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
- Few-shot generation
- UE5
- multimodal
- NLP
- 모션매칭
- WinAPI
- 오블완
- Unreal Engine
- 언리얼엔진
- 딥러닝
- Stat110
- ddpm
- motion matching
- ue5.4
- CNN
- GAN
- BERT
- animation retargeting
- Diffusion
- Font Generation
- RNN
- WBP
- userwidget
- 폰트생성
- dl
- Generative Model
- deep learning
- 디퓨전모델
- cv
- 생성모델
Archives
- Today
- Total
목록Inception score #GAN (1)
Deeper Learning
Inception score (IS)
Why Inception score GAN은 생성된 결과물의 performance의 측정이 어렵다. Generator와 Discriminator의 loss만을 보고 생성된 결과물의 performance를 측정하는 것은 정확하지 않다. 완벽한 가상의 Discriminator를 속일 수 있는 Generator가 존재한다면 측정이 가능하지만 현실적으로 불가능하다. Generated output에 대한 performance를 측정하기 위한 metrics가 필요하고 이를 위해 고안된 여러 가지 방법 중 하나가 Inception Score. Inception Score (IS) Inception score(이하 IS)는 pre-trained된 inception v3 모델을 사용하기 때문에 붙여진 이름이다. Imag..
AI/Deep Learning
2020. 12. 4. 14:41