Collections

134

Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer

Paper • 1701.06538 • Published Jan 23, 2017 • 4
Sparse Networks from Scratch: Faster Training without Losing Performance

Paper • 1907.04840 • Published Jul 10, 2019 • 3
ZeRO: Memory Optimizations Toward Training Trillion Parameter Models

Paper • 1910.02054 • Published Oct 4, 2019 • 4
A Mixture of h-1 Heads is Better than h Heads

Paper • 2005.06537 • Published May 13, 2020 • 2

2

Simple linear attention language models balance the recall-throughput tradeoff

Paper • 2402.18668 • Published Feb 28 • 18
Linear Transformers with Learnable Kernel Functions are Better In-Context Models

Paper • 2402.10644 • Published Feb 16 • 78
Repeat After Me: Transformers are Better than State Space Models at Copying

Paper • 2402.01032 • Published Feb 1 • 22
Zoology: Measuring and Improving Recall in Efficient Language Models

Paper • 2312.04927 • Published Dec 8, 2023 • 2

1

-

Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity

Paper • 2101.03961 • Published Jan 11, 2021 • 14

Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer

Sparse Networks from Scratch: Faster Training without Losing Performance

ZeRO: Memory Optimizations Toward Training Trillion Parameter Models

A Mixture of h-1 Heads is Better than h Heads

Simple linear attention language models balance the recall-throughput tradeoff

Linear Transformers with Learnable Kernel Functions are Better In-Context Models

Repeat After Me: Transformers are Better than State Space Models at Copying

Zoology: Measuring and Improving Recall in Efficient Language Models

SMOTE: Synthetic Minority Over-sampling Technique

Scikit-learn: Machine Learning in Python

Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translation

Distributed Representations of Sentences and Documents

MoE-LLaVA: Mixture of Experts for Large Vision-Language Models

DeepSeekMoE: Towards Ultimate Expert Specialization in Mixture-of-Experts Language Models

SwitchHead: Accelerating Transformers with Mixture-of-Experts Attention

Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity

Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity

Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer

Sparse Networks from Scratch: Faster Training without Losing Performance

ZeRO: Memory Optimizations Toward Training Trillion Parameter Models

A Mixture of h-1 Heads is Better than h Heads

Attention Is All You Need

BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding

RoBERTa: A Robustly Optimized BERT Pretraining Approach

DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter

QMoE: Practical Sub-1-Bit Compression of Trillion-Parameter Models

Pre-gated MoE: An Algorithm-System Co-Design for Fast and Scalable Mixture-of-Expert Inference

Towards MoE Deployment: Mitigating Inefficiencies in Mixture-of-Expert (MoE) Inference

EvoMoE: An Evolutional Mixture-of-Experts Training Framework via Dense-To-Sparse Gate

QMoE: Practical Sub-1-Bit Compression of Trillion-Parameter Models

Pre-gated MoE: An Algorithm-System Co-Design for Fast and Scalable Mixture-of-Expert Inference

Towards MoE Deployment: Mitigating Inefficiencies in Mixture-of-Expert (MoE) Inference

EvoMoE: An Evolutional Mixture-of-Experts Training Framework via Dense-To-Sparse Gate

The Era of 1-bit LLMs: All Large Language Models are in 1.58 Bits

Mixtral of Experts

Mistral 7B

Don't Make Your LLM an Evaluation Benchmark Cheater