No models found after download it

[txitxo0]

Please do not report issues with the model generating incorrect output. This includes any instance where the model responds with Failed to run: ... or outputs badly formatted responses. If you are having trouble getting the correct output from the model, please open a Discussion thread instead. Hi there, thanks for the project I read in the Readme this is not the best way to add llama.ccp to Home assistant as an Add-on, could you share this better way?

Describe the bug
Anyway, I am trying to play a little with this local ai models but I can't. After download a model, and the label "Model successfully saved to models/." appear, the dropdown menu to select the model is still having the same option "none" even after reload it. What can I do? I have already tried setting manually the models path into the add-on configuration, but the behaviour is the same

Expected behavior
After click on reload button, the model already dowloaded should appear to be selected

Logs
No related logs found:

Running on local URL:  http://0.0.0.0:7860

Downloading the model to models

  0%|          | 0.00  /5.59G  ?iB/s 
  0%|          | 6.29M /5.59G  61.9MiB/s
  0%|          | 12.6M /5.59G  62.0MiB/s
  0%|          | 19.9M /5.59G  66.4MiB/s
  0%|          | 27.3M /5.59G  68.7MiB/s
  1%|          | 34.6M /5.59G  70.1MiB/s
  1%|          | 41.9M /5.59G  70.7MiB/s
  1%|          | 49.3M /5.59G  17.9MiB/s
  1%|          | 54.5M /5.59G  19.1MiB/s
  1%|          | 62.9M /5.59G  26.5MiB/s
  1%|▏         | 70.3M /5.59G  33.1MiB/s
  1%|▏         | 77.6M /5.59G  25.0MiB/s
  2%|▏         | 88.1M /5.59G  35.4MiB/s
  2%|▏         | 96.5M /5.59G  42.3MiB/s
...
 99%|█████████▊| 5.51G /5.59G  70.3MiB/s
 99%|█████████▊| 5.52G /5.59G  69.5MiB/s
 99%|█████████▉| 5.53G /5.59G  69.8MiB/s
 99%|█████████▉| 5.54G /5.59G  69.4MiB/s
 99%|█████████▉| 5.54G /5.59G  69.5MiB/s
 99%|█████████▉| 5.55G /5.59G  69.7MiB/s
 99%|█████████▉| 5.56G /5.59G  69.1MiB/s
 99%|█████████▉| 5.56G /5.59G  69.5MiB/s
100%|█████████▉| 5.57G /5.59G  69.0MiB/s
100%|█████████▉| 5.58G /5.59G  68.9MiB/s
100%|█████████▉| 5.59G /5.59G  69.6MiB/s
100%|██████████| 5.59G /5.59G  47.0MiB/s

[acon96]

Please try the updated version that was just pushed (2024.04.07). I'm pretty sure this was a bug in the upstream project.

[txitxo0]

Thank you @acon96 ! Now there is a bug during installation. Could be the app dir variable?

[acon96]

Looks like there was a slight tweak in the installation process. Can you try again? the 2024.04.09 version should build properly now

[txitxo0]

Thank you for the fast fix @acon96 now the installation works like a charm.

But the problem to storage & choose the model is still there. The behaviour is the same. The log in thw web ui sais "Model successfully saved to models/." and in the addon nothing related, just the dowload progress.

After that, clicking on the reload button the model does not appear in the dropdown menu.

Further more, there are three paths cointining the models for the addos but no one have anything in there

• addon_configs/f459db47_text-generation-webui/models
• config/addon_configs/f459db47_text-generation-webui/models
• homeassistant/addon_configs/f459db47_text-generation-webui/models

Running HAOS 12.1 with HA core 2024.4.2 and supervisor 2024.4.2

[acon96]

Just confirmed it is a bug upstream in text-generation-webui and submitted a PR to fix it.

[acon96]

The new version of text-generation-webui has this issue fixed and I have updated the addon to use that version.