Commit Graph

1730 Commits

Author SHA1 Message Date
wenmeng.zwm
6ab85256af to [#50529030] fix easycv lr hook error
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/13069804
2023-06-26 21:54:01 +08:00
suluyan.sly
b6c40aaa87 [to #50538422] feat: make diff & recover for llama based weights
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/13055390
2023-06-26 19:41:19 +08:00
mulin.lyh
698c794070 [to #50537864]fix: fix select case issue
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/13055098
2023-06-25 22:44:29 +08:00
hemu.zp
1421629392 Support device_map for transformers model
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/13048856

* support device_map

* support device_map for T5

* fix bug
2023-06-25 20:15:11 +08:00
chenyafeng.cyf
29062d9f94 eres2net_aug v2
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/13046524

* eres2net_aug v2
2023-06-25 18:07:04 +08:00
wenmeng.zwm
674c67ce48 move data to lfs 2023-06-21 17:57:34 +08:00
tongmu.wh
f03c93cda5 add speaker diarization pipeline and improve some speaker pipelines
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12891685

* add new speaker diarization pipeline. improve the capability of speaker pipelines
2023-06-21 17:56:05 +08:00
xingjun.wxj
0db0ec5586 Merge code from github
1. Merge(add) daily regression from github PR (daily_regression.yaml)
2. Add lora stable diffusion from github PR
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/13010802
* fix: device arg not work, rename device to ngpu (#272)

* Correcting the lora stable diffusion example script (#300)

* add vad model and punc model in README.md 

add vad model and punc model

* Merge pull request #302 from modelscope/langgz-patch-1

add vad model and punc model in README.md

* add 1.6

* modify ignore

* Merge pull request #307 from modelscope/dev_rs_16

Merge release 1.6

* undo datetime to 2099

* Merge pull request #311 from modelscope/fix_master_version

undo datetime to 2099

* add daily regression workflow

* modify workflow name

* fix cron format issue

* lora trainer

* Merge pull request #315 from liuyhwangyh/add_regression_workflow

add daily regression workflow
2023-06-21 10:22:06 +08:00
lanjinpeng.ljp
bc5c16aa10 support third_party key in pipeline
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12940228

* support third_party key in pipeline
2023-06-20 19:25:24 +08:00
mulin.lyh
fa7562fd96 [to #50384694]fix: ci hang when build torch extension
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12976382
2023-06-19 17:56:04 +08:00
hemu.zp
ad5d1aeb62 Fix gpt3 finetune nan
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12981998
* fix gpt3 finetune nan
2023-06-19 11:25:27 +08:00
xingjun.wxj
cc3c384d5e Fix issues for downloading mplug-youku dataset
1. Optimize downloading meta-csv files for large-scale dataset like mPLUG-youku (> 1GB for meta csv mapping)
2. Add head and overall progress bar for NativeIterableDataset
3. Modify the try-catch info for oss_utils
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12952842
2023-06-15 15:42:21 +08:00
suluyan.sly
de491cfe6b [to #48855014] llama finetune + deepspeed
1. llama base finetune:使用trainer从llama finetune至alpaca,效果确认并提供example
2. deepspeed通用性完善:mpu解耦;deepspeed的训练信息支持从ms log透出(目前ms打印的log有误);支持从modelscope configuration.json 进行 deepspeed config 配置;deepspeed optimizer和lr_scheduler 初始化支持;解决deepspeed和ddp同时使用报错;解决保存ckpt时报错
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12651323
2023-06-14 10:00:23 +08:00
zhifu.gzf
8b4e9dcdfb ngpu bug and rm easyasr
修复ngpu指定无效的问题;移除easyasr,全部涉及模型都下架了;将funasr版本限制为>=0.6.0
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12933049
* ngpu bug and rm easyasr
2023-06-13 16:36:21 +08:00
mulin.lyh
d6bf9e80fb [to #50129142]feat: support pre build torch extension build image, first extension megatron_util 2023-06-12 11:03:28 +08:00
hemu.zp
96c2d42f09 Add StreamingMixin
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12445731
* StreamingMixin poc

