Mirrors/modelscope
1
0
Fork 0
mirror of https://github.com/modelscope/modelscope.git synced 2025-12-25 12:39:25 +01:00
Code Issues Packages Projects Releases Wiki Activity
1,543 Commits 263 Branches 106 Tags
7659b64cdca47164da16b1f77f8de01735bbfe0c
Commit Graph

1543 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
xingjun.wxj
7659b64cdc downward compatible with v1.3.x and lower version for task_datasets, to fix import errors: 
1. gopro_image_deblurring_dataset
2. reds_image_deblurring_dataset
3. sidd_image_denoising
4. video_summarization_dataset
5. ASRDataset
 2023-03-13 16:00:33 +08:00
wucong.lyb
8e88ff46fc change funasr version to 0.3.0 
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11956032
 2023-03-13 11:24:41 +08:00
lee.lcy
104213e4bf feat(thirdparty): add ADADET && add thirdparty arg for damoyolo trainer. 
feat(thirdparty): add ADADET && add thirdparty arg for damoyolo trainer.
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11957959

    * feat(thirdparty): add ADADET && add thirdparty arg for damoyolo trainer.
 2023-03-13 11:20:26 +08:00
lllcho.lc
38bcd54ee4 fix bug for install detecron2 
从github安装库有时会因为网络问题导致失败，故改成gitee
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11945563
 2023-03-10 15:14:21 +08:00
wucong.lyb
2286e7be08 fix bug：modify function generate_scp_from_url 
对于多通道url音频，目前maas-lib解析出来的bytes 每个通道的数据都混到了一起，因此做出修改：用户输入url音频，maas-lib负责下载到本地，然后把本地音频路径传给funasr，然后由funasr做音频解析
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11948914

    * modify function generate_scp_from_url
 2023-03-10 15:14:07 +08:00
hemu.zp
eecdd90412 add finetune_text_generation 
1. Add TrainingArgs and cli call form for text generation task
2. Fix dp+tp finetune bug
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11929345
 2023-03-10 15:08:56 +08:00
yzhao
bd12695d47 Update modelscope version to 1.4.0 2023-03-10 15:03:00 +08:00
yzhao
c505c6027e Merge branch 'master-merge-github20230310' 2023-03-10 14:58:34 +08:00
yzhao
f8a6362a22 pre-commit passed 2023-03-10 14:06:38 +08:00
kangxiaoyang.kxy
681afac803 add finetune for ddcolor image colorization 
图像上色模型添加finetune支持，本地测试已通过：python tests/test_image_colorization_trainer.py

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11874339
 2023-03-10 14:00:48 +08:00
yzhao
99e94bc2c2 Merge branch 'master-github' into master-merge-github20230310 2023-03-10 13:52:31 +08:00
james.wjg
1ce28bc4f9 cv/video_instance_segmentation 
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11891175
 2023-03-10 10:04:03 +08:00
hemu.zp
429cfee826 Support run text generation pipeline with args 
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11937122
 2023-03-10 09:48:10 +08:00
xingjun.wxj
e02a260c93 Refactor the task_datasets module 
Refactor the task_datasets module:

