lautel

/

MEG-Mistral-7B-Instruct-v0.1

like 0

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

How to fix it?

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

How to fix it?

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

How to fix it?

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

How to fix it?

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

How to fix it?

• "torch.device",
• "torch.float32",
• "transformers.trainer_utils.IntervalStrategy",
• "transformers.trainer_utils.HubStrategy",
• "accelerate.utils.dataclasses.DeepSpeedPlugin",
• "trl.trainer.sft_config.SFTConfig",
• "transformers.trainer_pt_utils.AcceleratorConfig",
• "accelerate.state.PartialState",
• "accelerate.utils.dataclasses.DistributedType",
• "transformers.training_args.OptimizerNames",
• "transformers.integrations.deepspeed.HfTrainerDeepSpeedConfig",
• "transformers.trainer_utils.SchedulerType"

How to fix it?