【深度学习】PyTorch常用代码段合集
來源 | 極市平臺,機器學習算法與自然語言處理
本文是PyTorch常用代碼段合集,涵蓋基本配置、張量處理、模型定義與操作、數據處理、模型訓練與測試等5個方面,還給出了多個值得注意的Tips,內容非常全面。
PyTorch最好的資料是官方文檔。本文是PyTorch常用代碼段,在參考資料[1](張皓:PyTorch Cookbook)的基礎上做了一些修補,方便使用時查閱。
01
基本配置
導入包和版本查詢
import torch import torch.nn as nn import torchvision print(torch.__version__) print(torch.version.cuda) print(torch.backends.cudnn.version()) print(torch.cuda.get_device_name(0))可復現性
在硬件設備(CPU、GPU)不同時,完全的可復現性無法保證,即使隨機種子相同。但是,在同一個設備上,應該保證可復現性。具體做法是,在程序開始的時候固定torch的隨機種子,同時也把numpy的隨機種子固定。
np.random.seed(0) torch.manual_seed(0) torch.cuda.manual_seed_all(0)torch.backends.cudnn.deterministic = True torch.backends.cudnn.benchmark = False顯卡設置
如果只需要一張顯卡
# Device configuration device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')如果需要指定多張顯卡,比如0,1號顯卡。
import os os.environ['CUDA_VISIBLE_DEVICES'] = '0,1'也可以在命令行運行代碼時設置顯卡:
CUDA_VISIBLE_DEVICES=0,1 python train.py清除顯存
torch.cuda.empty_cache()也可以使用在命令行重置GPU的指令
nvidia-smi --gpu-reset -i [gpu_id]02
張量(Tensor)處理
張量的數據類型
PyTorch有9種CPU張量類型和9種GPU張量類型。
張量基本信息
tensor = torch.randn(3,4,5) print(tensor.type()) # 數據類型 print(tensor.size()) # 張量的shape,是個元組 print(tensor.dim()) # 維度的數量命名張量
張量命名是一個非常有用的方法,這樣可以方便地使用維度的名字來做索引或其他操作,大大提高了可讀性、易用性,防止出錯。
# 在PyTorch 1.3之前,需要使用注釋 # Tensor[N, C, H, W] images = torch.randn(32, 3, 56, 56) images.sum(dim=1) images.select(dim=1, index=0)# PyTorch 1.3之后 NCHW = [‘N’, ‘C’, ‘H’, ‘W’] images = torch.randn(32, 3, 56, 56, names=NCHW) images.sum('C') images.select('C', index=0) # 也可以這么設置 tensor = torch.rand(3,4,1,2,names=('C', 'N', 'H', 'W')) # 使用align_to可以對維度方便地排序 tensor = tensor.align_to('N', 'C', 'H', 'W')數據類型轉換
# 設置默認類型,pytorch中的FloatTensor遠遠快于DoubleTensor torch.set_default_tensor_type(torch.FloatTensor)# 類型轉換 tensor = tensor.cuda() tensor = tensor.cpu() tensor = tensor.float() tensor = tensor.long()torch.Tensor與np.ndarray轉換
除了CharTensor,其他所有CPU上的張量都支持轉換為numpy格式然后再轉換回來。
ndarray = tensor.cpu().numpy() tensor = torch.from_numpy(ndarray).float() tensor = torch.from_numpy(ndarray.copy()).float() # If ndarray has negative stride.Torch.tensor與PIL.Image轉換
# pytorch中的張量默認采用[N, C, H, W]的順序,并且數據范圍在[0,1],需要進行轉置和規范化 # torch.Tensor -> PIL.Image image = PIL.Image.fromarray(torch.clamp(tensor*255, min=0, max=255).byte().permute(1,2,0).cpu().numpy()) image = torchvision.transforms.functional.to_pil_image(tensor) # Equivalently way# PIL.Image -> torch.Tensor path = r'./figure.jpg' tensor = torch.from_numpy(np.asarray(PIL.Image.open(path))).permute(2,0,1).float() / 255 tensor = torchvision.transforms.functional.to_tensor(PIL.Image.open(path)) # Equivalently waynp.ndarray與PIL.Image的轉換
image = PIL.Image.fromarray(ndarray.astype(np.uint8))ndarray = np.asarray(PIL.Image.open(path))從只包含一個元素的張量中提取值
value = torch.rand(1).item()張量形變
# 在將卷積層輸入全連接層的情況下通常需要對張量做形變處理, # 相比torch.view,torch.reshape可以自動處理輸入張量不連續的情況。 tensor = torch.rand(2,3,4) shape = (6, 4) tensor = torch.reshape(tensor, shape)打亂順序
tensor = tensor[torch.randperm(tensor.size(0))] # 打亂第一個維度水平翻轉
