AI Dev Impact Lab JA
← Topics ranking · 2026-11
GITHUB TOPIC

#from-scratch

GitHub repositories that have self-applied the topic "from-scratch" — a creator-tagged metadata that surfaces how AI projects describe themselves.

10
tagged repos
4
top 10 stars
4
with tool sigs
10
shown

REPOS Repos for #from-scratch (top 10 by stars)

Abhisingh18/Sutra-1.3B-Model

A 1.32B-parameter Mixture-of-Experts LLM trained from scratch in PyTorch - custom tokenizer, MLA attention, 18B-token pretraining, SFT and DPO. No pretrained weights, no HuggingFace Trainer.

HTML ★ 2 AI 70 Solo live ↗
atandra2000/GPT-OSS-Lite

Faithful from-scratch PyTorch reproduction of OpenAI's GPT-OSS architecture (sliding/full attention alternation, learned attention sinks, YaRN 128K, top-2-of-8 MoE), scaled to Chinchilla-optimal 502M total / 247M active training on a single A100 80GB

Python ★ 1 AI 90 2 sig
monkfromearth/monkway

Monk LLM Gateway: one API in front of every model. An LLM gateway and router built from scratch (proxy, routing, retry/fallback, caching, budgets), with an interactive course.

Astro ★ 1 AI 70 Solo 1 sig live ↗
aurimas13/Calculator-Agent

Calculator Agent: First AI agent, built from scratch. Claude using a calculator tool in a loop, with every step streamed live to a web UI. No agent framework.

Python ★ 0 AI 100 Solo live ↗
zch6a/build-your-own-ai

Build your own AI from scratch: tokenizer, autograd, GPT, inference, vector database, RAG, agent and MCP. Each a small, tested, readable module that runs on a laptop CPU with no API key.

Python ★ 0 AI 100
KeigoShimadaCC/llm-from-scratch

Mac-local LLM-from-scratch lab for building and evaluating a small decoder-only language model on Apple Silicon.

Python ★ 0 AI 90 2 sig
iqureshi123/tinyinfer#1335284915

An LLM inference engine written from scratch in Python and C++/Metal — tokenizer, forward pass, KV cache, and INT4 quantization implemented by hand. No PyTorch, no llama.cpp.

Python ★ 0 AI 70
iqureshi123/tinyinfer

An LLM inference engine written from scratch in Python and C++/Metal — tokenizer, forward pass, KV cache, and INT4 quantization implemented by hand. No PyTorch, no llama.cpp.

Python ★ 0 AI 70
PrayasPanda/tinytalk

A 14M-parameter GPT built from scratch in PyTorch hand-written attention, custom BPE tokenizer, no pretrained weights.

Python ★ 0 AI 70 Solo live ↗
DevxD98/JimmyLabs

JimmyLabs — a GPT-style language model built from scratch and fully documented on a MacBook Air M1 (8GB, Apple Silicon). Documentation-first educational research platform.

Python ★ 0 AI 60 1 sig

RELATED Other topics · full topics ranking →

#claude-code

3,787

#ai-agents

3,201

#llm

2,090

#mcp

1,866

#claude

1,854

#python

1,678

#developer-tools

1,530

#codex

1,376

#ai

1,181

Aggregated by case-insensitive match against topics_json of each repo's latest content snapshot. methodology