add single_model network and use intermediate api

[blacksheep-Aristotle]

PR types

New features

PR changes

Models

Description

添加llama ，qwen，gpt 单卡组网。
auto_trainer支持使用中层api & 运行脚本支持使用中层api

[paddle-bot]

Thanks for your contribution!

[jeff41404]

是否构造dp_config传入parallelize，不需要再单独进行 sharded_data_parallel

[codecov]

Codecov Report

Attention: Patch coverage is 16.83103% with 1265 lines in your changes missing coverage. Please review.

Project coverage is 52.32%. Comparing base (10a62c7) to head (bbc0a6d).
Report is 6 commits behind head on develop.

Files with missing lines	Patch %	Lines
paddlenlp/transformers/gpt/modeling_network.py	17.50%	443 Missing ⚠️
paddlenlp/transformers/llama/modeling_network.py	15.97%	426 Missing ⚠️
paddlenlp/transformers/qwen/modeling_network.py	17.18%	376 Missing ⚠️
paddlenlp/trainer/auto_trainer.py	0.00%	20 Missing ⚠️

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #9412      +/-   ##
===========================================
- Coverage    53.04%   52.32%   -0.73%     
===========================================
  Files          677      679       +2     
  Lines       108671   109432     +761     
===========================================
- Hits         57649    57256     -393     
- Misses       51022    52176    +1154     

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