4 projects
sykra
Hybrid 3:1 Gated DeltaNet and Attention Transformer for PyTorch and TPU.
seren-llm
Seren: A Self-Refining Decoder-Only LLM with GQA, RoPE, Flash Attention, and a novel two-pass self-refinement mechanism.
sift-llm
SIFT - Selective Intelligence Filtering Transformer (U-Net backbone)
syko-llm
None