readme example does not work for quantization pissa model

[chuangzhidan]

from trl import SFTTrainer
from datasets import load_dataset
from transformers import AutoTokenizer, AutoModelForCausalLM
from peft import PeftModel
MODEL_ID = "PiSSA-Llama-2-7b-hf-r128"
residual_model = AutoModelForCausalLM.from_pretrained(MODEL_ID,device_map="auto")
model = PeftModel.from_pretrained(residual_model, MODEL_ID, subfolder = "pissa_init", is_trainable=True)
tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)
dataset = load_dataset("imdb", split="train[:1%]") # Only use 1% of the dataset
trainer = SFTTrainer(
model=peft_model,
train_dataset=dataset,
dataset_text_field="text",
max_seq_length=128,
tokenizer=tokenizer,
)
trainer.train()
peft_model.save_pretrained("pissa-llama-2-7b-ft")

this example does not work for quantization model like

fxmeng/PiSSA-Llama-2-7B-r16-4bit-5iter, "fxmeng/PiSSA-Qwen2-7B-4bit-r128-5iter" and so on.

[chuangzhidan]

model = PeftModel.from_pretrained(residual_model, MODEL_ID, subfolder = "pissa_init", is_trainable=True) is a typo? should be：
peft_model = PeftModel.from_pretrained(residual_model, MODEL_ID, subfolder = "pissa_init", is_trainable=True)

because
trainer = SFTTrainer(
model=peft_model,
train_dataset=dataset,
dataset_text_field="text",
max_seq_length=128,
tokenizer=tokenizer,
)

also,those 2 parameters doesn't work anymore with newest trl-0.13.0:
dataset_text_field="text",
max_seq_length=128,

so ,how to change the code acccordingly?