ngxson / alpaca.cpp-webui Goto Github PK

[1] llama_model_load: loading model from 'F:\GPTProgram\alpaca.cpp-webui\bin\ggml-vicuna-7b-1.1-q4_0.bin' - please wait ...
[1] llama_model_load: invalid model file 'F:\GPTProgram\alpaca.cpp-webui\bin\ggml-vicuna-7b-1.1-q4_0.bin' (bad magic)
[1] main: failed to load model from 'F:\GPTProgram\alpaca.cpp-webui\bin\ggml-vicuna-7b-1.1-q4_0.bin'

bin and model has in right place and config file edited, webui run but can not response to any prompt.

Hello,
my computer is so slow that it is not possible to access the web UI while alpaca.cpp is running. So I tried to access the interface from another device. Is it possible to connect to the interface from another device in the same network? I have already tried it out with different combinations of the IP address and the Port(localhost:13000, 192.168.178.41:13000, 192.168.178.41:13001).

Thanks and Regards,
Paul

PS: Sorry for the bad English.

Why not add the ability to switch models, for example; change to gpt4all without having to edit the config.

This looks so good. I wish I could get it to work.
I'm running Windows 10 with docker and app store WSL.
I get this error through the web browser after loading through command line:

<Cannot find ggml model file or executable file under {YOUR_PROJECT_DIRECTORY}/bin

Please follow this guide to download the model:
https://github.com/ngxson/alpaca.cpp-webui#how-to-use

Then, make sure that these paths exist:
/app/bin/chat
/app/bin/ggml-alpaca-7b-q4.bin>

I created a folder named app with bin in it (to match the path above) and it still didn't work. I have ggml-alpaca-7b-q4.bin and chat in both a folder /bin and /app/bin just to cover my bases... but no luck.

I am trying to get this working on Windows 11 with GPT4All. There was an issue with setting the Port but I think I have resolved that by changing the package.json from
"proc_serve": "PORT=13000 next start",
to
"proc_serve": "set PORT=13000 react-scripts start",

however I now get the following exit code, I'm not a coder and can't figure out where the error is.

[0]
[0] > [email protected] proc_serve
[0] > set PORT=13000 react-scripts start
[0]
[1]
[1] > [email protected] proc_native
[1] > node ./utils/native.js
[1]
[0] npm run proc_serve exited with code 0
--> Sending SIGTERM to other processes..
[1] npm run proc_native exited with code 1