babylm-seqlen

/

opt-8192-warmup

like 0

Follow

BabyLM Sequence Length 4

• "transformers.trainer_utils.IntervalStrategy",
• "accelerate.utils.dataclasses.DistributedType",
• "transformers.training_args.OptimizerNames",
• "torch.device",
• "transformers.trainer_utils.SchedulerType",
• "transformers.trainer_pt_utils.AcceleratorConfig",
• "transformers.trainer_utils.HubStrategy",
• "transformers.trainer_utils.SaveStrategy",
• "transformers.training_args.TrainingArguments",
• "accelerate.state.PartialState"

How to fix it?