Hacker News new | past | comments | ask | show | jobs | submit login
Show HN: NanoXLSTM: minimal codebase for playing with xLSTM language models (github.com/jadechip)
3 points by jadechips 8 months ago | hide | past | favorite
nanoXLSTM is a minimal codebase for playing around with language models based on the xLSTM (extended Long Short-Term Memory) architecture from the awesome research paper: xLSTM: Extended Long Short-Term Memory and heavily inspired by Andrej Karpathy's nanoGPT.

*Note: Work in progress!!! I am working on improving the generated text.

No lofty goals here - just a simple codebase for tinkering with this innovative xLSTM technology!




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: