def print_mutation(results, hyp, save_dir, bucket): evolve_csv, results_csv, evolve_yaml = save_dir / 'evolve.csv', save_dir / 'results.csv', save_dir / 'hyp_evolve.yaml' keys = ('metrics/precision', 'metrics/recall', 'metrics/mAP_0.5', 'metrics/mAP_0.5:0.95', 'val/box_loss', 'val/obj_loss', 'val/cls_loss') + tuple(hyp.keys()) # [results + hyps] keys = tuple(x.strip() for x in keys) vals = results + tuple(hyp.values()) n = len(keys) # Download (optional) if bucket: url = f'gs://{bucket}/evolve.csv' if gsutil_getsize(url) > (os.path.getsize(evolve_csv) if os.path.exists(evolve_csv) else 0): os.system(f'gsutil cp {url} {save_dir}' ) # download evolve.csv if larger than local # Log to evolve.csv s = '' if evolve_csv.exists() else ( ('%20s,' * n % keys).rstrip(',') + '\n') # add header with open(evolve_csv, 'a') as f: f.write(s + ('%20.5g,' * n % vals).rstrip(',') + '\n') # Print to screen print(colorstr('evolve: ') + ', '.join(f'{x.strip():>20s}' for x in keys)) print(colorstr('evolve: ') + ', '.join(f'{x:20.5g}' for x in vals), end='\n\n\n') # Save yaml with open(evolve_yaml, 'w') as f: data = pd.read_csv(evolve_csv) data = data.rename(columns=lambda x: x.strip()) # strip keys i = np.argmax(fitness(data.values[:, :7])) # f.write(f'# YOLOv5 Hyperparameter Evolution Results\n' + f'# Best generation: {i}\n' + f'# Last generation: {len(data)}\n' + f'# ' + ', '.join(f'{x.strip():>20s}' for x in keys[:7]) + '\n' + f'# ' + ', '.join(f'{x:>20.5g}' for x in data.values[i, :7]) + '\n\n') yaml.safe_dump(hyp, f, sort_keys=False) if bucket: os.system( f'gsutil cp {evolve_csv} {evolve_yaml} gs://{bucket}') # upload
def print_mutation(hyp, results, yaml_file="hyp_evolved.yaml", bucket=""): # Print mutation results to evolve.txt (for use with train.py --evolve) a = "%10s" * len(hyp) % tuple(hyp.keys()) # hyperparam keys b = "%10.3g" * len(hyp) % tuple(hyp.values()) # hyperparam values c = ( "%10.4g" * len(results) % results ) # results (P, R, [email protected], [email protected]:0.95, val_losses x 3) print("\n%s\n%s\nEvolved fitness: %s\n" % (a, b, c)) if bucket: url = "gs://%s/evolve.txt" % bucket if gsutil_getsize(url) > ( os.path.getsize("evolve.txt") if os.path.exists("evolve.txt") else 0 ): os.system( "gsutil cp %s ." % url ) # download evolve.txt if larger than local with open("evolve.txt", "a") as f: # append result f.write(c + b + "\n") x = np.unique(np.loadtxt("evolve.txt", ndmin=2), axis=0) # load unique rows x = x[np.argsort(-fitness(x))] # sort np.savetxt("evolve.txt", x, "%10.3g") # save sort by fitness # Save yaml for i, k in enumerate(hyp.keys()): hyp[k] = float(x[0, i + 7]) with open(yaml_file, "w") as f: results = tuple(x[0, :7]) c = ( "%10.4g" * len(results) % results ) # results (P, R, [email protected], [email protected]:0.95, val_losses x 3) f.write( "# Hyperparameter Evolution Results\n# Generations: %g\n# Metrics: " % len(x) + c + "\n\n" ) yaml.dump(hyp, f, sort_keys=False) if bucket: os.system("gsutil cp evolve.txt %s gs://%s" % (yaml_file, bucket)) # upload
def print_mutation(hyp, results, yaml_file='hyp_evolved.yaml', bucket=''): # Print mutation results to evolve.txt (for use with train.py --evolve) a = '%10s' * len(hyp) % tuple( hyp.keys_data_aug()) # hyperparam keys_data_aug b = '%10.3g' * len(hyp) % tuple(hyp.values()) # hyperparam values c = '%10.4g' * len( results ) % results # results (P, R, [email protected], [email protected]:0.95, val_losses x 3) print('\n%s\n%s\nEvolved fitness: %s\n' % (a, b, c)) if bucket: url = 'gs://%s/evolve.txt' % bucket if gsutil_getsize(url) > (os.path.getsize('evolve.txt') if os.path.exists('evolve.txt') else 0): os.system('gsutil cp %s .' % url) # download evolve.txt if larger than local with open('evolve.txt', 'a') as f: # append result f.write(c + b + '\n') x = np.unique(np.loadtxt('evolve.txt', ndmin=2), axis=0) # load unique rows x = x[np.argsort(-fitness(x))] # sort np.savetxt('evolve.txt', x, '%10.3g') # save sort by fitness # Save yaml for i, k in enumerate(hyp.keys_data_aug()): hyp[k] = float(x[0, i + 7]) with open(yaml_file, 'w') as f: results = tuple(x[0, :7]) c = '%10.4g' * len( results ) % results # results (P, R, [email protected], [email protected]:0.95, val_losses x 3) f.write( '# Hyperparameter Evolution Results\n# Generations: %g\n# Metrics: ' % len(x) + c + '\n\n') yaml.dump(hyp, f, sort_keys=False) if bucket: os.system('gsutil cp evolve.txt %s gs://%s' % (yaml_file, bucket)) # upload
def plot_evolution(yaml_file='data/hyp.finetune.yaml'): # from utils.general import *; plot_evolution() # Plot hyperparameter evolution results in evolve.txt with open(yaml_file) as f: hyp = yaml.load(f, Loader=yaml.FullLoader) x = np.loadtxt('evolve.txt', ndmin=2) f = fitness(x) # weights = (f - f.min()) ** 2 # for weighted results plt.figure(figsize=(10, 12), tight_layout=True) matplotlib.rc('font', **{'size': 8}) for i, (k, v) in enumerate(hyp.items()): y = x[:, i + 7] # mu = (y * weights).sum() / weights.sum() # best weighted result mu = y[f.argmax()] # best single result plt.subplot(6, 5, i + 1) plt.scatter(y, f, c=hist2d(y, f, 20), cmap='viridis', alpha=.8, edgecolors='none') plt.plot(mu, f.max(), 'k+', markersize=15) plt.title('%s = %.3g' % (k, mu), fontdict={'size': 9}) # limit to 40 characters if i % 5 != 0: plt.yticks([]) print('%15s: %.3g' % (k, mu)) plt.savefig('evolve.png', dpi=200) print('\nPlot saved as evolve.png')
def plot_evolve(evolve_csv='path/to/evolve.csv'): # from utils.plots import *; plot_evolve() # Plot evolve.csv hyp evolution results evolve_csv = Path(evolve_csv) data = pd.read_csv(evolve_csv) keys = [x.strip() for x in data.columns] x = data.values f = fitness(x) j = np.argmax(f) # max fitness index plt.figure(figsize=(10, 12), tight_layout=True) matplotlib.rc('font', **{'size': 8}) for i, k in enumerate(keys[7:]): v = x[:, 7 + i] mu = v[j] # best single result plt.subplot(6, 5, i + 1) plt.scatter(v, f, c=hist2d(v, f, 20), cmap='viridis', alpha=.8, edgecolors='none') plt.plot(mu, f.max(), 'k+', markersize=15) plt.title('%s = %.3g' % (k, mu), fontdict={'size': 9}) # limit to 40 characters if i % 5 != 0: plt.yticks([]) print('%15s: %.3g' % (k, mu)) f = evolve_csv.with_suffix('.png') # filename plt.savefig(f, dpi=200) plt.close() print(f'Saved {f}')
def main(opt, callbacks=Callbacks()): # Checks if RANK in [-1, 0]: print_args(FILE.stem, opt) check_git_status() check_requirements(exclude=['thop']) # Resume if opt.resume and not check_wandb_resume( opt) and not opt.evolve: # resume an interrupted run ckpt = opt.resume if isinstance( opt.resume, str) else get_latest_run() # specified or most recent path assert os.path.isfile( ckpt), 'ERROR: --resume checkpoint does not exist' with open(Path(ckpt).parent.parent / 'opt.yaml', errors='ignore') as f: opt = argparse.Namespace(**yaml.safe_load(f)) # replace opt.cfg, opt.weights, opt.resume = '', ckpt, True # reinstate LOGGER.info(f'Resuming training from {ckpt}') else: opt.data, opt.cfg, opt.hyp, opt.weights, opt.project = \ check_file(opt.data), check_yaml(opt.cfg), check_yaml(opt.hyp), str(opt.weights), str(opt.project) # checks assert len(opt.cfg) or len( opt.weights), 'either --cfg or --weights must be specified' if opt.evolve: if opt.project == str( ROOT / 'runs/train' ): # if default project name, rename to runs/evolve opt.project = str(ROOT / 'runs/evolve') opt.exist_ok, opt.resume = opt.resume, False # pass resume to exist_ok and disable resume opt.save_dir = str( increment_path(Path(opt.project) / opt.name, exist_ok=opt.exist_ok)) # DDP mode device = select_device(opt.device, batch_size=opt.batch_size) if LOCAL_RANK != -1: msg = 'is not compatible with YOLOv5 Multi-GPU DDP training' assert not opt.image_weights, f'--image-weights {msg}' assert not opt.evolve, f'--evolve {msg}' assert opt.batch_size != -1, f'AutoBatch with --batch-size -1 {msg}, please pass a valid --batch-size' assert opt.batch_size % WORLD_SIZE == 0, f'--batch-size {opt.batch_size} must be multiple of WORLD_SIZE' assert torch.cuda.device_count( ) > LOCAL_RANK, 'insufficient CUDA devices for DDP command' torch.cuda.set_device(LOCAL_RANK) device = torch.device('cuda', LOCAL_RANK) dist.init_process_group( backend="nccl" if dist.is_nccl_available() else "gloo") # Train if not opt.evolve: train(opt.hyp, opt, device, callbacks) if WORLD_SIZE > 1 and RANK == 0: LOGGER.info('Destroying process group... ') dist.destroy_process_group() # Evolve hyperparameters (optional) else: # Hyperparameter evolution metadata (mutation scale 0-1, lower_limit, upper_limit) meta = { 'lr0': (1, 1e-5, 1e-1), # initial learning rate (SGD=1E-2, Adam=1E-3) 'lrf': (1, 0.01, 1.0), # final OneCycleLR learning rate (lr0 * lrf) 'momentum': (0.3, 0.6, 0.98), # SGD momentum/Adam beta1 'weight_decay': (1, 0.0, 0.001), # optimizer weight decay 'warmup_epochs': (1, 0.0, 5.0), # warmup epochs (fractions ok) 'warmup_momentum': (1, 0.0, 0.95), # warmup initial momentum 'warmup_bias_lr': (1, 0.0, 0.2), # warmup initial bias lr 'box': (1, 0.02, 0.2), # box loss gain 'cls': (1, 0.2, 4.0), # cls loss gain 'cls_pw': (1, 0.5, 2.0), # cls BCELoss positive_weight 'obj': (1, 0.2, 4.0), # obj loss gain (scale with pixels) 'obj_pw': (1, 0.5, 2.0), # obj BCELoss positive_weight 'iou_t': (0, 0.1, 0.7), # IoU training threshold 'anchor_t': (1, 2.0, 8.0), # anchor-multiple threshold 'anchors': (2, 2.0, 10.0), # anchors per output grid (0 to ignore) 'fl_gamma': (0, 0.0, 2.0), # focal loss gamma (efficientDet default gamma=1.5) 'hsv_h': (1, 0.0, 0.1), # image HSV-Hue augmentation (fraction) 'hsv_s': (1, 0.0, 0.9), # image HSV-Saturation augmentation (fraction) 'hsv_v': (1, 0.0, 0.9), # image HSV-Value augmentation (fraction) 'degrees': (1, 0.0, 45.0), # image rotation (+/- deg) 'translate': (1, 0.0, 0.9), # image translation (+/- fraction) 'scale': (1, 0.0, 0.9), # image scale (+/- gain) 'shear': (1, 0.0, 10.0), # image shear (+/- deg) 'perspective': (0, 0.0, 0.001), # image perspective (+/- fraction), range 0-0.001 'flipud': (1, 0.0, 1.0), # image flip up-down (probability) 'fliplr': (0, 0.0, 1.0), # image flip left-right (probability) 'mosaic': (1, 0.0, 1.0), # image mixup (probability) 'mixup': (1, 0.0, 1.0), # image mixup (probability) 'copy_paste': (1, 0.0, 1.0) } # segment copy-paste (probability) with open(opt.hyp, errors='ignore') as f: hyp = yaml.safe_load(f) # load hyps dict if 'anchors' not in hyp: # anchors commented in hyp.yaml hyp['anchors'] = 3 opt.noval, opt.nosave, save_dir = True, True, Path( opt.save_dir) # only val/save final epoch # ei = [isinstance(x, (int, float)) for x in hyp.values()] # evolvable indices evolve_yaml, evolve_csv = save_dir / 'hyp_evolve.yaml', save_dir / 'evolve.csv' if opt.bucket: os.system(f'gsutil cp gs://{opt.bucket}/evolve.csv {evolve_csv}' ) # download evolve.csv if exists for _ in range(opt.evolve): # generations to evolve if evolve_csv.exists( ): # if evolve.csv exists: select best hyps and mutate # Select parent(s) parent = 'single' # parent selection method: 'single' or 'weighted' x = np.loadtxt(evolve_csv, ndmin=2, delimiter=',', skiprows=1) n = min(5, len(x)) # number of previous results to consider x = x[np.argsort(-fitness(x))][:n] # top n mutations w = fitness(x) - fitness(x).min() + 1E-6 # weights (sum > 0) if parent == 'single' or len(x) == 1: # x = x[random.randint(0, n - 1)] # random selection x = x[random.choices(range(n), weights=w)[0]] # weighted selection elif parent == 'weighted': x = (x * w.reshape( n, 1)).sum(0) / w.sum() # weighted combination # Mutate mp, s = 0.8, 0.2 # mutation probability, sigma npr = np.random npr.seed(int(time.time())) g = np.array([meta[k][0] for k in hyp.keys()]) # gains 0-1 ng = len(meta) v = np.ones(ng) while all( v == 1 ): # mutate until a change occurs (prevent duplicates) v = (g * (npr.random(ng) < mp) * npr.randn(ng) * npr.random() * s + 1).clip(0.3, 3.0) for i, k in enumerate(hyp.keys()): # plt.hist(v.ravel(), 300) hyp[k] = float(x[i + 7] * v[i]) # mutate # Constrain to limits for k, v in meta.items(): hyp[k] = max(hyp[k], v[1]) # lower limit hyp[k] = min(hyp[k], v[2]) # upper limit hyp[k] = round(hyp[k], 5) # significant digits # Train mutation results = train(hyp.copy(), opt, device, callbacks) callbacks = Callbacks() # Write mutation results print_mutation(results, hyp.copy(), save_dir, opt.bucket) # Plot results plot_evolve(evolve_csv) LOGGER.info( f'Hyperparameter evolution finished {opt.evolve} generations\n' f"Results saved to {colorstr('bold', save_dir)}\n" f'Usage example: $ python train.py --hyp {evolve_yaml}')
def train( hyp, # path/to/hyp.yaml or hyp dictionary opt, device, callbacks): save_dir, epochs, batch_size, weights, single_cls, evolve, data, cfg, resume, noval, nosave, workers, freeze = \ Path(opt.save_dir), opt.epochs, opt.batch_size, opt.weights, opt.single_cls, opt.evolve, opt.data, opt.cfg, \ opt.resume, opt.noval, opt.nosave, opt.workers, opt.freeze # Directories w = save_dir / 'weights' # weights dir (w.parent if evolve else w).mkdir(parents=True, exist_ok=True) # make dir last, best = w / 'last.pt', w / 'best.pt' # Hyperparameters if isinstance(hyp, str): with open(hyp, errors='ignore') as f: hyp = yaml.safe_load(f) # load hyps dict LOGGER.info( colorstr('hyperparameters: ') + ', '.join(f'{k}={v}' for k, v in hyp.items())) # Save run settings if not evolve: with open(save_dir / 'hyp.yaml', 'w') as f: yaml.safe_dump(hyp, f, sort_keys=False) with open(save_dir / 'opt.yaml', 'w') as f: yaml.safe_dump(vars(opt), f, sort_keys=False) # Loggers data_dict = None if RANK in [-1, 0]: loggers = Loggers(save_dir, weights, opt, hyp, LOGGER) # loggers instance if loggers.wandb: data_dict = loggers.wandb.data_dict if resume: weights, epochs, hyp, batch_size = opt.weights, opt.epochs, opt.hyp, opt.batch_size # Register actions for k in methods(loggers): callbacks.register_action(k, callback=getattr(loggers, k)) # Config plots = not evolve # create plots cuda = device.type != 'cpu' init_seeds(1 + RANK) with torch_distributed_zero_first(LOCAL_RANK): data_dict = data_dict or check_dataset(data) # check if None train_path, val_path = data_dict['train'], data_dict['val'] nc = 1 if single_cls else int(data_dict['nc']) # number of classes names = ['item'] if single_cls and len( data_dict['names']) != 1 else data_dict['names'] # class names assert len( names ) == nc, f'{len(names)} names found for nc={nc} dataset in {data}' # check is_coco = isinstance(val_path, str) and val_path.endswith( 'coco/val2017.txt') # COCO dataset # Model check_suffix(weights, '.pt') # check weights pretrained = weights.endswith('.pt') if pretrained: with torch_distributed_zero_first(LOCAL_RANK): weights = attempt_download( weights) # download if not found locally ckpt = torch.load(weights, map_location='cpu' ) # load checkpoint to CPU to avoid CUDA memory leak model = Model(cfg or ckpt['model'].yaml, ch=3, nc=nc, anchors=hyp.get('anchors')).to(device) # create exclude = [ 'anchor' ] if (cfg or hyp.get('anchors')) and not resume else [] # exclude keys csd = ckpt['model'].float().state_dict( ) # checkpoint state_dict as FP32 csd = intersect_dicts(csd, model.state_dict(), exclude=exclude) # intersect model.load_state_dict(csd, strict=False) # load LOGGER.info( f'Transferred {len(csd)}/{len(model.state_dict())} items from {weights}' ) # report else: model = Model(cfg, ch=3, nc=nc, anchors=hyp.get('anchors')).to(device) # create # Freeze freeze = [ f'model.{x}.' for x in (freeze if len(freeze) > 1 else range(freeze[0])) ] # layers to freeze for k, v in model.named_parameters(): v.requires_grad = True # train all layers if any(x in k for x in freeze): LOGGER.info(f'freezing {k}') v.requires_grad = False # Image size gs = max(int(model.stride.max()), 32) # grid size (max stride) imgsz = check_img_size(opt.imgsz, gs, floor=gs * 2) # verify imgsz is gs-multiple # Batch size if RANK == -1 and batch_size == -1: # single-GPU only, estimate best batch size batch_size = check_train_batch_size(model, imgsz) loggers.on_params_update({"batch_size": batch_size}) # Optimizer nbs = 64 # nominal batch size accumulate = max(round(nbs / batch_size), 1) # accumulate loss before optimizing hyp['weight_decay'] *= batch_size * accumulate / nbs # scale weight_decay LOGGER.info(f"Scaled weight_decay = {hyp['weight_decay']}") g0, g1, g2 = [], [], [] # optimizer parameter groups for v in model.modules(): if hasattr(v, 'bias') and isinstance(v.bias, nn.Parameter): # bias g2.append(v.bias) if isinstance(v, nn.BatchNorm2d): # weight (no decay) g0.append(v.weight) elif hasattr(v, 'weight') and isinstance( v.weight, nn.Parameter): # weight (with decay) g1.append(v.weight) if opt.optimizer == 'Adam': optimizer = Adam(g0, lr=hyp['lr0'], betas=(hyp['momentum'], 0.999)) # adjust beta1 to momentum elif opt.optimizer == 'AdamW': optimizer = AdamW(g0, lr=hyp['lr0'], betas=(hyp['momentum'], 0.999)) # adjust beta1 to momentum else: optimizer = SGD(g0, lr=hyp['lr0'], momentum=hyp['momentum'], nesterov=True) optimizer.add_param_group({ 'params': g1, 'weight_decay': hyp['weight_decay'] }) # add g1 with weight_decay optimizer.add_param_group({'params': g2}) # add g2 (biases) LOGGER.info( f"{colorstr('optimizer:')} {type(optimizer).