1. Add new module modelscope.msdatasets.dataset_cls.custom_datasets.
2. Add new function: modelscope.msdatasets.ms_dataset.MsDataset.to_custom_dataset().
2. Add calling to_custom_dataset() func in MsDataset.load() to adapt new custom_datasets module.
3. Refactor the pipeline for loading custom dataset: 
	1) Only use MsDataset.load() function to load the custom datasets.
	2) Combine MsDataset.load() with class EpochBasedTrainer.
4. Add new entry func for building datasets in EpochBasedTrainer: see modelscope.trainers.trainer.EpochBasedTrainer.build_dataset()
5. Add new func to build the custom dataset from model configuration, see: modelscope.trainers.trainer.EpochBasedTrainer.build_dataset_from_cfg()
6. Add new registry function for building custom datasets, see: modelscope.msdatasets.dataset_cls.custom_datasets.builder.build_custom_dataset()
7. Refine the class SiameseUIETrainer to adapt the new custom_datasets module.
8. Add class TorchCustomDataset as a superclass for custom datasets classes.
9. To move modules/classes/functions:
	1) Move module msdatasets.audio to custom_datasets
	2) Move module msdatasets.cv to custom_datasets
	3) Move module bad_image_detecting to custom_datasets
	4) Move module damoyolo to custom_datasets
	5) Move module face_2d_keypoints to custom_datasets
	6) Move module hand_2d_keypoints to custom_datasets
	7) Move module human_wholebody_keypoint to custom_datasets
	8) Move module image_classification to custom_datasets
	9) Move module image_inpainting to custom_datasets
	10) Move module image_portrait_enhancement to custom_datasets
	11) Move module image_quality_assessment_degradation to custom_datasets
	12) Move module image_quality_assmessment_mos to custom_datasets
	13) Move class LanguageGuidedVideoSummarizationDataset to custom_datasets
	14) Move class MGeoRankingDataset to custom_datasets
	15) Move module movie_scene_segmentation custom_datasets
	16) Move module object_detection to custom_datasets
	17) Move module referring_video_object_segmentation to custom_datasets
	18) Move module sidd_image_denoising to custom_datasets
	19) Move module video_frame_interpolation to custom_datasets
	20) Move module video_stabilization to custom_datasets
	21) Move module video_super_resolution to custom_datasets
	22) Move class GoproImageDeblurringDataset to custom_datasets
	23) Move class EasyCVBaseDataset to custom_datasets
	24) Move class ImageInstanceSegmentationCocoDataset to custom_datasets
	25) Move class RedsImageDeblurringDataset to custom_datasets
	26) Move class TextRankingDataset to custom_datasets
	27) Move class VecoDataset to custom_datasets
	28) Move class VideoSummarizationDataset to custom_datasets
10. To delete modules/functions/classes:
	1) Del module task_datasets
	2) Del to_task_dataset() in EpochBasedTrainer
	3) Del build_dataset() in EpochBasedTrainer and renew a same name function.
11. Rename class Datasets to CustomDatasets in metainfo.py

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11872747
 2023-03-10 09:03:32 +08:00
yanwen.pyl
fc7daea9c2 add soonet for video temporal grounding 
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11810444
 2023-03-10 09:02:39 +08:00
zhangzhicheng.zzc
8a19e9645d [to #47860410]plugin with cli tool 
1. 支持 plugin方式接入外部 repo、github repo，本地repo，并进行外部插件管理
2. 支持allow_remote方式接入modelhub repo，该类型属于model 范畴不做额外插件管理
3. 支持cli 安装plugin相关

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11775456
 2023-03-09 23:07:13 +08:00
hemu.zp
2b1af959d5 Convert cfg during training 
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11900238
 2023-03-09 22:27:44 +08:00
jinmao.yk
4078abf488 add human reconstruction task 
单图人体重建任务 

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11778199

 * add human reconstruction task
 2023-03-09 21:58:48 +08:00
dave.ma
f493e33720 vidt模型代码评审 
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11873585

* vidt_v0
 2023-03-09 21:52:18 +08:00
wucong.lyb
d9b34daa79 add task: speech_timestamp && fix vad bugs(audio_fs is invalid) 
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11940358

* modify punctuation postprocess

* modify audio input types

* add timestamp args

* add task: speech_timestamp

* modidy timestamp outputs

* add args:seg_dict&split_with_space

* feat: Fix conflict, auto commit by WebIDE
 2023-03-09 21:51:04 +08:00
yinyueqin.yyq
b5fe11fea0 upload disco guided diffusion 
把https://huggingface.co/IDEA-CCNL/Taiyi-Diffusion-532M-Cyberpunk-Chinese和https://huggingface.co/IDEA-CCNL/Taiyi-Diffusion-532M-Nature-Chinese迁移到MaaS-lib上。该project基于disco diffusion+guided diffusion。

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11818412

* upload disco guided diffusion
 2023-03-09 21:45:33 +08:00
yuze.zyz
a0bc5549a1 trainer support parallel_groups 
Design doc: https://yuque.alibaba-inc.com/suluyan.sly/yh1rvu/yx0owblyebpa2b3l?singleDoc#flU3s

