中文版

主要更新

• 更新默认分支版本（default revision）策略，以提升仓库版本管理的一致性与易用性 #1546 #1553
• Hub API 新增 set_repo_visibility 接口，支持设置模型仓库可见性，并补充可见性参数校验和错误日志提示 #1564 #1568 #1569 #1570
• 新增 Ascend NPU 的 Dockerfile，便于在昇腾 NPU 环境中部署与使用 ModelScope #1550
• snapshot_download 新增 DEFAULT_MAX_WORKERS 参数配置，优化模型快照下载的并发行为 #1566

AIGC 相关能力增强

• 为 AIGC 的 create_model_tag 接口新增参数，支持更灵活的模型标签创建与管理 #1555
• AIGC 模型新增 base_model_sub_type 字段，便于更精细地描述与区分底座模型类型 #1563

Hub 与上传下载体验优化

• 修正 upload_folder 的默认忽略规则，避免重要文件被错误忽略 #1549
• 修复上传模式匹配（upload patterns）相关问题，提升文件上传的准确性 #1567

稳定性与工程改进

• 修复影响 hasattr 行为的实验类型定义问题，避免潜在属性检测异常 #1552
• 动态模块加载逻辑补丁修复，提升动态模块相关功能的稳定性 #1554
• 更新构建文件与 CI 工作流，改进整体构建发布流程 #1551

English Version

Key Updates

• Updated default revision strategy to improve consistency and usability of repository version management #1546 #1553
• Added set_repo_visibility in Hub API to control repository visibility, with input validation and detailed error logging #1564 #1568 #1569 #1570
• Added Ascend NPU Dockerfile to simplify deployment on Ascend NPU environments #1550
• Introduced DEFAULT_MAX_WORKERS in snapshot_download to better control concurrent download behavior #1566

AIGC Enhancements

• Added new arguments for AIGC create_model_tag for more flexible model tag creation and management #1555
• Added base_model_sub_type for AIGC models to describe and distinguish base model types more precisely #1563

Hub & Upload/Download Improvements

• Fixed default ignore patterns in upload_folder to avoid unintentionally skipping important files #1549
• Fixed upload patterns issues to improve accuracy of file uploads #1567

Stability & Engineering

• Fixed experiment type definition that affected hasattr behavior and could cause attribute detection issues #1552
• Patched dynamic module handling to improve stability of dynamic module related features #1554
• Updated build files and CI workflows to refine the overall build and release pipeline #1551

What's Changed

• Feat/update default revision by @alcholiclg in #1546
• Fix default ignore patterns for upload_folder by @wangxingjun778 in #1549
• Fix exp type which affects hasattr by @tastelikefeet in #1552
• Patch dynamic module by @tastelikefeet in #1554
• Feat: add new args for aigc create_model_tag by @wangxingjun778 in #1555
• [docker] Add Ascend NPU dockerfile by @baymax591 in #1550
• Update build file and workflow by @Yunnglin in #1551
• [WIP]Feat/update default revision by @alcholiclg in #1553
• Feat: add set_repo_visibility in hub api by @wangxingjun778 in #1564
• Add base_model_sub_type for aigc model by @wangxingjun778 in #1563
• Feat: add DEFAULT_MAX_WORKERS in snapshot_download func by @wangxingjun778 in #1566
• Fix/upload patterns by @wangxingjun778 in #1567
• fix set_repo_visibility by @suluyana in #1568
• Fix: add error log for invalid input visibility args by @wangxingjun778 in #1569
• fix: check tasks in HubApi.set_repo_visibility by @suluyana in #1570

New Contributors

• @baymax591 made their first contribution in #1550

Full Changelog: v1.32.0...v1.33.0

中文版

1. 支持使用 pyproject.toml 来安装 modelscope 库，保证与之前安装方式兼容 #1542
2. hub api 的 list_datasets 接口切换为 OpenAPI 端点 #1532
3. AIGC 模型支持 model_source 参数 #1530
4. ModelScope Hub 添加定期自动上传功能：自动定期上传本地文件夹到 ModelScope Hub，增量提交日志、训练产物等文件 #1480
5. AIGC 的 cover_images 支持 mp4 文件 #1533
6. 为 .safetensors 等单个模型文件添加内部云加速 #1529

English Version

1. Support using pyproject.toml to install the modelscope library, ensuring compatibility with previous installation methods #1542
2. Switch the list_datasets interface of hub API to OpenAPI endpoint #1532
3. AIGC models support the model_source parameter #1530
4. Add scheduled automatic upload functionality to ModelScope Hub: automatically upload local folders to ModelScope Hub at regular intervals, with incremental commits of logs, training artifacts, and other files #1480
5. AIGC's cover_images now support mp4 files #1533
6. Add internal cloud acceleration for single model files such as .safetensors #1529

