X · @teortaxesTex
· X / Twitter
Good guy DeepSeek gives us accelerated models The most interesting one here is Gemma4-12B, I presume vision included. Might be the best local model in…
Good guy DeepSeek gives us accelerated modelsThe most interesting one here is Gemma4-12B, I presume vision included. Might be the best local model in its weight class now, by some marginQwen 3.5 not included because DS[park] doesn't do linear attention I guessFlorian Brand: