SqueezeNet is a convolutional neural network that employs design strategies to reduce the number of parameters, notably with the use of fire modules that "squeeze" parameters using 1x1 convolutions.
Source: SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model sizePaper | Code | Results | Date | Stars |
---|
Task | Papers | Share |
---|---|---|
General Classification | 11 | 8.73% |
Object Detection | 10 | 7.94% |
Image Classification | 8 | 6.35% |
Classification | 8 | 6.35% |
Deep Learning | 5 | 3.97% |
Face Recognition | 4 | 3.17% |
Quantization | 4 | 3.17% |
Object | 4 | 3.17% |
Network Pruning | 3 | 2.38% |