[Inference]qwen2-a8w8c8 support use_fake_parameter

[ckl117]

PR types

New features

PR changes

Others

Description

qwen2-a8w8c8 support use_fake_parameter;
remove shift and smooth when use_fake_parameter=True；
qwen2 support MultiGPU inference;
set the default scale type to fp32;

[paddle-bot]

Thanks for your contribution!

[codecov]

Codecov Report

Attention: Patch coverage is 0% with 61 lines in your changes missing coverage. Please review.

Project coverage is 53.30%. Comparing base (9806293) to head (8fe4d67).
Report is 232 commits behind head on develop.

Files with missing lines	Patch %	Lines
...dlenlp/experimental/transformers/qwen2/modeling.py	0.00%	50 Missing ⚠️
...dlenlp/experimental/transformers/llama/modeling.py	0.00%	9 Missing ⚠️
paddlenlp/experimental/model_utils.py	0.00%	1 Missing ⚠️
paddlenlp/experimental/transformers/utils.py	0.00%	1 Missing ⚠️

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #9109      +/-   ##
===========================================
- Coverage    53.34%   53.30%   -0.04%     
===========================================
  Files          652      652              
  Lines       105401   105470      +69     
===========================================
  Hits         56222    56222              
- Misses       49179    49248      +69     

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