import os import sys import json import argparse import itertools import math import time import logging import torch from torch import nn, optim from torch.nn import functional as F from torch.utils.data import DataLoader from torch.utils.tensorboard import SummaryWriter import torch.multiprocessing as mp import torch.distributed as dist from torch.nn.parallel import DistributedDataParallel as DDP from torch.cuda.amp import autocast, GradScaler sys.path.append('../..') import modules.commons as commons import utils.utils as utils from dataset import DatasetConstructor from models import ( SynthesizerTrn, Discriminator ) from modules.losses import ( generator_loss, discriminator_loss, feature_loss, kl_loss, ) from preprocess.mel_processing import mel_spectrogram_torch, spec_to_mel_torch, spectrogram_torch torch.backends.cudnn.benchmark = True global_step = 0 use_cuda = torch.cuda.is_available() print("use_cuda, ", use_cuda) numba_logger = logging.getLogger('numba') numba_logger.setLevel(logging.WARNING) def main(): """Assume Single Node Multi GPUs Training Only""" hps = utils.get_hparams() os.environ['MASTER_ADDR'] = 'localhost' os.environ['MASTER_PORT'] = str(hps.train.port) if (torch.cuda.is_available()): n_gpus = torch.cuda.device_count() mp.spawn(run, nprocs=n_gpus, args=(n_gpus, hps,)) else: cpurun(0, 1, hps) def run(rank, n_gpus, hps): global global_step if rank == 0: logger = utils.get_logger(hps.train.save_dir) logger.info(hps.train) logger.info(hps.data) logger.info(hps.model) utils.check_git_hash(hps.train.save_dir) writer = SummaryWriter(log_dir=hps.train.save_dir) writer_eval = SummaryWriter(log_dir=os.path.join(hps.train.save_dir, "eval")) dist.init_process_group(backend='nccl', init_method='env://', world_size=n_gpus, rank=rank) torch.manual_seed(hps.train.seed) torch.cuda.set_device(rank) dataset_constructor = DatasetConstructor(hps, num_replicas=n_gpus, rank=rank) train_loader = dataset_constructor.get_train_loader() if rank == 0: valid_loader = dataset_constructor.get_valid_loader() net_g = SynthesizerTrn(hps).cuda(rank) net_d = Discriminator(hps, hps.model.use_spectral_norm).cuda(rank) optim_g = torch.optim.AdamW( net_g.parameters(), hps.train.learning_rate, betas=hps.train.betas, eps=hps.train.eps) optim_d = torch.optim.AdamW( net_d.parameters(), hps.train.learning_rate, betas=hps.train.betas, eps=hps.train.eps) net_g = DDP(net_g, device_ids=[rank], find_unused_parameters=True) net_d = DDP(net_d, device_ids=[rank], find_unused_parameters=True) try: _, _, _, epoch_str = utils.load_checkpoint(utils.latest_checkpoint_path(hps.train.save_dir, "G_*.pth"), net_g, optim_g) _, _, _, epoch_str = utils.load_checkpoint(utils.latest_checkpoint_path(hps.train.save_dir, "D_*.pth"), net_d, optim_d) global_step = (epoch_str - 1) * len(train_loader) except: epoch_str = 1 global_step = 0 scheduler_g = torch.optim.lr_scheduler.ExponentialLR(optim_g, gamma=hps.train.lr_decay, last_epoch=epoch_str - 2) scheduler_d = torch.optim.lr_scheduler.ExponentialLR(optim_d, gamma=hps.train.lr_decay, last_epoch=epoch_str - 2) for epoch in range(epoch_str, hps.train.epochs + 1): if rank == 0: train_and_evaluate(rank, epoch, hps, [net_g, net_d], [optim_g, optim_d], [scheduler_g, scheduler_d], [train_loader, valid_loader], logger, [writer, writer_eval]) else: