I am new and Confused about the correct step for the section of "getting started"

[RookyCoder]

I am thrilled to see this local AI coding assistant but I am unsure some of the installation step and Your help is appreicated.

1.)Clone a copy of "main" of "turbopilot" as item no. I shown in the above figure.

2.) I go for Option A: Download pre-quantized models and paste it to the models folder in main, i choose 6B ~4GB as item no.II and it looks like the same model in the example run command in step 4

3.) Download binary of TurboPilot Server and extract it to the root project folder.

4.) ./codegen-serve -t 6 -m ./models/codegen-6B-multi-ggml-4bit-quant.bin

I want to use it in VS Code with FauxPilot Plugin, also i am new to this too.
At the time of writing, it seems Venthe has accepted the PR so no need to install the "fauxpilot-1.1.5-ravenscroft.vsix" locally. Instead, i can download the plugin directly in vs code extension market.

5. After install "fauxpilot Plug-in", Ctrl+Shift+P - Preferences: Open User Settings (JSON)
   

================================= @ravenscroftj @
Q1: The suitable path of turbopilot-main ? Somewhere in Vscode folder such as installation path of fauxpilot?or Doesn't matter ?

Q2. Adminstrator mode of Cmd: execute step 4? Has to be the root project folder path?

Same go to Q3 where to call it?
C:\ A \ B \Downloads\TurbopilotALL\turbopilot-main\turbopilot-main

Q3. difficulties about the calling the API directly? I have no idea about these mysterious code or command. But i assume that it can be called in web browser URL and talks to the recipient in root project folder?

Running the requests in admin cmd http://localhost:18080/v1/engines/codegen/cohttp://localhost:18080/v1/engines/codegen/completions mpletions

Q4. Any confliction between other ai prompt offers similar feature or intelliSense that build in function in vs code?

Thanks

[ravenscroftj]

Hi @RookyCoder

I think the confusion is probably because I currently don't ship a windows version of turbocoder - I don't have a windows machine and haven't had time set up builds for windows - I might have a look over the weekend as it should be pretty straight forward given that llama.cpp does support windows.

When I add a codegen-serve.exe you'll just need to run that with the appropriate model name passed in.

You can call the API directly using any old HTTP client but I wouldn't recommend it as a beginner, you might as well stick to using Fauxpilot or similar.

There should not be any conflict between turbocoder and other features - Fauxpilot for VSCode uses the official VSCode extension points so it will complement other systems. I don't know if trying to use Fauxpilot and the official Copilot plugin at the same time would introduce some weirdness so I probably wouldn't recommend that.

I'll reply here again once I get windows builds working - although I don't have a windows machine so I'll have to rely on the community testing them for now.

[ravenscroftj]

I'm going to close this issue for now. Follow #24 for status updates on windows support.