文件最后提交记录最后更新时间
[Docs] Document corrections Co-authored-by: js1234567<jiangshuo9@h-partners.com> # message auto-generated for no-merge-commit merge: !2109 merge 2.3.0 into 2.3.0 [Docs] Document corrections Created-by: js1234567 Commit-by: js1234567 Merged-by: ascend-robot Description: ## Motivation Document corrections: 1. 添加2.3.0配套信息 2. 中英文标点问题 3. 链接版本更新 4. CANN8.5.0版本配置环境变量刷新, 涉及环境变量设置需全面排查修改 ## Modification Readme.md shells ## Self-test (Optional) If modifications to this PR may cause/fix function/accuracy/performance DTSs/issues, a self-inspection record needs to be attached. ## BC-breaking (Optional) If there are compatibility issues, such as dependencies on cann/torch_npu versions, they need to be explained in the PR. ## Checklist **Before PR**: - [x] The new code needs to comply with the Clean Code specification. - [x] The PR content is self-checked, and the expression can be clear and the writing standardized **After PR**: - [ ] CLA has been signed and all committers have signed the CLA in this PR. - [ ] The ci-pipeline is passed, Code Check is passed. See merge request: Ascend/MindSpeed-MM!21093 个月前
[Docs] Document corrections Co-authored-by: js1234567<jiangshuo9@h-partners.com> # message auto-generated for no-merge-commit merge: !2109 merge 2.3.0 into 2.3.0 [Docs] Document corrections Created-by: js1234567 Commit-by: js1234567 Merged-by: ascend-robot Description: ## Motivation Document corrections: 1. 添加2.3.0配套信息 2. 中英文标点问题 3. 链接版本更新 4. CANN8.5.0版本配置环境变量刷新, 涉及环境变量设置需全面排查修改 ## Modification Readme.md shells ## Self-test (Optional) If modifications to this PR may cause/fix function/accuracy/performance DTSs/issues, a self-inspection record needs to be attached. ## BC-breaking (Optional) If there are compatibility issues, such as dependencies on cann/torch_npu versions, they need to be explained in the PR. ## Checklist **Before PR**: - [x] The new code needs to comply with the Clean Code specification. - [x] The PR content is self-checked, and the expression can be clear and the writing standardized **After PR**: - [ ] CLA has been signed and all committers have signed the CLA in this PR. - [ ] The ci-pipeline is passed, Code Check is passed. See merge request: Ascend/MindSpeed-MM!21093 个月前
[Docs][mindspore] Update version info for branch 2.3.0 Co-authored-by: weixin_47897441<wuyouqi1@h-partners.com> # message auto-generated for no-merge-commit merge: !2013 merge 2.3.0 into 2.3.0 [Docs][mindspore] Update version info for branch 2.3.0 Created-by: weixin_47897441 Commit-by: weixin_47897441 Merged-by: ascend-robot Description: ## Motivation Update installation version info for the 2.3.0 commercial release branch. ## Modification Updated installation-related version details of the 2.3.0 commercial branch. ## Self-test (Optional) No functional/performance impacts; version info verified consistent with branch requirements. ## BC-breaking (Optional) No compatibility issues; no changes to CANN/torch_npu dependencies. ## Checklist **Before PR**: - [x] The new code needs to comply with the Clean Code specification. - [x] The PR content is self-checked, and the expression can be clear and the writing standardized **After PR**: - [x] CLA has been signed and all committers have signed the CLA in this PR. - [x] The ci-pipeline is passed, Code Check is passed. See merge request: Ascend/MindSpeed-MM!20134 个月前
docs: update branch 2.3.0 docs link, switch from master to 2.3.0 Co-authored-by: liyingxuan<liyingxuan3@huawei.com> # message auto-generated for no-merge-commit merge: !2241 merge 2.3.0 into 2.3.0 docs: update branch 2.3.0 docs link, switch from master to 2.3.0 Created-by: liyx616 Commit-by: liyingxuan Merged-by: ascend-robot Description: ## What this PR does / why we need it? 将分支2.3.0链接到master分支的文档全部修改为链接到2.3.0 ## Does this PR introduce any user-facing change? 修复了2.3.0分支的使用体验 ## How was this patch tested? 文档修改,不涉及 See merge request: Ascend/MindSpeed-MM!22412 个月前
readme.md

