Goog lenet
Download as PPTX, PDF0 likes17 views
This document discusses convolutional neural networks and provides details on GoogLeNet, an architecture that utilized deeper convolutions to achieve state-of-the-art results in image classification. It covers related work on CNNs and the Network in Network model. The architectural details section explains innovations in GoogLeNet, while the training methodology discusses hyperparameters like the optimizer, momentum, learning rate, and data splits. Results showed a 4% decrease in error rate using stochastic gradient descent.
1 of 10
Download to read offline
Goog lenet
- 1. Going Deeper with Convolutions Mar 27, 2021 HeeDae Kwon
- 2. Contents • 1. Introdution • 2. Realted Work • 3. Architectural Details • 4. GoogLeNet • 5. Training Methodology • 6. Results
- 3. Introduction
- 4. Realted Work • CNN • NIN(Network In Network
- 6. GoogLeNet
- 7. GoogLeNet
- 8. Training Methodology optimizer momentum Learning rate Test images Validation images Training images Error rate SGD 0.9 4% decrease Every 8 epochs 100K 50K 1.2M Top-1 Top-5
- 9. Results
- 10. 감사합니다