logoalt Hacker News

syntaxprtoday at 4:50 AM0 repliesview on HN

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