search

lm-sys / FastChat

An open platform for training, serving, and evaluating large language models. Release repo for Vicuna and Chatbot Arena.
Apache License 2.0
37.09k stars 4.57k forks source link

chatglm2-6b coming #1795

Closed Halfknow closed 1 year ago

Halfknow commented 1 year ago

chatglm2-6b has released with great updates bilingually, hope you guys can add this open llm in your website which will be really exciting. The github repo is here: https://github.com/THUDM/ChatGLM2-6B.

sdhjl2000 commented 1 year ago

start fastchat use new chatglm2-6b model,it wroks python3 -m fastchat.serve.model_worker --model-names "gpt-3.5-turbo,text-davinci-003,text-embedding-ada-002" --device mps --model-path /Users/XXX/Projects/6b/chatglm2-6b

but get error when token max than 2048,"openai.error.APIError: Invalid response object from API: '{"object":"error","message":"This model\'s maximum context length is 2048 tokens. However, you requested 2895 tokens (2383 in the messages, 512 in the completion). Please reduce the length of the messages or completion.","code":40303}' (HTTP response code was 400)"" accord to chatglm2 document,chatglm2-6b support more than 2048 token image

ye7love7 commented 1 year ago

can it used as embedding model?wo need embedding when use vector store