From 7ba79e4ee3f017de1e94ffc83ba17e1cae01f582 Mon Sep 17 00:00:00 2001 From: wailovet Date: Tue, 23 Apr 2024 02:33:59 +0800 Subject: [PATCH] 'update' --- mz_llama_cpp.py | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/mz_llama_cpp.py b/mz_llama_cpp.py index b94c878..0cb113a 100644 --- a/mz_llama_cpp.py +++ b/mz_llama_cpp.py @@ -157,7 +157,6 @@ def llama_cpp_messages(model_file, n_gpu_layers, chat_handler=None, messages=[], return "" result = choices[0].get("message", {}).get("content", "") - return result def get_schema_base_type(t): @@ -234,7 +233,9 @@ def llama_cpp_simple_interrogator_to_json(model_file, n_gpu_layers, use_system=T if options.get("max_tokens", None) is None: options["max_tokens"] = 2048 - return llama_cpp_messages(model_file, n_gpu_layers, None, messages, options=options) + result = llama_cpp_messages(model_file, n_gpu_layers, None, messages, options=options) + result = result.replace("\n", " ") + return result def llama_cpp_simple_interrogator(model_file, n_gpu_layers, use_system=True, system=None, question="", options={}): if options is None: @@ -273,7 +274,7 @@ def llama_cpp_simple_interrogator(model_file, n_gpu_layers, use_system=True, sys "content": question }, ] - return llama_cpp_messages(model_file, n_gpu_layers, None, messages, options={}) + return llama_cpp_messages(model_file, n_gpu_layers, None, messages, options=options) def llava_cpp_messages(model_file, n_gpu_layers, chat_handler, messages, options={}):