train_and_evaluate(rank, epoch, hps, [net_g, net_d], [optim_g, optim_d], [scheduler_g, scheduler_d], [train_loader, None], None, None) scheduler_g.step() scheduler_d.step() def cpurun(rank, n_gpus, hps): global global_step if rank == 0: logger = utils.get_logger(hps.train.save_dir) logger.info(hps.train) logger.info(hps.data) logger.info(hps.model) utils.check_git_hash(hps.train.save_dir) writer = SummaryWriter(log_dir=hps.train.save_dir) writer_eval = SummaryWriter(log_dir=os.path.join(hps.train.save_dir, "eval")) torch.manual_seed(hps.train.seed) dataset_constructor = DatasetConstructor(hps, num_replicas=n_gpus, rank=rank) train_loader = dataset_constructor.get_train_loader() if rank == 0: valid_loader = dataset_constructor.get_valid_loader() net_g = SynthesizerTrn(hps) net_d = Discriminator(hps, hps.model.use_spectral_norm) optim_g = torch.optim.AdamW( net_g.parameters(), hps.train.learning_rate, betas=hps.train.betas, eps=hps.train.eps) optim_d = torch.optim.AdamW( net_d.parameters(), hps.train.learning_rate, betas=hps.train.betas, eps=hps.train.eps) try: _, _, _, epoch_str = utils.load_checkpoint(utils.latest_checkpoint_path(hps.train.save_dir, "G_*.pth"), net_g, optim_g) _, _, _, epoch_str = utils.load_checkpoint(utils.latest_checkpoint_path(hps.train.save_dir, "D_*.pth"), net_g, optim_g) global_step = (epoch_str - 1) * len(train_loader) except: epoch_str = 1 global_step = 0 scheduler_g = torch.optim.lr_scheduler.ExponentialLR(optim_g, gamma=hps.train.lr_decay, last_epoch=epoch_str - 2) scheduler_d = torch.optim.lr_scheduler.ExponentialLR(optim_d, gamma=hps.train.lr_decay, last_epoch=epoch_str - 2) for epoch in range(epoch_str, hps.train.epochs + 1): train_and_evaluate(rank, epoch, hps, [net_g, net_d], [optim_g, optim_d], [scheduler_g, scheduler_d], [train_loader, valid_loader], logger, [writer, writer_eval]) scheduler_g.step() scheduler_d.step() def train_and_evaluate(rank, epoch, hps, nets, optims, schedulers, loaders, logger, writers): net_g, net_d = nets optim_g, optim_d = optims scheduler_g, scheduler_d = schedulers train_loader, eval_loader = loaders if writers is not None: writer, writer_eval = writers train_loader.sampler.set_epoch(epoch) global global_step net_g.train() net_d.train() for batch_idx, data_dict in enumerate(train_loader): phone = data_dict["phone"] pitchid = data_dict["pitchid"] dur = data_dict["dur"] slur = data_dict["slur"] gtdur = data_dict["gtdur"] mel = data_dict["mel"] f0 = data_dict["f0"] wav = data_dict["wav"] spkid = data_dict["spkid"] phone_lengths = data_dict["phone_lengths"] mel_lengths = data_dict["mel_lengths"] wav_lengths = data_dict["wav_lengths"] f0_lengths = data_dict["f0_lengths"] # data if (use_cuda): phone, phone_lengths = phone.cuda(rank, non_blocking=True), phone_lengths.cuda(rank, non_blocking=True) pitchid = pitchid.cuda(rank, non_blocking=True) dur = dur.cuda(rank, non_blocking=True) slur = slur.cuda(rank, non_blocking=True) gtdur = gtdur.cuda(rank, non_blocking=True) mel, mel_lengths = mel.cuda(rank, non_blocking=True), mel_lengths.cuda(rank, non_blocking=True) wav, wav_lengths = wav.cuda(rank, non_blocking=True), wav_lengths.cuda(rank, non_blocking=True) f0, f0_lengths = f0.cuda(rank, non_blocking=True), f0_lengths.cuda(rank, non_blocking=True) spkid = spkid.cuda(rank, non_blocking=True) # forward y_hat, ids_slice, LF0, y_ddsp, kl_div, predict_mel, mask = net_g(phone, phone_lengths, pitchid, dur, slur, gtdur, f0, mel, mel_lengths, spk_id=spkid) y_ddsp = y_ddsp.unsqueeze(1) # Discriminator y = commons.slice_segments(wav, ids_slice * hps.data.hop_size, hps.train.segment_size) # slice y_ddsp_mel = mel_spectrogram_torch( y_ddsp.squeeze(1), hps.data.n_fft, hps.data.acoustic_dim, hps.data.sample_rate, hps.data.hop_size, hps.data.win_size, hps.data.fmin, hps.data.fmax ) y_logspec = torch.log(spectrogram_torch( y.squeeze(1), hps.data.n_fft, hps.data.sample_rate, hps.data.hop_size, hps.data.win_size ) + 1e-7) y_ddsp_logspec = torch.log(spectrogram_torch( y_ddsp.squeeze(1), hps.data.n_fft, hps.data.sample_rate, hps.data.hop_size, hps.data.win_size ) + 1e-7) y_mel = mel_spectrogram_torch( y.squeeze(1), hps.data.n_fft, hps.data.acoustic_dim, hps.data.sample_rate, hps.data.hop_size, hps.data.win_size, hps.data.fmin, hps.data.fmax ) y_hat_mel = mel_spectrogram_torch( y_hat.squeeze(1), hps.data.n_fft, hps.data.acoustic_dim, hps.data.sample_rate, hps.data.hop_size, hps.data.win_size, hps.data.fmin, hps.data.fmax ) y_d_hat_r, y_d_hat_g, _, _ = net_d(y, y_hat.detach()) loss_disc, losses_disc_r, losses_disc_g = discriminator_loss(y_d_hat_r, y_d_hat_g) loss_disc_all = loss_disc optim_d.zero_grad() loss_disc_all.backward() grad_norm_d = commons.clip_grad_value_(net_d.parameters(), None) optim_d.step() # loss y_d_hat_r, y_d_hat_g, fmap_r, fmap_g = net_d(y, y_hat) loss_mel = F.l1_loss(y_mel, y_hat_mel) * 45 loss_mel_dsp = F.l1_loss(y_mel, y_ddsp_mel) * 45 loss_spec_dsp = F.l1_loss(y_logspec, y_ddsp_logspec) * 45 loss_mel_am = F.mse_loss(mel * mask, predict_mel * mask) # * 10 loss_fm = feature_loss(fmap_r, fmap_g) loss_gen, losses_gen = generator_loss(y_d_hat_g) loss_fm = loss_fm / 2 loss_gen = loss_gen / 2 loss_gen_all = loss_gen + loss_fm + loss_mel + loss_mel_dsp + kl_div + loss_mel_am + loss_spec_dsp loss_gen_all = loss_gen_all / hps.train.accumulation_steps loss_gen_all.backward() if ((global_step + 1) % hps.train.accumulation_steps == 0): grad_norm_g = commons.clip_grad_value_(net_g.parameters(), None) optim_g.step() optim_g.zero_grad() if rank == 0: if (global_step + 1) % (hps.train.accumulation_steps * 10) == 0: logger.info(["step&time", global_step, time.asctime(time.localtime(time.time()))]) logger.info(["mel&mel_dsp&spec_dsp: ", loss_mel, loss_mel_dsp, loss_spec_dsp]) logger.info(["adv&fm: ", loss_gen, loss_fm]) logger.info(["kl: ", kl_div]) logger.info(["am&dur: ", loss_mel_am]) if global_step % hps.train.log_interval == 0: lr = optim_g.param_groups[0]['lr'] losses = [loss_gen_all, loss_mel] logger.info('Train Epoch: {} [{:.0f}%]'.format( epoch, 100. * batch_idx / len(train_loader))) logger.info([x.item() for x in losses] + [global_step, lr]) scalar_dict = {"loss/total": loss_gen_all, "loss/mel": loss_mel, "loss/adv": loss_gen, "loss/fm": loss_fm, "loss/mel_ddsp": loss_mel_dsp, "loss/spec_ddsp": loss_spec_dsp, "loss/mel_am": loss_mel_am, "loss/kl_div": kl_div, "learning_rate": lr} utils.summarize( writer=writer, global_step=global_step, scalars=scalar_dict) if global_step % hps.train.eval_interval == 0: logger.info(['All training params(G): ', utils.count_parameters(net_g), ' M']) # print('Sub