train.py

# Copyright (C) 2021. Huawei Technologies Co., Ltd. All rights reserved.
# This program is free software; you can redistribute it and/or modify
# it under the terms of the MIT License.
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# MIT License for more details.

import numpy as np
from tqdm import tqdm

import torch
from torch.utils.data import DataLoader
from torch.utils.tensorboard import SummaryWriter

import params
from model import GradTTS, GradTTSSDP, GradTTSStft, GradTTSDependencyGraph
from data import TextMelDataset, TextMelBatchCollate, TextMelGraphDataset, TextMelGraphDatasetCollate
from utils import plot_tensor, save_plot
from text.symbols import symbols
import pickle


train_filelist_path = params.train_filelist_path
valid_filelist_path = params.valid_filelist_path
cmudict_path = params.cmudict_path
add_blank = params.add_blank

log_dir = params.log_dir
n_epochs = params.n_epochs
batch_size = params.batch_size
out_size = params.out_size
learning_rate = params.learning_rate
random_seed = params.seed

nsymbols = len(symbols) + 1 if add_blank else len(symbols)
n_enc_channels = params.n_enc_channels
filter_channels = params.filter_channels
filter_channels_dp = params.filter_channels_dp
n_enc_layers = params.n_enc_layers
enc_kernel = params.enc_kernel
enc_dropout = params.enc_dropout
n_heads = params.n_heads
window_size = params.window_size

n_feats = params.n_feats
n_fft = params.n_fft
sample_rate = params.sample_rate
hop_length = params.hop_length
win_length = params.win_length
f_min = params.f_min
f_max = params.f_max

dec_dim = params.dec_dim
beta_min = params.beta_min
beta_max = params.beta_max
pe_scale = params.pe_scale

stft_params = params.stft_config
stft_loss_params = params.stft_loss_config

segment_size = params.segment_size
hop_length = params.hop_length

synta_params = params.synta_params

ph_dict_size = 41

output_dir = "/mnt/Stuff/TTS/speech_synthesis_in_bangla-master/resources/data"
all_words_file = f'{output_dir}/all_words.txt'

all_words = set()

with open(all_words_file, 'r', encoding='utf-8') as f:
    for line in f:
        all_words.add(line.strip())

word_dict_size = len(all_words)

word_dict = {}

for i, word in enumerate(all_words):
    word_dict[word] = i + 1

if __name__ == "__main__":

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

    torch.manual_seed(random_seed)
    np.random.seed(random_seed)

    print('Initializing logger...')
    logger = SummaryWriter(log_dir=log_dir)
    
    """Dataloader for DependencyGraph model"""
    print('Initializing data loaders...')
    train_dataset = TextMelGraphDataset(synta_params, train_filelist_path, word_dict)
    batch_collate = TextMelGraphDatasetCollate(synta_params)
    loader = DataLoader(dataset=train_dataset, batch_size=batch_size,
                        collate_fn=batch_collate, drop_last=True,
                        num_workers=0, shuffle=False)
    
    test_dataset = TextMelGraphDataset(synta_params, valid_filelist_path, word_dict)
    test_loader = DataLoader(dataset=test_dataset, batch_size=batch_size,
                        collate_fn=batch_collate, drop_last=True,
                        num_workers=0, shuffle=False)

    # print('Initializing data loaders...')
    # train_dataset = TextMelDataset(train_filelist_path, add_blank,
    #                                n_fft, n_feats, sample_rate, hop_length,
    #                                win_length, f_min, f_max)
    # batch_collate = TextMelBatchCollate()
    # loader = DataLoader(dataset=train_dataset, batch_size=batch_size,
    #                     collate_fn=batch_collate, drop_last=True,
    #                     num_workers=4, shuffle=False)
    # test_dataset = TextMelDataset(valid_filelist_path, add_blank,
    #                               n_fft, n_feats, sample_rate, hop_length,
    #                               win_length, f_min, f_max)

    print('Initializing model...')

