Fine-Tuning/train_ms.py

#-*- coding: utf-8 -*-
import os
import json
import argparse
import itertools
import math
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
from torch.cuda.amp import autocast, GradScaler
import librosa
import logging

# logging.getLogger('numba').setLevel(logging.WARNING)

# logging.getLogger('matplotlib').setLevel(logging.WARNING)
# logging.getLogger('PIL').setLevel(logging.WARNING)

import commons
import utils
from data_utils import (
  TextAudioSpeakerLoader,
  TextAudioSpeakerCollate,
  DistributedBucketSampler
)
from models import (
  SynthesizerTrn,
  MultiPeriodDiscriminator,
)
from losses import (
  generator_loss,
  discriminator_loss,
  feature_loss,
  kl_loss
)
from mel_processing import mel_spectrogram_torch, spec_to_mel_torch
from text.symbols import symbols

torch.backends.cudnn.benchmark = True
global_step = 0

def main():
  """Оптимизирано за Windows и единична видеокарта (RTX 4060)"""
  assert torch.cuda.is_available(), "CPU тренирането не е разрешено за VITS."

  # Зареждаме параметрите от config.json
  hps = utils.get_hparams()
  
  # Стартираме директно без mp.spawn процеси, за да не гърми под Windows
  run(0, 1, hps)

def run(rank, n_gpus, hps):
  global global_step
  
  # Настройки за TensorBoard и Логване
  logger = utils.get_logger(hps.model_dir)
  logger.info(hps)
  utils.check_git_hash(hps.model_dir)
  writer = SummaryWriter(log_dir=hps.model_dir)
  writer_eval = SummaryWriter(log_dir=os.path.join(hps.model_dir, "eval"))

  torch.manual_seed(hps.train.seed)
  torch.cuda.set_device(rank)

  # Зареждане на डेटाсета
  train_dataset = TextAudioSpeakerLoader(hps.data.training_files, hps.data)
  train_sampler = DistributedBucketSampler(
      train_dataset,
      hps.train.batch_size,
      [32,300,400,500,600,700,800,900,1000],
      num_replicas=n_gpus,
      rank=rank,
      shuffle=True)
  
  collate_fn = TextAudioSpeakerCollate()
  
  # 🎯 FIX ЗА WINDOWS: num_workers е закован на 0, за да няма тихи сривове
  train_loader = DataLoader(train_dataset, num_workers=0, shuffle=False, pin_memory=True,
      collate_fn=collate_fn, batch_sampler=train_sampler)
  
  eval_dataset = TextAudioSpeakerLoader(hps.data.validation_files, hps.data)
  eval_loader = DataLoader(eval_dataset, num_workers=0, shuffle=False,
      batch_size=hps.train.batch_size, pin_memory=True,
      drop_last=False, collate_fn=collate_fn)

  # Създаване на Генератора и Дискриминатора
  net_g = SynthesizerTrn(
      len(symbols),
      hps.data.filter_length // 2 + 1,
      hps.train.segment_size // hps.data.hop_length,
      n_speakers=hps.data.n_speakers,
      **hps.model).cuda(rank)
      
  net_d = MultiPeriodDiscriminator(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)
# ---------------------------------------------
# Фикс за новите версии на PyTorch (unhashable dict)
# if hasattr(torch.nn.utils, 'remove_weight_norm'):
#     try:
#         import warnings
#         warnings.filterwarnings("ignore", category=UserWarning, message=".*weight_norm.*")
#     except:
#         pass
# ---------------------------------------------
  # Премахнати са DDP обвивките, които чупят Windows при единично GPU
  try:
    _, _, _, epoch_str = utils.load_checkpoint(utils.latest_checkpoint_path(hps.model_dir, "G_*.pth"), net_g, optim_g)
    _, _, _, epoch_str = utils.load_checkpoint(utils.latest_checkpoint_path(hps.model_dir, "D_*.pth"), net_d, optim_d)
    global_step = (epoch_str - 1) * len(train_loader)
    epoch_str = 1
    print(f"♻️ Намерена съществуваща точка! Продължаваме от епоха: {epoch_str}")
  except:
    print("✨ Не са намерени стари записи. Стартираме от Епоха 1 (чисто ново начало или базови тегла)...")
    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)

  scaler = GradScaler(enabled=hps.train.fp16_run)

