Upload trainer.py

QA_result/trainer.py

@@ -0,0 +1,1208 @@
+# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license
+"""
+Train a model on a dataset.
+
+Usage:
+    $ yolo mode=train model=yolo11n.pt data=coco8.yaml imgsz=640 epochs=100 batch=16
+"""
+
+import gc
+import math
+import os
+import subprocess
+import time
+import warnings
+from copy import copy, deepcopy
+from datetime import datetime, timedelta
+from pathlib import Path
+
+import numpy as np
+import torch
+import torch.nn.functional as F
+from torch import distributed as dist
+from torch import nn, optim
+
+from ultralytics import __version__
+from ultralytics.cfg import get_cfg, get_save_dir
+from ultralytics.data.utils import check_cls_dataset, check_det_dataset
+from ultralytics.nn.tasks import attempt_load_one_weight, attempt_load_weights
+from ultralytics.utils import (
+    DEFAULT_CFG,
+    LOCAL_RANK,
+    LOGGER,
+    RANK,
+    TQDM,
+    YAML,
+    callbacks,
+    clean_url,
+    colorstr,
+    emojis,
+)
+from ultralytics.utils.autobatch import check_train_batch_size
+from ultralytics.utils.checks import check_amp, check_file, check_imgsz, check_model_file_from_stem, print_args
+from ultralytics.utils.dist import ddp_cleanup, generate_ddp_command
+from ultralytics.utils.files import get_latest_run
+from ultralytics.utils.torch_utils import (
+    TORCH_2_4,
+    EarlyStopping,
+    ModelEMA,
+    autocast,
+    convert_optimizer_state_dict_to_fp16,
+    init_seeds,
+    one_cycle,
+    select_device,
+    strip_optimizer,
+    torch_distributed_zero_first,
+    unset_deterministic,
+)
+
+class CWDLoss(nn.Module):
+    """PyTorch version of `Channel-wise Distillation for Semantic Segmentation.
+    <https://arxiv.org/abs/2011.13256>`_.
+    """
+
+    def __init__(self, channels_s, channels_t, tau=1.0):
+        super().__init__()
+        self.tau = tau
+
+    def forward(self, y_s, y_t):
+        """Forward computation.
+        Args:
+            y_s (list): The student model prediction with
+                shape (N, C, H, W) in list.
+            y_t (list): The teacher model prediction with
+                shape (N, C, H, W) in list.
+        Return:
+            torch.Tensor: The calculated loss value of all stages.
+        """
+        assert len(y_s) == len(y_t)
+        losses = []
+
+        for idx, (s, t) in enumerate(zip(y_s, y_t)):
+            assert s.shape == t.shape
+            N, C, H, W = s.shape
+
+            # normalize in channel dimension
+            softmax_pred_T = F.softmax(t.view(-1, W * H) / self.tau, dim=1)
+
+            logsoftmax = torch.nn.LogSoftmax(dim=1)
+            cost = torch.sum(
+                softmax_pred_T * logsoftmax(t.view(-1, W * H) / self.tau) -
+                softmax_pred_T * logsoftmax(s.view(-1, W * H) / self.tau)) * (self.tau ** 2)
+
+            losses.append(cost / (C * N))
+        loss = sum(losses)
+        return loss
+
+class MGDLoss(nn.Module):
+    def __init__(self,
+                 student_channels,
+                 teacher_channels,
+                 alpha_mgd=0.00002,
+                 lambda_mgd=0.65,
+                 ):
+        super(MGDLoss, self).__init__()
+        self.alpha_mgd = alpha_mgd
+        self.lambda_mgd = lambda_mgd
+        device = 'cuda' if torch.cuda.is_available() else 'cpu'
+
+        self.generation = nn.ModuleList([
+            nn.Sequential(
+                nn.Conv2d(s_chan, t_chan, kernel_size=3, padding=1),
+                nn.ReLU(inplace=True),
+                nn.Conv2d(t_chan, t_chan, kernel_size=3, padding=1)
+            ).to(device) for s_chan, t_chan in zip(student_channels, teacher_channels)
+        ])
+
+    def forward(self, y_s, y_t, layer=None):
+        """Forward computation.
+        Args:
+            y_s (list): The student model prediction with
+                shape (N, C, H, W) in list.
+            y_t (list): The teacher model prediction with
+                shape (N, C, H, W) in list.
+        Return:
+            torch.Tensor: The calculated loss value of all stages.
+        """
+        losses = []
+        for idx, (s, t) in enumerate(zip(y_s, y_t)):
+            # print(s.shape)
+            # print(t.shape)
+            # assert s.shape == t.shape
+            if layer == "outlayer":
+                idx = -1
+            losses.append(self.get_dis_loss(s, t, idx) * self.alpha_mgd)
+        loss = sum(losses)
+        return loss
+
+    def get_dis_loss(self, preds_S, preds_T, idx):
+        loss_mse = nn.MSELoss(reduction='sum')
+        N, C, H, W = preds_T.shape
+
+        device = preds_S.device
+        mat = torch.rand((N, 1, H, W)).to(device)
+        mat = torch.where(mat > 1 - self.lambda_mgd, 0, 1).to(device)
+
+        masked_fea = torch.mul(preds_S, mat)
+        new_fea = self.generation[idx](masked_fea)
+
+        dis_loss = loss_mse(new_fea, preds_T) / N
+        return dis_loss
+
+
+class FeatureLoss(nn.Module):
+    def __init__(self, channels_s, channels_t, distiller='mgd', loss_weight=1.0):
+        super(FeatureLoss, self).__init__()
+        self.loss_weight = loss_weight
+        self.distiller = distiller
+
+        # Move all modules to same precision
+        device = 'cuda' if torch.cuda.is_available() else 'cpu'
+
+        # Convert to ModuleList and ensure consistent dtype
+        self.align_module = nn.ModuleList()
+        self.norm = nn.ModuleList()
+        self.norm1 = nn.ModuleList()
+
+        # Create alignment modules
+        for s_chan, t_chan in zip(channels_s, channels_t):
+            align = nn.Sequential(
+                nn.Conv2d(s_chan, t_chan, kernel_size=1, stride=1, padding=0),
+                nn.BatchNorm2d(t_chan, affine=False)
+            ).to(device)
+            self.align_module.append(align)
+
+        # Create normalization layers
+        for t_chan in channels_t:
+            self.norm.append(nn.BatchNorm2d(t_chan, affine=False).to(device))
+
+        for s_chan in channels_s:
+            self.norm1.append(nn.BatchNorm2d(s_chan, affine=False).to(device))
+
+        if distiller == 'mgd':
+            self.feature_loss = MGDLoss(channels_s, channels_t)
+        elif distiller == 'cwd':
+            self.feature_loss = CWDLoss(channels_s, channels_t)
+        else:
+            raise NotImplementedError
+
+    def forward(self, y_s, y_t):
+        min_len = min(len(y_s), len(y_t))
