Update README.md

README.md

@@ -2,6 +2,55 @@
 license: apache-2.0
 ---
 
-Train Config
-base_model: allganize/Llama-3-Alpha-Ko-8B-Instruct 
-model_type: AutoModelForCausalLM tokenizer_type: AutoTokenizer
+- Train Config
+- base_model: allganize/Llama-3-Alpha-Ko-8B-Instruct
+- model_type: AutoModelForCausalLM tokenizer_type: AutoTokenizer
+
+## HOW TO USE
+```python
+from transformers import AutoTokenizer, AutoModelForCausalLM
+
+model_id = "MRAIRR/minillama3_8b_all"
+
+tokenizer = AutoTokenizer.from_pretrained(model_id)
+model = AutoModelForCausalLM.from_pretrained(
+    model_id,
+    torch_dtype="auto",
+    device_map="auto",
+)
+
+PROMPT_TEMPLATE = """
+# 지시:
+당신은 인공지능 어시스턴트입니다. 사용자가 묻는 말에 친절하고 정확하게 답변하세요.
+"""
+
+messages = [
+    {"role": "system", "content":PROMPT_TEMPLATE},
+    {"role": "user", "content": "안녕? 내 이름은 현수 ㅎㅎ 만나서 반가워"},
+]
+
+input_ids = tokenizer.apply_chat_template(
+    messages,
+    add_generation_prompt=True,
+    return_tensors="pt"
+).to(model.device)
+
+terminators = [
+    tokenizer.eos_token_id,
+    tokenizer.convert_tokens_to_ids("<|eot_id|>")
+]
+
+outputs = model.generate(
+    input_ids,
+    max_new_tokens=256,
+    temperature = 0.3,
+    eos_token_id=terminators,
+    do_sample=True,
+    repetition_penalty=1.05,
+)
+response = outputs[0][input_ids.shape[-1]:]
+response_text = tokenizer.decode(response, skip_special_tokens=True)
+completion = '\n'.join(response_text.split("."))
+print(completion)
+```
+