Upload folder using huggingface_hub

train_sdxs_vae-Copy1.py

@@ -0,0 +1,627 @@
+# -*- coding: utf-8 -*-
+import os
+import math
+import re
+import torch
+import numpy as np
+import random
+import gc
+from datetime import datetime
+from pathlib import Path
+
+import torchvision.transforms as transforms
+import torch.nn.functional as F
+from torch.utils.data import DataLoader, Dataset
+from torch.optim.lr_scheduler import LambdaLR
+from diffusers import AutoencoderKL, AsymmetricAutoencoderKL
+# QWEN: импорт класса
+from diffusers import AutoencoderKLQwenImage
+from diffusers import AutoencoderKLWan
+
+from accelerate import Accelerator
+from PIL import Image, UnidentifiedImageError
+from tqdm import tqdm
+import bitsandbytes as bnb
+import wandb
+import lpips   # pip install lpips
+from FDL_pytorch import FDL_loss # pip install fdl-pytorch
+from collections import deque
+
+# --------------------------- Параметры ---------------------------
+ds_path            = "/workspace/d23"
+project            = "vae2"
+batch_size         = 1
+base_learning_rate = 6e-6
+min_learning_rate  = 7e-7
+num_epochs         = 1
+sample_interval_share = 30
+use_wandb          = False
+save_model         = True
+use_decay          = True
+optimizer_type     = "adam8bit"
+dtype              = torch.float32
+
+model_resolution   = 512
+high_resolution    = 1024
+limit              = 0
+save_barrier       = 1.3
+warmup_percent     = 0.005
+beta2              = 0.997
+eps                = 1e-8
+clip_grad_norm     = 2.0  # Увеличили с 1.0 для лучшей прорисовки деталей
+mixed_precision    = "no"
+gradient_accumulation_steps = 1
+generated_folder   = "samples"
+save_as            = "vae3"
+num_workers        = 0
+device = None
+torch.backends.cuda.matmul.allow_tf32 = True
+torch.backends.cudnn.allow_tf32 = True
+# Включение Flash Attention 2/SDPA #MAX_JOBS=4 pip install flash-attn --no-build-isolation
+torch.backends.cuda.enable_flash_sdp(True)
+torch.backends.cuda.enable_mem_efficient_sdp(True)
+torch.backends.cuda.enable_math_sdp(False) 
+
+# --- Режимы обучения ---
+# QWEN: учим только декодер
+train_decoder_only = True
+train_up_only      = False
+full_training      = False  # если True — учим весь VAE и добавляем KL (ниже)
+kl_ratio           = 0.0
+
+# Доли лоссов
+loss_ratios = {
+    "lpips": 0.60,
+    "fdl" :  0.15,   
+    "mse":   0.00,
+    "mae":   0.20,
+    "dssim": 0.05,  # <-- Заменён edge на DSSIM
+    "kl":    0.00,
+}
+median_coeff_steps = 250
+
+resize_long_side = 1280  # ресайз длинной стороны исходных картинок
+
+# QWEN: конфиг загрузки модели
+vae_kind      = "kl"  # "qwen" или "kl" (обычный)
+
+Path(generated_folder).mkdir(parents=True, exist_ok=True)
+
+accelerator = Accelerator(
+    mixed_precision=mixed_precision,
+    gradient_accumulation_steps=gradient_accumulation_steps
+)
+device = accelerator.device
+
+# reproducibility
+seed = int(datetime.now().strftime("%Y%m%d")) + 42
+torch.manual_seed(seed); np.random.seed(seed); random.seed(seed)
+torch.backends.cudnn.benchmark = False
+
+# --------------------------- WandB ---------------------------
+if use_wandb and accelerator.is_main_process:
+    wandb.init(project=project, config={