+        y_s = y_s[:min_len]
+        y_t = y_t[:min_len]
+
+        tea_feats = []
+        stu_feats = []
+
+        for idx, (s, t) in enumerate(zip(y_s, y_t)):
+            s = s.type(next(self.align_module[idx].parameters()).dtype)
+            t = t.type(next(self.align_module[idx].parameters()).dtype)
+
+            if self.distiller == "cwd":
+                s = self.align_module[idx](s)
+                stu_feats.append(s)
+                tea_feats.append(t.detach())
+            else:
+                t = self.norm[idx](t)  # ✅ Correct normalization
+                stu_feats.append(s)
+                tea_feats.append(t.detach())
+
+        loss = self.feature_loss(stu_feats, tea_feats)
+        return self.loss_weight * loss
+
+
+class DistillationLoss:
+    def __init__(self, models, modelt, distiller="CWDLoss"):
+        self.distiller = distiller
+        self.layers = ["6", "8", "13", "16", "19", "22"]
+        self.models = models
+        self.modelt = modelt
+
+        device = 'cuda' if torch.cuda.is_available() else 'cpu'
+        # ini warm up
+        with torch.no_grad():
+            dummy_input = torch.randn(1, 3, 640, 640)
+            _ = self.models(dummy_input.to(device))
+            _ = self.modelt(dummy_input.to(device))
+
+        self.channels_s = []
+        self.channels_t = []
+        self.teacher_module_pairs = []
+        self.student_module_pairs = []
+        self.remove_handle = []
+
+        self._find_layers()
+
+        self.distill_loss_fn = FeatureLoss(
+            channels_s=self.channels_s,
+            channels_t=self.channels_t,
+            distiller=distiller[:3]
+        )
+
+    def _find_layers(self):
+
+        self.channels_s = []
+        self.channels_t = []
+        self.teacher_module_pairs = []
+        self.student_module_pairs = []
+
+        for name, ml in self.modelt.named_modules():
+            if name is not None:
+                name = name.split(".")
+                # print(name)
+
+                if name[0] != "model":
+                    continue
+                if len(name) >= 3:
+                    if name[1] in self.layers:
+                        if "cv2" in name[2]:
+                            if hasattr(ml, 'conv'):
+                                self.channels_t.append(ml.conv.out_channels)
+                                self.teacher_module_pairs.append(ml)
+        # print()
+        for name, ml in self.models.named_modules():
+            if name is not None:
+                name = name.split(".")
+                # print(name)
+                if name[0] != "model":
+                    continue
+                if len(name) >= 3:
+                    if name[1] in self.layers:
+                        if "cv2" in name[2]:
+                            if hasattr(ml, 'conv'):
+                                self.channels_s.append(ml.conv.out_channels)
+                                self.student_module_pairs.append(ml)
+
+        nl = min(len(self.channels_s), len(self.channels_t))
+        self.channels_s = self.channels_s[-nl:]
+        self.channels_t = self.channels_t[-nl:]
+        self.teacher_module_pairs = self.teacher_module_pairs[-nl:]
+        self.student_module_pairs = self.student_module_pairs[-nl:]
+
+    def register_hook(self):
+        # Remove the existing hook if they exist
+        self.remove_handle_()
+
+        self.teacher_outputs = []
+        self.student_outputs = []
+
+        def make_student_hook(l):
+            def forward_hook(m, input, output):
+                if isinstance(output, torch.Tensor):
+                    out = output.clone()  # Clone to ensure we don't modify the original
+                    l.append(out)
+                else:
+                    l.append([o.clone() if isinstance(o, torch.Tensor) else o for o in output])
+            return forward_hook
+
+        def make_teacher_hook(l):
+            def forward_hook(m, input, output):
+                if isinstance(output, torch.Tensor):
+                    l.append(output.detach().clone())  # Detach and clone teacher outputs
+                else:
+                    l.append([o.detach().clone() if isinstance(o, torch.Tensor) else o for o in output])
+            return forward_hook
+
+        for ml, ori in zip(self.teacher_module_pairs, self.student_module_pairs):
+            self.remove_handle.append(ml.register_forward_hook(make_teacher_hook(self.teacher_outputs)))
+            self.remove_handle.append(ori.register_forward_hook(make_student_hook(self.student_outputs)))
+
+    def get_loss(self):
+        if not self.teacher_outputs or not self.student_outputs:
+            return torch.tensor(0.0, requires_grad=True)
+
+        if len(self.teacher_outputs) != len(self.student_outputs):
+            print(f"Warning: Mismatched outputs - Teacher: {len(self.teacher_outputs)}, Student: {len(self.student_outputs)}")
+            return torch.tensor(0.0, requires_grad=True)
+
+        quant_loss = self.distill_loss_fn(y_s=self.student_outputs, y_t=self.teacher_outputs)
+
+        if self.distiller != 'cwd':
+            quant_loss *= 0.3
+
+        self.teacher_outputs.clear()
+        self.student_outputs.clear()
+
+        return quant_loss
+
+    def remove_handle_(self):
+        for rm in self.remove_handle:
+            rm.remove()
+        self.remove_handle.clear()
+
+
+
+class BaseTrainer:
+    """
+    A base class for creating trainers.
+
+    This class provides the foundation for training YOLO models, handling the training loop, validation, checkpointing,
+    and various training utilities. It supports both single-GPU and multi-GPU distributed training.
+
+    Attributes:
+        args (SimpleNamespace): Configuration for the trainer.
+        validator (BaseValidator): Validator instance.
+        model (nn.Module): Model instance.
+        callbacks (defaultdict): Dictionary of callbacks.
+        save_dir (Path): Directory to save results.
+        wdir (Path): Directory to save weights.
+        last (Path): Path to the last checkpoint.
+        best (Path): Path to the best checkpoint.
+        save_period (int): Save checkpoint every x epochs (disabled if < 1).
+        batch_size (int): Batch size for training.
+        epochs (int): Number of epochs to train for.
+        start_epoch (int): Starting epoch for training.
