Quasar Attention Continuous-Time Memory Dynamics for Scaling Models Context to Millions of Tokens
☆29Mar 26, 2026Updated 6 months ago
Alternatives and similar repositories for quasar-attention
Users that are interested in quasar-attention are comparing it to the libraries listed below. We may earn a commission when you buy through links labeled 'Ad' on this page.
Sorting:
- [ACL 2026] Official repo for "Native Hybrid Attention for Efficient Sequence Modeling"☆33Sep 17, 2026Updated last month
- The official implementation of HybridNorm: Towards Stable and Efficient Transformer Training via Hybrid Normalization☆19Mar 7, 2025Updated last year
- My personal ML infrastructure☆13Nov 15, 2025Updated 11 months ago
- ☆11Mar 29, 2025Updated last year
- The official repository for SkyLadder: Better and Faster Pretraining via Context Window Scheduling☆44Dec 29, 2025Updated 9 months ago
- Deploy on Railway without the complexity - Free Credits Offer • AdConnect your repo and Railway handles the rest with instant previews. Quickly provision container image services, databases, and storage volumes.
- Official repository for the paper "Blending Complementary Memory Systems in Hybrid Quadratic-Linear Transformers"☆13Jun 3, 2025Updated last year
- Use the tokenizer in parallel to achieve superior acceleration☆20Mar 21, 2024Updated 2 years ago
- ☆23Updated this week
- train entropix like a champ!☆20Oct 10, 2024Updated 2 years ago
- look how they massacred my boy☆63Oct 16, 2024Updated 2 years ago
- ☆19Dec 4, 2025Updated 10 months ago
- Parallel Associative Scan for Language Models☆18Jan 8, 2024Updated 2 years ago
- manipulating cointegrated pairs to achieve a market-neutral strategy that outperforms indices☆11Jan 12, 2021Updated 5 years ago
- https://arxiv.org/abs/2610.13990☆33Oct 15, 2026Updated last week
- Deploy to Railway using AI coding agents - Free Credits Offer • AdUse Claude Code, Codex, OpenCode, and more. Autonomous software development now has the infrastructure to match with Railway.
- Efficient PScan implementation in PyTorch☆17Jan 2, 2024Updated 2 years ago
- Gecko Architecture☆19Jan 13, 2026Updated 9 months ago
- [Tech Report, NeurIPS 2026] Expanded Hyper-Connections☆75Updated this week
- LongAttn ：Selecting Long-context Training Data via Token-level Attention☆15Jul 16, 2025Updated last year
- ☆93Aug 18, 2024Updated 2 years ago
- ☆23Jun 12, 2025Updated last year
- Structured Primitives for Efficient Architecture Research☆21Dec 22, 2025Updated 10 months ago
- [NeurIPS '25] Multi-Token Prediction Needs Registers☆34Dec 14, 2025Updated 10 months ago
- ☆16Nov 5, 2018Updated 7 years ago
- AI Agents on DigitalOcean Gradient AI Platform • AdBuild production-ready AI agents using customizable tools or access multiple LLMs through a single endpoint. Create custom knowledge bases or connect external data.
- Sparse Attention with Linear Units☆20Apr 21, 2021Updated 5 years ago
- FastMCP 3.2 server and webapp wrapping the World Labs Marble API — generate navigable 3D worlds from text, images, and video☆24Oct 14, 2026Updated last week
- Combining SOAP and MUON☆24Feb 11, 2025Updated last year
- Compositional Muon release☆25Jun 5, 2026Updated 4 months ago
- ☆66Jun 12, 2025Updated last year
- ☆18Mar 10, 2023Updated 3 years ago
- Stick-breaking attention☆64Jul 1, 2025Updated last year
- A Tool to visualize NN training☆15Feb 21, 2026Updated 8 months ago
- The code of Advancing Expert Specialization for Better MoE (NeurIPS2025 oral)☆35Jan 22, 2026Updated 9 months ago
- GPU virtual machines on DigitalOcean Gradient AI • AdGet to production fast with high-performance AMD and NVIDIA GPUs you can spin up in seconds. The definition of operational simplicity.
- [ACL'21] Data for "An In-depth Study on Internal Structure of Chinese Words".☆14May 27, 2021Updated 5 years ago
- ✒️ ChatGPT as a writing partner.☆14Mar 6, 2023Updated 3 years ago
- ☆147Feb 4, 2026Updated 8 months ago
- ☆23Dec 8, 2022Updated 3 years ago
- Research without Re-search: Maximal Update Parametrization Yields Accurate Loss Prediction across Scales☆33Jul 17, 2023Updated 3 years ago
- Expanding linear RNN state-transition matrix eigenvalues to include negatives improves state-tracking tasks and language modeling without…☆22Mar 15, 2025Updated last year
- ☆69Mar 21, 2025Updated last year