Best of both, Structured and Unstructured Sparsity in Neural Networks
Schulte-Schüren C, Wagner S, Runge A, Bariamis D, Hammer B, Yoneki E, Nardi L (2023)
In: Proceedings of the 3rd Workshop on Machine Learning and Systems. New York, NY, USA: ACM: 104-108.
Konferenzbeitrag
| Veröffentlicht | Englisch
Download
Es wurden keine Dateien hochgeladen. Nur Publikationsnachweis!
Autor*in
Schulte-Schüren, ChristopherUniBi;
Wagner, Sven;
Runge, Armin;
Bariamis, Dimitrios;
Hammer, BarbaraUniBi 
;
Yoneki, Eiko;
Nardi, Luigi
;
Yoneki, Eiko;
Nardi, LuigiEinrichtung
Abstract / Bemerkung
Besides quantization, pruning has shown to be one of the most effective methods to reduce the inference time and required energy of Deep Neural Networks (DNNs). In this work, we propose a sparsity definition that reflects the number of saved operations by pruned parameters to guide the pruning process in order to save as many operations as possible. Based on this, we show the importance of the baseline model's size and quantify the overhead of unstructured sparsity for a commercial-of-the-shelf AI Hardware Accelerator (HWA) in terms of latency reductions. Furthermore, we show that a combination of both structured and unstructured sparsity can mitigate this effect.
Erscheinungsjahr
2023
Titel des Konferenzbandes
Proceedings of the 3rd Workshop on Machine Learning and Systems
Seite(n)
104-108
Konferenz
EuroMLSys '23: 3rd Workshop on Machine Learning and Systems
Konferenzort
Rome Italy
Konferenzdatum
2023-05-08 – 2023-05-08
ISBN
9798400700842
Page URI
https://pub.uni-bielefeld.de/record/2984048
Zitieren
Schulte-Schüren C, Wagner S, Runge A, et al. Best of both, Structured and Unstructured Sparsity in Neural Networks. In: Proceedings of the 3rd Workshop on Machine Learning and Systems. New York, NY, USA: ACM; 2023: 104-108.
Schulte-Schüren, C., Wagner, S., Runge, A., Bariamis, D., Hammer, B., Yoneki, E., & Nardi, L. (2023). Best of both, Structured and Unstructured Sparsity in Neural Networks. Proceedings of the 3rd Workshop on Machine Learning and Systems, 104-108. New York, NY, USA: ACM. https://doi.org/10.1145/3578356.3592583
Schulte-Schüren, Christopher, Wagner, Sven, Runge, Armin, Bariamis, Dimitrios, Hammer, Barbara, Yoneki, Eiko, and Nardi, Luigi. 2023. “Best of both, Structured and Unstructured Sparsity in Neural Networks”. In Proceedings of the 3rd Workshop on Machine Learning and Systems, 104-108. New York, NY, USA: ACM.
Schulte-Schüren, C., Wagner, S., Runge, A., Bariamis, D., Hammer, B., Yoneki, E., and Nardi, L. (2023). “Best of both, Structured and Unstructured Sparsity in Neural Networks” in Proceedings of the 3rd Workshop on Machine Learning and Systems (New York, NY, USA: ACM), 104-108.
Schulte-Schüren, C., et al., 2023. Best of both, Structured and Unstructured Sparsity in Neural Networks. In Proceedings of the 3rd Workshop on Machine Learning and Systems. New York, NY, USA: ACM, pp. 104-108.
C. Schulte-Schüren, et al., “Best of both, Structured and Unstructured Sparsity in Neural Networks”, Proceedings of the 3rd Workshop on Machine Learning and Systems, New York, NY, USA: ACM, 2023, pp.104-108.
Schulte-Schüren, C., Wagner, S., Runge, A., Bariamis, D., Hammer, B., Yoneki, E., Nardi, L.: Best of both, Structured and Unstructured Sparsity in Neural Networks. Proceedings of the 3rd Workshop on Machine Learning and Systems. p. 104-108. ACM, New York, NY, USA (2023).
Schulte-Schüren, Christopher, Wagner, Sven, Runge, Armin, Bariamis, Dimitrios, Hammer, Barbara, Yoneki, Eiko, and Nardi, Luigi. “Best of both, Structured and Unstructured Sparsity in Neural Networks”. Proceedings of the 3rd Workshop on Machine Learning and Systems. New York, NY, USA: ACM, 2023. 104-108.
Google Scholar
ISBN Suche