import torch
from torch.utils.tensorboard import SummaryWriter
writer = SummaryWriter(“logs”)
writer.add_image()
for i in range(100):
writer.add_scalar(“y=2x”,2*i,i)
writer.close()
tensorboard --logdir=logs --port=6007
端口可以指定
add_scalar
tag (string): Data identifier 图像标签
scalar_value (float or string/blobname): Value to save 图像y轴
global_step (int): Global step value to record 图像x轴
当打开多个log 图像形状会重叠
import torch
from PIL import Image
import numpy as np
from torch.utils.tensorboard import SummaryWriter
writer = SummaryWriter(“logs”)
image_path=‘练手数据集/train/ants_image/0013035.jpg’
img_PIL=Image.open(image_path)
img_array=np.array(img_PIL)
print(type(img_array))
print(img_array.shape)
writer.add_image(‘train’,img_array,1,dataformats=‘HWC’)
writer.add_image(‘test’,img_array,1,dataformats=‘HWC’)
for i in range(100):
writer.add_scalar(“y=2x”,2*i,i)
writer.close()
tag (string): Data identifier
img_tensor (torch.Tensor, numpy.array, or string/blobname): Image data
global_step (int): Global step value to record
参数1 就是 step 1 参数2 就是step2 相同的文件一定是step 1 到 step 2
读取和保存
结构+参数
model=torch.load(xx.pth)#这种读取同时有结构+参数
torch.save(model,“xx.pth”) xx.pth 是path
#只有参数
model.load_state_dict(torch.load(xx.pth))
torch.save(xx.state_dict(),“xx.pth”)
vgg16 =torchvision.model.vgg16(pretrain=False)
vgg16.load_state_dict(torch.load(xx.pth))
#第二种 只有参数 如果加载 需要模型