Osye's picture

5

Osye

Osye

·

AI & ML interests

None yet

Recent Activity

updated a dataset about 22 hours ago

published a dataset about 22 hours ago

reacted to s3nh's post with ❤️ 12 days ago

Existing methods — GPTQ, AWQ, llama.cpp's k-quants — minimize empirical loss heuristically. None of them prove they are optimal in any information-theoretic sense. ICRB-Q builds a quantization scheme that is provably optimal via the Cramér-Rao lower bound (CRB): no unbiased estimator of a weight can have lower variance than [F(θ)]⁻¹, where F is the Fisher information matrix.

View all activity

Organizations

Osye 's Spaces 1

Model Inspector