Claycable Blog
Technical writing on LoRA fine-tuning, training data, and building domain-specific AI models.
A deep dive into Low-Rank Adaptation — how it works mathematically, what rank and alpha actually mean, when to use QLoRA vs LoRA, and how to get real results from a small dataset.
Why 300 great pairs beat 3,000 mediocre ones. A practical guide to writing training pairs that produce consistent, domain-specific behavior — not a confused base model with new clothes.
From DreamBooth to Flux LoRA — how image fine-tuning actually works, what your dataset needs to look like, trigger words, choosing between Replicate and FAL.ai, and what makes a good image adapter.