Medusa: Simple LLM Inference Acceleration Framework with Multiple Decoding Heads
Paper
•
2401.10774
•
Published
•
50
Efficient LLM
-
-
APAR: LLMs Can Do Auto-Parallel Auto-Regressive Decoding
Paper • 2401.06761 • Published • 1 -
Infinite-LLM: Efficient LLM Service for Long Context with DistAttention and Distributed KVCache
Paper • 2401.02669 • Published • 10 -
MambaByte: Token-free Selective State Space Model
Paper • 2401.13660 • Published • 47 -
EAGLE: Speculative Sampling Requires Rethinking Feature Uncertainty
Paper • 2401.15077 • Published • 17 -
Small LLMs Are Weak Tool Learners: A Multi-LLM Agent
Paper • 2401.07324 • Published • 3 -
Hierarchical State Space Models for Continuous Sequence-to-Sequence Modeling
Paper • 2402.10211 • Published • 8 -
LongRoPE: Extending LLM Context Window Beyond 2 Million Tokens
Paper • 2402.13753 • Published • 104 -
Ouroboros: Speculative Decoding with Large Model Enhanced Drafting
Paper • 2402.13720 • Published • 4 -
Neural Network Diffusion
Paper • 2402.13144 • Published • 92 -
LongAlign: A Recipe for Long Context Alignment of Large Language Models
Paper • 2401.18058 • Published • 21 -
LongHeads: Multi-Head Attention is Secretly a Long Context Processor
Paper • 2402.10685 • Published • 1 -
LLM Maybe LongLM: Self-Extend LLM Context Window Without Tuning
Paper • 2401.01325 • Published • 24 -
E^2-LLM: Efficient and Extreme Length Extension of Large Language Models
Paper • 2401.06951 • Published • 23 -
InfLLM: Unveiling the Intrinsic Capacity of LLMs for Understanding Extremely Long Sequences with Training-Free Memory
Paper • 2402.04617 • Published • 4 -
Speculative Streaming: Fast LLM Inference without Auxiliary Models
Paper • 2402.11131 • Published • 41 -
Towards Optimal Learning of Language Models
Paper • 2402.17759 • Published • 16 -
When Scaling Meets LLM Finetuning: The Effect of Data, Model and Finetuning Method
Paper • 2402.17193 • Published • 23 -
GaLore: Memory-Efficient LLM Training by Gradient Low-Rank Projection
Paper • 2403.03507 • Published • 172 -
DenseMamba: State Space Models with Dense Hidden Connection for Efficient Large Language Models
Paper • 2403.00818 • Published • 13 -
LongNet: Scaling Transformers to 1,000,000,000 Tokens
Paper • 2307.02486 • Published • 79 -
Recurrent Drafter for Fast Speculative Decoding in Large Language Models
Paper • 2403.09919 • Published • 19 -
DiJiang: Efficient Large Language Models through Compact Kernelization
Paper • 2403.19928 • Published • 9 -
ReFT: Representation Finetuning for Language Models
Paper • 2404.03592 • Published • 70 -
Rethinking Optimization and Architecture for Tiny Language Models
Paper • 2402.02791 • Published • 12 -
MobileLLM: Optimizing Sub-billion Parameter Language Models for On-Device Use Cases
Paper • 2402.14905 • Published • 80 -
Mixture-of-Depths: Dynamically allocating compute in transformer-based language models
Paper • 2404.02258 • Published • 99 -
Pre-training Small Base LMs with Fewer Tokens
Paper • 2404.08634 • Published • 32 -
Scaling (Down) CLIP: A Comprehensive Analysis of Data, Architecture, and Training Strategies
Paper • 2404.08197 • Published • 26 -
Rho-1: Not All Tokens Are What You Need
Paper • 2404.07965 • Published • 79 -
Megalodon: Efficient LLM Pretraining and Inference with Unlimited Context Length
Paper • 2404.08801 • Published • 61 -
Skill-it! A Data-Driven Skills Framework for Understanding and Training Language Models
Paper • 2307.14430 • Published • 3 -
Compression Represents Intelligence Linearly
Paper • 2404.09937 • Published • 27 -
Self-Play Fine-Tuning Converts Weak Language Models to Strong Language Models
Paper • 2401.01335 • Published • 61