Upload folder using huggingface_hub
Browse files- app.py +59 -23
- requirements.txt +1 -1
app.py
CHANGED
@@ -14,11 +14,27 @@ def load_tts_model():
|
|
14 |
|
15 |
tts = load_tts_model()
|
16 |
|
17 |
-
|
18 |
-
|
19 |
-
|
20 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
21 |
|
|
|
|
|
|
|
22 |
with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as temp_audio:
|
23 |
temp_audio_path = temp_audio.name
|
24 |
|
@@ -31,28 +47,48 @@ def clone(text, audio_file, language, speaking_rate, pitch, volume,
|
|
31 |
|
32 |
return temp_audio_path
|
33 |
|
|
|
|
|
|
|
|
|
34 |
# Define Gradio interface
|
35 |
-
|
36 |
-
|
37 |
-
|
38 |
-
|
39 |
-
gr.
|
40 |
-
|
41 |
-
|
42 |
-
|
43 |
-
|
44 |
-
|
45 |
-
|
46 |
-
|
47 |
-
|
48 |
-
|
49 |
-
|
50 |
-
|
51 |
-
|
52 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
53 |
|
54 |
# Launch the interface
|
55 |
-
|
56 |
|
57 |
# Clean up temporary files (this will run after the Gradio server is closed)
|
58 |
for file in os.listdir():
|
|
|
14 |
|
15 |
tts = load_tts_model()
|
16 |
|
17 |
+
# Celebrity voices (example list, you may want to expand or modify this)
|
18 |
+
celebrity_voices = {
|
19 |
+
"Morgan Freeman": "path/to/morgan_freeman_sample.wav",
|
20 |
+
"Scarlett Johansson": "path/to/scarlett_johansson_sample.wav",
|
21 |
+
"David Attenborough": "path/to/david_attenborough_sample.wav",
|
22 |
+
}
|
23 |
+
|
24 |
+
def tts_generate(text, voice, language):
|
25 |
+
with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as temp_audio:
|
26 |
+
temp_audio_path = temp_audio.name
|
27 |
+
|
28 |
+
tts.tts_to_file(
|
29 |
+
text=text,
|
30 |
+
speaker_wav=celebrity_voices[voice],
|
31 |
+
language=language,
|
32 |
+
file_path=temp_audio_path
|
33 |
+
)
|
34 |
|
35 |
+
return temp_audio_path
|
36 |
+
|
37 |
+
def clone_voice(text, audio_file, language):
|
38 |
with tempfile.NamedTemporaryFile(delete=False, suffix=".wav") as temp_audio:
|
39 |
temp_audio_path = temp_audio.name
|
40 |
|
|
|
47 |
|
48 |
return temp_audio_path
|
49 |
|
50 |
+
# Placeholder function for Talking Image tab
|
51 |
+
def talking_image_placeholder():
|
52 |
+
return "Talking Image functionality not implemented yet."
|
53 |
+
|
54 |
# Define Gradio interface
|
55 |
+
with gr.Blocks() as demo:
|
56 |
+
gr.Markdown("# Advanced Voice Synthesis")
|
57 |
+
|
58 |
+
with gr.Tabs():
|
59 |
+
with gr.TabItem("TTS"):
|
60 |
+
with gr.Row():
|
61 |
+
tts_text = gr.Textbox(label="Text to speak")
|
62 |
+
tts_voice = gr.Dropdown(choices=list(celebrity_voices.keys()), label="Celebrity Voice")
|
63 |
+
tts_language = gr.Dropdown(["en", "es", "fr", "de", "it"], label="Language", value="en")
|
64 |
+
tts_generate_btn = gr.Button("Generate")
|
65 |
+
tts_output = gr.Audio(label="Generated Audio")
|
66 |
+
|
67 |
+
tts_generate_btn.click(
|
68 |
+
tts_generate,
|
69 |
+
inputs=[tts_text, tts_voice, tts_language],
|
70 |
+
outputs=tts_output
|
71 |
+
)
|
72 |
+
|
73 |
+
with gr.TabItem("Talking Image"):
|
74 |
+
gr.Markdown("Talking Image functionality coming soon!")
|
75 |
+
|
76 |
+
with gr.TabItem("Clone Voice"):
|
77 |
+
with gr.Row():
|
78 |
+
clone_text = gr.Textbox(label="Text to speak")
|
79 |
+
clone_audio = gr.Audio(label="Voice reference audio file", type="filepath")
|
80 |
+
clone_language = gr.Dropdown(["en", "es", "fr", "de", "it"], label="Language", value="en")
|
81 |
+
clone_generate_btn = gr.Button("Generate")
|
82 |
+
clone_output = gr.Audio(label="Generated Audio")
|
83 |
+
|
84 |
+
clone_generate_btn.click(
|
85 |
+
clone_voice,
|
86 |
+
inputs=[clone_text, clone_audio, clone_language],
|
87 |
+
outputs=clone_output
|
88 |
+
)
|
89 |
|
90 |
# Launch the interface
|
91 |
+
demo.launch()
|
92 |
|
93 |
# Clean up temporary files (this will run after the Gradio server is closed)
|
94 |
for file in os.listdir():
|
requirements.txt
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
-
|
2 |
torch
|
3 |
TTS
|
|
|
1 |
+
streamlit
|
2 |
torch
|
3 |
TTS
|