Changed to use argentinian female voice

README.md

@@ -1,13 +1,35 @@
 ---
-title: persian-tts-piper
+title: Piper showcasing trained argentinian female voice
 emoji: 🌍
-colorFrom: blue
+colorFrom: indigo
 colorTo: blue
 sdk: gradio
-sdk_version: 5.13.1
+sdk_version: 5.34.2
 app_file: app.py
+short-description: A sample of the Piper Model with Argentinian female voice
 pinned: false
 license: mit
+disable_embedding: true
+language:
+ - es
+tags:
+ - piper
+ - tts
+ - text-to-speech
+ - onnx
+ - español
+base_model:
+ - rhasspy/piper-voices
+models:
+ - gyroing/Persian-Piper-Model-gyro
+ - rhasspy/piper-voices
+ - larcanio/piper-voices
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+This is a sample showcasing how to use [Piper](https://github.com/rhasspy/piper) with [Gradio](https://github.com/gradio-app/gradio) to create a TTS app with Argentinian female voice.
+
+## References
+
+- [Piper](https://github.com/rhasspy/piper)
+- [Argentinian Female Voice](https://huggingface.co/larcanio/piper-voices)
+- [Persian Female Voice](https://huggingface.co/gyroing/Persian-Piper-Model-gyro)

app.py

@@ -3,12 +3,12 @@ import wave
 import numpy as np
 from io import BytesIO
 from huggingface_hub import hf_hub_download
-from piper import PiperVoice 
+from piper import PiperVoice
 from transformers import pipeline
 import typing
 
-model_path = hf_hub_download(repo_id="gyroing/Persian-Piper-Model-gyro", filename="fa_IR-gyro-medium.onnx")
-config_path = hf_hub_download(repo_id="gyroing/Persian-Piper-Model-gyro", filename="fa_IR-gyro-medium.onnx.json")
+model_path = hf_hub_download(repo_id="larcanio/piper-voices", filename="es_AR-daniela-high.onnx")
+config_path = hf_hub_download(repo_id="larcanio/piper-voices", filename="es_AR-daniela-high.json")
 voice = PiperVoice.load(model_path, config_path)
 
 
@@ -32,15 +32,24 @@ def synthesize_speech(text):
 
     return audio_data.tobytes(), None
 
+BANNER_TEXT = """
+# Demo en español argentino con Piper
+
+[***Piper***](https://huggingface.co/rhasspy/piper-voices/) es un modelo de abierto de Texto a Voz (TTS)
+que permite entrenarse con voz propia, destaca por no requerir conectarse a Internet y ofrecer resultados
+sin exigir GPU.  Inicialmente diseñado para Raspberri Pi.
+
+Este demo solo muestra español, puedes probar [voces en otros idiomas](https://rhasspy.github.io/piper-samples/).
+"""
+
 # Using Gradio Blocks
 with gr.Blocks(theme=gr.themes.Base()) as blocks:
-    gr.Markdown("# Persian Text to Speech Synthesizer")
-    gr.Markdown("Enter text to synthesize it into speech using Piper With Persian gyro Model :")
-    input_text = gr.Textbox(label=" ", rtl=True , text_align="right" )
-    output_audio = gr.Audio(label="Synthesized Speech", type="numpy")
-    output_text = gr.Textbox(label="Output Text", visible=False, rtl=True)  # This is the new text output component
-    submit_button = gr.Button("Synthesize")
+    gr.Markdown(BANNER_TEXT)
+    input_text = gr.Textbox(label=" ", placeholder="Introduce el texto a leer aquí")
+    output_audio = gr.Audio(label="Audio generado", type="numpy")
+    output_text = gr.Textbox(label="Tokens generados", visible=False)
+    submit_button = gr.Button("Genera audio")
 
     submit_button.click(synthesize_speech, inputs=input_text, outputs=[output_audio, output_text])
 # Run the app
-blocks.launch()
+blocks.launch()