Commit Graph

396 Commits

Author SHA1 Message Date
Yufeng
c390dc0c79 add Mglm (#5)
* mglm init

* add mglm requirements

Co-authored-by: Yufeng <zhuyufeng@gmail.com>
Co-authored-by: wenmeng.zwm <wenmeng.zwm@alibaba-inc.com>
2022-10-28 17:09:27 +08:00
xianzhe.xxz
88e8d4291a [to #42322933]"fix: set the eps and momentum of BN consistent with training"
To keep consistent between training and evaluation, change the eps and momentum of BN.
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10554451
2022-10-28 09:27:55 +08:00
menrui.mr
c7b0787049 修复初始化过程参数未生效问题
此前文生图模型没有加载configuration.json中的参数 影响默认配置
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10558026
2022-10-27 23:29:08 +08:00
hemu.zp
fa415d8720 [to #42322933] Fix bug for bloom and gpt_neo
1. 修复 bloom 和 gpt_neo 模型更新 transformers 4.23 后后处理报错的问题
2. 统一使用 ModelOutput 作为模型输出
3. gpt_neo checkpoint 已上线,修改 ut 为 level2
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10553103
2022-10-27 23:27:28 +08:00
Yingda Chen
374fd3090e [to #42322933]skip referring video tests since model is private 2022-10-27 20:23:51 +08:00
hemu.zp
d0f8547e7e [to #42322933] Fix gpt3 loading checkpoint after finetuning.
1. 修复GPT-3模型无法加载finetune保存的checkpoint的问题
2. 为GPT-3诗词生成模型添加 ut
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10537209
2022-10-26 20:58:00 +08:00
hemu.zp
e4a0e046f9 [to #42322933] Add ut for mplug and bloom
为新上线的 langboat/bloom-1b4-zh,damo/mplug_visual-question-answering_coco_base_zh,damo/mplug_image-captioning_coco_base_zh 三个模型添加 ut,test_level 设置为 2
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10524221
2022-10-26 16:19:20 +08:00
caorongyu.cry
3b8fb92c13 [to #42322933] debug header ids and header names
修复header_ids和header_names命名反了的问题
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10516557
2022-10-26 16:04:14 +08:00
ran.zhou
13f7e9ceca [to #42322933]SEA multilingual NLP (NER & word segmentation)
添加东南亚小语种NLP支持,包括:
1. 针对泰语,越南语NER的预处理
2. 基于XLMR-CRF架构的分词模型和pipeline
3. 针对泰语分词的预处理

添加了相应pipeline的unittest
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10492404
2022-10-26 14:52:22 +08:00
jiaqi.sjq
5190c7de11 [to #41669377] tts using default master revision model in UT
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10526747
2022-10-26 11:53:52 +08:00
yuanzheng.yuanzhen
bab54bbce8 [to #42322933]support uni fold
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10481410
2022-10-25 22:59:19 +08:00
yuze.zyz
c2da44b371 [to #42322933] remove dev model inference and fix some bugs
1. Change structbert dev revision to master revision
2. Fix bug:  Sample code failed because the updating of model configuration
3. Fix bug: Continue training regression failed
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10519992
2022-10-25 22:38:49 +08:00
zhangzhicheng.zzc
e1ab73b7d8 [to #42322933]support type str for for zero-shot labels' input
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10506320
2022-10-25 13:55:09 +08:00
hemu.zp
ffd834fc25 [to #42322933] Add bloom model
添加 bloom 模型
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10509187
2022-10-25 12:58:02 +08:00
yuze.zyz
605cd7f44a [to #42322933] NLP 1030 Refactor
Features:
1. Refactor the directory structure of nlp models. All model files are placed into either the model folder or the task_model folder
2. Refactor all the comments to google style
3. Add detail comments to important tasks and nlp models, to list the description of the model, and its preprocessor&trainer
4. Model Exporting now supports a direct all to TorchModelExporter(no need to derive from it)
5. Refactor model save_pretrained method to support direct running(independent from trainer)
6. Remove the judgement of Model in the pipeline base class, to support outer register models running in our pipelines
7. Nlp trainer now has a NLPTrainingArguments class , user can pass arguments into the dataclass, and use it as a normal cfg_modify_fn, to simplify the operation of modify cfg.
8. Merge the BACKBONES and the MODELS, so user can get a backbone with the Model.from_pretrained call
9. Model.from_pretrained now support a task argument, so user can use a backbone and load it with a specific task class.
10. Support Preprocessor.from_pretrained method
11. Add standard return classes to important nlp tasks, so some of the pipelines and the models are independent now, the return values of the models will always be tensors, and the pipelines will take care of the conversion to numpy and the following stuffs.
12. Split the file of the nlp preprocessors, to make the dir structure more clear.

