Create app.py

app.py

@@ -0,0 +1,27 @@
+import gradio as gr
+import torch
+from transformers import AutoTokenizer, AutoModelForCausalLM
+from peft import PeftModel
+
+base_id = "Qwen/Qwen2.5-7B-Instruct"
+adapter_id = "epinfomax/BizFlow-Summarizer-Ko"
+
+# CPU 환경(무료)이면 float32, GPU 환경이면 float16/bfloat16 사용
+device = "cuda" if torch.cuda.is_available() else "cpu"
+dtype = torch.float16 if device == "cuda" else torch.float32
+
+tokenizer = AutoTokenizer.from_pretrained(base_id)
+model = AutoModelForCausalLM.from_pretrained(base_id, torch_dtype=dtype)
+model = PeftModel.from_pretrained(model, adapter_id)
+model.to(device)
+
+def summarize(text):
+    messages =
+    prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
+    inputs = tokenizer(prompt, return_tensors="pt").to(device)
+    
+    outputs = model.generate(**inputs, max_new_tokens=512, temperature=0.3)
+    return tokenizer.decode(outputs[inputs.input_ids.shape[1]:], skip_special_tokens=True)
+
+iface = gr.Interface(fn=summarize, inputs="text", outputs="text", title="BizFlow 요약기")
+iface.launch()