# pytorch不支持tensor[::-1]這樣的負步長操作,水平翻轉可以通過張量索引實現 # 假設張量的維度為[N, D, H, W]. tensor = tensor[:,:,:,torch.arange(tensor.size(3) - 1, -1, -1).long()]復制張量
# Operation | New/Shared memory | Still in computation graph | tensor.clone() # | New | Yes | tensor.detach() # | Shared | No | tensor.detach.clone()() # | New | No |張量拼接
''' 注意torch.cat和torch.stack的區別在于torch.cat沿著給定的維度拼接, 而torch.stack會新增一維。例如當參數是3個10x5的張量,torch.cat的結果是30x5的張量, 而torch.stack的結果是3x10x5的張量。 ''' tensor = torch.cat(list_of_tensors, dim=0) tensor = torch.stack(list_of_tensors, dim=0)將整數標簽轉為one-hot編碼
# pytorch的標記默認從0開始 tensor = torch.tensor([0, 2, 1, 3]) N = tensor.size(0) num_classes = 4 one_hot = torch.zeros(N, num_classes).long() one_hot.scatter_(dim=1, index=torch.unsqueeze(tensor, dim=1), src=torch.ones(N, num_classes).long())得到非零元素
torch.nonzero(tensor) # index of non-zero elements torch.nonzero(tensor==0) # index of zero elements torch.nonzero(tensor).size(0) # number of non-zero elements torch.nonzero(tensor == 0).size(0) # number of zero elements判斷兩個張量相等
torch.allclose(tensor1, tensor2) # float tensor torch.equal(tensor1, tensor2) # int tensor張量擴展
# Expand tensor of shape 64*512 to shape 64*512*7*7. tensor = torch.rand(64,512) torch.reshape(tensor, (64, 512, 1, 1)).expand(64, 512, 7, 7)矩陣乘法
# Matrix multiplcation: (m*n) * (n*p) * -> (m*p). result = torch.mm(tensor1, tensor2)# Batch matrix multiplication: (b*m*n) * (b*n*p) -> (b*m*p) result = torch.bmm(tensor1, tensor2)# Element-wise multiplication. result = tensor1 * tensor2計算兩組數據之間的兩兩歐式距離
利用broadcast機制
dist = torch.sqrt(torch.sum((X1[:,None,:] - X2) ** 2, dim=2))03
模型定義和操作
一個簡單兩層卷積網絡的示例
# convolutional neural network (2 convolutional layers) class ConvNet(nn.Module):def __init__(self, num_classes=10):super(ConvNet, self).__init__()self.layer1 = nn.Sequential(nn.Conv2d(1, 16, kernel_size=5, stride=1, padding=2),nn.BatchNorm2d(16),nn.ReLU(),nn.MaxPool2d(kernel_size=2, stride=2))self.layer2 = nn.Sequential(nn.Conv2d(16, 32, kernel_size=5, stride=1, padding=2),nn.BatchNorm2d(32),nn.ReLU(),nn.MaxPool2d(kernel_size=2, stride=2))self.fc = nn.Linear(7*7*32, num_classes)def forward(self, x):out = self.layer1(x)out = self.layer2(out)out = out.reshape(out.size(0), -1)out = self.fc(out)return outmodel = ConvNet(num_classes).to(device)卷積層的計算和展示可以用這個網站輔助。
雙線性匯合(bilinear pooling)
X = torch.reshape(N, D, H * W) # Assume X has shape N*D*H*W X = torch.bmm(X, torch.transpose(X, 1, 2)) / (H * W) # Bilinear pooling assert X.size() == (N, D, D) X = torch.reshape(X, (N, D * D)) X = torch.sign(X) * torch.sqrt(torch.abs(X) + 1e-5) # Signed-sqrt normalization X = torch.nn.functional.normalize(X) # L2 normalization多卡同步 BN(Batch normalization)
當使用 torch.nn.DataParallel 將代碼運行在多張 GPU 卡上時,PyTorch 的 BN 層默認操作是各卡上數據獨立地計算均值和標準差,同步 BN 使用所有卡上的數據一起計算 BN 層的均值和標準差,緩解了當批量大小(batch size)比較小時對均值和標準差估計不準的情況,是在目標檢測等任務中一個有效的提升性能的技巧。
sync_bn = torch.nn.SyncBatchNorm(num_features, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)將已有網絡的所有BN層改為同步BN層
def convertBNtoSyncBN(module, process_group=None):'''Recursively replace all BN layers to SyncBN layer.Args:module[torch.nn.Module]. Network'''if isinstance(module, torch.nn.modules.batchnorm._BatchNorm):sync_bn = torch.nn.SyncBatchNorm(module.num_features, module.eps, module.momentum, module.affine, module.track_running_stats, process_group)sync_bn.running_mean = module.running_meansync_bn.running_var = module.running_varif module.affine:sync_bn.weight = module.weight.clone().detach()sync_bn.bias = module.bias.clone().detach()return sync_bnelse:for name, child_module in module.named_children():setattr(module, name) = convert_syncbn_model(child_module, process_group=process_group))return module類似 BN 滑動平均
