【Hackathon 4th No.104】generate API对齐HF

[zwhong714]

PR types

Application

PR changes

Description

提交RFC文档，sampling改写，新增contrastive search

[paddle-bot]

Thanks for your contribution!

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[gongel]

非常赞，感谢贡献，我们马上review哈。也辛苦签署下CLA

[zwhong714]

OK

[gongel]

改的很好！另外是否考虑做一些默认超参的对齐，比如我们这里的max_length和HF的max_new_tokens是一致的，HF的max_length会考虑输入的长度。

[zwhong714]

hello，这边交了一版，不知道还有什么问题没有

[gongel]

conflicts解决下哈
另外大批量CI挂了

[zwhong714]

目前无法处理的模型如下
无decoder hidden states
pytest PaddleNLP/tests/transformers/pegasus/test_modeling.py

multihead类型（OPT)，与paddle.nn 中multiheadAttention 不一致
pytest PaddleNLP/tests/transformers/opt/test_modeling.py

multihead类型(GPT)，与paddle.nn 中multiheadAttention 不一致
pytest PaddleNLP/tests/transformers/gpt/test_modeling.py

[gongel]

好的辛苦，目前改动文件太多，检查下是否代码风格的CI没过。

[zwhong714]

好的辛苦，目前改动文件太多，检查下是否代码风格的CI没过。

好像只有test_generation_utils, 和test_generation_utils 改了，其它文件不知道是不是拉取develop版本的时候出现的问题

[zwhong714]

这边我看日志，test和cli挂了好像就是因为那三个模型的问题

[sijunhe]

@zwhong714 您好，感谢您参与飞桨黑客松以及为这个PR付出的努力。这个PR我们看下来问题不大，基本可以合入, 但是因为我们目前的一些内部原因暂时不能合入。我们后续会找合适的机会合入这个PR。黑客松活动这边我们算您完成这个任务了~

[github-actions]

This Pull Request is stale because it has been open for 60 days with no activity. 当前Pull Request 60天内无活动，被标记为stale。

[github-actions]

This Pull Request is stale because it has been open for 60 days with no activity. 当前Pull Request 60天内无活动，被标记为stale。

[github-actions]

This Pull Request is stale because it has been open for 60 days with no activity. 当前Pull Request 60天内无活动，被标记为stale。

[paddle-bot]

Automatically closed by Paddle-bot.