update fp8 training

[lugimzzz]

PR types

Others

PR changes

APIs

Description

1.改用FP8改用forward fp8_fp8_half_gemm_fused，如果反向不使用FP8无需安装TE
2.反向进行padding，支持moe场景中seq_len不为fp8倍数场景
3.去除多余dtype view

[paddle-bot]

Thanks for your contribution!

[codecov]

Codecov Report

Attention: Patch coverage is 10.00000% with 18 lines in your changes missing coverage. Please review.

Project coverage is 46.93%. Comparing base (7309a5d) to head (683d5ef).
Report is 1 commits behind head on develop.

Files with missing lines	Patch %	Lines
paddlenlp/quantization/qat_utils.py	10.00%	18 Missing ⚠️

❌ Your patch status has failed because the patch coverage (10.00%) is below the target coverage (80.00%). You can increase the patch coverage or adjust the target coverage.
❌ Your project status has failed because the head coverage (46.93%) is below the target coverage (58.00%). You can increase the head coverage or adjust the target coverage.

Additional details and impacted files

@@           Coverage Diff            @@
##           develop   #10720   +/-   ##
========================================
  Coverage    46.93%   46.93%           
========================================
  Files          800      800           
  Lines       132741   132739    -2     
========================================
+ Hits         62304    62305    +1     
+ Misses       70437    70434    -3     

☔ 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.

[nepeplwu]

LGTM