Pytorch教程內(nèi)置模型源碼實(shí)現(xiàn)
翻譯自
https://pytorch.org/docs/stable/torchvision/models.html
主要講解了torchvision.models的使用
torchvision.models
torchvision.models中包含了如下模型
- AlexNet
- VGG
- ResNet
- SqueezeNet
- DenseNet
- Inception v3
隨機(jī)初始化模型
import torchvision.models as models resnet18 = models.resnet18() alexnet = models.alexnet() vgg16 = models.vgg16() squeezenet = models.squeezenet1_0() desnet = models.densenet161() inception =models.inception_v3()
使用預(yù)訓(xùn)練好的參數(shù)
pytorch提供了預(yù)訓(xùn)練的模型,使用torch.utils.model_zoo ,通過讓參數(shù)pretrained =True來構(gòu)建訓(xùn)練好的模型
方法如下
resnet18 = models.resnet18(pretrained=True) alexnet = models.alexnet(pretrained=True) squeezenet = models.squeezenet1_0(pretrained=True) vgg16 = models.vgg16(pretrained=True) densenet = models.densenet161(pretrained=True) inception = models.inception_v3(pretrained=True)
實(shí)例化一個(gè)預(yù)訓(xùn)練好的模型會(huì)自動(dòng)下載權(quán)重到緩存目錄,這個(gè)權(quán)重存儲(chǔ)路徑可以通過環(huán)境變量TORCH_MODEL_ZOO來指定,詳細(xì)的參考torch.utils.model_zoo.load_url() 這個(gè)函數(shù)
有的模型試驗(yàn)了不同的訓(xùn)練和評(píng)估,例如batch normalization。使用model.train()和model.eval()來轉(zhuǎn)換,查看train() or eval() 來了解更多細(xì)節(jié)
所有的預(yù)訓(xùn)練網(wǎng)絡(luò)希望使用相同的方式進(jìn)行歸一化,例如圖片是mini-batch形式的3通道RGB圖片(3HW),H和W最少是244,。 圖像必須加載到[0,1]范圍內(nèi),然后使用均值=[0.485,0.456,0.406]和std =[0.229, 0.224, 0.225]進(jìn)行歸一化。
您可以使用以下轉(zhuǎn)換來normalzie:
normalize = trainform.Normalize9mean = [0.485,0.456,0.406],std = [0.229,0.224,0.225])
在這里我們可以找到一個(gè)在Imagenet上的這樣的例子
https://github.com/pytorch/examples/blob/42e5b996718797e45c46a25c55b031e6768f8440/imagenet/main.py#L89-L101
目前這些模型的效果如下
下面是模型源碼的具體實(shí)現(xiàn),具體實(shí)現(xiàn)大家可以閱讀源碼
###ALEXNET torchvision.models.alexnet(pretrained=False, **kwargs)[SOURCE] AlexNet model architecture from the “One weird trick…” paper. Parameters: pretrained (bool) – If True, returns a model pre-trained on ImageNet ###VGG torchvision.models.vgg11(pretrained=False, **kwargs)[SOURCE] VGG 11-layer model (configuration “A”) Parameters: pretrained (bool) – If True, returns a model pre-trained on ImageNet torchvision.models.vgg11_bn(pretrained=False, **kwargs)[SOURCE] VGG 11-layer model (configuration “A”) with batch normalization Parameters: pretrained (bool) – If True, returns a model pre-trained on ImageNet torchvision.models.vgg13(pretrained=False, **kwargs)[SOURCE] VGG 13-layer model (configuration “B”) Parameters: pretrained (bool) – If True, returns a model pre-trained on ImageNet torchvision.models.vgg13_bn(pretrained=False, **kwargs)[SOURCE] VGG 13-layer model (configuration “B”) with batch normalization Parameters: pretrained (bool) – If True, returns a model pre-trained on ImageNet torchvision.models.vgg16(pretrained=False, **kwargs)[SOURCE] VGG 16-layer model (configuration “D”) Parameters: pretrained (bool) – If True, returns a model pre-trained on ImageNet torchvision.models.vgg16_bn(pretrained=False, **kwargs)[SOURCE] VGG 16-layer model (configuration “D”) with batch normalization Parameters: pretrained (bool) – If True, returns a model pre-trained on ImageNet torchvision.models.vgg19(pretrained=False, **kwargs)[SOURCE] VGG 19-layer model (configuration “E”) Parameters: pretrained (bool) – If True, returns a model pre-trained on ImageNet torchvision.models.vgg19_bn(pretrained=False, **kwargs)[SOURCE] VGG 19-layer model (configuration ‘E') with batch normalization Parameters: pretrained (bool) – If True, returns a model pre-trained on ImageNet RESNET torchvision.models.resnet18(pretrained=False, **kwargs)[SOURCE] Constructs a ResNet-18 model. Parameters: pretrained (bool) – If True, returns a model pre-trained on ImageNet torchvision.models.resnet34(pretrained=False, **kwargs)[SOURCE] Constructs a ResNet-34 