distily

/

distily_multi_attn_experiment_ortho

like 0

• "torch.bfloat16",
• "numpy.core.multiarray.scalar",
• "numpy.dtype"

How to fix it?

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

How to fix it?