app.py

# import gradio as gr

# def greet(name):
#     return "Hello " + name + "!!"

# demo = gr.Interface(fn=greet, inputs="text", outputs="text")
# demo.launch()


import torch
torch.manual_seed(0)
torch.backends.cudnn.benchmark = False
torch.backends.cudnn.deterministic = True

import random
random.seed(0)

import numpy as np
np.random.seed(0)

# load packages
import time
import random
import yaml
from munch import Munch
import numpy as np
import torch
from torch import nn
import torch.nn.functional as F
import torchaudio
import librosa
from nltk.tokenize import word_tokenize

from models import *
from utils import *
from text_utils import TextCleaner
textcleaner = TextCleaner()

from huggingface_hub import hf_hub_download
import os

import nltk
nltk.download('punkt')
nltk.download('punkt_tab')
print("nltk downloaded punkt and punkt_tab")

HF_TOKEN = os.getenv("HF_TOKEN", None)
hf_hub_download(
    repo_id="presencesw/tts", 
    filename="phoaudio_single_v1.pth", 
    local_dir="Models/phoaudio_single_v1", 
    local_dir_use_symlinks=False,
    token=HF_TOKEN
)

to_mel = torchaudio.transforms.MelSpectrogram(
    n_mels=80, n_fft=2048, win_length=1200, hop_length=300)
mean, std = -4, 4

def length_to_mask(lengths):
    mask = torch.arange(lengths.max()).unsqueeze(0).expand(lengths.shape[0], -1).type_as(lengths)
    mask = torch.gt(mask+1, lengths.unsqueeze(1))
    return mask

def preprocess(wave):
    wave_tensor = torch.from_numpy(wave).float()
    mel_tensor = to_mel(wave_tensor)
    mel_tensor = (torch.log(1e-5 + mel_tensor.unsqueeze(0)) - mean) / std
    return mel_tensor

def compute_style(path):
    wave, sr = librosa.load(path, sr=24000)
    audio, index = librosa.effects.trim(wave, top_db=30)
    if sr != 24000:
        audio = librosa.resample(audio, sr, 24000)
    mel_tensor = preprocess(audio).to(device)

    with torch.no_grad():
        ref_s = model.style_encoder(mel_tensor.unsqueeze(1))
        ref_p = model.predictor_encoder(mel_tensor.unsqueeze(1))

    return torch.cat([ref_s, ref_p], dim=1)

device = 'cuda' if torch.cuda.is_available() else 'cpu'

# load phonemizer
import phonemizer
# global_phonemizer = phonemizer.backend.EspeakBackend(language='en-us', preserve_punctuation=True,  with_stress=True)
vi_phonemizer = phonemizer.backend.EspeakBackend(language='vi', preserve_punctuation=True, with_stress=True, words_mismatch='ignore')

# config = yaml.safe_load(open("Models/LibriTTS/config.yml"))
config = yaml.safe_load(open("Models/phoaudio_single_v1/config_phoaudio_single_v1.yml"))

# load pretrained ASR model
ASR_config = config.get('ASR_config', False)
ASR_path = config.get('ASR_path', False)
text_aligner = load_ASR_models(ASR_path, ASR_config)

# load pretrained F0 model
F0_path = config.get('F0_path', False)
pitch_extractor = load_F0_models(F0_path)

# # load BERT model
# from Utils.PLBERT.util import load_plbert
# BERT_path = config.get('PLBERT_dir', False)
# plbert = load_plbert(BERT_path)
from Utils_extend_v1.PLBERT.util import load_plbert
BERT_path = config.get('PLBERT_dir', False)
plbert = load_plbert(BERT_path)

model_params = recursive_munch(config['model_params'])
model = build_model(model_params, text_aligner, pitch_extractor, plbert)
_ = [model[key].eval() for key in model]
_ = [model[key].to(device) for key in model]

params_whole = torch.load("Models/phoaudio_single_v1/phoaudio_single_v1.pth", map_location='cpu')
params = params_whole['net']

for key in model:
    if key in params:
        print('%s loaded' % key)
        try:
            model[key].load_state_dict(params[key])
        except:
            from collections import OrderedDict
            state_dict = params[key]
            new_state_dict = OrderedDict()
            for k, v in state_dict.items():
                name = k[7:] # remove `module.`
                new_state_dict[name] = v
            # load params
            model[key].load_state_dict(new_state_dict, strict=False)
#             except:
#                 _load(params[key], model[key])
_ = [model[key].eval() for key in model]

from Modules.diffusion.sampler import DiffusionSampler, ADPM2Sampler, KarrasSchedule

sampler = DiffusionSampler(
    model.diffusion.diffusion,
    sampler=ADPM2Sampler(),
    sigma_schedule=KarrasSchedule(sigma_min=0.0001, sigma_max=3.0, rho=9.0), # empirical parameters
    clamp=False
)

