Views
No views yet
Research checkpoint, not AGI. This compact model was trained from random initialization on a small, synthetic, auditable task. It is intended for reproducibility demonstrations and bounded evaluation only.
| Attribute | Value |
|---|---|
| Model type | tiny_cnn_classifier |
| Parameters | 56,710 |
| Training steps | 750 |
| Seed | 20260922 |
| Version | 0.1.0-research |
1{
2 "specialty": "vision",
3 "model_type": "tiny_cnn_classifier",
4 "parameter_count": 56710,
5 "train_steps": 750,
6 "final_train_loss": 0.006249,
7 "mean_train_loss": 0.467884,
8 "heldout_accuracy": 1.0,
9 "labels": [
10 "red_square",
11 "red_circle",
12 "green_square",
13 "green_circle",
14 "blue_square",
15 "blue_circle"
16 ],
17 "seed": 20260922,
18 "limitations": "Metrics apply only to procedural colored-shape images and do not demonstrate real-world visual understanding, OCR, face recognition, sensitive-attribute inference, or general multimodal ability."
19}train_portfolio.py. No private user data, credentials, raw images of people, or downloaded model weights are used. This makes provenance inspectable but also sharply limits capability and ecological validity.python3 train_portfolio.py --steps 80 --output artifacts from the project root. Record hardware, operating system, dependency versions, commit hash, and seed alongside any comparison.