reagent.training package

Subpackages

Submodules

reagent.training.c51_trainer module

class reagent.training.c51_trainer.C51Trainer(q_network, q_network_target, actions: List[str] = Field(name='actions',type=typing.List[str],default=<dataclasses._MISSING_TYPE object>,default_factory=<class 'list'>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), rl: reagent.core.parameters.RLParameters = Field(name='rl',type=<class 'reagent.core.parameters.RLParameters'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<class 'reagent.core.parameters.RLParameters'>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), double_q_learning: bool = True, minibatch_size: int = 1024, minibatches_per_step: int = 1, num_atoms: int = 51, qmin: float = -100, qmax: float = 200, optimizer: reagent.optimizer.union.Optimizer__Union = Field(name='optimizer',type=<class 'reagent.optimizer.union.Optimizer__Union'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<bound method Optimizer__Union.default of <class 'reagent.optimizer.union.Optimizer__Union'>>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD))

Bases: reagent.training.rl_trainer_pytorch.RLTrainerMixin, reagent.training.reagent_lightning_module.ReAgentLightningModule

Implementation of 51 Categorical DQN (C51)

See https://arxiv.org/abs/1707.06887 for details

argmax_with_mask(q_values, possible_actions_mask)

boost_rewards(rewards: torch.Tensor, actions: torch.Tensor) → torch.Tensor

configure_optimizers()

Choose what optimizers and learning-rate schedulers to use in your optimization. Normally you’d need one. But in the case of GANs or similar you might have multiple.

Returns

Any of these 6 options.

Single optimizer.
List or Tuple of optimizers.
Two lists - The first list has multiple optimizers, and the second has multiple LR schedulers (or multiple lr_scheduler_config).
Dictionary, with an "optimizer" key, and (optionally) a "lr_scheduler" key whose value is a single LR scheduler or lr_scheduler_config.
Tuple of dictionaries as described above, with an optional "frequency" key.
None - Fit will run without any optimizer.

The lr_scheduler_config is a dictionary which contains the scheduler and its associated configuration. The default configuration is shown below.

lr_scheduler_config = {
    # REQUIRED: The scheduler instance
    "scheduler": lr_scheduler,
    # The unit of the scheduler's step size, could also be 'step'.
    # 'epoch' updates the scheduler on epoch end whereas 'step'
    # updates it after a optimizer update.
    "interval": "epoch",
    # How many epochs/steps should pass between calls to
    # `scheduler.step()`. 1 corresponds to updating the learning
    # rate after every epoch/step.
    "frequency": 1,
    # Metric to to monitor for schedulers like `ReduceLROnPlateau`
    "monitor": "val_loss",
    # If set to `True`, will enforce that the value specified 'monitor'
    # is available when the scheduler is updated, thus stopping
    # training if not found. If set to `False`, it will only produce a warning
    "strict": True,
    # If using the `LearningRateMonitor` callback to monitor the
    # learning rate progress, this keyword can be used to specify
    # a custom logged name
    "name": None,
}

When there are schedulers in which the .step() method is conditioned on a value, such as the torch.optim.lr_scheduler.ReduceLROnPlateau scheduler, Lightning requires that the lr_scheduler_config contains the keyword "monitor" set to the metric name that the scheduler should be conditioned on.

Metrics can be made available to monitor by simply logging it using self.log('metric_to_track', metric_val) in your LightningModule.

Note

The frequency value specified in a dict along with the optimizer key is an int corresponding to the number of sequential batches optimized with the specific optimizer. It should be given to none or to all of the optimizers. There is a difference between passing multiple optimizers in a list, and passing multiple optimizers in dictionaries with a frequency of 1:

This is different from the frequency value specified in the lr_scheduler_config mentioned above.

def configure_optimizers(self):
    optimizer_one = torch.optim.SGD(self.model.parameters(), lr=0.01)
    optimizer_two = torch.optim.SGD(self.model.parameters(), lr=0.01)
    return [
        {"optimizer": optimizer_one, "frequency": 5},
        {"optimizer": optimizer_two, "frequency": 10},
    ]

In this example, the first optimizer will be used for the first 5 steps, the second optimizer for the next 10 steps and that cycle will continue. If an LR scheduler is specified for an optimizer using the lr_scheduler key in the above dict, the scheduler will only be updated when its optimizer is being used.

Examples:

# most cases. no learning rate scheduler
def configure_optimizers(self):
    return Adam(self.parameters(), lr=1e-3)

# multiple optimizer case (e.g.: GAN)
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    return gen_opt, dis_opt

# example with learning rate schedulers
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    dis_sch = CosineAnnealing(dis_opt, T_max=10)
    return [gen_opt, dis_opt], [dis_sch]

# example with step-based learning rate schedulers
# each optimizer has its own scheduler
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    gen_sch = {
        'scheduler': ExponentialLR(gen_opt, 0.99),
        'interval': 'step'  # called after each training step
    }
    dis_sch = CosineAnnealing(dis_opt, T_max=10) # called every epoch
    return [gen_opt, dis_opt], [gen_sch, dis_sch]

# example with optimizer frequencies
# see training procedure in `Improved Training of Wasserstein GANs`, Algorithm 1
# https://arxiv.org/abs/1704.00028
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    n_critic = 5
    return (
        {'optimizer': dis_opt, 'frequency': n_critic},
        {'optimizer': gen_opt, 'frequency': 1}
    )

Note

Some things to know:

Lightning calls .backward() and .step() on each optimizer and learning rate scheduler as needed.
If you use 16-bit precision (precision=16), Lightning will automatically handle the optimizers.
If you use multiple optimizers, training_step() will have an additional optimizer_idx parameter.
If you use torch.optim.LBFGS, Lightning handles the closure function automatically for you.
If you use multiple optimizers, gradients will be calculated only for the parameters of current optimizer at each training step.
If you need to control how often those optimizers step or override the default .step() schedule, override the optimizer_step() hook.

rl_parameters: reagent.core.parameters.RLParameters

train_step_gen(training_batch: reagent.core.types.DiscreteDqnInput, batch_idx: int): Implement training step as generator here

reagent.training.cem_trainer module

The Trainer for Cross-Entropy Method. The idea is that an ensemble of: world models are fitted to predict transitions and reward functions.

A cross entropy method-based planner will then plan the best next action based on simulation data generated by the fitted world models.

The idea is inspired by: https://arxiv.org/abs/1805.12114

class reagent.training.cem_trainer.CEMTrainer(cem_planner_network: reagent.models.cem_planner.CEMPlannerNetwork, world_model_trainers: List[reagent.training.world_model.mdnrnn_trainer.MDNRNNTrainer], parameters: reagent.core.parameters.CEMTrainerParameters)

Bases: reagent.training.reagent_lightning_module.ReAgentLightningModule

configure_optimizers()

Choose what optimizers and learning-rate schedulers to use in your optimization. Normally you’d need one. But in the case of GANs or similar you might have multiple.

Returns

Any of these 6 options.

Single optimizer.
List or Tuple of optimizers.
Two lists - The first list has multiple optimizers, and the second has multiple LR schedulers (or multiple lr_scheduler_config).
Dictionary, with an "optimizer" key, and (optionally) a "lr_scheduler" key whose value is a single LR scheduler or lr_scheduler_config.
Tuple of dictionaries as described above, with an optional "frequency" key.
None - Fit will run without any optimizer.

The lr_scheduler_config is a dictionary which contains the scheduler and its associated configuration. The default configuration is shown below.

lr_scheduler_config = {
    # REQUIRED: The scheduler instance
    "scheduler": lr_scheduler,
    # The unit of the scheduler's step size, could also be 'step'.
    # 'epoch' updates the scheduler on epoch end whereas 'step'
    # updates it after a optimizer update.
    "interval": "epoch",
    # How many epochs/steps should pass between calls to
    # `scheduler.step()`. 1 corresponds to updating the learning
    # rate after every epoch/step.
    "frequency": 1,
    # Metric to to monitor for schedulers like `ReduceLROnPlateau`
    "monitor": "val_loss",
    # If set to `True`, will enforce that the value specified 'monitor'
    # is available when the scheduler is updated, thus stopping
    # training if not found. If set to `False`, it will only produce a warning
    "strict": True,
    # If using the `LearningRateMonitor` callback to monitor the
    # learning rate progress, this keyword can be used to specify
    # a custom logged name
    "name": None,
}

When there are schedulers in which the .step() method is conditioned on a value, such as the torch.optim.lr_scheduler.ReduceLROnPlateau scheduler, Lightning requires that the lr_scheduler_config contains the keyword "monitor" set to the metric name that the scheduler should be conditioned on.

Metrics can be made available to monitor by simply logging it using self.log('metric_to_track', metric_val) in your LightningModule.

Note

The frequency value specified in a dict along with the optimizer key is an int corresponding to the number of sequential batches optimized with the specific optimizer. It should be given to none or to all of the optimizers. There is a difference between passing multiple optimizers in a list, and passing multiple optimizers in dictionaries with a frequency of 1:

This is different from the frequency value specified in the lr_scheduler_config mentioned above.

def configure_optimizers(self):
    optimizer_one = torch.optim.SGD(self.model.parameters(), lr=0.01)
    optimizer_two = torch.optim.SGD(self.model.parameters(), lr=0.01)
    return [
        {"optimizer": optimizer_one, "frequency": 5},
        {"optimizer": optimizer_two, "frequency": 10},
    ]

In this example, the first optimizer will be used for the first 5 steps, the second optimizer for the next 10 steps and that cycle will continue. If an LR scheduler is specified for an optimizer using the lr_scheduler key in the above dict, the scheduler will only be updated when its optimizer is being used.

Examples:

# most cases. no learning rate scheduler
def configure_optimizers(self):
    return Adam(self.parameters(), lr=1e-3)

# multiple optimizer case (e.g.: GAN)
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    return gen_opt, dis_opt

# example with learning rate schedulers
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    dis_sch = CosineAnnealing(dis_opt, T_max=10)
    return [gen_opt, dis_opt], [dis_sch]

# example with step-based learning rate schedulers
# each optimizer has its own scheduler
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    gen_sch = {
        'scheduler': ExponentialLR(gen_opt, 0.99),
        'interval': 'step'  # called after each training step
    }
    dis_sch = CosineAnnealing(dis_opt, T_max=10) # called every epoch
    return [gen_opt, dis_opt], [gen_sch, dis_sch]

# example with optimizer frequencies
# see training procedure in `Improved Training of Wasserstein GANs`, Algorithm 1
# https://arxiv.org/abs/1704.00028
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    n_critic = 5
    return (
        {'optimizer': dis_opt, 'frequency': n_critic},
        {'optimizer': gen_opt, 'frequency': 1}
    )

Note

Some things to know:

Lightning calls .backward() and .step() on each optimizer and learning rate scheduler as needed.
If you use 16-bit precision (precision=16), Lightning will automatically handle the optimizers.
If you use multiple optimizers, training_step() will have an additional optimizer_idx parameter.
If you use torch.optim.LBFGS, Lightning handles the closure function automatically for you.
If you use multiple optimizers, gradients will be calculated only for the parameters of current optimizer at each training step.
If you need to control how often those optimizers step or override the default .step() schedule, override the optimizer_step() hook.

train_step_gen(training_batch: reagent.core.types.MemoryNetworkInput, batch_idx: int): Implement training step as generator here

training: bool

reagent.training.cem_trainer.print_mdnrnn_losses(minibatch, model_index, losses) → None

reagent.training.discrete_crr_trainer module

class reagent.training.discrete_crr_trainer.DiscreteCRRTrainer(actor_network, actor_network_target, q1_network, q1_network_target, reward_network, q2_network=None, q2_network_target=None, q_network_cpe=None, q_network_cpe_target=None, metrics_to_score=None, evaluation: reagent.core.parameters.EvaluationParameters = Field(name=None,type=None,default=<dataclasses._MISSING_TYPE object>,default_factory=<class 'reagent.core.parameters.EvaluationParameters'>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=None), rl: reagent.core.parameters.RLParameters = Field(name='rl',type=<class 'reagent.core.parameters.RLParameters'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<class 'reagent.core.parameters.RLParameters'>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), double_q_learning: bool = True, q_network_optimizer: reagent.optimizer.union.Optimizer__Union = Field(name='q_network_optimizer',type=<class 'reagent.optimizer.union.Optimizer__Union'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<bound method Optimizer__Union.default of <class 'reagent.optimizer.union.Optimizer__Union'>>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), actor_network_optimizer: reagent.optimizer.union.Optimizer__Union = Field(name='actor_network_optimizer',type=<class 'reagent.optimizer.union.Optimizer__Union'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<bound method Optimizer__Union.default of <class 'reagent.optimizer.union.Optimizer__Union'>>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), use_target_actor: bool = False, actions: List[str] = Field(name='actions',type=typing.List[str],default=<dataclasses._MISSING_TYPE object>,default_factory=<class 'list'>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), delayed_policy_update: int = 1, beta: float = 1.0, entropy_coeff: float = 0.0, clip_limit: float = 10.0, max_weight: float = 20.0)

Bases: reagent.training.dqn_trainer_base.DQNTrainerBaseLightning

Critic Regularized Regression (CRR) algorithm trainer as described in https://arxiv.org/abs/2006.15134

compute_actor_loss(batch_idx, action, logged_action_probs, all_q_values, all_action_scores)

compute_target_q_values(next_state, rewards, not_terminal, next_q_values)

compute_td_loss(q_network, state, action, target_q_values)

configure_optimizers()

Choose what optimizers and learning-rate schedulers to use in your optimization. Normally you’d need one. But in the case of GANs or similar you might have multiple.

Returns

Any of these 6 options.

Single optimizer.
List or Tuple of optimizers.
Two lists - The first list has multiple optimizers, and the second has multiple LR schedulers (or multiple lr_scheduler_config).
Dictionary, with an "optimizer" key, and (optionally) a "lr_scheduler" key whose value is a single LR scheduler or lr_scheduler_config.
Tuple of dictionaries as described above, with an optional "frequency" key.
None - Fit will run without any optimizer.

The lr_scheduler_config is a dictionary which contains the scheduler and its associated configuration. The default configuration is shown below.

lr_scheduler_config = {
    # REQUIRED: The scheduler instance
    "scheduler": lr_scheduler,
    # The unit of the scheduler's step size, could also be 'step'.
    # 'epoch' updates the scheduler on epoch end whereas 'step'
    # updates it after a optimizer update.
    "interval": "epoch",
    # How many epochs/steps should pass between calls to
    # `scheduler.step()`. 1 corresponds to updating the learning
    # rate after every epoch/step.
    "frequency": 1,
    # Metric to to monitor for schedulers like `ReduceLROnPlateau`
    "monitor": "val_loss",
    # If set to `True`, will enforce that the value specified 'monitor'
    # is available when the scheduler is updated, thus stopping
    # training if not found. If set to `False`, it will only produce a warning
    "strict": True,
    # If using the `LearningRateMonitor` callback to monitor the
    # learning rate progress, this keyword can be used to specify
    # a custom logged name
    "name": None,
}

When there are schedulers in which the .step() method is conditioned on a value, such as the torch.optim.lr_scheduler.ReduceLROnPlateau scheduler, Lightning requires that the lr_scheduler_config contains the keyword "monitor" set to the metric name that the scheduler should be conditioned on.

Metrics can be made available to monitor by simply logging it using self.log('metric_to_track', metric_val) in your LightningModule.

Note

The frequency value specified in a dict along with the optimizer key is an int corresponding to the number of sequential batches optimized with the specific optimizer. It should be given to none or to all of the optimizers. There is a difference between passing multiple optimizers in a list, and passing multiple optimizers in dictionaries with a frequency of 1:

This is different from the frequency value specified in the lr_scheduler_config mentioned above.

def configure_optimizers(self):
    optimizer_one = torch.optim.SGD(self.model.parameters(), lr=0.01)
    optimizer_two = torch.optim.SGD(self.model.parameters(), lr=0.01)
    return [
        {"optimizer": optimizer_one, "frequency": 5},
        {"optimizer": optimizer_two, "frequency": 10},
    ]

In this example, the first optimizer will be used for the first 5 steps, the second optimizer for the next 10 steps and that cycle will continue. If an LR scheduler is specified for an optimizer using the lr_scheduler key in the above dict, the scheduler will only be updated when its optimizer is being used.

Examples:

# most cases. no learning rate scheduler
def configure_optimizers(self):
    return Adam(self.parameters(), lr=1e-3)

# multiple optimizer case (e.g.: GAN)
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    return gen_opt, dis_opt

# example with learning rate schedulers
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    dis_sch = CosineAnnealing(dis_opt, T_max=10)
    return [gen_opt, dis_opt], [dis_sch]

# example with step-based learning rate schedulers
# each optimizer has its own scheduler
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    gen_sch = {
        'scheduler': ExponentialLR(gen_opt, 0.99),
        'interval': 'step'  # called after each training step
    }
    dis_sch = CosineAnnealing(dis_opt, T_max=10) # called every epoch
    return [gen_opt, dis_opt], [gen_sch, dis_sch]

# example with optimizer frequencies
# see training procedure in `Improved Training of Wasserstein GANs`, Algorithm 1
# https://arxiv.org/abs/1704.00028
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    n_critic = 5
    return (
        {'optimizer': dis_opt, 'frequency': n_critic},
        {'optimizer': gen_opt, 'frequency': 1}
    )

Note

Some things to know:

Lightning calls .backward() and .step() on each optimizer and learning rate scheduler as needed.
If you use 16-bit precision (precision=16), Lightning will automatically handle the optimizers.
If you use multiple optimizers, training_step() will have an additional optimizer_idx parameter.
If you use torch.optim.LBFGS, Lightning handles the closure function automatically for you.
If you use multiple optimizers, gradients will be calculated only for the parameters of current optimizer at each training step.
If you need to control how often those optimizers step or override the default .step() schedule, override the optimizer_step() hook.

get_detached_model_outputs(state) → Tuple[torch.Tensor, None]

property q_network

rl_parameters: reagent.core.parameters.RLParameters

train_step_gen(training_batch: reagent.core.types.DiscreteDqnInput, batch_idx: int): IMPORTANT: the input action here is preprocessed according to the training_batch type, which in this case is DiscreteDqnInput. Hence, the preprocessor in the DiscreteDqnInputMaker class in the trainer_preprocessor.py is used, which converts acion taken to a one-hot representation.

validation_step(batch, batch_idx)

Operates on a single batch of data from the validation set. In this step you’d might generate examples or calculate anything of interest like accuracy.

# the pseudocode for these calls
val_outs = []
for val_batch in val_data:
    out = validation_step(val_batch)
    val_outs.append(out)
validation_epoch_end(val_outs)

Parameters

batch (Tensor | (Tensor, …) | [Tensor, …]) – The output of your DataLoader. A tensor, tuple or list.
batch_idx (int) – The index of this batch
dataloader_idx (int) – The index of the dataloader that produced this batch (only if multiple val dataloaders used)

Returns

Any object or value
None - Validation will skip to the next batch

# pseudocode of order
val_outs = []
for val_batch in val_data:
    out = validation_step(val_batch)
    if defined("validation_step_end"):
        out = validation_step_end(out)
    val_outs.append(out)
val_outs = validation_epoch_end(val_outs)

# if you have one val dataloader:
def validation_step(self, batch, batch_idx):
    ...


# if you have multiple val dataloaders:
def validation_step(self, batch, batch_idx, dataloader_idx):
    ...

Examples:

# CASE 1: A single validation dataset
def validation_step(self, batch, batch_idx):
    x, y = batch

    # implement your own
    out = self(x)
    loss = self.loss(out, y)

    # log 6 example images
    # or generated text... or whatever
    sample_imgs = x[:6]
    grid = torchvision.utils.make_grid(sample_imgs)
    self.logger.experiment.add_image('example_images', grid, 0)

    # calculate acc
    labels_hat = torch.argmax(out, dim=1)
    val_acc = torch.sum(y == labels_hat).item() / (len(y) * 1.0)

    # log the outputs!
    self.log_dict({'val_loss': loss, 'val_acc': val_acc})

If you pass in multiple val dataloaders, validation_step() will have an additional argument.

# CASE 2: multiple validation dataloaders
def validation_step(self, batch, batch_idx, dataloader_idx):
    # dataloader_idx tells you which dataset this is.
    ...

Note

If you don’t need to validate you don’t need to implement this method.

Note

When the validation_step() is called, the model has been put in eval mode and PyTorch gradients have been disabled. At the end of validation, the model goes back to training mode and gradients are enabled.

reagent.training.dqn_trainer module

class reagent.training.dqn_trainer.BCQConfig(drop_threshold: float = 0.1)

Bases: object

drop_threshold: float = 0.1

class reagent.training.dqn_trainer.DQNTrainer(q_network, q_network_target, reward_network, q_network_cpe=None, q_network_cpe_target=None, metrics_to_score=None, evaluation: reagent.core.parameters.EvaluationParameters = Field(name=None,type=None,default=<dataclasses._MISSING_TYPE object>,default_factory=<class 'reagent.core.parameters.EvaluationParameters'>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=None), imitator=None, actions: List[str] = Field(name='actions',type=typing.List[str],default=<dataclasses._MISSING_TYPE object>,default_factory=<class 'list'>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), rl: reagent.core.parameters.RLParameters = Field(name='rl',type=<class 'reagent.core.parameters.RLParameters'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<class 'reagent.core.parameters.RLParameters'>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), double_q_learning: bool = True, bcq: Optional[reagent.training.dqn_trainer.BCQConfig] = None, minibatch_size: int = 1024, minibatches_per_step: int = 1, optimizer: reagent.optimizer.union.Optimizer__Union = Field(name='optimizer',type=<class 'reagent.optimizer.union.Optimizer__Union'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<bound method Optimizer__Union.default of <class 'reagent.optimizer.union.Optimizer__Union'>>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD))

Bases: reagent.training.dqn_trainer_base.DQNTrainerBaseLightning

compute_discount_tensor(batch: reagent.core.types.DiscreteDqnInput, boosted_rewards: torch.Tensor)

compute_td_loss(batch: reagent.core.types.DiscreteDqnInput, boosted_rewards: torch.Tensor, discount_tensor: torch.Tensor)

configure_optimizers()

Choose what optimizers and learning-rate schedulers to use in your optimization. Normally you’d need one. But in the case of GANs or similar you might have multiple.

Returns

Any of these 6 options.

Single optimizer.
List or Tuple of optimizers.
Two lists - The first list has multiple optimizers, and the second has multiple LR schedulers (or multiple lr_scheduler_config).
Dictionary, with an "optimizer" key, and (optionally) a "lr_scheduler" key whose value is a single LR scheduler or lr_scheduler_config.
Tuple of dictionaries as described above, with an optional "frequency" key.
None - Fit will run without any optimizer.

The lr_scheduler_config is a dictionary which contains the scheduler and its associated configuration. The default configuration is shown below.

lr_scheduler_config = {
    # REQUIRED: The scheduler instance
    "scheduler": lr_scheduler,
    # The unit of the scheduler's step size, could also be 'step'.
    # 'epoch' updates the scheduler on epoch end whereas 'step'
    # updates it after a optimizer update.
    "interval": "epoch",
    # How many epochs/steps should pass between calls to
    # `scheduler.step()`. 1 corresponds to updating the learning
    # rate after every epoch/step.
    "frequency": 1,
    # Metric to to monitor for schedulers like `ReduceLROnPlateau`
    "monitor": "val_loss",
    # If set to `True`, will enforce that the value specified 'monitor'
    # is available when the scheduler is updated, thus stopping
    # training if not found. If set to `False`, it will only produce a warning
    "strict": True,
    # If using the `LearningRateMonitor` callback to monitor the
    # learning rate progress, this keyword can be used to specify
    # a custom logged name
    "name": None,
}

When there are schedulers in which the .step() method is conditioned on a value, such as the torch.optim.lr_scheduler.ReduceLROnPlateau scheduler, Lightning requires that the lr_scheduler_config contains the keyword "monitor" set to the metric name that the scheduler should be conditioned on.

Metrics can be made available to monitor by simply logging it using self.log('metric_to_track', metric_val) in your LightningModule.

Note

The frequency value specified in a dict along with the optimizer key is an int corresponding to the number of sequential batches optimized with the specific optimizer. It should be given to none or to all of the optimizers. There is a difference between passing multiple optimizers in a list, and passing multiple optimizers in dictionaries with a frequency of 1:

This is different from the frequency value specified in the lr_scheduler_config mentioned above.

def configure_optimizers(self):
    optimizer_one = torch.optim.SGD(self.model.parameters(), lr=0.01)
    optimizer_two = torch.optim.SGD(self.model.parameters(), lr=0.01)
    return [
        {"optimizer": optimizer_one, "frequency": 5},
        {"optimizer": optimizer_two, "frequency": 10},
    ]

In this example, the first optimizer will be used for the first 5 steps, the second optimizer for the next 10 steps and that cycle will continue. If an LR scheduler is specified for an optimizer using the lr_scheduler key in the above dict, the scheduler will only be updated when its optimizer is being used.

Examples:

# most cases. no learning rate scheduler
def configure_optimizers(self):
    return Adam(self.parameters(), lr=1e-3)

# multiple optimizer case (e.g.: GAN)
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    return gen_opt, dis_opt

# example with learning rate schedulers
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    dis_sch = CosineAnnealing(dis_opt, T_max=10)
    return [gen_opt, dis_opt], [dis_sch]

# example with step-based learning rate schedulers
# each optimizer has its own scheduler
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    gen_sch = {
        'scheduler': ExponentialLR(gen_opt, 0.99),
        'interval': 'step'  # called after each training step
    }
    dis_sch = CosineAnnealing(dis_opt, T_max=10) # called every epoch
    return [gen_opt, dis_opt], [gen_sch, dis_sch]

# example with optimizer frequencies
# see training procedure in `Improved Training of Wasserstein GANs`, Algorithm 1
# https://arxiv.org/abs/1704.00028
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    n_critic = 5
    return (
        {'optimizer': dis_opt, 'frequency': n_critic},
        {'optimizer': gen_opt, 'frequency': 1}
    )

Note

Some things to know:

Lightning calls .backward() and .step() on each optimizer and learning rate scheduler as needed.
If you use 16-bit precision (precision=16), Lightning will automatically handle the optimizers.
If you use multiple optimizers, training_step() will have an additional optimizer_idx parameter.
If you use torch.optim.LBFGS, Lightning handles the closure function automatically for you.
If you use multiple optimizers, gradients will be calculated only for the parameters of current optimizer at each training step.
If you need to control how often those optimizers step or override the default .step() schedule, override the optimizer_step() hook.

get_detached_model_outputs(state) → Tuple[torch.Tensor, Optional[torch.Tensor]]: Gets the q values from the model and target networks

rl_parameters: reagent.core.parameters.RLParameters

train_step_gen(training_batch: reagent.core.types.DiscreteDqnInput, batch_idx: int): Implement training step as generator here

validation_step(batch, batch_idx)

Operates on a single batch of data from the validation set. In this step you’d might generate examples or calculate anything of interest like accuracy.

