diff --git a/main_linear_probe.py b/main_linear_probe.py deleted file mode 100644 index 2dc7e4f..0000000 --- a/main_linear_probe.py +++ /dev/null @@ -1,414 +0,0 @@ -import argparse -import datetime -import json - -import numpy as np -import os -import time -from pathlib import Path - -import torch -import torch.backends.cudnn as cudnn -from torch.utils.tensorboard import SummaryWriter -from timm.models.layers import trunc_normal_ -from timm.data.mixup import Mixup - -import models_vit as models -import util.lr_decay as lrd -import util.misc as misc -from util.datasets import build_dataset -from util.pos_embed import interpolate_pos_embed -from util.misc import NativeScalerWithGradNormCount as NativeScaler -from huggingface_hub import hf_hub_download, login -from engine_finetune import train_one_epoch, evaluate - -import warnings -import faulthandler - -faulthandler.enable() -warnings.simplefilter(action='ignore', category=FutureWarning) - - -def get_args_parser(): - parser = argparse.ArgumentParser('MAE fine-tuning for image classification', add_help=False) - parser.add_argument('--batch_size', default=128, type=int, - help='Batch size per GPU (effective batch size is batch_size * accum_iter * # gpus') - parser.add_argument('--epochs', default=50, type=int) - parser.add_argument('--accum_iter', default=1, type=int, - help='Accumulate gradient iterations (for increasing the effective batch size under memory constraints)') - - # Model parameters - parser.add_argument('--model', default='vit_large_patch16', type=str, metavar='MODEL', - help='Name of model to train') - parser.add_argument('--input_size', default=256, type=int, - help='images input size') - parser.add_argument('--drop_path', type=float, default=0.2, metavar='PCT', - help='Drop path rate (default: 0.1)') - - # Optimizer parameters - parser.add_argument('--clip_grad', type=float, default=None, metavar='NORM', - help='Clip gradient norm (default: None, no clipping)') - parser.add_argument('--weight_decay', type=float, default=0.05, - help='weight decay (default: 0.05)') - parser.add_argument('--lr', type=float, default=None, metavar='LR', - help='learning rate (absolute lr)') - parser.add_argument('--blr', type=float, default=5e-3, metavar='LR', - help='base learning rate: absolute_lr = base_lr * total_batch_size / 256') - parser.add_argument('--layer_decay', type=float, default=0.65, - help='layer-wise lr decay from ELECTRA/BEiT') - parser.add_argument('--min_lr', type=float, default=1e-6, metavar='LR', - help='lower lr bound for cyclic schedulers that hit 0') - parser.add_argument('--warmup_epochs', type=int, default=10, metavar='N', - help='epochs to warmup LR') - - # Augmentation parameters - parser.add_argument('--color_jitter', type=float, default=None, metavar='PCT', - help='Color jitter factor (enabled only when not using Auto/RandAug)') - parser.add_argument('--aa', type=str, default='rand-m9-mstd0.5-inc1', metavar='NAME', - help='Use AutoAugment policy. "v0" or "original". " + "(default: rand-m9-mstd0.5-inc1)'), - parser.add_argument('--smoothing', type=float, default=0.1, - help='Label smoothing (default: 0.1)') - - # * Random Erase params - parser.add_argument('--reprob', type=float, default=0.25, metavar='PCT', - help='Random erase prob (default: 0.25)') - parser.add_argument('--remode', type=str, default='pixel', - help='Random erase mode (default: "pixel")') - parser.add_argument('--recount', type=int, default=1, - help='Random erase count (default: 1)') - parser.add_argument('--resplit', action='store_true', default=False, - help='Do not random erase first (clean) augmentation split') - - # * Mixup params - parser.add_argument('--mixup', type=float, default=0, - help='mixup alpha, mixup enabled if > 0.') - parser.add_argument('--cutmix', type=float, default=0, - help='cutmix alpha, cutmix enabled if > 0.') - parser.add_argument('--cutmix_minmax', type=float, nargs='+', default=None, - help='cutmix min/max ratio, overrides alpha and enables cutmix if set (default: None)') - parser.add_argument('--mixup_prob', type=float, default=1.0, - help='Probability of performing mixup or cutmix when either/both is enabled') - parser.add_argument('--mixup_switch_prob', type=float, default=0.5, - help='Probability of switching to cutmix when both mixup and cutmix enabled') - parser.add_argument('--mixup_mode', type=str, default='batch', - help='How to apply mixup/cutmix params. Per "batch", "pair", or "elem"') - - # * Finetuning params - parser.add_argument('--finetune', default='', type=str, - help='finetune from checkpoint') - parser.add_argument('--task', default='', type=str, - help='finetune from checkpoint') - parser.add_argument('--global_pool', action='store_true') - parser.set_defaults(global_pool=True) - parser.add_argument('--cls_token', action='store_false', dest='global_pool', - help='Use class token instead of global pool for classification') - - # Dataset parameters - parser.add_argument('--data_path', default='./data/', type=str, - help='dataset path') - parser.add_argument('--nb_classes', default=8, type=int, - help='number of the classification types') - parser.add_argument('--output_dir', default='./output_dir', - help='path where to save, empty for no saving') - parser.add_argument('--log_dir', default='./output_logs', - help='path where to tensorboard log') - parser.add_argument('--device', default='cuda', - help='device to use for training / testing') - parser.add_argument('--seed', default=0, type=int) - parser.add_argument('--resume', default='', - help='resume from checkpoint') - parser.add_argument('--start_epoch', default=0, type=int, metavar='N', - help='start epoch') - parser.add_argument('--eval', action='store_true', - help='Perform evaluation only') - parser.add_argument('--dist_eval', action='store_true', default=False, - help='Enabling distributed evaluation (recommended during training for faster monitor') - parser.add_argument('--num_workers', default=10, type=int) - parser.add_argument('--pin_mem', action='store_true', - help='Pin CPU memory in DataLoader for more efficient (sometimes) transfer to GPU.') - parser.set_defaults(pin_mem=True) - - # distributed training parameters - parser.add_argument('--world_size', default=1, type=int, - help='number of distributed processes') - parser.add_argument('--local_rank', default=-1, type=int) - parser.add_argument('--dist_on_itp', action='store_true') - parser.add_argument('--dist_url', default='env://', - help='url used to set up distributed training') - - # fine-tuning parameters - parser.add_argument('--savemodel', action='store_true', default=True, - help='Save model') - parser.add_argument('--norm', default='IMAGENET', type=str, help='Normalization method') - parser.add_argument('--enhance', action='store_true', default=False, help='Use enhanced data') - parser.add_argument('--datasets_seed', default=2026, type=int) - - return parser - - -def main(args, criterion): - if args.resume and not args.eval: - resume = args.resume - checkpoint = torch.load(args.resume, map_location='cpu') - print("Load checkpoint from: %s" % args.resume) - args = checkpoint['args'] - args.resume = resume - - misc.init_distributed_mode(args) - - print('job dir: {}'.format(os.path.dirname(os.path.realpath(__file__)))) - print("{}".format(args).replace(', ', ',\n')) - - device = torch.device(args.device) - - # fix the seed for reproducibility - seed = args.seed + misc.get_rank() - torch.manual_seed(seed) - np.random.seed(seed) - - cudnn.benchmark = True - - if args.model=='RETFound_mae': - model = models.__dict__[args.model]( - img_size=args.input_size, - num_classes=args.nb_classes, - drop_path_rate=args.drop_path, - global_pool=args.global_pool, - ) - else: - model = models.__dict__[args.model]( - num_classes=args.nb_classes, - drop_path_rate=args.drop_path, - args=args, - ) - - if args.finetune and not args.eval: - - print(f"Downloading pre-trained weights from: {args.finetune}") - - checkpoint_path = hf_hub_download( - repo_id=f'YukunZhou/{args.finetune}', - filename=f'{args.finetune}.pth', - ) - - checkpoint = torch.load(checkpoint_path, map_location='cpu') - print("Load pre-trained checkpoint from: %s" % args.finetune) - - if args.model!