NX-AI/xlstm
A recurrent neural network architecture extending LSTM with exponential gating and matrix memory for language modeling.
Not currently ranked — collecting fresh signals.
star history
xLSTM is a new RNN architecture based on the original LSTM that incorporates Exponential Gating with normalization and stabilization techniques along with a new Matrix Memory to overcome traditional LSTM limitations. The architecture demonstrates competitive performance on language modeling tasks compared to Transformers and State Space Models. The repository includes training code and a 7B parameter xLSTM Large model trained on 2.3T tokens.
Frequently asked
- What is NX-AI/xlstm?
- A recurrent neural network architecture extending LSTM with exponential gating and matrix memory for language modeling.
- Is xlstm open source?
- Yes — NX-AI/xlstm is open source, released under the Apache-2.0 license.
- What language is xlstm written in?
- NX-AI/xlstm is primarily written in Python.
- How popular is xlstm?
- NX-AI/xlstm has 2.2k stars on GitHub.
- Where can I find xlstm?
- NX-AI/xlstm is on GitHub at https://github.com/NX-AI/xlstm.