Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
from funasr import AutoModel
|
3 |
+
|
4 |
+
# Initialize the FunASR model with the specified components
|
5 |
+
model = AutoModel(
|
6 |
+
model="paraformer-zh", model_revision="v2.0.4",
|
7 |
+
vad_model="fsmn-vad", vad_model_revision="v2.0.4",
|
8 |
+
punc_model="ct-punc-c", punc_model_revision="v2.0.4",
|
9 |
+
# Uncomment the next line to enable speaker verification/diarization
|
10 |
+
# spk_model="cam++", spk_model_revision="v2.0.2",
|
11 |
+
)
|
12 |
+
|
13 |
+
def transcribe(audio_file):
|
14 |
+
# Processing the input audio file
|
15 |
+
result = model.generate(
|
16 |
+
input=audio_file.name,
|
17 |
+
batch_size_s=300,
|
18 |
+
hotword='魔搭' # This is an example keyword; replace or remove as necessary
|
19 |
+
)
|
20 |
+
return result
|
21 |
+
|
22 |
+
# Gradio interface setup
|
23 |
+
interface = gr.Interface(
|
24 |
+
fn=transcribe,
|
25 |
+
inputs=gr.inputs.Audio(source="microphone", type="filepath", label="Upload your audio in Mandarin"),
|
26 |
+
outputs="text",
|
27 |
+
title="FunASR Speech Recognition",
|
28 |
+
description="This Gradio app uses the paraformer-zh model for speech recognition with additional features like VAD and punctuation restoration."
|
29 |
+
)
|
30 |
+
|
31 |
+
# Run the Gradio app
|
32 |
+
if __name__ == "__main__":
|
33 |
+
interface.launch()
|