Szczegóły publikacji
Opis bibliograficzny
Power-of-two quantization for low bitwidth and hardware compliant neural networks / Dominika PRZEWŁOCKA-RUS, Syed Shakib Sarwar, H. Ekin Sumbul, Yuecheng Li, Barbara De Salvo // W: tinyML research symposium 2022 [Dokument elektroniczny] : 28 march 2022, San Jose. — Wersja do Windows. — Dane tekstowe. — [USA : tinyML Foundation], cop. 2022. — S. [1–7]. — Wymagania systemowe: Adobe Reader. — Tryb dostępu: https://cms.tinyml.org/wp-content/uploads/talks2022/2203.0502... [2022-06-23]. — Bibliogr. s. [7], Abstr. — Na stronie dod. link prezentacji: https://www.youtube.com/watch?v=doYRnyoGSvc [2022-06-23]
Autorzy (5)
- AGHPrzewłocka-Rus Dominika
- Sarwar Syed Shakib
- Ekin Sumbul H.
- Li Yuecheng
- de Salvo Barbara
Słowa kluczowe
Dane bibliometryczne
| ID BaDAP | 140733 |
|---|---|
| Data dodania do BaDAP | 2022-07-01 |
| Rok publikacji | 2022 |
| Typ publikacji | materiały konferencyjne (aut.) |
| Otwarty dostęp |  |
Abstract
Deploying Deep Neural Networks in low-power embedded devices for real time-constrained applications requires optimization of memory and computational complexity of the networks, usually by quantizing the weights. Most of the existing works employ linear quantization which causes considerable degradation in accuracy for weight bit widths lower than 8. Since the distribution of weights is usually non-uniform (with most weights concentrated around zero), other methods, such as logarithmic quantization, are more suitable as they are able to preserve the shape of the weight distribution more precise. Moreover, using base-2 logarithmic representation allows optimizing the multiplication by replacing it with bit shifting. In this paper, we explore non-linear quantization techniques for exploiting lower bit precision and identify favorable hardware implementation options. We developed the Quantization Aware Training (QAT) algorithm that allowed training of low bit width Power-of-Two (PoT) networks and achieved accuracies on par with state-of-the-art floating point models for different tasks. We explored PoT weight encoding techniques and investigated hardware designs of MAC units for three different quantization schemes - uniform, PoT and Additive-PoT (APoT) - to show the increased efficiency when using the proposed approach. Eventually, the experiments showed that for low bit width precision, non-uniform quantization performs better than uniform, and at the same time, PoT quantization vastly reduces the computational complexity of the neural network.
