Bopalv/dgpo-tiny-pure

DGPO Tiny Pure

Tiny 0.43M parameter language model trained with DGPO (Distillation-Guided Policy Optimization).

• • Python pickle model (1.8MB)
• • GGUF format

Usage:

import pickle
with open("dgpo_tiny.pkl", "rb") as f:
    model = pickle.load(f)