__name__} with parameter groups " f"{len(g0)} weight (no decay), {len(g1)} weight, {len(g2)} bias") del g0, g1, g2 # Scheduler if opt.cos_lr: lf = one_cycle(1, hyp['lrf'], epochs) # cosine 1->hyp['lrf'] else: lf = lambda x: (1 - x / epochs) * (1.0 - hyp['lrf']) + hyp['lrf' ] # linear scheduler = lr_scheduler.LambdaLR( optimizer, lr_lambda=lf) # plot_lr_scheduler(optimizer, scheduler, epochs) # EMA ema = ModelEMA(model) if RANK in [-1, 0] else None # Resume start_epoch, best_fitness = 0, 0.0 if pretrained: # Optimizer if ckpt['optimizer'] is not None: optimizer.load_state_dict(ckpt['optimizer']) best_fitness = ckpt['best_fitness'] # EMA if ema and ckpt.get('ema'): ema.ema.load_state_dict(ckpt['ema'].float().state_dict()) ema.updates = ckpt['updates'] # Epochs start_epoch = ckpt['epoch'] + 1 if resume: assert start_epoch > 0, f'{weights} training to {epochs} epochs is finished, nothing to resume.' if epochs < start_epoch: LOGGER.info( f"{weights} has been trained for {ckpt['epoch']} epochs. Fine-tuning for {epochs} more epochs." ) epochs += ckpt['epoch'] # finetune additional epochs del ckpt, csd # DP mode if cuda and RANK == -1 and torch.cuda.device_count() > 1: LOGGER.warning( 'WARNING: DP not recommended, use torch.distributed.run for best DDP Multi-GPU results.\n' 'See Multi-GPU Tutorial at https://github.com/ultralytics/yolov5/issues/475 to get started.' ) model = torch.nn.DataParallel(model) # SyncBatchNorm if opt.sync_bn and cuda and RANK != -1: model = torch.nn.SyncBatchNorm.convert_sync_batchnorm(model).to(device) LOGGER.info('Using SyncBatchNorm()') # Trainloader train_loader, dataset = create_dataloader( train_path, imgsz, batch_size // WORLD_SIZE, gs, single_cls, hyp=hyp, augment=True, cache=None if opt.cache == 'val' else opt.cache, rect=opt.rect, rank=LOCAL_RANK, workers=workers, image_weights=opt.image_weights, quad=opt.quad, prefix=colorstr('train: '), shuffle=True) mlc = int(np.concatenate(dataset.labels, 0)[:, 0].max()) # max label class nb = len(train_loader) # number of batches assert mlc < nc, f'Label class {mlc} exceeds nc={nc} in {data}. Possible class labels are 0-{nc - 1}' # Process 0 if RANK in [-1, 0]: val_loader = create_dataloader(val_path, imgsz, batch_size // WORLD_SIZE * 2, gs, single_cls, hyp=hyp, cache=None if noval else opt.cache, rect=True, rank=-1, workers=workers * 2, pad=0.5, prefix=colorstr('val: '))[0] if not resume: labels = np.concatenate(dataset.labels, 0) # c = torch.tensor(labels[:, 0]) # classes # cf = torch.bincount(c.long(), minlength=nc) + 1. # frequency # model._initialize_biases(cf.to(device)) if plots: plot_labels(labels, names, save_dir) # Anchors if not opt.noautoanchor: check_anchors(dataset, model=model, thr=hyp['anchor_t'], imgsz=imgsz) model.half().float() # pre-reduce anchor precision callbacks.run('on_pretrain_routine_end') # DDP mode if cuda and RANK != -1: model = DDP(model, device_ids=[LOCAL_RANK], output_device=LOCAL_RANK) # Model attributes nl = de_parallel( model).model[-1].nl # number of detection layers (to scale hyps) hyp['box'] *= 3 / nl # scale to layers hyp['cls'] *= nc / 80 * 3 / nl # scale to classes and layers hyp['obj'] *= (imgsz / 640)**2 * 3 / nl # scale to image size and layers hyp['label_smoothing'] = opt.label_smoothing model.nc = nc # attach number of classes to model model.hyp = hyp # attach hyperparameters to model model.class_weights = labels_to_class_weights( dataset.labels, nc).to(device) * nc # attach class weights model.names = names # Start training t0 = time.time() nw = max(round(hyp['warmup_epochs'] * nb), 1000) # number of warmup iterations, max(3 epochs, 1k iterations) # nw = min(nw, (epochs - start_epoch) / 2 * nb) # limit warmup to < 1/2 of training last_opt_step = -1 maps = np.zeros(nc) # mAP per class results = (0, 0, 0, 0, 0, 0, 0 ) # P, R, [email protected], [email protected], val_loss(box, obj, cls) scheduler.last_epoch = start_epoch - 1 # do not move scaler = amp.GradScaler(enabled=cuda) stopper = EarlyStopping(patience=opt.patience) compute_loss = ComputeLoss(model) # init loss class LOGGER.info( f'Image sizes {imgsz} train, {imgsz} val\n' f'Using {train_loader.num_workers * WORLD_SIZE} dataloader workers\n' f"Logging results to {colorstr('bold', save_dir)}\n" f'Starting training for {epochs} epochs...') for epoch in range( start_epoch, epochs ): # epoch ------------------------------------------------------------------ model.train() # Update image weights (optional, single-GPU only) if opt.image_weights: cw = model.class_weights.cpu().numpy() * ( 1 - maps)**2 / nc # class weights iw = labels_to_image_weights(dataset.labels, nc=nc, class_weights=cw) # image weights dataset.indices = random.choices(range(dataset.n), weights=iw, k=dataset.n) # rand weighted idx # Update mosaic border (optional) # b = int(random.uniform(0.25 * imgsz, 0.75 * imgsz + gs) // gs * gs) # dataset.mosaic_border = [b - imgsz, -b] # height, width borders mloss = torch.zeros(3, device=device) # mean losses if RANK != -1: train_loader.sampler.set_epoch(epoch) pbar = enumerate(train_loader) LOGGER.info( ('\n' + '%10s' * 7) % ('Epoch', 'gpu_mem', 'box', 'obj', 'cls', 'labels', 'img_size')) if RANK in [-1, 0]: pbar = tqdm( pbar, total=nb, bar_format='{l_bar}{bar:10}{r_bar}{bar:-10b}') # progress bar optimizer.zero_grad() for i, ( imgs, targets, paths, _ ) in pbar: # batch ------------------------------------------------------------- ni = i + nb * epoch # number integrated batches (since train start) imgs = imgs.to(device, non_blocking=True).float( ) / 255 # uint8 to float32, 0-255 to 0.0-1.0 # Warmup if ni <= nw: xi = [0, nw] # x interp # compute_loss.gr = np.interp(ni, xi, [0.0, 1.0]) # iou loss ratio (obj_loss = 1.0 or iou) accumulate = max( 1, np.interp(ni, xi, [1, nbs / batch_size]).round()) for j, x in enumerate(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, [ hyp['warmup_bias_lr'] if j == 2 else 0.0, x['initial_lr'] * lf(epoch) ]) if 'momentum' in x: x['momentum'] = np.interp( ni, xi, [hyp['warmup_momentum'], hyp['momentum']]) # Multi-scale if opt.multi_scale: sz = random.randrange(imgsz * 0.5, imgsz * 1.5 + gs) // gs * gs # size sf = sz / max(imgs.shape[2:]) # scale factor if sf != 1: ns = [math.ceil(x * sf / gs) * gs for x in imgs.shape[2:] ] # new shape (stretched to gs-multiple) imgs = nn.functional.interpolate(imgs, size=ns, mode='bilinear', align_corners=False) # Forward with amp.autocast(enabled=cuda): pred = model(imgs) # forward loss, loss_items = compute_loss( pred, targets.to(device)) # loss scaled by batch_size if RANK != -1: loss *= WORLD_SIZE # gradient averaged between devices in DDP mode if opt.quad: loss *= 4. # Backward scaler.scale(loss).backward() # Optimize if ni - last_opt_step >= accumulate: scaler.step(optimizer) # optimizer.step scaler.update() optimizer.zero_grad() if ema: ema.update(model) last_opt_step = ni # Log if RANK in [-1, 0]: mloss = (mloss * i + loss_items) / (i + 1 ) # update mean losses mem = f'{torch.cuda.memory_reserved() / 1E9 if torch.cuda.is_available() else 0:.3g}G' # (GB) pbar.set_description(('%10s' * 2 + '%10.4g' * 5) % (f'{epoch}/{epochs - 1}', mem, *mloss, targets.shape[0], imgs.shape[-1])) callbacks.run('on_train_batch_end', ni, model, imgs, targets, paths, plots, opt.sync_bn) if callbacks.stop_training: return # end batch ------------------------------------------------------------------------------------------------ # Scheduler lr = [x['lr'] for x in optimizer.param_groups] # for loggers scheduler.step() if RANK in [-1, 0]: # mAP callbacks.run('on_train_epoch_end', epoch=epoch) ema.update_attr(model, include=[ 'yaml', 'nc', 'hyp', 'names', 'stride', 'class_weights' ]) final_epoch = (epoch + 1 == epochs) or stopper.possible_stop if not noval or final_epoch: # Calculate mAP results, maps, _ = val.run(data_dict, batch_size=batch_size // WORLD_SIZE * 2, imgsz=imgsz, model=ema.ema, single_cls=single_cls, dataloader=val_loader, save_dir=save_dir, plots=False, callbacks=callbacks, compute_loss=compute_loss) # Update best mAP fi = fitness(np.array(results).reshape( 1, -1)) # weighted combination of [P, R, [email protected], [email protected]] if fi > best_fitness: best_fitness = fi log_vals = list(mloss) + list(results) + lr callbacks.run('on_fit_epoch_end', log_vals, epoch, best_fitness, fi) # Save model if (not nosave) or (final_epoch and not evolve): # if save ckpt = { 'epoch': epoch, 'best_fitness': best_fitness, 'model': deepcopy(de_parallel(model)).half(), 'ema': deepcopy(ema.ema).half(), 'updates': ema.updates, 'optimizer': optimizer.state_dict(), 'wandb_id': loggers.wandb.wandb_run.id if loggers.wandb else None, 'date': datetime.now().isoformat() } # Save last, best and delete torch.save(ckpt, last) if best_fitness == fi: torch.save(ckpt, best) if (epoch > 0) and (opt.save_period > 0) and (epoch % opt.save_period == 0): torch.save(ckpt, w / f'epoch{epoch}.pt') del ckpt callbacks.run('on_model_save', last, epoch, final_epoch, best_fitness, fi) # Stop Single-GPU if RANK == -1 and stopper(epoch=epoch, fitness=fi): break # Stop DDP TODO: known issues shttps://github.com/ultralytics/yolov5/pull/4576 # stop = stopper(epoch=epoch, fitness=fi) # if RANK == 0: # dist.broadcast_object_list([stop], 0) # broadcast 'stop' to all ranks # Stop DPP # with torch_distributed_zero_first(RANK): # if stop: # break # must break all DDP ranks # end epoch ---------------------------------------------------------------------------------------------------- # end training ----------------------------------------------------------------------------------------------------- if RANK in [-1, 0]: LOGGER.info( f'\n{epoch - start_epoch + 1} epochs completed in {(time.time() - t0) / 3600:.3f} hours.' ) for f in last, best: if f.exists(): strip_optimizer(f) # strip optimizers if f is best: LOGGER.info(f'\nValidating {f}...') results, _, _ = val.run( data_dict, batch_size=batch_size // WORLD_SIZE * 2, imgsz=imgsz, model=attempt_load(f, device).half(), iou_thres=0.65 if is_coco else 0.60, # best pycocotools results at 0.65 single_cls=single_cls, dataloader=val_loader, save_dir=save_dir, save_json=is_coco, verbose=True, plots=True, callbacks=callbacks, compute_loss=compute_loss) # val best model with plots if is_coco: callbacks.run('on_fit_epoch_end', list(mloss) + list(results) + lr, epoch, best_fitness, fi) callbacks.run('on_train_end', last, best, plots, epoch, results) LOGGER.info(f"Results saved to {colorstr('bold', save_dir)}") torch.cuda.empty_cache() return results
def main(opt): set_logging(RANK) if RANK in [-1, 0]: print( colorstr('train: ') + ', '.join(f'{k}={v}' for k, v in vars(opt).items())) check_git_status() check_requirements(exclude=['thop']) # Resume wandb_run = check_wandb_resume(opt) if opt.resume and not wandb_run: # resume an interrupted run ckpt = opt.resume if isinstance( opt.resume, str) else get_latest_run() # specified or most recent path assert os.path.isfile( ckpt), 'ERROR: --resume checkpoint does not exist' with open(Path(ckpt).parent.parent / 'opt.yaml') as f: opt = argparse.Namespace(**yaml.safe_load(f)) # replace opt.cfg, opt.weights, opt.resume = '', ckpt, True # reinstate logger.info('Resuming training from %s' % ckpt) else: # opt.hyp = opt.hyp or ('hyp.finetune.yaml' if opt.weights else 'hyp.scratch.yaml') opt.data, opt.cfg, opt.hyp = check_file(opt.data), check_file( opt.cfg), check_file(opt.hyp) # check files assert len(opt.cfg) or len( opt.weights), 'either --cfg or --weights must be specified' opt.img_size.extend( [opt.img_size[-1]] * (2 - len(opt.img_size))) # extend to 2 sizes (train, test) opt.name = 'evolve' if opt.evolve else opt.name opt.save_dir = str( increment_path(Path(opt.project) / opt.name, exist_ok=opt.exist_ok | opt.evolve)) # DDP mode device = select_device(opt.device, batch_size=opt.batch_size) if LOCAL_RANK != -1: from datetime import timedelta assert torch.cuda.device_count( ) > LOCAL_RANK, 'insufficient CUDA devices for DDP command' torch.cuda.set_device(LOCAL_RANK) device = torch.device('cuda', LOCAL_RANK) dist.init_process_group( backend="nccl" if dist.is_nccl_available() else "gloo", timeout=timedelta(seconds=60)) assert opt.batch_size % WORLD_SIZE == 0, '--batch-size must be multiple of CUDA device count' assert not opt.image_weights, '--image-weights argument is not compatible with DDP training' # Train if not opt.evolve: train(opt.hyp, opt, device) if WORLD_SIZE > 1 and RANK == 0: _ = [ print('Destroying process group... ', end=''), dist.destroy_process_group(), print('Done.') ] # Evolve hyperparameters (optional) else: # Hyperparameter evolution metadata (mutation scale 0-1, lower_limit, upper_limit) meta = { 'lr0': (1, 1e-5, 1e-1), # initial learning rate (SGD=1E-2, Adam=1E-3) 'lrf': (1, 0.01, 1.0), # final OneCycleLR learning rate (lr0 * lrf) 'momentum': (0.3, 0.6, 0.98), # SGD momentum/Adam beta1 'weight_decay': (1, 0.0, 0.001), # optimizer weight decay 'warmup_epochs': (1, 0.0, 5.0), # warmup epochs (fractions ok) 'warmup_momentum': (1, 0.0, 0.95), # warmup initial momentum 'warmup_bias_lr': (1, 0.0, 0.2), # warmup initial bias lr 'box': (1, 0.02, 0.2), # box loss gain 'cls': (1, 0.2, 4.0), # cls loss gain 'cls_pw': (1, 0.5, 2.0), # cls BCELoss positive_weight 'obj': (1, 0.2, 4.0), # obj loss gain (scale with pixels) 'obj_pw': (1, 0.5, 2.0), # obj BCELoss positive_weight 'iou_t': (0, 0.1, 0.7), # IoU training threshold 'anchor_t': (1, 2.0, 8.0), # anchor-multiple threshold 'anchors': (2, 2.0, 10.0), # anchors per output grid (0 to ignore) 'fl_gamma': (0, 0.0, 2.0), # focal loss gamma (efficientDet default gamma=1.5) 'hsv_h': (1, 0.0, 0.1), # image HSV-Hue augmentation (fraction) 'hsv_s': (1, 0.0, 0.9), # image HSV-Saturation augmentation (fraction) 'hsv_v': (1, 0.0, 0.9), # image HSV-Value augmentation (fraction) 'degrees': (1, 0.0, 45.0), # image rotation (+/- deg) 'translate': (1, 0.0, 0.9), # image translation (+/- fraction) 'scale': (1, 0.0, 0.9), # image scale (+/- gain) 'shear': (1, 0.0, 10.0), # image shear (+/- deg) 'perspective': (0, 0.0, 0.001), # image perspective (+/- fraction), range 0-0.001 'flipud': (1, 0.0, 1.0), # image flip up-down (probability) 'fliplr': (0, 0.0, 1.0), # image flip left-right (probability) 'mosaic': (1, 0.0, 1.0), # image mixup (probability) 'mixup': (1, 0.0, 1.0) } # image mixup (probability) with open(opt.hyp) as f: hyp = yaml.safe_load(f) # load hyps dict assert LOCAL_RANK == -1, 'DDP mode not implemented for --evolve' opt.notest, opt.nosave = True, True # only test/save final epoch # ei = [isinstance(x, (int, float)) for x in hyp.values()] # evolvable indices yaml_file = Path( opt.save_dir) / 'hyp_evolved.yaml' # save best result here if opt.bucket: os.system('gsutil cp gs://%s/evolve.txt .' % opt.bucket) # download evolve.txt if exists for _ in range(300): # generations to evolve if Path('evolve.txt').exists( ): # if evolve.txt exists: select best hyps and mutate # Select parent(s) parent = 'single' # parent selection method: 'single' or 'weighted' x = np.loadtxt('evolve.txt', ndmin=2) n = min(5, len(x)) # number of previous results to consider x = x[np.argsort(-fitness(x))][:n] # top n mutations w = fitness(x) - fitness(x).min() + 1E-6 # weights (sum > 0) if parent == 'single' or len(x) == 1: # x = x[random.randint(0, n - 1)] # random selection x = x[random.choices(range(n), weights=w)[0]] # weighted selection elif parent == 'weighted': x = (x * w.reshape( n, 1)).sum(0) / w.sum() # weighted combination # Mutate mp, s = 0.8, 0.2 # mutation probability, sigma npr = np.random npr.seed(int(time.time())) g = np.array([x[0] for x in meta.values()]) # gains 0-1 ng = len(meta) v = np.ones(ng) while all( v == 1 ): # mutate until a change occurs (prevent duplicates) v = (g * (npr.random(ng) < mp) * npr.randn(ng) * npr.random() * s + 1).clip(0.3, 3.0) for i, k in enumerate(hyp.keys()): # plt.hist(v.ravel(), 300) hyp[k] = float(x[i + 7] * v[i]) # mutate # Constrain to limits for k, v in meta.items(): hyp[k] = max(hyp[k], v[1]) # lower limit hyp[k] = min(hyp[k], v[2]) # upper limit hyp[k] = round(hyp[k], 5) # significant digits # Train mutation results = train(hyp.copy(), opt, device) # Write mutation results print_mutation(hyp.copy(), results, yaml_file, opt.bucket) # Plot results plot_evolution(yaml_file) print( f'Hyperparameter evolution complete. Best results saved as: {yaml_file}\n' f'Command to train a new model with these hyperparameters: $ python train.py --hyp {yaml_file}' )
