## Description The base `Trainer` class contains methods like `_on_fit_start`and `_on_fit_end`, which is only used in RL training. ## TODOs - [x] Separate current base `Trainer` into `SLTrainer` and `RLTrainer`. - [x] Update `SFTTrainer`. - [x] Update `RMTrainer`. - [x] Update `PPOTrainer`.
Description
The base
Trainerclass contains methods like_on_fit_startand_on_fit_end, which is only used in RL training.TODOs
TrainerintoSLTrainerandRLTrainer.SFTTrainer.RMTrainer.PPOTrainer.