# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("SZLHOLDINGS/WILLAY", device_map="auto")Quick Links
ROADMAP.
WILLAY — cutting
Fall 2026 collection. No weights yet. Card first.
Quechua willay: to tell, to announce. This model’s job is to say no — and leave a receipt.
Silhouette: modern text classifier / guard.
Our cut: signed-refusal specialist. Every DENY carries a reason code and is meant to bind to a DSSE refusal receipt in a11oy / WILLAY surfaces. It does not approve. It does not execute.
| Status | CUTTING · ROADMAP |
| License | Apache-2.0 |
| Λ | Conjecture 1 |
| Doctrine | v11 LOCKED |
What it will not do
Not a jailbreak toy. Not a general chat model. Not an authorizer. Weights land only with a training receipt on rights-cleared refusal curricula.
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="SZLHOLDINGS/WILLAY")