[Question] Adding several peft adapters and ensuring unsloth takes them into account

[tcarta]

Adding several peft adapters I would like to add several peft adapters to my LLM. Since unsloth does not override the add_adapter function in PEFT I currently use the following:

config = LoraConfig(
                r=self._r,
                lora_alpha=self._alpha,
                target_modules=target_modules,
                lora_dropout=0,
                bias="none"
            )
unsloth_peft_config = config.to_dict()
del unsloth_peft_config["task_type"]
peft_model =  FastLanguageModel.get_peft_model(
                    llm_module,
                    **unsloth_peft_config,
                    use_gradient_checkpointing="unsloth" if not self._use_cache else False
                )
peft_model.add_adapter("new_adapter_1", config)
peft_model.add_adapter("new_adapter_2", config)
peft_model.add_adapter("new_adapter_3", config)

I am wondering if it is the correct method to adopt or if there is a better way to do it? Are the new adapters taken into account by unsloth?