Show HN: Benchmarks of continual learning model trained for a month on a laptop

An open-source project called mini-AGI has trained a byte-level, continually learning language model for a month on a single 8 GB laptop GPU (an Nvidia RTX 3070), reaching 1.14 bits/byte on enwik9 after reading only 2.8 GB of data at an estimated cost of about $56. The model pages 397 experts (1.26B total parameters) from an NVMe drive while keeping 32 resident (~109M active) in 5.1 GB of VRAM, and has consumed roughly 35% of its ~8 GB training corpus. In online mode, where the model specializes as it moves through the evaluation set, mini-AGI scored 1.447 bits/byte on PG19 versus 1.995 frozen, though both trail MambaByte-972M's 0.883, which was trained on 75 GB of PG19 books.

A month ago, I shared mini-AGI, an experiment in training a byte-level, continually learning language model from scratch on a single 8 GB laptop GPU an RTX 3070 . To recap the setup briefly: it completely decouples model size from VRAM by paging experts to and from an NVMe drive currently 397 experts / 1.26B total parameters on disk, with only 32 resident / ~109M active in 5.1 GB of VRAM . It trains on a single stream at batch 1, with the attention/embedding trunk learning rate dampened to 0.1x to mitigate catastrophic forgetting. The training is still going and went only through ~35% of the training corpus selected at the beginning of the run about 8GB in total . Here is the current status: 1. enwik9 First 10^9 bytes of English Wikipedia : Model Params Data Read Cost enwik9 bits/byte ↓ Llama-3.2-1B 1.23B 36 TB 9T tok ~$900k 0.68 Chinchilla 1B 1.00B 5.6 TB ~$68k 0.90 GPT-2 XL 1.56B 40 GB ~$43k 1.07 GPT-2 Large 0.77B 40 GB ~$21k 1.12 mini-AGI online 1.26B 2.8 GB ~$56 1.14 GPT-2 Medium 0.35B 40 GB ~$10k 1.17 mini-AGI dynamic 1.26B 2.8 GB ~$56 1.23 mini-AGI frozen 1.26B 2.8 GB ~$56 1.29 GPT-2 Small 0.12B 40 GB ~$3k 1.33 2. PG19 100 Project Gutenberg books, 41 MB test split : Model Params Trained On Cost PG19 bits/byte ↓ MambaByte-972M 0.97B 75 GB PG19 ~$4k 0.883 MegaByte-1B 1.02B 80 GB PG19 ~$1.6k 1.000 Transformer-320M 0.32B 80 GB PG19 ~$1.6k 1.057 mini-AGI online 1.26B 2.8 GB 0 PG19 ~$56 1.447 mini-AGI frozen 1.26B 2.8 GB 0 PG19 ~$56 1.995 The published models were trained on PG19's own training books, 75 to 80 GB of them, and are scored on more books like those. mini-AGI has never been trained on PG19 and has read no Victorian novels, so for it these books are completely out of its domain. There are more details in the repository. The most important part is a difference between "frozen" and "online" attempts where we allow the model to specialize on the task as it goes through the evaluation dataset. Clearly showcasing the power of continual learning regime simply unavailable for any other typical LLM. Full methodology, and replication code is in the repo: https://github.com/volotat/mini-AGI https://github.com/volotat/mini-AGI Intermediate weights snapshot on Hugging Face: https://huggingface.co/Volotat/mini-AGI-undertrained https://huggingface.co/Volotat/mini-AGI-undertrained Comments URL: https://news.ycombinator.com/item?id=50219762 https://news.ycombinator.com/item?id=50219762 Points: 2 Comments: 0