Setting $PATH and Configuring Default AI Vendor/Patterns

[Mualjo]

Hello Everyone,

I am having a bit of a challenge when it comes to setting my PATH right for the command 'fabric' to work. I followed the instructions from the installation guide. After I reboot my terminal and computer I am having to do the exports for the environment variables all over again for the command. what do I have to do differently to make sure I don't have to keep repeating this process?

When I do have everything operational, I am looking to configure the ### Default AI Vendor and Model and Patterns. I am using Gemini as my AI Vendor, this is the list it gives me for the options i'm going to us for the tool:

Gemini

[1]	chat-bison-001
[2]	text-bison-001
[3]	embedding-gecko-001
[4]	gemini-1.0-pro-latest
[5]	gemini-1.0-pro
[6]	gemini-pro
[7]	gemini-1.0-pro-001
[8]	gemini-1.0-pro-vision-latest
[9]	gemini-pro-vision
[10]	gemini-1.5-pro-latest
[11]	gemini-1.5-pro-001
[12]	gemini-1.5-pro-002
[13]	gemini-1.5-pro
[14]	gemini-1.5-pro-exp-0801
[15]	gemini-1.5-pro-exp-0827
[16]	gemini-1.5-flash-latest
[17]	gemini-1.5-flash-001
[18]	gemini-1.5-flash-001-tuning
[19]	gemini-1.5-flash
[20]	gemini-1.5-flash-exp-0827
[21]	gemini-1.5-flash-002
[22]	gemini-1.5-flash-8b
[23]	gemini-1.5-flash-8b-001
[24]	gemini-1.5-flash-8b-latest
[25]	gemini-1.5-flash-8b-exp-0827
[26]	gemini-1.5-flash-8b-exp-0924
[27]	gemini-2.0-flash-exp
[28]	gemini-exp-1206
[29]	gemini-exp-1121
[30]	gemini-exp-1114
[31]	gemini-2.0-flash-thinking-exp
[32]	gemini-2.0-flash-thinking-exp-1219
[33]	learnlm-1.5-pro-experimental
[34]	embedding-001
[35]	text-embedding-004
[36]	aqa

but which option am I using for the Vendor and Model?

[NuCl34R]

Not sure what OS you are using but on a Linux system you have to add the environment variables in your ~/.bashrc file (home directory, make sure show hidden files is checked, or just 'nano .bashrc' when you open the terminal. Save the file, then type 'source .bashrc' to update with the changes, no need to reboot.

As for your 2nd question, you just enter the number next to the Model to use, since the newest version of Gemini is experimental (2.0 flash) then you can use that free of charge for some time, so use option 27, just type 27 when asked for model.