1. Add parallel_group field in trainer to support DP, TP, PP.
2. Move the construction of common hooks(except optimizer/lrscheduler hook) to trainer's init method to support after_init stage.
	after_init is to support DP, TP, PP's initializing
         https://aone.alibaba-inc.com/v2/workitem#viewIdentifier=1c46ee8637e0c978f115b6f7&openWorkitemIdentifier=48099986
3. Add before_eval/after_eval stage to support model wrapping.
	to solve the order problem of apex amp & ddp wrapping.
         https://aone.alibaba-inc.com/v2/workitem#viewIdentifier=1c46ee8637e0c978f115b6f7&openWorkitemIdentifier=48099986
4. Exporter supports lazy importing.
	https://aone.alibaba-inc.com/v2/workitem#viewIdentifier=1c46ee8637e0c978f115b6f7&openWorkitemIdentifier=48122780
5. Fold all megatron imports to megatron hook.
         https://aone.alibaba-inc.com/v2/workitem#viewIdentifier=1c46ee8637e0c978f115b6f7&openWorkitemIdentifier=48099986
6. Add compile method to TorchModel ,Pipeline,Trainer to support torch2.0
	https://aone.alibaba-inc.com/v2/workitem#viewIdentifier=1c46ee8637e0c978f115b6f7&openWorkitemIdentifier=46869415
7. Fix bug: Lrscheduler builder does not support torch2.0
8. Add callbacks for trainer
	https://aone.alibaba-inc.com/v2/workitem#viewIdentifier=1c46ee8637e0c978f115b6f7&openWorkitemIdentifier=48210342
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11849932
 2023-03-09 21:33:35 +08:00
yuanzhi.zyz
2ce41d4974 add ocr_reco_crnntrain 
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11824395
 2023-03-09 21:14:09 +08:00
bin.xue
281d3c630e [to #42322933] feat: kws support continue training from a checkpoint 
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11940446

* feat: kws support continue training from a checkpoint

* log: add loading model log
 2023-03-09 21:07:34 +08:00
yuze.zyz
a2bed42fe1 fix a CI case 
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11940845
 2023-03-09 19:55:22 +08:00
lllcho.lc
3e70ecfb37 add action detection finetune 
添加action-detection模型finetune支持

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11873651
 2023-03-09 16:54:19 +08:00
Louis
8abfffc7e5 import math to fix NameError (#147) 2023-03-09 15:35:32 +08:00
neo.dzh
86d5c2c5a8 fixbug for speaker_verification_pipeline and speaker_diarization_pipeline: re-write the default config with configure.json 
修复speaker_verification_pipeline和speaker_diarization_pipeline中configuration.json配置不生效的bug：
使用configuration.json中有的参数覆盖default

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11932226
 2023-03-09 15:20:53 +08:00
mulin.lyh
62f769877f [to #48291018]fix: fix data releate case failed bug 
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11925637
 2023-03-09 15:18:50 +08:00
xixing.tj
46799325c8 add ocr_detection_db training module 
新增ocr_detection dbnet训练代码

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11892455
 2023-03-09 15:11:27 +08:00
feiyu.gfy
f534fd3ae0 add lore lineness table recognition 
读光无线表格识别（lineless-table-recognition）

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11772358
 2023-03-09 15:09:53 +08:00
tastelikefeet
cf8d2d574f add conversion script for checkpoint files before version 1.3.1. (#161) 
* add conversion script for older checkpoint files.

* fix format
 2023-03-09 14:58:21 +08:00
jiaqi.sjq
8099cc9f60 [to #41669377] Refactoring by using kantts as third party wheel and add ptts finetune&infer support 
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11913929
 2023-03-09 11:32:44 +08:00
Patrick von Platen
2bf4ead3fb Correct DDIM sampling on GPU (#173) 2023-03-09 11:21:56 +08:00
ru.he
9307b071b7 Add the code for PEER model 
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11873170
 2023-03-09 01:26:27 +08:00
qize.yqz
0efc892196 add smoke and fire detection model using damoyolo 
新增垂域目标检测模型-基于DAMO_YOLO的烟火检测模型

