piotrpiekos / MoSALinks
User-friendly implementation of the Mixture-of-Sparse-Attention (MoSA). MoSA selects distinct tokens for each head with expert choice routing providing a content-based sparse attention mechanism.
☆21Updated last month
Alternatives and similar repositories for MoSA
Users that are interested in MoSA are comparing it to the libraries listed below
Sorting:
- [NeurIPS 2023 spotlight] Official implementation of HGRN in our NeurIPS 2023 paper - Hierarchically Gated Recurrent Neural Network for Se…☆65Updated last year
- HGRN2: Gated Linear RNNs with State Expansion☆55Updated 10 months ago
- ☆23Updated 8 months ago
- [NeurIPS 2022] Your Transformer May Not be as Powerful as You Expect (official implementation)☆34Updated last year
- Official repository of paper "RNNs Are Not Transformers (Yet): The Key Bottleneck on In-context Retrieval"☆27Updated last year
- Here we will test various linear attention designs.☆59Updated last year
- Curse-of-memory phenomenon of RNNs in sequence modelling☆19Updated last month
- ☆32Updated last year
- [ICLR 2023] "Sparse MoE as the New Dropout: Scaling Dense and Self-Slimmable Transformers" by Tianlong Chen*, Zhenyu Zhang*, Ajay Jaiswal…☆52Updated 2 years ago
- This is a simple torch implementation of the high performance Multi-Query Attention☆17Updated last year
- ☆32Updated last year
- ☆55Updated 11 months ago
- MambaFormer in-context learning experiments and implementation for https://arxiv.org/abs/2402.04248☆55Updated last year
- The open-source materials for paper "Sparsing Law: Towards Large Language Models with Greater Activation Sparsity".☆23Updated 7 months ago
- The this is the official implementation of "DAPE: Data-Adaptive Positional Encoding for Length Extrapolation"☆38Updated 8 months ago
- [ICLR 2025] Official Code Release for Explaining Modern Gated-Linear RNNs via a Unified Implicit Attention Formulation☆42Updated 3 months ago
- ☆14Updated 6 months ago
- Reference implementation of "Softmax Attention with Constant Cost per Token" (Heinsen, 2024)☆24Updated last year
- [ICML 2024] When Linear Attention Meets Autoregressive Decoding: Towards More Effective and Efficient Linearized Large Language Models☆31Updated last year
- DeciMamba: Exploring the Length Extrapolation Potential of Mamba (ICLR 2025)☆28Updated 2 months ago
- Stick-breaking attention☆57Updated last week
- Code for "Everybody Prune Now: Structured Pruning of LLMs with only Forward Passes"☆28Updated last year
- Official code for the paper "Attention as a Hypernetwork"☆39Updated last year
- ☆23Updated 2 months ago
- ☆15Updated last week
- Official repository for the paper "Approximating Two-Layer Feedforward Networks for Efficient Transformers"☆38Updated last week
- [NAACL 2025] A Closer Look into Mixture-of-Experts in Large Language Models☆52Updated 4 months ago
- Xmixers: A collection of SOTA efficient token/channel mixers☆11Updated 7 months ago
- ☆13Updated 2 months ago
- Official Code Repository for the paper "Key-value memory in the brain"☆26Updated 4 months ago