0
代码
代码
Issues
Pull Requests
讨论
Wiki
分析
0
ascend-robotascend-robot[bugfix] fix qwen-image-edit quant error
04166287创建于 2 天前历史提交
文件最后提交记录最后更新时间
Flux
[docs] fix bug about multimodal_sd example Co-authored-by: caishengcheng<caishengcheng@huawei.com> # message auto-generated for no-merge-commit merge: !422 merge docs into master [docs] fix bug about multimodal_sd example Created-by: caishengcheng Commit-by: caishengcheng Merged-by: ascend-robot Description: 感谢您贡献的Pull Request！ 在提交之前，请务必阅读 [CONTRIBUTING.md](https://gitcode.com/Ascend/msmodelslim/blob/master/CONTRIBUTING.md)。 Thanks for sending a pull request! BEFORE SUBMITTING, PLEASE READ [CONTRIBUTING.md](https://gitcode.com/Ascend/msmodelslim/blob/master/CONTRIBUTING.md). ## PR描述 (What this PR does / why we need it?) 因为多模态生成模型一键量化适配了quant_type参数，同步修改example相关描述 See merge request: Ascend/msmodelslim!4228 天前
HunYuanVideo
[bugfix] fix qwen-image-edit quant error Co-authored-by: caishengcheng<caishengcheng@huawei.com> # message auto-generated for no-merge-commit merge: !456 merge bugfix into master [bugfix] fix qwen-image-edit quant error Created-by: caishengcheng Commit-by: caishengcheng Merged-by: ascend-robot Description: 感谢您贡献的Pull Request！ 在提交之前，请务必阅读 [CONTRIBUTING.md](https://gitcode.com/Ascend/msmodelslim/blob/master/CONTRIBUTING.md)。 Thanks for sending a pull request! BEFORE SUBMITTING, PLEASE READ [CONTRIBUTING.md](https://gitcode.com/Ascend/msmodelslim/blob/master/CONTRIBUTING.md). ## PR描述 (What this PR does / why we need it?) 同步mindie仓对custom_op、register_fake接口的适配，解决qwen-image-edit量化报错的issue 参考链接：https://modelers.cn/models/MindIE/Qwen-Image-Edit-2509/blob/main/run_edit_2509.py See merge request: Ascend/msmodelslim!4562 天前
OpenSoraPlanV1_2
修复example、ssz、histogram资料问题 Co-authored-by: libarry<870390541@qq.com> # message auto-generated for no-merge-commit merge: !278 merge docs_bugfix into master 【docs】修复example、ssz、histogram资料问题 Created-by: libarry Commit-by: libarry Merged-by: ascend-robot Description: 感谢您贡献的Pull Request！ 在提交之前，请务必阅读 [CONTRIBUTING.md](https://gitcode.com/Ascend/msmodelslim/blob/master/CONTRIBUTING.md)。 Thanks for sending a pull request! BEFORE SUBMITTING, PLEASE READ [CONTRIBUTING.md](https://gitcode.com/Ascend/msmodelslim/blob/master/CONTRIBUTING.md). ## PR描述 (What this PR does / why we need it?) - 请明确说明您提交PR的变更内容。本部分旨在概述所做的变更，以及此PR是如何解决该问题的。请尽可能地提供有助于评审人员更高效、更快速完成检视审查的实用说明。 - 请说明为何需要这些更改，例如具体的使用场景或bug描述。 - 关联issue号（如果有）。 - Please clarify what changes you are proposing. The purpose of this section is to outline the changes and how this PR fixes the issue. If possible, please consider writing useful notes for better and faster reviews in your PR. - Please clarify why the changes are needed. For instance, the use case and bug description. - Related issue number (if any) ## 面向用户的变更 (Does this PR introduce _any_ user-facing change)? - 请注意，这里指的是**任何**面向用户的变更，包括但不限于API、用户界面或其他使用方式上的变更。 - Note that it means *any* user-facing change including all aspects such as API, interface or other behavior changes. ## 功能验证 (How was this patch tested?) 请确认CI已通过增量及存量的单元测试用例。 如果本次测试方式与常规单元测试不同，请详细说明您的测试步骤(最好提供完整的可复现的操作路径及关键截图)，以便Committer能够快速复现验证，也便于后续的维护。 如果未添加测试，请说明未添加的原因，以及为何难添加测试。 - [_] 功能自验 - [_] 本地自验截图(涉及个人标识符等敏感信息请注意脱敏) - [_] 新增/变更内容是否已新增/适配UT测试用例看护 CI passed with new added/existing test. If it was tested in a way different from regular unit tests, please clarify how you tested step by step, ideally copy and paste-able, so that other reviewers can test and check, and descendants can verify in the future. If tests were not added, please describe why they were not added and/or why it was difficult to add. - [_] Self-verification of the feature. - [_] Screenshot of local self-verification (please anonymize any sensitive information such as personal identifiers) - [_] Have new or modified unit test (UT) cases been added or adapted to cover the newly added or changed content? See merge request: Ascend/msmodelslim!2782 个月前
QwenImageEdit