+        device (torch.device): Device to use for training.
+        amp (bool): Flag to enable AMP (Automatic Mixed Precision).
+        scaler (amp.GradScaler): Gradient scaler for AMP.
+        data (str): Path to data.
+        ema (nn.Module): EMA (Exponential Moving Average) of the model.
+        resume (bool): Resume training from a checkpoint.
+        lf (nn.Module): Loss function.
+        scheduler (torch.optim.lr_scheduler._LRScheduler): Learning rate scheduler.
+        best_fitness (float): The best fitness value achieved.
+        fitness (float): Current fitness value.
+        loss (float): Current loss value.
+        tloss (float): Total loss value.
+        loss_names (list): List of loss names.
+        csv (Path): Path to results CSV file.
+        metrics (dict): Dictionary of metrics.
+        plots (dict): Dictionary of plots.
+
+    Methods:
+        train: Execute the training process.
+        validate: Run validation on the test set.
+        save_model: Save model training checkpoints.
+        get_dataset: Get train and validation datasets.
+        setup_model: Load, create, or download model.
+        build_optimizer: Construct an optimizer for the model.
+
+    Examples:
+        Initialize a trainer and start training
+        >>> trainer = BaseTrainer(cfg="config.yaml")
+        >>> trainer.train()
+    """
+
+    def __init__(self, cfg=DEFAULT_CFG, overrides=None, _callbacks=None):
+        """
+        Initialize the BaseTrainer class.
+
+        Args:
+            cfg (str, optional): Path to a configuration file.
+            overrides (dict, optional): Configuration overrides.
+            _callbacks (list, optional): List of callback functions.
+        """
+        self.args = get_cfg(cfg, overrides)
+        self.check_resume(overrides)
+        self.device = select_device(self.args.device, self.args.batch)
+        # Update "-1" devices so post-training val does not repeat search
+        self.args.device = os.getenv("CUDA_VISIBLE_DEVICES") if "cuda" in str(self.device) else str(self.device)
+        self.validator = None
+        self.metrics = None
+        self.plots = {}
+
+        if overrides:
+            self.teacher = overrides.get("teacher", None)
+            self.loss_type = overrides.get("distillation_loss", None)
+            if "teacher" in overrides:
+                overrides.pop("teacher")
+            if "distillation_loss" in overrides:
+                overrides.pop("distillation_loss")
+        else:
+            self.loss_type = None
+            self.teacher = None
+
+        init_seeds(self.args.seed + 1 + RANK, deterministic=self.args.deterministic)
+
+        # Dirs
+        self.save_dir = get_save_dir(self.args)
+        self.args.name = self.save_dir.name  # update name for loggers
+        self.wdir = self.save_dir / "weights"  # weights dir
+        if RANK in {-1, 0}:
+            self.wdir.mkdir(parents=True, exist_ok=True)  # make dir
+            self.args.save_dir = str(self.save_dir)
+            YAML.save(self.save_dir / "args.yaml", vars(self.args))  # save run args
+        self.last, self.best = self.wdir / "last.pt", self.wdir / "best.pt"  # checkpoint paths
+        self.save_period = self.args.save_period
+
+        self.batch_size = self.args.batch
+        self.epochs = self.args.epochs or 100  # in case users accidentally pass epochs=None with timed training
+        self.start_epoch = 0
+        if RANK == -1:
+            print_args(vars(self.args))
+
+        # Device
+        if self.device.type in {"cpu", "mps"}:
+            self.args.workers = 0  # faster CPU training as time dominated by inference, not dataloading
+
+        # Model and Dataset
+        self.model = check_model_file_from_stem(self.args.model)  # add suffix, i.e. yolo11n -> yolo11n.pt
+        with torch_distributed_zero_first(LOCAL_RANK):  # avoid auto-downloading dataset multiple times
+            self.data = self.get_dataset()
+
+        self.ema = None
+
+        # Optimization utils init
+        self.lf = None
+        self.scheduler = None
+
+        # Epoch level metrics
+        self.best_fitness = None
+        self.fitness = None
+        self.loss = None
+        self.tloss = None
+        self.loss_names = ["Loss"]
+        self.csv = self.save_dir / "results.csv"
+        self.plot_idx = [0, 1, 2]
+
+        # HUB
+        self.hub_session = None
+
+        # Callbacks
+        self.callbacks = _callbacks or callbacks.get_default_callbacks()
+        if RANK in {-1, 0}:
+            callbacks.add_integration_callbacks(self)
+
+    def add_callback(self, event: str, callback):
+        """Append the given callback to the event's callback list."""
+        self.callbacks[event].append(callback)
+
+    def set_callback(self, event: str, callback):
+        """Override the existing callbacks with the given callback for the specified event."""
+        self.callbacks[event] = [callback]
+
+    def run_callbacks(self, event: str):
+        """Run all existing callbacks associated with a particular event."""
+        for callback in self.callbacks.get(event, []):
+            callback(self)
+
+    def train(self):
+        """Allow device='', device=None on Multi-GPU systems to default to device=0."""
+        if isinstance(self.args.device, str) and len(self.args.device):  # i.e. device='0' or device='0,1,2,3'
+            world_size = len(self.args.device.split(","))
+        elif isinstance(self.args.device, (tuple, list)):  # i.e. device=[0, 1, 2, 3] (multi-GPU from CLI is list)
+            world_size = len(self.args.device)
+        elif self.args.device in {"cpu", "mps"}:  # i.e. device='cpu' or 'mps'
+            world_size = 0
+        elif torch.cuda.is_available():  # i.e. device=None or device='' or device=number
+            world_size = 1  # default to device 0
+        else:  # i.e. device=None or device=''
+            world_size = 0
+
+        # Run subprocess if DDP training, else train normally
+        if world_size > 1 and "LOCAL_RANK" not in os.environ:
+            # Argument checks
+            if self.args.rect:
+                LOGGER.warning("'rect=True' is incompatible with Multi-GPU training, setting 'rect=False'")
+                self.args.rect = False
+            if self.args.batch < 1.0:
+                LOGGER.warning(
+                    "'batch<1' for AutoBatch is incompatible with Multi-GPU training, setting default 'batch=16'"
+                )
+                self.args.batch = 16
+
+            # Command
+            cmd, file = generate_ddp_command(world_size, self)
+            try:
+                LOGGER.info(f"{colorstr('DDP:')} debug command {' '.join(cmd)}")
+                subprocess.run(cmd, check=True)
+            except Exception as e:
+                raise e
+            finally:
+                ddp_cleanup(self, str(file))
+
+        else:
+            self._do_train(world_size)
+
+    def _setup_scheduler(self):
+        """Initialize training learning rate scheduler."""
