卡在这不往下走了--Register to controller，用的chatglm2-6b-int4

[kevinchi8781]

当前使用的分词器：ChineseRecursiveTextSplitter 当前启动的LLM模型：['chatglm2-6b-int4'] @ cuda {'device': 'cuda', 'host': '127.0.0.1', 'infer_turbo': False, 'model_path': 'F:\LANG_CHATCHAT\models\THUDM\chatglm2-6b-int4', 'port': 20002} 当前Embbedings模型： m3e-base @ cuda ==============================Langchain-Chatchat Configuration==============================

2023-11-05 11:36:55 | INFO | model_worker | Loading the model ['chatglm2-6b-int4'] on worker abe432a0 ... 2023-11-05 11:36:55 | INFO | model_worker | Register to controller 2023-11-05 11:44:25 | INFO | stdout | INFO: 127.0.0.1:3915 - "GET / HTTP/1.1" 404 Not Found 2023-11-05 11:44:26 | INFO | stdout | INFO: 127.0.0.1:3915 - "GET / HTTP/1.1" 404 Not Found 2023-11-05 11:44:27 | INFO | stdout | INFO: 127.0.0.1:3915 - "GET / HTTP/1.1" 404 Not Found

访问http://127.0.0.1:20000/ 就提示 { "detail": "Not Found" }

cmd显示这个 2023-11-05 11:44:25 | INFO | stdout | INFO: 127.0.0.1:3915 - "GET / HTTP/1.1" 404 Not Found 2023-11-05 11:44:26 | INFO | stdout | INFO: 127.0.0.1:3915 - "GET / HTTP/1.1" 404 Not Found 2023-11-05 11:44:27 | INFO | stdout | INFO: 127.0.0.1:3915 - "GET / HTTP/1.1" 404 Not Found

[zRzRzRzRzRzRzR]

int4模型使用的量化版本比较老会有兼容性问题，我们的框架没有适配int4

[PedroNeal]

所以应该换一个模型是吗，像chatglm2-6b的

[zRzRzRzRzRzRzR]

换成glm3-6b就行了

[PedroNeal]

换了，但是还是一样的问题