JT-Ushio / MHA2MLALinks
Towards Economical Inference: Enabling DeepSeek's Multi-Head Latent Attention in Any Transformer-based LLMs
☆188Updated 2 months ago
Alternatives and similar repositories for MHA2MLA
Users that are interested in MHA2MLA are comparing it to the libraries listed below
Sorting:
- Parallel Scaling Law for Language Model — Beyond Parameter and Inference Time Scaling☆429Updated 3 months ago
- TransMLA: Multi-Head Latent Attention Is All You Need☆339Updated last month
- Skywork-MoE: A Deep Dive into Training Techniques for Mixture-of-Experts Language Models☆136Updated last year
- [ICML 2025] |TokenSwift: Lossless Acceleration of Ultra Long Sequence Generation☆113Updated 3 months ago
- Chain of Experts (CoE) enables communication between experts within Mixture-of-Experts (MoE) models☆220Updated last month
- [ICLR 2025] DuoAttention: Efficient Long-Context LLM Inference with Retrieval and Streaming Heads☆487Updated 6 months ago
- ☆199Updated 4 months ago
- A highly capable 2.4B lightweight LLM using only 1T pre-training data with all details.☆208Updated 3 weeks ago
- Simple extension on vLLM to help you speed up reasoning model without training.☆177Updated 2 months ago
- Unofficial implementation for the paper "Mixture-of-Depths: Dynamically allocating compute in transformer-based language models"☆169Updated last year
- Efficient triton implementation of Native Sparse Attention.☆208Updated 3 months ago
- Official implementation of "Fast-dLLM: Training-free Acceleration of Diffusion LLM by Enabling KV Cache and Parallel Decoding"☆374Updated this week
- [ICLR2025] Codebase for "ReMoE: Fully Differentiable Mixture-of-Experts with ReLU Routing", built on Megatron-LM.☆87Updated 8 months ago
- ☆79Updated 4 months ago
- SeerAttention: Learning Intrinsic Sparse Attention in Your LLMs☆147Updated 2 weeks ago
- [ICLR 2025] COAT: Compressing Optimizer States and Activation for Memory-Efficient FP8 Training☆227Updated 2 weeks ago
- Lightning Attention-2: A Free Lunch for Handling Unlimited Sequence Lengths in Large Language Models☆324Updated 6 months ago
- Official codebase for "Can 1B LLM Surpass 405B LLM? Rethinking Compute-Optimal Test-Time Scaling".☆269Updated 6 months ago
- Implementation of the LongRoPE: Extending LLM Context Window Beyond 2 Million Tokens Paper☆149Updated last year
- Fira: Can We Achieve Full-rank Training of LLMs Under Low-rank Constraint?☆115Updated 10 months ago
- [ICML'24] The official implementation of “Rethinking Optimization and Architecture for Tiny Language Models”☆123Updated 7 months ago
- ☆117Updated 2 months ago
- Block Transformer: Global-to-Local Language Modeling for Fast Inference (NeurIPS 2024)☆160Updated 4 months ago
- qwen-nsa☆74Updated 4 months ago
- MiroMind-M1 is a fully open-source series of reasoning language models built on Qwen-2.5, focused on advancing mathematical reasoning.☆202Updated last week
- siiRL: Shanghai Innovation Institute RL Framework for Advanced LLMs and Multi-Agent Systems☆168Updated last week
- Efficient LLM Inference over Long Sequences☆389Updated last month
- 🔥 A minimal training framework for scaling FLA models☆228Updated this week
- Tina: Tiny Reasoning Models via LoRA☆275Updated last week
- [ICML'24] Data and code for our paper "Training-Free Long-Context Scaling of Large Language Models"☆435Updated 10 months ago