Jeong-su

/

car

like 0

• "deepspeed.utils.tensor_fragment.fragment_address",
• "collections.OrderedDict",
• "torch.FloatStorage",
• "deepspeed.runtime.fp16.loss_scaler.LossScaler",
• "deepspeed.runtime.zero.config.ZeroStageEnum",
• "torch._utils._rebuild_tensor_v2"

How to fix it?

• "deepspeed.utils.tensor_fragment.fragment_address",
• "collections.OrderedDict",
• "torch.FloatStorage",
• "deepspeed.runtime.fp16.loss_scaler.LossScaler",
• "deepspeed.runtime.zero.config.ZeroStageEnum",
• "torch._utils._rebuild_tensor_v2"

How to fix it?

• "torch.BFloat16Storage",
• "torch.ByteStorage",
• "collections.OrderedDict",
• "torch.FloatStorage",
• "__builtin__.set",
• "torch._utils._rebuild_tensor_v2",
• "torch.Size"

How to fix it?

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

How to fix it?

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

How to fix it?

• "transformers.trainer_utils.IntervalStrategy",
• "accelerate.utils.dataclasses.DistributedType",
• "transformers.trainer_pt_utils.AcceleratorConfig",
• "transformers.trainer_utils.SchedulerType",
• "transformers.training_args.OptimizerNames",
• "transformers.integrations.deepspeed.HfDeepSpeedConfig",
• "accelerate.state.PartialState",
• "transformers.training_args_seq2seq.Seq2SeqTrainingArguments",
• "torch.device",
• "accelerate.utils.dataclasses.DeepSpeedPlugin",
• "torch.bfloat16",
• "transformers.trainer_utils.HubStrategy",
• "transformers.integrations.deepspeed.HfTrainerDeepSpeedConfig"

How to fix it?