Update app.py
Browse files
app.py
CHANGED
@@ -2,14 +2,15 @@ import gradio as gr
|
|
2 |
import numpy as np
|
3 |
import torch
|
4 |
from datasets import load_dataset
|
5 |
-
|
6 |
from transformers import pipeline
|
|
|
7 |
|
8 |
|
9 |
device = "cuda:0" if torch.cuda.is_available() else "cpu"
|
10 |
|
11 |
# load speech translation checkpoint
|
12 |
-
asr_pipe = pipeline("automatic-speech-recognition", model="openai/whisper-
|
13 |
|
14 |
|
15 |
|
@@ -27,8 +28,8 @@ def translate(audio):
|
|
27 |
def synthesise(text):
|
28 |
inputs = processor(text=text, return_tensors="pt")
|
29 |
with torch.no_grad():
|
30 |
-
|
31 |
-
return
|
32 |
|
33 |
|
34 |
def speech_to_speech_translation(audio):
|
|
|
2 |
import numpy as np
|
3 |
import torch
|
4 |
from datasets import load_dataset
|
5 |
+
|
6 |
from transformers import pipeline
|
7 |
+
from transformers import VitsModel, VitsTokenizer
|
8 |
|
9 |
|
10 |
device = "cuda:0" if torch.cuda.is_available() else "cpu"
|
11 |
|
12 |
# load speech translation checkpoint
|
13 |
+
asr_pipe = pipeline("automatic-speech-recognition", model="openai/whisper-base", device=device)
|
14 |
|
15 |
|
16 |
|
|
|
28 |
def synthesise(text):
|
29 |
inputs = processor(text=text, return_tensors="pt")
|
30 |
with torch.no_grad():
|
31 |
+
speech = model(inputs["input_ids"].to(device))
|
32 |
+
return speech.audio[0]
|
33 |
|
34 |
|
35 |
def speech_to_speech_translation(audio):
|