Logo image
Se connecter
Deep Nibble: A 4-bit Number Format for Efficient DNN Training and Inference in FPGA
Acte de colloque

Deep Nibble: A 4-bit Number Format for Efficient DNN Training and Inference in FPGA

Luiz F. H. Duarte, George B. Nardes, Wesley Grignani, Douglas R. Melo, Cesar A. Zeferino et IEEE
2024 37th SBC/SBMicro/IEEE Symposium on Integrated Circuits and Systems Design (SBCCI), pp.1-5
02/09/2024

Résumé

Accuracy Artificial neural networks Deep learning FPGA Hardware Acceleration Integrated circuit modeling Logarithmic Quantization Memory management Performance evaluation Quantization (signal) Stochastic processes Stochastic Rounding Throughput Training

Indicateurs

1 Consultations de la notice

Détails

Logo image