WFGY/specs
2025-09-27 17:07:43 +08:00
..
Smoothness_Lab Add files via upload 2025-06-15 14:28:16 +08:00
bbam.onnx Add files via upload 2025-06-12 15:22:09 +08:00
bbcr.onnx Add files via upload 2025-06-12 15:22:09 +08:00
bbmc.onnx Add files via upload 2025-06-12 15:22:09 +08:00
bbpf.onnx Add files via upload 2025-06-12 15:22:09 +08:00
README.md Update README.md 2025-09-27 17:07:43 +08:00

WFGY Public API Specs & ONNX Graphs

Version 1.0 · last updated 2025-06-12

WFGY is open by design.
To guarantee that each module can be inspected, audited, or re-implemented, we publish (1) a concise API spec and (2) a lightweight ONNX graph for every core component.
Download the .onnx files and open them in Netron or any ONNX viewer to see the exact computation graph—no black box, no hidden layers.

Module API spec ONNX graph
BBMC
Semantic Residue
bbmc_api.md onnx/bbmc.onnx
BBPF
Progression Flow
bbpf_api.md onnx/bbpf.onnx
BBCR
CollapseRebirth
bbcr_api.md onnx/bbcr.onnx
BBAM
Attention Modulator
bbam_api.md onnx/bbam.onnx

Re-exporting the graphs

All four graphs are intentionally tiny (~20 KB each) so they build in seconds and reveal the tensor flow without shipping proprietary weights.

# in repo root
python export_onnx.py

This script recreates the folder structure:

specs/
├── onnx/
│   ├── bbam.onnx
│   ├── bbcr.onnx
│   ├── bbmc.onnx
│   └── bbpf.onnx
└── *.md               # human-readable API specs

Feel free to replace these dummy graphs with full-precision weights if you need exact numerical reproduction.


License & Citation

The API descriptions and ONNX graphs are released under the MIT license. If you use WFGY in research, please cite the Zenodo release:

PSBigBig, “WFGY 1.0: A Self-Healing Variance Gate for LLMs,”

Any issue / PR is welcome—open science thrives on transparency.


---

### Will Colab runs change after adding this folder?

No. `specs/` is pure documentation.  
All examples (`example_01``08`), tests, and the SDK itself remain identical,
so you can keep running the original Colab notebook without any modification.

🧭 Explore More

Module Description Link
WFGY Core WFGY 2.0 engine is live: full symbolic reasoning architecture and math stack View →
Problem Map 1.0 Initial 16-mode diagnostic and symbolic fix framework View →
Problem Map 2.0 RAG-focused failure tree, modular fixes, and pipelines View →
Semantic Clinic Index Expanded failure catalog: prompt injection, memory bugs, logic drift View →
Semantic Blueprint Layer-based symbolic reasoning & semantic modulations View →
Benchmark vs GPT-5 Stress test GPT-5 with full WFGY reasoning suite View →
🧙‍♂️ Starter Village 🏡 New here? Lost in symbols? Click here and let the wizard guide you through Start →

👑 Early Stargazers: See the Hall of Fame
Engineers, hackers, and open source builders who supported WFGY from day one.

GitHub stars WFGY Engine 2.0 is already unlocked. Star the repo to help others discover it and unlock more on the Unlock Board.

WFGY Main   TXT OS   Blah   Blot   Bloc   Blur   Blow