Solution for choosing the GPT model to be used

lifan0127 / ai-research-assistant

Aria is Your AI Research Assistant Powered by GPT Large Language Models

GNU Affero General Public License v3.0

794 stars 33 forks source link

Solution for choosing the GPT model to be used #4

Open Edilson-R opened 1 year ago

Edilson-R commented 1 year ago

Hello @lifan0127. What a great evolution of the project that I'm following and you always diligently helping, again my thanks. It would be possible to put a field to choose the model that the plugin will access. This would be interesting for token economy and other jobs where it is not necessary to use GPT 3.5, or even update the model to choose GPT 4, for example.

Tenfleques commented 1 year ago

I vouched for this in the discussion. This will be great if one can also hookup a local LLM

lparolari commented 11 months ago

I'm not able to use the gpt-4-0613 model required by Aria (see the attachment). It seems a common issue although on July 6 OpenAI released GPT-4 for all users.

it would be nice to add an option that allows you to choose the model to use.

Thank you!

Screenshot_20230725_183530

lifan0127 commented 11 months ago

Hi @lparolari , right now this plugin heavily uses the new "function calling" feature in GPT-4-0613 (technically this feature is also available in GPT-3.5-Turbo-0613 but is far less reliable). I hope your access issue will be resolved soon by OpenAI.

In the future, I am definitely interested in exploring open-source LLM options such as Llama2.

11930958 commented 1 month ago

Facing following error despite I have GPT-4 and generated API key two times.