| 文件 | 最后提交记录 | 最后更新时间 |
|---|---|---|
docs(pytorch): updated doc for AIDD Co-authored-by: LQ1206<liuqian164@h-partners.com> # message auto-generated for no-merge-commit merge: !4768 merge master into master docs(pytorch): updated doc for AIDD Created-by: LQ1206 Commit-by: LQ1206 Merged-by: ascend-robot Description: ## What this PR does / why we need it? updated doc for AIDD ## Does this PR introduce any user-facing change? fix doc error ## How was this patch tested? No test See merge request: Ascend/MindSpeed-LLM!4768 | 10 天前 | |
feat: FSDP2 Longcat-Flash-Lite adaptation for the GEMM function Co-authored-by: wj<wangjin230@huawei.com> # message auto-generated for no-merge-commit merge: !4753 merge fsdp2-gemm into master feat: FSDP2 Longcat-Flash-Lite adaptation for the GEMM function Created-by: gcw_RxnYoBVv Commit-by: wj Merged-by: ascend-robot Description: ## What this PR does / why we need it? LongCat-Flash-Lite supports online weight conversion, Grouped GEMM (GMM) expert computation, and N-gram embedding memory optimization in the FSDP2 scenario. ## Does this PR introduce any user-facing change? The model can directly use the original HuggingFace weights to start FSDP2 training/inference, and reduce the risk of out-of-memory (OOM) caused by dense gradient accumulation in the backward phase of multiple N-gram embedding tables. ## How was this patch tested? Single-machine layer reduction precision alignment, with an absolute error of 0.004. See merge request: Ascend/MindSpeed-LLM!4753 | 19 小时前 | |
docs: fix AIDD review documentation issues Co-authored-by: xuwenyue<xuwenyue1@huawei.com> # message auto-generated for no-merge-commit merge: !4791 merge docs/fix-issue-1610-master into master docs: fix AIDD review documentation issues Created-by: xuwenyue Commit-by: xuwenyue Merged-by: ascend-robot Description: ## What this PR does / why we need it? Fix documentation formatting issues including FAQ markdown formatting, precision_issue table alignment, supported_models content updates, and quick_start improvements. ## Does this PR introduce any user-facing change? Yes, multiple documentation files are updated for better readability and accuracy. ## How was this patch tested? Reviewed the updated markdown content for correctness and formatting. Related issue: #1610 See merge request: Ascend/MindSpeed-LLM!4791 | 7 天前 | |
[pytorch][docs] add docs license Co-authored-by: leizhenzhen23<leizhenzhen4@huawei.com> # message auto-generated for no-merge-commit merge: !4240 merge llm_0225 into master [pytorch][docs] add docs license Created-by: leizhenzhen23 Commit-by: leizhenzhen23 Merged-by: ascend-robot Description: [pytorch][docs] add docs license See merge request: Ascend/MindSpeed-LLM!4240 | 4 个月前 | |
docs(pytorch): fix AIDD review document issues Co-authored-by: zhyebin01<zhangyebin@h-partners.com> # message auto-generated for no-merge-commit merge: !4806 merge master into master docs(pytorch): fix AIDD review document issues Created-by: zhyebin01 Commit-by: zhyebin01 Merged-by: ascend-robot Description: ## What this PR does / why we need it? fix AIDD review document issues ## Does this PR introduce any user-facing change? NA. ## How was this patch tested? doc-ci pass See merge request: Ascend/MindSpeed-LLM!4806 | 7 天前 | |
fix(pytorch): checkpoint_convert_hf_mcore.md sunset Co-authored-by: cjy840282<chenjingyi9@huawei.com> # message auto-generated for no-merge-commit merge: !4697 merge ckpt into master fix(pytorch): checkpoint_convert_hf_mcore.md sunset Created-by: cjy840282 Commit-by: cjy840282 Merged-by: ascend-robot Description: ## What this PR does / why we need it? Due to an upgrade in the transformers version, support for convert_ckpt_V1 has been discontinued. Running the example code provided in the checkpoint_convert_hf_mcore.md document for convert_ckpt.py now results in errors. Consequently, this PR has removed the documentation related to convert_ckpt_V1. ## Does this PR introduce any user-facing change? Guide users to use the convert_ckpt_V2, which offers better usability. ## How was this patch tested? This PR primarily involves documentation updates, with no code modifications. It has been tested and verified through manual review. See merge request: Ascend/MindSpeed-LLM!4697 | 19 天前 | |
docs: update README/quickstart/precision_issue/supported_models Co-authored-by: xuwenyue<xuwenyue1@huawei.com> # message auto-generated for no-merge-commit merge: !4550 merge fix_docs_0529 into master docs: update README/quickstart/precision_issue/supported_models Created-by: xuwenyue Commit-by: xuwenyue Merged-by: ascend-robot Description: ## What this PR does / why we need it? update README: 未来规划/社区会议/加入我们 move to bottom, 最新消息 add a details tag, update quickstart: add note add precision_issue.md update supported_models: remove mindspeed-mm supported models table ## Does this PR introduce any user-facing change? Please describe whether the PR will result in any user-facing usage changes. If there is related documentation, please specify its path. ## How was this patch tested? Please explain how to verify the correctness and effectiveness of this feature, as well as its usage constraints and limitations. See merge request: Ascend/MindSpeed-LLM!4550 | 1 个月前 | |
