Jethuestad
commited on
Commit
•
d659367
1
Parent(s):
0758a27
added gradio app
Browse files
app.py
ADDED
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
from transformers import pipeline
|
3 |
+
|
4 |
+
|
5 |
+
from datasets import DatasetDict, Dataset, load_dataset, Audio
|
6 |
+
from transformers import WhisperProcessor, WhisperForConditionalGeneration
|
7 |
+
|
8 |
+
def transcribe(audio):
|
9 |
+
# load model and processor
|
10 |
+
processor = WhisperProcessor.from_pretrained("openai/whisper-medium")
|
11 |
+
model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-medium")
|
12 |
+
|
13 |
+
ds = Dataset.from_dict({"audio": [audio]}).cast_column("audio", Audio())
|
14 |
+
ds = ds.cast_column("audio", Audio(sampling_rate=16_000))
|
15 |
+
input_speech = next(iter(ds))["audio"]["array"]
|
16 |
+
|
17 |
+
input_features = processor(input_speech, return_tensors="pt").input_features
|
18 |
+
forced_decoder_ids = processor.get_decoder_prompt_ids(language = "no", task = "transcribe")
|
19 |
+
|
20 |
+
predicted_ids = model.generate(input_features, forced_decoder_ids = forced_decoder_ids)
|
21 |
+
transcription = processor.batch_decode(predicted_ids, skip_special_tokens = True)
|
22 |
+
|
23 |
+
return transcription
|
24 |
+
|
25 |
+
|
26 |
+
gr.Interface(
|
27 |
+
title = "OpenAI Whisper ASR Gradio Norwegian Web UI",
|
28 |
+
fn=transcribe,
|
29 |
+
inputs=[
|
30 |
+
gr.inputs.Audio(type="filepath")
|
31 |
+
],
|
32 |
+
outputs=[
|
33 |
+
"textbox"
|
34 |
+
]
|
35 |
+
).launch()
|