Update app.py
Browse files
app.py
CHANGED
|
@@ -10,21 +10,21 @@ model, utils = torch.hub.load(repo_or_dir='snakers4/silero-vad',
|
|
| 10 |
|
| 11 |
(get_speech_timestamps, save_audio, read_audio, VADIterator, collect_chunks) = utils
|
| 12 |
|
| 13 |
-
def remove_silence(audio_path):
|
| 14 |
if audio_path is None:
|
| 15 |
return None
|
| 16 |
|
| 17 |
wav = read_audio(audio_path, sampling_rate=16000)
|
| 18 |
|
| 19 |
-
#
|
| 20 |
speech_timestamps = get_speech_timestamps(
|
| 21 |
wav,
|
| 22 |
model,
|
| 23 |
sampling_rate=16000,
|
| 24 |
-
threshold=
|
| 25 |
-
min_silence_duration_ms=
|
| 26 |
-
speech_pad_ms=
|
| 27 |
-
min_speech_duration_ms=50
|
| 28 |
)
|
| 29 |
|
| 30 |
if not speech_timestamps:
|
|
@@ -57,25 +57,36 @@ css = """
|
|
| 57 |
}
|
| 58 |
"""
|
| 59 |
|
| 60 |
-
# ---- Gradio Blocks (Advanced UI) ----
|
| 61 |
with gr.Blocks(css=css, title="AI Silence Remover") as iface:
|
| 62 |
|
| 63 |
gr.HTML('<div class="deepu-brand">✨ Made by Deepu ✨</div>')
|
| 64 |
|
| 65 |
-
gr.Markdown("<h1 style='text-align: center;'>🎙️ AI Silence Remover (
|
| 66 |
-
gr.Markdown("<p style='text-align: center;'>
|
| 67 |
|
| 68 |
with gr.Row():
|
| 69 |
-
|
| 70 |
-
|
| 71 |
-
|
| 72 |
-
|
| 73 |
-
|
| 74 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 75 |
|
| 76 |
gr.HTML('<div class="deepu-brand">✨ Made by Deepu ✨</div>')
|
| 77 |
|
| 78 |
-
submit_btn.click(fn=remove_silence, inputs=audio_in, outputs=audio_out)
|
| 79 |
|
| 80 |
if __name__ == "__main__":
|
| 81 |
iface.launch()
|
|
|
|
| 10 |
|
| 11 |
(get_speech_timestamps, save_audio, read_audio, VADIterator, collect_chunks) = utils
|
| 12 |
|
| 13 |
+
def remove_silence(audio_path, vad_threshold, min_silence, padding):
|
| 14 |
if audio_path is None:
|
| 15 |
return None
|
| 16 |
|
| 17 |
wav = read_audio(audio_path, sampling_rate=16000)
|
| 18 |
|
| 19 |
+
# Ab values sliders se aayengi!
|
| 20 |
speech_timestamps = get_speech_timestamps(
|
| 21 |
wav,
|
| 22 |
model,
|
| 23 |
sampling_rate=16000,
|
| 24 |
+
threshold=vad_threshold,
|
| 25 |
+
min_silence_duration_ms=int(min_silence),
|
| 26 |
+
speech_pad_ms=int(padding),
|
| 27 |
+
min_speech_duration_ms=50
|
| 28 |
)
|
| 29 |
|
| 30 |
if not speech_timestamps:
|
|
|
|
| 57 |
}
|
| 58 |
"""
|
| 59 |
|
| 60 |
+
# ---- Gradio Blocks (Advanced UI with Sliders) ----
|
| 61 |
with gr.Blocks(css=css, title="AI Silence Remover") as iface:
|
| 62 |
|
| 63 |
gr.HTML('<div class="deepu-brand">✨ Made by Deepu ✨</div>')
|
| 64 |
|
| 65 |
+
gr.Markdown("<h1 style='text-align: center;'>🎙️ AI Silence Remover (Pro Control 🎛️)</h1>")
|
| 66 |
+
gr.Markdown("<p style='text-align: center;'>Ab aap khud sliders se control karein ki kitni strict cutting karni hai!</p>")
|
| 67 |
|
| 68 |
with gr.Row():
|
| 69 |
+
with gr.Column():
|
| 70 |
+
audio_in = gr.Audio(type="filepath", label="1. Raw Audio Upload Karein")
|
| 71 |
+
|
| 72 |
+
gr.Markdown("### ⚙️ Advance Settings (Inhe adjust karein)")
|
| 73 |
+
# Slider 1: AI Strictness (Maine default 0.90 kar diya hai, pehle 0.80 tha)
|
| 74 |
+
vad_threshold = gr.Slider(minimum=0.1, maximum=0.99, value=0.90, step=0.05, label="AI Strictness (Jitna zyada, utni strict cutting)")
|
| 75 |
+
|
| 76 |
+
# Slider 2: Min Silence Duration
|
| 77 |
+
min_silence = gr.Slider(minimum=0, maximum=500, value=10, step=10, label="Min Silence Duration (ms) - 0 rakhein sharpest cut ke liye")
|
| 78 |
+
|
| 79 |
+
# Slider 3: Padding
|
| 80 |
+
padding = gr.Slider(minimum=0, maximum=200, value=0, step=10, label="Speech Padding (ms) - Aawaz ke aage-peeche ki safety jagah")
|
| 81 |
+
|
| 82 |
+
submit_btn = gr.Button("Cut Silence", variant="primary")
|
| 83 |
+
|
| 84 |
+
with gr.Column():
|
| 85 |
+
audio_out = gr.Audio(type="filepath", label="2. Processed Audio (Zero Gaps)")
|
| 86 |
|
| 87 |
gr.HTML('<div class="deepu-brand">✨ Made by Deepu ✨</div>')
|
| 88 |
|
| 89 |
+
submit_btn.click(fn=remove_silence, inputs=[audio_in, vad_threshold, min_silence, padding], outputs=audio_out)
|
| 90 |
|
| 91 |
if __name__ == "__main__":
|
| 92 |
iface.launch()
|