Duplicated from  mlfoundations-dev/DCFT-hero_run_2_fix_conversations-etash

open-thoughts

/

OpenThinker2-32B

like 57

Follow

OpenThoughts 849

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

How to fix it?