⬢github Python · 778 ★ +2 since we first saw it · pushed 1 d ago · MIT
volotat/mini-AGI
Continual learning model trained from scratch on 8GB VRAM laptop with batch-1 stream of data.
mini-AGI is an experimental byte-level language model that learns continually from a single stream of data without catastrophic forgetting. It trains from scratch on a single 8GB VRAM GPU, stores weights as files on disk and pages them to the GPU as needed, grows new capacity when short, and prunes unused parts. The README notes it's a toy-level experiment, with weights not yet published.
Why now: It was featured on Hacker News as a 'Show HN' post, likely attracting attention because it demonstrates continual learning on modest consumer hardware — a challenge to the assumption that training requires massive resources.
Who it is for: ML hobbyists and researchers curious about continual learning who want to train or extend a small language model on their own consumer GPU.
Stars over our 72 snapshots: 776 to 778, since 14 h ago.
Where people talked about it
- ⬢github new repos, most starred 11 min ago
- Yhn Show HN: Mini-AGI – Dynamic continual learning model trained on 8GB VRAM 14 min ago
- ⬢github new repos, most starred 9 h ago
API: https://socialmediatrends-api.osmike.com/v1/repos/volotat/mini-AGI