TensorBoard简单教程

TensorBoard简单教程

这里仅展示命令行怎么使用

安装

pip install tensorboard

使用

要启动 TensorBoard，打开终端或命令提示符并运行：

tensorboard --logdir=<directory_name>

将 directory_name 标记替换为保存数据的目录。 默认是“logs”。

1. 首先导入tensorboard

from torch.utils.tensorboard import SummaryWriter   

2. 实例化

log是保存日志路径结尾，保存日志的文件会生成且命名

def __init__(self,
	 log_dir: str = None,
	 comment: str = "",
	 purge_step: int = None,
	 max_queue: int = 10,
	 flush_secs: int = 120,
	 filename_suffix: str = "") -> Non

writer = SummaryWriter('./path/to/log')

3. 调用实例对象将需要的数据写入日志文件

def add_scalar(self,
	   tag: str,   # 可以理解为title
	   scalar_value: Any,   # 记录的值，任意类型
	   global_step: int = None,   # x轴坐标
	   walltime: float = None,
	   new_style: bool = False,
	   double_precision: bool = False) -> None

writer.add_scalar(tag, scalar_value, global_step=None, walltime=None)

示例代码

from torch.utils.tensorboard import SummaryWriter  
import numpy as np  
writer = SummaryWriter(log_dir='logs/test1', comment="aaaa")  
  
for n_iter in range(100):  
    writer.add_scalar('Loss/train', np.random.random(), n_iter)  
    writer.add_scalar('Loss/test', np.random.random(), n_iter)  
    writer.add_scalar('Accuracy/train', np.random.random(), n_iter)  
    writer.add_scalar('Accuracy/test', np.random.random(), n_iter)

控制台命令

tensorboard --logdir='./tensorboard/logs/test2'

成功的界面：

效果图：