Having trouble to download models in the cookbook

[fullthunes-ship-it]

i am having trouble to download a model from the cookbook, It says :

Download stopped before HuggingFace reported completion.
Suggested action: hit Reconnect to re-attach to the tmux session. If that fails, retry — HuggingFace resumes incomplete files when possible.

I tried to reconnect, to retry download, to check my tmux verions and to re download it but it still fails, any solutions ?

Thank you

[Michiel-VandeVelde]

Thanks for the report.

This usually means the download process exited before the HuggingFace download was fully completed. If reconnect/retry keeps failing, it may not be a tmux issue itself, but something causing the download command inside the tmux session to stop early.

A few things worth checking:

1. Disk space

   Make sure the target drive/cache has enough free space. Model downloads can need more temporary space than the final file size.

   df -h

2. HuggingFace access / token

   If the model is gated or requires accepting terms on HuggingFace, make sure you are logged in and have access to the model.

   huggingface-cli whoami

3. Partial/corrupt cache

   If the same download keeps failing at the same point, the partial cache may be stuck. Try removing only the partial/incomplete files for that model, then retry.

4. Network / timeout

   Unstable connections can cause this, especially with large GGUF/model files. Retrying should normally resume, but if it repeatedly stops, it may be worth testing with a smaller model first to confirm the downloader works.

5. tmux session state

   Check whether the download session still exists:

   tmux ls

   If the session is gone, the download process likely crashed or exited, so the next useful thing is the actual log output from the failed download.

Could you share:

• which OS you are on
• which model from the cookbook you tried to download
• the last 30–50 lines of the download log / tmux output
• how much free disk space you have

With that, we can tell whether this is a model access issue, disk/cache issue, network issue, or a bug in the cookbook downloader.

[fullthunes-ship-it]

Hi! First of all thank you for your fast reply. I relaunched Odysseus with the cmd in the folder location while prompting the command powershell -ExecutionPolicy Bypass -File .\launch-windows.ps1 and it launched Odysseus

Now somehow i have downloaded the model but it crashes when i launch it after a couple of seconds.

I haven't tried your solutions yet and i will as soon as possible. but where can i enter those commands ?

I'm using windows 11, have 278GB of free space and i am not able to paste the 50 last lines here. I tried downloading Qwen2.5-Math-7B-Instruct-GGUF.

And an other quick question, after i successfully downloaded and installed odysseus, how can i relaunch it efficiently ? By prompting the command via the powershell for the Native Windos like i did ? or by using docker ?

Thanks again

Kind regards

[fullthunes-ship-it]

actually, when i try to download other model, it returns me the same error as before.
And when a download is finished, it shows 0 MB as the size of it which is odd.
The gemma model was downloaded with Ollama

[Seivad12]

I am having the exact same problem and am still unsure how to fix it.
I am on windows 11
Was trying to download this model: Llama-3.2-1B-Instruct-GGUF
Free disk space is 738 GB

Also to note I had to get the help of Chatgpt to download the chat model because I kept getting this GGUF error when trying download directly from cookbook. If you have any insight on that it would be appreciated.

[Amir0234-afk]

I have faced the same issue and couldn't download a model. this issue happened to me mainly because my internet connection is rather unstable.
The way that I fixed it was to manually download the model from HuggingFace and adding it to ollama and letting Odysseus detect it.
if nothing worked try doing this.

[fullthunes-ship-it]

How can i download models from there ?

[Ashish2271]

just setup ollama locally

[Amir0234-afk]

How can i download models from there ?

find the model on HuggingFace and it's details
for example the direct download link to Qwen3-8b is:
https://huggingface.co/bartowski/Qwen_Qwen3-8B-GGUF/resolve/main/Qwen_Qwen3-8B-Q4_K_M.gguf

just place this in your URL or use a download manager

[Ashish2271]

i am having same issue with cookbook tried alot of things cant make it work.
i am using local ollama to use models it works for me .