+        if self.args.cos_lr:
+            self.lf = one_cycle(1, self.args.lrf, self.epochs)  # cosine 1->hyp['lrf']
+        else:
+            self.lf = lambda x: max(1 - x / self.epochs, 0) * (1.0 - self.args.lrf) + self.args.lrf  # linear
+        self.scheduler = optim.lr_scheduler.LambdaLR(self.optimizer, lr_lambda=self.lf)
+
+    def _setup_ddp(self, world_size):
+        """Initialize and set the DistributedDataParallel parameters for training."""
+        torch.cuda.set_device(RANK)
+        self.device = torch.device("cuda", RANK)
+        # LOGGER.info(f'DDP info: RANK {RANK}, WORLD_SIZE {world_size}, DEVICE {self.device}')
+        os.environ["TORCH_NCCL_BLOCKING_WAIT"] = "1"  # set to enforce timeout
+        dist.init_process_group(
+            backend="nccl" if dist.is_nccl_available() else "gloo",
+            timeout=timedelta(seconds=10800),  # 3 hours
+            rank=RANK,
+            world_size=world_size,
+        )
+
+    def _setup_train(self, world_size):
+        """Build dataloaders and optimizer on correct rank process."""
+        # Model
+        self.run_callbacks("on_pretrain_routine_start")
+        ckpt = self.setup_model()
+        self.model = self.model.to(self.device)
+
+        # Load teacher model to device
+        if self.teacher is not None:
+            for k, v in self.teacher.named_parameters():
+                v.requires_grad = True
+            self.teacher = self.teacher.to(self.device)
+
+        self.set_model_attributes()
+
+        # Freeze layers
+        freeze_list = (
+            self.args.freeze
+            if isinstance(self.args.freeze, list)
+            else range(self.args.freeze)
+            if isinstance(self.args.freeze, int)
+            else []
+        )
+        always_freeze_names = [".dfl"]  # always freeze these layers
+        freeze_layer_names = [f"model.{x}." for x in freeze_list] + always_freeze_names
+        self.freeze_layer_names = freeze_layer_names
+        for k, v in self.model.named_parameters():
+            # v.register_hook(lambda x: torch.nan_to_num(x))  # NaN to 0 (commented for erratic training results)
+            if any(x in k for x in freeze_layer_names):
+                LOGGER.info(f"Freezing layer '{k}'")
+                v.requires_grad = False
+            elif not v.requires_grad and v.dtype.is_floating_point:  # only floating point Tensor can require gradients
+                LOGGER.warning(
+                    f"setting 'requires_grad=True' for frozen layer '{k}'. "
+                    "See ultralytics.engine.trainer for customization of frozen layers."
+                )
+                v.requires_grad = True
+
+        # Check AMP
+        self.amp = torch.tensor(self.args.amp).to(self.device)  # True or False
+        if self.amp and RANK in {-1, 0}:  # Single-GPU and DDP
+            callbacks_backup = callbacks.default_callbacks.copy()  # backup callbacks as check_amp() resets them
+            self.amp = torch.tensor(check_amp(self.model), device=self.device)
+            callbacks.default_callbacks = callbacks_backup  # restore callbacks
+        if RANK > -1 and world_size > 1:  # DDP
+            dist.broadcast(self.amp.int(), src=0)  # broadcast from rank 0 to all other ranks; gloo errors with boolean
+        self.amp = bool(self.amp)  # as boolean
+        self.scaler = (
+            torch.amp.GradScaler("cuda", enabled=self.amp) if TORCH_2_4 else torch.cuda.amp.GradScaler(enabled=self.amp)
+        )
+        if world_size > 1:
+            self.model = nn.parallel.DistributedDataParallel(self.model, device_ids=[RANK], find_unused_parameters=True)
+            if self.teacher is not None:
+                self.teacher = nn.parallel.DistributedDataParallel(self.teacher, device_ids=[RANK])
+                temp = self.teacher.eval()
+
+        # Check imgsz
+        gs = max(int(self.model.stride.max() if hasattr(self.model, "stride") else 32), 32)  # grid size (max stride)
+        self.args.imgsz = check_imgsz(self.args.imgsz, stride=gs, floor=gs, max_dim=1)
+        self.stride = gs  # for multiscale training
+
+        # Batch size
+        if self.batch_size < 1 and RANK == -1:  # single-GPU only, estimate best batch size
+            self.args.batch = self.batch_size = self.auto_batch()
+
+
+        # Dataloaders
+        batch_size = self.batch_size // max(world_size, 1)
+        self.train_loader = self.get_dataloader(
+            self.data["train"], batch_size=batch_size, rank=LOCAL_RANK, mode="train"
+        )
+        if RANK in {-1, 0}:
+            # Note: When training DOTA dataset, double batch size could get OOM on images with >2000 objects.
+            self.test_loader = self.get_dataloader(
+                self.data.get("val") or self.data.get("test"),
+                batch_size=batch_size if self.args.task == "obb" else batch_size * 2,
+                rank=-1,
+                mode="val",
+            )
+            self.validator = self.get_validator()
+            metric_keys = self.validator.metrics.keys + self.label_loss_items(prefix="val")
+            self.metrics = dict(zip(metric_keys, [0] * len(metric_keys)))
+            self.ema = ModelEMA(self.model)
+            if self.args.plots:
+                self.plot_training_labels()
+
+        # Optimizer
+        self.accumulate = max(round(self.args.nbs / self.batch_size), 1)  # accumulate loss before optimizing
+        weight_decay = self.args.weight_decay * self.batch_size * self.accumulate / self.args.nbs  # scale weight_decay
+        iterations = math.ceil(len(self.train_loader.dataset) / max(self.batch_size, self.args.nbs)) * self.epochs
+        self.optimizer = self.build_optimizer(
+            model=self.model,
+            teacher=self.teacher,
+            name=self.args.optimizer,
+            lr=self.args.lr0,
+            momentum=self.args.momentum,
+            decay=weight_decay,
+            iterations=iterations,
+        )
+        # Scheduler
+        self._setup_scheduler()
+        self.stopper, self.stop = EarlyStopping(patience=self.args.patience), False
+        self.resume_training(ckpt)
+        self.scheduler.last_epoch = self.start_epoch - 1  # do not move
+        self.run_callbacks("on_pretrain_routine_end")
+
+    def _do_train(self, world_size=1):
+        """Train the model with the specified world size."""
