PyTorch都用代碼段合集

發布時間：2023-04-13 05:47:13

點擊上圖，核對考試細節

僅作學術研究社交，不都是本公眾號主張，侵權聯絡刪掉

轉載于：比如說 | 極市平臺

PyTorch最好的詳細資料是官方HTML。本文是PyTorch特指編譯支架段，在簡要[1](張皓：PyTorch Cookbook)的基礎上想到了一些修補，簡便用于時翻查。

『整體配置』

新增包和發行版查詢

import torch import torch.nn as nn import torchvision print(torch.曲在version曲在) print(torch.version.cuda) print(torch.backends.cudnn.version) print(torch.cuda.get_device_name(0))

可復現性

在硬件電源（CPU、GPU）不盡不盡相同時，完正因如此的可復現性能夠意味著，即使隨機葉子不盡相同。但是，在同一個電源上，應該意味著可復現性。具體想到法是，在處置程序開始的時候相同torch的隨機葉子，同時也把numpy的隨機葉子相同。

np.random.seed(0) torch.manual_seed(0) torch.cuda.manual_seed_all(0)

torch.backends.cudnn.deterministic = Truetorch.backends.cudnn.benchmark = False

DirectX另設

如果只只能一張DirectX

# Device configurationdevice = torch.device( 'cuda'iftorch.cuda.is_available else'cpu')

如果只能所選多張DirectX，比如0，1號DirectX。

import osos.environ['CUDA_VISIBLE_DEVICES'] = '0,1'

也可以在GUI開始運行編譯支架時另設DirectX：

CUDA_VISIBLE_DEVICES=0,1 python train.py

掃除顯存

torch.cuda.empty_cache

也可以用于在GUI重置GPU的解釋器

nvidia-smi ；還有gpu-reset -i [gpu_id]

『Tensor處置』

標量的統計數據特性

PyTorch有9種CPU標量特性和9種GPU標量特性。

標量整體反饋

tensor = torch.randn(3,4,5)print(tensor.type) # 統計數據特性print(tensor.size) # 標量的shape，是個個位print(tensor.dim) # 等價的為數

定名標量

標量定名是一個更加精確的步驟，這樣可以簡便地用于等價的拼法來想到索引或其他操控，得益于了通用性、易用性，不必要出錯。

# 在PyTorch 1.3此前，只能用于注解# 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)# 也可以這么另設tensor = torch.rand(3,4,1,2,names=('C', 'N', 'H', 'W'))# 用于align_to可以對等價簡便地排序tensor = tensor.align_to('N', 'C', 'H', 'W')

統計數據特性疊加

# 另設匹配特性，pytorch中會的FloatTensor遠遠很慢DoubleTensortorch.set_default_tensor_type(torch.FloatTensor)

# 特性疊加tensor = tensor.cudatensor = tensor.cputensor = tensor.floattensor = tensor.long

torch.Tensor與np.ndarray疊加

除了CharTensor，其他所有CPU上的標量都正因如此力支持疊加為numpy格式然后再疊加去找。

ndarray = tensor.cpu.numpytensor = torch.from_numpy(ndarray).floattensor = torch.from_numpy(ndarray.copy).float # If ndarray has negative stride.

Torch.tensor與PIL.Image疊加

# pytorch中會的標量匹配采用[N, C, H, W]的依序，并且統計數據適用范圍在[0,1]，只能透過轉置和制度本土化# torch.Tensor -> PIL.Imageimage = 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.Tensorpath = r'./figure.jpg'tensor = torch.from_numpy(np.asarray(PIL.Image.open(path))).permute(2,0,1).float / 255tensor = torchvision.transforms.functional.to_tensor(PIL.Image.open(path)) # Equivalently way

np.ndarray與PIL.Image的疊加

image = PIL.Image.fromarray(ndarray.astype(np.uint8))

ndarray = np.asarray(PIL.Image.open(path))