# the pseudocode for these calls
val_outs = []
for val_batch in val_data:
    out = validation_step(val_batch)
    val_outs.append(out)
validation_epoch_end(val_outs)

Parameters

batch (Tensor | (Tensor, …) | [Tensor, …]) – The output of your DataLoader. A tensor, tuple or list.
batch_idx (int) – The index of this batch
dataloader_idx (int) – The index of the dataloader that produced this batch (only if multiple val dataloaders used)

Returns

Any object or value
None - Validation will skip to the next batch

# pseudocode of order
val_outs = []
for val_batch in val_data:
    out = validation_step(val_batch)
    if defined("validation_step_end"):
        out = validation_step_end(out)
    val_outs.append(out)
val_outs = validation_epoch_end(val_outs)

# if you have one val dataloader:
def validation_step(self, batch, batch_idx):
    ...


# if you have multiple val dataloaders:
def validation_step(self, batch, batch_idx, dataloader_idx):
    ...

Examples:

# CASE 1: A single validation dataset
def validation_step(self, batch, batch_idx):
    x, y = batch

    # implement your own
    out = self(x)
    loss = self.loss(out, y)

    # log 6 example images
    # or generated text... or whatever
    sample_imgs = x[:6]
    grid = torchvision.utils.make_grid(sample_imgs)
    self.logger.experiment.add_image('example_images', grid, 0)

    # calculate acc
    labels_hat = torch.argmax(out, dim=1)
    val_acc = torch.sum(y == labels_hat).item() / (len(y) * 1.0)

    # log the outputs!
    self.log_dict({'val_loss': loss, 'val_acc': val_acc})

If you pass in multiple val dataloaders, validation_step() will have an additional argument.

# CASE 2: multiple validation dataloaders
def validation_step(self, batch, batch_idx, dataloader_idx):
    # dataloader_idx tells you which dataset this is.
    ...

Note

If you don’t need to validate you don’t need to implement this method.

Note

When the validation_step() is called, the model has been put in eval mode and PyTorch gradients have been disabled. At the end of validation, the model goes back to training mode and gradients are enabled.

reagent.training.dqn_trainer_base module

class reagent.training.dqn_trainer_base.DQNTrainerBaseLightning(rl_parameters: reagent.core.parameters.RLParameters, metrics_to_score=None, actions: Optional[List[str]] = None, evaluation_parameters: Optional[reagent.core.parameters.EvaluationParameters] = None)

Bases: reagent.training.dqn_trainer_base.DQNTrainerMixin, reagent.training.rl_trainer_pytorch.RLTrainerMixin, reagent.training.reagent_lightning_module.ReAgentLightningModule

boost_rewards(rewards: torch.Tensor, actions: torch.Tensor) → torch.Tensor

gather_eval_data(validation_step_outputs)

property num_actions: int

rl_parameters: reagent.core.parameters.RLParameters

validation_epoch_end(valid_step_outputs)

Called at the end of the validation epoch with the outputs of all validation steps.

# the pseudocode for these calls
val_outs = []
for val_batch in val_data:
    out = validation_step(val_batch)
    val_outs.append(out)
validation_epoch_end(val_outs)

Parameters: outputs – List of outputs you defined in validation_step(), or if there are multiple dataloaders, a list containing a list of outputs for each dataloader.
Returns: None

Note

If you didn’t define a validation_step(), this won’t be called.

Examples

With a single dataloader:

def validation_epoch_end(self, val_step_outputs):
    for out in val_step_outputs:
        ...

With multiple dataloaders, outputs will be a list of lists. The outer list contains one entry per dataloader, while the inner list contains the individual outputs of each validation step for that dataloader.

def validation_epoch_end(self, outputs):
    for dataloader_output_result in outputs:
        dataloader_outs = dataloader_output_result.dataloader_i_outputs

    self.log("final_metric", final_value)

validation_step(batch, batch_idx)

Operates on a single batch of data from the validation set. In this step you’d might generate examples or calculate anything of interest like accuracy.

# the pseudocode for these calls
val_outs = []
for val_batch in val_data:
    out = validation_step(val_batch)
    val_outs.append(out)
validation_epoch_end(val_outs)

Parameters

batch (Tensor | (Tensor, …) | [Tensor, …]) – The output of your DataLoader. A tensor, tuple or list.
batch_idx (int) – The index of this batch
dataloader_idx (int) – The index of the dataloader that produced this batch (only if multiple val dataloaders used)

Returns

Any object or value
None - Validation will skip to the next batch

# pseudocode of order
val_outs = []
for val_batch in val_data:
    out = validation_step(val_batch)
    if defined("validation_step_end"):
        out = validation_step_end(out)
    val_outs.append(out)
val_outs = validation_epoch_end(val_outs)

# if you have one val dataloader:
def validation_step(self, batch, batch_idx):
    ...


# if you have multiple val dataloaders:
def validation_step(self, batch, batch_idx, dataloader_idx):
    ...

Examples:

# CASE 1: A single validation dataset
def validation_step(self, batch, batch_idx):
    x, y = batch

    # implement your own
    out = self(x)
    loss = self.loss(out, y)

    # log 6 example images
    # or generated text... or whatever
    sample_imgs = x[:6]
    grid = torchvision.utils.make_grid(sample_imgs)
    self.logger.experiment.add_image('example_images', grid, 0)

    # calculate acc
    labels_hat = torch.argmax(out, dim=1)
    val_acc = torch.sum(y == labels_hat).item() / (len(y) * 1.0)

    # log the outputs!
    self.log_dict({'val_loss': loss, 'val_acc': val_acc})

If you pass in multiple val dataloaders, validation_step() will have an additional argument.

# CASE 2: multiple validation dataloaders
def validation_step(self, batch, batch_idx, dataloader_idx):
    # dataloader_idx tells you which dataset this is.
    ...

Note

If you don’t need to validate you don’t need to implement this method.

Note

When the validation_step() is called, the model has been put in eval mode and PyTorch gradients have been disabled. At the end of validation, the model goes back to training mode and gradients are enabled.

class reagent.training.dqn_trainer_base.DQNTrainerMixin

Bases: object

ACTION_NOT_POSSIBLE_VAL = -1000000000.0

get_max_q_values(q_values, possible_actions_mask)

get_max_q_values_with_target(q_values, q_values_target, possible_actions_mask)

Used in Q-learning update.

Parameters

q_values – PyTorch tensor with shape (batch_size, action_dim). Each row contains the list of Q-values for each possible action in this state.
q_values_target – PyTorch tensor with shape (batch_size, action_dim). Each row contains the list of Q-values from the target network for each possible action in this state.
possible_actions_mask – PyTorch tensor with shape (batch_size, action_dim). possible_actions[i][j] = 1 iff the agent can take action j from state i.

Returns a tensor of maximum Q-values for every state in the batch: and also the index of the corresponding action (which is used in evaluation_data_page.py, in create_from_tensors_dqn()).

reagent.training.imitator_training module

reagent.training.imitator_training.get_valid_actions_from_imitator(imitator, input, drop_threshold): Create mask for non-viable actions under the imitator.

reagent.training.multi_stage_trainer module

class reagent.training.multi_stage_trainer.MultiStageTrainer(trainers: List[reagent.training.reagent_lightning_module.ReAgentLightningModule], epochs: List[int], assign_reporter_function=None, flush_reporter_function=None, automatic_optimization=True)

Bases: reagent.training.reagent_lightning_module.ReAgentLightningModule

configure_optimizers()

Choose what optimizers and learning-rate schedulers to use in your optimization. Normally you’d need one. But in the case of GANs or similar you might have multiple.

Returns

Any of these 6 options.

Single optimizer.
List or Tuple of optimizers.
Two lists - The first list has multiple optimizers, and the second has multiple LR schedulers (or multiple lr_scheduler_config).
Dictionary, with an "optimizer" key, and (optionally) a "lr_scheduler" key whose value is a single LR scheduler or lr_scheduler_config.
Tuple of dictionaries as described above, with an optional "frequency" key.
None - Fit will run without any optimizer.

The lr_scheduler_config is a dictionary which contains the scheduler and its associated configuration. The default configuration is shown below.

lr_scheduler_config = {
    # REQUIRED: The scheduler instance
    "scheduler": lr_scheduler,
    # The unit of the scheduler's step size, could also be 'step'.
    # 'epoch' updates the scheduler on epoch end whereas 'step'
    # updates it after a optimizer update.
    "interval": "epoch",
    # How many epochs/steps should pass between calls to
    # `scheduler.step()`. 1 corresponds to updating the learning
    # rate after every epoch/step.
    "frequency": 1,
    # Metric to to monitor for schedulers like `ReduceLROnPlateau`
    "monitor": "val_loss",
    # If set to `True`, will enforce that the value specified 'monitor'
    # is available when the scheduler is updated, thus stopping
    # training if not found. If set to `False`, it will only produce a warning
    "strict": True,
    # If using the `LearningRateMonitor` callback to monitor the
    # learning rate progress, this keyword can be used to specify
    # a custom logged name
    "name": None,
}

When there are schedulers in which the .step() method is conditioned on a value, such as the torch.optim.lr_scheduler.ReduceLROnPlateau scheduler, Lightning requires that the lr_scheduler_config contains the keyword "monitor" set to the metric name that the scheduler should be conditioned on.

Metrics can be made available to monitor by simply logging it using self.log('metric_to_track', metric_val) in your LightningModule.

Note

The frequency value specified in a dict along with the optimizer key is an int corresponding to the number of sequential batches optimized with the specific optimizer. It should be given to none or to all of the optimizers. There is a difference between passing multiple optimizers in a list, and passing multiple optimizers in dictionaries with a frequency of 1:

This is different from the frequency value specified in the lr_scheduler_config mentioned above.

def configure_optimizers(self):
    optimizer_one = torch.optim.SGD(self.model.parameters(), lr=0.01)
    optimizer_two = torch.optim.SGD(self.model.parameters(), lr=0.01)
    return [
        {"optimizer": optimizer_one, "frequency": 5},
        {"optimizer": optimizer_two, "frequency": 10},
    ]

In this example, the first optimizer will be used for the first 5 steps, the second optimizer for the next 10 steps and that cycle will continue. If an LR scheduler is specified for an optimizer using the lr_scheduler key in the above dict, the scheduler will only be updated when its optimizer is being used.

Examples:

# most cases. no learning rate scheduler
def configure_optimizers(self):
    return Adam(self.parameters(), lr=1e-3)

# multiple optimizer case (e.g.: GAN)
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    return gen_opt, dis_opt

# example with learning rate schedulers
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    dis_sch = CosineAnnealing(dis_opt, T_max=10)
    return [gen_opt, dis_opt], [dis_sch]

# example with step-based learning rate schedulers
# each optimizer has its own scheduler
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    gen_sch = {
        'scheduler': ExponentialLR(gen_opt, 0.99),
        'interval': 'step'  # called after each training step
    }
    dis_sch = CosineAnnealing(dis_opt, T_max=10) # called every epoch
    return [gen_opt, dis_opt], [gen_sch, dis_sch]

# example with optimizer frequencies
# see training procedure in `Improved Training of Wasserstein GANs`, Algorithm 1
# https://arxiv.org/abs/1704.00028
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    n_critic = 5
    return (
        {'optimizer': dis_opt, 'frequency': n_critic},
        {'optimizer': gen_opt, 'frequency': 1}
    )

Note

Some things to know:

Lightning calls .backward() and .step() on each optimizer and learning rate scheduler as needed.
If you use 16-bit precision (precision=16), Lightning will automatically handle the optimizers.
If you use multiple optimizers, training_step() will have an additional optimizer_idx parameter.
If you use torch.optim.LBFGS, Lightning handles the closure function automatically for you.
If you use multiple optimizers, gradients will be calculated only for the parameters of current optimizer at each training step.
If you need to control how often those optimizers step or override the default .step() schedule, override the optimizer_step() hook.

property multi_stage_total_epochs

on_fit_end()

Called at the very end of fit.

If on DDP it is called on every process

on_fit_start()

Called at the very beginning of fit.

If on DDP it is called on every process

on_test_end(): Called at the end of testing.

on_test_start(): Called at the beginning of testing.

optimizer_step(epoch: int, batch_idx: int, optimizer, optimizer_idx: int, optimizer_closure, on_tpu: int = False, using_native_amp: int = False, using_lbfgs: int = False)

Override this method to adjust the default way the Trainer calls each optimizer. By default, Lightning calls step() and zero_grad() as shown in the example once per optimizer. This method (and zero_grad()) won’t be called during the accumulation phase when Trainer(accumulate_grad_batches != 1).

Parameters

epoch – Current epoch
batch_idx – Index of current batch
optimizer – A PyTorch optimizer
optimizer_idx – If you used multiple optimizers, this indexes into that list.
optimizer_closure – Closure for all optimizers. This closure must be executed as it includes the calls to training_step(), optimizer.zero_grad(), and backward().
on_tpu – True if TPU backward is required
using_native_amp – True if using native amp
using_lbfgs – True if the matching optimizer is torch.optim.LBFGS

Examples:

# DEFAULT
def optimizer_step(self, epoch, batch_idx, optimizer, optimizer_idx,
                   optimizer_closure, on_tpu, using_native_amp, using_lbfgs):
    optimizer.step(closure=optimizer_closure)

# Alternating schedule for optimizer steps (i.e.: GANs)
def optimizer_step(self, epoch, batch_idx, optimizer, optimizer_idx,
                   optimizer_closure, on_tpu, using_native_amp, using_lbfgs):
    # update generator opt every step
    if optimizer_idx == 0:
        optimizer.step(closure=optimizer_closure)

    # update discriminator opt every 2 steps
    if optimizer_idx == 1:
        if (batch_idx + 1) % 2 == 0 :
            optimizer.step(closure=optimizer_closure)
        else:
            # call the closure by itself to run `training_step` + `backward` without an optimizer step
            optimizer_closure()

    # ...
    # add as many optimizers as you want

Here’s another example showing how to use this for more advanced things such as learning rate warm-up:

# learning rate warm-up
def optimizer_step(
    self,
    epoch,
    batch_idx,
    optimizer,
    optimizer_idx,
    optimizer_closure,
    on_tpu,
    using_native_amp,
    using_lbfgs,
):
    # warm up lr
    if self.trainer.global_step < 500:
        lr_scale = min(1.0, float(self.trainer.global_step + 1) / 500.0)
        for pg in optimizer.param_groups:
            pg["lr"] = lr_scale * self.learning_rate

    # update params
    optimizer.step(closure=optimizer_closure)

set_reporter(reporter)

test_epoch_end(outputs)

Called at the end of a test epoch with the output of all test steps.

# the pseudocode for these calls
test_outs = []
for test_batch in test_data:
    out = test_step(test_batch)
    test_outs.append(out)
test_epoch_end(test_outs)

Parameters: outputs – List of outputs you defined in test_step_end(), or if there are multiple dataloaders, a list containing a list of outputs for each dataloader
Returns: None

Note

If you didn’t define a test_step(), this won’t be called.

Examples

With a single dataloader:

def test_epoch_end(self, outputs):
    # do something with the outputs of all test batches
    all_test_preds = test_step_outputs.predictions

    some_result = calc_all_results(all_test_preds)
    self.log(some_result)

With multiple dataloaders, outputs will be a list of lists. The outer list contains one entry per dataloader, while the inner list contains the individual outputs of each test step for that dataloader.

def test_epoch_end(self, outputs):
    final_value = 0
    for dataloader_outputs in outputs:
        for test_step_out in dataloader_outputs:
            # do something
            final_value += test_step_out

    self.log("final_metric", final_value)

test_step(*args, **kwargs)

Operates on a single batch of data from the test set. In this step you’d normally generate examples or calculate anything of interest such as accuracy.

# the pseudocode for these calls
test_outs = []
for test_batch in test_data:
    out = test_step(test_batch)
    test_outs.append(out)
test_epoch_end(test_outs)

Parameters

batch (Tensor | (Tensor, …) | [Tensor, …]) – The output of your DataLoader. A tensor, tuple or list.
batch_idx (int) – The index of this batch.
dataloader_idx (int) – The index of the dataloader that produced this batch (only if multiple test dataloaders used).

Returns

Any of.

# if you have one test dataloader:
def test_step(self, batch, batch_idx):
    ...


# if you have multiple test dataloaders:
def test_step(self, batch, batch_idx, dataloader_idx):
    ...

Examples:

# CASE 1: A single test dataset
def test_step(self, batch, batch_idx):
    x, y = batch

    # implement your own
    out = self(x)
    loss = self.loss(out, y)

    # log 6 example images
    # or generated text... or whatever
    sample_imgs = x[:6]
    grid = torchvision.utils.make_grid(sample_imgs)
    self.logger.experiment.add_image('example_images', grid, 0)

    # calculate acc
    labels_hat = torch.argmax(out, dim=1)
    test_acc = torch.sum(y == labels_hat).item() / (len(y) * 1.0)

    # log the outputs!
    self.log_dict({'test_loss': loss, 'test_acc': test_acc})

If you pass in multiple test dataloaders, test_step() will have an additional argument.

# CASE 2: multiple test dataloaders
def test_step(self, batch, batch_idx, dataloader_idx):
    # dataloader_idx tells you which dataset this is.
    ...

Note

If you don’t need to test you don’t need to implement this method.

Note

When the test_step() is called, the model has been put in eval mode and PyTorch gradients have been disabled. At the end of the test epoch, the model goes back to training mode and gradients are enabled.

training: bool

training_epoch_end(outputs)

Called at the end of the training epoch with the outputs of all training steps. Use this in case you need to do something with all the outputs returned by training_step().

# the pseudocode for these calls
train_outs = []
for train_batch in train_data:
    out = training_step(train_batch)
    train_outs.append(out)
training_epoch_end(train_outs)

Parameters: outputs – List of outputs you defined in training_step(). If there are multiple optimizers, it is a list containing a list of outputs for each optimizer. If using truncated_bptt_steps > 1, each element is a list of outputs corresponding to the outputs of each processed split batch.
Returns: None

Note

If this method is not overridden, this won’t be called.

def training_epoch_end(self, training_step_outputs):
    # do something with all training_step outputs
    for out in training_step_outputs:
        ...

training_step(batch, batch_idx: int, optimizer_idx: int = 0)

Here you compute and return the training loss and some additional metrics for e.g. the progress bar or logger.

Parameters

batch (Tensor | (Tensor, …) | [Tensor, …]) – The output of your DataLoader. A tensor, tuple or list.
batch_idx (int) – Integer displaying index of this batch
optimizer_idx (int) – When using multiple optimizers, this argument will also be present.
hiddens (Any) – Passed in if :paramref:`~pytorch_lightning.core.lightning.LightningModule.truncated_bptt_steps` > 0.

Returns

Any of.

Tensor - The loss tensor
dict - A dictionary. Can include any keys, but must include the key 'loss'
None - Training will skip to the next batch. This is only for automatic optimization.
This is not supported for multi-GPU, TPU, IPU, or DeepSpeed.

In this step you’d normally do the forward pass and calculate the loss for a batch. You can also do fancier things like multiple forward passes or something model specific.

Example:

def training_step(self, batch, batch_idx):
    x, y, z = batch
    out = self.encoder(x)
    loss = self.loss(out, x)
    return loss

If you define multiple optimizers, this step will be called with an additional optimizer_idx parameter.

# Multiple optimizers (e.g.: GANs)
def training_step(self, batch, batch_idx, optimizer_idx):
    if optimizer_idx == 0:
        # do training_step with encoder
        ...
    if optimizer_idx == 1:
        # do training_step with decoder
        ...

If you add truncated back propagation through time you will also get an additional argument with the hidden states of the previous step.

# Truncated back-propagation through time
def training_step(self, batch, batch_idx, hiddens):
    # hiddens are the hidden states from the previous truncated backprop step
    out, hiddens = self.lstm(data, hiddens)
    loss = ...
    return {"loss": loss, "hiddens": hiddens}

Note

The loss value shown in the progress bar is smoothed (averaged) over the last values, so it differs from the actual loss returned in train/validation step.

validation_epoch_end(outputs)

Called at the end of the validation epoch with the outputs of all validation steps.

# the pseudocode for these calls
val_outs = []
for val_batch in val_data:
    out = validation_step(val_batch)
    val_outs.append(out)
validation_epoch_end(val_outs)

Parameters: outputs – List of outputs you defined in validation_step(), or if there are multiple dataloaders, a list containing a list of outputs for each dataloader.
Returns: None

Note

If you didn’t define a validation_step(), this won’t be called.

Examples

With a single dataloader:

def validation_epoch_end(self, val_step_outputs):
    for out in val_step_outputs:
        ...

With multiple dataloaders, outputs will be a list of lists. The outer list contains one entry per dataloader, while the inner list contains the individual outputs of each validation step for that dataloader.

def validation_epoch_end(self, outputs):
    for dataloader_output_result in outputs:
        dataloader_outs = dataloader_output_result.dataloader_i_outputs

    self.log("final_metric", final_value)

validation_step(*args, **kwargs)

Operates on a single batch of data from the validation set. In this step you’d might generate examples or calculate anything of interest like accuracy.

# the pseudocode for these calls
val_outs = []
for val_batch in val_data:
    out = validation_step(val_batch)
    val_outs.append(out)
validation_epoch_end(val_outs)

Parameters

batch (Tensor | (Tensor, …) | [Tensor, …]) – The output of your DataLoader. A tensor, tuple or list.
batch_idx (int) – The index of this batch
dataloader_idx (int) – The index of the dataloader that produced this batch (only if multiple val dataloaders used)

Returns

Any object or value
None - Validation will skip to the next batch

# pseudocode of order
val_outs = []
for val_batch in val_data:
    out = validation_step(val_batch)
    if defined("validation_step_end"):
        out = validation_step_end(out)
    val_outs.append(out)
val_outs = validation_epoch_end(val_outs)

# if you have one val dataloader:
def validation_step(self, batch, batch_idx):
    ...


# if you have multiple val dataloaders:
def validation_step(self, batch, batch_idx, dataloader_idx):
    ...

Examples:

# CASE 1: A single validation dataset
def validation_step(self, batch, batch_idx):
    x, y = batch

    # implement your own
    out = self(x)
    loss = self.loss(out, y)

    # log 6 example images
    # or generated text... or whatever
    sample_imgs = x[:6]
    grid = torchvision.utils.make_grid(sample_imgs)
    self.logger.experiment.add_image('example_images', grid, 0)

    # calculate acc
    labels_hat = torch.argmax(out, dim=1)
    val_acc = torch.sum(y == labels_hat).item() / (len(y) * 1.0)

    # log the outputs!
    self.log_dict({'val_loss': loss, 'val_acc': val_acc})

If you pass in multiple val dataloaders, validation_step() will have an additional argument.

# CASE 2: multiple validation dataloaders
def validation_step(self, batch, batch_idx, dataloader_idx):
    # dataloader_idx tells you which dataset this is.
    ...

Note

If you don’t need to validate you don’t need to implement this method.

