EntropyDrop/SKING_DDJ_Dataset

SKING_DDJ Public Generation & Preference Dataset

This repository contains data from public skin-generation tasks performed on EntropyDrop (entropydrop_website) with models in the SKING_DDJ series (e.g., SKING_DDJ_v54, SKING_DDJ_v61, SKING_DDJ_v61b, SKING_DDJ_v66).

New imports are scoped to public generations from this model family with an explicit public_license of cc-by-nc-4.0 and is_deleted = false; private generations, unknown public licenses, platform manual in-browser edits, manual user uploads, and unrelated model families are excluded. Eligible records from legacy model directories may also be present. Annotation-time repairs created by this repository’s editor are stored separately as {id}_perfect.png, preserving the imported candidate images.

In addition to source inputs, intermediate edited images, and final 64×64 RGBA skin textures, this dataset includes multi-version comparative candidate results (e.g. v101, v103, v104, v105) and human preference annotations (_best.txt, _perfect.txt, _all_rejected.txt). Subject to the applicable asset licenses and non-commercial restriction, the dataset supports:

• Supervised Fine-Tuning (SFT) for 2D-to-UV and Text-to-Skin models;
• Reward Modeling (RM) and Direct Preference Optimization (DPO / RLHF);
• Error diagnosis, UV alignment verification, and generation pipeline benchmarking.

---

📂 Repository Structure

SKING_DDJ_Dataset/
├── README.md                           # Dataset documentation and specifications
├── LICENSE                             # Dataset license scope and CC BY-NC 4.0 legal text
├── dataset_layout.py                   # Shared skin naming and three-character prefix layout
├── find_results_without_source.py      # Scan & detect orphan result images lacking source images
├── results_without_source.txt          # Scan report generated by find_results_without_source.py
├── best_annotator/                     # Lightweight web annotation & preference ranking tool
│   ├── README.md                       # Annotator user manual & keyboard shortcuts
│   ├── app.py                          # Python 3.10+ standard-library HTTP server
│   ├── index.html                      # Vite source entry
│   ├── package.json / package-lock.json # Frontend dependencies and scripts
│   ├── vite.config.js                  # Build output and development API proxy
│   ├── src/                            # React UI and ported skin editor
│   ├── public/vendor/                  # Bundled skinview3d source assets
│   ├── static/                         # Production build served by Python (assets/ and vendor/)
│   ├── test_perfect.py                 # Perfect labels, repaired PNGs and rollback tests
│   ├── test_app.py                     # Unittest suite for annotator backend
│   └── test_ui.cjs                     # E2E UI flow test suite (via JSDOM)
└── entropydrop_website_generations/
    ├── SKING_DDJ_v54/                  # Every model uses the same three-character ID prefixes
    │   ├── 14C/                       # Complete skin groups whose IDs start with 14C
    │   ├── 2L4/                       # Complete skin groups whose IDs start with 2L4
    │   └── .../
    ├── SKING_DDJ_v61/                  # Generations & evaluations under model version v61
    ├── SKING_DDJ_v61b/                 # Generations & evaluations under model version v61b
    └── SKING_DDJ_v66/                  # Generations & evaluations under model version v66

Sample File Hierarchy

Each generation task is identified by a unique, stable <generation_id>. Within each <model_version> directory, files are organized as follows:

All model directories use <model_version>/<generation_id[:3]>/. The prefix preserves the first three characters exactly, including digits and letter case; IDs shorter than three characters use their entire ID. For example, 2L4DPU434KNC82YY under v54 is stored in SKING_DDJ_v54/2L4/. This rule also applies to legacy model directories and future models.

Read model directories recursively. All images, prompts, author metadata, annotations, repairs, and audit files for one skin remain together. Label files still contain candidate filenames without directory prefixes. The annotator aggregates filters and progress by model across all prefix directories, and also accepts a single model or prefix directory as its scan root. Readers remain compatible with older flat and one-character layouts.

entropydrop_website_generations/<model_version>/<generation_id[:3]>/
├── <generation_id>_source.png              # Original user input image
├── <generation_id>_edited.png              # Intermediate edited figure image
├── <generation_id>_result.png              # Primary / baseline generated 64x64 skin
├── <generation_id>_result_<version>.png    # Multi-checkpoint comparative result (e.g. _result_v104.png)
├── <generation_id>_prompt.txt              # User prompt text (omitted when empty)
├── <generation_id>_author.json             # Generating account: userId and nickName
├── <generation_id>_best.txt                # Human preference: selected best candidate filename
├── <generation_id>_perfect.txt             # Perfect label: original candidate filename, same as best
├── <generation_id>_perfect.png             # Optional annotation-time repaired 64x64 skin
├── <generation_id>_all_rejected.txt        # Human preference: rejection label and rationale
├── <generation_id>_result_<ver>.uv_review.json # UV texel alignment & quality audit report
├── <generation_id>_source_err.png          # Input image associated with a failed generation
├── <generation_id>_edited_err.png          # Intermediate image associated with a failed generation
└── <generation_id>_result_err.png          # Result image associated with a failed generation

Image files use .png, .jpg, .jpeg, or .webp matching their original stored format (skin results are always RGBA PNGs).

