transforms使用（pytroch框架入门]）

2、TensorBoard的使用

• 使用1 writer.add_scalar()

from torch.utils.tensorboard import SummaryWriter

writer=SummaryWriter("logs")
# writer.add_image()
# y=x
for i in range(100):
    writer.add_scalar("y=2x",3*i,i)#title 、y轴、x轴

writer.close

得到一个log

接着在terminal里面，输入
tensorboard --logdir=D:\cv_box\pytorch_learning\数据的加载.ipynb_checkpoints\logs --port=6007

双击蓝色地址，得到

• 使用2 add_image()只接受tensor格式

from torch.utils.tensorboard import SummaryWriter
from PIL import Image
import numpy as np

writer=SummaryWriter("logs")
image_path=r"pytorch_learning\dataset\train\daisy\5547758_eea9edfd54_n.jpg"
img_PIL=Image.open(image_path)#不是add_image形参的格式，则这里使用numpy来转换格式
img_array=np.array(img_PIL)

writer.add_image("train",img_array,1,dataformats='HWC')
# y=2x
for i in range(100):
    writer.add_scalar("y=2x",2*i,i)#title 、y轴、x轴

writer.close()

3、torchvision中的transforms

1、Transforms介绍

2、 常见的Transforms工具箱的使用

输入：  PIL            Image.open()
输出： tensor	      ToTensor()
作用： narrays       cv.imread（）

• 常见的Transform（一） Totensor类使用
  “”"Convert a PIL Image or numpy.ndarray to tensor. This transform does not support torchscript.

from torch.utils.tensorboard import SummaryWriter
from torchvision import transforms
from PIL import Image

img_path=r"D:\cv_box\pytorch_learning\dataset\train\daisy\5547758_eea9edfd54_n.jpg"
img=Image.open(img_path)
writer=SummaryWriter("logs")

# ToTensor
tensor_trans=transforms.ToTensor()
tensor_img=tensor_trans(img)

writer.add_image("Tensor_img",tensor_img)#只接受Tensor格式
writer.close()

• 常见的Transform（二） Normlize类
  “”"Normalize a tensor image with mean and standard deviation.

#	output[channel] = (input[channel] - mean[channel]) / std[channel]
# Normalize
print(img_tensor[0][0][0])
trans_norm=transforms.Normalize([0.5,0.5,0.5],[0.5,0.5,0.5])#均值、标准差
img_norm=trans_norm(img_tensor)
print(img_norm[0][0][0])
writer.add_image("Normalize",img_norm)

writer.close()

# Normalize
print(img_tensor[0][0][0])
trans_norm=transforms.Normalize([1,3,5],[3,2,1])#均值、标准差
img_norm=trans_norm(img_tensor)
print(img_norm[0][0][0])
writer.add_image("Normalize",img_norm,1)

writer.close()

• 常见的Transform（三） Resize类
  Resize the input image to the given size.

# Resize
print(img_PIL.size)# PIL Image or Tensor
trans_resize=transforms.Resize((512,512))
img_resize=trans_resize(img_PIL)#要求的格式PIL
print(img_resize)
img_resize_tensor=trans_totensor(img_resize)
print(img_resize_tensor)
writer.add_image("Resize_Totensor",img_resize_tensor,0)#tensorboard只接收tensor格式
writer.close()

• 常见的Transform（四） Compose类
  Composes several transforms together.

# Compose - resize -2
trans_resize_2=transforms.Resize(512)
# PIL320  >>   PIL512    >>    Tensor512
trans_compose=transforms.Compose([trans_resize_2,trans_totensor])#两步合成一步
img_resize_2_tensor=trans_compose(img_PIL)
writer.add_image("Resize_2_Totensor",img_resize_2_tensor)
print(img_resize_2_tensor)

writer.close()

• 常见的Transform（五） RandomCrop类
  Base class for a list of transformations with randomne

# RandomCrop
trans_randomcrop=transforms.RandomCrop(200)
# centercrop   >>    totensor
trans_compose_2=transforms.Compose([trans_randomcrop,trans_totensor])
for i in range(10):
    img_PIL_random_totensor=trans_compose_2(img_PIL)
    writer.add_image("RandomCrop_Totensor",img_PIL_random_totensor,i)

writer.close()

• 总结：
  主要关注函数输入输出类型
  关注函数的参数