juraj on Nostr: Tried distilled llama-70b deepseek r1 (not the big one yet). I'm quite disappointed. ...
Tried distilled llama-70b deepseek r1 (not the big one yet).
I'm quite disappointed. The thinking process is there, it is pretty good, but then the model does not follow its own advice and forgets things.
Unfortunately the big model does not run on my machine.
I'm quite disappointed. The thinking process is there, it is pretty good, but then the model does not follow its own advice and forgets things.
Unfortunately the big model does not run on my machine.