doc(fsdp2): add fsdp2 finetune doc Co-authored-by: tichang<tichang@huawei.com> # message auto-generated for no-merge-commit merge: !4756 merge ddoc/fsdp2_tune_doc into master doc(fsdp2): add fsdp2 finetune doc Created-by: tichang Commit-by: tichang Merged-by: ascend-robot Description: ## What this PR does / why we need it? Update new FSDP2 finetune doc. Delete origin FSDP2 train guide. ## Does this PR introduce any user-facing change? Just add a new document. no introducing any user-facing change. ## How was this patch tested? The content of the document has been all tested. See merge request: Ascend/MindSpeed-LLM!4756 | 2 天前 | |
docs(pytorch): update document Co-authored-by: LQ1206<liuqian164@h-partners.com> # message auto-generated for no-merge-commit merge: !4752 merge master into master docs(pytorch): update document Created-by: LQ1206 Commit-by: LQ1206 Merged-by: ascend-robot Description: ## What this PR does / why we need it? update document ## Does this PR introduce any user-facing change? change and update checkpoint document , update the decs. of introduction ## How was this patch tested? No test See merge request: Ascend/MindSpeed-LLM!4752 | 12 天前 | |
docs: Fix Style and Missing Image Reference in Project Guide Co-authored-by: leizhenzhen23<leizhenzhen4@huawei.com> # message auto-generated for no-merge-commit merge: !4821 merge docs/fix-project-guide-style-master into master docs: Fix Style and Missing Image Reference in Project Guide Created-by: leizhenzhen23 Commit-by: leizhenzhen23 Merged-by: ascend-robot Description: ## What this PR does / why we need it? Fix two minor documentation issues in docs/zh/project_guide.md: 1. Removed redundant spaces before enumeration commas in "dense、MoE 、SSM 、Linear" to align with document style. 2. Replaced "如下图所示" with "如下所示" on line 211, since no image is referenced in the document. ## Does this PR introduce any user-facing change? No. Only documentation text is modified. ## How was this patch tested? Verified by reading the modified file to confirm the changes. See merge request: Ascend/MindSpeed-LLM!4821 | 7 天前 | |
docs(pytorch): updated doc for AIDD Co-authored-by: LQ1206<liuqian164@h-partners.com> # message auto-generated for no-merge-commit merge: !4768 merge master into master docs(pytorch): updated doc for AIDD Created-by: LQ1206 Commit-by: LQ1206 Merged-by: ascend-robot Description: ## What this PR does / why we need it? updated doc for AIDD ## Does this PR introduce any user-facing change? fix doc error ## How was this patch tested? No test See merge request: Ascend/MindSpeed-LLM!4768 | 10 天前 | |
docs(pytorch): fix docs Co-authored-by: qyzqyz<quyueze@h-partners.com> # message auto-generated for no-merge-commit merge: !4812 merge master into master docs(pytorch): fix docs Created-by: qyzqyz Commit-by: qyzqyz Merged-by: ascend-robot Description: ## What this PR does / why we need it? some docs fix ## Does this PR introduce any user-facing change? No. Documentation and type annotation fixes only. No runtime behavior change. ## How was this patch tested? Confirmed no new spelling, formatting, or content errors introduced See merge request: Ascend/MindSpeed-LLM!4812 | 8 天前 | |
docs: add task entry guide and Qwen3 mcore workflow index Co-authored-by: zhouzirui1234<2490136803@qq.com> # message auto-generated for no-merge-commit merge: !4695 merge docs/task-entry-guide into master docs: add task entry guide and Qwen3 mcore workflow index Created-by: zhouzirui-2026 Commit-by: zhouzirui1234 Merged-by: ascend-robot Description: ## What this PR does / why we need it? Adds lightweight documentation to help users locate scripts, configuration sources, and implementation paths for common MindSpeed-LLM workflows. This PR includes: - [Doc] Add docs/zh/task_entry_guide.md to summarize task entry scripts, example scripts, configuration sources, and implementation paths. - [Doc] Add examples/mcore/qwen3/README.md as a Qwen3 mcore workflow index for data preprocessing, checkpoint conversion, training, inference, and evaluation examples. - [Doc] Update docs/zh/docs_guide.md to link the new task entry guide. Fixes #1545. ## Does this PR introduce any user-facing change? Yes. This PR adds documentation only. It provides an additional navigation guide for users who need to find workflow entry points and related example scripts. This PR does not modify training, inference, evaluation, data preprocessing, checkpoint conversion logic, example scripts, or configs. ## How was this patch tested? - Ran git diff --check. - Checked local Markdown links and referenced paths in the updated documents. - Checked that the PR branch is rebased on the latest master and has no merge conflicts. - Did not run large-scale training because this PR only changes documentation. See merge request: Ascend/MindSpeed-LLM!4695 | 4 天前 |
| 文件 | 最后提交记录 | 最后更新时间 |
|---|---|---|
| 10 天前 | ||
| 19 小时前 | ||
| 7 天前 | ||
| 4 个月前 | ||
| 7 天前 | ||
| 19 天前 | ||
| 1 个月前 | ||
| 2 天前 | ||
| 12 天前 | ||
| 7 天前 | ||
| 10 天前 | ||
| 8 天前 | ||
| 4 天前 |