Bugs Fixing:
1. Fix a bug that lr_scheduler can be called earlier than the optimizer's step
2. Fix a bug that the direct call of Pipelines (not from pipeline(xxx)) throws error
3. Fix a bug that the trainer will not call the correct TaskDataset class
4. Fix a bug that the internal loading of dataset will throws error in the trainer class
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10490585
2022-10-25 12:26:25 +08:00
siyang.ssy
6d51f44dc7 [to #42322933]fix input type for video embeding
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10506601
2022-10-25 12:11:28 +08:00
bin.xue
525fa3ea89 [to #42322933]test: use 'master' branch in training test
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10509580
2022-10-25 12:10:07 +08:00
caorongyu.cry
6178f46910 [to #42322933] add ut for multi threads
1. 修复multi thread引起的问题
2. 增加multi thread的unittest
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10502008
2022-10-25 09:49:02 +08:00
yingda.chen
de7b6a06e9 [to #42322933] remove revision usage for face detection
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10507910

    * [to #42322933] remove revision usage for face detection
2022-10-25 09:28:01 +08:00
zhangyanzhao.zyz
c4dbb69d65 [to #42322933]增加对text-ranking任务中文模型的单元测试,以方便得到官方模型打标。
增加对text-ranking任务中文模型的单元测试,以方便得到官方模型打标。
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10492754
2022-10-24 23:41:20 +08:00
yichang.zyc
35c612a642 [to #42322933]去除clip ut中的dev revision
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10507748

    * remove clip ut dev revision
2022-10-24 23:40:38 +08:00
ashui.cbh
e223c1b008 [to #42322933]merge master after demo service support
demo service 对接,修改输入接口为可调用的方式
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10502169
2022-10-24 18:47:01 +08:00
caorongyu.cry
35644fa0a7 [to #42322933] change star3 to space_T_cn
1. 合并star和star3框架
2. 修改star和star3的model type
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10492793
2022-10-23 20:25:24 +08:00
zhangzhicheng.zzc
182ba1768f [to #42322933]support multi tasks for part of speech
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10491994
2022-10-23 10:56:52 +08:00
yuze.zyz
707cbef013 [to #42322933]Fix bug in daily UT
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10491891
2022-10-22 23:25:18 +08:00
zhangzhicheng.zzc
2a87dee561 [to #42322933]support multi tasks-- will be failed, since configuration has not changed yet
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10492024
2022-10-22 21:09:15 +08:00
caorongyu.cry
9edfd7e50c [to #42322933] update tableqa params
1. 增加传入table_id
2. 将result和table的结构统一
3. 默认开启is_use_sqlite
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10492027
2022-10-22 20:33:49 +08:00
shichen.fsc
1854ceeb74 [to #42322933] Fix all asr models in UT with mistake model_id
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10491024
2022-10-22 20:30:45 +08:00
zhangyanzhao.zyz
824ee8232c [to #42322933]更新语义相关性任务英文名称为text ranking,修改对应变量名和类名
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10491951
2022-10-22 17:12:48 +08:00
yichang.zyc
683ee5bfed [to #42322933]use Tasks.ocr_recognition
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10490937
2022-10-22 17:01:03 +08:00
liugao.lg
baff7c5b64 [to #42322933]add ofa-ocr-recogniiton pipeline
新增ofa关于日常场景文字识别的任务,主要包括:
1、新增pipeline及task名称定义;
2、新增pipeline、task、model及prepreocess核心类方法的代码逻辑;
3、其它同步修正的小细节逻辑;
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10471089
2022-10-21 09:11:15 +08:00
hemu.zp
2b49b322a2 [to #42322933] Add palm ut
为以下三个模型补充 ut
damo/nlp_palm2.0_text-generation_chinese-large
damo/nlp_palm2.0_text-generation_commodity_chinese-base
damo/nlp_palm2.0_text-generation_weather_chinese-base
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10435599
2022-10-20 19:50:40 +08:00
xiangpeng.wxp
8ec90ccbf8 [to #42322933] Add uttest for en2fr and fr2en tasks
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10467797

