patrickvonplaten commited on
Commit
916a703
β€’
1 Parent(s): b5455c9
Files changed (4) hide show
  1. README.md +4 -4
  2. app.py +37 -0
  3. packages.txt +2 -0
  4. requirements.txt +5 -0
README.md CHANGED
@@ -1,8 +1,8 @@
1
  ---
2
- title: XLS R 2B 21 EN
3
- emoji: 🏒
4
- colorFrom: green
5
- colorTo: indigo
6
  sdk: gradio
7
  app_file: app.py
8
  pinned: false
 
1
  ---
2
+ title: XLS R 2B 22 EN
3
+ emoji: πŸ“Š
4
+ colorFrom: gray
5
+ colorTo: red
6
  sdk: gradio
7
  app_file: app.py
8
  pinned: false
app.py ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import gradio as gr
2
+ import librosa
3
+ from transformers import AutoFeatureExtractor, AutoTokenizer, SpeechEncoderDecoderModel
4
+
5
+ model_name = "facebook/wav2vec2-xls-r-2b-21-to-en",
6
+
7
+ feature_extractor = AutoFeatureExtractor.from_pretrained(model_name)
8
+ tokenizer = AutoTokenizer.from_pretrained(model_name, use_fast=False)
9
+ model = SpeechEncoderDecoderModel.from_pretrained(model_name)
10
+
11
+ def process_audio_file(file):
12
+ data, sr = librosa.load(file)
13
+ if sr != 16000:
14
+ data = librosa.resample(data, sr, 16000)
15
+ input_values = feature_extractor(data, return_tensors="pt").input_values
16
+ return input_values
17
+
18
+ def transcribe(file):
19
+ input_values = process_audio_file(file)
20
+
21
+ sequences = model.generate(input_values, num_beams=1, max_length=30)
22
+
23
+ transcription = tokenizer.batch_decode(sequences, skip_special_tokens=True)
24
+ return transcription[0]
25
+
26
+ iface = gr.Interface(
27
+ fn=transcribe,
28
+ inputs=[
29
+ gr.inputs.Audio(source="microphone", type='filepath'),
30
+ ],
31
+ outputs="text",
32
+ layout="horizontal",
33
+ theme="huggingface",
34
+ title="XLS-R 2B 21-to-EN Speech Translation",
35
+ description="A simple interface to translate from 21 spoken languages to written English.",
36
+ )
37
+ iface.launch()
packages.txt ADDED
@@ -0,0 +1,2 @@
 
 
 
1
+ libsndfile1
2
+ sox
requirements.txt ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ SoundFile==0.9.0.post1
2
+ librosa
3
+ sentencepiece
4
+ torch
5
+ git+git://github.com/huggingface/transformers