File size: 707 Bytes
fa52b2c
96b4603
9dd385e
2df411b
fa52b2c
2df411b
 
9dd385e
 
fa52b2c
 
9dd385e
 
dd6b086
9dd385e
 
 
fa52b2c
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
import gradio as gr
import torch
import scipy.io.wavfile as wavfile
from transformers import AutoProcessor, SeamlessM4TModel

tokenizer = AutoProcessor.from_pretrained("facebook/hf-seamless-m4t-medium")
model = SeamlessM4TModel.from_pretrained("facebook/hf-seamless-m4t-medium")

text = "some example text in the English language"

def greet(text):
    inputs = tokenizer(text, return_tensors="pt")
    with torch.no_grad():
        output = model(**inputs, decoder_input_ids=inputs["input_ids"]).waveform
        out = output[0]
        wavfile.write("tmp.wav", rate=16000, data=out)
        return open("tmp.wav", "rb").read()

iface = gr.Interface(fn=greet, inputs="text", outputs="audio")
iface.launch()