sparse-models

Star

Here are 8 public repositories matching this topic...

Language: All

Filter by language

All 8 Python 6 Jupyter Notebook 1 Rust 1

PharmaCore

reacherwu / PharmaCore

Star 8

AI drug discovery on Apple Silicon. De novo generation + drug repurposing. 100% local, no cloud, fully auditable.

machine-learning bioinformatics ai cheminformatics computational-biology pytorch drug-discovery apple-silicon protein-language-model sparse-models

Updated Apr 20, 2026
Python

SVECTOR-CORPORATION / SSMA

Sponsor

Star 2

Structured State Matrix Architecture (SSMA) is a high-performance framework designed for efficient sequence modeling, combining structured state space models with adaptive attention mechanisms.

machine-learning ai deep-learning architecture memory-efficient attention-mechanism self-attention ssma sparse-models

Updated Mar 17, 2025
Python

PERSPECTIVE v2 — A 1.05 trillion parameter sparse Mixture-of-Experts language model that runs on consumer hardware (4 GB VRAM + 32 GB RAM). Features O(1) perspective decay recurrence, 3D torus manifold routing, native ternary {-1,0,+1} weights, holographic distributed memory, and hard geometric safety constraints. Built in Rust.

nlp rust machine-learning research deep-learning neural-network transformer moe language-model ai-safety efficient-inference consumer-hardware mixture-of-experts on-device-ai ternary-weights llm sparse-models holographic-memory manifold-routing novel-architecture

Updated Feb 20, 2026
Rust

Cmouzouni / three-phases-moe

Star 1

Code and data for: Three Phases of Expert Routing — How Load Balance Evolves During MoE Training

deep-learning transformer moe research-paper load-balancing mixture-of-experts training-dynamics phase-transitions mean-field-games llm-architecture sparse-models expert-routing

Updated Apr 5, 2026
Python

originaonxi / moe-efficiency-study

Star 1

Research — Mixture-of-Experts efficiency analysis. Benchmarking sparse activation vs dense models on cost-per-token and quality.

benchmarking efficiency ai-research mixture-of-experts llm sparse-models

Updated Mar 17, 2026
Python

btboilerplate / Lasso_Regression

Star 1

Developed a Lasso Regression model applying L1 regularization for housing price prediction. The model achieved an R2 score of ~0.64 while performing implicit feature selection and maintaining strong predictive performance.

machine-learning feature-selection supervised-learning lasso-regression l1-regularization sparse-models