attention-kernel

Here are 2 public repositories matching this topic...

sachncs / laker

PyTorch kernel-regression library for spectrum cartography.

pytorch conjugate-gradient cccp kernel-regression preconditioning scientific-machine-learning attention-kernel spectrum-cartography convex-concave-procedure

Updated Sep 12, 2026
Python

ChaoyuWang04 / Halftone_QuantizedKVCache

Star

Data-driven INT8 KV-cache quantization: SQNR-guided granularity, near-lossless PPL (+0.2%), 0.50× memory, fused Triton int8 decode attention at 1.56× — gap to 2× fully explained.

cuda transformers triton quantization memory-bandwidth roofline kv-cache llm-inference qwen kernel-optimization attention-kernel in8

Updated Jul 18, 2026
Python

Add this topic to your repo

To associate your repository with the attention-kernel topic, visit your repo's landing page and select "manage topics."

Learn more