Skip to content

MeforgersDev/PulsarAttention

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1 Commit
 
 
 
 

Repository files navigation

PulsarAttention

We report the expected speedup (combined forward and backward pass) and memory savings achieved by PulsarAttention compared to PyTorch’s standard attention implementation. By improving memory access efficiency and compute utilization, PulsarAttention enables faster model execution with significantly reduced VRAM usage.

About

We report the expected speedup (combined forward and backward pass) and memory savings achieved by PulsarAttention compared to PyTorch’s standard attention implementation. By improving memory access efficiency and compute utilization, PulsarAttention enables faster model execution with significantly reduced VRAM usage.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors