add speed parameter

rust/src/example_onnx.rs

@@ -26,6 +26,10 @@ struct Args {
     #[arg(long, default_value = "5")]
     total_step: usize,
 
+    /// Speech speed factor (higher = faster)
+    #[arg(long, default_value = "1.05")]
+    speed: f32,
+
     /// Number of times to generate
     #[arg(long, default_value = "4")]
     n_test: usize,
@@ -53,6 +57,7 @@ fn main() -> Result<()> {
     // --- 1. Parse arguments --- //
     let args = Args::parse();
     let total_step = args.total_step;
+    let speed = args.speed;
     let n_test = args.n_test;
     let voice_style_paths = &args.voice_style;
     let text_list = &args.text;
@@ -85,11 +90,11 @@ fn main() -> Result<()> {
 
         let (wav, duration) = if batch {
             timer("Generating speech from text", || {
-                text_to_speech.batch(text_list, &style, total_step)
+                text_to_speech.batch(text_list, &style, total_step, speed)
             })?
         } else {
             let (w, d) = timer("Generating speech from text", || {
-                text_to_speech.call(&text_list[0], &style, total_step, 0.3)
+                text_to_speech.call(&text_list[0], &style, total_step, speed, 0.3)
             })?;
             (w, vec![d])
         };

rust/src/helper.rs

@@ -484,6 +484,7 @@ impl TextToSpeech {
         text_list: &[String],
         style: &Style,
         total_step: usize,
+        speed: f32,
     ) -> Result<(Vec<f32>, Vec<f32>)> {
         let bsz = text_list.len();
 
@@ -511,7 +512,12 @@ impl TextToSpeech {
         })?;
 
         let (_, duration_data) = dp_outputs["duration"].try_extract_tensor::<f32>()?;
-        let duration: Vec<f32> = duration_data.to_vec();
+        let mut duration: Vec<f32> = duration_data.to_vec();
+        
+        // Apply speed factor to duration
+        for dur in duration.iter_mut() {
+            *dur /= speed;
+        }
 
         // Encode text
         let style_ttl_value = Value::from_array(style.ttl.clone())?;
@@ -584,6 +590,7 @@ impl TextToSpeech {
         text: &str,
         style: &Style,
         total_step: usize,
+        speed: f32,
         silence_duration: f32,
     ) -> Result<(Vec<f32>, f32)> {
         let chunks = chunk_text(text, None);
@@ -592,7 +599,7 @@ impl TextToSpeech {
         let mut dur_cat: f32 = 0.0;
 
         for (i, chunk) in chunks.iter().enumerate() {
-            let (wav, duration) = self._infer(&[chunk.clone()], style, total_step)?;
+            let (wav, duration) = self._infer(&[chunk.clone()], style, total_step, speed)?;
             
             let dur = duration[0];
             let wav_len = (self.sample_rate as f32 * dur) as usize;
@@ -619,8 +626,9 @@ impl TextToSpeech {
         text_list: &[String],
         style: &Style,
         total_step: usize,
+        speed: f32,
     ) -> Result<(Vec<f32>, Vec<f32>)> {
-        self._infer(text_list, style, total_step)
+        self._infer(text_list, style, total_step, speed)
     }
 }