Public Timelines

For educational institutions For teachers For students Cabinet

For educational institutions For teachers For students Open cabinet

Create a timeline

Public timelines

About & Feedback Соглашение Приватность FAQ

Cabinet Get premium

Donate

8 март 2019 г. - ChengGong: $\mu$L2Q: An Ultra-Low Loss Quantization Method for DNN

Описание:

本文实现了多位宽的量化方案。相比于二值三值或者定点数量化等方案，本文首次从权值分布角度进行量化，实现了多位宽的高精度量化。实验表明再相同条件下，uL2Q可以实现更低精度下降。

Добавлено на ленту времени:

Quantization timeline

29 мар 2020

1

0

810

Дата:

8 март 2019 г.

Сейчас

~ 6 гг и 7 мес назад

About & Feedback Соглашение Приватность FAQ

Cabinet Get premium

Donate

The service accepts bank transfer (ACH, Wire) or cards (Visa, MasterCard, etc). Processed by Stripe.

Secured with SSL