training params(G): ', \ # 'text_encoder: ', utils.count_parameters(net_g.module.text_encoder), ' M, ', \ # 'decoder: ', utils.count_parameters(net_g.module.decoder), ' M, ', \ # 'mel_decoder: ', utils.count_parameters(net_g.module.mel_decoder), ' M, ', \ # 'dec: ', utils.count_parameters(net_g.module.dec), ' M, ', \ # 'dec_harm: ', utils.count_parameters(net_g.module.dec_harm), ' M, ', \ # 'dec_noise: ', utils.count_parameters(net_g.module.dec_noise), ' M, ', \ # 'posterior: ', utils.count_parameters(net_g.module.posterior_encoder), ' M, ', \ # ) evaluate(hps, net_g, eval_loader, writer_eval) utils.save_checkpoint(net_g, optim_g, hps.train.learning_rate, epoch, os.path.join(hps.train.save_dir, "G_{}.pth".format(global_step)), hps.train.eval_interval) utils.save_checkpoint(net_d, optim_d, hps.train.learning_rate, epoch, os.path.join(hps.train.save_dir, "D_{}.pth".format(global_step)), hps.train.eval_interval) net_g.train() global_step += 1 if rank == 0: logger.info('====> Epoch: {}'.format(epoch)) def evaluate(hps, generator, eval_loader, writer_eval): generator.eval() image_dict = {} audio_dict = {} with torch.no_grad(): for batch_idx, data_dict in enumerate(eval_loader): if batch_idx == 4: break phone = data_dict["phone"] pitchid = data_dict["pitchid"] dur = data_dict["dur"] slur = data_dict["slur"] gtdur = data_dict["gtdur"] mel = data_dict["mel"] f0 = data_dict["f0"] wav = data_dict["wav"] spkid = data_dict["spkid"] phone_lengths = data_dict["phone_lengths"] mel_lengths = data_dict["mel_lengths"] wav_lengths = data_dict["wav_lengths"] f0_lengths = data_dict["f0_lengths"] # data if (use_cuda): phone, phone_lengths = phone.cuda(0), phone_lengths.cuda(0) pitchid = pitchid.cuda(0) dur = dur.cuda(0) slur = slur.cuda(0) wav = wav.cuda(0) mel = mel.cuda(0) f0 = f0.cuda(0) gtdur = gtdur.cuda(0) spkid = spkid.cuda(0) # remove else phone = phone[:1] phone_lengths = phone_lengths[:1] pitchid = pitchid[:1] dur = dur[:1] slur = slur[:1] wav = wav[:1] mel = mel[:1] f0 = f0[:1] gtdur = gtdur[:1] spkid = spkid[:1] y_hat, y_harm, y_noise = generator.module.infer(phone, phone_lengths, pitchid, dur, slur, gtdur=gtdur, F0=f0, spk_id=spkid) spec = spectrogram_torch( wav.squeeze(1), hps.data.n_fft, hps.data.sample_rate, hps.data.hop_size, hps.data.win_size ) y_mel = mel_spectrogram_torch( wav.squeeze(1), hps.data.n_fft, hps.data.acoustic_dim, hps.data.sample_rate, hps.data.hop_size, hps.data.win_size, hps.data.fmin, hps.data.fmax ) y_hat_mel = mel_spectrogram_torch( y_hat.squeeze(1), hps.data.n_fft, hps.data.acoustic_dim, hps.data.sample_rate, hps.data.hop_size, hps.data.win_size, hps.data.fmin, hps.data.fmax ) image_dict.update({ f"gen/mel_{batch_idx}": utils.plot_spectrogram_to_numpy(y_hat_mel[0].cpu().numpy()), }) audio_dict.update( { f"gen/audio_{batch_idx}": y_hat[0, :, :], f"gen/harm_{batch_idx}": y_harm[0, :, :], "gen/noise": y_noise[0, :, :] }) # if global_step == 0: image_dict.update({f"gt/mel_{batch_idx}": utils.plot_spectrogram_to_numpy(mel[0].cpu().numpy())}) audio_dict.update({f"gt/audio_{batch_idx}": wav[0, :, :wav_lengths[0]]}) utils.summarize( writer=writer_eval, global_step=global_step, images=image_dict, audios=audio_dict, audio_sampling_rate=hps.data.sample_rate ) generator.train() if __name__ == "__main__": main()