Low-Rank Adapters Meet Neural Architecture Search for LLM Compression
Paper
•
2501.16372
•
Published
•
9
inference optimization
-
-
TAID: Temporally Adaptive Interpolated Distillation for Efficient Knowledge Transfer in Language Models
Paper • 2501.16937 • Published • 6 -
Matryoshka Quantization
Paper • 2502.06786 • Published • 30 -
Identifying Sensitive Weights via Post-quantization Integral
Paper • 2503.01901 • Published • 8 -
EAGLE-3: Scaling up Inference Acceleration of Large Language Models via Training-Time Test
Paper • 2503.01840 • Published • 5 -
SEAP: Training-free Sparse Expert Activation Pruning Unlock the Brainpower of Large Language Models
Paper • 2503.07605 • Published • 69 -
DistiLLM-2: A Contrastive Approach Boosts the Distillation of LLMs
Paper • 2503.07067 • Published • 32 -
Efficient Distillation of Classifier-Free Guidance using Adapters
Paper • 2503.07274 • Published • 4 -
RayFlow: Instance-Aware Diffusion Acceleration via Adaptive Flow Trajectories
Paper • 2503.07699 • Published • 5 -
QuartDepth: Post-Training Quantization for Real-Time Depth Estimation on the Edge
Paper • 2503.16709 • Published -
xKV: Cross-Layer SVD for KV-Cache Compression
Paper • 2503.18893 • Published • 4 -
LogQuant: Log-Distributed 2-Bit Quantization of KV Cache with Superior Accuracy Preservation
Paper • 2503.19950 • Published • 11 -
Optimal Stepsize for Diffusion Sampling
Paper • 2503.21774 • Published • 13 -
AdaptiVocab: Enhancing LLM Efficiency in Focused Domains through Lightweight Vocabulary Adaptation
Paper • 2503.19693 • Published • 75 -
Harnessing the Reasoning Economy: A Survey of Efficient Reasoning for Large Language Models
Paper • 2503.24377 • Published • 17 -
ShortV: Efficient Multimodal Large Language Models by Freezing Visual Tokens in Ineffective Layers
Paper • 2504.00502 • Published • 22 -
Distillation and Refinement of Reasoning in Small Language Models for Document Re-ranking
Paper • 2504.03947 • Published • 4 -
HybriMoE: Hybrid CPU-GPU Scheduling and Cache Management for Efficient MoE Inference
Paper • 2504.05897 • Published • 18 -
Accelerate Parallelizable Reasoning via Parallel Decoding within One Sequence
Paper • 2503.20533 • Published • 12 -
SpecReason: Fast and Accurate Inference-Time Compute via Speculative Reasoning
Paper • 2504.07891 • Published • 5 -
PRIMA.CPP: Speeding Up 70B-Scale LLM Inference on Low-Resource Everyday Home Clusters
Paper • 2504.08791 • Published • 132 -
Adaptive Computation Pruning for the Forgetting Transformer
Paper • 2504.06949 • Published • 3 -
70% Size, 100% Accuracy: Lossless LLM Compression for Efficient GPU Inference via Dynamic-Length Float
Paper • 2504.11651 • Published • 28 -
Learning Adaptive Parallel Reasoning with Language Models
Paper • 2504.15466 • Published • 42 -
MMInference: Accelerating Pre-filling for Long-Context VLMs via Modality-Aware Permutation Sparse Attention
Paper • 2504.16083 • Published • 9 -
Taming the Titans: A Survey of Efficient LLM Inference Serving
Paper • 2504.19720 • Published • 10 -
A Survey on Inference Engines for Large Language Models: Perspectives on Optimization and Efficiency
Paper • 2505.01658 • Published • 35 -
ReplaceMe: Network Simplification via Layer Pruning and Linear Transformations
Paper • 2505.02819 • Published • 24 -
RADLADS: Rapid Attention Distillation to Linear Attention Decoders at Scale
Paper • 2505.03005 • Published • 31 -
An Empirical Study of Qwen3 Quantization
Paper • 2505.02214 • Published • 23 -
A Token is Worth over 1,000 Tokens: Efficient Knowledge Distillation through Low-Rank Clone
Paper • 2505.12781 • Published • 2 -
QVGen: Pushing the Limit of Quantized Video Generative Models
Paper • 2505.11497 • Published • 4 -
Scaling Law for Quantization-Aware Training
Paper • 2505.14302 • Published • 72 -
WINA: Weight Informed Neuron Activation for Accelerating Large Language Model Inference
Paper • 2505.19427 • Published • 9 -
KVzip: Query-Agnostic KV Cache Compression with Context Reconstruction
Paper • 2505.23416 • Published • 9 -
CLaSp: In-Context Layer Skip for Self-Speculative Decoding
Paper • 2505.24196 • Published • 13 -
Accelerating Diffusion LLMs via Adaptive Parallel Decoding
Paper • 2506.00413 • Published • 6 -
DLP: Dynamic Layerwise Pruning in Large Language Models
Paper • 2505.23807 • Published • 4 -
POSS: Position Specialist Generates Better Draft for Speculative Decoding
Paper • 2506.03566 • Published • 6