#LoRA

An introduction to Gradience, a tool that quantifies whether a LoRA rank setting is excessive using singular value decomposition. In experiments on Mistral-7B, halving the rank improved accuracy.

Configuration for running a Qwen-Image-Layered LoRA that automatically separates facial parts on RunPod. Comparison of RTX 6000 Ada (48GB) and RTX PRO 6000 (96GB).

A hands-on log of three iterations of LoRA training on SeaArt. Summarizes model selection, parameter tuning, and lessons learned from failures.

How to train LoRA in the cloud on SeaArt and download a safetensors file for local generation.

Step-by-step guide to building a LoRA training environment on Windows 11 with an RTX 3060 Laptop (6 GB VRAM) using kohya_ss — from caption writing to VRAM-saving settings.

Introducing a Mac mini M4 Pro to build an in-house RAG system. A plan for setting up a LoRA training environment during downtime while waiting for specs to be finalized.