Views
No views yet
1import torch
2from transformers import VitsTokenizer, VitsModel, set_seed
3import scipy
4
5tokenizer = VitsTokenizer.from_pretrained("VIZINTZOR/VIZINTZOR/MMS-TTS-THAI-FEMALEV1",cache_dir="./mms")
6model = VitsModel.from_pretrained("VIZINTZOR/VIZINTZOR/MMS-TTS-THAI-FEMALEV1",cache_dir="./mms")
7
8inputs = tokenizer(text="สวัสดีค่ะ นี่คือเสียงพูดภาษาไทย", return_tensors="pt")
9
10set_seed(456) # make deterministic
11
12with torch.no_grad():
13 outputs = model(**inputs)
14
15waveform = outputs.waveform[0]
16
17# Convert PyTorch tensor to NumPy array
18waveform_array = waveform.numpy()
19
20scipy.io.wavfile.write("techno_output.wav", rate=model.config.sampling_rate, data=waveform_array)