【pytorch】tensorboard + transforms的使用

一、tensorboard的使用

1. 加载一张图片转化为tensor类型，并通过tenboard可视化

from PIL import Image
from torch.utils.tensorboard import SummaryWriter
from torchvision import transforms
 
img_path = "dataset/train/ants_image/0013035.jpg"
 
img = Image.open(img_path)
 
#print(img)
writer = SummaryWriter('logs')
 
tensor_trans = transforms.ToTensor()
tensor_img = tensor_trans(img)
 
writer.add_image("Tesnor_img",tensor_img)
 
#print(tensor_img)
 
writer.close()
 
 
#在Terminal中输入 tensorboard --logdir=logs

 二、transforms的常用函数

细心的小伙伴已发现了，上面已经使用了transforms了，我们在添加图片时，用到了ToTensor()这个函数。

ToTensor()函数：

将一个PIL类型转换成tensor类型；

#totensor
tensor_trans = transforms.ToTensor()
tensor_img = tensor_trans(img)
writer.add_image("to_Tesnor",tensor_img)
print(tensor_img)

Normalize() 函数：

用于归一化，使他的范数或者数值在一定的范围。

#normalize
trnas_norm = transforms.Normalize([0.5,0.5,0.5],[0.5,0.5,0.5])
img_norm = trnas_norm(tensor_img)
print(img_norm[0][0][0])
 
writer.add_image("normalize",img_norm)

Resize()函数: 

用来调整数组大小。

#resize
print(img)
trans_resize = transforms.Resize((512,512))
img_resize = trans_resize(img)
 
img_resize = tensor_trans(img_resize)
 
writer.add_image("resize",img_resize,0)
print(img_resize)

 

Compose()函数:

简单来说就是将各种操作就行联合起来进行操作，注意操作顺序。

#compose
trans_resize_2 = transforms.Resize(512)
 
trans_compose = transforms.Compose([trans_resize_2,tensor_trans])
img_resize_2 = trans_compose(img)
writer.add_image("resize",img_resize_2,1)

 RandomCrop()函数:

随机裁剪函数，看效果，这里只展示了一步。

#randomcrop
trans_random = transforms.RandomCrop(128)
 
trans_compose_2 = transforms.Compose([trans_random,tensor_trans])
 
for i in range(10):
    img_crop = trans_compose_2(img)
    writer.add_image("Randomcrop",img_crop,i)
 
 
writer.close()

---

tensorboard配合transforms就是pytorch学习中的两大利器