zaydzuhri / softpick-attentionLinks
Landing repository for the paper "Softpick: No Attention Sink, No Massive Activations with Rectified Softmax"
β84Updated 3 weeks ago
Alternatives and similar repositories for softpick-attention
Users that are interested in softpick-attention are comparing it to the libraries listed below
Sorting:
- πSmall Batch Size Training for Language Modelsβ62Updated this week
- β85Updated last year
- Pytorch implementation of the PEER block from the paper, Mixture of A Million Experts, by Xu Owen He at Deepmindβ128Updated last year
- H-Net Dynamic Hierarchical Architectureβ80Updated 3 weeks ago
- β102Updated 2 months ago
- The evaluation framework for training-free sparse attention in LLMsβ98Updated 3 months ago
- Tree Attention: Topology-aware Decoding for Long-Context Attention on GPU clustersβ130Updated 10 months ago
- The simplest, fastest repository for training/finetuning medium-sized GPTs.β164Updated 3 months ago
- Normalized Transformer (nGPT)β191Updated 10 months ago
- Tiled Flash Linear Attention library for fast and efficient mLSTM Kernels.β70Updated last month
- supporting pytorch FSDP for optimizersβ84Updated 9 months ago
- Mixture of A Million Expertsβ48Updated last year
- β58Updated last year
- Explorations into the recently proposed Taylor Series Linear Attentionβ100Updated last year
- The official github repo for "Diffusion Language Models are Super Data Learners".β116Updated this week
- research impl of Native Sparse Attention (2502.11089)β61Updated 7 months ago
- Collection of autoregressive model implementationβ86Updated 5 months ago
- NanoGPT-speedrunning for the poor T4 enjoyersβ72Updated 5 months ago
- β82Updated last year
- β64Updated 6 months ago
- DeMo: Decoupled Momentum Optimizationβ193Updated 10 months ago
- Official implementation of Phi-Mamba. A MOHAWK-distilled model (Transformers to SSMs: Distilling Quadratic Knowledge to Subquadratic Modeβ¦β116Updated last year
- Fork of Flame repo for training of some new stuff in developmentβ18Updated last month
- [ICML 2025] Roll the dice & look before you leap: Going beyond the creative limits of next-token predictionβ70Updated 4 months ago
- Repository for the Q-Filters method (https://arxiv.org/pdf/2503.02812)β35Updated 6 months ago
- Flash-Muon: An Efficient Implementation of Muon Optimizerβ189Updated 3 months ago
- RWKV-X is a Linear Complexity Hybrid Language Model based on the RWKV architecture, integrating Sparse Attention to improve the model's lβ¦β47Updated 2 months ago
- Code for NeurIPS 2024 Spotlight: "Scaling Laws and Compute-Optimal Training Beyond Fixed Training Durations"β84Updated 11 months ago
- Fast and memory-efficient exact attentionβ70Updated 7 months ago
- β89Updated last year