Install ollama.
-
wrote last edited by [email protected]
Install ollama. Install Deepseek-R1 1.5B with `ollama run deepseek-r1:1.5b` and now see what all the hype is about. It's just the smallest version but it runs quite fast on my laptop, so it is fun to play with.
-
replied to Jan Wildeboer π·:krulorange: last edited by
@jwildeboer it's fun!
-
replied to Jan Wildeboer π·:krulorange: last edited by
@jwildeboer it has become obligatory to ask it about Tiananmen square and then draw biased conclusions about the influence of state-obliged censorship for the whole dang thing.
That has become the lackmus-test of freedom, apparently.
-
replied to phoenixπ§ππππ last edited by
@phoenix You will soon see free models on huggingface that have been trained the R1 way but without such restrictions, so it's just a matter of time before this problem is gone.
-
replied to Jan Wildeboer π·:krulorange: last edited by
Or, even better, use ramalama to download, run or serve models in a container.
A project by Red Hat's engineers using containers.
-
replied to Pavol BabinΔΓ‘k last edited by
@scrool That will definitely be used in my homelab!
-
J [email protected] shared this topic
-
replied to Pavol BabinΔΓ‘k last edited by
@scrool @jwildeboer Sorry but every time I read ramalama i hear "ding dong" in my head .
P.S.: For the young people in the public, it's an old catchy song, one of those that gets to you and becomes an earworm. -
replied to ghostdancer last edited by
@ghostdancer @scrool It is just a rumour that #ramalama has a config option that plays a ding dong sound every time it starts a container