+        "batch_size": batch_size,
+        "base_learning_rate": base_learning_rate,
+        "num_epochs": num_epochs,
+        "optimizer_type": optimizer_type,
+        "model_resolution": model_resolution,
+        "high_resolution": high_resolution,
+        "gradient_accumulation_steps": gradient_accumulation_steps,
+        "train_decoder_only": train_decoder_only,
+        "full_training": full_training,
+        "kl_ratio": kl_ratio,
+        "vae_kind": vae_kind,
+    })
+
+# --------------------------- VAE ---------------------------
+def get_core_model(model):
+    m = model
+    # если модель уже обёрнута torch.compile
+    if hasattr(m, "_orig_mod"):
+        m = m._orig_mod
+    return m
+
+def is_video_vae(model) -> bool:
+    # WAN/Qwen — это видео-VAEs
+    if vae_kind in ("wan", "qwen"):
+        return True
+    # fallback по структуре (если понадобится)
+    try:
+        core = get_core_model(model)
+        enc = getattr(core, "encoder", None)
+        conv_in = getattr(enc, "conv_in", None)
+        w = getattr(conv_in, "weight", None)
+        if isinstance(w, torch.nn.Parameter):
+            return w.ndim == 5
+    except Exception:
+        pass
+    return False
+
+# загрузка
+if vae_kind == "qwen":
+    vae = AutoencoderKLQwenImage.from_pretrained("Qwen/Qwen-Image", subfolder="vae")
+else:
+    if vae_kind == "wan":
+        vae = AutoencoderKLWan.from_pretrained(project)
+    else:
+        # старое поведение (пример)
+        if model_resolution==high_resolution:
+            vae = AutoencoderKL.from_pretrained(project)
+        else:
+            vae = AsymmetricAutoencoderKL.from_pretrained(project)
+
+vae = vae.to(dtype)
+
+# torch.compile (опционально)
+if hasattr(torch, "compile"):
+    try:
+        vae = torch.compile(vae)
+    except Exception as e:
+        print(f"[WARN] torch.compile failed: {e}")
+
+# --------------------------- Freeze/Unfreeze ---------------------------
+core = get_core_model(vae)
+
+for p in core.parameters():
+    p.requires_grad = False
+
+unfrozen_param_names = []
+
+if full_training and not train_decoder_only:
+    for name, p in core.named_parameters():
+        p.requires_grad = True
+        unfrozen_param_names.append(name)
+        loss_ratios["kl"] = float(kl_ratio)
+        trainable_module = core
+else:
+    # учим только 0-й блок декодера + post_quant_conv
+    if hasattr(core, "decoder"):
+        if train_up_only:#hasattr(core.decoder, "up_blocks") and len(core.decoder.up_blocks) > 0:
+            # --- только 0-й up_block ---
+            for name, p in core.decoder.up_blocks[0].named_parameters():
+                p.requires_grad = True
+                unfrozen_param_names.append(f"{name}")
+        else:
+            print("Decoder — fallback to full decoder")
+            for name, p in core.decoder.named_parameters():
+                p.requires_grad = True
+                unfrozen_param_names.append(f"decoder.{name}")
+    if hasattr(core, "post_quant_conv"):
+        for name, p in core.post_quant_conv.named_parameters():
+            p.requires_grad = True
+            unfrozen_param_names.append(f"post_quant_conv.{name}")