    """ GradTTSSDP model """
    # model = GradTTSSDP(nsymbols, 1, None, n_enc_channels, filter_channels, filter_channels_dp,
    #                 n_heads, n_enc_layers, enc_kernel, enc_dropout, window_size, 
    #                 n_feats, dec_dim, beta_min, beta_max, pe_scale).to(device)

    """GradTTSStft model"""
    # model = GradTTSStft(nsymbols, 1, None, n_enc_channels, filter_channels, filter_channels_dp,
    #                 n_heads, n_enc_layers, enc_kernel, enc_dropout, window_size, 
    #                 n_feats, dec_dim, beta_min, beta_max, pe_scale, stft_params).to(device)
    # model.load_state_dict(torch.load("/mnt/Stuff/TTS/speech_synthesis_in_bangla-master/logs/multistream_stft_diffusion/grad_5.pt", map_location=lambda loc, storage: loc))

    """ GradTTSDependencyGraph model"""
    model = GradTTSDependencyGraph(nsymbols, 1, None, n_enc_channels, filter_channels, filter_channels_dp,
                    n_heads, n_enc_layers, enc_kernel, enc_dropout, window_size, 
                    n_feats, dec_dim, beta_min, beta_max, pe_scale, word_dict_size, ph_dict_size, synta_params).to(device)  

    # print('Number of encoder + duration predictor parameters: %.2fm' % (model.encoder.nparams/1e6))
    # print('Number of decoder parameters: %.2fm' % (model.decoder.nparams/1e6))
    # print('Total parameters: %.2fm' % (model.nparams/1e6))

    print('Initializing optimizer...')
    optimizer = torch.optim.Adam(params=model.parameters(), lr=learning_rate)

    print('Logging test batch...')
    test_batch = test_dataset.sample_test_batch(size=params.test_size)
    # for i, item in enumerate(test_batch):
        # mel = item['mel']
        # logger.add_image(f'image_{i}/ground_truth', plot_tensor(mel.squeeze()),
        #                  global_step=0, dataformats='HWC')
        # save_plot(mel.squeeze(), f'{log_dir}/original_{i}.png')

    print('Start training...')
    iteration = 0
    for epoch in range(1, n_epochs + 1):
        model.train()
        dur_losses = []
        prior_losses = []
        diff_losses = []
        context_losses = []
        stft_losses = []
        with tqdm(loader, total=len(train_dataset)//batch_size) as progress_bar:
            for batch_idx, batch in enumerate(progress_bar):
                model.zero_grad()
                x, x_lengths = batch['x'].to(device), batch['x_lengths'].to(device)
                y, y_lengths = batch['y'].to(device), batch['y_lengths'].to(device)
                word_tokens = batch['word_tokens'].to(device)
                mel2ph = batch['mel2ph'].to(device)
                mel2word = batch['mel2word'].to(device)
                ph2word = batch['ph2word'].to(device)
                graph_lst = batch['graph_lst']
                etypes_lst = batch['etypes_lst']

                """ for running GradTTSStft model """
                # y_enc, y_dec, attn, y_g_hat, y_mb_hat = model(x, x_lengths, n_timesteps=50)
                # dur_loss, prior_loss, diff_loss, stft_loss = model.compute_loss(x, x_lengths,
                #                                                      y, y_lengths,
                #                                                      y_mb_hat, **stft_loss_params,
                #                                                      out_size=out_size)
                # loss = sum([dur_loss, prior_loss, diff_loss, stft_loss])

                """GradTTSSDP"""
                # y_enc, y_dec, attn = model(x, x_lengths, n_timesteps=50)
                # dur_loss, prior_loss, diff_loss = model.compute_loss(x, x_lengths,
                #                                                      y, y_lengths,
                #                                                      out_size=out_size)

                """ for running DependencyGraph model """
                y_enc, y_dec, attn, ret = model(x, x_lengths, word_tokens, ph2word,
                                                mel2word, mel2ph, graph_lst, etypes_lst, n_timesteps=50)
                dur_loss, prior_loss, diff_loss = model.compute_loss(x, x_lengths,
                                                                     y, y_lengths, ret,
                                                                     out_size=out_size)
                loss = sum([dur_loss, prior_loss, diff_loss])


                loss.backward()

                enc_grad_norm = torch.nn.utils.clip_grad_norm_(model.encoder.parameters(),
                                                               max_norm=1)
                dec_grad_norm = torch.nn.utils.clip_grad_norm_(model.decoder.parameters(),
                                                               max_norm=1)
                optimizer.step()

                logger.add_scalar('training/duration_loss', dur_loss.item(),
                                  global_step=iteration)
                logger.add_scalar('training/prior_loss', prior_loss.item(),
                                  global_step=iteration)
                logger.add_scalar('training/diffusion_loss', diff_loss.item(),
                                  global_step=iteration)
                
