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 | 29 | 30 |
- multimodal
- Unreal Engine
- BERT
- 모션매칭
- Few-shot generation
- Font Generation
- cv
- ddpm
- 언리얼엔진
- deep learning
- Diffusion
- 딥러닝
- WinAPI
- 폰트생성
- Stat110
- 오블완
- NLP
- GAN
- userwidget
- UE5
- ue5.4
- dl
- motion matching
- 디퓨전모델
- CNN
- animation retargeting
- Generative Model
- WBP
- 생성모델
- RNN
- Today
- Total
목록Vision Transformer (2)
Deeper Learning

Hugo Touvron, Matthieu Cord, Matthijs Douze, Francisco Massa, Alexandre Sablayrolles, Herve Jegou. Facebook AI. Sorbonne University. (2020.12) Abstract Image task에 pure attention 기반 모델이 활용되지만 large dataset에서 pre-training이 필수적이며 이는 활용에 한계를 가져옴 single computer에서 3일 동안 ImageNet dataset만 학습하여 top-1 acc 83.1%를 달성 attention을 통해 학습하는 distillation token을 도입한 transformer를 위한 teacher-student strategy를 제시 ..

Zihang Dai, Hanxiao Liu, Quoc V. Le, Mingxing Tan Google Research, Brain Team (2021.06) Abstract Transformer 아직 vision task에서 SOTA convolutional network보다 성능이 떨어짐 Transformer는 더 큰 capacity를 가지지만 inductive bias의 부족으로 일반화 성능이 convolutional network에 비해 떨어짐 두 아키텍처의 장점을 결합하기 위해 hybrid model인 CoAtNets를 제시 CoAtNet의 key insights depthwise Convolution, Self-Attention은 간단한 relative attention을 통해 결합 가능 수직으..