A Step-by-step deployment guide with ollama

#16
by snowkylin - opened

Just wanna share my deployment process in case of need.

https://snowkylin.github.io/blogs/a-note-on-deepseek-r1.html

Unsloth AI org

how did u manage to run the model directly using the ollama run command? :)

did u merge the ggufs yourself?

how did u manage to run the model directly using the ollama run command? :)

did u merge the ggufs yourself?

Yes, I merged them by llama-gguf-splitin llama.cpp. You can find the detail here.

Yes, it's really sad that Ollama is the last one today which don't support model in parts - everyone should be warned, and main stop for me.
I tested even in Kobold-cpp app and it can support gguf in parts, not counting LM Studio or oobabooga.
Also kinda ComfyUI people complain that there's very low context window in Ollama API, like only 2048 and require hacks, i haven't tested but beware too.

Sign up or log in to comment