* update design

* Merge branch 'master' into feat/StreamingMixin

* add dicstr

* make postprocessor input consistent
2023-06-08 19:40:14 +08:00
wucong.lyb
a77d390238 fix extra_args
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12839071
* fix extra_args

* update extra_args
2023-06-08 19:39:05 +08:00
hejunjie.hjj
ae672b73df bug fix for single human parsing
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12869114
* fix bug: single human parsing, 当src_h, src_w相等时,new_h, new_w = 0, 0, 出现bug
2023-06-08 19:36:18 +08:00
zsl01670416
75c90d8a98 add model revision in training_args and modify dataset loading in finetune text classification
1.add parameter model_revision in training_args.py.
2.add parameter model_revision in kwargs for finetune_text_classification.py and finetune_text_generation.py.
3.modify dataset loading in finetune_text_classification.py for flex training.
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12869552
* add model revision in training_args and modify dataset loading in finetune text classification
2023-06-08 19:36:05 +08:00
mulin.lyh
25a7398aea docker file py38 and py37 compatible merge
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12722824
* debug

* add missing deps

* add stanza

* fix numpy issue

* numpy version

* tf case

* numpy <= 1.22.0

* fix tf 1.15.5 case

* add python38 and python37 compatible docker file

* cv add fvcore

* add easycv

* remove debug code

* fix style issue

* fix python3.7 python3.8 compatible issue

* comment

* modify run config

* scipy<=1.7.3

* revert scipy<=1.7.3

* fix compatible bug

* fix compatible bug

* add easycv

* update version: 1.6.1rc0
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12754942
fix python3.8 and python3.8 compatible issue

* numpy version change

* numpy version

* numpy version

* fix numpy version

* fix numpy version

* update version: 1.6.1rc0
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12754942
fix python3.8 and python3.8 compatible issue

* restore setup.py

* restore setup.py

* add build base image

* fix style

* numpy version fix

* optimize build image, split base and modelscope image

* fix style issue

* modify example path

* modify examples folder

* add SETUPTOOLS_USE_DISTUTILS=stdlib comments
2023-06-08 10:01:58 +08:00
pangda
73f8d39897 Add ner example for flextrain
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12835095
* update named_entity_recognition example
2023-06-05 15:17:45 +08:00
xixing.tj
1b7e0f50f4 add ocr detection new model db-nas
新增5M的DB-NAS ocr detection 文字检测模型
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12760623
* add ocr detection new model db-nas

* add comment
2023-05-31 21:32:46 +08:00
hemu.zp
898e3a42eb Fix bug for DeepspeedHook.register_processor
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12686921
2023-05-31 21:29:36 +08:00
yuanzhi.zyz
10c39b5ce1 add new ocr recognition model (LightweightEdge) and some functions
1. 增加了新轻量化端侧识别模型 LightweightEdge,并把原来CRNN和ConvNextViT的代码整理了
2. 增加batch inference支持
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12787905
2023-05-31 21:16:22 +08:00
zhangzhicheng.zzc
c7aede97ba add pattern to skip easycv.thirdparty
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12785287

* add pattern to remove

* update pattern to remove

* avoid import with build
2023-05-30 11:16:39 +08:00
yuze.zyz
003c33648a Fix CI: test merge dataset failed
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12771691
* fix CI

