Update app.py

app.py

@@ -57,11 +57,15 @@ def get_pipeline(model_name):
             tokenizer = AutoTokenizer.from_pretrained(local_path)
             model = AutoModelForCausalLM.from_pretrained(local_path)
 
+        # 修正 pad_token 缺失問題
+        if tokenizer.pad_token is None:
+            tokenizer.pad_token = tokenizer.eos_token
+
         generator = pipeline(
             "text-generation",
             model=model,
             tokenizer=tokenizer,
-            device= -1  # 強制 CPU
+            device=-1  # CPU 強制
         )
         _loaded_pipelines[model_name] = generator
     return _loaded_pipelines[model_name]
@@ -70,10 +74,11 @@ def call_local_inference(model_name, prompt, max_new_tokens=256):
     try:
         generator = get_pipeline(model_name)
         outputs = generator(
-            prompt, 
-            max_new_tokens=max_new_tokens, 
-            do_sample=True, 
-            temperature=0.7
+            prompt,
+            max_new_tokens=max_new_tokens,
+            do_sample=True,
+            temperature=0.7,
+            pad_token_id=generator.tokenizer.pad_token_id
         )
         return outputs[0]["generated_text"]
     except Exception as e:
@@ -115,7 +120,7 @@ def generate_article_progress(query, model_name, segments=5):
 # 5. Gradio 介面
 # -------------------------------
 with gr.Blocks() as demo:
-    gr.Markdown("# 佛教經論 RAG 系統 (CPU 免費版)")
+    gr.Markdown("# 📺 電視弘法視頻生成文章RAG系統(CPU免費版))")
     gr.Markdown("支援 DistilGPT2 / BTLM-3B / BART-Base，Auto 模式會自動選擇。")
 
     query_input = gr.Textbox(lines=2, placeholder="請輸入文章主題", label="文章主題")