docs: explain the format, both payloads, and the training run

README.md

@@ -1,3 +1,121 @@
 ---
 license: mit
+tags:
+  - inline-studio
+  - character
+  - lora
+  - flux2
+  - krea2
 ---
+
+# Unified Face Models
+
+One character file that works on more than one image model. `emmy-s5.char` carries FLUX.2
+reference images and a trained Krea 2 adapter in the same 107 MB container, so the same face
+renders on either model without keeping two sets of files in sync.
+
+Built and trained on the canvas in [Inline Studio](https://github.com/inlineresearch/Inline-Studio).
+
+## Why one file holds two things
+
+FLUX.2 klein and Krea 2 accept a character in completely different ways, so a single artifact
+cannot work for both.
+
+FLUX.2 klein has a reference channel. Reference images become tokens the model attends to at every
+denoising step, which is why a prompt can address them by position. Feeding it resized copies of
+the originals is enough, and no training is involved.
+
+Krea 2 has no reference channel. Its image input is ordinary img2img: the picture is encoded once
+into the starting latent and then denoised away, so there is nothing left for the model to hold on
+to. The only way to give Krea 2 a persistent identity is a trained LoRA.
+
+A `.char` stores identity once and compiles a payload per model family:
+
+```
+emmy-s5.char
+  manifest.json                        payload index, fingerprints, training record
+  refs/000..004.png                    the 5 reference images, immutable
+  derived/face_000..004.png            YuNet face crops at 512px
+  text/description.md                  the locked description, which is also the trigger
+  scoring/centroid_sface.json          128-d SFace identity centroid
+  scoring/centroid_dinov2-base.json    768-d DINOv2 subject centroid
+  scoring/embeds_*.json                per-reference embeddings
+  payloads/flux2-klein/ref_000..004.png    references resized onto FLUX.2's policy
+  payloads/krea2-lora/adapter.safetensors  the trained Krea 2 LoRA, 183 MB
+```
+
+The references and the description are the truth. Everything under `payloads/` and `scoring/` is
+cache and can be rebuilt from `refs/`. Adding a third model means adding a payload, not rebuilding
+the character.
+
+## What is in this one
+
+| | |
+| --- | --- |
+| References | 5 images, three at 2048px and two at 640px |
+| Trigger | `emmy4k woman with, natural unretouched skin, black hairs, fair skin` |
+| FLUX.2 payload | `flux2-klein`, 5 compiled references, max 1 MP, dimensions rounded to a multiple of 16 |
+| Krea 2 payload | `krea2-lora`, adapter fused at strength 1.0 |
+
+### Krea 2 training
+
+| | |
+| --- | --- |
+| Base | `krea2_raw_bf16.safetensors` (Krea 2 RAW) |
+| Steps | 500 |
+| Rank / alpha | 16 / 16 |
+| Learning rate | 1e-4, constant, Adam8bit |
+| Resolution | 512 |
+| Batch size | 1 |
+| Scope | full (attention + feed-forward) |
+| Caption dropout | 0.05 |
+
+Trained on RAW rather than Turbo on purpose. Turbo is step-distilled and trains badly; the adapter
+still loads on Turbo afterwards, which is faster and usually better. Every training image used the
+description above as its caption, so `emmy4k` is what the adapter binds to.
+
+The 500-step checkpoint beat the 800-step one on the same references. With five images that is 100
+passes each, and past roughly that point the adapter starts returning the training frames instead
+of the person in them.
+
+## Using it
+
+Drop the file in `models/characters/` and add a **Load Character** node, then wire it into the
+`character` input of a FLUX.2 or Krea 2 node.
+
+Inline Studio picks the payload for whichever model you wired it to. FLUX.2 gets the five
+references and a prompt line naming their positions. Krea 2 gets the adapter. Either way the
+description is prepended to your prompt, so type only what changes:
+
+```
+walking on a forest trail, half body shot
+```
+
+Not the description again. It is already there, and repeating it pushes the adapter harder toward
+the frames it trained on.
+
+For Krea 2, generate with **Krea 2 Turbo** at 8 steps and guidance 0. The adapter trained against
+RAW and loads onto Turbo as intended.
+
+## Building your own
+
+The full graph that produced this file, from five photos to a `.char` with both payloads, is
+published here:
+
+[FLUX.2 + Krea 2: multi-model portable consistent characters](https://inlinestudio.art/workflows/flux-2-krea-2-multi-model-portable-consistent-characters-training-only)
+
+Import it into Inline Studio, swap the reference images, and run the graph. It encodes the
+character, compiles the FLUX.2 references, builds a training dataset from the same references,
+trains the Krea 2 LoRA, attaches it, and writes the `.char`.
+
+## Links
+
+- [Inline Studio](https://github.com/inlineresearch/Inline-Studio): the node canvas and UI
+- [Inline Core](https://github.com/inlineresearch/Inline-Core): the generation and training engine
+- [inlinestudio.art](https://inlinestudio.art): downloads and guides
+- [Inline Registry](https://github.com/inlineresearch/Inline-Registry): the model and extension index
+
+## License
+
+MIT. The reference images are the character's own; check that you have the rights to any faces you
+train on before publishing a `.char`.