---

🏷️ File Suffixes & Schema

Suffix / Pattern	Type	Description	Example / Content
_source	Image	Original input image supplied to the generation task.	2L4DPU434KNC82YY_source.png
_edited	Image	Intermediate figure image cropped, segmented, or preprocessed between pipeline stages.	2L4DPU434KNC82YY_edited.png
_result	Image	Primary / baseline generated 64×64 Minecraft skin UV texture.	2L4DPU434KNC82YY_result.png
_result_<version>	Image	Candidate skin texture generated by an alternative model checkpoint or UV parser version (e.g. v101, v103, v104, v105).	2L4DPU434KNC82YY_result_v104.png
_prompt.txt	Text	Text prompt associated with the generation task (UTF-8, omitted if empty).	"cyberpunk knight with glowing visor"
_author.json	JSON	Generating account's ID and current public display name. Unknown fields are null.	{"userId": "USER_ID", "nickName": "Display name"}
_best.txt	Label	Filename of the human-annotated best skin result among available candidates.	2L4DPU434KNC82YY_result_v104.png\n
_perfect.txt	Label	Perfect-quality annotation; records the original candidate filename, identical to best.	2L4DPU434KNC82YY_result_v104.png\n
_perfect.png	Image	Optional repaired 64×64 PNG saved by the annotation editor; does not replace the original result.	Present only after repair submission.
_all_rejected.txt	Label	Rejection annotation when no candidate is satisfactory. Contains the primary failure reason.	result_unsatisfied or edited_unsatisfied
_result_<ver>.uv_review.json	JSON	Diagnostic audit report covering UV symmetry, hair/beard masks, and texel mismatch counts.	JSON metadata
_source_err	Image	Input image from a task that failed validation checks (e.g., non-compliant viewing angle).	NORESULT_source_err.jpg
_edited_err	Image	Intermediate image from a failed validation task.	ABC_edited_err.png
_result_err	Image	Result image from a failed validation task.	ABC_result_err.png

Not every entry contains every optional file. For example, text-to-skin pipelines may lack _source images, intermediate stages may be bypassed, or prompts may have been left blank. A missing optional file does not indicate corruption.

Each local skin group has one UTF-8 <generation_id>_author.json beside its images. userId comes from the generation's user_id; nickName comes from the matching account's username. The author identifies the platform account associated with the generation. Missing IDs or unavailable display names are stored as JSON null, without substituting an invented author.

---

🎯 Human Preference & Rejection Annotations

The dataset incorporates fine-grained human feedback collected via best_annotator:

1. Best Choice (_best.txt)

When annotators identify a clearly superior skin result among candidates (e.g. comparing baseline _result.png against _result_v101.png, _result_v104.png, etc.), {id}_best.txt is written. It contains solely the filename of the selected result:

2L4DPU434KNC82YY_result_v104.png

2. All Rejected (_all_rejected.txt)

When all candidate skin results fail quality expectations, annotators mark the entry as rejected, categorizing the root cause into {id}_all_rejected.txt:

Rejection Label	Meaning
result_unsatisfied	The input and intermediate edited images are acceptable, but all generated skin UV maps exhibit severe flaws, poor color mapping, artifacts, or broken layouts.
edited_unsatisfied	The failure originates earlier in the pipeline: the intermediate image (_edited.png) is poorly segmented, truncated, distorted, or hallucinated, preventing valid skin synthesis downstream.

3. Perfect Labels and Repaired Skins

“Select as best (perfect)” writes both {id}_best.txt and {id}_perfect.txt, each containing the original selected candidate filename followed by a newline. It does not create a new image.

“Repair and select as best (perfect)” opens the React skin editor with the intermediate edited image as a reference. Submission additionally writes {id}_perfect.png; the TXT files still identify the original candidate, not the repaired PNG. Original candidates remain unchanged. Opening the same perfect candidate again resumes its saved repair.

