Collection by
Mar 16
Mar 16
Collections
Discover the best community collections!
Collections including paper arxiv:2106.04426
Collection by
Mar 15
Mar 15
Collection by
Mar 16
Mar 16
-
SELF: Language-Driven Self-Evolution for Large Language Model
Paper • 2310.00533 • Published • 2 -
GrowLength: Accelerating LLMs Pretraining by Progressively Growing Training Length
Paper • 2310.00576 • Published • 2 -
A Pretrainer's Guide to Training Data: Measuring the Effects of Data Age, Domain Coverage, Quality, & Toxicity
Paper • 2305.13169 • Published • 3 -
Transformers Can Achieve Length Generalization But Not Robustly
Paper • 2402.09371 • Published • 12
Collection by
28 days ago
28 days ago
-
Unleashing the Power of Pre-trained Language Models for Offline Reinforcement Learning
Paper • 2310.20587 • Published • 15 -
JoMA: Demystifying Multilayer Transformers via JOint Dynamics of MLP and Attention
Paper • 2310.00535 • Published • 2 -
Does Circuit Analysis Interpretability Scale? Evidence from Multiple Choice Capabilities in Chinchilla
Paper • 2307.09458 • Published • 9 -
The Impact of Depth and Width on Transformer Language Model Generalization
Paper • 2310.19956 • Published • 9
Collection by
Mar 15
Mar 15
-
Turn Waste into Worth: Rectifying Top-k Router of MoE
Paper • 2402.12399 • Published • 2 -
CompeteSMoE -- Effective Training of Sparse Mixture of Experts via Competition
Paper • 2402.02526 • Published • 3 -
Buffer Overflow in Mixture of Experts
Paper • 2402.05526 • Published • 8 -
OpenMoE: An Early Effort on Open Mixture-of-Experts Language Models
Paper • 2402.01739 • Published • 26
Collection by
Apr 26
Apr 26
-
Non-asymptotic oracle inequalities for the Lasso in high-dimensional mixture of experts
Paper • 2009.10622 • Published • 1 -
MoE-LLaVA: Mixture of Experts for Large Vision-Language Models
Paper • 2401.15947 • Published • 46 -
MoE-Mamba: Efficient Selective State Space Models with Mixture of Experts
Paper • 2401.04081 • Published • 68 -
MoE-Infinity: Activation-Aware Expert Offloading for Efficient MoE Serving
Paper • 2401.14361 • Published • 2
Collection by
6 days ago
6 days ago
-
Measuring the Effects of Data Parallelism on Neural Network Training
Paper • 1811.03600 • Published • 2 -
Adafactor: Adaptive Learning Rates with Sublinear Memory Cost
Paper • 1804.04235 • Published • 2 -
EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks
Paper • 1905.11946 • Published • 3 -
Yi: Open Foundation Models by 01.AI
Paper • 2403.04652 • Published • 59
Collection by
Feb 27
Feb 27
-
Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer
Paper • 1701.06538 • Published • 4 -
Sparse Networks from Scratch: Faster Training without Losing Performance
Paper • 1907.04840 • Published • 3 -
ZeRO: Memory Optimizations Toward Training Trillion Parameter Models
Paper • 1910.02054 • Published • 3 -
A Mixture of h-1 Heads is Better than h Heads
Paper • 2005.06537 • Published • 2
Collection by
4 days ago
4 days ago
-
Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer
Paper • 1701.06538 • Published • 4 -
Sparse Networks from Scratch: Faster Training without Losing Performance
Paper • 1907.04840 • Published • 3 -
ZeRO: Memory Optimizations Toward Training Trillion Parameter Models
Paper • 1910.02054 • Published • 3 -
A Mixture of h-1 Heads is Better than h Heads
Paper • 2005.06537 • Published • 2
Collection by
4 days ago
4 days ago
-
QMoE: Practical Sub-1-Bit Compression of Trillion-Parameter Models
Paper • 2310.16795 • Published • 26 -
Pre-gated MoE: An Algorithm-System Co-Design for Fast and Scalable Mixture-of-Expert Inference
Paper • 2308.12066 • Published • 4 -
Towards MoE Deployment: Mitigating Inefficiencies in Mixture-of-Expert (MoE) Inference
Paper • 2303.06182 • Published • 1 -
EvoMoE: An Evolutional Mixture-of-Experts Training Framework via Dense-To-Sparse Gate
Paper • 2112.14397 • Published • 1