如果要實現類似 BN 滑動平均的操作,在 forward 函數中要使用原地(inplace)操作給滑動平均賦值。
class BN(torch.nn.Module)def __init__(self):...self.register_buffer('running_mean', torch.zeros(num_features))def forward(self, X):...self.running_mean += momentum * (current - self.running_mean)計算模型整體參數量
num_parameters = sum(torch.numel(parameter) for parameter in model.parameters())查看網絡中的參數
可以通過model.state_dict()或者model.named_parameters()函數查看現在的全部可訓練參數(包括通過繼承得到的父類中的參數)
params = list(model.named_parameters()) (name, param) = params[28] print(name) print(param.grad) print('-------------------------------------------------') (name2, param2) = params[29] print(name2) print(param2.grad) print('----------------------------------------------------') (name1, param1) = params[30] print(name1) print(param1.grad)模型可視化(使用pytorchviz)
szagoruyko/pytorchvizgithub.com
類似 Keras 的 model.summary() 輸出模型信息,使用pytorch-summary
sksq96/pytorch-summarygithub.com
模型權重初始化
注意 model.modules() 和 model.children() 的區別:model.modules() 會迭代地遍歷模型的所有子層,而 model.children() 只會遍歷模型下的一層。
# Common practise for initialization. for layer in model.modules():if isinstance(layer, torch.nn.Conv2d):torch.nn.init.kaiming_normal_(layer.weight, mode='fan_out',nonlinearity='relu')if layer.bias is not None:torch.nn.init.constant_(layer.bias, val=0.0)elif isinstance(layer, torch.nn.BatchNorm2d):torch.nn.init.constant_(layer.weight, val=1.0)torch.nn.init.constant_(layer.bias, val=0.0)elif isinstance(layer, torch.nn.Linear):torch.nn.init.xavier_normal_(layer.weight)if layer.bias is not None:torch.nn.init.constant_(layer.bias, val=0.0)# Initialization with given tensor. layer.weight = torch.nn.Parameter(tensor)提取模型中的某一層
modules()會返回模型中所有模塊的迭代器,它能夠訪問到最內層,比如self.layer1.conv1這個模塊,還有一個與它們相對應的是name_children()屬性以及named_modules(),這兩個不僅會返回模塊的迭代器,還會返回網絡層的名字。
# 取模型中的前兩層 new_model = nn.Sequential(*list(model.children())[:2] # 如果希望提取出模型中的所有卷積層,可以像下面這樣操作: for layer in model.named_modules():if isinstance(layer[1],nn.Conv2d):conv_model.add_module(layer[0],layer[1])部分層使用預訓練模型
注意如果保存的模型是 torch.nn.DataParallel,則當前的模型也需要是
model.load_state_dict(torch.load('model.pth'), strict=False)將在 GPU 保存的模型加載到 CPU
model.load_state_dict(torch.load('model.pth', map_location='cpu'))導入另一個模型的相同部分到新的模型
模型導入參數時,如果兩個模型結構不一致,則直接導入參數會報錯。用下面方法可以把另一個模型的相同的部分導入到新的模型中。
# model_new代表新的模型 # model_saved代表其他模型,比如用torch.load導入的已保存的模型 model_new_dict = model_new.state_dict() model_common_dict = {k:v for k, v in model_saved.items() if k in model_new_dict.keys()} model_new_dict.update(model_common_dict) model_new.load_state_dict(model_new_dict)04
數據處理
計算數據集的均值和標準差
import os import cv2 import numpy as np from torch.utils.data import Dataset from PIL import Imagedef compute_mean_and_std(dataset):# 輸入PyTorch的dataset,輸出均值和標準差mean_r = 0mean_g = 0mean_b = 0for img, _ in dataset:img = np.asarray(img) # change PIL Image to numpy arraymean_b += np.mean(img[:, :, 0])mean_g += np.mean(img[:, :, 1])mean_r += np.mean(img[:, :, 2])mean_b /= len(dataset)mean_g /= len(dataset)mean_r /= len(dataset)diff_r = 0diff_g = 0diff_b = 0N = 0for img, _ in dataset:img = np.asarray(img)diff_b += np.sum(np.power(img[:, :, 0] - mean_b, 2))diff_g += np.sum(np.power(img[:, :, 1] - mean_g, 2))diff_r += np.sum(np.power(img[:, :, 2] - mean_r, 2))N += np.prod(img[:, :, 0].shape)std_b = np.sqrt(diff_b / N)std_g = np.sqrt(diff_g / N)std_r = np.sqrt(diff_r / N)mean = (mean_b.item() / 255.0, mean_g.item() / 255.0, mean_r.item() / 255.0)std = (std_b.item() / 255.0, std_g.item() / 255.0, std_r.item() / 255.0)return mean, std得到視頻數據基本信息