+        if world_size > 1:
+            self._setup_ddp(world_size)
+        self._setup_train(world_size)
+
+        nb = len(self.train_loader)  # number of batches
+        nw = max(round(self.args.warmup_epochs * nb), 100) if self.args.warmup_epochs > 0 else -1  # warmup iterations
+        last_opt_step = -1
+        self.epoch_time = None
+        self.epoch_time_start = time.time()
+        self.train_time_start = time.time()
+        self.run_callbacks("on_train_start")
+        LOGGER.info(
+            f"Image sizes {self.args.imgsz} train, {self.args.imgsz} val\n"
+            f"Using {self.train_loader.num_workers * (world_size or 1)} dataloader workers\n"
+            f"Logging results to {colorstr('bold', self.save_dir)}\n"
+            f"Starting training for " + (f"{self.args.time} hours..." if self.args.time else f"{self.epochs} epochs...")
+        )
+        if self.args.close_mosaic:
+            base_idx = (self.epochs - self.args.close_mosaic) * nb
+            self.plot_idx.extend([base_idx, base_idx + 1, base_idx + 2])
+
+        # make loss
+        if self.teacher is not None:
+            distillation_loss = DistillationLoss(self.model, self.teacher, distiller=self.loss_type)
+
+        epoch = self.start_epoch
+        self.optimizer.zero_grad()  # zero any resumed gradients to ensure stability on train start
+        while True:
+            self.epoch = epoch
+            self.run_callbacks("on_train_epoch_start")
+            with warnings.catch_warnings():
+                warnings.simplefilter("ignore")  # suppress 'Detected lr_scheduler.step() before optimizer.step()'
+                self.scheduler.step()
+
+            self._model_train()
+            if RANK != -1:
+                self.train_loader.sampler.set_epoch(epoch)
+            pbar = enumerate(self.train_loader)
+            # Update dataloader attributes (optional)
+            if epoch == (self.epochs - self.args.close_mosaic):
+                self._close_dataloader_mosaic()
+                self.train_loader.reset()
+
+            if RANK in {-1, 0}:
+                LOGGER.info(self.progress_string())
+                pbar = TQDM(enumerate(self.train_loader), total=nb)
+            self.tloss = None
+            if self.teacher is not None:
+                distillation_loss.register_hook()
+            for i, batch in pbar:
+                self.run_callbacks("on_train_batch_start")
+                # Warmup
+                ni = i + nb * epoch
+                if ni <= nw:
+                    xi = [0, nw]  # x interp
+                    self.accumulate = max(1, int(np.interp(ni, xi, [1, self.args.nbs / self.batch_size]).round()))
+                    for j, x in enumerate(self.optimizer.param_groups):
+                        # Bias lr falls from 0.1 to lr0, all other lrs rise from 0.0 to lr0
+                        x["lr"] = np.interp(
+                            ni, xi, [self.args.warmup_bias_lr if j == 0 else 0.0, x["initial_lr"] * self.lf(epoch)]
+                        )
+                        if "momentum" in x:
+                            x["momentum"] = np.interp(ni, xi, [self.args.warmup_momentum, self.args.momentum])
+
+                # Forward
+                with autocast(self.amp):
+                    batch = self.preprocess_batch(batch)
+                    loss, self.loss_items = self.model(batch)
+                    self.loss = loss.sum()
+                    if RANK != -1:
+                        self.loss *= world_size
+                    self.tloss = (
+                        (self.tloss * i + self.loss_items) / (i + 1) if self.tloss is not None else self.loss_items
+                    )
+
+                # Add more distillation logic
+                if self.teacher is not None:
+                    distill_weight = ((1 - math.cos(i * math.pi / len(self.train_loader))) / 2) * (0.1 - 1) + 1
+                    with torch.no_grad():
+                        pred = self.teacher(batch['img'])
+
+                    self.d_loss = distillation_loss.get_loss()
+                    self.d_loss *- distill_weight
+                    self.loss += self.d_loss
+                # Backward
+                self.scaler.scale(self.loss).backward()
+
+                # Optimize - https://pytorch.org/docs/master/notes/amp_examples.html
+                if ni - last_opt_step >= self.accumulate:
+                    self.optimizer_step()
+                    last_opt_step = ni
+
+                    # Timed stopping
+                    if self.args.time:
+                        self.stop = (time.time() - self.train_time_start) > (self.args.time * 3600)
+                        if RANK != -1:  # if DDP training
+                            broadcast_list = [self.stop if RANK == 0 else None]
+                            dist.broadcast_object_list(broadcast_list, 0)  # broadcast 'stop' to all ranks
+                            self.stop = broadcast_list[0]
+                        if self.stop:  # training time exceeded
+                            break
+
+                # Log
+                if RANK in {-1, 0}:
+                    loss_length = self.tloss.shape[0] if len(self.tloss.shape) else 1
+                    pbar.set_description(
+                        ("%11s" * 2 + "%11.4g" * (2 + loss_length))
+                        % (
+                            f"{epoch + 1}/{self.epochs}",
+                            f"{self._get_memory():.3g}G",  # (GB) GPU memory util
+                            *(self.tloss if loss_length > 1 else torch.unsqueeze(self.tloss, 0)),  # losses
+                            batch["cls"].shape[0],  # batch size, i.e. 8
+                            batch["img"].shape[-1],  # imgsz, i.e 640
+                        )
+                    )
+                    self.run_callbacks("on_batch_end")
+                    if self.args.plots and ni in self.plot_idx:
+                        self.plot_training_samples(batch, ni)
+
+                self.run_callbacks("on_train_batch_end")
+
+            # More distillation logic
+            if self.teacher is not None:
+                distillation_loss.remove_handle_()
+
+
+            self.lr = {f"lr/pg{ir}": x["lr"] for ir, x in enumerate(self.optimizer.param_groups)}  # for loggers
+            self.run_callbacks("on_train_epoch_end")
+            if RANK in {-1, 0}:
+                final_epoch = epoch + 1 >= self.epochs
+                self.ema.update_attr(self.model, include=["yaml", "nc", "args", "names", "stride", "class_weights"])
+
+                # Validation
+                if self.args.val or final_epoch or self.stopper.possible_stop or self.stop:
+                    self._clear_memory(threshold=0.5)  # prevent VRAM spike
+                    self.metrics, self.fitness = self.validate()
+                self.save_metrics(metrics={**self.label_loss_items(self.tloss), **self.metrics, **self.lr})
+                self.stop |= self.stopper(epoch + 1, self.fitness) or final_epoch
+                if self.args.time:
+                    self.stop |= (time.time() - self.train_time_start) > (self.args.time * 3600)
+
+                # Save model
+                if self.args.save or final_epoch:
+                    self.save_model()
+                    self.run_callbacks("on_model_save")
+
+            # Scheduler
+            t = time.time()
+            self.epoch_time = t - self.epoch_time_start
+            self.epoch_time_start = t
+            if self.args.time:
+                mean_epoch_time = (t - self.train_time_start) / (epoch - self.start_epoch + 1)
+                self.epochs = self.args.epochs = math.ceil(self.args.time * 3600 / mean_epoch_time)
+                self._setup_scheduler()
+                self.scheduler.last_epoch = self.epoch  # do not move
+                self.stop |= epoch >= self.epochs  # stop if exceeded epochs
+            self.run_callbacks("on_fit_epoch_end")
+            self._clear_memory(0.5)  # clear if memory utilization > 50%
+
+            # Early Stopping
+            if RANK != -1:  # if DDP training
+                broadcast_list = [self.stop if RANK == 0 else None]
+                dist.broadcast_object_list(broadcast_list, 0)  # broadcast 'stop' to all ranks
+                self.stop = broadcast_list[0]
+            if self.stop:
+                break  # must break all DDP ranks
+            epoch += 1
+
+        if RANK in {-1, 0}:
+            # Do final val with best.pt
+            seconds = time.time() - self.train_time_start
+            LOGGER.info(f"\n{epoch - self.start_epoch + 1} epochs completed in {seconds / 3600:.3f} hours.")
