trl-lib

/

pythia-6.9b-deduped-tldr-online-dpo

like 0

Follow

TRL 38

• "transformers.trainer_utils.HubStrategy",
• "accelerate.utils.deepspeed.HfDeepSpeedConfig",
• "accelerate.utils.dataclasses.DistributedType",
• "transformers.training_args.OptimizerNames",
• "transformers.integrations.deepspeed.HfDeepSpeedConfig",
• "accelerate.utils.dataclasses.DeepSpeedPlugin",
• "transformers.trainer_pt_utils.AcceleratorConfig",
• "trl.trainer.online_dpo_config.OnlineDPOConfig",
• "transformers.trainer_utils.IntervalStrategy",
• "accelerate.state.PartialState",
• "transformers.trainer_utils.SchedulerType",
• "torch.device"

How to fix it?