volotat/mini-AGI
Continual learning model trained from scratch on 8GB VRAM laptop with batch-1 stream of data.
Older snapshot: check the original source for current information.
From the publisher
README & documentation
Source previewRead the project’s overview, installation instructions and usage examples. The original README is the source of truth.
mini-AGI - is a continual learning byte-level language model that assembles its own architecture, trains from scratch on a single 8 GB VRAM GPU, and keeps learning from everything it reads. It stores its weights as ordinary files on disk and pages them onto the card as it needs them, so the parameter count is bounded by free disk space rather than by VRAM. It grows new capacity while training when it runs short, prunes what nothing asks for, and reads through exactly the same code path it serves on. Targeted at a PC or laptop with at least an 8 GB VRAM GPU on the board. NOTE: as of now this is a small toy-level model. Do not expect a frontier level capabilities. This is rather a small experiment to show, that continual learning from the single stream of data without catastrophic forgetting is possible. Furthermore it is possible…
Read the full README ↗ · Preview checked 2026-11-02T16:12:06.867Z
Inside the original README — Document outline
- mini-AGI
- Motivation
- How the architecture works
- ...and the same thing while it writes
- How paging works
- How growth and pruning work
- How continual learning works
- This is a worst case, not a use case
- Reading your own files
- Benchmarks
- enwik9 against published models
- PG19 against published models
Headings are captured from the source. Links open the publisher’s document, not a locally hosted copy.
Links from the README
References supplied by the publisher, not independently verified endorsements. Check the destination before downloading files or entering credentials.
- its model card
- Training Deep Nets with Sublinear Memory Cost
- Training Compute-Optimal Large Language Models
- An Empirical Model of Large-Batch Training
- Muon is Scalable for LLM Training
- Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling
- Llama 3.2 model card
- The Smol Training Playbook
- lm-evaluation-harness
- TinyStories
- OpenHermes-2.5
- OpenThoughts-114k
Documentation belongs to its respective authors. Reported project/model license: MIT. A listing is not a grant of reuse or training rights. Confirm the document’s own terms at the source.
What this repository does
Continual learning model trained from scratch on 8GB VRAM laptop with batch-1 stream of data.
Repository facts
- Owner
- volotat
- Primary language
- Python
- Stars
- 1,205
- Forks
- 195
- Open issues + pull requests
- 15
- License
- MIT
- Archived
- No
- Default branch
- main
- Created
- 2026-09-19T00:25:59.000Z
- Last push
- 2026-10-07T04:16:57.000Z
Topics and intended use
No topics were included in the latest source metadata.
Review the README for scope, installation, examples and limitations. We do not run repository code or certify it.
Evaluate before installing
Review licensing and dependencies, inspect recent commits and unresolved issues, and test in an isolated environment before production use. Stars and forks alone cannot answer these questions.
Source and freshness
Source: GitHub. Metadata observed 2026-10-07T12:07:06.715Z. Daily imports are snapshots, not real-time monitoring.
Popularity and source listings do not establish security, suitability, licensing rights or benchmark performance.
Related repositories
EbookFoundation/free-programming-books
donnemartin/system-design-primer