* add teardown
2023-05-29 10:32:33 +08:00
xingjun.wxj
5b99bb24fe update version: 1.6.1rc0
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12754942
2023-05-25 20:28:36 +08:00
yuze.zyz
e2a20fd817 fix merge_hook
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12718423
* fix merge_hook
2023-05-25 20:05:04 +08:00
yuze.zyz
a7a3eb5dc5 Support flex train feature
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12737668
2023-05-25 19:39:24 +08:00
yuze.zyz
c08b924968 Fix ner tokenizer which won't accept kwargs
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12664244
* fix slow tokenizer
2023-05-25 18:55:47 +08:00
feiyu.gfy
f82361d71b 修正lineless_table_recognition功能遇到空白图片崩溃的bug
用户反馈lineless_table_recognition功能遇到空白图片会发生崩溃。
增加处理机制以防止崩溃。

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12741201
* fix bug
2023-05-25 18:28:05 +08:00
lee.lcy
a524e01e47 feat: support onnx export for domain_specific_object_detection
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12682688
2023-05-25 10:44:09 +08:00
xingjun.wxj
f2640a5a12 fix private dataset auth issue
1. Fix private datasets auth issue
2. Add arg: token (optional) in MsDataset.load() for FlexTrain
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12721569
2023-05-24 19:48:20 +08:00
yuze.zyz
d38fa563a8 Re commit lossing files for flextrain
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12674746
2023-05-19 10:48:33 +08:00
mulin.lyh
6abdce929d fix py38 issue
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12660340
2023-05-17 22:11:23 +08:00
chenyafeng.cyf
f6ea3eadea eres2net
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12640199
2023-05-16 22:28:20 +08:00
yuze.zyz
c6ef9024d7 Fix distributed hook to lazyimport; fix an import bug
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12638653
2023-05-16 21:37:12 +08:00
yuze.zyz
38fa4cceed fix transformer example and fix some bugs
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12626375
2023-05-16 15:27:17 +08:00
zsl01670416
71a80173e5 modify examples text_classification, text_generation, token_classification for improved trainer
1.text_classification/run_train.sh
2.text_generation/run_train_mt5.sh, run_train_palm.sh
3.token_classification/finetune_token_classification.py, run_train_mgeo.sh, run_train_structbert.sh
above files were modified to adapt to improved trainer
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12629010
2023-05-16 14:32:43 +08:00
hemu.zp
5804ad2dc1 update multi_modal_embedding example
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12626062
2023-05-16 14:31:26 +08:00
zhifu.gzf
6c8c3a53f8 fix bug in **kwargs duplicated
修复**kwargs中参数重复的bug
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12633542
2023-05-16 13:49:01 +08:00
wenmeng.zwm
3ac31290f2 add description for modelscope 2023-05-15 19:26:21 +08:00
yuze.zyz
e325862be8 add pop for base class parameters
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12622643
2023-05-15 17:46:32 +08:00
yeqinghao.yqh
b9c8c99776 Support mPLUG-Owl model.
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12610417
2023-05-15 16:32:46 +08:00
hemu.zp
1d28c5b730 mv mmdet to cv
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12614719
2023-05-15 11:48:29 +08:00
wenmeng.zwm
a631c17353 Merge branch dev/merge_github_master_0512 into master
Title: Merge github master 

Merge github master 0512

1. https://github.com/modelscope/modelscope/pull/297
2. https://github.com/modelscope/modelscope/pull/289
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12609811
2023-05-15 09:39:28 +08:00
lingcai.wl
0aaaaa13ad [to #49275037] remove demo check and fix service decoder
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12432683
2023-05-14 23:41:40 +08:00
xingjun.wang
98e899b8c2 merge master 2023-05-14 02:09:02 +08:00
xingjun.wang
bf2cd23aa2 cancel a ut 2023-05-14 02:06:11 +08:00
yuze.zyz
febc0365de Support FlexTrain and update the structure of trainer
1. Refactor training_args
2. Refactor hooks
3. Add train_id for push_to_hub
4. Support both output_dir/output_sub_dir for checkpoint_hooks
5. Support copy when hardlink fails when checkpointing
6. Support mixed dataset config file as a CLI argument
7. Add eval txt in output folder
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/12384253
* support the ignorance of file pattern
2023-05-13 12:12:04 +08:00