News for hackers. Stuff that matters.

A comment on Gemma 4 12B: A unified, encoder-free multimodal model. Parent comment

Re:Gemma 4 12B: A unified, encoder-free multimodal model by pseudollm on Wednesday June 03, 2026 @08:01PM (#48391855)
No there isn't - read the paper. It's just 40msec raw audio samples. Multiplied by one matrix to translate to 3800 input vector. That's it. The next 40 msec are fed in the next transformer input step. Without any positional encoding. Repeat ad infinitum
[ Parent | on HN ]

There are no comments on this comment yet.