Training the gym-hybrid using hppo

[ddharshan]

Hi I am new to DI-engine and trying to train the gym hybrid "thomashirtz" agent using HPPO. could you please explain me what are config,entry, envs folders are mentioning under DI-engine/dizoo/gym_hybrid/??

[PaParaZz1]

• envs: the wrapper which transform original gym env into DI-engine's format
• config: the configuration file which contains the setting of Algorithm + Env (nested dict) and the default main function to run this config
• entry: another main functions for running corresponding config

we suggest that you should implement your own main function for your demand, and you can imitate the following two files:

• dizoo/gym_hybrid/entry/gym_hybrid_ddpg_main.py(for env)
• ding/entry/serial_entry_onpolicy.py(for HPPO on policy algorithm)

[hugo921]

Could you provide a hppo main function file ,I am new to DI-engine and H-PPO, thank you very much.

• envs: the wrapper which transform original gym env into DI-engine's format
• config: the configuration file which contains the setting of Algorithm + Env (nested dict) and the default main function to run this config
• entry: another main functions for running corresponding config

we suggest that you should implement your own main function for your demand, and you can imitate the following two files:

• dizoo/gym_hybrid/entry/gym_hybrid_ddpg_main.py(for env)
• ding/entry/serial_entry_onpolicy.py(for HPPO on policy algorithm)