PyTorch自学&遇到的一些错误
为什么是false？
①检查GPU是否支持CUDA？
支持
理解Package结构及法宝函数的作用
pytorch就像一个工具箱
dir():打开操作，能看到里面有什么东西---->dir(torch)
help():说明书---->help(torch.cuda.is_available)
From torch.utils.data import Dataset
pytorch如何读取数据？
①Dataset
提供一种方式去获取数据及其label
如何获取每一个数据及其label
告诉我们总共有多少的数据
②Dataloader
为网络提供不同的数据形式
Tensorboard的使用
Global_step：x轴
Scalar_value：y轴
如何读取logs里面的文件
PS F:\HDU\PyTorchStudy> tensorboard --logdir=logs
指定端口：tensorboard --logdir=logs --port=6007
结果：有图像展示
图像变换，transform的使用
利用numpy.array()，对PIL图片进行转换
torchvision中的transforms:对图像进行一个变换
Transforms.py 工具箱
Totensor/resize/。。。
图片--->工具 --->结果
结果：
tensor([[[0.3137, 0.3137, 0.3137, ..., 0.3176, 0.3098, 0.2980],
[0.3176, 0.3176, 0.3176, ..., 0.3176, 0.3098, 0.2980],
[0.3216, 0.3216, 0.3216, ..., 0.3137, 0.3098, 0.3020],
...,
[0.3412, 0.3412, 0.3373, ..., 0.1725, 0.3725, 0.3529],
[0.3412, 0.3412, 0.3373, ..., 0.3294, 0.3529, 0.3294],
[0.3412, 0.3412, 0.3373, ..., 0.3098, 0.3059, 0.3294]],
[[0.5922, 0.5922, 0.5922, ..., 0.5961, 0.5882, 0.5765],
[0.5961, 0.5961, 0.5961, ..., 0.5961, 0.5882, 0.5765],
[0.6000, 0.6000, 0.6000, ..., 0.5922, 0.5882, 0.5804],
...,
[0.6275, 0.6275, 0.6235, ..., 0.3608, 0.6196, 0.6157],
[0.6275, 0.6275, 0.6235, ..., 0.5765, 0.6275, 0.5961],
[0.6275, 0.6275, 0.6235, ..., 0.6275, 0.6235, 0.6314]],
[[0.9137, 0.9137, 0.9137, ..., 0.9176, 0.9098, 0.8980],
[0.9176, 0.9176, 0.9176, ..., 0.9176, 0.9098, 0.8980],
[0.9216, 0.9216, 0.9216, ..., 0.9137, 0.9098, 0.9020],
...,
[0.9294, 0.9294, 0.9255, ..., 0.5529, 0.9216, 0.8941],
[0.9294, 0.9294, 0.9255, ..., 0.8863, 1.0000, 0.9137],
[0.9294, 0.9294, 0.9255, ..., 0.9490, 0.9804, 0.9137]]])
常见的Transforms
| 输入 |
PIL |
Image.open() |
| 输出 |
tensor |
ToTensor() |
| 作用 |
narrays |
Cv.imread() |
call的使用
__call__Hello zhangsan
Hellolisi
归一化：Normalize
output[channel] = (input[channel] - mean[channel]) / std[channel]
tensor(0.8863)
tensor(0.7725)
Resize()的使用
(500, 333)
tensor([[[0.8863, 0.8824, 0.8745, ..., 0.8392, 0.8392, 0.8392],
[0.8824, 0.8784, 0.8706, ..., 0.8392, 0.8392, 0.8392],
[0.8784, 0.8745, 0.8667, ..., 0.8353, 0.8353, 0.8353],
...,
[1.0000, 1.0000, 1.0000, ..., 0.8314, 0.8314, 0.8314],
[1.0000, 1.0000, 1.0000, ..., 0.8314, 0.8314, 0.8314],
[1.0000, 1.0000, 1.0000, ..., 0.8314, 0.8314, 0.8314]],
[[0.1451, 0.1412, 0.1333, ..., 0.8471, 0.8471, 0.8471],
[0.1451, 0.1412, 0.1333, ..., 0.8471, 0.8471, 0.8471],
[0.1490, 0.1451, 0.1373, ..., 0.8431, 0.8431, 0.8431],
...,
[0.8275, 0.8275, 0.8275, ..., 0.8549, 0.8549, 0.8549],
[0.8275, 0.8275, 0.8275, ..., 0.8549, 0.8549, 0.8549],
[0.8275, 0.8275, 0.8275, ..., 0.8549, 0.8549, 0.8549]],
[[0.1686, 0.1647, 0.1569, ..., 0.4275, 0.4275, 0.4275],
[0.1686, 0.1647, 0.1569, ..., 0.4275, 0.4275, 0.4275],
[0.1686, 0.1647, 0.1569, ..., 0.4235, 0.4235, 0.4235],
...,
[0.0078, 0.0078, 0.0078, ..., 0.4235, 0.4235, 0.4235],
[0.0078, 0.0078, 0.0078, ..., 0.4235, 0.4235, 0.4235],
[0.0078, 0.0078, 0.0078, ..., 0.4235, 0.4235, 0.4235]]])
Compose()用法
数据需要时transforms类型，所以得到Compose([transforms参数1,transforms参数2,…])
RandomCrop
关注输入和输出类型
多看官方文档
关注方法需要什么参数
不知道返回值的时候
*print(type())
*debug
DataLoader的使用
例：
Dataloader(batch_size=4)
Img0,target0 = dataset[0]
Img1,target1 = dataset[1]
Img2,target2 = dataset[2]
Img3,target3 = dataset[3]
Getitem():
Return img,target
神经网络的搭建nn.Module