Note

When the validation_step() is called, the model has been put in eval mode and PyTorch gradients have been disabled. At the end of validation, the model goes back to training mode and gradients are enabled.

reagent.training.parameters module

class reagent.training.parameters.C51TrainerParameters(actions: List[str] = <factory>, rl: reagent.core.parameters.RLParameters = <factory>, double_q_learning: bool = True, minibatch_size: int = 1024, minibatches_per_step: int = 1, num_atoms: int = 51, qmin: float = -100, qmax: float = 200, optimizer: reagent.optimizer.union.Optimizer__Union = <factory>)

Bases: object

actions: List[str]

asdict()

double_q_learning: bool = True

minibatch_size: int = 1024

minibatches_per_step: int = 1

num_atoms: int = 51

optimizer: reagent.optimizer.union.Optimizer__Union

qmax: float = 200

qmin: float = -100

rl: reagent.core.parameters.RLParameters

class reagent.training.parameters.CRRTrainerParameters(rl: reagent.core.parameters.RLParameters = <factory>, double_q_learning: bool = True, q_network_optimizer: reagent.optimizer.union.Optimizer__Union = <factory>, actor_network_optimizer: reagent.optimizer.union.Optimizer__Union = <factory>, use_target_actor: bool = False, actions: List[str] = <factory>, delayed_policy_update: int = 1, beta: float = 1.0, entropy_coeff: float = 0.0, clip_limit: float = 10.0, max_weight: float = 20.0)

Bases: object

actions: List[str]

actor_network_optimizer: reagent.optimizer.union.Optimizer__Union

asdict()

beta: float = 1.0

clip_limit: float = 10.0

delayed_policy_update: int = 1

double_q_learning: bool = True

entropy_coeff: float = 0.0

max_weight: float = 20.0

q_network_optimizer: reagent.optimizer.union.Optimizer__Union

rl: reagent.core.parameters.RLParameters

use_target_actor: bool = False

class reagent.training.parameters.DQNTrainerParameters(actions: List[str] = <factory>, rl: reagent.core.parameters.RLParameters = <factory>, double_q_learning: bool = True, bcq: Optional[reagent.training.dqn_trainer.BCQConfig] = None, minibatch_size: int = 1024, minibatches_per_step: int = 1, optimizer: reagent.optimizer.union.Optimizer__Union = <factory>)

Bases: object

actions: List[str]

asdict()

bcq: Optional[reagent.training.dqn_trainer.BCQConfig] = None

double_q_learning: bool = True

minibatch_size: int = 1024

minibatches_per_step: int = 1

optimizer: reagent.optimizer.union.Optimizer__Union

rl: reagent.core.parameters.RLParameters

class reagent.training.parameters.LinUCBTrainerParameters(num_actions: int = - 1, use_interaction_features: bool = True)

Bases: object

asdict()

num_actions: int = -1

use_interaction_features: bool = True

class reagent.training.parameters.PPOTrainerParameters(gamma: float = 0.9, optimizer: reagent.optimizer.union.Optimizer__Union = <factory>, optimizer_value_net: reagent.optimizer.union.Optimizer__Union = <factory>, actions: List[str] = <factory>, reward_clip: float = 1000000.0, normalize: bool = True, subtract_mean: bool = True, offset_clamp_min: bool = False, update_freq: int = 1, update_epochs: int = 1, ppo_batch_size: int = 1, ppo_epsilon: float = 0.2, entropy_weight: float = 0.0)

Bases: object

actions: List[str]

asdict()

entropy_weight: float = 0.0

gamma: float = 0.9

normalize: bool = True

offset_clamp_min: bool = False

optimizer: reagent.optimizer.union.Optimizer__Union

optimizer_value_net: reagent.optimizer.union.Optimizer__Union

ppo_batch_size: int = 1

ppo_epsilon: float = 0.2

reward_clip: float = 1000000.0

subtract_mean: bool = True

update_epochs: int = 1

update_freq: int = 1

class reagent.training.parameters.ParametricDQNTrainerParameters(rl: reagent.core.parameters.RLParameters = <factory>, double_q_learning: bool = True, minibatches_per_step: int = 1, optimizer: reagent.optimizer.union.Optimizer__Union = <factory>)

Bases: object

asdict()

double_q_learning: bool = True

minibatches_per_step: int = 1

optimizer: reagent.optimizer.union.Optimizer__Union

rl: reagent.core.parameters.RLParameters

class reagent.training.parameters.QRDQNTrainerParameters(actions: List[str] = <factory>, rl: reagent.core.parameters.RLParameters = <factory>, double_q_learning: bool = True, num_atoms: int = 51, minibatch_size: int = 1024, minibatches_per_step: int = 1, optimizer: reagent.optimizer.union.Optimizer__Union = <factory>, cpe_optimizer: reagent.optimizer.union.Optimizer__Union = <factory>)

Bases: object

actions: List[str]

asdict()

cpe_optimizer: reagent.optimizer.union.Optimizer__Union

double_q_learning: bool = True

minibatch_size: int = 1024

minibatches_per_step: int = 1

num_atoms: int = 51

optimizer: reagent.optimizer.union.Optimizer__Union

rl: reagent.core.parameters.RLParameters

class reagent.training.parameters.ReinforceTrainerParameters(gamma: float = 0.0, optimizer: reagent.optimizer.union.Optimizer__Union = <factory>, optimizer_value_net: reagent.optimizer.union.Optimizer__Union = <factory>, actions: List[str] = <factory>, off_policy: bool = False, reward_clip: float = 1000000.0, clip_param: float = 1000000.0, normalize: bool = True, subtract_mean: bool = True, offset_clamp_min: bool = False)

Bases: object

actions: List[str]

asdict()

clip_param: float = 1000000.0

gamma: float = 0.0

normalize: bool = True

off_policy: bool = False

offset_clamp_min: bool = False

optimizer: reagent.optimizer.union.Optimizer__Union

optimizer_value_net: reagent.optimizer.union.Optimizer__Union

reward_clip: float = 1000000.0

subtract_mean: bool = True

class reagent.training.parameters.RewardNetworkTrainerParameters(optimizer: reagent.optimizer.union.Optimizer__Union = <factory>, loss_type: reagent.training.reward_network_trainer.LossFunction = <LossFunction.MSE: 'MSE_Loss'>, reward_ignore_threshold: Optional[float] = None, weighted_by_inverse_propensity: bool = False)

Bases: object

asdict()

loss_type: reagent.training.reward_network_trainer.LossFunction = 'MSE_Loss'

optimizer: reagent.optimizer.union.Optimizer__Union

reward_ignore_threshold: Optional[float] = None

weighted_by_inverse_propensity: bool = False

class reagent.training.parameters.SACTrainerParameters(rl: reagent.core.parameters.RLParameters = <factory>, q_network_optimizer: reagent.optimizer.union.Optimizer__Union = <factory>, value_network_optimizer: reagent.optimizer.union.Optimizer__Union = <factory>, actor_network_optimizer: reagent.optimizer.union.Optimizer__Union = <factory>, alpha_optimizer: Optional[reagent.optimizer.union.Optimizer__Union] = <factory>, minibatch_size: int = 1024, entropy_temperature: float = 0.01, logged_action_uniform_prior: bool = True, target_entropy: float = -1.0, action_embedding_kld_weight: Optional[float] = None, apply_kld_on_mean: bool = False, action_embedding_mean: Optional[List[float]] = None, action_embedding_variance: Optional[List[float]] = None, crr_config: Optional[reagent.training.sac_trainer.CRRWeightFn] = None, backprop_through_log_prob: bool = True)

Bases: object

action_embedding_kld_weight: Optional[float] = None

action_embedding_mean: Optional[List[float]] = None

action_embedding_variance: Optional[List[float]] = None

actor_network_optimizer: reagent.optimizer.union.Optimizer__Union

alpha_optimizer: Optional[reagent.optimizer.union.Optimizer__Union]

apply_kld_on_mean: bool = False

asdict()

backprop_through_log_prob: bool = True

crr_config: Optional[reagent.training.sac_trainer.CRRWeightFn] = None

entropy_temperature: float = 0.01

logged_action_uniform_prior: bool = True

minibatch_size: int = 1024

q_network_optimizer: reagent.optimizer.union.Optimizer__Union

rl: reagent.core.parameters.RLParameters

target_entropy: float = -1.0

value_network_optimizer: reagent.optimizer.union.Optimizer__Union

class reagent.training.parameters.Seq2SlateTrainerParameters(params: reagent.core.parameters.Seq2SlateParameters = <factory>, policy_optimizer: reagent.optimizer.union.Optimizer__Union = <factory>, baseline_optimizer: reagent.optimizer.union.Optimizer__Union = <factory>, policy_gradient_interval: int = 1, print_interval: int = 100, calc_cpe: bool = False, reward_network: Optional[torch.nn.modules.module.Module] = None)

Bases: reagent.core.base_dataclass.BaseDataClass

asdict()

baseline_optimizer: reagent.optimizer.union.Optimizer__Union

calc_cpe: bool = False

params: reagent.core.parameters.Seq2SlateParameters

policy_gradient_interval: int = 1

policy_optimizer: reagent.optimizer.union.Optimizer__Union

print_interval: int = 100

reward_network: Optional[torch.nn.modules.module.Module] = None

class reagent.training.parameters.SlateQTrainerParameters(rl: reagent.core.parameters.RLParameters = <factory>, optimizer: reagent.optimizer.union.Optimizer__Union = <factory>, slate_opt_parameters: Optional[reagent.core.parameters.SlateOptParameters] = None, discount_time_scale: Optional[float] = None, single_selection: bool = True, next_slate_value_norm_method: reagent.training.slate_q_trainer.NextSlateValueNormMethod = <NextSlateValueNormMethod.NORM_BY_CURRENT_SLATE_SIZE: 'norm_by_current_slate_size'>, minibatch_size: int = 1024, evaluation: reagent.core.parameters.EvaluationParameters = <factory>)

Bases: object

asdict()

discount_time_scale: Optional[float] = None

evaluation: reagent.core.parameters.EvaluationParameters

minibatch_size: int = 1024

next_slate_value_norm_method: reagent.training.slate_q_trainer.NextSlateValueNormMethod = 'norm_by_current_slate_size'

optimizer: reagent.optimizer.union.Optimizer__Union

rl: reagent.core.parameters.RLParameters

single_selection: bool = True

slate_opt_parameters: Optional[reagent.core.parameters.SlateOptParameters] = None

class reagent.training.parameters.TD3TrainerParameters(rl: reagent.core.parameters.RLParameters = <factory>, q_network_optimizer: reagent.optimizer.union.Optimizer__Union = <factory>, actor_network_optimizer: reagent.optimizer.union.Optimizer__Union = <factory>, minibatch_size: int = 64, noise_variance: float = 0.2, noise_clip: float = 0.5, delayed_policy_update: int = 2, minibatches_per_step: int = 1)

Bases: object

actor_network_optimizer: reagent.optimizer.union.Optimizer__Union

asdict()

delayed_policy_update: int = 2

minibatch_size: int = 64

minibatches_per_step: int = 1

noise_clip: float = 0.5

noise_variance: float = 0.2

q_network_optimizer: reagent.optimizer.union.Optimizer__Union

rl: reagent.core.parameters.RLParameters

reagent.training.parametric_dqn_trainer module

class reagent.training.parametric_dqn_trainer.ParametricDQNTrainer(q_network, q_network_target, reward_network, rl: reagent.core.parameters.RLParameters = Field(name='rl', type=<class 'reagent.core.parameters.RLParameters'>, default=<dataclasses._MISSING_TYPE object>, default_factory=<class 'reagent.core.parameters.RLParameters'>, init=True, repr=True, hash=None, compare=True, metadata=mappingproxy({}), _field_type=_FIELD), double_q_learning: bool = True, minibatches_per_step: int = 1, optimizer: reagent.optimizer.union.Optimizer__Union = Field(name='optimizer', type=<class 'reagent.optimizer.union.Optimizer__Union'>, default=<dataclasses._MISSING_TYPE object>, default_factory=<bound method Optimizer__Union.default of <class 'reagent.optimizer.union.Optimizer__Union'>>, init=True, repr=True, hash=None, compare=True, metadata=mappingproxy({}), _field_type=_FIELD))

Bases: reagent.training.dqn_trainer_base.DQNTrainerMixin, reagent.training.rl_trainer_pytorch.RLTrainerMixin, reagent.training.reagent_lightning_module.ReAgentLightningModule

configure_optimizers()

Choose what optimizers and learning-rate schedulers to use in your optimization. Normally you’d need one. But in the case of GANs or similar you might have multiple.

Returns

Any of these 6 options.

Single optimizer.
List or Tuple of optimizers.
Two lists - The first list has multiple optimizers, and the second has multiple LR schedulers (or multiple lr_scheduler_config).
Dictionary, with an "optimizer" key, and (optionally) a "lr_scheduler" key whose value is a single LR scheduler or lr_scheduler_config.
Tuple of dictionaries as described above, with an optional "frequency" key.
None - Fit will run without any optimizer.

The lr_scheduler_config is a dictionary which contains the scheduler and its associated configuration. The default configuration is shown below.

lr_scheduler_config = {
    # REQUIRED: The scheduler instance
    "scheduler": lr_scheduler,
    # The unit of the scheduler's step size, could also be 'step'.
    # 'epoch' updates the scheduler on epoch end whereas 'step'
    # updates it after a optimizer update.
    "interval": "epoch",
    # How many epochs/steps should pass between calls to
    # `scheduler.step()`. 1 corresponds to updating the learning
    # rate after every epoch/step.
    "frequency": 1,
    # Metric to to monitor for schedulers like `ReduceLROnPlateau`
    "monitor": "val_loss",
    # If set to `True`, will enforce that the value specified 'monitor'
    # is available when the scheduler is updated, thus stopping
    # training if not found. If set to `False`, it will only produce a warning
    "strict": True,
    # If using the `LearningRateMonitor` callback to monitor the
    # learning rate progress, this keyword can be used to specify
    # a custom logged name
    "name": None,
}

When there are schedulers in which the .step() method is conditioned on a value, such as the torch.optim.lr_scheduler.ReduceLROnPlateau scheduler, Lightning requires that the lr_scheduler_config contains the keyword "monitor" set to the metric name that the scheduler should be conditioned on.

Metrics can be made available to monitor by simply logging it using self.log('metric_to_track', metric_val) in your LightningModule.

Note

The frequency value specified in a dict along with the optimizer key is an int corresponding to the number of sequential batches optimized with the specific optimizer. It should be given to none or to all of the optimizers. There is a difference between passing multiple optimizers in a list, and passing multiple optimizers in dictionaries with a frequency of 1:

This is different from the frequency value specified in the lr_scheduler_config mentioned above.

def configure_optimizers(self):
    optimizer_one = torch.optim.SGD(self.model.parameters(), lr=0.01)
    optimizer_two = torch.optim.SGD(self.model.parameters(), lr=0.01)
    return [
        {"optimizer": optimizer_one, "frequency": 5},
        {"optimizer": optimizer_two, "frequency": 10},
    ]

In this example, the first optimizer will be used for the first 5 steps, the second optimizer for the next 10 steps and that cycle will continue. If an LR scheduler is specified for an optimizer using the lr_scheduler key in the above dict, the scheduler will only be updated when its optimizer is being used.

Examples:

# most cases. no learning rate scheduler
def configure_optimizers(self):
    return Adam(self.parameters(), lr=1e-3)

# multiple optimizer case (e.g.: GAN)
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    return gen_opt, dis_opt

# example with learning rate schedulers
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    dis_sch = CosineAnnealing(dis_opt, T_max=10)
    return [gen_opt, dis_opt], [dis_sch]

# example with step-based learning rate schedulers
# each optimizer has its own scheduler
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    gen_sch = {
        'scheduler': ExponentialLR(gen_opt, 0.99),
        'interval': 'step'  # called after each training step
    }
    dis_sch = CosineAnnealing(dis_opt, T_max=10) # called every epoch
    return [gen_opt, dis_opt], [gen_sch, dis_sch]

# example with optimizer frequencies
# see training procedure in `Improved Training of Wasserstein GANs`, Algorithm 1
# https://arxiv.org/abs/1704.00028
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    n_critic = 5
    return (
        {'optimizer': dis_opt, 'frequency': n_critic},
        {'optimizer': gen_opt, 'frequency': 1}
    )

Note

Some things to know:

Lightning calls .backward() and .step() on each optimizer and learning rate scheduler as needed.
If you use 16-bit precision (precision=16), Lightning will automatically handle the optimizers.
If you use multiple optimizers, training_step() will have an additional optimizer_idx parameter.
If you use torch.optim.LBFGS, Lightning handles the closure function automatically for you.
If you use multiple optimizers, gradients will be calculated only for the parameters of current optimizer at each training step.
If you need to control how often those optimizers step or override the default .step() schedule, override the optimizer_step() hook.

get_detached_model_outputs(state, action) → Tuple[torch.Tensor, torch.Tensor]: Gets the q values from the model and target networks

train_step_gen(training_batch: reagent.core.types.ParametricDqnInput, batch_idx: int): Implement training step as generator here

reagent.training.ppo_trainer module

class reagent.training.ppo_trainer.PPOTrainer(policy: reagent.gym.policies.policy.Policy, gamma: float = 0.9, optimizer: reagent.optimizer.union.Optimizer__Union = Field(name='optimizer',type=<class 'reagent.optimizer.union.Optimizer__Union'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<bound method Optimizer__Union.default of <class 'reagent.optimizer.union.Optimizer__Union'>>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), optimizer_value_net: reagent.optimizer.union.Optimizer__Union = Field(name='optimizer_value_net',type=<class 'reagent.optimizer.union.Optimizer__Union'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<bound method Optimizer__Union.default of <class 'reagent.optimizer.union.Optimizer__Union'>>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), actions: List[str] = Field(name='actions',type=typing.List[str],default=<dataclasses._MISSING_TYPE object>,default_factory=<class 'list'>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), reward_clip: float = 1000000.0, normalize: bool = True, subtract_mean: bool = True, offset_clamp_min: bool = False, update_freq: int = 1, update_epochs: int = 1, ppo_batch_size: int = 1, ppo_epsilon: float = 0.2, entropy_weight: float = 0.0, value_net: Optional[reagent.models.base.ModelBase] = None)

Bases: reagent.training.reagent_lightning_module.ReAgentLightningModule

Proximal Policy Optimization (PPO). See https://arxiv.org/pdf/1707.06347.pdf This is the “clip” version of PPO. It does not include: - KL divergence - Bootstrapping with a critic model (our approach only works if full trajectories up to terminal state are fed in) Optionally, a value network can be trained and used as a baseline for rewards.

configure_optimizers()

Choose what optimizers and learning-rate schedulers to use in your optimization. Normally you’d need one. But in the case of GANs or similar you might have multiple.

Returns

Any of these 6 options.

Single optimizer.
List or Tuple of optimizers.
Two lists - The first list has multiple optimizers, and the second has multiple LR schedulers (or multiple lr_scheduler_config).
Dictionary, with an "optimizer" key, and (optionally) a "lr_scheduler" key whose value is a single LR scheduler or lr_scheduler_config.
Tuple of dictionaries as described above, with an optional "frequency" key.
None - Fit will run without any optimizer.

The lr_scheduler_config is a dictionary which contains the scheduler and its associated configuration. The default configuration is shown below.

lr_scheduler_config = {
    # REQUIRED: The scheduler instance
    "scheduler": lr_scheduler,
    # The unit of the scheduler's step size, could also be 'step'.
    # 'epoch' updates the scheduler on epoch end whereas 'step'
    # updates it after a optimizer update.
    "interval": "epoch",
    # How many epochs/steps should pass between calls to
    # `scheduler.step()`. 1 corresponds to updating the learning
    # rate after every epoch/step.
    "frequency": 1,
    # Metric to to monitor for schedulers like `ReduceLROnPlateau`
    "monitor": "val_loss",
    # If set to `True`, will enforce that the value specified 'monitor'
    # is available when the scheduler is updated, thus stopping
    # training if not found. If set to `False`, it will only produce a warning
    "strict": True,
    # If using the `LearningRateMonitor` callback to monitor the
    # learning rate progress, this keyword can be used to specify
    # a custom logged name
    "name": None,
}

When there are schedulers in which the .step() method is conditioned on a value, such as the torch.optim.lr_scheduler.ReduceLROnPlateau scheduler, Lightning requires that the lr_scheduler_config contains the keyword "monitor" set to the metric name that the scheduler should be conditioned on.

Metrics can be made available to monitor by simply logging it using self.log('metric_to_track', metric_val) in your LightningModule.

Note

The frequency value specified in a dict along with the optimizer key is an int corresponding to the number of sequential batches optimized with the specific optimizer. It should be given to none or to all of the optimizers. There is a difference between passing multiple optimizers in a list, and passing multiple optimizers in dictionaries with a frequency of 1:

This is different from the frequency value specified in the lr_scheduler_config mentioned above.

def configure_optimizers(self):
    optimizer_one = torch.optim.SGD(self.model.parameters(), lr=0.01)
    optimizer_two = torch.optim.SGD(self.model.parameters(), lr=0.01)
    return [
        {"optimizer": optimizer_one, "frequency": 5},
        {"optimizer": optimizer_two, "frequency": 10},
    ]

In this example, the first optimizer will be used for the first 5 steps, the second optimizer for the next 10 steps and that cycle will continue. If an LR scheduler is specified for an optimizer using the lr_scheduler key in the above dict, the scheduler will only be updated when its optimizer is being used.

Examples:

# most cases. no learning rate scheduler
def configure_optimizers(self):
    return Adam(self.parameters(), lr=1e-3)

# multiple optimizer case (e.g.: GAN)
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    return gen_opt, dis_opt

# example with learning rate schedulers
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    dis_sch = CosineAnnealing(dis_opt, T_max=10)
    return [gen_opt, dis_opt], [dis_sch]

# example with step-based learning rate schedulers
# each optimizer has its own scheduler
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    gen_sch = {
        'scheduler': ExponentialLR(gen_opt, 0.99),
        'interval': 'step'  # called after each training step
    }
    dis_sch = CosineAnnealing(dis_opt, T_max=10) # called every epoch
    return [gen_opt, dis_opt], [gen_sch, dis_sch]

# example with optimizer frequencies
# see training procedure in `Improved Training of Wasserstein GANs`, Algorithm 1
# https://arxiv.org/abs/1704.00028
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    n_critic = 5
    return (
        {'optimizer': dis_opt, 'frequency': n_critic},
        {'optimizer': gen_opt, 'frequency': 1}
    )

Note

Some things to know:

Lightning calls .backward() and .step() on each optimizer and learning rate scheduler as needed.
If you use 16-bit precision (precision=16), Lightning will automatically handle the optimizers.
If you use multiple optimizers, training_step() will have an additional optimizer_idx parameter.
If you use torch.optim.LBFGS, Lightning handles the closure function automatically for you.
If you use multiple optimizers, gradients will be calculated only for the parameters of current optimizer at each training step.
If you need to control how often those optimizers step or override the default .step() schedule, override the optimizer_step() hook.

get_optimizers()

training: bool

training_step(training_batch: Union[reagent.core.types.PolicyGradientInput, Dict[str, torch.Tensor]], batch_idx: int)

Here you compute and return the training loss and some additional metrics for e.g. the progress bar or logger.

Parameters

batch (Tensor | (Tensor, …) | [Tensor, …]) – The output of your DataLoader. A tensor, tuple or list.
batch_idx (int) – Integer displaying index of this batch
optimizer_idx (int) – When using multiple optimizers, this argument will also be present.
hiddens (Any) – Passed in if :paramref:`~pytorch_lightning.core.lightning.LightningModule.truncated_bptt_steps` > 0.

Returns

Any of.

Tensor - The loss tensor
dict - A dictionary. Can include any keys, but must include the key 'loss'
None - Training will skip to the next batch. This is only for automatic optimization.
This is not supported for multi-GPU, TPU, IPU, or DeepSpeed.

In this step you’d normally do the forward pass and calculate the loss for a batch. You can also do fancier things like multiple forward passes or something model specific.

Example:

def training_step(self, batch, batch_idx):
    x, y, z = batch
    out = self.encoder(x)
    loss = self.loss(out, x)
    return loss

If you define multiple optimizers, this step will be called with an additional optimizer_idx parameter.

# Multiple optimizers (e.g.: GANs)
def training_step(self, batch, batch_idx, optimizer_idx):
    if optimizer_idx == 0:
        # do training_step with encoder
        ...
    if optimizer_idx == 1:
        # do training_step with decoder
        ...

If you add truncated back propagation through time you will also get an additional argument with the hidden states of the previous step.

# Truncated back-propagation through time
def training_step(self, batch, batch_idx, hiddens):
    # hiddens are the hidden states from the previous truncated backprop step
    out, hiddens = self.lstm(data, hiddens)
    loss = ...
    return {"loss": loss, "hiddens": hiddens}

Note

The loss value shown in the progress bar is smoothed (averaged) over the last values, so it differs from the actual loss returned in train/validation step.

update_model()

reagent.training.qrdqn_trainer module

class reagent.training.qrdqn_trainer.QRDQNTrainer(q_network, q_network_target, metrics_to_score=None, reward_network=None, q_network_cpe=None, q_network_cpe_target=None, actions: List[str] = Field(name='actions',type=typing.List[str],default=<dataclasses._MISSING_TYPE object>,default_factory=<class 'list'>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), rl: reagent.core.parameters.RLParameters = Field(name='rl',type=<class 'reagent.core.parameters.RLParameters'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<class 'reagent.core.parameters.RLParameters'>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), double_q_learning: bool = True, num_atoms: int = 51, minibatch_size: int = 1024, minibatches_per_step: int = 1, optimizer: reagent.optimizer.union.Optimizer__Union = Field(name='optimizer',type=<class 'reagent.optimizer.union.Optimizer__Union'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<bound method Optimizer__Union.default of <class 'reagent.optimizer.union.Optimizer__Union'>>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), cpe_optimizer: reagent.optimizer.union.Optimizer__Union = Field(name='cpe_optimizer',type=<class 'reagent.optimizer.union.Optimizer__Union'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<bound method Optimizer__Union.default of <class 'reagent.optimizer.union.Optimizer__Union'>>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), evaluation: reagent.core.parameters.EvaluationParameters = Field(name=None,type=None,default=<dataclasses._MISSING_TYPE object>,default_factory=<class 'reagent.core.parameters.EvaluationParameters'>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=None))

Bases: reagent.training.dqn_trainer_base.DQNTrainerBaseLightning

Implementation of QR-DQN (Quantile Regression Deep Q-Network)

See https://arxiv.org/abs/1710.10044 for details

allow_zero_length_dataloader_with_multiple_devices: bool

argmax_with_mask(q_values, possible_actions_mask)

boost_rewards(rewards: torch.Tensor, actions: torch.Tensor) → torch.Tensor

configure_optimizers()

Choose what optimizers and learning-rate schedulers to use in your optimization. Normally you’d need one. But in the case of GANs or similar you might have multiple.

Returns

Any of these 6 options.

Single optimizer.
List or Tuple of optimizers.
Two lists - The first list has multiple optimizers, and the second has multiple LR schedulers (or multiple lr_scheduler_config).
Dictionary, with an "optimizer" key, and (optionally) a "lr_scheduler" key whose value is a single LR scheduler or lr_scheduler_config.
Tuple of dictionaries as described above, with an optional "frequency" key.
None - Fit will run without any optimizer.

The lr_scheduler_config is a dictionary which contains the scheduler and its associated configuration. The default configuration is shown below.

lr_scheduler_config = {
    # REQUIRED: The scheduler instance
    "scheduler": lr_scheduler,
    # The unit of the scheduler's step size, could also be 'step'.
    # 'epoch' updates the scheduler on epoch end whereas 'step'
    # updates it after a optimizer update.
    "interval": "epoch",
    # How many epochs/steps should pass between calls to
    # `scheduler.step()`. 1 corresponds to updating the learning
    # rate after every epoch/step.
    "frequency": 1,
    # Metric to to monitor for schedulers like `ReduceLROnPlateau`
    "monitor": "val_loss",
    # If set to `True`, will enforce that the value specified 'monitor'
    # is available when the scheduler is updated, thus stopping
    # training if not found. If set to `False`, it will only produce a warning
    "strict": True,
    # If using the `LearningRateMonitor` callback to monitor the
    # learning rate progress, this keyword can be used to specify
    # a custom logged name
    "name": None,
}

When there are schedulers in which the .step() method is conditioned on a value, such as the torch.optim.lr_scheduler.ReduceLROnPlateau scheduler, Lightning requires that the lr_scheduler_config contains the keyword "monitor" set to the metric name that the scheduler should be conditioned on.

Metrics can be made available to monitor by simply logging it using self.log('metric_to_track', metric_val) in your LightningModule.

Note

The frequency value specified in a dict along with the optimizer key is an int corresponding to the number of sequential batches optimized with the specific optimizer. It should be given to none or to all of the optimizers. There is a difference between passing multiple optimizers in a list, and passing multiple optimizers in dictionaries with a frequency of 1:

This is different from the frequency value specified in the lr_scheduler_config mentioned above.

def configure_optimizers(self):
    optimizer_one = torch.optim.SGD(self.model.parameters(), lr=0.01)
    optimizer_two = torch.optim.SGD(self.model.parameters(), lr=0.01)
    return [
        {"optimizer": optimizer_one, "frequency": 5},
        {"optimizer": optimizer_two, "frequency": 10},
    ]

In this example, the first optimizer will be used for the first 5 steps, the second optimizer for the next 10 steps and that cycle will continue. If an LR scheduler is specified for an optimizer using the lr_scheduler key in the above dict, the scheduler will only be updated when its optimizer is being used.