[bugfix] fix qwen-image-edit quant error Co-authored-by: caishengcheng<caishengcheng@huawei.com> # message auto-generated for no-merge-commit merge: !456 merge bugfix into master [bugfix] fix qwen-image-edit quant error Created-by: caishengcheng Commit-by: caishengcheng Merged-by: ascend-robot Description: 感谢您贡献的Pull Request！ 在提交之前，请务必阅读 [CONTRIBUTING.md](https://gitcode.com/Ascend/msmodelslim/blob/master/CONTRIBUTING.md)。 Thanks for sending a pull request! BEFORE SUBMITTING, PLEASE READ [CONTRIBUTING.md](https://gitcode.com/Ascend/msmodelslim/blob/master/CONTRIBUTING.md). ## PR描述 (What this PR does / why we need it?) 同步mindie仓对custom_op、register_fake接口的适配，解决qwen-image-edit量化报错的issue 参考链接：https://modelers.cn/models/MindIE/Qwen-Image-Edit-2509/blob/main/run_edit_2509.py See merge request: Ascend/msmodelslim!4562 天前
SD3
修复example、ssz、histogram资料问题 Co-authored-by: libarry<870390541@qq.com> # message auto-generated for no-merge-commit merge: !278 merge docs_bugfix into master 【docs】修复example、ssz、histogram资料问题 Created-by: libarry Commit-by: libarry Merged-by: ascend-robot Description: 感谢您贡献的Pull Request！ 在提交之前，请务必阅读 [CONTRIBUTING.md](https://gitcode.com/Ascend/msmodelslim/blob/master/CONTRIBUTING.md)。 Thanks for sending a pull request! BEFORE SUBMITTING, PLEASE READ [CONTRIBUTING.md](https://gitcode.com/Ascend/msmodelslim/blob/master/CONTRIBUTING.md). ## PR描述 (What this PR does / why we need it?) - 请明确说明您提交PR的变更内容。本部分旨在概述所做的变更，以及此PR是如何解决该问题的。请尽可能地提供有助于评审人员更高效、更快速完成检视审查的实用说明。 - 请说明为何需要这些更改，例如具体的使用场景或bug描述。 - 关联issue号（如果有）。 - Please clarify what changes you are proposing. The purpose of this section is to outline the changes and how this PR fixes the issue. If possible, please consider writing useful notes for better and faster reviews in your PR. - Please clarify why the changes are needed. For instance, the use case and bug description. - Related issue number (if any) ## 面向用户的变更 (Does this PR introduce _any_ user-facing change)? - 请注意，这里指的是**任何**面向用户的变更，包括但不限于API、用户界面或其他使用方式上的变更。 - Note that it means *any* user-facing change including all aspects such as API, interface or other behavior changes. ## 功能验证 (How was this patch tested?) 请确认CI已通过增量及存量的单元测试用例。 如果本次测试方式与常规单元测试不同，请详细说明您的测试步骤(最好提供完整的可复现的操作路径及关键截图)，以便Committer能够快速复现验证，也便于后续的维护。 如果未添加测试，请说明未添加的原因，以及为何难添加测试。 - [_] 功能自验 - [_] 本地自验截图(涉及个人标识符等敏感信息请注意脱敏) - [_] 新增/变更内容是否已新增/适配UT测试用例看护 CI passed with new added/existing test. If it was tested in a way different from regular unit tests, please clarify how you tested step by step, ideally copy and paste-able, so that other reviewers can test and check, and descendants can verify in the future. If tests were not added, please describe why they were not added and/or why it was difficult to add. - [_] Self-verification of the feature. - [_] Screenshot of local self-verification (please anonymize any sensitive information such as personal identifiers) - [_] Have new or modified unit test (UT) cases been added or adapted to cover the newly added or changed content? See merge request: Ascend/msmodelslim!2782 个月前
Wan2_1
[doc]资料整改 Co-authored-by: caishengcheng<caishengcheng@huawei.com> # message auto-generated for no-merge-commit merge: !252 merge doc into master [doc]资料整改 Created-by: caishengcheng Commit-by: caishengcheng Merged-by: ascend-robot Description: 感谢您贡献的Pull Request！ 在提交之前，请务必阅读 [CONTRIBUTING.md](https://gitcode.com/Ascend/msmodelslim/blob/master/CONTRIBUTING.md)。 Thanks for sending a pull request! BEFORE SUBMITTING, PLEASE READ [CONTRIBUTING.md](https://gitcode.com/Ascend/msmodelslim/blob/master/CONTRIBUTING.md). ## PR描述 (What this PR does / why we need it?) - 请明确说明您提交PR的变更内容。本部分旨在概述所做的变更，以及此PR是如何解决该问题的。请尽可能地提供有助于评审人员更高效、更快速完成检视审查的实用说明。 - 请说明为何需要这些更改，例如具体的使用场景或bug描述。 - 关联issue号（如果有）。 - Please clarify what changes you are proposing. The purpose of this section is to outline the changes and how this PR fixes the issue. If possible, please consider writing useful notes for better and faster reviews in your PR. - Please clarify why the changes are needed. For instance, the use case and bug description. - Related issue number (if any) ## 面向用户的变更 (Does this PR introduce _any_ user-facing change)? - 请注意，这里指的是**任何**面向用户的变更，包括但不限于API、用户界面或其他使用方式上的变更。 - Note that it means *any* user-facing change including all aspects such as API, interface or other behavior changes. ## 功能验证 (How was this patch tested?) 请确认CI已通过增量及存量的单元测试用例。 