What's Changed

• Feat/support commit scheduler by @alcholiclg in #1480
• Add INTRA_CLOUD_ACCELERATION for single model file downloading by @wangxingjun778 in #1529
• Feat: add model source for aigc model by @wangxingjun778 in #1530
• Add trust_remote_code for _download_additional_modules by @wangxingjun778 in #1521
• Update list datasets to OpenAPI by @Yunnglin in #1532
• [Feature] Add pyproject.toml by @Yunnglin in #1542
• Remove raise FileIntegrityError by @wangxingjun778 in #1517
• [WIP]change DEFAULT_MODEL_REVISION to master by @alcholiclg in #1544
• update ollama: granite4,deepseek-v3.1,gpt-oss by @xiaojianpinga in #1531
• Add mp4 support for AIGC cover_images by @wangxingjun778 in #1533
• fix daily regression by @Yunnglin in #1545

New Contributors

• @xiaojianpinga made their first contribution in #1531

Full Changelog: v1.31.0...v1.32.0

English Version

1. Integrate support for "official tags" of AIGC models. #1509
2. Extend the network timeout duration for model_revision checking. #1507
3. Update the ms-swift Docker build process. #1502
4. Fix some bugs.

中文版

1. 集成对 AIGC 模型“官方标签”的支持。 #1509
2. 延长 model_revision 检查的网络超时时间。 #1507
3. 更新 ms-swift Docker 构建流程。 #1502
4. 修复了一些bugs。

What's Changed

• [docker] update swift docker by @Jintao-Huang in #1502
• Update timeout for checking model revision by @wangxingjun778 in #1507
• Fix get_domain and endpoint by @wangxingjun778 in #1511
• Fix SyntaxError warning on MODELSCOPE_ASCII by @MichaelCurrie in #1512
• [fix] ans_dfsmn_pipeline.py resample by @weedge in #1510
• fix tokenizer download by @tastelikefeet in #1516
• Feat/add official tag by @wangxingjun778 in #1509

New Contributors

• @MichaelCurrie made their first contribution in #1512
• @weedge made their first contribution in #1510

Full Changelog: v1.30.0...v1.31.0

English Version

1. Added support for creating model tags via API or CLI. (@Koko-ry, #1487)
2. Extended repo_info functionality to include the siblings field, enabling better metadata inspection. (@alcholiclg, #1490)
3. Support for uploading AIGC model cover images in Base64 format. (@Koko-ry, #1492)
4. Fixed compatibility issues with models when using the latest version of Hugging Face Transformers. (@tastelikefeet, #1491)
5. Resolved linting errors in utility modules to improve code quality and consistency. (@wangxingjun778, #1493)
6. Corrected endpoint behavior for AI-generated content (AIGC) services. (@Koko-ry, #1494)
7. Fixed invalid revision reference in repo_info responses, ensuring accurate version tracking. (@alcholiclg, #1496)
8. Improved model tag descriptions by correcting formatting and enhancing clarity. (@Koko-ry, #1500)

中文版本

1. 新增通过 API 或 CLI 创建模型标签的功能。(@Koko-ry, #1487)
2. 扩展 repo_info 功能，支持返回 siblings 字段，提升元数据查看能力。(@alcholiclg, #1490)
3. 支持使用Base64格式上传AIGC模型的封面图。 (@Koko-ry, #1492)
4. 修复与最新版 Hugging Face Transformers 的模型兼容性问题。(@tastelikefeet, #1491)
5. 修复工具模块中的代码规范（lint）错误，提升代码质量与一致性。(@wangxingjun778, #1493)
6. 修正 AI 生成内容（AIGC）服务端点的行为逻辑。(@Koko-ry, #1494)
7. 修复 repo_info 响应中无效的版本号引用，确保版本追踪准确无误。(@alcholiclg, #1496)
8. 优化模型标签描述内容，修正格式并提升可读性。(@Koko-ry, #1500)

What's Changed

• fix models with the latest transformers by @tastelikefeet in #1491
• add create model tag by @Koko-ry in #1487
• add support for siblings field in repo_info by @alcholiclg in #1490
• Fix/upload base64 by @Koko-ry in #1492
• Fix utils lint by @wangxingjun778 in #1493
• fix aigc endpoint by @Koko-ry in #1494
• Fix invalid revision in repo_info by @alcholiclg in #1496
• Fix model tag des by @Koko-ry in #1500

Full Changelog: v1.29.2...v1.30.0

English Version

1. Support repo_info support for retrieving repository-level metadata (datasets, models) and list all commit info
2. Fix some bugs