Examples:

# most cases. no learning rate scheduler
def configure_optimizers(self):
    return Adam(self.parameters(), lr=1e-3)

# multiple optimizer case (e.g.: GAN)
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    return gen_opt, dis_opt

# example with learning rate schedulers
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    dis_sch = CosineAnnealing(dis_opt, T_max=10)
    return [gen_opt, dis_opt], [dis_sch]

# example with step-based learning rate schedulers
# each optimizer has its own scheduler
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    gen_sch = {
        'scheduler': ExponentialLR(gen_opt, 0.99),
        'interval': 'step'  # called after each training step
    }
    dis_sch = CosineAnnealing(dis_opt, T_max=10) # called every epoch
    return [gen_opt, dis_opt], [gen_sch, dis_sch]

# example with optimizer frequencies
# see training procedure in `Improved Training of Wasserstein GANs`, Algorithm 1
# https://arxiv.org/abs/1704.00028
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    n_critic = 5
    return (
        {'optimizer': dis_opt, 'frequency': n_critic},
        {'optimizer': gen_opt, 'frequency': 1}
    )

Note

Some things to know:

Lightning calls .backward() and .step() on each optimizer and learning rate scheduler as needed.
If you use 16-bit precision (precision=16), Lightning will automatically handle the optimizers.
If you use multiple optimizers, training_step() will have an additional optimizer_idx parameter.
If you use torch.optim.LBFGS, Lightning handles the closure function automatically for you.
If you use multiple optimizers, gradients will be calculated only for the parameters of current optimizer at each training step.
If you need to control how often those optimizers step or override the default .step() schedule, override the optimizer_step() hook.

get_detached_model_outputs(state: reagent.core.types.FeatureData) → Tuple[torch.Tensor, torch.Tensor]: Gets the q values from the model and target networks

huber(x)

precision: int

prepare_data_per_node: bool

rl_parameters: reagent.core.parameters.RLParameters

train_step_gen(training_batch: reagent.core.types.DiscreteDqnInput, batch_idx: int): Implement training step as generator here

training: bool

use_amp: bool

reagent.training.reagent_lightning_module module

class reagent.training.reagent_lightning_module.ReAgentLightningModule(automatic_optimization=True)

Bases: pytorch_lightning.core.lightning.LightningModule

increase_next_stopping_epochs(num_epochs: int)

on_epoch_end(): Called when either of train/val/test epoch ends.

on_test_batch_end(*args, **kwargs)

Called in the test loop after the batch.

Parameters

outputs – The outputs of test_step_end(test_step(x))
batch – The batched data as it is returned by the test DataLoader.
batch_idx – the index of the batch
dataloader_idx – the index of the dataloader

on_train_batch_end(*args, **kwargs)

Called in the training loop after the batch.

Parameters

outputs – The outputs of training_step_end(training_step(x))
batch – The batched data as it is returned by the training DataLoader.
batch_idx – the index of the batch
unused – Deprecated argument. Will be removed in v1.7.

on_validation_batch_end(*args, **kwargs)

Called in the validation loop after the batch.

Parameters

outputs – The outputs of validation_step_end(validation_step(x))
batch – The batched data as it is returned by the validation DataLoader.
batch_idx – the index of the batch
dataloader_idx – the index of the dataloader

optimizers(use_pl_optimizer: bool = True)

Returns the optimizer(s) that are being used during training. Useful for manual optimization.

Parameters: use_pl_optimizer – If True, will wrap the optimizer(s) in a LightningOptimizer for automatic handling of precision and profiling.
Returns: A single optimizer, or a list of optimizers in case multiple ones are present.

property reporter

set_clean_stop(clean_stop: bool)

set_reporter(reporter)

soft_update_result() → torch.Tensor: A dummy loss to trigger soft-update

property summary_writer: Accessor to TensorBoard’s SummaryWriter

train(*args)

Sets the module in training mode.

This has any effect only on certain modules. See documentations of particular modules for details of their behaviors in training/evaluation mode, if they are affected, e.g. Dropout, BatchNorm, etc.

Parameters: mode (bool) – whether to set training mode (True) or evaluation mode (False). Default: True.
Returns: self
Return type: Module

train_step_gen(training_batch, batch_idx: int): Implement training step as generator here

training: bool

training_step(batch, batch_idx: int, optimizer_idx: int = 0)

Here you compute and return the training loss and some additional metrics for e.g. the progress bar or logger.

Parameters

batch (Tensor | (Tensor, …) | [Tensor, …]) – The output of your DataLoader. A tensor, tuple or list.
batch_idx (int) – Integer displaying index of this batch
optimizer_idx (int) – When using multiple optimizers, this argument will also be present.
hiddens (Any) – Passed in if :paramref:`~pytorch_lightning.core.lightning.LightningModule.truncated_bptt_steps` > 0.

Returns

Any of.

Tensor - The loss tensor
dict - A dictionary. Can include any keys, but must include the key 'loss'
None - Training will skip to the next batch. This is only for automatic optimization.
This is not supported for multi-GPU, TPU, IPU, or DeepSpeed.

In this step you’d normally do the forward pass and calculate the loss for a batch. You can also do fancier things like multiple forward passes or something model specific.

Example:

def training_step(self, batch, batch_idx):
    x, y, z = batch
    out = self.encoder(x)
    loss = self.loss(out, x)
    return loss

If you define multiple optimizers, this step will be called with an additional optimizer_idx parameter.

# Multiple optimizers (e.g.: GANs)
def training_step(self, batch, batch_idx, optimizer_idx):
    if optimizer_idx == 0:
        # do training_step with encoder
        ...
    if optimizer_idx == 1:
        # do training_step with decoder
        ...

If you add truncated back propagation through time you will also get an additional argument with the hidden states of the previous step.

# Truncated back-propagation through time
def training_step(self, batch, batch_idx, hiddens):
    # hiddens are the hidden states from the previous truncated backprop step
    out, hiddens = self.lstm(data, hiddens)
    loss = ...
    return {"loss": loss, "hiddens": hiddens}

Note

The loss value shown in the progress bar is smoothed (averaged) over the last values, so it differs from the actual loss returned in train/validation step.

class reagent.training.reagent_lightning_module.StoppingEpochCallback(num_epochs)

Bases: pytorch_lightning.callbacks.base.Callback

We use this callback to control the number of training epochs in incremental training. Epoch & step counts are not reset in the checkpoint. If we were to set max_epochs on the trainer, we would have to keep track of the previous max_epochs and add to it manually. This keeps the infomation in one place.

Note that we need to set _cleanly_stopped back to True before saving the checkpoint. This is done in ModelManager.save_trainer().

on_pretrain_routine_end(trainer, pl_module): Called when the pretrain routine ends.

reagent.training.reagent_lightning_module.has_test_step_override(trainer_module: reagent.training.reagent_lightning_module.ReAgentLightningModule): Detect if a subclass of LightningModule has test_step overridden

reagent.training.reinforce_trainer module

class reagent.training.reinforce_trainer.ReinforceTrainer(policy: reagent.gym.policies.policy.Policy, gamma: float = 0.0, optimizer: reagent.optimizer.union.Optimizer__Union = Field(name='optimizer',type=<class 'reagent.optimizer.union.Optimizer__Union'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<bound method Optimizer__Union.default of <class 'reagent.optimizer.union.Optimizer__Union'>>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), optimizer_value_net: reagent.optimizer.union.Optimizer__Union = Field(name='optimizer_value_net',type=<class 'reagent.optimizer.union.Optimizer__Union'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<bound method Optimizer__Union.default of <class 'reagent.optimizer.union.Optimizer__Union'>>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), actions: List[str] = Field(name='actions',type=typing.List[str],default=<dataclasses._MISSING_TYPE object>,default_factory=<class 'list'>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), off_policy: bool = False, reward_clip: float = 1000000.0, clip_param: float = 1000000.0, normalize: bool = True, subtract_mean: bool = True, offset_clamp_min: bool = False, value_net: Optional[reagent.models.base.ModelBase] = None)

Bases: reagent.training.reagent_lightning_module.ReAgentLightningModule

allow_zero_length_dataloader_with_multiple_devices: bool

configure_optimizers()

Choose what optimizers and learning-rate schedulers to use in your optimization. Normally you’d need one. But in the case of GANs or similar you might have multiple.

Returns

Any of these 6 options.

Single optimizer.
List or Tuple of optimizers.
Two lists - The first list has multiple optimizers, and the second has multiple LR schedulers (or multiple lr_scheduler_config).
Dictionary, with an "optimizer" key, and (optionally) a "lr_scheduler" key whose value is a single LR scheduler or lr_scheduler_config.
Tuple of dictionaries as described above, with an optional "frequency" key.
None - Fit will run without any optimizer.

The lr_scheduler_config is a dictionary which contains the scheduler and its associated configuration. The default configuration is shown below.

lr_scheduler_config = {
    # REQUIRED: The scheduler instance
    "scheduler": lr_scheduler,
    # The unit of the scheduler's step size, could also be 'step'.
    # 'epoch' updates the scheduler on epoch end whereas 'step'
    # updates it after a optimizer update.
    "interval": "epoch",
    # How many epochs/steps should pass between calls to
    # `scheduler.step()`. 1 corresponds to updating the learning
    # rate after every epoch/step.
    "frequency": 1,
    # Metric to to monitor for schedulers like `ReduceLROnPlateau`
    "monitor": "val_loss",
    # If set to `True`, will enforce that the value specified 'monitor'
    # is available when the scheduler is updated, thus stopping
    # training if not found. If set to `False`, it will only produce a warning
    "strict": True,
    # If using the `LearningRateMonitor` callback to monitor the
    # learning rate progress, this keyword can be used to specify
    # a custom logged name
    "name": None,
}

When there are schedulers in which the .step() method is conditioned on a value, such as the torch.optim.lr_scheduler.ReduceLROnPlateau scheduler, Lightning requires that the lr_scheduler_config contains the keyword "monitor" set to the metric name that the scheduler should be conditioned on.

Metrics can be made available to monitor by simply logging it using self.log('metric_to_track', metric_val) in your LightningModule.

Note

The frequency value specified in a dict along with the optimizer key is an int corresponding to the number of sequential batches optimized with the specific optimizer. It should be given to none or to all of the optimizers. There is a difference between passing multiple optimizers in a list, and passing multiple optimizers in dictionaries with a frequency of 1:

This is different from the frequency value specified in the lr_scheduler_config mentioned above.

def configure_optimizers(self):
    optimizer_one = torch.optim.SGD(self.model.parameters(), lr=0.01)
    optimizer_two = torch.optim.SGD(self.model.parameters(), lr=0.01)
    return [
        {"optimizer": optimizer_one, "frequency": 5},
        {"optimizer": optimizer_two, "frequency": 10},
    ]

In this example, the first optimizer will be used for the first 5 steps, the second optimizer for the next 10 steps and that cycle will continue. If an LR scheduler is specified for an optimizer using the lr_scheduler key in the above dict, the scheduler will only be updated when its optimizer is being used.

Examples:

# most cases. no learning rate scheduler
def configure_optimizers(self):
    return Adam(self.parameters(), lr=1e-3)

# multiple optimizer case (e.g.: GAN)
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    return gen_opt, dis_opt

# example with learning rate schedulers
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    dis_sch = CosineAnnealing(dis_opt, T_max=10)
    return [gen_opt, dis_opt], [dis_sch]

# example with step-based learning rate schedulers
# each optimizer has its own scheduler
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    gen_sch = {
        'scheduler': ExponentialLR(gen_opt, 0.99),
        'interval': 'step'  # called after each training step
    }
    dis_sch = CosineAnnealing(dis_opt, T_max=10) # called every epoch
    return [gen_opt, dis_opt], [gen_sch, dis_sch]

# example with optimizer frequencies
# see training procedure in `Improved Training of Wasserstein GANs`, Algorithm 1
# https://arxiv.org/abs/1704.00028
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    n_critic = 5
    return (
        {'optimizer': dis_opt, 'frequency': n_critic},
        {'optimizer': gen_opt, 'frequency': 1}
    )

Note

Some things to know:

Lightning calls .backward() and .step() on each optimizer and learning rate scheduler as needed.
If you use 16-bit precision (precision=16), Lightning will automatically handle the optimizers.
If you use multiple optimizers, training_step() will have an additional optimizer_idx parameter.
If you use torch.optim.LBFGS, Lightning handles the closure function automatically for you.
If you use multiple optimizers, gradients will be calculated only for the parameters of current optimizer at each training step.
If you need to control how often those optimizers step or override the default .step() schedule, override the optimizer_step() hook.

precision: int

prepare_data_per_node: bool

train_step_gen(training_batch: reagent.core.types.PolicyGradientInput, batch_idx: int): Implement training step as generator here

training: bool

use_amp: bool

reagent.training.reward_network_trainer module

class reagent.training.reward_network_trainer.LossFunction(value)

Bases: enum.Enum

An enumeration.

BCELoss = 'BCE_Loss'

L1Loss = 'L1_Loss'

MSE = 'MSE_Loss'

SmoothL1Loss = 'SmoothL1_Loss'

class reagent.training.reward_network_trainer.RewardNetTrainer(reward_net: reagent.models.base.ModelBase, optimizer: reagent.optimizer.union.Optimizer__Union = Field(name='optimizer', type=<class 'reagent.optimizer.union.Optimizer__Union'>, default=<dataclasses._MISSING_TYPE object>, default_factory=<bound method Optimizer__Union.default of <class 'reagent.optimizer.union.Optimizer__Union'>>, init=True, repr=True, hash=None, compare=True, metadata=mappingproxy({}), _field_type=_FIELD), loss_type: reagent.training.reward_network_trainer.LossFunction = LossFunction.MSE, reward_ignore_threshold: Optional[float] = None, weighted_by_inverse_propensity: bool = False)

Bases: reagent.training.reagent_lightning_module.ReAgentLightningModule

allow_zero_length_dataloader_with_multiple_devices: bool

configure_optimizers()

Choose what optimizers and learning-rate schedulers to use in your optimization. Normally you’d need one. But in the case of GANs or similar you might have multiple.

Returns

Any of these 6 options.

Single optimizer.
List or Tuple of optimizers.
Two lists - The first list has multiple optimizers, and the second has multiple LR schedulers (or multiple lr_scheduler_config).
Dictionary, with an "optimizer" key, and (optionally) a "lr_scheduler" key whose value is a single LR scheduler or lr_scheduler_config.
Tuple of dictionaries as described above, with an optional "frequency" key.
None - Fit will run without any optimizer.

The lr_scheduler_config is a dictionary which contains the scheduler and its associated configuration. The default configuration is shown below.

lr_scheduler_config = {
    # REQUIRED: The scheduler instance
    "scheduler": lr_scheduler,
    # The unit of the scheduler's step size, could also be 'step'.
    # 'epoch' updates the scheduler on epoch end whereas 'step'
    # updates it after a optimizer update.
    "interval": "epoch",
    # How many epochs/steps should pass between calls to
    # `scheduler.step()`. 1 corresponds to updating the learning
    # rate after every epoch/step.
    "frequency": 1,
    # Metric to to monitor for schedulers like `ReduceLROnPlateau`
    "monitor": "val_loss",
    # If set to `True`, will enforce that the value specified 'monitor'
    # is available when the scheduler is updated, thus stopping
    # training if not found. If set to `False`, it will only produce a warning
    "strict": True,
    # If using the `LearningRateMonitor` callback to monitor the
    # learning rate progress, this keyword can be used to specify
    # a custom logged name
    "name": None,
}

When there are schedulers in which the .step() method is conditioned on a value, such as the torch.optim.lr_scheduler.ReduceLROnPlateau scheduler, Lightning requires that the lr_scheduler_config contains the keyword "monitor" set to the metric name that the scheduler should be conditioned on.

Metrics can be made available to monitor by simply logging it using self.log('metric_to_track', metric_val) in your LightningModule.

Note

The frequency value specified in a dict along with the optimizer key is an int corresponding to the number of sequential batches optimized with the specific optimizer. It should be given to none or to all of the optimizers. There is a difference between passing multiple optimizers in a list, and passing multiple optimizers in dictionaries with a frequency of 1:

This is different from the frequency value specified in the lr_scheduler_config mentioned above.

def configure_optimizers(self):
    optimizer_one = torch.optim.SGD(self.model.parameters(), lr=0.01)
    optimizer_two = torch.optim.SGD(self.model.parameters(), lr=0.01)
    return [
        {"optimizer": optimizer_one, "frequency": 5},
        {"optimizer": optimizer_two, "frequency": 10},
    ]

In this example, the first optimizer will be used for the first 5 steps, the second optimizer for the next 10 steps and that cycle will continue. If an LR scheduler is specified for an optimizer using the lr_scheduler key in the above dict, the scheduler will only be updated when its optimizer is being used.

Examples:

# most cases. no learning rate scheduler
def configure_optimizers(self):
    return Adam(self.parameters(), lr=1e-3)

# multiple optimizer case (e.g.: GAN)
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    return gen_opt, dis_opt

# example with learning rate schedulers
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    dis_sch = CosineAnnealing(dis_opt, T_max=10)
    return [gen_opt, dis_opt], [dis_sch]

# example with step-based learning rate schedulers
# each optimizer has its own scheduler
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    gen_sch = {
        'scheduler': ExponentialLR(gen_opt, 0.99),
        'interval': 'step'  # called after each training step
    }
    dis_sch = CosineAnnealing(dis_opt, T_max=10) # called every epoch
    return [gen_opt, dis_opt], [gen_sch, dis_sch]

# example with optimizer frequencies
# see training procedure in `Improved Training of Wasserstein GANs`, Algorithm 1
# https://arxiv.org/abs/1704.00028
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    n_critic = 5
    return (
        {'optimizer': dis_opt, 'frequency': n_critic},
        {'optimizer': gen_opt, 'frequency': 1}
    )

Note

Some things to know:

Lightning calls .backward() and .step() on each optimizer and learning rate scheduler as needed.
If you use 16-bit precision (precision=16), Lightning will automatically handle the optimizers.
If you use multiple optimizers, training_step() will have an additional optimizer_idx parameter.
If you use torch.optim.LBFGS, Lightning handles the closure function automatically for you.
If you use multiple optimizers, gradients will be calculated only for the parameters of current optimizer at each training step.
If you need to control how often those optimizers step or override the default .step() schedule, override the optimizer_step() hook.

precision: int

prepare_data_per_node: bool

train_step_gen(training_batch: reagent.core.types.PreprocessedRankingInput, batch_idx: int): Implement training step as generator here

training: bool

use_amp: bool

validation_epoch_end(outputs)

Called at the end of the validation epoch with the outputs of all validation steps.

# the pseudocode for these calls
val_outs = []
for val_batch in val_data:
    out = validation_step(val_batch)
    val_outs.append(out)
validation_epoch_end(val_outs)

Parameters: outputs – List of outputs you defined in validation_step(), or if there are multiple dataloaders, a list containing a list of outputs for each dataloader.
Returns: None

Note

If you didn’t define a validation_step(), this won’t be called.

Examples

With a single dataloader:

def validation_epoch_end(self, val_step_outputs):
    for out in val_step_outputs:
        ...

With multiple dataloaders, outputs will be a list of lists. The outer list contains one entry per dataloader, while the inner list contains the individual outputs of each validation step for that dataloader.

def validation_epoch_end(self, outputs):
    for dataloader_output_result in outputs:
        dataloader_outs = dataloader_output_result.dataloader_i_outputs

    self.log("final_metric", final_value)

validation_step(batch: reagent.core.types.PreprocessedRankingInput, batch_idx: int)

Operates on a single batch of data from the validation set. In this step you’d might generate examples or calculate anything of interest like accuracy.

# the pseudocode for these calls
val_outs = []
for val_batch in val_data:
    out = validation_step(val_batch)
    val_outs.append(out)
validation_epoch_end(val_outs)

Parameters

batch (Tensor | (Tensor, …) | [Tensor, …]) – The output of your DataLoader. A tensor, tuple or list.
batch_idx (int) – The index of this batch
dataloader_idx (int) – The index of the dataloader that produced this batch (only if multiple val dataloaders used)

Returns

Any object or value
None - Validation will skip to the next batch

# pseudocode of order
val_outs = []
for val_batch in val_data:
    out = validation_step(val_batch)
    if defined("validation_step_end"):
        out = validation_step_end(out)
    val_outs.append(out)
val_outs = validation_epoch_end(val_outs)

# if you have one val dataloader:
def validation_step(self, batch, batch_idx):
    ...


# if you have multiple val dataloaders:
def validation_step(self, batch, batch_idx, dataloader_idx):
    ...

Examples:

# CASE 1: A single validation dataset
def validation_step(self, batch, batch_idx):
    x, y = batch

    # implement your own
    out = self(x)
    loss = self.loss(out, y)

    # log 6 example images
    # or generated text... or whatever
    sample_imgs = x[:6]
    grid = torchvision.utils.make_grid(sample_imgs)
    self.logger.experiment.add_image('example_images', grid, 0)

    # calculate acc
    labels_hat = torch.argmax(out, dim=1)
    val_acc = torch.sum(y == labels_hat).item() / (len(y) * 1.0)

    # log the outputs!
    self.log_dict({'val_loss': loss, 'val_acc': val_acc})

If you pass in multiple val dataloaders, validation_step() will have an additional argument.

# CASE 2: multiple validation dataloaders
def validation_step(self, batch, batch_idx, dataloader_idx):
    # dataloader_idx tells you which dataset this is.
    ...

Note

If you don’t need to validate you don’t need to implement this method.

Note

When the validation_step() is called, the model has been put in eval mode and PyTorch gradients have been disabled. At the end of validation, the model goes back to training mode and gradients are enabled.

warm_start_components()

reagent.training.rl_trainer_pytorch module

class reagent.training.rl_trainer_pytorch.RLTrainerMixin

Bases: object

ACTION_NOT_POSSIBLE_VAL = -1000000000.0

property gamma: float

property maxq_learning: bool

property multi_steps: Optional[int]

rl_parameters: reagent.core.parameters.RLParameters

property rl_temperature: float

property tau: float

property use_seq_num_diff_as_time_diff: bool

reagent.training.sac_trainer module

class reagent.training.sac_trainer.CRRWeightFn(indicator_fn_threshold: Optional[float] = None, exponent_beta: Optional[float] = None, exponent_clamp: Optional[float] = None)

Bases: object

exponent_beta: Optional[float] = None

exponent_clamp: Optional[float] = None

get_weight_from_advantage(advantage)

indicator_fn_threshold: Optional[float] = None

class reagent.training.sac_trainer.SACTrainer(actor_network, q1_network, q2_network=None, value_network=None, rl: reagent.core.parameters.RLParameters = Field(name='rl',type=<class 'reagent.core.parameters.RLParameters'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<class 'reagent.core.parameters.RLParameters'>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), q_network_optimizer: reagent.optimizer.union.Optimizer__Union = Field(name='q_network_optimizer',type=<class 'reagent.optimizer.union.Optimizer__Union'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<bound method Optimizer__Union.default of <class 'reagent.optimizer.union.Optimizer__Union'>>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), value_network_optimizer: reagent.optimizer.union.Optimizer__Union = Field(name='value_network_optimizer',type=<class 'reagent.optimizer.union.Optimizer__Union'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<bound method Optimizer__Union.default of <class 'reagent.optimizer.union.Optimizer__Union'>>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), actor_network_optimizer: reagent.optimizer.union.Optimizer__Union = Field(name='actor_network_optimizer',type=<class 'reagent.optimizer.union.Optimizer__Union'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<bound method Optimizer__Union.default of <class 'reagent.optimizer.union.Optimizer__Union'>>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), alpha_optimizer: Optional[reagent.optimizer.union.Optimizer__Union] = Field(name='alpha_optimizer',type=typing.Optional[reagent.optimizer.union.Optimizer__Union],default=<dataclasses._MISSING_TYPE object>,default_factory=<bound method Optimizer__Union.default of <class 'reagent.optimizer.union.Optimizer__Union'>>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), minibatch_size: int = 1024, entropy_temperature: float = 0.01, logged_action_uniform_prior: bool = True, target_entropy: float = -1.0, action_embedding_kld_weight: Optional[float] = None, apply_kld_on_mean: bool = False, action_embedding_mean: Optional[List[float]] = None, action_embedding_variance: Optional[List[float]] = None, crr_config: Optional[reagent.training.sac_trainer.CRRWeightFn] = None, backprop_through_log_prob: bool = True)

Bases: reagent.training.rl_trainer_pytorch.RLTrainerMixin, reagent.training.reagent_lightning_module.ReAgentLightningModule

Soft Actor-Critic trainer as described in https://arxiv.org/pdf/1801.01290

The actor is assumed to implement reparameterization trick.

configure_optimizers()

Choose what optimizers and learning-rate schedulers to use in your optimization. Normally you’d need one. But in the case of GANs or similar you might have multiple.

Returns

Any of these 6 options.

Single optimizer.
List or Tuple of optimizers.
Two lists - The first list has multiple optimizers, and the second has multiple LR schedulers (or multiple lr_scheduler_config).
Dictionary, with an "optimizer" key, and (optionally) a "lr_scheduler" key whose value is a single LR scheduler or lr_scheduler_config.
Tuple of dictionaries as described above, with an optional "frequency" key.
None - Fit will run without any optimizer.

The lr_scheduler_config is a dictionary which contains the scheduler and its associated configuration. The default configuration is shown below.

lr_scheduler_config = {
    # REQUIRED: The scheduler instance
    "scheduler": lr_scheduler,
    # The unit of the scheduler's step size, could also be 'step'.
    # 'epoch' updates the scheduler on epoch end whereas 'step'
    # updates it after a optimizer update.
    "interval": "epoch",
    # How many epochs/steps should pass between calls to
    # `scheduler.step()`. 1 corresponds to updating the learning
    # rate after every epoch/step.
    "frequency": 1,
    # Metric to to monitor for schedulers like `ReduceLROnPlateau`
    "monitor": "val_loss",
    # If set to `True`, will enforce that the value specified 'monitor'
    # is available when the scheduler is updated, thus stopping
    # training if not found. If set to `False`, it will only produce a warning
    "strict": True,
    # If using the `LearningRateMonitor` callback to monitor the
    # learning rate progress, this keyword can be used to specify
    # a custom logged name
    "name": None,
}

When there are schedulers in which the .step() method is conditioned on a value, such as the torch.optim.lr_scheduler.ReduceLROnPlateau scheduler, Lightning requires that the lr_scheduler_config contains the keyword "monitor" set to the metric name that the scheduler should be conditioned on.

Metrics can be made available to monitor by simply logging it using self.log('metric_to_track', metric_val) in your LightningModule.

Note

The frequency value specified in a dict along with the optimizer key is an int corresponding to the number of sequential batches optimized with the specific optimizer. It should be given to none or to all of the optimizers. There is a difference between passing multiple optimizers in a list, and passing multiple optimizers in dictionaries with a frequency of 1:

This is different from the frequency value specified in the lr_scheduler_config mentioned above.

def configure_optimizers(self):
    optimizer_one = torch.optim.SGD(self.model.parameters(), lr=0.01)
    optimizer_two = torch.optim.SGD(self.model.parameters(), lr=0.01)
    return [
        {"optimizer": optimizer_one, "frequency": 5},
        {"optimizer": optimizer_two, "frequency": 10},
    ]

In this example, the first optimizer will be used for the first 5 steps, the second optimizer for the next 10 steps and that cycle will continue. If an LR scheduler is specified for an optimizer using the lr_scheduler key in the above dict, the scheduler will only be updated when its optimizer is being used.