如果本次测试方式与常规单元测试不同，请详细说明您的测试步骤(最好提供完整的可复现的操作路径及关键截图)，以便Committer能够快速复现验证，也便于后续的维护。 如果未添加测试，请说明未添加的原因，以及为何难添加测试。 - [_] 功能自验 - [_] 本地自验截图(涉及个人标识符等敏感信息请注意脱敏) - [_] 新增/变更内容是否已新增/适配UT测试用例看护 CI passed with new added/existing test. If it was tested in a way different from regular unit tests, please clarify how you tested step by step, ideally copy and paste-able, so that other reviewers can test and check, and descendants can verify in the future. If tests were not added, please describe why they were not added and/or why it was difficult to add. - [_] Self-verification of the feature. - [_] Screenshot of local self-verification (please anonymize any sensitive information such as personal identifiers) - [_] Have new or modified unit test (UT) cases been added or adapted to cover the newly added or changed content? See merge request: Ascend/msmodelslim!2522 个月前
Wan2_2
[docs] fix bug about multimodal_sd example Co-authored-by: caishengcheng<caishengcheng@huawei.com> # message auto-generated for no-merge-commit merge: !422 merge docs into master [docs] fix bug about multimodal_sd example Created-by: caishengcheng Commit-by: caishengcheng Merged-by: ascend-robot Description: 感谢您贡献的Pull Request！ 在提交之前，请务必阅读 [CONTRIBUTING.md](https://gitcode.com/Ascend/msmodelslim/blob/master/CONTRIBUTING.md)。 Thanks for sending a pull request! BEFORE SUBMITTING, PLEASE READ [CONTRIBUTING.md](https://gitcode.com/Ascend/msmodelslim/blob/master/CONTRIBUTING.md). ## PR描述 (What this PR does / why we need it?) 因为多模态生成模型一键量化适配了quant_type参数，同步修改example相关描述 See merge request: Ascend/msmodelslim!4228 天前
README.md
[doc]资料整改 Co-authored-by: caishengcheng<caishengcheng@huawei.com> # message auto-generated for no-merge-commit merge: !252 merge doc into master [doc]资料整改 Created-by: caishengcheng Commit-by: caishengcheng Merged-by: ascend-robot Description: 感谢您贡献的Pull Request！ 在提交之前，请务必阅读 [CONTRIBUTING.md](https://gitcode.com/Ascend/msmodelslim/blob/master/CONTRIBUTING.md)。 Thanks for sending a pull request! BEFORE SUBMITTING, PLEASE READ [CONTRIBUTING.md](https://gitcode.com/Ascend/msmodelslim/blob/master/CONTRIBUTING.md). ## PR描述 (What this PR does / why we need it?) - 请明确说明您提交PR的变更内容。本部分旨在概述所做的变更，以及此PR是如何解决该问题的。请尽可能地提供有助于评审人员更高效、更快速完成检视审查的实用说明。 - 请说明为何需要这些更改，例如具体的使用场景或bug描述。 - 关联issue号（如果有）。 - Please clarify what changes you are proposing. The purpose of this section is to outline the changes and how this PR fixes the issue. If possible, please consider writing useful notes for better and faster reviews in your PR. - Please clarify why the changes are needed. For instance, the use case and bug description. - Related issue number (if any) ## 面向用户的变更 (Does this PR introduce _any_ user-facing change)? - 请注意，这里指的是**任何**面向用户的变更，包括但不限于API、用户界面或其他使用方式上的变更。 - Note that it means *any* user-facing change including all aspects such as API, interface or other behavior changes. ## 功能验证 (How was this patch tested?) 请确认CI已通过增量及存量的单元测试用例。 如果本次测试方式与常规单元测试不同，请详细说明您的测试步骤(最好提供完整的可复现的操作路径及关键截图)，以便Committer能够快速复现验证，也便于后续的维护。 如果未添加测试，请说明未添加的原因，以及为何难添加测试。 - [_] 功能自验 - [_] 本地自验截图(涉及个人标识符等敏感信息请注意脱敏) - [_] 新增/变更内容是否已新增/适配UT测试用例看护 CI passed with new added/existing test. If it was tested in a way different from regular unit tests, please clarify how you tested step by step, ideally copy and paste-able, so that other reviewers can test and check, and descendants can verify in the future. If tests were not added, please describe why they were not added and/or why it was difficult to add. - [_] Self-verification of the feature. - [_] Screenshot of local self-verification (please anonymize any sensitive information such as personal identifiers) - [_] Have new or modified unit test (UT) cases been added or adapted to cover the newly added or changed content? See merge request: Ascend/msmodelslim!2522 个月前
utils.py
【msmodelslim】刷新文件头部声明和版权 Co-authored-by: caishengcheng<caishengcheng@huawei.com> # message auto-generated for no-merge-commit merge: !28 merge license into master 【msmodelslim】刷新文件头部声明和版权 Created-by: caishengcheng Commit-by: caishengcheng Merged-by: ascend-robot Description: 【msmodelslim】刷新文件头部声明和版权 See merge request: Ascend/msmodelslim!284 个月前
README.md

