https://x.com/osanseviero/status/1907824380496769366?t=4s0nHgTVy-ulbT2BxTp_ww&s=35
X (formerly Twitter)
Omar Sanseviero (@osanseviero) on X
I'm very excited to announce the release of Quantization-Aware Trained (QAT) Gemma 3🚀
🤏3x less VRAM
🔥while retaining almost same quality
🤖use today with llama.cpp and Ollama
Models in https://t.co/SRWzcS8NwK
🤏3x less VRAM
🔥while retaining almost same quality
🤖use today with llama.cpp and Ollama
Models in https://t.co/SRWzcS8NwK
