Openrl Versions Save

• add deepspeed support
• fix eval_callback bug
• fix petting zoo bugs
• add envpool to openrl

• fix arena petting zoo import error
• fix typo: change loss rate to lose rate
• add RandomAgent for Arena

• fix winning rate output bug of smac
• fix atari training bugs
• fix windows bugs
• add test atari

• add smacv2
• fix arena bugs

• support loading stable-baseline3's models from hugging face

• fix openrl random opponent bugs
• add gym_pybullet_drones env
• add A2C algorithm
• merge dm_control to gymnasium

• add JiDi evaluation support #217
• add google footbal evaluation support #218
• add google footbal evaluation support #218

add custom env examples

• add dm-control
• add snake environment

• add shuffle traj to offline env
• support log evaluation callback info to wandb
• add arena
• add set seed to arena, and test reproducibility
• fix leaky_relu twice written