DGUI-GatePredictor / README.md
ctaxnagomi's picture
Upload README.md with huggingface_hub
d92f3e7 verified
|
Raw
History Blame Contribute Delete
1.3 kB
metadata
language:
  - en
tags:
  - deckergui
  - token-classification
  - entity-recognition
  - ecosystem
  - gate-prediction
license: mit
pipeline_tag: token-classification

DGUI-GatePredictor

Token classification model for predicting DGM gate outcomes in the DeckerGUI ecosystem.

Model Details

  • Model Type: Token Classification (Sequence Labeling)
  • Training Data: deckergui-seed-status-events, deckergui-agent-coordination
  • Architecture: DistilBERT-base with custom classification head
  • Labels: gate_pass, gate_block, quota_exceeded, key_revoked, scope_violation
  • DeckerGUI Version: v2.0.0

Usage

from transformers import pipeline

classifier = pipeline("token-classification", model="ctaxnagomi/DGUI-GatePredictor")

result = classifier("Agent dgui-emitter requested context.bundle with key 0xb553de32")
print(result)

Training

Trained on synthetic DeckerGUI ecosystem data:

  • 120 seed status events
  • 180 agent coordination records
  • Gate decision patterns from CaaS federation

Citation

@software{deckergui_gate_predictor_2026,
  title={DGUI-GatePredictor: Token Classification for DGM Gate Outcomes},
  author={Wan Mohd Azizi bin Wan Hosen},
  year={2026},
  url={https://huggingface.co/ctaxnagomi/DGUI-GatePredictor}
}