def train( hyp, # path/to/hyp.yaml or hyp dictionary opt, device, ): save_dir, epochs, batch_size, weights, single_cls, evolve, data, cfg, resume, notest, nosave, workers, = \ opt.save_dir, opt.epochs, opt.batch_size, opt.weights, opt.single_cls, opt.evolve, opt.data, opt.cfg, \ opt.resume, opt.notest, opt.nosave, opt.workers # Directories save_dir = Path(save_dir) wdir = save_dir / 'weights' wdir.mkdir(parents=True, exist_ok=True) # make dir last = wdir / 'last.pt' best = wdir / 'best.pt' results_file = save_dir / 'results.txt' # Hyperparameters if isinstance(hyp, str): with open(hyp) as f: hyp = yaml.safe_load(f) # load hyps dict logger.info( colorstr('hyperparameters: ') + ', '.join(f'{k}={v}' for k, v in hyp.items())) # Save run settings with open(save_dir / 'hyp.yaml', 'w') as f: yaml.safe_dump(hyp, f, sort_keys=False) with open(save_dir / 'opt.yaml', 'w') as f: yaml.safe_dump(vars(opt), f, sort_keys=False) # Configure plots = not evolve # create plots cuda = device.type != 'cpu' init_seeds(1 + RANK) with open(data) as f: data_dict = yaml.safe_load(f) # data dict # Loggers loggers = {'wandb': None, 'tb': None} # loggers dict if RANK in [-1, 0]: # TensorBoard if not evolve: prefix = colorstr('tensorboard: ') logger.info( f"{prefix}Start with 'tensorboard --logdir {opt.project}', view at http://localhost:6006/" ) loggers['tb'] = SummaryWriter(str(save_dir)) # W&B opt.hyp = hyp # add hyperparameters run_id = torch.load(weights).get('wandb_id') if weights.endswith( '.pt') and os.path.isfile(weights) else None run_id = run_id if opt.resume else None # start fresh run if transfer learning wandb_logger = WandbLogger(opt, save_dir.stem, run_id, data_dict) loggers['wandb'] = wandb_logger.wandb if loggers['wandb']: data_dict = wandb_logger.data_dict weights, epochs, hyp = opt.weights, opt.epochs, opt.hyp # may update weights, epochs if resuming nc = 1 if single_cls else int(data_dict['nc']) # number of classes names = ['item'] if single_cls and len( data_dict['names']) != 1 else data_dict['names'] # class names assert len(names) == nc, '%g names found for nc=%g dataset in %s' % ( len(names), nc, data) # check is_coco = data.endswith('coco.yaml') and nc == 80 # COCO dataset # Model pretrained = weights.endswith('.pt') if pretrained: with torch_distributed_zero_first(RANK): weights = attempt_download( weights) # download if not found locally ckpt = torch.load(weights, map_location=device) # load checkpoint model = Model(cfg or ckpt['model'].yaml, ch=3, nc=nc, anchors=hyp.get('anchors')).to(device) # create exclude = [ 'anchor' ] if (cfg or hyp.get('anchors')) and not resume else [] # exclude keys state_dict = ckpt['model'].float().state_dict() # to FP32 state_dict = intersect_dicts(state_dict, model.state_dict(), exclude=exclude) # intersect model.load_state_dict(state_dict, strict=False) # load logger.info( 'Transferred %g/%g items from %s' % (len(state_dict), len(model.state_dict()), weights)) # report else: model = Model(cfg, ch=3, nc=nc, anchors=hyp.get('anchors')).to(device) # create with torch_distributed_zero_first(RANK): check_dataset(data_dict) # check train_path = data_dict['train'] test_path = data_dict['val'] # Freeze freeze = [] # parameter names to freeze (full or partial) for k, v in model.named_parameters(): v.requires_grad = True # train all layers if any(x in k for x in freeze): print('freezing %s' % k) v.requires_grad = False # Optimizer nbs = 64 # nominal batch size accumulate = max(round(nbs / batch_size), 1) # accumulate loss before optimizing hyp['weight_decay'] *= batch_size * accumulate / nbs # scale weight_decay logger.info(f"Scaled weight_decay = {hyp['weight_decay']}") pg0, pg1, pg2 = [], [], [] # optimizer parameter groups for k, v in model.named_modules(): if hasattr(v, 'bias') and isinstance(v.bias, nn.Parameter): pg2.append(v.bias) # biases if isinstance(v, nn.BatchNorm2d): pg0.append(v.weight) # no decay elif hasattr(v, 'weight') and isinstance(v.weight, nn.Parameter): pg1.append(v.weight) # apply decay if opt.adam: optimizer = optim.Adam(pg0, lr=hyp['lr0'], betas=(hyp['momentum'], 0.999)) # adjust beta1 to momentum else: optimizer = optim.SGD(pg0, lr=hyp['lr0'], momentum=hyp['momentum'], nesterov=True) optimizer.add_param_group({ 'params': pg1, 'weight_decay': hyp['weight_decay'] }) # add pg1 with weight_decay optimizer.add_param_group({'params': pg2}) # add pg2 (biases) logger.info('Optimizer groups: %g .bias, %g conv.weight, %g other' % (len(pg2), len(pg1), len(pg0))) del pg0, pg1, pg2 # Scheduler https://arxiv.org/pdf/1812.01187.pdf # https://pytorch.org/docs/stable/_modules/torch/optim/lr_scheduler.html#OneCycleLR if opt.linear_lr: lf = lambda x: (1 - x / (epochs - 1)) * (1.0 - hyp['lrf']) + hyp[ 'lrf'] # linear else: lf = one_cycle(1, hyp['lrf'], epochs) # cosine 1->hyp['lrf'] scheduler = lr_scheduler.LambdaLR(optimizer, lr_lambda=lf) # plot_lr_scheduler(optimizer, scheduler, epochs) # EMA ema = ModelEMA(model) if RANK in [-1, 0] else None # Resume start_epoch, best_fitness = 0, 0.0 if pretrained: # Optimizer if ckpt['optimizer'] is not None: optimizer.load_state_dict(ckpt['optimizer']) best_fitness = ckpt['best_fitness'] # EMA if ema and ckpt.get('ema'): ema.ema.load_state_dict(ckpt['ema'].float().state_dict()) ema.updates = ckpt['updates'] # Results if ckpt.get('training_results') is not None: results_file.write_text( ckpt['training_results']) # write results.txt # Epochs start_epoch = ckpt['epoch'] + 1 if resume: assert start_epoch > 0, '%s training to %g epochs is finished, nothing to resume.' % ( weights, epochs) if epochs < start_epoch: logger.info( '%s has been trained for %g epochs. Fine-tuning for %g additional epochs.' % (weights, ckpt['epoch'], epochs)) epochs += ckpt['epoch'] # finetune additional epochs del ckpt, state_dict # Image sizes gs = max(int(model.stride.max()), 32) # grid size (max stride) nl = model.model[ -1].nl # number of detection layers (used for scaling hyp['obj']) imgsz, imgsz_test = [check_img_size(x, gs) for x in opt.img_size ] # verify imgsz are gs-multiples # DP mode if cuda and RANK == -1 and torch.cuda.device_count() > 1: logging.warning( 'DP not recommended, instead use torch.distributed.run for best DDP Multi-GPU results.\n' 'See Multi-GPU Tutorial at https://github.com/ultralytics/yolov5/issues/475 to get started.' ) model = torch.nn.DataParallel(model) # SyncBatchNorm if opt.sync_bn and cuda and RANK != -1: model = torch.nn.SyncBatchNorm.convert_sync_batchnorm(model).to(device) logger.info('Using SyncBatchNorm()') # Trainloader dataloader, dataset = create_dataloader(train_path, imgsz, batch_size // WORLD_SIZE, gs, single_cls, hyp=hyp, augment=True, cache=opt.cache_images, rect=opt.rect, rank=RANK, workers=workers, image_weights=opt.image_weights, quad=opt.quad, prefix=colorstr('train: ')) mlc = np.concatenate(dataset.labels, 0)[:, 0].max() # max label class nb = len(dataloader) # number of batches assert mlc < nc, 'Label class %g exceeds nc=%g in %s. Possible class labels are 0-%g' % ( mlc, nc, data, nc - 1) # Process 0 if RANK in [-1, 0]: testloader = create_dataloader(test_path, imgsz_test, batch_size // WORLD_SIZE * 2, gs, single_cls, hyp=hyp, cache=opt.cache_images and not notest, rect=True, rank=-1, workers=workers, pad=0.5, prefix=colorstr('val: '))[0] if not resume: labels = np.concatenate(dataset.labels, 0) c = torch.tensor(labels[:, 0]) # classes # cf = torch.bincount(c.long(), minlength=nc) + 1. # frequency # model._initialize_biases(cf.to(device)) if plots: plot_labels(labels, names, save_dir, loggers) if loggers['tb']: loggers['tb'].add_histogram('classes', c, 0) # TensorBoard # Anchors if not opt.noautoanchor: check_anchors(dataset, model=model, thr=hyp['anchor_t'], imgsz=imgsz) model.half().float() # pre-reduce anchor precision # DDP mode if cuda and RANK != -1: model = DDP(model, device_ids=[LOCAL_RANK], output_device=LOCAL_RANK) # Model parameters hyp['box'] *= 3. / nl # scale to layers hyp['cls'] *= nc / 80. * 3. / nl # scale to classes and layers hyp['obj'] *= (imgsz / 640)**2 * 3. / nl # scale to image size and layers hyp['label_smoothing'] = opt.label_smoothing model.nc = nc # attach number of classes to model model.hyp = hyp # attach hyperparameters to model model.gr = 1.0 # iou loss ratio (obj_loss = 1.0 or iou) model.class_weights = labels_to_class_weights( dataset.labels, nc).to(device) * nc # attach class weights model.names = names # Start training t0 = time.time() nw = max(round(hyp['warmup_epochs'] * nb), 1000) # number of warmup iterations, max(3 epochs, 1k iterations) # nw = min(nw, (epochs - start_epoch) / 2 * nb) # limit warmup to < 1/2 of training last_opt_step = -1 maps = np.zeros(nc) # mAP per class results = (0, 0, 0, 0, 0, 0, 0 ) # P, R, [email protected], [email protected], val_loss(box, obj, cls) scheduler.last_epoch = start_epoch - 1 # do not move scaler = amp.GradScaler(enabled=cuda) compute_loss = ComputeLoss(model) # init loss class logger.info(f'Image sizes {imgsz} train, {imgsz_test} test\n' f'Using {dataloader.num_workers} dataloader workers\n' f'Logging results to {save_dir}\n' f'Starting training for {epochs} epochs...') for epoch in range( start_epoch, epochs ): # epoch ------------------------------------------------------------------ model.train() # Update image weights (optional) if opt.image_weights: # Generate indices if RANK in [-1, 0]: cw = model.class_weights.cpu().numpy() * ( 1 - maps)**2 / nc # class weights iw = labels_to_image_weights(dataset.labels, nc=nc, class_weights=cw) # image weights dataset.indices = random.choices( range(dataset.n), weights=iw, k=dataset.n) # rand weighted idx # Broadcast if DDP if RANK != -1: indices = (torch.tensor(dataset.indices) if