WaveletLM
wavelet-based, attention-free model with O(n log n) scaling. WaveletLM. WaveletLM is a wavelet-based, attention-free architecture that replaces self-attention with learned lifting wavelet decomposition, a Fast Walsh-Hadamard Transform, per-scale gated spectral mixing with SwiGLU activation, an inverse FWHT, and wavelet reconstruction.
오픈소스
반응 (1)
비슷한 프로덕트
설명이 비슷한 순
I built a Rust-powered Wavelet Matrix library for Python.
Wavelr
wavelr.comAI-native Decision Intelligence for renewable investments
Trim Transformer
github.comA transformer for physics models
This morning, we released a new Apache 2.0 licensed model on HuggingFace for detecting hallucinations in retrieval augmented generation (RAG) systems.
[ICML 2023] UPop: Unified and Progressive Pruning for Compressing Vision-Language Transformers