從只包含一個要素的標量中會提煉出值

value = torch.rand(1).item

標量拉伸

# 在將變換層疊加成正因如此連接層的情況下有時候只能對標量想到拉伸處置，# 相比torch.view，torch.reshape可以終端處置疊加成標量不連續的情況。tensor = torch.rand(2,3,4)shape = (6, 4)tensor = torch.reshape(tensor, shape)

打亂依序

tensor = tensor[torch.randperm(tensor.size(0))] # 打亂第一個等價

總體反轉

# pytorch不正因如此力支持tensor[::-1]這樣的負步長操控，總體反轉可以通過標量索引解決問題# 假設標量的等價為[N, D, H, W].tensor = tensor[:,:,:,torch.arange(tensor.size(3) - 1, -1, -1).long]

粘貼標量

# Operation | New/Shared memory | Still in computation graph |tensor.clone # | New | Yes |tensor.detach # | Shared | No |tensor.detach.clone # | New | No |

標量拼接

'''請注意torch.cat和torch.stack的區別在于torch.cat沿著給定的等價拼接，而torch.stack才會新增給定。例如當值是3個10x5的標量，torch.cat的結果是30x5的標量，而torch.stack的結果是3x10x5的標量。'''tensor = torch.cat(list_of_tensors, dim=0)tensor = torch.stack(list_of_tensors, dim=0)

將整數附加轉為one-hot編碼

# pytorch的標示匹配從0開始tensor = torch.tensor([0, 2, 1, 3])N = tensor.size(0)num_classes = 4one_hot = torch.zeros(N, num_classes).longone_hot.scatter_(dim=1, index=torch.unsqueeze(tensor, dim=1), src=torch.ones(N, num_classes).long)

得不到非零要素

torch.nonzero(tensor) # index of non-zero elementstorch.nonzero(tensor==0) # index of zero elementstorch.nonzero(tensor).size(0) # number of non-zero elementstorch.nonzero(tensor == 0).size(0) # number of zero elements

判斷兩個標量相等

torch.allclose(tensor1, tensor2) # float tensortorch.equal(tensor1, tensor2) # int tensor

標量拓展

# 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)

行列式自然數

# 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

近似值三組統計數據錯綜復雜的兩兩歐式距離

來透過broadcast前提

dist = torch.sqrt(torch.sum((X1[:,None,:] - X2) ** 2, dim=2))

『框架定義和操控』

一個精確兩層變換局域網的范例

# convolutional neural network (2 convolutional layers)class ConvNet(nn.Module):def 曲在init曲在(self, num_classes=10):super(ConvNet, self).曲在init曲在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 out

model = ConvNet(num_classes).to(device)

雙差分匯合（bilinear pooling）

X = torch.reshape(N, D, H * W) # Assume X has shape N*D*H*WX = torch.bmm(X, torch.transpose(X, 1, 2)) / (H * W) # Bilinear poolingassert 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 normalizationX = torch.nn.functional.normalize(X) # L2 normalization

多卡不間斷 BN（Batch normalization）

當用于 torch.nn.DataParallel 將編譯支架開始運行在多張 GPU 卡上時，PyTorch 的 BN 層匹配操控是各卡上統計數據獨立地近似值平方根和加權，不間斷 BN 用于所有卡上的統計數據兩兄弟近似值 BN 層的平方根和加權，緩解了當批量尺寸（batch size）比較同一時間對平方根和加權推估不準的情況，是在目的探測等任務中會一個有效的提升性能的高難度。

sync_bn = torch.nn.SyncBatchNorm(num_features, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)

將較早局域網的所有BN層改名不間斷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.detachsync_bn.bias = module.bias.clone.detachreturn sync_bnelse:for name, child_module in module.named_children:setattr(module, name) = convert_syncbn_model(child_module, process_group=process_group))return module

