Skip to yearly menu bar Skip to main content


Contributed Talk & Poster
in
Workshop: Workshop on Advancing Neural Network Training (WANT): Computational Efficiency, Scalability, and Resource Optimization

MatFormer: Nested Transformer for Elastic Inference

Fnu Devvrit · Sneha Kudugunta · Aditya Kusupati · Tim Dettmers · Kaifeng Chen · Inderjit Dhillon · Yulia Tsvetkov · Hannaneh Hajishirzi · Sham Kakade · Ali Farhadi · Prateek Jain

[ ] [ Project Page ]
Sat 16 Dec 9:37 a.m. PST — 9:42 a.m. PST

Abstract:

Transformer models are deployed in a wide range of settings, from multi-accelerator clusters to standalone mobile phones. The diverse inference constraints in these scenarios necessitate practitioners to train foundation models such as PaLM 2, Llama, & ViTs as a series of models of varying sizes. Due to significant training costs, only a select few model sizes are trained and supported, limiting more fine-grained control over relevant tradeoffs, including latency, cost, and accuracy. This work introduces MatFormer, a nested Transformer architecture designed to offer elasticity in a variety of deployment constraints. Each Feed Forward Network (FFN) block of a MatFormer model is jointly optimized with a few nested smaller FFN blocks. This training procedure allows for the Mix'n'Match of model granularities across layers -- i.e., a trained universal MatFormer model enables extraction of hundreds of accurate smaller models, which were never explicitly optimized. We empirically demonstrate MatFormer's effectiveness across different model classes (decoders & encoders), modalities (language & vision), and scales (up to 2.6B parameters). We find that a 2.6B decoder-only MatFormer language model (MatLM) allows us to extract smaller models spanning from 1.5B to 2.6B, each exhibiting comparable validation loss and one-shot downstream evaluations to their independently trained counterparts. Furthermore, we observe that smaller encoders extracted from a universal MatFormer-based ViT (MatViT) encoder preserve the metric-space structure for adaptive large-scale retrieval. Finally, we showcase that speculative decoding with the accurate and consistent submodels extracted from MatFormer can further reduce inference latency.

Chat is not available.