Saved state	_best.txt	_perfect.txt	_perfect.png	_all_rejected.txt
Unannotated / cleared	Absent	Absent	Absent	Absent
Best	Candidate filename	Absent	Absent	Absent
Perfect, without repair	Candidate filename	Same as best	Absent	Absent
Perfect, with repair	Candidate filename	Same as best	Repaired skin	Absent
Rejected	Absent	Absent	Absent	Rejection reason

Files are grouped by directory and generation ID. For a valid perfect annotation matching best with no rejection conflict, use the repaired PNG when present; otherwise use the original candidate named in best. A standalone perfect PNG is not sufficient evidence of a valid annotation. _edited remains the pipeline intermediate image and must not be confused with _perfect.png.

See the annotator file structure and state table for the source/build directory layout and scanner rules.

4. Mutual Exclusivity & Integrity

• _best.txt and _all_rejected.txt are strictly mutually exclusive: selecting a best candidate removes any prior _all_rejected.txt, and rejecting all candidates removes any prior _best.txt.
• Choosing ordinary best, rejecting, or clearing removes the previous perfect TXT and PNG. Choosing perfect without repair removes any previous repaired PNG. Repeated repair submission replaces the current repaired PNG.
• Individual files use temporary files and atomic replacement; revision checks cover best, rejected, perfect, and repaired-image contents. I/O failures trigger an attempt to restore prior files. This is not a cross-process or power-loss transaction; use one server instance per dataset directory.
• Skipping changes no files, including existing annotations.

---

🛠️ Tooling & Utilities

1. DDJ Best-Result Annotator (best_annotator/)

A self-contained web application for rapid side-by-side preference annotation.

• Features:
  • Zero external dependencies: runs on Python 3.10+ standard library.
  • Bundled 3D Minecraft skin renderer using skinview3d 3.4.2 (no CDN or Node required at runtime).
  • Side-by-side synchronized 3D preview and 2D UV layout for multiple candidate versions.
  • Keyboard shortcuts: 1–9 to pick the best result, S to skip temporarily, ← / → for navigation.
  • Completed-item subfilters for perfect, best (including perfect), best without perfect, and each rejection reason; filters and progress span every prefix directory within the selected model.
  • React skin editor with larger tool buttons and an edited-image reference panel.
• Usage:

  cd best_annotator
  python3 app.py --root ../entropydrop_website_generations --host 0.0.0.0 --port 8765
  

  Open http://localhost:8765 in a browser. See best_annotator/README.md for full details.
• Testing:

  python3 -m unittest discover -p "test_*.py"
  npm ci
  npm test
  

2. Orphan Result Scanner (find_results_without_source.py)

Scans directories recursively for generated skin results that have no matching {id}_source.png input image:

python3 find_results_without_source.py
# Or check a specific model version:
python3 find_results_without_source.py --version v104 --output missing_v104.txt

The scan results are published atomically to results_without_source.txt.

---

⚠️ Error Samples (_err)

Generations that failed automated validation (e.g. failing camera angle, face orientation, or figure geometry checks) are deliberately preserved with the _err suffix.

• The _err designation applies to the task validation outcome.
• A _source_err file represents the original input for a failed task; it does not necessarily imply that the user's source image was corrupt, but rather that the pipeline failed to satisfy output constraints with it.
• These samples provide valuable negative examples for out-of-distribution detection, input sanitization, and classifier training.

---

📄 License

The dataset license is Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0). See LICENSE for the scope and full terms, or the official license summary.

• Scope: public skin assets retain their source CC BY-NC 4.0 license. Project-owned annotations, repair contributions, metadata, and dataset compilation contributions are offered under CC BY-NC 4.0 to the extent the project holds the relevant rights. Original reference images, prompts, and other third-party material retain their own rights and permissions; inclusion does not grant additional rights over them.
• Permitted use: the license allows non-commercial copying, sharing, and adaptation, subject to its terms. When sharing, retain the supplied author information (including {skin_id}_author.json or equivalent attribution), source references and license notices, provide a license link, and indicate modifications.
• Commercial use: uses requiring permission under this license must be non-commercial. Commercial use requires separate authorization from the relevant rights holders. A commercial license held by a skin's generating account does not transfer to dataset recipients.
• No additional publication requirement: this dataset license does not add an obligation to publish software source code, training code, or model weights, or require adaptations to use the same license. The underlying material remains subject to its applicable non-commercial and attribution conditions.
• Other rights and software: statutory exceptions and limitations remain unaffected. This dataset license does not relicense software or third-party dependencies; their applicable licenses remain in effect. Previously valid license grants are not revoked by this update.