Cannot choose a llama model while running web-ui

[udada63]

As you see here there should be models I can use if it is downloaded correctly. But I see nothing here.

I do clearly see the file and there was no issue during the installation... and llama model path is visible too. Is anybody suffering the same symptom?

I am using windows 10 22H2 on AMD Ryzen 7700 CPU and Radeon 7900XTX GPU based system. Using Python 3.10.6 and Nodejs 18.15. Using Chrome Browser.

[vChase42]

Also getting this issue. The alpaca model has no issue showing up correctly.

In the console it shows that the llama models do exist. Is there a way I can query the llama models without the web ui?

[udada63]

Also getting this issue. The alpaca model has no issue showing up correctly.

In the console it shows that the llama models do exist. Is there a way I can query the llama models without the web ui?

Well... I just downloaded Alpaca too and I do see it. Maybe it's llama only issue?

[mattcurr-sas]

when in doubt side step the issue, I'll take a look at the socket when I'm not exhausted and see what is actually happening but in the meantime just fill it in manually with this script in index.ejs....put that all the way at the end and either call it with an onload or just drop the function call at the bottom like this "addLlamaModles();"

function addLlamaModels() { const llamaModels = ['llama.30B', 'llama.7B', 'llama.14B'];

llamaModels.forEach((model) => { if (!config.models.includes(model)) { config.models.push(model); } }); }

[NotendoAtendo]

Same issue here, the console clearly states that llama 7B is installed, yet the dropdown in the web UI is empty. Have this issue both on laptop and tabletop pc's.

[irondonut]

In my case the filename of llama seems to have changed so in the file

the_app\node_modules\dalai\index.js

I changed on line 370:

let e = await exists(path.resolve(modelsPath, modelFolder, 'ggml-model-f16.bin')) //'ggml-model-q4_0.bin'))