attention alignment

[cjw-d]

PR types

Others

PR changes

Others

Description

Align attention implementation with transformers

[paddle-bot]

Thanks for your contribution!

[codecov-commenter]

Codecov Report

❌ Patch coverage is 84.72222% with 11 lines in your changes missing coverage. Please review.
⚠️ Please upload report for BASE (develop@b1afef6). Learn more about missing BASE report.

Files with missing lines	Patch %	Lines
paddleformers/transformers/ernie4_5/modeling.py	50.00%	5 Missing ⚠️
paddleformers/transformers/deepseek_v2/modeling.py	75.00%	2 Missing ⚠️
paddleformers/transformers/glm4_moe/modeling.py	71.42%	2 Missing ⚠️
paddleformers/transformers/gpt_oss/modeling.py	75.00%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             develop    #2966   +/-   ##
==========================================
  Coverage           ?   31.44%           
==========================================
  Files              ?      420           
  Lines              ?    68566           
  Branches           ?        0           
==========================================
  Hits               ?    21560           
  Misses             ?    47006           
  Partials           ?        0           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.