# no-magic-ai/no-magic

Because `model.fit()` isn't an explanation

Repository: https://github.com/no-magic-ai/no-magic
Canonical: https://ross.abutalabs.com/products/no-magic
Homepage: https://no-magic-ai.github.io/
Language: Python
License: MIT
License Family: permissive
Topics: ai-algorithms, algorithms, no-dependencies, open-soruce
Last push: 2026-04-26T10:11:31+00:00

## Health v2 (maintenance only)
Score: 67/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 79, release rhythm 81, longevity 14
- inputs: {"age_days": 201, "days_push": 129, "days_rel": 129, "gap_med": 30.0, "n_releases_24m": 3}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1410, forks 108 (observed 2026-08-28T04:04:38.883938+00:00)

## What it is
A curated collection of 48 single-file, zero-dependency Python implementations of core AI algorithms, from GPT and LSTM to LoRA, quantization, and MCTS. Each script is a complete runnable program that trains a model from scratch, designed as an educational resource for understanding what frameworks like PyTorch do underneath.

## Use cases
- learn how transformers and GPT work from scratch
- understand what model.fit() does under the hood
- implement BPE tokenization without libraries
- study LoRA, DPO, and RLHF alignment algorithms
- learn quantization and KV-cache inference optimization
- understand RAG and vector search internals
- study agent algorithms like MCTS and ReAct
- teach deep learning concepts with runnable code

## When to choose
- you want to deeply understand AI algorithms rather than just call APIs
- you prefer dependency-free, runnable single-file examples
- you want a structured curriculum through modern ML concepts
- you're preparing for ML interviews or teaching

## When to avoid
- you need production-ready, performant model training
- you want GPU-accelerated training with ecosystem support
- you need a framework to build applications on top of

## Facets
- artifact type: learning-resource
- maturity: active
- function: machine-learning, deep-learning, llm-training, agent-framework, rag, nlp
- domain: machine-learning, deep-learning, large-language-models, artificial-intelligence, education, tutorials
- platform: python, cross-platform, cli
- tags: educational, zero-dependencies, single-file, algorithms-from-scratch, transformers, reinforcement-learning, quantization, microgpt, ai-agents

## Member repositories
- no-magic-ai/no-magic (main) score 67

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:38.883938+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 2026-08-30T04:38:29.180228+00:00, confidence not recorded.
  - readme: https://github.com/no-magic-ai/no-magic (fetched 2026-08-28T04:04:38.883938+00:00, sha 3a1535b27e8f)
  - homepage: https://no-magic-ai.github.io/ (fetched 2026-08-29T11:51:57.083884+00:00, sha 8d10333454ef)
- Data as of 2026-08-30T08:39:29.467469+00:00.