Preparing for system design interviews?  Try bugzed.com →

dsalt

JSON →
library 0.4.34 ·python
verified Jun 29, 2026

dsalt (Dynamic Sparse Attention with Landmark Tokens) is a high-performance Triton-based implementation of sparse attention for transformers. Version 0.4.34 supports PyTorch and provides fused kernels for landmark token selection and sparse attention computation, targeting long-context LLM inference and training. Released monthly.

total hits 9
actors 2 distinct systems
last hit 16d ago AhrefsBot
GPTBot
3
Humans
5

top countries 🇺🇸 United States · 🇸🇬 Singapore · 🇨🇦 Canada · 🇮🇩 Indonesia