+    trainable_module = core.decoder if hasattr(core, "decoder") else core
+
+
+print(f"[INFO] Разморожено параметров: {len(unfrozen_param_names)}. Первые 200 имён:")
+for nm in unfrozen_param_names[:10]:
+    print(" ", nm)
+
+# --------------------------- Датасет ---------------------------
+from torch.utils.data import Dataset
+from PIL import Image, UnidentifiedImageError
+import random
+import torchvision.transforms as transforms
+import os
+
+class PngFolderDataset(Dataset):
+    def __init__(self, root_dir, resolution=1024, min_exts=('.png',), limit=0):
+        self.resolution = resolution
+        self.paths = []
+
+        for root, _, files in os.walk(root_dir):
+            for f in files:
+                if f.lower().endswith(tuple(ext.lower() for ext in min_exts)):
+                    self.paths.append(os.path.join(root, f))
+
+        if limit:
+            self.paths = self.paths[:limit]
+
+        # фильтруем недопустимые картинки
+        valid = []
+        for p in self.paths:
+            try:
+                with Image.open(p) as img:
+                    img.verify()  # только метаданные
+                    w, h = img.size
+                    if w < resolution or h < resolution:
+                        continue
+                valid.append(p)
+            except (OSError, UnidentifiedImageError):
+                continue
+
+        self.paths = valid
+        if not self.paths:
+            raise RuntimeError("No valid images found")
+
+        random.shuffle(self.paths)
+        self.transform = transforms.ToTensor()  # конвертирует сразу [0,1] float32
+
+    def __len__(self):
+        return len(self.paths)
+
+    def __getitem__(self, idx):
+        p = self.paths[idx % len(self.paths)]
+        with Image.open(p) as img:
+            img = img.convert("RGB")
+            return img
+
+def random_crop(img, sz):
+    w, h = img.size
+    x = random.randint(0, max(1, img.width - sz))
+    y = random.randint(0, max(1, img.height - sz))
+    return img.crop((x, y, x + sz, y + sz))
+
+tfm = transforms.Compose([
+    transforms.ToTensor(),
+    transforms.Normalize([0.5, 0.5, 0.5], [0.5, 0.5, 0.5])
+])
+
+def collate_fn(batch):
+    imgs = []
+    for img in batch:
+        img = random_crop(img, high_resolution)
+        imgs.append(tfm(img))
+    return torch.stack(imgs)
+
+dataset = PngFolderDataset(ds_path, min_exts=('.png',), resolution=high_resolution, limit=limit)
+print("len(dataset)",len(dataset))
+if len(dataset) < batch_size:
+    raise RuntimeError(f"Not enough valid images ({len(dataset)}) to form a batch of size {batch_size}")
+
+dataloader = DataLoader(dataset, batch_size=batch_size, shuffle=True, collate_fn=collate_fn, num_workers=num_workers, pin_memory=True, drop_last=True)
+
+# --------------------------- Оптимизатор ---------------------------
+def get_param_groups(module, weight_decay=0.001):
+    no_decay_tokens = ("bias", "norm", "rms", "layernorm")
+    decay_params, no_decay_params = [], []
+    for n, p in module.named_parameters():
+        if not p.requires_grad:
+            continue
+        n_l = n.lower()
+        if any(t in n_l for t in no_decay_tokens):
+            no_decay_params.append(p)
