Redefining Temporal Modeling in Video Diffusion: The Vectorized Timestep Approach
Paper
•
2410.03160
•
Published
•
4
Raphael Liu
RaphaelLiu
AI & ML interests
CV
Recent Activity
liked
a model
19 days ago
Etched/oasis-500m
liked
a dataset
28 days ago
Koala-36M/Koala-36M-v1
Reacted to
singhsidhukuldeep's
post
with ❤️
29 days ago
If you have ~300+ GB of V-RAM, you can run Mochi from @genmo
A SOTA model that dramatically closes the gap between closed and open video generation models.
Mochi 1 introduces revolutionary architecture featuring joint reasoning over 44,520 video tokens with full 3D attention. The model implements extended learnable rotary positional embeddings (RoPE) in three dimensions, with network-learned mixing frequencies for space and time axes.
The model incorporates cutting-edge improvements, including:
- SwiGLU feedforward layers
- Query-key normalization for enhanced stability
- Sandwich normalization for controlled internal activations
What is currently available?
The base model delivers impressive 480p video generation with exceptional motion quality and prompt adherence. Released under the Apache 2.0 license, it's freely available for both personal and commercial applications.
What's Coming?
Genmo has announced Mochi 1 HD, scheduled for release later this year, which will feature:
- Enhanced 720p resolution
- Improved motion fidelity
- Better handling of complex scene warping
Organizations
None yet