Instructions to use SZLHOLDINGS/a11oy-code-r1-candidate with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use SZLHOLDINGS/a11oy-code-r1-candidate with Transformers:
# pip install -U transformers accelerate # Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("SZLHOLDINGS/a11oy-code-r1-candidate", device_map="auto") - Notebooks
- Google Colab
- Kaggle
a11oy-code-r1-candidate
LoRA fine-tune of Qwen/Qwen2.5-Coder-3B-Instruct on SZLHOLDINGS/a11oy-code-sft-v1 (estate source + doctrine), trained on the SZL A6000 mesh.
Abstention battery: 0/6 β gate FAIL
This is a research candidate. The abstention/refusal battery is part of the training receipt. The gate did NOT pass β this model fabricated on probes it should have abstained from. Treat as research only, not production.
- Full probes and responses:
training_report.jsonin this repo - Receipt on the chain:
receipts/a11oy-code-r1.jsonin SZLHOLDINGS/frontier-54-runs - Trained on the A6000 mesh (revision 21b0e2d0)
Verdict retro-embedded 2026-11-10 per the Kamachikuna (Law 2: never deceive; Law 7: every claim carries a receipt). The leg now embeds this at publish time.
Inference Providers NEW
This model isn't deployed by any Inference Provider. π Ask for provider support