Training config

TrainingConfig

class src.config.runs.training.arguments.TrainingConfig(_case_sensitive: bool | None = None, _env_prefix: str | None = None, _env_file: DotenvType | None = PosixPath('.'), _env_file_encoding: str | None = None, _env_nested_delimiter: str | None = None, _secrets_dir: str | Path | None = None, *, num_train_epochs: int = 10, per_device_train_batch_size: int = 4, per_device_eval_batch_size: int = 20, warmup_steps: int = 200, weight_decay: float = 0.01, load_best_model_at_end: bool = True, evaluation_strategy: str = 'epoch', save_strategy: str = 'epoch', dataloader_pin_memory: bool = True)[source]

Bases: Settings

class Config[source]

Bases: object

env_prefix = 'runs_training_args_'

_abc_impl = <_abc._abc_data object>

dataloader_pin_memory: bool

evaluation_strategy: str

load_best_model_at_end: bool

model_config: ClassVar[SettingsConfigDict] = {'arbitrary_types_allowed': True, 'case_sensitive': False, 'env_file': '.env', 'env_file_encoding': 'utf-8', 'env_nested_delimiter': None, 'env_prefix': 'runs_training_args_', 'extra': 'ignore', 'protected_namespaces': ('model_', 'settings_'), 'secrets_dir': None, 'validate_default': True}: Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[dict[str, FieldInfo]] = {'dataloader_pin_memory': FieldInfo(annotation=bool, required=False, default=True), 'evaluation_strategy': FieldInfo(annotation=str, required=False, default='epoch'), 'load_best_model_at_end': FieldInfo(annotation=bool, required=False, default=True), 'num_train_epochs': FieldInfo(annotation=int, required=False, default=10), 'per_device_eval_batch_size': FieldInfo(annotation=int, required=False, default=20), 'per_device_train_batch_size': FieldInfo(annotation=int, required=False, default=4), 'save_strategy': FieldInfo(annotation=str, required=False, default='epoch'), 'warmup_steps': FieldInfo(annotation=int, required=False, default=200), 'weight_decay': FieldInfo(annotation=float, required=False, default=0.01)}

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo].

This replaces Model.__fields__ from Pydantic V1.

num_train_epochs: int

per_device_eval_batch_size: int

per_device_train_batch_size: int

save_strategy: str

warmup_steps: int

weight_decay: float

DefaultConfig

class src.config.runs.training.default.DefaultConfig(_case_sensitive: bool | None = None, _env_prefix: str | None = None, _env_file: DotenvType | None = PosixPath('.'), _env_file_encoding: str | None = None, _env_nested_delimiter: str | None = None, _secrets_dir: str | Path | None = None, *, bert_model_id: str = 'GroNLP/bert-base-dutch-cased', distil_bert_model_id: str = 'Geotrend/distilbert-base-nl-cased', setfit_model_id: str = 'sentence-transformers/paraphrase-mpnet-base-v2', keep_negative_examples: bool = False)[source]