vis4d.config.typing

Type definitions for configuration files.

Classes

`DataConfig`([initial_dictionary, type_safe, ...])	Configuration for a data set.
`ExperimentConfig`([initial_dictionary, ...])	Configuration for an experiment.
`ExperimentParameters`([initial_dictionary, ...])	Parameters for an experiment.
`LrSchedulerConfig`([initial_dictionary, ...])	Configuration for a learning rate scheduler.
`OptimizerConfig`([initial_dictionary, ...])	Configuration for an optimizer.
`ParamGroupCfg`	Parameter group config.
`ParameterSweepConfig`([initial_dictionary, ...])	Configuration for a parameter sweep.

class ParamGroupCfg[source]

Parameter group config.

custom_keys

List of custom keys.

Type:: list[str]

lr_mult

Learning rate multiplier.

Type:: NotRequired[float]

decay_mult

Weight Decay multiplier.

Type:: NotRequired[float]

class DataConfig(initial_dictionary=None, type_safe=True, convert_dict=True, *, allow_dotted_keys=False, sort_keys=True)[source]

Configuration for a data set.

This data object is used to configure the training and test data of an experiment. In particular, the train_dataloader and test_dataloader need to be config dicts that can be instantiated as a dataloader.

train_dataloader

Configuration for the training dataloader.

Type:: ConfigDict

test_dataloader

Configuration for the test dataloader.

Type:: ConfigDict

Example

>>> from vis4d.config.types import DataConfig
>>> from vis4d.zoo.base import class_config
>>> from my_package.data import MyDataLoader
>>> cfg = DataConfig()
>>> cfg.train_dataloader = class_config(MyDataLoader, ...)

class LrSchedulerConfig(initial_dictionary=None, type_safe=True, convert_dict=True, *, allow_dotted_keys=False, sort_keys=True)[source]

Configuration for a learning rate scheduler.

scheduler

Configuration for the learning rate scheduler.

Type:: ConfigDict

begin

Begin epoch.

Type:: int

end

End epoch.

Type:: int

epoch_based

Whether the learning rate scheduler is epoch based or step based.

Type:: bool

convert_epochs_to_steps

Whether to convert the begin and end for a step based scheduler to steps automatically based on length of train dataloader. Enables users to set the iteration breakpoints as epochs. Defaults to False.

Type:: bool

convert_attributes

List of attributes in the scheduler that should be converted to steps. Defaults to None.

Type:: list[str] | None

class OptimizerConfig(initial_dictionary=None, type_safe=True, convert_dict=True, *, allow_dotted_keys=False, sort_keys=True)[source]

Configuration for an optimizer.

optimizer

Configuration for the optimizer.

Type:: ConfigDict

lr_scheduler

Configuration for the learning rate scheduler.

Type:: list[LrSchedulerConfig] | None

param_groups

Configuration for the parameter groups.

Type:: list[ParamGroupCfg] | None

class ExperimentParameters(initial_dictionary=None, type_safe=True, convert_dict=True)[source]

Parameters for an experiment.

samples_per_gpu

Number of samples per GPU.

Type:: int

workers_per_gpu

Number of workers per GPU.

Type:: int

class ExperimentConfig(initial_dictionary=None, type_safe=True, convert_dict=True)[source]

Configuration for an experiment.

This data object is used to configure an experiment. It contains the minimal required configuration to run an experiment. In particular, the data, model, optimizers, and loss need to be config dicts that can be instantiated as a data set, model, optimizer, and loss function, respectively.

work_dir

The working directory for the experiment.

Type:: str | FieldReference

experiment_name

The name of the experiment.

Type:: str | FieldReference

timestamp

The timestamp of the experiment.

Type:: str | FieldReference

version

The version of the experiment.

Type:: str | FieldReference

output_dir

The output directory for the experiment.

Type:: str | FieldReference

seed

The random seed for the experiment.

Type:: int | FieldReference

log_every_n_steps

The number of steps after which the logs should be written.

Type:: int | FieldReference

use_tf32

Whether to use tf32.

Type:: bool | FieldReference

benchmark

Whether to enable benchmarking.

Type:: bool | FieldReference

params

Configuration for the experiment parameters.

Type:: ExperimentParameters

data

Configuration for the dataset.

Type:: DataConfig

model

Configuration for the model.

Type:: FieldConfigDictOrRef

loss

Configuration for the loss function.

Type:: FieldConfigDictOrRef

optimizers

Configuration for the optimizers.

Type:: list[OptimizerConfig]

data_connector

Configuration for the data connector.

Type:: FieldConfigDictOrRef

callbacks

Configuration for the callbacks which are used in the engine.

Type:: list[FieldConfigDictOrRef]

class ParameterSweepConfig(initial_dictionary=None, type_safe=True, convert_dict=True)[source]

Configuration for a parameter sweep.

Confguration object for a parameter sweep. It contains the minimal required configuration to run a parameter sweep.

method

Sweep method that should be used (e.g. grid)

Type:: str

sampling_args

Arguments that should be passed to the sweep method. E.g. for grid, this would be a list of tuples of the form (parameter_name, parameter_values).

Type:: list[tuple[str, Any]]

suffix

Suffix that should be appended to the output directory. This will be interpreted as a string template and can contain references to the sampling_args. E.g. “lr_{lr:.2e}_bs_{batch_size}”.

Type:: str