feat: sync tensorboard(x)

[Zeyi-Lin]

Description

支持同步tensorboardX和pytorch自带的tensorboard。

交互：

from tensorboardX import SummaryWriter
# from torch.utils.tensorboard import SummaryWriter
import numpy as np
import swanlab

swanlab.sync_tensorboardX()
# swanlab.sync_tensorboard_torch()

writer = SummaryWriter('runs/example')

for i in range(100):
    scalar_value = np.random.rand()
    writer.add_scalar('random_scalar', scalar_value, i)

writer.close()

This pull request adds functionality for syncing TensorBoard logs with SwanLab. The most important changes include importing and defining new synchronization functions, updating the __init__.py files to include these new functions, and adding a test script for the new functionality.

New synchronization functions:

• swanlab/sync/tensorboard.py: Added sync_tensorboardX and sync_tensorboard_torch functions to enable synchronization of TensorBoard logs with SwanLab.

Updates to initialization files:

• swanlab/__init__.py: Updated imports and __all__ to include sync_tensorboardX and sync_tensorboard_torch. [1] [2]
• swanlab/sync/__init__.py: Updated imports and __all__ to include sync_tensorboardX and sync_tensorboard_torch.

Testing:

• test/tensorboard_sync.py: Added a test script to verify the functionality of sync_tensorboard_torch with SwanLab.

Closes: #786