Examples:

# most cases. no learning rate scheduler
def configure_optimizers(self):
    return Adam(self.parameters(), lr=1e-3)

# multiple optimizer case (e.g.: GAN)
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    return gen_opt, dis_opt

# example with learning rate schedulers
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    dis_sch = CosineAnnealing(dis_opt, T_max=10)
    return [gen_opt, dis_opt], [dis_sch]

# example with step-based learning rate schedulers
# each optimizer has its own scheduler
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    gen_sch = {
        'scheduler': ExponentialLR(gen_opt, 0.99),
        'interval': 'step'  # called after each training step
    }
    dis_sch = CosineAnnealing(dis_opt, T_max=10) # called every epoch
    return [gen_opt, dis_opt], [gen_sch, dis_sch]

# example with optimizer frequencies
# see training procedure in `Improved Training of Wasserstein GANs`, Algorithm 1
# https://arxiv.org/abs/1704.00028
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    n_critic = 5
    return (
        {'optimizer': dis_opt, 'frequency': n_critic},
        {'optimizer': gen_opt, 'frequency': 1}
    )

Note

Some things to know:

Lightning calls .backward() and .step() on each optimizer and learning rate scheduler as needed.
If you use 16-bit precision (precision=16), Lightning will automatically handle the optimizers.
If you use multiple optimizers, training_step() will have an additional optimizer_idx parameter.
If you use torch.optim.LBFGS, Lightning handles the closure function automatically for you.
If you use multiple optimizers, gradients will be calculated only for the parameters of current optimizer at each training step.
If you need to control how often those optimizers step or override the default .step() schedule, override the optimizer_step() hook.

rl_parameters: reagent.core.parameters.RLParameters

train_step_gen(training_batch: reagent.core.types.PolicyNetworkInput, batch_idx: int): IMPORTANT: the input action here is assumed to match the range of the output of the actor.

reagent.training.slate_q_trainer module

class reagent.training.slate_q_trainer.NextSlateValueNormMethod(value)

Bases: enum.Enum

The Q value of the current slate item is the sum of the item’s short-term reward and the normalized sum of all item Q-values on the next slate. We can normalize the sum by either the current slate size (NORM_BY_CURRENT_SLATE_SIZE) or the next slate size (NORM_BY_NEXT_SLATE_SIZE). This enum distinguishes between these two different ways of normalizing the next slate value.

NORM_BY_CURRENT_SLATE_SIZE = 'norm_by_current_slate_size'

NORM_BY_NEXT_SLATE_SIZE = 'norm_by_next_slate_size'

class reagent.training.slate_q_trainer.SlateQTrainer(q_network, q_network_target, slate_size, rl: reagent.core.parameters.RLParameters = Field(name='rl', type=<class 'reagent.core.parameters.RLParameters'>, default=<dataclasses._MISSING_TYPE object>, default_factory=<function SlateQTrainer.<lambda>>, init=True, repr=True, hash=None, compare=True, metadata=mappingproxy({}), _field_type=_FIELD), optimizer: reagent.optimizer.union.Optimizer__Union = Field(name='optimizer', type=<class 'reagent.optimizer.union.Optimizer__Union'>, default=<dataclasses._MISSING_TYPE object>, default_factory=<bound method Optimizer__Union.default of <class 'reagent.optimizer.union.Optimizer__Union'>>, init=True, repr=True, hash=None, compare=True, metadata=mappingproxy({}), _field_type=_FIELD), slate_opt_parameters: Optional[reagent.core.parameters.SlateOptParameters] = None, discount_time_scale: Optional[float] = None, single_selection: bool = True, next_slate_value_norm_method: reagent.training.slate_q_trainer.NextSlateValueNormMethod = NextSlateValueNormMethod.NORM_BY_CURRENT_SLATE_SIZE, minibatch_size: int = 1024, evaluation: reagent.core.parameters.EvaluationParameters = Field(name='evaluation', type=<class 'reagent.core.parameters.EvaluationParameters'>, default=<dataclasses._MISSING_TYPE object>, default_factory=<function SlateQTrainer.<lambda>>, init=True, repr=True, hash=None, compare=True, metadata=mappingproxy({}), _field_type=_FIELD))

Bases: reagent.training.rl_trainer_pytorch.RLTrainerMixin, reagent.training.reagent_lightning_module.ReAgentLightningModule

configure_optimizers()

Choose what optimizers and learning-rate schedulers to use in your optimization. Normally you’d need one. But in the case of GANs or similar you might have multiple.

Returns

Any of these 6 options.

Single optimizer.
List or Tuple of optimizers.
Two lists - The first list has multiple optimizers, and the second has multiple LR schedulers (or multiple lr_scheduler_config).
Dictionary, with an "optimizer" key, and (optionally) a "lr_scheduler" key whose value is a single LR scheduler or lr_scheduler_config.
Tuple of dictionaries as described above, with an optional "frequency" key.
None - Fit will run without any optimizer.

The lr_scheduler_config is a dictionary which contains the scheduler and its associated configuration. The default configuration is shown below.

lr_scheduler_config = {
    # REQUIRED: The scheduler instance
    "scheduler": lr_scheduler,
    # The unit of the scheduler's step size, could also be 'step'.
    # 'epoch' updates the scheduler on epoch end whereas 'step'
    # updates it after a optimizer update.
    "interval": "epoch",
    # How many epochs/steps should pass between calls to
    # `scheduler.step()`. 1 corresponds to updating the learning
    # rate after every epoch/step.
    "frequency": 1,
    # Metric to to monitor for schedulers like `ReduceLROnPlateau`
    "monitor": "val_loss",
    # If set to `True`, will enforce that the value specified 'monitor'
    # is available when the scheduler is updated, thus stopping
    # training if not found. If set to `False`, it will only produce a warning
    "strict": True,
    # If using the `LearningRateMonitor` callback to monitor the
    # learning rate progress, this keyword can be used to specify
    # a custom logged name
    "name": None,
}

When there are schedulers in which the .step() method is conditioned on a value, such as the torch.optim.lr_scheduler.ReduceLROnPlateau scheduler, Lightning requires that the lr_scheduler_config contains the keyword "monitor" set to the metric name that the scheduler should be conditioned on.

Metrics can be made available to monitor by simply logging it using self.log('metric_to_track', metric_val) in your LightningModule.

Note

The frequency value specified in a dict along with the optimizer key is an int corresponding to the number of sequential batches optimized with the specific optimizer. It should be given to none or to all of the optimizers. There is a difference between passing multiple optimizers in a list, and passing multiple optimizers in dictionaries with a frequency of 1:

This is different from the frequency value specified in the lr_scheduler_config mentioned above.

def configure_optimizers(self):
    optimizer_one = torch.optim.SGD(self.model.parameters(), lr=0.01)
    optimizer_two = torch.optim.SGD(self.model.parameters(), lr=0.01)
    return [
        {"optimizer": optimizer_one, "frequency": 5},
        {"optimizer": optimizer_two, "frequency": 10},
    ]

In this example, the first optimizer will be used for the first 5 steps, the second optimizer for the next 10 steps and that cycle will continue. If an LR scheduler is specified for an optimizer using the lr_scheduler key in the above dict, the scheduler will only be updated when its optimizer is being used.

Examples:

# most cases. no learning rate scheduler
def configure_optimizers(self):
    return Adam(self.parameters(), lr=1e-3)

# multiple optimizer case (e.g.: GAN)
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    return gen_opt, dis_opt

# example with learning rate schedulers
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    dis_sch = CosineAnnealing(dis_opt, T_max=10)
    return [gen_opt, dis_opt], [dis_sch]

# example with step-based learning rate schedulers
# each optimizer has its own scheduler
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    gen_sch = {
        'scheduler': ExponentialLR(gen_opt, 0.99),
        'interval': 'step'  # called after each training step
    }
    dis_sch = CosineAnnealing(dis_opt, T_max=10) # called every epoch
    return [gen_opt, dis_opt], [gen_sch, dis_sch]

# example with optimizer frequencies
# see training procedure in `Improved Training of Wasserstein GANs`, Algorithm 1
# https://arxiv.org/abs/1704.00028
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    n_critic = 5
    return (
        {'optimizer': dis_opt, 'frequency': n_critic},
        {'optimizer': gen_opt, 'frequency': 1}
    )

Note

Some things to know:

Lightning calls .backward() and .step() on each optimizer and learning rate scheduler as needed.
If you use 16-bit precision (precision=16), Lightning will automatically handle the optimizers.
If you use multiple optimizers, training_step() will have an additional optimizer_idx parameter.
If you use torch.optim.LBFGS, Lightning handles the closure function automatically for you.
If you use multiple optimizers, gradients will be calculated only for the parameters of current optimizer at each training step.
If you need to control how often those optimizers step or override the default .step() schedule, override the optimizer_step() hook.

rl_parameters: reagent.core.parameters.RLParameters

train_step_gen(training_batch: reagent.core.types.SlateQInput, batch_idx: int): Implement training step as generator here

reagent.training.td3_trainer module

class reagent.training.td3_trainer.TD3Trainer(actor_network, q1_network, q2_network=None, rl: reagent.core.parameters.RLParameters = Field(name='rl', type=<class 'reagent.core.parameters.RLParameters'>, default=<dataclasses._MISSING_TYPE object>, default_factory=<class 'reagent.core.parameters.RLParameters'>, init=True, repr=True, hash=None, compare=True, metadata=mappingproxy({}), _field_type=_FIELD), q_network_optimizer: reagent.optimizer.union.Optimizer__Union = Field(name='q_network_optimizer', type=<class 'reagent.optimizer.union.Optimizer__Union'>, default=<dataclasses._MISSING_TYPE object>, default_factory=<bound method Optimizer__Union.default of <class 'reagent.optimizer.union.Optimizer__Union'>>, init=True, repr=True, hash=None, compare=True, metadata=mappingproxy({}), _field_type=_FIELD), actor_network_optimizer: reagent.optimizer.union.Optimizer__Union = Field(name='actor_network_optimizer', type=<class 'reagent.optimizer.union.Optimizer__Union'>, default=<dataclasses._MISSING_TYPE object>, default_factory=<bound method Optimizer__Union.default of <class 'reagent.optimizer.union.Optimizer__Union'>>, init=True, repr=True, hash=None, compare=True, metadata=mappingproxy({}), _field_type=_FIELD), minibatch_size: int = 64, noise_variance: float = 0.2, noise_clip: float = 0.5, delayed_policy_update: int = 2, minibatches_per_step: int = 1)

Bases: reagent.training.rl_trainer_pytorch.RLTrainerMixin, reagent.training.reagent_lightning_module.ReAgentLightningModule

Twin Delayed Deep Deterministic Policy Gradient algorithm trainer as described in https://arxiv.org/pdf/1802.09477

configure_optimizers()

Choose what optimizers and learning-rate schedulers to use in your optimization. Normally you’d need one. But in the case of GANs or similar you might have multiple.

Returns

Any of these 6 options.

Single optimizer.
List or Tuple of optimizers.
Two lists - The first list has multiple optimizers, and the second has multiple LR schedulers (or multiple lr_scheduler_config).
Dictionary, with an "optimizer" key, and (optionally) a "lr_scheduler" key whose value is a single LR scheduler or lr_scheduler_config.
Tuple of dictionaries as described above, with an optional "frequency" key.
None - Fit will run without any optimizer.

The lr_scheduler_config is a dictionary which contains the scheduler and its associated configuration. The default configuration is shown below.

lr_scheduler_config = {
    # REQUIRED: The scheduler instance
    "scheduler": lr_scheduler,
    # The unit of the scheduler's step size, could also be 'step'.
    # 'epoch' updates the scheduler on epoch end whereas 'step'
    # updates it after a optimizer update.
    "interval": "epoch",
    # How many epochs/steps should pass between calls to
    # `scheduler.step()`. 1 corresponds to updating the learning
    # rate after every epoch/step.
    "frequency": 1,
    # Metric to to monitor for schedulers like `ReduceLROnPlateau`
    "monitor": "val_loss",
    # If set to `True`, will enforce that the value specified 'monitor'
    # is available when the scheduler is updated, thus stopping
    # training if not found. If set to `False`, it will only produce a warning
    "strict": True,
    # If using the `LearningRateMonitor` callback to monitor the
    # learning rate progress, this keyword can be used to specify
    # a custom logged name
    "name": None,
}

When there are schedulers in which the .step() method is conditioned on a value, such as the torch.optim.lr_scheduler.ReduceLROnPlateau scheduler, Lightning requires that the lr_scheduler_config contains the keyword "monitor" set to the metric name that the scheduler should be conditioned on.

Metrics can be made available to monitor by simply logging it using self.log('metric_to_track', metric_val) in your LightningModule.

Note

The frequency value specified in a dict along with the optimizer key is an int corresponding to the number of sequential batches optimized with the specific optimizer. It should be given to none or to all of the optimizers. There is a difference between passing multiple optimizers in a list, and passing multiple optimizers in dictionaries with a frequency of 1:

This is different from the frequency value specified in the lr_scheduler_config mentioned above.

def configure_optimizers(self):
    optimizer_one = torch.optim.SGD(self.model.parameters(), lr=0.01)
    optimizer_two = torch.optim.SGD(self.model.parameters(), lr=0.01)
    return [
        {"optimizer": optimizer_one, "frequency": 5},
        {"optimizer": optimizer_two, "frequency": 10},
    ]

In this example, the first optimizer will be used for the first 5 steps, the second optimizer for the next 10 steps and that cycle will continue. If an LR scheduler is specified for an optimizer using the lr_scheduler key in the above dict, the scheduler will only be updated when its optimizer is being used.

Examples:

# most cases. no learning rate scheduler
def configure_optimizers(self):
    return Adam(self.parameters(), lr=1e-3)

# multiple optimizer case (e.g.: GAN)
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    return gen_opt, dis_opt

# example with learning rate schedulers
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    dis_sch = CosineAnnealing(dis_opt, T_max=10)
    return [gen_opt, dis_opt], [dis_sch]

# example with step-based learning rate schedulers
# each optimizer has its own scheduler
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    gen_sch = {
        'scheduler': ExponentialLR(gen_opt, 0.99),
        'interval': 'step'  # called after each training step
    }
    dis_sch = CosineAnnealing(dis_opt, T_max=10) # called every epoch
    return [gen_opt, dis_opt], [gen_sch, dis_sch]

# example with optimizer frequencies
# see training procedure in `Improved Training of Wasserstein GANs`, Algorithm 1
# https://arxiv.org/abs/1704.00028
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    n_critic = 5
    return (
        {'optimizer': dis_opt, 'frequency': n_critic},
        {'optimizer': gen_opt, 'frequency': 1}
    )

Note

Some things to know:

Lightning calls .backward() and .step() on each optimizer and learning rate scheduler as needed.
If you use 16-bit precision (precision=16), Lightning will automatically handle the optimizers.
If you use multiple optimizers, training_step() will have an additional optimizer_idx parameter.
If you use torch.optim.LBFGS, Lightning handles the closure function automatically for you.
If you use multiple optimizers, gradients will be calculated only for the parameters of current optimizer at each training step.
If you need to control how often those optimizers step or override the default .step() schedule, override the optimizer_step() hook.

rl_parameters: reagent.core.parameters.RLParameters

train_step_gen(training_batch: reagent.core.types.PolicyNetworkInput, batch_idx: int): IMPORTANT: the input action here is assumed to be preprocessed to match the range of the output of the actor.

reagent.training.utils module

reagent.training.utils.discounted_returns(rewards: torch.Tensor, gamma: float = 0) → torch.Tensor: Perform rollout to compute reward to go and do a baseline subtraction.

reagent.training.utils.gen_permutations(seq_len: int, num_action: int) → torch.Tensor: generate all seq_len permutations for a given action set the return shape is (SEQ_LEN, PERM_NUM, ACTION_DIM)

reagent.training.utils.rescale_actions(actions: torch.Tensor, new_min: torch.Tensor, new_max: torch.Tensor, prev_min: torch.Tensor, prev_max: torch.Tensor) → torch.Tensor: Scale from [prev_min, prev_max] to [new_min, new_max]

reagent.training.utils.whiten(x: torch.Tensor, subtract_mean: bool) → torch.Tensor

Module contents

class reagent.training.BanditRewardNetTrainer(reward_net: reagent.models.base.ModelBase, optimizer: reagent.optimizer.union.Optimizer__Union = Field(name=None, type=None, default=<dataclasses._MISSING_TYPE object>, default_factory=<bound method Optimizer__Union.default of <class 'reagent.optimizer.union.Optimizer__Union'>>, init=True, repr=True, hash=None, compare=True, metadata=mappingproxy({}), _field_type=None), loss_type: reagent.training.reward_network_trainer.LossFunction = LossFunction.MSE, reward_ignore_threshold: Optional[float] = None, weighted_by_inverse_propensity: bool = False)

Bases: reagent.training.reagent_lightning_module.ReAgentLightningModule

allow_zero_length_dataloader_with_multiple_devices: bool

configure_optimizers()

Choose what optimizers and learning-rate schedulers to use in your optimization. Normally you’d need one. But in the case of GANs or similar you might have multiple.

Returns

Any of these 6 options.

Single optimizer.
List or Tuple of optimizers.
Two lists - The first list has multiple optimizers, and the second has multiple LR schedulers (or multiple lr_scheduler_config).
Dictionary, with an "optimizer" key, and (optionally) a "lr_scheduler" key whose value is a single LR scheduler or lr_scheduler_config.
Tuple of dictionaries as described above, with an optional "frequency" key.
None - Fit will run without any optimizer.

The lr_scheduler_config is a dictionary which contains the scheduler and its associated configuration. The default configuration is shown below.

lr_scheduler_config = {
    # REQUIRED: The scheduler instance
    "scheduler": lr_scheduler,
    # The unit of the scheduler's step size, could also be 'step'.
    # 'epoch' updates the scheduler on epoch end whereas 'step'
    # updates it after a optimizer update.
    "interval": "epoch",
    # How many epochs/steps should pass between calls to
    # `scheduler.step()`. 1 corresponds to updating the learning
    # rate after every epoch/step.
    "frequency": 1,
    # Metric to to monitor for schedulers like `ReduceLROnPlateau`
    "monitor": "val_loss",
    # If set to `True`, will enforce that the value specified 'monitor'
    # is available when the scheduler is updated, thus stopping
    # training if not found. If set to `False`, it will only produce a warning
    "strict": True,
    # If using the `LearningRateMonitor` callback to monitor the
    # learning rate progress, this keyword can be used to specify
    # a custom logged name
    "name": None,
}

When there are schedulers in which the .step() method is conditioned on a value, such as the torch.optim.lr_scheduler.ReduceLROnPlateau scheduler, Lightning requires that the lr_scheduler_config contains the keyword "monitor" set to the metric name that the scheduler should be conditioned on.

Metrics can be made available to monitor by simply logging it using self.log('metric_to_track', metric_val) in your LightningModule.

Note

The frequency value specified in a dict along with the optimizer key is an int corresponding to the number of sequential batches optimized with the specific optimizer. It should be given to none or to all of the optimizers. There is a difference between passing multiple optimizers in a list, and passing multiple optimizers in dictionaries with a frequency of 1:

This is different from the frequency value specified in the lr_scheduler_config mentioned above.

def configure_optimizers(self):
    optimizer_one = torch.optim.SGD(self.model.parameters(), lr=0.01)
    optimizer_two = torch.optim.SGD(self.model.parameters(), lr=0.01)
    return [
        {"optimizer": optimizer_one, "frequency": 5},
        {"optimizer": optimizer_two, "frequency": 10},
    ]

In this example, the first optimizer will be used for the first 5 steps, the second optimizer for the next 10 steps and that cycle will continue. If an LR scheduler is specified for an optimizer using the lr_scheduler key in the above dict, the scheduler will only be updated when its optimizer is being used.

Examples:

# most cases. no learning rate scheduler
def configure_optimizers(self):
    return Adam(self.parameters(), lr=1e-3)

# multiple optimizer case (e.g.: GAN)
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    return gen_opt, dis_opt

# example with learning rate schedulers
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    dis_sch = CosineAnnealing(dis_opt, T_max=10)
    return [gen_opt, dis_opt], [dis_sch]

# example with step-based learning rate schedulers
# each optimizer has its own scheduler
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    gen_sch = {
        'scheduler': ExponentialLR(gen_opt, 0.99),
        'interval': 'step'  # called after each training step
    }
    dis_sch = CosineAnnealing(dis_opt, T_max=10) # called every epoch
    return [gen_opt, dis_opt], [gen_sch, dis_sch]

# example with optimizer frequencies
# see training procedure in `Improved Training of Wasserstein GANs`, Algorithm 1
# https://arxiv.org/abs/1704.00028
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    n_critic = 5
    return (
        {'optimizer': dis_opt, 'frequency': n_critic},
        {'optimizer': gen_opt, 'frequency': 1}
    )

Note

Some things to know:

Lightning calls .backward() and .step() on each optimizer and learning rate scheduler as needed.
If you use 16-bit precision (precision=16), Lightning will automatically handle the optimizers.
If you use multiple optimizers, training_step() will have an additional optimizer_idx parameter.
If you use torch.optim.LBFGS, Lightning handles the closure function automatically for you.
If you use multiple optimizers, gradients will be calculated only for the parameters of current optimizer at each training step.
If you need to control how often those optimizers step or override the default .step() schedule, override the optimizer_step() hook.

precision: int

prepare_data_per_node: bool

train_step_gen(training_batch: reagent.core.types.BanditRewardModelInput, batch_idx: int): Implement training step as generator here

training: bool

use_amp: bool

validation_epoch_end(outputs)

Called at the end of the validation epoch with the outputs of all validation steps.

# the pseudocode for these calls
val_outs = []
for val_batch in val_data:
    out = validation_step(val_batch)
    val_outs.append(out)
validation_epoch_end(val_outs)

Parameters: outputs – List of outputs you defined in validation_step(), or if there are multiple dataloaders, a list containing a list of outputs for each dataloader.
Returns: None

Note

If you didn’t define a validation_step(), this won’t be called.

Examples

With a single dataloader:

def validation_epoch_end(self, val_step_outputs):
    for out in val_step_outputs:
        ...

With multiple dataloaders, outputs will be a list of lists. The outer list contains one entry per dataloader, while the inner list contains the individual outputs of each validation step for that dataloader.

def validation_epoch_end(self, outputs):
    for dataloader_output_result in outputs:
        dataloader_outs = dataloader_output_result.dataloader_i_outputs

    self.log("final_metric", final_value)

validation_step(batch: reagent.core.types.BanditRewardModelInput, batch_idx: int)

Operates on a single batch of data from the validation set. In this step you’d might generate examples or calculate anything of interest like accuracy.

# the pseudocode for these calls
val_outs = []
for val_batch in val_data:
    out = validation_step(val_batch)
    val_outs.append(out)
validation_epoch_end(val_outs)

Parameters

batch (Tensor | (Tensor, …) | [Tensor, …]) – The output of your DataLoader. A tensor, tuple or list.
batch_idx (int) – The index of this batch
dataloader_idx (int) – The index of the dataloader that produced this batch (only if multiple val dataloaders used)

Returns

Any object or value
None - Validation will skip to the next batch

# pseudocode of order
val_outs = []
for val_batch in val_data:
    out = validation_step(val_batch)
    if defined("validation_step_end"):
        out = validation_step_end(out)
    val_outs.append(out)
val_outs = validation_epoch_end(val_outs)

# if you have one val dataloader:
def validation_step(self, batch, batch_idx):
    ...


# if you have multiple val dataloaders:
def validation_step(self, batch, batch_idx, dataloader_idx):
    ...

Examples:

# CASE 1: A single validation dataset
def validation_step(self, batch, batch_idx):
    x, y = batch

    # implement your own
    out = self(x)
    loss = self.loss(out, y)

    # log 6 example images
    # or generated text... or whatever
    sample_imgs = x[:6]
    grid = torchvision.utils.make_grid(sample_imgs)
    self.logger.experiment.add_image('example_images', grid, 0)

    # calculate acc
    labels_hat = torch.argmax(out, dim=1)
    val_acc = torch.sum(y == labels_hat).item() / (len(y) * 1.0)

    # log the outputs!
    self.log_dict({'val_loss': loss, 'val_acc': val_acc})

If you pass in multiple val dataloaders, validation_step() will have an additional argument.

# CASE 2: multiple validation dataloaders
def validation_step(self, batch, batch_idx, dataloader_idx):
    # dataloader_idx tells you which dataset this is.
    ...

Note

If you don’t need to validate you don’t need to implement this method.

Note

When the validation_step() is called, the model has been put in eval mode and PyTorch gradients have been disabled. At the end of validation, the model goes back to training mode and gradients are enabled.

class reagent.training.C51Trainer(q_network, q_network_target, actions: List[str] = Field(name='actions',type=typing.List[str],default=<dataclasses._MISSING_TYPE object>,default_factory=<class 'list'>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), rl: reagent.core.parameters.RLParameters = Field(name='rl',type=<class 'reagent.core.parameters.RLParameters'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<class 'reagent.core.parameters.RLParameters'>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), double_q_learning: bool = True, minibatch_size: int = 1024, minibatches_per_step: int = 1, num_atoms: int = 51, qmin: float = -100, qmax: float = 200, optimizer: reagent.optimizer.union.Optimizer__Union = Field(name='optimizer',type=<class 'reagent.optimizer.union.Optimizer__Union'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<bound method Optimizer__Union.default of <class 'reagent.optimizer.union.Optimizer__Union'>>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD))

Bases: reagent.training.rl_trainer_pytorch.RLTrainerMixin, reagent.training.reagent_lightning_module.ReAgentLightningModule

Implementation of 51 Categorical DQN (C51)

See https://arxiv.org/abs/1707.06887 for details

allow_zero_length_dataloader_with_multiple_devices: bool

argmax_with_mask(q_values, possible_actions_mask)

boost_rewards(rewards: torch.Tensor, actions: torch.Tensor) → torch.Tensor

configure_optimizers()

Choose what optimizers and learning-rate schedulers to use in your optimization. Normally you’d need one. But in the case of GANs or similar you might have multiple.

Returns

Any of these 6 options.

Single optimizer.
List or Tuple of optimizers.
Two lists - The first list has multiple optimizers, and the second has multiple LR schedulers (or multiple lr_scheduler_config).
Dictionary, with an "optimizer" key, and (optionally) a "lr_scheduler" key whose value is a single LR scheduler or lr_scheduler_config.
Tuple of dictionaries as described above, with an optional "frequency" key.
None - Fit will run without any optimizer.

