더북(TheBook)

다음은 모델을 학습시킨 결과입니다.

Iteration: 500. Loss: 1.661692500114441. Accuracy: 43.59000015258789
Iteration: 1000. Loss: 0.8945671319961548. Accuracy: 76.19999694824219
Iteration: 1500. Loss: 0.29147762060165405. Accuracy: 89.7300033569336
Iteration: 2000. Loss: 0.23627924919128418. Accuracy: 93.51000213623047
Iteration: 2500. Loss: 0.03288724273443222. Accuracy: 95.05000305175781
Iteration: 3000. Loss: 0.030374949797987938. Accuracy: 95.81999969482422
Iteration: 3500. Loss: 0.16210567951202393. Accuracy: 96.33999633789062
Iteration: 4000. Loss: 0.1930878460407257. Accuracy: 96.19000244140625
Iteration: 4500. Loss: 0.05172012746334076. Accuracy: 97.0
Iteration: 5000. Loss: 0.1390017569065094. Accuracy: 97.26000213623047
Iteration: 5500. Loss: 0.08090303093194962. Accuracy: 97.62000274658203
Iteration: 6000. Loss: 0.1048836037516594. Accuracy: 97.69000244140625
Iteration: 6500. Loss: 0.07984019815921783. Accuracy: 97.80000305175781
Iteration: 7000. Loss: 0.10250381380319595. Accuracy: 97.55999755859375
Iteration: 7500. Loss: 0.06477993726730347. Accuracy: 97.86000061035156
Iteration: 8000. Loss: 0.10547631978988647. Accuracy: 97.80000305175781
Iteration: 8500. Loss: 0.042811520397663116. Accuracy: 98.0199966430664
Iteration: 9000. Loss: 0.04198891296982765. Accuracy: 98.22000122070312

LSTM 셀을 실행했을 때의 정확도와 유사합니다. 즉, LSTM 셀을 사용하든, GRU 셀을 사용하든 정확도가 비슷하게 나왔습니다. 시계열 처리 관련한 모델(RNN, LSTM, GRU) 중 어떤 것이 더 좋다고 말할 수는 없습니다. 따라서 주어진 데이터셋을 다양한 모델에 적용하여 최적의 모델을 찾는 것이 중요합니다.

신간 소식 구독하기
뉴스레터에 가입하시고 이메일로 신간 소식을 받아 보세요.