determined.TrialContext¶

The TrialContext provides useful methods for writing Trial subclasses. All Trial subclasses receive a TrialContext object as an argument to their __init__() method:

PyTorchTrial subclasses receive a plain PyTorchTrialContext.
TFKerasTrial subclasses receive a TFKerasTrialContext.
EstimatorTrial subclasses receive an EstimatorTrialContext.

`determined.TrialContext`¶

class determined.TrialContext(env: determined._env_context.EnvContext, hvd_config: determined.horovod.HorovodContext)¶

A base class that all TrialContexts will inherit from. The context passed to the User’s Trial.__init__() will inherit from this class.

TrialContext always has an instance of DistributedContext accessible as the distributed attribute for information related to distributed training.

classmethod from_config(config: Dict[str, Any]) → determined._train_context._TrainContext¶

Create an context object suitable for debugging outside of Determined.

An example for a subclass of PyTorchTrial:

config = { ... }
context = det.pytorch.PyTorchTrialContext.from_config(config)
my_trial = MyPyTorchTrial(context)

train_ds = my_trial.build_training_data_loader()
for epoch_idx in range(3):
    for batch_idx, batch in enumerate(train_ds):
        metrics = my_trial.train_batch(batch, epoch_idx, batch_idx)
        ...

An example for a subclass of TFKerasTrial:

config = { ... }
context = det.keras.TFKerasTrialContext.from_config(config)
my_trial = tf_keras_one_var_model.OneVarTrial(context)

model = my_trial.build_model()
model.fit(my_trial.build_training_data_loader())
eval_metrics = model.evaluate(my_trial.build_validation_data_loader())

Parameters: config – An experiment config file, in dictionary form.

get_data_config() → Dict[str, Any]¶: Return the data configuration.

get_experiment_config() → Dict[str, Any]¶: Return the experiment configuration.

get_experiment_id() → int¶: Return the experiment ID of the current trial.

get_global_batch_size() → int¶: Return the global batch size.

get_hparam(name: str) → Any¶: Return the current value of the hyperparameter with the given name.

get_hparams() → Dict[str, Any]¶: Return a dictionary of hyperparameter names to values.

get_per_slot_batch_size() → int¶: Return the per-slot batch size. When a model is trained with a single GPU, this is equal to the global batch size. When multi-GPU training is used, this is equal to the global batch size divided by the number of GPUs used to train the model.

get_stop_requested() → bool¶: Return whether a trial stoppage has been requested.

get_trial_id() → int¶: Return the trial ID of the current trial.

set_stop_requested(stop_requested: bool) → None¶: Set a flag to request a trial stoppage. When this flag is set to True, we finish the step, checkpoint, then exit.