# Conflicts:
#	modelscope/metrics/__init__.py

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10400324

添加基于 Megatron-v3 的 GPT3 tensor 并行的推理代码
复用 DistributedPipeline 与 megatron-util
适用模型：1.3B/2.7B/13B 参数的 GPT-3 预训练生成模型
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10416721

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10427797

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10434107
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10434107

    * [Update] update finetune_result_upload

* [Update] rename finetune_result_upload to model_dir_upload

* Merge branch 'master' into feat/upload_ckpt

* Merge branch 'master' into feat/upload_ckpt

* [Fix] fix import error

* [Fix] fix import error

* Merge branch 'master' into feat/upload_ckpt

* [Update] changes name to upload_folder and using tempfile to save repo

* Merge branch 'master' into feat/upload_ckpt

* [Fix] fix commit

* Merge branch 'master' into feat/upload_ckpt

* [Fix] fix format

* Merge branch 'master' into feat/upload_ckpt

* [Fix] add uuid after model created from upload ut

新增video-object-detection 算法
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10247489

1. 添加 gpt_neo 模型，因 checkpoint 归属于 Langboat 还未上传到模型库，已线下完成测试
2. 添加 text-generation task models 与 head，后续会将 gpt3，palm 等已上线文本生成模型统一为 backbone + head 结构的 task models
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10404249

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10423383

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10386689

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10412829

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10421413

    * fix outputs

默认输出为array的，通道格式统一为BGR格式，本次修改是为了与这个格式一致
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10422508

* image_captioninig support single image and dict input
* image_style_transfer use dict input

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10417330

修改output的结构，直接返回可转化成json format的结构
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10415403

上传和下载支持多文件操作

添加2d手部关键点检测finetune功能
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10371710

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10383770

    * T5 support translate

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10371244
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10371244

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10275823

1. 调整face_detection的文件层级(scrfd与其余新增face_detection方法平级);
2. 增加极大脸/旋转脸的检测方法，更新了新模型；
3. 支持读入数据集并finetune和eval；
4. 新增card_detection模型，支持读入datasethub数据集并finetune
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10244540

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10319306

相比于master上的tableqa，做出了如下修复：
1. 修复了schema linking中的问题。
2. 同时设置了有history sql和没有history sql的两种输入
3. 增加了sqlite执行逻辑，可以返回sql执行结果
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10365114

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10338265

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10362437

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10111615

video summarization:
1. 添加 license & header; 
2. 修改 output for demo service
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10260946

Link https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10360754#tab=detail
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10360754

统一关键点检测输出key的名字
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10359335

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10198228

    * [to #44847108] add sparsity hook (pst algorithm)

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10217723

    * add image_body_reshaping code

# Conflicts:
#	modelscope/models/multi_modal/ofa/utils/__init__.py

Add pipelines for the following new models:
- [Multilingual Quality Estimation](https://modelscope.cn/models/damo/nlp_translation_quality_estimation_multilingual/summary)
- [Automatic Post-Editing (En-De)](https://modelscope.cn/models/damo/nlp_automatic_post_editing_for_translation_en2de/summary)
- [Domain classification (Zh)](https://modelscope.cn/models/damo/nlp_domain_classification_chinese/summary)
- [Style classification (Zh)](https://modelscope.cn/models/damo/nlp_style_classification_chinese/summary)
- [Style classification (En)](https://modelscope.cn/models/damo/nlp_style_classification_english/summary)
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10315370

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10233235

* add input definition for pipeline


* support multiple input format for single pipeline


*  change param for body_3d_keypoints

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10252583

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10202117

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10263904

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10260332

Add patch shift transformer model for action recognition task.
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10282964

以下算法进行了header变更：
modelscope.models.cv.cmdssl_video_embedding
modelscope.models.cv.action_recognition 
modelscope.models.cv.animal_recognition
modelscope.models.multi_modal.multi_stage_diffusion
modelscope.models.multi_modal.gemm

modelscope.pipelines.cv.live_category_pipeline
modelscope.pipelines.cv.video_category_pipeline
modelscope.models.cv.image_to_image_translation
modelscope.models.cv.image_to_image_generation

modelscope.models.cv.video_inpainting
modelscope.models.multi_modal.diffusion
modelscope.models.multi_modal.team
modelscope.models.cv.shop_segmentation
modelscope.models.cv.text_driven_segmentation
modelscope.models.cv.action_recognition 


modelscope.models.cv.face_emotion
modelscope.models.cv.hand_static
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10268474

1.新增支持原始bert模型（非easynlp的 backbone prefix版本）
2.支持bert的在sequence classification/fill mask /token classification上的backbone head形式
3.统一了sequence classification几个任务的pipeline到一个类
4.fill mask 支持backbone head形式
5.token classification的几个子任务（ner，word seg， part of speech）的preprocessor 统一到了一起TokenClassificationPreprocessor
6. sequence classification的几个子任务（single classification， pair classification）的preprocessor 统一到了一起SequenceClassificationPreprocessor
7. 改动register中 cls的group_key 赋值位置，之前的group_key在多个decorators的情况下，会被覆盖，obj_cls的group_key信息不正确
8. 基于backbone head形式将 原本group_key和 module同名的情况尝试做调整，如下在modelscope/pipelines/nlp/sequence_classification_pipeline.py 中 
原本
 @PIPELINES.register_module(
    Tasks.sentiment_classification, module_name=Pipelines.sentiment_classification)
改成
@PIPELINES.register_module(
    Tasks.text_classification, module_name=Pipelines.sentiment_classification)
相应的configuration.json也有改动，这样的改动更符合任务和pipline（子任务）的关系。
8. 其他相应改动为支持上述功能
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10041463

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10244616