中文版本

1. 支持获取repo级别的详细信息，以及获取所有的提交记录
2. 修复了一些bug

What's Changed

• Fix/upload commit by @wangxingjun778 in #1459
• Fix: add error log for validate_blob and upload_blob by @wangxingjun778 in #1482
• Fix:save_cached_files for sync dumping by @wangxingjun778 in #1484
• Feat/support repo info by @alcholiclg in #1483

New Contributors

• @alcholiclg made their first contribution in #1483

Full Changelog: v1.29.1...v1.29.2

English Version

1. Support arg parser when python version >= 3.12, thanks to the contribution of @yaqiangsun
2. Fix some bugs, e.g. VLLM cannot download from ModelScope of deepseek/internvl/internlm models

中文版本

1. 兼容python>=3.12时的arg parser，感谢@yaqiangsun的贡献
2. 修复了一些问题，例如VLLM无法从ModelScope下载deepseek和浦江实验室的模型

What's Changed

• Fix intermixed missing in cli_argument_parser.py by @yaqiangsun in #1437
• Fix/aigc weight by @Koko-ry in #1464
• Fix trust_remote_code by @wangxingjun778 in #1462
• fix:file already exist by @Koko-ry in #1465
• Merge 1.29 to master by @tastelikefeet in #1469
• support patch intern and deepseek by @tastelikefeet in #1475
• skip aigc test by @tastelikefeet in #1476

New Contributors

• @yaqiangsun made their first contribution in #1437

Full Changelog: v1.29.0...v1.29.1

English Version

New Features

1. Added support for AIGC model creation interface:

modelscope create <repo_id> --token <token> --visibility <public/private/internal> --aigc --model_path <model_path> --aigc_type <Checkpoint/LoRA/VAE> --base_model_type <base_model_type>

2. Added support for using ModelScope MCP server. Users can interface with MCP plaza through the following method:

from modelscope.hub.mcp_api import MCPApi
server = MCPApi().get_mcp_server('@modelcontextprotocol/fetch')
...

3. Added support for repository creation command:

modelscope create owner/repo_name --token ms-xxx --repo_type model --visibility private --chinese_name 测试模型1 --license MIT

Bug Fixes

1. Fixed some security issues
2. Fixed issues where some models could not run on higher Python versions. These models will now throw exceptions directly, guiding users to use lower Python versions
3. Added additional trust_remote_code validation for models that need to run external code or use torch.load during loading

中文版本

新功能

1. 支持了AIGC模型的创建接口:

modelscope create <repo_id> --token <token> --visibility <public/private/internal> --aigc --model_path <model_path> --aigc_type <Checkpoint/LoRA/VAE> --base_model_type <base_model_type>

3. 增加了对ModelScope MCP server的使用支持，用户可以通过多个接口与MCP广场交互，例如：

from modelscope.hub.mcp_api import MCPApi
server = MCPApi().get_mcp_server('@modelcontextprotocol/fetch')
...
...

4. 支持了创建repo的命令：

modelscope create owner/repo_name --token ms-xxx --repo_type model --visibility private --chinese_name 测试模型1 --license MIT

Bug修复

1. 修复了一些安全问题
2. 修复了部分模型在高版本python下无法运行的问题，这些模型会直接抛出异常，指引用户使用更低版本的python
3. 对部分模型在加载时需要运行外部code或者使用torch.load的情况增加了额外的trust_remote_code判断

What's Changed

• fix pyyaml according to: https://github.com/Anchor0221/CVE-2025-50460 by @tastelikefeet in #1428
• 解决使用uv tool的依赖的问题 by @Tian-Jionglu in #1427
• Merge release 1.28 by @hjh0119 in #1401
• ollama template: qwen3-coder, gemma3n, mistral-small3.2, magistral, devstral by @suluyana in #1434
• add ms-agent into dockerfile by @tastelikefeet in #1436
• fix large log in scan cache test by @Yunnglin in #1439
• Fix: add revision arg for upload_file by @wangxingjun778 in #1440
• fix bug by @tastelikefeet in #1441
• Feat: add atomic capabilities - MCPApi by @Koko-ry in #1426
• Add create repo cli by @wangxingjun778 in #1445
• feat: sentence_embedding pipeline by @suluyana in #1435
• Feat: Add AIGC model creation support to create_model by @Koko-ry in #1431
• merge release1.28 by @wangxingjun778 in #1448
• fix: TestTinynasDamoyoloTrainerSingleGPU ut by @wangxingjun778 in #1449
• fix: fix mcp api ut by @wangxingjun778 in #1450
• fix: fix ci by @wangxingjun778 in #1452
• fix tensorflow cannot load gpu by @tastelikefeet in #1456
• Update default version of python by @tastelikefeet in #1457
• Add create cli warning by @wangxingjun778 in #1458