多模态生成模型量化说明

模型介绍

SD3 Stable Diffusion 3, 由stability.ai发布的强大的文本到图像模型，在多主题提示、图像质量和拼写功能方面的性能得到了大幅提升。

Open-Sora-Plan v1.2 是一个开源的多模态视频生成模型，由北大-兔展AIGC联合实验室共同发起，专注于高效视频生成任务。

Flux.1是由 Black Forest Labs 开发的一款开源的 120 亿参数的图像生成模型，它能够根据文本描述生成高质量的图像。

HunyuanVideo 是腾讯发布的一种新颖的开源视频基础模型，它在视频生成方面的性能可与领先的闭源模型相媲美，甚至优于领先的闭源模型。

Wan2.1 是阿里巴巴发布的一套全面且开放的视频基础模型，它突破了视频生成的界限。支持文本到视频(T2V)、图像到视频(I2V)、文本到图像(T2I)等多种生成任务。

Wan2.2 是阿里巴巴在 Wan 系列上的新一代开源视频基础模型，面向更高质量、更可控的影视级视频生成；在 Wan2.1 的基础上进一步扩充训练数据与能力，并引入面向视频扩散的 混合专家（MoE） 等设计，在保持开放生态的同时提升生成效率与观感。支持 文本到视频（T2V）、图像到视频（I2V） 以及 文本+图像到视频（TI2V） 等多种模式。

