Depthwise Convolution
الالتفاف بالعمق
عملية التفافية تطبّق مرشّحاً مكانياً واحداً مستقلاً على كل قناة من قنوات الدخل. لا تمزج بين القنوات بل ترشّح كل قناة على حدة، مما يجعلها رخيصة حوسبياً لكنها تحتاج لخطوة مزج لاحقة.
A convolution operation that applies a single independent spatial filter to each input channel. It does not mix channels — it filters each channel separately — making it computationally cheap but requiring a subsequent mixing step.
Also translated asالالتفاف العمقي، الالتفاف لكل قناة
First appears in this corpus in: MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications (2017)
Appears in these papers
- A ConvNet for the 2020s2022in the sky ✦
- A ConvNet for the 2020s2022in the sky ✦
- MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications2017in the sky ✦
- Efficiently Modeling Long Sequences with Structured State Spaces2022in the sky ✦
- SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers2021in the sky ✦
- SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers2021in the sky ✦