So sánh phương pháp
Xem các phương pháp đã chọn cạnh nhau; những hàng khác biệt được làm nổi bật.
| MobileNet: Mạng nơ-ron tích chập hiệu quả cho thị giác di động× | Chưng cất tri thức× | |
|---|---|---|
| Lĩnh vực | Học sâu | Học sâu |
| Họ | Machine learning | Machine learning |
| Năm ra đời≠ | 2017 | 2015 |
| Người khởi xướng≠ | Andrew Howard et al. (Google) | Hinton, G., Vinyals, O. & Dean, J. |
| Loại≠ | Lightweight CNN architecture | Neural network compression (teacher–student) |
| Công trình gốc≠ | Howard, A. G., et al. (2017). MobileNets: Efficient convolutional neural networks for mobile vision applications. arXiv preprint. link ↗ | Hinton, G., Vinyals, O. & Dean, J. (2015). Distilling the Knowledge in a Neural Network. NeurIPS Deep Learning Workshop. link ↗ |
| Tên gọi khác | MobileNets, Depthwise Separable CNN, Efficient Mobile Vision Network, Mobil Evrişimli Sinir Ağı | Bilgi Damıtma (Knowledge Distillation), bilgi damıtma, teacher-student distillation, model distillation |
| Liên quan≠ | 2 | 5 |
| Tóm tắt≠ | MobileNet is a family of lightweight convolutional neural network architectures introduced by Howard et al. at Google in 2017. It is designed to run image classification, object detection, and other vision tasks directly on mobile devices and embedded systems with limited computational budgets. By replacing standard convolutions with depthwise separable convolutions and exposing two global hyperparameters, MobileNet dramatically reduces multiply-add operations and model size while retaining competitive accuracy. | Knowledge Distillation is a model-compression technique, introduced by Geoffrey Hinton and colleagues in 2015, that trains a small student model using the soft-label outputs of a large teacher model. Distilled models such as DistilBERT and TinyBERT reach roughly 97% of the larger model's performance while running far faster. |
| ScholarGateBộ dữ liệu ↗ |
|
|