Skip to content

Files

Latest commit

bbb2e33 · Aug 2, 2024

History

History
13 lines (10 loc) · 903 Bytes

File metadata and controls

13 lines (10 loc) · 903 Bytes

Speed

Below are the results of inferring one frame on an RTX 4090 GPU using the native PyTorch framework with torch.compile:

Model Parameters(M) Model Size(MB) Inference(ms)
Appearance Feature Extractor 0.84 3.3 0.82
Motion Extractor 28.12 108 0.84
Spade Generator 55.37 212 7.59
Warping Module 45.53 174 5.21
Stitching and Retargeting Modules 0.23 2.3 0.31

Note: The values for the Stitching and Retargeting Modules represent the combined parameter counts and total inference time of three sequential MLP networks.