frob

https://github.com/ollama/ollama/blob/main/docs/modelfile.md Or pull from HF using the `Use this model` feature in the upper right of the model page. ``` ollama pull hf.co/unsloth/DeepSeek-R1-Distill-Qwen-14B-GGUF:Q4_K_M ```

The source of all templating for models is the Jinja template stored in the `chat_template` in the GGUF file of a model. If you can find the safetensors versions of...

There's nothing wrong with the create command. If you don't supply a template then ollama will use a default. In some cases, ollama will recognize the embedded `chat_template` and substitute...

Probably, but I haven't used one. Some of the constructs used in the Jinja templates (namespaces, external arguments) don't easily translate into the Go templating language used in ollama, so...

Correct. Other users have made available modified deepseek templates that support tools, see https://ollama.com/search?c=tools&q=deepseek .

[Server logs](https://github.com/ollama/ollama/blob/main/docs/troubleshooting.md#how-to-troubleshoot-issues) from the failures will help in debugging.

Turning wordwrap on and off should only affect the CLI, it's weird that it causes the server to crash. Can you add some server logs?

I wasn't able to duplicate this. Looking at the calls to llama.cpp, there's no difference depending on `set` operations in the CLI. However, from the logs, these failures are out...

@thany has hit the nail on the head, there's no way you are going to load a 231G model in to 32G RAM + 12 to 24G VRAM. You could...

falcon 180b is 101G, so yes, that will fit in 98G RAM with some spillover to swap. If you create another 150G swapfile, then you could run lama3.1:405b, but it...