Views
No views yet
from transformers import AutoModelForCausalLM, AutoTokenizer
model_name = "jaeyong2/Pretrain-Recommandation-Preview"
model = AutoModelForCausalLM.from_pretrained(model_name)
tokenizer = AutoTokenizer.from_pretrained(model_name)
prompt = """
You are a recommendation system AI.
You input a list of items and a persona from the user.
From the list, you recommend the item most appropriate for the persona, along with a reason why.
(If no suitable item is found, you don't recommend it.)
""".strip()
content ="""
Item list: [Men's all-in-one skincare set, Mineral sunscreen with UV protection, Deep moisturizing body oil, Retinol-based wrinkle cream]
persona :A woman in her late 20s with sensitive skin who works in an office. She prefers natural ingredients and spends a lot of time outdoors.
""".strip()
system = {"role":"system", "content":prompt}
user = {"role":"user", "content":content}
messages = [system, user]
text = tokenizer.apply_chat_template(
messages,
tokenize=False,
add_generation_prompt=True,
enable_thinking=False # Switches between thinking and non-thinking modes. Default is True.
)
model_inputs = tokenizer([text], return_tensors="pt").to(model.device)
# conduct text completion
generated_ids = model.generate(
**model_inputs,
max_new_tokens=32768
)
output_ids = generated_ids[0][len(model_inputs.input_ids[0]):].tolist()
content = tokenizer.decode(output_ids, skip_special_tokens=True).strip("\n")
print(content)selected_item : Mineral sunscreen with UV protection
reason :Persona 2, a woman with sensitive skin and outdoor activities, would benefit from a mineral sunscreen with UV protection. It aligns with her preference for natural ingredients (mineral-based) and her need for protective UV rays, which are essential for outdoor workers. The product also matches Persona 1's likely focus on practical, non-toxic skincare products.