koreashin/siftformer-1.5

SiftFormer 1.5

Classes

Index	Class
0	정상 (normal)
1	졸음 (drowsy)
2	주의분산 (distracted)
3	폭행 (violence)

Input

• Shape: [B, T, C, H, W]
• T = 30 frames
• H = W = 224
• C = 3 (RGB)
• ImageNet normalization (mean [0.485, 0.456, 0.406], std [0.229, 0.224, 0.225])
• dtype: float32 or bfloat16