+        else:
+            decay_params.append(p)
+    return [
+        {"params": decay_params, "weight_decay": weight_decay},
+        {"params": no_decay_params, "weight_decay": 0.0},
+    ]
+
+def create_optimizer(name, param_groups):
+    if name == "adam8bit":
+        return bnb.optim.AdamW8bit(param_groups, lr=base_learning_rate, betas=(0.9, beta2), eps=eps)
+    raise ValueError(name)
+
+param_groups = get_param_groups(get_core_model(vae), weight_decay=0.001)
+optimizer = create_optimizer(optimizer_type, param_groups)
+
+# --------------------------- LR schedule ---------------------------
+batches_per_epoch = len(dataloader)
+steps_per_epoch = int(math.ceil(batches_per_epoch / float(gradient_accumulation_steps)))
+total_steps = steps_per_epoch * num_epochs
+
+def lr_lambda(step):
+    if not use_decay:
+        return 1.0
+    x = float(step) / float(max(1, total_steps))
+    warmup = float(warmup_percent)
+    min_ratio = float(min_learning_rate) / float(base_learning_rate)
+    if x < warmup:
+        return min_ratio + (1.0 - min_ratio) * (x / warmup)
+    decay_ratio = (x - warmup) / (1.0 - warmup)
+    return min_ratio + 0.5 * (1.0 - min_ratio) * (1.0 + math.cos(math.pi * decay_ratio))
+
+scheduler = LambdaLR(optimizer, lr_lambda)
+
+# Подготовка
+dataloader, vae, optimizer, scheduler = accelerator.prepare(dataloader, vae, optimizer, scheduler)
+trainable_params = [p for p in vae.parameters() if p.requires_grad]
+
+# fdl
+fdl_loss = FDL_loss()
+fdl_loss = fdl_loss.to(accelerator.device)
+
+# --------------------------- LPIPS и DSSIM ---------------------------
+_lpips_net = None
+def _get_lpips():
+    global _lpips_net
+    if _lpips_net is None:
+        _lpips_net = lpips.LPIPS(net='vgg', verbose=False).eval().to(accelerator.device).eval()
+    return _lpips_net
+
+# --- DSSIM Implementation (замена Sobel) ---
+def _gaussian_kernel(window_size, sigma, device, dtype):
+    coords = torch.arange(window_size, dtype=dtype, device=device) - (window_size - 1) / 2
+    k = torch.exp(-coords**2 / (2 * sigma**2))
+    return k / k.sum()
+
+def _ssim(x, y, window_size=11, sigma=1.5):
+    """Вычисляет SSIM для тензоров в диапазоне [-1, 1]"""
+    C = x.shape[1]
+    k = _gaussian_kernel(window_size, sigma, x.device, x.dtype)
+    win = (k.view(1, 1, -1, 1) * k.view(1, 1, 1, -1)).expand(C, 1, window_size, window_size).contiguous()
+    
+    mu1 = F.conv2d(x, win, padding=window_size//2, groups=C)
+    mu2 = F.conv2d(y, win, padding=window_size//2, groups=C)
+    mu1_sq = mu1.pow(2)
+    mu2_sq = mu2.pow(2)
+    mu1_mu2 = mu1 * mu2
+    
+    sigma1_sq = F.conv2d(x*x, win, padding=window_size//2, groups=C) - mu1_sq
+    sigma2_sq = F.conv2d(y*y, win, padding=window_size//2, groups=C) - mu2_sq
+    sigma12 = F.conv2d(x*y, win, padding=window_size//2, groups=C) - mu1_mu2
+    
+    L = 2.0  # диапазон для [-1, 1] -> 2
+    C1 = (0.01 * L)**2
+    C2 = (0.03 * L)**2
+    
+    num = (2*mu1_mu2 + C1) * (2*sigma12 + C2)
+    den = (mu1_sq + mu2_sq + C1) * (sigma1_sq + sigma2_sq + C2)
+    ssim_map = num / den
+    return ssim_map.mean()
+
+def dssim_loss(x, y):
+    """DSSIM = 1 - SSIM. Отлично сохраняет локальную структуру штрихов."""