RANK == 0 else torch.zeros(dataset.n)).int() dist.broadcast(indices, 0) if RANK != 0: dataset.indices = indices.cpu().numpy() # Update mosaic border # b = int(random.uniform(0.25 * imgsz, 0.75 * imgsz + gs) // gs * gs) # dataset.mosaic_border = [b - imgsz, -b] # height, width borders mloss = torch.zeros(4, device=device) # mean losses if RANK != -1: dataloader.sampler.set_epoch(epoch) pbar = enumerate(dataloader) logger.info( ('\n' + '%10s' * 8) % ('Epoch', 'gpu_mem', 'box', 'obj', 'cls', 'total', 'labels', 'img_size')) if RANK in [-1, 0]: pbar = tqdm(pbar, total=nb) # progress bar optimizer.zero_grad() for i, ( imgs, targets, paths, _ ) in pbar: # batch ------------------------------------------------------------- ni = i + nb * epoch # number integrated batches (since train start) imgs = imgs.to(device, non_blocking=True).float( ) / 255.0 # uint8 to float32, 0-255 to 0.0-1.0 # Warmup if ni <= nw: xi = [0, nw] # x interp # model.gr = np.interp(ni, xi, [0.0, 1.0]) # iou loss ratio (obj_loss = 1.0 or iou) accumulate = max( 1, np.interp(ni, xi, [1, nbs / batch_size]).round()) for j, x in enumerate(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, [ hyp['warmup_bias_lr'] if j == 2 else 0.0, x['initial_lr'] * lf(epoch) ]) if 'momentum' in x: x['momentum'] = np.interp( ni, xi, [hyp['warmup_momentum'], hyp['momentum']]) # Multi-scale if opt.multi_scale: sz = random.randrange(imgsz * 0.5, imgsz * 1.5 + gs) // gs * gs # size sf = sz / max(imgs.shape[2:]) # scale factor if sf != 1: ns = [math.ceil(x * sf / gs) * gs for x in imgs.shape[2:] ] # new shape (stretched to gs-multiple) imgs = F.interpolate(imgs, size=ns, mode='bilinear', align_corners=False) # Forward with amp.autocast(enabled=cuda): pred = model(imgs) # forward loss, loss_items = compute_loss( pred, targets.to(device)) # loss scaled by batch_size if RANK != -1: loss *= WORLD_SIZE # gradient averaged between devices in DDP mode if opt.quad: loss *= 4. # Backward scaler.scale(loss).backward() # Optimize if ni - last_opt_step >= accumulate: scaler.step(optimizer) # optimizer.step scaler.update() optimizer.zero_grad() if ema: ema.update(model) last_opt_step = ni # Print if RANK in [-1, 0]: mloss = (mloss * i + loss_items) / (i + 1 ) # update mean losses mem = '%.3gG' % (torch.cuda.memory_reserved() / 1E9 if torch.cuda.is_available() else 0) # (GB) s = ('%10s' * 2 + '%10.4g' * 6) % (f'{epoch}/{epochs - 1}', mem, *mloss, targets.shape[0], imgs.shape[-1]) pbar.set_description(s) # Plot if plots and ni < 3: f = save_dir / f'train_batch{ni}.jpg' # filename Thread(target=plot_images, args=(imgs, targets, paths, f), daemon=True).start() if loggers['tb'] and ni == 0: # TensorBoard with warnings.catch_warnings(): warnings.simplefilter( 'ignore') # suppress jit trace warning loggers['tb'].add_graph( torch.jit.trace(de_parallel(model), imgs[0:1], strict=False), []) elif plots and ni == 10 and loggers['wandb']: wandb_logger.log({ 'Mosaics': [ loggers['wandb'].Image(str(x), caption=x.name) for x in save_dir.glob('train*.jpg') if x.exists() ] }) # end batch ------------------------------------------------------------------------------------------------ # Scheduler lr = [x['lr'] for x in optimizer.param_groups] # for loggers scheduler.step() # DDP process 0 or single-GPU if RANK in [-1, 0]: # mAP ema.update_attr(model, include=[ 'yaml', 'nc', 'hyp', 'gr', 'names', 'stride', 'class_weights' ]) final_epoch = epoch + 1 == epochs if not notest or final_epoch: # Calculate mAP wandb_logger.current_epoch = epoch + 1 results, maps, _ = test.run(data_dict, batch_size=batch_size // WORLD_SIZE * 2, imgsz=imgsz_test, model=ema.ema, single_cls=single_cls, dataloader=testloader, save_dir=save_dir, save_json=is_coco and final_epoch, verbose=nc < 50 and final_epoch, plots=plots and final_epoch, wandb_logger=wandb_logger, compute_loss=compute_loss) # Write with open(results_file, 'a') as f: f.write(s + '%10.4g' * 7 % results + '\n') # append metrics, val_loss # Log tags = [ 'train/box_loss', 'train/obj_loss', 'train/cls_loss', # train loss 'metrics/precision', 'metrics/recall', 'metrics/mAP_0.5', 'metrics/mAP_0.5:0.95', 'val/box_loss', 'val/obj_loss', 'val/cls_loss', # val loss 'x/lr0', 'x/lr1', 'x/lr2' ] # params for x, tag in zip(list(mloss[:-1]) + list(results) + lr, tags): if loggers['tb']: loggers['tb'].add_scalar(tag, x, epoch) # TensorBoard if loggers['wandb']: wandb_logger.log({tag: x}) # W&B # Update best mAP fi = fitness(np.array(results).reshape( 1, -1)) # weighted combination of [P, R, [email protected], [email protected]] if fi > best_fitness: best_fitness = fi wandb_logger.end_epoch(best_result=best_fitness == fi) # Save model if (not nosave) or (final_epoch and not evolve): # if save ckpt = { 'epoch': epoch, 'best_fitness': best_fitness, 'training_results': results_file.read_text(), 'model': deepcopy(de_parallel(model)).half(), 'ema': deepcopy(ema.ema).half(), 'updates': ema.updates, 'optimizer': optimizer.state_dict(), 'wandb_id': wandb_logger.wandb_run.id if loggers['wandb'] else None } # Save last, best and delete torch.save(ckpt, last) if best_fitness == fi: torch.save(ckpt, best) if loggers['wandb']: if ((epoch + 1) % opt.save_period == 0 and not final_epoch) and opt.save_period != -1: wandb_logger.log_model(last.parent, opt, epoch, fi, best_model=best_fitness == fi) del ckpt # end epoch ---------------------------------------------------------------------------------------------------- # end training ----------------------------------------------------------------------------------------------------- if RANK in [-1, 0]: logger.info( f'{epoch - start_epoch + 1} epochs completed in {(time.time() - t0) / 3600:.3f} hours.\n' ) if plots: plot_results(save_dir=save_dir) # save as results.png if loggers['wandb']: files = [ 'results.png', 'confusion_matrix.png', *[f'{x}_curve.png' for x in ('F1', 'PR', 'P', 'R')] ] wandb_logger.log({ "Results": [ loggers['wandb'].Image(str(save_dir / f), caption=f) for f in files if (save_dir / f).exists() ] }) if not evolve: if is_coco: # COCO dataset for m in [last, best ] if best.exists() else [last]: # speed, mAP tests results, _, _ = test.run( data_dict, batch_size=batch_size // WORLD_SIZE * 2, imgsz=imgsz_test, conf_thres=0.001, iou_thres=0.7, model=attempt_load(m, device).half(), single_cls=single_cls, dataloader=testloader, save_dir=save_dir, save_json=True, plots=False) # Strip optimizers for f in last, best: if f.exists(): strip_optimizer(f) # strip optimizers if loggers['wandb']: # Log the stripped model loggers['wandb'].log_artifact( str(best if best.exists() else last), type='model', name='run_' + wandb_logger.wandb_run.id + '_model', aliases=['latest', 'best', 'stripped']) wandb_logger.finish_run() torch.cuda.empty_cache() return results
def train(hyp, opt, device, tb_writer=None, wandb=None): logger.info(colorstr('hyperparameters: ') + ', '.join(f'{k}={v}' for k, v in hyp.items())) save_dir, epochs, batch_size, total_batch_size, weights, rank = \ Path(opt.save_dir), opt.epochs, opt.batch_size, opt.total_batch_size, opt.weights, opt.global_rank # Directories wdir = save_dir / 'weights' wdir.mkdir(parents=True, exist_ok=True) # make dir last = wdir / 'last.pt' best = wdir / 'best.pt' results_file = save_dir / 'results.txt' # Save run settings with open(save_dir / 'hyp.yaml', 'w') as f: yaml.dump(hyp, f, sort_keys=False) with open(save_dir / 'opt.yaml', 'w') as f: yaml.dump(vars(opt), f, sort_keys=False) # Configure plots = not opt.evolve # create plots cuda = device.type != 'cpu' init_seeds(2 + rank) with open(opt.data) as f: data_dict = yaml.load(f, Loader=yaml.SafeLoader) # data dict with torch_distributed_zero_first(rank): check_dataset(data_dict) # check train_path = data_dict['train'] test_path = data_dict['val'] nc1 = 1 if opt.single_cls else int(data_dict['nc1']) # number of classes names1 = ['item'] if opt.single_cls and len(data_dict['names1']) != 1 else data_dict['names1'] # class names assert len(names1) == nc1, '%g names found for nc=%g dataset in %s' % (len(names1), nc1, opt.data) # check nc2 = 1 if opt.single_cls else int(data_dict['nc2']) # number of classes # edit names2 = ['item'] if opt.single_cls and len(data_dict['names2']) != 1 else data_dict['names2'] # class names # edit assert len(names2) == nc2, '%g names found for nc=%g dataset in %s' % (len(names2), nc2, opt.data) # check # edit # Model pretrained = weights.endswith('.pt') if pretrained: with torch_distributed_zero_first(rank): attempt_download(weights) # download if not found locally ckpt = torch.load(weights, map_location=device) # load checkpoint if hyp.get('anchors'): ckpt['model'].yaml['anchors'] = round(hyp['anchors']) # force