Spaces:
Running
Running
Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
from transformers import WhisperProcessor, WhisperForConditionalGeneration
|
3 |
+
import torch
|
4 |
+
|
5 |
+
# モデルとプロセッサの読み込み
|
6 |
+
model_name = "openai/whisper-large-v3"
|
7 |
+
processor = WhisperProcessor.from_pretrained(model_name)
|
8 |
+
model = WhisperForConditionalGeneration.from_pretrained(model_name)
|
9 |
+
|
10 |
+
# 音声ファイルを文字起こしする関数
|
11 |
+
def transcribe_audio(audio):
|
12 |
+
# 音声を処理
|
13 |
+
audio_input = processor(audio, return_tensors="pt", sampling_rate=16000)
|
14 |
+
|
15 |
+
# モデルによる文字起こし
|
16 |
+
with torch.no_grad():
|
17 |
+
predicted_ids = model.generate(input_ids=audio_input.input_ids)
|
18 |
+
|
19 |
+
# 文字起こし結果のデコード
|
20 |
+
transcription = processor.decode(predicted_ids[0], skip_special_tokens=True)
|
21 |
+
|
22 |
+
return transcription
|
23 |
+
|
24 |
+
# Gradioのインターフェース作成
|
25 |
+
interface = gr.Interface(
|
26 |
+
fn=transcribe_audio,
|
27 |
+
inputs=gr.Audio(source="microphone", type="filepath"),
|
28 |
+
outputs="text",
|
29 |
+
live=True
|
30 |
+
)
|
31 |
+
|
32 |
+
# インターフェースの起動
|
33 |
+
interface.launch()
|