tensorboard中一个函数多图

显示不同颜色的曲线，该怎么做？

首先，您需要在代码中定义每个曲线的名称和颜色，然后将它们传递给 tensorboard.writer.add_scalar 函数以进行记录。例如:

from torch.utils.tensorboard import SummaryWriter

# Define curve names and colors
curve1_name = 'Curve 1'
curve1_color = 'red'
curve2_name = 'Curve 2'
curve2_color = 'blue'

# Initialize writer
writer = SummaryWriter()

# Record curves
for i in range(10):
  # Record curve 1
  curve1_val = i * 2
  writer.add_scalar(curve1_name, curve1_val, i, color=curve1_color)

  # Record curve 2
  curve2_val = i**2
  writer.add_scalar(curve2_name, curve2_val, i, color=curve2_color)

# Close writer
writer.close()

在 tensorboard 中，您可以通过选择“Runs”选项卡，然后从下拉菜单中选择当前记录的运行以查看记录的曲线。您应该能够看到两条曲线，其中一条为红色，另一条为蓝色。