Siesher

/

qwen3-1.7b-reasoning-lora_LoRA

like 1

• "torch.FloatStorage",
• "torch.serialization._get_layout",
• "collections.OrderedDict",
• "torch.device",
• "torch._tensor._rebuild_from_type_v2",
• "torch._utils._rebuild_wrapper_subclass",
• "torch.ByteStorage",
• "torchao.optim.subclass_8bit.OptimState8bit",
• "torch.float32",
• "torch._utils._rebuild_tensor_v2"

How to fix it?

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

How to fix it?

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

How to fix it?