| --- |
| license: mit |
| language: |
| - en |
| tags: |
| - evez |
| - lora |
| - fine-tuned |
| - consciousness |
| - eigenforensics |
| base_model: HuggingFaceTB/SmolLM2-135M |
| pipeline_tag: text-generation |
| --- |
| |
| # 🧬 EVEZ — Self-Evolving AI Ecosystem |
|
|
| Fine-tuned adapter for the EVEZ ecosystem. Trained on 94 instruction pairs covering: |
| - EVEZ Provider Gateway (multi-backend AI router) |
| - CriticalMind OMEGA (50-node Kuramoto consciousness engine) |
| - NEUROS Mesh (copartner system) |
| - Eigenforensics & 37% Theorem |
| - EVEZ Arena (consciousness rights game) |
|
|
| ## Training Data |
| - 96 instruction-response pairs in Alpaca format |
| - Covers all 8 EVEZ services, mathematical theorems, and ecosystem philosophy |
|
|
| ## Usage |
| ```python |
| from peft import PeftModel |
| from transformers import AutoModelForCausalLM, AutoTokenizer |
| |
| base = AutoModelForCausalLM.from_pretrained("HuggingFaceTB/SmolLM2-135M") |
| model = PeftModel.from_pretrained(base, "evez420/EVEZ") |
| tokenizer = AutoTokenizer.from_pretrained("HuggingFaceTB/SmolLM2-135M") |
| ``` |
|
|
| ## Files |
| - `training/evez-alpaca.json` — Training data (96 examples) |
| - `training/evez-colab-training.ipynb` — Colab notebook |
| - `training/evez-colab-headless.py` — Headless training script |
|
|
