support qwen template.

README.md

@@ -30,6 +30,9 @@ Supervised Finetuning, RLHF(Reward Modeling and Reinforcement Learning) and DPO(
 - DPO方法来自论文[Direct Preference Optimization:Your Language Model is Secretly a Reward Model](https://arxiv.org/pdf/2305.18290.pdf)
 
 ## 🔥 News
+
+[2024/04/17] v1.9版本：支持了 **[ORPO](https://arxiv.org/abs/2403.07691)**，详细用法请参照 `run_orpo.sh`。详见[Release-v1.9](https://github.com/shibing624/MedicalGPT/releases/tag/1.9.0)
+
 [2024/01/26] v1.8版本：支持微调Mixtral混合专家MoE模型 **[Mixtral 8x7B](https://huggingface.co/mistralai/Mixtral-8x7B-v0.1)**。详见[Release-v1.8](https://github.com/shibing624/MedicalGPT/releases/tag/1.8.0)
 
 [2024/01/14] v1.7版本：新增检索增强生成(RAG)的基于文件问答[ChatPDF](https://github.com/shibing624/ChatPDF)功能，代码`chatpdf.py`，可以基于微调后的LLM结合知识库文件问答提升行业问答准确率。详见[Release-v1.7](https://github.com/shibing624/MedicalGPT/releases/tag/1.7.0)

requirements.txt

@@ -6,5 +6,5 @@ tqdm
 tensorboard
 tqdm>=4.47.0
 peft~=0.10.0
-accelerate~=0.21.0
+accelerate~=0.27.2
 trl~=0.8.3

supervised_finetuning.py

@@ -699,6 +699,22 @@ def register_conv_template(template: Conversation):
     )
 )
 
+"""Qwen template
+source: https://huggingface.co/Qwen/CodeQwen1.5-7B-Chat/blob/main/tokenizer_config.json#L18
+Supports: https://huggingface.co/Qwen/CodeQwen1.5-7B-Chat
+"""
+register_conv_template(
+    Conversation(
+        name="qwen",
+        system_prompt="<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n",
+        messages=[],
+        roles=("user", "assistant"),
+        prompt="<|im_start|>user\n{query}<|im_end|>\n<|im_start|>assistant\n",
+        sep="\n",
+        stop_str="<|im_end|>",
+    )
+)
+
 
 def get_conv_template(name: str) -> Conversation:
     """Get a conversation template."""