AltHub

FareedKhan-dev/train-llm-from-scratch Alternatives

LLM training from scratch / educational

Educational PyTorch codebase that trains an LLM from scratch, from raw text to an aligned reasoning model.

9.4kPythonMITactive Β· last push 2mo agoGitHub

Drop-in peers

Same problem, same approach. Swapping one for another is a config change, not a rewrite.

datawhalechina/happy-llmGitHub

Same from-scratch LLM tutorial approach; choose it for a maintained Chinese-language walkthrough.

πŸ“š δ»Žι›ΆεΌ€ε§‹ζž„ε»Ίε€§ζ¨‘εž‹

33.1k
Jupyter Notebook
NOASSERTION
active2mo ago
rasbt/reasoning-from-scratchGitHub

Choose when your goal is specifically a step-by-step reasoning-model training walkthrough in PyTorch.

Implement a reasoning LLM in PyTorch from scratch, step by step

5.3k
Jupyter Notebook
Apache-2.0
active23d ago
raiyanyahya/how-to-train-your-gptGitHub

Choose it if you want an ultra-commented, beginner-oriented from-scratch GPT implementation.

Build a modern LLM from scratch. Every line commented. Explained like we are five.

3.4k
Jupyter Notebook
MIT
active1mo ago
rasbt/LLMs-from-scratchGitHub

Choose this for a broader, continuously updated textbook-style from-scratch LLM implementation.

Implement a ChatGPT-like LLM in PyTorch from scratch, step by step

105.3k
Jupyter Notebook
NOASSERTION
active23d ago

Same job, different approach

Solves the same problem with a different architecture or at a different layer. Expect to rewrite the integration.

Lightning-AI/litgptGitHub

Choose when you need a production-grade training/finetuning framework with many model recipes instead of hand-written educational code.

20+ high-performance LLMs with recipes to pretrain, finetune and deploy at scale.

13.7k
Python
Apache-2.0
active26d ago

Listed as an alternative to

These projects were analysed and named train-llm-from-scratch among their alternatives. The relationship is not symmetric β€” how train-llm-from-scratch rates them is a separate judgement, made when train-llm-from-scratch is analysed in its own right.

jingyaogong/minimindGitHub

calls train-llm-from-scratch β€œChoose this if you want a concise end-to-end pipeline that includes data download and text generation with minimal dependencies.”

🧠 Train a 64M-parameter LLM from scratch in just 2h!

61.7kPythonactive
stas00/ml-engineeringGitHub

calls train-llm-from-scratch β€œFor beginners wanting a single end-to-end small LLM training tutorial, but narrower than the seed's operational breadth.”

Machine Learning Engineering Open Book

18.7kPythonactive
Lightning-AI/litgptGitHub

calls train-llm-from-scratch β€œminimal from-scratch training script for small-scale models, lacks distributed and finetuning features”

20+ high-performance LLMs with recipes to pretrain, finetune and deploy at scale.

13.7kPythonactive