Spaces:
Paused
Paused
Update app.py
Browse files
app.py
CHANGED
|
@@ -1,14 +1,27 @@
|
|
| 1 |
import gradio as gr
|
| 2 |
import spaces
|
| 3 |
import torch
|
|
|
|
| 4 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5 |
zero = torch.Tensor([0]).cuda()
|
| 6 |
print(zero.device) # <-- 'cpu' 🤔
|
| 7 |
-
|
| 8 |
@spaces.GPU
|
| 9 |
-
def
|
| 10 |
-
|
| 11 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 12 |
|
| 13 |
-
demo = gr.Interface(fn=
|
| 14 |
demo.launch()
|
|
|
|
| 1 |
import gradio as gr
|
| 2 |
import spaces
|
| 3 |
import torch
|
| 4 |
+
from transformers import AutoTokenizer
|
| 5 |
|
| 6 |
+
|
| 7 |
+
|
| 8 |
+
token=os.environ.get("key_")
|
| 9 |
+
tokenizer = AutoTokenizer.from_pretrained("wasmdashai/vtk",token=token)
|
| 10 |
+
|
| 11 |
+
model=VitsModel.from_pretrained("wasmdashai/vtk",token=token).cuda()
|
| 12 |
zero = torch.Tensor([0]).cuda()
|
| 13 |
print(zero.device) # <-- 'cpu' 🤔
|
| 14 |
+
import torch
|
| 15 |
@spaces.GPU
|
| 16 |
+
def modelspeech(text):
|
| 17 |
+
|
| 18 |
+
|
| 19 |
+
inputs = tokenizer(text, return_tensors="pt").cuda()
|
| 20 |
+
with torch.no_grad():
|
| 21 |
+
wav = model(input_ids=inputs["input_ids"]).waveform.cpu().numpy().reshape(-1)#.detach()
|
| 22 |
+
|
| 23 |
+
return model.config.sampling_rate,wav#remove_noise_nr(wav)
|
| 24 |
+
|
| 25 |
|
| 26 |
+
demo = gr.Interface(fn=modelspeech, inputs=["text", outputs=["audio"])
|
| 27 |
demo.launch()
|