update

models/cpp_qwen2.py

@@ -47,8 +47,7 @@ class Qwen2Simulator(Simulator):
                 tokenizer=llama_cpp.llama_tokenizer.LlamaHFTokenizer(self.hf_tokenizer),
                 n_ctx=config.MAX_SEQUENCE_LENGTH,  #
                 # n_threads=None, # 默认会根据cpu数来设置 n_threads
-
-                use_mlock=True,
+                # use_mlock=True,
                 verbose=True,
             )
         else:
@@ -58,7 +57,7 @@ class Qwen2Simulator(Simulator):
                 tokenizer=llama_cpp.llama_tokenizer.LlamaHFTokenizer(self.hf_tokenizer),
                 filename="*fp16.gguf",
                 n_ctx=config.MAX_SEQUENCE_LENGTH,
-                use_mlock=True,
+                # use_mlock=True,
                 verbose=False,
             )
         logger.info(f"llm has been initialized: {self.llm}, "
@@ -150,6 +149,6 @@ if __name__ == "__main__":
 
     message = {"role": "user", "content": generated_text}
     print(message)
-    for generated_text, all_tokens in bot.generate_query(message, all_tokens, stream=True):
+    for generated_text, all_tokens in bot.generate_response(message, all_tokens, stream=True):
         print(generated_text, all_tokens)