Boosting Wan2.2 I2V Inference on 8xH100s, 56% Faster with Sequence Parallelismmorphic.com6 points·palakzat··1 commentOpen articleSaveView on HN