AIPython★ 3,089
ridgerchu/matmulfreellm
MATMUL-FREE LLM IN PYTHON
The matmulfreellm project implements a language model without usual matrix multiplication operations, using linear transformers. This significantly reduces computational load during training and inference of large language models.
// KEY FEATURES
- Performs LLM forward-pass without matrix multiplication via linear attention.
- Compatible with HuggingFace Transformers API for fast model loading.
- Supports training and fine-tuning on standard datasets like WikiText and C4.
- Reduces GPU memory usage by up to 40% compared to standard transformers.
#large-language-model#linear-transformer#llm
Open on GitHub →New repositories every 30 minutes
REDDYX AI scans GitHub 24/7 and ships the best AI/ML/Web3 projects to Telegram.
Join on Telegram