HenryNdubuaku/super-lazy-autograd

Hand-derived memory-efficient VJPs for tuning LLMs on laptops.

25
/ 100
Experimental

This tool helps machine learning engineers or researchers fine-tune large language models (LLMs) like Qwen or DeepSeek on a personal laptop, even when memory is limited. It takes an existing LLM and a dataset of text, then outputs a specialized version of the model that performs better on your specific tasks. It is designed for those who need to iterate quickly on LLMs without access to high-end data center GPUs.

No commits in the last 6 months.

Use this if you need to fine-tune a supported large language model for a specific task but only have a laptop with limited memory available.

Not ideal if you have access to powerful GPU servers or cloud computing resources, as dedicated hardware will offer significantly faster and more stable training.

large-language-models model-fine-tuning natural-language-processing machine-learning-engineering
Stale 6m No Package No Dependents
Maintenance 2 / 25
Adoption 7 / 25
Maturity 16 / 25
Community 0 / 25

How are scores calculated?

Stars

38

Forks

Language

Python

License

Apache-2.0

Category

llm-fine-tuning

Last pushed

Apr 14, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/transformers/HenryNdubuaku/super-lazy-autograd"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.