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}
}