Terminal GPT (Open AI's API implemented into a Python program that can be accessed from terminal)
[TWO STEP Installation guide ] -- Terminal GPT [Version 1.2] -- February 15th, 2023
- [Step 1]: TO INSTALL, OPEN 'install_TmGPT.sh' with Terminal.
- [Step 2]: Open the TmGPT.py file and paste your own free, custom OpenAI API key where it says "pasteYourKeyHere" [DONE]: TmGPT is now installed!
Type 'TmGPT' in Terminal to use and run GTAI's TerminalGPT! Enjoy!
[Three OPTIONAL Tips]:
The three tips provided below are all [optional], but may further help you understand how to use the program and the OpenAI API.
-
[TIP #1]: You can change the API KEY in the .py file to your own custom API key from OpenAI. This will greatly increase the effectiveness and efficiency of the program. This is HIGHLY SUGGESTED (If you are referrencing the program from git, you must add you own key. If you would prefer not to, just download the TmGPT.zip from my site, gtaio.io
-
[TIP #2] : Changing the ENGINE ID...(default set to davinci003) Five options: [davinci002/03, babbage001,currie001, ada001, or cpu001] can and will greatly alter TWO things: 1.) how the AI perceives the prompt (Example, cpu is best for arithmetic and math, davinci and ada will be better at reasoning, creative responses, etc.) 2.) the type of responses provided from the OpenAI API
-
[TIP #3:] Lastly, with the DaVinci engine in particular, changing the 'max_tokens' value (on a scale of 1 to 3000) will also GREATLY alter the intelligence of the program and the length and complexity of the responses provided. The higher the value of the max tokens, the 'better' and 'richer' the depth of response will be. Alternatively, the lower the value, the 'less' dense the information and simpler the responses.
Any advice, recommendations, or if you are looking to help out/get involved, contact me at [email protected]!
This project is open-sourced, all I ask is if you like or enjoy the program, then please share it!
Terminal GPT [Version 1.2] -- Updated: February 15th, 2023 -- GTAI.io