Update app.py

app.py

@@ -1,25 +1,26 @@
 import os
-from datasets import load_dataset
 from TTS.api import TTS
+from datasets import load_dataset, Audio
 import gradio as gr
 
-# Auto-accept Coqui license
+# Automatically accept Coqui license
 os.environ["COQUI_TOS_AGREED"] = "1"
 
-# Load dataset
+# Load dataset from Hugging Face
 dataset = load_dataset("Emmylahot12/nnamdi", split="train")
+dataset = dataset.cast_column("audio", Audio())
 
-# Check if dataset is valid
-if dataset is None or len(dataset) == 0 or "audio" not in dataset[0] or dataset[0]["audio"] is None:
+# Validate audio exists
+if not dataset or not dataset[0]["audio"] or not dataset[0]["audio"]["path"]:
     raise ValueError("Dataset is empty or audio is missing")
 
-# Get reference audio path
+# Reference voice sample path
 voice_sample_path = dataset[0]["audio"]["path"]
 
-# Initialize TTS
-tts = TTS("tts_models/multilingual/multi-dataset/xtts_v2", gpu=False)
+# Initialize TTS with XTTS v2 model (CPU mode)
+tts = TTS(model_name="tts_models/multilingual/multi-dataset/xtts_v2", gpu=False)
 
-# Synthesize speech
+# TTS synthesis function
 def synthesize(text, language="en"):
     output_path = "output.wav"
     tts.tts_to_file(
@@ -31,12 +32,16 @@ def synthesize(text, language="en"):
     return output_path
 
 # Gradio UI
-gr.Interface(
+interface = gr.Interface(
     fn=synthesize,
     inputs=[
-        gr.Textbox(label="Enter text"),
-        gr.Dropdown(choices=["en", "fr", "es"], label="Language", value="en")
+        gr.Textbox(label="Enter Text"),
+        gr.Dropdown(choices=["en", "fr", "es", "de"], label="Language")
     ],
-    outputs=gr.Audio(label="Generated Audio"),
-    title="Nnamdi TTS App"
-).launch()
+    outputs=gr.Audio(label="Generated Speech"),
+    title="Nnamdi TTS - Powered by Coqui XTTS v2",
+    description="Custom voice from dataset: Emmylahot12/nnamdi"
+)
+
+if __name__ == "__main__":
+    interface.launch()