Update README.md

README.md

@@ -18,8 +18,11 @@ base_model:
 <!-- Provide a quick summary of what the model is/does. -->
 
 ## Model Details
-This model is a Reasoner version of the phi-4 model by employing open-r1, which mimics the Distill methodology of Deepseek-R1.
-(Deepseek-R1のDistill手法を模倣した、open-r1を採用して、phi-4モデルを Reasonerにしたモデルです)
+This model is a Reasoner version of the phi-4 model by employing open-r1, which mimics the Distill methodology of Deepseek-R1. In particular, since it is specialized for Japanese, answers will be given in Japanese in principle.
+It can be made more flexible by interweaving English.
+
+Deepseek-R1のDistill手法を模倣した、open-r1を採用して、phi-4モデルを Reasonerにしたモデルです。特に日本語に特化させているため、原則日本語で回答します。
+英語を織り交ぜることでより柔軟にすることが可能です。
 
 ## Example Output
 ```
@@ -69,6 +72,48 @@ pip install --upgrade transformers accelerate datasets trl
 ### Predict(using AutoModelForCausalLM)
 ```python
 
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+
+
+model_name = "AXCXEPT/phi-4-open-R1-Distill-EZOv1"
+
+model = AutoModelForCausalLM.from_pretrained(
+    model_name,
+    torch_dtype="auto",
+    device_map="auto"
+)
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+
+prompt = f"""
+ある父と子の年齢に関する問題があります。条件は次の通りです：
+
+父の現在の年齢は息子の現在の年齢の3倍です。
+6年前、父の年齢は息子の年齢の5倍でした。
+父と息子の年齢を求めてください。
+"""
+messages = [
+    {"role": "system", "content": "You are an excellent AI."},
+    {"role": "user", "content": prompt}
+]
+text = tokenizer.apply_chat_template(
+    messages,
+    tokenize=False,
+    add_generation_prompt=True
+)
+model_inputs = tokenizer([text], return_tensors="pt").to(model.device)
+
+generated_ids = model.generate(
+    **model_inputs,
+    max_new_tokens=2048
+)
+generated_ids = [
+    output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
+]
+
+response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
+
+print(response)
 ```
 -------------------
 ### Predict(using vllm)