import cv2 video = cv2.VideoCapture(mp4_path) height = int(video.get(cv2.CAP_PROP_FRAME_HEIGHT)) width = int(video.get(cv2.CAP_PROP_FRAME_WIDTH)) num_frames = int(video.get(cv2.CAP_PROP_FRAME_COUNT)) fps = int(video.get(cv2.CAP_PROP_FPS)) video.release()TSN 每段(segment)采樣一幀視頻
K = self._num_segments if is_train:if num_frames > K:# Random index for each segment.frame_indices = torch.randint(high=num_frames // K, size=(K,), dtype=torch.long)frame_indices += num_frames // K * torch.arange(K)else:frame_indices = torch.randint(high=num_frames, size=(K - num_frames,), dtype=torch.long)frame_indices = torch.sort(torch.cat((torch.arange(num_frames), frame_indices)))[0] else:if num_frames > K:# Middle index for each segment.frame_indices = num_frames / K // 2frame_indices += num_frames // K * torch.arange(K)else:frame_indices = torch.sort(torch.cat(( torch.arange(num_frames), torch.arange(K - num_frames))))[0] assert frame_indices.size() == (K,) return [frame_indices[i] for i in range(K)]常用訓練和驗證數據預處理
其中 ToTensor 操作會將 PIL.Image 或形狀為 H×W×D,數值范圍為 [0, 255] 的 np.ndarray 轉換為形狀為 D×H×W,數值范圍為 [0.0, 1.0] 的 torch.Tensor。
train_transform = torchvision.transforms.Compose([torchvision.transforms.RandomResizedCrop(size=224,scale=(0.08, 1.0)),torchvision.transforms.RandomHorizontalFlip(),torchvision.transforms.ToTensor(),torchvision.transforms.Normalize(mean=(0.485, 0.456, 0.406),std=(0.229, 0.224, 0.225)),])val_transform = torchvision.transforms.Compose([torchvision.transforms.Resize(256),torchvision.transforms.CenterCrop(224),torchvision.transforms.ToTensor(),torchvision.transforms.Normalize(mean=(0.485, 0.456, 0.406),std=(0.229, 0.224, 0.225)), ])05
模型訓練和調試
分類模型訓練代碼
# Loss and optimizer criterion = nn.CrossEntropyLoss() optimizer = torch.optim.Adam(model.parameters(), lr=learning_rate)# Train the model total_step = len(train_loader) for epoch in range(num_epochs):for i ,(images, labels) in enumerate(train_loader):images = images.to(device)labels = labels.to(device)# Forward passoutputs = model(images)loss = criterion(outputs, labels)# Backward and optimizeroptimizer.zero_grad()loss.backward()optimizer.step()if (i+1) % 100 == 0:print('Epoch: [{}/{}], Step: [{}/{}], Loss: {}'.format(epoch+1, num_epochs, i+1, total_step, loss.item()))分類模型測試代碼
# Test the model model.eval() # eval mode(batch norm uses moving mean/variance #instead of mini-batch mean/variance) with torch.no_grad():correct = 0total = 0for images, labels in test_loader:images = images.to(device)labels = labels.to(device)outputs = model(images)_, predicted = torch.max(outputs.data, 1)total += labels.size(0)correct += (predicted == labels).sum().item()print('Test accuracy of the model on the 10000 test images: {} %'.format(100 * correct / total))自定義loss
繼承torch.nn.Module類寫自己的loss。