+    return 1.0 - _ssim(x, y)
+    
+class MedianLossNormalizer:
+    def __init__(self, desired_ratios: dict, window_steps: int):
+        s = sum(desired_ratios.values())
+        self.ratios = {k: (v / s) if s > 0 else 0.0 for k, v in desired_ratios.items()}
+        self.buffers = {k: deque(maxlen=window_steps) for k in self.ratios.keys()}
+        self.window = window_steps
+
+    def update_and_total(self, abs_losses: dict):
+        for k, v in abs_losses.items():
+            if k in self.buffers:
+                self.buffers[k].append(float(v.detach().abs().cpu()))
+        meds = {k: (np.median(self.buffers[k]) if len(self.buffers[k]) > 0 else 1.0) for k in self.buffers}
+        coeffs = {k: (self.ratios[k] / max(meds[k], 1e-12)) for k in self.ratios}
+        total = sum(coeffs[k] * abs_losses[k] for k in abs_losses if k in coeffs)
+        return total, coeffs, meds
+
+if full_training and not train_decoder_only:
+    loss_ratios["kl"] = float(kl_ratio)
+normalizer = MedianLossNormalizer(loss_ratios, median_coeff_steps)
+
+# --------------------------- Сэмплы ---------------------------
+@torch.no_grad()
+def get_fixed_samples(n=3):
+    idx = random.sample(range(len(dataset)), min(n, len(dataset)))
+    pil_imgs = [dataset[i] for i in idx]
+    tensors = []
+    for img in pil_imgs:
+        img = random_crop(img, high_resolution)
+        tensors.append(tfm(img))
+    return torch.stack(tensors).to(accelerator.device, dtype)
+
+fixed_samples = get_fixed_samples()
+
+@torch.no_grad()
+def _to_pil_uint8(img_tensor: torch.Tensor) -> Image.Image:
+    arr = ((img_tensor.float().clamp(-1, 1) + 1.0) * 127.5).clamp(0, 255).byte().cpu().numpy().transpose(1, 2, 0)
+    return Image.fromarray(arr)
+
+
+@torch.no_grad()
+def generate_and_save_samples(step=None):
+    try:
+        #temp_vae = accelerator.unwrap_model(vae).eval()
+        if hasattr(vae, "module"):
+            # Если это DDP или DistributedDataParallel
+            unwrapped_vae = vae.module
+        else:
+            unwrapped_vae = vae
+        
+        # Если испол��зовался torch.compile, достаем оригинал
+        if hasattr(unwrapped_vae, "_orig_mod"):
+            temp_vae = unwrapped_vae._orig_mod
+        else:
+            temp_vae = unwrapped_vae
+        
+        temp_vae = temp_vae.eval()
+        lpips_net = _get_lpips()
+        with torch.no_grad():
+            orig_high = fixed_samples
+            orig_low = F.interpolate(
+                orig_high,
+                size=(model_resolution, model_resolution),
+                mode="bilinear",
+                align_corners=False
+            )
+            model_dtype = next(temp_vae.parameters()).dtype
+            orig_low = orig_low.to(dtype=model_dtype)
+
+            # Encode/decode с учётом видео-режима
+            if is_video_vae(temp_vae):
+                x_in = orig_low.unsqueeze(2)  # [B,3,1,H,W]
+                enc = temp_vae.encode(x_in)
+                latents_mean = enc.latent_dist.mean
+                dec = temp_vae.decode(latents_mean).sample  # [B,3,1,H,W]
+                rec = dec.squeeze(2)  # [B,3,H,W]
+            else:
+                enc = temp_vae.encode(orig_low)
+                latents_mean = enc.latent_dist.mean
+                rec = temp_vae.decode(latents_mean).sample
+
+        # Подгон размеров, если надо
+        #if rec.shape[-2:] != orig_high.shape[-2:]:
+        #    rec = F.interpolate(rec, size=orig_high.shape[-2:], mode="bilinear", align_corners=False)
+
+        # Сохраняем все real/decoded
+        for i in range(rec.shape[0]):
+            real_img = _to_pil_uint8(orig_high[i])
+            dec_img  = _to_pil_uint8(rec[i])
+            real_img.save(f"{generated_folder}/sample_real_{i}.png")
+            dec_img.save(f"{generated_folder}/sample_decoded_{i}.png")
+
+        # LPIPS
+        lpips_scores = []
+        for i in range(rec.shape[0]):
+            orig_full = orig_high[i:i+1].to(torch.float32)
+            rec_full  = rec[i:i+1].to(torch.float32)
+            #if rec_full.shape[-2:] != orig_full.shape[-2:]:
+            #    rec_full = F.interpolate(rec_full, size=orig_full.shape[-2:], mode="bilinear", align_corners=False)
+            lpips_val = lpips_net(orig_full, rec_full).item()
+            lpips_scores.append(lpips_val)
+        avg_lpips = float(np.mean(lpips_scores))
+
+        # W&B логирование
+        if use_wandb and accelerator.is_main_process:
+            log_data = {"lpips_mean": avg_lpips}
+            for i in range(rec.shape[0]):
+                log_data[f"sample/real_{i}"] = wandb.Image(f"{generated_folder}/sample_real_{i}.png", caption=f"real_{i}")
+                log_data[f"sample/decoded_{i}"] = wandb.Image(f"{generated_folder}/sample_decoded_{i}.png", caption=f"decoded_{i}")
+            wandb.log(log_data, step=step)
+
+    finally:
+        gc.collect()
+        torch.cuda.empty_cache()
+
+
+if accelerator.is_main_process and save_model:
+    print("Генерация сэмплов до старта обучения...")