autoanchor model = Model(opt.cfg or ckpt['model'].yaml, ch=3, nc1=nc1, nc2=nc2).to(device) # create # edit exclude = ['anchor'] if opt.cfg or hyp.get('anchors') else [] # exclude keys state_dict = ckpt['model'].float().state_dict() # to FP32 state_dict = intersect_dicts(state_dict, model.state_dict(), exclude=exclude) # intersect model.load_state_dict(state_dict, strict=False) # load logger.info('Transferred %g/%g items from %s' % (len(state_dict), len(model.state_dict()), weights)) # report else: model = Model(opt.cfg, ch=3, nc1=nc1, nc2=nc2).to(device) # create # Freeze freeze = [] # parameter names to freeze (full or partial) for k, v in model.named_parameters(): v.requires_grad = True # train all layers if any(x in k for x in freeze): print('freezing %s' % k) v.requires_grad = False # Optimizer nbs = 64 # nominal batch size accumulate = max(round(nbs / total_batch_size), 1) # accumulate loss before optimizing hyp['weight_decay'] *= total_batch_size * accumulate / nbs # scale weight_decay logger.info(f"Scaled weight_decay = {hyp['weight_decay']}") pg0, pg1, pg2 = [], [], [] # optimizer parameter groups for k, v in model.named_modules(): if hasattr(v, 'bias') and isinstance(v.bias, nn.Parameter): pg2.append(v.bias) # biases if isinstance(v, nn.BatchNorm2d): pg0.append(v.weight) # no decay elif hasattr(v, 'weight') and isinstance(v.weight, nn.Parameter): pg1.append(v.weight) # apply decay if opt.adam: optimizer = optim.Adam(pg0, lr=hyp['lr0'], betas=(hyp['momentum'], 0.999)) # adjust beta1 to momentum else: optimizer = optim.SGD(pg0, lr=hyp['lr0'], momentum=hyp['momentum'], nesterov=True) optimizer.add_param_group({'params': pg1, 'weight_decay': hyp['weight_decay']}) # add pg1 with weight_decay optimizer.add_param_group({'params': pg2}) # add pg2 (biases) logger.info('Optimizer groups: %g .bias, %g conv.weight, %g other' % (len(pg2), len(pg1), len(pg0))) del pg0, pg1, pg2 # Scheduler https://arxiv.org/pdf/1812.01187.pdf # https://pytorch.org/docs/stable/_modules/torch/optim/lr_scheduler.html#OneCycleLR if opt.linear_lr: lf = lambda x: (1 - x / (epochs - 1)) * (1.0 - hyp['lrf']) + hyp['lrf'] # linear else: lf = one_cycle(1, hyp['lrf'], epochs) # cosine 1->hyp['lrf'] scheduler = lr_scheduler.LambdaLR(optimizer, lr_lambda=lf) # plot_lr_scheduler(optimizer, scheduler, epochs) # Logging if rank in [-1, 0] and wandb and wandb.run is None: opt.hyp = hyp # add hyperparameters wandb_run = wandb.init(config=opt, resume="allow", project=opt.wandb if opt.project == 'runs/train' else Path(opt.project).stem, name=save_dir.stem, id=ckpt.get('wandb_id') if 'ckpt' in locals() else None) loggers = {'wandb': wandb} # loggers dict # Resume start_epoch, best_fitness = 0, 0.0 if pretrained: # Optimizer if ckpt['optimizer'] is not None: optimizer.load_state_dict(ckpt['optimizer']) best_fitness = ckpt['best_fitness'] # Results if ckpt.get('training_results') is not None: with open(results_file, 'w') as file: file.write(ckpt['training_results']) # write results.txt # Epochs start_epoch = ckpt['epoch'] + 1 if opt.resume: assert start_epoch > 0, '%s training to %g epochs is finished, nothing to resume.' % (weights, epochs) if epochs < start_epoch: logger.info('%s has been trained for %g epochs. Fine-tuning for %g additional epochs.' % (weights, ckpt['epoch'], epochs)) epochs += ckpt['epoch'] # finetune additional epochs del ckpt, state_dict # Image sizes gs = int(model.stride.max()) # grid size (max stride) nl = model.model[-1].nl # number of detection layers (used for scaling hyp['obj']) imgsz, imgsz_test = [check_img_size(x, gs) for x in opt.img_size] # verify imgsz are gs-multiples # DP mode if cuda and rank == -1 and torch.cuda.device_count() > 1: model = torch.nn.DataParallel(model) # SyncBatchNorm if opt.sync_bn and cuda and rank != -1: model = torch.nn.SyncBatchNorm.convert_sync_batchnorm(model).to(device) logger.info('Using SyncBatchNorm()') # EMA ema = ModelEMA(model) if rank in [-1, 0] else None # DDP mode if cuda and rank != -1: model = DDP(model, device_ids=[opt.local_rank], output_device=opt.local_rank) # Trainloader (labels: cls1, cls2, xyxy) dataloader, dataset = create_dataloader(train_path, imgsz, batch_size, gs, opt, hyp=hyp, augment=True, cache=opt.cache_images, rect=opt.rect, rank=rank, world_size=opt.world_size, workers=opt.workers, image_weights=opt.image_weights, quad=opt.quad, prefix=colorstr('train: ')) nb = len(dataloader) # number of batches mlc1 = np.concatenate(dataset.labels, 0)[:, 0].max() # max label class 1 # edit assert mlc1 < nc1, 'Label class 1 %g exceeds nc1=%g in %s. Possible class 1 labels are 0-%g'\ % (mlc1, nc1, opt.data, nc1 - 1) # edit mlc2 = np.concatenate(dataset.labels, 0)[:, 1].max() # max label class 2 # edit assert mlc2 < nc2, 'Label class 2 %g exceeds nc2=%g in %s. Possible class 2 labels are 0-%g'\ % (mlc2, nc2, opt.data, nc2 - 1) # edit # Process 0 if rank in [-1, 0]: ema.updates = start_epoch * nb // accumulate # set EMA updates testloader = create_dataloader(test_path, imgsz_test, batch_size * 2, gs, opt, # testloader hyp=hyp, cache=opt.cache_images and not opt.notest, rect=True, rank=-1, world_size=opt.world_size, workers=opt.workers, pad=0.5, prefix=colorstr('val: '))[0] if not opt.resume: labels = np.concatenate(dataset.labels, 0) c1 = torch.tensor(labels[:, 0]) # classes 1 c2 = torch.tensor(labels[:, 1]) # classes 2 # edit # cf = torch.bincount(c.long(), minlength=nc) + 1. # frequency # model._initialize_biases(cf.to(device)) if plots: plot_labels(labels[:, 0], labels[:, 2:], save_dir, loggers, suffix="_1") plot_labels(labels[:, 1], labels[:, 2:], save_dir, loggers, suffix="_2") # edit if tb_writer: tb_writer.add_histogram('classes 1', c1, 0) tb_writer.add_histogram('classes 1', c2, 0) # edit # Anchors if not opt.noautoanchor: check_anchors(dataset, model=model, thr=hyp['anchor_t'], imgsz=imgsz) # Model parameters hyp['box'] *= 3. / nl # scale to layers hyp['cls1'] = hyp['cls'] * nc1 / 80. * 3. / nl # scale to classes and layers # edit hyp['cls2'] = hyp['cls'] * nc2 / 80. * 3. / nl # scale to classes and layers # edit hyp['obj'] *= (imgsz / 640) ** 2 * 3. / nl # scale to image size and layers model.nc1 = nc1 # attach number of classes to model # edit model.nc2 = nc2 # attach number of classes to model # edit model.hyp = hyp # attach hyperparameters to model model.gr = 1.0 # iou loss ratio (obj_loss = 1.0 or iou) model.class1_weights, model.class2_weights = labels_to_class_weights(dataset.labels, nc1, nc2) # attach class weights # edit model.class1_weights = model.class1_weights.to(device) * nc1 # edit model.class2_weights = model.class1_weights.to(device) * nc2 # edit model.names1 = names1 # edit model.names2 = names2 # edit # Start training t0 = time.time() nw = max(round(hyp['warmup_epochs'] * nb), 1000) # number of warmup iterations, max(3 epochs, 1k iterations) # nw = min(nw, (epochs - start_epoch) / 2 * nb) # limit warmup to < 1/2 of training maps1 = np.zeros(nc1) # mAP per class 1 # edit maps2 = np.zeros(nc2) # mAP per class 2 # edit results1 = (0, 0, 0, 0, 0, 0, 0, 0) # P_1, R_1, [email protected]_1, [email protected]_1, val_loss(box, obj, cls1, cls2) # edit results2 = (0, 0, 0, 0, 0, 0, 0, 0) # P_2, R_2, [email protected]_2, [email protected]_2, val_loss(box, obj, cls1, cls2) # edit scheduler.last_epoch = start_epoch - 1 # do not move scaler = amp.GradScaler(enabled=cuda) compute_loss = ComputeLoss(model) # init loss class logger.info(f'Image sizes {imgsz} train, {imgsz_test} test\n' f'Using {dataloader.num_workers} dataloader workers\n' f'Logging results to {save_dir}\n' f'Starting training for {epochs} epochs...') for epoch in range(start_epoch, epochs): # epoch ------------------------------------------------------------------ model.train() # Update image weights (optional) if opt.image_weights: # Generate indices if rank in [-1, 0]: cw1 = model.class1_weights.cpu().numpy() * (1 - maps1) ** 2 / nc1 # class weights # edit cw2 = model.class2_weights.cpu().numpy() * (1 - maps2) ** 2 / nc2 # class weights # edit iw1 = labels_to_image_weights(dataset.labels, nc=nc1, class_weights=cw1) # image weights # edit iw2 = labels_to_image_weights(dataset.labels, nc=nc2, class_weights=cw2) # image weights # edit dataset.indices1 = random.choices(range(dataset.n), weights=iw1, k=dataset.n) # rand weighted idx # edit dataset.indices2 = random.choices(range(dataset.n), weights=iw2, k=dataset.n) # rand weighted idx # edit # Broadcast if DDP if rank != -1: indices = (torch.tensor(dataset.indices) if rank == 0 else torch.zeros(dataset.n)).int() dist.broadcast(indices, 0) if rank != 0: dataset.indices = indices.cpu().numpy() # Update