Abstract: Transformer-based video generation models have demonstrated significant potential in content creation. However, the current state-of-the-art model employing “ 3 D full attention” encounters ...
Abstract: As deep learning models are increasingly being deployed on resource-constrained edge devices, the need to develop techniques to make the model more energy efficient without sacrificing its ...
Oaken is an accleration solution that achieves high accuracy and high performance simultaneously through co-designing algorithm and hardware, leveraging online ...
[2024-11-12]: Support for sageattn_varlen is available now. For SageAttention V1 in Triton (slower than SageAttention V2/V2++/V3), refer to SageAttention-1 branch and install using pip: pip install ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results