+    generate_and_save_samples(0)
+
+accelerator.wait_for_everyone()
+
+# --------------------------- Тренировка ---------------------------
+progress = tqdm(total=total_steps, disable=not accelerator.is_local_main_process)
+global_step = 0
+min_loss = float("inf")
+sample_interval = max(1, total_steps // max(1, sample_interval_share * num_epochs))
+
+for epoch in range(num_epochs):
+    vae.train()
+    batch_losses, batch_grads = [], []
+    track_losses = {k: [] for k in loss_ratios.keys()}
+
+    for imgs in dataloader:
+        with accelerator.accumulate(vae):
+            imgs = imgs.to(accelerator.device)
+
+            if high_resolution != model_resolution:
+                imgs_low = F.interpolate(imgs, size=(model_resolution, model_resolution),mode="area") # mode="bilinear", align_corners=False)
+            else:
+                imgs_low = imgs
+
+            model_dtype = next(vae.parameters()).dtype
+            imgs_low_model = imgs_low.to(dtype=model_dtype) if imgs_low.dtype != model_dtype else imgs_low
+
+            # Вместо: current_vae = accelerator.unwrap_model(vae)
+            unwrapped = vae.module if hasattr(vae, "module") else vae
+            current_vae = getattr(unwrapped, "_orig_mod", unwrapped)
+
+
+            # QWEN: encode/decode с T=1
+            if is_video_vae(current_vae):
+                x_in = imgs_low_model.unsqueeze(2)             # [B,3,1,H,W]
+                enc = current_vae.encode(x_in)
+                latents = enc.latent_dist.mean if train_decoder_only else enc.latent_dist.sample()
+                dec = current_vae.decode(latents).sample               # [B,3,1,H,W]
+                rec = dec.squeeze(2)                           # [B,3,H,W]
+            else:
+                enc = current_vae.encode(imgs_low_model)
+                latents = enc.latent_dist.mean if train_decoder_only else enc.latent_dist.sample()
+                rec = current_vae.decode(latents).sample
+
+            rec_f32 = rec.to(torch.float32)
+            imgs_f32 = imgs.to(torch.float32)
+
+            # --- Расчёт лоссов ---
+            abs_losses = {
+                "mae":   F.l1_loss(rec_f32, imgs_f32),
+                "mse":   torch.tensor(0.0, device=accelerator.device, dtype=torch.float32), # Отключён, экономим вычисления
+                "lpips": _get_lpips()(rec_f32, imgs_f32).mean(),
+                "fdl":   fdl_loss(rec_f32, imgs_f32),
+                "dssim": dssim_loss(rec_f32, imgs_f32), # <-- DSSIM вместо Edge
+            }
+
+            if full_training and not train_decoder_only:
+                mean   = enc.latent_dist.mean
+                logvar = enc.latent_dist.logvar
+                kl = -0.5 * torch.mean(1 + logvar - mean.pow(2) - logvar.exp())
+                abs_losses["kl"] = kl
+            else:
+                abs_losses["kl"] = torch.tensor(0.0, device=accelerator.device, dtype=torch.float32)
+
+            total_loss, coeffs, meds = normalizer.update_and_total(abs_losses)
+
+            if torch.isnan(total_loss) or torch.isinf(total_loss):
+                raise RuntimeError("NaN/Inf loss")
+
+            accelerator.backward(total_loss)
+
+            grad_norm = torch.tensor(0.0, device=accelerator.device)
+            if accelerator.sync_gradients:
+                grad_norm = accelerator.clip_grad_norm_(trainable_params, clip_grad_norm)
+                optimizer.step()
+                scheduler.step()
