# arman-bd/guppylm

A ~9M parameter LLM that talks like a small fish.

Repository: https://github.com/arman-bd/guppylm
Canonical: https://ross.abutalabs.com/products/guppylm
Language: Python
License Family: other
Last push: 2026-04-15T21:23:03+00:00

## Health v2 (maintenance only)
Score: 49/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 77, release rhythm 35, longevity 11
- inputs: {"age_days": 157, "days_push": 140, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, young, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3435, forks 304 (observed 2026-08-28T04:08:04.902610+00:00)

## What it is
GuppyLM is a ~9M parameter language model trained from scratch to talk like a small fish, built as an educational project demonstrating the full LLM pipeline: data generation, tokenizer, architecture, training loop, and inference. It ships with Colab notebooks, a browser demo, and a Hugging Face model and dataset.

## Use cases
- learn how to train your own language model from scratch
- understand the full LLM pipeline without a PhD or GPU cluster
- train a tiny transformer in a Colab notebook in minutes
- demo a small language model in the browser
- teach students how tokenizers, training loops, and inference work
- build a toy chatbot with a custom personality

## When to choose
- you want a hands-on educational walkthrough of building an LLM end to end
- you need a tiny, fast model that runs on free Colab or in-browser via WebAssembly
- you're teaching or learning how language models actually work under the hood

## When to avoid
- you need a production-quality LLM for real tasks like writing or coding
- you require strong reasoning, long context, or general knowledge
- you need a model with an established license for commercial use

## Facets
- artifact type: learning-resource
- maturity: active
- function: llm-training, llm-inference, machine-learning, deep-learning, data-generation
- domain: large-language-models, machine-learning, deep-learning, education, tutorials
- platform: python, browser
- tags: educational, tiny-language-model, from-scratch, colab-notebook, toy-model, transformer, beginner-friendly, web, gpu

## Member repositories
- arman-bd/guppylm (main) score 49

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:04.902610+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:37:42.481726+00:00, confidence not recorded.
  - readme: https://github.com/arman-bd/guppylm (fetched 2026-08-28T04:08:04.902610+00:00, sha 0e73b79b371a)
- Data as of 2026-08-30T08:39:29.467469+00:00.