def inference(text, ref_s, alpha = 0.3, beta = 0.7, diffusion_steps=5, embedding_scale=1):
    text = text.strip()
    ps = vi_phonemizer.phonemize([text])
    ps = ps[0].replace("(en)", "").replace("(vi)", "").strip()
    # ps = word_tokenize(temp)
    # ps = word_tokenize(ps[0])
    # ps = ' '.join(ps)
    # ps = ps.replace(" .", ".").replace(" ,", ",").replace(" !", "!").replace(" ?", "?").replace(" :", ":").replace(" ;", ";")
    # ps  = ps.replace('( en ) ', '(en)').replace(' ( vi )', '(vi)')
    # print(f"Phonemizer: {ps}")
    tokens = textcleaner(ps)
    # print(f"TextCleaner: {tokens}")
    tokens.insert(0, 0)
    tokens = torch.LongTensor(tokens).to(device).unsqueeze(0)

    with torch.no_grad():
        input_lengths = torch.LongTensor([tokens.shape[-1]]).to(device)
        text_mask = length_to_mask(input_lengths).to(device)

        t_en = model.text_encoder(tokens, input_lengths, text_mask)
        bert_dur = model.bert(tokens, attention_mask=(~text_mask).int())
        d_en = model.bert_encoder(bert_dur).transpose(-1, -2)

        s_pred = sampler(noise = torch.randn((1, 256)).unsqueeze(1).to(device),
                                          embedding=bert_dur,
                                          embedding_scale=embedding_scale,
                                            features=ref_s, # reference from the same speaker as the embedding
                                             num_steps=diffusion_steps).squeeze(1)


        s = s_pred[:, 128:]
        ref = s_pred[:, :128]

        ref = alpha * ref + (1 - alpha)  * ref_s[:, :128]
        s = beta * s + (1 - beta)  * ref_s[:, 128:]

        d = model.predictor.text_encoder(d_en,
                                         s, input_lengths, text_mask)

        x, _ = model.predictor.lstm(d)
        duration = model.predictor.duration_proj(x)

        duration = torch.sigmoid(duration).sum(axis=-1)
        pred_dur = torch.round(duration.squeeze()).clamp(min=1)


        pred_aln_trg = torch.zeros(input_lengths, int(pred_dur.sum().data))
        c_frame = 0
        for i in range(pred_aln_trg.size(0)):
            pred_aln_trg[i, c_frame:c_frame + int(pred_dur[i].data)] = 1
            c_frame += int(pred_dur[i].data)

        # encode prosody
        en = (d.transpose(-1, -2) @ pred_aln_trg.unsqueeze(0).to(device))
        if model_params.decoder.type == "hifigan":
            asr_new = torch.zeros_like(en)
            asr_new[:, :, 0] = en[:, :, 0]
            asr_new[:, :, 1:] = en[:, :, 0:-1]
            en = asr_new

        F0_pred, N_pred = model.predictor.F0Ntrain(en, s)

        asr = (t_en @ pred_aln_trg.unsqueeze(0).to(device))
        if model_params.decoder.type == "hifigan":
            asr_new = torch.zeros_like(asr)
            asr_new[:, :, 0] = asr[:, :, 0]
            asr_new[:, :, 1:] = asr[:, :, 0:-1]
            asr = asr_new
        # print(f"asr shape: {asr.shape}")
        # print(f"F0_pred shape: {F0_pred.shape}")
        # print(f"N_pred shape: {N_pred.shape}")
        # print(f"ref shape: {ref.shape}")
        out = model.decoder(asr,
                                F0_pred, N_pred, ref.squeeze().unsqueeze(0))


    return out.squeeze().cpu().numpy()[..., :-50], ps # weird pulse at the end of the model, need to be fixed later

voice_ref = "voice_ref/Tram_audio.wav"

import gradio as gr

def generate_voice(input_string):
    noise = torch.randn(1,1,256).to(device)
    ref = compute_style(voice_ref)
    start = time.time()
    wav, phonemizer = inference(input_string, ref, alpha=0.3, beta=0.7, diffusion_steps=5, embedding_scale=1)
    wav = wav.astype(np.float32)
    rtf = (time.time() - start) / (len(wav) / 24000)
    return (24000, wav), rtf, phonemizer

# demo = gr.Interface(fn=greet, inputs="text", outputs="text")
with gr.Blocks() as demo:
    # with gr.Row():
    #     text_input = gr.Textbox(value="Đây là một ví dụ về tổng hợp giọng nói.")
    #     audio_output = gr.Audio()
    # with gr.Row():
    #     run_button = gr.Button(value="generate voice")
    #     rtf_log = gr.Number(label="Real Time Factor")
    with gr.Row():
        with gr.Column():
            text_input = gr.Textbox(value="Đây là một ví dụ về tổng hợp giọng nói.", label="Input text")
            phonemizer = gr.Textbox(label="Phonemizer of the input text")
            run_button = gr.Button(value="Generate voice")
        with gr.Column():
            audio_output = gr.Audio(label="Generated audio")
            rtf_log = gr.Number(label="Real Time Factor")

    # run_button.click(fn=generate_voice, inputs=[text_input], outputs=[audio_output])
    run_button.click(fn=generate_voice, inputs=[text_input], outputs=[audio_output, rtf_log, phonemizer],)

if __name__ == "__main__":
    demo.launch(server_port=7860, server_name="0.0.0.0")