Update app.py

app.py

@@ -23,16 +23,16 @@ SAMPLE_RATE = feature_extractor.sampling_rate
 SEED = 42
 
 default_text = "La voix humaine est un instrument de musique au-dessus de tous les autres."
-default_description = "The voice speaks slowly with a very noisy background, carrying a low-pitch tone and displaying a touch of expressiveness and animation. The sound is very distant, adding an air of intrigue."
+default_description = "A male voice speaks slowly with a very noisy background, displaying a touch of expressiveness and animation. The sound is very distant, adding an air of intrigue."
 examples = [
 [
 "La voix humaine est un instrument de musique au-dessus de tous les autres.",
-"A male voice speaks slowly with a very noisy background, carrying a low-pitch tone and displaying a touch of expressiveness and animation. The sound is very distant, adding an air of intrigue.",
+"A male voice speaks slowly with a very noisy background, displaying a touch of expressiveness and animation. The sound is very distant, adding an air of intrigue.",
 None,
 ],
 [
 "Tout ce qu'un homme est capable d'imaginer, d'autres hommes seront capables de le réaliser.",
-"A female voice delivers a slightly expressive and animated speech with a moderate speed. The recording features a low-pitch voice and slight background noise, creating a close-sounding audio experience.",
+"A male voice delivers a slightly expressive and animated speech with a moderate speed. The recording features a low-pitch voice, creating a close-sounding audio experience.",
 None,
 ],
 [
@@ -139,7 +139,7 @@ with gr.Blocks(css=css) as block:
 high-fidelity text-to-speech (TTS) models.</p> 
 <p>The model demonstrated here, French Parler-TTS <a href="https://huggingface.co/PHBJT/french_parler_tts_mini_v0.1">Mini v0.1 French</a>, 
 has been fine-tuned on a French dataset. It generates high-quality speech with features that can be controlled using a simple text prompt (e.g. gender, background noise, speaking rate, pitch and reverberation). 
-Due to limitations on the dataset, this model might underperform for female voices.</p>
+Due to limitations on the dataset, this model might underperform for female voices (we recomand using male voices only).</p>
 
 <p>By default, Parler-TTS generates 🎲 random male voice characteristics. To ensure 🎯 <b>speaker consistency</b> across generations, try to use consistent descriptions in your prompts.</p>
 <p><b>Note:</b> do NOT specify the nationnality of the speaker it will cause inconsistent audio generation (do: "a male speaker", don't: "a french male speaker") </p>