多種不同 BN 滑動平均

如果要解決問題多種不同 BN 滑動平均的操控，在 forward 表達式中會要用于原地（inplace）操控給滑動平均賦值。

class BN(torch.nn.Module)def 曲在init曲在(self):...self.register_buffer('running_mean', torch.zeros(num_features))

def forward(self, X):...self.running_mean += momentum * (current - self.running_mean)

近似值框架整體值量

num_parameters = sum(torch.numel(parameter) for parameter in model.parameters)

核對局域網中會的值

可以通過model.state_dict或者model.named_parameters表達式核對那時候的正因如此部可體能訓練值（除此以外通過傳給得不到的父類中會的值）

params = list(model.named_parameters)(name, param) = params[28]print(name)print(param.grad)print('；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有-')(name2, param2) = params[29]print(name2)print(param2.grad)print('；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有')(name1, param1) = params[30]print(name1)print(param1.grad)

框架統計數據處置（用于pytorchviz）

szagoruyko/pytorchvizgithub.com

多種不同 Keras 的 model.summary 反向框架反饋，用于pytorch-summary

sksq96/pytorch-summarygithub.com

框架二階初始本土化

請注意 model.modules 和 model.children 的區別：model.modules 才會插值地二叉樹框架的所有子層，而 model.children 只才會二叉樹框架下的一層。

# 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)

提煉出框架中會的某一層

modules才會離開框架中會所有接口的插值支架，它能夠訪問到最包覆，比如self.layer1.conv1這個接口，還有一個與它們相對應的是name_children一般來說以及named_modules,這兩個不僅才會離開接口的插值支架，還才會離開局域網層的拼法。

# 取框架中會的前兩層new_model = nn.Sequential(*list(model.children)[:2] # 如果希望提煉出出框架中會的所有變換層，可以像后面這樣操控：for layer in model.named_modules:if isinstance(layer[1],nn.Conv2d):conv_model.add_module(layer[0],layer[1])

之外層用于實體能訓練框架

請注意如果留存的框架是 torch.nn.DataParallel，則也就是說的框架也只能是

model.load_state_dict(torch.load('model.pth'), strict=False)

將在 GPU 留存的框架加載到 CPU

model.load_state_dict(torch.load('model.pth', map_location='cpu'))

新增另一個框架的不盡相同之外到屬于自己框架

框架新增值時，如果兩個框架結構不贊同，則直接新增值才會報錯。用后面步驟可以把另一個框架的不盡相同的之外新增到屬于自己框架中會。

# model_new都是屬于自己框架# model_saved都是其他框架，比如用torch.load新增的已留存的框架model_new_dict = model_new.state_dictmodel_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)

『統計數據處置』

近似值統計數據集的平方根和加權

import osimport cv2import numpy as npfrom torch.utils.data import Datasetfrom PIL import Image

def compute_mean_and_std(dataset):# 疊加成PyTorch的dataset，反向平方根和加權mean_r = 0mean_g = 0mean_b = 0

for 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 = 0

N = 0

for 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

得不到視頻統計數據整體反饋

import cv2video = 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）濾波一幀視頻

K = self._num_segmentsif 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)]

特指體能訓練和有效性統計數據實處置

其中會 ToTensor 操控才會將 PIL.Image 或形狀為 H×W×D，參數適用范圍為 [0, 255] 的 np.ndarray 疊加為形狀為 D×H×W，參數適用范圍為 [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)), ])

『框架體能訓練和測試』

分類框架體能訓練編譯支架

# Loss and optimizercriterion = nn.CrossEntropyLossoptimizer = torch.optim.Adam(model.parameters, lr=learning_rate)

# Train the modeltotal_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_gradloss.backwardoptimizer.step

if (i+1) % 100 == 0:print('Epoch: [{}/{}], Step: [{}/{}], Loss: {}'.format(epoch+1, num_epochs, i+1, total_step, loss.item))