model. Parameters: pretrained (bool) – If True, returns a model pre-trained on ImageNet torchvision.models.resnet50(pretrained=False, **kwargs)[SOURCE] Constructs a ResNet-50 model. Parameters: pretrained (bool) – If True, returns a model pre-trained on ImageNet torchvision.models.resnet101(pretrained=False, **kwargs)[SOURCE] Constructs a ResNet-101 model. Parameters: pretrained (bool) – If True, returns a model pre-trained on ImageNet torchvision.models.resnet152(pretrained=False, **kwargs)[SOURCE] Constructs a ResNet-152 model. Parameters: pretrained (bool) – If True, returns a model pre-trained on ImageNet SQUEEZENET torchvision.models.squeezenet1_0(pretrained=False, **kwargs)[SOURCE] SqueezeNet model architecture from the “SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size” paper. Parameters: pretrained (bool) – If True, returns a model pre-trained on ImageNet torchvision.models.squeezenet1_1(pretrained=False, **kwargs)[SOURCE] SqueezeNet 1.1 model from the official SqueezeNet repo. SqueezeNet 1.1 has 2.4x less computation and slightly fewer parameters than SqueezeNet 1.0, without sacrificing accuracy. Parameters: pretrained (bool) – If True, returns a model pre-trained on ImageNet DENSENET torchvision.models.densenet121(pretrained=False, **kwargs)[SOURCE] Densenet-121 model from “Densely Connected Convolutional Networks” Parameters: pretrained (bool) – If True, returns a model pre-trained on ImageNet torchvision.models.densenet169(pretrained=False, **kwargs)[SOURCE] Densenet-169 model from “Densely Connected Convolutional Networks” Parameters: pretrained (bool) – If True, returns a model pre-trained on ImageNet torchvision.models.densenet161(pretrained=False, **kwargs)[SOURCE] Densenet-161 model from “Densely Connected Convolutional Networks” Parameters: pretrained (bool) – If True, returns a model pre-trained on ImageNet torchvision.models.densenet201(pretrained=False, **kwargs)[SOURCE] Densenet-201 model from “Densely Connected Convolutional Networks” Parameters: pretrained (bool) – If True, returns a model pre-trained on ImageNet INCEPTION V3 torchvision.models.inception_v3(pretrained=False, **kwargs)[SOURCE] Inception v3 model architecture from “Rethinking the Inception Architecture for Computer Vision”. Parameters: pretrained (bool) – If True, returns a model pre-trained on ImageNet
以上就是Pytorch教程內(nèi)置模型源碼實(shí)現(xiàn)的詳細(xì)內(nèi)容,更多關(guān)于Pytorch內(nèi)置模型的資料請(qǐng)關(guān)注腳本之家其它相關(guān)文章!
相關(guān)文章
Python使用numpy模塊創(chuàng)建數(shù)組操作示例
這篇文章主要介紹了Python使用numpy模塊創(chuàng)建數(shù)組操作,結(jié)合實(shí)例形式分析了Python使用numpy模塊實(shí)現(xiàn)數(shù)組的創(chuàng)建、賦值、修改、打印等相關(guān)操作技巧與注意事項(xiàng),需要的朋友可以參考下2018-06-06centos 安裝Python3 及對(duì)應(yīng)的pip教程詳解
這篇文章主要介紹了centos 安裝Python3 及對(duì)應(yīng)的pip的教程,本文給大家介紹的非常詳細(xì),具有一定的參考借鑒價(jià)值,需要的朋友可以參考下2019-06-06Python input函數(shù)實(shí)現(xiàn)獲取鍵盤輸入的字符串流程講解
這篇文章主要介紹了Python input函數(shù)實(shí)現(xiàn)獲取鍵盤輸入的字符串流程,input()是Python的內(nèi)置函數(shù),用于從控制臺(tái)讀取用戶輸入的內(nèi)容。input()函數(shù)總是以字符串的形式來處理用戶輸入的內(nèi)容,所以用戶輸入的內(nèi)容可以包含任何字符2023-01-01Python中tkinter開發(fā)的常用29種功能用法總結(jié)
tkinter(Tk?interface)是Python的標(biāo)準(zhǔn)GUl庫(kù),支持跨平臺(tái)的GUl程序開發(fā),本文為大家整理了tkinter開發(fā)時(shí)常用的29種功能用法,希望對(duì)大家有所幫助2023-05-05Python pandas軸旋轉(zhuǎn)stack和unstack的使用說明
這篇文章主要介紹了Python pandas軸旋轉(zhuǎn)stack和unstack的使用說明,具有很好的參考價(jià)值,希望對(duì)大家有所幫助。一起跟隨小編過來看看吧2021-03-03pyinstaller打包opencv和numpy程序運(yùn)行錯(cuò)誤解決
這篇文章主要介紹了pyinstaller打包opencv和numpy程序運(yùn)行錯(cuò)誤解決,文中通過示例代碼介紹的非常詳細(xì),對(duì)大家的學(xué)習(xí)或者工作具有一定的參考學(xué)習(xí)價(jià)值,需要的朋友可以參考下2019-08-08Python操作lxml庫(kù)實(shí)戰(zhàn)之Xpath篇
XPath是一門在XML文檔中查找信息的語言,下面這篇文章主要給大家介紹了關(guān)于Python操作lxml庫(kù)實(shí)戰(zhàn)之Xpath篇的相關(guān)資料,文中通過示例代碼介紹的非常詳細(xì),需要的朋友可以參考下2022-12-12