+                optimizer.zero_grad(set_to_none=True)
+                global_step += 1
+                progress.update(1)
+
+            if accelerator.is_main_process:
+                try:
+                    current_lr = optimizer.param_groups[0]["lr"]
+                except Exception:
+                    current_lr = scheduler.get_last_lr()[0]
+
+                batch_losses.append(total_loss.detach().item())
+                batch_grads.append(float(grad_norm.detach().cpu().item()) if isinstance(grad_norm, torch.Tensor) else float(grad_norm))
+                for k, v in abs_losses.items():
+                    track_losses[k].append(float(v.detach().item()))
+
+                if use_wandb and accelerator.sync_gradients:
+                    log_dict = {
+                        "total_loss": float(total_loss.detach().item()),
+                        "learning_rate": current_lr,
+                        "epoch": epoch,
+                        "grad_norm": batch_grads[-1],
+                    }
+                    for k, v in abs_losses.items():
+                        log_dict[f"loss_{k}"] = float(v.detach().item())
+                    for k in coeffs:
+                        log_dict[f"coeff_{k}"] = float(coeffs[k])
+                        log_dict[f"median_{k}"] = float(meds[k])
+                    wandb.log(log_dict, step=global_step)
+
+            if global_step > 0 and global_step % sample_interval == 0:
+                if accelerator.is_main_process:
+                    generate_and_save_samples(global_step)
+                accelerator.wait_for_everyone()
+
+                n_micro = sample_interval * gradient_accumulation_steps
+                avg_loss = float(np.mean(batch_losses[-n_micro:])) if len(batch_losses) >= n_micro else float(np.mean(batch_losses)) if batch_losses else float("nan")
+                avg_grad = float(np.mean(batch_grads[-n_micro:])) if len(batch_grads) >= 1 else float(np.mean(batch_grads)) if batch_grads else 0.0
+
+                if accelerator.is_main_process:
+                    print(f"Epoch {epoch} step {global_step} loss: {avg_loss:.6f}, grad_norm: {avg_grad:.6f}, lr: {current_lr:.9f}")
+                    if save_model and avg_loss < min_loss * save_barrier:
+                        min_loss = avg_loss
+                        unwrapped = vae.module if hasattr(vae, "module") else vae
+                        current_vae = getattr(unwrapped, "_orig_mod", unwrapped)
+                        current_vae.save_pretrained(save_as)
+                    if use_wandb:
+                        wandb.log({"interm_loss": avg_loss, "interm_grad": avg_grad}, step=global_step)
+
+    if accelerator.is_main_process:
+        epoch_avg = float(np.mean(batch_losses)) if batch_losses else float("nan")
+        print(f"Epoch {epoch} done, avg loss {epoch_avg:.6f}")
+        if use_wandb:
+            wandb.log({"epoch_loss": epoch_avg, "epoch": epoch + 1}, step=global_step)
+
+# --------------------------- Финальное сохранение ---------------------------
+if accelerator.is_main_process:
+    print("Training finished – saving final model")
+    if save_model:
+        unwrapped = vae.module if hasattr(vae, "module") else vae
+        current_vae = getattr(unwrapped, "_orig_mod", unwrapped)
+        current_vae.save_pretrained(save_as)
+
+accelerator.free_memory()
+if torch.distributed.is_initialized():
+    torch.distributed.destroy_process_group()
+print("Готово!")