+            self.final_eval()
+            if self.args.plots:
+                self.plot_metrics()
+            self.run_callbacks("on_train_end")
+        self._clear_memory()
+
+        # Distill logic
+        if self.teacher is not None:
+            distillation_loss.remove_handle_()
+        unset_deterministic()
+
+        self.run_callbacks("teardown")
+
+    def auto_batch(self, max_num_obj=0):
+        """Calculate optimal batch size based on model and device memory constraints."""
+        return check_train_batch_size(
+            model=self.model,
+            imgsz=self.args.imgsz,
+            amp=self.amp,
+            batch=self.batch_size,
+            max_num_obj=max_num_obj,
+        )  # returns batch size
+
+    def _get_memory(self, fraction=False):
+        """Get accelerator memory utilization in GB or as a fraction of total memory."""
+        memory, total = 0, 0
+        if self.device.type == "mps":
+            memory = torch.mps.driver_allocated_memory()
+            if fraction:
+                return __import__("psutil").virtual_memory().percent / 100
+        elif self.device.type != "cpu":
+            memory = torch.cuda.memory_reserved()
+            if fraction:
+                total = torch.cuda.get_device_properties(self.device).total_memory
+        return ((memory / total) if total > 0 else 0) if fraction else (memory / 2**30)
+
+    def _clear_memory(self, threshold: float = None):
+        """Clear accelerator memory by calling garbage collector and emptying cache."""
+        if threshold:
+            assert 0 <= threshold <= 1, "Threshold must be between 0 and 1."
+            if self._get_memory(fraction=True) <= threshold:
+                return
+        gc.collect()
+        if self.device.type == "mps":
+            torch.mps.empty_cache()
+        elif self.device.type == "cpu":
+            return
+        else:
+            torch.cuda.empty_cache()
+
+    def read_results_csv(self):
+        """Read results.csv into a dictionary using pandas."""
+        import pandas as pd  # scope for faster 'import ultralytics'
+
+        return pd.read_csv(self.csv).to_dict(orient="list")
+
+    def _model_train(self):
+        """Set model in training mode."""
+        self.model.train()
+        # Freeze BN stat
+        for n, m in self.model.named_modules():
+            if any(filter(lambda f: f in n, self.freeze_layer_names)) and isinstance(m, nn.BatchNorm2d):
+                m.eval()
+
+    def save_model(self):
+        """Save model training checkpoints with additional metadata."""
+        import io
+
+        # Serialize ckpt to a byte buffer once (faster than repeated torch.save() calls)
+        buffer = io.BytesIO()
+        torch.save(
+            {
+                "epoch": self.epoch,
+                "best_fitness": self.best_fitness,
+                "model": None,  # resume and final checkpoints derive from EMA
+                "ema": deepcopy(self.ema.ema).half(),
+                "updates": self.ema.updates,
+                "optimizer": convert_optimizer_state_dict_to_fp16(deepcopy(self.optimizer.state_dict())),
+                "train_args": vars(self.args),  # save as dict
+                "train_metrics": {**self.metrics, **{"fitness": self.fitness}},
+                "train_results": self.read_results_csv(),
+                "date": datetime.now().isoformat(),
+                "version": __version__,
+                "license": "AGPL-3.0 (https://ultralytics.com/license)",
+                "docs": "https://docs.ultralytics.com",
+            },
+            buffer,
+        )
+        serialized_ckpt = buffer.getvalue()  # get the serialized content to save
+
+        # Save checkpoints
+        self.last.write_bytes(serialized_ckpt)  # save last.pt
+        if self.best_fitness == self.fitness:
+            self.best.write_bytes(serialized_ckpt)  # save best.pt
+        if (self.save_period > 0) and (self.epoch % self.save_period == 0):
+            (self.wdir / f"epoch{self.epoch}.pt").write_bytes(serialized_ckpt)  # save epoch, i.e. 'epoch3.pt'
+        # if self.args.close_mosaic and self.epoch == (self.epochs - self.args.close_mosaic - 1):
+        #    (self.wdir / "last_mosaic.pt").write_bytes(serialized_ckpt)  # save mosaic checkpoint
+
+    def get_dataset(self):
+        """
+        Get train and validation datasets from data dictionary.
+
+        Returns:
+            (dict): A dictionary containing the training/validation/test dataset and category names.