The lr_scheduler_config is a dictionary which contains the scheduler and its associated configuration. The default configuration is shown below.

lr_scheduler_config = {
    # REQUIRED: The scheduler instance
    "scheduler": lr_scheduler,
    # The unit of the scheduler's step size, could also be 'step'.
    # 'epoch' updates the scheduler on epoch end whereas 'step'
    # updates it after a optimizer update.
    "interval": "epoch",
    # How many epochs/steps should pass between calls to
    # `scheduler.step()`. 1 corresponds to updating the learning
    # rate after every epoch/step.
    "frequency": 1,
    # Metric to to monitor for schedulers like `ReduceLROnPlateau`
    "monitor": "val_loss",
    # If set to `True`, will enforce that the value specified 'monitor'
    # is available when the scheduler is updated, thus stopping
    # training if not found. If set to `False`, it will only produce a warning
    "strict": True,
    # If using the `LearningRateMonitor` callback to monitor the
    # learning rate progress, this keyword can be used to specify
    # a custom logged name
    "name": None,
}

When there are schedulers in which the .step() method is conditioned on a value, such as the torch.optim.lr_scheduler.ReduceLROnPlateau scheduler, Lightning requires that the lr_scheduler_config contains the keyword "monitor" set to the metric name that the scheduler should be conditioned on.

Metrics can be made available to monitor by simply logging it using self.log('metric_to_track', metric_val) in your LightningModule.

Note

The frequency value specified in a dict along with the optimizer key is an int corresponding to the number of sequential batches optimized with the specific optimizer. It should be given to none or to all of the optimizers. There is a difference between passing multiple optimizers in a list, and passing multiple optimizers in dictionaries with a frequency of 1:

This is different from the frequency value specified in the lr_scheduler_config mentioned above.

def configure_optimizers(self):
    optimizer_one = torch.optim.SGD(self.model.parameters(), lr=0.01)
    optimizer_two = torch.optim.SGD(self.model.parameters(), lr=0.01)
    return [
        {"optimizer": optimizer_one, "frequency": 5},
        {"optimizer": optimizer_two, "frequency": 10},
    ]

In this example, the first optimizer will be used for the first 5 steps, the second optimizer for the next 10 steps and that cycle will continue. If an LR scheduler is specified for an optimizer using the lr_scheduler key in the above dict, the scheduler will only be updated when its optimizer is being used.

Examples:

# most cases. no learning rate scheduler
def configure_optimizers(self):
    return Adam(self.parameters(), lr=1e-3)

# multiple optimizer case (e.g.: GAN)
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    return gen_opt, dis_opt

# example with learning rate schedulers
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    dis_sch = CosineAnnealing(dis_opt, T_max=10)
    return [gen_opt, dis_opt], [dis_sch]

# example with step-based learning rate schedulers
# each optimizer has its own scheduler
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    gen_sch = {
        'scheduler': ExponentialLR(gen_opt, 0.99),
        'interval': 'step'  # called after each training step
    }
    dis_sch = CosineAnnealing(dis_opt, T_max=10) # called every epoch
    return [gen_opt, dis_opt], [gen_sch, dis_sch]

# example with optimizer frequencies
# see training procedure in `Improved Training of Wasserstein GANs`, Algorithm 1
# https://arxiv.org/abs/1704.00028
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    n_critic = 5
    return (
        {'optimizer': dis_opt, 'frequency': n_critic},
        {'optimizer': gen_opt, 'frequency': 1}
    )

Note

Some things to know:

Lightning calls .backward() and .step() on each optimizer and learning rate scheduler as needed.
If you use 16-bit precision (precision=16), Lightning will automatically handle the optimizers.
If you use multiple optimizers, training_step() will have an additional optimizer_idx parameter.
If you use torch.optim.LBFGS, Lightning handles the closure function automatically for you.
If you use multiple optimizers, gradients will be calculated only for the parameters of current optimizer at each training step.
If you need to control how often those optimizers step or override the default .step() schedule, override the optimizer_step() hook.

precision: int

prepare_data_per_node: bool

rl_parameters: reagent.core.parameters.RLParameters

train_step_gen(training_batch: reagent.core.types.DiscreteDqnInput, batch_idx: int): Implement training step as generator here

training: bool

use_amp: bool

class reagent.training.C51TrainerParameters(actions: List[str] = <factory>, rl: reagent.core.parameters.RLParameters = <factory>, double_q_learning: bool = True, minibatch_size: int = 1024, minibatches_per_step: int = 1, num_atoms: int = 51, qmin: float = -100, qmax: float = 200, optimizer: reagent.optimizer.union.Optimizer__Union = <factory>)

Bases: object

actions: List[str]

asdict()

double_q_learning: bool = True

minibatch_size: int = 1024

minibatches_per_step: int = 1

num_atoms: int = 51

optimizer: reagent.optimizer.union.Optimizer__Union

qmax: float = 200

qmin: float = -100

rl: reagent.core.parameters.RLParameters

class reagent.training.CEMTrainer(cem_planner_network: reagent.models.cem_planner.CEMPlannerNetwork, world_model_trainers: List[reagent.training.world_model.mdnrnn_trainer.MDNRNNTrainer], parameters: reagent.core.parameters.CEMTrainerParameters)

Bases: reagent.training.reagent_lightning_module.ReAgentLightningModule

allow_zero_length_dataloader_with_multiple_devices: bool

configure_optimizers()

Choose what optimizers and learning-rate schedulers to use in your optimization. Normally you’d need one. But in the case of GANs or similar you might have multiple.

Returns

Any of these 6 options.

Single optimizer.
List or Tuple of optimizers.
Two lists - The first list has multiple optimizers, and the second has multiple LR schedulers (or multiple lr_scheduler_config).
Dictionary, with an "optimizer" key, and (optionally) a "lr_scheduler" key whose value is a single LR scheduler or lr_scheduler_config.
Tuple of dictionaries as described above, with an optional "frequency" key.
None - Fit will run without any optimizer.

The lr_scheduler_config is a dictionary which contains the scheduler and its associated configuration. The default configuration is shown below.

lr_scheduler_config = {
    # REQUIRED: The scheduler instance
    "scheduler": lr_scheduler,
    # The unit of the scheduler's step size, could also be 'step'.
    # 'epoch' updates the scheduler on epoch end whereas 'step'
    # updates it after a optimizer update.
    "interval": "epoch",
    # How many epochs/steps should pass between calls to
    # `scheduler.step()`. 1 corresponds to updating the learning
    # rate after every epoch/step.
    "frequency": 1,
    # Metric to to monitor for schedulers like `ReduceLROnPlateau`
    "monitor": "val_loss",
    # If set to `True`, will enforce that the value specified 'monitor'
    # is available when the scheduler is updated, thus stopping
    # training if not found. If set to `False`, it will only produce a warning
    "strict": True,
    # If using the `LearningRateMonitor` callback to monitor the
    # learning rate progress, this keyword can be used to specify
    # a custom logged name
    "name": None,
}

When there are schedulers in which the .step() method is conditioned on a value, such as the torch.optim.lr_scheduler.ReduceLROnPlateau scheduler, Lightning requires that the lr_scheduler_config contains the keyword "monitor" set to the metric name that the scheduler should be conditioned on.

Metrics can be made available to monitor by simply logging it using self.log('metric_to_track', metric_val) in your LightningModule.

Note

The frequency value specified in a dict along with the optimizer key is an int corresponding to the number of sequential batches optimized with the specific optimizer. It should be given to none or to all of the optimizers. There is a difference between passing multiple optimizers in a list, and passing multiple optimizers in dictionaries with a frequency of 1:

This is different from the frequency value specified in the lr_scheduler_config mentioned above.

def configure_optimizers(self):
    optimizer_one = torch.optim.SGD(self.model.parameters(), lr=0.01)
    optimizer_two = torch.optim.SGD(self.model.parameters(), lr=0.01)
    return [
        {"optimizer": optimizer_one, "frequency": 5},
        {"optimizer": optimizer_two, "frequency": 10},
    ]

In this example, the first optimizer will be used for the first 5 steps, the second optimizer for the next 10 steps and that cycle will continue. If an LR scheduler is specified for an optimizer using the lr_scheduler key in the above dict, the scheduler will only be updated when its optimizer is being used.

Examples:

# most cases. no learning rate scheduler
def configure_optimizers(self):
    return Adam(self.parameters(), lr=1e-3)

# multiple optimizer case (e.g.: GAN)
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    return gen_opt, dis_opt

# example with learning rate schedulers
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    dis_sch = CosineAnnealing(dis_opt, T_max=10)
    return [gen_opt, dis_opt], [dis_sch]

# example with step-based learning rate schedulers
# each optimizer has its own scheduler
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    gen_sch = {
        'scheduler': ExponentialLR(gen_opt, 0.99),
        'interval': 'step'  # called after each training step
    }
    dis_sch = CosineAnnealing(dis_opt, T_max=10) # called every epoch
    return [gen_opt, dis_opt], [gen_sch, dis_sch]

# example with optimizer frequencies
# see training procedure in `Improved Training of Wasserstein GANs`, Algorithm 1
# https://arxiv.org/abs/1704.00028
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    n_critic = 5
    return (
        {'optimizer': dis_opt, 'frequency': n_critic},
        {'optimizer': gen_opt, 'frequency': 1}
    )

Note

Some things to know:

Lightning calls .backward() and .step() on each optimizer and learning rate scheduler as needed.
If you use 16-bit precision (precision=16), Lightning will automatically handle the optimizers.
If you use multiple optimizers, training_step() will have an additional optimizer_idx parameter.
If you use torch.optim.LBFGS, Lightning handles the closure function automatically for you.
If you use multiple optimizers, gradients will be calculated only for the parameters of current optimizer at each training step.
If you need to control how often those optimizers step or override the default .step() schedule, override the optimizer_step() hook.

precision: int

prepare_data_per_node: bool

train_step_gen(training_batch: reagent.core.types.MemoryNetworkInput, batch_idx: int): Implement training step as generator here

training: bool

use_amp: bool

class reagent.training.CRRTrainerParameters(rl: reagent.core.parameters.RLParameters = <factory>, double_q_learning: bool = True, q_network_optimizer: reagent.optimizer.union.Optimizer__Union = <factory>, actor_network_optimizer: reagent.optimizer.union.Optimizer__Union = <factory>, use_target_actor: bool = False, actions: List[str] = <factory>, delayed_policy_update: int = 1, beta: float = 1.0, entropy_coeff: float = 0.0, clip_limit: float = 10.0, max_weight: float = 20.0)

Bases: object

actions: List[str]

actor_network_optimizer: reagent.optimizer.union.Optimizer__Union

asdict()

beta: float = 1.0

clip_limit: float = 10.0

delayed_policy_update: int = 1

double_q_learning: bool = True

entropy_coeff: float = 0.0

max_weight: float = 20.0

q_network_optimizer: reagent.optimizer.union.Optimizer__Union

rl: reagent.core.parameters.RLParameters

use_target_actor: bool = False

class reagent.training.DQNTrainer(q_network, q_network_target, reward_network, q_network_cpe=None, q_network_cpe_target=None, metrics_to_score=None, evaluation: reagent.core.parameters.EvaluationParameters = Field(name=None,type=None,default=<dataclasses._MISSING_TYPE object>,default_factory=<class 'reagent.core.parameters.EvaluationParameters'>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=None), imitator=None, actions: List[str] = Field(name='actions',type=typing.List[str],default=<dataclasses._MISSING_TYPE object>,default_factory=<class 'list'>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), rl: reagent.core.parameters.RLParameters = Field(name='rl',type=<class 'reagent.core.parameters.RLParameters'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<class 'reagent.core.parameters.RLParameters'>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), double_q_learning: bool = True, bcq: Optional[reagent.training.dqn_trainer.BCQConfig] = None, minibatch_size: int = 1024, minibatches_per_step: int = 1, optimizer: reagent.optimizer.union.Optimizer__Union = Field(name='optimizer',type=<class 'reagent.optimizer.union.Optimizer__Union'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<bound method Optimizer__Union.default of <class 'reagent.optimizer.union.Optimizer__Union'>>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD))

Bases: reagent.training.dqn_trainer_base.DQNTrainerBaseLightning

allow_zero_length_dataloader_with_multiple_devices: bool

compute_discount_tensor(batch: reagent.core.types.DiscreteDqnInput, boosted_rewards: torch.Tensor)

compute_td_loss(batch: reagent.core.types.DiscreteDqnInput, boosted_rewards: torch.Tensor, discount_tensor: torch.Tensor)

configure_optimizers()

Choose what optimizers and learning-rate schedulers to use in your optimization. Normally you’d need one. But in the case of GANs or similar you might have multiple.

Returns

Any of these 6 options.

Single optimizer.
List or Tuple of optimizers.
Two lists - The first list has multiple optimizers, and the second has multiple LR schedulers (or multiple lr_scheduler_config).
Dictionary, with an "optimizer" key, and (optionally) a "lr_scheduler" key whose value is a single LR scheduler or lr_scheduler_config.
Tuple of dictionaries as described above, with an optional "frequency" key.
None - Fit will run without any optimizer.

The lr_scheduler_config is a dictionary which contains the scheduler and its associated configuration. The default configuration is shown below.

lr_scheduler_config = {
    # REQUIRED: The scheduler instance
    "scheduler": lr_scheduler,
    # The unit of the scheduler's step size, could also be 'step'.
    # 'epoch' updates the scheduler on epoch end whereas 'step'
    # updates it after a optimizer update.
    "interval": "epoch",
    # How many epochs/steps should pass between calls to
    # `scheduler.step()`. 1 corresponds to updating the learning
    # rate after every epoch/step.
    "frequency": 1,
    # Metric to to monitor for schedulers like `ReduceLROnPlateau`
    "monitor": "val_loss",
    # If set to `True`, will enforce that the value specified 'monitor'
    # is available when the scheduler is updated, thus stopping
    # training if not found. If set to `False`, it will only produce a warning
    "strict": True,
    # If using the `LearningRateMonitor` callback to monitor the
    # learning rate progress, this keyword can be used to specify
    # a custom logged name
    "name": None,
}

When there are schedulers in which the .step() method is conditioned on a value, such as the torch.optim.lr_scheduler.ReduceLROnPlateau scheduler, Lightning requires that the lr_scheduler_config contains the keyword "monitor" set to the metric name that the scheduler should be conditioned on.

Metrics can be made available to monitor by simply logging it using self.log('metric_to_track', metric_val) in your LightningModule.

Note

The frequency value specified in a dict along with the optimizer key is an int corresponding to the number of sequential batches optimized with the specific optimizer. It should be given to none or to all of the optimizers. There is a difference between passing multiple optimizers in a list, and passing multiple optimizers in dictionaries with a frequency of 1:

This is different from the frequency value specified in the lr_scheduler_config mentioned above.

def configure_optimizers(self):
    optimizer_one = torch.optim.SGD(self.model.parameters(), lr=0.01)
    optimizer_two = torch.optim.SGD(self.model.parameters(), lr=0.01)
    return [
        {"optimizer": optimizer_one, "frequency": 5},
        {"optimizer": optimizer_two, "frequency": 10},
    ]

In this example, the first optimizer will be used for the first 5 steps, the second optimizer for the next 10 steps and that cycle will continue. If an LR scheduler is specified for an optimizer using the lr_scheduler key in the above dict, the scheduler will only be updated when its optimizer is being used.

Examples:

# most cases. no learning rate scheduler
def configure_optimizers(self):
    return Adam(self.parameters(), lr=1e-3)

# multiple optimizer case (e.g.: GAN)
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    return gen_opt, dis_opt

# example with learning rate schedulers
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    dis_sch = CosineAnnealing(dis_opt, T_max=10)
    return [gen_opt, dis_opt], [dis_sch]

# example with step-based learning rate schedulers
# each optimizer has its own scheduler
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    gen_sch = {
        'scheduler': ExponentialLR(gen_opt, 0.99),
        'interval': 'step'  # called after each training step
    }
    dis_sch = CosineAnnealing(dis_opt, T_max=10) # called every epoch
    return [gen_opt, dis_opt], [gen_sch, dis_sch]

# example with optimizer frequencies
# see training procedure in `Improved Training of Wasserstein GANs`, Algorithm 1
# https://arxiv.org/abs/1704.00028
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    n_critic = 5
    return (
        {'optimizer': dis_opt, 'frequency': n_critic},
        {'optimizer': gen_opt, 'frequency': 1}
    )

Note

Some things to know:

Lightning calls .backward() and .step() on each optimizer and learning rate scheduler as needed.
If you use 16-bit precision (precision=16), Lightning will automatically handle the optimizers.
If you use multiple optimizers, training_step() will have an additional optimizer_idx parameter.
If you use torch.optim.LBFGS, Lightning handles the closure function automatically for you.
If you use multiple optimizers, gradients will be calculated only for the parameters of current optimizer at each training step.
If you need to control how often those optimizers step or override the default .step() schedule, override the optimizer_step() hook.

get_detached_model_outputs(state) → Tuple[torch.Tensor, Optional[torch.Tensor]]: Gets the q values from the model and target networks

precision: int

prepare_data_per_node: bool

rl_parameters: reagent.core.parameters.RLParameters

train_step_gen(training_batch: reagent.core.types.DiscreteDqnInput, batch_idx: int): Implement training step as generator here

training: bool

use_amp: bool

validation_step(batch, batch_idx)

Operates on a single batch of data from the validation set. In this step you’d might generate examples or calculate anything of interest like accuracy.

# the pseudocode for these calls
val_outs = []
for val_batch in val_data:
    out = validation_step(val_batch)
    val_outs.append(out)
validation_epoch_end(val_outs)

Parameters

batch (Tensor | (Tensor, …) | [Tensor, …]) – The output of your DataLoader. A tensor, tuple or list.
batch_idx (int) – The index of this batch
dataloader_idx (int) – The index of the dataloader that produced this batch (only if multiple val dataloaders used)

Returns

Any object or value
None - Validation will skip to the next batch

# pseudocode of order
val_outs = []
for val_batch in val_data:
    out = validation_step(val_batch)
    if defined("validation_step_end"):
        out = validation_step_end(out)
    val_outs.append(out)
val_outs = validation_epoch_end(val_outs)

# if you have one val dataloader:
def validation_step(self, batch, batch_idx):
    ...


# if you have multiple val dataloaders:
def validation_step(self, batch, batch_idx, dataloader_idx):
    ...

Examples:

# CASE 1: A single validation dataset
def validation_step(self, batch, batch_idx):
    x, y = batch

    # implement your own
    out = self(x)
    loss = self.loss(out, y)

    # log 6 example images
    # or generated text... or whatever
    sample_imgs = x[:6]
    grid = torchvision.utils.make_grid(sample_imgs)
    self.logger.experiment.add_image('example_images', grid, 0)

    # calculate acc
    labels_hat = torch.argmax(out, dim=1)
    val_acc = torch.sum(y == labels_hat).item() / (len(y) * 1.0)

    # log the outputs!
    self.log_dict({'val_loss': loss, 'val_acc': val_acc})

If you pass in multiple val dataloaders, validation_step() will have an additional argument.

# CASE 2: multiple validation dataloaders
def validation_step(self, batch, batch_idx, dataloader_idx):
    # dataloader_idx tells you which dataset this is.
    ...

Note

If you don’t need to validate you don’t need to implement this method.

Note

When the validation_step() is called, the model has been put in eval mode and PyTorch gradients have been disabled. At the end of validation, the model goes back to training mode and gradients are enabled.

class reagent.training.DQNTrainerParameters(actions: List[str] = <factory>, rl: reagent.core.parameters.RLParameters = <factory>, double_q_learning: bool = True, bcq: Optional[reagent.training.dqn_trainer.BCQConfig] = None, minibatch_size: int = 1024, minibatches_per_step: int = 1, optimizer: reagent.optimizer.union.Optimizer__Union = <factory>)

Bases: object

actions: List[str]

asdict()

bcq: Optional[reagent.training.dqn_trainer.BCQConfig] = None

double_q_learning: bool = True

minibatch_size: int = 1024

minibatches_per_step: int = 1

optimizer: reagent.optimizer.union.Optimizer__Union

rl: reagent.core.parameters.RLParameters

class reagent.training.DiscreteCRRTrainer(actor_network, actor_network_target, q1_network, q1_network_target, reward_network, q2_network=None, q2_network_target=None, q_network_cpe=None, q_network_cpe_target=None, metrics_to_score=None, evaluation: reagent.core.parameters.EvaluationParameters = Field(name=None,type=None,default=<dataclasses._MISSING_TYPE object>,default_factory=<class 'reagent.core.parameters.EvaluationParameters'>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=None), rl: reagent.core.parameters.RLParameters = Field(name='rl',type=<class 'reagent.core.parameters.RLParameters'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<class 'reagent.core.parameters.RLParameters'>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), double_q_learning: bool = True, q_network_optimizer: reagent.optimizer.union.Optimizer__Union = Field(name='q_network_optimizer',type=<class 'reagent.optimizer.union.Optimizer__Union'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<bound method Optimizer__Union.default of <class 'reagent.optimizer.union.Optimizer__Union'>>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), actor_network_optimizer: reagent.optimizer.union.Optimizer__Union = Field(name='actor_network_optimizer',type=<class 'reagent.optimizer.union.Optimizer__Union'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<bound method Optimizer__Union.default of <class 'reagent.optimizer.union.Optimizer__Union'>>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), use_target_actor: bool = False, actions: List[str] = Field(name='actions',type=typing.List[str],default=<dataclasses._MISSING_TYPE object>,default_factory=<class 'list'>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), delayed_policy_update: int = 1, beta: float = 1.0, entropy_coeff: float = 0.0, clip_limit: float = 10.0, max_weight: float = 20.0)

Bases: reagent.training.dqn_trainer_base.DQNTrainerBaseLightning

Critic Regularized Regression (CRR) algorithm trainer as described in https://arxiv.org/abs/2006.15134

allow_zero_length_dataloader_with_multiple_devices: bool

compute_actor_loss(batch_idx, action, logged_action_probs, all_q_values, all_action_scores)

compute_target_q_values(next_state, rewards, not_terminal, next_q_values)

compute_td_loss(q_network, state, action, target_q_values)

configure_optimizers()

Choose what optimizers and learning-rate schedulers to use in your optimization. Normally you’d need one. But in the case of GANs or similar you might have multiple.

Returns

Any of these 6 options.

Single optimizer.
List or Tuple of optimizers.
Two lists - The first list has multiple optimizers, and the second has multiple LR schedulers (or multiple lr_scheduler_config).
Dictionary, with an "optimizer" key, and (optionally) a "lr_scheduler" key whose value is a single LR scheduler or lr_scheduler_config.
Tuple of dictionaries as described above, with an optional "frequency" key.
None - Fit will run without any optimizer.

The lr_scheduler_config is a dictionary which contains the scheduler and its associated configuration. The default configuration is shown below.

lr_scheduler_config = {
    # REQUIRED: The scheduler instance
    "scheduler": lr_scheduler,
    # The unit of the scheduler's step size, could also be 'step'.
    # 'epoch' updates the scheduler on epoch end whereas 'step'
    # updates it after a optimizer update.
    "interval": "epoch",
    # How many epochs/steps should pass between calls to
    # `scheduler.step()`. 1 corresponds to updating the learning
    # rate after every epoch/step.
    "frequency": 1,
    # Metric to to monitor for schedulers like `ReduceLROnPlateau`
    "monitor": "val_loss",
    # If set to `True`, will enforce that the value specified 'monitor'
    # is available when the scheduler is updated, thus stopping
    # training if not found. If set to `False`, it will only produce a warning
    "strict": True,
    # If using the `LearningRateMonitor` callback to monitor the
    # learning rate progress, this keyword can be used to specify
    # a custom logged name
    "name": None,
}

When there are schedulers in which the .step() method is conditioned on a value, such as the torch.optim.lr_scheduler.ReduceLROnPlateau scheduler, Lightning requires that the lr_scheduler_config contains the keyword "monitor" set to the metric name that the scheduler should be conditioned on.

Metrics can be made available to monitor by simply logging it using self.log('metric_to_track', metric_val) in your LightningModule.

Note

The frequency value specified in a dict along with the optimizer key is an int corresponding to the number of sequential batches optimized with the specific optimizer. It should be given to none or to all of the optimizers. There is a difference between passing multiple optimizers in a list, and passing multiple optimizers in dictionaries with a frequency of 1:

This is different from the frequency value specified in the lr_scheduler_config mentioned above.

def configure_optimizers(self):
    optimizer_one = torch.optim.SGD(self.model.parameters(), lr=0.01)
    optimizer_two = torch.optim.SGD(self.model.parameters(), lr=0.01)
    return [
        {"optimizer": optimizer_one, "frequency": 5},
        {"optimizer": optimizer_two, "frequency": 10},
    ]

In this example, the first optimizer will be used for the first 5 steps, the second optimizer for the next 10 steps and that cycle will continue. If an LR scheduler is specified for an optimizer using the lr_scheduler key in the above dict, the scheduler will only be updated when its optimizer is being used.

Examples:

# most cases. no learning rate scheduler
def configure_optimizers(self):
    return Adam(self.parameters(), lr=1e-3)

# multiple optimizer case (e.g.: GAN)
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    return gen_opt, dis_opt

# example with learning rate schedulers
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    dis_sch = CosineAnnealing(dis_opt, T_max=10)
    return [gen_opt, dis_opt], [dis_sch]

# example with step-based learning rate schedulers
# each optimizer has its own scheduler
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    gen_sch = {
        'scheduler': ExponentialLR(gen_opt, 0.99),
        'interval': 'step'  # called after each training step
    }
    dis_sch = CosineAnnealing(dis_opt, T_max=10) # called every epoch
    return [gen_opt, dis_opt], [gen_sch, dis_sch]

# example with optimizer frequencies
# see training procedure in `Improved Training of Wasserstein GANs`, Algorithm 1
# https://arxiv.org/abs/1704.00028
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    n_critic = 5
    return (
        {'optimizer': dis_opt, 'frequency': n_critic},
        {'optimizer': gen_opt, 'frequency': 1}
    )

Note

Some things to know:

Lightning calls .backward() and .step() on each optimizer and learning rate scheduler as needed.
If you use 16-bit precision (precision=16), Lightning will automatically handle the optimizers.
If you use multiple optimizers, training_step() will have an additional optimizer_idx parameter.
If you use torch.optim.LBFGS, Lightning handles the closure function automatically for you.
If you use multiple optimizers, gradients will be calculated only for the parameters of current optimizer at each training step.
If you need to control how often those optimizers step or override the default .step() schedule, override the optimizer_step() hook.

get_detached_model_outputs(state) → Tuple[torch.Tensor, None]

precision: int

prepare_data_per_node: bool

property q_network

rl_parameters: reagent.core.parameters.RLParameters

train_step_gen(training_batch: reagent.core.types.DiscreteDqnInput, batch_idx: int): IMPORTANT: the input action here is preprocessed according to the training_batch type, which in this case is DiscreteDqnInput. Hence, the preprocessor in the DiscreteDqnInputMaker class in the trainer_preprocessor.py is used, which converts acion taken to a one-hot representation.

training: bool

use_amp: bool

validation_step(batch, batch_idx)

Operates on a single batch of data from the validation set. In this step you’d might generate examples or calculate anything of interest like accuracy.

