trinity.trainer package

Subpackages

• trinity.trainer.verl package
  • Submodules
    • trinity.trainer.verl.dp_actor module
      • DataParallelPPOActor
    • trinity.trainer.verl.fsdp_workers module
      • ActorRolloutRefWorker
      • CriticWorker
  • Module contents

Submodules

• trinity.trainer.trainer module
  • Trainer
    • Trainer.__init__()
    • Trainer.prepare()
    • Trainer.train()
    • Trainer.train_step()
    • Trainer.need_sync()
    • Trainer.need_save()
    • Trainer.sync_weight()
    • Trainer.save_checkpoint()
    • Trainer.shutdown()
    • Trainer.train_step_num
    • Trainer.is_alive()
    • Trainer.get_actor()
  • TrainEngineWrapper
    • TrainEngineWrapper.prepare()
    • TrainEngineWrapper.train_step_num
    • TrainEngineWrapper.train_step()
    • TrainEngineWrapper.save_checkpoint()
    • TrainEngineWrapper.sync_weight()
    • TrainEngineWrapper.upload_state_dict()
    • TrainEngineWrapper.save_state_dict()
  • get_trainer_wrapper()
• trinity.trainer.verl_trainer module
  • CheckpointMonitor
    • CheckpointMonitor.__init__()
    • CheckpointMonitor.update_latest_checkpoint_step()
    • CheckpointMonitor.update_latest_state_dict_step()
    • CheckpointMonitor.register_thread_count()
    • CheckpointMonitor.monitor_step()
    • CheckpointMonitor.notify_started()
    • CheckpointMonitor.notify_finished()
    • CheckpointMonitor.get_actor()
  • VerlPPOTrainerWrapper
    • VerlPPOTrainerWrapper.__init__()
    • VerlPPOTrainerWrapper.init_workers()
    • VerlPPOTrainerWrapper.train_step_num
    • VerlPPOTrainerWrapper.prepare()
    • VerlPPOTrainerWrapper.save_state_dict()
    • VerlPPOTrainerWrapper.upload_state_dict()
    • VerlPPOTrainerWrapper.train_step()
    • VerlPPOTrainerWrapper.save_checkpoint()
    • VerlPPOTrainerWrapper.sync_weight()
    • VerlPPOTrainerWrapper.post_process_batch()

Module contents

class trinity.trainer.Trainer(config: Config)

Bases: object

Consume the experience and train the model.

__init__(config: Config) → None

async prepare() → None

Prepare the trainer.

async train() → str

Train the model.

async train_step(exps: Experiences) → Dict

Train one step.

Returns:

Whether to continue training. Dict: Metrics of the training step.

Return type:

bool

async need_sync() → bool

Whether to sync the model weight.

need_save() → bool

Whether to save the checkpoint.

async sync_weight() → Dict

Sync the model weight.

save_checkpoint(block_until_saved: bool = False, save_as_hf: bool = False) → Dict

async shutdown() → None

property train_step_num: int

Get the current training step number.

async is_alive() → bool

Check if the trainer is alive.

classmethod get_actor(config: Config)

Get a Ray actor for the trainer.

class trinity.trainer.TrainEngineWrapper

Bases: ABC

A wrapper class to wrap various training engines.

abstract prepare() → None

Do some preparation before training started.

abstract property train_step_num: int

Get the current training step number.

abstract train_step(batch: Experiences) → Dict

Training one step.

Parameters:

batch (Experiences) – A batch of experiences to train.

Returns:

Metrics of the training step.

Return type:

Dict

abstract save_checkpoint(block_until_saved: bool = False, save_as_hf: bool = False) → None

Save the checkpoint.

abstract sync_weight() → None

Sync the model weight.

abstract upload_state_dict() → None

Upload the state dict to Synchronizer.

abstract save_state_dict() → None

Only save the model state dict for Synchronizer.

trinity.trainer.get_trainer_wrapper(config: Config) → TrainEngineWrapper

Get a trainer wrapper.