feat: sync wandb

[Zeyi-Lin]

Description

实现在训练时同步wandb的实验到swanlab，只需在脚本的最开头增加，即可实现跟踪：

import swanlab

swanlab.sync_wandb()

测试代码：

import wandb
import random
import swanlab

swanlab.sync_wandb()
# swanlab.init(project="sync_wandb")

wandb.init(
  project="test",
  config={"a": 1, "b": 2},
  name="test",
  )

epochs = 10
offset = random.random() / 5
for epoch in range(2, epochs):
  acc = 1 - 2 ** -epoch - random.random() / epoch - offset
  loss = 2 ** -epoch + random.random() / epoch + offset

  wandb.log({"acc": acc, "loss": loss})

Closes: #784

Details

This pull request introduces a new feature to integrate wandb (Weights & Biases) with swanlab by adding a synchronization mechanism. The changes include importing necessary modules, defining the sync_wandb function, and updating the package's initialization and export lists.

Key changes include:

Integration with wandb:

• swanlab/sync/wandb.py: Added the sync_wandb function to synchronize wandb with swanlab, including patching wandb.init, wandb.log, and wandb.finish to integrate with swanlab's logging and configuration management.

Module and package updates:

• swanlab/__init__.py: Imported sync_wandb and updated the export list to include sync_wandb. [1] [2]
• swanlab/sync/__init__.py: Added import and export for sync_wandb to ensure it is accessible from the swanlab.sync module.