[GoogleML] Adam Optimizer
2023. 9. 20. 22:23ใArtificialIntelligence/2023GoogleMLBootcamp
Gradient Descent with Momentum
RMSprop
์ด๊ฒ์ด ์ด๋ป๊ฒ ๋์ํ ์ ์๋๊ฐ?
์ธ๋ก ๋ฐฉํฅ์ผ๋ก๋ ์ ๊ฒ,
๊ฐ๋ก ๋ฐฉํฅ์ผ๋ก๋ ๋ง์ด update ๋์ด์ผ ํ๋ค! (๋ง์ด ๊ฐ์ผ ํ๋ค)
Adam Optimization Algorithm
Learning Rate Decay
์ด๋ฐ์๋ ํฌ๊ฒ ํ ํ
๋์ค์๋ learning rate ํฌ๊ธฐ๋ฅผ ์ค์ธ๋ค
The Problem of Local Optima
์ค๋ซ๋์ ๊ธฐ์ธ๊ธฐ๊ฐ 0์ ๊ฐ๊น์ด, plateau๋ ํ์ต์ ๋ฐฉํดํ๋ ์์ธ์ด ๋๋ค
'ArtificialIntelligence > 2023GoogleMLBootcamp' ์นดํ ๊ณ ๋ฆฌ์ ๋ค๋ฅธ ๊ธ
[GoogleML] Batch Normalization (0) | 2023.09.21 |
---|---|
[GoogleML] Hyperparameter Tuning (1) | 2023.09.20 |
[GoogleML] Optimization Algorithms (0) | 2023.09.20 |
[GoogleML] Optimization Problem (0) | 2023.09.13 |
[GoogleML] Regularizing Neural Network (0) | 2023.09.12 |