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11880693
 2023-03-09 01:25:06 +08:00
lingchen.zlm
a5b9d3378f [to #42322933] add generative multimodal embedding model RLEG 
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11867878

* add generative multimodal embedding model RLEG

* remove useless import in rleg model
 2023-03-09 01:17:58 +08:00
gongbiao.gb
13752fa0c0 add vop_se for text video retrival 
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11719262
 2023-03-09 01:14:47 +08:00
lanjinpeng.ljp
a10e59c8f3 add ProContEXT model for video single object tracking 
支持ProContEXT视频单目标跟踪-通用领域模型

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11904797
 2023-03-09 01:12:58 +08:00
lee.lcy
621539f6b6 [to #42322933] add video streaming perception models longshortnet 
add video streaming perception models longshortnet

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11908652
 2023-03-09 01:09:40 +08:00
mulin.lyh
e7b9b1ca24 [to #47783159]fix: simply snapshot_download log 
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11908733
 2023-03-09 01:03:26 +08:00
shimin.ysm
dc53c42cea image depth estimation BTS model code license solve 
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11920166
 2023-03-09 00:58:18 +08:00
wangjiuniu.wjn
0ca0a8c134 add text-to-video-synthesis 
文本生成视频（text-to-video-synthesis）代码

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11767775
 2023-03-09 00:57:43 +08:00
wangmeng.xwm
a17598b13d add MAN for image-quality-assessment 
add MAN model for image-quality-assessment

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11890708

* add MAN for image-quality-assessment

* update and fix bugs in init file
 2023-03-09 00:44:55 +08:00
bin.xue
de67aa28e6 [to #42322933] feat: optimize kws pipeline and training conf 
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11897822
 2023-03-09 00:37:51 +08:00
liugao.lg
c28fd09d42 add more cli tools 
提供独立repo库快速接入modelscope的小工具，模板创建，模型上传，一键导入、版本管理等命令

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11757517
 2023-03-09 00:17:11 +08:00
wucong.lyb
9825c9d469 modify audio input types && punc postprocess 
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11899315

* modify punctuation postprocess

* modify audio input types

* add timestamp args
 2023-03-09 00:08:19 +08:00
mulin.lyh
a8365bb17f [to #48269383]fix: onnx protobuf version conflict 
Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11924445
 2023-03-08 21:58:43 +08:00
zeyinzi.jzyz
bf3a2b6c09 support vision efficient tuning finetune 
## 查看改动点 ↓↓↓
### vision efficient tuning finetune
- Model模块改造成适配训练的
- Model模块在支持训练同时向下兼容之前发布的modecard
- Pipline兼容modelcard加载的preprocessor或直接定义的
- 添加 ImageClassificationPreprocessor （非mmcv版本）
- 添加 VisionEfficientTuningTrainer
- ~~添加 opencv_transforms==0.0.6~~ (以源代码引入必要)

### Modelcard
- test pipeline和trainer合并到一起
- 新增3个模型的test
- 新增demo service

### 公共组件
- ms_dataset.py: fix warning, [UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or xxx]
- preprocessor添加common：ToNumpy、Rename、Identity
- preprocessor common对于dict进行key判断再取值。
- ~~修复learning rate在iter级别变化的逻辑。~~ （本次不做了）
- ~~修复非dist状态下train data没有进行shuffle的bug。~~ (Master已有人改了)
- 修复训练时调用util中非cv包的异常 zhconv。

### 其他
- 为防止新引入的preprocessor模块在config中被原代码加载，导致在其他人做CI时会报错；所以暂时没有添加新的tag，等CR完成后，会进行打tag再rerun CI。
        Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11762108

* support vision efficient tuning finetune

* update test case

* update shuffle on IterableDataset

* update bitfit & sidetuning

* compatible with base trainer
 2023-03-08 16:42:23 +08:00
xingjun.wxj
8298a3c31d add force redownload for nerf dataset in ut 
add force redownload for nerf dataset in ut

Link: https://code.alibaba-inc.com/Ali-MaaS/MaaS-lib/codereview/11915271
 2023-03-08 16:23:58 +08:00