first attempt

app.py

@@ -1,7 +1,40 @@
+import sys
+sys.path.append("neutts-air")
+from neuttsair.neutts import NeuTTSAir
 import gradio as gr
+import spaces
 
-def greet(name):
-    return "Hello " + name + "!!"
+# load model
+tts = NeuTTSAir(
+    backbone_repo=backbone,
+    backbone_device="gpu",
+    codec_repo="neuphonic/neucodec",
+    codec_device="gpu"
+)
 
-demo = gr.Interface(fn=greet, inputs="text", outputs="text")
-demo.launch()
+@spaces.GPU()
+def infer(ref_text, ref_audio_path, gen_text):
+
+    gr.Info("Starting inference request!")
+    gr.Info("Encoding reference...")
+    ref_codes = tts.encode_reference(ref_audio_path)
+
+    gr.Info(f"Generating audio for input text: {input_text}")
+    wav = tts.infer(input_text, ref_codes, ref_text)
+
+    return (24_000, wav)
+
+demo = gr.Interface(
+    fn=infer,
+    inputs=[
+        gr.Textbox(label="Reference Text"),
+        gr.Audio(source="upload", type="filepath", label="Reference Audio"),
+        gr.Textbox(label="Text to Generate"),
+    ],
+    outputs=gr.Audio(type="numpy", label="Generated Speech"),
+    title="NeuTTS-Air",
+    description="Upload a reference audio sample, provide the reference text, and enter new text to synthesize."
+)
+
+if __name__ == "__main__":
+    demo.launch()

neutts-air

@@ -0,0 +1 @@
+Subproject commit ededc7d354b05cb6d245c2a8563e04c5f8ac12a2

packages.txt

@@ -0,0 +1 @@
+espeak

requirements.txt

@@ -0,0 +1,19 @@
+datasets==4.0.0
+librosa==0.11.0
+neucodec>=0.0.3
+numpy==2.2.6
+pandas==2.3.2
+phonemizer==3.3.0
+requests==2.32.5
+scipy>=1.15
+soundfile==0.13.1
+torch==2.8.0
+torchao==0.13.0
+torchaudio==2.8.0
+torchtune==0.6.1
+tqdm==4.67.1
+transformers==4.56.1
+vector-quantize-pytorch==1.17.8
+resemble-perth==1.0.1
+accelerate==1.10.1
+gradio