                """ Omit if not running GradTTSStft """
                # logger.add_scalar('training/stft_loss', stft_loss.item(),
                #                   global_step=iteration)

                logger.add_scalar('training/encoder_grad_norm', enc_grad_norm,
                                  global_step=iteration)
                logger.add_scalar('training/decoder_grad_norm', dec_grad_norm,
                                  global_step=iteration)
                
                dur_losses.append(dur_loss.item())
                prior_losses.append(prior_loss.item())
                diff_losses.append(diff_loss.item())

                """ Omit if not running GradTTSStft """
                # stft_losses.append(stft_loss.item())
                
                if batch_idx % 5 == 0:
                    """ For GradTTSStft """
                    # msg = f'Epoch: {epoch}, iteration: {iteration} | dur_loss: {dur_loss.item()}, prior_loss: {prior_loss.item()}, diff_loss: {diff_loss.item()}, stft_loss: {stft_loss.item()}'

                    """ For others """
                    msg = f'Epoch: {epoch}, iteration: {iteration} | dur_loss: {dur_loss.item()}, prior_loss: {prior_loss.item()}, diff_loss: {diff_loss.item()}'

                    progress_bar.set_description(msg)
                
                iteration += 1

                break

        log_msg = 'Epoch %d: duration loss = %.3f ' % (epoch, np.mean(dur_losses))
        log_msg += '| prior loss = %.3f ' % np.mean(prior_losses)
        log_msg += '| diffusion loss = %.3f\n' % np.mean(diff_losses)

        """ Omit if not using GradTTSStft """
        log_msg += '| stft loss = %.3f\n' % np.mean(stft_losses)


        with open(f'{log_dir}/train.log', 'a') as f:
            f.write(log_msg)

        if epoch % params.save_every > 0:
            continue

        model.eval()
        print('Synthesis...')
        with torch.no_grad():
            with tqdm(test_loader, total=len(test_dataset)//batch_size) as progress_bar:
                for batch_idx, item in enumerate(progress_bar):
                    x, x_lengths = item['x'].to(device), item['x_lengths'].to(device)
                    word_tokens = item['word_tokens'].to(device)
                    mel2ph = item['mel2ph'].to(device)
                    mel2word = item['mel2word'].to(device)
                    ph2word = item['ph2word'].to(device)
                    graph_lst = item['graph_lst']
                    etypes_lst = item['etypes_lst']
                    # x = item['x'].to(torch.long).unsqueeze(0).to(device)
                    # x_lengths = torch.LongTensor([x.shape[-1]]).to(device)

                    """ For GradTTSStft """
                    # y_enc, y_dec, attn, _, _ = model(x, x_lengths, n_timesteps=50)

                    """Dependency graph model"""
                    y_enc, y_dec, attn, ret = model(x, x_lengths, word_tokens, ph2word,
                                                    mel2word, mel2ph, graph_lst, etypes_lst, n_timesteps=50)
                    """ For other models """
                    # y_enc, y_dec, attn = model(x, x_lengths, n_timesteps=50)


                    # logger.add_image(f'image_{i}/generated_enc',
                    #                  plot_tensor(y_enc.squeeze().cpu()),
                    #                  global_step=iteration, dataformats='HWC')
                    # logger.add_image(f'image_{i}/generated_dec',
                    #                  plot_tensor(y_dec.squeeze().cpu()),
                    #                  global_step=iteration, dataformats='HWC')
                    # logger.add_image(f'image_{i}/alignment',
                    #                  plot_tensor(attn.squeeze().cpu()),
                    #                  global_step=iteration, dataformats='HWC')
                    save_plot(y_enc.squeeze().cpu(), 
                            f'{log_dir}/generated_enc_{i}.png')
                    save_plot(y_dec.squeeze().cpu(), 
                            f'{log_dir}/generated_dec_{i}.png')
                    save_plot(attn.squeeze().cpu(), 
                            f'{log_dir}/alignment_{i}.png')
                    
                    break

        ckpt = model.state_dict()
        torch.save(ckpt, f=f"{log_dir}/grad_{epoch}.pt")