refactor(generator): 优化生成逻辑

khaosz/inference/generator.py

@@ -118,7 +118,6 @@ class ChatGenerator(GeneratorCore):
         
         ids = self.tokenizer.encode(build_prompt(query, history))
         input_ids = torch.tensor([ids], device=device, dtype=torch.long)
-        cpy_history = history.copy()
         
         start_cache_pos = len(ids)
         cur_cache_pos = 0
@@ -132,9 +131,8 @@ class ChatGenerator(GeneratorCore):
         )
         
         response = self.tokenizer.decode(ids[start_cache_pos:])
-        cpy_history.append((query, response))
         
-        return response, cpy_history
+        return response
     
 
 class StreamGenerator(GeneratorCore):
@@ -278,7 +276,7 @@ class RetrievalGenerator(GeneratorCore):
             history = []
             
         retrieved = "\n".join([f"{idx + 1}. {key}" for idx, key in enumerate(retrieved)]) if retrieved else ""
-        retrieved_query = f"{retrieved}<eos>\n\n根据以上内容回答: {query}" if retrieved else query
+        retrieved_query = f"{retrieved}\n\n{query}" if retrieved else query
         parameter = ModelParameter(self.model, self.tokenizer, self.config)
         
         return ChatGenerator(parameter).generate(