Spaces:
Sleeping
Sleeping
rexsimiloluwah
commited on
Commit
•
bcb1eec
1
Parent(s):
cefcfbf
added app for automatic speech recognition
Browse files- app.py +13 -4
- requirements.txt +8 -0
- tasks/__init__.py +0 -0
- tasks/asr.py +47 -0
app.py
CHANGED
@@ -1,7 +1,16 @@
|
|
1 |
import gradio as gr
|
2 |
|
3 |
-
|
4 |
-
|
|
|
|
|
5 |
|
6 |
-
|
7 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
import gradio as gr
|
2 |
|
3 |
+
from tasks.asr import (
|
4 |
+
mic_transcribe_interface,
|
5 |
+
file_transcribe_interface
|
6 |
+
)
|
7 |
|
8 |
+
app = gr.Blocks()
|
9 |
+
|
10 |
+
with app:
|
11 |
+
gr.TabbedInterface(
|
12 |
+
[mic_transcribe_interface, file_transcribe_interface],
|
13 |
+
["Transcribe from Microphone", "Transcribe from Audio File"]
|
14 |
+
)
|
15 |
+
|
16 |
+
app.launch(share=True)
|
requirements.txt
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
torch
|
2 |
+
librosa
|
3 |
+
soundfile
|
4 |
+
transformers
|
5 |
+
pillow
|
6 |
+
numpy
|
7 |
+
requests
|
8 |
+
matplotlib
|
tasks/__init__.py
ADDED
File without changes
|
tasks/asr.py
ADDED
@@ -0,0 +1,47 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import librosa
|
2 |
+
import numpy as np
|
3 |
+
import gradio as gr
|
4 |
+
import soundfile as sf
|
5 |
+
from transformers import pipeline
|
6 |
+
|
7 |
+
# Load the pipeline
|
8 |
+
model = pipeline(
|
9 |
+
task="automatic-speech-recognition",
|
10 |
+
model="distil-whisper/distil-small.en"
|
11 |
+
)
|
12 |
+
|
13 |
+
def transcribe_audio(filepath):
|
14 |
+
"""Transcribe audio to text"""
|
15 |
+
audio, sample_rate = sf.read(filepath)
|
16 |
+
audio_mono = librosa.to_mono(np.transpose(audio))
|
17 |
+
|
18 |
+
# resample the audio
|
19 |
+
audio_16KHz = librosa.resample(
|
20 |
+
audio_mono,
|
21 |
+
orig_sr=sample_rate,
|
22 |
+
target_sr=16000
|
23 |
+
)
|
24 |
+
|
25 |
+
output = model(
|
26 |
+
audio_16KHz,
|
27 |
+
chunk_length_s=30,
|
28 |
+
batch_size=4,
|
29 |
+
)
|
30 |
+
|
31 |
+
return output["text"]
|
32 |
+
|
33 |
+
mic_transcribe_interface = gr.Interface(
|
34 |
+
fn=transcribe_audio,
|
35 |
+
inputs=gr.Audio(sources="microphone", type="filepath"),
|
36 |
+
outputs=gr.Textbox(label="Transcription", lines=3),
|
37 |
+
allow_flagging="never",
|
38 |
+
title="Transcribe Audio from your Microphone"
|
39 |
+
)
|
40 |
+
|
41 |
+
file_transcribe_interface = gr.Interface(
|
42 |
+
fn=transcribe_audio,
|
43 |
+
inputs=gr.Audio(sources="upload", type="filepath"),
|
44 |
+
outputs=gr.Textbox(label="Transcription", lines=3),
|
45 |
+
allow_flagging="never",
|
46 |
+
title="Transcribe Audio from a File"
|
47 |
+
)
|