Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
|
@@ -1 +1,29 @@
|
|
| 1 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import gradio as gr
|
| 2 |
+
from transformers import WhisperProcessor, FlaxWhisperForConditionalGeneration
|
| 3 |
+
|
| 4 |
+
# Load Whisper model
|
| 5 |
+
model_name = "openai/whisper-large"
|
| 6 |
+
processor = WhisperProcessor.from_pretrained(model_name)
|
| 7 |
+
model = FlaxWhisperForConditionalGeneration.from_pretrained(model_name)
|
| 8 |
+
|
| 9 |
+
def transcribe(audio):
|
| 10 |
+
# Preprocess audio
|
| 11 |
+
inputs = processor(audio, return_tensors="np", sampling_rate=16000)
|
| 12 |
+
input_features = inputs.input_features
|
| 13 |
+
|
| 14 |
+
# Generate transcription
|
| 15 |
+
predicted_ids = model.generate(input_features)
|
| 16 |
+
transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True)[0]
|
| 17 |
+
|
| 18 |
+
return transcription
|
| 19 |
+
|
| 20 |
+
# Gradio Interface
|
| 21 |
+
interface = gr.Interface(
|
| 22 |
+
fn=transcribe,
|
| 23 |
+
inputs=gr.Audio(source="upload", type="numpy"),
|
| 24 |
+
outputs="text",
|
| 25 |
+
title="Whisper JAX Transcription",
|
| 26 |
+
description="Upload an audio file to transcribe using Whisper JAX."
|
| 27 |
+
)
|
| 28 |
+
|
| 29 |
+
interface.launch()
|