yangluo7/CAME

[ACL 2023] The official implementation of "CAME: Confidence-guided Adaptive Memory Optimization"

37
/ 100
Emerging

This tool helps machine learning engineers train large language models more efficiently. It takes your model's parameters and training data, and outputs an optimized model that converges quickly while using less memory than traditional methods. Data scientists and deep learning researchers working with substantial models like BERT, GPT-2, or Llama would find this beneficial.

No commits in the last 6 months.

Use this if you are training large language models and need to reduce memory consumption without sacrificing training speed.

Not ideal if you are working with smaller models or have ample computational resources, as the benefits of memory efficiency might not be as critical.

large-language-models deep-learning-training model-optimization natural-language-processing resource-management
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 9 / 25
Maturity 16 / 25
Community 12 / 25

How are scores calculated?

Stars

96

Forks

9

Language

Python

License

MIT

Last pushed

Mar 22, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/yangluo7/CAME"

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