class MyLoss(torch.nn.Moudle):def __init__(self):super(MyLoss, self).__init__()def forward(self, x, y):loss = torch.mean((x - y) ** 2)return loss標簽平滑(label smoothing)
寫一個label_smoothing.py的文件,然后在訓練代碼里引用,用LSR代替交叉熵損失即可。label_smoothing.py內容如下:
import torch import torch.nn as nnclass LSR(nn.Module):def __init__(self, e=0.1, reduction='mean'):super().__init__()self.log_softmax = nn.LogSoftmax(dim=1)self.e = eself.reduction = reductiondef _one_hot(self, labels, classes, value=1):"""Convert labels to one hot vectorsArgs:labels: torch tensor in format [label1, label2, label3, ...]classes: int, number of classesvalue: label value in one hot vector, default to 1Returns:return one hot format labels in shape [batchsize, classes]"""one_hot = torch.zeros(labels.size(0), classes)#labels and value_added size must matchlabels = labels.view(labels.size(0), -1)value_added = torch.Tensor(labels.size(0), 1).fill_(value)value_added = value_added.to(labels.device)one_hot = one_hot.to(labels.device)one_hot.scatter_add_(1, labels, value_added)return one_hotdef _smooth_label(self, target, length, smooth_factor):"""convert targets to one-hot format, and smooththem.Args:target: target in form with [label1, label2, label_batchsize]length: length of one-hot format(number of classes)smooth_factor: smooth factor for label smoothReturns:smoothed labels in one hot format"""one_hot = self._one_hot(target, length, value=1 - smooth_factor)one_hot += smooth_factor / (length - 1)return one_hot.to(target.device)def forward(self, x, target):if x.size(0) != target.size(0):raise ValueError('Expected input batchsize ({}) to match target batch_size({})'.format(x.size(0), target.size(0)))if x.dim() < 2:raise ValueError('Expected input tensor to have least 2 dimensions(got {})'.format(x.size(0)))if x.dim() != 2:raise ValueError('Only 2 dimension tensor are implemented, (got {})'.format(x.size()))smoothed_target = self._smooth_label(target, x.size(1), self.e)x = self.log_softmax(x)loss = torch.sum(- x * smoothed_target, dim=1)if self.reduction == 'none':return losselif self.reduction == 'sum':return torch.sum(loss)elif self.reduction == 'mean':return torch.mean(loss)else:raise ValueError('unrecognized option, expect reduction to be one of none, mean, sum')或者直接在訓練文件里做label smoothing
for images, labels in train_loader:images, labels = images.cuda(), labels.cuda()N = labels.size(0)# C is the number of classes.smoothed_labels = torch.full(size=(N, C), fill_value=0.1 / (C - 1)).cuda()smoothed_labels.scatter_(dim=1, index=torch.unsqueeze(labels, dim=1), value=0.9)score = model(images)log_prob = torch.nn.functional.log_softmax(score, dim=1)loss = -torch.sum(log_prob * smoothed_labels) / Noptimizer.zero_grad()loss.backward()optimizer.step()Mixup訓練
beta_distribution = torch.distributions.beta.Beta(alpha, alpha) for images, labels in train_loader:images, labels = images.cuda(), labels.cuda()# Mixup images and labels.lambda_ = beta_distribution.sample([]).item()index = torch.randperm(images.size(0)).cuda()mixed_images = lambda_ * images + (1 - lambda_) * images[index, :]label_a, label_b = labels, labels[index]# Mixup loss.scores = model(mixed_images)loss = (lambda_ * loss_function(scores, label_a)+ (1 - lambda_) * loss_function(scores, label_b))optimizer.zero_grad()loss.backward()optimizer.step()L1 正則化