mosaic border # b = int(random.uniform(0.25 * imgsz, 0.75 * imgsz + gs) // gs * gs) # dataset.mosaic_border = [b - imgsz, -b] # height, width borders mloss = torch.zeros(5, device=device) # mean losses # edit if rank != -1: dataloader.sampler.set_epoch(epoch) pbar = enumerate(dataloader) logger.info(('\n' + '%10s' * 9) % ('Epoch', 'gpu_mem', 'box', 'obj', 'cls1', 'cls2', 'total', 'targets', 'img_size')) # edit if rank in [-1, 0]: pbar = tqdm(pbar, total=nb) # progress bar optimizer.zero_grad() for i, (imgs, targets, paths, _) in pbar: # batch ------------------------------------------------------------- # print(i, targets) # todo ni = i + nb * epoch # number integrated batches (since train start) imgs = imgs.to(device, non_blocking=True).float() / 255.0 # uint8 to float32, 0-255 to 0.0-1.0 # Warmup if ni <= nw: xi = [0, nw] # x interp # model.gr = np.interp(ni, xi, [0.0, 1.0]) # iou loss ratio (obj_loss = 1.0 or iou) accumulate = max(1, np.interp(ni, xi, [1, nbs / total_batch_size]).round()) for j, x in enumerate(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, [hyp['warmup_bias_lr'] if j == 2 else 0.0, x['initial_lr'] * lf(epoch)]) if 'momentum' in x: x['momentum'] = np.interp(ni, xi, [hyp['warmup_momentum'], hyp['momentum']]) # Multi-scale if opt.multi_scale: sz = random.randrange(imgsz * 0.5, imgsz * 1.5 + gs) // gs * gs # size sf = sz / max(imgs.shape[2:]) # scale factor if sf != 1: ns = [math.ceil(x * sf / gs) * gs for x in imgs.shape[2:]] # new shape (stretched to gs-multiple) imgs = F.interpolate(imgs, size=ns, mode='bilinear', align_corners=False) # Forward with amp.autocast(enabled=cuda): pred = model(imgs) # forward # loss: loss * batch_size; loss_items: lbox, lobj, lcls1, lcls2, loss loss, loss_items = compute_loss(pred, targets.to(device)) # loss scaled by batch_size if rank != -1: loss *= opt.world_size # gradient averaged between devices in DDP mode if opt.quad: loss *= 4. # Backward scaler.scale(loss).backward() # Optimize if ni % accumulate == 0: scaler.step(optimizer) # optimizer.step scaler.update() optimizer.zero_grad() if ema: ema.update(model) # Print if rank in [-1, 0]: mloss = (mloss * i + loss_items) / (i + 1) # update mean losses mem = '%.3gG' % (torch.cuda.memory_reserved() / 1E9 if torch.cuda.is_available() else 0) # (GB) s = ('%10s' * 2 + '%10.4g' * 7) %\ ('%g/%g' % (epoch, epochs - 1), mem, *mloss, targets.shape[0], imgs.shape[-1]) # edit pbar.set_description(s) # Plot if plots and ni < 3: f = save_dir / f'train_batch{ni}.jpg' # filename Thread(target=plot_images, args=(imgs, targets, paths, f), daemon=True).start() # if tb_writer: # tb_writer.add_image(f, result, dataformats='HWC', global_step=epoch) # tb_writer.add_graph(model, imgs) # add model to tensorboard elif plots and ni == 10 and wandb: wandb.log({"Mosaics": [wandb.Image(str(x), caption=x.name) for x in save_dir.glob('train*.jpg') if x.exists()]}, commit=False) # end batch ------------------------------------------------------------------------------------------------ # end epoch ---------------------------------------------------------------------------------------------------- # Scheduler lr = [x['lr'] for x in optimizer.param_groups] # for tensorboard scheduler.step() # DDP process 0 or single-GPU if rank in [-1, 0]: # mAP if ema: ema.update_attr(model, include=['yaml', 'nc1', 'nc2', 'hyp', 'gr', 'names1', 'names2', 'stride', 'class_weights']) # edit final_epoch = epoch + 1 == epochs if not opt.notest or final_epoch: # Calculate mAP maps1, maps2, times, results1, results2 = test.test(opt.data, batch_size=batch_size * 2, imgsz=imgsz_test, model=ema.ema, single_cls=opt.single_cls, dataloader=testloader, save_dir=save_dir, verbose=nc1 < 50 and nc2 < 50 and final_epoch, # edit plots=plots and final_epoch, log_imgs=opt.log_imgs if wandb else 0, compute_loss=compute_loss) # Write # print("before write results:", results1, results2) # todo with open(results_file, 'a') as f: # epoch/epochs, mem, train_loss(box, obj, cls1, cls2, loss), targets.shape[0], imgs.shape[-1] # edit # P_1, R_1, [email protected]_1, [email protected]_1, P_2, R_2, [email protected]_2, [email protected]_2, val_loss(box, obj, cls1, cls2) f.write(s + '%10.4g' * 4 % results1[:4] + '%10.4g' * 8 % results2 + '\n') # edit if len(opt.name) and opt.bucket: os.system('gsutil cp %s gs://%s/results/results%s.txt' % (results_file, opt.bucket, opt.name)) # Log tags = ['train/box_loss', 'train/obj_loss', 'train/cls_loss_1', 'train/cls_loss_2', # train loss # edit 'metrics/precision_1', 'metrics/recall_1', 'metrics/mAP_0.5_1', 'metrics/mAP_0.5:0.95_1', 'metrics/precision_2', 'metrics/recall_2', 'metrics/mAP_0.5_2', 'metrics/mAP_0.5:0.95_2', # edit 'val/box_loss', 'val/obj_loss', 'val/cls_loss_1', 'val/cls_loss_2', # val loss # edit 'x/lr0', 'x/lr1', 'x/lr2'] # params for x, tag in zip(list(mloss[:-1]) + list(results1)[:4] + list(results2) + lr, tags): # edit if tb_writer: tb_writer.add_scalar(tag, x, epoch) # tensorboard if wandb: wandb.log({tag: x}, step=epoch, commit=tag == tags[-1]) # W&B # Update best mAP # weighted combination of [P, R, [email protected], [email protected]] for class 1 and class 2 fi = fitness(np.array(results1).reshape(1, -1), np.array(results2).reshape(1, -1)) # edit if fi > best_fitness: best_fitness = fi # Save model save = (not opt.nosave) or (final_epoch and not opt.evolve) if save: with open(results_file, 'r') as f: # create checkpoint ckpt = {'epoch': epoch, 'best_fitness': best_fitness, 'training_results': f.read(), 'model': ema.ema, 'optimizer': None if final_epoch else optimizer.state_dict(), 'wandb_id': wandb_run.id if wandb else None} # Save last, best and delete torch.save(ckpt, last) if best_fitness == fi: torch.save(ckpt, best) del ckpt # end epoch ---------------------------------------------------------------------------------------------------- # end training if rank in [-1, 0]: # Strip optimizers final = best if best.exists() else last # final model for f in [last, best]: if f.exists(): strip_optimizer(f) # strip optimizers if opt.bucket: os.system(f'gsutil cp {final} gs://{opt.bucket}/weights') # upload # Plots if plots: plot_results(save_dir=save_dir) # save as results.png if wandb: files = ['results.png', 'confusion_matrix_1.png', 'confusion_matrix_2.png', *[f'{x}_curve_{y}.png' for x in ('F1', 'PR', 'P', 'R') for y in ('1', '2')]] # edit wandb.log({"Results": [wandb.Image(str(save_dir / f), caption=f) for f in files if (save_dir / f).exists()]}) if opt.log_artifacts: wandb.log_artifact(artifact_or_path=str(final), type='model', name=save_dir.stem) # Test best.pt logger.info('%g epochs completed in %.3f hours.\n' % (epoch - start_epoch + 1, (time.time() - t0) / 3600)) if opt.data.endswith('coco.yaml') and nc1 == 80: # if COCO for conf, iou, save_json in ([0.25, 0.45, False], [0.001, 0.65, True]): # speed, mAP tests maps1, maps2, times, results1, results2 = test.test(opt.data, batch_size=batch_size * 2, imgsz=imgsz_test, conf_thres=conf, iou_thres=iou, model=attempt_load(final, device).half(), single_cls=opt.single_cls, dataloader=testloader, save_dir=save_dir, save_json=save_json, plots=False) else: dist.destroy_process_group() wandb.run.finish() if wandb and wandb.run else None torch.cuda.empty_cache() return results1, results2
'mixup': (1, 0.0, 1.0)} # image mixup (probability) assert opt.local_rank == -1, 'DDP mode not implemented for --evolve' opt.notest, opt.nosave = True, True # only test/save final epoch # ei = [isinstance(x, (int, float)) for x in hyp.values()] # evolvable indices yaml_file = Path(opt.save_dir) / 'hyp_evolved.yaml' # save best result here if opt.bucket: os.system('gsutil cp gs://%s/evolve.txt .' % opt.bucket) # download evolve.txt if exists for _ in range(300): # generations to evolve if Path('evolve.txt').exists(): # if evolve.txt exists: select best hyps and mutate # Select parent(s) parent = 'single' # parent selection method: 'single' or 'weighted' x = np.loadtxt('evolve.txt', ndmin=2) n = min(5, len(x)) # number of previous results to consider x = x[np.argsort(-fitness(x))][:n] # top n mutations w = fitness(x) - fitness(x).min() # weights if parent == 'single' or len(x) == 1: # x = x[random.randint(0, n - 1)] # random selection x = x[random.choices(range(n), weights=w)[0]] # weighted selection elif parent == 'weighted': x = (x * w.reshape(n, 1)).sum(0) / w.sum() # weighted combination # Mutate mp, s = 0.8, 0.2 # mutation probability, sigma npr = np.random npr.seed(int(time.time())) g = np.array([x[0] for x in meta.values()]) # gains 0-1 ng = len(meta) v = np.ones(ng) while all(v == 1): # mutate until a change occurs (prevent duplicates)