Research

Single-page index that maps the nine pipeline stages to specific scripts, lists key results (model, optimal checkpoint per character, ArcFace deltas, optimal CFG/Steps/Shift), and shows the file tree so anyone joining the project can orient in one read. Anchor doc.

The longest document in the project. Synthesizes consensus from FPHam (500+ LoRAs trained), djdante's verified Klein 9B run, CivitAI threads, and Reddit/SD community findings. Distinguishes what to keep constant (identity) versus what to vary (angle, distance, expression, lighting, wardrobe), and explains why caption-only diversity cannot replace image-level diversity.

Built on djdante's published Klein 9B training config (29 chest-up images, simple captions, ~4000 steps). Documents the trigger word strategy, caption philosophy, hyperparameter choices, and the deliberate departures we made for our two-character project on RTX 5090.

Per-checkpoint ArcFace identity similarity, measured every 100 steps for both actors. Picks the optimum at 2200 steps for the female actor (AVG 0.732) and 1500 for the male, beating the HuggingFace text-only reference (0.5090 baseline). Documents the underfit / optimal / overfit transition shape.

Single-run write-up with timestamps, decisions, deviations from plan, and final selected checkpoint. Functions as both a record for future runs and a reproducibility note for anyone wanting to repeat the experiment.

Cross-tab of training configurations across runs. Surfaces the tradeoffs between learning rate, optimizer, EMA, batch size, gradient accumulation, and quantization choice. Used to justify the final picked config over alternatives.

Companion document distilling the dataset bible into a checklist for actor onboarding sessions: pose list, lighting requirements, wardrobe variety, do/don't examples. The on-the-ground operations layer above the bible.