l1_regularization = torch.nn.L1Loss(reduction='sum') loss = ... # Standard cross-entropy loss for param in model.parameters():loss += torch.sum(torch.abs(param)) loss.backward()不對偏置項進行權重衰減(weight decay)
pytorch里的weight decay相當于l2正則
bias_list = (param for name, param in model.named_parameters() if name[-4:] == 'bias') others_list = (param for name, param in model.named_parameters() if name[-4:] != 'bias') parameters = [{'parameters': bias_list, 'weight_decay': 0}, {'parameters': others_list}] optimizer = torch.optim.SGD(parameters, lr=1e-2, momentum=0.9, weight_decay=1e-4)梯度裁剪(gradient clipping)
torch.nn.utils.clip_grad_norm_(model.parameters(), max_norm=20)得到當前學習率
# If there is one global learning rate (which is the common case). lr = next(iter(optimizer.param_groups))['lr']# If there are multiple learning rates for different layers. all_lr = [] for param_group in optimizer.param_groups:all_lr.append(param_group['lr'])另一種方法,在一個batch訓練代碼里,當前的lr是optimizer.param_groups[0]['lr']
學習率衰減
# Reduce learning rate when validation accuarcy plateau. scheduler = torch.optim.lr_scheduler.ReduceLROnPlateau(optimizer, mode='max', patience=5, verbose=True) for t in range(0, 80):train(...)val(...)scheduler.step(val_acc)# Cosine annealing learning rate. scheduler = torch.optim.lr_scheduler.CosineAnnealingLR(optimizer, T_max=80) # Reduce learning rate by 10 at given epochs. scheduler = torch.optim.lr_scheduler.MultiStepLR(optimizer, milestones=[50, 70], gamma=0.1) for t in range(0, 80):scheduler.step() train(...)val(...)# Learning rate warmup by 10 epochs. scheduler = torch.optim.lr_scheduler.LambdaLR(optimizer, lr_lambda=lambda t: t / 10) for t in range(0, 10):scheduler.step()train(...)val(...)優化器鏈式更新
從1.4版本開始,torch.optim.lr_scheduler 支持鏈式更新(chaining),即用戶可以定義兩個 schedulers,并交替在訓練中使用。
import torch from torch.optim import SGD from torch.optim.lr_scheduler import ExponentialLR, StepLR model = [torch.nn.Parameter(torch.randn(2, 2, requires_grad=True))] optimizer = SGD(model, 0.1) scheduler1 = ExponentialLR(optimizer, gamma=0.9) scheduler2 = StepLR(optimizer, step_size=3, gamma=0.1) for epoch in range(4):print(epoch, scheduler2.get_last_lr()[0])optimizer.step()scheduler1.step()scheduler2.step()模型訓練可視化
PyTorch可以使用tensorboard來可視化訓練過程。
安裝和運行TensorBoard。
pip install tensorboard tensorboard --logdir=runs使用SummaryWriter類來收集和可視化相應的數據,放了方便查看,可以使用不同的文件夾,比如'Loss/train'和'Loss/test'。
from torch.utils.tensorboard import SummaryWriter import numpy as npwriter = SummaryWriter()for n_iter in range(100):writer.add_scalar('Loss/train', np.random.random(), n_iter)writer.add_scalar('Loss/test', np.random.random(), n_iter)writer.add_scalar('Accuracy/train', np.random.random(), n_iter)writer.add_scalar('Accuracy/test', np.random.random(), n_iter)保存與加載斷點
注意為了能夠恢復訓練,我們需要同時保存模型和優化器的狀態,以及當前的訓練輪數。