# the pseudocode for these calls
val_outs = []
for val_batch in val_data:
    out = validation_step(val_batch)
    val_outs.append(out)
validation_epoch_end(val_outs)

Parameters

batch (Tensor | (Tensor, …) | [Tensor, …]) – The output of your DataLoader. A tensor, tuple or list.
batch_idx (int) – The index of this batch
dataloader_idx (int) – The index of the dataloader that produced this batch (only if multiple val dataloaders used)

Returns

Any object or value
None - Validation will skip to the next batch

# pseudocode of order
val_outs = []
for val_batch in val_data:
    out = validation_step(val_batch)
    if defined("validation_step_end"):
        out = validation_step_end(out)
    val_outs.append(out)
val_outs = validation_epoch_end(val_outs)

# if you have one val dataloader:
def validation_step(self, batch, batch_idx):
    ...


# if you have multiple val dataloaders:
def validation_step(self, batch, batch_idx, dataloader_idx):
    ...

Examples:

# CASE 1: A single validation dataset
def validation_step(self, batch, batch_idx):
    x, y = batch

    # implement your own
    out = self(x)
    loss = self.loss(out, y)

    # log 6 example images
    # or generated text... or whatever
    sample_imgs = x[:6]
    grid = torchvision.utils.make_grid(sample_imgs)
    self.logger.experiment.add_image('example_images', grid, 0)

    # calculate acc
    labels_hat = torch.argmax(out, dim=1)
    val_acc = torch.sum(y == labels_hat).item() / (len(y) * 1.0)

    # log the outputs!
    self.log_dict({'val_loss': loss, 'val_acc': val_acc})

If you pass in multiple val dataloaders, validation_step() will have an additional argument.

# CASE 2: multiple validation dataloaders
def validation_step(self, batch, batch_idx, dataloader_idx):
    # dataloader_idx tells you which dataset this is.
    ...

Note

If you don’t need to validate you don’t need to implement this method.

Note

When the validation_step() is called, the model has been put in eval mode and PyTorch gradients have been disabled. At the end of validation, the model goes back to training mode and gradients are enabled.

class reagent.training.MDNRNNTrainer(memory_network: reagent.models.world_model.MemoryNetwork, params: reagent.core.parameters.MDNRNNTrainerParameters, cum_loss_hist: int = 100)

Bases: reagent.training.reagent_lightning_module.ReAgentLightningModule

Trainer for MDN-RNN

allow_zero_length_dataloader_with_multiple_devices: bool

configure_optimizers()

Choose what optimizers and learning-rate schedulers to use in your optimization. Normally you’d need one. But in the case of GANs or similar you might have multiple.

Returns

Any of these 6 options.

Single optimizer.
List or Tuple of optimizers.
Two lists - The first list has multiple optimizers, and the second has multiple LR schedulers (or multiple lr_scheduler_config).
Dictionary, with an "optimizer" key, and (optionally) a "lr_scheduler" key whose value is a single LR scheduler or lr_scheduler_config.
Tuple of dictionaries as described above, with an optional "frequency" key.
None - Fit will run without any optimizer.

The lr_scheduler_config is a dictionary which contains the scheduler and its associated configuration. The default configuration is shown below.

lr_scheduler_config = {
    # REQUIRED: The scheduler instance
    "scheduler": lr_scheduler,
    # The unit of the scheduler's step size, could also be 'step'.
    # 'epoch' updates the scheduler on epoch end whereas 'step'
    # updates it after a optimizer update.
    "interval": "epoch",
    # How many epochs/steps should pass between calls to
    # `scheduler.step()`. 1 corresponds to updating the learning
    # rate after every epoch/step.
    "frequency": 1,
    # Metric to to monitor for schedulers like `ReduceLROnPlateau`
    "monitor": "val_loss",
    # If set to `True`, will enforce that the value specified 'monitor'
    # is available when the scheduler is updated, thus stopping
    # training if not found. If set to `False`, it will only produce a warning
    "strict": True,
    # If using the `LearningRateMonitor` callback to monitor the
    # learning rate progress, this keyword can be used to specify
    # a custom logged name
    "name": None,
}

When there are schedulers in which the .step() method is conditioned on a value, such as the torch.optim.lr_scheduler.ReduceLROnPlateau scheduler, Lightning requires that the lr_scheduler_config contains the keyword "monitor" set to the metric name that the scheduler should be conditioned on.

Metrics can be made available to monitor by simply logging it using self.log('metric_to_track', metric_val) in your LightningModule.

Note

The frequency value specified in a dict along with the optimizer key is an int corresponding to the number of sequential batches optimized with the specific optimizer. It should be given to none or to all of the optimizers. There is a difference between passing multiple optimizers in a list, and passing multiple optimizers in dictionaries with a frequency of 1:

This is different from the frequency value specified in the lr_scheduler_config mentioned above.

def configure_optimizers(self):
    optimizer_one = torch.optim.SGD(self.model.parameters(), lr=0.01)
    optimizer_two = torch.optim.SGD(self.model.parameters(), lr=0.01)
    return [
        {"optimizer": optimizer_one, "frequency": 5},
        {"optimizer": optimizer_two, "frequency": 10},
    ]

In this example, the first optimizer will be used for the first 5 steps, the second optimizer for the next 10 steps and that cycle will continue. If an LR scheduler is specified for an optimizer using the lr_scheduler key in the above dict, the scheduler will only be updated when its optimizer is being used.

Examples:

# most cases. no learning rate scheduler
def configure_optimizers(self):
    return Adam(self.parameters(), lr=1e-3)

# multiple optimizer case (e.g.: GAN)
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    return gen_opt, dis_opt

# example with learning rate schedulers
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    dis_sch = CosineAnnealing(dis_opt, T_max=10)
    return [gen_opt, dis_opt], [dis_sch]

# example with step-based learning rate schedulers
# each optimizer has its own scheduler
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    gen_sch = {
        'scheduler': ExponentialLR(gen_opt, 0.99),
        'interval': 'step'  # called after each training step
    }
    dis_sch = CosineAnnealing(dis_opt, T_max=10) # called every epoch
    return [gen_opt, dis_opt], [gen_sch, dis_sch]

# example with optimizer frequencies
# see training procedure in `Improved Training of Wasserstein GANs`, Algorithm 1
# https://arxiv.org/abs/1704.00028
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    n_critic = 5
    return (
        {'optimizer': dis_opt, 'frequency': n_critic},
        {'optimizer': gen_opt, 'frequency': 1}
    )

Note

Some things to know:

Lightning calls .backward() and .step() on each optimizer and learning rate scheduler as needed.
If you use 16-bit precision (precision=16), Lightning will automatically handle the optimizers.
If you use multiple optimizers, training_step() will have an additional optimizer_idx parameter.
If you use torch.optim.LBFGS, Lightning handles the closure function automatically for you.
If you use multiple optimizers, gradients will be calculated only for the parameters of current optimizer at each training step.
If you need to control how often those optimizers step or override the default .step() schedule, override the optimizer_step() hook.

get_loss(training_batch: reagent.core.types.MemoryNetworkInput, state_dim: Optional[int] = None)

Compute losses:: GMMLoss(next_state, GMMPredicted) / (STATE_DIM + 2) + MSE(reward, predicted_reward) + BCE(not_terminal, logit_not_terminal)
The STATE_DIM + 2 factor is here to counteract the fact that the GMMLoss scales: approximately linearly with STATE_DIM, dim of states. All losses are averaged both on the batch and the sequence dimensions (the two first dimensions).

Parameters

training_batch – training_batch has these fields: - state: (SEQ_LEN, BATCH_SIZE, STATE_DIM) torch tensor - action: (SEQ_LEN, BATCH_SIZE, ACTION_DIM) torch tensor - reward: (SEQ_LEN, BATCH_SIZE) torch tensor - not-terminal: (SEQ_LEN, BATCH_SIZE) torch tensor - next_state: (SEQ_LEN, BATCH_SIZE, STATE_DIM) torch tensor
state_dim – the dimension of states. If provided, use it to normalize gmm loss

Returns

dictionary of losses, containing the gmm, the mse, the bce and the averaged loss.

precision: int

prepare_data_per_node: bool

test_step(training_batch: reagent.core.types.MemoryNetworkInput, batch_idx: int)

Operates on a single batch of data from the test set. In this step you’d normally generate examples or calculate anything of interest such as accuracy.

# the pseudocode for these calls
test_outs = []
for test_batch in test_data:
    out = test_step(test_batch)
    test_outs.append(out)
test_epoch_end(test_outs)

Parameters

batch (Tensor | (Tensor, …) | [Tensor, …]) – The output of your DataLoader. A tensor, tuple or list.
batch_idx (int) – The index of this batch.
dataloader_idx (int) – The index of the dataloader that produced this batch (only if multiple test dataloaders used).

Returns

Any of.

# if you have one test dataloader:
def test_step(self, batch, batch_idx):
    ...


# if you have multiple test dataloaders:
def test_step(self, batch, batch_idx, dataloader_idx):
    ...

Examples:

# CASE 1: A single test dataset
def test_step(self, batch, batch_idx):
    x, y = batch

    # implement your own
    out = self(x)
    loss = self.loss(out, y)

    # log 6 example images
    # or generated text... or whatever
    sample_imgs = x[:6]
    grid = torchvision.utils.make_grid(sample_imgs)
    self.logger.experiment.add_image('example_images', grid, 0)

    # calculate acc
    labels_hat = torch.argmax(out, dim=1)
    test_acc = torch.sum(y == labels_hat).item() / (len(y) * 1.0)

    # log the outputs!
    self.log_dict({'test_loss': loss, 'test_acc': test_acc})

If you pass in multiple test dataloaders, test_step() will have an additional argument.

# CASE 2: multiple test dataloaders
def test_step(self, batch, batch_idx, dataloader_idx):
    # dataloader_idx tells you which dataset this is.
    ...

Note

If you don’t need to test you don’t need to implement this method.

Note

When the test_step() is called, the model has been put in eval mode and PyTorch gradients have been disabled. At the end of the test epoch, the model goes back to training mode and gradients are enabled.

train_step_gen(training_batch: reagent.core.types.MemoryNetworkInput, batch_idx: int): Implement training step as generator here

training: bool

use_amp: bool

validation_step(training_batch: reagent.core.types.MemoryNetworkInput, batch_idx: int)

Operates on a single batch of data from the validation set. In this step you’d might generate examples or calculate anything of interest like accuracy.

# the pseudocode for these calls
val_outs = []
for val_batch in val_data:
    out = validation_step(val_batch)
    val_outs.append(out)
validation_epoch_end(val_outs)

Parameters

batch (Tensor | (Tensor, …) | [Tensor, …]) – The output of your DataLoader. A tensor, tuple or list.
batch_idx (int) – The index of this batch
dataloader_idx (int) – The index of the dataloader that produced this batch (only if multiple val dataloaders used)

Returns

Any object or value
None - Validation will skip to the next batch

# pseudocode of order
val_outs = []
for val_batch in val_data:
    out = validation_step(val_batch)
    if defined("validation_step_end"):
        out = validation_step_end(out)
    val_outs.append(out)
val_outs = validation_epoch_end(val_outs)

# if you have one val dataloader:
def validation_step(self, batch, batch_idx):
    ...


# if you have multiple val dataloaders:
def validation_step(self, batch, batch_idx, dataloader_idx):
    ...

Examples:

# CASE 1: A single validation dataset
def validation_step(self, batch, batch_idx):
    x, y = batch

    # implement your own
    out = self(x)
    loss = self.loss(out, y)

    # log 6 example images
    # or generated text... or whatever
    sample_imgs = x[:6]
    grid = torchvision.utils.make_grid(sample_imgs)
    self.logger.experiment.add_image('example_images', grid, 0)

    # calculate acc
    labels_hat = torch.argmax(out, dim=1)
    val_acc = torch.sum(y == labels_hat).item() / (len(y) * 1.0)

    # log the outputs!
    self.log_dict({'val_loss': loss, 'val_acc': val_acc})

If you pass in multiple val dataloaders, validation_step() will have an additional argument.

# CASE 2: multiple validation dataloaders
def validation_step(self, batch, batch_idx, dataloader_idx):
    # dataloader_idx tells you which dataset this is.
    ...

Note

If you don’t need to validate you don’t need to implement this method.

Note

When the validation_step() is called, the model has been put in eval mode and PyTorch gradients have been disabled. At the end of validation, the model goes back to training mode and gradients are enabled.

class reagent.training.MultiStageTrainer(trainers: List[reagent.training.reagent_lightning_module.ReAgentLightningModule], epochs: List[int], assign_reporter_function=None, flush_reporter_function=None, automatic_optimization=True)

Bases: reagent.training.reagent_lightning_module.ReAgentLightningModule

allow_zero_length_dataloader_with_multiple_devices: bool

configure_optimizers()

Choose what optimizers and learning-rate schedulers to use in your optimization. Normally you’d need one. But in the case of GANs or similar you might have multiple.

Returns

Any of these 6 options.

Single optimizer.
List or Tuple of optimizers.
Two lists - The first list has multiple optimizers, and the second has multiple LR schedulers (or multiple lr_scheduler_config).
Dictionary, with an "optimizer" key, and (optionally) a "lr_scheduler" key whose value is a single LR scheduler or lr_scheduler_config.
Tuple of dictionaries as described above, with an optional "frequency" key.
None - Fit will run without any optimizer.

The lr_scheduler_config is a dictionary which contains the scheduler and its associated configuration. The default configuration is shown below.

lr_scheduler_config = {
    # REQUIRED: The scheduler instance
    "scheduler": lr_scheduler,
    # The unit of the scheduler's step size, could also be 'step'.
    # 'epoch' updates the scheduler on epoch end whereas 'step'
    # updates it after a optimizer update.
    "interval": "epoch",
    # How many epochs/steps should pass between calls to
    # `scheduler.step()`. 1 corresponds to updating the learning
    # rate after every epoch/step.
    "frequency": 1,
    # Metric to to monitor for schedulers like `ReduceLROnPlateau`
    "monitor": "val_loss",
    # If set to `True`, will enforce that the value specified 'monitor'
    # is available when the scheduler is updated, thus stopping
    # training if not found. If set to `False`, it will only produce a warning
    "strict": True,
    # If using the `LearningRateMonitor` callback to monitor the
    # learning rate progress, this keyword can be used to specify
    # a custom logged name
    "name": None,
}

When there are schedulers in which the .step() method is conditioned on a value, such as the torch.optim.lr_scheduler.ReduceLROnPlateau scheduler, Lightning requires that the lr_scheduler_config contains the keyword "monitor" set to the metric name that the scheduler should be conditioned on.

Metrics can be made available to monitor by simply logging it using self.log('metric_to_track', metric_val) in your LightningModule.

Note

The frequency value specified in a dict along with the optimizer key is an int corresponding to the number of sequential batches optimized with the specific optimizer. It should be given to none or to all of the optimizers. There is a difference between passing multiple optimizers in a list, and passing multiple optimizers in dictionaries with a frequency of 1:

This is different from the frequency value specified in the lr_scheduler_config mentioned above.

def configure_optimizers(self):
    optimizer_one = torch.optim.SGD(self.model.parameters(), lr=0.01)
    optimizer_two = torch.optim.SGD(self.model.parameters(), lr=0.01)
    return [
        {"optimizer": optimizer_one, "frequency": 5},
        {"optimizer": optimizer_two, "frequency": 10},
    ]

In this example, the first optimizer will be used for the first 5 steps, the second optimizer for the next 10 steps and that cycle will continue. If an LR scheduler is specified for an optimizer using the lr_scheduler key in the above dict, the scheduler will only be updated when its optimizer is being used.

Examples:

# most cases. no learning rate scheduler
def configure_optimizers(self):
    return Adam(self.parameters(), lr=1e-3)

# multiple optimizer case (e.g.: GAN)
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    return gen_opt, dis_opt

# example with learning rate schedulers
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    dis_sch = CosineAnnealing(dis_opt, T_max=10)
    return [gen_opt, dis_opt], [dis_sch]

# example with step-based learning rate schedulers
# each optimizer has its own scheduler
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    gen_sch = {
        'scheduler': ExponentialLR(gen_opt, 0.99),
        'interval': 'step'  # called after each training step
    }
    dis_sch = CosineAnnealing(dis_opt, T_max=10) # called every epoch
    return [gen_opt, dis_opt], [gen_sch, dis_sch]

# example with optimizer frequencies
# see training procedure in `Improved Training of Wasserstein GANs`, Algorithm 1
# https://arxiv.org/abs/1704.00028
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    n_critic = 5
    return (
        {'optimizer': dis_opt, 'frequency': n_critic},
        {'optimizer': gen_opt, 'frequency': 1}
    )

Note

Some things to know:

Lightning calls .backward() and .step() on each optimizer and learning rate scheduler as needed.
If you use 16-bit precision (precision=16), Lightning will automatically handle the optimizers.
If you use multiple optimizers, training_step() will have an additional optimizer_idx parameter.
If you use torch.optim.LBFGS, Lightning handles the closure function automatically for you.
If you use multiple optimizers, gradients will be calculated only for the parameters of current optimizer at each training step.
If you need to control how often those optimizers step or override the default .step() schedule, override the optimizer_step() hook.

property multi_stage_total_epochs

on_fit_end()

Called at the very end of fit.

If on DDP it is called on every process

on_fit_start()

Called at the very beginning of fit.

If on DDP it is called on every process

on_test_end(): Called at the end of testing.

on_test_start(): Called at the beginning of testing.

optimizer_step(epoch: int, batch_idx: int, optimizer, optimizer_idx: int, optimizer_closure, on_tpu: int = False, using_native_amp: int = False, using_lbfgs: int = False)

Override this method to adjust the default way the Trainer calls each optimizer. By default, Lightning calls step() and zero_grad() as shown in the example once per optimizer. This method (and zero_grad()) won’t be called during the accumulation phase when Trainer(accumulate_grad_batches != 1).

Parameters

epoch – Current epoch
batch_idx – Index of current batch
optimizer – A PyTorch optimizer
optimizer_idx – If you used multiple optimizers, this indexes into that list.
optimizer_closure – Closure for all optimizers. This closure must be executed as it includes the calls to training_step(), optimizer.zero_grad(), and backward().
on_tpu – True if TPU backward is required
using_native_amp – True if using native amp
using_lbfgs – True if the matching optimizer is torch.optim.LBFGS

Examples:

# DEFAULT
def optimizer_step(self, epoch, batch_idx, optimizer, optimizer_idx,
                   optimizer_closure, on_tpu, using_native_amp, using_lbfgs):
    optimizer.step(closure=optimizer_closure)

# Alternating schedule for optimizer steps (i.e.: GANs)
def optimizer_step(self, epoch, batch_idx, optimizer, optimizer_idx,
                   optimizer_closure, on_tpu, using_native_amp, using_lbfgs):
    # update generator opt every step
    if optimizer_idx == 0:
        optimizer.step(closure=optimizer_closure)

    # update discriminator opt every 2 steps
    if optimizer_idx == 1:
        if (batch_idx + 1) % 2 == 0 :
            optimizer.step(closure=optimizer_closure)
        else:
            # call the closure by itself to run `training_step` + `backward` without an optimizer step
            optimizer_closure()

    # ...
    # add as many optimizers as you want

Here’s another example showing how to use this for more advanced things such as learning rate warm-up:

# learning rate warm-up
def optimizer_step(
    self,
    epoch,
    batch_idx,
    optimizer,
    optimizer_idx,
    optimizer_closure,
    on_tpu,
    using_native_amp,
    using_lbfgs,
):
    # warm up lr
    if self.trainer.global_step < 500:
        lr_scale = min(1.0, float(self.trainer.global_step + 1) / 500.0)
        for pg in optimizer.param_groups:
            pg["lr"] = lr_scale * self.learning_rate

    # update params
    optimizer.step(closure=optimizer_closure)

precision: int

prepare_data_per_node: bool

set_reporter(reporter)

test_epoch_end(outputs)

Called at the end of a test epoch with the output of all test steps.

# the pseudocode for these calls
test_outs = []
for test_batch in test_data:
    out = test_step(test_batch)
    test_outs.append(out)
test_epoch_end(test_outs)

Parameters: outputs – List of outputs you defined in test_step_end(), or if there are multiple dataloaders, a list containing a list of outputs for each dataloader
Returns: None

Note

If you didn’t define a test_step(), this won’t be called.

Examples

With a single dataloader:

def test_epoch_end(self, outputs):
    # do something with the outputs of all test batches
    all_test_preds = test_step_outputs.predictions

    some_result = calc_all_results(all_test_preds)
    self.log(some_result)

With multiple dataloaders, outputs will be a list of lists. The outer list contains one entry per dataloader, while the inner list contains the individual outputs of each test step for that dataloader.

def test_epoch_end(self, outputs):
    final_value = 0
    for dataloader_outputs in outputs:
        for test_step_out in dataloader_outputs:
            # do something
            final_value += test_step_out

    self.log("final_metric", final_value)

test_step(*args, **kwargs)

Operates on a single batch of data from the test set. In this step you’d normally generate examples or calculate anything of interest such as accuracy.

# the pseudocode for these calls
test_outs = []
for test_batch in test_data:
    out = test_step(test_batch)
    test_outs.append(out)
test_epoch_end(test_outs)

Parameters

batch (Tensor | (Tensor, …) | [Tensor, …]) – The output of your DataLoader. A tensor, tuple or list.
batch_idx (int) – The index of this batch.
dataloader_idx (int) – The index of the dataloader that produced this batch (only if multiple test dataloaders used).

Returns

Any of.

# if you have one test dataloader:
def test_step(self, batch, batch_idx):
    ...


# if you have multiple test dataloaders:
def test_step(self, batch, batch_idx, dataloader_idx):
    ...

Examples:

# CASE 1: A single test dataset
def test_step(self, batch, batch_idx):
    x, y = batch

    # implement your own
    out = self(x)
    loss = self.loss(out, y)

    # log 6 example images
    # or generated text... or whatever
    sample_imgs = x[:6]
    grid = torchvision.utils.make_grid(sample_imgs)
    self.logger.experiment.add_image('example_images', grid, 0)

    # calculate acc
    labels_hat = torch.argmax(out, dim=1)
    test_acc = torch.sum(y == labels_hat).item() / (len(y) * 1.0)

    # log the outputs!
    self.log_dict({'test_loss': loss, 'test_acc': test_acc})

If you pass in multiple test dataloaders, test_step() will have an additional argument.

# CASE 2: multiple test dataloaders
def test_step(self, batch, batch_idx, dataloader_idx):
    # dataloader_idx tells you which dataset this is.
    ...

Note

If you don’t need to test you don’t need to implement this method.

Note

When the test_step() is called, the model has been put in eval mode and PyTorch gradients have been disabled. At the end of the test epoch, the model goes back to training mode and gradients are enabled.

training: bool

training_epoch_end(outputs)

Called at the end of the training epoch with the outputs of all training steps. Use this in case you need to do something with all the outputs returned by training_step().

# the pseudocode for these calls
train_outs = []
for train_batch in train_data:
    out = training_step(train_batch)
    train_outs.append(out)
training_epoch_end(train_outs)

Parameters: outputs – List of outputs you defined in training_step(). If there are multiple optimizers, it is a list containing a list of outputs for each optimizer. If using truncated_bptt_steps > 1, each element is a list of outputs corresponding to the outputs of each processed split batch.
Returns: None

Note

If this method is not overridden, this won’t be called.

def training_epoch_end(self, training_step_outputs):
    # do something with all training_step outputs
    for out in training_step_outputs:
        ...

training_step(batch, batch_idx: int, optimizer_idx: int = 0)

Here you compute and return the training loss and some additional metrics for e.g. the progress bar or logger.

Parameters

batch (Tensor | (Tensor, …) | [Tensor, …]) – The output of your DataLoader. A tensor, tuple or list.
batch_idx (int) – Integer displaying index of this batch
optimizer_idx (int) – When using multiple optimizers, this argument will also be present.
hiddens (Any) – Passed in if :paramref:`~pytorch_lightning.core.lightning.LightningModule.truncated_bptt_steps` > 0.

Returns

Any of.

Tensor - The loss tensor
dict - A dictionary. Can include any keys, but must include the key 'loss'
None - Training will skip to the next batch. This is only for automatic optimization.
This is not supported for multi-GPU, TPU, IPU, or DeepSpeed.

In this step you’d normally do the forward pass and calculate the loss for a batch. You can also do fancier things like multiple forward passes or something model specific.

Example:

def training_step(self, batch, batch_idx):
    x, y, z = batch
    out = self.encoder(x)
    loss = self.loss(out, x)
    return loss

If you define multiple optimizers, this step will be called with an additional optimizer_idx parameter.

# Multiple optimizers (e.g.: GANs)
def training_step(self, batch, batch_idx, optimizer_idx):
    if optimizer_idx == 0:
        # do training_step with encoder
        ...
    if optimizer_idx == 1:
        # do training_step with decoder
        ...

If you add truncated back propagation through time you will also get an additional argument with the hidden states of the previous step.

# Truncated back-propagation through time
def training_step(self, batch, batch_idx, hiddens):
    # hiddens are the hidden states from the previous truncated backprop step
    out, hiddens = self.lstm(data, hiddens)
    loss = ...
    return {"loss": loss, "hiddens": hiddens}

Note

The loss value shown in the progress bar is smoothed (averaged) over the last values, so it differs from the actual loss returned in train/validation step.

use_amp: bool

validation_epoch_end(outputs)

Called at the end of the validation epoch with the outputs of all validation steps.

# the pseudocode for these calls
val_outs = []
for val_batch in val_data:
    out = validation_step(val_batch)
    val_outs.append(out)
validation_epoch_end(val_outs)

Parameters: outputs – List of outputs you defined in validation_step(), or if there are multiple dataloaders, a list containing a list of outputs for each dataloader.
Returns: None

Note

If you didn’t define a validation_step(), this won’t be called.

Examples

With a single dataloader:

def validation_epoch_end(self, val_step_outputs):
    for out in val_step_outputs:
        ...

With multiple dataloaders, outputs will be a list of lists. The outer list contains one entry per dataloader, while the inner list contains the individual outputs of each validation step for that dataloader.

def validation_epoch_end(self, outputs):
    for dataloader_output_result in outputs:
        dataloader_outs = dataloader_output_result.dataloader_i_outputs

    self.log("final_metric", final_value)

validation_step(*args, **kwargs)

Operates on a single batch of data from the validation set. In this step you’d might generate examples or calculate anything of interest like accuracy.