+        """
+        try:
+            if self.args.task == "classify":
+                data = check_cls_dataset(self.args.data)
+            elif self.args.data.rsplit(".", 1)[-1] in {"yaml", "yml"} or self.args.task in {
+                "detect",
+                "segment",
+                "pose",
+                "obb",
+            }:
+                data = check_det_dataset(self.args.data)
+                if "yaml_file" in data:
+                    self.args.data = data["yaml_file"]  # for validating 'yolo train data=url.zip' usage
+        except Exception as e:
+            raise RuntimeError(emojis(f"Dataset '{clean_url(self.args.data)}' error ❌ {e}")) from e
+        if self.args.single_cls:
+            LOGGER.info("Overriding class names with single class.")
+            data["names"] = {0: "item"}
+            data["nc"] = 1
+        return data
+
+    def setup_model(self):
+        """
+        Load, create, or download model for any task.
+
+        Returns:
+            (dict): Optional checkpoint to resume training from.
+        """
+        if isinstance(self.model, torch.nn.Module):  # if model is loaded beforehand. No setup needed
+            return
+
+        cfg, weights = self.model, None
+        ckpt = None
+        if str(self.model).endswith(".pt"):
+            weights, ckpt = attempt_load_one_weight(self.model)
+            cfg = weights.yaml
+        elif isinstance(self.args.pretrained, (str, Path)):
+            weights, _ = attempt_load_one_weight(self.args.pretrained)
+        self.model = self.get_model(cfg=cfg, weights=weights, verbose=RANK == -1)  # calls Model(cfg, weights)
+        return ckpt
+
+    def optimizer_step(self):
+        """Perform a single step of the training optimizer with gradient clipping and EMA update."""
+        self.scaler.unscale_(self.optimizer)  # unscale gradients
+        torch.nn.utils.clip_grad_norm_(self.model.parameters(), max_norm=10.0)  # clip gradients
+        self.scaler.step(self.optimizer)
+        self.scaler.update()
+        self.optimizer.zero_grad()
+        if self.ema:
+            self.ema.update(self.model)
+
+    def preprocess_batch(self, batch):
+        """Allow custom preprocessing model inputs and ground truths depending on task type."""
+        return batch
+
+    def validate(self):
+        """
+        Run validation on test set using self.validator.
+
+        Returns:
+            metrics (dict): Dictionary of validation metrics.
+            fitness (float): Fitness score for the validation.
+        """
+        metrics = self.validator(self)
+        fitness = metrics.pop("fitness", -self.loss.detach().cpu().numpy())  # use loss as fitness measure if not found
+        if not self.best_fitness or self.best_fitness < fitness:
+            self.best_fitness = fitness
+        return metrics, fitness
+
+    def get_model(self, cfg=None, weights=None, verbose=True):
+        """Get model and raise NotImplementedError for loading cfg files."""
+        raise NotImplementedError("This task trainer doesn't support loading cfg files")
+
+    def get_validator(self):
+        """Return a NotImplementedError when the get_validator function is called."""
+        raise NotImplementedError("get_validator function not implemented in trainer")
+
+    def get_dataloader(self, dataset_path, batch_size=16, rank=0, mode="train"):
+        """Return dataloader derived from torch.data.Dataloader."""
+        raise NotImplementedError("get_dataloader function not implemented in trainer")
+
+    def build_dataset(self, img_path, mode="train", batch=None):
+        """Build dataset."""
+        raise NotImplementedError("build_dataset function not implemented in trainer")
+
+    def label_loss_items(self, loss_items=None, prefix="train"):
+        """
+        Return a loss dict with labelled training loss items tensor.
+
+        Note:
+            This is not needed for classification but necessary for segmentation & detection
+        """
+        return {"loss": loss_items} if loss_items is not None else ["loss"]
+
+    def set_model_attributes(self):
+        """Set or update model parameters before training."""
+        self.model.names = self.data["names"]
+
+    def build_targets(self, preds, targets):
+        """Build target tensors for training YOLO model."""
+        pass
+
+    def progress_string(self):
+        """Return a string describing training progress."""
+        return ""
+
+    # TODO: may need to put these following functions into callback
+    def plot_training_samples(self, batch, ni):
+        """Plot training samples during YOLO training."""
+        pass
+
+    def plot_training_labels(self):
+        """Plot training labels for YOLO model."""
+        pass
+
+    def save_metrics(self, metrics):
+        """Save training metrics to a CSV file."""
+        keys, vals = list(metrics.keys()), list(metrics.values())
+        n = len(metrics) + 2  # number of cols
+        s = "" if self.csv.exists() else (("%s," * n % tuple(["epoch", "time"] + keys)).rstrip(",") + "\n")  # header
+        t = time.time() - self.train_time_start
+        with open(self.csv, "a", encoding="utf-8") as f:
+            f.write(s + ("%.6g," * n % tuple([self.epoch + 1, t] + vals)).rstrip(",") + "\n")
+
+    def plot_metrics(self):
+        """Plot and display metrics visually."""
+        pass
+
+    def on_plot(self, name, data=None):
+        """Register plots (e.g. to be consumed in callbacks)."""
+        path = Path(name)
+        self.plots[path] = {"data": data, "timestamp": time.time()}
+
+    def final_eval(self):
+        """Perform final evaluation and validation for object detection YOLO model."""
+        ckpt = {}
+        for f in self.last, self.best:
+            if f.exists():
+                if f is self.last:
+                    ckpt = strip_optimizer(f)
+                elif f is self.best:
+                    k = "train_results"  # update best.pt train_metrics from last.pt
+                    strip_optimizer(f, updates={k: ckpt[k]} if k in ckpt else None)
+                    LOGGER.info(f"\nValidating {f}...")
+                    self.validator.args.plots = self.args.plots
+                    self.metrics = self.validator(model=f)
+                    self.metrics.pop("fitness", None)
+                    self.run_callbacks("on_fit_epoch_end")
+
+    def check_resume(self, overrides):
+        """Check if resume checkpoint exists and update arguments accordingly."""
+        resume = self.args.resume
+        if resume:
+            try:
+                exists = isinstance(resume, (str, Path)) and Path(resume).exists()
+                last = Path(check_file(resume) if exists else get_latest_run())
+
+                # Check that resume data YAML exists, otherwise strip to force re-download of dataset
+                ckpt_args = attempt_load_weights(last).args
+                if not isinstance(ckpt_args["data"], dict) and not Path(ckpt_args["data"]).exists():
+                    ckpt_args["data"] = self.args.data
+
+                resume = True
+                self.args = get_cfg(ckpt_args)
+                self.args.model = self.args.resume = str(last)  # reinstate model
+                for k in (
+                    "imgsz",
+                    "batch",
+                    "device",
+                    "close_mosaic",
+                ):  # allow arg updates to reduce memory or update device on resume
+                    if k in overrides:
+                        setattr(self.args, k, overrides[k])
+
+            except Exception as e:
+                raise FileNotFoundError(
+                    "Resume checkpoint not found. Please pass a valid checkpoint to resume from, "
+                    "i.e. 'yolo train resume model=path/to/last.pt'"
+                ) from e
+        self.resume = resume
+
+    def resume_training(self, ckpt):
+        """Resume YOLO training from given epoch and best fitness."""
