repositories Search Results · repo:Bruce-Lee-LY/decoding_attention language:C++

Filter by

• • Code... (...)results
  • Issues
    results
  • Pull requests
    results
  • Discussions
    results
  • Commits
    results
  • Packages
    results
  • Wikis
    results

• Advanced

  • Owner
  • Size
  • Number of followers
  • Number of forks
  • Number of stars
  • Date created
  • Date pushed
  • Topic
  • License
  • Archived
  • Public
  • Private
• Advanced search

Code... (...)

0 files

 (63 ms)

0 files

inBruce-Lee-LY/decoding_attention (press backspace or delete to remove)

Sort by: Best match

Filter

Bruce-Lee-LY/decoding_attention

Decoding Attention is specially optimized for MHA, MQA, GQA and MLA using CUDA core for the decoding stage of LLM inference.

gpu

cuda

inference

nvidia

mha

• C++
·
• 38
·
• Updated
  26 days ago

Star

Sponsor open source projects you depend on

Contributors are working behind the scenes to make open source better for everyone—give them the help and recognition they deserve.Explore sponsorable projects

ProTip! 

Press the

/

key to activate the search input again and adjust your query.

Sponsor open source projects you depend on

Contributors are working behind the scenes to make open source better for everyone—give them the help and recognition they deserve.Explore sponsorable projects

ProTip! 

Press the

/

key to activate the search input again and adjust your query.