    * add uttest for en2fr and fr2en tasks
2022-10-20 17:35:27 +08:00
shichen.fsc
1483c64638 [to #42322933] Fix ASR error when resample failed, and add all asr models UT, add apply-cmvn for pytorch models
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10465241
2022-10-20 12:54:37 +08:00
xianzhe.xxz
865397763e [to #42322933]add damoyolo model in tinynas-object-detection
接入damyolo系列检测模型
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10377688
2022-10-18 16:53:29 +08:00
shuying.shu
cb570d586c add referring video object segmentation pipeline
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10400324
2022-10-18 16:10:10 +08:00
hemu.zp
3b1f1a0252 [to #42322933] Add GPT3 tensor parallel inference
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
2022-10-18 15:58:33 +08:00
leyuan.hjy
172522d196 [to #42322933]video-object-detection init
新增video-object-detection 算法
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10247489
2022-10-17 20:58:23 +08:00
hemu.zp
271e2a2a99 [to #42322933] Add gpt_neo model
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
2022-10-17 20:54:29 +08:00
jiangnana.jnn
7720ae50e2 return dict values when input single sample for easycv pipeline
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10423383
2022-10-17 20:30:42 +08:00
shichen.fsc
542c4ce1b3 [to #42322933] Fix bug in KWS when setting customized keyword
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10412829
2022-10-17 14:07:05 +08:00
wendi.hwd
674e1a7878 [to #42322933]cv/cvdet_fix_outputs->master fix outputs
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10421413

    * fix outputs
2022-10-17 14:06:07 +08:00
wenqi.oywq
88a7599efb [to #42322933]change output channels from RGB to BGR, to consistent with demo-service
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
2022-10-17 14:05:12 +08:00
wenmeng.zwm
f6e542cdcb refine pipeline input to support demo service
* 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
2022-10-17 10:40:08 +08:00
caorongyu.cry
202fcdf298 [to #42322933] change tableqa output
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
2022-10-14 23:11:19 +08:00
zhangzhicheng.zzc
0eb823b764 [to #42322933] support t5_with_translation
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10383770

    * T5 support translate
2022-10-13 10:52:40 +08:00
jiaqi.sjq
144ffee2cf [to #42322933] Add explict model id in tts UT
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10371244
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10371244
2022-10-13 10:16:07 +08:00
yuxiang.tyx
2989492bc0 [to #42322933]更新face_detection_scrfd模型并支持finetune, 新增card_detection模型
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
2022-10-12 19:58:50 +08:00
shouzhou.bx
4cb5f8a2cd [to #42322933] add human whole body model and image object detection auto model
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/10319306
2022-10-12 19:53:14 +08:00
caorongyu.cry
7145990054 [to #42322933] reivse model problem and remove history sql for demo
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
2022-10-12 15:18:35 +08:00