Not training. Transposing rows/columns of matrices to group 128 parameters with similar (shared...

syntaxpr • yesterday at 4:50 AM • 1 reply • view on HN

Not training. Transposing rows/columns of matrices to group 128 parameters with similar (shared) scale factor. Qwen-3 model.

MarsIronPI • yesterday at 2:10 PM

I'm not sure what you mean. Could you please elaborate?

alt Hacker News