AI & ML interests
None defined yet.
Recent Activity
View all activity
Four FastH3 4-Step Preview v1 T2VA checkpoints from FastVideo: VSA and dense attention, with data-free and synthetic-data recipes.
-
FastVideo/FastVideo-FastH3-4-step-Preview-v1-VSA-DataFree
Text-to-Video • 35B • Updated • 858k • 327 -
FastVideo/FastVideo-FastH3-4-step-Preview-v1-VSA-Synthetic-Step1300
Text-to-Video • 35B • Updated • 7 • 4 -
FastVideo/FastVideo-FastH3-4-step-Preview-v1-VSA-Synthetic-Step1900
Text-to-Video • 35B • Updated • 1.23k • 2 -
FastVideo/FastVideo-FastH3-4-step-Preview-v1-Dense-DataFree
Text-to-Video • 33B • Updated • 810 • 2
models trained with video sparse attention: https://arxiv.org/abs/2505.13389 and distillation
Four FastH3 4-Step Preview v1 T2VA checkpoints from FastVideo: VSA and dense attention, with data-free and synthetic-data recipes.
-
FastVideo/FastVideo-FastH3-4-step-Preview-v1-VSA-DataFree
Text-to-Video • 35B • Updated • 858k • 327 -
FastVideo/FastVideo-FastH3-4-step-Preview-v1-VSA-Synthetic-Step1300
Text-to-Video • 35B • Updated • 7 • 4 -
FastVideo/FastVideo-FastH3-4-step-Preview-v1-VSA-Synthetic-Step1900
Text-to-Video • 35B • Updated • 1.23k • 2 -
FastVideo/FastVideo-FastH3-4-step-Preview-v1-Dense-DataFree
Text-to-Video • 33B • Updated • 810 • 2
FastMetal QAD INT8 text-to-video models for Apple Silicon
models trained with video sparse attention: https://arxiv.org/abs/2505.13389 and distillation