[LLM] Add Aquila2 model

[ftgreat]

PR types

New features

PR changes

Models

Description

Add Aquila2 model

[paddle-bot]

Thanks for your contribution!

[DrownFish19]

LGTM

[codecov]

Codecov Report

Attention: Patch coverage is 15.87697% with 1012 lines in your changes missing coverage. Please review.

Project coverage is 52.38%. Comparing base (801682d) to head (4bfeb7b).
Report is 2 commits behind head on develop.

Files with missing lines	Patch %	Lines
paddlenlp/transformers/aquila2/modeling.py	13.81%	786 Missing ⚠️
paddlenlp/transformers/aquila2/tokenizer.py	22.85%	108 Missing ⚠️
paddlenlp/transformers/aquila2/fusion_ops.py	17.43%	90 Missing ⚠️
paddlenlp/transformers/aquila2/configuration.py	28.20%	28 Missing ⚠️

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #9692      +/-   ##
===========================================
+ Coverage    52.34%   52.38%   +0.04%     
===========================================
  Files          726      723       -3     
  Lines       117265   113500    -3765     
===========================================
- Hits         61377    59461    -1916     
+ Misses       55888    54039    -1849     

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

[ZHUI]

上传个小模型，写一下 model，Tokenizer的测试吧。模型加载，前向运行。Tokenizer加载，文本转id

[DrownFish19]

在PaddleNLP/tests下辛苦增加一下单元测试，可参考https://github.com/PaddlePaddle/PaddleNLP/tree/develop/tests/transformers/qwen2 实现，重点验证模型初始化和运行能力。