MindSpore后端支持说明

MindSpeed-MM已支持接入华为自研AI套件MindSpore,旨在提供华为全栈易用的端到端的多模态模型训练解决方案,以此获得更极致的性能体验。MindSpore后端提供了一套对标PyTorch的API,用户无需进行额外代码适配即可无缝切换。


🔥🔥🔥Latest News

🚀🚀🚀 MindSpore后端已支持CogVideoXInternVL2Qwen2.5VLQwen25OmniQwen3VL模型。

版本配套

MindSpeed-MM + MindSpore后端的依赖配套如下表,安装步骤参考基础安装指导

依赖软件
昇腾NPU驱动固件 在研版本
昇腾 CANN 在研版本
MindSpore 2.7.2
Python >=3.10

环境部署

具体部署步骤请查看部署文档

快速上手

快速上手操作请查看快速上手文档


模型/特性介绍

已支持特性概览

模型 \ 特性 TP TP-SP VPP PP CP EP Distributed Optimizer Recomputation LoRA
CogVideoX系列-T2V CP (Ulysses)
CogVideoX系列-I2V CP (Ulysses)
InternVL2-2B
InternVL2-8B
Qwen2.5VL-7B
Qwen2.5VL-72B
Qwen25Omni-7B
Qwen3vl-8B
Qwen3vl-30B

备注:

支持模型

MindSpeed MM (MindSpore后端)模型支持列表
模型任务 模型 参数量 任务 集群 精度格式 NPU性能 参考性能 平均序列长度 支持情况
多模态生成
CogVideoX-T2V 5B 预训练 1x8 BF16 0.46 (SPS) /
CogVideoX-I2V 5B 预训练 1x8 BF16 0.46 (SPS) /
CogVideoX 1.5-T2V 5B 预训练 1x8 BF16 2.09 (SPS) /
5B Lora微调 1x8 BF16 3.03 (SPS) / 支持中
CogVideoX 1.5-I2V 5B 预训练 1x8 BF16 2.01 (SPS) /
5B Lora微调 1x8 BF16 3.92 (SPS) / 支持中
Wan2.1-I2V 1.3B 预训练 1x8 BF16 0.61 (SPS) /
Wan2.1-T2V 1.3B 预训练 1x8 BF16 0.73 (SPS) /
1.3B Lora微调 1x8 BF16 / / 支持中
多模态理解 InternVL 2.0 2B 微调 1x8 BF16 22.46 (SPS) /
8B 微调 1x8 BF16 11.00 (SPS) /
26B 微调 1x8 BF16 3.26 (SPS) / 支持中
76B 全参微调 8x16 BF16 191 (TPS) / 支持中
Qwen2.5-VL 3B 微调 1x8 BF16 21.79 (SPS) 563 支持中
7B 微调 1x8 BF16 12.67 (SPS) 563
32B 微调 2x8 BF16 / 563 支持中
72B 微调 8x8 BF16 256.28 (TPS) 563
Qwen25Omni 7B 微调 1x8 BF16 / /
Qwen3VL 8B 微调 1x8 BF16 / /
30B 微调 1x16 BF16 / /

特性规划

工具使用

昇腾Profiling采集工具

MindSpeed MM集成了昇腾profiling采集工具,以提供对模型运行情况的分析。该工具能够依照配置采集模型的算子、显存等关键信息,同时支持动静态两种采集方式,协助开发者分析模型瓶颈,并可根据实际场景需求选择使用。

具体方法见 README 的profiling章节。

MindStudio Insight性能分析工具

针对大模型集群场景的性能调优,这里推荐一款优秀的可视化调优工具MindStudio Insight。 MindStudio Insight提供了包括Timeline视图、通信分析、计算耗时等的可视化呈现,以便用户分析潜在的性能瓶颈,并指导如何采取措施消除或减少这些瓶颈。

具体使用方法见《MindStudio Insight操作指南》,下载地址《MindStudio Insight》