New Contributors

• @Tian-Jionglu made their first contribution in #1427

Full Changelog: v1.28.2...v1.29.0

V1.28.2 Release Note

English Version

1. Hotfix datasets features for datasets >=3.0.0, <=3.6.0 #1447

中文版本

1. 修复datasets features兼容性问题 #1447

Full Changelog: v1.28.1...v1.28.2

V1.28.1 Release Note

English Version

1. Fix use_cookies for get_model_files in delete_files: #1424
2. Add MODELSCOPE_CREDENTIALS_PATH env for user-defined authorization path: #1423
3. Skip UT TextToSpeechSambertHifigan16kPipelineTest: #1420
4. Add weak file lock for snapshot_download: #1417
5. Fix RCE issue for plugins (add trust_remote_code) for from_pretrained function: #1415
6. Fix get_dataset_infos to avoid list all pages for dataset: #1414

中文版本

1. 修复delete_files函数中的get_model_files调用，当需要鉴权的场景，需要明确传入use_cookies参数 : #1424
2. 增加环境变量MODELSCOPE_CREDENTIALS_PATH，支持用户自定义配置本地授权的存储路径（cookies、session等信息）: #1423
3. 跳过测试用例TextToSpeechSambertHifigan16kPipelineTest（用例失效）: #1420
4. 在snapshot_download中增加weak file lock: #1417
5. 在from_pretrained中，增加 trust_remote_code`参数，避免插件RCE问题: #1415
6. 修复get_dataset_infos函数中的分页问题：#1414

What's Changed

• Fix use_cookies in get_model_files by @wangxingjun778 #1424
• Add MODELSCOPE_CREDENTIALS_PATH for user-defined authorization path by @wangxingjun778 #1423
• Skip UT TextToSpeechSambertHifigan16kPipelineTest temporarily by @Koko-ry #1420
• weak file lock by @tastelikefeet #1417
• Fix RCE issue for plugins by @Koko-ry #1415
• Fix dataset infos by @wangxingjun778 #1414

Full Changelog: v1.28.0...v1.28.1

English Version

• Added Hub download progress callback: Supports custom callback functions for download progress updates. #1380
• Added environment variable configuration: Introduced the SIZE_THRESHOLD_TO_ENFORCE_LFS environment variable to configure the file size threshold for enforcing LFS usage. #1383
• Added wildcard batch deletion: Supports batch deletion of files in model or dataset repositories using wildcards. #1384
• Optimized VAE processing: Improved VAE chunking workflow and reduced VRAM consumption. #1378
• Added Hub upload environment variables: Added environment variable configurations for Hub uploads. #1389
• Compatibility improvement: Added compatibility support for datasets library version 3.6. #1393
• Bug fix: Resolved an issue where AutoProcessor and AutoTokenizer might download incomplete files when using from_pretrained. #1398

中文版本

• 新增 Hub 下载进度回调支持： 支持自定义下载进度更新回调函数。 #1380
• 新增环境变量配置： 添加 SIZE_THRESHOLD_TO_ENFORCE_LFS 环境变量，用于配置强制使用 LFS 的文件大小阈值。 #1383
• 新增通配符批量删除功能： 支持使用通配符批量删除模型或数据集仓库中的文件。 #1384
• 优化 VAE 处理： 优化 VAE 分块处理流程并减少显存占用。#1378
• 新增 Hub 上传环境变量配置： 增加 Hub 上传相关的环境变量配置选项。 #1389
• 兼容性提升： 兼容 datasets 库 3.6 版本。 #1393
• 问题修复： 修复 AutoProcessor 和 AutoTokenizer 使用 from_pretrained 方法时可能出现的文件下载不完整问题。#1398

What's Changed

• [hub] support ProgressCallback by @Jintao-Huang in #1380
• Add env SIZE_THRESHOLD_TO_ENFORCE_LFS by @wangxingjun778 in #1383
• Feat deletefiles by @Koko-ry in #1384
• Merge release 1.27 by @hjh0119 in #1392
• perf(vae): optimize tiling performance and memory usage by @MengAiDev in #1378
• Add upload check envs by @wangxingjun778 in #1389
• Upgrade datasets by @wangxingjun778 in #1393
• Fix ignore files of AutoProcessor by @tastelikefeet in #1398

New Contributors

• @Koko-ry made their first contribution in #1384
• @hjh0119 made their first contribution in #1392
• @MengAiDev made their first contribution in #1378

Full Changelog: v1.27.1...v1.28.0