stacking-babylm-2025

/

Continual_LR_only_1024

Like 0

Follow

Stacking Knowledge: Gradual Stacking on BabyLM 2025 Corpus 1

• "typing.Dict",
• "omegaconf.nodes.IntegerNode",
• "src.config.ContinualPretrainingParams",
• "__builtin__.long",
• "src.config.InfiniteLrSchedulerParams",
• "omegaconf.dictconfig.DictConfig",
• "transformers.training_args.TrainingArguments",
• "omegaconf.nodes.BooleanNode",
• "torch.device",
• "src.config.TokenizerParams",
• "src.config.DataCurriculumParams",
• "transformers.trainer_utils.SchedulerType",
• "collections.defaultdict",
• "omegaconf.nodes.StringNode",
• "__builtin__.unicode",
• "__builtin__.bool",
• "typing.List",
• "src.config.PacingFunctionParams",
• "src.config.ModelParams",
• "__builtin__.dict",
• "omegaconf.nodes.AnyNode",
• "src.config.BabyLMConfig",
• "src.config.TrainerParams",
• "transformers.trainer_pt_utils.AcceleratorConfig",
• "transformers.trainer_utils.SaveStrategy",
• "omegaconf.nodes.FloatNode",
• "transformers.trainer_utils.HubStrategy",
• "transformers.training_args.OptimizerNames",
• "src.config.GradualStackingParams",
• "src.config.ExperimentParams",
• "src.config.DatasetParams",
• "src.config.DataPreprocessingParams",
• "omegaconf.listconfig.ListConfig",
• "accelerate.utils.dataclasses.DistributedType",
• "typing.Any",
• "omegaconf.base.Metadata",
• "__builtin__.float",
• "accelerate.state.PartialState",
• "transformers.trainer_utils.IntervalStrategy",
• "omegaconf.base.ContainerMetadata"

How to fix it?