Initial commit for Arabic Dialect ID demo
Browse files- app.py +52 -0
- requirements.txt +4 -0
app.py
ADDED
@@ -0,0 +1,52 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
from transformers import pipeline
|
3 |
+
import numpy as np
|
4 |
+
|
5 |
+
# Load the model
|
6 |
+
model_id = "badrex/mms-300m-arabic-dialect-identifier" # Replace with your model ID
|
7 |
+
classifier = pipeline("audio-classification", model=model_id)
|
8 |
+
|
9 |
+
# Define dialect names for better display
|
10 |
+
dialect_mapping = {
|
11 |
+
"MSA": "Modern Standard Arabic",
|
12 |
+
"Egyptian": "Egyptian Arabic",
|
13 |
+
"Gulf": "Gulf Arabic",
|
14 |
+
"Levantine": "Levantine Arabic",
|
15 |
+
"Maghrebi": "Maghrebi Arabic"
|
16 |
+
}
|
17 |
+
|
18 |
+
def predict_dialect(audio, sr):
|
19 |
+
# Process the audio input
|
20 |
+
if len(audio.shape) > 1:
|
21 |
+
audio = audio.mean(axis=1) # Convert stereo to mono
|
22 |
+
|
23 |
+
# Classify the dialect
|
24 |
+
predictions = classifier({"sampling_rate": sr, "raw": audio})
|
25 |
+
|
26 |
+
# Format results for display
|
27 |
+
results = {}
|
28 |
+
for pred in predictions:
|
29 |
+
dialect_name = dialect_mapping.get(pred['label'], pred['label'])
|
30 |
+
results[dialect_name] = float(pred['score'])
|
31 |
+
|
32 |
+
return results
|
33 |
+
|
34 |
+
# Create the Gradio interface
|
35 |
+
demo = gr.Interface(
|
36 |
+
fn=predict_dialect,
|
37 |
+
inputs=gr.Audio(type="numpy", label="Upload or Record Audio"),
|
38 |
+
outputs=gr.Label(num_top_classes=5, label="Predicted Dialect"),
|
39 |
+
title="Arabic Dialect Identifier",
|
40 |
+
description="""This demo identifies Arabic dialects from speech audio.
|
41 |
+
Upload an audio file or record your voice speaking Arabic to see which dialect it matches.
|
42 |
+
The model identifies: Modern Standard Arabic (MSA), Egyptian, Gulf, Levantine, and Maghrebi dialects.""",
|
43 |
+
examples=[
|
44 |
+
# Optional: Add example audio files here if you have them
|
45 |
+
# ["examples/msa_example.wav"],
|
46 |
+
# ["examples/egyptian_example.wav"],
|
47 |
+
],
|
48 |
+
allow_flagging="never"
|
49 |
+
)
|
50 |
+
|
51 |
+
# Launch the app
|
52 |
+
demo.launch()
|
requirements.txt
ADDED
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
1 |
+
gradio>=3.50.2
|
2 |
+
transformers>=4.36.0
|
3 |
+
torch>=2.0.0
|
4 |
+
librosa>=0.10.1
|