+        if ckpt is None or not self.resume:
+            return
+        best_fitness = 0.0
+        start_epoch = ckpt.get("epoch", -1) + 1
+        if ckpt.get("optimizer", None) is not None:
+            self.optimizer.load_state_dict(ckpt["optimizer"])  # optimizer
+            best_fitness = ckpt["best_fitness"]
+        if self.ema and ckpt.get("ema"):
+            self.ema.ema.load_state_dict(ckpt["ema"].float().state_dict())  # EMA
+            self.ema.updates = ckpt["updates"]
+        assert start_epoch > 0, (
+            f"{self.args.model} training to {self.epochs} epochs is finished, nothing to resume.\n"
+            f"Start a new training without resuming, i.e. 'yolo train model={self.args.model}'"
+        )
+        LOGGER.info(f"Resuming training {self.args.model} from epoch {start_epoch + 1} to {self.epochs} total epochs")
+        if self.epochs < start_epoch:
+            LOGGER.info(
+                f"{self.model} has been trained for {ckpt['epoch']} epochs. Fine-tuning for {self.epochs} more epochs."
+            )
+            self.epochs += ckpt["epoch"]  # finetune additional epochs
+        self.best_fitness = best_fitness
+        self.start_epoch = start_epoch
+        if start_epoch > (self.epochs - self.args.close_mosaic):
+            self._close_dataloader_mosaic()
+
+    def _close_dataloader_mosaic(self):
+        """Update dataloaders to stop using mosaic augmentation."""
+        if hasattr(self.train_loader.dataset, "mosaic"):
+            self.train_loader.dataset.mosaic = False
+        if hasattr(self.train_loader.dataset, "close_mosaic"):
+            LOGGER.info("Closing dataloader mosaic")
+            self.train_loader.dataset.close_mosaic(hyp=copy(self.args))
+
+    def build_optimizer(self, model, teacher=None, name="auto", lr=0.001, momentum=0.9, decay=1e-5, iterations=1e5):
+        """
+        Construct an optimizer for the given model.
+
+        Args:
+            model (torch.nn.Module): The model for which to build an optimizer.
+            name (str, optional): The name of the optimizer to use. If 'auto', the optimizer is selected
+                based on the number of iterations.
+            lr (float, optional): The learning rate for the optimizer.
+            momentum (float, optional): The momentum factor for the optimizer.
+            decay (float, optional): The weight decay for the optimizer.
+            iterations (float, optional): The number of iterations, which determines the optimizer if
+                name is 'auto'.
+
+        Returns:
+            (torch.optim.Optimizer): The constructed optimizer.
+        """
+        g = [], [], []  # optimizer parameter groups
+        bn = tuple(v for k, v in nn.__dict__.items() if "Norm" in k)  # normalization layers, i.e. BatchNorm2d()
+        if name == "auto":
+            LOGGER.info(
+                f"{colorstr('optimizer:')} 'optimizer=auto' found, "
+                f"ignoring 'lr0={self.args.lr0}' and 'momentum={self.args.momentum}' and "
+                f"determining best 'optimizer', 'lr0' and 'momentum' automatically... "
+            )
+            nc = self.data.get("nc", 10)  # number of classes
+            lr_fit = round(0.002 * 5 / (4 + nc), 6)  # lr0 fit equation to 6 decimal places
+            name, lr, momentum = ("SGD", 0.01, 0.9) if iterations > 10000 else ("AdamW", lr_fit, 0.9)
+            self.args.warmup_bias_lr = 0.0  # no higher than 0.01 for Adam
+
+        for module_name, module in model.named_modules():
+            for param_name, param in module.named_parameters(recurse=False):
+                fullname = f"{module_name}.{param_name}" if module_name else param_name
+                if "bias" in fullname:  # bias (no decay)
+                    g[2].append(param)
+                elif isinstance(module, bn) or "logit_scale" in fullname:  # weight (no decay)
+                    # ContrastiveHead and BNContrastiveHead included here with 'logit_scale'
+                    g[1].append(param)
+                else:  # weight (with decay)
+                    g[0].append(param)
+
+        if teacher is not None:
+            for v in teacher.modules():
+                if hasattr(v, 'bias') and isinstance(v.bias, nn.Parameter):
+                    g[2].append(v.bias)
+                if isinstance(v, bn):  # weight (no decay)
+                    g[1].append(v.weight)
+                elif hasattr(v, 'weight') and isinstance(v.weight, nn.Parameter):  # weight (with decay)
+                    g[0].append(v.weight)
+        optimizers = {"Adam", "Adamax", "AdamW", "NAdam", "RAdam", "RMSProp", "SGD", "auto"}
+        name = {x.lower(): x for x in optimizers}.get(name.lower())
+        if name in {"Adam", "Adamax", "AdamW", "NAdam", "RAdam"}:
+            optimizer = getattr(optim, name, optim.Adam)(g[2], lr=lr, betas=(momentum, 0.999), weight_decay=0.0)
+        elif name == "RMSProp":
+            optimizer = optim.RMSprop(g[2], lr=lr, momentum=momentum)
+        elif name == "SGD":
+            optimizer = optim.SGD(g[2], lr=lr, momentum=momentum, nesterov=True)
+        else:
+            raise NotImplementedError(
+                f"Optimizer '{name}' not found in list of available optimizers {optimizers}. "
+                "Request support for addition optimizers at https://github.com/ultralytics/ultralytics."
+            )
+
+        optimizer.add_param_group({"params": g[0], "weight_decay": decay})  # add g0 with weight_decay
+        optimizer.add_param_group({"params": g[1], "weight_decay": 0.0})  # add g1 (BatchNorm2d weights)
+        LOGGER.info(
+            f"{colorstr('optimizer:')} {type(optimizer).__name__}(lr={lr}, momentum={momentum}) with parameter groups "
+            f"{len(g[1])} weight(decay=0.0), {len(g[0])} weight(decay={decay}), {len(g[2])} bias(decay=0.0)"
+        )
+        return optimizer