『您的瀏覽器不支援JavaScript功能,若網頁功能無法正常使用時,請開啟瀏覽器JavaScript狀態』

跳到主要內容區塊

工業技術研究院

:::

技術名稱: 硬體架構感知深度學習模型量化技術

技術簡介

提供DNN於重訓練過程考慮量化處理對精準度之影響,並透過data driven方式來適應晶片硬體運算架構,保持DNN模型精確度

Abstract

Support the DNN retraining to consider the impact of all quantization processing on accuracy, and adapt to the hardware computing architecture through the data driven method for keep the accuracy of DNN model.

技術規格

相較原始模型(FP32)可提升推論速度達1.7倍,精確度損失小於1.8%

Technical Specification

Compared with the original model (FP32), the inference speed can be increased by 1.7 times, and the accuracy loss is less than 1.8%

技術特色

相較現有量化技術僅可調整act./weight bit,本案可再支援partial sum bit縮減,使量化處理過程得以考慮硬體運算架構。

應用範圍

智慧影像監控、行車物件偵測等需達即時效能之需求。

接受技術者具備基礎建議(設備)

1.具備windows或linux作業系統的電腦 2.需安裝python與tensorflow

接受技術者具備基礎建議(專業)

1.具備深度學習基礎 2.具備DNN模型優化觀念

技術分類 資訊

聯絡資訊

聯絡人:徐福裕 技術推廣組

電話:+886-3-5914566 或 Email:ben.hsu@itri.org.tw

客服專線:+886-800-45-8899

傳真:+886-3-5820240

[{"text":"企業網","weight":13.0},{"text":"材化所","weight":11.5},{"text":"機械所","weight":10.0},{"text":"綠能所","weight":9.4},{"text":"生醫所","weight":8.0},{"text":"半導體","weight":6.2},{"text":"南分院","weight":5.0},{"text":"太陽能","weight":5.0},{"text":"課程","weight":5.0},{"text":"遠紅外線","weight":5.0},{"text":"雷射","weight":4.0},{"text":"LED","weight":4.0},{"text":"LED可見光","weight":3.0},{"text":"5G","weight":3.0},{"text":"工研人","weight":3.0},{"text":"電光所","weight":3.0},{"text":"綠能與環境研究所","weight":3.0},{"text":"機械","weight":3.0},{"text":"資通所","weight":2.0},{"text":"面板","weight":2.0},{"text":"文字轉語音","weight":2.0},{"text":"實習","weight":2.0},{"text":"無人機","weight":2.0},{"text":"生醫","weight":2.0},{"text":"3D","weight":2.0},{"text":"v2x","weight":2.0},{"text":"員工","weight":2.0},{"text":"地圖","weight":2.0},{"text":"太陽光電","weight":2.0},{"text":"材料與化工研究所","weight":1.0}]