From 545104ba705b6d7ea5bf1c9bdb955a3508141338 Mon Sep 17 00:00:00 2001
From: ViperEkura <3081035982@qq.com>
Date: Wed, 29 Jul 2026 21:23:20 +0800
Subject: [PATCH] feat: add Alpaca-GPT4-ZH SFT export script

---
 .../sft_alpaca_gpt4_zh.py                     | 25 +++++++++++++++++++
 1 file changed, 25 insertions(+)
 create mode 100644 scripts/supervised_finetuning/sft_alpaca_gpt4_zh.py

diff --git a/scripts/supervised_finetuning/sft_alpaca_gpt4_zh.py b/scripts/supervised_finetuning/sft_alpaca_gpt4_zh.py
new file mode 100644
index 0000000..a04205b
--- /dev/null
+++ b/scripts/supervised_finetuning/sft_alpaca_gpt4_zh.py
@@ -0,0 +1,25 @@
+from datasets import load_dataset
+from pipeline import export_dataset
+
+
+def process_func(input_dict: dict):
+    instruction = input_dict["instruction"]
+    inp = input_dict.get("input", "")
+    if inp:
+        content = instruction + "\n" + inp
+    else:
+        content = instruction
+    return {"messages": [
+        {"role": "user", "content": content},
+        {"role": "assistant", "content": input_dict["output"]},
+    ]}
+
+
+if __name__ == "__main__":
+    dataset = load_dataset("llm-wizard/alpaca-gpt4-data-zh")
+    export_dataset(
+        dataset=dataset["train"],
+        output_dir="./dataset",
+        output_prefix="alpaca-gpt4-data-zh",
+        process_func=process_func,
+    )