wls04

/

rm-biased

like 0

• "torch.ByteStorage",
• "collections.OrderedDict",
• "numpy.dtype",
• "numpy.ndarray",
• "_codecs.encode",
• "torch._utils._rebuild_tensor_v2",
• "numpy._core.multiarray._reconstruct"

How to fix it?

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

How to fix it?