Code

Block Attention for Bert model#210

Code

Open

quic-amitraj wants to merge 7 commits intoquic:mainquic/efficient-transformers:mainfrom quic-amitraj:block_attentionquic-amitraj/efficient-transformers:block_attentionCopy head branch name to clipboard

+144-10

Commits

Commits on Jan 10, 2025

• Block Attention for Bert model

  Show description for e0eb624
  quic-amitraj

  committed
  e0eb624

  Copy full SHA for e0eb624

• Minor-fix-1

  Show description for a5ce814
  quic-amitraj

  committed
  a5ce814

  Copy full SHA for a5ce814

• Update Automodel to handle block attention

  Show description for 7706904
  quic-amitraj

  committed
  7706904

  Copy full SHA for 7706904

• Addedd test for the Block Attention

  Show description for 0d99ad2
  quic-amitraj

  committed
  0d99ad2

  Copy full SHA for 0d99ad2

• Added copyright

  Show description for 750bc87
  quic-amitraj

  committed
  750bc87

  Copy full SHA for 750bc87

• Approach changed

  Show description for b6b232a
  quic-amitraj

  committed
  b6b232a

  Copy full SHA for b6b232a

• Code clean-1

  Show description for 20b4b85
  quic-amitraj

  committed
  20b4b85

  Copy full SHA for 20b4b85