start_epoch = 0 # Load checkpoint. if resume: # resume為參數,第一次訓練時設為0,中斷再訓練時設為1model_path = os.path.join('model', 'best_checkpoint.pth.tar')assert os.path.isfile(model_path)checkpoint = torch.load(model_path)best_acc = checkpoint['best_acc']start_epoch = checkpoint['epoch']model.load_state_dict(checkpoint['model'])optimizer.load_state_dict(checkpoint['optimizer'])print('Load checkpoint at epoch {}.'.format(start_epoch))print('Best accuracy so far {}.'.format(best_acc))# Train the model for epoch in range(start_epoch, num_epochs): ... # Test the model...# save checkpointis_best = current_acc > best_accbest_acc = max(current_acc, best_acc)checkpoint = {'best_acc': best_acc,'epoch': epoch + 1,'model': model.state_dict(),'optimizer': optimizer.state_dict(),}model_path = os.path.join('model', 'checkpoint.pth.tar')best_model_path = os.path.join('model', 'best_checkpoint.pth.tar')torch.save(checkpoint, model_path)if is_best:shutil.copy(model_path, best_model_path)提取 ImageNet 預訓練模型某層的卷積特征
# VGG-16 relu5-3 feature. model = torchvision.models.vgg16(pretrained=True).features[:-1] # VGG-16 pool5 feature. model = torchvision.models.vgg16(pretrained=True).features # VGG-16 fc7 feature. model = torchvision.models.vgg16(pretrained=True) model.classifier = torch.nn.Sequential(*list(model.classifier.children())[:-3]) # ResNet GAP feature. model = torchvision.models.resnet18(pretrained=True) model = torch.nn.Sequential(collections.OrderedDict(list(model.named_children())[:-1]))with torch.no_grad():model.eval()conv_representation = model(image)提取 ImageNet 預訓練模型多層的卷積特征
class FeatureExtractor(torch.nn.Module):"""Helper class to extract several convolution features from the givenpre-trained model.Attributes:_model, torch.nn.Module._layers_to_extract, list<str> or set<str>Example:>>> model = torchvision.models.resnet152(pretrained=True)>>> model = torch.nn.Sequential(collections.OrderedDict(list(model.named_children())[:-1]))>>> conv_representation = FeatureExtractor(pretrained_model=model,layers_to_extract={'layer1', 'layer2', 'layer3', 'layer4'})(image)"""def __init__(self, pretrained_model, layers_to_extract):torch.nn.Module.__init__(self)self._model = pretrained_modelself._model.eval()self._layers_to_extract = set(layers_to_extract)def forward(self, x):with torch.no_grad():conv_representation = []for name, layer in self._model.named_children():x = layer(x)if name in self._layers_to_extract:conv_representation.append(x)return conv_representation微調全連接層
model = torchvision.models.resnet18(pretrained=True) for param in model.parameters():param.requires_grad = False model.fc = nn.Linear(512, 100) # Replace the last fc layer optimizer = torch.optim.SGD(model.fc.parameters(), lr=1e-2, momentum=0.9, weight_decay=1e-4)以較大學習率微調全連接層,較小學習率微調卷積層
model = torchvision.models.resnet18(pretrained=True) finetuned_parameters = list(map(id, model.fc.parameters())) conv_parameters = (p for p in model.parameters() if id(p) not in finetuned_parameters) parameters = [{'params': conv_parameters, 'lr': 1e-3}, {'params': model.fc.parameters()}] optimizer = torch.optim.SGD(parameters, lr=1e-2, momentum=0.9, weight_decay=1e-4)06
其他注意事項
不要使用太大的線性層。因為nn.Linear(m,n)使用的是的內存,線性層太大很容易超出現有顯存。
不要在太長的序列上使用RNN。因為RNN反向傳播使用的是BPTT算法,其需要的內存和輸入序列的長度呈線性關系。
model(x) 前用 model.train() 和 model.eval() 切換網絡狀態。
不需要計算梯度的代碼塊用 with torch.no_grad() 包含起來。
model.eval() 和 torch.no_grad() 的區別在于,model.eval() 是將網絡切換為測試狀態,例如 BN 和dropout在訓練和測試階段使用不同的計算方法。torch.no_grad() 是關閉 PyTorch 張量的自動求導機制,以減少存儲使用和加速計算,得到的結果無法進行 loss.backward()。
model.zero_grad()會把整個模型的參數的梯度都歸零, 而optimizer.zero_grad()只會把傳入其中的參數的梯度歸零.
