Fix DPO finetuning example (#12313)

python/llm/example/GPU/LLM-Finetuning/DPO/README.md

@@ -17,8 +17,9 @@ conda create -n llm python=3.11
 conda activate llm
 # below command will install intel_extension_for_pytorch==2.1.10+xpu as default
 pip install --pre --upgrade ipex-llm[xpu] --extra-index-url https://pytorch-extension.intel.com/release-whl/stable/xpu/us/
-pip install transformers==4.36.0 datasets
-pip install trl peft==0.10.0
+pip install datasets
+pip install peft==0.10.0
+pip install 'trl<0.9'
 # Note, if you don't want to reinstall BNBs dependencies, append the `--no-deps` flag!
 pip install --no-deps --force-reinstall 'https://github.com/bitsandbytes-foundation/bitsandbytes/releases/download/continuous-release_multi-backend-refactor/bitsandbytes-0.44.1.dev0-py3-none-manylinux_2_24_x86_64.whl'
 ```

python/llm/example/GPU/LLM-Finetuning/DPO/dpo_finetuning.py

@@ -34,12 +34,12 @@
 import os
 import torch
 
+from ipex_llm.transformers.qlora import get_peft_model, prepare_model_for_kbit_training
+from ipex_llm.transformers import AutoModelForCausalLM
 import transformers
 from transformers import AutoTokenizer, TrainingArguments, BitsAndBytesConfig
 from datasets import load_dataset
 from peft import LoraConfig
-from ipex_llm.transformers.qlora import get_peft_model, prepare_model_for_kbit_training
-from ipex_llm.transformers import AutoModelForCausalLM
 from trl import DPOTrainer
 import argparse