='RETFound_mae': - checkpoint_model = checkpoint['teacher'] - else: - checkpoint_model = checkpoint['model'] - - checkpoint_model = {k.replace("backbone.", ""): v for k, v in checkpoint_model.items()} - checkpoint_model = {k.replace("mlp.w12.", "mlp.fc1."): v for k, v in checkpoint_model.items()} - checkpoint_model = {k.replace("mlp.w3.", "mlp.fc2."): v for k, v in checkpoint_model.items()} - - state_dict = model.state_dict() - for k in ['head.weight', 'head.bias']: - if k in checkpoint_model and checkpoint_model[k].shape != state_dict[k].shape: - print(f"Removing key {k} from pretrained checkpoint") - del checkpoint_model[k] - - # interpolate position embedding - interpolate_pos_embed(model, checkpoint_model) - - # load pre-trained model - msg = model.load_state_dict(checkpoint_model, strict=False) - - trunc_normal_(model.head.weight, std=2e-5) - - dataset_train = build_dataset(is_train='train', args=args) - dataset_val = build_dataset(is_train='val', args=args) - dataset_test = build_dataset(is_train='test', args=args) - - - if True: # args.distributed: - num_tasks = misc.get_world_size() - global_rank = misc.get_rank() - if not args.eval: - sampler_train = torch.utils.data.DistributedSampler( - dataset_train, num_replicas=num_tasks, rank=global_rank, shuffle=True - ) - print("Sampler_train = %s" % str(sampler_train)) - if args.dist_eval: - if len(dataset_val) % num_tasks != 0: - print( - 'Warning: Enabling distributed evaluation with an eval dataset not divisible by process number. ' - 'This will slightly alter validation results as extra duplicate entries are added to achieve ' - 'equal num of samples per-process.') - sampler_val = torch.utils.data.DistributedSampler( - dataset_val, num_replicas=num_tasks, rank=global_rank, - shuffle=True) # shuffle=True to reduce monitor bias - else: - sampler_val = torch.utils.data.SequentialSampler(dataset_val) - - if args.dist_eval: - if len(dataset_test) % num_tasks != 0: - print('Warning: Enabling distributed evaluation with an eval dataset not divisible by process number. ' - 'This will slightly alter validation results as extra duplicate entries are added to achieve ' - 'equal num of samples per-process.') - sampler_test = torch.utils.data.DistributedSampler( - dataset_test, num_replicas=num_tasks, rank=global_rank, - shuffle=True) # shuffle=True to reduce monitor bias - else: - sampler_test = torch.utils.data.SequentialSampler(dataset_test) - - if global_rank == 0 and args.log_dir is not None and not args.eval: - os.makedirs(args.log_dir, exist_ok=True) - log_writer = SummaryWriter(log_dir=os.path.join(args.log_dir, args.task)) - else: - log_writer = None - - if not args.eval: - data_loader_train = torch.utils.data.DataLoader( - dataset_train, sampler=sampler_train, - batch_size=args.batch_size, - num_workers=args.num_workers, - pin_memory=args.pin_mem, - drop_last=True, - ) - - print(f'len of train_set: {len(data_loader_train) * args.batch_size}') - - data_loader_val = torch.utils.data.DataLoader( - dataset_val, sampler=sampler_val, - batch_size=args.batch_size, - num_workers=args.num_workers, - pin_memory=args.pin_mem, - drop_last=False - ) - - data_loader_test = torch.utils.data.DataLoader( - dataset_test, sampler=sampler_test, - batch_size=args.batch_size, - num_workers=args.num_workers, - pin_memory=args.pin_mem, - drop_last=False - ) - - mixup_fn = None - mixup_active = args.mixup > 0 or args.cutmix > 0. or args.cutmix_minmax is not None - if mixup_active: - print("Mixup is activated!") - mixup_fn = Mixup( - mixup_alpha=args.mixup, cutmix_alpha=args.cutmix, cutmix_minmax=args.cutmix_minmax, - prob=args.mixup_prob, switch_prob=args.mixup_switch_prob, mode=args.mixup_mode, - label_smoothing=args.smoothing, num_classes=args.nb_classes) - - if args.resume and args.eval: - checkpoint = torch.load(args.resume, map_location='cpu') - print("Load checkpoint from: %s" % args.resume) - model.load_state_dict(checkpoint['model']) - - model.to(device) - model_without_ddp = model - - n_parameters = sum(p.numel() for p in model.parameters() if p.requires_grad) - print('number of model params (M): %.2f' % (n_parameters / 1.e6)) - - eff_batch_size = args.batch_size * args.accum_iter * misc.get_world_size() - - if args.lr is None: # only base_lr is specified - args.lr = args.blr * eff_batch_size / 256 - - print("base lr: %.2e" % (args.lr * 256 / eff_batch_size)) - print("actual lr: %.2e" % args.lr) - - print("accumulate grad iterations: %d" % args.accum_iter) - print("effective batch size: %d" % eff_batch_size) - - if args.distributed: - model = torch.nn.parallel.DistributedDataParallel(model, device_ids=[args.gpu],find_unused_parameters=True) - model_without_ddp = model.module - - no_weight_decay = model_without_ddp.no_weight_decay() if hasattr(model_without_ddp, 'no_weight_decay') else [] - param_groups = lrd.param_groups_lrd(model_without_ddp, args.weight_decay, - no_weight_decay_list=no_weight_decay, - layer_decay=args.layer_decay - ) - optimizer = torch.optim.AdamW(param_groups, lr=args.lr) - loss_scaler = NativeScaler() - - print("criterion = %s" % str(criterion)) - - misc.load_model(args=args, model_without_ddp=model_without_ddp, optimizer=optimizer, loss_scaler=loss_scaler) - - for name, param in model.named_parameters(): - if 'head' in name: - param.requires_grad = True - else: - param.requires_grad = False - - - if args.eval: - if 'epoch' in checkpoint: - print("Test with the best model at epoch = %d" % checkpoint['epoch']) - test_stats, auc_roc = evaluate(data_loader_test, model, device, args, epoch=0, mode='test', - num_class=args.nb_classes, log_writer=log_writer) - exit(0) - - print(f"Start training for {args.epochs} epochs") - start_time = time.time() - max_score = 0.0 - best_epoch = 0 - for epoch in range(args.start_epoch, args.epochs): - if args.distributed: - data_loader_train.sampler.set_epoch(epoch) - - train_stats = train_one_epoch( - model, criterion, data_loader_train, - optimizer, device, epoch, loss_scaler, - args.clip_grad, mixup_fn, - log_writer=log_writer, - args=args - ) - - val_stats, val_score = evaluate(data_loader_val, model, device, args, epoch, mode='val', - num_class=args.nb_classes, log_writer=log_writer) - if max_score < val_score: - max_score = val_score - best_epoch = epoch - if args.output_dir and args.savemodel: - misc.save_model( - args=args, model=model, model_without_ddp=model_without_ddp, optimizer=optimizer, - loss_scaler=loss_scaler, epoch=epoch, mode='best') - print("Best epoch = %d, Best score = %.4f" % (best_epoch, max_score)) - - - if epoch == (args.epochs - 1): - checkpoint = torch.load(os.path.join(args.output_dir, args.task, 'checkpoint-best.pth'), map_location='cpu') - model.load_state_dict(checkpoint['model'], strict=False) - model.to(device) - print("Test with the best model, epoch = %d:" % checkpoint['epoch']) - test_stats, auc_roc = evaluate(data_loader_test, model, device, args, -1, mode='test', - num_class=args.nb_classes, log_writer=None) - - if log_writer is not None: - log_writer.add_scalar('loss/val', val_stats['loss'], epoch) - - log_stats = {**{f'train_{k}': v for k, v in train_stats.items()}, - 'epoch': epoch, - 'n_parameters': n_parameters} - - if args.output_dir and misc.is_main_process(): - if log_writer is not None: - log_writer.flush() - with open(os.path.join(args.output_dir, args.task, "log.txt"), mode="a", encoding="utf-8") as f: - f.write(json.dumps(log_stats) + "\n") - - total_time = time.time() - start_time - total_time_str = str(datetime.timedelta(seconds=int(total_time))) - print('Training time {}'.format(total_time_str)) - - -if __name__ == '__main__': - args = get_args_parser() - args = args.parse_args() - - criterion = torch.nn.CrossEntropyLoss() - - if args.output_dir: - Path(args.output_dir).mkdir(parents=True, exist_ok=True) - main(args, criterion) - -