conftrace_

adaptive ramping optimizer (q-ramping)

1 paper

Explore in graph

Co-occurring keywords · papers mentioning both

ema quantizer (q-ema) (1) fp4 precision (1) microscaling (mx) data format (1) weight oscillation (1) vision transformer (1)

Papers

Oscillation-Reduced MXFP4 Training for Vision Transformers ICML 2025