  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], scaler, [train_loader, eval_loader], logger, [writer, writer_eval])
    scheduler_g.step()
    scheduler_d.step()

def train_and_evaluate(rank, epoch, hps, nets, optims, schedulers, scaler, loaders, logger, writers):
  net_g, net_d = nets
  optim_g, optim_d = optims
  scheduler_g, scheduler_d = schedulers
  train_loader, eval_loader = loaders
  writer, writer_eval = writers

  train_loader.batch_sampler.set_epoch(epoch)
  global global_step

  net_g.train()
  net_d.train()

# epoch_id = 1  # Насилствено нулиране на брояча за епох
  
  for batch_idx, (x, x_lengths, spec, spec_lengths, y, y_lengths, speakers) in enumerate(train_loader):
    x, x_lengths = x.cuda(rank, non_blocking=True), x_lengths.cuda(rank, non_blocking=True)
    spec, spec_lengths = spec.cuda(rank, non_blocking=True), spec_lengths.cuda(rank, non_blocking=True)
    y, y_lengths = y.cuda(rank, non_blocking=True), y_lengths.cuda(rank, non_blocking=True)
    speakers = speakers.cuda(rank, non_blocking=True)

    # Обучение на Дискриминатора (Съдията)
    with autocast(enabled=hps.train.fp16_run):
      y_hat, l_length, attn, ids_slice, x_mask, z_mask,\
      (z, z_p, m_p, logs_p, m_q, logs_q) = net_g(x, x_lengths, spec, spec_lengths, speakers)

      mel = spec_to_mel_torch(spec, hps.data.filter_length, hps.data.n_mel_channels, hps.data.sampling_rate, hps.data.mel_fmin, hps.data.mel_fmax)
      y_mel = commons.slice_segments(mel, ids_slice, hps.train.segment_size // hps.data.hop_length)
      y_hat_mel = mel_spectrogram_torch(y_hat.squeeze(1), hps.data.filter_length, hps.data.n_mel_channels, hps.data.sampling_rate, hps.data.hop_length, hps.data.win_length, hps.data.mel_fmin, hps.data.mel_fmax)
      y = commons.slice_segments(y, ids_slice * hps.data.hop_length, hps.train.segment_size)

      y_d_hat_r, y_d_hat_g, _, _ = net_d(y, y_hat.detach())
      with autocast(enabled=False):
        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()
    scaler.scale(loss_disc_all).backward()
    scaler.unscale_(optim_d)
    grad_norm_d = commons.clip_grad_value_(net_d.parameters(), None)
    scaler.step(optim_d)

    # Обучение на Генератора (Гласа)
    with autocast(enabled=hps.train.fp16_run):
      y_d_hat_r, y_d_hat_g, fmap_r, fmap_g = net_d(y, y_hat)
      with autocast(enabled=False):
        loss_dur = torch.sum(l_length.float())
        loss_mel = F.l1_loss(y_mel, y_hat_mel) * hps.train.c_mel
        loss_kl = kl_loss(z_p, logs_q, m_p, logs_p, z_mask) * hps.train.c_kl
        loss_fm = feature_loss(fmap_r, fmap_g)
        loss_gen, losses_gen = generator_loss(y_d_hat_g)
        loss_gen_all = loss_gen + loss_fm + loss_mel + loss_dur + loss_kl
        
    optim_g.zero_grad()
    scaler.scale(loss_gen_all).backward()
    scaler.unscale_(optim_g)
    grad_norm_g = commons.clip_grad_value_(net_g.parameters(), None)
    scaler.step(optim_g)
    scaler.update()

    # Логове и извеждане на информация
    if global_step % hps.train.log_interval == 0:
      lr = optim_g.param_groups[0]['lr']
      losses = [loss_disc, loss_gen, loss_fm, loss_mel, loss_dur, loss_kl]
      logger.info('Train Epoch: {} [{:.0f}%] | Step: {} | Loss G: {:.4f} | Loss D: {:.4f}'.format(
        epoch, 100. * batch_idx / len(train_loader), global_step, loss_gen_all.item(), loss_disc_all.item()))

      print('Train Epoch: {} [{:.0f}%] | Step: {} | Loss G: {:.4f} | Loss D: {:.4f}'.format(
        epoch, 100. * batch_idx / len(train_loader), global_step, loss_gen_all.item(), loss_disc_all.item()))
      
      scalar_dict = {"loss/g/total": loss_gen_all, "loss/d/total": loss_disc_all, "learning_rate": lr, "grad_norm_d": grad_norm_d, "grad_norm_g": grad_norm_g}
      scalar_dict.update({"loss/g/fm": loss_fm, "loss/g/mel": loss_mel, "loss/g/dur": loss_dur, "loss/g/kl": loss_kl})
      utils.summarize(writer=writer, global_step=global_step, scalars=scalar_dict)

    # Валидация и Автоматичен запис на контролни точки
    if global_step % hps.train.eval_interval == 0 and global_step > 0:
      evaluate(hps, net_g, eval_loader, writer_eval)
      utils.save_checkpoint(net_g, optim_g, hps.train.learning_rate, epoch, os.path.join(hps.model_dir, "G_{}.pth".format(global_step)))
      utils.save_checkpoint(net_d, optim_d, hps.train.learning_rate, epoch, os.path.join(hps.model_dir, "D_{}.pth".format(global_step)))
      
      # Автоматично триене на много стари точки, за да не се пълни диска
      old_g = os.path.join(hps.model_dir, "G_{}.pth".format(global_step-10000))
      old_d = os.path.join(hps.model_dir, "D_{}.pth".format(global_step-10000))
      if os.path.exists(old_g): os.remove(old_g)
      if os.path.exists(old_d): os.remove(old_d)
      
    global_step += 1
    
  logger.info('====> Епоха {} приключи.'.format(epoch))
  print('====> Епоха {} приключи. ', format(epoch)), 'приключи.'
# print('Стъпки:', global_step, '====> Епоха {} приключи. ', format(epoch)), 'приключи.'

def evaluate(hps, generator, eval_loader, writer_eval):
    generator.eval()
    with torch.no_grad():
      for batch_idx, (x, x_lengths, spec, spec_lengths, y, y_lengths, speakers) in enumerate(eval_loader):
        x, x_lengths = x.cuda(0), x_lengths.cuda(0)
        spec, spec_lengths = spec.cuda(0), spec_lengths.cuda(0)
        y, y_lengths = y.cuda(0), y_lengths.cuda(0)
        speakers = speakers.cuda(0)
        break
        
      # Генериране на тестово аудио за TensorBoard
      y_hat, attn, mask, *_ = generator.infer(x, x_lengths, speakers, max_len=1000)
      y_hat_lengths = mask.sum([1,2]).long() * hps.data.hop_length

      mel = spec_to_mel_torch(spec, hps.data.filter_length, hps.data.n_mel_channels, hps.data.sampling_rate, hps.data.mel_fmin, hps.data.mel_fmax)
      y_hat_mel = mel_spectrogram_torch(y_hat.squeeze(1).float(), hps.data.filter_length, hps.data.n_mel_channels, hps.data.sampling_rate, hps.data.hop_length, hps.data.win_length, hps.data.mel_fmin, hps.data.mel_fmax)
      
    image_dict = {"gen/mel": utils.plot_spectrogram_to_numpy(y_hat_mel[0].cpu().numpy())}
    audio_dict = {"gen/audio": y_hat[0,:,:y_hat_lengths[0]]}
    
    utils.summarize(writer=writer_eval, global_step=global_step, images=image_dict, audios=audio_dict, audio_sampling_rate=hps.data.sampling_rate)
    generator.train()

if __name__ == "__main__":
  main()