HackerReposA FIELD GUIDE TO OPEN SOURCE
ProSponsor the shelf♧ 0
← Back to top all-time
volotat
mini-AGI
Python2026–2026
volotat / mini-AGI

mini-AGI

Continual learning model trained from scratch on 8GB VRAM laptop with batch-1 stream of data.

3Distinct people
4Mentions
0Explicit recommenders
1.3KGitHub stars
REPOSITORY AT A GLANCE

What to know before exploring

What it is

Continual learning model trained from scratch on 8GB VRAM laptop with batch-1 stream of data.

GitHub

Where it is used

No documented product use has been verified yet.

What people are saying

No recommendation phrase was found in the indexed comments.

No classified warning is attached to this page. Read the discussions before treating silence as approval.

Also in these discussions

No other repository is repeatedly linked in the same discussions.

This week on Hacker News

Nov 9–Nov 16 UTC · Direct repository links and separately labeled related project stories. Indexed through 11/16/2026, 2:19:09 PM UTC.

No direct repository links were found in the indexed HN discussions over the past seven days. Older discussions are available below.

Mentions over time

HN mentions by week4 dated mentions

Observed mentions: 2026-09-21 04:42:37 UTC to 2026-10-24 10:09:27 UTC. Weeks start Monday, UTC.

01232026-09-21 00:00:00 UTC to 2026-09-28 00:00:00 UTC (exclusive): 3 mentions; 1 submissions, 2 commentsSep 212026-09-28 00:00:00 UTC to 2026-10-05 00:00:00 UTC (exclusive): 0 mentions; 0 submissions, 0 comments2026-10-05 00:00:00 UTC to 2026-10-12 00:00:00 UTC (exclusive): 0 mentions; 0 submissions, 0 commentsOct 52026-10-12 00:00:00 UTC to 2026-10-19 00:00:00 UTC (exclusive): 0 mentions; 0 submissions, 0 comments2026-10-19 00:00:00 UTC to 2026-10-26 00:00:00 UTC (exclusive): 1 mentions; 1 submissions, 0 commentsOct 19

Counts come from the repository’s dated source evidence. Index cutoff: 2026-11-02 08:16:59 UTC.

View mention counts and date ranges
UTC ranges include their start and exclude their end. Select a period to filter the source evidence.
PeriodStart (UTC)End (exclusive, UTC)MentionsSubmissionsComments
Sep 212026-09-21 00:00:00 UTC2026-09-28 00:00:00 UTC312
Sep 282026-09-28 00:00:00 UTC2026-10-05 00:00:00 UTC000
Oct 52026-10-05 00:00:00 UTC2026-10-12 00:00:00 UTC000
Oct 122026-10-12 00:00:00 UTC2026-10-19 00:00:00 UTC000
Oct 192026-10-19 00:00:00 UTC2026-10-26 00:00:00 UTC110

What Hacker News says

3 sources
I built a smaller scale replication. Most of your claims prove out but the lever is essentially Biderman et als "learn less forget less". My first pass actually made this mistake initially and the findings didn't replicate. After controlling for step density I replicated your findings. Here is a write up: https://huggingface.co/spaces/dreddnafious/mini-agi-replicat... At the bottom of my write up I added some ideas to pin down the actual dominant parameters. I also added a pr to fix an issue in the codebase: PR: https://github.com/volotat/mini-AGI/pull/20 What it fixes: a crash in upstream's GradSNR meter (issue #19), a diagnostic that tracks how much of the gradient is signal versus noise.
In Show HN: Mini-AGI – Dynamic continual learning model trained on 8GB VRAM
Using the term AGI and not including any performance analysis. My AI calls it: "massive marketing overreach". Somebody called this slop in the comments. As a professor who published on continual learning I'm leaning towards agreement[1]. It lacks any substance. No relation to related work, no description of algorithm, no ablation study, just hand-waving that we're feeding some data and "Chess is not forgotten". This "how-continual-learning-works" markdown text is not an algorithm [2]. [1] https://arxiv.org/abs/2301.12530 [2] https://github.com/volotat/mini-AGI/#how-continual-learning-...
In Show HN: Mini-AGI – Dynamic continual learning model trained on 8GB VRAM
volotat·Sep 21, 2026Submission

Show HN: Mini-AGI – Dynamic continual learning model trained on 8GB VRAM

Sorry for the pretentious name, I know, I know.. It just contains all the pieces I would like to see a AGI model to have, and I can't stand the temptation. Before throwing rocks at me, please take a glance at the Readme, and I hope it will cover your mood a little bit. So, first of all it does work and you can see the sample from the whole training run here: https://raw.githubusercontent.com/volotat/mini-AGI/refs/head... Here is the scaling law graph I have so far, and it looks very promising: https://github.com/volotat/mini-AGI/blob/main/assets/scaling... The model was built under my deep dissatisfaction so we cannot really train even moderately big models (1B+ scale) on the consumer's hardware. We can inference and fine-tune them for sure, but I would like to have full control over what the model sees over the training run, so it is ful
Read the rest
ly aligned with my interests, not some corporations. I was thinking about for some time and come up with two interesting ideas I thought worth pursuing: MoE with a lot of experts that gets added and pruned from the model while it trains, where only a small subset of of experts are actually in use at any particular moment + batch 1 training on the single continuous stream of data. First allows us to be bounded only by the disk space in terms of number of parameters and load and unload experts only when they are needed. The second (if figured out and it turns out to be doable) allows us to get aways with small VRAM capacity because we do not need to store big randomized batches and their respective gradients. I started brainstorming with Claude and after some time we found an approach that seems to be promising, and low and behold, a few weeks pass and you can see the results yourself. Obviously, I did use AI in the process of making this project and I am pretty sure it would be completely impossible for me to do something like this without it, so I hope it is more than justified. The model is still running over the first of 7.8B characters corpus I selected for training, so the weights are not out yet, and it's about a couple weeks of waiting until they are cooked at the current reading speed. And yeah, the model just read continuous interleaved passages from the dataset, each by 32K characters long each as a single stream. Just as you or I would do. The set up seems to be really simple so you can git clone the project, run it and observe everything for yourself. Thanks for your attention.
277 points · Open discussion

MAKE YOURSELF AT HOME

A few shortcuts

/
Search the collection
?
Open this guide
Esc
Close dialogs and navigation

Your saved repositories and interests stay in this browser.