分類框架測試編譯支架

# Test the modelmodel.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))

定制loss

傳給torch.nn.Module類寫自己的loss。

class MyLoss(torch.nn.Moudle):def 曲在init曲在(self):super(MyLoss, self).曲在init曲在

def forward(self, x, y):loss = torch.mean((x - y) ** 2)return loss

附加紋理（label smoothing）

寫一個label_smoothing.py的文件，然后在體能訓練編譯支架之中引用，用LSR代替交叉熵損失均可。label_smoothing.py細節如下：

import torchimport torch.nn as nn

class LSR(nn.Module):

def 曲在init曲在(self, e=0.1, reduction='mean'):super.曲在init曲在

self.log_softmax = nn.LogSoftmax(dim=1)self.e = eself.reduction = reduction

def _one_hot(self, labels, classes, value=1):"""Convert labels to one hot vectors

Args:labels: torch tensor in format [label1, label2, label3, ...]classes: int, number of classesvalue: label value in one hot vector, default to 1

Returns: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_hot

def _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 smooth

Returns: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 loss

elif 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')

或者直接在體能訓練文件之中想到label smoothing

for images, labels in train_loader:images, labels = images.cuda, labels.cudaN = labels.size(0)# C is the number of classes.smoothed_labels = torch.full(size=(N, C), fill_value=0.1 / (C - 1)).cudasmoothed_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_gradloss.backwardoptimizer.step

Mixup體能訓練

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([]).itemindex = torch.randperm(images.size(0)).cudamixed_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_gradloss.backwardoptimizer.step

L1 下述本土化

l1_regularization = torch.nn.L1Loss(reduction='sum')loss = ... # Standard cross-entropy lossfor param in model.parameters:loss += torch.sum(torch.abs(param))loss.backward

不對回授項透過二階震蕩（weight decay）

pytorch之中的weight decay相當于l2下述

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)

溫度梯度拼接（gradient clipping）

torch.nn.utils.clip_grad_norm_(model.parameters, max_norm=20)

得不到也就是說努力學習所部

# 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'])

另一種步驟，在一個batch體能訓練編譯支架之中，也就是說的lr是optimizer.param_groups[0]['lr']

努力學習所部震蕩

# 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.steptrain(...)val(...)

最優本土化支架鏈式更加新

從1.4發行版開始，torch.optim.lr_scheduler 正因如此力支持鏈式更加新（chaining），即Gmail可以定義兩個 schedulers，并交替在體能訓練中會用于。

import torchfrom torch.optim import SGDfrom torch.optim.lr_scheduler import ExponentialLR, StepLRmodel = [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.stepscheduler1.stepscheduler2.step

框架體能訓練統計數據處置

PyTorch可以用于tensorboard來統計數據處置體能訓練過程。

安裝和開始運行TensorBoard。

pip install tensorboardtensorboard ；還有logdir=runs

用于SummaryWriter類來采集和統計數據處置反之亦然的統計數據，放了簡便核對，可以用于不盡不盡相同的頁面，比如'Loss/train'和'Loss/test'。

from torch.utils.tensorboard import SummaryWriterimport numpy as np

writer = 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)

留存與加載斷點

請注意為了能夠恢復體能訓練，我們只能同時留存框架和最優本土化支架的穩定狀態，以及也就是說的體能訓練輪數。

提煉出 ImageNet 實體能訓練框架某層的變換形態

# 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.evalconv_representation = model(image)

提煉出 ImageNet 實體能訓練框架多層的變換形態

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 or set

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 曲在init曲在(self, pretrained_model, layers_to_extract):torch.nn.Module.曲在init曲在(self)self._model = pretrained_modelself._model.evalself._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

這兩項正因如此連接層

model = torchvision.models.resnet18(pretrained=True)for param in model.parameters:param.requires_grad = Falsemodel.fc = nn.Linear(512, 100) # Replace the last fc layeroptimizer = torch.optim.SGD(model.fc.parameters, lr=1e-2, momentum=0.9, weight_decay=1e-4)

以較小努力學習所部這兩項正因如此連接層，較小努力學習所部這兩項變換層

model = torchvision.models.resnet18(pretrained= True) finetuned_parameters = list(map(id, model.fc.parameters))conv_parameters = (p forp inmodel.parameters ifid(p) notinfinetuned_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)

