Boosting Wan2.2 I2V Inference on 8xH100s, 56% Faster with Sequence Parallelism

6 pointsposted 18 hours ago
by palakzat
(morphic.com)

1 Comments