torch.nn.CrossEntropyLoss 的輸入不需要經過 Softmax。torch.nn.CrossEntropyLoss 等價于 torch.nn.functional.log_softmax + torch.nn.NLLLoss。
loss.backward() 前用 optimizer.zero_grad() 清除累積梯度。
torch.utils.data.DataLoader 中盡量設置 pin_memory=True,對特別小的數據集如 MNIST 設置 pin_memory=False 反而更快一些。num_workers 的設置需要在實驗中找到最快的取值。
用 del 及時刪除不用的中間變量,節約 GPU 存儲。
使用 inplace 操作可節約 GPU 存儲,如
x = torch.nn.functional.relu(x, inplace=True)減少 CPU 和 GPU 之間的數據傳輸。例如如果你想知道一個 epoch 中每個 mini-batch 的 loss 和準確率,先將它們累積在 GPU 中等一個 epoch 結束之后一起傳輸回 CPU 會比每個 mini-batch 都進行一次 GPU 到 CPU 的傳輸更快。
使用半精度浮點數 half() 會有一定的速度提升,具體效率依賴于 GPU 型號。需要小心數值精度過低帶來的穩定性問題。
時常使用 assert tensor.size() == (N, D, H, W) 作為調試手段,確保張量維度和你設想中一致。
除了標記 y 外,盡量少使用一維張量,使用 n*1 的二維張量代替,可以避免一些意想不到的一維張量計算結果。
統計代碼各部分耗時
with torch.autograd.profiler.profile(enabled=True, use_cuda=False) as profile: ...print(profile)# 或者在命令行運行python -m torch.utils.bottleneck main.py使用TorchSnooper來調試PyTorch代碼,程序在執行的時候,就會自動 print 出來每一行的執行結果的 tensor 的形狀、數據類型、設備、是否需要梯度的信息。
# pip install torchsnooperimport torchsnooper# 對于函數,使用修飾器@torchsnooper.snoop()# 如果不是函數,使用 with 語句來激活 TorchSnooper,把訓練的那個循環裝進 with 語句中去。with torchsnooper.snoop(): 原本的代碼https://github.com/zasdfgbnm/TorchSnoopergithub.com
模型可解釋性,使用captum庫:https://captum.ai/captum.ai
參考資料
張皓:PyTorch Cookbook(常用代碼段整理合集),https://zhuanlan.zhihu.com/p/59205847?
PyTorch官方文檔和示例
https://pytorch.org/docs/stable/notes/faq.html
https://github.com/szagoruyko/pytorchviz
https://github.com/sksq96/pytorch-summary
其他
本站qq群955171419,加入微信群請掃碼:
總結
以上是生活随笔為你收集整理的【深度学习】PyTorch常用代码段合集的全部內容,希望文章能夠幫你解決所遇到的問題。
- 上一篇: mybatis jdbcType 对应j
- 下一篇: C3 粒子闪烁效果