Instructions to use Frontiers/alien-si-unified with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use Frontiers/alien-si-unified with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("unsloth/Qwen3-14B-unsloth-bnb-4bit") model = PeftModel.from_pretrained(base_model, "Frontiers/alien-si-unified") - Notebooks
- Google Colab
- Kaggle
Alien SI Unified (TIES merge)
Architect / creator: Barbu Mario Mihael
First concrete Alien SI artifact produced under the "acquire, don't pretrain" strategy:
a TIES merge of two expert LoRA adapters (alien-si-arsenal, alien-si-fusion) on
Qwen3-14B (r=32, alpha=64, all MLP+attention projections).
- Method: TIES (trim -> elect sign -> mean), density=0.5
- Source adapters: alien-si-arsenal, alien-si-fusion
- Tensors: 560 | NaN: none | std: 0.00485 | nonzero: 69.4%
- Feasible on free Kaggle CPU (33GB RAM), no GPU quota consumed
Part of the Alien SI program: modular cognitive architecture with an independent evaluator and a bounded recursive self-improvement loop.
- Downloads last month
- -
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support