Article
I put a 24 dimensional lattice in a CUDA kernel to run Qwen3-4B in 2.6 GB
Pier-Jean
• Malandrino PRO
Pier-Jean
AI & ML interests
Document parsing
Quantization
Recent Activity
liked a model about 11 hours ago
convaiinnovations/laya liked a Space 4 days ago
convaiinnovations/laya-demo authored a paper 22 days ago
Unfolding the Leech Lattice: Fused Multi-Shell Decoding and VRAM Layouts for 2-Bit LLM WeightsOrganizations
published an article 27 days ago
published an article 6 months ago
Article
Docling Studio — Open-Source Visual Inspection for Docling Pipelines
Pier-Jean
• published an article over 1 year ago
Article
The Delegated Chain of Thought Architecture
Pier-Jean
•