LLM inference in C/C++ with changes from Prism-ML to support 1Bit models for DGX Spark variants
Repositories
bmaltais repositories
Mixed-precision quantization for LLMs. Every layer refracts into a different format based on its sensitivity. Native compressed-tensors export, validated on Qwen3.6-35B-A3B MoE with MTP speculative decoding.
Tool to automatically generate Terraform files for your Azure subscription
quick cropping utility to grab 1:1 ratio sections from a folder of images
Qwen3-VL Image caption custom nodes
Interactive Terraform visualization. State and configuration explorer.
A GUI to help with manual tagging and cropping
opinionated bayesian optimisation for stable diffusion models block merge
WebUI extension for ControlNet
Merge models with separate rate for each 25 U-Net block (input, middle, output). Extension for Stable Diffusion UI by AUTOMATIC1111
CLI tool for managing AI agent skills across multiple agents
AI coding agent optimized for small LLMs. 87% benchmark with 4B-active model.