# the pseudocode for these calls
val_outs = []
for val_batch in val_data:
    out = validation_step(val_batch)
    val_outs.append(out)
validation_epoch_end(val_outs)

Parameters

batch (Tensor | (Tensor, …) | [Tensor, …]) – The output of your DataLoader. A tensor, tuple or list.
batch_idx (int) – The index of this batch
dataloader_idx (int) – The index of the dataloader that produced this batch (only if multiple val dataloaders used)

Returns

Any object or value
None - Validation will skip to the next batch

# pseudocode of order
val_outs = []
for val_batch in val_data:
    out = validation_step(val_batch)
    if defined("validation_step_end"):
        out = validation_step_end(out)
    val_outs.append(out)
val_outs = validation_epoch_end(val_outs)

# if you have one val dataloader:
def validation_step(self, batch, batch_idx):
    ...


# if you have multiple val dataloaders:
def validation_step(self, batch, batch_idx, dataloader_idx):
    ...

Examples:

# CASE 1: A single validation dataset
def validation_step(self, batch, batch_idx):
    x, y = batch

    # implement your own
    out = self(x)
    loss = self.loss(out, y)

    # log 6 example images
    # or generated text... or whatever
    sample_imgs = x[:6]
    grid = torchvision.utils.make_grid(sample_imgs)
    self.logger.experiment.add_image('example_images', grid, 0)

    # calculate acc
    labels_hat = torch.argmax(out, dim=1)
    val_acc = torch.sum(y == labels_hat).item() / (len(y) * 1.0)

    # log the outputs!
    self.log_dict({'val_loss': loss, 'val_acc': val_acc})

If you pass in multiple val dataloaders, validation_step() will have an additional argument.

# CASE 2: multiple validation dataloaders
def validation_step(self, batch, batch_idx, dataloader_idx):
    # dataloader_idx tells you which dataset this is.
    ...

Note

If you don’t need to validate you don’t need to implement this method.

Note

When the validation_step() is called, the model has been put in eval mode and PyTorch gradients have been disabled. At the end of validation, the model goes back to training mode and gradients are enabled.

class reagent.training.PPOTrainer(policy: reagent.gym.policies.policy.Policy, gamma: float = 0.9, optimizer: reagent.optimizer.union.Optimizer__Union = Field(name='optimizer',type=<class 'reagent.optimizer.union.Optimizer__Union'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<bound method Optimizer__Union.default of <class 'reagent.optimizer.union.Optimizer__Union'>>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), optimizer_value_net: reagent.optimizer.union.Optimizer__Union = Field(name='optimizer_value_net',type=<class 'reagent.optimizer.union.Optimizer__Union'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<bound method Optimizer__Union.default of <class 'reagent.optimizer.union.Optimizer__Union'>>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), actions: List[str] = Field(name='actions',type=typing.List[str],default=<dataclasses._MISSING_TYPE object>,default_factory=<class 'list'>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), reward_clip: float = 1000000.0, normalize: bool = True, subtract_mean: bool = True, offset_clamp_min: bool = False, update_freq: int = 1, update_epochs: int = 1, ppo_batch_size: int = 1, ppo_epsilon: float = 0.2, entropy_weight: float = 0.0, value_net: Optional[reagent.models.base.ModelBase] = None)

Bases: reagent.training.reagent_lightning_module.ReAgentLightningModule

Proximal Policy Optimization (PPO). See https://arxiv.org/pdf/1707.06347.pdf This is the “clip” version of PPO. It does not include: - KL divergence - Bootstrapping with a critic model (our approach only works if full trajectories up to terminal state are fed in) Optionally, a value network can be trained and used as a baseline for rewards.

allow_zero_length_dataloader_with_multiple_devices: bool

configure_optimizers()

Choose what optimizers and learning-rate schedulers to use in your optimization. Normally you’d need one. But in the case of GANs or similar you might have multiple.

Returns

Any of these 6 options.

Single optimizer.
List or Tuple of optimizers.
Two lists - The first list has multiple optimizers, and the second has multiple LR schedulers (or multiple lr_scheduler_config).
Dictionary, with an "optimizer" key, and (optionally) a "lr_scheduler" key whose value is a single LR scheduler or lr_scheduler_config.
Tuple of dictionaries as described above, with an optional "frequency" key.
None - Fit will run without any optimizer.

The lr_scheduler_config is a dictionary which contains the scheduler and its associated configuration. The default configuration is shown below.

lr_scheduler_config = {
    # REQUIRED: The scheduler instance
    "scheduler": lr_scheduler,
    # The unit of the scheduler's step size, could also be 'step'.
    # 'epoch' updates the scheduler on epoch end whereas 'step'
    # updates it after a optimizer update.
    "interval": "epoch",
    # How many epochs/steps should pass between calls to
    # `scheduler.step()`. 1 corresponds to updating the learning
    # rate after every epoch/step.
    "frequency": 1,
    # Metric to to monitor for schedulers like `ReduceLROnPlateau`
    "monitor": "val_loss",
    # If set to `True`, will enforce that the value specified 'monitor'
    # is available when the scheduler is updated, thus stopping
    # training if not found. If set to `False`, it will only produce a warning
    "strict": True,
    # If using the `LearningRateMonitor` callback to monitor the
    # learning rate progress, this keyword can be used to specify
    # a custom logged name
    "name": None,
}

When there are schedulers in which the .step() method is conditioned on a value, such as the torch.optim.lr_scheduler.ReduceLROnPlateau scheduler, Lightning requires that the lr_scheduler_config contains the keyword "monitor" set to the metric name that the scheduler should be conditioned on.

Metrics can be made available to monitor by simply logging it using self.log('metric_to_track', metric_val) in your LightningModule.

Note

The frequency value specified in a dict along with the optimizer key is an int corresponding to the number of sequential batches optimized with the specific optimizer. It should be given to none or to all of the optimizers. There is a difference between passing multiple optimizers in a list, and passing multiple optimizers in dictionaries with a frequency of 1:

This is different from the frequency value specified in the lr_scheduler_config mentioned above.

def configure_optimizers(self):
    optimizer_one = torch.optim.SGD(self.model.parameters(), lr=0.01)
    optimizer_two = torch.optim.SGD(self.model.parameters(), lr=0.01)
    return [
        {"optimizer": optimizer_one, "frequency": 5},
        {"optimizer": optimizer_two, "frequency": 10},
    ]

In this example, the first optimizer will be used for the first 5 steps, the second optimizer for the next 10 steps and that cycle will continue. If an LR scheduler is specified for an optimizer using the lr_scheduler key in the above dict, the scheduler will only be updated when its optimizer is being used.

Examples:

# most cases. no learning rate scheduler
def configure_optimizers(self):
    return Adam(self.parameters(), lr=1e-3)

# multiple optimizer case (e.g.: GAN)
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    return gen_opt, dis_opt

# example with learning rate schedulers
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    dis_sch = CosineAnnealing(dis_opt, T_max=10)
    return [gen_opt, dis_opt], [dis_sch]

# example with step-based learning rate schedulers
# each optimizer has its own scheduler
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    gen_sch = {
        'scheduler': ExponentialLR(gen_opt, 0.99),
        'interval': 'step'  # called after each training step
    }
    dis_sch = CosineAnnealing(dis_opt, T_max=10) # called every epoch
    return [gen_opt, dis_opt], [gen_sch, dis_sch]

# example with optimizer frequencies
# see training procedure in `Improved Training of Wasserstein GANs`, Algorithm 1
# https://arxiv.org/abs/1704.00028
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    n_critic = 5
    return (
        {'optimizer': dis_opt, 'frequency': n_critic},
        {'optimizer': gen_opt, 'frequency': 1}
    )

Note

Some things to know:

Lightning calls .backward() and .step() on each optimizer and learning rate scheduler as needed.
If you use 16-bit precision (precision=16), Lightning will automatically handle the optimizers.
If you use multiple optimizers, training_step() will have an additional optimizer_idx parameter.
If you use torch.optim.LBFGS, Lightning handles the closure function automatically for you.
If you use multiple optimizers, gradients will be calculated only for the parameters of current optimizer at each training step.
If you need to control how often those optimizers step or override the default .step() schedule, override the optimizer_step() hook.

get_optimizers()

precision: int

prepare_data_per_node: bool

training: bool

training_step(training_batch: Union[reagent.core.types.PolicyGradientInput, Dict[str, torch.Tensor]], batch_idx: int)

Here you compute and return the training loss and some additional metrics for e.g. the progress bar or logger.

Parameters

batch (Tensor | (Tensor, …) | [Tensor, …]) – The output of your DataLoader. A tensor, tuple or list.
batch_idx (int) – Integer displaying index of this batch
optimizer_idx (int) – When using multiple optimizers, this argument will also be present.
hiddens (Any) – Passed in if :paramref:`~pytorch_lightning.core.lightning.LightningModule.truncated_bptt_steps` > 0.

Returns

Any of.

Tensor - The loss tensor
dict - A dictionary. Can include any keys, but must include the key 'loss'
None - Training will skip to the next batch. This is only for automatic optimization.
This is not supported for multi-GPU, TPU, IPU, or DeepSpeed.

In this step you’d normally do the forward pass and calculate the loss for a batch. You can also do fancier things like multiple forward passes or something model specific.

Example:

def training_step(self, batch, batch_idx):
    x, y, z = batch
    out = self.encoder(x)
    loss = self.loss(out, x)
    return loss

If you define multiple optimizers, this step will be called with an additional optimizer_idx parameter.

# Multiple optimizers (e.g.: GANs)
def training_step(self, batch, batch_idx, optimizer_idx):
    if optimizer_idx == 0:
        # do training_step with encoder
        ...
    if optimizer_idx == 1:
        # do training_step with decoder
        ...

If you add truncated back propagation through time you will also get an additional argument with the hidden states of the previous step.

# Truncated back-propagation through time
def training_step(self, batch, batch_idx, hiddens):
    # hiddens are the hidden states from the previous truncated backprop step
    out, hiddens = self.lstm(data, hiddens)
    loss = ...
    return {"loss": loss, "hiddens": hiddens}

Note

The loss value shown in the progress bar is smoothed (averaged) over the last values, so it differs from the actual loss returned in train/validation step.

update_model()

use_amp: bool

class reagent.training.PPOTrainerParameters(gamma: float = 0.9, optimizer: reagent.optimizer.union.Optimizer__Union = <factory>, optimizer_value_net: reagent.optimizer.union.Optimizer__Union = <factory>, actions: List[str] = <factory>, reward_clip: float = 1000000.0, normalize: bool = True, subtract_mean: bool = True, offset_clamp_min: bool = False, update_freq: int = 1, update_epochs: int = 1, ppo_batch_size: int = 1, ppo_epsilon: float = 0.2, entropy_weight: float = 0.0)

Bases: object

actions: List[str]

asdict()

entropy_weight: float = 0.0

gamma: float = 0.9

normalize: bool = True

offset_clamp_min: bool = False

optimizer: reagent.optimizer.union.Optimizer__Union

optimizer_value_net: reagent.optimizer.union.Optimizer__Union

ppo_batch_size: int = 1

ppo_epsilon: float = 0.2

reward_clip: float = 1000000.0

subtract_mean: bool = True

update_epochs: int = 1

update_freq: int = 1

class reagent.training.ParametricDQNTrainer(q_network, q_network_target, reward_network, rl: reagent.core.parameters.RLParameters = Field(name='rl', type=<class 'reagent.core.parameters.RLParameters'>, default=<dataclasses._MISSING_TYPE object>, default_factory=<class 'reagent.core.parameters.RLParameters'>, init=True, repr=True, hash=None, compare=True, metadata=mappingproxy({}), _field_type=_FIELD), double_q_learning: bool = True, minibatches_per_step: int = 1, optimizer: reagent.optimizer.union.Optimizer__Union = Field(name='optimizer', type=<class 'reagent.optimizer.union.Optimizer__Union'>, default=<dataclasses._MISSING_TYPE object>, default_factory=<bound method Optimizer__Union.default of <class 'reagent.optimizer.union.Optimizer__Union'>>, init=True, repr=True, hash=None, compare=True, metadata=mappingproxy({}), _field_type=_FIELD))

Bases: reagent.training.dqn_trainer_base.DQNTrainerMixin, reagent.training.rl_trainer_pytorch.RLTrainerMixin, reagent.training.reagent_lightning_module.ReAgentLightningModule

configure_optimizers()

Choose what optimizers and learning-rate schedulers to use in your optimization. Normally you’d need one. But in the case of GANs or similar you might have multiple.

Returns

Any of these 6 options.

Single optimizer.
List or Tuple of optimizers.
Two lists - The first list has multiple optimizers, and the second has multiple LR schedulers (or multiple lr_scheduler_config).
Dictionary, with an "optimizer" key, and (optionally) a "lr_scheduler" key whose value is a single LR scheduler or lr_scheduler_config.
Tuple of dictionaries as described above, with an optional "frequency" key.
None - Fit will run without any optimizer.

The lr_scheduler_config is a dictionary which contains the scheduler and its associated configuration. The default configuration is shown below.

lr_scheduler_config = {
    # REQUIRED: The scheduler instance
    "scheduler": lr_scheduler,
    # The unit of the scheduler's step size, could also be 'step'.
    # 'epoch' updates the scheduler on epoch end whereas 'step'
    # updates it after a optimizer update.
    "interval": "epoch",
    # How many epochs/steps should pass between calls to
    # `scheduler.step()`. 1 corresponds to updating the learning
    # rate after every epoch/step.
    "frequency": 1,
    # Metric to to monitor for schedulers like `ReduceLROnPlateau`
    "monitor": "val_loss",
    # If set to `True`, will enforce that the value specified 'monitor'
    # is available when the scheduler is updated, thus stopping
    # training if not found. If set to `False`, it will only produce a warning
    "strict": True,
    # If using the `LearningRateMonitor` callback to monitor the
    # learning rate progress, this keyword can be used to specify
    # a custom logged name
    "name": None,
}

When there are schedulers in which the .step() method is conditioned on a value, such as the torch.optim.lr_scheduler.ReduceLROnPlateau scheduler, Lightning requires that the lr_scheduler_config contains the keyword "monitor" set to the metric name that the scheduler should be conditioned on.

Metrics can be made available to monitor by simply logging it using self.log('metric_to_track', metric_val) in your LightningModule.

Note

The frequency value specified in a dict along with the optimizer key is an int corresponding to the number of sequential batches optimized with the specific optimizer. It should be given to none or to all of the optimizers. There is a difference between passing multiple optimizers in a list, and passing multiple optimizers in dictionaries with a frequency of 1:

This is different from the frequency value specified in the lr_scheduler_config mentioned above.

def configure_optimizers(self):
    optimizer_one = torch.optim.SGD(self.model.parameters(), lr=0.01)
    optimizer_two = torch.optim.SGD(self.model.parameters(), lr=0.01)
    return [
        {"optimizer": optimizer_one, "frequency": 5},
        {"optimizer": optimizer_two, "frequency": 10},
    ]

In this example, the first optimizer will be used for the first 5 steps, the second optimizer for the next 10 steps and that cycle will continue. If an LR scheduler is specified for an optimizer using the lr_scheduler key in the above dict, the scheduler will only be updated when its optimizer is being used.

Examples:

# most cases. no learning rate scheduler
def configure_optimizers(self):
    return Adam(self.parameters(), lr=1e-3)

# multiple optimizer case (e.g.: GAN)
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    return gen_opt, dis_opt

# example with learning rate schedulers
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    dis_sch = CosineAnnealing(dis_opt, T_max=10)
    return [gen_opt, dis_opt], [dis_sch]

# example with step-based learning rate schedulers
# each optimizer has its own scheduler
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    gen_sch = {
        'scheduler': ExponentialLR(gen_opt, 0.99),
        'interval': 'step'  # called after each training step
    }
    dis_sch = CosineAnnealing(dis_opt, T_max=10) # called every epoch
    return [gen_opt, dis_opt], [gen_sch, dis_sch]

# example with optimizer frequencies
# see training procedure in `Improved Training of Wasserstein GANs`, Algorithm 1
# https://arxiv.org/abs/1704.00028
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    n_critic = 5
    return (
        {'optimizer': dis_opt, 'frequency': n_critic},
        {'optimizer': gen_opt, 'frequency': 1}
    )

Note

Some things to know:

Lightning calls .backward() and .step() on each optimizer and learning rate scheduler as needed.
If you use 16-bit precision (precision=16), Lightning will automatically handle the optimizers.
If you use multiple optimizers, training_step() will have an additional optimizer_idx parameter.
If you use torch.optim.LBFGS, Lightning handles the closure function automatically for you.
If you use multiple optimizers, gradients will be calculated only for the parameters of current optimizer at each training step.
If you need to control how often those optimizers step or override the default .step() schedule, override the optimizer_step() hook.

get_detached_model_outputs(state, action) → Tuple[torch.Tensor, torch.Tensor]: Gets the q values from the model and target networks

train_step_gen(training_batch: reagent.core.types.ParametricDqnInput, batch_idx: int): Implement training step as generator here

class reagent.training.ParametricDQNTrainerParameters(rl: reagent.core.parameters.RLParameters = <factory>, double_q_learning: bool = True, minibatches_per_step: int = 1, optimizer: reagent.optimizer.union.Optimizer__Union = <factory>)

Bases: object

asdict()

double_q_learning: bool = True

minibatches_per_step: int = 1

optimizer: reagent.optimizer.union.Optimizer__Union

rl: reagent.core.parameters.RLParameters

class reagent.training.QRDQNTrainer(q_network, q_network_target, metrics_to_score=None, reward_network=None, q_network_cpe=None, q_network_cpe_target=None, actions: List[str] = Field(name='actions',type=typing.List[str],default=<dataclasses._MISSING_TYPE object>,default_factory=<class 'list'>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), rl: reagent.core.parameters.RLParameters = Field(name='rl',type=<class 'reagent.core.parameters.RLParameters'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<class 'reagent.core.parameters.RLParameters'>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), double_q_learning: bool = True, num_atoms: int = 51, minibatch_size: int = 1024, minibatches_per_step: int = 1, optimizer: reagent.optimizer.union.Optimizer__Union = Field(name='optimizer',type=<class 'reagent.optimizer.union.Optimizer__Union'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<bound method Optimizer__Union.default of <class 'reagent.optimizer.union.Optimizer__Union'>>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), cpe_optimizer: reagent.optimizer.union.Optimizer__Union = Field(name='cpe_optimizer',type=<class 'reagent.optimizer.union.Optimizer__Union'>,default=<dataclasses._MISSING_TYPE object>,default_factory=<bound method Optimizer__Union.default of <class 'reagent.optimizer.union.Optimizer__Union'>>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=_FIELD), evaluation: reagent.core.parameters.EvaluationParameters = Field(name=None,type=None,default=<dataclasses._MISSING_TYPE object>,default_factory=<class 'reagent.core.parameters.EvaluationParameters'>,init=True,repr=True,hash=None,compare=True,metadata=mappingproxy({}),_field_type=None))

Bases: reagent.training.dqn_trainer_base.DQNTrainerBaseLightning

Implementation of QR-DQN (Quantile Regression Deep Q-Network)

See https://arxiv.org/abs/1710.10044 for details

allow_zero_length_dataloader_with_multiple_devices: bool

argmax_with_mask(q_values, possible_actions_mask)

boost_rewards(rewards: torch.Tensor, actions: torch.Tensor) → torch.Tensor

configure_optimizers()

Choose what optimizers and learning-rate schedulers to use in your optimization. Normally you’d need one. But in the case of GANs or similar you might have multiple.

Returns

Any of these 6 options.

Single optimizer.
List or Tuple of optimizers.
Two lists - The first list has multiple optimizers, and the second has multiple LR schedulers (or multiple lr_scheduler_config).
Dictionary, with an "optimizer" key, and (optionally) a "lr_scheduler" key whose value is a single LR scheduler or lr_scheduler_config.
Tuple of dictionaries as described above, with an optional "frequency" key.
None - Fit will run without any optimizer.

The lr_scheduler_config is a dictionary which contains the scheduler and its associated configuration. The default configuration is shown below.

lr_scheduler_config = {
    # REQUIRED: The scheduler instance
    "scheduler": lr_scheduler,
    # The unit of the scheduler's step size, could also be 'step'.
    # 'epoch' updates the scheduler on epoch end whereas 'step'
    # updates it after a optimizer update.
    "interval": "epoch",
    # How many epochs/steps should pass between calls to
    # `scheduler.step()`. 1 corresponds to updating the learning
    # rate after every epoch/step.
    "frequency": 1,
    # Metric to to monitor for schedulers like `ReduceLROnPlateau`
    "monitor": "val_loss",
    # If set to `True`, will enforce that the value specified 'monitor'
    # is available when the scheduler is updated, thus stopping
    # training if not found. If set to `False`, it will only produce a warning
    "strict": True,
    # If using the `LearningRateMonitor` callback to monitor the
    # learning rate progress, this keyword can be used to specify
    # a custom logged name
    "name": None,
}

When there are schedulers in which the .step() method is conditioned on a value, such as the torch.optim.lr_scheduler.ReduceLROnPlateau scheduler, Lightning requires that the lr_scheduler_config contains the keyword "monitor" set to the metric name that the scheduler should be conditioned on.

Metrics can be made available to monitor by simply logging it using self.log('metric_to_track', metric_val) in your LightningModule.

Note

The frequency value specified in a dict along with the optimizer key is an int corresponding to the number of sequential batches optimized with the specific optimizer. It should be given to none or to all of the optimizers. There is a difference between passing multiple optimizers in a list, and passing multiple optimizers in dictionaries with a frequency of 1:

This is different from the frequency value specified in the lr_scheduler_config mentioned above.

def configure_optimizers(self):
    optimizer_one = torch.optim.SGD(self.model.parameters(), lr=0.01)
    optimizer_two = torch.optim.SGD(self.model.parameters(), lr=0.01)
    return [
        {"optimizer": optimizer_one, "frequency": 5},
        {"optimizer": optimizer_two, "frequency": 10},
    ]

In this example, the first optimizer will be used for the first 5 steps, the second optimizer for the next 10 steps and that cycle will continue. If an LR scheduler is specified for an optimizer using the lr_scheduler key in the above dict, the scheduler will only be updated when its optimizer is being used.

Examples:

# most cases. no learning rate scheduler
def configure_optimizers(self):
    return Adam(self.parameters(), lr=1e-3)

# multiple optimizer case (e.g.: GAN)
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    return gen_opt, dis_opt

# example with learning rate schedulers
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    dis_sch = CosineAnnealing(dis_opt, T_max=10)
    return [gen_opt, dis_opt], [dis_sch]

# example with step-based learning rate schedulers
# each optimizer has its own scheduler
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    gen_sch = {
        'scheduler': ExponentialLR(gen_opt, 0.99),
        'interval': 'step'  # called after each training step
    }
    dis_sch = CosineAnnealing(dis_opt, T_max=10) # called every epoch
    return [gen_opt, dis_opt], [gen_sch, dis_sch]

# example with optimizer frequencies
# see training procedure in `Improved Training of Wasserstein GANs`, Algorithm 1
# https://arxiv.org/abs/1704.00028
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    n_critic = 5
    return (
        {'optimizer': dis_opt, 'frequency': n_critic},
        {'optimizer': gen_opt, 'frequency': 1}
    )

Note

Some things to know:

Lightning calls .backward() and .step() on each optimizer and learning rate scheduler as needed.
If you use 16-bit precision (precision=16), Lightning will automatically handle the optimizers.
If you use multiple optimizers, training_step() will have an additional optimizer_idx parameter.
If you use torch.optim.LBFGS, Lightning handles the closure function automatically for you.
If you use multiple optimizers, gradients will be calculated only for the parameters of current optimizer at each training step.
If you need to control how often those optimizers step or override the default .step() schedule, override the optimizer_step() hook.

get_detached_model_outputs(state: reagent.core.types.FeatureData) → Tuple[torch.Tensor, torch.Tensor]: Gets the q values from the model and target networks

huber(x)

precision: int

prepare_data_per_node: bool

rl_parameters: reagent.core.parameters.RLParameters

train_step_gen(training_batch: reagent.core.types.DiscreteDqnInput, batch_idx: int): Implement training step as generator here

training: bool

use_amp: bool

class reagent.training.QRDQNTrainerParameters(actions: List[str] = <factory>, rl: reagent.core.parameters.RLParameters = <factory>, double_q_learning: bool = True, num_atoms: int = 51, minibatch_size: int = 1024, minibatches_per_step: int = 1, optimizer: reagent.optimizer.union.Optimizer__Union = <factory>, cpe_optimizer: reagent.optimizer.union.Optimizer__Union = <factory>)

Bases: object

actions: List[str]

asdict()

cpe_optimizer: reagent.optimizer.union.Optimizer__Union

double_q_learning: bool = True

minibatch_size: int = 1024

minibatches_per_step: int = 1

num_atoms: int = 51

optimizer: reagent.optimizer.union.Optimizer__Union

rl: reagent.core.parameters.RLParameters

class reagent.training.ReAgentLightningModule(automatic_optimization=True)

Bases: pytorch_lightning.core.lightning.LightningModule

allow_zero_length_dataloader_with_multiple_devices: bool

increase_next_stopping_epochs(num_epochs: int)

on_epoch_end(): Called when either of train/val/test epoch ends.

on_test_batch_end(*args, **kwargs)

Called in the test loop after the batch.

Parameters

outputs – The outputs of test_step_end(test_step(x))
batch – The batched data as it is returned by the test DataLoader.
batch_idx – the index of the batch
dataloader_idx – the index of the dataloader

on_train_batch_end(*args, **kwargs)

Called in the training loop after the batch.

Parameters

outputs – The outputs of training_step_end(training_step(x))
batch – The batched data as it is returned by the training DataLoader.
batch_idx – the index of the batch
unused – Deprecated argument. Will be removed in v1.7.

on_validation_batch_end(*args, **kwargs)

Called in the validation loop after the batch.

Parameters

outputs – The outputs of validation_step_end(validation_step(x))
batch – The batched data as it is returned by the validation DataLoader.
batch_idx – the index of the batch
dataloader_idx – the index of the dataloader

optimizers(use_pl_optimizer: bool = True)

Returns the optimizer(s) that are being used during training. Useful for manual optimization.

Parameters: use_pl_optimizer – If True, will wrap the optimizer(s) in a LightningOptimizer for automatic handling of precision and profiling.
Returns: A single optimizer, or a list of optimizers in case multiple ones are present.

precision: int

prepare_data_per_node: bool

property reporter

set_clean_stop(clean_stop: bool)