500 error-ollama

#39
by XtaZa - opened

C:\Users\user>ollama run hf.co/HauhauCS/Qwen3.5-9B-Uncensored-HauhauCS-Aggressive:Q4_K_M
pulling manifest
pulling 2ca636d9e81d: 100% ▕██████████████████████████████████████████████████████████▏ 5.6 GB
pulling 05f662501f8b: 100% ▕██████████████████████████████████████████████████████████▏ 921 MB
pulling e8b41bd7e9bc: 100% ▕██████████████████████████████████████████████████████████▏ 481 B
verifying sha256 digest
writing manifest
success
Error: 500 Internal Server Error: unable to load model: D:\Ollama agent\blobs\sha256-2ca636d9e81d3d23ca9b60c234fe185d30ec082eeba69ce770fdb0c76559a4f5

Short version

If ollama run hf.co/... fails on Windows 11 with:

Error: 500 Internal Server Error: unable to load model

this worked for me:

  1. Download the main .gguf model file manually from Hugging Face in the browser.
  2. Create a Modelfile that points directly to that local .gguf.
  3. Import it with ollama create.
  4. Run the local model instead of hf.co/....

Commands:

cd D:\AI\ollama
notepad Modelfile
ollama create hauhau-qwen35 -f .\Modelfile
ollama list
ollama show hauhau-qwen35
ollama run hauhau-qwen35

Example Modelfile:

FROM D:\AI\ollama\Qwen3.5-9B-Uncensored-HauhauCS-Aggressive-BF16.gguf
PARAMETER temperature 1
PARAMETER top_k 20
PARAMETER top_p 0.95
PARAMETER presence_penalty 1.5

That solved the issue for me.


Full version

I had this problem on Windows 11 when trying to run the model directly from Hugging Face with:

ollama run hf.co/...

Ollama downloaded the files successfully, but then failed with:

Error: 500 Internal Server Error: unable to load model

What worked for me was not using hf.co/... directly.

Instead, I downloaded the actual .gguf model file manually from the Hugging Face page, saved it locally, and created a local Ollama model from that file.

Steps

  1. Download the main .gguf file manually from Hugging Face using your browser.

Example path:

D:\AI\ollama\Qwen3.5-9B-Uncensored-HauhauCS-Aggressive-BF16.gguf
  1. Create a Modelfile like this:
FROM D:\AI\ollama\Qwen3.5-9B-Uncensored-HauhauCS-Aggressive-BF16.gguf
PARAMETER temperature 1
PARAMETER top_k 20
PARAMETER top_p 0.95
PARAMETER presence_penalty 1.5
  1. Run:
cd D:\AI\ollama
ollama create hauhau-qwen35 -f .\Modelfile
ollama list
ollama show hauhau-qwen35
ollama run hauhau-qwen35

Result

After that, the model loaded correctly and also worked fine in the Ollama UI.

Note

For me, the direct hf.co/... route failed with the 500 error, but importing the local .gguf through Modelfile worked immediately.

If someone else hits the same issue on Windows, this workaround may help.

Sign up or log in to comment