Flash-MSA: Accelerating Million-Token Training with Sparse Attention Kernelsnanduruganesh.github.io16 points·by rawsh·4 godziny temu·0 comments
Zyphra ZAYA1-base: First large-scale model trained on AMDzyphra.com6 points·by rawsh·8 miesięcy temu·1 comments
Debugging divergence between engine and transformers logprobs for RLgist.github.com2 points·by rawsh·10 miesięcy temu·0 comments