『其他請注意事項』

絕須要于太大的差分層。因為nn.Linear(m,n)用于的是O(mn)的內存，差分層太大很不易大于現有顯存。

絕不在過長的序列上用于RNN。因為RNN反向傳播用于的是BPTT算法，其只能的內存和疊加成序列的長度方形差分關系。

model(x) 前用 model.train 和 model.eval 切換局域網穩定狀態。

不只能近似值溫度梯度的編譯支架塊用 with torch.no_grad 包含出去。

model.eval 和 torch.no_grad 的區別在于，model.eval 是將局域網切換為測試穩定狀態，例如 BN 和dropout在體能訓練和測試版用于不盡不盡相同的近似值步驟。torch.no_grad 是關閉 PyTorch 標量的終端切線前提，以減緩加載用于和加速近似值，得不到的結果能夠透過 loss.backward。

model.zero_grad才會把整個框架的值的溫度梯度都歸零, 而optimizer.zero_grad只才會把傳入其中會的值的溫度梯度歸零.

torch.nn.CrossEntropyLoss 的疊加成不只能經過 Softmax。torch.nn.CrossEntropyLoss 等價于 torch.nn.functional.log_softmax + torch.nn.NLLLoss。

loss.backward 前用 optimizer.zero_grad 掃除再加溫度梯度。

torch.utils.data.DataLoader 中會盡量另設 pin_memory=True，對特別小的統計數據集如 MNIST 另設 pin_memory=False 反而更加快一些。num_workers 的另設只能在實驗中會看到極快的取值。

用 del 及時刪掉須要的中會間表達式，減省 GPU 加載。

用于 inplace 操控可減省 GPU 加載，如

x = torch.nn.functional.relu(x, inplace=True)

減緩 CPU 和 GPU 錯綜復雜的統計數據數據傳輸。例如如果你想知道一個 epoch 中會每個 mini-batch 的 loss 和準確所部，必先將它們再加在 GPU 中會等一個 epoch 落幕之后兩兄弟數據傳輸回 CPU 才會比每個 mini-batch 都透過一次 GPU 到 CPU 的數據傳輸更加快。

用于半精確度二進制 half 才會有一定的低速提升，具體效所部意味著 GPU 標準型。只能小心參數精確度過低帶來的穩定性原因。

時特指于 assert tensor.size == (N, D, H, W) 作為復用手段，確保標量等價和你設想中會贊同。

除了標示 y 外，盡量少用于給定標量，用于 n*1 的二維標量代替，可以不必要一些意想不到的給定標量近似值結果。

匯總編譯支架各之外用時

with torch.autograd.profiler.profile(enabled=True, use_cuda=False) as profile: ...print(profile)# 或者在GUI開始運行python -m torch.utils.bottleneck main.py

用于TorchSnooper來復用PyTorch編譯支架，處置程序在指派的時候，就才會終端 print 出來每一行的指派結果的 tensor 的形狀、統計數據特性、電源、是否只能溫度梯度的反饋。

# pip install torchsnooperimport torchsnooper# 對于表達式，用于修飾支架@torchsnooper.snoop# 如果不是表達式，用于 with 語句來激活 TorchSnooper，把體能訓練的那個循環裝進 with 語句中會去。with torchsnooper.snoop: 原本的代

框架可解釋性，用于captum奎：

簡要

張皓：PyTorch Cookbook（特指編譯支架段采集合集），

PyTorch官方HTML和范例

其他

；還有；還有；還有；還有；還有；還有；還有；還有；還有-

END

；還有；還有；還有；還有；還有；還有；還有；還有；還有；還有

腸胃不舒服吃什么藥飲食要注意，若不能緩解，及時就醫檢查
慢性結膜炎有哪些癥狀
米氟來特片是什么藥？能長期服用嗎？

猜你會喜歡的