Global Average Pooling
التجميع المتوسط العام
عملية تحسب متوسط كل خريطة سمات مكانياً لتنتج قيمة واحدة لكل قناة، بدلاً من طبقة كاملة الاتصال.
An operation that averages each feature map spatially to a single value per channel, replacing fully connected layers.
Also translated asتجميع المتوسط الشامل، تجميع متوسط عام
First appears in this corpus in: Going Deeper with Convolutions (2014)
Appears in these papers
- ALIGN: Scaling Up Visual and Vision-Language Representation Learning with Noisy Text Supervision2021in the sky ✦
- A ConvNet for the 2020s2022in the sky ✦
- EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks2019in the sky ✦
- EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks2019in the sky ✦
- Going Deeper with Convolutions2014in the sky ✦
- Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization2017in the sky ✦
- MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications2017in the sky ✦
- Deep Residual Learning for Image Recognition2015in the sky ✦
- Deep Residual Learning for Image Recognition2015in the sky ✦
- Squeeze-and-Excitation Networks2018in the sky ✦
- Squeeze-and-Excitation Networks2018in the sky ✦