collaborators

5 papers

cs.LG2026

The Recurrent Transformer: Greater Effective Depth and Efficient Decoding

Costin-Andrei Oncescu, Depen Morwani, Samy Jelassi +3

Transformers process tokens in parallel but are temporally shallow: at position , each layer attends to key-value pairs computed based on the previous layer, yielding a depth ca…

cs.LG2025

Flash Inference: Near Linear Time Inference for Long Convolution Sequence Models and Beyond

Costin-Andrei Oncescu, Sanket Purandare, Stratos Idreos +1

While transformers have been at the core of most recent advancements in sequence generative models, their computational cost remains quadratic in sequence length. Several subquadra…

cs.LG2025

Opportunistic Expert Activation: Batch-Aware Expert Routing for Faster Decode Without Retraining

Costin-Andrei Oncescu, Qingyang Wu, Wai Tong Chung +5

An increasing number of LLMs employ Mixture-of-Experts (MoE) architectures where the feed-forward layer is replaced by a pool of experts and each token only activates a small subse…

cs.LG2025

Seesaw: Accelerating Training by Balancing Learning Rate and Batch Size Scheduling

Alexandru Meterez, Depen Morwani, Jingfeng Wu +3

Increasing the batch size during training -- a ''batch ramp'' -- is a promising strategy to accelerate large language model pretraining. While for SGD, doubling the batch size can…

cs.LG2025

A Simplified Analysis of SGD for Linear Regression with Weight Averaging

Alexandru Meterez, Depen Morwani, Costin-Andrei Oncescu +3

Theoretically understanding stochastic gradient descent (SGD) in overparameterized models has led to the development of several optimization algorithms that are widely used in prac…