SIDIX: archived research build (Apr-Aug 2026)
Collection
Adapters of SIDIX, a self-hosted AI agent project whose research continued as MiganCore. Archived, not maintained. • 2 items • Updated
SIDIX persona v1 — archived experiment (April 2026)
Status: experimental, never used in production, archived. Despite the repository name, this is a plain LoRA adapter, not DoRA: DoRA was dropped before training.
What it is
- Base: Qwen/Qwen2.5-7B-Instruct (Apache-2.0).
- Method: LoRA, 3 epochs, one rented A4000 GPU (86 minutes).
Training data
7,500 persona question–answer pairs (6,750 train / 750 validation), assembled from templates (opener × body × closer) by a program generator. The generator, including its templates, was written with the help of an AI coding assistant. No model generated the rows at training time.
Evaluation
No pre-registered evaluation was published for this adapter.
More
- SIDIX archive: fahmiwol/sidix, full source at tag
v0.7.0. - Successor research: fahmiwol/migancore.
License: MIT for the adapter; the base model is Apache-2.0. Author: Fahmi Ghani (fahmiwol@gmail.com).
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support