250x250
반응형
Notice
Recent Posts
Recent Comments
Link
일 | 월 | 화 | 수 | 목 | 금 | 토 |
---|---|---|---|---|---|---|
1 | ||||||
2 | 3 | 4 | 5 | 6 | 7 | 8 |
9 | 10 | 11 | 12 | 13 | 14 | 15 |
16 | 17 | 18 | 19 | 20 | 21 | 22 |
23 | 24 | 25 | 26 | 27 | 28 |
Tags
- requests
- subdag
- XAI
- airflow subdag
- 유튜브 API
- Airflow
- youtube data
- hadoop
- UDF
- login crawling
- gather_nd
- 상관관계
- GenericGBQException
- BigQuery
- Retry
- correlation
- tensorflow text
- Counterfactual Explanations
- 공분산
- grad-cam
- spark udf
- API
- integrated gradient
- TensorFlow
- API Gateway
- session 유지
- GCP
- flask
- top_k
- chatGPT
Archives
- Today
- Total
목록grad-cam (1)
데이터과학 삼학년
Grad-CAM
Grad-CAM(Gradient-weighted Class Activation Mapping) 딥러닝 모델이 어떤 부분을 보고 특정 클래스를 판단했는지를 시각화하는 기술 -> 이미지에서 주로 사용 -> 텍스트에서도 사용 가능 이를 통해 모델이 어떤 부분을 주로 활용하는지를 알 수 있어 모델의 해석성(interpretability)을 높일 수 있음 Grad-CAM은 기존의 Class Activation Mapping(CAM)을 발전시킨 기술로, CAM은 Global Average Pooling(GAP)을 사용하여 클래스에 대한 중요도를 계산 이와 달리 Grad-CAM은 전체적인 특성 맵의 중요도가 아닌 클래스에 대한 중요도를 계산 Grad-CAM 계산 단계 1. 모델의 gradient를 계산 - 모델의 g..
Explainable AI
2023. 3. 24. 11:47