Training with Quantization Noise for Extreme Model Compression - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2021

Training with Quantization Noise for Extreme Model Compression

Abstract

We tackle the problem of producing compact models, maximizing their accuracy for a given model size. A standard solution is to train networks with Quantization Aware Training, where the weights are quantized during training and the gradients approximated with the Straight-Through Estimator. In this paper, we extend this approach to work with extreme compression methods where the approximations introduced by STE are severe. Our proposal is to only quantize a different random subset of weights during each forward, allowing for unbiased gradients to flow through the other weights. Controlling the amount of noise and its form allows for extreme compression rates while maintaining the performance of the original model. As a result we establish new state-of-the-art compromises between accuracy and model size both in natural language processing and image classification. For example, applying our method to state-of-the-art Transformer and ConvNet architectures, we can achieve 82.5% accuracy on MNLI by compressing RoBERTa to 14 MB and 80.0% top-1 accuracy on ImageNet by compressing an EfficientNet-B3 to 3.3 MB.

Dates and versions

hal-03136442 , version 1 (09-02-2021)

Identifiers

Cite

Pierre Stock, Angela Fan, Benjamin Graham, Edouard Grave, Rémi Gribonval, et al.. Training with Quantization Noise for Extreme Model Compression. ICLR 2021 - International Conference on Learning Representations, May 2021, Vienna, Austria. ⟨hal-03136442⟩
179 View
0 Download

Altmetric

Share

Gmail Mastodon Facebook X LinkedIn More