ExplosionNuclear

/

GPT2-small-BERT-modified-loss

like 0

• "accelerate.utils.dataclasses.DistributedType",
• "accelerate.state.PartialState",
• "transformers.training_args.OptimizerNames",
• "__builtin__.list",
• "transformers.trainer_utils.HubStrategy",
• "__builtin__.dict",
• "transformers.trainer_utils.IntervalStrategy",
• "omegaconf.listconfig.ListConfig",
• "omegaconf.dictconfig.DictConfig",
• "omegaconf.base.Metadata",
• "transformers.trainer_utils.SaveStrategy",
• "collections.defaultdict",
• "transformers.trainer_utils.SchedulerType",
• "trl.trainer.sft_config.SFTConfig",
• "omegaconf.nodes.AnyNode",
• "typing.Any",
• "transformers.trainer_pt_utils.AcceleratorConfig",
• "omegaconf.base.ContainerMetadata",
• "torch.device",
• "__builtin__.long"

How to fix it?