
+	defer sherpa.DeleteOfflineTts(tts)
+
+	log.Println("Model created!")
+
+	log.Println("Start generating!")
+
+	audio := tts.Generate(text, sid)
+
+	log.Println("Done!")
+
+	ok := audio.Save(filename)
+	if ok != 1 {
+		log.Fatalf("Failed to write", filename)
+	}
+
+}

go-api-examples/non-streaming-tts/run-vits-ljs.sh

@@ -0,0 +1,14 @@
+#!/usr/bin/env bash
+
+# please refer to
+# https://k2-fsa.github.io/sherpa/onnx/tts/pretrained_models/vits.html#ljspeech-english-single-speaker
+# to download the model before you run this script
+
+./non-streaming-tts \
+  --vits-model=./vits-ljs/vits-ljs.onnx \
+  --vits-lexicon=./vits-ljs/lexicon.txt \
+  --vits-tokens=./vits-ljs/tokens.txt \
+  --sid=0 \
+  --debug=1 \
+  --output-filename=./vits-ljs.wav \
+  "Liliana, the most beautiful and lovely assistant of our team!"

go-api-examples/non-streaming-tts/run-vits-vctk.sh

@@ -0,0 +1,16 @@
+#!/usr/bin/env bash
+
+# please refer to
+# https://k2-fsa.github.io/sherpa/onnx/tts/pretrained_models/vits.html#vctk-english-multi-speaker-109-speakers
+# to download the model before you run this script
+
+for sid in 0 10 108; do
+./non-streaming-tts \
+  --vits-model=./vits-vctk/vits-vctk.onnx \
+  --vits-lexicon=./vits-vctk/lexicon.txt \
+  --vits-tokens=./vits-vctk/tokens.txt \
+  --sid=0 \
+  --debug=1 \
+  --output-filename=./kennedy-$sid.wav \
+  'Ask not what your country can do for you; ask what you can do for your country.'
+done

go-api-examples/non-streaming-tts/run-vits-zh-aishell3.sh

@@ -0,0 +1,16 @@
+#!/usr/bin/env bash
+
+# please refer to
+# https://k2-fsa.github.io/sherpa/onnx/tts/pretrained_models/vits.html#aishell3-chinese-multi-speaker-174-speakers
+# to download the model before you run this script
+
+for sid in 10 33 99; do
+./non-streaming-tts \
+  --vits-model=./vits-zh-aishell3/vits-aishell3.onnx \
+  --vits-lexicon=./vits-zh-aishell3/lexicon.txt \
+  --vits-tokens=./vits-zh-aishell3/tokens.txt \
+  --sid=10 \
+  --debug=1 \
+  --output-filename=./liliana-$sid.wav \
+  "林美丽最美丽、最漂亮、最可爱！"
+done