Interaction
Scene Memory
Current Frame
Output
t = 0
t = 1
t = 2
t = 3
Causal DiT
Zero
Zero
Zero
x0rgb_00.png
rgb_00.png
cΔv0velocity_00.png
cpos0position_00.png
ctrack0track_00.png
x1rgb_04.png
rgb_04.png
cΔv1velocity_04.png
cpos1position_04.png
ctrack1track_04.png
x2rgb_08.png
rgb_08.png
cΔv2velocity_08.png
cpos2position_08.png
ctrack2track_08.png
x3rgb_12.png
rgb_12.png
Update memory
Update memory
1.0×