Qwen-Image-Edit 是阿里巴巴通义千问团队基于 Qwen-Image 图像基础模型推出的开源图像编辑模型，兼顾语义级改动（如风格、构图、物体增删与替换）与外观级细节控制。支持中英文画面内文字的精准修改。

使用前准备

支持的模型版本与量化策略

模型系列 模型版本 HuggingFace链接 W8A8 W8A16 W4A16 W4A4 稀疏量化 KV Cache Attention 时间步量化 FA3量化 量化命令
SD3 SD3-Medium SD3-Medium ✅ W8A8静态量化
Open-Sora-Plan Open-Sora-Plan v1.2 Open-Sora-Plan v1.2 ✅ W8A8静态量化
FLUX FLUX.1-dev FLUX.1-dev ✅ ✅ ✅ ✅ W8A8静态量化 / W8A8分时间步量化 / FA3+W8A8动态量化 / 异常值抑制+W8A8动态量化
HunyuanVideo HunyuanVideo HunyuanVideo ✅ ✅ ✅ ✅ W8A8静态量化 / W8A8分时间步量化 / FA3+W8A8动态量化 / 异常值抑制+W8A8动态量化
Wan2.1 Wan2.1-T2V-14B Wan2.1-T2V-14B ✅ W8A8动态量化
Wan2.2 Wan2.2-T2V-A14B Wan2.2-T2V-A14B ✅ ✅ FA3+W8A8动态量化
Wan2.2 Wan2.2-I2V-A14B Wan2.2-I2V-A14B ✅ ✅ FA3+W8A8动态量化
Wan2.2 Wan2.2-TI2V-5B Wan2.2-TI2V-5B ✅ ✅ FA3+W8A8动态量化
Qwen-Image-Edit Qwen-Image-Edit-2509 Qwen-Image-Edit-2509 ✅ ✅ FA3+W8A8动态量化

说明：

  • ✅ 表示该量化策略已通过msModelSlim官方验证，功能完整、性能稳定，建议优先采用。
  • 空格表示该量化策略暂未通过msModelSlim官方验证，用户可根据实际需求进行配置尝试，但量化效果和功能稳定性无法得到官方保证。
  • 点击量化命令列中的链接可跳转到对应的具体量化命令。
  • 其中FLUX.1-dev、HunyuanVideo、Wan2.2、Qwen-Image-Edit-2509支持在昇腾950代际产品上运行的mxfp8量化，详情点击查看具体量化命令。

使用示例

使用量化前，需要加载模型和校准数据，其中加载模型依赖于diffusers库（如SD3-Medium）或多模态生成模型魔乐社区推理工程仓（如Open-Sora-Plan v1.2、Flux.1-dev、HunyuanVideo、Wan2.1），请先确保依据推理工程仓可以正常进行浮点推理。

SD3-Medium W8A8静态量化

请参考SD3-Medium 量化使用说明

Open-Sora-Plan v1.2 W8A8静态量化

请参考Open-Sora-Plan V1.2 量化使用说明

FLUX.1-dev W8A8静态量化

请参考FLUX.1-dev 量化使用说明

FLUX.1-dev W8A8分时间步量化

请参考FLUX.1-dev 量化使用说明

FLUX.1-dev FA3+W8A8动态量化

请参考FLUX.1-dev 量化使用说明

FLUX.1-dev 异常值抑制+W8A8动态量化

请参考FLUX.1-dev 量化使用说明

HunyuanVideo W8A8静态量化

请参考HunyuanVideo 量化使用说明

HunyuanVideo W8A8分时间步量化

请参考HunyuanVideo 量化使用说明

HunyuanVideo FA3+W8A8动态量化

请参考HunyuanVideo 量化使用说明

HunyuanVideo 异常值抑制+W8A8动态量化

请参考HunyuanVideo 量化使用说明

Wan2.1 W8A8动态量化

请参考Wan2.1 量化使用说明

Wan2.2 FA3+W8A8动态量化

请参考Wan2.2 量化使用说明

Qwen-Image-Edit-2509 FA3+W8A8动态量化

请参考Qwen Image Edit 量化使用说明